Compare commits

...
Author SHA1 Message Date
Evgeny @ SimpleX Chat 81f0c2be45 test, fix 2026-06-11 15:09:17 +00:00
Evgeny @ SimpleX Chat 8b7fc88cb8 agent test 2026-06-08 11:55:59 +00:00
Evgeny @ SimpleX Chat 5f1216bf07 agent layer functions (untested) 2026-06-01 18:24:20 +00:00
Evgeny @ SimpleX Chat 7f95b4540d smp web: agent store functions and agent infrastructure 2026-05-31 19:52:44 +00:00
Evgeny @ SimpleX Chat 910a922e0d fix store 2026-05-31 14:59:47 +00:00
Evgeny @ SimpleX Chat ad50ca8644 fixes 2026-05-31 06:43:05 +00:00
Evgeny @ SimpleX Chat 3bc7a8c0a8 smp web: agent store 2026-05-30 22:31:10 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> e56c12ab3c smp web: client with tests (#1782)
* smp web: client with tests

* fixes

* support batching and SMP proxy

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-05-22 09:59:31 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 34e3d30c78 smp web: implement protocol encodings and encryption end-to-end (#1778)
* smp web: protocol encodings and x3dh

* fix

* strnup761 compiled to wasm

* AES-256-GCM, comatibility tests

* core of double ratchet

* PQ double ratchet

* test typescript ratchets

* agent encoding/encryption stack with test

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-05-16 18:18:45 +01:00
Evgeny Poberezkin a0af7377ab Merge branch 'master' into smp-web 2026-05-11 18:25:19 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 3f40febe60 smp web: spike (#1745)
* smp web: initial setup

* encoding/decoding of LGET/LNK

* SMP over websocket handshake works

* parse short connection links

* hkdf for short links

* refactor hkdf

* decrypt link data

* fetch link data via websocket

* signed challenge in web handshake

* test: skip smp-web tests when TS not built

* block encryption

* gitignore

* parse FixedLinkData

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-05-11 18:24:54 +01:00
sh fd298ae328 xftp: web page tweaks (#1772) 2026-05-06 09:24:18 +01:00
Evgeny Poberezkin 1f173abf6d 6.5.1.0 2026-05-01 17:57:19 +01:00
Evgeny 21f4597dad xftp: backwards compatible file header decoding (#1768) 2026-05-01 13:32:19 +01:00
Evgeny Poberezkin ba6af65c54 6.5.0.17 2026-04-29 20:31:42 +01:00
Evgeny Poberezkin 858fac7f4f 6.5.0.16 2026-04-21 20:26:51 +01:00
Evgeny Poberezkin 3536a156d1 Merge branch 'master' into smp-web 2026-04-20 13:43:10 +01:00
Evgeny Poberezkin 90432a44b4 tests: fix test compilation 2026-04-20 13:42:25 +01:00
sh 1e1f897c79 core: use = as INI key-value separator (#1767)
* core: use = as INI key-value separator

* core: update docker entrypoints for = INI separator

* core: update INI separator in README and test scripts
2026-04-20 09:22:14 +01:00
sh 0dc2940eff ci: add xftp-server postgres binaries (#1766) 2026-04-17 15:57:30 +01:00
sh 8833e5c1b5 xftp-server: support postgresql backend (#1755)
* xftp: add PostgreSQL backend design spec

* update doc

* adjust styling

* add implementation plan

* refactor: move usedStorage from FileStore to XFTPEnv

* refactor: add getUsedStorage, getFileCount, expiredFiles store functions

* refactor: change file store operations from STM to IO

* refactor: extract FileStoreClass typeclass, move STM impl to Store.STM

* refactor: make XFTPEnv and server polymorphic over FileStoreClass

* feat: add PostgreSQL store skeleton with schema migration

* feat: implement PostgresFileStore operations

* feat: add PostgreSQL INI config, store dispatch, startup validation

* feat: add database import/export CLI commands

* test: add PostgreSQL backend tests

* fix: map ForeignKeyViolation to AUTH in addRecipient

When a file is concurrently deleted while addRecipient runs, the FK
constraint on recipients.sender_id raises ForeignKeyViolation. Previously
this propagated as INTERNAL; now it returns AUTH (file not found).

* fix: only decrement usedStorage for uploaded files on expiration

expireServerFiles unconditionally subtracted file_size from usedStorage
for every expired file, including files that were never uploaded (no
file_path). Since reserve only increments usedStorage during upload,
expiring never-uploaded files caused usedStorage to drift negative.

* fix: handle setFilePath error in receiveServerFile

setFilePath result was discarded with void. If it failed (file deleted
concurrently, or double-upload where file_path IS NULL guard rejected
the second write), the server still reported FROk, incremented stats,
and left usedStorage permanently inflated. Now the error is checked:
on failure, reserved storage is released and AUTH is returned.

* fix: escape double quotes in COPY CSV status field

The status field (e.g. "blocked,reason=spam,notice={...}") is quoted in
CSV for COPY protocol, but embedded double quotes from BlockingInfo
notice (JSON) were not escaped. This could break CSV parsing during
import. Now double quotes are escaped as "" per CSV spec.

* fix: reject upload to blocked file in Postgres setFilePath

In Postgres mode, getFile returns a snapshot TVar for fileStatus. If a
file is blocked between getFile and setFilePath, the stale status check
passes but the upload should be rejected. Added status = 'active' to
the UPDATE WHERE clause so blocked files cannot receive uploads.

* fix: add CHECK constraint on file_size > 0

Prevents negative or zero file_size values at the database level.
Without this, corrupted data from import or direct DB access could
cause incorrect storage accounting (getUsedStorage sums file_size,
and expiredFiles casts to Word32 which wraps negative values).

* fix: check for existing data before database import

importFileStore now checks if the target database already contains
files and aborts with an error. Previously, importing into a non-empty
database would fail mid-COPY on duplicate primary keys, leaving the
database in a partially imported state.

* fix: clean up disk file when setFilePath fails in receiveServerFile

When setFilePath fails (file deleted or blocked concurrently, or
duplicate upload), the uploaded file was left orphaned on disk with
no DB record pointing to it. Now the file is removed on failure,
matching the cleanup in the receiveChunk error path.

* fix: check storeAction result in deleteOrBlockServerFile_

The store action result (deleteFile/blockFile) was discarded with void.
If the DB row was already deleted by a concurrent operation, the
function still decremented usedStorage, causing drift. Now the error
propagates via ExceptT, skipping the usedStorage adjustment.

* fix: check deleteFile result in expireServerFiles

deleteFile result was discarded with void. If a concurrent delete
already removed the file, deleteFile returned AUTH but usedStorage
was still decremented — causing double-decrement drift. Now the
usedStorage adjustment and filesExpired stat only run on success.

* refactor: merge STM store into Store.hs, parameterize server tests

- Move STMFileStore and its FileStoreClass instance from Store/STM.hs
  back into Store.hs — the separate file was unnecessary indirection
  for the always-present default implementation.

- Parameterize xftpFileTests over store backend using HSpec SpecWith
  pattern (following SMP's serverTests approach). The same 11 tests
  now run against both memory and PostgreSQL backends via a bracket
  parameter, eliminating all *Pg test duplicates.

- Extract shared run* functions (runTestFileChunkDeliveryAddRecipients,
  runTestWrongChunkSize, runTestFileChunkExpiration, runTestFileStorageQuota)
  from inlined test bodies.

* refactor: clean up per good-code review

- Remove internal helpers from Postgres.hs export list (withDB, withDB',
  handleDuplicate, assertUpdated, withLog are not imported externally)
- Replace local isNothing_ with Data.Maybe.isNothing in Env.hs
- Consolidate duplicate/unused imports in XFTPStoreTests.hs
- Add file_path IS NULL and status guards to STM setFilePath, matching
  the Postgres implementation semantics

* test: parameterize XFTP server, agent and CLI tests over store backend

- xftpTest/xftpTest2/xftpTest4/xftpTestN now take XFTPTestBracket as
  first argument, enabling the same test to run against both memory
  and PostgreSQL backends.

- xftpFileTests (server tests), xftpAgentFileTests (agent tests), and
  xftpCLIFileTests (CLI tests) are SpecWith-parameterized suites that
  receive the bracket from HSpec's before combinator.

- Test.hs runs each parameterized suite twice: once with
  xftpMemoryBracket, once with xftpPostgresBracket (CPP-guarded).

- STM-specific tests (store log restore/replay) stay in memory-only
  xftpAgentTests. SNI/CORS tests stay in memory-only xftpServerTests.

* refactor: remove dead test wrappers after parameterization

Remove old non-parameterized test wrapper functions that were
superseded by the store-backend-parameterized test suites.
All test bodies (run* and _ functions) are preserved and called
from the parameterized specs. Clean up unused imports.

* feat: add manual tests and guide

* refactor: merge file_size CHECK into initial migration

* refactor: extract rowToFileRec shared by getFile sender/recipient paths

* refactor: parameterize XFTPServerConfig over store type

Embed XFTPStoreConfig s as serverStoreCfg field, matching SMP's
ServerConfig. runXFTPServer and newXFTPServerEnv now take a single
XFTPServerConfig s. Restore verifyCmd local helper structure.

* refactor: minimize diff in tests

Restore xftpServerTests and xftpAgentTests bodies to match master
byte-for-byte (only type signatures change for XFTPTestBracket
parameterization); inline the runTestXXX helpers that were split
on this branch.

* refactor: restore getFile position to match master

* refactor: rename withSTMFile back to withFile

* refactor: close store log inside closeFileStore for STM backend

Move STM store log close responsibility into closeFileStore to
match PostgresFileStore, removing the asymmetry where only PG's
close was self-contained.

STMFileStore holds the log in a TVar populated by newXFTPServerEnv
after readWriteFileStore; stopServer no longer needs the explicit
withFileLog closeStoreLog call. Writes still go through XFTPEnv.storeLog
via withFileLog (unchanged).

* refactor: rename XFTPTestBracket to XFTPTestServer

* fix: move file_size check from PG schema to store log import

* refactor: use SQL-standard type names in XFTP schema

* perf: batch expired file deletions with deleteFiles

* refactor: stream export instead of loading recipients into memory

* refactor: parameterize XFTP store with FSType singleton dispatch

* refactor: minimize diff per review feedback

* refactor: use types over strings, deduplicate parser

* refactor: always parse database store type, fail at startup

* fix compilation without postgresql

* refactor: always parse database store type, fail at startup
2026-04-16 09:06:04 +01:00
Evgeny 95b17ada27 lib: fix incorrect encoding of Signature (incompatible with decoding, but never used together) - breaks backward compatibility for remote control connections (#1765)
* lib: fix incorrect StrEncoding of Signature (it was not compatible with decoding, but was never used)

* align encoding with used in links (breaks backward compatibility)
2026-04-15 15:11:06 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 43cdf55f3b lib: add JSON instance to Signature type (#1764)
Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-04-14 19:47:20 +01:00
Evgeny Poberezkin bc5ea42bec 6.5.0.15 2026-04-11 19:39:50 +01:00
Evgeny Poberezkin 0933cbcb9c agent: add compression api 2026-04-11 18:14:48 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> f2dafd983b agent: export decompressedSize (#1763)
Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-04-11 17:28:00 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 34c0909c1a agent: drop message after N reception attempts (#1762)
* agent: drop message after N reception attempts

* test

* increase count for message expiration

* fix migration

* update schema

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-04-11 16:24:30 +01:00
Evgeny Poberezkin 97802a30fc 6.5.0.14 2026-04-04 17:28:23 +01:00
Evgeny b82cf7d001 xftp: remove page (#1761) 2026-04-03 10:47:52 +01:00
spaced4ndy 9bc0c70fa0 agent: getConnLinkPrivKey (#1759) 2026-04-02 15:22:44 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 0741583f78 agent: read queues in batches for subscriptions (#1758)
* agent: read queues in batches for subscriptions

* resubscribe in batches too

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-04-01 16:07:17 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> f8f172f32f agent: fix race when pending subscriptions are never subscribed (#1756)
* agent: fix race when pending subscriptions are never subscribed

* small agent

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-03-31 19:16:54 +01:00
spaced4ndy 9c07ddff3c agent: allow to use existing connId for getConnShortLinkAsync (#1752) 2026-03-30 09:48:31 +00:00
Evgeny Poberezkin 50b71d3e56 6.5.0.12 2026-03-29 07:54:48 +01:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> a1b762992b agent: pass key and link ID when preparing group link (#1754)
* agent: pass key and link ID when preparing group link

* binding

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-03-28 20:29:48 +00:00
Evgeny Poberezkin 1e0093be9a docs: update whitepaper 2026-03-26 20:48:47 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 01fe841e3c smp: allow websocket connections on the same port (#1738)
* smp: allow websocket connections on the same port

* remove logs

* diff

* fix

* merge functions

* refactor

* remove unused

* refactor

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-03-20 20:01:21 +00:00
sh 1a12ee0a5a xftp-web: version bump to 0.3.0 (#1742) 2026-03-20 11:43:43 +00:00
sh efcef2d1fd xftp-web: add postgres schema cleanup for integration tests (#1741)
Stale postgres schema leaked pending XFTP operations between
cross-language tests, causing N-1 of N tests to fail.
2026-03-20 09:58:56 +00:00
Evgeny 963d7b2f75 fix small bugs (#1740)
* fix small bugs

* toChunks
2026-03-20 08:59:38 +00:00
Evgeny Poberezkin 4e4e0a4f42 6.5.0.11 2026-03-17 09:26:02 +00:00
sh 082a6c6f22 web: serve on-the-fly compressed gzip static files (#1735)
* web: serve pre-compressed gzip static files

* web: compress static files on the fly instead of pre-compressed
2026-03-16 09:08:43 +00:00
shandEvgeny dc2921e4ce xftp-server: embed file download widget in XFTP server web page (#1733)
* xftp-server: embed file download widget in XFTP server web page

When a URL has a hash fragment (>50 chars), the server page shows the
file download UI instead of the server info page. Embeds xftp-web
assets (JS, CSS, crypto worker) and protocol overlay with matching
website content. Overlay renders below the server navbar.

* xftp-server: fix overlay scroll lock, remove extra margin, fix dark SVG

* xftp-server: move file transfer widget to standalone /file page

* web: collapse all repeated Nothing sections in render

section_ only collapsed the first occurrence of a section when content
was Nothing, leaving subsequent sections with the same label intact.
This caused SMP server pages to show raw <x-xftpConfig> tags.

* xftp-server: update bundled css/js

* xftp-server: move file.html to xftp-server, rename xftp bundle dir

* web: remove unused server-info wrapper div

* refactor

* fix

---------

Co-authored-by: Evgeny <evgeny@poberezkin.com>
2026-03-13 16:00:02 +00:00
sh 5a32e729e0 xftp-web: add "Upload your file" link after download completes (#1736)
* xftp-web: add "Upload your file" link after download completes

* xftp-web: call initUpload directly instead of hashchange dispatch
2026-03-13 15:01:15 +00:00
shandEvgeny Poberezkin 782cacfb3c fix: using simplexmq as dependency (move embedFile to executables) (#1734)
* web: parameterize generateSite, remove Embedded from library

Move embedFile/embedDir out of the library so it works when
simplexmq is consumed as a dependency. generateSite now accepts
mediaContent, wellKnown, and linkHtml as parameters.

* smp-server, xftp-server: embed static files in executables

Add shared apps/common/Embedded.hs with TH splices, update SMPWeb
and XFTPWeb to pass embedded content to generateSite, move
file-embed dependency from library to executables and test suite.

* refactor

* add export, move common files to Web subfolder

* fix .cabal

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2026-03-12 17:05:00 +00:00
sh 328d3b941a xftp-web: use XFTP server domain in share link, verify on download (#1732)
* xftp-web: use XFTP server domain in share link, verify on download

Use a random XFTP server host from the file description as the link
origin instead of the current page domain. On download, verify
that the hosting domain matches one of the servers in the description.

* xftp-web: use first description server as link origin

* xftp-web: show wrong server error in download UI instead of blank page
2026-03-11 15:59:26 +00:00
sh b6b87a6323 xftp-web: disable minification in vite build (#1731) 2026-03-11 09:08:23 +00:00
sh 33454444a4 xftp-web: new version (#1728)
* xftp-web: remove flux from preset

* xftp-web: new version
2026-03-09 14:26:40 +00:00
Evgeny Poberezkin 1449e6418c 6.5.0.10 2026-03-09 12:25:08 +00:00
sh 3a627a9599 tests: use correct web CA certificate in CLI static files test (#1727) 2026-03-09 12:23:03 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> d2d834ad16 agent: validate destination relay certificate, allow 3-4 certificate chains (#1717)
Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-03-09 12:22:42 +00:00
shandEvgeny Poberezkin 437cdde4a5 xftp: add web page for server information (#1724)
* xftp: add web page for server information

* web: rename XFTP.Web to XFTPWeb, remove XFTP subdirectory

* refactor(xftp): remove storage quota from web page

* refactor

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2026-03-09 08:44:28 +00:00
shandEvgeny Poberezkin eed1bf14c6 web: extract shared web module from smp-server (#1723)
* web: extract shared web module from smp-server

Move web serving infrastructure (warp, static files, HTML templating)
from apps/smp-server/web/Static.hs into library modules:
- Simplex.Messaging.Server.Web (generic web infra + templating)
- Simplex.Messaging.Server.Web.Embedded (TH-embedded assets)

Move static assets from apps/smp-server/static/ to
src/Simplex/Messaging/Server/Web/.

Move EmbeddedWebParams/WebHttpsParams from Server.Main to Server.Web.

Keep SMP-specific rendering (serverInformation) in apps/smp-server/SMP/Web.hs.

generateSite is now generic: takes pre-rendered HTML + link page paths,
enabling reuse by XFTP and NTF servers.

* web: add tests for templating engine

Tests for render, section_, item_, and timedTTLText functions
in Simplex.Messaging.Server.Web module.

* web: add serverInfoSubsts, serveStaticPageH2, safe port parsing

* web: rename SMP.Web to SMPWeb, remove SMP subdirectory

* fix(web): section_ collapsing sections with Just "" content

Commit e48bedea ("servers: fix server pages when source code is not
specified") changed section_ to treat Just "" the same as Nothing -
collapsing the section. The intent was to handle the sourceCode case
(empty string when not specified), but the guard
`not (B.null content)` also broke operator, admin, complaints, and
hosting - all of which legitimately use Just "" as a
section-present marker.

Before (correct):
  Nothing -> before <> next
  Just content -> before <> item_ label content inside <> ...

After (broken):
  Just content | not (B.null content) -> ...
  _ -> before <> next

Restore the original behavior: only Nothing collapses a section.

* refactor

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2026-03-09 08:42:38 +00:00
Evgeny @ SimpleX Chat 313e96513c docs: correction to governance process (#1725) 2026-03-08 17:27:56 +00:00
Evgeny c6e3a4d80f add missing exports (#1722)
* add missing exports

* fix dependency
2026-03-04 07:31:46 +00:00
Evgeny Poberezkin 0e67647e90 6.5.0.9 2026-03-02 18:14:37 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> f3408d9bb6 explicit exports (#1719)
* explicit exports

* more empty exports

* add exports

* reorder

* use correct ControlProtocol type for xftp router

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-03-02 17:34:01 +00:00
sh 3ecb901e3e xftp-web: fix build and Playwright test failures (#1720) 2026-03-02 16:22:36 +00:00
EvgenyEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>shumsh
f6aca47604 xftp: implementation of XFTP client as web page (#1708)
* xftp: implementation of XFTP client as web page (rfc, low level functions)

* protocol, file descriptions, more cryptogrpahy, handshake encoding, etc.

* xftp server changes to support web slients: SNI-based certificate choice, CORS headers, OPTIONS request

* web handshake

* test for xftp web handshake

* xftp-web client functions, fix transmission encoding

* support description "redirect" in agent.ts and cross-platform compatibility tests (Haskell <> TypeScript)

* rfc: web transport

* client transport abstraction

* browser environment

* persistent client sessions

* move rfcs

* web page plan

* improve plan

* webpage implementation (not tested)

* fix test

* fix test 2

* fix test 3

* fixes and page test plan

* allow sending xftp client hello after handshake - for web clients that dont know if established connection exists

* page tests pass

* concurrent and padded hellos in the server

* update TS client to pad hellos

* fix tests

* preview:local

* local preview over https

* fixed https in the test page

* web test cert fixtures

* debug logging in web page and server

* remove debug logging in server/browser, run preview xftp server via cabal run to ensure the latest code is used

* debug logging for page sessions

* add plan

* improve error handling, handle browser reconnections/re-handshake

* fix

* debugging

* opfs fallback

* delete test screenshot

* xftp CLI to support link

* fix encoding for XFTPServerHandshake

* support redirect file descriptions in xftp CLI receive

* refactor CLI redirect

* xftp-web: fixes and multi-server upload (#1714)

* fix: await sodium.ready in crypto/keys.ts (+ digest.ts StateAddress cast)

* multi-server parallel upload, remove pickRandomServer

* fix worker message race: wait for ready signal before posting messages

* suppress vite build warnings: emptyOutDir, externals, chunkSizeWarningLimit

* fix Haskell web tests: use agent+server API, wrap server in array, suppress debug logs

* remove dead APIs: un-export connectXFTP, delete closeXFTP

* fix TypeScript errors in check:web (#1716)

- client.ts: cast globalThis.process to any for browser tsconfig,
  suppress node:http2 import, use any for Buffer/chunks, cast fetch body
- crypto.worker.ts: cast sha512_init() return to StateAddress

* fix: serialize worker message processing to prevent OPFS handle race

async onmessage allows interleaved execution at await points.
When downloadFileRaw fetches chunks from multiple servers in parallel,
concurrent handleDecryptAndStore calls both see downloadWriteHandle
as null and race on createSyncAccessHandle for the same file,
causing intermittent NoModificationAllowedError.

Chain message handlers on a promise queue so each runs to completion
before the next starts.

* xftp-web: prepare for npm publishing (#1715)

* prepare package.json for npm publishing

Remove private flag, add description/license/repository/publishConfig,
rename postinstall to pretest, add prepublishOnly, set files and main.

* stable output filenames in production build

* fix repository url format, expand files array

* embeddable component: scoped CSS, dark mode, i18n, events, share

- worker output to assets/ for single-directory deployment
- scoped all CSS under #app, removed global resets
- dark mode via .dark ancestor class
- progress ring reads colors from CSS custom properties
- i18n via window.__XFTP_I18N__ with t() helper
- configurable mount element via data-xftp-app attribute
- optional hashchange listener (data-no-hashchange)
- completion events: xftp:upload-complete, xftp:download-complete
- enhanced file-too-large error mentioning SimpleX app
- native share button via navigator.share

* deferred init and runtime server configuration

- data-defer-init attribute skips auto-initialization
- window.__XFTP_SERVERS__ overrides baked-in server list

* use relative base path for relocatable build output

* xftp-web: retry resets to default state, use innerHTML for errors

* xftp-web: only enter download mode for valid XFTP URIs in hash

* xftp-web: render UI before WASM is ready

Move sodium.ready await after UI initialization so the upload/download
interface appears instantly. WASM is only needed when user triggers
an actual upload or download. Dispatch xftp:ready event once WASM loads.

* xftp-web: CLS placeholder HTML and embedder CSS selectors

Add placeholder HTML to index.html so the page renders a styled card
before JS executes, preventing layout shift. Use a <template> element
with an inline script to swap to the download placeholder when the URL
hash indicates a file download. Auto-compute CSP SHA-256 hashes for
inline scripts in the vite build plugin.

Change all CSS selectors from #app to :is(#app, [data-xftp-app]) so
styles apply when the widget is embedded with data-xftp-app attribute.

* xftp-web: progress ring overhaul

Rewrite progress ring with smooth lerp animation, green checkmark on
completion, theme reactivity via MutationObserver, and per-phase color
variables (encrypt/upload/download/decrypt).

Show honest per-phase progress: each phase animates 0-100% independently
with a ring color change between phases. Add decrypt progress callback
from the web worker so the decryption phase tracks real chunk processing
instead of showing an indeterminate spinner.

Snap immediately on phase reset (0) and completion (1) to avoid
lingering partial progress. Clean up animation and observers via
destroy() in finally blocks.

* xftp-web: single progress ring for upload, simplify ring color

* xftp-web: single progress ring for download

* feat(xftp-web): granular progress for encrypt/decrypt phases

Add byte-level progress callbacks to encryptFile, decryptChunks,
and sha512Streaming by processing data in 256KB segments. Worker
reports fine-grained progress across all phases (encrypt+hash+write
for upload, read+hash+decrypt for download). Progress ring gains
fillTo method for smooth ease-out animation during minimum display
delays. Encrypt/decrypt phases fill their weighted regions (0-15%
and 85-99%) with real callbacks, with fillTo covering remaining
time when work finishes under the 1s minimum for files >= 100KB.

* rename package

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
Co-authored-by: shum <github.shum@liber.li>
Co-authored-by: sh <37271604+shumvgolove@users.noreply.github.com>
2026-03-02 09:57:46 +00:00
Evgeny Poberezkin 50ae1e1c3e docs: update governance process 2026-02-24 18:05:30 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 483ac674fb agent: fix possible deadlocks of queue overloading when processing messages (#1713)
* agent: fix possibly deadlocks of queue overloading when processing messages

* simplify

* refactor

* refactor

* refactor 2

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-02-24 16:47:00 +00:00
EvgenyandEvgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com> 3dba9c2b1e rfc: SimpleX Network Consortium governance (#1718)
* rfc: SimpleX Network Consortium governance

* update

* add draft label

---------

Co-authored-by: Evgeny @ SimpleX Chat <259188159+evgeny-simplex@users.noreply.github.com>
2026-02-22 18:58:09 +00:00
Evgeny Poberezkin 8fdc0703bc 6.5.0.8 2026-01-28 23:12:19 +00:00
Evgeny Poberezkin d7b90b8415 Merge branch 'stable' 2026-01-28 22:52:33 +00:00
Evgeny Poberezkin 9346b85c3f 6.4.8.0 2026-01-28 21:58:32 +00:00
Evgeny 3c5ec8d9a1 agent: improve error handling (#1707)
* agent: improve error handling

* simplify

* report critical error when subscriber crashes

* fix test
2026-01-28 21:54:41 +00:00
Evgeny d10e05b796 agent: split creating connection to two steps to prepare connection link in advance (#1704)
* agent: split creating connection to two steps to prepare connection link in advance

* linkEntityId, newOwnerAuth

* simplify
2026-01-27 10:54:13 +00:00
Evgeny 66cc06738e agent: fix tests for short links (#1703) 2026-01-26 00:11:50 +00:00
Evgeny Poberezkin 0f3b8a4a16 docs: update contributing 2026-01-24 18:05:01 +00:00
Evgeny a1596ed234 docs: contributing guide (#1702) 2026-01-24 13:33:17 +00:00
spaced4ndyandEvgeny Poberezkin 89b81d151f agent: getConnShortLinkAsync; implement joinConnectionAsync for contact connections; narrow setConnShortLinkAsync only to Contact connections (#1694)
* agent: getConnShortLinkAsync

* enable all tests

* comment

* comment

* join conn async for contact URI (wip)

* fix test

* remove enableNtfs param

* FixedLinkData

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2026-01-20 08:40:35 +00:00
Evgeny 3e5b654109 agent: support multiple link owners in link data (#1701)
* agent: support multiple link owners in link data

* fix
2026-01-19 22:08:11 +00:00
spaced4ndyandEvgeny Poberezkin ca26c69937 db: withSavepoint; agent: correctly handle errors in createWithRandomId for postgres (#1693)
* db: withSavepoint; agent: correctly handle errors in createWithRandomId

* comment

* refactor

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2026-01-15 14:59:45 +00:00
Evgeny Poberezkin 58212c421a 6.5.0.7 2026-01-13 20:55:38 +00:00
Evgenyandspaced4ndy 1000107259 agent: clean up old records in batches (#1692)
* agent: clean up old records in batches

* update

* update

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>

* stabilize tests

---------

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>
2026-01-13 19:11:46 +00:00
Evgeny 6aadcf1f3f agent: lock rows for concurrent queries in PostgreSQL (#1688)
* agent: lock rows for concurrent queries in PostgreSQL

* fix race conditions in workers

* refactor
2026-01-08 11:09:58 +00:00
spaced4ndy 07604a146f agent: async command to set connection short link (setConnShortLinkAsync) (#1680) 2026-01-07 15:12:52 +00:00
Ritiek Malhotra 4c782d3191 docker: build multi-arch images for x86-64 and arm64 (#1540) 2026-01-07 10:49:21 +00:00
Evgeny c4b687ba64 agent: reset schema after changes (#1690) 2026-01-05 20:37:50 +00:00
Evgeny a7b43b1a3e agent: use strict tables (#1686)
* agent: use strict tables

* migrate existing tables to strict

* test: verify that all tables are strict

* fix column types for device_token and ntf_mode

* fix encodings and column types for ntf_sub_action and ntf_sub_smp_action

* update schema

* remove debug.trace

* log
2026-01-03 17:19:18 +00:00
Evgeny d6df769799 ntf server: additional tests (#1687) 2026-01-02 14:07:13 +00:00
Evgeny Poberezkin 5f73d1e629 6.5.0.6 2025-12-23 18:00:26 +00:00
Evgeny Poberezkin 70d1b99fb4 Merge branch 'stable' 2025-12-23 14:18:07 +00:00
Evgeny Poberezkin 2ea98db9d8 6.4.7.0 2025-12-23 12:42:46 +00:00
Evgeny bbe1c716e6 xrcp: treat any 127.x.x.x IP address as local (#1682) 2025-12-23 12:42:05 +00:00
Evgeny Poberezkin 77ac452190 6.5.0.5 2025-12-22 21:56:25 +00:00
Evgeny ea70575275 agent: use server cert hash when selecting server queues to subscribe (#1679) 2025-12-15 11:48:17 +00:00
Evgeny 49e9ce1649 agent: skip subscribing servers without queues (#1678) 2025-12-14 23:01:21 +00:00
Evgeny Poberezkin 2ca440dd2d 6.5.0.4 2025-12-10 21:09:33 +00:00
spaced4ndyandEvgeny Poberezkin 92a9579e69 agent: option to add SQLite aggregates to DB connection (#1673)
* agent: add build_relations_vector function to sqlite

* update aggregate

* use static aggregate

* remove relations

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-12-04 08:58:14 +00:00
Evgeny cf9b7e5b6a agent: option to add SQLite functions to DB connection (#1674)
* agent: option to add SQLite functions to DB connection

* add module
2025-12-03 16:10:35 +00:00
spaced4ndyandEvgeny Poberezkin 3016b929b4 agent: make createConnection and setConnShortLink apis support setting all link data fields (#1665)
* agent: make createConnection and setConnShortLink apis support setting all link data fields

* add functions

* refactor

* refactor

* fix tests

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-11-05 07:08:33 +00:00
Evgeny Poberezkin 1ae3e8d0be 6.5.0.3 2025-10-23 08:34:11 +01:00
Evgeny 1329fc726f smp: support client notices (#1659)
* agent: support client notices

* improve

* fix, test

* rename

* cleanup

* send and process notices in more cases

* dont delete

* dont remove notice on other permanent errors

* dont remove notice if there is no notice ID in queue

* add server to error

* allow deleting

* only use notice if key hash matches
2025-10-17 18:34:59 +01:00
spaced4ndyandEvgeny 234aeb81dd agent: add subscription status to connection stats (#1658)
* agent: add subscription status to connection stats

* wip

* conn status

* format

* refactor

* refactor

* m

* shorter

* shorter

---------

Co-authored-by: Evgeny <evgeny@poberezkin.com>
2025-10-13 17:53:28 +01:00
spaced4ndyandEvgeny Poberezkin 1dbc15b2e6 agent: sync connections (#1654)
* agent: sync subscriptions

* remove comment

* add shouldDelete flag

* compare api

* remove instance

* query

* rename

* refactor

* functor

* JSON instances

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-10-10 05:41:42 +00:00
Evgeny 80aa56cbcc agent: track which queues need subscribing for iOS NSE (#1657)
* agent: track which queues need subscribing for iOS NSE

* fix down migration

* fix, cleanup
2025-10-09 18:33:27 +01:00
Evgeny 9cda20381f agent: subscribe all connections (#1655)
* agent: subscribe all connections

* query, version

* BoolInt

* add query to errors

* Revert "add query to errors"

This reverts commit 32a1f7fe11.

* fix optional field

* version

* limit number of in-flight subscriptions to 35000
2025-10-09 13:43:48 +01:00
Evgeny 318ddf692a smp server: import/export to/from PostgreSQL with one command, deprecation notice for journal storage (#1653) 2025-10-09 13:31:57 +01:00
Evgeny 11a4859f8e agent: batch processing of subscription results and errors (#1652)
* agent: batch processing of subscription results and errors

* run agent tests with in-memory server storage

* version

* non empty errors

* size
2025-10-07 14:49:38 +01:00
Evgeny a3d1a72eb0 agent: optimize subscriptions memory usage (#1651)
* agent: optimize subscriptions memory usage more (do not store subscribed queues in memory) WIP

* use new session subscriptions data

* version

* remove old data structure

* remove version

* batch deletions

* test TSessionSubs

* comment
2025-10-06 15:21:17 +01:00
Evgeny Poberezkin 17b71cf673 Merge branch 'stable' 2025-10-05 22:03:58 +01:00
Evgeny 779222d1a7 agent: optimize subscriptions memory usage (#1649)
* agent: optimize subscriptions memory usage

* version

* remove typeclass

* disable subscription to test memory usage

* disable delivery and commands

* disable reading connections

* enable subscriptions

* RcvQueueCred

* RcvQueueCred postgres

* fix

* optimize keys

* version

* rename

* comments
2025-10-05 13:35:54 +01:00
Evgeny Poberezkin 538dcb6a4c 6.4.6.0 2025-10-04 14:49:42 +01:00
Evgeny Poberezkin 98115b6db1 6.5.0.2 2025-09-30 12:37:28 +01:00
Evgeny 9854caca33 agent: optimize subscriptions (#1645)
* agent: optimize subscriptions

* simplify

* clean up
2025-09-30 12:34:42 +01:00
spaced4ndy c8b551dcf7 agent: fix file reception error in postgres (#1647) 2025-09-30 08:32:45 +01:00
Evgeny 99f40ae109 support for additional database migrations (#1644) 2025-09-27 11:21:38 +01:00
sh 3a9381b966 ci/scripts: execute git version if chmod fails (#1643) 2025-09-23 09:08:56 +01:00
Evgeny Poberezkin 54a2a6c905 6.5.0.1 2025-09-17 09:04:03 +01:00
Evgeny 7b7616ce7e agent: batch loading of connections with PostgreSQL client (#1639)
* agent: batch loading of connections with PostgreSQL client

* batch more

* optimize getPendingCommandServers

* fix Bool conversion

* enable all tests

* cleanup
2025-09-16 14:28:06 +01:00
sh b020a08ea0 ci: disable aarch64 tests (#1638) 2025-09-15 09:02:34 +01:00
Evgeny Poberezkin 2737e7d602 6.5.0.0.1 2025-09-15 07:52:17 +01:00
Evgeny d3754b36bc smp server: do not use queue cache with PostgreSQL message storage (#1637)
* smp server: do not use queue cache with PostgreSQL message storage

* fix loading queues via notifier IDs
2025-09-15 07:51:27 +01:00
Evgeny Poberezkin 112cd9d5f4 Merge branch 'stable' 2025-09-14 14:42:05 +01:00
Evgeny 8a7991a376 agent: option to set SQLite database backup path (#1636)
* agent: option to set SQLite database backup path

* fix test compilation
2025-09-14 12:41:49 +01:00
Evgeny 6f4d0f5377 Merge pull request #1630 from simplex-chat/db-messages
Feature branch: PostgreSQL message store
2025-09-13 19:24:59 +01:00
Evgeny 9cfdae3494 smp server: uniterruptible mask for DB operations (#1635) 2025-09-13 19:08:07 +01:00
Evgeny Poberezkin 62183699db 6.5.0.0 2025-09-13 11:37:59 +01:00
Evgeny 9be9a76b42 smp server: improve message expiration (#1634)
* smp server: limit by time the queues to export journal messages for

* pass queue/msg thresholds separately

* reset db connection on errors

* Revert "smp server: limit by time the queues to export journal messages for"

This reverts commit d3bc0cba4b.

* fix test compilation

* flag to expire messages

* improve test

* expire messages newer than quota
2025-09-13 11:35:34 +01:00
Evgeny Poberezkin 80f7be6dd7 Merge branch 'master' into db-messages 2025-09-13 09:31:23 +01:00
Evgeny 83721240a4 reset db connection on errors (#1633) 2025-09-13 09:30:01 +01:00
Evgeny 6c66cf367a smp server: set message counts correctly after import (#1632) 2025-09-12 13:08:11 +01:00
Evgeny a137d01c90 smp server: improve counts management and expiration for postgres message store (#1631)
* smp server: improve counts management and expiration for postgres message store

* one update

* fix expiration
2025-09-12 11:39:35 +01:00
Evgeny bac6ea6e91 smp server: store messages in PostgreSQL (#1622)
* smp server: store messages in PostgreSQL

* stored procedures to write and to expire messages

* function to export messages

* move all message functions to PostgreSQL, remove delete trigger

* comments

* import messages to db

* fix message import, add export

* fix export

* fix export

* fix compilation flags

* import messages line by line

* fix server start with database storage

* fix compilation

* comments
2025-09-11 20:22:55 +01:00
Evgeny 0c1030cf02 smp server: faster export (#1626)
* smp server: faster export

* flush

* export messages with PostgreSQL database

* remove flush

* batch message writes

* Revert "batch message writes"

This reverts commit 61fb3c3b9e.

* remove $!

* fast journal export

* another approach

* Revert "smp server: remove dependency of message size on the version (#1627)"

This reverts commit 8fea15245a.

* style

* faster?

* cleanup

* cleanup

* refactor

* refactor

* concurrent read messages

* Revert "concurrent read messages"

This reverts commit 05a32e6491.

* concurrent read/write

* parameter to export to/import from another message log file

* Revert "parameter to export to/import from another message log file"

This reverts commit 4e88b035c4.

* Revert "concurrent read/write"

This reverts commit a8eab1f7e3.
2025-09-10 18:22:58 +01:00
Evgeny Poberezkin 23aff6b0b1 Revert "smp server: remove dependency of message size on the version (#1627)"
This reverts commit 8fea15245a.
2025-09-09 09:56:48 +01:00
sh 3335cd5500 ci: add aarch64 builds (#1624) 2025-09-09 08:55:37 +01:00
spaced4ndyandEvgeny Poberezkin a4f049d8da agent: parameterize withWork, getWorkItem with StoreError; parameterized Binary for SQLite (#1617)
* agent: parameterize withWork StoreError

* getWorkItem

* export

* binary

* remove handleWrkErr AnyStoreError constraint

* put AnyError in AnyStoreError constraint

* move typeclass

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-09-08 16:38:08 +01:00
Evgeny 8fea15245a smp server: remove dependency of message size on the version (#1627) 2025-09-08 11:28:42 +01:00
sh 42a2c1fc57 ci: fix git in docker (#1623) 2025-09-08 09:41:20 +01:00
Evgeny Poberezkin 7e98b3103f 6.4.5.2 2025-09-07 15:13:19 +01:00
Evgeny Poberezkin 2a61085f07 backward compatible encoding for missing NetworkError 2025-09-07 15:04:06 +01:00
Evgeny 4386dd8b5a smp server: option to expire messages when new message is sent (now off by default) (#1625)
* smp server: option to expire messages when new message is sent (now off by default)

* fix test
2025-09-07 09:02:32 +01:00
Evgeny Poberezkin 50ddd630be 6.4.5.1 2025-09-03 11:12:20 +01:00
Evgeny cb3250e7b4 servers: better socket leak prevention during TLS handshake, add NetworkError type to better diagnose connection errors (#1619)
* servers: better socket leak prevention during TLS handshake

* log tcp connection errors

* more detailed network error

* log full address

* rename error

* add encodings for NetworkError

* refactor

* comment

* bind

* style

* remove parameters of NETWORK error from encoding
2025-09-02 16:07:37 +01:00
Evgeny Poberezkin 0319addd2b 6.4.5.0 2025-09-01 18:57:09 +01:00
Evgeny 77bf76e1f9 smp server: fix in-memory server not restoring queue/service associations after 2+ restarts (#1618)
* smp server: fix in-memory server not restoring queue/service associations after 2+ restarts

* fix test, do not reuse database
2025-09-01 17:45:31 +01:00
Evgeny beafac1f73 agent: make agent workers usable from other contexts (#1614) 2025-08-29 08:33:55 +01:00
Evgeny Poberezkin a2d777bda0 6.4.4.1 2025-08-23 20:05:10 +01:00
Evgeny e48bedeaf2 servers: fix server pages when source code is not specified, include commit SHA in printed version and in web page. (#1608)
* smp server: fix server pages when source code is not specified

* servers: include git commit in version

* flexible alpn

* fix test

* fix ghc 8.10.7 build
2025-08-23 19:59:00 +01:00
Evgeny Poberezkin a2d35281b2 6.4.4.0 2025-08-18 16:22:24 +01:00
spaced4ndy 46035af9a3 agent: add context to SEMsgNotFound error (#1604) 2025-08-18 10:59:57 +01:00
Evgeny Poberezkin 4b7fc34fe3 enable all tests 2025-08-17 10:40:12 +01:00
Evgeny 96e8b4a146 smp server: wrap all queries in transactions (#1603)
* smp server: wrap all queries in transactions

* fix test

* fix schema test
2025-08-17 10:38:20 +01:00
Evgeny 2cedb66667 agent: chat relay address type for short links (#1602) 2025-08-15 07:24:37 +01:00
Evgeny e345671c76 agent: extend xrcp certificate validity 1 hour in the past, to allow out of sync clocks (#1601) 2025-08-14 22:48:11 +01:00
Evgeny 86fb2cddc5 smp server: support short SimpleX addresses in server information / pages (#1600) 2025-08-12 07:31:32 +01:00
Evgeny Poberezkin 931c533a3d 6.4.3.0 2025-08-05 12:30:28 +01:00
Evgeny 79ba60e3ad agent: update contact request server hosts to match server in short link (#1597) 2025-08-05 12:29:50 +01:00
Evgeny fb477b24d7 smp server: support short link URI as queue identifier in control port commands (#1596) 2025-08-05 08:18:39 +01:00
Evgeny Poberezkin 9f263e8f3e docs: changelog 2025-08-02 09:21:44 +01:00
Evgeny Poberezkin db325cb81f 6.4.2.0 2025-07-31 09:21:11 +01:00
Evgeny b167d01f8a rfc: using the same profile from multiple devices (#1588) 2025-07-30 20:54:03 +01:00
Evgenyandsh f4e7469f96 xftp server: prometheus metrics (#1595)
* xftp server: prometheus metrics

* update

Co-authored-by: sh <37271604+shumvgolove@users.noreply.github.com>

---------

Co-authored-by: sh <37271604+shumvgolove@users.noreply.github.com>
2025-07-30 16:18:49 +01:00
Evgeny 4647d69d4b smp server: do not include previously blocked queues in stats, prevent leak of client threads (#1593) 2025-07-30 11:12:57 +01:00
sh 9ab071d62c ci: fix reproduce-schedule (#1594) 2025-07-30 09:17:28 +01:00
Evgeny Poberezkin f4c09ac51f Merge branch 'stable' 2025-07-29 09:09:36 +01:00
Evgeny Poberezkin 846be50f72 6.4.1.0 2025-07-21 15:22:47 +01:00
Evgeny c0f357d817 servers: control port session improvements (#1591)
* servers: prohibit changing role during control port session

* quota for blocked queues

* allow disabling blocking and quota

* fix test

* fix INI file
2025-07-21 15:19:43 +01:00
Evgeny 40fc09a93d smp server: additional stat counter for ntf credentials created together with the queue (#1589)
* smp server: additional stat counter for ntf credentials created together with the queue

* fix prometheus

* fix test

* fix qSub
2025-07-18 09:26:25 +01:00
Evgeny 2a90a2c552 smp protocol: create notification credentials via NEW command that creates the queue (#1586)
* smp protocol: create notification credentials via NEW command that creates the queue

* create ntf subscription for queues created with ntf credetials

* do not create ntf credentials when switching connection to another queue
2025-07-16 16:46:45 +01:00
sh fc581bf729 github/workflows: fix reproduce schedule (#1587) 2025-07-16 11:42:14 +01:00
Evgeny Poberezkin b6ea025333 docs: update build badge 2025-07-15 10:03:27 +01:00
Evgeny Poberezkin d3e2d9ac5b docs: changelog v6.4 2025-07-13 15:20:26 +01:00
Evgeny Poberezkin 85a7fbf538 6.4.0.12 2025-07-12 14:19:32 +01:00
Evgeny 62733ef4c1 smp server: refactor subscriptions and delivery in order to always response SOK on subscription with an optional message to follow. (#1573)
* smp server: refactor subscriptions and delivery

* metric for time between MSG and ACK

* cleanup

* refactor pattern match for ghc 8.10.7

* time buckets

* split max time metric

* histogram

* fix
2025-07-12 14:18:38 +01:00
Evgeny 99e59b73a3 agent: fix message delivery in case one of the connections has no snd queue for any reason - it could break delivery to all connections (#1585)
* agent: fix message delivery in case one of the connections has no snd queue for any reason - it could break delivery to all connections

* simplify

* comment

* refactor, also postpone failing on ratchet sync send prohibited errors

* postpone failing on connection errors to allow subsequent connections succeed
2025-07-10 10:21:12 +01:00
Evgeny 384a3352cf ntf server: more efficient status update query (#1584) 2025-07-08 23:01:57 +01:00
Evgeny Poberezkin 7a04fdff2e 6.4.0.11.3 2025-07-07 18:50:31 +01:00
Evgeny ba3c75e58c smp server: correctly track if ntf service is subscribed and total subscribed queues count (fixes race condition between NSUB and NSUBS from notification server) (#1583) 2025-07-07 18:49:47 +01:00
Evgeny Poberezkin 1062ccc5c3 6.4.0.11.2 2025-07-07 10:07:54 +01:00
Evgeny 36f05e272e agent: support different timeouts for interactive and background requests (#1582)
* agent: support different timeouts for interactive and background requests

* fix tests

* use one constructor for the first request and for retries
2025-07-07 09:38:52 +01:00
Evgeny Poberezkin 660c4293f0 6.4.0.11.1 2025-07-05 11:35:27 +01:00
spaced4ndyandEvgeny Poberezkin 1b8613d767 build: do not exclude postgres modules for client_library flag (#1579)
* don't exclude postgres modules for client_library flag

* fix build with client and server postgres flag

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-07-04 17:31:57 +01:00
Evgeny bfa52c4ba5 docs: update RFC for short links to match the implementation (#1581) 2025-07-03 18:09:02 +01:00
spaced4ndy c5eb66038b agent: allow to accept contact requests after address is deleted (#1580) 2025-07-03 09:29:39 +00:00
Evgeny Poberezkin a46edd60f0 6.4.0.11 2025-06-30 21:50:21 +01:00
Evgeny b4bcfd325b agent: better handling errors during connection handshake retries (#1578)
* agent: handle invitation connection handshake errors

* fix/test retries for connecting via address
2025-06-25 19:06:00 +01:00
Evgeny 976bd3a389 agent: short link encodings (#1577) 2025-06-24 10:06:44 +01:00
Evgeny Poberezkin bf27c846da 6.4.0.10.1 2025-06-20 13:49:17 +01:00
Evgeny 455360205c smp server: prometheus histogram for message confirmation times (ACK) (#1575)
* time buckets

* split max time metric

* histogram

* histogram for confirmed delivery times

* gaugehistogram

* fix created, _ in gauge_histogram

* remove comments

* fix metrics
2025-06-20 13:46:59 +01:00
Evgeny 79c67f2026 agent: JSON encoding for connection mode (#1576) 2025-06-19 11:21:29 +01:00
Evgeny c8928626fc ntf server: do not log subscription status updates on subscribing/unsubscribing (#1574) 2025-06-19 09:15:07 +01:00
Evgeny b47d28a22a smp server: prometheus metrics for delievered and not acknowledged messages (to monitor stuck deliveries) (#1572) 2025-06-17 21:10:09 +01:00
Evgeny c5b7d3c7af agent: use flexible (small/large) string encoding for contact link data, for backward/forward compatibility (#1569)
* agent: use large string encoding for contact link data

* size-dependent user link data encoding

* tag large user data encoding with \255
2025-06-16 11:11:59 +01:00
Evgeny Poberezkin d950012530 6.4.0.10 2025-06-14 21:06:54 +01:00
Evgeny Poberezkin 27b1f48929 Merge branch 'stable' 2025-06-14 21:05:59 +01:00
Evgeny Poberezkin 3d62a383d5 6.4.0.3.1 2025-06-14 14:19:31 +01:00
Evgeny 6ac7101f4f agent: make decoding for short link data forward compatible (#1568) 2025-06-14 14:13:14 +01:00
Evgeny 65cc19842c servers: restore constant time for absent queue/wrong key (regression in 6.4-beta.9) (#1567) 2025-06-14 07:10:51 +01:00
Evgeny 656f290660 agent: fix updating short invitation link data (#1566) 2025-06-13 21:27:46 +01:00
Evgeny Poberezkin 643c3c3b3e 6.4.0.9 2025-06-13 09:48:20 +01:00
Evgeny da37384335 smp server: batch commands (#1559)
* protocol: refactor types and encoding

* clean

* smp server: batch commands (#1560)

* smp server: batch commands verification into one DB transaction

* ghc 8.10.7

* flatten transmission tuples

* diff

* only use batch logic if there is more than one transmission

* func

* reset NTF service when adding notifier

* version

* Revert "smp server: use separate database pool for reading queues and creating service records (#1561)"

This reverts commit 3df2425162.

* version

* Revert "version"

This reverts commit d80a6b74c5.
2025-06-12 23:05:04 +01:00
Evgeny 1658048c2c Revert "smp server: use separate database pool for reading queues and creating service records (#1561)" (#1564)
This reverts commit 3df2425162.
2025-06-12 11:04:43 +01:00
Evgeny 27d38518e1 agent: use PQ keys in contact request data inside link container (but not in contact request link); use PQ keys in invitations sent to contact addresses (#1563)
* agent: use PQ keys in contact request data inside link container (but not in contact request link); use PQ keys in invitations sent to contact addresses

* do not use PQ keys in the link with old address versions
2025-06-11 18:22:37 +01:00
Evgeny cf8088ac6a smp server: do not require ALPN for HTTP to fix Android asset links, do not use port 443 in server-server SMP connections (#1562)
* log alpn

* always use HTTP when SNI is sent, regardless of ALPN

* decide credential based on SNI

* do not use web port in SMP/Ntf servers connecting to SMP servers

* simpler

* refactor

* fix
2025-06-10 22:09:07 +01:00
Evgeny Poberezkin 1e82104224 6.4.0.8 2025-06-09 14:20:25 +01:00
Evgeny 46ff37c362 ntf server: additional statistics (#1558)
* ntf server: additional statistics

* version

* fix stats

* add stats to track notifications without active token

* refactor

* fix stats parser

* version
2025-06-09 14:14:27 +01:00
Evgeny 3df2425162 smp server: use separate database pool for reading queues and creating service records (#1561) 2025-06-09 12:47:35 +01:00
Evgeny 5241f5fe5e rfc: client certificates for servers using SMP protocol as clients (opertors' chat relays, notification servers, service bots) (#1534)
* rfc: client certificates for high volume clients (opertors' chat relays, notification servers, service bots)

* client certificates types (WIP)

* parameterize Transport

* protocol/schema/api changes

* agent API

* rename command

* agent subscriptions return local ClientServiceId to chat

* verify transmissions

* fix receiving client certificates, refactor

* ntf server: remove shared queue for all notification subscriptions (#1543)

* ntf server: remove shared queue for all notification subscriptions

* wait for subscriber with timeout

* safer

* refactor

* log

* remove unused

* WIP service subscriptions and associations, refactor

* process service subscriptions

* rename

* simplify switching subscriptions

* SMP service handshake with additional server handshake response

* notification delivery and STM persistence for services

* smp server: database storage, store log, fix encoding for STORE error, replace String with Text in locks and error

* stats

* more stats

* rename SMP commands

* service subscriptions in ntf server agent (tests fail)

* fix

* refactor

* exports

* subscribe ntf server as service for associated queues

* test ntf service connection, fix SOKS response, fix service associations not removed in STM storage

* INI option to support services

* ntf server: downgrade subscriptions when service is no longer supported, track counts of subscribed queues

* smp protocol: include service certificate fingerprint in the string signed over with entity key (TODO two tests fail)

* fix test

* ntf server prometheus stats, use Int64 in SOKS/ENDS responses (to avoid conversions), additional error status for ntf subscription

* update RFC

* refactor useServiceAuth to avoid ad hoc decisions about which commands use service signatures, and to prohibit service signatures on other commands

* remove duplicate service signature syntax check from checkCredentials, it is checked in verifyTransmission

* service errors, todos

* fix checkCredentials in ntf server, service errors

* refactor service auth

* refactor

* service agent: store returned queue count instead of expected

* refactor serverThread

* refactor serviceSig

* rename

* refactor, rename, test repeat NSUB service association

* respond with error to SUBS

* smp server: export/import service records between database and store log

* comment

* comments

* ghc 8.10.7
2025-06-06 08:03:47 +01:00
Evgeny 8e86c97a13 servers: include supported ALPNs in server transport config (#1557) 2025-06-03 13:35:35 +01:00
Evgeny 90e8c3adf6 smp agent: use static RNG function to avoid dynamic C stub created by Haskell FFI wrapper (#1556) 2025-06-02 20:11:27 +01:00
Evgeny 56851365b1 servers: parameterize SMP agent to optimize subscriptions (#1555)
* servers: parameterize SMP agent to optimize subscriptions

* refactor ClientCommand to avoid additional map in the notification server agent

* rename
2025-06-01 16:14:52 +01:00
Evgeny a9814bb6d3 smp server: parameterize server by store type to avoid boxing/unboxing client handles (#1554)
* smp server: parameterize server by store type to avoid boxing/unboxing client handles

* ghc 8.10.7
2025-05-29 22:17:28 +01:00
Evgeny Poberezkin 3ad8bd15a6 6.4.0.7 2025-05-29 13:14:09 +01:00
spaced4ndy 4c33d8ac43 agent: To and FromField instances for AConnShortLink (#1553) 2025-05-29 10:59:36 +01:00
spaced4ndyandEvgeny Poberezkin a94ca62624 agent: allow changing invitation link data before it is secured (#1552)
* agent: setInvitationShortLink api

* Eq instance

* allow changing link data on server, refactor

* fix

* encodings

* remove link data after connection

* Revert "encodings"

This reverts commit f8e254cca9.

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-05-29 10:44:16 +01:00
Evgeny 53b72469b6 refactor types (#1551) 2025-05-27 22:31:02 +01:00
Evgeny Poberezkin f80ed32a06 6.4.0.6 2025-05-25 17:28:39 +01:00
Evgeny 07eaf9157b smp server: allow getting and deleting short links for the old contact queues (#1549)
* smp server: allow getting and deleting short links for the old contact queues

* fix verifaction of legacy contact queues

* test
2025-05-25 17:03:02 +01:00
Evgeny 56ea2fdd56 refactor types for DB entity (#1548) 2025-05-24 18:19:11 +01:00
Evgeny ffecd4a17a parameterize transport by peer type (client/server) (#1545)
* parameterize transport by peer type (client/server)

* LogDebug level when test is retried

* support "flipped" HTTP2, fix test retry to avoid retrying pending tests

* move sync to the end of the tests
2025-05-24 14:34:22 +01:00
Evgeny Poberezkin dae649fb87 6.4.0.5 2025-05-24 14:21:53 +01:00
Evgeny 57a77f75c1 smp server: support adding link data to contact addresses created before July 2024 (#1547) 2025-05-24 14:20:25 +01:00
spaced4ndy 18e73b8aa7 agent: pass CRClientData to setContactShortLink (#1546)
* agent: pass CRClientData to setContactShortLink

* fix

* fix
2025-05-23 18:21:36 +01:00
Evgeny af9ca59e51 smp server: optimize concurrency and memory usage, refactor (#1544)
* smp server: optimize concurrency and memory usage, refactor

* hide clients IntMap

* reduce STM contention

* comment

* version

* correct stats for subscriptions

* version

* comment

* remove subscribed clients from map

* version

* optimze, refactor

* version

* debug test

* enable all tests

* remove test logs

* retry failed tests with debug logging

* increase test timeout

* sync between tests
2025-05-23 12:52:18 +01:00
Evgeny Poberezkin d352d518c2 6.4.0.4 2025-05-21 08:56:01 +01:00
Evgeny f0dc600016 ntf server: remove shared queue for all notification subscriptions (#1543)
* ntf server: remove shared queue for all notification subscriptions

* wait for subscriber with timeout

* safer

* refactor

* log
2025-05-21 08:55:32 +01:00
Evgeny Poberezkin f44ea0a6d8 6.4.0.3 2025-05-15 21:01:49 +01:00
Evgeny f7d31d4c02 Merge pull request #1524 from simplex-chat/ntf-storage
ntf server: PostgreSQL database storage (feature branch)
2025-05-15 21:00:58 +01:00
Evgeny b90e25a3a5 ntf server: fix repeat token registration when it became invalid (regression) (#1539) 2025-05-15 19:56:48 +01:00
Evgeny Poberezkin cf4b9f669d agent: always enable periodic notifications 2025-05-15 15:16:44 +01:00
Evgeny Poberezkin e417d35cce Merge branch 'master' into ntf-storage 2025-05-14 22:20:55 +01:00
Evgeny Poberezkin deaec3cce2 6.4.0.2 2025-05-07 07:24:06 +01:00
Evgeny 7bbd99644a agent: fix query that updates last message timestamps for notification processing (#1536) 2025-05-06 23:19:58 +01:00
Evgeny cb59a449dd agent: return error and message absence differently when getting notification messages (#1535)
* agent: return error and message absence differently when getting notification messages

* fix test

* mapM

* inline nse functions, release lock on error or no message
2025-05-06 16:20:01 +01:00
Evgeny a632eea75b protocol: parsing for UTCTime to stop on "," and ";" (#1532) 2025-05-04 21:11:06 +01:00
Evgeny 3d10c9bf9e agent: handle cases when last message ts is not set for notifications; set last ts for "stale" notifications when messages expired and queue is empty, to prevent repeated processing (#1531)
* agent: handle cases when last message ts is not set for notifications; set last ts for "stale" notifications when messages expired and queue is empty, to prevent repeated processing

* only log errors if they exist

* only set last ts for queue that delivered notification
2025-05-02 11:21:39 +01:00
Evgeny 2f0cdc40af ntf server: fix returned last notifications (#1530) 2025-04-30 14:57:04 +01:00
Evgeny 0a3d014f5d ntf server: priority db pool for notifications and client commands, concurrent resubscriptions, CLI options (#1529)
* ntf server: option to skip specific tokens when importing store log

* logs

* batch later

* subscribe in parallel, bigger queues

* ntf server: test schema and migrations

* optimize subscriptions

* log level

* resubscribe before starting

* better subscribing

* pooled concurrent resubscriptions

* 5

* async resubscribe

* typo

* priority pool

* db batch size
2025-04-30 11:40:32 +01:00
Evgeny 7d0115daec ntf server, agent: send all periodic notifications from one thread, only to old active clients or new clients with periodic notification mode (#1528)
* ntf server, agent: send all periodic notifications from one thread, only to old active clients or new clients with periodic notification mode

* send different type via subscription queues

* option to compact store log on start
2025-04-28 12:18:55 +01:00
Evgeny f024ab1c3f ntf server: prometheus metrics (#1527)
* ntf server: save prometheus stats

* info metrics

* fix test
2025-04-27 13:25:44 +01:00
Evgeny f4bc1f0926 ntf server: skip duplicates when importing tokens and subscriptions (#1526)
* ntf server: skip duplicates when importing tokens and subscriptions

* skip imported last notifications when no token or subscription present

* fix skipped imported notifications count

* all tests

* fix test
2025-04-26 23:37:55 +01:00
Evgeny 42dbb887f7 ntf server: allow retries when creating subscriptions, prohibit subscriptions with the same queue but another notifier key or token (#1525)
* ntf server: allow retries when creating subscriptions, prohibit subscriptions with the same queue but another notifier key or token

* sync files in the test

* refactor
2025-04-25 19:41:18 +01:00
Evgeny 850d2fa423 ntf server: PostgreSQL database storage (#1519)
* ntf server: PostgreSQL database storage

* ntf server: import/export stubs

* ntf server postgres db functions

* some notifications tests pass

* notifications tests pass

* import/export notification store logs

* fix ntf server CLI

* log in parralel

* update subscription statuses using executeMany

* fix import/export

* refactor

* fix queries

* prohibit token_id and subscription_id updates, dont update tokens and subscriptions on conflict, improve server insertion, remove duplicate tokens for import, remove subscriptions without tokens for import

* comment

* remame

* increase test delay
2025-04-25 16:47:39 +01:00
Evgeny 08b84deba4 agent: option to use web port by default for preset servers only (#1523)
* agent: option to use web port by default for preset servers only

* shorten/restore short links in agent, add encodings for SMP web port setting

* decouple preset domains from preset servers for short links

* refactor, rename
2025-04-24 17:11:52 +01:00
Evgeny Poberezkin ec5a60430d 6.4.0.1 2025-04-22 22:01:16 +01:00
Evgeny e4d4b51c59 smp server: fix/test database import (#1521) 2025-04-22 21:59:27 +01:00
Evgeny Poberezkin 6b60f8bab6 Revert "ntf server: optimize in-memory storage (#1516)"
This reverts commit afb338a41a.
2025-04-21 21:01:24 +01:00
shum aa9b93eee5 ci: retry only in stable or release 2025-04-21 19:06:04 +01:00
Evgeny afb338a41a ntf server: optimize in-memory storage (#1516)
* ntf server: optimize in-memory storage

* test

* ntf server: fix store log parser for token status
2025-04-21 17:12:16 +01:00
Evgenyandshum 1e29f7c811 ci: adjust output, retry tests on failure, include hashes in release notes (#1520)
* scripts/reproduce-builds: rename with prefix and adjust output

* ci: rerun tests and include hashes in release

* ci: fix variables and really skip 8.10.7

* ci: fix prepare step

* ci: setup swap for 8.10.7

* ci: fix upload and formatting

* ci: properly format hashes

* ci: simplify

---------

Co-authored-by: shum <shum@liber.li>
2025-04-21 17:05:31 +01:00
Evgeny Poberezkin d11be15295 ntf server: fix store log parser for token status 2025-04-16 13:01:14 +01:00
Evgeny Poberezkin 305f79d2a6 6.4.0.0 2025-04-14 10:52:05 +01:00
Evgeny Poberezkin af988f774e Merge branch 'stable' 2025-04-14 10:51:34 +01:00
Evgeny Poberezkin 82e389298a docs: CHANGELOG 2025-04-14 10:51:23 +01:00
Evgeny c784d5ce0c Merge pull request #1495 from simplex-chat/short-links
smp protocol: short 1-time invitations and contact address links
2025-04-14 09:58:09 +01:00
Evgeny c2c4730953 smp server: serve .well-known folder via server pages (#1514) 2025-04-13 22:02:48 +01:00
Evgeny Poberezkin ccdd8e1775 agent: additional encodings for short links 2025-04-13 13:43:27 +01:00
Evgeny Poberezkin 307a784174 Merge branch 'master' into short-links 2025-04-13 10:51:32 +01:00
Evgeny Poberezkin 9abc0fa88d 6.3.2.0 2025-04-11 23:22:48 +01:00
Evgeny 76aad61f00 smp server: update server page for short links to correctly change URI for apps (#1511)
* smp server: update server page for short links to correctly change URI for apps

* change page titles
2025-04-11 23:18:23 +01:00
Evgeny 37ce109009 smp server: .well-known folder for server pages for SimpleX apps to handle preset server links (#1510)
* smp server: .well-known folder for server pages for SimpleX apps to handle preset server links

* add short link group path

* test
2025-04-11 18:30:43 +01:00
Evgeny a66163dc46 agent: additional short link types and encodings (#1509) 2025-04-11 12:29:44 +01:00
Evgeny Poberezkin bdf8bf093c smp server: make symlinks relative 2025-04-10 20:58:33 +01:00
Evgenyandsh b7a9542213 smp server: short links and owners for channels (#1506)
* smp server: short links and owners for channels

* types

* support mutliple rcv keys

* fix down migration, test/create server schema dump

* reduce schema dump

* parameterize type for link data by connection type

* return full connection link data

* test version

* change short link encoding

* test: print pg_dump output

* server pages, link encoding

* fix connection request when queue data and sender ID are created for old servers

* test, change pattern

* ci: install postgresql tools in runner (#1507)

* ci: install postgresql tools in runner

* ci: docker shell abort on error

* fix pattern for ghc 8.10.7

* patch ConnReqUriData SMP encoding to preserve queue mode after decoding

* test for RKEY

* fix/test store log with RKEY

---------

Co-authored-by: sh <37271604+shumvgolove@users.noreply.github.com>
2025-04-10 19:09:47 +01:00
Evgeny Poberezkin 3a3f9fd51e Merge branch 'master' into short-links 2025-04-10 11:51:37 +01:00
Evgeny 6dc9d76ed3 smp agent: handle client/agent version downgrades after connection was established (#1508) 2025-04-10 11:50:19 +01:00
Evgeny a2a4b80af4 agent: padded encryption for link data, shorten/restore short links, tests (#1505)
* agent: padded encryption for link data, tests

* lambda

* test short links via proxy

* tests: server persistence with short links

* rfc: group links

* shorten, restore, test short links encoding

* rfc
2025-04-05 07:45:39 +01:00
Evgeny 7ec0ae3bb5 agent: types and encodings to use from the app (#1504)
* agent: types and encodings to use from the app

* use action forks
2025-04-03 16:31:22 +01:00
Evgeny 2c5530c9f0 agent: binary protocol encoding for connection request (#1503)
* agent: binary protocol encoding for connection request

* enable tests

* test
2025-04-01 22:32:49 +01:00
Evgeny 94ee3ceced agent: replace sndSecure with queueMode in queue URIs and records (#1502)
* agent: replace sndSecure with queueMode in queue URIs and records

* fix test
2025-03-31 16:54:29 +01:00
Evgeny Poberezkin dd67de4d71 Merge branch 'master' into short-links 2025-03-31 16:54:07 +01:00
Evgeny 285fd93c32 smp server: notices about server pages and port 443, enable store log by default (#1501)
* smp server: notices about server pages and port 443, enable store log by default

* enable store log by default for ntf and xftp servers, fix tests
2025-03-31 16:41:11 +01:00
Evgeny 56bec06856 smp protocol: remove creating notifications from NEW command, new ideas for notifications (#1500) 2025-03-30 19:56:16 +01:00
Evgeny 04cbed90fb agent: set/update and delete contact short link data (#1499)
* agent: set/update and delete contact short link data

* delete contact link data, tests

* comments

* type signature for GHC 8.10.7
2025-03-30 11:17:25 +01:00
Evgeny c1a6647f19 agent: join connection when 1-time invitation short link is already secured (#1496)
* agent: join connection when 1-time invitation short link is already secured

* do not pass short link to join

* delete short link record after connection
2025-03-29 09:42:56 +00:00
Evgeny Poberezkin 1dd677eec2 Merge branch 'master' into short-links 2025-03-28 19:50:18 +00:00
Evgeny 7636bc7491 smp server: remove locks for deleted queues, additional statistics for objects in memory (#1498)
* smp server: remove locks for deleted queues, additional statistics for objects in memory

* version

* reduce queue cache usage

* less caching, refactor

* comments

* revert version
2025-03-28 18:51:54 +00:00
sh 79adb83782 ci: add reproduce builds workflow (#1497)
* ci: add reproduce builds workflow

* ci: trigger webhook and build every day

* ci: change secrets
2025-03-27 16:06:32 +00:00
Evgeny b83d897650 smp protocol: short links and other changes from RFC (#1489)
* smp protocol: short links types and other changes from RFC

* add fields for queue link ID and data

* create queue and ntf credentials with NEW command

* all tests

* simplfiy types, update rfc

* update rfc

* include SenderId in NEW request in case queue data is sent

* store queue data and generate link ID if needed

* update rfc

* agent API and types

* SMP commands and persistence for short links

* SMP client functions for short links

* agent client functions for short links

* create rcv queue with short link (TODO secret_box)

* encryption and encoding for link data, postgres client migration

* test creating short link

* get link and data, tests

* comments

* type signature
2025-03-26 17:26:27 +00:00
Evgeny Poberezkin 0c3b25706a Merge branch 'stable' 2025-03-23 18:18:07 +00:00
sh c3f57beafd scripts: fix script for reproducible builds (#1492)
* scripts: fix script for reproducible builds

* scripts: add automatic hash comparasion + compute file with checksums

* scripts: silence cleanup

* ci/scripts: enable parallel build (#1493)
2025-03-23 18:17:27 +00:00
Evgeny Poberezkin aace3fd2fb 6.3.1.0 2025-03-22 13:51:55 +00:00
sh 2e67ed9c4c ci/scripts: refactor reproducible builds (#1490)
* ci: build cache, remove 20.04, minor refactor

* scripts: update build reproducibility script

* script: add postgresql reproducibility

* script: no cache when building image

* scripts: force symlinks

* scripts: also download prebuilt postgresql binary

* scripts: add missing app var
2025-03-22 13:27:11 +00:00
Evgeny 614fa2b163 smp server: reduce queue expiration/idle intervals, skip expiring very old queues (#1488)
* smp server: reduce idle queue interval and queue expiration interval

* only expire recent queues (not very old)

* fix

* version
2025-03-20 08:57:47 +00:00
Evgeny 903e96bdfa rfc: smp queue protocol changes (#1485)
* rfc: smp queue protocol changes

* questions

* update

* update

* correction
2025-03-19 10:47:57 +00:00
Evgeny 5c0adcbbff Merge pull request #1467 from simplex-chat/postgres
SMP server PostgreSQL storage - feature branch (merge, dont squash)
2025-03-18 16:48:25 +00:00
sh b2f16eeff4 ci: seperate builds (pg/standard) (#1486) 2025-03-18 09:42:53 +00:00
Evgeny 6db79808aa smp server: use COPY to import store log to postgres db, improve concurrency and error handling (#1487)
* smp server: use COPY to import store log to postgres db

* compact queues when importing to postgres

* mempty

* version

* handle errors while expiring, mask async exceptions while getting queue

* whitespace

* version
2025-03-18 09:40:22 +00:00
Evgeny f4b55bfc0c smp server: CLI option to compact store log with PostgreSQL database (#1484)
* smp server: CLI option to compact store log with PostgreSQL database

* version

* fix test
2025-03-16 18:47:02 +00:00
Evgeny fe64d42db1 smp server: split postgres support to a separate executable, to not require postgres library in the main binary (#1482)
* smp server: split postgres support to a separate executable, to not require postgres library in the main binary

* comments

* enable server_postgres flag by default, add CPP option to test

* refactor

* change default for server_postgres to False

* diff
2025-03-16 11:37:25 +00:00
Evgeny 1b5a9f3b0c smp server: do not cache all queues from database while processing expirations (#1483)
* smp server: expire only active queues

* version

* do not cache all queues while processing expirations

* refactor

* foldWithOptions_

* version

* use shared lock when expiring all queues

* use TMVar

* comment

* rename

* remove fold options

* do not create locks in the Map for temporarily loaded queues

* fix

* revert version
2025-03-15 20:11:30 +00:00
Evgeny fdf8bd7ee2 smp server: mask database store operations from async exception for state consistency (#1481)
* smp server: mask database store operations from async exception for state consistency

* remove unused, comments
2025-03-13 13:57:46 +00:00
Evgeny 019a32a623 smp server: optionally append store log with postgres storage (without loading and compacting, for debugging during migration) (#1480)
* smp server: optionally maintain store log with postgres storage (without loading and compacting, for debugging during migration)

* refactor

* remove comment
2025-03-13 10:25:56 +00:00
Evgeny d44f09d111 tests: pass AStoreType to agent test as a parameter (#1479)
* tests: run agent tests with PostgreSQL SMP servers

* agent tests with postgres database

* enable tests

* fix store log tests

* fix test
2025-03-12 21:47:44 +00:00
Evgeny 6b5de2c51b smp server: exception on invalid migration, check queue storage settings on server start (#1478)
* smp server: check queue storage settings on server start

* fix incorrect postgres migration not throwing exception
2025-03-10 17:28:26 +00:00
shum 9410fb6f16 docker: add libpq dep 2025-03-10 10:03:44 +00:00
Evgeny 7b42aaa132 smp server: expire messages in postgres database, mark queues as deleted, combine tables (#1471)
* smp server: expire messages in postgres database

* tty

* fail if nothing updated in db

* remove old deleted queues

* index

* fix tests
2025-03-10 09:31:50 +00:00
Evgeny Poberezkin e4b9aa9746 Merge branch 'master' into postgres 2025-03-09 21:58:39 +00:00
Evgeny Poberezkin aa26a55df4 Merge branch 'master' into stable 2025-03-09 21:53:03 +00:00
sh 6e505f5c0b scripts: add script to reproduce binaries locally (#1477) 2025-03-08 20:51:20 +00:00
Evgeny Poberezkin a491a1d878 6.3.0.8 2025-03-07 14:30:00 +00:00
shandEvgeny Poberezkin 36f5539b9a ci: introduce reproducible builds (#1476)
* ci: introduce reproducible builds

* ci: return 20.04

* smp server: increase timing test threshold

* ci: test outside docker

* ci: fix test step

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-03-07 14:15:18 +00:00
Evgeny 1a2afe8bfd agent: fix JSON encoding for protocol errors to be compatible with iOS (#1475) 2025-03-06 08:02:27 +00:00
Evgeny Poberezkin 9fece9ce3d 6.3.0.7 2025-03-02 22:45:07 +00:00
Evgeny Poberezkin 2e2ede5968 Merge branch 'master' into postgres 2025-03-02 22:35:28 +00:00
Evgeny 205d4ead1c smp server: remove store log backups when server starts (#1472) 2025-02-27 07:39:05 +00:00
Evgenyandspaced4ndy 80a070a8ea smp server: update STM cache only after PostgreSQL update (#1470)
* refactor journal store

* update postgres store methods to update STM cache only after DB update

* todos

* drop

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>

* fix GHC 8.10.7

---------

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>
2025-02-26 11:43:08 +00:00
Evgeny 172540984c agent, smp server: PostgreSQL connection pool (#1468)
* agent, smp server: PostgreSQL connection pool

* simplify, create all connections before start

* rename

* remove imports
2025-02-25 15:33:54 +00:00
Evgenyandspaced4ndy 4dc40bd795 smp server: PostgreSQL queue store (#1448)
* smp server: queue store typeclass

* parameterize JournalMsgStore

* typeclass for queue store

* postgres WIP

* compiles, passes tests

* remove StoreType

* split migrations

* progress

* addQueueRec

* reduce type spaghetti

* remove addQueue from typeclass definition

* getQueue

* test postgres storage in SMP server

* fix schema

* comment

* import queues to postgresql

* import queues to postgresql

* log

* fix test

* counts

* ci: test smp server with postgres backend (#1463)

* ci: test smp server with postgres backend

* postgres service

* attempt

* attempt

* empty

* empty

* PGHOST attempt

* PGHOST + softlink attempt

* only softlink attempt

* working attempt (PGHOST)

* remove env var

* empty

* do not start server without DB schema, do not import when schema exists

* export database

* enable all tests, disable two tests

* option for migration confirmation

* comments

---------

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>
2025-02-24 15:47:27 +00:00
Evgeny Poberezkin f9d7b1eebc 6.3.0.6 2025-02-23 19:42:16 +00:00
Evgeny ffbc733d58 smp server: remove duplicate progress log (#1466) 2025-02-23 19:40:49 +00:00
Evgeny 2286726d72 smp server: start options maintenance and skip-warnings (#1465)
* smp server: start options `maintenance` and `skip-warnings`

* ignore invalid parsing of the last lines

* parsingErr

* fix
2025-02-22 19:26:03 +00:00
spaced4ndy 1b8110a332 xftp server: restore file status from log (#1461)
* xftp server: restore file blocking info from log

* fix parse

* rework

* update

* rename
2025-02-21 22:43:56 +00:00
sh dad7e1b60c simplex-servers-update: download scripts from tag (#1457)
* simplex-servers-update: download scripts from tag

* set remote version when re-executing the update script

* safeguard variables

* additional checks
2025-02-20 09:26:03 +00:00
Evgeny 72c2ddcf57 agent: move migrations (#1459) 2025-02-18 23:39:29 +00:00
Evgeny a75e138965 smp server: remove empty queues journals when expiring messages of idle queues (#1458)
* smp server: remove empty queues journals when expiring messages of idle queues

* remove unnecessary update

* ci: update action

* rename
2025-02-18 20:04:58 +00:00
Evgeny fa319d798a smp server: remove empty journals when opening message queue (#1456)
* smp server: remove empty journals when opening message queue

* update, do not backup state

* test

* version

* do not close queue state when queue is opened for writing

* comment

* quota = 4

* refactor openMsgQueue to prevent extra state backups

* use interval in config

* version, expire backups after 5 min

* refactor

* test
2025-02-17 23:11:34 +00:00
Evgeny Poberezkin c192339af9 6.3.0.5 2025-02-15 15:36:44 +00:00
Evgeny fa67d128d1 agent: fix deleting messages after delivery to avoid deleting shared message bodies (#1455)
* agent: fix deleting messages after delivery to avoid deleting shared message bodies

* fix, comments

* rename

* comment
2025-02-15 12:40:10 +00:00
Evgeny 55ff581655 SMP client: dont block on writing to sending queues (#1454)
* SMP client: dont block on writing to sending queues

* only fork if full
2025-02-15 11:31:05 +00:00
sh 560b257af7 scripts: simplex-servers-update menu to update only server binaries etc. (#1445)
* scripts/simplex-servers-update: refactor

* simplex-servers-update: don't check empty variable

* simplex-servers-update: minor decoration change

* simplex-servers-update: add new line when reexecuting script

* simplex-servers-update: refactor menus

* simplex-servers-update: fix re-executing

* simplex-servers-update: fix menus

* simplex-servers-update: do not rerended menu every time

* simplex-servers-update: update menus logic

* simplex-servers-update: ask permission to update only if versions differ

* simplex-servers-update: skip empty variable

* simplex-servers-update: include while loop

* simplex-servers-update: update re-executing skip logic

* simplex-servers-update: ignore empty

* simplex-servers-update: s/while/if

* simplex-servers-update: ignore empty variable

* simplex-servers-update: update skipping scripts

* simplex-servers-update: minor fixes

* simplex-servers-update: add spaces

* simplex-servers-update: safeguard files and add cleanup logic

* simplex-servers-update: update menus

* simplex-servers-update: fix binary check

* simplex-servers-update: adjust erasing lines

* simplex-servers-update: simplify bin update logic

* simplex-servers-update: fix binary path

* simplex-servers-update: format errors

* scripts/main/simplex-servers-update: initial refactor2

* simplex-servers-update: incr update

* simplex-servers-update: initial dynamic menu

* simplex-servers-update: misc fixes

* simplex-servers-update: fix VER

* simplex-servers-update: add VER check

* simplex-servers-update: abort early if there's no options available

* simplex-servers-update: add new helper function and update_misc

* simplex-servers-update: pass selection

* simplex-servers-update: non-interactive usage

* simplex-servers-update: add help menu

* simplex-servers-update: misc fixes and new binary update logic

* simplex-servers-update: local path check

* simplex-servers-update: fix change check

* simplex-servers-update: OLF -> OLD

* simplex-servers-update: always return code from functions

* simplex-servers-update: new menu flow

* simplex-servers-update: fixes

* simplex-servers-update: download function + fixes

* simplex-servers-update: s/bins/binaries

* simplex-servers-update: download message

* simplex-servers-update: cleanup
2025-02-15 10:31:19 +00:00
spaced4ndyandEvgeny Poberezkin 7ac80bffcb agent: store shared message body only once (when it is the same across messages when batching) (#1453)
* agent: store shared message body only once (when it is the same across messages when batching)

* rename

* refactor

* refactor

* save bodies and messages in single transaction

* comment

* comment

* comment

* box

* mapME

* box

* ValueOrRef

* remove instances

* refactor

* comments

* test

* refactor

* mapAccumLM compatibility with ghc 8.10.7

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-02-14 18:01:40 +00:00
spaced4ndyandEvgeny Poberezkin 0d8a1a2879 agent: encrypt messages on delivery (#1446)
* agent: save message body once (plan, schema)

* split

* new type

* bs

* encrypt on delivery

* schema

* fix test

* check pad size

* rename

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-02-14 12:35:18 +00:00
Evgeny Poberezkin bd97cb0449 6.3.0.4 2025-02-10 15:12:54 +00:00
Evgeny a58d3540ad agent: refactor migrations (#1452)
* agent: refactor migrations

* fix postgres
2025-02-07 17:17:53 +00:00
Evgenyandspaced4ndy 5dbe6337ea ntf server: additional statistics, new invalid token reasons (#1451)
* agent: check ntf token status on registration

* remove check

* update on check

* refactor

* version

* fix

* ntf server: additional statistics

* swap

* version

* more stats

* test, verify invalid

* rename

* exclude test token from stats

* increase delay

* handle invalid token in retry, more reasons

* focus tests

* disable new tests in CI

* fix

---------

Co-authored-by: spaced4ndy <8711996+spaced4ndy@users.noreply.github.com>
2025-02-07 12:19:11 +00:00
spaced4ndyandEvgeny Poberezkin b633f89c1a agent: check ntf token status on registration (#1450)
* agent: check ntf token status on registration

* remove check

* update on check

* refactor

* version

* fix

* test, verify invalid

* rename

* increase delay

* disable new tests in CI

* fix

---------

Co-authored-by: Evgeny Poberezkin <evgeny@poberezkin.com>
2025-02-07 11:36:29 +00:00
Evgeny 944a22a2fb ntf server: record token invalidation reason, add date of the last token activity (#1449)
* ntf server: record token invalidation reason, add date of the last token activity

* update time

* rename

* optional

* include token ID in delivery error

* version

* protocol version

* fix, log error
2025-02-06 17:11:35 +00:00
Evgeny ce24f83b64 refactor STM queues (#1447) 2025-02-05 12:04:27 +00:00
sh be81fe1f74 scripts/systemd: update services (#1440) 2025-01-24 13:37:56 +00:00
387 changed files with 75332 additions and 8933 deletions
+44
View File
@@ -0,0 +1,44 @@
name: 'Set Swap Space'
description: 'Add moar swap'
branding:
icon: 'crop'
color: 'orange'
inputs:
swap-size-gb:
description: 'Swap space to create, in Gigabytes.'
required: false
default: '10'
runs:
using: "composite"
steps:
- name: Swap space report before modification
shell: bash
run: |
echo "Memory and swap:"
free -h
echo
swapon --show
echo
- name: Set Swap
shell: bash
run: |
export SWAP_FILE=$(swapon --show=NAME | tail -n 1)
echo "Swap file: $SWAP_FILE"
if [ -z "$SWAP_FILE" ]; then
SWAP_FILE=/opt/swapfile
else
sudo swapoff $SWAP_FILE
sudo rm $SWAP_FILE
fi
sudo fallocate -l ${{ inputs.swap-size-gb }}G $SWAP_FILE
sudo chmod 600 $SWAP_FILE
sudo mkswap $SWAP_FILE
sudo swapon $SWAP_FILE
- name: Swap space report after modification
shell: bash
run: |
echo "Memory and swap:"
free -h
echo
swapon --show
echo
+260 -56
View File
@@ -10,62 +10,25 @@ on:
pull_request:
jobs:
build:
name: build-${{ matrix.os }}-${{ matrix.ghc }}
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
include:
- os: ubuntu-20.04
platform_name: 20_04-x86-64
ghc: "8.10.7"
- os: ubuntu-20.04
platform_name: 20_04-x86-64
ghc: "9.6.3"
- os: ubuntu-22.04
platform_name: 22_04-x86-64
ghc: "9.6.3"
# =============================
# Create release
# =============================
# Create release, but only if it's triggered by tag push.
# On pull requests/commits push, this job will always complete.
maybe-release:
runs-on: ubuntu-latest
steps:
- name: Clone project
if: startsWith(github.ref, 'refs/tags/v')
uses: actions/checkout@v3
- name: Setup Haskell
uses: haskell-actions/setup@v2
with:
ghc-version: ${{ matrix.ghc }}
cabal-version: "3.10.1.0"
- name: Cache dependencies
uses: actions/cache@v2
with:
path: |
~/.cabal/store
dist-newstyle
key: ${{ matrix.os }}-${{ hashFiles('cabal.project', 'simplexmq.cabal') }}
- name: Build
shell: bash
run: cabal build --enable-tests
- name: Test
timeout-minutes: 40
shell: bash
run: cabal test --test-show-details=direct
- name: Prepare binaries
if: startsWith(github.ref, 'refs/tags/v')
shell: bash
run: |
mv $(cabal list-bin smp-server) smp-server-ubuntu-${{ matrix.platform_name}}
mv $(cabal list-bin ntf-server) ntf-server-ubuntu-${{ matrix.platform_name}}
mv $(cabal list-bin xftp-server) xftp-server-ubuntu-${{ matrix.platform_name}}
mv $(cabal list-bin xftp) xftp-ubuntu-${{ matrix.platform_name}}
- name: Build changelog
if: startsWith(github.ref, 'refs/tags/v')
id: build_changelog
uses: mikepenz/release-changelog-builder-action@v1
if: startsWith(github.ref, 'refs/tags/v')
uses: simplex-chat/release-changelog-builder-action@v5
with:
configuration: .github/changelog_conf.json
failOnError: true
@@ -75,8 +38,8 @@ jobs:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Create release
if: startsWith(github.ref, 'refs/tags/v') && matrix.ghc != '8.10.7'
uses: softprops/action-gh-release@v1
if: startsWith(github.ref, 'refs/tags/v')
uses: simplex-chat/action-gh-release@v2
with:
body: |
See full changelog [here](https://github.com/simplex-chat/simplexmq/blob/master/CHANGELOG.md).
@@ -86,10 +49,251 @@ jobs:
prerelease: true
files: |
LICENSE
smp-server-ubuntu-${{ matrix.platform_name}}
ntf-server-ubuntu-${{ matrix.platform_name}}
xftp-server-ubuntu-${{ matrix.platform_name}}
xftp-ubuntu-${{ matrix.platform_name}}
fail_on_unmatched_files: true
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
# =============================
# Main build job
# =============================
build:
name: "ubuntu-${{ matrix.os }}-${{ matrix.arch }}, GHC: ${{ matrix.ghc }}"
needs: maybe-release
env:
apps: "smp-server xftp-server ntf-server xftp"
runs-on: ${{ matrix.runner }}
services:
postgres:
image: postgres:15
env:
POSTGRES_HOST_AUTH_METHOD: trust # Allows passwordless access
options: >-
--health-cmd pg_isready
--health-interval 10s
--health-timeout 5s
--health-retries 5
ports:
# Maps tcp port 5432 on service container to the host
- 5432:5432
strategy:
fail-fast: false
matrix:
include:
- os: 22.04
os_underscore: 22_04
arch: x86-64
runner: "ubuntu-22.04"
ghc: "8.10.7"
should_run: ${{ !(github.ref == 'refs/heads/stable' || startsWith(github.ref, 'refs/tags/v')) }}
- os: 22.04
os_underscore: 22_04
arch: x86-64
runner: "ubuntu-22.04"
ghc: "9.6.3"
should_run: true
- os: 24.04
os_underscore: 24_04
arch: x86-64
runner: "ubuntu-24.04"
ghc: "9.6.3"
should_run: true
- os: 22.04
os_underscore: 22_04
arch: aarch64
runner: "ubuntu-22.04-arm"
ghc: "9.6.3"
should_run: true
- os: 24.04
os_underscore: 24_04
arch: aarch64
runner: "ubuntu-24.04-arm"
ghc: "9.6.3"
should_run: true
steps:
- name: Clone project
if: matrix.should_run == true
uses: actions/checkout@v3
- name: Set up Docker Buildx
if: matrix.should_run == true
uses: simplex-chat/docker-setup-buildx-action@v3
- name: Setup swap
if: matrix.ghc == '8.10.7' && matrix.should_run == true
uses: ./.github/actions/swap
with:
swap-size-gb: 20
- name: Install PostgreSQL 15 client tools
if: matrix.os == '22.04' && matrix.should_run == true
shell: bash
run: |
# Import the repository signing key
sudo install -d /usr/share/postgresql-common/pgdg
sudo curl -o /usr/share/postgresql-common/pgdg/apt.postgresql.org.asc --fail https://www.postgresql.org/media/keys/ACCC4CF8.asc
# Add the PostgreSQL APT repository
sudo sh -c 'echo "deb [signed-by=/usr/share/postgresql-common/pgdg/apt.postgresql.org.asc] https://apt.postgresql.org/pub/repos/apt $(lsb_release -cs)-pgdg main" > /etc/apt/sources.list.d/pgdg.list'
# Update repository and install postgresql tools
sudo apt update
sudo apt -y install postgresql-client-15
- name: Build and cache Docker image
if: matrix.should_run == true
uses: simplex-chat/docker-build-push-action@v6
with:
context: .
load: true
file: Dockerfile.build
tags: build/${{ matrix.os }}:latest
build-args: |
TAG=${{ matrix.os }}
GHC=${{ matrix.ghc }}
- name: Cache dependencies
if: matrix.should_run == true
uses: actions/cache@v4
with:
path: |
~/.cabal/store
dist-newstyle
key: ubuntu-${{ matrix.os }}-${{ matrix.arch }}-ghc${{ matrix.ghc }}-${{ hashFiles('cabal.project', 'simplexmq.cabal') }}
- name: Start container
if: matrix.should_run == true
shell: bash
run: |
docker run -t -d \
--device /dev/fuse \
--cap-add SYS_ADMIN \
--security-opt apparmor:unconfined \
--name builder \
-v ~/.cabal:/root/.cabal \
-v /home/runner/work/_temp:/home/runner/work/_temp \
-v ${{ github.workspace }}:/project \
build/${{ matrix.os }}:latest
- name: Build smp-server, xftp-server (postgresql) and tests
if: matrix.should_run == true
shell: docker exec -t builder sh -eu {0}
run: |
chmod -fR 777 ~/.cabal ./dist-newstyle || :; git config --global --add safe.directory '*'
cabal clean
cabal update
cabal build --jobs=$(nproc) --enable-tests -fserver_postgres
mkdir -p /out
for i in smp-server xftp-server simplexmq-test; do
bin=$(find /project/dist-newstyle -name "$i" -type f -executable)
chmod +x "$bin"
mv "$bin" /out/
done
strip /out/smp-server /out/xftp-server
- name: Copy simplexmq-test from container
if: matrix.should_run == true
shell: bash
run: |
docker cp builder:/out/simplexmq-test .
- name: Copy smp-server, xftp-server (postgresql) from container and prepare it
if: startsWith(github.ref, 'refs/tags/v') && matrix.should_run == true
id: prepare-postgres
shell: bash
run: |
printf 'bins<<EOF\n' > bins.output
printf 'hashes<<EOF\n' > hashes.output
for i in smp-server xftp-server; do
name="${i}-postgres-ubuntu-${{ matrix.os_underscore }}-${{ matrix.arch }}"
docker cp builder:/out/$i $name
path="${{ github.workspace }}/$name"
hash="SHA2-256($name)= $(openssl sha256 $path | cut -d' ' -f 2)"
printf '%s\n' "$path" >> bins.output
printf '%s\n\n' "$hash" >> hashes.output
done
printf 'EOF\n' >> bins.output
printf 'EOF\n' >> hashes.output
cat bins.output >> "$GITHUB_OUTPUT"
cat hashes.output >> "$GITHUB_OUTPUT"
- name: Build everything else (standard)
if: matrix.should_run == true
shell: docker exec -t builder sh -eu {0}
run: |
cabal build --jobs=$(nproc)
mkdir -p /out
for i in ${{ env.apps }}; do
bin=$(find /project/dist-newstyle -name "$i" -type f -executable)
strip "$bin"
chmod +x "$bin"
mv "$bin" /out/
done
- name: Copy binaries from container and prepare them
id: prepare-regular
if: startsWith(github.ref, 'refs/tags/v') && matrix.should_run == true
shell: bash
run: |
docker cp builder:/out .
printf 'bins<<EOF\n' > bins.output
printf 'hashes<<EOF\n' > hashes.output
for i in ${{ env.apps }}; do
name="$i-ubuntu-${{ matrix.os_underscore }}-${{ matrix.arch }}"
mv ./out/$i ./$name
path="${{ github.workspace }}/$name"
hash="SHA2-256($name)= $(openssl sha256 $path | cut -d' ' -f 2)"
printf '%s\n' "$path" >> bins.output
printf '%s\n\n' "$hash" >> hashes.output
done
printf 'EOF\n' >> bins.output
printf 'EOF\n' >> hashes.output
cat bins.output >> "$GITHUB_OUTPUT"
cat hashes.output >> "$GITHUB_OUTPUT"
- name: Upload binaries
if: startsWith(github.ref, 'refs/tags/v') && matrix.should_run == true
uses: simplex-chat/action-gh-release@v2
with:
append_body: true
prerelease: true
fail_on_unmatched_files: true
body: |
${{ steps.prepare-regular.outputs.hashes }}
${{ steps.prepare-postgres.outputs.hashes }}
files: |
${{ steps.prepare-regular.outputs.bins }}
${{ steps.prepare-postgres.outputs.bins }}
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Test
if: matrix.should_run == true && matrix.arch == 'x86-64'
timeout-minutes: 120
shell: bash
env:
PGHOST: localhost
run: |
i=1
attempts=1
${{ (github.ref == 'refs/heads/stable' || startsWith(github.ref, 'refs/tags/v')) }} && attempts=3
while [ "$i" -le "$attempts" ]; do
if ./simplexmq-test; then
break
else
echo "Attempt $i failed, retrying..."
i=$((i + 1))
sleep 1
fi
done
if [ "$i" -gt "$attempts" ]; then
echo "All "$attempts" attempts failed."
exit 1
fi
+11 -3
View File
@@ -22,14 +22,14 @@ jobs:
uses: actions/checkout@v4
- name: Log in to Docker Hub
uses: docker/login-action@v3
uses: simplex-chat/docker-login-action@v3
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_PASSWORD }}
- name: Extract metadata for Docker image
id: meta
uses: docker/metadata-action@v5
uses: simplex-chat/docker-metadata-action@v5
with:
images: ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.app }}
flavor: |
@@ -39,9 +39,17 @@ jobs:
type=semver,pattern=v{{major}}.{{minor}}
type=semver,pattern=v{{major}}
- name: Set up QEMU
uses: docker/setup-qemu-action@v3
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Build and push Docker image
uses: docker/build-push-action@v6
uses: simplex-chat/docker-build-push-action@v6
with:
context: .
platforms: linux/amd64,linux/arm64
push: true
build-args: |
APP=${{ matrix.app }}
+45
View File
@@ -0,0 +1,45 @@
name: Reproduce latest release
on:
workflow_dispatch:
schedule:
- cron: '0 2 * * *' # every day at 02:00 night
jobs:
reproduce:
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v3
- name: Get latest release
shell: bash
run: |
curl --proto '=https' \
--tlsv1.2 \
-sSf -L \
'https://api.github.com/repos/simplex-chat/simplexmq/releases/latest' \
2>/dev/null | \
grep -i "tag_name" | \
awk -F \" '{print "TAG="$4}' >> $GITHUB_ENV
- name: Execute reproduce script
run: |
${GITHUB_WORKSPACE}/scripts/simplexmq-reproduce-builds.sh "$TAG" || :
- name: Check if build has been reproduced
env:
url: ${{ secrets.STATUS_SIMPLEX_WEBHOOK_URL }}
user: ${{ secrets.STATUS_SIMPLEX_WEBHOOK_USER }}
pass: ${{ secrets.STATUS_SIMPLEX_WEBHOOK_PASS }}
run: |
if [ -f "${GITHUB_WORKSPACE}/${TAG}-simplexmq/_sha256sums" ]; then
exit 0
else
curl --proto '=https' --tlsv1.2 -sSf \
-u "${user}:${pass}" \
-H 'Content-Type: application/json' \
-d '{"title": "👾 GitHub: Runner", "description": "⛔️ '"$TAG"' did not reproduce."}' \
"$url"
exit 1
fi
+1
View File
@@ -11,3 +11,4 @@ cabal.project.local~
.hpc/
*.tix
.coverage
@@ -0,0 +1,53 @@
# XFTPClientAgent Pattern
## TOC
1. Executive Summary
2. Changes: client.ts
3. Changes: agent.ts
4. Changes: test/browser.test.ts
5. Verification
## Executive Summary
Add `XFTPClientAgent` — a per-server connection pool matching the Haskell pattern. The agent caches `XFTPClient` instances by server URL. All orchestration functions (`uploadFile`, `downloadFile`, `deleteFile`) take `agent` as first parameter and use `getXFTPServerClient(agent, server)` instead of calling `connectXFTP` directly. Connections stay open on success; the caller creates and closes the agent.
`connectXFTP` and `closeXFTP` stay exported (used by `XFTPWebTests.hs` Haskell tests). The `browserClients` hack, per-function `connections: Map`, and `getOrConnect` are deleted.
## Changes: client.ts
**Add** after types section: `XFTPClientAgent` interface, `newXFTPAgent`, `getXFTPServerClient`, `closeXFTPServerClient`, `closeXFTPAgent`.
**Delete**: `browserClients` Map and all `isNode` browser-cache checks in `connectXFTP` and `closeXFTP`.
**Revert `closeXFTP`** to unconditional `c.transport.close()` (browser transport.close() is already a no-op).
`connectXFTP` stays exported (backward compat) but becomes a raw low-level function — no caching.
## Changes: agent.ts
**Imports**: replace `connectXFTP`/`closeXFTP` with `getXFTPServerClient`/`closeXFTPAgent` etc.
**Re-export** from agent.ts: `newXFTPAgent`, `closeXFTPAgent`, `XFTPClientAgent`.
**`uploadFile`**: add `agent: XFTPClientAgent` as first param. Replace `connectXFTP``getXFTPServerClient`. Remove `finally { closeXFTP }`. Pass `agent` to `uploadRedirectDescription`.
**`uploadRedirectDescription`**: change from `(client, server, innerFd)` to `(agent, server, innerFd)`. Get client via `getXFTPServerClient`.
**`downloadFile`**: add `agent` param. Delete local `connections: Map`. Replace `getOrConnect``getXFTPServerClient`. Remove finally cleanup. Pass `agent` to `downloadWithRedirect`.
**`downloadWithRedirect`**: add `agent` param. Same replacements. Remove try/catch cleanup. Recursive call passes `agent`.
**`deleteFile`**: add `agent` param. Same pattern.
**Delete**: `getOrConnect` function entirely.
## Changes: test/browser.test.ts
Create agent before operations, pass to upload/download, close in finally.
## Verification
1. `npx vitest --run` — browser round-trip test passes
2. No remaining `browserClients`, `getOrConnect`, or per-function `connections: Map` locals
3. `connectXFTP` and `closeXFTP` still exported (XFTPWebTests.hs compat)
4. All orchestration functions take `agent` as first param
+137 -1
View File
@@ -1,3 +1,139 @@
# 6.5.1
Version 6.5.1.0
XFTP client:
- backwards compatible file header decoding.
# 6.5.0
Version 6.5.0.17
SMP agent:
- improve subscriptions
- reduce memory usage and retries during initial subscription (#1758)
- fix race resulting in pending subscriptions never subscribed (#1756)
- batch processing of subscription results and errors (#1652)
- reduce memory usage of active subscriptions.
- drop message after N reception attempts (#1762)
- fix possible deadlocks of queue overloading when processing messages (#1713)
- improved APIs for short link management and creation.
- support multiple link owners in link data (#1701)
SMP server:
- store messages in PostgreSQL (#1622).
- reduce memory usage with PostgreSQL database - do not use queue cache (#1637)
- fix in-memory server not restoring queue/service associations after 2+ restarts (#1618)
XFTP server:
- support PostgreSQL database.
- add server page.
- support uploads from web clients.
Servers:
- better socket leak prevention during TLS handshake, NetworkError type to bette diagnose connection errors (#1619)
- use "=" as default INI key-value separator (#1767)
# 6.4.4
Servers:
- fix server pages when source code is not specified.
- include commit SHA in printed version and in web page (#1608).
SMP server:
- support short SimpleX addresses in server information page (#1600).
- wrap all queries in transactions (#1603).
SMP agent:
- chat relay address type for short links (#1602).
- extend xrcp certificate validity 1 hour in the past, to allow out of sync clocks (#1601).
# 6.4.3
SMP agent:
- fix some connection errors by updating contact request server hosts to match server in short link (#1597).
SMP server:
- support short link URI as queue identifier in control port commands (#1596).
# 6.4.2
SMP server:
- fix memory leak when connection interrupts straight after client connects.
- do not include repeated queue blocking into stats/quota.
XFTP server:
- prometheus metrics
# 6.4.1
SMP protocol:
- create notification credentials via NEW command that creates the queue (#1586)
SMP server:
- control port session improvements (#1591)
- additional stat counter for ntf credentials created together with the queue (#1589)
# 6.4.0
SMP protocol (server/client):
- support associated queue data and short connection links (see [RFC](./rfcs/2025-03-16-smp-queues.md)).
- service certificates to optimize subscriptions.
SMP agent:
- support retries for interactive connection handshakes.
- use web port 443 by default for preset servers.
- use static RNG function to avoid creating dynamic C stubs when generating sntrup keys (it was detected as Dynamic Code Loading in GrapheneOS).
- different timeouts for interactive and background operations.
Ntf server:
- PostgreSQL storage.
- Prometheus metrics.
- use service certificates.
- fix repeat token registration.
# 6.3.2
Servers:
- enable store log by default (#1501).
SMP server:
- reduce memory usage (#1498)
SMP agent:
- handle client/agent version downgrades after connection was established (#1508).
# 6.3.1
Servers:
- handle ECONNABORTED error on client connections.
- reproducible builds.
- blocking records for content moderation.
- update script (simplex-servers-update) downloads scripts from the specified or the latest stable tag.
SMP server:
- support for PostgreSQL database for queue records for higher traffic servers.
- fix old clients sending messages to new servers (#1443)
- remove empty journals when opening message queues and expiring idle queues (#1456, #1458).
- additional start options (#1465):
- `maintenance` to run all start/stop operations without starting server.
- `skip-warnings` to ignore the last corrupted line in store log (can happen on abnormal termination).
Ntf server:
- record date of last token activity, to allow expiring inactive tokens.
- additional token invalidation reasons in logs.
SMP agent:
- store message sent to multiple connections only once, to reduce storage when sending to groups (#1453).
- encrypt messages on delivery, to reduce database writes (#1446).
- don't block method calls on congested sockets for better concurrency (#1454).
- check notification token status on client connection.
- option to skip SQLite vacuum on migrations.
# 6.3.0
SMP agent: fix joining connection after failure by using the same ratchet.
# 6.2.2
SMP server:
@@ -32,7 +168,7 @@ Servers: more reliable restoring of state.
SMP server: reduced memory usage and faster start.
Notifications: compensate for iOS notifications being droppted by Apple while device is offline (#1378):
Notifications: compensate for iOS notifications being dropped by Apple while device is offline (#1378):
- Ntf server: send multiple SMP notifications in one iOS notification.
- Agent: get multiple messages for one iOS notification.
+31
View File
@@ -0,0 +1,31 @@
# syntax=docker/dockerfile:1.7.0-labs
ARG TAG=24.04
FROM ubuntu:${TAG} AS build
### Build stage
ARG GHC=9.6.3
ARG CABAL=3.14.1.1
# Install curl, git and and simplexmq dependencies
RUN apt-get update && apt-get install -y curl libpq-dev git sqlite3 libsqlite3-dev build-essential libgmp3-dev zlib1g-dev llvm llvm-dev libnuma-dev libssl-dev
# Specify bootstrap Haskell versions
ENV BOOTSTRAP_HASKELL_GHC_VERSION=${GHC}
ENV BOOTSTRAP_HASKELL_CABAL_VERSION=${CABAL}
# Do not install Stack
ENV BOOTSTRAP_HASKELL_INSTALL_NO_STACK=true
ENV BOOTSTRAP_HASKELL_INSTALL_NO_STACK_HOOK=true
# Install ghcup
RUN curl --proto '=https' --tlsv1.2 -sSf https://get-ghcup.haskell.org | BOOTSTRAP_HASKELL_NONINTERACTIVE=1 sh
# Adjust PATH
ENV PATH="/root/.cabal/bin:/root/.ghcup/bin:$PATH"
# Set both as default
RUN ghcup set ghc "${GHC}" && \
ghcup set cabal "${CABAL}"
WORKDIR /project
+7 -7
View File
@@ -1,6 +1,6 @@
# SimpleXMQ
[![GitHub build](https://github.com/simplex-chat/simplexmq/workflows/build/badge.svg)](https://github.com/simplex-chat/simplexmq/actions?query=workflow%3Abuild)
[![GitHub build](https://github.com/simplex-chat/simplexmq/actions/workflows/build.yml/badge.svg)](https://github.com/simplex-chat/simplexmq/actions/workflows/build.yml)
[![GitHub release](https://img.shields.io/github/v/release/simplex-chat/simplexmq)](https://github.com/simplex-chat/simplexmq/releases)
📢 SimpleXMQ v1 is released - with many security, privacy and efficiency improvements, new functionality - see [release notes](https://github.com/simplex-chat/simplexmq/releases/tag/v1.0.0).
@@ -33,7 +33,7 @@ To initialize the server use `smp-server init -n <fqdn>` (or `smp-server init --
SMP server uses in-memory persistence with an optional append-only log of created queues that allows to re-start the server without losing the connections. This log is compacted on every server restart, permanently removing suspended and removed queues.
To enable store log, initialize server using `smp-server -l` command, or modify `smp-server.ini` created during initialization (uncomment `enable: on` option in the store log section). Use `smp-server --help` for other usage tips.
To enable store log, initialize server using `smp-server -l` command, or modify `smp-server.ini` created during initialization (uncomment `enable = on` option in the store log section). Use `smp-server --help` for other usage tips.
Starting from version 2.3.0, when store log is enabled, the server would also enable saving undelivered messages on exit and restoring them on start. This can be disabled via a separate setting `restore_messages` in `smp-server.ini` file. Saving messages would only work if the server is stopped with SIGINT signal (keyboard interrupt), if it is stopped with SIGTERM signal the messages would not be saved.
@@ -116,7 +116,7 @@ On Linux, you can deploy smp and xftp server using Docker. This will download im
2. Run your Docker container.
- `smp-server`
You must change **your_ip_or_domain**. `-e "pass=password"` is optional variable to password-protect your `smp` server:
```sh
docker run -d \
@@ -129,7 +129,7 @@ On Linux, you can deploy smp and xftp server using Docker. This will download im
```
- `xftp-server`
You must change **your_ip_or_domain** and **maximum_storage**.
```sh
docker run -d \
@@ -187,7 +187,7 @@ On Linux, you can build smp server using Docker.
3. Run your Docker container.
- `smp-server`
You must change **your_ip_or_domain**. `-e "pass=password"` is optional variable to password-protect your `smp` server:
```sh
docker run -d \
@@ -200,7 +200,7 @@ On Linux, you can build smp server using Docker.
```
- `xftp-server`
You must change **your_ip_or_domain** and **maximum_storage**.
```sh
docker run -d \
@@ -247,7 +247,7 @@ On Linux, you can build smp server using Docker.
`xftp-server`
```sh
cabal list-bin exe:xftp-server
cabal list-bin exe:xftp-server
```
- Initialize SMP server with `smp-server init [-l] -n <fqdn>` or `smp-server init [-l] --ip <ip>` - depending on how you initialize it, either FQDN or IP will be used for server's address.
+15
View File
@@ -0,0 +1,15 @@
{-# LANGUAGE TemplateHaskell #-}
module Web.Embedded where
import Data.FileEmbed (embedDir, embedFile)
import Simplex.Messaging.Server.Web (EmbeddedContent (..))
embeddedContent :: EmbeddedContent
embeddedContent =
EmbeddedContent
{ indexHtml = $(embedFile "apps/common/Web/static/index.html"),
linkHtml = $(embedFile "apps/common/Web/static/link.html"),
mediaContent = $(embedDir "apps/common/Web/static/media/"),
wellKnown = $(embedDir "apps/common/Web/static/.well-known/")
}
@@ -0,0 +1,49 @@
{
"applinks": {
"details": [
{
"appIDs": [
"5NN7GUYB6T.chat.simplex.app"
],
"components": [
{
"/": "/contact/*"
},
{
"/": "/contact"
},
{
"/": "/invitation/*"
},
{
"/": "/invitation"
},
{
"/": "/a/*"
},
{
"/": "/a"
},
{
"/": "/c/*"
},
{
"/": "/c"
},
{
"/": "/g/*"
},
{
"/": "/g"
},
{
"/": "/i/*"
},
{
"/": "/i"
}
]
}
]
}
}
@@ -0,0 +1,16 @@
[
{
"relation": [
"delegate_permission/common.handle_all_urls"
],
"target": {
"namespace": "android_app",
"package_name": "chat.simplex.app",
"sha256_cert_fingerprints": [
"5E:3E:DC:C2:00:FB:A8:D5:F4:88:F3:CA:4C:32:5B:05:78:C5:6A:9C:03:A1:CC:B5:92:9C:D7:5C:7E:57:E2:4D",
"3C:52:C4:FD:3C:AD:1C:07:C9:B0:0A:70:80:E3:58:FA:B9:FE:FC:B8:AF:5A:EC:14:77:65:F1:6D:0F:21:AD:85",
"AE:C1:95:DC:FD:46:14:BD:3A:91:EC:26:D1:D5:14:C8:75:71:C5:CC:8D:CF:48:08:3F:92:83:14:3C:A2:B9:A6"
]
}
}
]
@@ -105,6 +105,13 @@
class="text-[16px] leading-[26px] tracking-[0.01em] nav-link-text text-black dark:text-white before:bg-black dark:before:bg-white">Server
information</span></a>
</li>
<!-- <x-xftpConfig>
<li class="nav-link relative"><a href="/file"
class="flex items-center justify-between gap-2 lg:py-5 whitespace-nowrap"><span
class="text-[16px] leading-[26px] tracking-[0.01em] nav-link-text text-black dark:text-white before:bg-black dark:before:bg-white">File
transfer</span></a>
</li>
</x-xftpConfig> -->
</ul><a target="_blank" href="https://github.com/simplex-chat/simplex-chat#help-us-with-donations"
class="whitespace-nowrap flex items-center gap-1 self-center text-white dark:text-black text-[16px] font-medium tracking-[0.02em] rounded-[34px] bg-primary-light dark:bg-primary-dark py-3 lg:py-2 px-20 lg:px-5 mb-16 lg:mb-0">Donate</a>
</div>
@@ -223,11 +230,14 @@
<table id="public-info">
<tr class="text-grey-black dark:text-white text-base">
<td>Server version:</td>
<td>${version}</td>
<td>${version}<x-commit> / <a href="${commitSourceCode}/commit/${commit}" target="_blank">${shortCommit}</a></x-commit></td>
</tr>
<tr class="text-grey-black dark:text-white text-base">
<td>Source code:</td>
<td><a href="${sourceCode}" target="_blank">${sourceCode}</a></td>
<td>
<x-sourceCode><a href="${sourceCode}" target="_blank">${sourceCode}</a></x-sourceCode>
<x-noSourceCode>add to ${iniFileName} (required by <a href="https://github.com/simplex-chat/simplexmq/blob/stable/LICENSE" target="_blank">AGPLv3</a>)</x-noSourceCode>
</td>
</tr>
<x-website>
<tr class="text-grey-black dark:text-white text-base">
@@ -314,6 +324,7 @@
<h2 class="text-[30px] mb-[20px] leading-[28px] text-[#606C71] dark:text-white font-bold max-w-[475px]">
Configuration</h2>
<table id="config">
<x-smpConfig>
<tr class="text-grey-black dark:text-white text-base">
<td>Persistence:</td>
<td>${persistence}</td>
@@ -334,6 +345,25 @@
<td>Basic auth enabled:</td>
<td>${basicAuthEnabled}</td>
</tr>
</x-smpConfig>
<x-xftpConfig>
<tr class="text-grey-black dark:text-white text-base">
<td>File expiration:</td>
<td>${fileExpiration}</td>
</tr>
<tr class="text-grey-black dark:text-white text-base">
<td>Stats enabled:</td>
<td>${statsEnabled}</td>
</tr>
<tr class="text-grey-black dark:text-white text-base">
<td>New uploads allowed:</td>
<td>${newUploadsAllowed}</td>
</tr>
<tr class="text-grey-black dark:text-white text-base">
<td>Basic auth enabled:</td>
<td>${basicAuthEnabled}</td>
</tr>
</x-xftpConfig>
</table>
</div>
</div>
@@ -142,8 +142,7 @@
class="hidden xl:block h-screen pt-[66px] bg-white dark:bg-gradient-radial-mobile dark:lg:bg-gradient-radial">
<div class="container m-auto h-full flex items-center justify-between px-5">
<div class="flex flex-col items-start justify-center w-full">
<h1 class="text-[38px] leading-[43px] font-bold max-w-[500px] mb-[30px] primary-header-contact">You received a
1-time link to connect on SimpleX Chat</h1>
<h1 class="text-[38px] leading-[43px] font-bold max-w-[500px] mb-[30px] primary-header-contact">This is a one-time link of the SimpleX&nbsp;network user</h1>
<h2
class="text-[20px] leading-[28px] text-[#606C71] dark:text-white font-bold max-w-[475px] mb-[80px] secondary-header-contact">
Scan the QR code with the SimpleX Chat app on your phone or tablet.</h2>
@@ -184,10 +183,8 @@
class="block xl:hidden pt-[106px] py-[90px] bg-white dark:bg-gradient-radial-mobile dark:lg:bg-gradient-radial">
<div class="container m-auto px-5">
<div class="flex flex-col items-center">
<h1 class="text-[28px] font-bold text-center max-w-[602px] mb-[40px] primary-header-contact">You received a
1-time link to connect on SimpleX Chat</h1>
<p class="text-[20px] leading-[28px] text-grey-black dark:text-white font-medium mb-[30px]">To make a
connection:</p>
<h1 class="text-[28px] font-bold text-center max-w-[602px] mb-[40px] primary-header-contact">This is a one-time link of the SimpleX&nbsp;network user</h1>
<p class="text-[20px] leading-[28px] text-grey-black dark:text-white font-medium mb-[30px]">To make a connection:</p>
<div
class="flex flex-col justify-center items-center p-4 w-full max-w-[468px] min-h-[131px] rounded-[30px] border-[1px] border-[#A8B0B4] dark:border-white border-opacity-60 mb-6 relative">
<p class="text-xl font-medium text-grey-black dark:text-white mb-4">Install SimpleX app</p>
@@ -506,13 +503,17 @@
const url = window.location.href
const messageElements = document.getElementsByClassName('primary-header-contact')
if (url.includes('/invitation')) {
for (let element of messageElements) {
element.textContent = 'You received a 1-time link to connect on SimpleX Chat'
}
} else {
for (let element of messageElements) {
element.textContent = 'You received an address to connect on SimpleX Chat'
for (let element of messageElements) {
if (url.includes('/g') || url.includes('&data=%7B%22groupLinkId%22%3A')) {
element.innerHTML = 'This is a public group address on SimpleX&nbsp;network'
} else if (url.includes('/a') || url.includes('/contact')) {
element.innerHTML = 'This is a public address of the SimpleX&nbsp;network user'
} else if (url.includes('/i') || url.includes('/invitation')) {
element.innerHTML = 'This is a one-time link of the SimpleX&nbsp;network user'
} else if (url.includes('/c')) {
element.innerHTML = 'This is a public channel address on SimpleX&nbsp;network'
} else if (url.includes('/r')) {
element.innerHTML = 'This is a chat relay address on SimpleX&nbsp;network'
}
}
</script>

Before

Width:  |  Height:  |  Size: 18 KiB

After

Width:  |  Height:  |  Size: 18 KiB

Before

Width:  |  Height:  |  Size: 16 KiB

After

Width:  |  Height:  |  Size: 16 KiB

@@ -20,7 +20,18 @@
parsedURI.pathname = "/" + action
connURI = parsedURI.toString()
console.log("connection URI: ", connURI)
mobileConnURIanchor.href = "simplex:" + parsedURI.pathname + parsedURI.hash
const hash = parsedURI.hash
const hostname = parsedURI.hostname
let appURI = "simplex:" + parsedURI.pathname
appURI += action.length > 1 // not short link
? hash
: !hash.includes("?") // otherwise add server hostname
? hash + "?h=" + hostname // no parameters
: !hash.includes("?h=") && !hash.includes("&h=")
? hash + "&h=" + hostname // no "h" parameter
: hash.replace(/([?&])h=([^&]+)/, `$1h=${hostname},$2`) // add as the first hostname to "h" parameter
mobileConnURIanchor.href = appURI
console.log("app URI: ", appURI)
connURIel.innerText = "/c " + connURI
for (const connQRCode of connQRCodes) {
try {

Before

Width:  |  Height:  |  Size: 289 KiB

After

Width:  |  Height:  |  Size: 289 KiB

Before

Width:  |  Height:  |  Size: 12 KiB

After

Width:  |  Height:  |  Size: 12 KiB

Before

Width:  |  Height:  |  Size: 1.1 KiB

After

Width:  |  Height:  |  Size: 1.1 KiB

Before

Width:  |  Height:  |  Size: 6.3 KiB

After

Width:  |  Height:  |  Size: 6.3 KiB

Before

Width:  |  Height:  |  Size: 7.2 KiB

After

Width:  |  Height:  |  Size: 7.2 KiB

Before

Width:  |  Height:  |  Size: 8.1 KiB

After

Width:  |  Height:  |  Size: 8.1 KiB

Before

Width:  |  Height:  |  Size: 1.1 KiB

After

Width:  |  Height:  |  Size: 1.1 KiB

Before

Width:  |  Height:  |  Size: 1.2 KiB

After

Width:  |  Height:  |  Size: 1.2 KiB

Before

Width:  |  Height:  |  Size: 632 B

After

Width:  |  Height:  |  Size: 632 B

Before

Width:  |  Height:  |  Size: 3.5 KiB

After

Width:  |  Height:  |  Size: 3.5 KiB

Before

Width:  |  Height:  |  Size: 18 KiB

After

Width:  |  Height:  |  Size: 18 KiB

-1
View File
@@ -15,7 +15,6 @@ logCfg = LogConfig {lc_file = Nothing, lc_stderr = True}
main :: IO ()
main = do
setLogLevel LogInfo
cfgPath <- getEnvPath "NTF_SERVER_CFG_PATH" defaultCfgPath
logPath <- getEnvPath "NTF_SERVER_LOG_PATH" defaultLogPath
withGlobalLogging logCfg $ ntfServerCLI cfgPath logPath
+4 -4
View File
@@ -2,8 +2,9 @@ module Main where
import Control.Logger.Simple
import Simplex.Messaging.Server.CLI (getEnvPath)
import Simplex.Messaging.Server.Main
import qualified Static
import Simplex.Messaging.Server.Main (smpServerCLI_)
import Simplex.Messaging.Server.Web (serveStaticFiles, attachStaticAndWS)
import SMPWeb (smpGenerateSite)
defaultCfgPath :: FilePath
defaultCfgPath = "/etc/opt/simplex"
@@ -16,7 +17,6 @@ logCfg = LogConfig {lc_file = Nothing, lc_stderr = True}
main :: IO ()
main = do
setLogLevel LogDebug
cfgPath <- getEnvPath "SMP_SERVER_CFG_PATH" defaultCfgPath
logPath <- getEnvPath "SMP_SERVER_LOG_PATH" defaultLogPath
withGlobalLogging logCfg $ smpServerCLI_ Static.generateSite Static.serveStaticFiles Static.attachStaticFiles cfgPath logPath
withGlobalLogging logCfg $ smpServerCLI_ smpGenerateSite serveStaticFiles attachStaticAndWS cfgPath logPath
+43
View File
@@ -0,0 +1,43 @@
{-# LANGUAGE NamedFieldPuns #-}
{-# LANGUAGE OverloadedStrings #-}
module SMPWeb
( smpGenerateSite,
serverInformation,
) where
import Data.ByteString (ByteString)
import Data.String (fromString)
import Web.Embedded (embeddedContent)
import Simplex.Messaging.Encoding.String (strEncode)
import Simplex.Messaging.Server.Information
import Simplex.Messaging.Server.Main (simplexmqSource)
import qualified Simplex.Messaging.Server.Web as Web
import Simplex.Messaging.Server.Web (render, serverInfoSubsts, timedTTLText)
import Simplex.Messaging.Transport.Client (TransportHost (..))
smpGenerateSite :: ServerInformation -> Maybe TransportHost -> FilePath -> IO ()
smpGenerateSite si onionHost path =
Web.generateSite embeddedContent (serverInformation si onionHost) smpLinkPages path
smpLinkPages :: [String]
smpLinkPages = ["contact", "invitation", "a", "c", "g", "r", "i"]
serverInformation :: ServerInformation -> Maybe TransportHost -> ByteString
serverInformation ServerInformation {config, information} onionHost = render (Web.indexHtml embeddedContent) substs
where
substs = [("smpConfig", Just "y"), ("xftpConfig", Nothing)] <> substConfig <> serverInfoSubsts simplexmqSource information <> [("onionHost", strEncode <$> onionHost), ("iniFileName", Just "smp-server.ini")]
substConfig =
[ ( "persistence",
Just $ case persistence config of
SPMMemoryOnly -> "In-memory only"
SPMQueues -> "Queues"
SPMMessages -> "Queues and messages"
),
("messageExpiration", Just $ maybe "Never" (fromString . timedTTLText) $ messageExpiration config),
("statsEnabled", Just . yesNo $ statsEnabled config),
("newQueuesAllowed", Just . yesNo $ newQueuesAllowed config),
("basicAuthEnabled", Just . yesNo $ basicAuthEnabled config)
]
yesNo True = "Yes"
yesNo False = "No"
@@ -1 +0,0 @@
../link.html
@@ -1 +0,0 @@
../link.html
-221
View File
@@ -1,221 +0,0 @@
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE NamedFieldPuns #-}
{-# LANGUAGE OverloadedStrings #-}
module Static where
import Control.Logger.Simple
import Control.Monad
import Data.ByteString (ByteString)
import qualified Data.ByteString.Char8 as B
import Data.Char (toUpper)
import Data.IORef (readIORef)
import Data.Maybe (fromMaybe)
import Data.String (fromString)
import Data.Text.Encoding (encodeUtf8)
import Network.Socket (getPeerName)
import Network.Wai (Application)
import qualified Network.Wai.Application.Static as S
import qualified Network.Wai.Handler.Warp as W
import qualified Network.Wai.Handler.Warp.Internal as WI
import qualified Network.Wai.Handler.WarpTLS as WT
import Simplex.Messaging.Encoding.String (strEncode)
import Simplex.Messaging.Server (AttachHTTP)
import Simplex.Messaging.Server.Information
import Simplex.Messaging.Server.Main (EmbeddedWebParams (..), WebHttpsParams (..))
import Simplex.Messaging.Transport (simplexMQVersion)
import Simplex.Messaging.Transport.Client (TransportHost (..))
import Simplex.Messaging.Util (tshow)
import Static.Embedded as E
import System.Directory (createDirectoryIfMissing)
import System.FilePath
import UnliftIO.Concurrent (forkFinally)
import UnliftIO.Exception (bracket, finally)
serveStaticFiles :: EmbeddedWebParams -> IO ()
serveStaticFiles EmbeddedWebParams {webStaticPath, webHttpPort, webHttpsParams} = do
forM_ webHttpPort $ \port -> flip forkFinally (\e -> logError $ "HTTP server crashed: " <> tshow e) $ do
logInfo $ "Serving static site on port " <> tshow port
W.runSettings (mkSettings port) (S.staticApp $ S.defaultFileServerSettings webStaticPath)
forM_ webHttpsParams $ \WebHttpsParams {port, cert, key} -> flip forkFinally (\e -> logError $ "HTTPS server crashed: " <> tshow e) $ do
logInfo $ "Serving static site on port " <> tshow port <> " (TLS)"
WT.runTLS (WT.tlsSettings cert key) (mkSettings port) app
where
app = staticFiles webStaticPath
mkSettings port = W.setPort port warpSettings
-- | Prepare context and prepare HTTP handler for TLS connections that already passed TLS.handshake and ALPN check.
attachStaticFiles :: FilePath -> (AttachHTTP -> IO ()) -> IO ()
attachStaticFiles path action =
-- Initialize global internal state for http server.
WI.withII warpSettings $ \ii -> do
action $ \socket cxt -> do
-- Initialize internal per-connection resources.
addr <- getPeerName socket
withConnection addr cxt $ \(conn, transport) ->
withTimeout ii conn $ \th ->
-- Run Warp connection handler to process HTTP requests for static files.
WI.serveConnection conn ii th addr transport warpSettings app
where
app = staticFiles path
-- from warp-tls
withConnection socket cxt = bracket (WT.attachConn socket cxt) (terminate . fst)
-- from warp
withTimeout ii conn =
bracket
(WI.registerKillThread (WI.timeoutManager ii) (WI.connClose conn))
WI.cancel
-- shared clean up
terminate conn = WI.connClose conn `finally` (readIORef (WI.connWriteBuffer conn) >>= WI.bufFree)
warpSettings :: W.Settings
warpSettings = W.setGracefulShutdownTimeout (Just 1) W.defaultSettings
staticFiles :: FilePath -> Application
staticFiles root = S.staticApp settings
where
settings = (S.defaultFileServerSettings root)
{ S.ssListing = Nothing
}
generateSite :: ServerInformation -> Maybe TransportHost -> FilePath -> IO ()
generateSite si onionHost sitePath = do
createDirectoryIfMissing True sitePath
B.writeFile (sitePath </> "index.html") $ serverInformation si onionHost
createDirectoryIfMissing True $ sitePath </> "media"
forM_ E.mediaContent $ \(path, bs) -> B.writeFile (sitePath </> "media" </> path) bs
createDirectoryIfMissing True $ sitePath </> "contact"
B.writeFile (sitePath </> "contact" </> "index.html") E.linkHtml
createDirectoryIfMissing True $ sitePath </> "invitation"
B.writeFile (sitePath </> "invitation" </> "index.html") E.linkHtml
logInfo $ "Generated static site contents at " <> tshow sitePath
serverInformation :: ServerInformation -> Maybe TransportHost -> ByteString
serverInformation ServerInformation {config, information} onionHost = render E.indexHtml substs
where
substs = substConfig <> maybe [] substInfo information <> [("onionHost", strEncode <$> onionHost)]
substConfig =
[ ( "persistence",
Just $ case persistence config of
SPMMemoryOnly -> "In-memory only"
SPMQueues -> "Queues"
SPMMessages -> "Queues and messages"
),
("messageExpiration", Just $ maybe "Never" (fromString . timedTTLText) $ messageExpiration config),
("statsEnabled", Just . yesNo $ statsEnabled config),
("newQueuesAllowed", Just . yesNo $ newQueuesAllowed config),
("basicAuthEnabled", Just . yesNo $ basicAuthEnabled config)
]
yesNo True = "Yes"
yesNo False = "No"
substInfo spi =
concat
[ basic,
maybe [("usageConditions", Nothing), ("usageAmendments", Nothing)] conds (usageConditions spi),
maybe [("operator", Nothing)] operatorE (operator spi),
maybe [("admin", Nothing)] admin (adminContacts spi),
maybe [("complaints", Nothing)] complaints (complaintsContacts spi),
maybe [("hosting", Nothing)] hostingE (hosting spi),
server
]
where
basic =
[ ("sourceCode", Just . encodeUtf8 $ sourceCode spi),
("version", Just $ B.pack simplexMQVersion),
("website", encodeUtf8 <$> website spi)
]
conds ServerConditions {conditions, amendments} =
[ ("usageConditions", Just $ encodeUtf8 conditions),
("usageAmendments", encodeUtf8 <$> amendments)
]
operatorE Entity {name, country} =
[ ("operator", Just ""),
("operatorEntity", Just $ encodeUtf8 name),
("operatorCountry", encodeUtf8 <$> country)
]
admin ServerContactAddress {simplex, email, pgp} =
[ ("admin", Just ""),
("adminSimplex", strEncode <$> simplex),
("adminEmail", encodeUtf8 <$> email),
("adminPGP", encodeUtf8 . pkURI <$> pgp),
("adminPGPFingerprint", encodeUtf8 . pkFingerprint <$> pgp)
]
complaints ServerContactAddress {simplex, email, pgp} =
[ ("complaints", Just ""),
("complaintsSimplex", strEncode <$> simplex),
("complaintsEmail", encodeUtf8 <$> email),
("complaintsPGP", encodeUtf8 . pkURI <$> pgp),
("complaintsPGPFingerprint", encodeUtf8 . pkFingerprint <$> pgp)
]
hostingE Entity {name, country} =
[ ("hosting", Just ""),
("hostingEntity", Just $ encodeUtf8 name),
("hostingCountry", encodeUtf8 <$> country)
]
server =
[ ("serverCountry", encodeUtf8 <$> serverCountry spi),
("hostingType", (\s -> maybe s (\(c, rest) -> toUpper c `B.cons` rest) $ B.uncons s) . strEncode <$> hostingType spi)
]
-- Copy-pasted from simplex-chat Simplex.Chat.Types.Preferences
{-# INLINE timedTTLText #-}
timedTTLText :: (Integral i, Show i) => i -> String
timedTTLText 0 = "0 sec"
timedTTLText ttl = do
let (m', s) = ttl `quotRem` 60
(h', m) = m' `quotRem` 60
(d', h) = h' `quotRem` 24
(mm, d) = d' `quotRem` 30
unwords $
[mms mm | mm /= 0]
<> [ds d | d /= 0]
<> [hs h | h /= 0]
<> [ms m | m /= 0]
<> [ss s | s /= 0]
where
ss s = show s <> " sec"
ms m = show m <> " min"
hs 1 = "1 hour"
hs h = show h <> " hours"
ds 1 = "1 day"
ds 7 = "1 week"
ds 14 = "2 weeks"
ds d = show d <> " days"
mms 1 = "1 month"
mms mm = show mm <> " months"
-- | Rewrite source with provided substitutions
render :: ByteString -> [(ByteString, Maybe ByteString)] -> ByteString
render src = \case
[] -> src
(label, content') : rest -> render (section_ label content' src) rest
-- | Rewrite section content inside @<x-label>...</x-label>@ markers.
-- Markers are always removed when found. Closing marker is mandatory.
-- If content is absent, whole section is removed.
-- Section content is delegated to `item_`. If no sections found, the whole source is delegated.
section_ :: ByteString -> Maybe ByteString -> ByteString -> ByteString
section_ label content' src =
case B.breakSubstring startMarker src of
(_, "") -> item_ label (fromMaybe "" content') src -- no section, just replace items
(before, afterStart') ->
-- found section start, search for end too
case B.breakSubstring endMarker $ B.drop (B.length startMarker) afterStart' of
(_, "") -> error $ "missing section end: " <> show endMarker
(inside, next') ->
let next = B.drop (B.length endMarker) next'
in case content' of
Nothing -> before <> next -- collapse section
Just content -> before <> item_ label content inside <> section_ label content' next
where
startMarker = "<x-" <> label <> ">"
endMarker = "</x-" <> label <> ">"
-- | Replace all occurences of @${label}@ with provided content.
item_ :: ByteString -> ByteString -> ByteString -> ByteString
item_ label content' src =
case B.breakSubstring marker src of
(done, "") -> done
(before, after') -> before <> content' <> item_ label content' (B.drop (B.length marker) after')
where
marker = "${" <> label <> "}"
-15
View File
@@ -1,15 +0,0 @@
{-# LANGUAGE TemplateHaskell #-}
module Static.Embedded where
import Data.FileEmbed (embedDir, embedFile)
import Data.ByteString (ByteString)
indexHtml :: ByteString
indexHtml = $(embedFile "apps/smp-server/static/index.html")
linkHtml :: ByteString
linkHtml = $(embedFile "apps/smp-server/static/link.html")
mediaContent :: [(FilePath, ByteString)]
mediaContent = $(embedDir "apps/smp-server/static/media/")
+4 -2
View File
@@ -1,8 +1,10 @@
module Main where
import Control.Logger.Simple
import Simplex.FileTransfer.Server.Main (xftpServerCLI_)
import Simplex.Messaging.Server.CLI (getEnvPath)
import Simplex.FileTransfer.Server.Main
import Simplex.Messaging.Server.Web (serveStaticFiles)
import XFTPWeb (xftpGenerateSite)
defaultCfgPath :: FilePath
defaultCfgPath = "/etc/opt/simplex-xftp"
@@ -18,4 +20,4 @@ main = do
setLogLevel LogDebug -- change to LogError in production
cfgPath <- getEnvPath "XFTP_SERVER_CFG_PATH" defaultCfgPath
logPath <- getEnvPath "XFTP_SERVER_LOG_PATH" defaultLogPath
withGlobalLogging logCfg $ xftpServerCLI cfgPath logPath
withGlobalLogging logCfg $ xftpServerCLI_ xftpGenerateSite serveStaticFiles cfgPath logPath
+67
View File
@@ -0,0 +1,67 @@
{-# LANGUAGE NamedFieldPuns #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE TemplateHaskell #-}
module XFTPWeb
( xftpGenerateSite,
xftpServerInformation,
) where
import Control.Monad (forM_)
import qualified Data.ByteString.Char8 as B
import Data.ByteString (ByteString)
import Data.FileEmbed (embedDir, embedFile)
import Data.Maybe (isJust)
import Data.String (fromString)
import Web.Embedded (embeddedContent)
import Simplex.FileTransfer.Server.Env (XFTPServerConfig (..))
import Simplex.Messaging.Encoding.String (strEncode)
import Simplex.Messaging.Server.Expiration (ExpirationConfig (..))
import Simplex.Messaging.Server.Information (ServerPublicInfo)
import Simplex.Messaging.Server.Main (simplexmqSource)
import qualified Simplex.Messaging.Server.Web as Web
import Simplex.Messaging.Server.Web (render, serverInfoSubsts, timedTTLText)
import Simplex.Messaging.Transport.Client (TransportHost (..))
import System.Directory (createDirectoryIfMissing)
import System.FilePath ((</>))
xftpWebContent :: [(FilePath, ByteString)]
xftpWebContent = $(embedDir "apps/xftp-server/static/xftp-web-bundle/")
xftpMediaContent :: [(FilePath, ByteString)]
xftpMediaContent = $(embedDir "apps/xftp-server/static/media/")
xftpFilePageHtml :: ByteString
xftpFilePageHtml = $(embedFile "apps/xftp-server/static/file.html")
xftpGenerateSite :: XFTPServerConfig s -> Maybe ServerPublicInfo -> Maybe TransportHost -> FilePath -> IO ()
xftpGenerateSite cfg info onionHost path = do
let substs = xftpSubsts cfg info onionHost
Web.generateSite embeddedContent (render (Web.indexHtml embeddedContent) substs) [] path
let xftpDir = path </> "xftp-web-bundle"
mediaDir = path </> "media"
fileDir = path </> "file"
filePage xftpDir xftpWebContent
filePage mediaDir xftpMediaContent
createDirectoryIfMissing True fileDir
B.writeFile (fileDir </> "index.html") $ render xftpFilePageHtml substs
where
filePage dir content_ = do
createDirectoryIfMissing True dir
forM_ content_ $ \(fp, content) -> B.writeFile (dir </> fp) content
xftpServerInformation :: XFTPServerConfig s -> Maybe ServerPublicInfo -> Maybe TransportHost -> ByteString
xftpServerInformation cfg info onionHost = render (Web.indexHtml embeddedContent) (xftpSubsts cfg info onionHost)
xftpSubsts :: XFTPServerConfig s -> Maybe ServerPublicInfo -> Maybe TransportHost -> [(ByteString, Maybe ByteString)]
xftpSubsts XFTPServerConfig {fileExpiration, logStatsInterval, allowNewFiles, newFileBasicAuth} information onionHost =
[("smpConfig", Nothing), ("xftpConfig", Just "y")] <> substConfig <> serverInfoSubsts simplexmqSource information <> [("onionHost", strEncode <$> onionHost), ("iniFileName", Just "file-server.ini")]
where
substConfig =
[ ("fileExpiration", Just $ maybe "Never" (fromString . timedTTLText . ttl) fileExpiration),
("statsEnabled", Just . yesNo $ isJust logStatsInterval),
("newUploadsAllowed", Just . yesNo $ allowNewFiles),
("basicAuthEnabled", Just . yesNo $ isJust newFileBasicAuth)
]
yesNo True = "Yes"
yesNo False = "No"
File diff suppressed because one or more lines are too long
@@ -0,0 +1,115 @@
<svg width="440" height="520" viewBox="-20 0 440 520" fill="none" xmlns="http://www.w3.org/2000/svg">
<!-- Sender browser -->
<rect x="120" y="16" width="160" height="56" rx="10" stroke="#70F0F9" stroke-width="1.5"/>
<text x="200" y="40" text-anchor="middle" font-family="system-ui, sans-serif" font-size="13" font-weight="600" fill="#70F0F9">Sender's browser</text>
<text x="200" y="56" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" fill="rgba(112,240,249,0.7)">encrypts file</text>
<!-- Arrow down from sender to chunks -->
<line x1="200" y1="72" x2="200" y2="120" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<!-- Chunks row -->
<rect x="112" y="120" width="176" height="40" rx="8" fill="none" stroke="#70F0F9" stroke-width="1" stroke-dasharray="4 3"/>
<text x="200" y="145" text-anchor="middle" font-family="system-ui, sans-serif" font-size="12" fill="#70F0F9">encrypted chunks</text>
<!-- Arrows from chunks to routers -->
<line x1="152" y1="160" x2="80" y2="220" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<line x1="200" y1="160" x2="200" y2="220" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<line x1="248" y1="160" x2="320" y2="220" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<!-- Router 1 (SimpleX) -->
<rect x="20" y="220" width="120" height="56" rx="6" fill="none" stroke="#70F0F9" stroke-width="1.5"/>
<g transform="translate(28, 227)">
<rect width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="6" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="12" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<circle cx="11" cy="2" r="1" fill="#70F0F9"/>
<circle cx="11" cy="8" r="1" fill="#70F0F9"/>
<circle cx="11" cy="14" r="1" fill="#70F0F9"/>
</g>
<text x="80" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#70F0F9">SimpleX</text>
<text x="80" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="rgba(112,240,249,0.7)">XFTP router</text>
<!-- Router 2 (Flux) -->
<rect x="155" y="220" width="90" height="56" rx="6" fill="none" stroke="#70F0F9" stroke-width="1.5"/>
<g transform="translate(163, 227)">
<rect width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="6" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="12" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<circle cx="11" cy="2" r="1" fill="#70F0F9"/>
<circle cx="11" cy="8" r="1" fill="#70F0F9"/>
<circle cx="11" cy="14" r="1" fill="#70F0F9"/>
</g>
<text x="200" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#70F0F9">Flux</text>
<text x="200" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="rgba(112,240,249,0.7)">XFTP router</text>
<!-- Router 3 (SimpleX) -->
<rect x="260" y="220" width="120" height="56" rx="6" fill="none" stroke="#70F0F9" stroke-width="1.5"/>
<g transform="translate(268, 227)">
<rect width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="6" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<rect y="12" width="14" height="4" rx="1" fill="rgba(112,240,249,0.5)"/>
<circle cx="11" cy="2" r="1" fill="#70F0F9"/>
<circle cx="11" cy="8" r="1" fill="#70F0F9"/>
<circle cx="11" cy="14" r="1" fill="#70F0F9"/>
</g>
<text x="320" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#70F0F9">SimpleX</text>
<text x="320" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="rgba(112,240,249,0.7)">XFTP router</text>
<!-- Arrows from routers down -->
<line x1="80" y1="276" x2="152" y2="336" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<line x1="200" y1="276" x2="200" y2="336" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<line x1="320" y1="276" x2="248" y2="336" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<!-- Re-encrypt label -->
<text x="330" y="310" text-anchor="start" font-family="system-ui, sans-serif" font-size="10" fill="rgba(112,240,249,0.7)">re-encrypted</text>
<text x="330" y="322" text-anchor="start" font-family="system-ui, sans-serif" font-size="10" fill="rgba(112,240,249,0.7)">per recipient</text>
<!-- Chunks row (download) -->
<rect x="112" y="336" width="176" height="40" rx="8" fill="none" stroke="#70F0F9" stroke-width="1" stroke-dasharray="4 3"/>
<text x="200" y="361" text-anchor="middle" font-family="system-ui, sans-serif" font-size="12" fill="#70F0F9">encrypted chunks</text>
<!-- Arrow down to recipient -->
<line x1="200" y1="376" x2="200" y2="424" stroke="#70F0F9" stroke-width="1.5" marker-end="url(#arrowC)"/>
<!-- Recipient browser -->
<rect x="120" y="424" width="160" height="56" rx="10" stroke="#70F0F9" stroke-width="1.5"/>
<text x="200" y="448" text-anchor="middle" font-family="system-ui, sans-serif" font-size="13" font-weight="600" fill="#70F0F9">Recipient's browser</text>
<text x="200" y="464" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" fill="rgba(112,240,249,0.7)">decrypts file</text>
<!-- Key path (dashed, side) -->
<path d="M120 44 L8 44 L8 452 L120 452" stroke="#70F0F9" stroke-width="1.5" stroke-dasharray="6 4" fill="none" marker-end="url(#arrowC)"/>
<text x="-6" y="240" text-anchor="middle" font-family="system-ui, sans-serif" font-size="10" fill="#70F0F9" transform="rotate(-90 -6 240)">key in URL fragment - never sent to page server or data router</text>
<!-- Closed padlock: encryption (between sender and chunks) -->
<g transform="translate(192, 88)">
<path d="M4,7 V4 C4,1.2 12,1.2 12,4 V7" stroke="#60a5fa" stroke-width="1.5" fill="none" stroke-linecap="round"/>
<rect x="2" y="7" width="12" height="9" rx="2" fill="#60a5fa"/>
<circle cx="8" cy="12" r="1.2" fill="#0B2A59"/>
</g>
<!-- Open padlock: decryption (between chunks and recipient) -->
<g transform="translate(192, 392)">
<path d="M4,7 V4 C4,1.2 12,1.2 12,4 V2" stroke="#60a5fa" stroke-width="1.5" fill="none" stroke-linecap="round"/>
<rect x="2" y="7" width="12" height="9" rx="2" fill="#60a5fa"/>
<circle cx="8" cy="12" r="1.2" fill="#0B2A59"/>
</g>
<!-- Key icon on dashed line -->
<g transform="translate(8, 410)">
<circle cx="0" cy="0" r="6" stroke="#FBBF24" stroke-width="2" fill="#FBBF24"/>
<circle cx="0" cy="0" r="2" fill="#0B2A59"/>
<line x1="6" y1="0" x2="16" y2="0" stroke="#FBBF24" stroke-width="2"/>
<line x1="14" y1="0" x2="14" y2="4" stroke="#FBBF24" stroke-width="2"/>
<line x1="11" y1="0" x2="11" y2="3.5" stroke="#FBBF24" stroke-width="2"/>
</g>
<!-- Annotation: no shared IDs -->
<text x="200" y="510" text-anchor="middle" font-family="system-ui, sans-serif" font-size="10" fill="rgba(112,240,249,0.7)">Each file fragment uses unique anonymous credentials - no shared identifiers</text>
<defs>
<marker id="arrowC" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse">
<path d="M 0 0 L 10 5 L 0 10 z" fill="#70F0F9"/>
</marker>
</defs>
</svg>

After

Width:  |  Height:  |  Size: 7.2 KiB

@@ -0,0 +1,130 @@
<svg width="440" height="520" viewBox="-20 0 440 520" fill="none" xmlns="http://www.w3.org/2000/svg">
<!-- Sender browser -->
<rect x="120" y="16" width="160" height="56" rx="10" fill="url(#gBox)" stroke="#606C71" stroke-width="1.5"/>
<text x="200" y="40" text-anchor="middle" font-family="system-ui, sans-serif" font-size="13" font-weight="600" fill="#fff">Sender's browser</text>
<text x="200" y="56" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" fill="rgba(255,255,255,0.8)">encrypts file</text>
<!-- Arrow down from sender to chunks -->
<line x1="200" y1="72" x2="200" y2="120" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<!-- Chunks row -->
<rect x="112" y="120" width="176" height="40" rx="8" fill="#f0f7ff" stroke="#0053D0" stroke-width="1" stroke-dasharray="4 3"/>
<text x="200" y="145" text-anchor="middle" font-family="system-ui, sans-serif" font-size="12" fill="#0053D0">encrypted chunks</text>
<!-- Arrows from chunks to routers -->
<line x1="152" y1="160" x2="80" y2="220" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<line x1="200" y1="160" x2="200" y2="220" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<line x1="248" y1="160" x2="320" y2="220" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<!-- Router 1 (SimpleX) -->
<rect x="20" y="220" width="120" height="56" rx="6" fill="#f0f4f8" stroke="#606C71" stroke-width="1.5"/>
<g transform="translate(28, 227)">
<rect width="14" height="4" rx="1" fill="#606C71"/>
<rect y="6" width="14" height="4" rx="1" fill="#606C71"/>
<rect y="12" width="14" height="4" rx="1" fill="#606C71"/>
<circle cx="11" cy="2" r="1" fill="#53C1FF"/>
<circle cx="11" cy="8" r="1" fill="#53C1FF"/>
<circle cx="11" cy="14" r="1" fill="#53C1FF"/>
</g>
<text x="80" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#3F484B">SimpleX</text>
<text x="80" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="#606C71">XFTP router</text>
<!-- Router 2 (Flux) -->
<rect x="155" y="220" width="90" height="56" rx="6" fill="#f0f4f8" stroke="#606C71" stroke-width="1.5"/>
<g transform="translate(163, 227)">
<rect width="14" height="4" rx="1" fill="#606C71"/>
<rect y="6" width="14" height="4" rx="1" fill="#606C71"/>
<rect y="12" width="14" height="4" rx="1" fill="#606C71"/>
<circle cx="11" cy="2" r="1" fill="#53C1FF"/>
<circle cx="11" cy="8" r="1" fill="#53C1FF"/>
<circle cx="11" cy="14" r="1" fill="#53C1FF"/>
</g>
<text x="200" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#3F484B">Flux</text>
<text x="200" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="#606C71">XFTP router</text>
<!-- Router 3 (SimpleX) -->
<rect x="260" y="220" width="120" height="56" rx="6" fill="#f0f4f8" stroke="#606C71" stroke-width="1.5"/>
<g transform="translate(268, 227)">
<rect width="14" height="4" rx="1" fill="#606C71"/>
<rect y="6" width="14" height="4" rx="1" fill="#606C71"/>
<rect y="12" width="14" height="4" rx="1" fill="#606C71"/>
<circle cx="11" cy="2" r="1" fill="#53C1FF"/>
<circle cx="11" cy="8" r="1" fill="#53C1FF"/>
<circle cx="11" cy="14" r="1" fill="#53C1FF"/>
</g>
<text x="320" y="244" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" font-weight="600" fill="#3F484B">SimpleX</text>
<text x="320" y="258" text-anchor="middle" font-family="system-ui, sans-serif" font-size="9" fill="#606C71">XFTP router</text>
<!-- Arrows from routers down -->
<line x1="80" y1="276" x2="152" y2="336" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<line x1="200" y1="276" x2="200" y2="336" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<line x1="320" y1="276" x2="248" y2="336" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<!-- Re-encrypt label -->
<text x="330" y="310" text-anchor="start" font-family="system-ui, sans-serif" font-size="10" fill="#606C71">re-encrypted</text>
<text x="330" y="322" text-anchor="start" font-family="system-ui, sans-serif" font-size="10" fill="#606C71">per recipient</text>
<!-- Chunks row (download) -->
<rect x="112" y="336" width="176" height="40" rx="8" fill="#f0f7ff" stroke="#0053D0" stroke-width="1" stroke-dasharray="4 3"/>
<text x="200" y="361" text-anchor="middle" font-family="system-ui, sans-serif" font-size="12" fill="#0053D0">encrypted chunks</text>
<!-- Arrow down to recipient -->
<line x1="200" y1="376" x2="200" y2="424" stroke="#606C71" stroke-width="1.5" marker-end="url(#arrowG)"/>
<!-- Recipient browser -->
<rect x="120" y="424" width="160" height="56" rx="10" fill="url(#gBox)" stroke="#606C71" stroke-width="1.5"/>
<text x="200" y="448" text-anchor="middle" font-family="system-ui, sans-serif" font-size="13" font-weight="600" fill="#fff">Recipient's browser</text>
<text x="200" y="464" text-anchor="middle" font-family="system-ui, sans-serif" font-size="11" fill="rgba(255,255,255,0.8)">decrypts file</text>
<!-- Key path (dashed, side) -->
<path d="M120 44 L8 44 L8 452 L120 452" stroke="#0053D0" stroke-width="1.5" stroke-dasharray="6 4" fill="none" marker-end="url(#arrowB)"/>
<text x="-6" y="240" text-anchor="middle" font-family="system-ui, sans-serif" font-size="10" fill="#0053D0" transform="rotate(-90 -6 240)">key in URL fragment - never sent to page server or data router</text>
<!-- Closed padlock: encryption (between sender and chunks) -->
<g transform="translate(192, 88)">
<path d="M4,7 V4 C4,1.2 12,1.2 12,4 V7" stroke="#0053D0" stroke-width="1.5" fill="none" stroke-linecap="round"/>
<rect x="2" y="7" width="12" height="9" rx="2" fill="#0053D0"/>
<circle cx="8" cy="12" r="1.2" fill="#fff"/>
</g>
<!-- Open padlock: decryption (between chunks and recipient) -->
<g transform="translate(192, 392)">
<path d="M4,7 V4 C4,1.2 12,1.2 12,4 V2" stroke="#0053D0" stroke-width="1.5" fill="none" stroke-linecap="round"/>
<rect x="2" y="7" width="12" height="9" rx="2" fill="#0053D0"/>
<circle cx="8" cy="12" r="1.2" fill="#fff"/>
</g>
<!-- Key icon on dashed line -->
<g transform="translate(8, 410)">
<circle cx="0" cy="0" r="6" stroke="#D97706" stroke-width="2" fill="#D97706"/>
<circle cx="0" cy="0" r="2" fill="#fff"/>
<line x1="6" y1="0" x2="16" y2="0" stroke="#D97706" stroke-width="2"/>
<line x1="14" y1="0" x2="14" y2="4" stroke="#D97706" stroke-width="2"/>
<line x1="11" y1="0" x2="11" y2="3.5" stroke="#D97706" stroke-width="2"/>
</g>
<!-- Annotation: no shared IDs -->
<text x="200" y="510" text-anchor="middle" font-family="system-ui, sans-serif" font-size="10" fill="#606C71">Each file fragment uses unique anonymous credentials - no shared identifiers</text>
<defs>
<linearGradient id="gBox" x1="120" y1="16" x2="280" y2="72" gradientUnits="userSpaceOnUse">
<stop stop-color="#0053D0"/>
<stop offset="1" stop-color="#53C1FF"/>
</linearGradient>
<linearGradient id="gSrv1" x1="20" y1="220" x2="140" y2="276" gradientUnits="userSpaceOnUse">
<stop stop-color="#0053D0"/>
<stop offset="1" stop-color="#53C1FF"/>
</linearGradient>
<linearGradient id="gSrv2" x1="155" y1="220" x2="245" y2="276" gradientUnits="userSpaceOnUse">
<stop stop-color="#0053D0"/>
<stop offset="1" stop-color="#53C1FF"/>
</linearGradient>
<marker id="arrowG" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse">
<path d="M 0 0 L 10 5 L 0 10 z" fill="#606C71"/>
</marker>
<marker id="arrowB" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse">
<path d="M 0 0 L 10 5 L 0 10 z" fill="#0053D0"/>
</marker>
</defs>
</svg>

After

Width:  |  Height:  |  Size: 7.8 KiB

File diff suppressed because one or more lines are too long
@@ -0,0 +1,145 @@
#app, [data-xftp-app] {
font-family: system-ui, -apple-system, sans-serif;
color: #333;
width: 100%;
max-width: 480px;
padding: 16px;
box-sizing: border-box;
--xftp-ring-fg: #3b82f6;
}
:is(#app, [data-xftp-app]) .card {
background: #fff;
border-radius: 12px;
padding: 32px 24px;
box-shadow: 0 1px 3px rgba(0,0,0,.1);
text-align: center;
}
:is(#app, [data-xftp-app]) h1 {
font-size: 1.25rem;
font-weight: 600;
margin-bottom: 24px;
}
:is(#app, [data-xftp-app]) .stage { margin-top: 16px; }
/* Drop zone */
:is(#app, [data-xftp-app]) .drop-zone {
border: 2px dashed #ccc;
border-radius: 8px;
padding: 32px 16px;
transition: border-color .15s, background .15s;
}
:is(#app, [data-xftp-app]) .drop-zone.drag-over {
border-color: #3b82f6;
background: #eff6ff;
}
/* Buttons */
:is(#app, [data-xftp-app]) .btn {
display: inline-block;
padding: 10px 24px;
border: none;
border-radius: 6px;
background: #3b82f6;
color: #fff;
font-size: .9rem;
font-weight: 500;
cursor: pointer;
transition: background .15s;
}
:is(#app, [data-xftp-app]) .btn:hover { background: #2563eb; }
:is(#app, [data-xftp-app]) .btn-secondary { background: #6b7280; }
:is(#app, [data-xftp-app]) .btn-secondary:hover { background: #4b5563; }
/* Hints */
:is(#app, [data-xftp-app]) .hint { color: #999; font-size: .85rem; margin-top: 8px; }
:is(#app, [data-xftp-app]) .expiry { margin-top: 12px; }
/* Progress */
:is(#app, [data-xftp-app]) .progress-ring { display: block; margin: 0 auto 12px; }
:is(#app, [data-xftp-app]) #upload-status,
:is(#app, [data-xftp-app]) #dl-status { font-size: .9rem; color: #666; margin-bottom: 12px; }
/* Share link row */
:is(#app, [data-xftp-app]) .link-row {
display: flex;
gap: 8px;
margin-top: 12px;
}
:is(#app, [data-xftp-app]) .link-row input {
flex: 1;
padding: 8px 10px;
border: 1px solid #ccc;
border-radius: 6px;
font-size: .85rem;
background: #f9fafb;
}
/* Upload link */
:is(#app, [data-xftp-app]) .upload-link {
margin-top: 12px;
color: #3b82f6;
font-size: .9rem;
text-decoration: none;
cursor: pointer;
}
:is(#app, [data-xftp-app]) .upload-link:not([hidden]) {
display: inline-block;
}
:is(#app, [data-xftp-app]) .upload-link:hover { text-decoration: underline; }
/* Messages */
:is(#app, [data-xftp-app]) .success { color: #16a34a; font-weight: 600; }
:is(#app, [data-xftp-app]) .error { color: #dc2626; font-weight: 500; margin-bottom: 12px; }
/* Security note */
:is(#app, [data-xftp-app]) .security-note {
margin-top: 20px;
padding: 12px;
background: #f0fdf4;
border-radius: 6px;
font-size: .8rem;
color: #555;
text-align: left;
}
:is(#app, [data-xftp-app]) .security-note p + p { margin-top: 6px; }
:is(#app, [data-xftp-app]) .security-note a { color: #3b82f6; text-decoration: none; }
:is(#app, [data-xftp-app]) .security-note a:hover { text-decoration: underline; }
/* ── Dark mode ─────────────────────────────────── */
.dark :is(#app, [data-xftp-app]) {
color: #e5e7eb;
--xftp-ring-bg: #374151;
--xftp-ring-fg: #60a5fa;
--xftp-ring-text: #e5e7eb;
--xftp-ring-done: #4ade80;
}
.dark :is(#app, [data-xftp-app]) .card {
background: #1f2937;
box-shadow: 0 1px 3px rgba(0,0,0,.4);
}
.dark :is(#app, [data-xftp-app]) .drop-zone { border-color: #4b5563; }
.dark :is(#app, [data-xftp-app]) .drop-zone.drag-over {
border-color: #60a5fa;
background: rgba(59,130,246,.15);
}
.dark :is(#app, [data-xftp-app]) .btn-secondary { background: #4b5563; }
.dark :is(#app, [data-xftp-app]) .btn-secondary:hover { background: #374151; }
.dark :is(#app, [data-xftp-app]) .hint { color: #9ca3af; }
.dark :is(#app, [data-xftp-app]) #upload-status,
.dark :is(#app, [data-xftp-app]) #dl-status { color: #9ca3af; }
.dark :is(#app, [data-xftp-app]) .link-row input {
background: #374151;
border-color: #4b5563;
color: #e5e7eb;
}
.dark :is(#app, [data-xftp-app]) .success { color: #4ade80; }
.dark :is(#app, [data-xftp-app]) .error { color: #f87171; }
.dark :is(#app, [data-xftp-app]) .security-note {
background: rgba(34,197,94,.1);
color: #d1d5db;
}
.dark :is(#app, [data-xftp-app]) .upload-link { color: #60a5fa; }
.dark :is(#app, [data-xftp-app]) .security-note a { color: #60a5fa; }
File diff suppressed because one or more lines are too long
+1
View File
@@ -4,6 +4,7 @@ packages: .
-- packages: . ../http2
-- packages: . ../network-transport
-- uncomment two sections below to run tests with coverage
-- package *
-- coverage: True
-- library-coverage: True
+104
View File
@@ -0,0 +1,104 @@
# Coding and building
This file provides guidance on coding style and approaches and on building the code.
## Code Security
When designing code and planning implementations:
- Apply adversarial thinking, and consider what may happen if one of the communicating parties is malicious.
- Formulate an explicit threat model for each change - who can do which undesirable things and under which circumstances.
## Code Quality Standards
Haskell client and server code serves as system specification, not just implementation — we use type-driven design to reflect the business domain in types. Quality, conciseness, and clarity of Haskell code are critical.
## Code Style, Formatting and Approaches
The project uses **fourmolu** for Haskell code formatting. Configuration is in `fourmolu.yaml`.
**Key formatting rules:**
- 2-space indentation
- Trailing function arrows, commas, and import/export style
- Record brace without space: `{field = value}`
- Single newline between declarations
- Never use unicode symbols
- Inline `let` style with right-aligned `in`
**Format code before committing:**
```bash
# Format a single file
fourmolu -i src/Simplex/Messaging/Protocol.hs
```
Some files that use CPP language extension cannot be formatted as a whole, so individual code fragments need to be formatted.
**Follow existing code patterns:**
- Match the style of surrounding code
- Use qualified imports with short aliases (e.g., `import qualified Data.ByteString.Char8 as B`)
- Use record syntax for types with multiple fields
- Prefer explicit pattern matching over partial functions
**Comments policy:**
- Avoid redundant comments that restate what the code already says
- Only comment on non-obvious design decisions or tricky implementation details
- Function names and type signatures should be self-documenting
- Do not add comments like "wire format encoding" (Encoding class is always wire format) or "check if X" when the function name already says that
- Assume a competent Haskell reader
**Diff and refactoring:**
- Avoid unnecessary changes and code movements
- Never do refactoring unless it substantially reduces cost of solving the current problem, including the cost of refactoring
- Aim to minimize the code changes - do what is minimally required to solve users' problems
**Document and code structure:**
- **Never move existing code or sections around** - add new content at appropriate locations without reorganizing existing structure.
- When adding new sections to documents, continue the existing numbering scheme.
- Minimize diff size - prefer small, targeted changes over reorganization.
**Code analysis and review:**
- Trace data flows end-to-end: from origin, through storage/parameters, to consumption. Flag values that are discarded and reconstructed from partial data (e.g. extracted from a URI missing original fields) — this is usually a bug.
- Read implementations of called functions, not just signatures — if duplication involves a called function, check whether decomposing it resolves the duplication.
- Do not save time on analysis. Read every function in the data flow even when the interface seems clear — wrong assumptions about internals are the main source of missed bugs.
### Haskell Extensions
- `StrictData` enabled by default
- Use STM for safe concurrency
- Assume concurrency in PostgreSQL queries
- Comprehensive warning flags with strict pattern matching
## Build Commands
```bash
# Standard build
cabal build
# Fast build
cabal build --ghc-options -O0
# Build specific executables
cabal build exe:smp-server exe:xftp-server exe:ntf-server exe:xftp
# Build with PostgreSQL server support
cabal build -fserver_postgres
# Client-only library build (no server code)
cabal build -fclient_library
# Find binary location
cabal list-bin exe:smp-server
```
### Cabal Flags
- `swift`: Enable Swift JSON format
- `client_library`: Build without server code
- `client_postgres`: Use PostgreSQL instead of SQLite for agent persistence
- `server_postgres`: PostgreSQL support for server queue/notification store
## External Dependencies
Custom forks specified in `cabal.project`:
- `aeson`, `hs-socks` (SimpleX forks)
- `direct-sqlcipher`, `sqlcipher-simple` (encrypted SQLite)
- `warp`, `warp-tls` (HTTP server)
+105
View File
@@ -0,0 +1,105 @@
# SimpleXMQ repository
This file provides guidance on the project structure to help working with code in this repository.
## Project Overview
SimpleXMQ is a Haskell message broker implementing unidirectional (simplex) queues for privacy-preserving messaging.
Key components:
- **SimpleX Messaging Protocol**: SMP protocol definition and encodings ([code](../src/Simplex/Messaging/Protocol.hs), [transport code](../src/Simplex/Messaging/Transport.hs), [spec](../protocol/simplex-messaging.md)).
- **SMP Server**: Message broker with TLS, in-memory queues, optional persistence ([main code](../src/Simplex/Messaging/Server.hs), [all code files](../src/Simplex/Messaging/Server/), [executable](../apps/smp-server/)). For proxying SMP commands the server uses [lightweight SMP client](../src/Simplex/Messaging/Client/Agent.hs).
- **SMP Client**: Functional API with STM-based message delivery ([code](../src/Simplex/Messaging/Client.hs)).
- **SMP Agent**: High-level duplex connections via multiple simplex queues with E2E encryption ([code](../src/Simplex/Messaging/Agent.hs)). Implements Agent-to-agent protocol ([code](../src/Simplex/Messaging/Agent/Protocol.hs), [spec](../protocol/agent-protocol.md)) via intermediary agent client ([code](../src/Simplex/Messaging/Agent/Client.hs)).
- **XFTP**: SimpleX File Transfer Protocol, server and CLI client ([code](../src/Simplex/FileTransfer/), [spec](../protocol/xftp.md)).
- **XRCP**: SimpleX Remote Control Protocol ([code](`../src/Simplex/RemoteControl/`), [spec](../protocol/xrcp.md)).
- **Notifications**: Push notifications server requires PostgreSQL ([code](../src/Simplex/Messaging/Notifications), [executable](../apps/ntf-server/)). Client protocol is used for clients to communicate with the server ([code](../src/Simplex/Messaging/Notifications/Protocol.hs), [spec](../protocol/push-notifications.md)). For subscribing to SMP notifications the server uses [lightweight SMP client](../src/Simplex/Messaging/Client/Agent.hs).
## Architecture
For general overview see `../protocol/overview-tjr.md`.
SMP Protocol Layers:
```
TLS Transport → SMP Protocol → Agent Protocol → Application protocol
```
XFTP Protocol Layers:
```
TLS Transport (HTTP2 encoding) → XFTP Protocol → Out-of-band file descriptions
```
## Key Patterns
1. **Persistence**: All queue state managed via Software Transactional Memory or via PostgreSQL
- `Simplex.Messaging.Server.MsgStore.STM` - in-memory messages
- `Simplex.Messaging.Server.QueueStore.STM` - in-memory queue state
- `Simplex.Messaging.Server.MsgStore.Postgres` - message storage
- `Simplex.Messaging.Server.QueueStore.Postgres` - queue storage
2. **Append-Only Store Log**: Optional persistence via journal for in-memory storage
- `Simplex.Messaging.Server.StoreLog` - queue creation log
- Compacted on restart
3. **Agent Storage**:
- SQLite (default) or PostgreSQL
- Migrations in `src/Simplex/Messaging/Agent/Store/{SQLite,Postgres}/Migrations/`
4. **Protocol Versioning**: All layers support version negotiation
- `Simplex.Messaging.Version` - version range utilities
5. **Double Ratchet E2E**: Per-connection encryption
- `Simplex.Messaging.Crypto.Ratchet`
- SNTRUP761 post-quantum KEM (`src/Simplex/Messaging/Crypto/SNTRUP761/`)
## Source Layout
```
src/Simplex/
├── Messaging/
│ ├── Agent.hs # Main agent (~210KB)
│ ├── Server.hs # SMP server (~130KB)
│ ├── Client.hs # Client API (~65KB)
│ ├── Protocol.hs # Protocol types (~77KB)
│ ├── Crypto.hs # E2E encryption (~52KB)
│ ├── Transport.hs # Transport encoding over TLS
│ ├── Agent/Store/ # SQLite/Postgres persistence
│ ├── Server/ # Server internals (QueueStore, MsgStore, Control)
│ └── Notifications/ # Push notification system
├── FileTransfer/ # XFTP implementation for file transfers
└── RemoteControl/ # XRCP implementation for device discovery & control
```
## Protocol Documentation
- `protocol/overview-tjr.md`: SMP protocols stack overview
- `protocol/simplex-messaging.md`: SMP protocol spec (v19)
- `protocol/agent-protocol.md`: Agent protocol spec (v7)
- `protocol/xftp.md`: File transfer protocol
- `protocol/xrcp.md`: Remote control protocol
- `rfcs/`: Design RFCs for features
## Testing
```bash
# Run all tests
cabal test --test-show-details=streaming
# Run specific test group (uses HSpec)
cabal test --test-option=--match="/Core tests/Encryption tests/"
# Run single test
cabal test --test-option=--match="/SMP client agent/functional API/"
```
Tests require PostgreSQL running on `localhost:5432` when using `-fserver_postgres` or `-fclient_postgres`.
Test files are in `tests/` with structure:
- `Test.hs`: Main runner
- `AgentTests/`: Agent protocol and connection tests
- `CoreTests/`: Crypto, encoding, storage tests
- `ServerTests.hs`: SMP server tests
- `XFTPServerTests.hs`: File transfer tests
+23
View File
@@ -0,0 +1,23 @@
# Contributing to SimpleX repositories
## Focus on user problems
We do not make code changes to improve code - any change must address a specific user problem or request.
## Discuss the plans as early as possible
Please discuss the problem you want to solve and your detailed implementation plan with the project team prior to contributing, to avoid wasted time and additional changes. Acceptance of your contribution depends on your willingness and ability to iterate the proposed contribution to achieve the required quality level, coding style, test coverage, and alignment with user requirements as they are understood by the project team.
## Follow project structure, coding style and approaches
./PROJECT.md has information about the structure of this `simplexmq` repository.
./CODE.md has details about general requirements common for `simplexmq` and `simplex-chat` repositories.
This files can be used with LLM prompts, e.g. if you use Claude Code you can create CLAUDE.md file in project root importing content from these files:
```markdown
@README.md
@contributing/PROJECT.md
@contributing/CODE.md
```
-23
View File
@@ -1,23 +0,0 @@
common:
corrId - random BS, used as CbNonce
entityId - p2r tlsUniq
# setup
s->p: "proxy", uri, auth?
# unless connected
p->r: "p_handshake"
p<-r: "r_key", tls-signed dh pub
s<-r: "r_key", tls-signed dh pub # reply entityId contains tlsUniq
# working
s ; generate random dh priv, make shared secret
s->p: s2r("forward", random dh pub, SEND command blob)
p->r: p2r("forward", random dh pub, s2r("forward", ...)))
r->c@ "msg", ...
p<-r: p2r("r_res", s2r("ok" / "error", error))
s<-p@ s2r("ok" / "error", error)
# expired
p<-r@ p2r("error", "key expired")
s<-p@ "error", "key expired"
s ; reconnect
@@ -0,0 +1,472 @@
# XFTP Server PostgreSQL Backend
## Overview
Add PostgreSQL backend support to xftp-server, following the SMP server pattern. Supports bidirectional migration between STM (in-memory with StoreLog) and PostgreSQL backends.
## Goals
- PostgreSQL-backed file metadata storage as an alternative to STM + StoreLog
- Polymorphic server code via `FileStoreClass` typeclass with IO-based methods (following `QueueStoreClass` pattern)
- Bidirectional migration: StoreLog <-> PostgreSQL via CLI commands
- Shared `server_postgres` cabal flag (same flag enables both SMP and XFTP Postgres support)
- INI-based backend selection at runtime
## Architecture
### FileStoreClass Typeclass
IO-based typeclass following the `QueueStoreClass` pattern — each method is a self-contained IO action, with the implementation responsible for its own atomicity (STM backend wraps in `atomically`, Postgres backend uses database transactions):
```haskell
class FileStoreClass s where
type FileStoreConfig s
-- Lifecycle
newFileStore :: FileStoreConfig s -> IO s
closeFileStore :: s -> IO ()
-- File operations
addFile :: s -> SenderId -> FileInfo -> RoundedFileTime -> ServerEntityStatus -> IO (Either XFTPErrorType ())
setFilePath :: s -> SenderId -> FilePath -> IO (Either XFTPErrorType ())
addRecipient :: s -> SenderId -> FileRecipient -> IO (Either XFTPErrorType ())
getFile :: s -> SFileParty p -> XFTPFileId -> IO (Either XFTPErrorType (FileRec, C.APublicAuthKey))
deleteFile :: s -> SenderId -> IO (Either XFTPErrorType ())
blockFile :: s -> SenderId -> BlockingInfo -> Bool -> IO (Either XFTPErrorType ())
deleteRecipient :: s -> RecipientId -> FileRec -> IO ()
ackFile :: s -> RecipientId -> IO (Either XFTPErrorType ())
-- Expiration (with LIMIT for Postgres; called in a loop until empty)
expiredFiles :: s -> Int64 -> Int -> IO [(SenderId, Maybe FilePath, Word32)]
-- Storage and stats (for init-time computation)
getUsedStorage :: s -> IO Int64
getFileCount :: s -> IO Int
```
- STM backend: each method wraps its STM transaction in `atomically` internally.
- Postgres backend: each method runs its query via `withDB` / database connection internally.
No polymorphic monad or `runStore` dispatcher needed — unlike `MsgStoreClass`, XFTP file operations are individually atomic and don't require grouping multiple operations into backend-dependent transactions.
### PostgresFileStore Data Type
```haskell
data PostgresFileStore = PostgresFileStore
{ dbStore :: DBStore,
dbStoreLog :: Maybe (StoreLog 'WriteMode)
}
```
- `dbStore` — connection pool created via `createDBStore`, runs schema migrations on init.
- `dbStoreLog` — optional parallel log file (enabled by `db_store_log` INI setting). When present, every mutation (`addFile`, `setFilePath`, `deleteFile`, `blockFile`, `addRecipient`, `ackFile`) also writes to this log via a `withLog` wrapper. `withLog` is called AFTER the DB operation succeeds (so the log reflects committed state only). Log write failures are non-fatal (logged as warnings, do not fail the DB operation). This provides an audit trail and enables recovery via export.
`closeFileStore` for Postgres calls `closeDBStore` (closes connection pool) then `mapM_ closeStoreLog dbStoreLog` (flushes and closes the parallel log). For STM, it closes the storeLog. Called from a `finally` block during server shutdown, matching SMP's `stopServer``closeMsgStore``closeQueueStore` pattern.
### STMFileStore Type
After extracting from current `Store.hs`, `STMFileStore` retains the file and recipient maps but no longer owns `usedStorage` (moved to `XFTPEnv`):
```haskell
data STMFileStore = STMFileStore
{ files :: TMap SenderId FileRec,
recipients :: TMap RecipientId (SenderId, RcvPublicAuthKey)
}
```
`closeFileStore` for STM is a no-op (TMaps are garbage-collected; the env-level `storeLog` is closed separately by the server).
### Error Handling
Postgres operations follow SMP's `withDB` / `handleDuplicate` pattern:
```haskell
withDB :: Text -> PostgresFileStore -> (DB.Connection -> IO (Either XFTPErrorType a)) -> ExceptT XFTPErrorType IO a
withDB op st action =
ExceptT $ E.try (withTransaction (dbStore st) action) >>= either logErr pure
where
logErr :: E.SomeException -> IO (Either XFTPErrorType a)
logErr e = logError ("STORE: " <> err) $> Left INTERNAL
where
err = op <> ", withDB, " <> tshow e
handleDuplicate :: SqlError -> IO (Either XFTPErrorType a)
handleDuplicate e = case constraintViolation e of
Just (UniqueViolation _) -> pure $ Left DUPLICATE_
_ -> E.throwIO e
```
- All DB operations wrapped in `withDB` — catches exceptions, logs, returns `INTERNAL`.
- Unique constraint violations caught by `handleDuplicate` and mapped to `DUPLICATE_`.
- UPDATE operations verified with `assertUpdated` — returns `AUTH` if 0 rows affected (matching SMP pattern, prevents silent failures when WHERE clause doesn't match).
- Critical sections (DB write + TVar update) wrapped in `uninterruptibleMask_` to prevent async exceptions from leaving inconsistent state between DB and TVars.
### FileRec and TVar Fields
`FileRec` retains its `TVar` fields (matching SMP's `PostgresQueue` pattern):
```haskell
data FileRec = FileRec
{ senderId :: SenderId,
fileInfo :: FileInfo,
filePath :: TVar (Maybe FilePath),
recipientIds :: TVar (Set RecipientId),
createdAt :: RoundedFileTime,
fileStatus :: TVar ServerEntityStatus
}
```
- **STM backend**: TVars are the source of truth, as currently.
- **Postgres backend**: `getFile` reads from DB and creates a `FileRec` with fresh TVars populated from the DB row (matching SMP's `mkQ` pattern — `newTVarIO` per load). Mutation methods (`setFilePath`, `blockFile`, etc.) update both the DB (persistence) and the TVars (in-session consistency). The `recipientIds` TVar is initialized to `S.empty` — no subquery needed because no server code reads `recipientIds` directly; all recipient operations go through the typeclass methods (`addRecipient`, `deleteRecipient`, `ackFile`), which query the `recipients` table for Postgres.
### usedStorage Ownership
`usedStorage :: TVar Int64` moves from the store to `XFTPEnv`. The store typeclass does **not** manage `usedStorage` — it only provides `getUsedStorage` for init-time computation.
- **STM init**: StoreLog replay calls `setFilePath` (which only sets the filePath TVar — the STM `setFilePath` implementation is changed to **not** update `usedStorage`). Similarly, STM `deleteFile` (Store.hs line 117) and `blockFile` (line 125) are changed to **not** update `usedStorage` — the server handles all `usedStorage` adjustments externally. After replay, `getUsedStorage` computes the sum over all file sizes (matching current `countUsedStorage` behavior).
- **Postgres init**: `getUsedStorage` executes `SELECT COALESCE(SUM(file_size), 0) FROM files`.
- **Runtime**: Server manages `usedStorage` TVar directly for reserve/commit/rollback during uploads, and adjusts after `deleteFile`/`blockFile` calls.
**Note on `getUsedStorage` semantics**: The current STM `countUsedStorage` sums all file sizes unconditionally (including files without `filePath` set, i.e., created but not yet uploaded). The Postgres `getUsedStorage` matches this: `SELECT SUM(file_size) FROM files` (no `WHERE file_path IS NOT NULL`). In practice, orphaned files (created but never uploaded) are rare and short-lived (expired within 48h), so the difference is negligible. A future improvement could filter by `file_path IS NOT NULL` in both backends to reflect actual disk usage more accurately.
### Server.hs Refactoring
`Server.hs` becomes polymorphic over `FileStoreClass s`. Since all typeclass methods are IO, call sites replace `atomically` with direct IO calls to the store.
**Call sites requiring changes** (exhaustive list):
1. **`receiveServerFile`** (line 563): `atomically $ writeTVar filePath (Just fPath)``setFilePath store senderId fPath`. The `reserve` logic (line 551-555) stays as direct TVar manipulation on `usedStorage` from `XFTPEnv`.
2. **`verifyXFTPTransmission`** (line 453): `atomically $ verify =<< getFile st party fId` — the `getFile` call and subsequent `readTVar fileStatus` are in a single `atomically` block. Refactored to: `getFile st party fId` (IO), then `readTVarIO (fileStatus fr)` from the returned `FileRec` (safe for both backends — STM TVar is the source of truth, Postgres TVar is a fresh snapshot from DB).
3. **`retryAdd`** (line 516): Signature `XFTPFileId -> STM (Either XFTPErrorType a)``XFTPFileId -> IO (Either XFTPErrorType a)`. The `atomically` call (line 520) replaced with `liftIO`.
4. **`deleteOrBlockServerFile_`** (line 620): Parameter `FileStore -> STM (Either XFTPErrorType ())``FileStoreClass s => s -> IO (Either XFTPErrorType ())`. The `atomically` call (line 626) removed — the store method is already IO. After the store action, server adjusts `usedStorage` TVar in `XFTPEnv` based on `fileInfo.size`.
5. **`ackFileReception`** (line 605): `atomically $ deleteRecipient st rId fr``deleteRecipient st rId fr`.
6. **Control port `CPDelete`/`CPBlock`** (lines 371, 377): `atomically $ getFile fs SFRecipient fileId``getFile fs SFRecipient fileId`.
7. **`expireServerFiles`** (line 636): Replace per-file `expiredFilePath` iteration with batched `expiredFiles st old batchSize`, which returns `[(SenderId, Maybe FilePath, Word32)]` — the `Word32` file size is needed so the server can adjust the `usedStorage` TVar after each deletion. Called in a loop until the returned list is empty. The `itemDelay` between files applies to the deletion loop over each batch, not the query itself. STM backend ignores the batch size limit (returns all expired files from TMap scan); Postgres uses `LIMIT`.
8. **`restoreServerStats`** (line 694): `FileStore {files, usedStorage} <- asks store` accesses store fields directly. Refactored to: `usedStorage` from `XFTPEnv` via `asks usedStorage`, file count via `getFileCount store`. STM: `M.size <$> readTVarIO files`. Postgres: `SELECT COUNT(*) FROM files`.
### Store Config Selection
GADT in `Env.hs`:
```haskell
data XFTPStoreConfig s where
XSCMemory :: Maybe FilePath -> XFTPStoreConfig STMFileStore
#if defined(dbServerPostgres)
XSCDatabase :: PostgresFileStoreCfg -> XFTPStoreConfig PostgresFileStore
#endif
```
`XFTPEnv` becomes polymorphic:
```haskell
data XFTPEnv s = XFTPEnv
{ config :: XFTPServerConfig,
store :: s,
usedStorage :: TVar Int64,
storeLog :: Maybe (StoreLog 'WriteMode),
...
}
```
The `M` monad (`ReaderT (XFTPEnv s) IO`) and all functions in `Server.hs` gain `FileStoreClass s =>` constraints.
**StoreLog lifecycle per backend:**
- **STM mode**: `storeLog = Just sl` (current behavior — append-only log for persistence and recovery).
- **Postgres mode**: `storeLog = Nothing` (main storeLog disabled — Postgres is the source of truth). The optional parallel `dbStoreLog` inside `PostgresFileStore` provides audit/recovery if enabled via `db_store_log` INI setting.
The existing `withFileLog` pattern in Server.hs continues to work unchanged — it maps over `Maybe (StoreLog 'WriteMode)`, which is `Nothing` in Postgres mode so the calls become no-ops.
### Main.hs Store Type Dispatch
The `Start` CLI command gains a `--confirm-migrations` flag (default `MCConsole` — manual prompt, matching SMP's `StartOptions`). For automated deployments, `--confirm-migrations up` auto-applies forward migrations. The import command uses `MCYesUp` (always auto-apply).
Following SMP's existential dispatch pattern (`AStoreType` + `run`), `Main.hs` selects the store type from INI config and dispatches to the polymorphic server:
```haskell
runServer ini = do
let storeType = fromRight "memory" $ lookupValue "STORE_LOG" "store_files" ini
case storeType of
"memory" -> run $ XSCMemory (enableStoreLog $> storeLogFilePath)
"database" ->
#if defined(dbServerPostgres)
run $ XSCDatabase PostgresFileStoreCfg {..}
#else
exitError "server not compiled with Postgres support"
#endif
_ -> exitError $ "Invalid store_files value: " <> storeType
where
run :: FileStoreClass s => XFTPStoreConfig s -> IO ()
run storeCfg = do
env <- newXFTPServerEnv storeCfg config
runReaderT (xftpServer config) env
```
**`newXFTPServerEnv` refactored signature:**
```haskell
newXFTPServerEnv :: FileStoreClass s => XFTPStoreConfig s -> XFTPServerConfig -> IO (XFTPEnv s)
newXFTPServerEnv storeCfg config = do
(store, storeLog) <- case storeCfg of
XSCMemory storeLogPath -> do
st <- newFileStore ()
sl <- mapM (`readWriteFileStore` st) storeLogPath
pure (st, sl)
XSCDatabase dbCfg -> do
st <- newFileStore dbCfg
pure (st, Nothing) -- main storeLog disabled for Postgres
usedStorage <- newTVarIO =<< getUsedStorage store
...
pure XFTPEnv {config, store, usedStorage, storeLog, ...}
```
### Startup Config Validation
Following SMP's `checkMsgStoreMode` pattern, `Main.hs` validates config before starting:
- **`store_files=database` + StoreLog file exists** (without `db_store_log=on`): Error — "StoreLog file present but store_files is `database`. Use `xftp-server database import` to migrate, or set `db_store_log: on`."
- **`store_files=database` + schema doesn't exist**: Error — "Create schema in PostgreSQL or use `xftp-server database import`."
- **`store_files=memory` + Postgres schema exists**: Warning — "Postgres schema exists but store_files is `memory`. Data in Postgres will not be used."
- **Binary compiled without `server_postgres` + `store_files=database`**: Error — "Server not compiled with Postgres support."
## Module Structure
```
src/Simplex/FileTransfer/Server/
Store.hs -- FileStoreClass typeclass + shared types (FileRec, FileRecipient, etc.)
Store/
STM.hs -- STMFileStore (extracted from current Store.hs)
Postgres.hs -- PostgresFileStore [CPP-guarded]
Postgres/
Migrations.hs -- Schema migrations [CPP-guarded]
Config.hs -- PostgresFileStoreCfg [CPP-guarded]
StoreLog.hs -- Unchanged (interchange format for both backends + migration)
Env.hs -- XFTPStoreConfig GADT, polymorphic XFTPEnv
Main.hs -- Store selection, migration CLI commands
Server.hs -- Polymorphic over FileStoreClass
```
## PostgreSQL Schema
Initial migration (`20260325_initial`):
```sql
CREATE TABLE files (
sender_id BYTEA NOT NULL PRIMARY KEY,
file_size INT4 NOT NULL,
file_digest BYTEA NOT NULL,
sender_key BYTEA NOT NULL,
file_path TEXT,
created_at INT8 NOT NULL,
status TEXT NOT NULL DEFAULT 'active'
);
CREATE TABLE recipients (
recipient_id BYTEA NOT NULL PRIMARY KEY,
sender_id BYTEA NOT NULL REFERENCES files ON DELETE CASCADE,
recipient_key BYTEA NOT NULL
);
CREATE INDEX idx_recipients_sender_id ON recipients (sender_id);
CREATE INDEX idx_files_created_at ON files (created_at);
```
- `file_size` is `INT4` matching `Word32` in `FileInfo.size`
- `sender_key` and `recipient_key` stored as `BYTEA` using binary encoding via `C.encodePubKey` / `C.decodePubKey` (matching SMP's `ToField`/`FromField` instances for `APublicAuthKey` — includes algorithm type tag in the binary format)
- `file_path` nullable (set after upload completes via `setFilePath`)
- `ON DELETE CASCADE` for recipients when file is hard-deleted
- `created_at` stores rounded epoch seconds (1-hour precision, `RoundedFileTime`)
- `status` as TEXT via `StrEncoding` (`ServerEntityStatus`: `EntityActive`, `EntityBlocked info`, `EntityOff`)
- Hard deletes (no `deleted_at` column)
- No PL/pgSQL functions needed; `setFilePath` uses `WHERE file_path IS NULL` to prevent duplicate uploads (the `UPDATE` itself acquires a row-level lock)
- `used_storage` computed on startup: `SELECT COALESCE(SUM(file_size), 0) FROM files` (matches STM `countUsedStorage` — all files, see usedStorage Ownership section)
### Migrations Module
Following SMP's `QueueStore/Postgres/Migrations.hs` pattern:
```haskell
module Simplex.FileTransfer.Server.Store.Postgres.Migrations
( xftpServerMigrations,
)
where
import Data.List (sortOn)
import Data.Text (Text)
import Simplex.Messaging.Agent.Store.Shared
import Text.RawString.QQ (r)
xftpSchemaMigrations :: [(String, Text, Maybe Text)]
xftpSchemaMigrations =
[ ("20260325_initial", m20260325_initial, Nothing)
]
xftpServerMigrations :: [Migration]
xftpServerMigrations = sortOn name $ map migration xftpSchemaMigrations
where
migration (name, up, down) = Migration {name, up, down = down}
m20260325_initial :: Text
m20260325_initial =
[r|
CREATE TABLE files (
sender_id BYTEA NOT NULL PRIMARY KEY,
...
);
|]
```
The `Migration` type (from `Simplex.Messaging.Agent.Store.Shared`) has fields `{name :: String, up :: Text, down :: Maybe Text}`. Initial migration has `Nothing` for `down`. Future migrations should include `Just down_migration` for rollback support. Called via `createDBStore dbOpts xftpServerMigrations (MigrationConfig confirmMigrations Nothing)`.
### Postgres Operations
Key query patterns:
- **`addFile`**: `INSERT INTO files (...) VALUES (...)`, return `DUPLICATE_` on unique violation.
- **`setFilePath`**: `UPDATE files SET file_path = ? WHERE sender_id = ? AND file_path IS NULL`, verified with `assertUpdated` (returns `AUTH` if 0 rows affected — file not found or already uploaded). The `WHERE file_path IS NULL` prevents duplicate uploads; the `UPDATE` acquires a row lock implicitly. Only persists the path; `usedStorage` managed by server.
- **`addRecipient`**: `INSERT INTO recipients (...)`, plus check for duplicates. No need for `recipientIds` TVar update — Postgres derives it from the table.
- **`getFile`** (sender): `SELECT ... FROM files WHERE sender_id = ?`, returns auth key from `sender_key` column.
- **`getFile`** (recipient): `SELECT f.*, r.recipient_key FROM recipients r JOIN files f ON ... WHERE r.recipient_id = ?`.
- **`deleteFile`**: `DELETE FROM files WHERE sender_id = ?` (recipients cascade).
- **`blockFile`**: `UPDATE files SET status = ? WHERE sender_id = ?`. When `deleted = True`, the server adjusts `usedStorage` externally (matching current STM behavior where `blockFile` only updates status and storage, not `filePath`).
- **`expiredFiles`**: `SELECT sender_id, file_path, file_size FROM files WHERE created_at + ? < ? LIMIT ?` — batched query replaces per-file iteration, includes `file_size` for `usedStorage` adjustment. Called in a loop until no rows returned.
## INI Configuration
New keys in `[STORE_LOG]` section:
```ini
[STORE_LOG]
enable: on
store_files: memory # memory | database
db_connection: postgresql://xftp@/xftp_server_store
db_schema: xftp_server
db_pool_size: 10
db_store_log: off
expire_files_hours: 48
```
`store_files` selects the backend (`store_files` rather than `store_queues` because XFTP stores files, not queues):
- `memory` -> `XSCMemory` (current behavior)
- `database` -> `XSCDatabase` (requires `server_postgres` build flag)
### INI Template Generation (`xftp-server init`)
The `iniFileContent` function in `Main.hs` must be updated to generate the new keys in the `[STORE_LOG]` section. Following SMP's `iniDbOpts` pattern with `optDisabled'` (prefixes `"# "` when value equals default), Postgres keys are generated commented out by default:
```ini
[STORE_LOG]
enable: on
# File storage mode: `memory` or `database` (PostgreSQL).
store_files: memory
# Database connection settings for PostgreSQL database (`store_files: database`).
# db_connection: postgresql://xftp@/xftp_server_store
# db_schema: xftp_server
# db_pool_size: 10
# Write database changes to store log file
# db_store_log: off
expire_files_hours: 48
```
Reuses `iniDBOptions` from `Simplex.Messaging.Server.CLI` for runtime parsing (falls back to defaults when keys are commented out or missing). `enableDbStoreLog'` pattern (`settingIsOn "STORE_LOG" "db_store_log"`) controls `dbStoreLogPath`.
### PostgresFileStoreCfg
```haskell
data PostgresFileStoreCfg = PostgresFileStoreCfg
{ dbOpts :: DBOpts,
dbStoreLogPath :: Maybe FilePath,
confirmMigrations :: MigrationConfirmation
}
```
No `deletedTTL` (hard deletes).
### Default DB Options
```haskell
defaultXFTPDBOpts :: DBOpts
defaultXFTPDBOpts =
DBOpts
{ connstr = "postgresql://xftp@/xftp_server_store",
schema = "xftp_server",
poolSize = 10,
createSchema = False
}
```
## Migration CLI
Bidirectional migration via StoreLog as interchange format:
```
xftp-server database import [--database DB_CONN] [--schema DB_SCHEMA] [--pool-size N]
xftp-server database export [--database DB_CONN] [--schema DB_SCHEMA] [--pool-size N]
```
No `--table` flag needed (unlike SMP which has queues/messages/all) — XFTP has a single entity type (files + recipients, always migrated together).
CLI options reuse `dbOptsP` parser from `Simplex.Messaging.Server.CLI`.
### Import (StoreLog -> PostgreSQL)
1. Confirm: prompt user with database connection details and StoreLog path
2. Read and replay StoreLog into temporary `STMFileStore`
3. Connect to PostgreSQL, run schema migrations (`createSchema = True`, `confirmMigrations = MCYesUp`)
4. Batch-insert file records into `files` table using PostgreSQL COPY protocol (matching SMP's `batchInsertQueues` pattern for performance). Progress reported every 10k files.
5. Batch-insert recipient records into `recipients` table using COPY protocol
6. Verify counts: `SELECT COUNT(*) FROM files` / `recipients` — warn if mismatch
7. Rename StoreLog to `.bak` (prevents accidental re-import, preserves original for rollback)
8. Report counts
### Export (PostgreSQL -> StoreLog)
1. Confirm: prompt user with database connection details and output path. Fail if output file already exists.
2. Connect to PostgreSQL
3. Open new StoreLog file for writing
4. Fold over all file records, writing per file (in this order, matching existing `writeFileStore`): `AddFile` (with `ServerEntityStatus` — this preserves `EntityBlocked` state), `AddRecipients`, then `PutFile` (if `file_path` is set)
5. Report counts
Note: `AddFile` carries `ServerEntityStatus` which includes `EntityBlocked info`, so blocking state is preserved through export/import without needing separate `BlockFile` log entries.
File data on disk is untouched by migration — only metadata moves between backends.
## Cabal Integration
Shared `server_postgres` flag. New Postgres modules added to existing conditional block:
```cabal
if flag(server_postgres)
cpp-options: -DdbServerPostgres
exposed-modules:
...existing SMP modules...
Simplex.FileTransfer.Server.Store.Postgres
Simplex.FileTransfer.Server.Store.Postgres.Migrations
Simplex.FileTransfer.Server.Store.Postgres.Config
```
CPP guards (`#if defined(dbServerPostgres)`) in:
- `Store.hs` — Postgres `FromField`/`ToField` instances for XFTP-specific types if needed
- `Env.hs``XSCDatabase` constructor
- `Main.hs` — database CLI commands, store selection for `database` mode, Postgres imports
- `Server.hs` — Postgres-specific imports if needed
## Testing
- **Parameterized server tests**: Existing `xftpServerTests` refactored to accept a store type parameter (following SMP's `SpecWith (ASrvTransport, AStoreType)` pattern). The same server tests run against both STM and Postgres backends — STM tests run unconditionally, Postgres tests added under `#if defined(dbServerPostgres)` with `postgressBracket` for database lifecycle (drop → create → test → drop).
- **Unit tests**: `PostgresFileStore` operations — add/get/delete/block/expire, duplicate detection, auth errors
- **Migration round-trip**: STM store → export to StoreLog → import to Postgres → export back → verify StoreLog equality (including blocked file status)
- **Tests location**: in `tests/` alongside existing XFTP tests, guarded by `server_postgres` CPP flag
- **Test database**: PostgreSQL on `localhost:5432`, using a dedicated `xftp_server_test` schema (dropped and recreated per test run via `postgressBracket`, following SMP's test database lifecycle pattern)
- **Test fixtures**: `testXFTPStoreDBOpts :: DBOpts` with `createSchema = True`, `confirmMigrations = MCYesUp`, in `tests/XFTPClient.hs`
@@ -0,0 +1,648 @@
# XFTP PostgreSQL Backend — Implementation Plan
> **For agentic workers:** REQUIRED: Use superpowers-extended-cc:subagent-driven-development (if subagents available) or superpowers-extended-cc:executing-plans to implement this plan. Steps use checkbox (`- [ ]`) syntax for tracking.
**Goal:** Add PostgreSQL backend support to xftp-server as an alternative to STM + StoreLog, with bidirectional migration.
**Architecture:** Introduce `FileStoreClass` typeclass (IO-based, following `QueueStoreClass` pattern). Extract current STM store into `Store/STM.hs`, make `Server.hs` polymorphic, then add `Store/Postgres.hs` behind `server_postgres` CPP flag. `usedStorage` moves from store to `XFTPEnv` so the server manages quota tracking externally.
**Tech Stack:** Haskell, postgresql-simple, STM, fourmolu, cabal with CPP flags
**Design spec:** `plans/2026-03-25-xftp-postgres-backend-design.md`
---
## File Structure
**Existing files modified:**
- `src/Simplex/FileTransfer/Server/Store.hs` — rewritten: becomes typeclass + shared types
- `src/Simplex/FileTransfer/Server/Env.hs` — polymorphic `XFTPEnv s`, `XFTPStoreConfig` GADT
- `src/Simplex/FileTransfer/Server.hs` — polymorphic over `FileStoreClass s`
- `src/Simplex/FileTransfer/Server/StoreLog.hs` — update for IO store functions
- `src/Simplex/FileTransfer/Server/Main.hs` — INI config, dispatch, CLI commands
- `simplexmq.cabal` — new modules
- `tests/XFTPClient.hs` — Postgres test fixtures
- `tests/Test.hs` — Postgres test group
**New files created:**
- `src/Simplex/FileTransfer/Server/Store/STM.hs``STMFileStore` (extracted from current `Store.hs`)
- `src/Simplex/FileTransfer/Server/Store/Postgres.hs``PostgresFileStore` [CPP-guarded]
- `src/Simplex/FileTransfer/Server/Store/Postgres/Config.hs``PostgresFileStoreCfg` [CPP-guarded]
- `src/Simplex/FileTransfer/Server/Store/Postgres/Migrations.hs` — schema SQL [CPP-guarded]
- `tests/CoreTests/XFTPStoreTests.hs` — Postgres store unit tests [CPP-guarded]
---
## Task 1: Move `usedStorage` from `FileStore` to `XFTPEnv`
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Store.hs`
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- Modify: `src/Simplex/FileTransfer/Server.hs`
- [ ] **Step 1: Remove `usedStorage` from `FileStore` in `Store.hs`**
1. Remove `usedStorage :: TVar Int64` field from `FileStore` record (line 47).
2. Remove `usedStorage <- newTVarIO 0` from `newFileStore` (line 75) and drop the field from the record construction (line 76).
3. In `setFilePath` (line 92-97): remove `modifyTVar' (usedStorage st) (+ fromIntegral (size fileInfo))` — keep only `writeTVar filePath (Just fPath)`. Change pattern from `\FileRec {fileInfo, filePath}` to `\FileRec {filePath}` (fileInfo is now unused — `-Wunused-matches` error).
4. In `deleteFile` (line 112-119): remove `modifyTVar' usedStorage $ subtract (fromIntegral $ size fileInfo)`. Change outer pattern match from `FileStore {files, recipients, usedStorage}` to `FileStore {files, recipients}`. Change inner pattern from `Just FileRec {fileInfo, recipientIds}` to `Just FileRec {recipientIds}` (`fileInfo` is now unused — `-Wunused-matches` error).
5. In `blockFile` (line 122-127): remove `when deleted $ modifyTVar' usedStorage $ subtract (fromIntegral $ size fileInfo)`. Change pattern match from `st@FileStore {usedStorage}` to `st`. The `deleted` parameter and `fileInfo` in the inner pattern become unused — prefix with `_` or remove from pattern to avoid `-Wunused-matches`.
- [ ] **Step 2: Add `usedStorage` to `XFTPEnv` in `Env.hs`**
1. Add `usedStorage :: TVar Int64` field to `XFTPEnv` record (between `store` and `storeLog`, line 93).
2. In `newXFTPServerEnv` (line 112-126): replace lines 117-118:
```
used <- countUsedStorage <$> readTVarIO (files store)
atomically $ writeTVar (usedStorage store) used
```
with:
```
usedStorage <- newTVarIO =<< countUsedStorage <$> readTVarIO (files store)
```
3. Add `usedStorage` to the `pure XFTPEnv {..}` construction.
- [ ] **Step 3: Update all `usedStorage` access sites in `Server.hs`**
1. Line 552: `us <- asks $ usedStorage . store` → `us <- asks usedStorage`.
2. Line 569: `us <- asks $ usedStorage . store` → `us <- asks usedStorage`.
3. Line 639: `usedStart <- readTVarIO $ usedStorage st` → `usedStart <- readTVarIO =<< asks usedStorage`.
4. Line 647: `usedEnd <- readTVarIO $ usedStorage st` → `usedEnd <- readTVarIO =<< asks usedStorage`.
5. Line 694: `FileStore {files, usedStorage} <- asks store` → split into `FileStore {files} <- asks store` and `usedStorage <- asks usedStorage`.
6. In `deleteOrBlockServerFile_` (line 620): after `void $ atomically $ storeAction st`, add usedStorage adjustment — `us <- asks usedStorage` then `atomically $ modifyTVar' us $ subtract (fromIntegral $ size fileInfo)` when file had a path (check `path` from `readTVarIO filePath` earlier in the function).
- [ ] **Step 4: Build and verify**
Run: `cabal build`
- [ ] **Step 5: Run existing tests**
Run: `cabal test --test-show-details=streaming --test-option=--match="/XFTP/"`
- [ ] **Step 6: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs
git add src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs
git commit -m "refactor(xftp): move usedStorage from FileStore to XFTPEnv"
```
---
## Task 2: Add `getUsedStorage`, `getFileCount`, `expiredFiles` functions
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Store.hs`
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- Modify: `src/Simplex/FileTransfer/Server.hs`
- [ ] **Step 1: Add three new functions to `Store.hs`**
1. Add to exports: `getUsedStorage`, `getFileCount`, `expiredFiles`.
2. Remove `expiredFilePath` from exports AND delete the function definition (dead code → `-Wunused-binds` error). Also remove `($>>=)` from import `Simplex.Messaging.Util (ifM, ($>>=))` → `Simplex.Messaging.Util (ifM)` — `$>>=` was only used by `expiredFilePath`.
3. Add import: `qualified Data.Map.Strict as M` (needed for `M.foldl'` in `getUsedStorage` and `M.toList` in `expiredFiles`).
4. Implement:
```haskell
getUsedStorage :: FileStore -> IO Int64
getUsedStorage FileStore {files} =
M.foldl' (\acc FileRec {fileInfo = FileInfo {size}} -> acc + fromIntegral size) 0 <$> readTVarIO files
getFileCount :: FileStore -> IO Int
getFileCount FileStore {files} = M.size <$> readTVarIO files
expiredFiles :: FileStore -> Int64 -> Int -> IO [(SenderId, Maybe FilePath, Word32)]
expiredFiles FileStore {files} old _limit = do
fs <- readTVarIO files
fmap catMaybes . forM (M.toList fs) $ \(sId, FileRec {fileInfo = FileInfo {size}, filePath, createdAt = RoundedSystemTime createdAt}) ->
if createdAt + fileTimePrecision < old
then do
path <- readTVarIO filePath
pure $ Just (sId, path, size)
else pure Nothing
```
5. Add imports: `Data.Maybe (catMaybes)`, `Data.Word (Word32)` (note: `qualified Data.Map.Strict as M` already added in item 3).
- [ ] **Step 2: Replace `countUsedStorage` in `Env.hs`**
1. Replace `countUsedStorage <$> readTVarIO (files store)` with `getUsedStorage store` in `newXFTPServerEnv`.
2. Remove `countUsedStorage` function definition and its export.
3. Remove `qualified Data.Map.Strict as M` import if no longer used.
- [ ] **Step 3: Update `restoreServerStats` in `Server.hs` to use `getFileCount`**
In `restoreServerStats` (line 694-696): replace `FileStore {files} <- asks store` and `_filesCount <- M.size <$> readTVarIO files` with `st <- asks store` and `_filesCount <- liftIO $ getFileCount st` (eliminates the `FileStore` pattern match — `files` binding no longer needed).
- [ ] **Step 4: Replace `expireServerFiles` iteration in `Server.hs`**
1. Replace the body of `expireServerFiles` (lines 636-660). Remove `files' <- readTVarIO (files st)` and the `forM_ (M.keys files')` loop.
2. New body: call `expiredFiles st old 10000` in a loop. For each `(sId, filePath_, fileSize)` in returned list: apply `itemDelay`, remove disk file if present, call `atomically $ deleteFile st sId`, adjust `usedStorage` TVar by `fileSize`, increment `filesExpired` stat. Loop until `expiredFiles` returns `[]`.
3. Remove `Data.Map.Strict` import from Server.hs if no longer needed (was used for `M.size` and `M.keys` — now replaced by `getFileCount` and `expiredFiles`).
- [ ] **Step 5: Build and verify**
Run: `cabal build`
- [ ] **Step 6: Run existing tests**
Run: `cabal test --test-show-details=streaming --test-option=--match="/XFTP/"`
- [ ] **Step 7: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs
git add src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs
git commit -m "refactor(xftp): add getUsedStorage, getFileCount, expiredFiles store functions"
```
---
## Task 3: Change `Store.hs` functions from STM to IO
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Store.hs`
- Modify: `src/Simplex/FileTransfer/Server.hs`
- Modify: `src/Simplex/FileTransfer/Server/StoreLog.hs`
- [ ] **Step 1: Change all Store.hs function signatures from STM to IO**
For each of: `addFile`, `setFilePath`, `addRecipient`, `getFile`, `deleteFile`, `blockFile`, `deleteRecipient`, `ackFile`:
1. Change return type from `STM (Either XFTPErrorType ...)` to `IO (Either XFTPErrorType ...)` (or `STM ()` to `IO ()` for `deleteRecipient`).
2. Wrap the function body in `atomically $ do ...`.
3. Keep `withFile` and `newFileRec` as internal STM helpers (called inside the `atomically` blocks).
- [ ] **Step 2: Update Server.hs call sites — remove `atomically` wrappers**
1. Line 563 (`receiveServerFile`): change `atomically $ writeTVar filePath (Just fPath)` → add `st <- asks store` then `void $ liftIO $ setFilePath st senderId fPath` (design call site #1 — `store` is not in scope in `receiveServerFile`'s `receive` helper, so bind via `asks`; `void` avoids `-Wunused-do-bind` warning on the `Either` result).
2. Line 453 (`verifyXFTPTransmission`): split `atomically $ verify =<< getFile st party fId` into: `liftIO (getFile st party fId)` (IO→M lift), then pattern match on result, use `readTVarIO (fileStatus fr)` instead of `readTVar`.
3. Lines 371, 377 (control port `CPDelete`/`CPBlock`): change `ExceptT $ atomically $ getFile fs SFRecipient fileId` → `ExceptT $ liftIO $ getFile fs SFRecipient fileId` (inside `unliftIO u $ do` block which runs in M monad — `liftIO` required to lift IO into M).
4. Line 508 (`addFile` in `createFile`): the `ExceptT $ addFile st sId file ts EntityActive` — `addFile` is now IO, `ExceptT` wraps IO directly. Remove any `atomically`.
5. Line 514 (`addRecipient`): same — `ExceptT . addRecipient st sId` works directly in IO.
6. Line 516 (`retryAdd`): change parameter type from `(XFTPFileId -> STM (Either XFTPErrorType a))` to `(XFTPFileId -> IO (Either XFTPErrorType a))`. Line 520: change `atomically (add fId)` to `liftIO (add fId)`.
7. Line 605 (`ackFileReception`): change `atomically $ deleteRecipient st rId fr` to `liftIO $ deleteRecipient st rId fr`.
8. Line 620 (`deleteOrBlockServerFile_`): change third parameter type from `(FileStore -> STM (Either XFTPErrorType ()))` to `(FileStore -> IO (Either XFTPErrorType ()))`. Line 626: change `void $ atomically $ storeAction st` to `void $ liftIO $ storeAction st`.
9. `expireServerFiles` `delete` helper: change `atomically $ deleteFile st sId` to `liftIO $ deleteFile st sId` (deleteFile is now IO; `liftIO` required because the helper runs in M monad, not IO).
- [ ] **Step 3: Update `StoreLog.hs` — remove `atomically` from replay**
In `readFileStore` (line 93), function `addToStore`:
1. Change `atomically (addToStore lr)` to `addToStore lr` — store functions are now IO.
2. The `addToStore` body calls `addFile`, `setFilePath`, `deleteFile`, `blockFile`, `ackFile` — all IO now, no `atomically` needed.
3. For `AddRecipients`: `runExceptT $ mapM_ (ExceptT . addRecipient st sId) rcps` — `addRecipient` returns `IO (Either ...)`, so `ExceptT . addRecipient st sId` works directly.
- [ ] **Step 4: Build and verify**
Run: `cabal build`
- [ ] **Step 5: Run existing tests**
Run: `cabal test --test-show-details=streaming --test-option=--match="/XFTP/"`
- [ ] **Step 6: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/StoreLog.hs
git add src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/StoreLog.hs
git commit -m "refactor(xftp): change file store operations from STM to IO"
```
---
## Task 4: Extract `FileStoreClass` typeclass, move STM impl to `Store/STM.hs`
**Files:**
- Rewrite: `src/Simplex/FileTransfer/Server/Store.hs`
- Create: `src/Simplex/FileTransfer/Server/Store/STM.hs`
- Modify: `src/Simplex/FileTransfer/Server/StoreLog.hs`
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- Modify: `src/Simplex/FileTransfer/Server.hs`
- Modify: `simplexmq.cabal`
- [ ] **Step 1: Create `Store/STM.hs` — move all implementation code**
1. Create directory `src/Simplex/FileTransfer/Server/Store/`.
2. Create `src/Simplex/FileTransfer/Server/Store/STM.hs`.
3. Move from `Store.hs`: `FileStore` data type (rename to `STMFileStore`), all function implementations, internal helpers (`withFile`, `newFileRec`), all STM-specific imports.
4. Rename all `FileStore` references to `STMFileStore` in the new file.
5. Module declaration: `module Simplex.FileTransfer.Server.Store.STM` exporting only `STMFileStore (..)` — do NOT export standalone functions (`addFile`, `setFilePath`, etc.) to avoid name collisions with the typeclass methods from `Store.hs`.
- [ ] **Step 2: Rewrite `Store.hs` as the typeclass module**
1. Add `{-# LANGUAGE TypeFamilies #-}` pragma to `Store.hs` (required for `type FileStoreConfig s` associated type).
2. Keep in `Store.hs`: `FileRec (..)`, `FileRecipient (..)`, `RoundedFileTime`, `fileTimePrecision` definitions and their `StrEncoding` instance.
3. Add `FileStoreClass` typeclass:
```haskell
class FileStoreClass s where
type FileStoreConfig s
-- Lifecycle
newFileStore :: FileStoreConfig s -> IO s
closeFileStore :: s -> IO ()
-- File operations
addFile :: s -> SenderId -> FileInfo -> RoundedFileTime -> ServerEntityStatus -> IO (Either XFTPErrorType ())
setFilePath :: s -> SenderId -> FilePath -> IO (Either XFTPErrorType ())
addRecipient :: s -> SenderId -> FileRecipient -> IO (Either XFTPErrorType ())
getFile :: s -> SFileParty p -> XFTPFileId -> IO (Either XFTPErrorType (FileRec, C.APublicAuthKey))
deleteFile :: s -> SenderId -> IO (Either XFTPErrorType ())
blockFile :: s -> SenderId -> BlockingInfo -> Bool -> IO (Either XFTPErrorType ())
deleteRecipient :: s -> RecipientId -> FileRec -> IO ()
ackFile :: s -> RecipientId -> IO (Either XFTPErrorType ())
-- Expiration
expiredFiles :: s -> Int64 -> Int -> IO [(SenderId, Maybe FilePath, Word32)]
-- Stats
getUsedStorage :: s -> IO Int64
getFileCount :: s -> IO Int
```
4. Do NOT re-export from `Store/STM.hs` — this would create a circular module dependency (Store.hs imports Store/STM.hs, Store/STM.hs imports Store.hs). Consumers must import `Store.STM` directly where they need `STMFileStore`.
5. Remove all STM-specific imports that are no longer needed.
- [ ] **Step 3: Add `FileStoreClass` instance in `Store/STM.hs`**
1. Import `FileStoreClass` from `Simplex.FileTransfer.Server.Store`.
2. Inline all implementations directly in the instance body (do NOT delegate to standalone functions — the standalone names collide with typeclass method names, causing ambiguous occurrences for importers):
```haskell
instance FileStoreClass STMFileStore where
type FileStoreConfig STMFileStore = ()
newFileStore () = do
files <- TM.emptyIO
recipients <- TM.emptyIO
pure STMFileStore {files, recipients}
closeFileStore _ = pure ()
addFile st sId fileInfo createdAt status = atomically $ ...
setFilePath st sId fPath = atomically $ ...
-- ... (each method's body is the existing function body, inlined)
```
3. Remove the standalone top-level function definitions — they are now instance methods. Keep only `withFile` and `newFileRec` as internal helpers used by the instance methods.
- [ ] **Step 4: Update importers**
1. `Env.hs`: add `import Simplex.FileTransfer.Server.Store.STM (STMFileStore (..))`. Change `FileStore` → `STMFileStore` in `XFTPEnv` type and `newXFTPServerEnv`. Change `store <- newFileStore` to `store <- newFileStore ()` (typeclass method now takes `FileStoreConfig STMFileStore` which is `()`). Keep `import Simplex.FileTransfer.Server.Store` for `FileRec`, `FileRecipient`, `FileStoreClass`, etc.
2. `Server.hs`: add `import Simplex.FileTransfer.Server.Store.STM`. Change `FileStore` → `STMFileStore` in any explicit type annotations. Import `FileStoreClass` from `Simplex.FileTransfer.Server.Store`.
3. `StoreLog.hs`: add `import Simplex.FileTransfer.Server.Store.STM` to access concrete `STMFileStore` type and store functions used during log replay. Change `FileStore` → `STMFileStore` in `readWriteFileStore` and `writeFileStore` parameter types.
- [ ] **Step 5: Update cabal file**
Add `Simplex.FileTransfer.Server.Store.STM` to `exposed-modules` in the `!flag(client_library)` section, alongside existing XFTP server modules.
- [ ] **Step 6: Build and verify**
Run: `cabal build`
- [ ] **Step 7: Run existing tests**
Run: `cabal test --test-show-details=streaming --test-option=--match="/XFTP/"`
- [ ] **Step 8: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Store/STM.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/StoreLog.hs
git add src/Simplex/FileTransfer/Server/Store.hs src/Simplex/FileTransfer/Server/Store/STM.hs src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/StoreLog.hs simplexmq.cabal
git commit -m "refactor(xftp): extract FileStoreClass typeclass, move STM impl to Store.STM"
```
---
## Task 5: Make `XFTPEnv` and `Server.hs` polymorphic over `FileStoreClass`
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- Modify: `src/Simplex/FileTransfer/Server.hs`
- Modify: `src/Simplex/FileTransfer/Server/Main.hs`
- Modify: `tests/XFTPClient.hs` (if it calls `runXFTPServerBlocking` directly)
- [ ] **Step 1: Make `XFTPEnv` polymorphic in `Env.hs`**
1. Add `XFTPStoreConfig` GADT: `data XFTPStoreConfig s where XSCMemory :: Maybe FilePath -> XFTPStoreConfig STMFileStore`.
2. Change `data XFTPEnv` to `data XFTPEnv s` — field `store :: FileStore` becomes `store :: s`.
3. Change `newXFTPServerEnv :: XFTPServerConfig -> IO XFTPEnv` to `newXFTPServerEnv :: FileStoreClass s => XFTPStoreConfig s -> XFTPServerConfig -> IO (XFTPEnv s)`.
4. Pattern match on `XSCMemory storeLogPath` in `newXFTPServerEnv` body. Create store via `newFileStore ()`, storeLog via `mapM (`readWriteFileStore` st) storeLogPath`.
- [ ] **Step 2: Make `Server.hs` polymorphic**
1. Change `type M a = ReaderT XFTPEnv IO a` to `type M s a = ReaderT (XFTPEnv s) IO a`.
2. Add `FileStoreClass s =>` constraint to all functions using `M s a`. Use `forall s.` in signatures of functions that have `where`-block bindings with `M s` type annotations — `ScopedTypeVariables` requires explicit `forall` to bring `s` into scope for inner type signatures (matching SMP's `smpServer :: forall s. MsgStoreClass s => ...` pattern). Full list: `xftpServer`, `processRequest`, `verifyXFTPTransmission`, `processXFTPRequest` and all its `where`-bound functions (`createFile`, `addRecipients`, `receiveServerFile`, `sendServerFile`, `deleteServerFile`, `ackFileReception`, `retryAdd`, `addFileRetry`, `addRecipientRetry`), `deleteServerFile_`, `blockServerFile`, `deleteOrBlockServerFile_`, `expireServerFiles`, `randomId`, `getFileId`, `withFileLog`, `incFileStat`, `saveServerStats`, `restoreServerStats`, `randomDelay` (inside `#ifdef slow_servers` CPP block). Also update `encodeXftp` (line 236) and `runCPClient` (line 339) which use explicit `ReaderT XFTPEnv IO` instead of the `M` alias — change to `ReaderT (XFTPEnv s) IO`.
3. Change `runXFTPServerBlocking` and `runXFTPServer` to take `XFTPStoreConfig s` parameter.
4. Add `closeFileStore store` call to the server shutdown path (in the `finally` block or `stopServer` equivalent — after saving stats, before logging "Server stopped"). This ensures Postgres connection pool and `dbStoreLog` are properly closed. For STM this is a no-op.
- [ ] **Step 3: Update `Main.hs` dispatch**
1. In `runServer`: construct `XSCMemory (enableStoreLog $> storeLogFilePath)`.
2. Add dispatch function that calls the updated `runXFTPServer` (which creates `started` internally):
```haskell
run :: FileStoreClass s => XFTPStoreConfig s -> IO ()
run storeCfg = runXFTPServer storeCfg serverConfig
```
3. Call `run` with the `XSCMemory` config.
- [ ] **Step 4: Update test helper if needed**
If `tests/XFTPClient.hs` calls `runXFTPServerBlocking` directly, update the call to pass an `XSCMemory` config. Check the `withXFTPServer` / `serverBracket` helper.
- [ ] **Step 5: Build and verify**
Run: `cabal build && cabal build test:simplexmq-test`
- [ ] **Step 6: Run existing tests**
Run: `cabal test --test-show-details=streaming --test-option=--match="/XFTP/"`
- [ ] **Step 7: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/Main.hs
git add src/Simplex/FileTransfer/Server/Env.hs src/Simplex/FileTransfer/Server.hs src/Simplex/FileTransfer/Server/Main.hs tests/XFTPClient.hs simplexmq.cabal
git commit -m "refactor(xftp): make XFTPEnv and server polymorphic over FileStoreClass"
```
---
## Task 6: Add Postgres config, migrations, and store skeleton
**Files:**
- Create: `src/Simplex/FileTransfer/Server/Store/Postgres/Config.hs`
- Create: `src/Simplex/FileTransfer/Server/Store/Postgres/Migrations.hs`
- Create: `src/Simplex/FileTransfer/Server/Store/Postgres.hs`
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- Modify: `simplexmq.cabal`
- [ ] **Step 1: Create `Store/Postgres/Config.hs`**
```haskell
module Simplex.FileTransfer.Server.Store.Postgres.Config
( PostgresFileStoreCfg (..),
defaultXFTPDBOpts,
)
where
import Simplex.Messaging.Agent.Store.Postgres.Options (DBOpts (..))
import Simplex.Messaging.Agent.Store.Shared (MigrationConfirmation)
data PostgresFileStoreCfg = PostgresFileStoreCfg
{ dbOpts :: DBOpts,
dbStoreLogPath :: Maybe FilePath,
confirmMigrations :: MigrationConfirmation
}
defaultXFTPDBOpts :: DBOpts
defaultXFTPDBOpts =
DBOpts
{ connstr = "postgresql://xftp@/xftp_server_store",
schema = "xftp_server",
poolSize = 10,
createSchema = False
}
```
- [ ] **Step 2: Create `Store/Postgres/Migrations.hs`**
Full migration module with `xftpServerMigrations :: [Migration]` and `m20260325_initial` containing CREATE TABLE SQL for `files` and `recipients` tables plus indexes. Follow SMP's `QueueStore/Postgres/Migrations.hs` pattern exactly: tuple list → `sortOn name . map migration`.
- [ ] **Step 3: Create `Store/Postgres.hs` with stub instance**
1. Define `PostgresFileStore` with `dbStore :: DBStore` and `dbStoreLog :: Maybe (StoreLog 'WriteMode)`.
2. `instance FileStoreClass PostgresFileStore` with `error "not implemented"` for all methods except `newFileStore` (calls `createDBStore` + opens `dbStoreLog`) and `closeFileStore` (closes both). `type FileStoreConfig PostgresFileStore = PostgresFileStoreCfg`.
3. Add `withDB`, `handleDuplicate`, `assertUpdated`, `withLog` helpers.
- [ ] **Step 4: Add `XSCDatabase` GADT constructor in `Env.hs` (CPP-guarded)**
```haskell
#if defined(dbServerPostgres)
import Simplex.FileTransfer.Server.Store.Postgres (PostgresFileStore)
import Simplex.FileTransfer.Server.Store.Postgres.Config (PostgresFileStoreCfg)
#endif
data XFTPStoreConfig s where
XSCMemory :: Maybe FilePath -> XFTPStoreConfig STMFileStore
#if defined(dbServerPostgres)
XSCDatabase :: PostgresFileStoreCfg -> XFTPStoreConfig PostgresFileStore
#endif
```
- [ ] **Step 5: Update cabal**
Add to existing `if flag(server_postgres)` block:
```
Simplex.FileTransfer.Server.Store.Postgres
Simplex.FileTransfer.Server.Store.Postgres.Config
Simplex.FileTransfer.Server.Store.Postgres.Migrations
```
- [ ] **Step 6: Build both ways**
Run: `cabal build && cabal build -fserver_postgres`
- [ ] **Step 7: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store/Postgres.hs src/Simplex/FileTransfer/Server/Store/Postgres/Config.hs src/Simplex/FileTransfer/Server/Env.hs
git add src/Simplex/FileTransfer/Server/Store/Postgres.hs src/Simplex/FileTransfer/Server/Store/Postgres/Config.hs src/Simplex/FileTransfer/Server/Store/Postgres/Migrations.hs src/Simplex/FileTransfer/Server/Env.hs simplexmq.cabal
git commit -m "feat(xftp): add PostgreSQL store skeleton with schema migration"
```
---
## Task 7: Implement `PostgresFileStore` operations
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Store/Postgres.hs`
- [ ] **Step 1: Implement `addFile`**
`INSERT INTO files (sender_id, file_size, file_digest, sender_key, file_path, created_at, status) VALUES (?,?,?,?,NULL,?,?)`. Catch unique violation with `handleDuplicate` → `DUPLICATE_`. Call `withLog "addFile"` after.
- [ ] **Step 2: Implement `getFile`**
For `SFSender`: `SELECT ... FROM files WHERE sender_id = ?`. Construct `FileRec` with `newTVarIO` per TVar field. `recipientIds = S.empty`.
For `SFRecipient`: `SELECT f.*, r.recipient_key FROM recipients r JOIN files f ON r.sender_id = f.sender_id WHERE r.recipient_id = ?`.
- [ ] **Step 3: Implement `setFilePath`**
`UPDATE files SET file_path = ? WHERE sender_id = ? AND file_path IS NULL`. Use `assertUpdated`. Call `withLog "setFilePath"`.
- [ ] **Step 4: Implement `addRecipient`**
`INSERT INTO recipients (recipient_id, sender_id, recipient_key) VALUES (?,?,?)`. `handleDuplicate` → `DUPLICATE_`. Call `withLog "addRecipient"`.
- [ ] **Step 5: Implement `deleteFile`, `blockFile`**
`deleteFile`: `DELETE FROM files WHERE sender_id = ?` (CASCADE). `withLog "deleteFile"`.
`blockFile`: `UPDATE files SET status = ? WHERE sender_id = ?`. `assertUpdated`. `withLog "blockFile"`.
- [ ] **Step 6: Implement `deleteRecipient`, `ackFile`**
`deleteRecipient`: `DELETE FROM recipients WHERE recipient_id = ?`. `withLog "deleteRecipient"`.
`ackFile`: same + return `Left AUTH` if 0 rows.
- [ ] **Step 7: Implement `expiredFiles`, `getUsedStorage`, `getFileCount`**
`expiredFiles`: `SELECT sender_id, file_path, file_size FROM files WHERE created_at + ? < ? LIMIT ?`.
`getUsedStorage`: `SELECT COALESCE(SUM(file_size), 0) FROM files`.
`getFileCount`: `SELECT COUNT(*) FROM files`.
- [ ] **Step 8: Add `ToField`/`FromField` instances**
For `RoundedFileTime` (Int64 wrapper), `ServerEntityStatus` (Text via StrEncoding), `C.APublicAuthKey` (Binary via `encodePubKey`/`decodePubKey`). Check SMP's `QueueStore/Postgres.hs` for existing instances to import.
- [ ] **Step 9: Wrap mutation operations in `uninterruptibleMask_`**
Operations that combine a DB write with a TVar update (e.g., `getFile` constructs `FileRec` with `newTVarIO`) must be wrapped in `E.uninterruptibleMask_` to prevent async exceptions from leaving inconsistent state. Follow SMP's `addQueue_`, `deleteStoreQueue` pattern.
- [ ] **Step 10: Build**
Run: `cabal build -fserver_postgres`
- [ ] **Step 11: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Store/Postgres.hs
git add src/Simplex/FileTransfer/Server/Store/Postgres.hs
git commit -m "feat(xftp): implement PostgresFileStore operations"
```
---
## Task 8: Add INI config, Main.hs dispatch, startup validation
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Main.hs`
- Modify: `src/Simplex/FileTransfer/Server/Env.hs`
- [ ] **Step 1: Update `iniFileContent` in `Main.hs`**
Add to `[STORE_LOG]` section: `store_files: memory`, commented-out `db_connection`, `db_schema`, `db_pool_size`, `db_store_log` keys. Follow SMP's `optDisabled'` pattern for commented defaults.
- [ ] **Step 2: Add `StartOptions` and `--confirm-migrations` flag**
```haskell
data StartOptions = StartOptions
{ confirmMigrations :: MigrationConfirmation
}
```
Add to `Start` command parser with default `MCConsole`. Thread through to `runServer`.
- [ ] **Step 3: Add store_files INI parsing and CPP-guarded Postgres dispatch**
In `runServer`: read `store_files` from INI (`fromRight "memory" $ lookupValue "STORE_LOG" "store_files" ini`). Add `"database"` branch (CPP-guarded) that constructs `PostgresFileStoreCfg` using `iniDBOptions ini defaultXFTPDBOpts` and `enableDbStoreLog'` pattern. Non-postgres build: `exitError`.
- [ ] **Step 4: Add `XSCDatabase` branch in `newXFTPServerEnv` (`Env.hs`)**
CPP-guarded pattern match on `XSCDatabase dbCfg`: `newFileStore dbCfg`, `storeLog = Nothing`.
- [ ] **Step 5: Add startup config validation**
Add `checkFileStoreMode` (CPP-guarded) before `run`: validate conflicting storeLog file + database mode, missing schema, etc. per design doc.
- [ ] **Step 6: Build both ways**
Run: `cabal build && cabal build -fserver_postgres`
- [ ] **Step 7: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Main.hs src/Simplex/FileTransfer/Server/Env.hs
git add src/Simplex/FileTransfer/Server/Main.hs src/Simplex/FileTransfer/Server/Env.hs
git commit -m "feat(xftp): add PostgreSQL INI config, store dispatch, startup validation"
```
---
## Task 9: Add database import/export CLI commands
**Files:**
- Modify: `src/Simplex/FileTransfer/Server/Main.hs`
- [ ] **Step 1: Add `Database` CLI command (CPP-guarded)**
Add `Database StoreCmd DBOpts` constructor to `CliCommand`. Add `database` subcommand parser with `import`/`export` subcommands + `dbOptsP defaultXFTPDBOpts`.
- [ ] **Step 2: Implement `importFileStoreToDatabase`**
1. `confirmOrExit` with database details.
2. Create temporary `STMFileStore`, replay StoreLog via `readWriteFileStore`.
3. Create `PostgresFileStore` with `createSchema = True`, `confirmMigrations = MCYesUp`.
4. Batch-insert files using PostgreSQL COPY protocol. Progress every 10k.
5. Batch-insert recipients using COPY protocol.
6. Verify counts: `SELECT COUNT(*)` — warn on mismatch.
7. Rename StoreLog to `.bak`.
8. Report counts.
- [ ] **Step 3: Implement `exportDatabaseToStoreLog`**
1. `confirmOrExit`. Fail if output file exists.
2. Create `PostgresFileStore` from config.
3. Open StoreLog for writing.
4. Fold over file records: write `AddFile` (with status), `AddRecipients`, `PutFile` per file.
5. Close StoreLog, report counts.
- [ ] **Step 4: Build**
Run: `cabal build -fserver_postgres`
- [ ] **Step 5: Format and commit**
```bash
fourmolu -i src/Simplex/FileTransfer/Server/Main.hs
git add src/Simplex/FileTransfer/Server/Main.hs
git commit -m "feat(xftp): add database import/export CLI commands"
```
---
## Task 10: Add Postgres tests
**Files:**
- Modify: `tests/XFTPClient.hs`
- Modify: `tests/Test.hs`
- Create: `tests/CoreTests/XFTPStoreTests.hs`
- [ ] **Step 1: Add test fixtures in `tests/XFTPClient.hs`**
```haskell
testXFTPStoreDBOpts :: DBOpts
testXFTPStoreDBOpts =
DBOpts
{ connstr = "postgresql://test_xftp_server_user@/test_xftp_server_db",
schema = "xftp_server_test",
poolSize = 10,
createSchema = True
}
```
Add `testXFTPDBConnectInfo :: ConnectInfo` matching the connection string.
- [ ] **Step 2: Add Postgres server test group in `tests/Test.hs`**
CPP-guarded block that runs existing `xftpServerTests` with Postgres store config, wrapped in `postgressBracket testXFTPDBConnectInfo`. Parameterize `withXFTPServer` to accept store config if needed.
- [ ] **Step 3: Create `tests/CoreTests/XFTPStoreTests.hs` — unit tests**
Test `PostgresFileStore` operations directly:
- `addFile` + `getFile SFSender` round-trip.
- `addFile` duplicate → `DUPLICATE_`.
- `getFile` nonexistent → `AUTH`.
- `setFilePath` + verify `WHERE file_path IS NULL` guard.
- `addRecipient` + `getFile SFRecipient` round-trip.
- `deleteFile` cascades recipients.
- `blockFile` + verify status.
- `expiredFiles` batch semantics.
- `getUsedStorage`, `getFileCount` correctness.
- [ ] **Step 4: Add migration round-trip test**
Create `STMFileStore` with test data (files + recipients + blocked status) → export to StoreLog → import to Postgres → export back → compare StoreLog files byte-for-byte.
- [ ] **Step 5: Build and run tests**
```bash
cabal build -fserver_postgres test:simplexmq-test
cabal test --test-show-details=streaming --test-option=--match="/XFTP/" -fserver_postgres
```
- [ ] **Step 6: Format and commit**
```bash
fourmolu -i tests/CoreTests/XFTPStoreTests.hs tests/XFTPClient.hs
git add tests/CoreTests/XFTPStoreTests.hs tests/XFTPClient.hs tests/Test.hs
git commit -m "test(xftp): add PostgreSQL backend tests"
```
+86 -227
View File
@@ -1,4 +1,4 @@
Revision 2, 2024-06-22
Revision 4, 2026-03-09
Evgeny Poberezkin
@@ -8,16 +8,17 @@ Evgeny Poberezkin
- [Introduction](#introduction)
- [What is SimpleX](#what-is-simplex)
- [Network model](#network-model)
- [Applications](#applications)
- [SimpleX objectives](#simplex-objectives)
- [In Comparison](#in-comparison)
- [Technical Details](#technical-details)
- [Trust in Servers](#trust-in-servers)
- [Client -> Server Communication](#client---server-communication)
- [Trust in Routers](#trust-in-routers)
- [Client -> Router Communication](#client---router-communication)
- [2-hop Onion Message Routing](#2-hop-onion-message-routing)
- [SimpleX Messaging Protocol](#simplex-messaging-protocol)
- [SimpleX Agents](#simplex-agents)
- [Encryption Primitives Used](#encryption-primitives-used)
- [Threat model](#threat-model)
- [Security](#security)
- [Acknowledgements](#acknowledgements)
@@ -27,27 +28,27 @@ Evgeny Poberezkin
SimpleX as a whole is a platform upon which applications can be built. [SimpleX Chat](https://github.com/simplex-chat/simplex-chat) is one such application that also serves as an example and reference application.
- [SimpleX Messaging Protocol](./simplex-messaging.md) (SMP) is a protocol to send messages in one direction to a recipient, relying on a server in-between. The messages are delivered via uni-directional queues created by recipients.
- SMP protocol allows to send message via a SMP server playing proxy role using 2-hop onion routing (referred to as "private routing" in messaging clients) to protect transport information of the sender (IP address and session) from the server chosen (and possibly controlled) by the recipient.
- [SimpleX Messaging Protocol](./simplex-messaging.md) (SMP) is a protocol to send messages in one direction to a recipient, relying on a router in-between. The messages are delivered via uni-directional queues created by recipients.
- SMP protocol allows to send message via a SMP router playing proxy role using 2-hop onion routing (referred to as "private routing" in messaging clients) to protect transport information of the sender (IP address and session) from the router chosen (and possibly controlled) by the recipient.
- SMP runs over a transport protocol (shown below as TLS) that provides integrity, server authentication, confidentiality, and transport channel binding.
- A SimpleX Server is one of those servers.
- A SimpleX router is one of those routers.
- The SimpleX Network is the term used for the collective of SimpleX Servers that facilitate SMP.
- The SimpleX Network is the term used for the collective of SimpleX routers that facilitate SMP.
- SimpleX Client libraries speak SMP to SimpleX Servers and provide a low-level API not generally intended to be used by applications.
- SimpleX Client libraries speak SMP to SimpleX routers and provide a low-level API not generally intended to be used by applications.
- SimpleX Agents interface with SimpleX Clients to provide a more high-level API intended to be used by applications. Typically they are embedded as libraries, but can also be abstracted into local services.
- SimpleX Agents communicate with other agents inside e2e encrypted envelopes provided by SMP protocol - the syntax and semantics of the messages exchanged by the agent are defined by [SMP agent protocol](./agent-protocol.md)
*Diagram showing the SimpleX Chat app, with logical layers of the chat application interfacing with a SimpleX Agent library, which in turn interfaces with a SimpleX Client library. The Client library in turn speaks the Messaging Protocol to a SimpleX Server.*
*Diagram showing the SimpleX Chat app, with logical layers of the chat application interfacing with a SimpleX Agent library, which in turn interfaces with a SimpleX Client library. The Client library in turn speaks the Messaging Protocol to a SimpleX router.*
```
User's Computer Internet Third-Party Server
User's Computer Internet Third-Party Router
------------------ | ---------------------- | -------------------------
| |
SimpleX Chat | |
@@ -57,11 +58,43 @@ SimpleX as a whole is a platform upon which applications can be built. [SimpleX
+----------------+ | |
| SimpleX Agent | | |
+----------------+ -------------- TLS ---------------- +----------------+
| SimpleX Client | ------ SimpleX Messaging Protocol ------> | SimpleX Server |
| SimpleX Client | ------ SimpleX Messaging Protocol ------> | SimpleX router |
+----------------+ ----------------------------------- +----------------+
| |
```
#### Network model
SimpleX is a general-purpose packet routing network built on top of the Internet. Network endpoints — end-user devices, automated services, AI-enabled applications, IoT devices — exchange data packets through SimpleX network nodes (SMP routers), which accept, buffer, and deliver packets. Each router operates independently and can be operated by any party on standard computing hardware.
SimpleX routers use resource-based addressing: each address identifies a resource on a router, similar to how the World Wide Web addresses resources via URLs. Internet routers, by comparison, use endpoint-based addressing, where IP addresses identify destination devices. Because of this design, SimpleX network participants do not need globally unique addresses to communicate.
SimpleX network has two resource-based addressing schemes:
- *Messaging queues* ([SMP](./simplex-messaging.md)). A queue is a unidirectional, ordered sequence of fixed-size data packets (16,384 bytes each). Each queue has a resource address on a specific router, gated by cryptographic credentials that separately authorize sending and receiving.
- *Data packets* ([XFTP](./xftp.md)). A data packet is an individually addressed block in one of the standard sizes. Each packet has a unique resource address on a specific router, gated by cryptographic credentials. Data packet addressing is more efficient for delivery of larger payloads than queues.
Packet delivery follows a two-router path. The sending endpoint submits a packet to a first router, which forwards it to a second router, where the receiving endpoint retrieves it. The sending endpoint's IP address is known only to the first router; the receiving endpoint's IP address is known only to the second router. See [2-hop Onion Message Routing](#2-hop-onion-message-routing) for details.
Routers buffer packets between submission and retrieval — from seconds to days, enabling asynchronous delivery when endpoints are online at different times. Packets are removed after delivery or after a configured expiration period.
#### Applications
Applications currently using SimpleX network:
- **SimpleX Chat** — a peer-to-peer messenger using SimpleX network as a transport layer, in the same way that communication applications use WebRTC, Tor, i2p, or Nym. All communication logic — contacts, conversations, groups, message formats, end-to-end encryption — runs on endpoint devices.
- **IoT devices** — using the SimpleX queue protocol directly for sensor data collection and device control.
- **AI-based services** — automated services built on the SimpleX Chat application core.
- **Secure monitoring and control systems** — applications for equipment monitoring and control, including robotics, using the network for command delivery and telemetry collection.
[SimpleGo](https://simplego.dev), developed by an independent organization, is a microcontroller-based device running a SimpleX Chat-compatible messenger directly on a microcontroller without a general-purpose operating system. Running over 20 days on a single battery charge, it demonstrates the energy efficiency of resource-based addressing: the device receives packets without continuous polling. A microcontroller-based router implementation that functions simultaneously as a WiFi router is also in development.
#### SimpleX objectives
1. Provide messaging infrastructure for distributed applications. This infrastructure needs to have the following qualities:
@@ -70,7 +103,7 @@ SimpleX as a whole is a platform upon which applications can be built. [SimpleX
- Privacy: protect against traffic correlation attacks to determine the contacts that the users communicate with.
- Reliability: the messages should be delivered even if some participating network servers or receiving clients fail, with at least once delivery guarantee.
- Reliability: the messages should be delivered even if some participating network routers or receiving clients fail, with "at least once" delivery guarantee.
- Integrity: the messages sent in one direction are ordered in a way that sender and recipient agree on; the recipient can detect when a message was removed or changed.
@@ -78,63 +111,63 @@ SimpleX as a whole is a platform upon which applications can be built. [SimpleX
- Low latency: the delay introduced by the network should not be higher than 100ms-1s in addition to the underlying TCP network latency.
2. Provide better communication security and privacy than the alternative instant messaging solutions. In particular SimpleX provides better privacy of metadata (who talks to whom and when) and better security against active network attackers and malicious servers.
2. Provide better communication security and privacy than the alternative instant messaging solutions. In particular SimpleX provides better privacy of metadata (who talks to whom and when) and better security against active network attackers and malicious routers.
3. Balance user experience with privacy requirements, prioritizing experience of mobile device users.
#### In Comparison
SimpleX network has a design similar to P2P networks, but unlike most P2P networks it consists of clients and servers without depending on any centralized component.
SimpleX network has a design similar to P2P networks, but unlike most P2P networks it consists of clients and routers without depending on any centralized component.
In comparison to more traditional messaging applications (e.g. WhatsApp, Signal, Telegram) the key differences of SimpleX network are:
- participants do not need to have globally unique addresses to communicate, instead they use redundant unidirectional (simplex) messaging queues, with a separate set of queues for each contact.
- connection requests are passed out-of-band, non-optionally protecting key exchange against man-in-the-middle attack.
- simple message queues provided by network servers are used by the clients to create more complex communication scenarios, such as duplex one-to-one communication, transmitting files, group communication without central servers, and content/communication channels.
- simple message queues provided by network routers are used by the clients to create more complex communication scenarios, such as duplex one-to-one communication, transmitting files, group communication without central routers, and content/communication channels.
- servers do not store any user information (no user profiles or contacts, or messages once they are delivered), and primarily use in-memory persistence.
- routers do not store any user information (no user profiles or contacts, or messages once they are delivered), and primarily use in-memory persistence.
- users can change servers with minimal disruption - even after an in-use server disappears, simply by changing the configuration on which servers the new queues are created.
- users can change routers with minimal disruption - even after an in-use router disappears, simply by changing the configuration on which routers the new queues are created.
## Technical Details
#### Trust in Servers
#### Trust in Routers
Clients communicate directly with servers (but not with other clients) using SimpleX Messaging Protocol (SMP) running over some transport protocol that provides integrity, server authentication, confidentiality, and transport channel binding. By default, we assume this transport protocol is TLS.
Clients communicate directly with routers (but not with other clients) using SimpleX Messaging Protocol (SMP) running over some transport protocol that provides integrity, server authentication, confidentiality, and transport channel binding. By default, we assume this transport protocol is TLS.
Users use multiple servers, and choose where to receive their messages. Accordingly, they send messages to their communication partners' chosen servers either directly, if this is a known/trusted server, or via another SMP server providing proxy functionality to protect IP address and session of the sender.
Users use multiple routers, and choose where to receive their messages. Accordingly, they send messages to their communication partners' chosen routers either directly, if this is a known/trusted router, or via another SMP router providing proxy functionality to protect IP address and session of the sender.
Although end-to-end encryption is always present, users place a degree of trust in servers they connect to. This trust decision is very similar to a user's choice of email provider; however the trust placed in a SimpleX server is significantly less. Notably, there is no re-used identifier or credential between queues on the same (or different) servers. While a user *may* re-use a transport connection to fetch messages from multiple queues, or connect to a server from the same IP address, both are choices a user may opt into to break the promise of un-correlatable queues.
Although end-to-end encryption is always present, users place a degree of trust in routers they connect to. This trust decision is very similar to a user's choice of email provider; however the trust placed in a SimpleX router is significantly less. Notably, there is no re-used identifier or credential between queues on the same (or different) routers. While a user *may* re-use a transport connection to fetch messages from multiple queues, or connect to a router from the same IP address, both are choices a user may opt into to break the promise of un-correlatable queues.
Users may trust a server because:
Users may trust a router because:
- They deploy and control the servers themselves from the available open-source code. This has the trade-offs of strong trust in the server but limited metadata obfuscation to a passive network observer. Techniques such as noise traffic, traffic mixing (incurring latency), and using an onion routing transport protocol can mitigate that.
- They deploy and control the routers themselves from the available open-source code. This has the trade-offs of strong trust in the router but limited metadata obfuscation to a passive network observer. Techniques such as noise traffic, traffic mixing (incurring latency), and using an onion routing transport protocol can mitigate that.
- They use servers from a trusted commercial provider. The more clients the provider has, the less metadata about the communication times is leaked to the network observers.
- They use routers from a trusted commercial provider. The more clients the provider has, the less metadata about the communication times is leaked to the network observers.
By default, servers do not retain access logs, and permanently delete messages and queues when requested. Messages persist only in memory until they cross a threshold of time, typically on the order of days.[0] There is still a risk that a server maliciously records all queues and messages (even though encrypted) sent via the same transport connection to gain a partial knowledge of the users communications graph and other meta-data.
By default, routers do not retain access logs, and permanently delete messages and queues when requested. Messages persist in memory or in a database until they cross a threshold of time, typically on the order of days.[0] There is still a risk that a router maliciously records all queues and messages (even though encrypted) sent via the same transport connection to gain a partial knowledge of the user's communications graph and other meta-data.
SimpleX supports measures (managed transparently to the user at the agent level) to mitigate the trust placed in servers. These include rotating the queues in use between users, noise traffic, supporting overlay networks such as Tor, and isolating traffic to different queues to different transport connections (and Tor circuits, if Tor is used).
SimpleX supports measures (managed transparently to the user at the agent level) to mitigate the trust placed in routers. These include rotating the queues in use between users, noise traffic, supporting overlay networks such as Tor, and isolating traffic to different queues to different transport connections (and Tor circuits, if Tor is used).
[0] While configurable by servers, a minimum value is enforced by the default software. SimpleX Agents can provide redundant routing over queues to mitigate against message loss.
[0] While configurable by routers, a minimum value is enforced by the default software. SimpleX Agents can provide redundant routing over queues to mitigate against message loss.
#### Client -> Server Communication
#### Client -> Router Communication
Utilizing TLS grants the SimpleX Messaging Protocol (SMP) server authentication and metadata protection to a passive network observer. But SMP does not rely on the transport protocol for message confidentiality or client authentication. The SMP protocol itself provides end-to-end confidentiality, authentication, and integrity of messages between communicating parties.
Servers have long-lived, self-signed, offline certificates whose hash is pre-shared with clients over secure channels - either provided with the client library or provided in the secure introduction between clients, as part of the server address. The offline certificate signs an online certificate used in the transport protocol handshake. [0]
Routers have long-lived, self-signed, offline certificates whose hash is pre-shared with clients over secure channels - either provided with the client library or provided in the secure introduction between clients, as part of the router address. The offline certificate signs an online certificate used in the transport protocol handshake. [0]
If the transport protocol's confidentiality is broken, incoming and outgoing messages to the server cannot be correlated by message contents. Additionally, because of encryption at the SMP layer, impersonating the server is not sufficient to pass (and therefore correlate) a message from a sender to recipient - the only attack possible is to drop the messages. Only by additionally *compromising* the server can one pass and correlate messages.
If the transport protocol's confidentiality is broken, incoming and outgoing messages to the router cannot be correlated by message contents. Additionally, because of encryption at the SMP layer, impersonating the router is not sufficient to pass (and therefore correlate) a message from a sender to recipient - the only attack possible is to drop the messages. Only by additionally *compromising* the router can one pass and correlate messages.
It's important to note that the SMP protocol does not do server authentication. Instead we rely upon the fact that an attacker who tricks the transport protocol into authenticating the server incorrectly cannot do anything with the SMP messages except drop them.
It's important to note that the SMP protocol does not do server authentication. Instead we rely upon the fact that an attacker who tricks the transport protocol into authenticating the router incorrectly cannot do anything with the SMP messages except drop them.
After the connection is established, the client sends blocks of a fixed size 16KB, and the server replies with the blocks of the same size to reduce metadata observable to a network adversary. The protocol has been designed to make traffic correlation attacks difficult, adapting ideas from Tor, remailers, and more general onion and mix networks. It does not try to replace Tor though - SimpleX servers can be deployed as onion services and SimpleX clients can communicate with servers over Tor to further improve participants privacy.
After the connection is established, the client sends blocks of a fixed size 16KB, and the router replies with the blocks of the same size to reduce metadata observable to a network adversary. The protocol has been designed to make traffic correlation attacks difficult, adapting ideas from Tor, remailers, and more general onion and mix networks. It does not try to replace Tor though - SimpleX routers can be deployed as onion services and SimpleX clients can communicate with routers over Tor to further improve participants privacy.
By using fixed-size blocks, oversized for the expected content, the vast majority of traffic is uniform in nature. When enough traffic is transiting a server simultaneously, the server acts as a low-latency mix node. We can't rely on this behavior to make a security claim, but we have engineered to take advantage of it when we can. As mentioned, this holds true even if the transport connection is compromised.
By using fixed-size blocks, oversized for the expected content, the vast majority of traffic is uniform in nature. When enough traffic is transiting a router simultaneously, the router acts as a low-latency mix node. We can't rely on this behavior to make a security claim, but we have engineered to take advantage of it when we can. As mentioned, this holds true even if the transport connection is compromised.
The protocol does not protect against attacks targeted at particular users with known identities - e.g., if the attacker wants to prove that two known users are communicating, they can achieve it by observing their local traffic. At the same time, it substantially complicates large-scale traffic correlation, making determining the real user identities much less effective.
@@ -143,39 +176,39 @@ The protocol does not protect against attacks targeted at particular users with
#### 2-hop Onion Message Routing
As SimpleX Messaging Protocol servers providing messaging queues are chosen by the recipients, in case senders connect to these servers directly the server owners (who potentially can be the recipients themselves) can learn senders' IP addresses (if Tor is not used) and which other queues on the same server are accessed by the user in the same transport connection (even if Tor is used).
As SimpleX Messaging Protocol routers providing messaging queues are chosen by the recipients, in case senders connect to these routers directly the router owners (who potentially can be the recipients themselves) can learn senders' IP addresses (if Tor is not used) and which other queues on the same router are accessed by the user in the same transport connection (even if Tor is used).
While the clients support isolating the messages sent to different queues into different transport connections (and Tor circuits), this is not practical, as it consumes additional traffic and system resources.
To mitigate this problem SimpleX Messaging Protocol servers support 2-hop onion message routing when the SMP server chosen by the sender forwards the messages to the servers chosen by the recipients, thus protecting both the senders IP addresses and sessions, even if connection isolation and Tor are not used.
To mitigate this problem SimpleX Messaging Protocol routers support 2-hop onion message routing when the SMP router chosen by the sender forwards the messages to the routers chosen by the recipients, thus protecting both the senders IP addresses and sessions, even if connection isolation and Tor are not used.
The design of 2-hop onion message routing prevents these potential attacks:
- MITM by proxy (SMP server that forwards the messages).
- MITM by proxy (SMP router that forwards the messages).
- Identification by the proxy which and how many queues the sender sends messages to (as messages are additionally e2e encrypted between the sender and the destination SMP server).
- Identification by the proxy which and how many queues the sender sends messages to (as messages are additionally e2e encrypted between the sender and the destination SMP router).
- Correlation of messages sent to different queues via the same user session (as random correlation IDs and keys are used for each message).
See more details about 2-hop onion message routing design in [SimpleX Messaging Protocol](./simplex-messaging.md#proxying-sender-commands)
Also see [Threat model](#threat-model)
Also see [Security](./security.md)
#### SimpleX Messaging Protocol
SMP is initialized with an in-person or out-of-band introduction message, where Alice provides Bob with details of a server (including IP address or host name, port, and hash of the long-lived offline certificate), a queue ID, and Alice's public keys to agree e2e encryption. These introductions are similar to the PANDA key-exchange, in that if observed, the adversary can race to establish the communication channel instead of the intended participant. [0]
SMP is initialized with an in-person or out-of-band introduction message, where Alice provides Bob with details of a router (including IP address or host name, port, and hash of the long-lived offline certificate), a queue ID, and Alice's public keys to agree e2e encryption. These introductions are similar to the PANDA key-exchange, in that if observed, the adversary can race to establish the communication channel instead of the intended participant. [0]
Because queues are uni-directional, Bob provides an identically-formatted introduction message to Alice over Alice's now-established receiving queue.
When setting up a queue, the server will create separate sender and recipient queue IDs (provided to Alice during set-up and Bob during initial connection). Additionally, during set-up Alice will perform a DH exchange with the server to agree upon a shared secret. This secret will be used to re-encrypt Bob's incoming message before Alice receives it, creating the anti-correlation property earlier-described should the transport encryption be compromised.
When setting up a queue, the router will create separate sender and recipient queue IDs (provided to Alice during set-up and Bob during initial connection). Additionally, during set-up Alice will perform a DH exchange with the router to agree upon a shared secret. This secret will be used to re-encrypt Bob's incoming message before Alice receives it, creating the anti-correlation property earlier-described should the transport encryption be compromised.
[0] Users can additionally create public 'contact queues' that are only used to receive connection requests.
[0] Users can additionally create public 'contact queues' that are only used to receive connection requests.
#### SimpleX Agents
SimpleX agents provide higher-level operations compared to SimpleX Clients, who are primarily concerned with creating queues and communicating with servers using SMP. Agent operations include:
SimpleX agents provide higher-level operations compared to SimpleX Clients, who are primarily concerned with creating queues and communicating with routers using SMP. Agent operations include:
- Managing sets of bi-directional, redundant queues for communication partners
@@ -186,195 +219,21 @@ SimpleX agents provide higher-level operations compared to SimpleX Clients, who
- Noise traffic
#### Encryption Primitives Used
## Security
- Ed25519 or Curve25519 to authorize/verify commands to SMP servers (authorization algorithm is set via client/server configuration).
- Curve25519 for DH exchange to agree:
- the shared secret between server and recipient (to encrypt message bodies - it avoids shared cipher-text in sender and recipient traffic)
- the shared secret between sender and recipient (to encrypt messages end-to-end in each queue - it avoids shared cipher-text in redundant queues).
- [NaCl crypto_box](https://nacl.cr.yp.to/box.html) encryption scheme (curve25519xsalsa20poly1305) for message body encryption between server and recipient and for E2E per-queue encryption.
- SHA256 to validate server offline certificates.
- [double ratchet](https://signal.org/docs/specifications/doubleratchet/) protocol for end-to-end message encryption between the agents:
- Curve448 keys to agree shared secrets required for double ratchet initialization (using [X3DH](https://signal.org/docs/specifications/x3dh/) key agreement with 2 ephemeral keys for each side),
- AES-GCM AEAD cipher,
- SHA512-based HKDF for key derivation.
For encryption primitives, threat model, and detailed security analysis, see [Security](./security.md).
SimpleX provides these security properties:
## Threat Model
- **End-to-end encryption** using Double Ratchet algorithm with forward secrecy and post-quantum cryptography.
#### Global Assumptions
- **No shared identifiers** across connections — contacts cannot prove they communicate with the same user.
- A user protects their local database and key material.
- The user's application is authentic, and no local malware is running.
- The cryptographic primitives in use are not broken.
- A user's choice of servers is not directly tied to their identity or otherwise represents distinguishing information about the user.
- The user's client uses 2-hop onion message routing.
- **Sender deniability** — neither routers nor recipients can cryptographically prove message origin.
#### A passive adversary able to monitor the traffic of one user
- **Transport metadata protection** — fixed-size blocks, 2-hop onion routing, and optional connection isolation frustrate traffic correlation.
*can:*
- identify that and when a user is using SimpleX.
- determine which servers the user receives the messages from.
- observe how much traffic is being sent, and make guesses as to its purpose.
*cannot:*
- see who sends messages to the user and who the user sends the messages to.
- determine the servers used by users' contacts.
#### A passive adversary able to monitor a set of senders and recipients
*can:*
- identify who and when is using SimpleX.
- learn which SimpleX Messaging Protocol servers are used as receive queues for which users.
- learn when messages are sent and received.
- perform traffic correlation attacks against senders and recipients and correlate senders and recipients within the monitored set, frustrated by the number of users on the servers.
- observe how much traffic is being sent, and make guesses as to its purpose
*cannot, even in case of a compromised transport protocol:*
- perform traffic correlation attacks with any increase in efficiency over a non-compromised transport protocol
#### SimpleX Messaging Protocol server
*can:*
- learn when a queue recipient is online
- know how many messages are sent via the queue (although some may be noise or not content messages).
- learn which messages would trigger notifications even if a user does not use [push notifications](./push-notifications.md).
- perform the correlation of the queue used to receive messages (matching multiple queues to a single user) via either a re-used transport connection, user's IP Address, or connection timing regularities.
- learn a recipient's IP address, track them through other IP addresses they use to access the same queue, and infer information (e.g. employer) based on the IP addresses, as long as Tor is not used.
- drop all future messages inserted into a queue, detectable only over other, redundant queues.
- lie about the state of a queue to the recipient and/or to the sender (e.g. suspended or deleted when it is not).
- spam a user with invalid messages.
*cannot:*
- undetectably add, duplicate, or corrupt individual messages.
- undetectably drop individual messages, so long as a subsequent message is delivered.
- learn the contents or type of messages.
- distinguish noise messages from regular messages except via timing regularities.
- compromise the users' end-to-end encryption with an active attack.
- learn a sender's IP address, track them through other IP addresses they use to access the same queue, and infer information (e.g. employer) based on the IP addresses, even if Tor is not used (provided messages are sent via proxy SMP server).
- perform senders' queue correlation (matching multiple queues to a single sender) via either a re-used transport connection, user's IP Address, or connection timing regularities, unless it has additional information from the proxy SMP server (provided messages are sent via proxy SMP server).
#### SimpleX Messaging Protocol server that proxies the messages to another SMP server
*can:*
- learn a sender's IP address, as long as Tor is not used.
- learn when a sender with a given IP address is online.
- know how many messages are sent from a given IP address and to a given destination SMP server.
- drop all messages from a given IP address or to a given destination server.
- unless destination SMP server detects repeated public DH keys of senders, replay messages to a destination server within a single session, causing either duplicate message delivery (which will be detected and ignored by the receiving clients), or, when receiving client is not connected to SMP server, exhausting capacity of destination queues used within the session.
*cannot:*
- perform queue correlation (matching multiple queues to a single user), unless it has additional information from the destination SMP server.
- undetectably add, duplicate, or corrupt individual messages.
- undetectably drop individual messages, so long as a subsequent message is delivered.
- learn the contents or type of messages.
- learn which messages would trigger notifications.
- learn the destination queues of messages.
- distinguish noise messages from regular messages except via timing regularities.
- compromise the user's end-to-end encryption with another user via an active attack.
- compromise the user's end-to-end encryption with the destination SMP servers via an active attack.
#### An attacker who obtained Alice's (decrypted) chat database
*can:*
- see the history of all messages exchanged by Alice with her communication partners.
- see shared profiles of contacts and groups.
- surreptitiously receive new messages sent to Alice via existing queues; until communication queues are rotated or the Double-Ratchet advances forward.
- prevent Alice from receiving all new messages sent to her - either surreptitiously by emptying the queues regularly or overtly by deleting them.
- send messages from the user to their contacts; recipients will detect it as soon as the user sends the next message, because the previous message hash wont match (and potentially wont be able to decrypt them in case they dont keep the previous ratchet keys).
*cannot:*
- impersonate a sender and send messages to the user whose database was stolen. Doing so requires also compromising the server (to place the message in the queue, that is possible until the Double-Ratchet advances forward) or the user's device at a subsequent time (to place the message in the database).
- undetectably communicate at the same time as Alice with her contacts. Doing so would result in the contact getting different messages with repeated IDs.
- undetectably monitor message queues in realtime without alerting the user they are doing so, as a second subscription request unsubscribes the first and notifies the second.
#### A users contact
*can:*
- spam the user with messages.
- forever retain messages from the user.
*cannot:*
- cryptographically prove to a third-party that a message came from a user (assuming the users device is not seized).
- prove that two contacts they have is the same user.
- cannot collaborate with another of the user's contacts to confirm they are communicating with the same user.
#### An attacker who observes Alice showing an introduction message to Bob
*can:*
- Impersonate Bob to Alice.
*cannot:*
- Impersonate Alice to Bob.
#### An attacker with Internet access
*can:*
- Denial of Service SimpleX messaging servers.
- spam a user's public “contact queue” with connection requests.
*cannot:*
- send messages to a user who they are not connected with.
- enumerate queues on a SimpleX server.
- **Out-of-band key exchange** — connection requests passed outside the network protect against MITM attacks.
## Acknowledgements
+215
View File
@@ -0,0 +1,215 @@
Revision 1, 2026-03-09
# SimpleX Network: Security
This document describes the cryptographic primitives and threat model for the SimpleX network. For a general introduction, see [SimpleX: messaging and application platform](./overview-tjr.md).
## Table of contents
- [Encryption primitives](#encryption-primitives)
- [Threat model](#threat-model)
- [Global Assumptions](#global-assumptions)
- [A passive adversary able to monitor the traffic of one user](#a-passive-adversary-able-to-monitor-the-traffic-of-one-user)
- [A passive adversary able to monitor a set of senders and recipients](#a-passive-adversary-able-to-monitor-a-set-of-senders-and-recipients)
- [SimpleX Messaging Protocol router](#simplex-messaging-protocol-router)
- [SimpleX Messaging Protocol router that proxies the messages to another SMP router](#simplex-messaging-protocol-router-that-proxies-the-messages-to-another-smp-router)
- [An attacker who obtained Alice's (decrypted) chat database](#an-attacker-who-obtained-alices-decrypted-chat-database)
- [A user's contact](#a-users-contact)
- [An attacker who observes Alice showing an introduction message to Bob](#an-attacker-who-observes-alice-showing-an-introduction-message-to-bob)
- [An attacker with Internet access](#an-attacker-with-internet-access)
## Encryption primitives
- **Router command authorization**: X25519 DH-based authenticated encryption (SMP v7+), providing sender deniability. Ed25519 signatures used for recipient commands and notifier commands.
- **Per-queue key agreement**: Curve25519 DH exchange to agree:
- the shared secret between router and recipient (to encrypt message bodies — avoids shared ciphertext in sender and recipient traffic),
- the shared secret between sender and recipient (to encrypt messages end-to-end in each queue — avoids shared ciphertext in redundant queues).
- **SMP-layer encryption**: [NaCl crypto_box](https://nacl.cr.yp.to/box.html) (curve25519xsalsa20poly1305) for message body encryption between router and recipient, and for e2e per-queue encryption.
- **Certificate validation**: SHA256 to validate router offline certificates.
- **End-to-end encryption**: [Double ratchet](https://signal.org/docs/specifications/doubleratchet/) protocol:
- Curve448 keys for shared secret agreement via [X3DH](https://signal.org/docs/specifications/x3dh/) with 2 ephemeral keys per side,
- optional [SNTRUP761](https://ntruprime.cr.yp.to/) post-quantum KEM running in parallel with the DH ratchet (see [PQDR](./pqdr.md)), providing post-quantum forward secrecy,
- AES-GCM AEAD cipher,
- SHA512-based HKDF for key derivation.
## Threat Model
### Global Assumptions
- A user protects their local database and key material.
- The user's application is authentic, and no local malware is running.
- The cryptographic primitives in use are not broken.
- A user's choice of routers is not directly tied to their identity or otherwise represents distinguishing information about the user.
- The user's client uses 2-hop onion message routing.
### A passive adversary able to monitor the traffic of one user
*can:*
- identify that and when a user is using SimpleX.
- determine which routers the user receives messages from.
- observe how much traffic is being sent, and make guesses as to its purpose.
*cannot:*
- see who sends messages to the user and who the user sends messages to.
- determine the routers used by users' contacts.
### A passive adversary able to monitor a set of senders and recipients
*can:*
- identify who and when is using SimpleX.
- learn which SimpleX Messaging Protocol routers are used as receive queues for which users.
- learn when messages are sent and received.
- perform traffic correlation attacks against senders and recipients and correlate senders and recipients within the monitored set, frustrated by the number of users on the routers.
- observe how much traffic is being sent, and make guesses as to its purpose.
*cannot, even in case of a compromised transport protocol:*
- perform traffic correlation attacks with any increase in efficiency over a non-compromised transport protocol.
### SimpleX Messaging Protocol router
*can:*
- learn when a queue recipient is online.
- know how many messages are sent via the queue (although some may be noise or not content messages).
- learn which messages would trigger notifications even if a user does not use [push notifications](./push-notifications.md).
- perform the correlation of the queue used to receive messages (matching multiple queues to a single user) via either a re-used transport connection, user's IP Address, or connection timing regularities.
- learn a recipient's IP address, track them through other IP addresses they use to access the same queue, and infer information (e.g. employer) based on the IP addresses, as long as Tor is not used.
- drop all future messages inserted into a queue, detectable only over other, redundant queues.
- lie about the state of a queue to the recipient and/or to the sender (e.g. suspended or deleted when it is not).
- spam a user with invalid messages.
*cannot:*
- undetectably add, duplicate, or corrupt individual messages.
- undetectably drop individual messages, so long as a subsequent message is delivered.
- learn the contents or type of messages.
- distinguish noise messages from regular messages except via timing regularities.
- compromise the users' end-to-end encryption with an active attack.
- learn a sender's IP address, track them through other IP addresses they use to access the same queue, and infer information (e.g. employer) based on the IP addresses, even if Tor is not used (provided messages are sent via proxy SMP router).
- perform senders' queue correlation (matching multiple queues to a single sender) via either a re-used transport connection, user's IP Address, or connection timing regularities, unless it has additional information from the proxy SMP router (provided messages are sent via proxy SMP router).
### SimpleX Messaging Protocol router that proxies the messages to another SMP router
*can:*
- learn a sender's IP address, as long as Tor is not used.
- learn when a sender with a given IP address is online.
- know how many messages are sent from a given IP address and to a given destination SMP router.
- drop all messages from a given IP address or to a given destination router.
- unless destination SMP router detects repeated public DH keys of senders, replay messages to a destination router within a single session, causing either duplicate message delivery (which will be detected and ignored by the receiving clients), or, when receiving client is not connected to SMP router, exhausting capacity of destination queues used within the session.
*cannot:*
- perform queue correlation (matching multiple queues to a single user), unless it has additional information from the destination SMP router.
- undetectably add, duplicate, or corrupt individual messages.
- undetectably drop individual messages, so long as a subsequent message is delivered.
- learn the contents or type of messages.
- learn which messages would trigger notifications.
- learn the destination queues of messages.
- distinguish noise messages from regular messages except via timing regularities.
- compromise the user's end-to-end encryption with another user via an active attack.
- compromise the user's end-to-end encryption with the destination SMP routers via an active attack.
### An attacker who obtained Alice's (decrypted) chat database
*can:*
- see the history of all messages exchanged by Alice with her communication partners.
- see shared profiles of contacts and groups.
- surreptitiously receive new messages sent to Alice via existing queues; until communication queues are rotated or the Double-Ratchet advances forward.
- prevent Alice from receiving all new messages sent to her - either surreptitiously by emptying the queues regularly or overtly by deleting them.
- send messages from the user to their contacts; recipients will detect it as soon as the user sends the next message, because the previous message hash won't match (and potentially won't be able to decrypt them in case they don't keep the previous ratchet keys).
*cannot:*
- impersonate a sender and send messages to the user whose database was stolen. Doing so requires also compromising the router (to place the message in the queue, that is possible until the Double-Ratchet advances forward) or the user's device at a subsequent time (to place the message in the database).
- undetectably communicate at the same time as Alice with her contacts. Doing so would result in the contact getting different messages with repeated IDs.
- undetectably monitor message queues in realtime without alerting the user they are doing so, as a second subscription request unsubscribes the first and notifies the first.
### A user's contact
*can:*
- spam the user with messages.
- forever retain messages from the user.
*cannot:*
- cryptographically prove to a third-party that a message came from a user (assuming the user's device is not seized).
- prove that two contacts they have is the same user.
- cannot collaborate with another of the user's contacts to confirm they are communicating with the same user.
### An attacker who observes Alice showing an introduction message to Bob
*can:*
- Impersonate Bob to Alice.
*cannot:*
- Impersonate Alice to Bob.
### An attacker with Internet access
*can:*
- Denial of Service SimpleX messaging routers.
- spam a user's public "contact queue" with connection requests.
*cannot:*
- send messages to a user who they are not connected with.
- enumerate queues on a SimpleX router.
+290
View File
@@ -0,0 +1,290 @@
# Protocol changes for creating and connecting to SMP queues
## Problems
This change is related to these problems:
- differentiating queue retention time,
- supporting MITM-resistant short connection links,
This RFC is based on the previous discussions about short links, blob storage and notifications ([1](./2024-06-21-short-links.md), [2](./2024-09-09-smp-blobs.md), [3](./2024-11-25-queue-blobs-2.md), [4](./2024-09-25-ios-notifications-2.md)).
SMP protocol supports two types of queues - queues to send messages (messaging queues) and queues to send invitations to connect (contact queues). While SMP protocol was originally "unaware" of these queue types, it could differentiate it by message flow, and with the recent addition of SKEY command to allow securing the queue by the sender this difference became persistent.
Simply designating queue types would allow to use this information to decide for how long to retain queues, and potentially extending it:
- unsecured 1-time invitation queues with sndSecure (support of securing by sender) - e.g., 3 months.
- contact address queues without sndSecure - e.g., 3 years without activity.
- Possibly, "queues" that prohibit messages and used only as blob storage - they would be used to store group profiles and super-peer addresses for the group.
## Design objectives
We want to achieve these objectives for short links and associated queue data:
1. no possibility to provide incorrect SenderId inside link data (e.g. from another queue).
2. link data cannot be accessed by the server unless it has the link.
3. prevent MITM attack by the server, including the server that obtained the link.
4. prevent changing of connection request by the user (to prevent MITM via break-in attack in the originating client).
5. for one-time links, prevent accessing link data by link observers who did not compromise the server.
6. allow changing the user-defined part of link data.
7. avoid changing the link when user-defined part of link data changes, while preventing MITM attack by the server on user-defined part, even if it has the link.
8. retain the quality that it is impossible to check the existence of secured queue from having any of its temporary visible IDs (sender ID and link ID in 1-time invitations) - it requires that these IDs remain server-generated (contrary to the previous RFCs).
To achieve these objectives the queue data will include fixed (immutable) and user-defined (mutable) parts.
Fixed part would include:
- full connection request (the current long link with all keys, including PQ keys). This includes SenderId that must match server response.
- public signature key to verify mutable part of link data.
Signed mutable part would include:
- any links to chat relays that should be contacted instead of this queue (not in this RFC), to allow delegating group connections and contact request connections to prevent spam, hiding online presence, etc.
- and user-defined data - user profile or group profile, chat preferences, welcome message, etc.
The link itself should include both the key and auth tag from the encryption of immutable part. Accessing one-time link data should require providing sender key and signing the command (`LKEY`).
## Solution
Current NEW and NKEY commands:
```haskell
NEW :: RcvPublicAuthKey -> RcvPublicDhKey -> Maybe BasicAuth -> SubscriptionMode -> SenderCanSecure -> Command Recipient
-- | Queue IDs and keys, returned in IDS response
data QueueIdsKeys = QIK
{ rcvId :: RecipientId,
sndId :: SenderId,
rcvPublicDhKey :: RcvPublicDhKey,
sndSecure :: SenderCanSecure
}
```
Proposed NEW command replaces SenderCanSecure with QueueMode, adds link data, and combines NKEY command:
```haskell
NEW :: NewQueueRequest -> Command Recipient
data NewQueueReq = NewQueueReq
{ rcvAuthKey :: RcvPublicAuthKey,
rcvDhKey :: RcvPublicDhKey,
auth_ :: Maybe BasicAuth,
subMode :: SubscriptionMode,
queueReqData :: Maybe QueueReqData,
}
-- QRMessaging implies that sender can secure the queue.
-- LinkId is not used with QRMessaging, to prevent the possibility of checking when connection is established by re-using the same link ID when creating another queue the creating would have to fail if it is used.
-- LinkId is required with QRContact, to have shorter link - it will be derived from the link_uri. And in this case we do not need to prevent checks that this queue exists.
data QueueReqData
= QRMessaging (Maybe (SenderId, QueueLinkData))
| QRContact (Maybe (LinkId, (SenderId, QueueLinkData)))
-- SenderId should be computed client-side as the first 24 bytes of sha3-384(correlation_id),
-- The server must verify it and reject if it is not.
-- It allows to include sender ID inside encrypted associated link data as part of full connection URI without requesting it from the server, but prevents checking if a given sender ID exists (queue creation would fail for a duplicate sender ID), as sha3-384 derivation is not reversible.
type QueueLinkData = (EncFixedLinkData, EncUserDataBytes)
type EncFixedLinkData = ByteString
type EncUserDataBytes = ByteString
-- We need to use binary encoding for ConnectionRequestUri to reduce its size
-- The clients would reject changed immutable data and
-- ConnectionRequestUri where server or SenderId of the queue do not match.
data FixedLinkData c = FixedLinkData
{ agentVRange :: VersionRangeSMPA,
rootKey :: C.PublicKeyEd25519,
connReq :: ConnectionRequestUri c
}
data ConnLinkData c where
InvitationLinkData :: VersionRangeSMPA -> UserLinkData -> ConnLinkData 'CMInvitation
ContactLinkData ::
{ agentVRange :: VersionRangeSMPA,
-- direct connection via connReq in fixed data is allowed.
direct :: Bool,
-- additional owner keys to sign changes of mutable data.
owners :: [OwnerAuth],
-- alternative addresses of chat relays that receive requests for this contact address.
relays :: [ConnShortLink 'CMContact],
userData :: UserLinkData
} -> ConnLinkData 'CMContact
newtype UserLinkData = UserLinkData ByteString
-- | Updated queue IDs and keys, returned in IDS response
data QueueIdsKeys = QIK
{ rcvId :: RecipientId, -- server-generated
sndId :: SenderId, -- server-generated
rcvPublicDhKey :: RcvPublicDhKey,
sndSecure :: SenderCanSecure, -- possibly, can be removed? or implied?
linkId :: Maybe LinkId -- server-generated
}
```
In addition to that we add the command allowing to update and also to retrieve and secure the queue and get link data in one request, to have only one request:
```haskell
-- This command allows to set all data or to update mutable part of contact address queue.
-- This command should fail on queues that support sndSecure and also on new queues created with QRMessaging.
-- This should fail if LinkId or immutable part of data is changed with the update, but will succeed if only mutable part is updated, so it can be retried.
-- Entity ID is RecipientId.
-- The response to this command is `OK`.
LSET :: LinkId -> QueueLinkData -> Command Recipient
-- Delete should link and associated data
-- Entity ID is RecipientId
LDEL :: Command Recipient
-- To be used with 1-time links.
-- Sender's key provided on the first request prevents observers from undetectably accessing 1-time link data.
-- If queue mode is QRContact (and queue does NOT allow sndSecure) the command will fail, same as SKEY.
-- Once queue is secured, the key must be the same in subsequent requests - to allow retries in case of network failures, and to prevent passive attacks.
-- The difference with securing queues is that queues allow sending unsecured messages to queues that allow sndSecure (for backwards compatibility), and 1-time links will NOT allow retrieving link data without securing the queue at the same time, preventing undetected access by observers.
-- Entity ID is LinkId
LKEY :: SndPublicAuthKey -> Command Sender
-- If queue mode is QRMessaging the command will fail.
-- Entity ID is LinkId
LGET :: Command Sender
-- Response to LGET and LSET
-- Entity ID is the same as in the command
LNK :: SenderId -> QueueLinkData -> BrokerMsg
```
To both include sender_id into the full link before the server response, and to prevent "oracle attack" when a failure to create the queue with the supplied `sender_id` can be used as a proof of queue existence, it is proposed that `sender_id` is computed client-side as the first 24 bytes of 48 in `sha3-384(correlation_id)` and validated server-side, where `corelation_id` is the transmission correlation ID.
To allow retries, every time the command is sent a new random `correlation_id` and new `sender_id` (and for contact queue, also `link_id`, which would be random as it is derived from hash of fixed link data that includes a random signature key) should be used on each attempt, because other IDs would be generated randomly on the server, and in case the previous command succeeded on the server but failed to be communicated to the client, the retry will fail if the same ID is used.
Alternative solutions that would allow retries that were considered and rejected:
- additional request to save queue data, after `sender_id` is returned by the server. The scenarios that require short links are interactive - creating user addresses and 1-time invitations - so making two requests instead of one would make the UX worse.
- include empty sender_id in the immutable data and have it replaced by the accepting party with `sender_id` received in `LINK` response - both a weird design, and might create possibility for some attacks via server, especially for contact addresses.
- making NEW commands idempotent. Doing it would require generating all IDs client-side, not only `sender_id`. It increases complexity, and it is not really necessary as the only scenarios when retries are needed are async NEW commands, that do not require short links. For future short links of chat relays the retries are much less likely, as chat relays will have good network connections.
## Algorithm to prepare and to interpret queue link data.
For contact addresses this approach follows the design proposed in [Short links](./2024-06-21-short-links.md) RFC - when link id is derived from the same random binary as key. For 1-time invitations link ID is independent and server-generated, to prevent existence checks (oracle attack).
This scheme results in 32 byte binary size for contact addresses and 56 bytes for 1-time invitation links.
For fixed link data.
1. Generate random `nonce` (also used as a correlation ID for server command) and signature key (public `rootKey` included in fixed data).
2. Compute sender ID from `nonce` as the first 24 bytes of sha3-384 of `nonce`.
3. Generate other keys for queue address, including queue e2e encryption keys and double ratchet connection e2e encryption keys.
4. Construct the full connection address to be included in fixed data.
5. `link_key = SHA3-256(fixed_data)` - used as part of the link, and to derive the key to encrypt content.
6. HKDF:
1) contact address: `(link_id, key) = HKDF(link_key, 56 bytes)`.
2) 1-time invitation: `key = HKDF(link_key, 32 bytes)`, `link-id` - server-generated.
7. Encrypt: `(ct1, tag1) = secret_box(fixed_data, key, nonce1)`, where `nonce1` is a random nonce
5. Store: `(nonce1, ct1, tag1)` stored as fixed link data.
For mutable user data:
1. Random `nonce2` and the same key are used.
2. Sign `user_data` with key included in `fixed_data`.
3. Encrypt: `(ct2, tag2) = secret_box(signed_used_data, key, nonce2)`.
4. Store: `(nonce2, ct2, tag2)`
Link recipient:
1. Receives `link_key` in the link, for 1-time invitations also `link_id`.
2. HKDF:
1) contact address: `(link_id, key) = HKDF(link_key, 56 bytes)`.
2) 1-time invitation: `key = HKDF(link_key, 32 bytes)`.
3. Retrieves via `link_id`: `(nonce1, ct1, tag1)` and `(nonce2, ct2, tag2)`:
1) contact address: `LGET` command, that allows retrieving link data multiple times.
2) 1-time invitation: `LKEY` command, that non-optionally secures the queue, and only allows repeated link data retrievals if the same sender's key is provided and signs the transmission. This prevents link data retrieval by link observers.
4. Decrypt: `(signature1, fixed_data) = decrypt (nonce1, ct1, tag1)`.
5. Verify: `SHA3-256(fixed_data) == link_key`, abort if not.
6. Decrypt: `(signature2, used_data) = decrypt(nonce2, ct2, tag2)`.
7. Verify signatures using key in the fixed data, abort if they don't match.
While using content hash as encryption key is unconventional, it is not completely unusual - e.g., it is used in convergent encryption (although in our case using random nonce makes it not convergent, but other use cases suggest that this approach preserves encryption security). It is particularly acceptable for our use case, as `fixed_data` contains mostly random keys.
## Threat model
**Compromised SMP server**
can:
- delete link data.
- hide link data selectively for some or for all requests.
cannot:
- undetectably replace link data, even if it has the link (objective 3).
- access unencrypted link data, whether it was or was not accessed by the accepting party, provided it has no link (objective 2).
- observe IP addresses of the users accessing link data, if private routing is used.
**Passive observer who observed short link**:
can:
- access original unencrypted link data for contact address links.
cannot:
- undetectably access observed 1-time link data, accessing the link would make the link inaccessible to the sender (objective 5).
- undetectably check the existence of messaging queue or 1-time link (objective 8).
- replace or delete the link data.
**Queue owner who did not compromise the server**:
cannot:
- redirect connecting user to another queue, on the same or on another server (objective 1).
- replace connection request in the link (objective 4).
## Correlation of design objectives with design elements
1. The presence of `SenderId` in `LNK` response from the server.
2. Encryption of link data with crypto_box.
3. Deriving encryption key from the hash of fixed data prevents it being modified by the server - any change would be detected and rejected by the client, as the hash of fixed data won't match the link. Signature verification with the key from fixed data, and signing of mutable data prevents server modification of mutable data.
4. No server command to change fixed data once it's set. Also, changing fixed data would require changing the link.
5. 1-time link data can only be accessed with `LKEY` command, that while allows retries to mitigate network failures, will require the same key for retries.
6. `LSET` command.
7. The link is derived from fixed data only, so it does not change when mutable link data changes. Mutable part is signed preventing server MITM attacks.
8. SenderId is derived from request correlation ID, so it cannot be arbitrary defined to check existence of some known queue. LinkId for 1-time invitation is generated server-side, so it cannot be provided by the client when creating the queues to check if these IDs are used.
## Syntax for short links
The syntax:
```abnf
shortConnectionLink = uriAuthority "/" linkUri [ "?" param *( "&" param ) ]
uriAuthority = %s"https://" smpServerHost / "simplex:" ; using simplex: scheme requires including host in the parameter hostParam
smpServerHost = <hostname> ; RFC1123, RFC5891
linkUri = %s"i#" oneTimeLink / contactType "#" contactLink
contactType = %s"a" / %s"g" / %s"c" ; contact / group / channel address, respectively
oneTimeLink = <base64url(linkId)> "/" <base64url(linkKey)> ; 56 bytes / 75 base64 encoded characters
contactLink = <base64url(linkKey)> ; 32 bytes / 43 base64 encoded characters
; linkId - 192 bits/24 bytes
; linkKey - 256 bits/32 bytes
param = hostsParam / portParam / certHashParam
hostsParam = %s"h=" host *("," host) ; additional hostnames, e.g. onion
portParam = %s"p=" 1*DIGIT ; server port
certHashParam = %s"c=" <base64url(server offline certificate fingerprint)>
```
To have shorter links fingerprint and additional server hostnames do not need to be specified for pre-configured servers, even if they are disabled - they can be used from the client code. Any user defined servers will require including additional hosts and server fingerprint.
Example one-time link for preset server (104 characters):
```
https://smp12.simplex.im/i#abcdefghij0123456789abcdefghij01/23456789abcdefghij0123456789abcdefghij01234
```
Example contact link for preset server (71 characters):
```
https://smp12.simplex.im/c#abcdefghij0123456789abcdefghij0123456789abc
```
Example contact link for user-defined server (with fingerprint, but without onion hostname - 117 characters):
```
https://smp1.example.com/c#abcdefghij0123456789abcdefghij0123456789abc?c=0YuTwO05YJWS8rkjn9eLJDjQhFKvIYd8d4xG8X1blIU
```
Example contact link for user-defined server (with fingerprint ant onion hostname - 182 characters):
```
https://smp1.example.com/c#abcdefghij0123456789abcdefghij0123456789abc?c=0YuTwO05YJWS8rkjn9eLJDjQhFKvIYd8d4xG8X1blIU&h=beccx4yfxxbvyhqypaavemqurytl6hozr47wfc7uuecacjqdvwpw2xid.onion
```
For the links to work in the browser the servers must provide server pages.
+93
View File
@@ -0,0 +1,93 @@
# New notifications protocol
## Problem
iOS notifications have these problems:
- iOS notification service crashes exceeding memory limit. This is being addressed by changes in GHC RTS.
- there is a large number of connections, because each member in a group requires individual connection. This will improve with chat relays when each group would require 2-3 connections.
- some notification may be not shown if notification with reply/mention is skipped, and instead some other message is delivered, which may be muted. This would not improve without some changes, as notifications may be skipped anyway.
- client devices delay communication with ntf server because it is done in background, and by that time the app may be suspended.
- notification server represents a bottleneck, as it has to be owned by the app vendor, and the current design when ntf server subscribes to notifications scales very badly.
This RFC is based on the previous [RFC related to notifications](./2024-09-25-ios-notifications-2.md).
## Solution
As notification server has to know client token and currently it associates subscriptions with this token anyway, we are not gaining any privacy and security by using per-subscription keys - both authorization and encryption keys of notification subscription can be dropped.
We still need to store the list of queue IDs associated with the token on the notification server, but we do not need any per-queue keys on the notification server, and we don't need subscriptions - it's effectively a simple set of IDs, with no other information.
In this case, when queue is created the client would supply notifier ID - it has to be derived from correlation ID, to prevent existense check (see previous RFC). As we also supply sender ID, instead of deriving it as sha3-192 of correlation ID, they both can be derived as sha3-384 and split to two IDs - 24 bytes each.
The notification server will maintain a rotating list of server keys with the latest key communicated to the client every time the token is registered and checked. The keys would expire after, say, 1 week or 1 month, and removed from notification server on expiration.
The packet containing association between notifier queue ID and token will be crypto_box encrypted using key agreement between identified notification server master key and an ephemeral per packet (effectively, per-queue) client-key.
Deleting the queue may also include encrypted packet that would verify that the client deleted the queue.
Instead of notification server subscribing to the notifications creating a lot of traffic for the queues without messages, the SMP server would push notifications via NTF server connection (whether via NTF or via SMP protocol). This could be used as a mechanism to migrate existing queues when with the next subscription the notification server would communicate it's address to SMP server and this association would be stored together with the queue.
## Protocol design
Additional/changed SMP commands:
```haskell
-- register notification server
-- should be signed with server key
NSRV :: NtfServerCreds -> Command NtfServer
-- response
NSID :: NtfServerId -> BrokerMsg
-- to communicate which server is responsible for the queue
-- should be signed with queue key
NSUB :: Maybe NtfServerId -> Command Notifier
-- subscribe to notificaions from all queues associated with the server
-- should be signed with server key
-- entity ID - NtfServerId
NSSUB :: Command NtfServer
data NtfServerCreds = NtfServerCreds
{ server :: NtfServer,
-- NTF server certificate chain that should match fingerpring in address
cert :: X.CertificateChain,
-- server autorizatio key to sign server subscription requests
authKey :: X.SignedExact X.PubKey
}
-- entity ID is recipient ID
NSKEY :: NtfSubscription -> Command Recipient
data NtfSubscription = NtfSubscription
-- key to encrypt notifications e2e with the client
{ ntfPubDbKey :: RcvNtfPublicDhKey,
ntfServer :: NtfServer,
-- should be linked to correlation ID to prevent existense check
-- the ID sent to notification server could be its hash?
ntfId :: NotifierId,
encNtfTokenAssoc :: EncDataBytes
}
-- before the encryption - equivalent to NSUB command, but without key to authorize requests to specific queue
data NtfTokenAssoc = NtfTokenAssoc
{ signature :: SignatureEd25519,
tknId :: NtfTokenId,
ntfQueue :: SMPQueueNtf
}
```
SMP server will need to maintain the list of Ntf servers and their credentials, and when NSSUB arrives to make only one subscription. When message arrives it would deliver notification to the correct connection via queue / ntf server association.
Ntf server needs to maintain three indices to the same data:
- `(smpServer, queueId) -> tokenId` - to deliver notification to the correct token
- `tokenId -> [smpServer -> [queueId]]` - to remove all queues when token is removed, and to store/update these associations effficiently - store log may have one compact line per token (after compacting), or per token/server combination.
- `[smpServer]` - array of SMP servers to subscribe to.
## Mention notifications
Currently we are marking messages with T (true) for messages that require notifications and F (false) for messages that don't require. Sender does not know whether the recipient has notifications disabled, enabled or in mentions-only mode.
The proposal is to:
- add additional values to this metadata, e.g. 2 (priority) and 3 (high priority) (and T/F could be sent as 0/1 respectively) - that is, to deliver notifications even if notifications are generally disabled (they can still be further filtered by the client).
- instead of deleting notification credentials when notifications are disabled - which is costly - communicate to SMP server the change of notificaion priority level, e.g. the client could set minimal notification priority to deliver notifications, where 0 would mean disabling it completely, 1 enable for all, 2 for priority 2+, 3 for priority 3. The downside here is that it could be used for timing correlation of queues in the group, but it already can be used on bulk deletions of ntf credentials for these queues and when sending messages.
+159
View File
@@ -0,0 +1,159 @@
# Using short links as group links
## Problem
To use the short links for groups these problems has to be / can be solved:
1. recognizing link as a group link.
2. permanent link with the ability to change chat relays.
3. binding owners signatures to the link.
4. allowing to add/remove owners, both to share ownership and for reliability in case of one owner losing keys/access.
While current short links solve problems 1-3 (via contact type, and via extension of user data in the link), the problem 4 is solved only partially.
We could include the current list of root owners in the user data, and we could send any history of ownership changes from this baseline as a short blockchain on joining the group, we still requrie one master owner to retain access to the queue associated with the group.
## Possible solution approaches
1. "Kick this can down the road" - ignore this problem until there is a namespace, and a group name can be associated with multiple queues.
Pros: simple and reasonable, and it suggests postponing multisig for owners too. The users can still see the list of owners and their keys in user data of the link, and receive admin roster signed by owners on joining.
Cons: if this "master owner" loses the access to the device, no further changes to group profile will be possible.
2. The queue access can be shared by sharing the key and recipient IDs with all owners.
The problems:
- preventing MITM attack between owners (this protect exists for other solutions too).
- protecting these credentials from chat relays. So somehow there should be direct key agreement between members allowing to send e2e encrypted message inaccessible to chat relays.
Pros: simpler than alternatives, and still provides protection against losing the key.
Cons:
- quite clunky, and requires the new primitive anyway (e2e encryption).
- no multisig
This could possibly be evolved into the requirement to have a direct connection with other owners, and verifying the security code before they have access to group.
3. Allow "joint management" of SMP queues.
SMP servers can support multiple recipients for contact queues:\
- subscription would be possible to the "subscriber recipient".
- all other changes (update data, change subscriber recipient, add or remove recipients) would require multiple recipient signatures on SMP command in line with n-of-m multisig rules, that the command sender would have to collect out-of-band (from SMP protocol point of view).
Pros: allows joint ownership, and protects from losing access to master owner device.
Cons:
- complicates queue abstraction with approach that is not needed for most queues.
- still retains the server as a single point of failure.
4. Introduce "group" as a new type of entity managed by SMP servers.
SMP servers would provide a separate set of commands for managing group records that would include in an encrypted container:
- the group profile
- the list of chat relay links
- the list of owner member IDs with their public keys
- multisig rules
- alternative group entity locations
- possibly, a globally unique group identity (as the hash of the initial/seed group data).
While the server domain would be used as the hostname in group link, it may contain alternative hosts (not just hostnames of the same server), both in the link and in the group record data.
Pros: separates additional complexity to where it is needed, allowing reliability and redundancy for group ownership.
Cons: complexity, coupling between SMP and chat protocol.
## Design for channel/group as a separate queue mode
Option 1.
A queue mode "channel" when owners are represented by their individual queues (either a separate mode, or a submode of "channel", or just normal contact address queues). In this case sending message to channel queue would broadcast message to queue owners, without exposing even the number of owners.
Pros:
- allows chat relays to send messages to all owners (e.g., channel can be secured with the list of snd keys, one per relay).
- quite easy to evolve from the current design.
- extensible.
Cons:
- close to "solution in search of a problem".
- does not require data model changes - channel queue would simply have a list of owner "recipient IDs", and each owner queue would also point to channel.
Option 2.
Also a separate queue mode "channel", but instead of having a linked owner queues, it would simply maintain a list of owner keys to maintain the data. In this case, messages cannot be sent to this "queue" at all.
Pros:
- simpler design.
- we could allow sending messages to it too, with the "main" owner receiving them. This could be negotiated in the protocol.
- it may be easier to migrate the current groups, as the admin link would be this queue (although for public groups in directory it would have to be recreated anyway).
- Possibly, when queue is created there should be a flag whether it should accept unsigned messages - then contact addresses would be created with unsigned messages ON, messages queues, once SKEY is universally supported, with unsigned messages OFF, and channel queues with unsigned messages OFF too for new public queues.
Cons:
- if no messages are accepted, this is not even a queue.
- no way to directly contact owners (maybe it is not a downside, as for relays there would be a communication channel anyway as part of the group).
Option 2 looks more simple and attractive, implementing server broadcast for SMP seems unnecessary, as while it could have been used for simple groups, it does not solve such problems as spam and pre-moderation anyway - it requires a higher level protocol.
The command to update owner keys would be `RKEY` with the list of keys, and we can make `NEW` accept multiple keys too, although the use case here is less clear.
## Multiple owners managing queue data.
Option 1: Use the same keys in SMP as when signing queue data.
Option 2: Use different keys.
The value here could be that the server could validate these signatures too, and also maintain the chain of key changes. While tempting, it is probably unnecessary, and this chain of ownership is better to be maintained on chat relay level, as there are no size constraints on the size of this chain. Also, it is better for metadata privacy to not couple transport and chat protocol keys.
We still need to bind the mutable data updates to the "genesis" signature key (the one included in the immutable data).
The proposed design:
- when mutable data is signed by genesis key, then it is bound, and no changes is needed.
- mutable data may be signed by the key of the new owner, in which case mutable part itself must contain the binding. We could also use ring signature to sign the mutable data, concealing which owner signed the data - that would increase the signature size from 64 bytes to `32 * (n + 1)` bytes.
Current mutable data:
```haskell
data UserLinkData = UserLinkData
{ agentVRange :: VersionRangeSMPA,
userData :: ConnInfo
}
```
Proposed mutable data:
```haskell
data UserLinkData = UserLinkData
{ agentVRange :: VersionRangeSMPA,
owners :: [OwnerInfo]
userData :: ConnInfo
}
type OwnerId = ByteString
data OwnerInfo = OwnerInfo
{ ownerId :: OwnerId, -- unique in the list, application specific - e.g., MemberId
ownerKey :: PublicKeyEd25519,
-- owner signature of sender ID,
-- confirms that the owner agreed with being the owner,
-- prevents a member being added as an owner without consent.
ownerSig :: SignatureEd25519,
-- owner authorization, sig(ownerId || ownerKey, prevKey), where prevKey is either a "genesis key" or some other key previously signed by the genesis key.
authOwnerId :: OwnerId, -- null for "genesis"
authOwnerSig :: SignatureEd25519
}
```
The size of the OwnerInfo record encoding is:
- ownerId: 1 + 12
- ownerKey: 1 + 32
- ownerSig: 1 + 64
- ownerAuthId: 1 + 12
- ownerAuthSig: 1 + 64
~189 bytes, so we should practically limit the number of owners to say 8 - 1 original + 7 addiitonal. Original creator could use a different key as a "genesis" key, to conceal creator identity from other members, and it needs to include the record with memberId anyway.
The structure is simplified, and it does not allow arbitrary ownership changes. Its purpose is not to comprehensively manage ownership changes - while it is possible with a generic blockchain, it seems not appropriate at this stage, - but rather to ensure access continuity and that the server cannot modify the data (although nothing prevents the server from removing the data completely or from serving the previous version of the data).
For example it would only allow any given owner to remove subsequenty added owners, preserving the group link and identity, but it won't allow removing owners that signed this owner authorization. So owners are not equal, with the creator having the highest rank and being able to remove all additional owners, and owners authorise by creator can remove all other owners but themselves and creator, and so on - they have to maintain the chain that authorized themselves, at least. We could explicitely include owner rank into OwnerInfo, or we could require that they are sorted by rank, or the rank can be simply derived from signatures.
When additional owners want to be added to the group, they would have to provide any of the current owners:
- the key for SMP commands authorization - this will be passed to SMP server together with other keys. There could be either RKEY to pass all keys (some risk to miss some, or of race conditions), or RADD/RGET/RDEL to add and remove recipient keys, which has no risk of race conditions.
- the signature of the immutable data by their member key included in their profile.
- the current owner would then include their member key into the queue data, and update it with LSET command. In any case there should be some simple consensus protocol between owners for owner changes, and it has to be maintained as a blockchain by owners and by chat relays, as otherwise it may lead to race conditions with LSET command.
Potentially, there could be one command to update keys and link data, so that they are consistent.
+142
View File
@@ -0,0 +1,142 @@
# Service certificates for high volume servers and services connecting to SMP servers
## Problem
The absense of user and client identification benefits privacy, but it requires separately authorizing subscription for each messaging queue, that doesn't scale when a high volume server or service acts as a client for SMP server even for the current traffic and network size.
These servers/services include:
- operators' chat relays (aka super-peers),
- notification servers,
- high-traffic service chat bots,
- high-traffic business support clients.
The future chat relays would reduce the number of subscriptions required for the usual clients, by replacing connections with each group member to 1-3 connections with chat relays per group/community, it would shift the burden to the chat relays, that are also clients.
Self-hosted chat relays may want to retain privacy, so they will not use client certificates, but this privacy is not needed (and counter-productive) for the chat relays provided by network operators.
Even today, directory service subscribing to all queues may take 15-20 minutes, which is experienced as downtime by the end users.
Notification servers also acting as clients to messaging servers also take 15-20 minutes to subscribe to all notifications, during which time notifications are not delivered.
Not only these subscription take a lot of time, they also consume a large amount of memory both in the clients and in the servers, as association between clients and queues is currently session-scoped and not persisted anywhere (and it should not be, because end-users' clients do need privacy).
## Solution
High volume "clients" (operators' chat relays, directory service, SimpleX Chat team support client, SimpleX Status bot, etc.) that don't need privacy will identify themselves to the messaging servers at a point of connection by providing client sertificate, both in TLS handshake and in SMP handshake (the same certificate must be provided).
All the new queues and subscriptions made in this session will be creating a permanent association of the messaging queue with the client, and on subsequent reconnections the client can "subscribe" to all their queues with a single client subscription command.
This will save a lot of time subscribing and resubscribing on server and client restarts, servers' bandwidth, servers' traffic spikes, and memory of both clients and servers.
## Protocol
An ephemeral per-session signature key signed by long-term client certificate is used for client authorization this session signature key will be passed in SMP handshake.
To transition existing queues, the subscription command will have to be double-signed - by the queue key, and then by client key.
When server receives such "hand-over" subscription it would create a permanent association between the client certificate and the queue, and on subsequent re-connections the client can subscribe to all the existing queues still associated with the client with one command.
The server will respond to the client with the number of queues it was subscribed to - it would both inform the client that it has to re-connect in case of interruption, and can be used for client and server statistics.
When client creates a new queue, it would also sign the request with both keys, per-queue and client's. Other queue operations (e.g., deletion, or changing associated queue data for short links) would still require two signatures, both the queue key and the client key.
The open question is whether there is any value in allowing to remove the association between the client and the queue. Probably not, as threat model should assume that the server would retain this information, and the use-case for users controlling their servers is narrow.
## Protocol connection handshake
Currently, the types for handshakes are:
```haskell
data ServerHandshake = ServerHandshake
{ smpVersionRange :: VersionRangeSMP,
sessionId :: SessionId,
-- pub key to agree shared secrets for command authorization and entity ID encryption.
-- todo C.PublicKeyX25519
authPubKey :: Maybe (X.CertificateChain, X.SignedExact X.PubKey)
}
data ClientHandshake = ClientHandshake
{ -- | agreed SMP server protocol version
smpVersion :: VersionSMP,
-- | server identity - CA certificate fingerprint
keyHash :: C.KeyHash,
-- | pub key to agree shared secret for entity ID encryption, shared secret for command authorization is agreed using per-queue keys.
authPubKey :: Maybe C.PublicKeyX25519,
-- | Whether connecting client is a proxy server (send from SMP v12).
-- This property, if True, disables additional transport encrytion inside TLS.
-- (Proxy server connection already has additional encryption, so this layer is not needed there).
proxyServer :: Bool
}
```
`ServerHandshake` already contains `authPubKey` with the server certificate chain and the signed key for connection encryption and creating a shared secret for denable authorization (with client entity key) and session encryption layer.
`ClientHandshake` contains only ephemeral `authPubKey` to compute a shared secret for session encryption layer, so we need an additional field for an optional client certificate:
```haskell
serviceCertKey :: Maybe (X.CertificateChain, X.SignedExact X.PubKey)
```
Certificate here defines client identity. The actual key to be used to sign commands is session-scoped, and is signed by the certificate key. In case of notification server it MUST be the same certificate that is used for server TLS connections.
For operators' clients we may optionally include operators' certificate in the chain, and that would allow servers to identify operators if either wants to. This would improve end-user security, as not only the server would validate that its certificate matches the address, but it would also validate that it is operated by SimpleX Chat or by Flux, preventing any server impersonation (e.g., via DNS manipulations) - the client could then report that the files are hosted on SimpleX Chat servers, but then can stop and show additional warning in case certificate does not match the domain - same as the browsers do with CA stores in the client.
## Protocol transmissions
Each transport block can contain one or several protocol transmissions.
Each transmission has this structure:
```abnf
transmission = authenticator authorized
; authenticator - Ed25519 signature for recipients or X25519 authenticator for senders, to provide repudiation.
; authenticator authorizes the rest of the transmission.
authorized = sessId corrId entityId command.
; sessId is tls-unique channel binding, its presense in the transmission prevents replay attacks.
```
The proposed change would replace authenticator with exactly one or two authenticators, where the first one will remain resource-level authorization (queue key), and the optional second one will be client authorization with the client key.
```abnf
authenticator = queue_authenticator ("0" / "1" service_authenticator)
; "0" and "1" characters (digit characters, not x00 or x01) are conventionally used for Maybe types in the protocol.
```
In case service_authenticator is present, queue_authenticator should authorize over `fingerprint authorized` (concatenation of service identity certificate fingerprint and the rest of the transmission).
All queues created with client key will have to be double-authorized with both the queue key and the client key - both the client and the server would have to maintain this knowledge, whether the queue is associated with the client or not.
Asymmetric retries have to be supported - the first request creating this association may succeed on the server and timeout on the client.
## Subscription
To subscribe to all associated queues the client has to send a single command authorized with the client key passed in handshake.
The command and response:
```haskell
SUBS :: Command Recipient -- to enable all client subscriptions, empty entity ID in the transmission, signed by client key - it must be the same as was used in handover subscription signature.
NSUBS :: Command Recipient -- notification subscription
SOK :: Maybe ServiceId -- new subscription response
SOKS :: Int64 -> BrokerMsg -- response from the server, includes the number of subscribed queues
ENDS :: Int64 -> BrokerMsg -- when another session subscribes with the same certificate
```
Open questions:
- What should used as an entity ID for `SUBS` transmission - certificate fingerprint or an empty string?
- Should there be a command to get the list of all associated queues? It is likely to be useful for debugging?
- What should happen when `SUB` is sent for a single already associated queue? What if it is signed with the correct session key, but that is different from existing association? The current approach is that once associated, this associaiton would require authorization for single subscriptions, with the same certificate as already associated.
## Ephemeral client-session association
This was considered to reduce costs for the usual clients to re-subscribe. Currently it's a big problem, because of groups, and with transition to chat relays it won't be.
For some very busy end-user clients it may help.
Given that server has access to an ephemeral association between recipient client session and queues anyway (even with clients connecting via Tor, unless per-connection transport isolation is used), introducing `sessionPubKey` to allow resubscription to the previously subscribed queues may reduce the traffic. This won't change threat model as the server would only keep this association in memory, and not persist it. Clients on another hand may safely persist this association for fast resubscription on client restarts.
This is not planned for the forseable future, as migrating to chat relays would solve most of the problem.
Assuming an average active user has 20 contacts and 20 groups, and they would need ~3 subscriptions for each (for redundancy), so about 120 subscription to reconnect. The single 16kb transport block allows to send ~136 subscriptions. Which means that ephemeral sessions would create no value for clients at all, unless they are super active.
Further, improving transport efficiency for super-active non-identified clients may help network abuse, so ephemeral sessions may have negative value.
+104
View File
@@ -0,0 +1,104 @@
# Using the same profile from multiple devices
## Problem
Double Ratchet algorithm makes it hard to send/receive messages sent to the user from different devices, as each message changes the state of Double Ratchet keys, and these state changes must be strictly sequential and they cannot be reversed (although skipping is possible).
Traditional approach for multi-device converts each direct conversation into a group, where each device participates as a member. Likewise, for group conversations each device also participates as a member. While these members *look* as if they are the same user to others, a very simple client app modification may show device ID for each message, and the communication peers, both in direct chats and in groups would know how many devices a user has and which device the user sent the message from. In addition to that, with this approach communication peers can send different messages to different devices (it can be prevented by provider who would request that only message key is encrypted with DR, while the encrypted message is the same) or withheld from some devices (it cannot be prevented by provider, as it cannot add key to the communication in case it is missing, and cannot withhold the message completely too). These opens various vectors for targeted attacks, e.g.:
- tracking movements of the user: once each devices is identified as "desk" and "phone" it would allow to know where the user is at a given time.
- manipulating information by sending messages to one device (to have proof it was sent) and withholding from others, or sending different messages if the protocol allows it.
In addition to that, the specific implementation of this approach in Signal compromises break-in recovery property (aka post-compromise security) of Double-Ratchet algorithm, making its design ineffective - the only reason to have the second ratchet in DR algorithm is to provide break-in recovery, without it a much simpler design with a single ratchet is sufficient. See [this paper](https://eprint.iacr.org/2021/626.pdf) for details.
While this limitation can be addressed with notifications when a new device is added and per-device keys, we still find the remaining attack vectors on user security and privacy to be unacceptable, and opening unsuspecting users to various criminal actions - and it is wrong to say that would only affect security conscious users, and most people would not be affected by these risks. Allowing potential criminals in groups to know which device you are currently using is a real risk for all users.
Another approach was offered by Threema that is ["mediator" server](https://threema.com/en/blog/md-architectural-overview) where the state of encryption ratchets is stored server-side. While it protects the user from their communication peers, it increases required level of trust to the servers, and in case of SimpleX network it would expose the knowledge of who communicates to whom. So while the idea of server-side storage of encryption state is promising, it has to be per-connection, to retain "no-accounts" property of SimpleX messaging network.
Also see [FAQ](https://simplex.chat/faq/#why-cant-i-use-the-same-profile-on-different-devices) and [this issue](https://github.com/simplex-chat/simplex-chat/issues/444#issuecomment-3066968358).
## Proposed solution
One of the ideas presented in FAQ - to store the state of Double Ratchet algorithm in the encrypted container on the server seems promising. The RFC develops this idea.
### Considerations for the design
1. The largest ratchet state size with the current implementation is less than 8kb (which is achieved when both sides shared PQ keys and ciphertexts), so while it cannot fit in the same transport blocks together with sent and received messages, it would fit in one transport block.
2. Protocol commands and events may be changed (even if at the cost of slightly reducing message size) can fit the hash of the ratchet state (32 bytes sha256 would be sufficient), so that the client can determine whether it has the most recent ratchet state or if it needs to retrieve the latest copy. Message size reduction won't affect the users because we use compression, and there is a substantial reserve.
3. Client commands that modify ratchet state would include the hash of the previous ratchet state so that the server can reject or ignore the command in case the previous ratchet state is different or in case command is repeated in case of lost response).
4. The client does not need to retrieve message state for each encryption and decryption operation - it can "speculatively" use the ratchet state it has, and receive correct ratchet state in the "error" response after attempting encryption based on incorrect ratchet state.
## Proposed protocol design
Ratchet state will be stored on the same server that stores message queue, as part of message queue record. 8kb is a sufficient size for this blob (the actual max size is 7800 bytes). The server would also store the hashes of the current and, possibly, the previous ratchet states (TBC).
While ratchet is used for duplex connection, the connection still has primary queue, and with redundancy the same ratchet state can be stored on all secondary queues.
Ratchet state will be encrypted using secret_box - a symmetric encryption scheme, so PQ-resistant. If ratchet state is stored on more than one server, it has to be encrypted with a different key for each server.
Questions: how to rotate the key used to store ratchet? Should key used to encrypt ratchet rotate at the same time when queue is rotated? The latter is a logical option, as it prevents additional complexity and solves the problem anyway. A possible option is to have "ratchet version" that will be used to advance the key used to encrypt ratchet via HKDF.
Security considerations: the scheme may reduce break-in recovery to the points queues are rotated, unless there is some randomness mixed-in into the key derivation (the key used to encrypt ratchet state). But including randomness would defeat the purpose, as other devices wouldn't be able to access the ratchets. Another approach would be to have each device use its own key for encryption, and encrypt to all keys of all devices (or to encrypt key, to avoid size increase). Having multiple encryptions would show how many devices use the queue, but servers already can observe it, so it is a better tradeoff. Another idea would be to rotate the key used to authorize queue commands - we already support multiple recipient keys, and it can be used for multi-device scenario. That would partially mitigate break-in attacks as the attacker who obtained the key from ratchet state would be able to decrypt it, but won't be able to decrypt it (the attacker collusion with the server is not mitigated). Yet another idea would be for each party (device) to share its private (or encapsulation) key and to have a symmetric key (used to encrypt the ratchet state) encrypted (encapsulated) separately for each device. This would reduce the size of the stored data to `ratchet size` + `encrypted key size` * N, so even in case of PQ encryption (e.g. sntrup) the size required to store the ratchet would be under transport block size, while limiting it to say 4-8 devices, which is sufficient.
To participate in multi-device scheme the devices would join the usual group that will be used to share public (encapsulation) device keys and to communicate updates to conversations that were received by the currently "active" device. "Active" means the device that received or sent and processed the message, and while only one device can receive messages from a given queue, device "active" state may be determined per queue, allowing concurrent usage.
The scheme must be resilient to state updates being lost, and in case of direct messages it would result in some messages not being shown (or shown as skipped), while conversation preference and profile updates can be re-requested from peers, while the current profile of the user would become the latest. Likewise, for groups state updates ca be requested from super-peers or for decentralized groups - from owners. Maintaining chat state consistency is an important consideration, but is not a focus of this RFC - the focus is managing message delivery and DR encryption for multiple devices. Other multi-device schemes have the same issues with state consistency. Partially, the profile state consistency can be improved by using a single shared queue (or set of queues) to store user's profile and chat preferences to synchronize profile updates asynchronously between the devices.
## The protocol to send the message
`rsi` - ratchet state on device `i`.
`enc(rs)` - current authoritative ratchet state on the server.
`pt` and `ct` - plaintext and ciphertext messages.
Encryption is a state transition function ratchetEnc: `(ct, rs') = ratchetEnc(pt, rs)`.
1. Device encrypts the message using the stored ratchet state: `(ct, rsi') = ratchetEnc(pt, rsi)`
2. Device sends modified encrypted ratchet state and the hash of the previous encrypted state to the server that stores the queue: `RSET (hash(enc(rsi)), enc(rsi'))`.
3. If the hash of the previous state matches state stored on the server (`hash(enc(rsi)) == hash(enc(rs))`), the server updates the state and responds with `ratchet_ok` (that may include the current state or it's hash, for validation). If the hash is different, the server responds with `bad_ratchet(enc(rs))` message that includes the correct ratchet state. These updates must be atomic. In this case device has to update the local ratchet state (provided it can decrypt it), and repeat encryption attempt. If device cannot decrypt the provided ratchet state, it means that the connection is disrupted (possibly, device is removed from device group, but missed the notifications).
4. After successful state update in primary receiving queue, the device would update it in secondary receiving queues.
5. Device sends encrypted message as usual, via proxy that must be different both from the server that stores the ratchet and from the destination server.
6. Device broadcasts sent message and new ratchet state to other devices in the device group.
This protocol is simple, and it minimizes requests when sending the message to one additional request to update ratchet state in most cases, only requiring two requests when device state was not updated via device group prior to message sending attempt.
## The protocol to receive the message
Decryption is also a state transition function: `(pt, rs') = ratchetDec(ct, rs)`
1. Server sends the message to the device (can be in response to SUB or ACK commands, or with active subscription). Pushed message would include the hash of the currently stored ratchet state: `hash(enc(rs))`.
2. If device has the ratchet state with the same hash (`hash(enc(rs)) == hash(enc(rsi))`), it decrypts the message: `(pt, rsi') = ratchetDec(ct, rsi)`.
3. If device has ratchet state with a different hash, it requests ratchet from the server with additional protocol command `RGET` with response `RCHT (enc(rs))` and updates the local state.
4. Device decrypts the message `(pt, rsi') = ratchetDec(ct, rsi)` and processes it as usual.
5. Device sends acknowledgement to the server as usual, but now it includes the new ratchet state and the hash of the previous state: `ACK msgId (hash(enc(rsi)), enc(rsi'))`
6. The server compares ratchet state with stored state hash, and in case it matches it processes `ACK` and responds with `OK` as usual (or `NO_MSG` in case msgId is incorrect, also as usual - it would happen in repeated ACK requests). If ratchet state hash does not match, the server would respond with `bad_ratchet(enc(rs))` - which means that the message was already processed by another device and ratchet was advanced. This is a complex scenario, as the client has to either revert the change from message processing or somehow combine the change with the updates communicated via device group (as a side note, device group can simply re-broadcast messages, not state updates, but it will result in state divergence between devices when different messages are lost).
Unlike sending messages, this flow does not require any additional requests in most cases, only requiring requesting message state reconciliation when the same message was received and processed by more than one client, but it does not require re-acknowledgement.
## Challenges
This is an idea of the design rather than the actual design, as it requires more thinking about:
- how to handle concurrent ratchet state updates,
- "active" status transitions per queue,
- avoiding concurrent subscriptions to queues from multiple devices,
- state updates and synchronization between devices,
- handling skipped messages,
- costs to update ratchets in bulk send scenario - this scheme would substantially increase costs of preparing large broadcasts, and it makes this scheme not acceptable for chat relays. Which means that "profile" on desktop used as chat relay won't be synched to other devices.
- etc.
## Advantages
The communication peers won't know how many devices the user has, and which device was used to send the message. Also, the communication peers won't be able to send different messages to different user's devices, or to withhold messages from some devices.
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,154 @@
# XFTP Server: SNI, CORS, and Web Support
Implementation details for Phase 3 of `rfcs/2026-01-30-send-file-page.md` (sections 6.1-6.4).
## 1. Overview
The XFTP server is extended to support web browser clients by:
1. **SNI-based TLS certificate switching** — Present a CA-issued web certificate (e.g., Let's Encrypt) to browsers, while continuing to present the self-signed XFTP identity certificate to native clients.
2. **CORS headers** — Add CORS response headers on SNI connections so browsers allow cross-origin XFTP requests.
3. **Configuration**`[WEB]` INI section for HTTPS cert/key paths; opt-in (commented out by default).
Web handshake (challenge-response identity proof, §6.3 of parent RFC) is not yet implemented and will be added separately.
## 2. SNI Certificate Switching
### 2.1 Reusing the SMP Pattern
The SMP server already implements SNI-based certificate switching via `TLSServerCredential` and `runTransportServerState_` (see `rfcs/2024-09-15-shared-port.md`). The XFTP server applies the same pattern with one key difference: both native and web XFTP clients use HTTP/2 transport, whereas SMP switches between raw SMP protocol and HTTP entirely.
### 2.2 Approach
When `httpServerCreds` is configured, the XFTP server bypasses `runHTTP2Server` and uses `runTransportServerState_` directly to obtain the per-connection `sniUsed` flag. It then sets up HTTP/2 manually on each TLS connection using `withHTTP2` (same internals as `runHTTP2ServerWith_`). The `sniUsed` flag is captured in the closure and shared by all HTTP/2 requests on that connection.
When `httpServerCreds` is absent, the existing `runHTTP2Server` path is unchanged.
```
Native client (no SNI) ──TLS──> XFTP identity cert ──HTTP/2──> processRequest (no CORS)
Browser client (SNI) ──TLS──> Web CA cert ──HTTP/2──> processRequest (+ CORS)
```
### 2.3 Certificate Chain
The web certificate file (e.g., `web.crt`) must contain the full chain: leaf certificate followed by the signing CA certificate. `loadServerCredential` uses `T.credentialLoadX509Chain` which reads all PEM blocks from the file.
The client validates the chain by comparing `idCert` fingerprint (the CA cert, second in the 2-cert chain) against the known `keyHash`. This is the same validation as for XFTP identity certificates — the CA that signed the web cert must match the XFTP server's identity.
## 3. CORS Support
### 3.1 Design
CORS headers are only added when both conditions are true:
- `addCORSHeaders` is `True` in `TransportServerConfig` (set in XFTP `Main.hs`)
- `sniUsed` is `True` for the current TLS connection
This ensures native clients never see CORS headers.
### 3.2 Response Headers
All POST responses on SNI connections include:
```
Access-Control-Allow-Origin: *
Access-Control-Expose-Headers: *
```
### 3.3 OPTIONS Preflight
OPTIONS requests are intercepted at the HTTP/2 dispatch level, before `processRequest`. This is necessary because `processRequest` rejects bodies that don't match `xftpBlockSize`.
Preflight response:
```
HTTP/2 200
Access-Control-Allow-Origin: *
Access-Control-Allow-Methods: POST, OPTIONS
Access-Control-Allow-Headers: *
Access-Control-Max-Age: 86400
```
### 3.4 Security
`Access-Control-Allow-Origin: *` is safe because:
- All XFTP commands require Ed25519 authentication (per-chunk keys from file description).
- No cookies or browser credentials are involved.
- File content is end-to-end encrypted.
## 4. Configuration
### 4.1 INI Template
```ini
[WEB]
# cert: /etc/opt/simplex-xftp/web.crt
# key: /etc/opt/simplex-xftp/web.key
```
Commented out by default — web support is opt-in.
### 4.2 Behavior
- `[WEB]` section not configured: silently ignored, server operates normally for native clients only.
- `[WEB]` section configured with valid cert/key paths: SNI + CORS enabled.
- `[WEB]` section configured with missing cert files: warning + continue (non-fatal, unlike SMP where it is fatal).
## 5. Files Modified
### 5.1 `src/Simplex/Messaging/Transport/Server.hs`
Added `addCORSHeaders :: Bool` field to `TransportServerConfig`. Updated `mkTransportServerConfig` to accept the new parameter. All existing SMP call sites pass `False`.
### 5.2 `src/Simplex/Messaging/Transport/HTTP2/Server.hs`
- Extracted `expireInactiveClient` from `runHTTP2ServerWith_`'s `where` clause to a module-level function.
- Parameterized `runHTTP2ServerWith_`: setup type changed from `((TLS p -> IO ()) -> a)` to `(((Bool, TLS p) -> IO ()) -> a)`, callback from `HTTP2ServerFunc` to `Bool -> HTTP2ServerFunc`. The `Bool` is the per-connection `sniUsed` flag, threaded through `H.run` to the callback.
- Extended `runHTTP2Server` with `Maybe T.Credential` parameter for SNI web certificate. Its setup uses `runTransportServerState_` with `TLSServerCredential`, which naturally provides `(sniUsed, tls)` pairs matching the new `runHTTP2ServerWith_` setup type.
- Adapted `runHTTP2ServerWith` (client-side HTTP/2, no SNI): wraps its setup to inject `(False, tls)` and its callback with `const`.
- Updated `getHTTP2Server` (test helper) to pass `Nothing` for httpCreds.
### 5.3 `src/Simplex/FileTransfer/Server/Env.hs`
- Added `httpCredentials :: Maybe ServerCredentials` to `XFTPServerConfig`.
- Added `httpServerCreds :: Maybe T.Credential` to `XFTPEnv`.
- `newXFTPServerEnv` loads HTTP credentials when configured.
### 5.4 `src/Simplex/FileTransfer/Server/Main.hs`
- Added `[WEB]` section to INI template.
- Added `httpCredentials` parsing from INI `[WEB]` section (`cert` and `key` fields).
- Set `addCORSHeaders = isJust httpCredentials_` in transport config (conditional on web cert presence).
### 5.5 `src/Simplex/FileTransfer/Server.hs`
Core server changes:
- `runServer` calls `runHTTP2Server` with `httpCreds_` and a `\sniUsed -> handleRequest (sniUsed && addCORSHeaders transportConfig)` callback. TLS params are `defaultSupportedParamsHTTPS` when web creds present, `defaultSupportedParams` otherwise. SNI routing, HTTP/2 setup, and client expiration are handled inside `runHTTP2Server`.
- `XFTPTransportRequest` carries `addCORS :: Bool` field, threaded through to `sendXFTPResponse`.
- `sendXFTPResponse` conditionally includes CORS headers based on `addCORS`.
- OPTIONS requests on SNI connections return CORS preflight headers before reaching `processRequest`.
- Helper functions: `corsHeaders` (response headers), `corsPreflightHeaders` (preflight headers).
### 5.6 `tests/XFTPClient.hs`
- Added `httpCredentials = Nothing` to `testXFTPServerConfig`.
- Added `testXFTPServerConfigSNI` with web cert config and `addCORSHeaders = True`.
- Added `withXFTPServerSNI` helper.
### 5.7 `tests/XFTPServerTests.hs`
Added SNI and CORS tests as a subsection within `xftpServerTests` (6 tests):
1. **SNI cert selection** — Connect with SNI + `h2` ALPN, verify RSA web certificate is presented.
2. **Non-SNI cert selection** — Connect without SNI + `xftp/1` ALPN, verify Ed448 XFTP certificate is presented.
3. **CORS headers** — SNI POST request includes `Access-Control-Allow-Origin: *` and `Access-Control-Expose-Headers: *`.
4. **OPTIONS preflight** — SNI OPTIONS request returns all CORS preflight headers.
5. **No CORS without SNI** — Non-SNI POST request has no CORS headers.
6. **File chunk delivery** — Full XFTP file chunk upload/download through SNI-enabled server verifying no regression.
## 6. Remaining Work
- **Web handshake** (§6.3 of parent RFC): Challenge-response identity proof for SNI connections. The server detects web clients via the `sniUsed` flag and expects a 32-byte challenge in the first POST body (non-empty, unlike standard handshake). Response includes full cert chain + signature over `(challenge ++ sessionId)`.
- **Static page serving** (§6.5 of parent RFC): Optional serving of the web page HTML/JS bundle on GET requests.
@@ -0,0 +1,246 @@
# Web Handshake — Challenge-Response Identity Proof
RFC §6.3: Server proves XFTP identity to web clients independently of TLS CA infrastructure.
## 1. Protocol
**Standard handshake** (unchanged):
```
Client → empty POST → Server
Server → padded {vRange, sessionId, authPubKey, Nothing} → Client
Client → padded {version, keyHash, Nothing} → Server
Server → empty → Client
```
**Web handshake** (SNI connection, non-empty hello):
```
Client → padded {32 random bytes} → Server
Server → padded {vRange, sessionId, authPubKey, Just sigBytes} → Client
sigBytes = signatureBytes(sign(identityLeafKey, challenge <> sessionId))
Client validates:
1. chainIdCaCerts(authPubKey.certChain) → CCValid {leafCert, idCert}
2. SHA-256(idCert) == keyHash (server identity)
3. verify(leafCert.pubKey, sigBytes, challenge <> sessionId) (challenge-response)
4. verify(leafCert.pubKey, signedPubKey.signature, signedPubKey.objectDer) (DH key auth)
Client → padded {version, keyHash, Just challenge} → Server
Server verifies: echoed challenge == stored challenge from step 1
Server → empty → Client
```
**Detection**: `sniUsed` per-connection flag. Non-empty hello allowed only when `sniUsed`. Empty hello with SNI → standard handshake.
**Why both steps 3 and 4**: Native clients verify `signedPubKey` using the TLS peer certificate (`serverKey` from `getServerVerifyKey`), which is the XFTP identity cert in non-SNI connections — TLS provides this binding. Web clients cannot access TLS peer certificate data (browser API limitation; TLS presents the web CA cert but provides no API to extract it). So web clients must verify at the application layer using `authPubKey.certChain`, which always contains the XFTP identity chain regardless of which cert TLS used. Step 3 proves the server holds its identity key *right now* (freshness via random challenge). Step 4 proves the DH session key was signed by the identity key holder (prevents MITM key substitution). Together they give web clients some assurance native clients get from TLS, except channel binding for commands.
## 2. Type Changes — `src/Simplex/FileTransfer/Transport.hs`
### `XFTPServerHandshake` (line 114)
Add field: `webIdentityProof :: Maybe ByteString` — raw Ed448 signature bytes (114 bytes), or `Nothing` for standard handshake. No record needed — the cert chain is already in `authPubKey.certChain`.
### `Encoding XFTPServerHandshake` (line 136)
- `smpEncode`: append `smpEncode webIdentityProof`
- `smpP`: `Tail compat`, if non-empty `eitherToMaybe $ smpDecode compat`
Backward compat: old clients ignore via `Tail _compat`; new client + old server → empty compat → `Nothing`.
### `XFTPClientHandshake` (line 121)
Add field: `webChallenge :: Maybe ByteString`
### `Encoding XFTPClientHandshake` (line 128)
Same `Tail compat` pattern as server handshake.
### Export list
Both types use `(..)` export — new fields auto-exported.
## 3. Server Changes — `src/Simplex/FileTransfer/Server.hs`
### `XFTPTransportRequest` (line 88)
Add field: `sniUsed :: SNICredentialUsed` (`Bool` from `Transport.Server`). Add import.
### `Handshake` (line 117)
`HandshakeSent C.PrivateKeyX25519``HandshakeSent C.PrivateKeyX25519 (Maybe ByteString)` — stores 32-byte web challenge or `Nothing`.
### `runServer` handler (line 145161)
- Pass `sniUsed` into request construction (line 154)
- SNI-first routing: when `sniUsed`, always route to `xftpServerHandshakeV1` (web ALPN `h2` would otherwise fall to `_` catch-all)
### `xftpServerHandshakeV1` (line 162)
- Destructure `sniUsed` from request
- Match `HandshakeSent pk challenge_``processClientHandshake pk challenge_`
### `processHello` (line 171)
- Branch `(sniUsed, B.null bodyHead)`:
- `(_, True)` → standard: `challenge_ = Nothing`
- `(True, False)` → web: unpad, verify 32 bytes, `challenge_ = Just`
- `(False, False)``throwE HANDSHAKE`
- Store: `HandshakeSent pk challenge_`
- Compute: `webIdentityProof = C.signatureBytes . C.sign serverSignKey . (<> sessionId) <$> challenge_`
- Construct `XFTPServerHandshake` with `webIdentityProof`
### `processClientHandshake` (line 183)
- Accept `challenge_` parameter
- Decode `webChallenge` from `XFTPClientHandshake`
- Add: `unless (challenge_ == webChallenge) $ throwE HANDSHAKE`
(standard: both `Nothing` → passes)
## 4. Native Client — `src/Simplex/FileTransfer/Client.hs`
### `xftpClientHandshakeV1` (line 142)
Add `webChallenge = Nothing` in `sendClientHandshake` call.
No other changes — parser handles new fields via `Tail`, native client ignores `webIdentityProof`.
## 5. TypeScript Changes (DONE except Ed448)
Sections 5.1 and 5.2 are implemented. Section 5.3 needs Ed448 support.
## 10. Ed448 Support via `@noble/curves`
**Problem**: Production servers use Ed448 certificates (default). `identity.ts` only supports Ed25519 via libsodium. libsodium has no Ed448 support and never will.
**Solution**: Add `@noble/curves` dependency for Ed448 verification only. All other crypto stays with libsodium.
### 10.1 `xftp-web/package.json` — Add dependency
```json
"dependencies": {
"libsodium-wrappers-sumo": "^0.7.13",
"@noble/curves": "^1.9.7"
}
```
Use v1.x (supports both CJS and ESM). v2.x is ESM-only with `.js` extension requirement.
### 10.2 `xftp-web/src/crypto/keys.ts` — Ed448 DER constants and decode
Add Ed448 SPKI DER prefix (12 bytes, same prefix length as Ed25519):
```
30 43 30 05 06 03 2b 65 71 03 3a 00
```
| Property | Ed25519 | Ed448 |
|----------|---------|-------|
| OID | `2b 65 70` | `2b 65 71` |
| SPKI prefix | `30 2a ...` | `30 43 ...` |
| Raw key size | 32 bytes | 57 bytes |
| SPKI total | 44 bytes | 69 bytes |
| Signature size | 64 bytes | 114 bytes |
New functions:
- `decodePubKeyEd448(der: Uint8Array): Uint8Array` — 69 bytes → 57 bytes raw
- `encodePubKeyEd448(raw: Uint8Array): Uint8Array` — 57 bytes → 69 bytes DER
- `verifyEd448(publicKey: Uint8Array, sig: Uint8Array, msg: Uint8Array): boolean` — uses `ed448.verify(sig, msg, publicKey)` from `@noble/curves/ed448`
Note: `@noble/curves` parameter order is `(signature, message, publicKey)`, not `(publicKey, signature, message)`.
### 10.3 `xftp-web/src/crypto/identity.ts` — Algorithm-agnostic verification
Replace `extractCertEd25519Key` + hardcoded Ed25519 `verify` with algorithm detection:
1. `extractCertPublicKeyInfo(certDer)` → SPKI DER (already exists, works for any algorithm)
2. Detect algorithm from SPKI: byte at offset 8 is `0x70` (Ed25519) or `0x71` (Ed448)
3. Extract raw key with appropriate decoder
4. Verify signatures with appropriate function
```typescript
type CertKeyAlgorithm = 'ed25519' | 'ed448'
function detectKeyAlgorithm(spki: Uint8Array): CertKeyAlgorithm {
if (spki.length === 44 && spki[8] === 0x70) return 'ed25519'
if (spki.length === 69 && spki[8] === 0x71) return 'ed448'
throw new Error("unsupported certificate key algorithm")
}
```
`verifyIdentityProof` changes:
- Extract SPKI from leaf cert
- Detect algorithm → choose `decodePubKeyEd25519`/`decodePubKeyEd448` and `verify`/`verifyEd448`
- Both challenge signature and DH key signature use the same leaf key + algorithm
Remove `extractCertEd25519Key` (replaced by generic path). Keep `extractCertPublicKeyInfo` (already generic).
### 10.4 `xftp-web/src/protocol/handshake.ts` — Comment update
`SignedKey.signature` comment: "raw Ed25519 signature bytes (64 bytes)" → "raw signature bytes (Ed25519: 64, Ed448: 114)"
### 10.5 Tests — `tests/XFTPWebTests.hs`
**Integration test**: Switch from `withXFTPServerEd25519SNI` (Ed25519 fixtures) to `withXFTPServerSNI` (default Ed448 fixtures). Update fingerprint source from `tests/fixtures/ed25519/ca.crt` to `tests/fixtures/ca.crt`.
Optionally add a second integration test with Ed25519 to cover both paths, or rely on existing unit tests for Ed25519 coverage.
### 10.6 Implementation order
1. `npm install @noble/curves` in `xftp-web/`
2. `keys.ts` — Ed448 constants, decode, encode, verifyEd448
3. `identity.ts` — algorithm detection, generic verification
4. `handshake.ts` — comment fix
5. `XFTPWebTests.hs` — switch integration test to Ed448
6. Build TS + run all tests
## 6. Haskell Integration Test — `tests/XFTPServerTests.hs`
Add `testWebHandshake` to "XFTP SNI and CORS" describe block.
1. `withXFTPServerSNI` — server with web credentials
2. Connect with SNI + `h2` ALPN
3. Send padded 32-byte challenge
4. Decode `XFTPServerHandshake`, assert `webIdentityProof` is `Just`
5. `chainIdCaCerts` on `authPubKey.certChain``CCValid {leafCert, idCert}`
6. Verify `SHA-256(idCert) == keyHash`
7. Extract `leafCert` public key, verify challenge signature
8. Verify `signedPubKey` signature using `leafCert` key (DH key auth)
9. Send `XFTPClientHandshake` with `webChallenge = Just challenge`
10. Assert empty response
Imports: `XFTPServerHandshake (..)`, `XFTPClientHandshake (..)`, `ChainCertificates (..)`, `chainIdCaCerts`.
## 7. TS Tests — `tests/XFTPWebTests.hs`
### Unit tests
- **`decodeServerHandshake` with proof**: Haskell-encode with `Just sigBytes`, TS-decode, verify bytes match.
- **`encodeClientHandshake` with challenge**: TS-encode, compare with Haskell-encoded.
- **`chainIdCaCerts`**: 2/3/4-cert chains return correct positions.
- **`caFingerprint` (fixed)**: matches `sha256(idCert)` for 2 and 3-cert chains.
### Integration test
Node.js inline script against `withXFTPServerSNI`:
1. Connect with SNI via `http2.connect`
2. Send padded challenge, decode `XFTPServerHandshake` with TS
3. `verifyIdentityProof` — full chain validation + challenge sig + DH key sig
4. Send client handshake with echoed challenge
5. Assert empty response
## 8. Implementation Order
1. `Transport.hs``Maybe` fields + encoding instances
2. `Server.hs``sniUsed`, challenge in `Handshake`, `processHello`, `processClientHandshake`, SNI routing
3. `Client.hs``webChallenge = Nothing`
4. Build: `cabal build --ghc-options -O0`
5. Run existing SNI/CORS tests
6. `XFTPServerTests.hs``testWebHandshake`
7. `handshake.ts` — types, decoding, `chainIdCaCerts`, fix `caFingerprint`
8. `crypto/identity.ts` — Node.js verification functions
9. `XFTPWebTests.hs` — unit + integration tests
10. Build TS + run all tests
## 9. Verification
```bash
cd xftp-web && npm install && npm run build && cd ..
cabal test --ghc-options=-O0 --test-option='--match=/XFTP/XFTP server/XFTP SNI and CORS/' --test-show-details=streaming
cabal test --ghc-options=-O0 --test-option='--match=/XFTP Web Client/' --test-show-details=streaming
```
@@ -0,0 +1,208 @@
# Plan: Browser ↔ Haskell File Transfer Tests
## Table of Contents
1. Goal
2. Current State
3. Implementation
4. Success Criteria
5. Files
6. Order
## 1. Goal
Run browser upload/download tests in headless Chromium via Vitest, proving fetch-based transport works in real browser environment.
## 2. Current State
- `client.ts`: Transport abstraction done — http2 for Node, fetch for browser ✓
- `agent.ts`: Uses `node:crypto` (randomBytes) and `node:zlib` (deflateRawSync/inflateRawSync) — **won't run in browser**
- `XFTPWebTests.hs`: Cross-language tests exist (Haskell calls TS via Node.js) ✓
## 3. Implementation
### 3.1 Make agent.ts isomorphic
| Current (Node.js only) | Isomorphic replacement |
|------------------------|------------------------|
| `import crypto from "node:crypto"` | Remove import |
| `import zlib from "node:zlib"` | `import pako from "pako"` |
| `crypto.randomBytes(32)` | `crypto.getRandomValues(new Uint8Array(32))` |
| `zlib.deflateRawSync(buf)` | `pako.deflateRaw(buf)` |
| `zlib.inflateRawSync(buf)` | `pako.inflateRaw(buf)` |
Note: `crypto.getRandomValues` available in both browser and Node.js (globalThis.crypto).
### 3.2 Vitest browser mode setup
`package.json` additions:
```json
"devDependencies": {
"vitest": "^3.0.0",
"@vitest/browser": "^3.0.0",
"playwright": "^1.50.0",
"@types/pako": "^2.0.3"
},
"dependencies": {
"pako": "^2.1.0"
}
```
`vitest.config.ts`:
```typescript
import {defineConfig} from 'vitest/config'
import {readFileSync} from 'fs'
import {createHash} from 'crypto'
// Compute fingerprint from ca.crt (same as Haskell's loadFileFingerprint)
const caCert = readFileSync('../tests/fixtures/ca.crt')
const fingerprint = createHash('sha256').update(caCert).digest('base64url')
const serverAddr = `xftp://${fingerprint}@localhost:7000`
export default defineConfig({
define: {
'import.meta.env.XFTP_SERVER': JSON.stringify(serverAddr)
},
test: {
browser: {
enabled: true,
provider: 'playwright',
instances: [{browser: 'chromium'}],
headless: true,
providerOptions: {
launch: {ignoreHTTPSErrors: true}
}
},
globalSetup: './test/globalSetup.ts'
}
})
```
### 3.3 Server startup
`test/globalSetup.ts`:
```typescript
import {spawn, ChildProcess} from 'child_process'
import {resolve, join} from 'path'
import {mkdtempSync, writeFileSync, copyFileSync} from 'fs'
import {tmpdir} from 'os'
let server: ChildProcess | null = null
export async function setup() {
const fixtures = resolve(__dirname, '../../tests/fixtures')
// Create temp directories
const cfgDir = mkdtempSync(join(tmpdir(), 'xftp-cfg-'))
const logDir = mkdtempSync(join(tmpdir(), 'xftp-log-'))
const filesDir = mkdtempSync(join(tmpdir(), 'xftp-files-'))
// Copy certificates to cfgDir (xftp-server expects ca.crt, server.key, server.crt there)
copyFileSync(join(fixtures, 'ca.crt'), join(cfgDir, 'ca.crt'))
copyFileSync(join(fixtures, 'server.key'), join(cfgDir, 'server.key'))
copyFileSync(join(fixtures, 'server.crt'), join(cfgDir, 'server.crt'))
// Write INI config file
const iniContent = `[STORE_LOG]
enable: off
[TRANSPORT]
host: localhost
port: 7000
[FILES]
path: ${filesDir}
[WEB]
cert: ${join(fixtures, 'web.crt')}
key: ${join(fixtures, 'web.key')}
`
writeFileSync(join(cfgDir, 'file-server.ini'), iniContent)
// Spawn xftp-server with env vars
server = spawn('cabal', ['exec', 'xftp-server', '--', 'start'], {
env: {
...process.env,
XFTP_SERVER_CFG_PATH: cfgDir,
XFTP_SERVER_LOG_PATH: logDir
},
stdio: ['ignore', 'pipe', 'pipe']
})
// Wait for "Listening on port 7000..."
await waitForServerReady(server)
}
export async function teardown() {
server?.kill('SIGTERM')
await new Promise(r => setTimeout(r, 500))
}
function waitForServerReady(proc: ChildProcess): Promise<void> {
return new Promise((resolve, reject) => {
const timeout = setTimeout(() => reject(new Error('Server start timeout')), 15000)
proc.stdout?.on('data', (data: Buffer) => {
if (data.toString().includes('Listening on port')) {
clearTimeout(timeout)
resolve()
}
})
proc.stderr?.on('data', (data: Buffer) => {
console.error('[xftp-server]', data.toString())
})
proc.on('error', reject)
proc.on('exit', (code) => {
clearTimeout(timeout)
if (code !== 0) reject(new Error(`Server exited with code ${code}`))
})
})
}
```
Server env vars (from `apps/xftp-server/Main.hs` + `getEnvPath`):
- `XFTP_SERVER_CFG_PATH` — directory containing `file-server.ini` and certs (`ca.crt`, `server.key`, `server.crt`)
- `XFTP_SERVER_LOG_PATH` — directory for logs
### 3.4 Browser test
`test/browser.test.ts`:
```typescript
import {test, expect} from 'vitest'
import {encryptFileForUpload, uploadFile, downloadFile} from '../src/agent.js'
import {parseXFTPServer} from '../src/protocol/address.js'
const server = parseXFTPServer(import.meta.env.XFTP_SERVER)
test('browser upload + download round-trip', async () => {
const data = new Uint8Array(50000)
crypto.getRandomValues(data)
const encrypted = encryptFileForUpload(data, 'test.bin')
const {rcvDescription} = await uploadFile(server, encrypted)
const {content} = await downloadFile(rcvDescription)
expect(content).toEqual(data)
})
```
## 4. Success Criteria
1. `npm run build` — agent.ts compiles without node: imports
2. `cabal test --test-option='--match=/XFTP Web Client/'` — existing Node.js tests still pass
3. `npm run test:browser` — browser round-trip test passes in headless Chromium
## 5. Files to Create/Modify
**Modify:**
- `xftp-web/package.json` — add vitest, @vitest/browser, playwright, pako, @types/pako
- `xftp-web/src/agent.ts` — replace node:crypto, node:zlib with isomorphic alternatives
**Create:**
- `xftp-web/vitest.config.ts` — browser mode config
- `xftp-web/test/globalSetup.ts` — xftp-server lifecycle
- `xftp-web/test/browser.test.ts` — browser round-trip test
## 6. Order of Implementation
1. **Add pako dependency**`npm install pako @types/pako`
2. **Make agent.ts isomorphic** — replace node:crypto, node:zlib
3. **Verify Node.js tests pass**`cabal test --test-option='--match=/XFTP Web Client/'`
4. **Set up Vitest** — add devDeps, create vitest.config.ts
5. **Create globalSetup.ts** — write INI config, spawn xftp-server
6. **Write browser test** — upload + download round-trip
7. **Verify browser test passes**`npm run test:browser`
@@ -0,0 +1,920 @@
# Browser Transport & Web Worker Architecture
## TOC
1. Executive Summary
2. Transport: fetch() API
3. Architecture: Environment Abstraction
4. Web Worker Implementation
5. OPFS Implementation
6. Implementation Plan
7. Testing Strategy
## 1. Executive Summary
Adapt `client.ts` from `node:http2` to `fetch()` API for isomorphic Node.js/browser support. Add environment abstraction layer so the same upload/download pipeline works with or without Web Workers and with or without OPFS. In browsers, crypto runs in a Web Worker to keep UI responsive; in Node.js tests, crypto runs directly.
**Key architectural constraint:** Existing crypto functions (`encryptFile`, `decryptChunks`, etc.) remain unchanged. The abstraction layer wraps them, choosing execution context (direct vs Worker) and storage (memory vs OPFS) based on environment.
**Scope:**
- Replace `node:http2` with `fetch()` in `client.ts`
- Add `CryptoBackend` abstraction with three implementations
- Create Web Worker that calls existing crypto functions
- Add OPFS storage for large files in browser
**Out of scope:** Web page UI (Phase 5 in main RFC).
## 2. Transport: fetch() API
### 2.1 Current State
`client.ts` uses `node:http2`:
```typescript
import http2 from "node:http2"
const session = http2.connect(url)
const stream = session.request({':method': 'POST', ':path': '/'})
stream.write(commandBlock)
stream.end(chunkData)
```
### 2.2 Target State
Isomorphic `fetch()` (Node.js 18+ and browsers):
```typescript
const response = await fetch(url, {
method: 'POST',
body: concatStreams(commandBlock, chunkData),
duplex: 'half', // Required for streaming request body
})
const reader = response.body!.getReader()
```
### 2.3 Key Differences
| Aspect | node:http2 | fetch() |
|--------|-----------|---------|
| Session management | Explicit `session.connect()` / `session.close()` | Per-request (HTTP/2 connection reuse is automatic) |
| Streaming upload | `stream.write()` chunks | `ReadableStream` body + `duplex: 'half'` |
| Streaming download | `stream.on('data')` | `response.body.getReader()` |
| Connection pooling | Manual | Automatic per origin |
### 2.4 API Changes
```typescript
// Before (node:http2)
export interface XFTPClient {
session: http2.ClientHttp2Session
thParams: THParams
server: XFTPServer
}
// After (fetch)
export interface XFTPClient {
baseUrl: string // "https://host:port"
thParams: THParams
server: XFTPServer
}
```
`connectXFTP()` performs handshake via fetch, returns `XFTPClient` with `baseUrl`.
Subsequent commands use `fetch(client.baseUrl, ...)`.
### 2.5 Handshake via fetch()
**TLS session binding:** Multiple fetch() requests to the same origin reuse the HTTP/2 connection, which means they share the same TLS session. The server's `sessionId` (derived from TLS channel binding) remains consistent across the handshake round-trips and subsequent commands.
```typescript
async function connectXFTP(server: XFTPServer): Promise<XFTPClient> {
const baseUrl = `https://${server.host}:${server.port}`
// Round-trip 1: challenge → server handshake + identity proof
const challenge = crypto.getRandomValues(new Uint8Array(32))
const req1 = pad(encodeWebClientHello(challenge), xftpBlockSize)
const resp1 = await fetch(baseUrl, {method: 'POST', body: req1})
const reader = resp1.body!.getReader()
const serverBlock = await readExactly(reader, xftpBlockSize)
const serverHs = decodeServerHandshake(unPad(serverBlock))
const proofBody = await readRemaining(reader)
verifyIdentityProof(server.keyHash, challenge, serverHs.sessionId, proofBody)
// Round-trip 2: client handshake → server ack
const clientHs = encodeClientHandshake({xftpVersion: 3, keyHash: server.keyHash})
const req2 = pad(clientHs, xftpBlockSize)
await fetch(baseUrl, {method: 'POST', body: req2})
return {baseUrl, thParams: {sessionId: serverHs.sessionId, ...}, server}
}
```
### 2.6 Command Execution
```typescript
async function sendXFTPCommand(
client: XFTPClient,
key: Uint8Array,
entityId: Uint8Array,
cmd: Uint8Array,
chunkData?: Uint8Array
): Promise<{response: Uint8Array, body?: ReadableStream}> {
const block = xftpEncodeAuthTransmission(client.thParams, key, entityId, cmd)
const reqBody = chunkData
? concatBytes(block, chunkData)
: block
const resp = await fetch(client.baseUrl, {
method: 'POST',
body: reqBody,
duplex: 'half',
})
const reader = resp.body!.getReader()
const responseBlock = await readExactly(reader, xftpBlockSize)
const parsed = xftpDecodeTransmission(responseBlock)
// For FGET: remaining body is encrypted chunk
const hasMore = await peekReader(reader)
return {
response: parsed,
body: hasMore ? wrapAsStream(reader) : undefined
}
}
```
## 3. Architecture: Environment Abstraction
### 3.1 Core Principle
**Existing crypto functions remain unchanged.** The functions `encryptFile()`, `decryptChunks()`, `sha512()`, etc. in `crypto/file.ts` and `crypto/digest.ts` are pure computation — they take input bytes and produce output bytes. They have no knowledge of Workers, OPFS, or execution context.
The abstraction layer sits between `agent.ts` (upload/download orchestration) and these crypto functions:
```
┌─────────────────────────────────────────────────────────────────────┐
│ agent.ts (upload/download orchestration) │
│ - Unchanged logic: encrypt → chunk → upload → build description │
│ - Calls CryptoBackend interface, not crypto functions directly │
├─────────────────────────────────────────────────────────────────────┤
│ CryptoBackend interface (env.ts) │
│ - Abstract interface for encrypt/decrypt/readChunk/writeChunk │
│ - Factory function selects implementation based on environment │
├──────────────┬──────────────────────┬───────────────────────────────┤
│ DirectMemory │ WorkerMemory │ WorkerOPFS │
│ Backend │ Backend │ Backend │
│ (Node.js) │ (Browser, ≤50MB) │ (Browser, >50MB) │
├──────────────┼──────────────────────┼───────────────────────────────┤
│ Calls crypto │ Posts to Worker, │ Posts to Worker, │
│ functions │ Worker calls crypto │ Worker calls crypto, │
│ directly │ functions, returns │ streams through OPFS │
│ │ via postMessage │ │
├──────────────┴──────────────────────┴───────────────────────────────┤
│ crypto/file.ts, crypto/digest.ts (unchanged) │
│ - encryptFile(), decryptChunks(), sha512(), etc. │
│ - Pure functions, no environment dependencies │
└─────────────────────────────────────────────────────────────────────┘
```
### 3.2 CryptoBackend Interface
```typescript
// env.ts
export interface CryptoBackend {
// Encrypt file, store result (in memory or OPFS depending on backend)
encrypt(
data: Uint8Array,
fileName: string,
onProgress?: (done: number, total: number) => void
): Promise<EncryptResult>
// Decrypt from stored encrypted data
decrypt(
key: Uint8Array,
nonce: Uint8Array,
size: number,
onProgress?: (done: number, total: number) => void
): Promise<DecryptResult>
// Read chunk from stored encrypted data (for upload)
readChunk(offset: number, size: number): Promise<Uint8Array>
// Write chunk to storage (for download, before decrypt)
writeChunk(data: Uint8Array, offset: number): Promise<void>
// Clean up temporary storage
cleanup(): Promise<void>
}
export interface EncryptResult {
digest: Uint8Array // SHA-512 of encrypted data
key: Uint8Array // Generated encryption key
nonce: Uint8Array // Generated nonce
chunkSizes: number[] // Chunk sizes for upload
totalSize: number // Total encrypted size
}
export interface DecryptResult {
header: FileHeader // Extracted file header (fileName, etc.)
content: Uint8Array // Decrypted file content
}
```
### 3.3 Backend Implementations
**DirectMemoryBackend** (Node.js):
```typescript
class DirectMemoryBackend implements CryptoBackend {
private encryptedData: Uint8Array | null = null
async encrypt(data: Uint8Array, fileName: string, onProgress?): Promise<EncryptResult> {
const key = randomBytes(32)
const nonce = randomBytes(24)
// Call existing crypto function directly
this.encryptedData = encryptFile(data, fileName, key, nonce, onProgress)
const digest = sha512(this.encryptedData)
const chunkSizes = prepareChunkSizes(this.encryptedData.length)
return { digest, key, nonce, chunkSizes, totalSize: this.encryptedData.length }
}
async decrypt(key, nonce, size, onProgress): Promise<DecryptResult> {
// Call existing crypto function directly
return decryptChunks([this.encryptedData!], key, nonce, size, onProgress)
}
async readChunk(offset: number, size: number): Promise<Uint8Array> {
return this.encryptedData!.slice(offset, offset + size)
}
async writeChunk(data: Uint8Array, offset: number): Promise<void> {
if (!this.encryptedData) this.encryptedData = new Uint8Array(offset + data.length)
this.encryptedData.set(data, offset)
}
async cleanup(): Promise<void> {
this.encryptedData = null
}
}
```
**WorkerMemoryBackend** and **WorkerOPFSBackend** are similar but post messages to a Web Worker instead of calling crypto directly. The Worker then calls the same `encryptFile()`, `decryptChunks()` functions. See §4 for Worker implementation details.
### 3.4 Factory Function
```typescript
// env.ts
export function createCryptoBackend(fileSize: number): CryptoBackend {
const hasWorker = typeof Worker !== 'undefined'
const hasOPFS = typeof navigator?.storage?.getDirectory !== 'undefined'
const isLargeFile = fileSize > 50 * 1024 * 1024
if (hasWorker && hasOPFS && isLargeFile) {
return new WorkerOPFSBackend() // Browser + large file
} else if (hasWorker) {
return new WorkerMemoryBackend() // Browser + small file
} else {
return new DirectMemoryBackend() // Node.js
}
}
```
### 3.5 Usage in agent.ts
```typescript
// agent.ts - upload orchestration (simplified)
export async function uploadFile(
server: XFTPServer,
fileData: Uint8Array,
fileName: string,
onProgress?: ProgressCallback
): Promise<string> {
// Create backend based on environment
const backend = createCryptoBackend(fileData.length)
try {
// Encrypt (runs in Worker in browser, directly in Node)
const enc = await backend.encrypt(fileData, fileName, onProgress)
// Upload chunks (same code regardless of backend)
const client = await connectXFTP(server)
const sentChunks = []
let offset = 0
for (const size of enc.chunkSizes) {
const chunk = await backend.readChunk(offset, size)
const sent = await uploadChunk(client, chunk, enc.digest)
sentChunks.push(sent)
offset += size
}
// Build description and URI
const fd = buildFileDescription(enc, sentChunks)
return encodeFileDescriptionURI(fd)
} finally {
await backend.cleanup()
}
}
```
The key point: `uploadFile()` logic is identical regardless of whether crypto runs in a Worker or directly. The `CryptoBackend` abstraction hides that detail.
### 3.6 Why This Matters for Testing
- **Layer 1 tests** (per-function): Call `encryptFile()`, `decryptChunks()` directly via Node — unchanged
- **Layer 2 tests** (full flow): Call `uploadFile()`, `downloadFile()` in Node — uses `DirectMemoryBackend`, same code path as browser except for Worker
- **Layer 3 tests** (browser): Call `uploadFile()`, `downloadFile()` in Playwright — uses `WorkerMemoryBackend` or `WorkerOPFSBackend`
All three layers exercise the same crypto functions. The only difference is execution context.
## 4. Web Worker Implementation
### 4.1 Why Web Worker
File encryption (XSalsa20-Poly1305) is sequential and CPU-bound:
- 100 MB file ≈ 1-2 seconds of continuous computation
- Running on main thread blocks UI (no progress updates, frozen page)
- Chunking into async microtasks adds complexity and still causes jank
Web Worker runs crypto in parallel thread. Main thread stays responsive.
### 4.2 Architecture
```
┌─────────────────────────────────────────────────────────────┐
│ Main Thread │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────────────┐ │
│ │ UI (upload/ │ │ Progress │ │ Network (fetch) │ │
│ │ download) │ │ display │ │ │ │
│ └──────┬──────┘ └──────▲──────┘ └──────────▲──────────┘ │
│ │ │ │ │
│ │ postMessage │ progress │ encrypted │
│ ▼ │ events │ chunks │
├─────────────────────────────────────────────────────────────┤
│ Web Worker │
│ ┌─────────────────────────────────────────────────────────┐│
│ │ Crypto Pipeline ││
│ │ - encryptFile() with progress callbacks ││
│ │ - decryptChunks() with progress callbacks ││
│ │ - OPFS read/write for temp storage ││
│ └─────────────────────────────────────────────────────────┘│
└─────────────────────────────────────────────────────────────┘
```
### 4.3 Message Protocol
**Main → Worker:**
```typescript
type WorkerRequest =
// Encrypt file, store result in OPFS (large) or memory (small)
| {type: 'encrypt', file: File, fileName: string, useOPFS: boolean}
// Read encrypted chunk from OPFS for upload
| {type: 'readChunk', offset: number, size: number}
// Write downloaded chunk to OPFS for later decryption
| {type: 'writeChunk', data: ArrayBuffer, offset: number}
// Decrypt from OPFS or provided chunks
| {type: 'decrypt', key: Uint8Array, nonce: Uint8Array, size: number, chunks?: ArrayBuffer[]}
// Delete OPFS temp files
| {type: 'cleanup'}
| {type: 'cancel'}
```
**Worker → Main:**
```typescript
type WorkerResponse =
| {type: 'progress', phase: 'encrypt' | 'decrypt', done: number, total: number}
// For OPFS: encData is empty, data lives in OPFS temp file
| {type: 'encrypted', encData: ArrayBuffer | null, digest: Uint8Array, key: Uint8Array, nonce: Uint8Array, chunkSizes: number[]}
| {type: 'chunk', data: ArrayBuffer} // Response to readChunk
| {type: 'chunkWritten'} // Response to writeChunk
| {type: 'decrypted', header: FileHeader, content: ArrayBuffer}
| {type: 'cleaned'} // Response to cleanup
| {type: 'error', message: string}
```
### 4.4 Worker Implementation
```typescript
// crypto.worker.ts
import {encryptFile, encryptFileStreaming, decryptChunks, decryptFromOPFS} from './crypto/file.js'
import {sha512} from './crypto/digest.js'
import {prepareChunkSizes} from './protocol/chunks.js'
let opfsHandle: FileSystemSyncAccessHandle | null = null
self.onmessage = async (e: MessageEvent<WorkerRequest>) => {
const req = e.data
if (req.type === 'encrypt') {
const key = crypto.getRandomValues(new Uint8Array(32))
const nonce = crypto.getRandomValues(new Uint8Array(24))
if (req.useOPFS) {
// Large file: stream through OPFS to avoid memory pressure
const root = await navigator.storage.getDirectory()
const fileHandle = await root.getFileHandle('encrypted-temp', {create: true})
opfsHandle = await fileHandle.createSyncAccessHandle()
// Stream encrypt: read 64KB from File, encrypt, write to OPFS
const digest = await encryptFileStreaming(
req.file,
req.fileName,
key,
nonce,
opfsHandle,
(done, total) => self.postMessage({type: 'progress', phase: 'encrypt', done, total})
)
const encSize = opfsHandle.getSize()
const chunkSizes = prepareChunkSizes(encSize)
self.postMessage({
type: 'encrypted',
encData: null, // Data in OPFS, not memory
digest, key, nonce, chunkSizes
})
} else {
// Small file: in-memory is fine
const source = new Uint8Array(await req.file.arrayBuffer())
const encData = encryptFile(source, req.fileName, key, nonce, (done, total) => {
self.postMessage({type: 'progress', phase: 'encrypt', done, total})
})
const digest = sha512(encData)
const chunkSizes = prepareChunkSizes(encData.length)
self.postMessage({
type: 'encrypted',
encData: encData.buffer,
digest, key, nonce, chunkSizes
}, [encData.buffer])
}
}
if (req.type === 'readChunk') {
// Read chunk from OPFS for upload
const chunk = new Uint8Array(req.size)
opfsHandle!.read(chunk, {at: req.offset})
self.postMessage({type: 'chunk', data: chunk.buffer}, [chunk.buffer])
}
if (req.type === 'writeChunk') {
// Write downloaded chunk to OPFS
if (!opfsHandle) {
const root = await navigator.storage.getDirectory()
const fileHandle = await root.getFileHandle('download-temp', {create: true})
opfsHandle = await fileHandle.createSyncAccessHandle()
}
opfsHandle.write(new Uint8Array(req.data), {at: req.offset})
self.postMessage({type: 'chunkWritten'})
}
if (req.type === 'decrypt') {
let result
if (req.chunks) {
// Small file: chunks provided in memory
const chunks = req.chunks.map(b => new Uint8Array(b))
result = decryptChunks(chunks, req.key, req.nonce, req.size, (done, total) => {
self.postMessage({type: 'progress', phase: 'decrypt', done, total})
})
} else {
// Large file: read from OPFS
result = decryptFromOPFS(opfsHandle!, req.key, req.nonce, req.size, (done, total) => {
self.postMessage({type: 'progress', phase: 'decrypt', done, total})
})
}
self.postMessage({
type: 'decrypted',
header: result.header,
content: result.content.buffer
}, [result.content.buffer])
}
if (req.type === 'cleanup') {
if (opfsHandle) {
opfsHandle.close()
opfsHandle = null
}
const root = await navigator.storage.getDirectory()
try { await root.removeEntry('encrypted-temp') } catch {}
try { await root.removeEntry('download-temp') } catch {}
self.postMessage({type: 'cleaned'})
}
}
```
### 4.5 Main Thread Wrapper
```typescript
// crypto-worker.ts (main thread)
export class CryptoWorker {
private worker: Worker
private pending: Map<string, {resolve: Function, reject: Function}> = new Map()
private onProgress?: (done: number, total: number) => void
constructor() {
this.worker = new Worker(new URL('./crypto.worker.js', import.meta.url), {type: 'module'})
this.worker.onmessage = (e) => this.handleMessage(e.data)
}
async encrypt(file: File, onProgress?: (done: number, total: number) => void): Promise<EncryptedFileInfo> {
const useOPFS = file.size > 50 * 1024 * 1024 // 50 MB threshold
return new Promise((resolve, reject) => {
this.pending.set('encrypt', {resolve, reject})
this.onProgress = onProgress
this.worker.postMessage({type: 'encrypt', file, fileName: file.name, useOPFS})
})
}
async decrypt(
chunks: Uint8Array[],
key: Uint8Array,
nonce: Uint8Array,
size: number,
onProgress?: (done: number, total: number) => void
): Promise<DownloadResult> {
return new Promise((resolve, reject) => {
this.pending.set('decrypt', {resolve, reject})
this.onProgress = onProgress
this.worker.postMessage({
type: 'decrypt',
chunks: chunks.map(c => c.buffer),
key, nonce, size
}, chunks.map(c => c.buffer))
})
}
private handleMessage(msg: WorkerResponse) {
if (msg.type === 'progress') {
this.onProgress?.(msg.done, msg.total)
} else if (msg.type === 'encrypted') {
this.pending.get('encrypt')?.resolve({
encData: msg.encData ? new Uint8Array(msg.encData) : null, // null when using OPFS
digest: msg.digest,
key: msg.key,
nonce: msg.nonce,
chunkSizes: msg.chunkSizes
})
} else if (msg.type === 'decrypted') {
this.pending.get('decrypt')?.resolve({
header: msg.header,
content: new Uint8Array(msg.content)
})
} else if (msg.type === 'error') {
// Reject all pending
for (const p of this.pending.values()) p.reject(new Error(msg.message))
}
}
}
```
## 5. OPFS Implementation
### 5.1 Purpose
For files approaching 100 MB, holding encrypted data in memory while uploading creates memory pressure. OPFS provides temporary file storage:
- Write encrypted data to OPFS as it's generated
- Read chunks from OPFS for upload
- Delete after upload completes
### 5.2 When to Use
- Files > 50 MB: Use OPFS
- Files ≤ 50 MB: In-memory (simpler, no OPFS overhead)
Threshold is configurable.
### 5.3 OPFS API
```typescript
// In Web Worker (synchronous API for performance)
const root = await navigator.storage.getDirectory()
const fileHandle = await root.getFileHandle('encrypted-temp', {create: true})
const accessHandle = await fileHandle.createSyncAccessHandle()
// Write encrypted chunks as they're generated
accessHandle.write(encryptedChunk, {at: offset})
// Read chunk for upload
const chunk = new Uint8Array(chunkSize)
accessHandle.read(chunk, {at: chunkOffset})
// Cleanup
accessHandle.close()
await root.removeEntry('encrypted-temp')
```
### 5.4 Upload Flow with OPFS
```
1. Main: user drops file
2. Main → Worker: {type: 'encrypt', file}
3. Worker:
- Create OPFS temp file
- Encrypt 64KB at a time, write to OPFS
- Post progress every 64KB
- Compute digest
- Return {digest, key, nonce, chunkSizes} (data stays in OPFS)
4. Main: for each chunk:
- Main → Worker: {type: 'readChunk', offset, size}
- Worker: read from OPFS, return chunk
- Main: upload chunk via fetch()
5. Main → Worker: {type: 'cleanup'}
6. Worker: delete OPFS temp file
```
### 5.5 Download Flow with OPFS
```
1. Main: parse URL, get FileDescription
2. Main: for each chunk:
- Download via fetch()
- Main → Worker: {type: 'writeChunk', data, offset}
- Worker: write to OPFS temp file
3. Main → Worker: {type: 'decrypt', key, nonce, size}
4. Worker:
- Read from OPFS
- Decrypt, verify auth tag
- Return {header, content}
5. Main: trigger browser download
6. Main → Worker: {type: 'cleanup'}
```
## 6. Implementation Plan
### 6.1 Phase A: fetch() Transport
**Goal:** Replace `node:http2` with `fetch()` in `client.ts`. All existing Node.js tests pass.
1. Rewrite `connectXFTP()` to use fetch() for handshake
2. Rewrite `sendXFTPCommand()` to use fetch()
3. Update `createXFTPChunk`, `uploadXFTPChunk`, `downloadXFTPChunk`, etc.
4. Remove `node:http2` import
5. Run existing Haskell integration tests — must pass
**Files:** `client.ts`
### 6.2 Phase B: Environment Abstraction + Web Worker
**Goal:** Add `CryptoBackend` abstraction (§3) so the same code works in Node (direct) and browser (Worker).
1. Create `env.ts` with `CryptoBackend` interface and `createCryptoBackend()` factory (as specified in §3)
2. Implement `DirectMemoryBackend` for Node.js
3. Create `crypto.worker.ts` that imports and calls existing crypto functions
4. Implement `WorkerMemoryBackend` for browser
5. Update `agent.ts` to use `createCryptoBackend()` instead of direct crypto calls
6. Existing tests pass (now using `DirectMemoryBackend`)
**Files:** `env.ts`, `crypto.worker.ts`, `agent.ts`
### 6.3 Phase C: OPFS Backend
**Goal:** Large files (>50 MB) use OPFS for temp storage in browser.
1. Implement `WorkerOPFSBackend` — uses OPFS sync API in worker
2. Add OPFS helpers in worker: read/write to temp file
3. Factory function now returns `WorkerOPFSBackend` for large files
4. Same `agent.ts` code works — only backend implementation differs
**Files:** `env.ts`, `crypto.worker.ts`
### 6.4 Phase D: Browser Testing
**Goal:** Verify everything works in real browsers.
1. Create minimal test HTML page
2. Test upload flow in Chrome, Firefox, Safari
3. Test download flow
4. Test progress reporting
5. Test cancellation
6. Test error handling (network failure, invalid file)
## 7. Testing Strategy
### 7.1 Test Layers
The `CryptoBackend` abstraction (§3) enables testing at multiple levels without code duplication:
```
┌─────────────────────────────────────────────────────────────────┐
│ Layer 3: Browser Integration (Playwright) │
│ - Web Worker message passing │
│ - OPFS read/write │
│ - Progress UI updates │
│ - Real browser fetch() with CORS │
├─────────────────────────────────────────────────────────────────┤
│ Layer 2: Full Flow (Haskell-driven, Node.js) │
│ - fetch() transport against real xftp-server │
│ - Upload: encrypt → chunk → upload → build description │
│ - Download: parse → download → verify → decrypt │
│ - Cross-language: TS upload ↔ Haskell download (and vice versa) │
├─────────────────────────────────────────────────────────────────┤
│ Layer 1: Per-Function (Haskell-driven, Node.js) │
│ - 172 existing tests │
│ - Byte-identical output vs Haskell functions │
└─────────────────────────────────────────────────────────────────┘
```
### 7.2 Layer 1: Per-Function Tests (Existing)
Existing Haskell-driven tests in `XFTPWebTests.hs`. Each test calls a TypeScript function via Node and compares output with Haskell.
```bash
cabal test --ghc-options -O0 --test-option='--match=/XFTP Web Client/'
```
All 172 tests must pass. No changes needed for browser transport work.
### 7.3 Layer 2: Full Flow Tests (Node.js + fetch)
Haskell-driven integration tests using Node.js native fetch(). These test the complete upload/download flow without Worker/OPFS.
```haskell
-- XFTPWebTests.hs (extends existing test file)
it "fetch transport: upload and download round-trip" $ do
withXFTPServer testXFTPServerConfigSNI $ \server -> do
-- TypeScript uploads via fetch(), returns URI
uri <- jsOut $ callTS "src/agent" "uploadFileTest" serverAddrHex <> testFileHex
-- TypeScript downloads via fetch()
content <- jsOut $ callTS "src/agent" "downloadFileTest" uriHex
content `shouldBe` testFileContent
it "fetch transport: TS upload, Haskell download" $ do
withXFTPServer testXFTPServerConfigSNI $ \server -> do
uri <- jsOut $ callTS "src/agent" "uploadFileTest" serverAddrHex <> testFileHex
-- Haskell agent downloads using existing xftp CLI pattern
outPath <- withAgent 1 agentCfg initAgentServers testDB $ \a -> do
rfId <- xftpReceiveFile' a 1 uri Nothing
waitRfDone a
content <- B.readFile outPath
content `shouldBe` testFileContent
```
**What this tests:**
- fetch() handshake (challenge-response, TLS session binding)
- fetch() command execution (FNEW, FPUT, FGET, FACK)
- Streaming request/response bodies
- Full encrypt → upload → download → decrypt flow
**What this doesn't test:**
- Web Worker message passing
- OPFS storage
- Browser-specific fetch() behavior (CORS preflight, etc.)
### 7.4 Layer 3: Browser Integration Tests (Playwright)
Playwright tests run in real browsers, testing browser-specific functionality.
**Test infrastructure:**
```
xftp-web/
├── test/
│ ├── browser.test.ts # Playwright test file
│ └── test-server.ts # Spawns xftp-server for tests
└── test-page/
├── index.html # Minimal test UI
└── test-harness.ts # Exposes test functions to window
```
**Running browser tests:**
```bash
cd xftp-web
npm run test:browser # Spawns xftp-server, runs Playwright
```
**Test cases:**
```typescript
// test/browser.test.ts
import { test, expect } from '@playwright/test'
import { spawn } from 'child_process'
let serverProcess: ChildProcess
test.beforeAll(async () => {
// Spawn xftp-server with SNI cert for browser TLS
serverProcess = spawn('xftp-server', ['start', '-c', 'test-config.ini'])
await waitForServer()
})
test.afterAll(async () => {
serverProcess.kill()
})
test('small file upload/download (in-memory)', async ({ page }) => {
await page.goto('/test-page/')
const result = await page.evaluate(async () => {
const data = new Uint8Array(1024 * 1024) // 1 MB
crypto.getRandomValues(data)
const file = new File([data], 'small.bin')
const uri = await window.xftp.uploadFile(file)
const downloaded = await window.xftp.downloadFile(uri)
return {
uploadedSize: data.length,
downloadedSize: downloaded.length,
match: arraysEqual(data, downloaded),
usedOPFS: window.xftp.lastUploadUsedOPFS
}
})
expect(result.match).toBe(true)
expect(result.usedOPFS).toBe(false) // Small file, no OPFS
})
test('large file upload/download (OPFS)', async ({ page }) => {
await page.goto('/test-page/')
const result = await page.evaluate(async () => {
const data = new Uint8Array(60 * 1024 * 1024) // 60 MB
crypto.getRandomValues(data)
const file = new File([data], 'large.bin')
const uri = await window.xftp.uploadFile(file)
const downloaded = await window.xftp.downloadFile(uri)
return {
match: arraysEqual(data, downloaded),
usedOPFS: window.xftp.lastUploadUsedOPFS
}
})
expect(result.match).toBe(true)
expect(result.usedOPFS).toBe(true) // Large file, used OPFS
})
test('progress events fire during upload', async ({ page }) => {
await page.goto('/test-page/')
const progressEvents = await page.evaluate(async () => {
const events: number[] = []
const data = new Uint8Array(10 * 1024 * 1024) // 10 MB
const file = new File([data], 'progress.bin')
await window.xftp.uploadFile(file, (done, total) => {
events.push(done / total)
})
return events
})
expect(progressEvents.length).toBeGreaterThan(1)
expect(progressEvents[progressEvents.length - 1]).toBe(1) // 100% at end
})
test('Web Worker keeps UI responsive', async ({ page }) => {
await page.goto('/test-page/')
// Start upload and measure main thread responsiveness
const result = await page.evaluate(async () => {
const data = new Uint8Array(50 * 1024 * 1024) // 50 MB
const file = new File([data], 'responsive.bin')
let frameCount = 0
let uploadDone = false
// Count animation frames during upload
function countFrames() {
frameCount++
if (!uploadDone) requestAnimationFrame(countFrames)
}
requestAnimationFrame(countFrames)
const start = performance.now()
await window.xftp.uploadFile(file)
uploadDone = true
const elapsed = performance.now() - start
// If main thread was blocked, frameCount would be very low
const expectedFrames = (elapsed / 1000) * 30 // ~30 fps minimum
return { frameCount, expectedFrames, elapsed }
})
// Should maintain reasonable frame rate (Worker offloaded crypto)
expect(result.frameCount).toBeGreaterThan(result.expectedFrames * 0.5)
})
```
### 7.5 Cross-Browser Matrix
| Browser | fetch streaming | Web Worker | OPFS sync | Status |
|---------|----------------|------------|-----------|--------|
| Chrome 105+ | ✓ | ✓ | ✓ | Primary target |
| Firefox 111+ | ✓ | ✓ | ✓ | Supported |
| Safari 16.4+ | ✓ | ✓ | ✓ | Supported |
| Edge 105+ | ✓ | ✓ | ✓ | Supported (Chromium) |
Playwright tests run against Chrome by default. CI can run against all browsers.
### 7.6 Test Execution Summary
| Phase | Test Layer | Command | What's Verified |
|-------|-----------|---------|-----------------|
| A | Layer 1 + 2 | `cabal test --test-option='--match=/XFTP Web Client/'` | fetch() transport, full flow |
| B | Layer 3 | `npm run test:browser` | Worker message passing, progress |
| C | Layer 3 | `npm run test:browser` | OPFS storage for large files |
| D | Layer 3 | `npm run test:browser -- --project=firefox,webkit` | Cross-browser |
@@ -0,0 +1,772 @@
# Send File Web Page — Implementation Plan
## TOC
1. Executive Summary
2. Architecture
3. CryptoBackend & Web Worker
4. Server Configuration
5. Page Structure & UI
6. Upload Flow
7. Download Flow
8. Build & Dev Setup
9. agent.ts Changes
10. Testing
11. Files
12. Implementation Order
## 1. Executive Summary
Build a static web page for browser-based XFTP file transfer (Phase 5 of master RFC). The page supports upload (drag-drop → encrypt → upload → shareable link) and download (open link → download → decrypt → save). Crypto runs in a Web Worker; large files use OPFS temp storage.
Two build variants:
- **Local**: single test server at `localhost:7000` (development/testing)
- **Production**: 12 preset XFTP servers (6 SimpleX + 6 Flux)
Uses Vite for bundling (already a dependency via vitest). No CSS framework — plain CSS per RFC spec.
## 2. Architecture
```
xftp-web/
├── src/ # Library (existing, targeted changes)
│ ├── agent.ts # Modified: uploadFile readChunk, downloadFileRaw
│ ├── client.ts # Modified: downloadXFTPChunkRaw
│ ├── crypto/ # Unchanged
│ ├── download.ts # Unchanged
│ └── protocol/
│ └── description.ts # Fix: SHA-256 → SHA-512 comment on digest field
├── web/ # Web page (new)
│ ├── index.html # Entry point (CSP meta tag)
│ ├── main.ts # Router + sodium.ready init
│ ├── upload.ts # Upload UI + orchestration
│ ├── download.ts # Download UI + orchestration
│ ├── progress.ts # Circular progress canvas component
│ ├── servers.ts # Server list (build-time configured, imports servers.json)
│ ├── servers.json # Preset server addresses (shared with vite.config.ts)
│ ├── crypto-backend.ts # CryptoBackend interface + WorkerBackend
│ ├── crypto.worker.ts # Web Worker: encrypt/decrypt/OPFS
│ └── style.css # Minimal styling
├── vite.config.ts # Page build config (new)
├── tsconfig.web.json # IDE/CI type-check for web/ (new)
├── tsconfig.worker.json # IDE/CI type-check for worker (new)
├── playwright.config.ts # Page E2E test config (new)
├── vitest.config.ts # Test config (existing)
├── .gitignore # Existing (add dist-web/)
└── test/ # Tests (existing + new page test)
```
Data flow:
```
┌───────────────────────────────────────────┐
│ Main Thread │
│ │
│ Upload: upload.ts ──► agent.ts ──► fetch()│
│ Download: download.ts ──► agent.ts ──► fetch()
│ │ │
│ postMessage HTTP/2 │
│ ▼ ▼
│ ┌─────────────────┐ ┌──────────┐│
│ │ Web Worker │ │ XFTP ││
│ │ crypto.worker.ts │ │ Server ││
│ │ ┌─────────────┐ │ └──────────┘│
│ │ │ OPFS temp │ │ │
│ │ └─────────────┘ │ │
│ └─────────────────┘ │
└───────────────────────────────────────────┘
```
Both upload and download use `agent.ts` for orchestration (connection pooling, parallel chunk transfers, redirect handling). Upload uses a `readChunk` callback for Worker data access. Download uses an `onRawChunk` callback to route raw encrypted chunks to the Worker for decryption (see §7.2). ACK is the caller's responsibility — `downloadFileRaw` returns the resolved `FileDescription` without ACKing, so the caller can verify integrity before acknowledging.
## 3. CryptoBackend & Web Worker
### 3.1 Interface
```typescript
// crypto-backend.ts
export interface CryptoBackend {
// Upload: encrypt file, store encrypted data in OPFS
encrypt(data: Uint8Array, fileName: string,
onProgress?: (done: number, total: number) => void
): Promise<EncryptResult>
// Upload: read encrypted chunk from OPFS (called by agent.ts via readChunk callback)
readChunk(offset: number, size: number): Promise<Uint8Array>
// Download: transit-decrypt raw chunk and store in OPFS
decryptAndStoreChunk(
dhSecret: Uint8Array, nonce: Uint8Array,
body: Uint8Array, digest: Uint8Array, chunkNo: number
): Promise<void>
// Download: verify digest + file-level decrypt all stored chunks
// Only needs size/digest/key/nonce — not the full FileDescription (avoids sending private keys to Worker)
verifyAndDecrypt(params: {size: number, digest: Uint8Array, key: Uint8Array, nonce: Uint8Array}
): Promise<{header: FileHeader, content: Uint8Array}>
cleanup(): Promise<void>
}
// Structurally identical to EncryptedFileMetadata from agent.ts (§9.1).
// Kept separate to avoid crypto-backend.ts importing from agent.ts
// (which would pull in node:http2 via client.ts, breaking Worker bundling).
// TypeScript structural typing makes them assignment-compatible.
export interface EncryptResult {
digest: Uint8Array
key: Uint8Array
nonce: Uint8Array
chunkSizes: number[]
}
```
### 3.2 Factory
```typescript
export function createCryptoBackend(): CryptoBackend {
if (typeof Worker === 'undefined') {
throw new Error('Web Workers required — update your browser')
}
return new WorkerBackend()
}
```
The Worker always uses OPFS for temp storage (single code path — no memory/disk branching). OPFS I/O overhead is negligible relative to crypto and network time. Each Worker session creates a unique directory in OPFS root named `session-<Date.now()>-<crypto.randomUUID()>`, containing `upload.bin` and `download.bin` as needed. `cleanup()` deletes the entire session directory. On Worker startup (before processing messages), sweep OPFS root and delete any `session-*` directories whose embedded timestamp (parsed from the name) is older than 1 hour — this handles stale files from crashed tabs. The OPFS API does not expose directory timestamps, so the name-encoded timestamp is the only reliable mechanism. This prevents cross-tab collisions and unbounded OPFS growth.
### 3.3 Worker message protocol
Every request carries a numeric `id`. Responses carry the same `id`. WorkerBackend maintains a `Map<number, {resolve, reject}>` to match responses to pending promises.
Main → Worker (fields marked `†` are Transferable — arrive as `ArrayBuffer` in Worker, must be wrapped with `new Uint8Array(...)` before use):
- `{id: number, type: 'encrypt', data†: ArrayBuffer, fileName: string}` — encrypt file, store in OPFS
- `{id: number, type: 'readChunk', offset: number, size: number}` — read encrypted chunk from OPFS
- `{id: number, type: 'decryptAndStoreChunk', dhSecret: Uint8Array, nonce: Uint8Array, body†: ArrayBuffer, chunkDigest: Uint8Array, chunkNo: number}` — transit-decrypt + store in OPFS. `chunkDigest` is the per-chunk SHA-256 digest (verified by `decryptReceivedChunk`). Distinct from the file-level SHA-512 digest in `verifyAndDecrypt`.
- `{id: number, type: 'verifyAndDecrypt', size: number, digest: Uint8Array, key: Uint8Array, nonce: Uint8Array}` — verify digest + file-level decrypt all chunks. Only the four fields needed for verification/decryption are sent — not the full `FileDescription`, which contains private replica keys that the Worker doesn't need.
- `{id: number, type: 'cleanup'}` — delete OPFS temp files
Worker → Main (fields marked `†` are Transferable):
- `{id: number, type: 'progress', done: number, total: number}` — encryption/decryption progress (fire-and-forget, no promise)
- `{id: number, type: 'encrypted', digest: Uint8Array, key: Uint8Array, nonce: Uint8Array, chunkSizes: number[]}` — all fields structured-cloned (not transferred)
- `{id: number, type: 'chunk', data†: ArrayBuffer}` — readChunk response
- `{id: number, type: 'stored'}` — decryptAndStore acknowledgment
- `{id: number, type: 'decrypted', header: FileHeader, content†: ArrayBuffer}` — verifyAndDecrypt response
- `{id: number, type: 'cleaned'}`
- `{id: number, type: 'error', message: string}` — rejects the pending promise for this `id`
All messages carrying large `ArrayBuffer` payloads use `postMessage(msg, [transferables])` to transfer ownership instead of structured-clone copying. Only `ArrayBuffer` can be transferred — `Uint8Array`, `number[]`, and other types are always structured-cloned. This applies to: `encrypt` request (`data`), `readChunk` response (`data`), `decryptAndStoreChunk` request (`body`), and `verifyAndDecrypt` response (`content`). The `WorkerBackend` implementation must ensure the transferred `ArrayBuffer` covers the full `Uint8Array` — if `byteOffset !== 0` or `byteLength !== buffer.byteLength`, slice first: `data.buffer.slice(data.byteOffset, data.byteOffset + data.byteLength)`. This is required for `decryptAndStore` request bodies: `sendXFTPCommand` returns `body = fullResp.subarray(XFTP_BLOCK_SIZE)`, which has `byteOffset = XFTP_BLOCK_SIZE`. Other payloads are full-buffer views (§6 step 3 creates `new Uint8Array(await file.arrayBuffer())`; Worker responses allocate fresh buffers) but `WorkerBackend` should guard unconditionally.
### 3.4 Worker internals
**Imports:** The Worker imports directly from `libsodium-wrappers-sumo` (for `await sodium.ready`), `src/crypto/file.js` (`encryptFile`, `encodeFileHeader`, `decryptChunks`), `src/crypto/digest.js` (`sha512`), `src/protocol/chunks.js` (`prepareChunkSizes`, `fileSizeLen`, `authTagSize`), `src/protocol/encoding.js` (`concatBytes`), and `src/download.js` (`decryptReceivedChunk`). `download.js` directly imports `src/protocol/client.js` (for `decryptTransportChunk`). These transitively pull in `src/crypto/secretbox.js`, `src/crypto/keys.js`, and `src/crypto/padding.js`. None of these import `src/agent.ts` or `src/client.ts` — those pull in `node:http2` via dynamic import which would break Worker bundling. Vite tree-shakes the transitive deps automatically. Note: `download.js``protocol/client.js``crypto/keys.js` transitively pulls in `@noble/curves` (~50-80KB). This is unavoidable since `decryptTransportChunk` needs `dh` from `keys.js`. If Worker bundle size becomes a concern, `decryptReceivedChunk` could be refactored out of `download.js` into a separate module that doesn't import `protocol/client.js`.
**ArrayBuffer → Uint8Array conversion:** All Transferable fields arrive in the Worker as `ArrayBuffer`. The Worker's message handler must wrap them before passing to library functions: `new Uint8Array(msg.data)` for encrypt, `new Uint8Array(msg.body)` for decryptAndStore. Non-transferred fields (`dhSecret`, `nonce`, `digest`, `chunkSizes`) arrive as their original types (`Uint8Array` / `number[]`) via structured clone.
The Worker's encrypt handler calls the same functions as `encryptFileForUpload` in agent.ts (key/nonce generation → `encryptFile``sha512``prepareChunkSizes`). This is not reimplementation — it's calling the same library functions from a different entry point.
**Libsodium init:** Both the Worker and the main thread must `await sodium.ready` before calling any crypto functions that use libsodium. The Worker does this once on startup before processing messages. The main thread needs it before `connectXFTP` (which uses libsodium via `verifyIdentityProof`) and before `downloadXFTPChunkRaw` (which uses libsodium via `generateX25519KeyPair` + `dh`). In practice, `main.ts` calls `await sodium.ready` at page load, before any XFTP calls.
Encrypt (mirrors `encryptFileForUpload` in agent.ts):
1. Generate key (32B) + nonce (24B) via `crypto.getRandomValues`
2. `fileHdr = encodeFileHeader({fileName, fileExtra: null})`
3. `fileSize = BigInt(fileHdr.length + source.length)`
4. `payloadSize = Number(fileSize) + fileSizeLen + authTagSize`
5. `chunkSizes = prepareChunkSizes(payloadSize)`
6. `encSize = BigInt(chunkSizes.reduce((a, b) => a + b, 0))`
7. `encData = encryptFile(source, fileHdr, key, nonce, fileSize, encSize)`
8. `digest = sha512(encData)` — note: the `digest` field comment in `FileDescription` in `description.ts` says "SHA-256" but the actual hash is SHA-512 everywhere (`sha512` in agent.ts and download.ts). Fix the comment during implementation.
9. Open OPFS upload file via `createSyncAccessHandle`, write `encData`, flush, close handle. Null out `encData` reference.
10. Reopen the same OPFS file with `createSyncAccessHandle` as a persistent read handle (stored on the Worker module scope). This handle is used by all subsequent `readChunk` calls and closed on `cleanup`.
11. Post back `{digest, key, nonce, chunkSizes}` (no encData transfer — data stays in OPFS)
readChunk:
- Use the persistent read handle: `handle.read(buf, {at: offset})` → return slice as transferable ArrayBuffer. OPFS allows only one `FileSystemSyncAccessHandle` per file; the persistent handle avoids per-call open/close overhead.
decryptAndStoreChunk (removes transport encryption only — stored data is still file-level encrypted):
1. `decryptReceivedChunk(dhSecret, nonce, new Uint8Array(body), chunkDigest)` → transit-decrypted chunk data (still file-level encrypted — only the transport layer is removed). Argument order matches signature `(dhSecret, cbNonce, encData, expectedDigest)` from download.ts. `body` arrives as `ArrayBuffer` via Transferable and must be wrapped; `dhSecret`, `nonce`, `chunkDigest` arrive as `Uint8Array` via structured clone.
2. On first call, open the OPFS download temp file via `createSyncAccessHandle` and store as a persistent write handle. Record `{chunkNo, size: decrypted.length}` in an in-memory `chunkMeta: Map<number, {offset: number, size: number}>` — offset is the running sum of sizes for chunks stored so far (chunks may arrive out of order with `concurrency > 1`, so offset is assigned as `currentFileOffset`, then `currentFileOffset += size`)
3. Write decrypted chunk to the persistent handle at the recorded offset
verifyAndDecrypt (mirrors size/digest checks in agent.ts `downloadFile`):
1. Close the persistent download write handle (flush first), then reopen as a read handle. Read each chunk from OPFS into a `Uint8Array[]` array, ordered by `chunkNo`: for each entry in `chunkMeta` sorted by `chunkNo`, `handle.read(buf, {at: offset})` with the recorded offset and size
2. Concatenate for verification: `combined = concatBytes(...chunks)`
3. Verify total size: `combined.length === params.size`
4. Verify SHA-512 digest: `sha512(combined)` matches `params.digest`
5. Decrypt: `decryptChunks(BigInt(params.size), chunks, params.key, params.nonce)``params.size` is the encrypted file size (`fd.size` = `sum(chunkSizes)` = `decryptChunks`' first param `encSize`). Called directly instead of via `processDownloadedFile` (which expects a full `FileDescription`). Pass the original `chunks` array (not `combined`), as `decryptChunks` handles concatenation internally.
6. Delete OPFS download temp file
7. Return `{header, content}` via transferable ArrayBuffer
### 3.5 Browser requirements
The page requires a modern browser with Web Worker and OPFS support:
- Chrome 102+, Firefox 114+, Safari 15.2+ (Workers + OPFS + ES module Workers — Firefox added module Worker support in 114)
- If Worker or OPFS is unavailable, the page shows an error message rather than falling back silently.
No `DirectBackend` is needed — the page is browser-only, and tests run in vitest browser mode (real Chromium). The existing library tests (`test/browser.test.ts`) test the crypto/upload/download pipeline directly without Workers.
## 4. Server Configuration
### 4.1 Server lists
`web/servers.json` — single source of truth for preset server addresses (imported by both `servers.ts` and `vite.config.ts`):
```json
{
"simplex": [
"xftp://da1aH3nOT-9G8lV7bWamhxpDYdJ1xmW7j3JpGaDR5Ug=@xftp1.simplex.im",
"xftp://5vog2Imy1ExJB_7zDZrkV1KDWi96jYFyy9CL6fndBVw=@xftp2.simplex.im",
"xftp://PYa32DdYNFWi0uZZOprWQoQpIk5qyjRJ3EF7bVpbsn8=@xftp3.simplex.im",
"xftp://k_GgQl40UZVV0Y4BX9ZTyMVqX5ZewcLW0waQIl7AYDE=@xftp4.simplex.im",
"xftp://-bIo6o8wuVc4wpZkZD3tH-rCeYaeER_0lz1ffQcSJDs=@xftp5.simplex.im",
"xftp://6nSvtY9pJn6PXWTAIMNl95E1Kk1vD7FM2TeOA64CFLg=@xftp6.simplex.im"
],
"flux": [
"xftp://92Sctlc09vHl_nAqF2min88zKyjdYJ9mgxRCJns5K2U=@xftp1.simplexonflux.com",
"xftp://YBXy4f5zU1CEhnbbCzVWTNVNsaETcAGmYqGNxHntiE8=@xftp2.simplexonflux.com",
"xftp://ARQO74ZSvv2OrulRF3CdgwPz_AMy27r0phtLSq5b664=@xftp3.simplexonflux.com",
"xftp://ub2jmAa9U0uQCy90O-fSUNaYCj6sdhl49Jh3VpNXP58=@xftp4.simplexonflux.com",
"xftp://Rh19D5e4Eez37DEE9hAlXDB3gZa1BdFYJTPgJWPO9OI=@xftp5.simplexonflux.com",
"xftp://0AznwoyfX8Od9T_acp1QeeKtxUi676IBIiQjXVwbdyU=@xftp6.simplexonflux.com"
]
}
```
`web/servers.ts`:
```typescript
import {parseXFTPServer, type XFTPServer} from '../src/protocol/address.js'
import presets from './servers.json'
declare const __XFTP_SERVERS__: string[]
const serverAddresses: string[] = typeof __XFTP_SERVERS__ !== 'undefined'
? __XFTP_SERVERS__
: [...presets.simplex, ...presets.flux]
export function getServers(): XFTPServer[] {
return serverAddresses.map(parseXFTPServer)
}
export function pickRandomServer(servers: XFTPServer[]): XFTPServer {
return servers[Math.floor(Math.random() * servers.length)]
}
```
### 4.2 Build-time injection
`vite.config.ts` defines `__XFTP_SERVERS__`:
- `mode === 'local'`: `["xftp://<test-fingerprint>@localhost:7000"]`
- `mode === 'production'`: not defined → falls through to hardcoded list
### 4.3 Assumption
Production XFTP servers must have `[WEB]` section configured with a CA-signed certificate for browser TLS. Without this, browsers will reject the self-signed XFTP identity cert. The local test server uses `tests/fixtures/` certs which Chromium accepts via `ignoreHTTPSErrors`.
## 5. Page Structure & UI
### 5.1 Routing
`main.ts` checks `window.location.hash` once on page load:
- Hash present → download mode
- Hash absent → upload mode
No `hashchange` listener — the shareable link opens in a new tab. Simple page-load routing.
### 5.2 Upload UI states
1. **Landing**: Drag-drop zone centered, file picker button, size limit note
2. **Uploading**: Circular progress (canvas), percentage, cancel button
3. **Complete**: Shareable link (input + copy button), "Install SimpleX" CTA
4. **Error**: Error message + retry button. On server-unreachable, auto-retry with exponential backoff (1s, 2s, 4s, up to 3 attempts) before showing the error state.
### 5.3 Download UI states
1. **Ready**: Approximate file size displayed (encrypted size from `fd.size` or `fd.redirect.size` — see §7 step 2; file name is unavailable — it's inside the encrypted content), download button
2. **Downloading**: Circular progress, percentage
3. **Complete**: Browser save dialog triggered automatically
4. **Error**: Error message (expired, corrupted, unreachable)
### 5.4 Security summary (RFC §7.4)
Both upload-complete and download-ready states display a brief non-technical security summary:
- Files are encrypted in the browser before upload — the server never sees file contents.
- The link contains the decryption key in the hash fragment, which the browser never sends to any server.
- For maximum security, use the SimpleX app.
### 5.5 File expiry
Display on upload-complete state: "Files are typically available for 48 hours." This is an approximation — actual expiry depends on each XFTP server's `[STORE_LOG]` retention configuration. The 48-hour figure matches the current preset server defaults.
### 5.6 Styling
Plain CSS, no framework. White background, centered content, responsive. Circular progress via `<canvas>` (arc drawing, percentage text in center).
File size limit: 100MB. Displayed on upload page.
### 5.7 CSP
`index.html` includes a `<meta>` Content-Security-Policy tag with a build-time placeholder:
```html
<meta http-equiv="Content-Security-Policy"
content="default-src 'self'; worker-src 'self' blob:; style-src 'self' 'unsafe-inline'; connect-src __CSP_CONNECT_SRC__;">
```
Vite's `transformIndexHtml` hook (in `vite.config.ts`) replaces `__CSP_CONNECT_SRC__` at build time with origins derived from the server list:
- Local mode: `https://localhost:7000`
- Production: `https://xftp1.simplex.im:443 https://xftp2.simplex.im:443 ...` (all 12 servers)
## 6. Upload Flow
`web/upload.ts`:
1. User drops/picks file → `File` object
2. Validate `file.size <= 100 * 1024 * 1024` — show error if exceeded
3. Read file: `new Uint8Array(await file.arrayBuffer())` — note: after `backend.encrypt()` transfers the buffer to the Worker, `fileData` is detached (zero-length). Peak memory is ~2× file size (main thread holds original until transfer, Worker holds encrypted copy before OPFS write). Acceptable for the 100MB limit; do not raise the limit without considering memory implications.
4. Create `CryptoBackend` via factory
5. Create `XFTPClientAgent`
6. `backend.encrypt(fileData, file.name, onProgress)``EncryptResult`
- Encryption progress shown on canvas (Worker posts progress messages)
7. Pick one random server from configured list (V1: all chunks to same server)
8. Call `uploadFile(agent, server, metadata, {onProgress, readChunk: (off, sz) => backend.readChunk(off, sz)})`:
- `metadata` = `{digest, key, nonce, chunkSizes}` from EncryptResult
- Network progress shown on canvas
- Returns `{rcvDescription, sndDescription, uri}`
9. Construct full URL: `window.location.origin + window.location.pathname + '#' + uri`
10. Display link, copy button
11. Cleanup: `backend.cleanup()`, `closeXFTPAgent(agent)`
**Cancel:** User can abort via cancel button. Sets an `AbortController` signal that:
- Sends `{type: 'cleanup'}` to Worker
- Closes the XFTPClientAgent (drops HTTP/2 connections)
- Resets UI to landing state
## 7. Download Flow
`web/download.ts`:
1. Parse `window.location.hash.slice(1)``decodeDescriptionURI(fragment)``FileDescription`
2. Display file size (`fd.size` bytes, formatted human-readable). Note: `fd.size` is the encrypted size (slightly larger than plaintext due to padding + auth tag). The plaintext size is not available until decryption — display it as an approximate file size. If `fd.redirect !== null`, size comes from `fd.redirect.size` (which is the inner encrypted size).
3. User clicks "Download"
4. Create `CryptoBackend` and `XFTPClientAgent`
5. Call `downloadFileRaw(agent, fd, onRawChunk, {onProgress, concurrency: 3})`:
- `onRawChunk` forwards each raw chunk to the Worker: `backend.decryptAndStoreChunk(raw.dhSecret, raw.nonce, raw.body, raw.digest, raw.chunkNo)`
- `downloadFileRaw` handles redirect resolution internally (see §7.1), parallel downloads, and connection pooling
- Returns the resolved `FileDescription` (inner fd for redirect case, original fd otherwise)
6. `backend.verifyAndDecrypt({size: resolvedFd.size, digest: resolvedFd.digest, key: resolvedFd.key, nonce: resolvedFd.nonce})``{header, content}`
- Verifies size + SHA-512 digest + file-level decryption inside Worker. Only the four needed fields are sent — private replica keys stay on the main thread.
7. ACK: `ackFileChunks(agent, resolvedFd)` — best-effort, after verification succeeds
8. Sanitize `header.fileName` before use: strip path separators (`/`, `\`), replace null/control characters (U+0000-U+001F, U+007F), strip Unicode bidi override characters (U+202A-U+202E, U+2066-U+2069 — prevents `doc.pdf.exe` appearing as `doc.exe.pdf`), limit length to 255 chars. The filename is user-controlled (set by the uploader) and arrives via decrypted content. Then trigger browser save: `new Blob([content])``<a download="${sanitizedName}">` click
9. Cleanup: `backend.cleanup()`, `closeXFTPAgent(agent)`
### 7.1 Redirect handling
Handled inside `downloadFileRaw` in agent.ts — the web page doesn't see it. When `fd.redirect !== null`:
1. Download redirect chunks via `downloadXFTPChunkRaw` (parallel, same as regular chunks)
2. Transit-decrypt + verify + file-level decrypt on main thread (redirect data is always small — a few KB of YAML, so main thread decryption is fine)
3. Parse YAML → inner `FileDescription`, validate against `fd.redirect.{size, digest}`
4. ACK redirect chunks (best-effort)
5. Continue downloading inner description's chunks, calling `onRawChunk` for each
### 7.2 Architecture note: download refactoring
Both upload and download use `agent.ts` for orchestration. The key difference is where the crypto/network split happens:
- **Upload**: agent.ts reads encrypted chunks from the Worker via `readChunk` callback, sends them over the network.
- **Download**: agent.ts receives raw encrypted responses from the network via `downloadXFTPChunkRaw` (DH key exchange + network only, no decryption), passes them to the web page via `onRawChunk` callback, which routes them to the Worker for transit decryption.
This split keeps all expensive crypto off the main thread. Transit decryption uses a custom JS Salsa20 implementation (`xorKeystream` in secretbox.ts) that would block the UI for ~50-200ms on a 4MB chunk. File-level decryption (`decryptChunks`) is similarly expensive. Both happen in the Worker.
The cheap operations stay on the main thread: DH key exchange (`generateX25519KeyPair` + `dh` — ~1ms via libsodium WASM), XFTP command encoding/decoding, connection management.
## 8. Build & Dev Setup
### 8.1 vite.config.ts (new, separate from vitest.config.ts)
```typescript
import {defineConfig, type Plugin} from 'vite'
import {readFileSync} from 'fs'
import {createHash} from 'crypto'
import presets from './web/servers.json'
function parseHost(addr: string): string {
const m = addr.match(/@(.+)$/)
if (!m) throw new Error('bad server address: ' + addr)
const host = m[1].split(',')[0]
return host.includes(':') ? host : host + ':443'
}
function cspPlugin(servers: string[]): Plugin {
const origins = servers.map(s => 'https://' + parseHost(s)).join(' ')
return {
name: 'csp-connect-src',
transformIndexHtml: {
order: 'pre',
handler(html, ctx) {
if (ctx.server) {
// Dev mode: remove CSP meta tag entirely — Vite HMR needs inline scripts
return html.replace(/<meta\s[^>]*?Content-Security-Policy[\s\S]*?>/i, '')
}
return html.replace('__CSP_CONNECT_SRC__', origins)
}
}
}
}
export default defineConfig(({mode}) => {
const define: Record<string, string> = {}
let servers: string[]
if (mode === 'local') {
const pem = readFileSync('../tests/fixtures/ca.crt', 'utf-8')
const der = Buffer.from(pem.replace(/-----[^-]+-----/g, '').replace(/\s/g, ''), 'base64')
const fp = createHash('sha256').update(der).digest('base64')
.replace(/\+/g, '-').replace(/\//g, '_')
servers = [`xftp://${fp}@localhost:7000`]
define['__XFTP_SERVERS__'] = JSON.stringify(servers)
} else {
servers = [...presets.simplex, ...presets.flux]
}
return {
root: 'web',
build: {outDir: '../dist-web'},
define,
worker: {format: 'es'},
plugins: [cspPlugin(servers)],
}
})
```
### 8.2 package.json scripts
```json
"dev": "vite --mode local",
"build:local": "vite build --mode local",
"build:prod": "vite build --mode production",
"preview": "vite preview",
"check:web": "tsc -p tsconfig.web.json --noEmit && tsc -p tsconfig.worker.json --noEmit"
```
Note: `check:web` type-checks `src/` twice (once per config) — acceptable for this small library.
Add `vite` as an explicit devDependency (`^6.0.0` — matching the version vitest 3.x depends on transitively). Relying on transitive resolution is fragile across package managers.
### 8.3 TypeScript configuration
The existing `tsconfig.json` has `rootDir: "src"` and `include: ["src/**/*.ts"]` — this is for library compilation only (output to `dist/`). Vite handles `web/` TypeScript compilation independently via esbuild, so the main tsconfig is unchanged. `web/*.ts` files import from `../src/*.js` using relative paths.
Add two tsconfigs for `web/` type-checking — split by environment to avoid type pollution between DOM and WebWorker globals:
`tsconfig.web.json` — main-thread files (DOM globals: `document`, `window`, etc.):
```json
{
"extends": "./tsconfig.json",
"compilerOptions": {
"rootDir": ".",
"noEmit": true,
"types": [],
"moduleResolution": "bundler",
"lib": ["ES2022", "DOM"]
},
"include": ["web/**/*.ts", "src/**/*.ts"],
"exclude": ["web/crypto.worker.ts"]
}
```
`tsconfig.worker.json` — Worker file (`self`, `FileSystemSyncAccessHandle`, etc.):
```json
{
"extends": "./tsconfig.json",
"compilerOptions": {
"rootDir": ".",
"noEmit": true,
"types": [],
"moduleResolution": "bundler",
"lib": ["ES2022", "WebWorker"]
},
"include": ["web/crypto.worker.ts", "src/**/*.ts"]
}
```
Both configs set `"types": []` to prevent auto-inclusion of `@types/node` and `"moduleResolution": "bundler"` for Vite-compatible resolution (JSON imports, `.js` extension mapping). The base config's `"moduleResolution": "node"` would cause false type errors on `import ... from './servers.json'`. Both override `@types/node`, which would pollute DOM/WebWorker environments with Node.js globals (`process`, `Buffer`, etc.). This means `src/client.ts`'s dynamic `import("node:http2")` will produce a type error in these configs. This is acceptable — `src/client.ts` provides `createNodeTransport` which is never used in browser code (Vite tree-shakes it out), and full `src/` type-checking is handled by the base `tsconfig.json`. If the error is distracting, add `src/client.ts` to both configs' `exclude` arrays.
Both extend the library tsconfig (inheriting `strict`, `module`, etc.) and include `src/**/*.ts` so imports from `../src/*.js` resolve. `"noEmit": true` means they're only used for type-checking — Vite handles actual compilation. The inherited `"exclude": ["node_modules", "dist", "test"]` intentionally excludes `test/` — test files are type-checked by their own vitest/playwright configs, not by `check:web`.
### 8.4 Dev workflow
`npm run dev` → Vite dev server at `localhost:5173`, configured for local test server. Start `xftp-server` on port 7000 separately (or via the existing globalSetup).
Note: The CSP meta tag's `default-src 'self'` blocks Vite's injected HMR inline scripts in dev mode. The `cspPlugin` handles this by removing the entire CSP `<meta>` tag in serve mode (dev server), so HMR works without restrictions. Production builds always have the correct CSP.
## 9. Library Changes (agent.ts + client.ts)
Changes to support the web page: upload `readChunk` callback, download `onRawChunk` callback with parallel chunk downloads.
### 9.1 Type changes
Split the existing `EncryptedFileInfo` (which currently has `encData`, `digest`, `key`, `nonce`, `chunkSizes` as direct fields) into a metadata-only base and an extension:
```typescript
// Metadata-only variant (no encData — data lives in Worker/OPFS)
export interface EncryptedFileMetadata {
digest: Uint8Array
key: Uint8Array
nonce: Uint8Array
chunkSizes: number[]
}
// Full variant (existing, extends metadata with data)
export interface EncryptedFileInfo extends EncryptedFileMetadata {
encData: Uint8Array
}
```
### 9.2 uploadFile signature change
Replace positional optional params with an options bag. Add optional `readChunk`. When provided, `encrypted.encData` is not accessed.
```typescript
export interface UploadOptions {
onProgress?: (uploaded: number, total: number) => void
redirectThreshold?: number
readChunk?: (offset: number, size: number) => Promise<Uint8Array>
}
export async function uploadFile(
agent: XFTPClientAgent,
server: XFTPServer,
encrypted: EncryptedFileMetadata,
options?: UploadOptions
): Promise<UploadResult>
```
Inside `uploadFile`:
- Chunk read: if `options?.readChunk` is provided, use it. Otherwise, verify `'encData' in encrypted` at runtime (throws `"uploadFile: readChunk required when encData is absent"` if missing), then use `(off, sz) => Promise.resolve((encrypted as EncryptedFileInfo).encData.subarray(off, off + sz))`. This guards against calling `uploadFile` with `EncryptedFileMetadata` but no `readChunk`. For each chunk, call `readChunk(offset, size)` once and use the returned `Uint8Array` for both `getChunkDigest(chunkData)` and `uploadXFTPChunk(..., chunkData)` — do not call `readChunk` twice per chunk.
- Progress total: `const total = encrypted.chunkSizes.reduce((a, b) => a + b, 0)` — replaces `encrypted.encData.length` (line 129) since `EncryptedFileMetadata` has no `encData`. The values are identical: `encData.length === sum(chunkSizes)`.
- `buildDescription` parameter type: change from `EncryptedFileInfo` to `EncryptedFileMetadata` — it only accesses `chunkSizes`, `digest`, `key`, `nonce` (not `encData`).
`uploadRedirectDescription` (internal) is unchanged — redirect descriptions are always small and created in-memory by `encryptFileForUpload`.
### 9.3 Backward compatibility
The signature change from positional params `(agent, server, encrypted, onProgress?, redirectThreshold?)` to `(agent, server, encrypted, options?)` is a breaking change for callers that pass `onProgress` or `redirectThreshold`. In practice, the only callers are the browser test (which passes no options — no change needed) and the web page (new code). `EncryptedFileInfo` extends `EncryptedFileMetadata`, so existing callers that pass `EncryptedFileInfo` work without change.
### 9.4 client.ts: downloadXFTPChunkRaw
Split `downloadXFTPChunk` at the network/crypto boundary. The new function does DH key exchange and network I/O but skips transit decryption:
```typescript
export interface RawChunkResponse {
dhSecret: Uint8Array
nonce: Uint8Array
body: Uint8Array
}
export async function downloadXFTPChunkRaw(
c: XFTPClient, rpKey: Uint8Array, fId: Uint8Array
): Promise<RawChunkResponse> {
const {publicKey, privateKey} = generateX25519KeyPair()
const cmd = encodeFGET(encodePubKeyX25519(publicKey))
const {response, body} = await sendXFTPCommand(c, rpKey, fId, cmd)
if (response.type !== "FRFile") throw new Error("unexpected response: " + response.type)
const dhSecret = dh(response.rcvDhKey, privateKey)
return {dhSecret, nonce: response.nonce, body}
}
```
`RawChunkResponse` contains only what client.ts produces (DH secret, nonce, encrypted body). The chunk metadata (`chunkNo`, `digest`) is added by agent.ts when constructing `RawDownloadedChunk` (see §9.5).
The existing `downloadXFTPChunk` is refactored to call `downloadXFTPChunkRaw` + `decryptReceivedChunk`:
```typescript
export async function downloadXFTPChunk(
c: XFTPClient, rpKey: Uint8Array, fId: Uint8Array, digest?: Uint8Array
): Promise<Uint8Array> {
const {dhSecret, nonce, body} = await downloadXFTPChunkRaw(c, rpKey, fId)
return decryptReceivedChunk(dhSecret, nonce, body, digest ?? null)
}
```
### 9.5 agent.ts: downloadFileRaw, ackFileChunks, RawDownloadedChunk
New type combining client.ts's `RawChunkResponse` with chunk metadata from agent.ts:
```typescript
export interface RawDownloadedChunk {
chunkNo: number
dhSecret: Uint8Array
nonce: Uint8Array
body: Uint8Array
digest: Uint8Array
}
```
New function providing download orchestration with a raw chunk callback. Handles connection pooling, parallel downloads, redirect resolution, and progress. Does **not** ACK — the caller ACKs after verification.
```typescript
export interface DownloadRawOptions {
onProgress?: (downloaded: number, total: number) => void
concurrency?: number // max parallel chunk downloads, default 1
}
export async function downloadFileRaw(
agent: XFTPClientAgent,
fd: FileDescription,
onRawChunk: (chunk: RawDownloadedChunk) => Promise<void>,
options?: DownloadRawOptions
): Promise<FileDescription>
```
Returns the resolved `FileDescription` — for redirect files this is the inner fd, for non-redirect files this is the original fd. The caller uses this for verification and ACK.
Internal structure:
1. Validate `fd` via `validateFileDescription` (may double-validate if caller already validated via `decodeDescriptionURI` — harmless)
2. If `fd.redirect !== null`: resolve redirect on main thread (redirect data is small):
a. Download redirect chunks via `downloadXFTPChunk` (not raw — main thread decryption is fine for a few KB)
b. Verify size + digest, `processDownloadedFile` → YAML bytes
c. Parse inner `FileDescription`, validate against `fd.redirect.{size, digest}`
d. ACK redirect chunks (best-effort — redirect chunks are small and separate from the file chunks)
e. Replace `fd` with inner description
3. Pre-connect: call `getXFTPServerClient(agent, server)` for each unique server before launching concurrent workers. This ensures the client connection exists in the agent's map, avoiding a race condition where multiple concurrent workers all see the client as missing and each call `connectXFTP` independently (leaking all but the last connection). Known limitation: if a connection drops mid-download and multiple workers attempt reconnection simultaneously, the same TOCTOU race reappears. This is a pre-existing issue in `getXFTPServerClient`; a proper fix (per-key connection promise) is out of scope for this plan but should be tracked for follow-up.
4. Download file chunks in parallel (concurrency-limited via sliding window):
- Create a queue of chunk indices `[0, 1, ..., N-1]`. Launch `min(concurrency, N)` async workers, each pulling the next index from the queue until empty. Each worker loops: pull index → derive key → `getXFTPServerClient``downloadXFTPChunkRaw``await onRawChunk(...)` → update progress → next index. `await Promise.all(workers)` to wait for completion.
- For each chunk: derive key (`decodePrivKeyEd25519``ed25519KeyPairFromSeed`), get client (`getXFTPServerClient`), call `downloadXFTPChunkRaw`, `await onRawChunk(...)` with result + `chunkNo` + `chunk.digest`
- Each concurrency slot awaits its `onRawChunk` before starting the next download on that slot. With `concurrency > 1`, multiple `onRawChunk` calls may be in-flight concurrently (one per slot). The Worker handles this correctly — messages are queued and processed sequentially.
- Update progress after each chunk: `downloaded += chunk.chunkSize; onProgress?.(downloaded, resolvedFd.size)` — both values use encrypted sizes for consistency
5. Return the resolved `fd`
New helper for ACKing after verification:
```typescript
export async function ackFileChunks(
agent: XFTPClientAgent, fd: FileDescription
): Promise<void> {
for (const chunk of fd.chunks) {
const replica = chunk.replicas[0]
if (!replica) continue
try {
const client = await getXFTPServerClient(agent, parseXFTPServer(replica.server))
const seed = decodePrivKeyEd25519(replica.replicaKey)
const kp = ed25519KeyPairFromSeed(seed)
await ackXFTPChunk(client, kp.privateKey, replica.replicaId)
} catch (_) {}
}
}
```
The existing `downloadFile` is refactored to use `downloadFileRaw` internally:
```typescript
export async function downloadFile(
agent: XFTPClientAgent,
fd: FileDescription,
onProgress?: (downloaded: number, total: number) => void
): Promise<DownloadResult> {
const chunks: Uint8Array[] = []
const resolvedFd = await downloadFileRaw(agent, fd, async (raw) => {
chunks[raw.chunkNo - 1] = decryptReceivedChunk(
raw.dhSecret, raw.nonce, raw.body, raw.digest
)
}, {onProgress})
// verify + file-level decrypt using resolvedFd (inner fd for redirect case)
const combined = chunks.length === 1 ? chunks[0] : concatBytes(...chunks)
if (combined.length !== resolvedFd.size) throw new Error("downloadFile: file size mismatch")
const digest = sha512(combined)
if (!digestEqual(digest, resolvedFd.digest)) throw new Error("downloadFile: file digest mismatch")
// processDownloadedFile re-concatenates chunks internally — this mirrors the
// existing downloadFile pattern (verify on concatenated data, then pass chunks
// array to decryptChunks which concatenates again). Acceptable overhead for
// correctness: verification must happen on transit-decrypted data before
// file-level decryption transforms it.
const result = processDownloadedFile(resolvedFd, chunks)
await ackFileChunks(agent, resolvedFd)
return result
}
```
Existing callers retain serial behavior (`concurrency` defaults to 1). The web page opts into parallelism by passing `concurrency: 3`. The browser test (`test/browser.test.ts`) continues to work unchanged. The chunks array is initialized empty (`[]`) and populated by sparse index assignment (`chunks[raw.chunkNo - 1] = ...`), so it correctly handles both redirect and non-redirect cases regardless of the outer fd's chunk count. `digestEqual` is an existing module-private helper in agent.ts (line 327) that performs constant-time byte comparison.
### 9.6 Backward compatibility (download)
`downloadFile` signature is unchanged — existing callers are unaffected. The refactoring adds `downloadFileRaw`, `ackFileChunks`, and `RawDownloadedChunk` as new exports from agent.ts, and `downloadXFTPChunkRaw` + `RawChunkResponse` as new exports from client.ts.
## 10. Testing
### 10.1 Existing tests (unchanged)
- `npm run test:browser` — vitest browser round-trip (library-level)
- `cabal test --test-option='--match=/XFTP Web Client/'` — Haskell per-function tests
### 10.2 New: page E2E test
Add `test/page.spec.ts` using `@playwright/test` (not vitest browser mode — vitest tests run IN the browser and can't control page navigation; Playwright tests run in Node.js and control the browser). Add `@playwright/test` as a devDependency.
Add `playwright.config.ts` at the project root (`xftp-web/`):
- `webServer: { command: 'vite build --mode local && vite preview', url: 'http://localhost:4173', reuseExistingServer: !process.env.CI }` — the `url` property tells Playwright to wait until the preview server is ready before running tests
- `use.ignoreHTTPSErrors: true` (test server uses self-signed cert)
- `use.launchOptions: { args: ['--ignore-certificate-errors'] }` — required because Playwright's `ignoreHTTPSErrors` only affects page navigation, not `fetch()` calls from in-page JavaScript. Without this flag, the page's `createBrowserTransport` fetch to `https://localhost:7000` would fail TLS validation.
- `globalSetup`: `'./test/globalSetup.ts'` (starts xftp-server, shared with vitest)
```typescript
import {test, expect} from '@playwright/test'
test('page upload + download round-trip', async ({page}) => {
await page.goto(PAGE_URL)
// Set file input via page.setInputFiles()
// Wait for upload link to appear: page.waitForSelector('[data-testid="share-link"]')
// Extract hash from link text
// Navigate to PAGE_URL + '#' + hash
// Wait for download complete state
// Verify file was offered for save (check download event)
})
```
Add script: `"test:page": "playwright test test/page.spec.ts"`
This tests the real bundle including Worker loading, OPFS, and CSP. The existing `test/browser.test.ts` continues to test the library-level pipeline (vitest browser mode, no Workers).
### 10.3 Manual testing
`npm run dev` → open `localhost:5173` in browser → drag file → get link → open link in new tab → download. Requires xftp-server running on port 7000 (local mode).
## 11. Files
**Create:**
- `xftp-web/web/index.html` — page entry point (includes CSP meta tag)
- `xftp-web/web/main.ts` — router + libsodium init
- `xftp-web/web/upload.ts` — upload UI + orchestration
- `xftp-web/web/download.ts` — download UI + orchestration
- `xftp-web/web/progress.ts` — circular progress canvas component
- `xftp-web/web/servers.json` — preset server addresses (shared by servers.ts and vite.config.ts)
- `xftp-web/web/servers.ts` — server configuration (imports servers.json)
- `xftp-web/web/crypto-backend.ts` — CryptoBackend interface + WorkerBackend + factory
- `xftp-web/web/crypto.worker.ts` — Web Worker implementation
- `xftp-web/web/style.css` — styles
- `xftp-web/vite.config.ts` — page build config (CSP generation, server list)
- `xftp-web/tsconfig.web.json` — IDE/CI type-checking for `web/` main-thread files (DOM)
- `xftp-web/tsconfig.worker.json` — IDE/CI type-checking for `web/crypto.worker.ts` (WebWorker)
- `xftp-web/playwright.config.ts` — Playwright E2E test config (webServer, globalSetup)
- `xftp-web/test/page.spec.ts` — page E2E test (Playwright)
**Modify:**
- `xftp-web/src/agent.ts` — add `EncryptedFileMetadata` type, `uploadFile` options bag with `readChunk`, `downloadFileRaw` with `onRawChunk` callback + parallel downloads, `ackFileChunks`, `RawDownloadedChunk` type, refactor `downloadFile` on top of `downloadFileRaw`, add `import {decryptReceivedChunk} from "./download.js"` (needed by refactored `downloadFile`)
- `xftp-web/src/client.ts` — add `downloadXFTPChunkRaw`, `RawChunkResponse` type, refactor `downloadXFTPChunk` to use raw variant
- `xftp-web/package.json` — add dev/build/check:web/test:page scripts, add `vite` + `@playwright/test` devDeps
- `xftp-web/src/protocol/description.ts` — fix stale "SHA-256" comment on `FileDescription.digest` to "SHA-512"
- `xftp-web/.gitignore` — add `dist-web/`
## 12. Implementation Order
1. **Library refactoring**`client.ts`: add `downloadXFTPChunkRaw`; `agent.ts`: add `downloadFileRaw` + parallel downloads, `uploadFile` options bag with `readChunk`; refactor existing `downloadFile` on top of `downloadFileRaw`. Run existing tests to verify no regressions.
2. **Vite config + HTML shell**`vite.config.ts`, `index.html`, `main.ts`, verify dev server works
3. **Server config**`servers.ts` with both local and production server lists
4. **CryptoBackend + Worker** — interface, WorkerBackend, Worker implementation, OPFS logic
5. **Upload flow**`upload.ts` with drag-drop, encrypt via Worker, upload via agent, show link
6. **Download flow**`download.ts` with URL parsing, download via agent `downloadFileRaw`, Worker decrypt, browser save
7. **Progress component**`progress.ts` canvas drawing
8. **Styling**`style.css`
9. **Testing** — page E2E test, manual browser verification
10. **Build scripts**`build:local`, `build:prod` in package.json
@@ -0,0 +1,53 @@
# XFTPClientAgent Pattern
## TOC
1. Executive Summary
2. Changes: client.ts
3. Changes: agent.ts
4. Changes: test/browser.test.ts
5. Verification
## Executive Summary
Add `XFTPClientAgent` — a per-server connection pool matching the Haskell pattern. The agent caches `XFTPClient` instances by server URL. All orchestration functions (`uploadFile`, `downloadFile`, `deleteFile`) take `agent` as first parameter and use `getXFTPServerClient(agent, server)` instead of calling `connectXFTP` directly. Connections stay open on success; the caller creates and closes the agent.
`connectXFTP` and `closeXFTP` stay exported (used by `XFTPWebTests.hs` Haskell tests). The `browserClients` hack, per-function `connections: Map`, and `getOrConnect` are deleted.
## Changes: client.ts
**Add** after types section: `XFTPClientAgent` interface, `newXFTPAgent`, `getXFTPServerClient`, `closeXFTPServerClient`, `closeXFTPAgent`.
**Delete**: `browserClients` Map and all `isNode` browser-cache checks in `connectXFTP` and `closeXFTP`.
**Revert `closeXFTP`** to unconditional `c.transport.close()` (browser transport.close() is already a no-op).
`connectXFTP` stays exported (backward compat) but becomes a raw low-level function — no caching.
## Changes: agent.ts
**Imports**: replace `connectXFTP`/`closeXFTP` with `getXFTPServerClient`/`closeXFTPAgent` etc.
**Re-export** from agent.ts: `newXFTPAgent`, `closeXFTPAgent`, `XFTPClientAgent`.
**`uploadFile`**: add `agent: XFTPClientAgent` as first param. Replace `connectXFTP``getXFTPServerClient`. Remove `finally { closeXFTP }`. Pass `agent` to `uploadRedirectDescription`.
**`uploadRedirectDescription`**: change from `(client, server, innerFd)` to `(agent, server, innerFd)`. Get client via `getXFTPServerClient`.
**`downloadFile`**: add `agent` param. Delete local `connections: Map`. Replace `getOrConnect``getXFTPServerClient`. Remove finally cleanup. Pass `agent` to `downloadWithRedirect`.
**`downloadWithRedirect`**: add `agent` param. Same replacements. Remove try/catch cleanup. Recursive call passes `agent`.
**`deleteFile`**: add `agent` param. Same pattern.
**Delete**: `getOrConnect` function entirely.
## Changes: test/browser.test.ts
Create agent before operations, pass to upload/download, close in finally.
## Verification
1. `npx vitest --run` — browser round-trip test passes
2. No remaining `browserClients`, `getOrConnect`, or per-function `connections: Map` locals
3. `connectXFTP` and `closeXFTP` still exported (XFTPWebTests.hs compat)
4. All orchestration functions take `agent` as first param
@@ -0,0 +1,859 @@
# XFTP Web Page E2E Tests Plan
## Table of Contents
1. [Executive Summary](#1-executive-summary)
2. [Test Infrastructure](#2-test-infrastructure)
3. [Test Infrastructure - Page Objects](#3-test-infrastructure---page-objects)
4. [Upload Flow Tests](#4-upload-flow-tests)
5. [Download Flow Tests](#5-download-flow-tests)
6. [Edge Cases](#6-edge-cases)
7. [Implementation Order](#7-implementation-order)
8. [Test Utilities](#8-test-utilities)
---
## 1. Executive Summary
This document specifies comprehensive Playwright E2E tests for the XFTP web page. The existing test (`page.spec.ts`) performs a basic upload/download round-trip. This plan extends coverage to:
- **Upload flow**: File selection (picker + drag-drop), validation, progress, cancellation, link sharing, error handling
- **Download flow**: Invalid link handling, download button, progress, file save, error states
- **Edge cases**: Boundary file sizes, special characters, network failures, multi-chunk files with redirect, UI information display
**Key constraints**:
- Tests run against a local XFTP server (started via `globalSetup.ts`)
- Server port is dynamic (read from `/tmp/xftp-test-server.port`)
- Browser uses `--ignore-certificate-errors` for self-signed certs
- OPFS and Web Workers are required (Chromium supports both)
**Test file location**: `/code/simplexmq/xftp-web/test/page.spec.ts`
**Architecture**: Tests use the Page Object Model pattern to encapsulate UI interactions, making tests read as domain-specific scenarios rather than raw Playwright API calls.
---
## 2. Test Infrastructure
### 2.1 Current Setup
```
xftp-web/
├── playwright.config.ts # Playwright config (webServer, globalSetup)
├── test/
│ ├── globalSetup.ts # Starts xftp-server, writes port to PORT_FILE
│ ├── page.spec.ts # E2E tests (to be extended)
│ └── pages/ # Page Objects (new)
│ ├── UploadPage.ts
│ └── DownloadPage.ts
```
### 2.2 Prerequisites
- `globalSetup.ts` starts the XFTP server and writes port to `PORT_FILE`
- Tests must read the port dynamically: `readFileSync(PORT_FILE, 'utf-8').trim()`
- Vite builds and serves the page at `http://localhost:4173`
---
## 3. Test Infrastructure - Page Objects
Page Objects encapsulate page-specific selectors and actions, providing a clean API for tests. This follows the standard Page Object Model pattern used in simplex-chat and most professional test suites.
### 3.1 UploadPage
```typescript
// test/pages/UploadPage.ts
import {Page, Locator, expect} from '@playwright/test'
export class UploadPage {
readonly page: Page
readonly dropZone: Locator
readonly fileInput: Locator
readonly progressStage: Locator
readonly progressCanvas: Locator
readonly statusText: Locator
readonly cancelButton: Locator
readonly completeStage: Locator
readonly shareLink: Locator
readonly copyButton: Locator
readonly errorStage: Locator
readonly errorMessage: Locator
readonly retryButton: Locator
readonly expiryNote: Locator
readonly securityNote: Locator
constructor(page: Page) {
this.page = page
this.dropZone = page.locator('#drop-zone')
this.fileInput = page.locator('#file-input')
this.progressStage = page.locator('#upload-progress')
this.progressCanvas = page.locator('#progress-container canvas')
this.statusText = page.locator('#upload-status')
this.cancelButton = page.locator('#cancel-btn')
this.completeStage = page.locator('#upload-complete')
this.shareLink = page.locator('[data-testid="share-link"]')
this.copyButton = page.locator('#copy-btn')
this.errorStage = page.locator('#upload-error')
this.errorMessage = page.locator('#error-msg')
this.retryButton = page.locator('#retry-btn')
this.expiryNote = page.locator('.expiry')
this.securityNote = page.locator('.security-note')
}
async goto() {
await this.page.goto('http://localhost:4173')
}
async selectFile(name: string, content: Buffer, mimeType = 'application/octet-stream') {
await this.fileInput.setInputFiles({name, mimeType, buffer: content})
}
async selectTextFile(name: string, content: string) {
await this.selectFile(name, Buffer.from(content, 'utf-8'), 'text/plain')
}
async selectLargeFile(name: string, sizeBytes: number) {
// Create large file in browser to avoid memory issues in test process
await this.page.evaluate(({name, size}) => {
const input = document.getElementById('file-input') as HTMLInputElement
const buffer = new ArrayBuffer(size)
new Uint8Array(buffer).fill(0x55)
const file = new File([buffer], name, {type: 'application/octet-stream'})
const dt = new DataTransfer()
dt.items.add(file)
input.files = dt.files
input.dispatchEvent(new Event('change', {bubbles: true}))
}, {name, size: sizeBytes})
}
async dragDropFile(name: string, content: Buffer) {
// Drag-drop uses same file input handler internally
await this.selectFile(name, content)
}
async waitForEncrypting(timeout = 10_000) {
await expect(this.statusText).toContainText('Encrypting', {timeout})
}
async waitForUploading(timeout = 30_000) {
await expect(this.statusText).toContainText('Uploading', {timeout})
}
async waitForShareLink(timeout = 60_000): Promise<string> {
await expect(this.shareLink).toBeVisible({timeout})
return await this.shareLink.inputValue()
}
async clickCopy() {
await this.copyButton.click()
await expect(this.copyButton).toContainText('Copied!')
}
async clickCancel() {
await this.cancelButton.click()
}
async clickRetry() {
await this.retryButton.click()
}
async expectError(messagePattern: string | RegExp) {
await expect(this.errorStage).toBeVisible()
await expect(this.errorMessage).toContainText(messagePattern)
}
async expectDropZoneVisible() {
await expect(this.dropZone).toBeVisible()
}
async expectProgressVisible() {
await expect(this.progressStage).toBeVisible()
await expect(this.progressCanvas).toBeVisible()
}
async expectCompleteWithExpiry() {
await expect(this.completeStage).toBeVisible()
await expect(this.expiryNote).toContainText('48 hours')
}
async expectSecurityNote() {
await expect(this.securityNote).toBeVisible()
await expect(this.securityNote).toContainText('encrypted')
}
getHashFromLink(url: string): string {
return new URL(url).hash
}
}
```
### 3.2 DownloadPage
```typescript
// test/pages/DownloadPage.ts
import {Page, Locator, expect, Download} from '@playwright/test'
export class DownloadPage {
readonly page: Page
readonly readyStage: Locator
readonly downloadButton: Locator
readonly progressStage: Locator
readonly progressCanvas: Locator
readonly statusText: Locator
readonly errorStage: Locator
readonly errorMessage: Locator
readonly retryButton: Locator
readonly securityNote: Locator
constructor(page: Page) {
this.page = page
this.readyStage = page.locator('#dl-ready')
this.downloadButton = page.locator('#dl-btn')
this.progressStage = page.locator('#dl-progress')
this.progressCanvas = page.locator('#dl-progress-container canvas')
this.statusText = page.locator('#dl-status')
this.errorStage = page.locator('#dl-error')
this.errorMessage = page.locator('#dl-error-msg')
this.retryButton = page.locator('#dl-retry-btn')
this.securityNote = page.locator('.security-note')
}
async goto(hash: string) {
await this.page.goto(`http://localhost:4173${hash}`)
}
async gotoWithLink(fullUrl: string) {
const hash = new URL(fullUrl).hash
await this.goto(hash)
}
async expectFileReady() {
await expect(this.readyStage).toBeVisible()
await expect(this.downloadButton).toBeVisible()
}
async expectFileSizeDisplayed() {
await expect(this.readyStage).toContainText(/\d+(?:\.\d+)?\s*(?:KB|MB|B)/)
}
async clickDownload(): Promise<Download> {
const downloadPromise = this.page.waitForEvent('download')
await this.downloadButton.click()
return downloadPromise
}
async waitForDownloading(timeout = 30_000) {
await expect(this.statusText).toContainText('Downloading', {timeout})
}
async waitForDecrypting(timeout = 30_000) {
await expect(this.statusText).toContainText('Decrypting', {timeout})
}
async expectProgressVisible() {
await expect(this.progressStage).toBeVisible()
await expect(this.progressCanvas).toBeVisible()
}
async expectInitialError(messagePattern: string | RegExp) {
// For malformed links - error shown in card without #dl-error stage
await expect(this.page.locator('.card .error')).toBeVisible()
await expect(this.page.locator('.card .error')).toContainText(messagePattern)
}
async expectRuntimeError(messagePattern: string | RegExp) {
// For runtime download errors - uses #dl-error stage
await expect(this.errorStage).toBeVisible()
await expect(this.errorMessage).toContainText(messagePattern)
}
async expectSecurityNote() {
await expect(this.securityNote).toBeVisible()
await expect(this.securityNote).toContainText('encrypted')
}
}
```
### 3.3 Test Fixtures
```typescript
// test/fixtures.ts
import {test as base} from '@playwright/test'
import {UploadPage} from './pages/UploadPage'
import {DownloadPage} from './pages/DownloadPage'
import {readFileSync} from 'fs'
// Extend Playwright test with page objects
export const test = base.extend<{
uploadPage: UploadPage
downloadPage: DownloadPage
}>({
uploadPage: async ({page}, use) => {
const uploadPage = new UploadPage(page)
await uploadPage.goto()
await use(uploadPage)
},
downloadPage: async ({page}, use) => {
await use(new DownloadPage(page))
},
})
export {expect} from '@playwright/test'
// Test data helpers
export function createTestContent(size: number, fill = 0x41): Buffer {
return Buffer.alloc(size, fill)
}
export function createTextContent(text: string): Buffer {
return Buffer.from(text, 'utf-8')
}
export function uniqueFileName(base: string, ext = 'txt'): string {
return `${base}-${Date.now()}.${ext}`
}
```
---
## 4. Upload Flow Tests
### 4.1 File Selection - File Picker Button
**Test ID**: `upload-file-picker`
```typescript
test('upload via file picker button', async ({uploadPage}) => {
await uploadPage.expectDropZoneVisible()
await uploadPage.selectTextFile('picker-test.txt', 'test content ' + Date.now())
await uploadPage.waitForEncrypting()
await uploadPage.waitForUploading()
const link = await uploadPage.waitForShareLink()
expect(link).toMatch(/^http:\/\/localhost:\d+\/#/)
})
```
### 4.2 File Selection - Drag and Drop
**Test ID**: `upload-drag-drop`
```typescript
test('upload via drag and drop', async ({uploadPage}) => {
await uploadPage.dragDropFile('dragdrop-test.txt', createTextContent('drag drop test'))
await uploadPage.expectProgressVisible()
const link = await uploadPage.waitForShareLink()
expect(link).toContain('#')
})
```
### 4.3 File Size Validation - Too Large
**Test ID**: `upload-file-too-large`
```typescript
test('upload rejects file over 100MB', async ({uploadPage}) => {
await uploadPage.selectLargeFile('large.bin', 100 * 1024 * 1024 + 1)
await uploadPage.expectError('too large')
await uploadPage.expectError('100 MB')
})
```
### 4.4 File Size Validation - Empty File
**Test ID**: `upload-file-empty`
```typescript
test('upload rejects empty file', async ({uploadPage}) => {
await uploadPage.selectFile('empty.txt', Buffer.alloc(0))
await uploadPage.expectError('empty')
})
```
### 4.5 Progress Display
**Test ID**: `upload-progress-display`
```typescript
test('upload shows progress during encryption and upload', async ({uploadPage}) => {
await uploadPage.selectFile('progress-test.bin', createTestContent(500 * 1024))
await uploadPage.expectProgressVisible()
await uploadPage.waitForEncrypting()
await uploadPage.waitForUploading()
await uploadPage.waitForShareLink()
})
```
### 4.6 Cancel Button
**Test ID**: `upload-cancel`
```typescript
test('cancel button aborts upload and returns to landing', async ({uploadPage}) => {
await uploadPage.selectFile('cancel-test.bin', createTestContent(1024 * 1024))
await uploadPage.expectProgressVisible()
await uploadPage.clickCancel()
await uploadPage.expectDropZoneVisible()
await expect(uploadPage.shareLink).toBeHidden()
})
```
### 4.7 Share Link Display and Copy
**Test ID**: `upload-share-link-copy`
```typescript
test('share link copy button works', async ({uploadPage, context}) => {
await context.grantPermissions(['clipboard-read', 'clipboard-write'])
await uploadPage.selectTextFile('copy-test.txt', 'copy test content')
const link = await uploadPage.waitForShareLink()
await uploadPage.clickCopy()
// Verify clipboard (may fail in headless)
try {
const clipboardText = await uploadPage.page.evaluate(() => navigator.clipboard.readText())
expect(clipboardText).toBe(link)
} catch {
// Clipboard API may not be available
}
})
```
### 4.8 Error Handling and Retry
**Test ID**: `upload-error-retry`
```typescript
test('error state shows retry button', async ({uploadPage}) => {
await uploadPage.selectFile('error-test.txt', Buffer.alloc(0))
await uploadPage.expectError('empty')
await expect(uploadPage.retryButton).toBeVisible()
})
```
---
## 5. Download Flow Tests
### 5.1 Invalid Link Handling - Malformed Hash
**Test ID**: `download-invalid-hash-malformed`
```typescript
test('download shows error for malformed hash', async ({downloadPage}) => {
await downloadPage.goto('#not-valid-base64!!!')
await downloadPage.expectInitialError(/[Ii]nvalid|corrupted/)
await expect(downloadPage.downloadButton).not.toBeVisible()
})
```
### 5.2 Invalid Link Handling - Valid Base64 but Invalid Structure
**Test ID**: `download-invalid-hash-structure`
```typescript
test('download shows error for invalid structure', async ({downloadPage}) => {
await downloadPage.goto('#AAAA')
await downloadPage.expectInitialError(/[Ii]nvalid|corrupted/)
})
```
### 5.3 Download Button Click
**Test ID**: `download-button-click`
```typescript
test('download button initiates download', async ({uploadPage, downloadPage}) => {
// Upload first
await uploadPage.selectTextFile('dl-btn-test.txt', 'download test content')
const link = await uploadPage.waitForShareLink()
// Navigate to download
await downloadPage.gotoWithLink(link)
await downloadPage.expectFileReady()
// Click download
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe('dl-btn-test.txt')
})
```
### 5.4 Progress Display
**Test ID**: `download-progress-display`
```typescript
test('download shows progress', async ({uploadPage, downloadPage}) => {
await uploadPage.selectFile('dl-progress.bin', createTestContent(500 * 1024))
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const downloadPromise = downloadPage.clickDownload()
await downloadPage.expectProgressVisible()
await downloadPage.waitForDownloading()
await downloadPromise
})
```
### 5.5 File Save Verification
**Test ID**: `download-file-save`
```typescript
test('downloaded file content matches upload', async ({uploadPage, downloadPage}) => {
const content = 'verification content ' + Date.now()
const fileName = 'verify.txt'
await uploadPage.selectTextFile(fileName, content)
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe(fileName)
const path = await download.path()
if (path) {
const downloadedContent = (await import('fs')).readFileSync(path, 'utf-8')
expect(downloadedContent).toBe(content)
}
})
```
---
## 6. Edge Cases
### 6.1 Very Small Files
**Test ID**: `edge-small-file`
```typescript
test('upload and download 1-byte file', async ({uploadPage, downloadPage}) => {
await uploadPage.selectFile('tiny.bin', Buffer.from([0x42]))
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe('tiny.bin')
const path = await download.path()
if (path) {
const content = (await import('fs')).readFileSync(path)
expect(content.length).toBe(1)
expect(content[0]).toBe(0x42)
}
})
```
### 6.2 Files Near 100MB Limit
**Test ID**: `edge-near-limit`
```typescript
test.slow()
test('upload file at exactly 100MB', async ({uploadPage}) => {
await uploadPage.selectLargeFile('exactly-100mb.bin', 100 * 1024 * 1024)
// Should succeed (not show error)
await expect(uploadPage.errorStage).toBeHidden({timeout: 5000})
await uploadPage.expectProgressVisible()
// Wait for completion (may take a while)
await uploadPage.waitForShareLink(300_000)
})
```
### 6.3 Special Characters in Filename
**Test ID**: `edge-special-chars-filename`
```typescript
test('upload and download file with unicode filename', async ({uploadPage, downloadPage}) => {
const fileName = 'test-\u4e2d\u6587-\u0420\u0443\u0441\u0441\u043a\u0438\u0439.txt'
await uploadPage.selectTextFile(fileName, 'unicode filename test')
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe(fileName)
})
test('upload and download file with spaces', async ({uploadPage, downloadPage}) => {
const fileName = 'my document (final) v2.txt'
await uploadPage.selectTextFile(fileName, 'spaces test')
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe(fileName)
})
test('filename with path separators is sanitized', async ({uploadPage, downloadPage}) => {
await uploadPage.selectTextFile('../../../etc/passwd', 'path traversal test')
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).not.toContain('/')
expect(download.suggestedFilename()).not.toContain('\\')
})
```
### 6.4 Network Errors (Mocked)
**Test ID**: `edge-network-error`
```typescript
test('upload handles network error gracefully', async ({uploadPage}) => {
// Intercept and abort POST requests
await uploadPage.page.route('**/localhost:*', route => {
if (route.request().method() === 'POST') {
route.abort('failed')
} else {
route.continue()
}
})
await uploadPage.selectTextFile('network-error.txt', 'network error test')
await uploadPage.expectError(/.+/) // Any error message
})
```
### 6.5 Binary File Content Integrity
**Test ID**: `edge-binary-content`
```typescript
test('binary file with all byte values', async ({uploadPage, downloadPage}) => {
// Create buffer with all 256 byte values
const buffer = Buffer.alloc(256)
for (let i = 0; i < 256; i++) buffer[i] = i
await uploadPage.selectFile('all-bytes.bin', buffer)
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
const path = await download.path()
if (path) {
const content = (await import('fs')).readFileSync(path)
expect(content.length).toBe(256)
for (let i = 0; i < 256; i++) {
expect(content[i]).toBe(i)
}
}
})
```
### 6.6 Multiple Concurrent Downloads
**Test ID**: `edge-concurrent-downloads`
```typescript
test('concurrent downloads from same link', async ({browser}) => {
const context = await browser.newContext({ignoreHTTPSErrors: true})
const page1 = await context.newPage()
const upload = new UploadPage(page1)
await upload.goto()
await upload.selectTextFile('concurrent.txt', 'concurrent download test')
const link = await upload.waitForShareLink()
const hash = upload.getHashFromLink(link)
// Open two tabs and download concurrently
const page2 = await context.newPage()
const page3 = await context.newPage()
const dl2 = new DownloadPage(page2)
const dl3 = new DownloadPage(page3)
await dl2.goto(hash)
await dl3.goto(hash)
const [download2, download3] = await Promise.all([
dl2.clickDownload(),
dl3.clickDownload()
])
expect(download2.suggestedFilename()).toBe('concurrent.txt')
expect(download3.suggestedFilename()).toBe('concurrent.txt')
await context.close()
})
```
### 6.7 Redirect File Handling (Multi-chunk)
**Test ID**: `edge-redirect-file`
```typescript
test.slow()
test('upload and download multi-chunk file with redirect', async ({uploadPage, downloadPage}) => {
// Use ~5MB file to get multiple chunks
await uploadPage.selectLargeFile('multi-chunk.bin', 5 * 1024 * 1024)
const link = await uploadPage.waitForShareLink(120_000)
await downloadPage.gotoWithLink(link)
const download = await downloadPage.clickDownload()
expect(download.suggestedFilename()).toBe('multi-chunk.bin')
const path = await download.path()
if (path) {
const stat = (await import('fs')).statSync(path)
expect(stat.size).toBe(5 * 1024 * 1024)
}
})
```
### 6.8 UI Information Display
**Test ID**: `edge-ui-info`
```typescript
test('upload complete shows expiry and security note', async ({uploadPage}) => {
await uploadPage.selectTextFile('ui-test.txt', 'ui test')
await uploadPage.waitForShareLink()
await uploadPage.expectCompleteWithExpiry()
await uploadPage.expectSecurityNote()
})
test('download page shows file size and security note', async ({uploadPage, downloadPage}) => {
await uploadPage.selectFile('size-test.bin', createTestContent(1024))
const link = await uploadPage.waitForShareLink()
await downloadPage.gotoWithLink(link)
await downloadPage.expectFileSizeDisplayed()
await downloadPage.expectSecurityNote()
})
```
---
## 7. Implementation Order
### Phase 1: Core Infrastructure (Priority: High)
1. Create `test/pages/UploadPage.ts` with Page Object
2. Create `test/pages/DownloadPage.ts` with Page Object
3. Create `test/fixtures.ts` with extended test function
4. Refactor existing test to use Page Objects
### Phase 2: Core Happy Path (Priority: High)
5. `upload-file-picker` - Basic upload via file picker
6. `download-button-click` - Basic download
7. `download-file-save` - Content verification
### Phase 3: Validation (Priority: High)
8. `upload-file-too-large` - Size validation
9. `upload-file-empty` - Empty file validation
10. `download-invalid-hash-malformed` - Invalid link handling
11. `download-invalid-hash-structure` - Invalid structure handling
### Phase 4: Progress and Cancel (Priority: Medium)
12. `upload-progress-display` - Progress visibility
13. `upload-cancel` - Cancel functionality
14. `download-progress-display` - Download progress
### Phase 5: Link Sharing (Priority: Medium)
15. `upload-share-link-copy` - Copy button functionality
16. `upload-drag-drop` - Drag-drop upload
### Phase 6: Edge Cases (Priority: Low)
17. `edge-small-file` - 1-byte file
18. `edge-special-chars-filename` - Unicode/special characters
19. `edge-binary-content` - Binary content integrity
20. `edge-near-limit` - 100MB file (slow test)
21. `edge-network-error` - Network error handling
### Phase 7: Error Recovery and Advanced (Priority: Low)
22. `upload-error-retry` - Retry after error
23. `edge-concurrent-downloads` - Concurrent access
24. `edge-redirect-file` - Multi-chunk file with redirect (slow)
25. `edge-ui-info` - Expiry message, security notes
---
## 8. Test Utilities
### 8.1 Shared Test Setup
```typescript
// test/page.spec.ts
import {test, expect, createTestContent, createTextContent, uniqueFileName} from './fixtures'
test.describe('Upload Flow', () => {
test('upload via file picker', async ({uploadPage}) => {
// Tests use uploadPage fixture which navigates automatically
})
})
test.describe('Download Flow', () => {
test('download works', async ({uploadPage, downloadPage}) => {
// Both pages available via fixtures
})
})
test.describe('Edge Cases', () => {
// Edge case tests
})
```
### 8.2 File Structure
```
xftp-web/test/
├── fixtures.ts # Playwright fixtures with page objects
├── pages/
│ ├── UploadPage.ts # Upload page object
│ └── DownloadPage.ts # Download page object
├── page.spec.ts # All E2E tests
└── globalSetup.ts # Server startup (existing)
```
---
## Appendix: Test Matrix
| Test ID | Category | Priority | Estimated Time | Dependencies |
|---------|----------|----------|----------------|--------------|
| upload-file-picker | Upload | High | 30s | - |
| upload-drag-drop | Upload | Medium | 30s | - |
| upload-file-too-large | Upload | High | 5s | - |
| upload-file-empty | Upload | High | 5s | - |
| upload-progress-display | Upload | Medium | 45s | - |
| upload-cancel | Upload | Medium | 30s | - |
| upload-share-link-copy | Upload | Medium | 30s | - |
| upload-error-retry | Upload | Low | 30s | - |
| download-invalid-hash-malformed | Download | High | 5s | - |
| download-invalid-hash-structure | Download | High | 5s | - |
| download-button-click | Download | High | 45s | upload |
| download-progress-display | Download | Medium | 60s | upload |
| download-file-save | Download | High | 45s | upload |
| edge-small-file | Edge | Low | 30s | - |
| edge-near-limit | Edge | Low | 300s | - |
| edge-special-chars-filename | Edge | Low | 30s | - |
| edge-network-error | Edge | Low | 45s | - |
| edge-binary-content | Edge | Low | 30s | - |
| edge-concurrent-downloads | Edge | Low | 60s | upload |
| edge-redirect-file | Edge | Low | 120s | - |
| edge-ui-info | Edge | Low | 60s | upload |
**Total estimated time**: ~18 minutes (excluding 100MB and 5MB tests)
@@ -0,0 +1,221 @@
# XFTP Web Hello Header — Session Re-handshake for Browser Connection Reuse
## 1. Problem Statement
Browser HTTP/2 connection pooling reuses TLS connections across page navigations (same origin = same connection pool). The XFTP server maintains per-TLS-connection session state in `TMap SessionId Handshake` keyed by `tlsUniq tls`. When a browser navigates from the upload page to the download page (or reloads), the new page sends a fresh ClientHello on the reused HTTP/2 connection. The server is already in `HandshakeAccepted` state for that connection, so it routes the request to `processRequest`, which expects a 16384-byte command block but receives a 34-byte ClientHello → `ERR BLOCK`.
**Root cause**: The server cannot distinguish a ClientHello from a command on an already-handshaked connection because both arrive on the same HTTP/2 connection (same `tlsUniq`), and there is no content-level discriminator (ClientHello is unpadded, but the server never gets to parse it — the size check in `processRequest` rejects it first).
**Browser limitation**: `fetch()` provides zero control over HTTP/2 connection pooling. There is no browser API to force a new connection or detect connection reuse before a request is sent.
## 2. Solution Summary
Add an HTTP header `xftp-web-hello` to web ClientHello requests. When the server sees this header on an already-handshaked connection (`HandshakeAccepted` state), it re-runs `processHello` **reusing the existing session keys** (same X25519 key pair from the original handshake). The client then completes the normal handshake flow (sends ClientHandshake, receives ack) and proceeds with commands.
Key properties:
- Server reuses existing `serverPrivKey` — no new key material generated on re-handshake, so `thAuth` remains consistent with any in-flight commands on concurrent HTTP/2 streams.
- Header is only checked when `sniUsed` is true (web/browser connections). Native XFTP clients are unaffected.
- CORS preflight already allows all headers (`Access-Control-Allow-Headers: *`).
- Web clients always send this header on ClientHello — it's harmless on first connection (`Nothing` state) and enables re-handshake on reused connections (`HandshakeAccepted` state).
## 3. Detailed Technical Design
### 3.1 Server change: parameterize `processHello` (`src/Simplex/FileTransfer/Server.hs`)
The entire server change is parameterizing the existing `processHello` with `Maybe C.PrivateKeyX25519`. Zero new functions.
#### Current code (lines 165-191):
```haskell
xftpServerHandshakeV1 chain serverSignKey sessions
XFTPTransportRequest {thParams = thParams0@THandleParams {sessionId}, reqBody = HTTP2Body {bodyHead}, sendResponse, sniUsed, addCORS} = do
s <- atomically $ TM.lookup sessionId sessions
r <- runExceptT $ case s of
Nothing -> processHello
Just (HandshakeSent pk) -> processClientHandshake pk
Just (HandshakeAccepted thParams) -> pure $ Just thParams
either sendError pure r
where
processHello = do
challenge_ <-
if
| B.null bodyHead -> pure Nothing
| sniUsed -> do
XFTPClientHello {webChallenge} <- liftHS $ smpDecode bodyHead
pure webChallenge
| otherwise -> throwE HANDSHAKE
(k, pk) <- atomically . C.generateKeyPair =<< asks random
atomically $ TM.insert sessionId (HandshakeSent pk) sessions
-- ...build and send ServerHandshake...
pure Nothing
```
#### After (diff is ~10 lines):
```haskell
xftpServerHandshakeV1 chain serverSignKey sessions
XFTPTransportRequest {thParams = thParams0@THandleParams {sessionId}, request, reqBody = HTTP2Body {bodyHead}, sendResponse, sniUsed, addCORS} = do
-- ^^^^^^^ bind request
s <- atomically $ TM.lookup sessionId sessions
r <- runExceptT $ case s of
Nothing -> processHello Nothing
Just (HandshakeSent pk) -> processClientHandshake pk
Just (HandshakeAccepted thParams)
| webHello -> processHello (serverPrivKey <$> thAuth thParams)
| otherwise -> pure $ Just thParams
either sendError pure r
where
webHello = sniUsed && any (\(t, _) -> tokenKey t == "xftp-web-hello") (fst $ H.requestHeaders request)
processHello pk_ = do
challenge_ <-
if
| B.null bodyHead -> pure Nothing
| sniUsed -> do
XFTPClientHello {webChallenge} <- liftHS $ smpDecode bodyHead
pure webChallenge
| otherwise -> throwE HANDSHAKE
(k, pk) <- maybe
(atomically . C.generateKeyPair =<< asks random)
(\pk -> pure (C.publicKey pk, pk))
pk_
atomically $ TM.insert sessionId (HandshakeSent pk) sessions
-- ...rest unchanged...
pure Nothing
```
#### What changes:
1. **Bind `request`** in the `XFTPTransportRequest` pattern (+1 field)
2. **Add `webHello`** binding in `where` clause (1 line) — checks header only when `sniUsed`
3. **Add `pk_` parameter** to `processHello` (change signature)
4. **Replace key generation** with `maybe` that generates fresh keys when `pk_ = Nothing`, or derives public from existing private when `pk_ = Just pk` (3 lines replace 1 line)
5. **Add guard** in `HandshakeAccepted` branch (2 lines replace 1 line)
6. **Call site** `Nothing -> processHello Nothing` (+1 word)
7. **One import** added: `Network.HPACK.Token (tokenKey)`
#### Imports to add:
```haskell
import Network.HPACK.Token (tokenKey)
```
`OverloadedStrings` (already enabled in Server.hs) provides the `IsString` instance for `CI ByteString`, so `tokenKey t == "xftp-web-hello"` works without importing `Data.CaseInsensitive`. Verified on Hackage: `requestHeaders :: Request -> HeaderTable`, `tokenKey :: Token -> CI ByteString`.
### 3.2 Re-handshake flow
When `webHello` is true in `HandshakeAccepted` state:
1. `processHello (serverPrivKey <$> thAuth thParams)` is called with `Just pk` (existing private key)
2. `(k, pk) <- pure (C.publicKey pk, pk)` — reuses same key pair, no generation
3. `TM.insert sessionId (HandshakeSent pk) sessions` — transitions state back to `HandshakeSent` with same `pk`
4. Server sends `ServerHandshake` response (same format as initial handshake)
5. Client sends `ClientHandshake` on next stream → enters `Just (HandshakeSent pk) -> processClientHandshake pk` → normal flow
6. `processClientHandshake` stores `HandshakeAccepted thParams` with same `serverPrivKey = pk`
### 3.3 Web client change (`xftp-web/src/client.ts`)
Add optional `headers?` parameter to `Transport.post()`, thread it through `fetch()` and `session.request()`, and pass `{"xftp-web-hello": "1"}` in the ClientHello call in `connectXFTP`.
### 3.4 What does NOT change
- **CORS**: Already has `Access-Control-Allow-Headers: *` (Server.hs:106).
- **Native Haskell client**: Uses `[]` headers. No header = existing behavior.
- **Protocol wire format**: ClientHello, ServerHandshake, ClientHandshake, commands — all unchanged.
- **`processRequest`**, **`processClientHandshake`**, **`sendError`**, **`encodeXftp`** — unchanged.
### 3.5 Haskell test (`tests/XFTPServerTests.hs`)
Add `testWebReHandshake` next to the existing `testWebHandshake` (line 504). It reuses the same SNI + HTTP/2 setup pattern, performs a full handshake, then sends a second ClientHello with the `xftp-web-hello` header on the same connection and verifies the server responds with a valid ServerHandshake (same `sessionId`), then completes the second handshake.
```haskell
-- Register in xftpServerTests (after line 86):
it "should re-handshake on same connection with xftp-web-hello header" testWebReHandshake
-- Test (after testWebHandshake):
testWebReHandshake :: Expectation
testWebReHandshake =
withXFTPServerSNI $ \_ -> do
Fingerprint fp <- loadFileFingerprint "tests/fixtures/ca.crt"
let keyHash = C.KeyHash fp
cfg = defaultTransportClientConfig {clientALPN = Just ["h2"], useSNI = True}
runTLSTransportClient defaultSupportedParamsHTTPS Nothing cfg Nothing "localhost" xftpTestPort (Just keyHash) $ \(tls :: TLS 'TClient) -> do
let h2cfg = HC.defaultHTTP2ClientConfig {HC.bodyHeadSize = 65536}
h2 <- either (error . show) pure =<< HC.attachHTTP2Client h2cfg (THDomainName "localhost") xftpTestPort mempty 65536 tls
g <- C.newRandom
-- First handshake (same as testWebHandshake)
challenge1 <- atomically $ C.randomBytes 32 g
let helloReq1 = H2.requestBuilder "POST" "/" [] $ byteString (smpEncode (XFTPClientHello {webChallenge = Just challenge1}))
resp1 <- either (error . show) pure =<< HC.sendRequest h2 helloReq1 (Just 5000000)
shs1 <- either error pure $ smpDecode =<< C.unPad (bodyHead (HC.respBody resp1))
let XFTPServerHandshake {sessionId = sid1} = shs1
clientHsPadded <- either (error . show) pure $ C.pad (smpEncode (XFTPClientHandshake {xftpVersion = VersionXFTP 1, keyHash})) xftpBlockSize
resp1b <- either (error . show) pure =<< HC.sendRequest h2 (H2.requestBuilder "POST" "/" [] $ byteString clientHsPadded) (Just 5000000)
B.length (bodyHead (HC.respBody resp1b)) `shouldBe` 0
-- Second handshake on same connection with xftp-web-hello header
challenge2 <- atomically $ C.randomBytes 32 g
let helloReq2 = H2.requestBuilder "POST" "/" [("xftp-web-hello", "1")] $ byteString (smpEncode (XFTPClientHello {webChallenge = Just challenge2}))
resp2 <- either (error . show) pure =<< HC.sendRequest h2 helloReq2 (Just 5000000)
shs2 <- either error pure $ smpDecode =<< C.unPad (bodyHead (HC.respBody resp2))
let XFTPServerHandshake {sessionId = sid2} = shs2
sid2 `shouldBe` sid1 -- same TLS connection → same sessionId
-- Complete second handshake
resp2b <- either (error . show) pure =<< HC.sendRequest h2 (H2.requestBuilder "POST" "/" [] $ byteString clientHsPadded) (Just 5000000)
B.length (bodyHead (HC.respBody resp2b)) `shouldBe` 0
```
The only difference from `testWebHandshake`: the second `helloReq2` passes `[("xftp-web-hello", "1")]` instead of `[]`. The test verifies:
1. Server responds with `ServerHandshake` (not `ERR BLOCK`)
2. Same `sessionId` (same TLS connection)
3. Second `ClientHandshake` completes with empty ACK
## 4. Implementation Plan
### Step 1: Server — parameterize `processHello`
Apply the diff from Section 3.1 to `src/Simplex/FileTransfer/Server.hs`.
### Step 2: Test — add `testWebReHandshake`
Add the test from Section 3.5 to `tests/XFTPServerTests.hs`.
### Step 3: Client — add `xftp-web-hello` header
Add optional `headers?` to `Transport.post()`, pass `{"xftp-web-hello": "1"}` on ClientHello in `connectXFTP`.
### Step 4: Test
Run Haskell tests (`cabal test`) and E2E Playwright tests (`npx playwright test` in `xftp-web/`).
## 5. Race Condition Analysis
### Single-tab navigation (the common case)
1. Upload page completes, all fetch() requests finish
2. Browser navigates to download page (or reloads)
3. All upload-page fetches are aborted on page unload
4. Download page sends ClientHello with `xftp-web-hello` header
5. Server is in `HandshakeAccepted``processHello (Just pk)``HandshakeSent pk` (same key)
6. No concurrent streams → no race
**Safe.**
### Multi-tab (edge case)
Tab A (upload) and Tab B (download) share the same HTTP/2 connection.
1. Tab A has active command streams (e.g., FPUT upload in progress)
2. Tab B sends ClientHello with header
3. Server reads `HandshakeAccepted` atomically for both streams
4. Tab A's stream already has its `thParams` snapshot → proceeds with `processRequest` using old `thParams`
5. Tab B's stream triggers `processHello (Just pk)` → stores `HandshakeSent pk` (same pk!)
6. Tab A's in-progress FPUT continues with snapshot `thParams` → completes normally (same `serverPrivKey`)
7. Tab A's NEXT command reads `HandshakeSent` from TMap → enters `processClientHandshake` → fails (command body ≠ ClientHandshake format) → HANDSHAKE error
**Tab A's in-flight commands succeed. Tab A's subsequent commands fail with HANDSHAKE error.** This is the inherent multi-tab problem — unavoidable with per-connection session state and HTTP/2 connection sharing. The failure is clean (HANDSHAKE error, not silent corruption).
## 6. Security Considerations
- **No new key material**: Re-handshake reuses existing `serverPrivKey`. No opportunity for key confusion or downgrade.
- **Identity re-verification**: Server re-signs the web challenge with its long-term signing key. Client verifies identity again.
- **Header cannot escalate privileges**: The header only triggers re-handshake (which the server was already capable of doing on first connection). It does not bypass any authentication.
- **Timing**: Re-handshake takes the same code path as initial handshake, so timing side-channels are unchanged.
@@ -0,0 +1,948 @@
# XFTP Web Error Handling and Connection Resilience
## 1. Problem Statement
The XFTP web client is fundamentally fragile: any transient error (browser opening a new HTTP/2 connection, network hiccup, server restart) causes an unrecoverable failure with a cryptic error message. There is no retry logic, no fetch timeout, no error categorization, and the upload uses a single server instead of distributing chunks across preset servers. This makes the app frustrating — it works most of the time but fails unpredictably, which is worse than being completely broken.
### Confirmed root cause (from diagnostic logs)
When the browser opens a new HTTP/2 connection mid-operation, the new connection has a different TLS SessionId with no handshake state in the server's `TMap SessionId Handshake`. The server's `Nothing` branch in `xftpServerHandshakeV1` (Server.hs:169) unconditionally calls `processHello`, which tries to decode the command body as `XFTPClientHello`, fails, and sends a raw padded "HANDSHAKE" error string. The client cannot parse this as a proper transmission (first byte 'H' = 72 is read as batch count), producing `"expected batch count 1, got 72"`.
Server log confirming the SessionId change:
```
DEBUG dispatch: Accepted+command sessId="ZSo1GGETgIvjbB7CWHbvGPpbMjx_b2IlC1eTI6aKfqc="
...20 successful commands...
DEBUG dispatch: Nothing sessId="mJC7Sck9xxW5UsXoPGoUWduuHghSVgf6CnD6ZC6SBhU=" webHello=False
```
### Why re-handshake is required (cannot be made optional)
1. **SessionId is baked into signed command data.** `encodeAuthTransmission` signs `concat(encode(sessionId), tInner)` with Ed25519. Server's `tDecodeServer` (Protocol.hs:2242) verifies `sessId == sessionId`. New connection = different sessionId = signature mismatch.
2. **Server generates per-session DH keys.** `processHello` creates fresh X25519 keypair stored in `HandshakeSent`. For SMP browser clients (future), `verifyCmdAuth` (Protocol.hs:1322) requires the matching `serverPrivKey` from `thAuth`.
3. **This applies to both XFTP and future SMP browser clients** — the session management approach is the same.
### Why multiple preset servers cannot work
Upload (`agent.ts:105-157`) takes a single `server: XFTPServer` parameter and uploads ALL chunks to it. `web/upload.ts:133` calls `pickRandomServer(servers)` which selects ONE random server from all presets. The multi-server preset configuration is pointless — only one server is ever used per upload. The design intent (RFC section 11.6: "upload in parallel to 8 randomly selected servers") is not implemented. This must be fixed in Phase 2 (section 3.7).
## 2. Solution Summary
### Phase 1: Error handling and connection resilience
1. **Server: strict dispatch for allowed protocol combinations** — reject all invalid combinations
2. **Client: automatic retry with re-handshake** on SESSION/HANDSHAKE errors
3. **Client: fetch timeout** with configurable duration
4. **UI: error categorization and retry** — auto-retry temporary, human-readable permanent
5. **Client: connection state with Promise-based lock and per-server queues**`ServerConnection` with `client: Promise<XFTPClient>` + `queue: Promise<void>`
6. **Client: fix cache key** — include keyHash
### Phase 2: Multi-server upload (after Phase 1)
7. **Multi-server upload with server selection and failover** — distribute chunks across servers, retry FNEW on different server if one fails
## 3. Detailed Technical Design
### 3.1 Server: strict dispatch for allowed protocol combinations
**Principle:** Everything not explicitly done by existing Haskell/TS clients is prohibited. It is better to fail on impossible combinations than to be permissive — permissiveness complicates debugging and creates attack vectors via unexpected behaviors.
**Allowed behaviors by client type:**
| Client | SNI | webHello header | Hello body | When |
|--------|-----|----------------|------------|------|
| Haskell | No | No | Empty | New connection only |
| Web | Yes | Yes | Non-empty (XFTPClientHello) | New OR existing connection |
**Minimal surgical change.** The existing dispatch (Server.hs:169-189) already correctly handles `HandshakeSent` and `HandshakeAccepted` — their guards cover all valid and invalid combinations. The ONLY missing case is `Nothing` + web client sending a command on a stale session.
`processHello` (Server.hs:194-217) already internally routes: `B.null bodyHead` → Haskell hello, `sniUsed` → web hello decode, else → HANDSHAKE. For stale web sessions, it currently tries to decode a command body as `XFTPClientHello`, fails, and throws HANDSHAKE. The fix: detect this case BEFORE calling processHello and throw SESSION instead, so the client knows to re-handshake (not that its hello was malformed).
**Change: add one guard to `Nothing` branch, remove debug logging.**
```haskell
-- Before (1 line):
Nothing -> processHello Nothing
-- After (3 lines):
Nothing
| sniUsed && not webHello -> throwE SESSION -- web command on stale session
| otherwise -> processHello Nothing -- normal hello (web or Haskell)
```
`throwE SESSION` is caught by `either sendError pure r` (line 190). `sendError` pads `smpEncode SESSION` = `"SESSION"` (Transport.hs:298) to `xftpBlockSize`. The client's padded error detection (section 3.2) catches this as a retriable error and triggers re-handshake. SESSION is a valid `XFTPErrorType` constructor (Transport.hs:225) — no new helpers needed.
**All other branches remain unchanged.** `HandshakeSent` guards (`webHello` → processHello, `otherwise` → processClientHandshake with body size check inside) are correct. `HandshakeAccepted` guards (`webHello`, `webHandshake`, `otherwise` → command) are correct.
### 3.2 Client: automatic retry with re-handshake
**Location:** `sendXFTPCommand` in `client.ts`
**Design:** Retry loop inside `sendXFTPCommand`. Maximum 3 attempts. On retriable error, close old client, re-handshake, retry.
**Error classification:**
| Error | Type | Retriable? | Human-readable message |
|-------|------|-----------|----------------------|
| Padded "HANDSHAKE" | Temporary | Yes (auto) | "Connection interrupted, reconnecting..." |
| Padded "SESSION" | Temporary | Yes (auto) | "Session expired, reconnecting..." |
| `FRErr SESSION` | Temporary | Yes (auto) | "Session expired, reconnecting..." |
| `FRErr HANDSHAKE` | Temporary | Yes (auto) | "Connection interrupted, reconnecting..." |
| `fetch()` TypeError | Temporary | Yes (auto) | "Network error, retrying..." |
| AbortError (timeout) | Temporary | Yes (auto) | "Server timeout, retrying..." |
| `FRErr AUTH` | Permanent | No | "File is invalid, expired, or has been removed" |
| `FRErr NO_FILE` | Permanent | No | "File not found — it may have expired" |
| `FRErr SIZE` | Permanent | No | "File size exceeds server limit" |
| `FRErr QUOTA` | Permanent | No | "Server storage quota exceeded" |
| `FRErr BLOCKED` | Permanent | No | "File has been blocked by server" |
| `FRErr DIGEST` | Permanent | No | "File integrity check failed" |
| `FRErr INTERNAL` | Permanent | No | "Server internal error" |
| `CMD *` | Permanent | No | "Protocol error" |
**Retry behavior:**
- Auto-retry up to 3 times for temporary errors, transparent to user
- After 3 failures: show human-readable error with diagnosis, offer manual retry button
- Permanent errors: show human-readable error immediately, NO manual retry button (user can reload page)
**Implementation:**
```typescript
async function sendXFTPCommand(
agent: XFTPClientAgent,
server: XFTPServer,
privateKey: Uint8Array,
entityId: Uint8Array,
cmdBytes: Uint8Array,
chunkData?: Uint8Array,
maxRetries: number = 3
): Promise<{response: FileResponse, body: Uint8Array}> {
let clientP = getXFTPServerClient(agent, server)
let client = await clientP
for (let attempt = 1; attempt <= maxRetries; attempt++) {
try {
return await sendXFTPCommandOnce(client, privateKey, entityId, cmdBytes, chunkData)
} catch (e) {
if (!isRetriable(e)) {
// Permanent error (AUTH, NO_FILE, etc.) — connection is fine, don't touch it
throw categorizeError(e)
}
if (attempt === maxRetries) {
// Retriable error exhausted — connection is bad, remove stale promise
removeStaleConnection(agent, server, clientP)
throw categorizeError(e)
}
clientP = reconnectClient(agent, server)
client = await clientP
}
}
throw new Error("unreachable")
}
```
**`sendXFTPCommandOnce`** — renamed from current `sendXFTPCommand`. Two changes:
1. **Padded error detection** (before `decodeTransmission`):
```typescript
// After getting respBlock, before decodeTransmission:
const raw = blockUnpad(respBlock)
if (raw.length < 20) {
const text = new TextDecoder().decode(raw)
if (/^[A-Z_]+$/.test(text)) {
throw new XFTPRetriableError(text) // "HANDSHAKE" or "SESSION"
}
}
```
2. **FRErr classification** (replaces current unconditional throw):
```typescript
// After decodeResponse, instead of throw new Error("Server error: " + err.type):
if (response.type === "FRErr") {
const err = response.err
if (err.type === "SESSION" || err.type === "HANDSHAKE") {
throw new XFTPRetriableError(err.type)
}
throw new XFTPPermanentError(err.type, humanReadableMessage(err))
}
```
### 3.3 Client: fetch timeout
**Location:** `createBrowserTransport` and `createNodeTransport` in `client.ts`
**Design:** `AbortController` with configurable timeout on every `fetch()`.
```typescript
interface TransportConfig {
timeoutMs: number // default 30000, lower for tests
}
function createBrowserTransport(baseUrl: string, config: TransportConfig): Transport {
return {
async post(body: Uint8Array, headers?: Record<string, string>): Promise<Uint8Array> {
const controller = new AbortController()
const timer = setTimeout(() => controller.abort(), config.timeoutMs)
try {
const resp = await fetch(effectiveUrl, {
method: "POST", headers, body,
signal: controller.signal
})
if (!resp.ok) throw new Error(`Server request failed: ${resp.status}`)
return new Uint8Array(await resp.arrayBuffer())
} finally {
clearTimeout(timer)
}
},
close() {}
}
}
```
For Node.js transport, use `setTimeout` on the HTTP/2 request stream.
Default: 30s for production, 5s for tests. Threaded through `connectXFTP``createTransport`.
### 3.4 UI: error categorization and retry
**Behavior (Option D):**
- **Temporary errors:** Auto-retry loop (3 attempts). After 3 failures, show human-readable diagnosis with manual retry button. Diagnosis examples: "Server timeout — the server may be temporarily unavailable", "Connection interrupted — your network may be unstable".
- **Permanent errors:** Show human-readable error immediately, NO retry button. User can reload page if they want to retry. Examples: "File is invalid, expired, or has been removed" (AUTH), "File not found" (NO_FILE).
**Current UI retry buttons:**
- `upload.ts:73-75` — retry calls `startUpload(pendingFile)` from scratch
- `download.ts:60` — retry calls `startDownload()` from scratch
**Improvement:** Track uploaded/downloaded chunk indices. On manual retry, skip completed chunks:
```typescript
// Upload: track which chunks completed
const completedChunks: Set<number> = new Set()
for (let i = 0; i < specs.length; i++) {
if (completedChunks.has(i)) continue
// ... create + upload chunk
completedChunks.add(i)
}
// Download: already naturally resumable — each chunk is independent
```
### 3.5 Client: connection state with Promise-based lock and per-server queues
**Design:** Each server gets a `ServerConnection` record containing a `Promise<XFTPClient>` (the connection lock) and a `Promise<void>` (the sequential command queue). The `XFTPClientAgent` maps server keys to these records.
The promise IS the lock — every consumer awaits the same promise. When reconnect is needed, the promise is replaced atomically.
```typescript
interface ServerConnection {
client: Promise<XFTPClient> // resolves to connected client; replaced on reconnect
queue: Promise<void> // tail of sequential command chain
}
interface XFTPClientAgent {
connections: Map<string, ServerConnection>
}
function newXFTPAgent(): XFTPClientAgent {
return {connections: new Map()}
}
```
**Connection lifecycle — `getXFTPServerClient` and `reconnectClient`:**
```typescript
function getXFTPServerClient(agent: XFTPClientAgent, server: XFTPServer): Promise<XFTPClient> {
const key = formatXFTPServer(server)
let conn = agent.connections.get(key)
if (!conn) {
const p = connectXFTP(server)
conn = {client: p, queue: Promise.resolve()}
agent.connections.set(key, conn)
// On connection failure, remove from map so next call retries
p.catch(() => {
const cur = agent.connections.get(key)
if (cur && cur.client === p) agent.connections.delete(key)
})
}
return conn.client
}
function reconnectClient(agent: XFTPClientAgent, server: XFTPServer): Promise<XFTPClient> {
const key = formatXFTPServer(server)
const old = agent.connections.get(key)
// Close old client (fire-and-forget)
old?.client.then(c => c.transport.close(), () => {})
// Replace with new connection promise — all concurrent callers will await this
// Queue survives reconnect — pending operations stay ordered
const p = connectXFTP(server)
const conn: ServerConnection = {client: p, queue: old?.queue ?? Promise.resolve()}
agent.connections.set(key, conn)
p.catch(() => {
const cur = agent.connections.get(key)
if (cur && cur.client === p) agent.connections.delete(key)
})
return p
}
function closeXFTPServerClient(agent: XFTPClientAgent, server: XFTPServer): void {
const key = formatXFTPServer(server)
const conn = agent.connections.get(key)
if (conn) {
agent.connections.delete(key)
conn.client.then(c => c.transport.close(), () => {})
}
}
function closeXFTPAgent(agent: XFTPClientAgent): void {
for (const conn of agent.connections.values()) {
conn.client.then(c => c.transport.close(), () => {})
}
agent.connections.clear()
}
```
**Precise semantics:**
1. `getXFTPServerClient(agent, server)` — returns existing `conn.client` promise if present, otherwise creates a new `ServerConnection` with fresh connection and empty queue
2. When error detected, first caller calls `reconnectClient` which replaces `conn.client` with a new connection promise. The queue is preserved across reconnect.
3. All concurrent callers awaiting the OLD promise receive the error
4. They then call `getXFTPServerClient` which returns the NEW promise
5. If reconnection fails, auto-cleanup (`p.catch(() => delete)`) removes the entry so the next caller starts fresh
**Stale error cleanup rule:** When a caller exhausts retries for a retriable error, it removes the failed entry from the map (only if no concurrent caller has already replaced it via `reconnectClient`). This prevents the next caller from receiving a stale rejected promise. Permanent errors (AUTH, NO_FILE, etc.) do NOT remove the connection — the transport is fine, only the command failed.
```typescript
function removeStaleConnection(
agent: XFTPClientAgent, server: XFTPServer, failedP: Promise<XFTPClient>
): void {
const key = formatXFTPServer(server)
const conn = agent.connections.get(key)
// Only remove if current promise is the one that failed — not if already replaced by reconnect
if (conn && conn.client === failedP) {
agent.connections.delete(key)
failedP.then(c => c.transport.close(), () => {})
}
}
```
**Per-server sequential queue:** `queue` is a `Promise<void>` — the tail of the sequential operation chain. Each new operation `.then()`s onto it. It's `void` because callers hold their own typed promises; the queue only tracks completion order:
```typescript
async function enqueueCommand<T>(
agent: XFTPClientAgent,
server: XFTPServer,
fn: () => Promise<T> // no client param — fn uses command wrappers (agent+server)
): Promise<T> {
const key = formatXFTPServer(server)
// Ensure connection exists (with auto-cleanup on failure)
await getXFTPServerClient(agent, server)
const conn = agent.connections.get(key)! // guaranteed to exist after getXFTPServerClient
// Chain onto the queue — fn runs after previous operation completes
let resolve_: (v: T) => void, reject_: (e: any) => void
const result = new Promise<T>((res, rej) => { resolve_ = res; reject_ = rej })
conn.queue = conn.queue.then(
() => fn().then(resolve_!, reject_!),
() => fn().then(resolve_!, reject_!)
).then(() => {}, () => {}) // swallow errors in the chain
return result
}
```
Commands to the same server execute one at a time via the queue. Commands to different servers execute concurrently because each has its own queue. `enqueueCommand` provides sequencing; `sendXFTPCommand` (called inside `fn` via command wrappers) provides retry. They compose as: `enqueueCommand` sequences calls to wrappers that internally use `sendXFTPCommand`.
**Download change:** Group chunks by server, process each server's chunks sequentially, servers in parallel. Uses `for` loop for per-server sequencing (same pattern as Stage 2 upload). `enqueueCommand` is available for cases where different callers target the same server.
```typescript
const byServer = new Map<string, FileChunk[]>()
for (const chunk of resolvedFd.chunks) {
const srv = chunk.replicas[0]?.server ?? ""
if (!byServer.has(srv)) byServer.set(srv, [])
byServer.get(srv)!.push(chunk)
}
await Promise.all([...byServer.entries()].map(async ([srv, chunks]) => {
const server = parseXFTPServer(srv)
for (const chunk of chunks) {
const seed = decodePrivKeyEd25519(chunk.replicas[0].replicaKey)
const kp = ed25519KeyPairFromSeed(seed)
const raw = await downloadXFTPChunkRaw(agent, server, kp.privateKey, chunk.replicas[0].replicaId)
await onRawChunk({chunkNo: chunk.chunkNo, dhSecret: raw.dhSecret, nonce: raw.nonce, body: raw.body, digest: chunk.digest})
downloaded += chunk.chunkSize
onProgress?.(downloaded, resolvedFd.size)
}
}))
```
### 3.6 Fix cache key
**Bug:** `getXFTPServerClient` (client.ts:110) uses `"https://" + server.host + ":" + server.port` as cache key, ignoring `keyHash`. Two servers with same host:port but different keyHash share a cached connection, bypassing identity verification.
**Fix:** Use `formatXFTPServer(server)` as cache key (includes keyHash). Already available in `protocol/address.ts:52-54`.
```typescript
// Before:
const key = "https://" + server.host + ":" + server.port
// After:
const key = formatXFTPServer(server)
```
Note: With the redesign in 3.5, the cache key fix is inherent — the `connections` Map uses `formatXFTPServer(server)` everywhere.
### 3.7 Phase 2: Multi-server upload with server selection and failover
**Problem:** Current upload (`agent.ts:105-157`) takes a single `server: XFTPServer` and uploads ALL chunks to it. The 12 preset servers (6 SimpleX + 6 Flux) are pointless — only one is ever used.
**Design goal:** Distribute chunks across servers. Retry FNEW on a different server if one fails. Once working servers are found, prefer them (heuristic: server unlikely to fail mid-process, more likely to be broken initially due to maintenance/downtime).
**Reference implementation:** Haskell `Agent.hs:457-486` (`createChunk` / `createWithNextSrv`) + `Client.hs:2335-2385` (`getNextServer_` / `withNextSrv`).
#### Haskell algorithm summary
Two-stage architecture:
1. **Allocate stage (serial per file in Haskell):** For each chunk, call FNEW on a randomly-selected server. If FNEW fails, pick a different server and retry. Track tried hosts to avoid retrying the same server. After all chunks are assigned to servers, spawn one upload worker per server.
2. **Upload stage (parallel per server):** Each server worker uploads its assigned chunks sequentially (FPUT). On FPUT failure, retry on the same server with backoff (because the chunk replica already exists on that server). No server failover for FPUT.
Server selection constraints (hierarchical, `getNextServer_` Client.hs:2335-2350):
1. Prefer servers from unused operators (operator diversity)
2. Prefer servers with unused hosts (host diversity)
3. Random pick from the most-constrained candidate set
4. If all exhausted, reset tried set and start over
#### Web client adaptation
The web client doesn't have operators or a database. Simplified algorithm with two stages:
**Stage 1 — Allocate:** Create chunk records on servers (FNEW). Unlike Haskell which is serial here, web FNEW runs concurrently within a concurrency limit. FNEW is a small command — concurrent FNEW on the same connection is not a problem, and concurrent FNEW across servers improves upload startup time.
**Stage 2 — Upload:** Upload chunk data (FPUT). Parallel across servers, sequential per server (reuses per-server queues from 3.5). FPUT retries on the same server with backoff — no server rotation because the chunk replica already exists on that server. Stage 2 reads chunk data by offset (via `readChunk`), so `SentChunk` must be extended with `chunkOffset: number` (from ChunkSpec).
```typescript
interface UploadState {
untriedServers: XFTPServer[] // servers not yet attempted — initially all servers
workingServers: XFTPServer[] // servers that succeeded FNEW
}
const MAX_FNEW_ATTEMPTS = 5 // per chunk: try up to 5 different servers
async function uploadFile(
agent: XFTPClientAgent,
allServers: XFTPServer[],
encrypted: EncryptedFileMetadata,
options?: UploadOptions
): Promise<UploadResult> {
const state: UploadState = {untriedServers: [...allServers], workingServers: []}
const specs = prepareChunkSpecs(encrypted.chunkSizes)
const concurrency = options?.concurrency ?? 4
// Stage 1: Allocate — concurrent FNEW within concurrency limit
const sentChunks: SentChunk[] = new Array(specs.length)
const queue = specs.map((spec, i) => ({spec, chunkNo: i + 1, index: i}))
let idx = 0
async function allocateWorker() {
while (idx < queue.length) {
const item = queue[idx++]
const {server, chunk} = await createChunkWithFailover(
agent, allServers, state, concurrency, item.spec, item.chunkNo
)
sentChunks[item.index] = chunk
}
}
const allocateWorkers = Array.from(
{length: Math.min(concurrency, queue.length)},
() => allocateWorker()
)
await Promise.all(allocateWorkers)
// Stage 2: Upload — parallel across servers, sequential per server
// readChunk reads from the encrypted file by offset (same as Phase 1 uploadFile)
let uploaded = 0
const total = encrypted.chunkSizes.reduce((a, b) => a + b, 0)
const byServer = groupBy(sentChunks, c => formatXFTPServer(c.server))
await Promise.all([...byServer.entries()].map(async ([srvKey, chunks]) => {
for (const chunk of chunks) {
const chunkData = await readChunk(chunk.chunkOffset, chunk.chunkSize)
await uploadXFTPChunk(agent, chunk.server, chunk.senderKey, chunk.senderId, chunkData)
uploaded += chunk.chunkSize
options?.onProgress?.(uploaded, total)
}
}))
return buildDescriptions(encrypted, sentChunks)
}
```
**`createChunkWithFailover`** — server selection with per-chunk retry limit:
```typescript
async function createChunkWithFailover(
agent: XFTPClientAgent,
allServers: XFTPServer[],
state: UploadState,
concurrency: number,
spec: ChunkSpec,
chunkNo: number
): Promise<{server: XFTPServer, chunk: SentChunk}> {
const maxAttempts = Math.min(allServers.length, MAX_FNEW_ATTEMPTS)
for (let attempt = 0; attempt < maxAttempts; attempt++) {
const server = pickServer(allServers, state, concurrency)
try {
const chunk = await createAndPrepareChunk(agent, server, spec, chunkNo)
// Success — add to working set (if not already there)
if (!state.workingServers.some(s => formatXFTPServer(s) === formatXFTPServer(server))) {
state.workingServers.push(server)
}
return {server, chunk}
} catch (e) {
// Remove from working if it was there
state.workingServers = state.workingServers.filter(
s => formatXFTPServer(s) !== formatXFTPServer(server)
)
if (attempt === maxAttempts - 1) throw e
}
}
throw new Error("unreachable")
}
```
**`pickServer`** — two-list selection:
```typescript
function pickServer(
allServers: XFTPServer[],
state: UploadState,
concurrency: number
): XFTPServer {
// Once enough working servers found, only use those
if (state.workingServers.length >= concurrency) {
return randomPick(state.workingServers)
}
// Still exploring — pick from untried
if (state.untriedServers.length > 0) {
const idx = Math.floor(Math.random() * state.untriedServers.length)
return state.untriedServers.splice(idx, 1)[0] // remove from untried
}
// All tried — reset untried to non-working servers and retry
state.untriedServers = allServers.filter(
s => !state.workingServers.some(w => formatXFTPServer(w) === formatXFTPServer(s))
)
if (state.untriedServers.length > 0) {
const idx = Math.floor(Math.random() * state.untriedServers.length)
return state.untriedServers.splice(idx, 1)[0]
}
// Every server is working — pick any working
return randomPick(state.workingServers)
}
```
**Algorithm:** Two lists — `untriedServers` (initially all) and `workingServers` (initially empty). When `workingServers.length < concurrency`, pick from `untriedServers` (removing on pick). On FNEW success, add to `workingServers`. On FNEW failure, server is already removed from `untriedServers`; remove from `workingServers` if present. When `untriedServers` is empty, reset it to all non-working servers. Once `workingServers.length >= concurrency`, pick randomly only from `workingServers`.
**Termination condition:** Each chunk tries at most `min(serverCount, 5)` different servers. If all attempts fail, the chunk fails and the upload fails with the last error. Rationale: if 5 out of 12 servers are down, something systemic is wrong and continuing is unlikely to help. Timeouts count as failures — the timed-out server is removed from working and a different server is picked next.
**Key differences from Haskell:**
- No operator concept — just host diversity via random selection
- No database — state tracked in-memory during upload
- FNEW runs concurrently (Haskell is serial) — improves startup time
- FNEW is cheap and retried with server rotation; FPUT retries on same server
**Download changes (also Phase 2):** Default concurrency should be 4 (matching Haskell). Download already groups by server in 3.5. If `replicas[0]` download fails, try `replicas[1]`, `replicas[2]`, etc. (fallback across replicas).
## 4. Implementation Plan
### Phase 1: Error handling and connection resilience
Steps are ordered by dependency and should be implemented one by one.
#### Step 1: Fix cache key (3.6)
- Change cache key to `formatXFTPServer(server)` in `getXFTPServerClient` and `closeXFTPServerClient`
- Add import for `formatXFTPServer`
- Run existing tests to verify no regression
#### Step 2: Typed error detection for padded server errors (3.2 client-side)
- Add `XFTPRetriableError` class
- In `sendXFTPCommand`, detect padded error strings before `decodeTransmission`
- Classify `FRErr` responses as retriable or permanent with human-readable messages
- Run existing tests
#### Step 3: Fetch timeout (3.3)
- Add `TransportConfig` with `timeoutMs`
- Thread config through `createTransport``connectXFTP` → command wrappers
- Add `AbortController` to browser `fetch()` and `setTimeout` to Node.js HTTP/2
- Add vitest test: timeout triggers after configured duration
- Run existing tests
#### Step 4: Connection state with Promise-based lock and per-server queues (3.5)
- Introduce `ServerConnection` record: `{client: Promise<XFTPClient>, queue: Promise<void>}`
- Replace `XFTPClientAgent.clients: Map<string, XFTPClient>` with `connections: Map<string, ServerConnection>`
- Implement `reconnectClient` — replaces `conn.client` with new promise, preserves queue
- Implement `enqueueCommand` — chains operation onto server's queue
- Implement `removeStaleConnection` — removes entry only if current promise is the failed one
- Auto-cleanup: `p.catch(() => delete)` removes failed connections so next caller starts fresh
- Adapt `closeXFTPServerClient` and `closeXFTPAgent`
- Add vitest tests:
- Concurrent calls to same server produce single connection
- Failed promise is cleaned up, next caller gets fresh connection
#### Step 5: Automatic retry in sendXFTPCommand (3.2)
- Add retry loop with reconnect
- Change `sendXFTPCommand` signature: takes `agent + server` instead of `client`; export it (needed by tests and by agent.ts callers)
- Rename current `sendXFTPCommand``sendXFTPCommandOnce` (private); add padded error detection + FRErr classification (throw `XFTPRetriableError` for SESSION/HANDSHAKE, `XFTPPermanentError` for AUTH/NO_FILE/etc.)
- All command wrappers (`createXFTPChunk`, `uploadXFTPChunk`, etc.) pass agent + server
- Update agent.ts call sites: remove `getXFTPServerClient` calls before command wrappers (in `uploadFile`, `uploadRedirectDescription`, `downloadFileRaw`, `resolveRedirect`, `deleteFile`)
- Max 3 retries for retriable errors, immediate throw for permanent
- On retriable error: call `reconnectClient` and retry. On retriable error exhausted: call `removeStaleConnection` to clean up. On permanent error: throw immediately without touching connection
- Add vitest tests:
- Server started with delay → first attempt fails, retry succeeds
- 3 retries exhausted → error propagates with human-readable message
- Non-retriable error (AUTH) → no retry, immediate failure
#### Step 6: Server-side stale session handling (3.1)
- Add one guard to `Nothing` branch: `sniUsed && not webHello -> throwE SESSION`
- Remove debug `hPutStrLn stderr` lines (all 6 occurrences in dispatch)
- All other branches unchanged
- Run Haskell tests + Playwright tests
#### Step 7: Download with per-server grouping
- Modify `downloadFileRaw` to group chunks by server, sequential within each server (`for` loop), parallel across servers (`Promise.all`)
- Add vitest test: concurrent downloads from different servers run in parallel
#### Step 8: UI error improvements (3.4)
- Temporary errors: auto-retry loop (3 attempts), then show human-readable diagnosis + manual retry button
- Permanent errors: show human-readable error, NO retry button
- Manual retry resumes from last successful chunk (not full restart)
#### Step 9: Remove debug logging
- Remove all `console.log('[DEBUG ...]')` and `hPutStrLn stderr "DEBUG ..."` lines
- Keep `console.error('[XFTP] ...')` error logging
### Phase 2: Multi-server upload
Implement after Phase 1 is complete and tested.
#### Step 10: Multi-server upload with failover (3.7)
- Extend `SentChunk` with `chunkOffset: number` (from ChunkSpec) and `server: XFTPServer` (assigned during allocate) — Stage 2 reads data by offset and groups chunks by server
- Change `uploadFile` signature: takes `allServers: XFTPServer[]` instead of single `server`
- Implement `UploadState` with `untriedServers` and `workingServers`
- Implement `createChunkWithFailover` and `pickServer`: two-list selection (untried → working once enough found), max `min(serverCount, 5)` attempts per chunk
- Allocate stage: concurrent FNEW within concurrency limit (default 4)
- Upload stage: parallel across servers, sequential per server (reuse queue from Step 7)
- Update `web/upload.ts`: pass `getServers()` instead of `pickRandomServer(getServers())`
- Update description building: each chunk references its actual server
- Add vitest tests:
- File split across N servers (verify different servers in description)
- One server down → chunks redistributed to others
- All servers down → error after exhausting 5 attempts per chunk
#### Step 11: Download concurrency and replica fallback
- Change default download concurrency from 1 to 4
- If `replicas[0]` download fails, try `replicas[1]`, `replicas[2]`, etc.
- Uses per-server queues from Step 7
## 5. Testing Plan
### Principle
Prefer low-level vitest tests over Playwright E2E. Each new function gets one focused test. Pure functions tested without mocks; connection management tested with mock `connectXFTP`; server behavior tested with real server. Total: 13 tests across 4 files.
Tests A-C run in browser context (`@vitest/browser` with Chromium headless), configured in `vitest.config.ts`. Test D (integration) requires a separate Node.js vitest config since it uses `node:http2`. Existing `globalSetup.ts` provides a real XFTP server for integration tests.
### Test file A: `test/errors.test.ts` — pure, no server
Tests error classification and padded error detection (Steps 2, 5).
**T1. `isRetriable` classifies errors correctly**
```typescript
// Retriable:
expect(isRetriable(new XFTPRetriableError("SESSION"))).toBe(true)
expect(isRetriable(new XFTPRetriableError("HANDSHAKE"))).toBe(true)
expect(isRetriable(new TypeError("fetch failed"))).toBe(true) // network error
expect(isRetriable(Object.assign(new Error(), {name: "AbortError"}))).toBe(true) // timeout
// Not retriable:
expect(isRetriable(new XFTPPermanentError("AUTH", "..."))).toBe(false)
expect(isRetriable(new XFTPPermanentError("NO_FILE", "..."))).toBe(false)
expect(isRetriable(new XFTPPermanentError("INTERNAL", "..."))).toBe(false)
```
**T2. `categorizeError` produces human-readable messages**
```typescript
// categorizeError receives thrown errors (from sendXFTPCommandOnce or transport)
const e = categorizeError(new XFTPPermanentError("AUTH", "File is invalid, expired, or has been removed"))
expect(e.message).toContain("expired")
// Verify every permanent error type maps to a non-empty human-readable message
for (const errType of ["AUTH", "NO_FILE", "SIZE", "QUOTA", "BLOCKED", "DIGEST", "INTERNAL"]) {
expect(humanReadableMessage({type: errType}).length).toBeGreaterThan(0)
}
// Retriable errors also get human-readable messages after exhaustion
const re = categorizeError(new XFTPRetriableError("SESSION"))
expect(re.message).toContain("expired") // "Session expired, reconnecting..."
```
**T3. Padded error detection extracts error string from padded block**
```typescript
import {blockPad, blockUnpad} from '../src/protocol/transmission.js'
// Simulate server sending padded "SESSION"
const padded = blockPad(new TextEncoder().encode("SESSION"))
const raw = blockUnpad(padded)
expect(raw.length).toBeLessThan(20)
expect(new TextDecoder().decode(raw)).toBe("SESSION")
// Normal transmission block (batch count + large-encoded data) is NOT a short string
const sessionId = new Uint8Array(32) // dummy
const normalBlock = encodeTransmission(sessionId, new Uint8Array(0), new Uint8Array(0), encodePING())
const normalRaw = blockUnpad(normalBlock)
expect(normalRaw.length).toBeGreaterThan(20) // not mistaken for padded error
```
### Test file B: `test/connection.test.ts` — mock connectXFTP, no server
Tests connection management functions (Steps 4, 5). Uses `vi.mock` to replace `connectXFTP` with a controllable promise factory.
**T4. `getXFTPServerClient` coalesces concurrent calls**
```typescript
// Mock connectXFTP to return a deferred promise
const {promise, resolve} = promiseWithResolvers<XFTPClient>()
vi.mocked(connectXFTP).mockReturnValueOnce(promise)
const agent = newXFTPAgent()
const p1 = getXFTPServerClient(agent, server)
const p2 = getXFTPServerClient(agent, server)
expect(p1).toBe(p2) // same promise, single connection
resolve(mockClient)
expect(await p1).toBe(mockClient)
```
**T5. `getXFTPServerClient` auto-cleans failed connections**
```typescript
vi.mocked(connectXFTP).mockReturnValueOnce(Promise.reject(new Error("down")))
const agent = newXFTPAgent()
const p1 = getXFTPServerClient(agent, server)
await expect(p1).rejects.toThrow("down")
// After microtask, entry is removed
await new Promise(r => setTimeout(r, 0))
expect(agent.connections.has(formatXFTPServer(server))).toBe(false)
// Next call creates fresh connection
vi.mocked(connectXFTP).mockReturnValueOnce(Promise.resolve(mockClient))
const p2 = getXFTPServerClient(agent, server)
expect(p2).not.toBe(p1)
```
**T6. `removeStaleConnection` respects promise identity**
```typescript
const agent = newXFTPAgent()
const p1 = Promise.resolve(mockClient)
agent.connections.set(key, {client: p1, queue: Promise.resolve()})
// Replace with reconnect
const p2 = Promise.resolve(mockClient2)
agent.connections.set(key, {client: p2, queue: Promise.resolve()})
// removeStaleConnection with old promise does NOT remove new entry
removeStaleConnection(agent, server, p1)
expect(agent.connections.has(key)).toBe(true)
expect(agent.connections.get(key)!.client).toBe(p2)
// removeStaleConnection with current promise removes it
removeStaleConnection(agent, server, p2)
expect(agent.connections.has(key)).toBe(false)
```
**T7. `reconnectClient` replaces promise but preserves queue**
```typescript
const agent = newXFTPAgent()
const origQueue = Promise.resolve()
agent.connections.set(key, {client: Promise.resolve(mockClient), queue: origQueue})
vi.mocked(connectXFTP).mockReturnValueOnce(Promise.resolve(mockClient2))
reconnectClient(agent, server)
const conn = agent.connections.get(key)!
expect(await conn.client).toBe(mockClient2) // new client
expect(conn.queue).toBe(origQueue) // queue preserved
```
**T8. Retry loop: retriable error triggers reconnect, permanent error does not**
Mock approach: `vi.mock('../src/client.js')` to mock `connectXFTP` (exported). `reconnectClient` is not exported — its behavior is controlled indirectly via `connectXFTP` mock (it calls `connectXFTP` internally). Verify retry count via `connectXFTP` call count. Note: vitest module mocking may need adjustment depending on ESM transform behavior — if intra-module calls bypass the mock, extract `connectXFTP` to a separate module or use dependency injection for testing.
```typescript
// Script: first connectXFTP returns client whose post throws retriable,
// second connectXFTP (from reconnect) returns client whose post succeeds
vi.mocked(connectXFTP)
.mockResolvedValueOnce({
...mockClient,
transport: { post: async () => { throw new XFTPRetriableError("SESSION") }, close: () => {} }
})
.mockResolvedValueOnce({
...mockClient,
transport: { post: async () => okResponseBlock, close: () => {} }
})
const agent = newXFTPAgent()
const result = await sendXFTPCommand(agent, server, dummyKey, dummyId, encodePING())
expect(result.response.type).toBe("FROk")
expect(vi.mocked(connectXFTP)).toHaveBeenCalledTimes(2) // initial + 1 reconnect
// Reset — all 3 retries exhausted: connectXFTP called 3 times (initial + 2 reconnects)
vi.mocked(connectXFTP).mockClear()
vi.mocked(connectXFTP).mockResolvedValue({
...mockClient,
transport: { post: async () => { throw new XFTPRetriableError("SESSION") }, close: () => {} }
})
const agent2 = newXFTPAgent()
await expect(sendXFTPCommand(agent2, server, dummyKey, dummyId, encodePING()))
.rejects.toThrow(/reconnecting|expired/)
expect(vi.mocked(connectXFTP)).toHaveBeenCalledTimes(3) // initial + 2 reconnects
// Reset — permanent error: connectXFTP called once (initial only, no reconnect)
vi.mocked(connectXFTP).mockClear()
vi.mocked(connectXFTP).mockResolvedValue({
...mockClient,
transport: { post: async () => authErrorBlock, close: () => {} }
})
const agent3 = newXFTPAgent()
await expect(sendXFTPCommand(agent3, server, dummyKey, dummyId, encodePING()))
.rejects.toThrow(/expired/)
expect(vi.mocked(connectXFTP)).toHaveBeenCalledTimes(1) // initial only, no reconnect
```
### Test file C: `test/server-selection.test.ts` — pure, no server
Tests `pickServer` state machine (Step 10). Determinism: seed `Math.random` or test invariants not specific picks.
**T9. `pickServer` picks from untried when working < concurrency**
```typescript
const servers = [s1, s2, s3, s4, s5]
const state: UploadState = {untriedServers: [...servers], workingServers: []}
const picked = pickServer(servers, state, 4)
// picked is from untried, and was removed from untried
expect(state.untriedServers.length).toBe(4)
expect(state.untriedServers).not.toContainEqual(picked)
```
**T10. `pickServer` picks only from working when working >= concurrency**
```typescript
const state: UploadState = {
untriedServers: [s5], // still has untried
workingServers: [s1, s2, s3, s4]
}
const picked = pickServer(servers, state, 4)
// Must pick from working, NOT from untried
expect([s1, s2, s3, s4]).toContainEqual(picked)
expect(state.untriedServers.length).toBe(1) // untried unchanged
```
**T11. `pickServer` resets untried when exhausted**
```typescript
const state: UploadState = {
untriedServers: [], // all tried
workingServers: [s1, s2] // only 2 working, concurrency=4
}
const picked = pickServer(servers, state, 4)
// Should have reset untried to non-working servers and picked from them
expect([s3, s4, s5]).toContainEqual(picked)
expect(state.untriedServers.length).toBe(2) // 3 non-working minus 1 picked
```
### Test file D: `test/integration.test.ts` — real server, Node.js mode
Requires separate vitest config with `browser: {enabled: false}` since these tests use `node:http2` directly. Alternatively, add `test/vitest.node.config.ts` that includes only `test/integration.test.ts` and runs in Node.js.
**T12. Stale session returns padded SESSION error (requires Step 6)**
```typescript
import http2 from 'node:http2'
// Connect and handshake normally via the client
const client = await connectXFTP(server)
// Create a raw HTTP/2 session (new TLS SessionId, no handshake state on server)
const session = http2.connect(client.baseUrl, {rejectUnauthorized: false})
// Build a dummy command block using the old client's sessionId.
// Content doesn't matter — server detects stale session before parsing command.
const dummyKey = new Uint8Array(64) // Ed25519 private key (dummy)
const dummyId = new Uint8Array(24) // entity ID (dummy)
const cmdBlock = encodeAuthTransmission(client.sessionId, new Uint8Array(0), dummyId, encodePING(), dummyKey)
const resp = await new Promise<Uint8Array>((resolve, reject) => {
const req = session.request({":method": "POST", ":path": "/"})
const chunks: Buffer[] = []
req.on("data", (c: Buffer) => chunks.push(c))
req.on("end", () => resolve(new Uint8Array(Buffer.concat(chunks))))
req.on("error", reject)
req.end(Buffer.from(cmdBlock))
})
// Server should return padded "SESSION" (not crash, not "HANDSHAKE")
const raw = blockUnpad(resp.subarray(0, XFTP_BLOCK_SIZE))
expect(new TextDecoder().decode(raw)).toBe("SESSION")
session.close()
closeXFTP(client)
```
**T13. Fetch timeout fires within configured duration**
```typescript
// connectXFTP with 1ms timeout — handshake requires multiple round trips,
// so even on localhost it will exceed 1ms and trigger abort
await expect(
connectXFTP(server, {timeoutMs: 1})
).rejects.toThrow(/abort|timeout/i)
```
### What existing tests already cover (no new tests needed)
| Behavior | Covered by |
|----------|-----------|
| Cache key fix (Step 1) | Existing round-trip test — uses `formatXFTPServer` after refactor |
| Basic upload/download | 24 Playwright tests + 1 vitest browser test |
| File size limits, unicode filenames | Playwright edge case tests |
| Server startup/teardown | `globalSetup.ts` / `globalTeardown.ts` |
| Handshake + identity verification | `connectXFTP` in existing round-trip test |
### Test ordering
Tests must be added alongside their implementation step:
- **Step 2**: Add T1, T2, T3 (test/errors.test.ts)
- **Step 3**: Add T13 (test/integration.test.ts) — requires Node.js vitest config
- **Step 4**: Add T4, T5, T6, T7 (test/connection.test.ts)
- **Step 5**: Add T8 (test/connection.test.ts)
- **Step 6**: Add T12 (test/integration.test.ts) — requires server change + Node.js vitest config
- **Step 10**: Add T9, T10, T11 (test/server-selection.test.ts)
## 6. Context for Implementation Sessions
### Files to re-read on session start
**TypeScript (xftp-web/src/):**
- `client.ts``XFTPClient`, `XFTPClientAgent`, `getXFTPServerClient`, `closeXFTPServerClient`, `connectXFTP`, `sendXFTPCommand`, `createBrowserTransport`, `createNodeTransport`, all command wrappers
- `agent.ts``uploadFile`, `downloadFileRaw`, `downloadFile`, `resolveRedirect`, `encryptFileForUpload`
- `protocol/transmission.ts``encodeAuthTransmission`, `decodeTransmission`, `blockPad`, `blockUnpad`
- `protocol/commands.ts``XFTPErrorType`, `FileResponse`, `decodeResponse`, `decodeXFTPError`
- `protocol/handshake.ts``decodeServerHandshake` (padded error detection heuristic)
- `protocol/address.ts``XFTPServer`, `parseXFTPServer`, `formatXFTPServer`
- `web/upload.ts` — UI error handling, retry button
- `web/download.ts` — UI error handling, retry button
- `web/servers.ts``getServers`, `pickRandomServer`
**TypeScript (xftp-web/test/):**
- `browser.test.ts` — vitest Node.js test template (uses real Haskell server)
- `globalSetup.ts` — server startup, config generation, port file
- `page.spec.ts` — Playwright page tests
**Haskell (reference for multi-server):**
- `src/Simplex/FileTransfer/Agent.hs``createChunk` (lines 457-486, allocate stage), `runXFTPSndPrepareWorker` (lines 391-430, serial allocate in Haskell), `runXFTPSndWorker` (lines 494-548, per-server upload worker)
- `src/Simplex/Messaging/Agent/Client.hs``getNextServer_` (lines 2335-2350), `withNextSrv` (lines 2366-2385), `pickServer` (lines 2309-2314)
**Haskell (server):**
- `src/Simplex/FileTransfer/Server.hs``xftpServerHandshakeV1` (lines 165-244), `processRequest` (lines 403-435)
- `src/Simplex/Messaging/Protocol.hs``tDecodeServer` (lines 2239-2265) — sessionId verification at line 2242
### Key design constraints
1. `tDecodeServer` (Protocol.hs:2242) verifies `sessId == sessionId` — commands signed with old sessionId WILL fail on new connection
2. Server generates per-session DH key in `processHello` (Server.hs:207) — cannot be shared across sessions
3. `fetch()` provides zero control over HTTP/2 connection reuse — browser decides
4. `xftp-web-hello` header is only checked in dispatch (Server.hs:192), NOT inside `processHello`
5. Handshake-phase errors are raw padded strings; command-phase errors are proper ERR transmissions
6. Ed25519 signature verification (`TASignature` path, Protocol.hs:1314) does NOT use `thAuth` — but SMP will
7. Reconnect must re-handshake to get new sessionId AND new server DH key
8. The new `throwE SESSION` guard (Step 6) sends a raw padded "SESSION" string — no sessionId framing. Client detects this via padded error heuristic (section 3.2), not via sessionId mismatch
9. FNEW is cheap (creates chunk record on server) — retry with different server on failure
10. FPUT retries on same server (chunk replica already exists there) — close connection + backoff
## 7. Plan Maintenance
This plan must be updated as implementation proceeds:
- Mark completed steps with date
- Record any deviations from the plan with rationale
- Add new issues discovered during implementation
- Update file references if code moves
@@ -0,0 +1,327 @@
# CLI-Web Link Compatibility
## Problem
CLI and web clients are isolated: CLI outputs `.xftp` description files, web outputs
`https://host/#<encoded>` links. A file uploaded via one cannot be downloaded via the other.
## Solution Summary
Make CLI produce and consume web-compatible links so that:
- CLI `send` always outputs a web link (in addition to `.xftp` files)
- CLI `recv` accepts a web link URL as input (alternative to `.xftp` file path)
- Browser can download files uploaded by CLI and vice versa
The web page host is derived from the XFTP server address - the server that hosts the file
also hosts the download page. Making XFTP servers actually serve the web page is a separate
concern (not covered here), but the link format anticipates it.
The YAML file description format is already identical between CLI and web.
The only gap is the URI encoding layer: DEFLATE-raw compression + base64url + URL structure.
## Current State
### Web link format
```
https://<xftp-server-host>/#<base64url(deflateRaw(YAML))>
```
Encoding chain (agent.ts:64-68):
1. `encodeFileDescription(fd)` -> YAML string
2. `TextEncoder.encode(yaml)` -> bytes
3. `pako.deflateRaw(bytes)` -> compressed
4. `base64urlEncode(compressed)` -> URI fragment (no `#`)
For multi-chunk files exceeding ~400 chars in URI, a redirect description is uploaded:
the real file description is encrypted, uploaded as a separate XFTP file, and a smaller
"redirect" description (pointing to it) is put in the URI.
### CLI file format
```
xftp send FILE -> writes rcv1.xftp (raw YAML), snd.xftp.private
xftp recv FILE.xftp -> reads raw YAML from file
```
No URI support. No compression. No redirect descriptions.
### Existing Haskell `FileDescriptionURI`
`Description.hs:243-266` defines a `simplex:/file#/?desc=<URL-encoded raw YAML>` format.
This is the SimpleX Chat app format - NOT the web page format. It uses URL-encoded raw YAML
(no DEFLATE compression), and has a different URL structure.
## Detailed Tech Design
### 1. File Header (Filename) Compatibility
The filename is carried **inside the encrypted file data**, not in the file description YAML.
Both CLI and web use the same `FileHeader` structure and binary encoding - full interop.
#### FileHeader type
Haskell (`Types.hs:36-46`):
```haskell
data FileHeader = FileHeader { fileName :: Text, fileExtra :: Maybe Text }
instance Encoding FileHeader where
smpEncode FileHeader {fileName, fileExtra} = smpEncode (fileName, fileExtra)
```
TypeScript (`crypto/file.ts:11-24`):
```typescript
interface FileHeader { fileName: string; fileExtra: string | null }
function encodeFileHeader(hdr: FileHeader): Uint8Array {
return concatBytes(encodeString(hdr.fileName), encodeMaybe(encodeString, hdr.fileExtra))
}
```
Both produce identical binary: `[1-byte UTF-8 length][fileName bytes]['0']` (for null fileExtra).
Max filename: 255 UTF-8 bytes (1-byte length prefix).
#### Encrypted file structure
Both CLI and web produce the same encrypted stream:
```
XSalsa20-Poly1305 encrypted:
[8-byte Int64 fileSize] [FileHeader] [file content] ['#' padding]
+ [16-byte auth tag]
Where fileSize = len(FileHeader) + len(file content)
```
The 8-byte length prefix and padding are handled identically:
- Haskell: `Crypto.hs:43-56` (`encryptFile`) / `Crypto.hs:81-87` (`decryptFirstChunk`)
- TypeScript: `crypto/file.ts:51-70` (`encryptFile`) / `crypto/file.ts:81-94` (`decryptChunks`)
On decryption, `unPadLazy`/`splitLen` strips the 8-byte length prefix, then `parseFileHeader`
extracts the filename from the remaining decrypted bytes (up to 1024 bytes examined, both sides).
#### CLI upload: sets real filename (ok)
`Client/Main.hs:246-247,273`:
```haskell
let (_, fileNameStr) = splitFileName filePath
fileName = T.pack fileNameStr
...
fileHdr = smpEncode FileHeader {fileName, fileExtra = Nothing}
```
Extracts the actual filename from the path and embeds it in the encrypted header.
#### CLI download: uses filename from header (ok)
`Crypto.hs:62-66` (single chunk) / `Crypto.hs:72-74` (multi-chunk):
```haskell
(FileHeader {fileName}, rest) <- parseFileHeader decryptedContent
destFile <- withExceptT FTCEFileIOError $ getDestFile fileName
```
`Client/Main.hs:435-441` (`getFilePath`):
- If output dir specified: saves to `<dir>/<fileName>`
- If no dir: saves to `~/Downloads/<fileName>`
The filename from the decrypted header determines the output file name.
#### Web upload: sets real filename (ok)
`upload.ts:121` -> `agent.ts:86`:
```typescript
const fileHdr = encodeFileHeader({fileName, fileExtra: null})
```
Where `fileName` comes from `file.name` (browser File API).
#### Web download: uses filename from header (ok)
`download.ts:97,102`:
```typescript
const fileName = sanitizeFileName(header.fileName)
a.download = encodeURIComponent(fileName)
```
The web client additionally sanitizes the filename (strips path separators, control chars,
bidi overrides, limits to 255 chars).
#### Web redirect description: empty filename (correct)
`agent.ts:193`: `encryptFileForUpload(yamlBytes, "")` - redirect descriptions use empty filename
because they are internal artifacts, not user files. This is handled correctly on both sides:
the redirect content is decrypted and parsed as YAML, not saved as a file.
#### Cross-client interop: fully compatible (ok)
| Scenario | Filename flow | Status |
|----------|--------------|--------|
| CLI upload -> CLI download | `splitFileName` -> header -> `getDestFile` | Works |
| Web upload -> Web download | `File.name` -> header -> `sanitizeFileName` | Works |
| CLI upload -> Web download | `splitFileName` -> header -> `sanitizeFileName` | **Compatible** |
| Web upload -> CLI download | `File.name` -> header -> `getDestFile` | **Compatible** |
The binary encoding is identical (smpEncode). No changes needed for filename interop.
The CLI should consider adding filename sanitization similar to the web client for safety.
### 2. Web Link Host Derivation
The web page URL domain comes from the XFTP server address, not from a CLI flag:
- **Non-redirected description**: use the server host of the first chunk's first replica.
E.g., `xftp://abc=@xftp1.simplex.im` -> `https://xftp1.simplex.im/#<encoded>`
- **Redirected description**: use the server host of the redirect chunk (the outer description's
chunk that stores the encrypted inner description).
The server address format is `xftp://<keyhash>@<host>[,<host2>,...][:<port>]`.
The web link uses `https://<host>` (port 443 implied).
This means the CLI does not need a `--web-url` flag - the server address fully determines
the link. The XFTP server serving the web page is a separate deployment concern.
### 3. Web URI Encoding/Decoding in Haskell
Add two functions (new module or in `Description.hs`):
```haskell
-- Encode file description as web URI fragment (no leading #)
encodeWebURI :: FileDescription 'FRecipient -> ByteString
-- 1. Y.encode . encodeFileDescription -> YAML bytes
-- 2. deflateRaw (raw DEFLATE, no zlib/gzip header) via zlib package
-- 3. base64url encode (with padding, matching Data.ByteString.Base64.URL)
-- Decode web URI fragment (no leading #) to file description
decodeWebURI :: ByteString -> Either String (ValidFileDescription 'FRecipient)
-- 1. base64url decode
-- 2. inflateRaw (raw DEFLATE decompress)
-- 3. Y.decodeEither' -> YAMLFileDescription -> FileDescription
-- 4. validateFileDescription
-- Build full web link from file description
-- Extracts server host from first chunk replica (or redirect chunk)
fileWebLink :: FileDescription 'FRecipient -> (String, ByteString)
-- Returns (webHost, uriFragment)
-- Caller assembles: "https://" <> webHost <> "/#" <> uriFragment
```
**Dependency**: Add `zlib` to `simplexmq.cabal` (for raw DEFLATE).
The codebase already has `zstd` for message compression - `zlib` is standard and small.
The `zlib` Haskell package provides `Codec.Compression.Zlib.Raw` for raw DEFLATE
(no header/trailer), matching `pako.deflateRaw()` / `pako.inflateRaw()`.
### 4. Redirect Description Support
The CLI currently does NOT create redirect descriptions. For single-server single-recipient
uploads, most file descriptions fit in a reasonable URI even for multi-chunk files. But for
large files (many chunks x long server hostnames), the URI can exceed practical limits.
**Approach**: Match the web client threshold.
- After encoding the URI, if `length > 400` and chunks > 1, upload a redirect description.
- The redirect upload uses the same XFTP upload flow: encrypt YAML -> upload as file -> create
outer description pointing to it.
- This matches `agent.ts:152-155` exactly.
- The redirect chunk's server becomes the web link host.
For CLI download from a redirect URI, the existing `cliReceiveFile` needs extension:
- After decoding the file description, check `redirect` field.
- If present: download and decrypt the redirect chunks first to get the inner description,
then download the actual file using the inner description.
- The web client already does this (`resolveRedirect` in agent.ts:320-346).
### 5. CLI Command Changes
#### `xftp send` - always output web link
```
xftp send FILE [DIR] [-n COUNT] [-s SERVERS]
```
- Upload file as usual
- Generate web link: `https://<server-host>/#<encodeWebURI(rcvDescription)>`
- If URI exceeds threshold, upload redirect description first
- Print web link to stdout (in addition to `.xftp` file paths)
- Only generates link for the first recipient (web links are single-recipient)
**Output change**:
```
Sender file description: ./file.xftp/snd.xftp.private
Pass file descriptions to the recipient(s):
./file.xftp/rcv1.xftp
Web link:
https://xftp1.simplex.im/#eJy0VduO2zYQ...
```
#### `xftp recv` - accept URL as input
```
xftp recv <FILE.xftp | URL> [DIR]
```
- If input starts with `http://` or `https://`, extract hash fragment after `#`
- Decode: base64url -> inflateRaw -> YAML -> FileDescription
- Resolve redirect if present
- Download and decrypt as usual
The URL must be quoted on the command line (`"https://...#..."`) because `#` is a shell
comment character when unquoted.
Implementation: modify `receiveP` parser to accept URL, add `decodeWebURI` path in
`cliReceiveFile` alongside existing `getFileDescription'`.
### 6. YAML Format Compatibility
Already identical. The web `description.ts` explicitly matches Haskell `Data.Yaml` output:
- Same field names (alphabetical key order)
- Same base64url encoding for binary fields (with `=` padding)
- Same server replica colon-delimited format: `chunkNo:replicaId:replicaKey[:digest][:chunkSize]`
- Same size encoding (`kb`/`mb`/`gb` suffixes)
- Same redirect structure
**Verification**: The Playwright test suite already tests upload->download round-trips.
Adding a cross-client test (CLI upload -> web download, or web upload -> CLI download) would
validate interop end-to-end.
### 7. Server Compatibility
No server changes needed. Both clients use the same XFTP protocol (FGET, FPUT, FNEW, FACK, FDEL).
The web client adds `xftp-web-hello: 1` header for the hello handshake, but the actual file
operations are identical wire-format.
The only consideration: CLI uses native HTTP/2 (via `http2` Haskell package), web uses
browser `fetch()` API over HTTP/2. Both produce identical XFTP protocol frames.
**Note**: Making XFTP servers actually serve the web download page at `https://<host>/` is a
separate deployment/infrastructure task. This plan only establishes the link format convention
so that links are ready to work once servers serve the page.
## Implementation Plan
### Phase 1: Web URI codec in Haskell
1. Add `zlib` dependency to `simplexmq.cabal`
2. Add `encodeWebURI` / `decodeWebURI` / `fileWebLink` to `Simplex.FileTransfer.Description`
(or a new `Simplex.FileTransfer.Description.WebURI` module)
3. `fileWebLink` extracts host from first chunk's first replica server address
4. Add unit tests: encode a known FileDescription, verify output matches web client encoding
5. Add round-trip test: encode -> decode -> compare
### Phase 2: CLI `recv` accepts URL
1. Modify `ReceiveOptions` to accept `Either FilePath WebURL` for `fileDescription`
2. In `cliReceiveFile`: if URL, extract fragment after `#`, call `decodeWebURI`
3. Add redirect resolution: if `redirect /= Nothing`, download redirect chunks,
decrypt, parse inner description, then proceed with download
4. Test: upload via web page -> copy link -> `xftp recv <link>`
### Phase 3: CLI `send` outputs web link
1. After upload, call `fileWebLink` to get (host, fragment)
2. If fragment exceeds threshold, upload redirect description first, rebuild link
3. Print `https://<host>/#<fragment>` to stdout
4. Test: `xftp send FILE` -> open link in browser -> download
### Phase 4: Cross-client integration test
1. Add test: CLI send -> extract link from stdout -> Playwright browser download -> verify
2. Add test: Playwright browser upload -> extract link -> CLI recv -> verify
3. These can be shell-script or Haskell test-suite tests that spawn both clients
+415
View File
@@ -0,0 +1,415 @@
# Fix subQ deadlock: blocking writeTBQueue inside connLock
## Problem
Users report that message reception silently and permanently stops across all connections, with no error alerts. The app appears functional but no messages arrive. Recovery requires restart.
Root cause: a deadlock between worker threads holding `connLock` and the `agentSubscriber` (sole `subQ` reader).
### The deadlock mechanism
`subQ` (`TBQueue ATransmission`, capacity 4096 on mobile / 1024 on desktop) is the single pipeline between the agent layer and the chat layer. The `agentSubscriber` thread (`Commands.hs:4373`) is its **sole reader**.
Three code sites hold `connLock` and call blocking `writeTBQueue subQ` without a fullness check. When `subQ` is full, these block while holding the lock. If `agentSubscriber` simultaneously needs the same `connLock` (via `sendMessagesB_``withConnLocks`), it blocks too — creating a circular wait:
- **Worker**: holds `connLock(X)`, waits for `subQ` space (needs `agentSubscriber` to read)
- **agentSubscriber**: sole `subQ` reader, waits for `connLock(X)` (needs worker to release)
- **Result**: permanent silent deadlock — no exception, no alert, all connections blocked
### Confirmed deadlock scenarios
**Scenario 1**: Delivery worker during queue rotation test
```
Delivery worker: agentSubscriber (sole subQ reader):
withConnLock(X) [2187] readTBQueue subQ → processAgentMessageConn
...DB operations... → sendPendingGroupMessages (on CON/SENT/QCONT)
notify → writeTBQueue subQ [2238] → batchSendConnMessages → deliverMessagesB
[BLOCKED — subQ full] → withAgent sendMessagesB [synchronous]
→ sendMessagesB_ → withConnLocks({..X..}) [1708]
[BLOCKED — connLock(X) held]
```
**Scenario 2**: Async command worker during message ACK with notification
```
Async cmd worker: agentSubscriber (sole subQ reader):
tryWithLock "ICAck" [1930→1824] readTBQueue subQ → processAgentMessageConn
→ withConnLock(X) → sendPendingGroupMessages
→ ack → ackQueueMessage [1899] → sendMessagesB_ → withConnLocks({..X..})
→ sendMsgNtf [2381] [BLOCKED — connLock(X) held]
→ writeTBQueue subQ [2386]
[BLOCKED — subQ full]
```
**Scenario 3**: Synchronous `ackMessage'` API (same mechanism as Scenario 2 but from external API caller)
```
ackMessage' caller: agentSubscriber (sole subQ reader):
withConnLock(X) [2254] → sendMessagesB_ → withConnLocks({..X..})
→ ack → ackQueueMessage [2267] [BLOCKED — connLock(X) held]
→ sendMsgNtf [2381]
→ writeTBQueue subQ [2386]
[BLOCKED — subQ full]
```
### ConnId overlap verified
No guard prevents a connection undergoing queue rotation (AM_QTEST_) or ACK processing from being included in `sendMessagesB_`'s batch. During these operations, the connection has `connStatus == ConnReady`, passing all filters in `memberSendAction`.
### Cascade amplification
Once any single deadlock triggers, `subQ` never drains. ALL other threads that attempt `writeTBQueue subQ` block progressively — their locks are held forever too. The entire threading system freezes within seconds.
### Affected code sites (blocking `writeTBQueue subQ` inside `connLock`)
| Site | File | Lock line | Write line | Events written |
|------|------|-----------|------------|----------------|
| `runSmpQueueMsgDelivery::notify` | Agent.hs | 2187 | 2238 | SWITCH SPCompleted, ERR INTERNAL |
| `runSmpQueueMsgDelivery::internalErr/notifyDel` | Agent.hs | 2187 | 2238 (via notifyDel→notify) | ERR INTERNAL + delMsg |
| `ackQueueMessage::sendMsgNtf` | Agent.hs | 2254 or 1930 | 2386 | MSGNTF |
### Safe patterns that already exist in the codebase
1. **`isFullTBQueue` + pending TVar** (used at `runCommandProcessing` lines 1782-1784/1937, and `runProcessSMP` lines 3027-3029/3216):
```haskell
-- Before processing (e.g. line 1782):
pending <- newTVarIO []
-- During processing — safe notify (e.g. line 1937):
notify cmd =
let t = (corrId, connId, AEvt (sAEntity @e) cmd)
in atomically $ ifM (isFullTBQueue subQ) (modifyTVar' pendingCmds (t :)) (writeTBQueue subQ t)
-- After processing — flush (e.g. line 1784):
mapM_ (atomically . writeTBQueue subQ) . reverse =<< readTVarIO pending
```
2. **`nonBlockingWriteTBQueue`** (used at Client.hs:789, NtfSubSupervisor.hs:507):
```haskell
nonBlockingWriteTBQueue q x = do
sent <- atomically $ tryWriteTBQueue q x
unless sent $ void $ forkIO $ atomically $ writeTBQueue q x
```
Note: `nonBlockingWriteTBQueue` does NOT preserve ordering — the spawned background thread may complete out of order relative to subsequent direct writes from the same calling thread.
### Exhaustive proof: no other deadlock scenarios exist
All 15 `withConnLock` sites in Agent.hs were analyzed. Only 3 write to `subQ`:
| withConnLock site | Writes subQ? | Safe? |
|-------------------|-------------|-------|
| switchConnectionAsync' (899) | No | ✓ |
| setConnShortLinkAsync' (995) | No | ✓ |
| setConnShortLink' (1031) | No | ✓ |
| deleteConnShortLink' (1075) | No | ✓ |
| allowConnection' (1407) | No | ✓ |
| acceptContact' (1417) | No | ✓ |
| sendMessagesB_ (1708, `withConnLocks`) | No | ✓ |
| tryWithLock/runSmpCommand (1930) | Yes (1937) | ✓ — `isFullTBQueue` check |
| tryMoveableWithLock/runSmpCommand (1931) | Yes (1937) | ✓ — `isFullTBQueue` check |
| **runSmpQueueMsgDelivery AM_QTEST_ (2187)** | **Yes (2238)** | **✗ — DEADLOCK** |
| **ackMessage' (2254)** | **Yes (2386)** | **✗ — DEADLOCK** |
| switchConnection' (2298) | No | ✓ |
| abortConnectionSwitch' (2328) | No | ✓ |
| synchronizeRatchet' (2351) | No | ✓ |
| suspendConnection' (2390) | No | ✓ |
| **processSMP (3037)** | Yes (3216) | ✓ — `isFullTBQueue` check |
Note: `processSMP` (line 3037) holds `connLock` and its local `notify` (line 3216) writes to `subQ`, but it uses the safe `isFullTBQueue` pattern. Its `ack` (line 3196) uses `enqueueCmd` (DB-only), NOT `ackQueueMessage`. The actual `ackQueueMessage` runs later from the async command worker via ICAck/ICAckDel.
Other lock pairs checked — no circular dependencies:
- `connLock × DB MVar`: DB never acquires connLock
- `entityLock × connLock`: consistent ordering (entity first in chat, conn in agent)
- `connLock(X) × connLock(Y)`: single agentSubscriber thread, one `withConnLocks` at a time
---
## Deadlock call graph: agentSubscriber → connLock
All deadlock paths require `agentSubscriber` to synchronously acquire `connLock`. Exhaustive analysis shows that **every such path converges on a single agent function**: `sendMessagesB_``withConnLocks` (Agent.hs:1708). No other agent API function called synchronously from the agentSubscriber acquires connLock.
Verified (FACT): `ackMessageAsync``enqueueCommand` only (no connLock). `toggleConnectionNtfs` → no lock. `deleteConnectionAsync``deleteLock` not `connLock`. `joinConnectionAsync``withInvLock` not `connLock`.
Also verified (FACT): `Lock = TMVar Text` (Lock.hs:24) is **non-reentrant** — double acquisition on the same thread deadlocks.
### All 22 trigger paths
Every path goes through `deliverMessage`/`deliverMessages`/`deliverMessagesB``withAgent sendMessagesB``sendMessagesB_``withConnLocks`:
| # | Trigger | Chat function | ConnIds locked | Risk |
|---|---------|--------------|----------------|------|
| 1 | Group CON (Invitee) | `introduceToAll` → broadcast XGrpMemNew | **ALL member connIds** | **HIGHEST** |
| 2 | Group MSG XGrpLinkAcpt | `introduceToRemaining` → broadcast | **ALL member connIds** | **HIGHEST** |
| 3 | Group CON (Invitee) | `sendIntroductions` → batch intros to new member | new member connId | Medium |
| 4 | Group CON (Invitee) | `sendHistory` → batch to new member | new member connId | Medium |
| 5 | Group CON | `sendPendingGroupMessages` | member connId | Medium |
| 6 | Group SENT | `sendPendingGroupMessages` | member connId | Medium |
| 7 | Group QCONT | `sendPendingGroupMessages` | member connId | Medium |
| 8 | Group CON (PendingReview) | `introduceToModerators` → to moderators | moderator connIds | Medium |
| 9 | Group CON (PreMember) | `sendXGrpMemCon` → to host | host connId | Low |
| 10 | Group CON (PreMember) | `probeMatchingMemberContact` → probes + hashes | member + N matching connIds | Medium |
| 11 | Direct CON | `probeMatchingMembers` → probes + hashes | contact + N matching connIds | Medium |
| 12 | Direct JOINED | `sendAutoReply` | contact connId | Low |
| 13 | Group JOINED | `sendGroupAutoReply` | member connId | Low |
| 14 | Group INV | `sendXGrpMemInv` → to host | host connId | Low |
| 15 | Group INV (legacy) | `sendGrpInvitation` → to contact | contact connId | Low |
| 16 | Group MSG XGrpMemInv | `xGrpMemInv``sendGroupMemberMessage` | re-member connId | Low |
| 17 | Group MSG XGrpMemDel | `forwardToMember` | deleted member connId | Low |
| 18 | Group MSG XGrpLinkMem | `probeMatchingMemberContact` | member + N matching connIds | Medium |
| 19 | Group MSG (dup relay) | `saveGroupRcvMsg` error → `sendDirectMemberMessage` | forwarder connId | Low |
| 20 | SFDONE | `sendFileDescriptions` → to recipients | recipient connIds | Medium |
| 21 | Group MSG XGrpLinkAcpt | `sendHistory` → to accepted member | accepted member connId | Medium |
| 22 | Direct MSG (autoAccept) | `autoAcceptFile` → inline accept reply | contact connId | Low (test-only config) |
### Key observations
1. **Single bottleneck**: All 22 paths converge on `sendMessagesB_``withConnLocks` (Agent.hs:1708). The deadlock is between this lock acquisition and any worker thread holding `connLock` + blocking on `writeTBQueue subQ`.
2. **Highest-risk paths** (#1, #2): Broadcasting to ALL group members in `introduceToAll` / `introduceToRemaining` acquires `withConnLocks` on ALL member connIds in a single batch. For large groups, this holds the agentSubscriber thread for a long time, during which subQ fills, which causes worker threads holding connLock on any of those connIds to deadlock.
3. **Medium-risk paths** (#5-7): `sendPendingGroupMessages` fires on every CON/SENT/QCONT. These are frequent and lock the member's connId, which is the SAME connId that a delivery worker or ACK worker may hold while writing to subQ.
---
## Analysis: `withConnLocks` in `sendMessagesB_`
### FACT: the lock protects ratchet encryption state
`sendMessagesB_` (Agent.hs:1708-1713) acquires `withConnLocks` and executes:
1. **`getConn_`** — reads connection metadata, send queues from DB
2. **`setConnPQSupport`** — updates PQ encryption flag per connection
3. **`enqueueMessagesB`** → `enqueueMessageB``storeSentMsg_` which calls:
- **`updateSndIds`** (AgentStore.hs:899) — increments `internalSndId` (sequential send counter)
- **`agentRatchetEncryptHeader`** (Agent.hs:3698) — reads current ratchet via `getRatchetForUpdate`, encrypts message header via `rcEncryptHeader`, writes advanced ratchet state via `updateRatchet`
- **`createSndMsg`** + **`createSndMsgDelivery`** — inserts message and delivery records
All operations run within `unsafeWithStore``withTransaction` (single DB transaction per batch).
### FACT: the lock CANNOT be removed
Without `withConnLocks`, concurrent `sendMessagesB_` calls targeting the same connection would:
- Read the same ratchet state, both encrypt, one overwrite the other → **ratchet desync** (unrecoverable)
- Get duplicate `internalSndId` values → **message ID collision**
- Race on `setConnPQSupport` → **PQ state inconsistency**
The lock serializes ALL operations on the connection's encryption state. Removing it would introduce data corruption.
Note: `sendMessage` (singular, line 530) uses the same `sendMessagesB_` function — there is no lock-free send path.
### Eliminated strategies
- **Strategy C (remove lock)**: The lock protects ratchet encryption. Removing it causes unrecoverable ratchet desync. Eliminated.
- **Strategy A (async dispatch)**: All 22 chat-layer callers use `deliverMessagesB` return values (delivery IDs, PQ state) synchronously. `forkIO` loses results. Eliminated.
- **Strategy W (isFullTBQueue + pending TVar)**: The existing pattern (lines 1937, 3216) buffers events in a local TVar and flushes after lock release. Between lock release and flush, another thread can acquire the same connLock and write events to subQ — reordering events within the same connection. This trades a visible deadlock for invisible ordering bugs. Eliminated.
- **Strategy O (per-connection overflow queues)**: Bounded overflow queues with "drop when full" were analyzed. Drop consequences are unacceptable at 5 of 6 write sites — CONF, INFO, CON cause permanent connection failure after ACK; INV loses connection invitations; SENT/MERR leave messages stuck forever. Unbounded overflow defeats backpressure. Eliminated.
---
## Solution: move subQ writes outside connLock
### Root cause
The `writeTBQueue subQ` calls at the 3 deadlock sites are inside `connLock` by accident of code structure, not necessity. `connLock` protects ratchet encryption state and DB consistency. The `notify` calls write informational events to `subQ` — they do not modify any state that `connLock` protects.
Moving the writes outside the lock scope eliminates the deadlock: blocking `writeTBQueue subQ` without holding `connLock` is safe — agentSubscriber is free to acquire the lock, process events, and drain `subQ`.
### Why reordering doesn't matter at these sites
The chat layer handlers for the 3 deadlock site events do NOT advance the ratchet:
| Event | Chat handler | Calls sendMessagesB_? |
|-------|-------------|----------------------|
| SWITCH SPCompleted | Creates internal chat item, updates UI | **No** |
| ERR INTERNAL | Logs error to view | **No** |
| MSGNTF | `toView CEvtNtfMessage` → empty output | **No** |
Events that DO trigger ratchet advances (CON, SENT, QCONT → `sendPendingGroupMessages``sendMessagesB_`) are all already written OUTSIDE `connLock` in the current code.
Ratchet state lives in the DB, not in subQ events. agentSubscriber processes events sequentially regardless of arrival order. The SENT-before-SWITCH race already exists in the current code (new queue worker writes SENT outside connLock while old queue worker writes SWITCH inside connLock).
### Fix: Site 1 — `runSmpQueueMsgDelivery` AM_QTEST_ (line 2187)
Restructure `withConnLock` to return the event, write outside.
**Current code** (Agent.hs:2187-2214):
```haskell
AM_QTEST_ -> withConnLock c connId "runSmpQueueMsgDelivery AM_QTEST_" $ do
withStore' c $ \db -> setSndQueueStatus db sq Active
SomeConn _ conn <- withStore c (`getConn` connId)
case conn of
DuplexConnection cData' rqs sqs -> do
let addr = qAddress sq
case findQ addr sqs of
Just SndQueue {dbReplaceQueueId = Just replacedId, primary} ->
case removeQP (\sq' -> dbQId sq' == replacedId && not (sameQueue addr sq')) sqs of
Nothing -> internalErr msgId "sent QTEST: queue not found in connection"
Just (sq', sq'' : sqs') -> do
checkSQSwchStatus sq' SSSendingQTEST
atomically $ TM.delete (qAddress sq') $ smpDeliveryWorkers c
withStore' c $ \db -> do
when primary $ setSndQueuePrimary db connId sq
deletePendingMsgs db connId sq'
deleteConnSndQueue db connId sq'
let sqs'' = sq'' :| sqs'
conn' = DuplexConnection cData' rqs sqs''
cStats <- connectionStats c conn'
notify $ SWITCH QDSnd SPCompleted cStats -- DEADLOCK
_ -> internalErr msgId "sent QTEST: ..." -- DEADLOCK (via notifyDel → notify)
_ -> internalErr msgId "sent QTEST: ..." -- DEADLOCK
_ -> internalErr msgId "QTEST sent not in duplex ..." -- DEADLOCK
```
**New code:**
```haskell
AM_QTEST_ -> do
evt_ <- withConnLock c connId "runSmpQueueMsgDelivery AM_QTEST_" $ do
withStore' c $ \db -> setSndQueueStatus db sq Active
SomeConn _ conn <- withStore c (`getConn` connId)
case conn of
DuplexConnection cData' rqs sqs -> do
let addr = qAddress sq
case findQ addr sqs of
Just SndQueue {dbReplaceQueueId = Just replacedId, primary} ->
case removeQP (\sq' -> dbQId sq' == replacedId && not (sameQueue addr sq')) sqs of
Nothing -> pure $ Left "sent QTEST: queue not found in connection"
Just (sq', sq'' : sqs') -> do
checkSQSwchStatus sq' SSSendingQTEST
atomically $ TM.delete (qAddress sq') $ smpDeliveryWorkers c
withStore' c $ \db -> do
when primary $ setSndQueuePrimary db connId sq
deletePendingMsgs db connId sq'
deleteConnSndQueue db connId sq'
let sqs'' = sq'' :| sqs'
conn' = DuplexConnection cData' rqs sqs''
cStats <- connectionStats c conn'
pure $ Right $ SWITCH QDSnd SPCompleted cStats
_ -> pure $ Left "sent QTEST: there is only one queue in connection"
_ -> pure $ Left "sent QTEST: queue not in connection or not replacing another queue"
_ -> pure $ Left "QTEST sent not in duplex connection"
-- subQ write is now OUTSIDE connLock — blocking writeTBQueue is safe
case evt_ of
Right evt -> notify evt
Left err -> internalErr msgId err
```
All DB operations remain inside the lock. Only `notify`/`internalErr` (which write to subQ) move outside. `internalErr` calls `notifyDel` = `notify >> delMsg` — both `notify` (subQ write) and `delMsg` (`deleteSndMsgDelivery`, keyed on unique msgId) are safe outside the lock. The existing double-delete pattern (`delMsg` inside `internalErr` + `delMsgKeep` at line 2216) is preserved.
### Fix: Sites 2 & 3 — `ackQueueMessage::sendMsgNtf` (line 2386)
Change `ackQueueMessage` to return the MSGNTF event instead of writing it to subQ. Callers write to subQ after releasing connLock.
**Current code** (Agent.hs:2371-2386):
```haskell
ackQueueMessage :: AgentClient -> RcvQueue -> SMP.MsgId -> AM ()
ackQueueMessage c rq@RcvQueue {userId, connId, server} srvMsgId = do
atomically $ incSMPServerStat c userId server ackAttempts
tryAllErrors (sendAck c rq srvMsgId) >>= \case
Right _ -> sendMsgNtf ackMsgs
Left (SMP _ SMP.NO_MSG) -> sendMsgNtf ackNoMsgErrs
Left e -> ...
where
sendMsgNtf stat = do
atomically $ incSMPServerStat c userId server stat
whenM (liftIO $ hasGetLock c rq) $ do
atomically $ releaseGetLock c rq
brokerTs_ <- eitherToMaybe <$> tryAllErrors (withStore c $ \db -> getRcvMsgBrokerTs db connId srvMsgId)
atomically $ writeTBQueue (subQ c) ("", connId, AEvt SAEConn $ MSGNTF srvMsgId brokerTs_)
```
**New code** — return `Maybe ATransmission` instead of writing:
```haskell
ackQueueMessage :: AgentClient -> RcvQueue -> SMP.MsgId -> AM (Maybe ATransmission)
ackQueueMessage c rq@RcvQueue {userId, connId, server} srvMsgId = do
atomically $ incSMPServerStat c userId server ackAttempts
tryAllErrors (sendAck c rq srvMsgId) >>= \case
Right _ -> sendMsgNtf ackMsgs
Left (SMP _ SMP.NO_MSG) -> sendMsgNtf ackNoMsgErrs
Left e -> ... >> pure Nothing
where
sendMsgNtf stat = do
atomically $ incSMPServerStat c userId server stat
ifM (liftIO $ hasGetLock c rq)
(do atomically $ releaseGetLock c rq
brokerTs_ <- eitherToMaybe <$> tryAllErrors (withStore c $ \db -> getRcvMsgBrokerTs db connId srvMsgId)
pure $ Just ("", connId, AEvt SAEConn $ MSGNTF srvMsgId brokerTs_))
(pure Nothing)
```
**Caller 1: `ackMessage'`** (Agent.hs:2253-2267) — return event from `withConnLock`, write after:
```haskell
ackMessage' c connId msgId rcptInfo_ = do
t_ <- withConnLock c connId "ackMessage" $ do
SomeConn _ conn <- withStore c (`getConn` connId)
case conn of
DuplexConnection {} -> do
t_ <- ack
sendRcpt conn
del
pure t_
RcvConnection {} -> do
t_ <- ack
del
pure t_
SndConnection {} -> throwE $ CONN SIMPLEX "ackMessage"
ContactConnection {} -> throwE $ CMD PROHIBITED "ackMessage: ContactConnection"
NewConnection _ -> throwE $ CMD PROHIBITED "ackMessage: NewConnection"
-- subQ write is OUTSIDE connLock
case t_ of
Just t -> atomically $ writeTBQueue (subQ c) t
Nothing -> pure ()
```
**Caller 2: `ICAck` / `ICAckDel`** (Agent.hs:1823-1824) — inline `tryWithLock` as `tryCommand` + `withConnLock`, write subQ between the two scopes:
`tryWithLock name = tryCommand . withConnLock c connId name` — by inlining, the subQ write can be placed outside `withConnLock` but inside `tryCommand` (retaining retry/error handling).
```haskell
ICAck rId srvMsgId -> withServer $ \srv ->
tryCommand $ do
t_ <- withConnLock c connId "ICAck" $ ack srv rId srvMsgId
-- subQ write is OUTSIDE connLock — cannot deadlock with agentSubscriber
forM_ t_ $ atomically . writeTBQueue subQ
ICAckDel rId srvMsgId msgId -> withServer $ \srv ->
tryCommand $ do
t_ <- withConnLock c connId "ICAckDel" $ do
t_ <- ack srv rId srvMsgId
withStore' c (\db -> deleteMsg db connId msgId)
pure t_
-- subQ write is OUTSIDE connLock — cannot deadlock with agentSubscriber
forM_ t_ $ atomically . writeTBQueue subQ
```
Where `ack` now returns `AM (Maybe ATransmission)`:
```haskell
ack srv rId srvMsgId = do
rq <- withStore c $ \db -> getRcvQueue db connId srv rId
ackQueueMessage c rq srvMsgId
```
All subQ writes for MSGNTF are now outside connLock. FIFO ordering is preserved — no `nonBlockingWriteTBQueue`, no forked threads. The same thread that held the lock writes to subQ sequentially after releasing it.
### Race analysis
Window between connLock release and subQ write at Site 1:
| Thread | Can acquire connLock(X)? | Writes subQ? | Consequence |
|--------|-------------------------|-------------|-------------|
| agentSubscriber via sendMessagesB_ | Yes | **No** (encrypts only) | No race |
| processSMP for connId X | Yes | Yes (pending flush) | MSG before SWITCH — cosmetic |
| runCommandProcessing for connId X | Yes | Yes (pending flush) | Command response before SWITCH — cosmetic |
| New queue delivery worker | No (SENT outside lock) | Yes | SENT before SWITCH — cosmetic, **already exists in current code** |
All races are cosmetic UI ordering. None affect ratchet state, protocol correctness, or message delivery.
### Summary of changes
| File | Change | Lines affected |
|------|--------|---------------|
| Agent.hs | Restructure AM_QTEST_ to return event from `withConnLock`, write outside | ~2187-2214 |
| Agent.hs | Change `ackQueueMessage` return type to `AM (Maybe ATransmission)`, return event instead of writing | ~2371-2386 |
| Agent.hs | `ackMessage'`: return event from `withConnLock`, write outside | ~2253-2267 |
| Agent.hs | `ICAck`/`ICAckDel`: inline `tryCommand` + `withConnLock`, write subQ between scopes | ~1823-1824 |
| Agent.hs | `ack` helper: propagate new return type | ~1899-1901 |
No new data structures. No new modules. No changes to other write sites (1937, 3216 — already safe). ~25 lines changed total.
+299
View File
@@ -0,0 +1,299 @@
# SMP Agent for Browser — Web Widget Infrastructure
## 1. Problem & Goal
The SimpleX web widget needs to create duplex connections, send and receive encrypted messages, and handle the full SMP agent lifecycle — all running in the browser. This requires a TypeScript implementation of the SMP protocol stack: encoding, transport, client, and agent layers, mirroring the Haskell implementation in simplexmq.
This document covers the protocol infrastructure that lives in the simplexmq repository (`smp-web/`). The widget UI and chat-layer semantics (contact addresses, business addresses, group links) live in simplex-chat.
## 2. Architecture
Four layers, mirroring the Haskell codebase:
```
┌─────────────────────────────────────────────────────────────┐
│ Agent Layer │
│ Duplex connections, X3DH key agreement, double ratchet, │
│ message delivery, queue rotation, connection lifecycle │
├─────────────────────────────────────────────────────────────┤
│ Client Layer │
│ Connection pool (per server), command/response correlation, │
│ reconnection, backoff │
├─────────────────────────────────────────────────────────────┤
│ Transport Layer │
│ WebSocket, SMP handshake, block framing (16384 bytes), │
│ block encryption (X25519 DH + SbChainKeys) │
├─────────────────────────────────────────────────────────────┤
│ Protocol Layer │
│ SMP commands (NEW, KEY, SUB, SEND, ACK, etc.), │
│ binary encoding, transmission format │
├─────────────────────────────────────────────────────────────┤
│ Shared (from xftp-web) │
│ encoding.ts, secretbox.ts, padding.ts, keys.ts, digest.ts │
└─────────────────────────────────────────────────────────────┘
▼ WebSocket (TLS via browser)
┌───────────────┐
│ SMP Server │
│ (SNI → Warp │
│ → WS upgrade)│
└───────────────┘
```
### Core Principle: Mirror Haskell Structure
TypeScript code mirrors the Haskell module hierarchy as closely as possible. Each Haskell module has a corresponding TypeScript file, placed in the same relative path. Functions keep the same names. This enables:
- Easy cross-reference between codebases
- Sync as protocol evolves
- Code review by people who know the Haskell side
- Byte-for-byte testing of corresponding functions
### File Structure
```
smp-web/
├── src/
│ ├── protocol.ts ← SMP commands, transmission format
│ ├── protocol/
│ │ └── types.ts ← protocol types
│ ├── version.ts ← version range negotiation
│ ├── transport.ts ← handshake, block framing, THandle
│ ├── transport/
│ │ └── websockets.ts ← WebSocket connection
│ ├── client.ts ← connection pool, correlation, reconnect
│ ├── crypto/
│ │ ├── ratchet.ts ← double ratchet
│ │ └── shortLink.ts ← HKDF, link data decrypt
│ └── agent/
│ ├── protocol.ts ← connection types, link data parsing
│ └── client.ts ← connection lifecycle, message delivery
├── package.json
└── tsconfig.json
```
Encoding and crypto primitives are imported directly from xftp-web (npm dependency). New files are only created where SMP-specific logic is needed.
### Haskell Module → TypeScript File Mapping
| Haskell Module | TypeScript File | Source |
|---|---|---|
| `Simplex.Messaging.Encoding` | xftp-web `protocol/encoding.ts` | import directly |
| `Simplex.Messaging.Crypto` | xftp-web `crypto/*` | import directly |
| `Simplex.Messaging.Protocol` | `protocol.ts` | new |
| `Simplex.Messaging.Protocol.Types` | `protocol/types.ts` | new |
| `Simplex.Messaging.Version` | `version.ts` | new |
| `Simplex.Messaging.Transport` | `transport.ts` | new |
| `Simplex.Messaging.Transport.WebSockets` | `transport/websockets.ts` | new |
| `Simplex.Messaging.Client` | `client.ts` | new |
| `Simplex.Messaging.Crypto.Ratchet` | `crypto/ratchet.ts` | new |
| `Simplex.Messaging.Crypto.ShortLink` | `crypto/shortLink.ts` | new |
| `Simplex.Messaging.Agent.Protocol` | `agent/protocol.ts` | new |
| `Simplex.Messaging.Agent.Client` | `agent/client.ts` | new |
Function names in TypeScript match Haskell names (camelCase preserved). When a Haskell function is `smpClientHandshake`, TypeScript has `smpClientHandshake`. When Haskell has `contactShortLinkKdf`, TypeScript has `contactShortLinkKdf`.
## 3. Relationship to xftp-web
xftp-web (`simplexmq-2/xftp-web/`) is a production TypeScript XFTP client. smp-web reuses its foundations:
**Reused directly (npm dependency)**:
- `protocol/encoding.ts` — Decoder class, Word16/Word32/Int64, ByteString, Large, Bool, Maybe, List encoding
- `crypto/secretbox.ts` — XSalsa20-Poly1305 (cbEncrypt/cbDecrypt, streaming)
- `crypto/padding.ts` — Block padding (2-byte length prefix + `#` fill)
- `crypto/keys.ts` — Ed25519, X25519 key generation, signing, DH, DER encoding
- `crypto/digest.ts` — SHA-256, SHA-512
- `crypto/identity.ts` — X.509 certificate chain parsing, signature verification
**New in smp-web**:
- SMP protocol commands and transmission format
- SMP handshake (different from XFTP handshake)
- WebSocket transport (XFTP uses HTTP/2 fetch)
- SMP client with queue-based correlation
- Agent layer (connections, ratchet, message processing)
- Short link operations (HKDF-SHA512, link data parsing)
**Same build pattern**:
- TypeScript strict, ES2022 modules
- `tsc``dist/`
- Haskell tests via `callNode` (same function from XFTPWebTests)
- Each TypeScript function verified byte-for-byte against Haskell
## 4. Server Changes
### Done
- `attachStaticAndWS` — unified HTTP + WebSocket handler via `wai-websockets`
- SNI-based routing: browser (SNI) → Warp → WebSocket upgrade → SMP over WS; native (no SNI) → SMP over TLS
- `acceptWSConnection` — constructs `WS 'TServer` from TLS connection + Warp PendingConnection, preserves peer cert chain
- `AttachHTTP` takes `TLS 'TServer` (not raw Context), enabling proper cert chain forwarding
- Test: `testWebSocketAndTLS` verifies native TLS and WebSocket clients on same port
### Remaining
- CORS headers for cross-origin widget embedding (pattern available in XFTP server)
- Server CLI configuration for enabling/disabling WebSocket support per port
## 5. Build Approach
Bottom-up, function-by-function. Each TypeScript function tested against its Haskell counterpart before building the next.
**Test infrastructure**: `SMPWebTests.hs` reuses `callNode`, `jsOut`, `jsUint8` from `XFTPWebTests.hs` (generalized, not copied).
**Pattern**: for each function:
1. Implement in TypeScript
2. Write Haskell test that calls it via `callNode`
3. Compare output byte-for-byte with Haskell reference
4. Also test cross-language: Haskell encodes → TypeScript decodes, and vice versa
## 6. Implementation Phases
### Phase 1: Protocol Encoding + Handshake
Foundation layer. SMP-specific binary encoding and handshake.
**Functions**:
- SMP transmission format: `[auth ByteString][corrId ByteString][entityId ByteString][command]`
- `encodeTransmission` / `parseTransmission`
- `parseSMPServerHandshake` — versionRange, sessionId, authPubKey (CertChainPubKey)
- `encodeSMPClientHandshake` — version, keyHash, authPubKey, proxyServer, clientService
- Server certificate chain verification (reuse xftp-web identity.ts)
- Version negotiation
**Key encoding details**:
- `authPubKey` uses `encodeAuthEncryptCmds`: Nothing → empty (0 bytes), Just → raw smpEncode (NOT Maybe 0/1 prefix)
- `proxyServer`: Bool 'T'/'F' (v14+)
- `clientService`: Maybe '0'/'1' (v16+)
### Phase 2: SMP Commands
All commands needed for messaging.
**Sender**: SKEY, SEND
**Receiver**: NEW, KEY, SUB, ACK, OFF, DEL
**Link**: LGET
**Common**: PING
**For each command**: encode function + decode function for its response, tested against Haskell.
### Phase 3: Transport
WebSocket connection with SMP block framing.
**Functions**:
- WebSocket connect (`wss://` URL)
- Block send/receive (16384-byte binary frames)
- SMP handshake over WebSocket
- Block encryption: X25519 DH → HKDF-SHA512 → SbChainKeys → per-block XSalsa20-Poly1305
**Block encryption flow**:
1. Client generates ephemeral X25519 keypair, sends public key in handshake
2. Server sends its signed DH key in handshake
3. Both sides compute DH shared secret
4. `sbcInit(sessionId, dhSecret)` → two 32-byte chain keys (HKDF-SHA512)
5. Each block: `sbcHkdf(chainKey)` → ephemeral key + nonce, advance chain
6. Encrypt/decrypt with XSalsa20-Poly1305, blockSize-16 padding target
### Phase 4: Client
Connection management layer.
**Functions**:
- Connection pool: one WebSocket per SMP server
- Command/response correlation via corrId
- Send queue + receive queue (ABQueue pattern from simplexmq-js)
- Automatic reconnection with exponential backoff
- Timeout handling
### Phase 5: Agent — Connection Establishment
Duplex SMP connections with X3DH key agreement.
**Functions**:
- Create receive queue (NEW)
- Join connection via invitation URI
- X3DH key agreement
- Send confirmation (SKEY + SEND)
- Complete handshake (HELLO exchange)
- Connection state machine
### Phase 6: Agent — Double Ratchet
Message encryption/decryption.
**Functions**:
- Signal double ratchet implementation
- Header encryption
- Ratchet state management
- Key derivation (HKDF)
- Message sequence + hash chain verification
### Phase 7: Agent — Message Delivery
Send and receive messages through established connections.
**Functions**:
- Send path: encrypt → encode agent envelope → SEND → handle OK/delivery receipt
- Receive path: SUB → receive MSG → decrypt → verify → ACK
- Delivery receipts
- Message acknowledgment
### Phase 8: Short Links
Entry point for the widget — parse short link, fetch profile.
**Functions**:
- Parse short link URI (contact, group, business address types)
- HKDF key derivation (SHA-512): `contactShortLinkKdf`
- LGET command → LNK response
- Decrypt link data (XSalsa20-Poly1305)
- Parse FixedLinkData, ConnLinkData, UserLinkData
- Extract profile JSON
## 7. Persistence
Agent state (keys, ratchet, connections, messages) must persist across page reloads.
**Open question**: storage backend.
Options:
- **IndexedDB directly** — universal browser support, async API, no additional dependencies. Downside: key-value semantics, no SQL queries, manual indexing.
- **SQLite in browser** — sql.js (WASM-compiled SQLite) or wa-sqlite (with OPFS backend for persistence). Upside: matches Haskell agent's SQLite storage, schema can mirror `Simplex.Messaging.Agent.Store.SQLite`. Downside: additional dependency, WASM bundle size.
- **OPFS + SQLite** — Origin Private File System for durable storage, SQLite for structured access. Best durability, but limited browser support (no Safari private browsing).
**Decision criteria**: how closely we want to mirror the Haskell agent's storage schema, bundle size budget, browser compatibility requirements.
## 8. Testing Strategy
### Unit Tests (per function)
Haskell tests in `SMPWebTests.hs` using `callNode` pattern:
- TypeScript function called via Node.js subprocess
- Output compared byte-for-byte with Haskell reference
- Cross-language tests: encode in one language, decode in the other
### Integration Tests
Against live SMP server (spawned by test setup, same pattern as xftp-web globalSetup.ts):
- WebSocket connect + handshake
- Command round-trips (NEW, KEY, SUB, SEND, ACK)
- Message delivery through server
- Reconnection after disconnect
### Browser Tests
Vitest + Playwright (same as xftp-web):
- Full connection lifecycle in browser environment
- WebSocket transport in real browser
- Persistence round-trips
## 9. Security Model
Same principles as xftp-web:
- **TLS via browser** — browser handles certificate validation for WSS connections
- **SNI routing** — browser connections use SNI, routed to Warp + WebSocket handler
- **Server identity** — verified via certificate chain in SMP handshake (keyHash from short link or known servers)
- **Block encryption** — X25519 DH + SbChainKeys provides forward secrecy per block, on top of TLS
- **End-to-end encryption** — double ratchet between agent peers, server sees only encrypted blobs
- **No server-side secrets** — all keys derived and stored client-side
- **CORS** — required for cross-origin widget embedding, safe because SMP requires auth on every command
- **CSP** — strict content security policy for widget page
**Threat model**: same as xftp-web. Primary risk is page substitution (malicious JS). Mitigated by HTTPS, CSP, SRI, and optionally IPFS hosting with published fingerprints.
@@ -0,0 +1,359 @@
# SMP Agent Web: Spike Plan
Revision 4, 2026-03-20
Parent RFC: [2026-03-20-smp-agent-web.md](../2026-03-20-smp-agent-web.md)
## Revision History
- **Rev 4**: Aligned with RFC. Restructured as bottom-up build plan with per-function Haskell tests. Router WebSocket support done. File structure mirrors Haskell modules.
- **Rev 3**: Fixed multiple encoding errors discovered during audit (see encoding details below).
## Objective
Fetch and display business/contact profile from a SimpleX short link URI, via WebSocket to SMP router. This is the first milestone of the SMP agent web implementation — it proves the protocol encoding, transport, crypto, and data parsing layers work end-to-end.
The spike is not throwaway code. It is the beginning of the `smp-web/` TypeScript library, built bottom-up with each function tested against its Haskell counterpart.
## What This Proves
- WebSocket transport to SMP router works from browser
- SMP protocol encoding is correct (binary format, not ASCII)
- SMP handshake works (version negotiation, server certificate parsing)
- Crypto is compatible (HKDF-SHA512, XSalsa20-Poly1305)
- Short link data parsing matches Haskell (FixedLinkData, ConnLinkData, profile)
## Success Criteria
Haskell test creates a short link, TypeScript fetches and decodes it via WebSocket, profile data matches.
## Protocol Flow
```
1. Parse short link URI
https://simplex.chat/c#<linkKey>?h=hosts&p=port&c=keyHash
→ server, linkKey
2. Derive keys (HKDF-SHA512)
linkKey → (linkId, sbKey)
3. WebSocket connect
wss://server:443 (TLS handled by browser)
4. SMP handshake
← SMPServerHandshake {sessionId, smpVersionRange, authPubKey}
→ SMPClientHandshake {smpVersion, keyHash, authPubKey=Nothing, proxyServer=False, clientService=Nothing}
5. Send LGET
→ [empty auth][corrId][linkId]["LGET"]
6. Receive LNK
← [auth][corrId][linkId]["LNK" space senderId encFixedData encUserData]
7. Decrypt
XSalsa20-Poly1305 with sbKey
→ FixedLinkData, ConnLinkData (with profile JSON)
8. Display profile
```
Note: spike sends `authPubKey=Nothing` so block encryption is not used (blocks are padded only). Block encryption is added in steps 12-13.
## Build Approach
Bottom-up, function-by-function. Each TypeScript function tested against its Haskell counterpart via `callNode` — the same pattern used in xftp-web (see `XFTPWebTests.hs`).
**Project location**: `simplexmq-2/smp-web/`
**Tests**: `simplexmq-2/tests/SMPWebTests.hs` — reuses `callNode`/`jsOut`/`jsUint8` from XFTPWebTests (generalized, not copied)
**xftp-web**: npm dependency via `file:../xftp-web` (encoding, crypto, padding imported directly). Note: libsodium-wrappers-sumo is xftp-web's dependency; tests must init the same sodium instance that xftp-web's secretbox uses. If xftp-web is ever published to npm, libsodium should become a peerDependency.
**File structure**: mirrors Haskell module hierarchy (see RFC section 2)
**Pattern for each function**:
1. Check if xftp-web already implements it (or something close). If so, import and reuse — export from xftp-web if not yet exported. Only write new code when no existing implementation covers the need.
2. Implement in TypeScript, in the file corresponding to its Haskell module
3. Write Haskell test that calls it via `callNode`
4. Compare output byte-for-byte with Haskell reference
5. Cross-language: Haskell encodes → TypeScript decodes, and vice versa
### Parsing Approach
All binary parsing uses xftp-web's `Decoder` class — the same class, not a copy. `Decoder` tracks position over a `Uint8Array`, throws on malformed input, returns subarray views (zero-copy).
SMP command parsing follows the same pattern as xftp-web's `decodeResponse` in `commands.ts`: `readTag` reads bytes until space or end, switch dispatches on the tag string, fields are parsed sequentially with `Decoder` methods (`decodeBytes`, `decodeLarge`, `decodeBool`, etc.).
**Prerequisite xftp-web change**: `readTag` and `readSpace` in xftp-web's `commands.ts` need to be exported so smp-web can import them.
### WebSocket Transport Approach
WebSocket transport follows the simplexmq-js `WSTransport` pattern:
- `WebSocket` connects to `wss://` URL with `binaryType = 'arraybuffer'`
- `onmessage` enqueues received frames into an `ABQueue` (async bounded queue with backpressure)
- `onclose` closes the queue (sentinel-based)
- `readBlock()` dequeues one frame, validates it is exactly 16384 bytes
- `sendBlock(data)` sends one 16384-byte binary frame
The `ABQueue` class from simplexmq-js provides backpressure via semaphores and clean async iteration. It can be included in smp-web or extracted as a shared utility.
The SMP transport layer wraps WebSocket transport:
- Receives raw blocks → unpad → parse transmission
- Encodes transmission → pad → send as block
- After handshake, if block encryption is active: decrypt before unpad, encrypt after pad
## Encoding Reference
Binary encoding rules (from `Simplex.Messaging.Encoding`):
| Type | Format |
|------|--------|
| `Word16` | 2 bytes big-endian |
| `Word32` | 4 bytes big-endian |
| `ByteString` | 1-byte length + bytes (max 255) |
| `Large` | 2-byte length (BE) + bytes (max 65535) |
| `Bool` | 'T' (0x54) or 'F' (0x46) |
| `Maybe a` | '0' (0x30) for Nothing, '1' (0x31) + value for Just |
| `smpEncodeList` | 1-byte count + items |
| `UserLinkData` | ByteString if ≤254 bytes, else 0xFF + Large |
**Critical**: `encodeAuthEncryptCmds Nothing` = empty (0 bytes), NOT 'F' or '0'.
**Transmission format** (binary, NOT ASCII with spaces):
```
[auth ByteString][corrId ByteString][entityId ByteString][command bytes]
```
For v7+ (`implySessId = True`): sessionId is NOT sent on wire, but is prepended to the `authorized` data for signature verification. For unauthenticated commands (LGET), this doesn't apply.
**Block framing**: `pad(transmission, 16384)` = `[2-byte BE length][message][padding with '#' (0x23)]`
## Server Changes — DONE
WebSocket support on the same port as native TLS is implemented and tested.
- `attachStaticAndWS` — unified HTTP + WebSocket handler via `wai-websockets`
- SNI routing: browser (SNI) → Warp → WebSocket upgrade → SMP over WS
- `acceptWSConnection` — constructs `WS 'TServer` from `TLS 'TServer` + PendingConnection
- Test: `testWebSocketAndTLS` in `ServerTests.hs`
**Remaining**: CORS headers for cross-origin widget embedding.
## Implementation Steps
Each step produces working, tested code. Steps 1-11 work without block encryption. Steps 12-13 add it.
### Step 1: Project Setup + xftp-web Changes
**smp-web setup**:
- Create `smp-web/` with `package.json` (xftp-web + `@noble/hashes` as dependencies), `tsconfig.json` (ES2022, strict, same as xftp-web)
- Build: `tsc``dist/`
**xftp-web change**:
- Export `readTag` and `readSpace` from `commands.ts` (currently unexported) so smp-web can import them
**Test infrastructure**:
- Create `SMPWebTests.hs`, reusing `callNode`/`jsOut`/`jsUint8` from XFTPWebTests (generalize shared utilities into a common test module, not copy)
- First test: import `decodeBytes` from xftp-web, encode a ByteString, verify output matches Haskell `smpEncode`
### Step 2: SMP Transmission Encode/Decode
**File**: `protocol.ts`
**Haskell reference**: `Simplex.Messaging.Protocol``encodeTransmission_`, `transmissionP`
**Implementation**:
- `encodeTransmission(corrId, entityId, command)`: `concatBytes(encodeBytes(emptyAuth), encodeBytes(corrId), encodeBytes(entityId), command)` — unsigned, empty auth byte (0x00)
- `decodeTransmission(data)`: sequential Decoder — `decodeBytes` for auth, corrId, entityId, then `takeAll` for command bytes
- Pad/unpad: reuse xftp-web `blockPad`/`blockUnpad` (same 2-byte length prefix + '#' padding, same 16384 block size)
**Tests**: encode in TypeScript → decode in Haskell (`transmissionP`), encode in Haskell (`encodeTransmission_`) → decode in TypeScript. Byte-for-byte match.
### Step 3: SMP Handshake Parse/Encode
**File**: `transport.ts`
**Haskell reference**: `Simplex.Messaging.Transport``SMPServerHandshake`, `SMPClientHandshake`
**Implementation**:
- `parseSMPServerHandshake(d: Decoder)`: `decodeWord16` × 2 for versionRange, `decodeBytes` for sessionId. For authPubKey: if `maxVersion >= 7` and bytes remaining, parse `CertChainPubKey` (reuse xftp-web `identity.ts` for X.509 cert chain parsing and signature extraction). If no bytes remain, authPubKey is absent (encodeAuthEncryptCmds encoded Nothing as empty).
- `encodeSMPClientHandshake(...)`: `concatBytes(encodeWord16(version), encodeBytes(keyHash), authPubKeyBytes, encodeBool(proxyServer), encodeMaybe(encodeService, clientService))`. Where authPubKey: empty bytes for Nothing, `encodeBytes(pubkey)` for Just. proxyServer only for v14+, clientService only for v16+.
**Tests**: Haskell encodes `SMPServerHandshake` → TypeScript parses, all fields match. TypeScript encodes `SMPClientHandshake` → Haskell parses via `smpP`.
### Step 4: LGET Command Encode
**File**: `protocol.ts`
**Haskell reference**: `Simplex.Messaging.Protocol``LGET` command encoding
**Implementation**:
- `encodeLGET()`: returns `ascii("LGET")` — 4 bytes, no parameters. The LinkId is carried as entityId in the transmission (step 2), not in the command body.
- Full LGET block: `blockPad(encodeTransmission(corrId, linkId, encodeLGET()), 16384)`
**Tests**: encode full LGET block in TypeScript, Haskell unpad + `transmissionP` + `parseProtocol` decodes as `LGET` with correct corrId and linkId.
### Step 5: LNK Response Parse
**File**: `protocol.ts`
**Haskell reference**: `Simplex.Messaging.Protocol``LNK` response encoding (line 1834)
**Implementation**:
- `decodeResponse(d: Decoder)`: `readTag(d)` → switch dispatch (same pattern as xftp-web `decodeResponse`)
- For `"LNK"`: `readSpace(d)`, `decodeBytes(d)` for senderId, `decodeLarge(d)` for encFixedData, `decodeLarge(d)` for encUserData
- Also handle `"ERR"` responses for error reporting
**Tests**: Haskell encodes `LNK senderId (encFixed, encUser)` → TypeScript `decodeResponse` parses. All fields match byte-for-byte.
### Step 6: Short Link URI Parse
**File**: `agent/protocol.ts`
**Haskell reference**: `Simplex.Messaging.Agent.Protocol``ConnShortLink` StrEncoding instance (lines 1599-1612)
**Implementation**:
- `parseShortLink(uri)`: regex to extract scheme (https/simplex), type char (c/g/a), linkKey (base64url, 43 chars → 32 bytes), query params (h=hosts, p=port, c=keyHash)
- `base64UrlDecode(s)`: pad to multiple of 4, replace `-``+`, `_``/`, decode
- Returns `{scheme, connType, server: {hosts, port, keyHash}, linkKey}`
**Tests**: Haskell `strEncode` a `ConnShortLink` → TypeScript `connShortLinkStrP` parses. All fields match. Test multiple formats: with/without query params, different type chars.
**Done**. Function: `connShortLinkStrP` in `agent/protocol.ts`. Uses `base64urlDecode` from xftp-web `description.ts`.
**Future**:
- Add long link parsing (`ConnectionRequestUri`) and an either-parser that handles both short and long links.
- Add `restoreShortLink`: preset servers are shortened to host-only (`SMPServerOnlyHost` - no port, no keyHash). After parsing, `restoreShortLink` looks up the full server by hostname from a preset servers list. Without this, connections to preset servers will fail. See `Agent/Protocol.hs:1692`.
### Step 7: HKDF Key Derivation
**File**: `crypto/shortLink.ts`
**Haskell reference**: `Simplex.Messaging.Crypto.ShortLink``contactShortLinkKdf` (line 48)
**Implementation**:
- `contactShortLinkKdf(linkKey)`: `hkdf(sha512, linkKey, new Uint8Array(0), "SimpleXContactLink", 56)` using `@noble/hashes/hkdf` + `@noble/hashes/sha512`. Split result: first 24 bytes = linkId, remaining 32 bytes = sbKey.
**Note**: Haskell `C.hkdf` uses SHA-512, not SHA3-256.
**Tests**: given known linkKey bytes, TypeScript and Haskell produce identical linkId and sbKey.
### Step 8: Link Data Decrypt
**File**: `crypto/shortLink.ts`
**Haskell reference**: `Simplex.Messaging.Crypto.ShortLink``decryptLinkData` (lines 100-120)
**Implementation**:
- `decryptLinkData(sbKey, encFixedData, encUserData)`:
1. For each EncDataBytes: `Decoder``decodeBytes(d)` for nonce (24 bytes), `decodeTail(d)` for ciphertext (includes Poly1305 tag)
2. `cbDecrypt(sbKey, nonce, ciphertext)` via xftp-web `secretbox.ts`
3. From decrypted plaintext: `decodeBytes(d)` for signature (1-byte len 0x40 + 64 bytes), `decodeTail(d)` for actual data
4. Return both plaintext data blobs (signature verification skipped for spike)
**Tests**: Haskell `encodeSignLinkData` + `sbEncrypt` with known key/nonce → TypeScript decrypts → plaintext matches.
### Step 9: ConnLinkData Parse
**File**: `agent/protocol.ts`
**Haskell reference**: `Simplex.Messaging.Agent.Protocol``ConnLinkData`, `UserContactData`, `OwnerAuth`, `ConnShortLink`, `ProtocolServer` Encoding instances
**Implementation** (proper decoding, not skipping):
- `decodeConnLinkData(d)`: `anyByte` for connectionMode ('C'=Contact), `decodeWord16` × 2 for agentVRange, then `decodeUserContactData`
- `decodeUserContactData(d)`: `decodeBool` for direct, `smpListP(decodeOwnerAuth, d)` for owners, `smpListP(decodeConnShortLink, d)` for relays, `decodeUserLinkData(d)` for userData
- `decodeOwnerAuth(d)`: `decodeBytes` for outer wrapper, then parse inner: `(ownerId, ownerKey, authOwnerSig)` all as ByteStrings
- `decodeConnShortLink(d)`: `anyByte` for mode, then Contact: `(ctTypeChar, srv, linkKey)` or Invitation: `(srv, linkId, linkKey)`
- `decodeProtocolServer(d)`: `decodeBytes` for scheme+keyHash, `decodeBytes` for host, `decodeBytes` for port — need to verify exact encoding
- `decodeUserLinkData(d)`: first byte 0xFF → `decodeLarge`; otherwise it's the 1-byte length of a ByteString
- `parseProfile(userData)`: check first byte for 'X' (0x58, zstd compressed) — if so, decompress; otherwise `JSON.parse` directly
**Tests**: Haskell encodes `ContactLinkData` with known values → TypeScript decodes → all fields match.
**FixedLinkData**: deferred to step 15. `linkConnReq` (ConnectionRequestUri) is NOT length-prefixed in the tuple encoding — it requires full parsing. FixedLinkData is also needed to validate mutable data signature using rootKey.
### Step 15: FixedLinkData Parse
**File**: `agent/protocol.ts`
**Haskell reference**: `Simplex.Messaging.Agent.Protocol``FixedLinkData`, `ConnectionRequestUri`, `ConnReqUriData` Encoding instances
**Implementation**:
- `decodeFixedLinkData(d)`: `decodeWord16` × 2 for agentVRange, `decodeBytes` for rootKey (32 bytes Ed25519), then parse `ConnectionRequestUri` (mode byte + `ConnReqUriData`), optional `decodeBytes` for linkEntityId
- `decodeConnectionRequestUri(d)`: full parsing of `ConnReqUriData` including SMP queue URIs
- Needed for: connecting to the contact, and validating mutable data signature with rootKey
**Tests**: Haskell encodes full `FixedLinkData` → TypeScript decodes → rootKey and linkConnReq fields match.
### Step 10: WebSocket Transport
**File**: `transport/websockets.ts`
**Pattern reference**: simplexmq-js `WSTransport` + `ABQueue`
**Implementation**:
- `ABQueue<T>` class: semaphore-based async bounded queue (from simplexmq-js `queue.ts` — reimplement or include as utility). `enqueue`/`dequeue`/`close`, sentinel-based close, async iterator.
- `connectWS(url)`: `new WebSocket(url)`, `binaryType = 'arraybuffer'`, `onmessage` enqueues `Uint8Array` frames into ABQueue, `onclose` closes queue, `onerror` closes socket. Returns transport handle on `onopen`.
- `readBlock(transport)`: dequeue one frame, verify `byteLength === 16384`, return `Uint8Array`
- `sendBlock(transport, data)`: `ws.send(data)`, verify `data.length === 16384`
- `smpHandshake(transport, keyHash)`: `readBlock``blockUnpad``parseSMPServerHandshake` → negotiate version → `encodeSMPClientHandshake``blockPad``sendBlock`. Returns `{sessionId, version}`.
**Integration test**: spawn test SMP server with web credentials (reuse `cfgWebOn` from SMPClient.hs), connect via WebSocket from Node.js, complete handshake, verify sessionId received.
### Step 11: End-to-End Integration
Wire steps 6-10 together: `parseShortLink``contactShortLinkKdf``connectWS``smpHandshake` → encode LGET block → `sendBlock``readBlock``blockUnpad``decodeTransmission``decodeResponse``decryptLinkData``decodeFixedLinkData` + `decodeConnLinkData``parseProfile`.
**Test**: Haskell creates a contact address with short link (using agent), TypeScript fetches and decodes it via WebSocket. Profile displayName matches. This is the full spike proof: browser can fetch a SimpleX contact profile via SMP protocol.
### Step 12: Server Certificate Verification
**File**: `transport.ts`
**Approach**: client sends a random challenge in an HTTP header on the WebSocket upgrade request. Server includes the signed challenge in the handshake response. Client verifies the signature using the server's certificate chain.
**Implementation**:
- Generate 32-byte random challenge, send as HTTP header (e.g. `smp-web-challenge`) on WebSocket upgrade
- Parse `CertChainPubKey` from server handshake (already parsed in step 3 as `authPubKey`)
- Verify certificate chain fingerprint matches `keyHash` (reuse xftp-web `caFingerprint`)
- Verify challenge signature (reuse xftp-web `identity.ts``extractCertPublicKeyInfo`, signature verification)
- Requires server-side change: detect the challenge header on WebSocket connections, sign `challenge || sessionId` with server key, include proof in handshake
**Tests**: connect to test server, verify challenge-response succeeds. Connect with wrong keyHash, verify rejection.
### Step 13: Block Encryption (DH + SbChainKeys)
**File**: `transport.ts`
**Haskell reference**: `Simplex.Messaging.Crypto``sbcInit`, `sbcHkdf`; `Simplex.Messaging.Transport``tPutBlock`, `tGetBlock`
**Implementation**:
- `generateX25519KeyPair()`, `dh(peerPub, ownPriv)` — reuse from xftp-web `keys.ts`
- `sbcInit(sessionId, dhSecret)`: `hkdf(sha512, dhSecret, sessionId, "SimpleXSbChainInit", 64)` → split at 32: `(sndChainKey, rcvChainKey)`. Note client swaps send/receive keys vs server (line 858 Transport.hs).
- `sbcHkdf(chainKey)`: `hkdf(sha512, chainKey, "", "SimpleXSbChain", 88)` → split: 32 bytes new chainKey, 32 bytes sbKey, 24 bytes nonce. Returns `{sbKey, nonce, nextChainKey}`.
- `encryptBlock(state, block)`: `sbcHkdf``cryptoBox(sbKey, nonce, pad(block, blockSize - 16))` → 16-byte tag + ciphertext
- `decryptBlock(state, block)`: `sbcHkdf` → split tag (first 16 bytes) + ciphertext → `cryptoBoxOpen``unpad`
**Tests**: Haskell and TypeScript DH with same keys → identical chain keys. Haskell encrypts block → TypeScript decrypts (and vice versa). Chain key advances identically after each block.
### Step 14: Full Handshake with Auth
**File**: `transport.ts`
**Haskell reference**: `Simplex.Messaging.Transport``smpClientHandshake` (lines 792-842)
**Implementation**:
- Update `smpHandshake` to generate ephemeral X25519 keypair and include public key in `encodeSMPClientHandshake` as authPubKey
- Compute DH: `dh(serverDhPub, clientPrivKey)` → shared secret
- `sbcInit(sessionId, dhSecret)` → chain keys (with client-side swap)
- All subsequent `readBlock`/`sendBlock` go through `decryptBlock`/`encryptBlock`
**Tests**: full handshake with real server, block encryption active, exchange encrypted commands. Haskell sends encrypted response → TypeScript decrypts correctly.
## Haskell Code References
### Handshake
- `Simplex.Messaging.Transport``smpClientHandshake`, `smpServerHandshake`, `SMPServerHandshake`, `SMPClientHandshake`
- `encodeAuthEncryptCmds` — Nothing → empty, Just → raw smpEncode
### Protocol
- `Simplex.Messaging.Protocol``LGET`, `LNK`, `encodeTransmission_`, `transmissionP`
- Block: `pad`/`unPad` in `Simplex.Messaging.Crypto`
### Short Links
- `Simplex.Messaging.Crypto.ShortLink``contactShortLinkKdf`, `decryptLinkData`
- `Simplex.Messaging.Agent.Protocol``ConnShortLink`, `FixedLinkData`, `ConnLinkData`, `UserLinkData`
### Block Encryption
- `Simplex.Messaging.Crypto``sbcInit`, `sbcHkdf`, `sbEncrypt`, `sbDecrypt`, `dh'`
- `Simplex.Messaging.Transport``blockEncryption`, `TSbChainKeys`, `tPutBlock`, `tGetBlock`
@@ -0,0 +1,355 @@
# SMP Client for Browser
**Parent**: [SMP Agent Web Spike](./2026-03-20-smp-agent-web-spike.md)
**Depends on**: Spike 1 (merged) — transport, ratchet, encoding, per-queue E2E
## Context
The encoding spike proved all four encryption layers work cross-language. The next implementable and testable piece is the SMP client — the layer that sends commands, correlates responses by CorrId, authenticates with entity keys, and exposes typed async functions.
Faithful transpilation of `Simplex.Messaging.Client` (Client.hs). Transport is WebSocket (already working), protocol logic is identical to Haskell.
## Encoding path (per command)
Traced from `sendSMPMessage` through every function call:
```
1. encodeTransmission_(v, (corrId, entityId, command))
→ smpEncode(corrId, entityId) <> encodeProtocol(v, cmd)
Already have as encodeTransmission() in protocol.ts — update in place
2. encodeTransmissionForAuth(thParams, transmission)
→ tForAuth = smpEncode(sessionId) <> encodeTransmission_(...)
→ tToSend = encodeTransmission_(...) [when implySessId=true, which is always true for v>=7]
Note: implySessId means tToSend omits sessionId, but tForAuth includes it (for signing)
3. authTransmission(thAuth, serviceAuth=false, maybePrivKey, nonce, tForAuth)
→ thAuth contains serverPubKey (X25519) from handshake
→ maybePrivKey is Nothing for unauthenticated commands (LGET, SEND without key)
→ Nothing privKey: no auth, encode empty ByteString
→ Just X25519 privKey: TAAuthenticator(cbAuthenticate(serverPubKey, privKey, nonce, tForAuth))
→ Just Ed25519 privKey: TASignature(sign(privKey, tForAuth))
Note: nonce IS the CorrId (same 24 bytes used for both)
Note: serviceAuth is always false for browser client (no service certificates)
4. tEncodeAuth(serviceAuth=false, maybeAuth)
→ Nothing: smpEncode("") [1-byte 0x00]
→ Just (TAAuthenticator s, _): smpEncode(s) [1-byte len + 80 bytes]
→ Just (TASignature sig, _): smpEncode(signatureBytes sig) [1-byte len + 64 bytes]
Note: TAuthorizations = (TransmissionAuth, Maybe serviceSig) — serviceSig always Nothing for us
5. tEncode(serviceAuth, (auth, tToSend))
→ tEncodeAuth(auth) <> tToSend
6. tEncodeBatch1(serviceAuth, sentRawTransmission)
→ lenEncode(1) + smpEncode(Large(tEncode(...)))
Single-command batch. Always used when batch=true (v7+).
7. batchTransmissions_(blockSize, transmissions)
→ Pack multiple Large-wrapped transmissions into ≤blockSize blocks
→ Count byte prefix, up to 255 per block
→ blockSize' = blockSize - 19 (2 pad + 1 count + 16 auth tag)
```
## Parsing path (per received block)
```
1. tParse(thParams, blockBytes)
→ batch=true: parse count byte, then N Large-wrapped transmissions
→ Each: transmissionP(thParams) parses:
- authenticator (ByteString, 1-byte len + data) — ignored by client
- rest = authorized bytes
- re-parse authorized: corrId (ByteString) + entityId (ByteString) + command (rest)
- if implySessId=true: sessionId not in wire format, prepended from thParams for verification
→ Returns RawTransmission{authenticator, corrId, entityId, command}
2. tDecodeClient(thParams, rawTransmission)
→ Verify sessId matches (skipped when implySessId=true)
→ parseProtocol(v, command) → Either ErrorType BrokerMsg
→ Return (corrId, entityId, Right msg | Left err)
3. clientResp classification (Client.hs:708-712):
→ Left err (parse error) → PCEResponseError
→ Right msg, protocolError msg = Just err → PCEProtocolError (ERR response)
→ Right msg, protocolError msg = Nothing → Right msg (success)
4. Process: lookup corrId in pendingCommands
→ Found: resolve Promise with clientResp
→ Not found (empty corrId = server push): deliver to event callback
```
## Functions to implement
### Crypto (`src/crypto.ts` — extend)
| Function | Haskell | Implementation |
|---|---|---|
| `sha512Hash(msg)` | `Crypto.hs:1016` | `sha512(msg)` from `@noble/hashes/sha512` |
| `cbAuthenticator(serverPubKey, entityPrivKey, nonce, msg)` | `Crypto.hs:1367` | `cryptoBox(dh(serverPubKey, privKey), nonce, sha512Hash(msg))` → 80 bytes (16 tag + 64 hash) |
`cryptoBox` and `dh` already available from xftp-web. `sha512` from `@noble/hashes`.
Not needed in spike: `cbDecryptNoPad` (only used by `cbVerify` and proxy commands).
Ed25519 signing: `crypto_sign_detached` from libsodium (already loaded and initialized via xftp-web for secretbox — no second implementation needed).
Not needed: `cbVerify` (server-side only).
### Transport update (`src/transport/websockets.ts` — update)
**Gap: `connectSMP` must return `serverPubKey`** (raw X25519 public key bytes from the handshake). Currently it computes the DH secret and derives block keys, but discards the server's raw public key. The client needs it for `cbAuthenticate` on every command.
Update `SMPConnection` to include:
```typescript
interface SMPConnection {
ws: WebSocket
sessionId: Uint8Array
smpVersion: number
sndKey: Uint8Array | null
rcvKey: Uint8Array | null
serverPubKey: Uint8Array | null // raw X25519 public key — needed for command auth
}
```
### Protocol encoding (`src/protocol.ts` — update existing)
Update `encodeTransmission`, `encodeBatch`, `decodeTransmission` in place — these were spike throwaway. Replace with auth-aware versions and update existing tests accordingly.
| Function | Haskell ref | Notes |
|---|---|---|
| `encodeTransmission_(v, corrId, entityId, command)` | `Protocol.hs:2194` | Update existing `encodeTransmission`. Also fix `encodeNEW`: QueueReqData should be `Just (QRMessaging Nothing)` not `Nothing`, and rename `sndAuthKey` param to `basicAuth` (it's server auth, not a crypto key) |
| `encodeTransmissionForAuth(sessionId, corrId, entityId, command)` | `Protocol.hs:2186` | Returns `{tForAuth, tToSend}`. `implySessId` always true for v>=7 |
| `authTransmission(serverPubKey, maybePrivKey, nonce, tForAuth)` | `Client.hs:1372` | `maybePrivKey` is `{type: "x25519"|"ed25519", key} | null`. Null for unauthenticated commands. X25519 → cbAuthenticator. Ed25519 → sign. |
| `tEncodeAuth(auth)` | `Protocol.hs:507` | Handles null, authenticator (80 bytes), signature (64 bytes) |
| `tEncode(auth, tToSend)` | `Protocol.hs:2171` | `tEncodeAuth(auth) + tToSend` |
| `tEncodeBatch1(auth, tToSend)` | `Protocol.hs:2179` | `[count=1] + Large(tEncode(...))` |
| `tEncodeForBatch(auth, tToSend)` | `Protocol.hs:2175` | `Large(tEncode(...))` |
| `batchTransmissions(blockSize, transmissions)` | `Protocol.hs:2151` | Pack into ≤(blockSize-19)-byte blocks, count prefix |
| `transmissionP(sessionId, block)` | `Protocol.hs:1629` | Skip auth bytes (1-byte len + data), parse corrId + entityId + command from rest. `implySessId`=true (sessionId not in wire, no need to verify on client side), `serviceAuth`=false (no serviceSig to skip) |
| `tParse(sessionId, block)` | `Protocol.hs:2211` | Parse count, N×Large, each through `transmissionP` |
| `tDecodeClient(sessionId, version, rawTransmission)` | `Protocol.hs:2256` | Parse command bytes → typed BrokerMsg |
| `encodePING()` | | PING command for keepalive |
Update `decodeResponse`:
- Add `SOK` (subscribe response with optional serviceId, returned by SUB in v19)
- Add `INFO` (queue info response, for `getSMPQueueInfo`)
- Improve `ERR` parsing: currently reads just the tag string. Need to parse structured `ErrorType` (at minimum AUTH, QUOTA, NO_MSG, INTERNAL) for proper error handling in the client
### Client (`src/client.ts` — new)
```typescript
interface SMPClient {
sessionId: Uint8Array
smpVersion: number
serverPubKey: Uint8Array // for cbAuthenticate
// Core: send pre-encoded command, correlate response
// Lower-level than Haskell's sendProtocolCommand — takes pre-encoded command bytes
// privKey: {type: "x25519", key} | {type: "ed25519", key} | null
// Rejects with PCEProtocolError (ERR response), PCEResponseError (parse fail), PCEResponseTimeout
sendCommand(privKey: AuthKey | null, entityId: Uint8Array, command: Uint8Array): Promise<BrokerMsg>
// High-level commands (keys are DER-encoded unless noted)
// authKeyPair: {publicKey, privateKey, type: "x25519"} — public goes in NEW encoding, private for auth
createQueue(authKeyPair, dhKey, subMode): Promise<QueueIdsKeys>
subscribeQueue(privKey, rcvId): Promise<void> // SUB can return MSG (queued message) → pushed to onMessage
sendMessage(privKey, sndId, flags, msg): Promise<void> // privKey can be null (before queue secured)
ackMessage(privKey, rcvId, msgId): Promise<void> // ACK can return MSG → pushed to onMessage
secureQueue(privKey, rcvId, senderKey): Promise<void>
secureSndQueue(privKey, sndId): Promise<void>
getQueueLink(linkId): Promise<{senderId, linkData}>
getQueueInfo(privKey, queueId): Promise<QueueInfo>
deleteQueue(privKey, rcvId): Promise<void>
suspendQueue(privKey, rcvId): Promise<void>
close(): void
}
function createSMPClient(
url: string,
keyHash: Uint8Array,
onMessage: (entityId: Uint8Array, msg: BrokerMsg) => void,
onDisconnected: () => void,
wsOptions?: object,
): Promise<SMPClient>
```
Internally:
- `connectSMP` for WebSocket + handshake (existing, updated to return serverPubKey)
- `Map<string, {resolve, reject}>` for hex(corrId) → Promise correlation
- WebSocket `onmessage`: `receiveEncryptedBlock``tParse` → for each transmission: `tDecodeClient` → classify via `protocolError` → correlate by corrId or push to `onMessage`
- `sendCommand`: generate random 24-byte corrId/nonce → `encodeTransmissionForAuth``authTransmission``tEncodeBatch1``sendEncryptedBlock` → return Promise resolved by correlator
- `setInterval` ping: send PING, count timeouts, close after N consecutive
- Timeout per command: `setTimeout` on pending Promise, reject with PCEResponseTimeout
**Message delivery model:**
All MSGs reach `onMessage` regardless of how they arrive. Three sources:
1. **Server push** (empty corrId): receive handler calls `onMessage` directly
2. **SUB response**: `sendCommand` resolves with MSG → `subscribeQueue` pushes to `onMessage`, returns success to caller
3. **ACK response**: same — `ackMessage` pushes to `onMessage`, returns success
High-level functions never expose MSG to their callers. This mirrors Haskell's `processSUBResponse_` (Client.hs:858-862) and `ackSMPMessage` (Client.hs:1042-1044) which both call `writeSMPMessage` to forward MSGs to msgQ and return OK-equivalent.
### Client REPL (`smp-web/tests/client-repl.ts` — new, separate from ratchet-repl.ts)
Separate REPL process holding a WebSocket connection + SMP client state. Same stdin/stdout line protocol approach, different state and commands.
**Message queue:** The REPL maintains an internal `Message[]` queue. The SMPClient's `onMessage` callback pushes to this queue. MSGs arrive here from three sources: server pushes (no corrId), SUB responses, and ACK responses — all handled identically by the client internals. The `RECV` command dequeues from this queue (or waits with timeout).
**Concurrency:** Unlike the ratchet REPL (pure, no network), the client REPL receives messages concurrently with stdin. This works because Node's event loop handles WebSocket `onmessage` events between readline callbacks — no explicit threading needed.
```
CONNECT <url> <keyHashHex> [wsOptions]
→ Creates SMPClient, returns "ok"
NEW <rcvAuthKeyHex> <rcvDhKeyHex> [subMode]
→ createQueue (defaults: no basic auth, SMSubscribe, QRMessaging, no ntf creds)
→ returns "ok: <rcvIdHex> <sndIdHex> <srvDhKeyHex>"
SUB <rcvIdHex> <rcvPrivKeyHex>
→ subscribeQueue, returns "ok"
SEND <sndIdHex> <sndPrivKeyHex> <bodyHex>
→ sendMessage, returns "ok"
ACK <rcvIdHex> <rcvPrivKeyHex> <msgIdHex>
→ ackMessage, returns "ok"
KEY <rcvIdHex> <rcvPrivKeyHex> <senderKeyHex>
→ secureQueue, returns "ok"
SKEY <sndIdHex> <sndPrivKeyHex>
→ secureSndQueue, returns "ok"
LGET <linkIdHex>
→ getQueueLink, returns "ok: <senderIdHex> <linkDataHex>"
RECV [timeoutMs]
→ Dequeue next server-pushed MSG, returns "ok: <entityIdHex> <msgIdHex> <bodyHex>"
→ Times out with "error: timeout" if no message arrives
```
### Polymorphic testing
Same pattern as ratchet tests: `TestPeer` sum type with `TestPeerHS` / `TestPeerJS` dispatch. For SMP client tests:
```haskell
data TestSMPClient
= TestClientHS SMPClient
| TestClientJS Handle Handle ProcessHandle -- stdin, stdout, process
-- Dispatch functions
tcCreateQueue :: TestSMPClient -> ... -> IO QueueIdsKeys
tcSubscribe :: TestSMPClient -> ... -> IO ()
tcSendMessage :: TestSMPClient -> ... -> IO ()
tcReceiveMessage :: TestSMPClient -> IO (EntityId, MsgId, ByteString)
tcSecureQueue :: TestSMPClient -> ... -> IO ()
tcAckMessage :: TestSMPClient -> ... -> IO ()
```
Then the same test function runs against HS↔HS, HS↔JS, JS↔HS, JS↔JS peer combinations. The test creates two clients (one receiver, one sender) on the same SMP server, creates a queue, exchanges keys, sends messages — proving protocol compatibility.
## Tests
### Unit tests (callNode, no server)
1. `sha512Hash` — same input → same output as Haskell
2. `cbAuthenticator` — same serverPubKey + entityPrivKey + nonce + message → same 80 bytes as Haskell
3. `encodeTransmissionForAuth` — same sessionId + corrId + entityId + command (encoded at v19) → same `{tForAuth, tToSend}` as Haskell
4. `authTransmission` with X25519 key — same keys + nonce + tForAuth → same authenticated bytes as Haskell
5. `authTransmission` with Ed25519 key — same key + tForAuth → same signature bytes as Haskell
6. `authTransmission` with no key (Nothing) — produces empty auth, matches Haskell
7. `tEncodeBatch1` — same auth + transmission → same block bytes as Haskell
8. `tParse` + `tDecodeClient` — TS parses Haskell-encoded response block, extracts corrId + entityId + typed response
9. `batchTransmissions` — given N transmissions, produces same batch boundaries and block bytes as Haskell
### Integration tests (with SMP server, using REPL)
10. JS client connects, sends PING, receives PONG
11. JS client creates queue (NEW → IDS)
12. JS receiver creates queue + subscribes, JS sender sends message, receiver gets MSG
13. Full handshake: create queue → secure (KEY) → subscribe → send → receive MSG → ack
### Polymorphic integration tests
14. Same test function, peer combinations:
- HS sender, JS receiver
- JS sender, HS receiver
- JS sender, JS receiver
## Implementation order
1. Transport update — `connectSMP` returns `serverPubKey`
2. Crypto additions — `sha512Hash`, `cbAuthenticator`, Ed25519 `sign`
3. Protocol encoding updates — `encodeTransmissionForAuth`, `authTransmission`, `tEncode`, `tEncodeBatch1`, `batchTransmissions`, `encodePING`
4. Protocol parsing updates — `transmissionP`, `tParse`, `tDecodeClient`, update `decodeResponse` (add `SOK`, `INFO`, structured `ERR`)
5. Unit tests for steps 1-4
6. Client core — `createSMPClient`, `sendCommand`, corrId correlation, receive dispatch, ping
7. High-level command functions
8. Client REPL
9. Integration tests with server
10. Polymorphic test wiring
## Files
| File | Action |
|---|---|
| `smp-web/src/transport/websockets.ts` | Update `connectSMP` to return `serverPubKey` |
| `smp-web/src/crypto.ts` | Add `sha512Hash`, `cbAuthenticator`, Ed25519 `sign` |
| `smp-web/src/protocol.ts` | Update transmission encoding/parsing, add auth, batching |
| `smp-web/src/client.ts` | New — SMP client |
| `smp-web/tests/client-repl.ts` | New — SMP client REPL for integration tests |
| `tests/SMPWebTests.hs` | Unit + integration tests |
## Scope
### Client spike (this plan)
Core client: connect, auth, send/receive, correlate, ping. High-level commands: NEW, SUB, KEY, SKEY, SEND, ACK, OFF, DEL, LGET, GET, QUE (getSMPQueueInfo). Single-command path. Tests against real server.
### Client MVP (next, after spike)
- Proxy commands (PRXY, PFWD, PRES) — essential for privacy, users must not connect directly to untrusted servers
- Batch subscribe (subscribeSMPQueues) — needed for groups
- `reverseNonce` — needed for proxy
- Batch delete (deleteSMPQueues)
### Post-MVP
| What | Why |
|---|---|
| Notification commands (NKEY, NDEL, NSUB) | Value only with webpush support |
| Service certificates (serviceAuth, serviceSig) | Browser doesn't use |
| Stream commands (streamSubscribeSMPQueues) | Not used in Haskell client either |
| NetworkConfig, SOCKS, host mode, transport selection | Browser connects via WebSocket directly |
| Queue link management (LSET, LDEL, LKEY) | Only needed to create links, not join them |
| `cbVerify` | Server-side only |
## Haskell references
- `Client.hs:179-200` — ProtocolClient, PClient types
- `Client.hs:248``type SMPClient = ProtocolClient SMPVersion ErrorType BrokerMsg`
- `Client.hs:506-512` — Request type
- `Client.hs:628-642` — client connection, handshake, raceAny_ [send, process, receive, monitor]
- `Client.hs:644-658` — send loop, receive loop
- `Client.hs:660-678` — monitor/ping loop
- `Client.hs:680-719` — process loop, processMsg (corrId correlation, clientResp classification)
- `Client.hs:810-828` — createSMPQueue
- `Client.hs:833-836` — subscribeSMPQueue
- `Client.hs:938-939` — secureSMPQueue
- `Client.hs:1027-1031` — sendSMPMessage
- `Client.hs:1040-1045` — ackSMPMessage (note: ACK can return MSG)
- `Client.hs:1239-1243` — okSMPCommand pattern
- `Client.hs:1300-1326` — sendProtocolCommand_, sendRecv, size check, tEncodeBatch1
- `Client.hs:1333-1344` — getResponse, timeout handling
- `Client.hs:1349-1370` — mkTransmission_, CorrId=nonce, encodeTransmissionForAuth, authTransmission
- `Client.hs:1372-1391` — authTransmission, authenticate (X25519 vs Ed25519), service sig
- `Protocol.hs:488-525` — RawTransmission, TransmissionAuth, TAuthorizations, tEncodeAuth
- `Protocol.hs:1629-1643` — transmissionP
- `Protocol.hs:2129-2198` — batching, tEncode, tEncodeBatch1, batchTransmissions_
- `Protocol.hs:2207-2267` — tGetClient, tParse, tDecodeClient
- `Crypto.hs:1016` — sha512Hash
- `Crypto.hs:1296-1298` — cbEncryptNoPad (= cryptoBox without padding)
- `Crypto.hs:1330-1331` — cbDecryptNoPad
- `Crypto.hs:1366-1371` — cbAuthenticate, cbVerify
@@ -0,0 +1,238 @@
# SMP Client MVP: Proxy + Batching — Transpilation Plan
**Parent**: [SMP Client Spike](./2026-05-17-smp-client.md)
## Rule
Every TypeScript function is a faithful transpilation of a specific Haskell function at specific lines. Same name, same steps, same call chain. No inferences, no approximations. Each entry below gives the exact source to transpile from.
## Crypto functions
### `reverseNonce` → transpile `Crypto.hs:1409-1410`
```haskell
reverseNonce (CryptoBoxNonce s) = CryptoBoxNonce (B.reverse s)
```
TS: `function reverseNonce(nonce: Uint8Array): Uint8Array` — reverse the 24 bytes.
### `cbDecryptNoPad` → transpile `Crypto.hs:1330-1331`
```haskell
cbDecryptNoPad (DhSecretX25519 secret) = sbDecryptNoPad_ secret
```
Which is `sbDecryptNoPad_` from secretbox. xftp-web's `cbDecrypt` does decrypt+unpad. Need decrypt without unpad — extract tag(16) + cipher, decrypt, verify tag, return raw (no unpad). Use xftp-web's `sbInit`/`sbDecryptChunk`/`sbAuth` directly.
## Protocol encoding functions
### `encodeProtocolServer` → transpile `Protocol.hs:1264-1266`
```haskell
smpEncode ProtocolServer {host, port, keyHash} = smpEncode (host, port, keyHash)
```
Where:
- `host :: NonEmpty TransportHost``smpEncodeList` (1-byte count + items)
- Each `TransportHost``smpEncode (strEncode host)``encodeBytes(ascii(hostname))` (`Transport/Client.hs:77-78`)
- `port :: ServiceName` = ByteString → `encodeBytes(port)`
- `keyHash :: KeyHash` = ByteString → `encodeBytes(keyHash)`
File: `src/protocol.ts`
### `encodePRXY` → transpile `Protocol.hs:1710`
```haskell
PRXY host auth_ -> e (PRXY_, ' ', host, auth_)
```
= `"PRXY " + smpEncode(server) + smpEncode(Maybe BasicAuth)`
Where `Maybe BasicAuth` = `encodeMaybe(encodeBytes, auth)`.
### `encodePFWD` → transpile `Protocol.hs:1711`
```haskell
PFWD fwdV pubKey (EncTransmission s) -> e (PFWD_, ' ', fwdV, pubKey, Tail s)
```
= `"PFWD " + encodeWord16(version) + encodeBytes(pubKeyDer) + encTransmission` (Tail = no length prefix)
### `decodePKEY` → transpile `Protocol.hs:1894`
```haskell
PKEY_ -> PKEY <$> _smpP <*> smpP <*> smpP
```
= space + `decodeBytes(d)` (sessionId) + `decodeVersionRange(d)` + `decodeCertChainPubKey(d)`
`VersionRange` encoding (`Version.hs`): `smpEncode (minVersion, maxVersion)` = two Word16.
`CertChainPubKey` encoding (`Transport.hs:663-667`): `smpEncode (encodeCertChain chain, SignedObject signedPubKey)``encodeCertChain` is `Large`-encoded DER bytes, `SignedObject` is `Large`-encoded DER bytes.
### `decodePRES` → transpile `Protocol.hs:1896`
```haskell
PRES_ -> PRES <$> (EncResponse . unTail <$> _smpP)
```
= space + rest of bytes (Tail) → `EncResponse`
### Add to `decodeResponse`: `PKEY` and `PRES` cases.
## Client functions
### `sendProtocolCommands` → transpile `Client.hs:1262-1278`
Call chain:
1. `mapM (mkTransmission c) cs` — for each command: generate corrId, encode, auth, register pending request
2. `batchTransmissions' thParams` — pack into blocks
3. `mapM (sendBatch c nm) bs` — send each block, collect responses
4. `validate` — verify response count matches command count
In TS: `mkTransmission` = the existing `sendCommand` logic (corrId generation, `encodeTransmissionForAuth`, `authTransmission`) but separated into encode+register vs send+await. Need to refactor `sendCommand` to split these.
### `batchTransmissions'` → transpile `Protocol.hs:2135-2148`
Already have `batchTransmissions` in protocol.ts that does `batchTransmissions_`. Need `batchTransmissions'` which wraps with `tEncodeForBatch` before batching. Currently the TS `batchTransmissions` takes pre-encoded Large-wrapped bytes. Need to match the Haskell call chain exactly:
```haskell
batchTransmissions' params ts
| batch = batchTransmissions_ bSize $ L.map (first $ fmap $ tEncodeForBatch serviceAuth) ts
```
### `sendBatch` → transpile `Client.hs:1285-1298`
Three cases:
- `TBError`: return error response
- `TBTransmissions s n rs`: send block `s`, await all `n` responses concurrently
- `TBTransmission s r`: send block `s`, await one response
In browser: "concurrently" = all promises pending simultaneously, resolved by `onBlock` handler as responses arrive.
### `subscribeSMPQueues` → transpile `Client.hs:840-845`
```haskell
subscribeSMPQueues c qs = do
liftIO $ enablePings c
sendProtocolCommands c NRMBackground cs >>= mapM (processSUBResponse c)
where
cs = L.map (\(rId, rpKey) -> (rId, Just rpKey, Cmd SRecipient SUB)) qs
```
### `processSUBResponse` → transpile `Client.hs:854-862`
```haskell
processSUBResponse c (Response rId r) = pure r $>>= processSUBResponse_ c rId
processSUBResponse_ c rId = \case
OK -> pure $ Right Nothing
SOK serviceId_ -> pure $ Right serviceId_
cmd@MSG {} -> writeSMPMessage c rId cmd $> Right Nothing
r' -> pure . Left $ unexpectedResponse r'
```
MSG → push to `onMessage`, return success. Same pattern as single subscribe.
### `deleteSMPQueues` → transpile `Client.hs:1062-1065`
```haskell
deleteSMPQueues = okSMPCommands DEL
```
Uses `okSMPCommands` (`Client.hs:1245-1253`) which calls `sendProtocolCommands` and checks each response is OK.
### `connectSMPProxiedRelay` → transpile `Client.hs:1069-1093`
Call chain:
1. Send `PRXY relayServ proxyAuth` to proxy (via `sendProtocolCommand_`, entityId = NoEntity)
2. Receive `PKEY sessionId versionRange certChainPubKey`
3. Check version compatibility
4. `validateRelay chain key` — validate cert chain against relay's keyHash, extract X25519 key
5. Return `ProxiedRelay {sessionId, version, auth, relayKey}`
`validateRelay` (`Client.hs:1085-1093`):
1. `chainIdCaCerts chain` → extract leaf, id, ca certs
2. Check `Fingerprint kh == getFingerprint idCert SHA256`
3. `x509validate caCert (hostName, port) chain`
4. Extract server key from leaf cert
5. Verify signed key against server key
In browser: we already have `verifyIdentityProof` and `extractSignedKey` from xftp-web. Need to adapt for relay validation where we receive the cert chain in the PKEY response (DER-encoded, not from TLS handshake).
### `proxySMPCommand` → transpile `Client.hs:1157-1206`
Call chain:
1. Construct `serverThParams` = `smpTHParamsSetVersion v proxyThParams {sessionId, thAuth = serverThAuth}`
- `serverThAuth = thAuth proxyThParams with peerServerPubKey = relayKey`
2. Generate ephemeral X25519 keypair: `(cmdPubKey, cmdPrivKey)`
3. `cmdSecret = dh(relayKey, cmdPrivKey)`
4. Generate random nonce (also used as corrId)
5. `encodeTransmissionForAuth serverThParams (CorrId corrId, sId, Cmd sParty command)` — encode as if sending to relay
6. `authTransmission serverThAuth False spKey nonce tForAuth` — authenticate with entity key against relay
7. `batchTransmissions serverThParams [Right (auth, tToSend)]` — batch into single block
8. `cbEncrypt cmdSecret nonce batchBlock paddedProxiedTLength``EncTransmission`
9. Send `PFWD version cmdPubKey encTransmission` to proxy (entityId = sessionId)
10. Receive `PRES (EncResponse encResponse)`
11. `cbDecrypt cmdSecret (reverseNonce nonce) encResponse` — decrypt relay's response
12. `tParse serverThParams decrypted` — parse as relay's response
13. `tDecodeClient serverThParams parsed` — decode command
14. Classify: `Right (ERR e)` → throw PCEProtocolError, `Right r` → return Right r, `Left e` → throw PCEResponseError
Error wrapping (`Client.hs:1200-1206`): proxy-level errors (from PFWD response itself) → `ProxyClientError` returned as `Left`. Relay-level errors (inside PRES) → `PCEProtocolError` thrown.
### `paddedProxiedTLength``Protocol.hs:306-307` = 16226
## Constants
```
paddedProxiedTLength = 16226 -- Protocol.hs:306
serviceCertsSMPVersion = 16 -- Transport.hs:213
```
## Testing
### Unit tests (callNode, no server)
Each encoding function tested byte-for-byte against Haskell:
1. `reverseNonce` — reverse known bytes, compare
2. `encodeProtocolServer` — encode known server, compare with `smpEncode @SMPServer`
3. `encodePRXY` — encode PRXY command, compare with `encodeProtocol v (Cmd SProxiedClient (PRXY srv auth))`
4. `encodePFWD` — encode PFWD command, compare with `encodeProtocol v (Cmd SProxiedClient (PFWD v pk et))`
5. `batchTransmissions` with multiple commands — same batch boundaries as `batchTransmissions_` in Haskell
### Integration tests (with two SMP servers, from SMPProxyTests.hs pattern)
6. `connectProxiedRelay` — JS connects to proxy, sends PRXY for relay, gets PKEY, validates cert, extracts key
7. `proxySMPMessage` — JS sends SEND via proxy to relay, HS receiver gets MSG
8. Full proxy roundtrip — JS creates queue on relay via proxy, sends message via proxy, HS receives
### Batch tests
9. `subscribeSMPQueues` — JS batch-subscribes to N queues, verifies all subscribed
10. `deleteSMPQueues` — JS batch-deletes N queues
## Implementation order
1. `reverseNonce`, `cbDecryptNoPad`
2. `encodeProtocolServer`, `encodePRXY`, `encodePFWD`
3. `decodePKEY`, `decodePRES`, update `decodeResponse`
4. Unit tests for steps 1-3
5. Refactor `sendCommand` → split into `mkTransmission` (encode+register) and send
6. `sendProtocolCommands`, `sendBatch`
7. `subscribeSMPQueues`, `deleteSMPQueues`
8. Batch integration tests
9. `connectSMPProxiedRelay` (cert validation, PRXY/PKEY)
10. `proxySMPCommand`, `proxySMPMessage`
11. Proxy integration tests
## Files
| File | Action |
|---|---|
| `smp-web/src/crypto.ts` | `reverseNonce`, `cbDecryptNoPad` |
| `smp-web/src/protocol.ts` | `encodeProtocolServer`, `encodePRXY`, `encodePFWD`, `decodePKEY`, `decodePRES` |
| `smp-web/src/client.ts` | `sendProtocolCommands`, `sendBatch`, `subscribeSMPQueues`, `deleteSMPQueues`, `connectSMPProxiedRelay`, `proxySMPCommand` |
| `smp-web/tests/client-repl.ts` | Proxy + batch REPL commands |
| `tests/SMPWebTests.hs` | Tests |
## Haskell source — exact lines to transpile
| TS function | Haskell function | File:lines |
|---|---|---|
| `reverseNonce` | `reverseNonce` | `Crypto.hs:1409-1410` |
| `cbDecryptNoPad` | `cbDecryptNoPad` / `sbDecryptNoPad_` | `Crypto.hs:1330-1331` |
| `encodeProtocolServer` | `instance Encoding (ProtocolServer p)` | `Protocol.hs:1264-1266` |
| `encodePRXY` | `encodeProtocol v (PRXY ...)` | `Protocol.hs:1710` |
| `encodePFWD` | `encodeProtocol v (PFWD ...)` | `Protocol.hs:1711` |
| `decodePKEY` | `protocolP v PKEY_` | `Protocol.hs:1894` |
| `decodePRES` | `protocolP v PRES_` | `Protocol.hs:1896` |
| `sendProtocolCommands` | `sendProtocolCommands` | `Client.hs:1262-1278` |
| `sendBatch` | `sendBatch` | `Client.hs:1285-1298` |
| `subscribeSMPQueues` | `subscribeSMPQueues` | `Client.hs:840-845` |
| `processSUBResponse` | `processSUBResponse` + `processSUBResponse_` | `Client.hs:854-862` |
| `deleteSMPQueues` | `deleteSMPQueues` via `okSMPCommands` | `Client.hs:1062-1065, 1245-1253` |
| `connectSMPProxiedRelay` | `connectSMPProxiedRelay` | `Client.hs:1069-1093` |
| `validateRelay` | `validateRelay` (inside `connectSMPProxiedRelay`) | `Client.hs:1085-1093` |
| `proxySMPCommand` | `proxySMPCommand` | `Client.hs:1157-1206` |
| `proxyOKSMPCommand` | `proxyOKSMPCommand` | `Client.hs:1150-1155` |
| `smpTHParamsSetVersion` | `smpTHParamsSetVersion` | `Transport.hs:921-926` |
| `batchTransmissions'` | `batchTransmissions'` | `Protocol.hs:2135-2148` |
| `batchTransmissions_` | `batchTransmissions_` | `Protocol.hs:2150-2169` |
@@ -0,0 +1,195 @@
# Agent for Browser: Transpilation Breakdown
**Parent**: [SMP Client MVP](./2026-05-20-client-mvp.md)
**Depends on**: SMP Client (complete, 96 tests), encoding/encryption spike (complete)
## Rule
Every TypeScript function is a faithful transpilation of a specific Haskell function. Same name, same steps, same call chain. No inferences.
## Scope
The web widget JOINS connections (never creates addresses). It sends and receives messages. It handles the connection handshake. It does NOT create invitations, manage notifications, transfer files, or do remote control.
## Architecture difference from Haskell
Haskell agent uses SQLite + multiple background threads (subscriber, delivery workers, cleanup manager, NTF supervisor). Browser agent uses IndexedDB + event-driven architecture (WebSocket onmessage, Promises, no threads).
The protocol logic is identical. The concurrency model differs. The store interface differs. The transpilation focuses on the protocol logic.
## Breakdown into testable pieces
### Piece 1: Agent protocol types (Agent/Protocol.hs)
Already partially done (AgentMsgEnvelope, AgentMessage, APrivHeader, AMessage). What's missing for the handshake:
| Type | Haskell location | What's needed |
|------|-----------------|---------------|
| `SMPQueueInfo` | `Agent/Protocol.hs:1310-1327` | Binary encode/decode — version-dependent, complex |
| `SMPQueueUri` | `Agent/Protocol.hs:1344-1431` | Binary encode/decode + string encode/decode |
| `SMPQueueAddress` | `Agent/Protocol.hs:1350-1356` | `{smpServer, senderId, dhPublicKey, queueMode}` |
| `ConnectionRequestUri` | `Agent/Protocol.hs:1436-1441` | Binary encode/decode: `CRInvitationUri` + `CRContactUri` |
| `ConnReqUriData` | `Agent/Protocol.hs:1728-1734` | Binary encode/decode: `{crAgentVRange, crSmpQueues, crClientData}` |
| `SMPConfirmation` | `Agent/Protocol.hs:798-810` | Not wire-encoded — internal data structure for confirmation handling |
| `E2ERatchetParams` | `Crypto/Ratchet.hs:223-241` | Already have encode/decode in ratchet.ts — need to verify completeness |
**Test**: each encode/decode function tested byte-for-byte against Haskell via callNode.
### Piece 2: Connection handshake — joinConnection (Agent.hs)
The join flow, transpiled step by step:
```
joinConnection (Agent.hs:~1200-1300)
1. Parse ConnectionRequestUri (already have URI parsing)
2. Create RcvQueue on a selected server (newRcvQueue — Agent/Client.hs:1373)
- Generate X25519 DH keypair for queue
- Generate X25519/Ed25519 auth keypair
- Call createSMPQueue on SMP client
- Get back rcvId, sndId, srvDhKey
3. Store connection + queue in database
4. Generate X448 E2E ratchet params (generateRcvE2EParams — already have)
5. Build ConnInfo (profile data)
6. Encrypt ConnInfo with ratchet → encConnInfo
7. Build AgentConfirmation envelope
8. Wrap in ClientMessage + per-queue E2E encrypt → ClientMsgEnvelope
9. Send via SMP SEND to the contact address queue
10. Subscribe to own receive queue (SUB)
11. Return connection ID
```
Each step is independently testable. The full flow is an integration test.
**Key Haskell functions to transpile:**
| Function | File:lines | What it does |
|----------|-----------|--------------|
| `joinConnection` | `Agent.hs:~1200` | Top-level join |
| `joinConn` | `Agent.hs:~1230` | Internal join logic |
| `newRcvQueue` | `Agent/Client.hs:1373-1420` | Create queue on server |
| `sendConfirmation` | `Agent/Client.hs:1788-1794` | Encrypt+send confirmation |
| `sendInvitation` | `Agent/Client.hs:1796-1806` | Encrypt+send invitation |
| `mkAgentConfirmation` | `Agent.hs:~3700` | Build confirmation envelope |
| `agentCbEncrypt` | `Agent/Client.hs:2074-2082` | Per-queue E2E encrypt (already have) |
### Piece 3: Message processing — subscriber (Agent.hs)
Incoming message handling:
```
subscriber (Agent.hs:2912-2919)
→ reads from msgQ (populated by SMP client's onMessage callback)
→ processSMPTransmissions (Agent.hs:2997-3297)
→ for each transmission:
→ STEvent (server push MSG):
→ decryptClientMessage (per-queue E2E decrypt)
→ parse AgentMsgEnvelope
→ for AgentMsgEnvelope 'M':
→ agentRatchetDecrypt (double ratchet decrypt)
→ parse AgentMessage
→ dispatch on AMessage type:
→ HELLO: complete handshake
→ A_MSG body: deliver to user
→ A_RCVD: delivery receipt
→ QADD/QKEY/QUSE/QTEST: queue switching (skip for MVP)
→ EREADY: ratchet sync (skip for MVP)
```
**Key Haskell functions to transpile:**
| Function | File:lines | What it does |
|----------|-----------|--------------|
| `processSMPTransmissions` | `Agent.hs:2997-3297` | Top-level message dispatcher |
| `decryptClientMessage` | `Agent.hs:3282-3293` | Per-queue E2E decrypt + parse envelope |
| `agentRatchetDecrypt` | `Agent.hs:3757-3767` | Ratchet decrypt + store update |
| `helloMsg` | `Agent.hs:~3400` | Process HELLO (complete handshake) |
| `smpConfirmation` | `Agent.hs:~3500` | Process received confirmation |
| `smpInvitation` | `Agent.hs:~3600` | Process received invitation |
### Piece 4: Message sending — sendMessage (Agent.hs)
```
sendMessage (Agent.hs:~1500)
→ enqueueMessageB
→ agentRatchetEncryptHeader (get ratchet encrypt key)
→ rcEncryptMsg (encrypt message body)
→ store encrypted message in DB
→ createSndMsgDelivery (link to queue)
→ delivery worker sends via SMP SEND
```
**Key Haskell functions to transpile:**
| Function | File:lines | What it does |
|----------|-----------|--------------|
| `sendMessage` | `Agent.hs:~1500` | Top-level send |
| `enqueueMessageB` | `Agent.hs:~2020-2060` | Encode + encrypt + store |
| `agentRatchetEncrypt` | `Agent.hs:3742-3746` | Ratchet encrypt |
| `agentRatchetEncryptHeader` | `Agent.hs:3748-3754` | Ratchet encrypt header |
| `encodeAgentMsgStr` | `Agent.hs:2050-2054` | Encode AgentMessage to bytes |
| `runSmpQueueMsgDelivery` | `Agent.hs:2092-2220` | Delivery worker — read from DB, send via SMP |
### Piece 5: Message acknowledgment (Agent.hs)
```
ackMessage (Agent.hs:~1550)
→ withStore: mark message as acknowledged
→ send ACK to SMP server
→ optionally send delivery receipt (A_RCVD)
```
### Piece 6: Store interface (IndexedDB)
The agent uses ~50 distinct store operations. For the web widget MVP, we need:
| Store operation | What it does | Used by |
|----------------|--------------|---------|
| `createConnection` | Create connection record | joinConnection |
| `getConn` | Get connection by ID | all operations |
| `updateRcvIds` | Increment receive IDs | message processing |
| `createRcvMsg` | Store received message | message processing |
| `getRatchetForUpdate` | Get ratchet state for modify | encrypt/decrypt |
| `updateRatchet` | Store updated ratchet | encrypt/decrypt |
| `getSkippedMsgKeys` | Get skipped message keys | ratchet decrypt |
| `createSndMsg` | Store sent message | sendMessage |
| `createSndMsgDelivery` | Link message to queue | sendMessage |
| `getPendingQueueMsg` | Get next message to send | delivery worker |
| `updateSndMsgStatus` | Update delivery status | delivery worker |
| `deleteMsg` | Delete after ACK | ackMessage |
## Implementation order
1. **Piece 1**: Agent protocol types — `SMPQueueInfo`, `SMPQueueUri`, `ConnectionRequestUri`, `ConnReqUriData` encode/decode. Test each against Haskell.
2. **Piece 6**: Store interface — define TypeScript interface matching the needed operations. Implement with in-memory Map first (for testing), IndexedDB later.
3. **Piece 4**: Message sending — `sendMessage` + `agentRatchetEncrypt` + delivery. Test: encrypt in TS, decrypt in HS.
4. **Piece 3**: Message processing — `processSMPTransmissions` + `agentRatchetDecrypt`. Test: encrypt in HS, decrypt in TS.
5. **Piece 2**: Connection handshake — `joinConnection`. Test: TS joins, HS accepts, messages flow.
6. **Piece 5**: Message acknowledgment — `ackMessage`. Test: full roundtrip with ack.
Each piece is independently testable. Piece 1 uses callNode (no server). Pieces 3-5 use the SMP client REPL + real server. Piece 2 is the integration test that ties everything together.
## What to skip
| Feature | Why skip | Haskell functions |
|---------|----------|-------------------|
| Queue switching | Additive, not needed for MVP | `switchConnection`, QADD/QKEY/QUSE/QTEST handling |
| Ratchet sync | MVP shows error, suggests reconnecting | `synchronizeRatchet`, EREADY handling |
| Notifications | No webpush yet | All NTF functions |
| File transfer | Separate protocol | All XFTP functions |
| Remote control | Not in scope | All RC functions |
| Client notices | Server admin feature | `processClientNotices` |
| Cleanup manager | Can do manual cleanup | `cleanupManager` |
| Server management | Configured at init | `setProtocolServers`, `testProtocolServer` |
| Connection creation | Widget only joins | `createConnection`, short links |
| Delivery receipts | Can add later | A_RCVD handling, receipt sending |
| Multiple receive queues | Single queue per connection for MVP | Queue replacement logic |
## Files
| File | Action |
|------|--------|
| `smp-web/src/agent/protocol.ts` | Extend with SMPQueueInfo, ConnectionRequestUri, etc. |
| `smp-web/src/agent/store.ts` | New — store interface + in-memory implementation |
| `smp-web/src/agent/agent.ts` | New — agent logic: join, send, receive, ack |
| `smp-web/tests/agent-repl.ts` | New — agent REPL for testing (or extend client-repl) |
| `tests/SMPWebTests.hs` | Agent tests |
@@ -0,0 +1,260 @@
# Agent API Inventory for Web Widget
Every exported function from `Agent.hs`, classified as MVP / post-MVP / skip, with reasoning.
## Context
The web widget:
- Joins existing connections via addresses hardcoded in the widget or sent as simplex links
- Sends and receives messages
- Does NOT create addresses, invitation links, or short links
- Does NOT transfer files (post-MVP)
- Does NOT manage notifications (post-MVP)
- Does NOT rotate queues
- Must accept ratchet re-sync initiated by the other side (but does not initiate)
## Lifecycle
| Function | MVP? | Reason |
|----------|------|--------|
| `getSMPAgentClient` | YES | Initialize agent — required |
| `getSMPAgentClient_` | NO | Variant with extra params, not needed |
| `disconnectAgentClient` | YES | Clean shutdown — required |
| `disposeAgentClient` | NO | Hard shutdown — disconnectAgentClient is sufficient |
| `resumeAgentClient` | NO | Widget doesn't suspend/resume — runs while page is open |
| `foregroundAgent` | NO | Mobile-only concept |
| `suspendAgent` | NO | Mobile-only concept |
## User management
| Function | MVP? | Reason |
|----------|------|--------|
| `createUser` | YES | Widget needs at least one user to own connections |
| `deleteUser` | NO | Widget doesn't delete users — page reload is cleanup |
## Connection creation (widget never creates)
| Function | MVP? | Reason |
|----------|------|--------|
| `createConnection` | NO | Widget joins, never creates |
| `createConnectionAsync` | NO | Same |
| `prepareConnectionLink` | NO | For creating links |
| `createConnectionForLink` | NO | For creating links |
| `setConnShortLink` | NO | For creating short links |
| `setConnShortLinkAsync` | NO | Same |
| `deleteConnShortLink` | NO | For managing short links |
| `getConnShortLink` | NO | For reading short links — widget uses hardcoded address |
| `getConnShortLinkAsync` | NO | Same |
| `getConnLinkPrivKey` | NO | For link management |
| `deleteLocalInvShortLink` | NO | For link cleanup |
| `changeConnectionUser` | NO | Widget has one user |
## Joining connections
| Function | MVP? | Reason |
|----------|------|--------|
| `prepareConnectionToJoin` | YES | Create connection record before join — prevents race with incoming confirmation |
| `joinConnection` | YES | Core function — join via address URI |
| `joinConnectionAsync` | NO | Async variant — widget can use sync joinConnection with await |
| `connRequestPQSupport` | YES | Determine PQ support from connection request — needed for join |
## Handshake (accepting incoming)
| Function | MVP? | Reason |
|----------|------|--------|
| `allowConnection` | YES | Allow connection after receiving CONF — part of handshake |
| `allowConnectionAsync` | NO | Async variant |
| `acceptContact` | YES | Accept contact request (for group join flow) |
| `acceptContactAsync` | NO | Async variant |
| `prepareConnectionToAccept` | YES | Prepare for accept — same race prevention as prepareConnectionToJoin |
| `rejectContact` | NO | Widget doesn't reject — it always accepts |
## Subscription
| Function | MVP? | Reason |
|----------|------|--------|
| `subscribeConnection` | YES | Subscribe to receive messages on one connection |
| `subscribeConnections` | YES | Batch subscribe — needed after page reload |
| `subscribeAllConnections` | YES | Subscribe everything for a user — simplest for widget |
| `resubscribeConnection` | YES | Resubscribe after network recovery |
| `resubscribeConnections` | YES | Batch resubscribe |
| `getConnectionMessages` | NO | Fetch stored messages — widget processes messages as they arrive |
| `getNotificationConns` | NO | Push notification management |
| `subscribeClientService` | NO | Service certificate management |
## Messaging
| Function | MVP? | Reason |
|----------|------|--------|
| `sendMessage` | YES | Send a message — core function |
| `sendMessages` | NO | Batch send — sendMessage is sufficient for MVP |
| `sendMessagesB` | NO | Batch send with error handling — optimization |
| `ackMessage` | YES | Acknowledge received message — required for protocol correctness |
| `ackMessageAsync` | NO | Async variant |
## Queue management
| Function | MVP? | Reason |
|----------|------|--------|
| `switchConnection` | NO | Queue rotation — not needed |
| `switchConnectionAsync` | NO | Same |
| `abortConnectionSwitch` | NO | Cancel rotation |
| `getConnectionQueueInfo` | NO | Debug info |
| `suspendConnection` | NO | Widget deletes or ignores, doesn't suspend |
## Ratchet
| Function | MVP? | Reason |
|----------|------|--------|
| `synchronizeRatchet` | NO | Widget doesn't initiate ratchet sync. But MUST handle incoming EREADY — that's in processSMPTransmissions, not a separate API call. |
| `getConnectionRatchetAdHash` | NO | Verification UI not in widget |
## Connection cleanup
| Function | MVP? | Reason |
|----------|------|--------|
| `deleteConnection` | YES | User may want to delete a conversation |
| `deleteConnectionAsync` | NO | Async variant |
| `deleteConnections` | NO | Batch delete — single delete sufficient |
| `deleteConnectionsAsync` | NO | Same |
| `getConnectionServers` | NO | Info only |
| `compareConnections` | NO | Database sync tool |
| `syncConnections` | NO | Database sync tool |
## Server configuration
| Function | MVP? | Reason |
|----------|------|--------|
| `setProtocolServers` | NO | Widget initialized with servers, doesn't change them |
| `checkUserServers` | NO | Admin function |
| `testProtocolServer` | NO | Admin function |
| `setNtfServers` | NO | No notifications in MVP |
| `setNetworkConfig` | NO | Widget uses default network config |
| `setUserNetworkInfo` | NO | Widget doesn't track network state changes |
| `reconnectAllServers` | NO | Widget handles reconnection via SMP client |
| `reconnectSMPServer` | NO | Same |
## Notifications (all post-MVP)
| Function | MVP? | Reason |
|----------|------|--------|
| `registerNtfToken` | NO | No webpush yet |
| `verifyNtfToken` | NO | Same |
| `checkNtfToken` | NO | Same |
| `deleteNtfToken` | NO | Same |
| `getNtfToken` | NO | Same |
| `getNtfTokenData` | NO | Same |
| `toggleConnectionNtfs` | NO | Same |
## File transfer (all post-MVP)
| Function | MVP? | Reason |
|----------|------|--------|
| `xftpStartWorkers` | NO | Post-MVP |
| `xftpStartSndWorkers` | NO | Same |
| `xftpReceiveFile` | NO | Same |
| `xftpDeleteRcvFile` | NO | Same |
| `xftpDeleteRcvFiles` | NO | Same |
| `xftpSendFile` | NO | Same |
| `xftpSendDescription` | NO | Same |
| `xftpDeleteSndFileInternal` | NO | Same |
| `xftpDeleteSndFilesInternal` | NO | Same |
| `xftpDeleteSndFileRemote` | NO | Same |
| `xftpDeleteSndFilesRemote` | NO | Same |
## Remote control (all skip)
| Function | MVP? | Reason |
|----------|------|--------|
| `rcNewHostPairing` | NO | Not in scope |
| `rcConnectHost` | NO | Same |
| `rcConnectCtrl` | NO | Same |
| `rcDiscoverCtrl` | NO | Same |
## Debug/stats (all skip)
| Function | MVP? | Reason |
|----------|------|--------|
| `getAgentSubsTotal` | NO | Debug |
| `getAgentServersSummary` | NO | Debug |
| `resetAgentServersStats` | NO | Debug |
| `execAgentStoreSQL` | NO | Debug |
| `getAgentMigrations` | NO | Debug |
| `debugAgentLocks` | NO | Debug |
| `getAgentSubscriptions` | NO | Debug |
| `logConnection` | NO | Debug |
| `withAgentEnv` | NO | Test utility |
## Summary
### Corrections after reviewing simplex-chat-2 Subscriber.hs + Commands.hs
The widget handles business chats (groups). Group flows trigger agent calls the widget doesn't initiate but must support:
- Member introductions create connections asynchronously → `createConnectionAsync`
- Members join via introductions → `joinConnectionAsync`, `prepareConnectionToJoin`
- Members leave/deleted → `deleteConnectionAsync`, `deleteConnectionsAsync`
- Group messages go to all members → `sendMessages`
- All message acks use async variant → `ackMessageAsync`
- Accepting group invitations → `allowConnectionAsync`
- `acceptContact` — used in contact request acceptance flow (APIAcceptContactRequest in Commands.hs)
- `toggleConnectionNtfs` — used when CON received for group member (Subscriber.hs:850)
### MVP functions (27 of 90):
**Lifecycle:** `getSMPAgentClient`, `disconnectAgentClient`
**User:** `createUser`
**Join:** `prepareConnectionToJoin`, `joinConnection`, `joinConnectionAsync`, `connRequestPQSupport`
**Handshake:** `allowConnection`, `allowConnectionAsync`, `acceptContact`, `acceptContactAsync`, `prepareConnectionToAccept`
**Connection creation (for group member flows):** `createConnectionAsync`, `deleteConnectionAsync`, `deleteConnectionsAsync`
**Subscribe:** `subscribeConnection`, `subscribeConnections`, `subscribeAllConnections`, `resubscribeConnection`, `resubscribeConnections`
**Message:** `sendMessage`, `sendMessages`, `ackMessage`, `ackMessageAsync`
**Cleanup:** `deleteConnection`
**Notification toggle:** `toggleConnectionNtfs`
**Internal (not exported but required):** `subscriber`/`processSMPTransmissions` (message processing), `runSmpQueueMsgDelivery` (delivery worker), all encryption/decryption functions, store operations for the above.
### Message types to handle in processSMPTransmissions:
| AMessage | Handle? | Reason |
|----------|---------|--------|
| `HELLO` | YES | Complete handshake |
| `A_MSG body` | YES | Deliver message to user |
| `A_RCVD receipts` | YES | Process delivery receipts (show checkmarks) |
| `A_QCONT addr` | NO | Queue continuation after quota — skip |
| `QADD qs` | NO | Queue rotation — skip |
| `QKEY qs` | NO | Queue rotation — skip |
| `QUSE qs` | NO | Queue rotation — skip |
| `QTEST qs` | NO | Queue rotation — skip |
| `EREADY msgId` | ACCEPT | Must handle incoming (don't initiate) — reset ratchet sync state |
### AgentMsgEnvelope types to handle:
| Variant | Handle? | Reason |
|---------|---------|--------|
| `AgentConfirmation` | YES | Handshake — received when peer confirms |
| `AgentMsgEnvelope` | YES | Normal encrypted messages |
| `AgentInvitation` | YES | Received when joining contact address |
| `AgentRatchetKey` | ACCEPT | Must handle incoming ratchet key — don't initiate |
### Store operations needed:
Based on the 18 MVP functions, the store needs (rough count):
- Connection CRUD: ~8 operations
- Queue CRUD: ~6 operations
- Ratchet state: ~4 operations (get, update, skipped keys)
- Message storage: ~6 operations (create rcv/snd msg, update status, delete)
- Message delivery: ~4 operations (create delivery, get pending, update status)
- User: ~2 operations (create, get)
- Confirmation: ~3 operations (create, get, delete)
**Estimated: ~33 store operations.** This is what determines whether SQLite WASM or IndexedDB direct is more practical.
@@ -0,0 +1,121 @@
# Agent Store: IndexedDB Design
**Parent**: [Agent Plan](./2026-05-22-agent.md)
## Schema
IndexedDB object stores, mapped from SQLite tables. Each store mirrors the Haskell schema from `agent_schema.sql`.
### Object stores needed (16)
```
users
key: userId (autoincrement)
fields: deleted
connections
key: connId (Uint8Array)
fields: connMode, lastInternalMsgId, lastInternalRcvMsgId, lastInternalSndMsgId,
lastExternalSndMsgId, lastRcvMsgHash, lastSndMsgHash, smpAgentVersion,
duplexHandshake, enableNtfs, deleted, userId, ratchetSyncState, pqSupport
rcv_queues
key: [host, port, rcvId] (compound)
index: [connId], [host, port, sndId]
fields: connId, rcvPrivateKey, rcvDhSecret, e2ePrivKey, e2eDhSecret, sndId, sndKey,
status, smpClientVersion, rcvQueueId, rcvPrimary, replaceRcvQueueId, queueMode,
serverKeyHash, lastBrokerTs
snd_queues
key: [host, port, sndId] (compound)
index: [connId]
fields: connId, sndPrivateKey, e2eDhSecret, status, smpClientVersion,
sndPublicKey, e2ePubKey, sndQueueId, sndPrimary, queueMode, serverKeyHash
messages
key: [connId, internalId] (compound)
index: [connId]
fields: internalTs, internalRcvId, internalSndId, msgType, msgBody, msgFlags, pqEncryption
rcv_messages
key: [connId, internalRcvId] (compound)
index: [connId, internalId]
fields: internalId, externalSndId, brokerId, brokerTs, internalHash,
externalPrevSndHash, integrity, userAck, rcvQueueId, receiveAttempts
snd_messages
key: [connId, internalSndId] (compound)
index: [connId, internalId]
fields: internalId, internalHash, previousMsgHash, retryIntSlow, retryIntFast,
rcptInternalId, rcptStatus, msgEncryptKey, paddedMsgLen, sndMessageBodyId
snd_message_deliveries
key: sndMessageDeliveryId (autoincrement)
index: [connId, sndQueueId]
fields: connId, sndQueueId, internalId, failed
snd_message_bodies
key: sndMessageBodyId (autoincrement)
fields: agentMsg
conn_confirmations
key: confirmationId (Uint8Array)
index: [connId]
fields: connId, e2eSndPubKey, senderKey, ratchetState, senderConnInfo,
accepted, ownConnInfo, smpReplyQueues, smpClientVersion
conn_invitations
key: invitationId (Uint8Array)
index: [contactConnId]
fields: contactConnId, crInvitation, recipientConnInfo, accepted, ownConnInfo
ratchets
key: connId (Uint8Array)
fields: x3dhPrivKey1, x3dhPrivKey2, ratchetState, e2eVersion,
x3dhPubKey1, x3dhPubKey2, pqPrivKem, pqPubKem
skipped_messages
key: skippedMessageId (autoincrement)
index: [connId]
fields: connId, headerKey, msgN, msgKey
servers
key: [host, port] (compound)
fields: keyHash
commands
key: commandId (autoincrement)
index: [connId], [host, port]
fields: connId, host, port, corrId, commandTag, command, agentVersion, serverKeyHash, failed
encrypted_rcv_message_hashes
key: id (autoincrement)
index: [connId, hash]
fields: connId, hash, createdAt
```
## Interface
TypeScript interface matching the ~60 store operations. Each method maps to a specific Haskell function in `AgentStore.hs`.
The interface will be defined in `src/agent/store.ts`. Implementation in `src/agent/store-idb.ts` (IndexedDB).
## Implementation approach
1. Define the TypeScript interface first — every method name matches the Haskell function name
2. Implement with IndexedDB transactions
3. Test each operation in isolation before wiring to agent
IndexedDB transactions map to SQLite transactions — both are ACID within a single store/table. Cross-store atomicity in IndexedDB requires putting multiple stores in one transaction, which is supported.
## Key differences from SQLite
1. **No SQL joins** — denormalize where needed, or do application-level joins
2. **No AUTO INCREMENT guaranteed ordering** — use explicit counters
3. **Blob keys** — IndexedDB supports ArrayBuffer keys natively
4. **Compound keys** — IndexedDB supports array keys: `[host, port, rcvId]`
5. **Indexes** — must be declared upfront in `onupgradeneeded`
## Testing
Each store operation tested by: write data, read it back, verify it matches. No server needed — pure store tests using `fake-indexeddb` in Node.js.
@@ -0,0 +1,666 @@
# Agent Client Middle Layer: Transpilation Plan
**Parent**: [Agent Plan](./2026-05-22-agent.md)
**Depends on**: Store (complete, 98 tests), SMP Client (complete, 99 tests), Ratchet (complete), Agent Protocol Types (complete)
## Rule
Every TypeScript function is a faithful transpilation of a specific Haskell function. Same name, same steps, same call chain. No inferences, no simplifications, no "browser-friendly" shortcuts. The concurrency primitives differ (Promises vs STM, event callbacks vs TBQueue), but the logic, state transitions, and decision paths must be identical.
## Architecture mapping
| Haskell | TypeScript | Notes |
|---------|-----------|-------|
| `TVar a` | mutable variable (object property) | Single-threaded, no atomicity needed |
| `TMap k v` | `Map<K, V>` | No STM, direct mutation |
| `TBQueue a` | `ABQueue<T>` | `subQ` for user events, `msgQ` for server messages |
| `TMVar a` | `Promise` + resolver, or flag | For worker doWork signaling |
| `STM` transaction | synchronous code block | Single-threaded JS, no races |
| `forkIO` / `async` | `setTimeout(0)` / microtask | Event loop scheduling |
| `Worker` thread | delivery loop function | Triggered by `submitPendingMsg`, runs via microtask |
| `ReaderT Env IO` (AM') | closure over agent state | Config + store + DRG captured in closure |
| `ExceptT AgentErrorType` (AM) | thrown errors / Result type | TBD: throw vs return Either |
## Files to create
| File | Purpose |
|------|---------|
| `smp-web/src/agent/queue.ts` | Sem, ABQueue (copied from simplex-chat) |
| `smp-web/src/agent/tmvar.ts` | TMVar — single-cell blocking variable |
| `smp-web/src/agent/session.ts` | SessionVar, getSessVar (Promise-based) |
| `smp-web/src/agent/retry.ts` | RetryInterval, RetryInterval2, withRetryLock2 |
| `smp-web/src/agent/subscriptions.ts` | TSessionSubs transpilation |
| `smp-web/src/agent/client.ts` | AgentClient state, session management, worker infrastructure, queue operations |
| `smp-web/src/agent/agent.ts` | Top-level agent API (joinConnection, sendMessage, etc.) |
| `smp-web/tests/agent-repl.ts` | Agent-level REPL for cross-language testing |
## Implementation order
Each step produces a testable artifact.
### Step 1: Pure infrastructure (TS-only tests)
- `queue.ts` — Sem, ABQueue (copied verbatim from simplex-chat)
- `tmvar.ts` — TMVar (new)
- `session.ts` — SessionVar, getSessVar, removeSessVar, tryReadSessVar
- `retry.ts` — RetryInterval types + nextRetryDelay + withRetryInterval + withRetryLock2
- `subscriptions.ts` — TSessionSubs (all 22 functions)
- **Test**: TS unit tests for each module — no server needed
### Step 2: AgentClient state + worker infrastructure (TS-only tests)
- AgentClient record, newAgentClient
- Worker, newWorker, getAgentWorker, runWorkerAsync, waitForWork, hasWorkToDo, withWork
- AgentOpState, operation bracket, suspend/resume
- Locking (withConnLock, withInvLock)
- Server selection (userServers, pickServer, getNextServer, withNextSrv)
- Store wrappers (withStore, withStore', storeError)
- **Test**: TS tests — create agent client, test worker lifecycle, test server selection
### Step 3a: SMP session management + direct queue operations
- getSMPServerClient, smpConnectClient, waitForSMPClient
- agentCbEncrypt, agentCbEncryptOnce, agentCbDecrypt (using existing ClientMsgEnvelope from protocol.ts)
- sendAgentMessage, sendConfirmation, sendInvitation
- secureQueue, secureSndQueue, sendAck
- decryptSMPMessage
- newRcvQueue
- subscribeQueues, subscribeServerQueues
- addNewQueueSubscription
- **Test**: TS-only round-trip test for agentCbEncrypt (encode → decode → decrypt), then agent-repl cross-language tests (TS creates queue → Haskell sends → TS receives, TS sends → Haskell receives)
### Step 3b: Disconnect handling + resubscription
- smpClientDisconnected (full: remove proxied relays, notify DOWN, trigger resubscribe)
- resubscribeSMPSession, resubscribeSessQueues
- processSubResults (partition into failed/subscribed)
- subscribeSessQueues_ (batch SUB + process results)
- **Test**: agent-repl: establish connection → kill WebSocket → verify subs move to pending → reconnect → verify resubscribed
### Step 3c: Proxy operations
- getSMPProxyClient (get/create proxied relay session)
- withProxySession (bracket for proxied operations)
- sendOrProxySMPMessage (decide direct vs proxy, delegate)
- sendOrProxySMPCommand (decide direct vs proxy for SKEY etc)
- ipAddressProtected, shouldUseProxy logic
- withClient_, withClient, withSMPClient, withLogClient_
- **Test**: agent-repl: TS sends via proxy → Haskell receives
### Step 4: Agent message flow (cross-language end-to-end)
- agentRatchetEncrypt, agentRatchetEncryptHeader, agentRatchetDecrypt
- encodeAgentMsgStr
- enqueueMessageB, storeConfirmation, enqueueConfirmation
- submitPendingMsg, getDeliveryWorker, runSmpQueueMsgDelivery
- enqueueCommand, runCommandProcessing
- **Test via agent-repl**: TS encrypts agent message → Haskell decrypts, Haskell encrypts → TS decrypts
### Step 5: Connection handshake + full agent API (cross-language end-to-end)
- newConnToJoin, joinConn, joinConnSrv, startJoinInvitation
- compatibleInvitationUri, compatibleContactUri
- secureConfirmQueue(Async), agentSecureSndQueue
- mkAgentConfirmation, createReplyQueue, newRcvConnSrv, createRcvQueue
- newSndQueue, connectReplyQueues
- allowConnection'
- processSMPTransmissions, subscriber
- decryptClientMessage, agentClientMsg
- smpConfirmation, helloMsg, smpInvitation
- sendMessage', sendMessagesB_
- ackMessage', ackQueueMessage
- subscribeConnection(s)
- **Test**: TS joins invitation URI created by Haskell agent → handshake completes → messages flow both ways → ack
---
## Piece -1: Concurrency primitives
### `Sem` and `ABQueue` — copy from simplex-chat
Copy verbatim from `/code/simplex-chat/packages/simplex-chat-client/typescript/src/queue.ts` into `smp-web/src/agent/queue.ts`.
`Sem` — counting semaphore. `wait()` blocks if permits=0. `signal()` increments and wakes a waiter.
`ABQueue` — async bounded queue. Two semaphores (enq for items, deq for slots). Backpressure on full. Close via sentinel. Implements AsyncIterator.
Used for:
- `subQ` — agent events to user. Agent writes, user reads via `dequeue()` loop or async iterator.
- `msgQ` — WebSocket onmessage enqueues, subscriber loop dequeues and calls `processSMPTransmissions`.
- Queues prevent deadlock: without them, processing a received message that triggers a send, which triggers another event, could cause unbounded reentrancy in single-threaded JS.
### `TMVar<T>` — new, in `smp-web/src/agent/tmvar.ts`
Single-cell mutable variable, empty or full. Blocking take/put/read.
```typescript
class TMVar<T> {
private val: T | undefined
private full: boolean
private takeQ: Array<(v: T) => void> = [] // waiters for value to appear
private putQ: Array<(v: T) => void> = [] // waiters for cell to empty
static empty<T>(): TMVar<T> // create empty
static new<T>(v: T): TMVar<T> // create full
take(): Promise<T> // block until full, take value, leave empty
put(v: T): Promise<void> // block until empty, put value
read(): Promise<T> // block until full, return value without taking
tryTake(): T | undefined // non-blocking take
tryPut(v: T): boolean // non-blocking put, returns false if full
tryRead(): T | undefined // non-blocking read
isEmpty(): boolean
}
```
Used for:
- `doWork :: TMVar ()` — worker signal. `waitForWork` = `read()`. `noWorkToDo` = `tryTake()`. `hasWorkToDo` = `tryPut(undefined)`.
- `action :: TMVar (Maybe ThreadId)` — worker running state. `runWorkerAsync` takes, checks, starts async loop.
- Retry lock in `withRetryLock2`.
The doWork race condition: worker clears FIRST (`tryTake`), THEN checks store. If work found, re-sets (`tryPut`). Any signal arriving during the store check (via `await` yielding to onmessage → `hasWorkToDo`) stays set because it happened after the clear. If we did read-then-clear-if-empty, the clear could swallow a signal set between the store check and the clear.
### Locks — `Sem(1)` or Promise chain
For `withConnLock`, `withInvLock`: `Map<string, Lock>`. Each Lock is either:
- `Sem(1)` — acquire = `wait()`, release = `signal()`, wrap in try/finally
- Or Promise chain (each `withLock` appends to previous promise)
`Sem(1)` is simpler and correct. Wrap in helper:
```typescript
async function withLock(locks: Map<string, Sem>, key: string, fn: () => Promise<T>): Promise<T> {
let sem = locks.get(key)
if (!sem) { sem = new Sem(1); locks.set(key, sem) }
await sem.wait()
try { return await fn() } finally { sem.signal() }
}
```
### SessionVar — Promise with exposed resolver
`SessionVar` tracks pending protocol client connections. First caller creates a Promise, connects, resolves it. Subsequent callers await the same Promise.
```typescript
interface SessionVar<T> {
id: number
ts: number
promise: Promise<T>
resolve: (v: T) => void
reject: (e: Error) => void
value: T | undefined // set after resolve, for tryRead
}
```
`getSessVar`: if key exists in Map, return Right (existing). Else create new with unresolved Promise, insert, return Left (new).
`removeSessVar`: delete if ID matches.
`tryReadSessVar`: return `value` if set.
No TMVar needed — Promise coalesces reads naturally.
---
## Piece 0: SessionVar (`session.ts`)
Transpile from `Simplex/Messaging/Session.hs` (43 lines).
| Function | Haskell lines | Purpose |
|----------|--------------|---------|
| `SessionVar` type | 18-22 | `{sessionVar: TMVar a, sessionVarId: number, sessionVarTs: Date}` |
| `getSessVar` | 24-33 | Get existing or create new empty session var for key |
| `removeSessVar` | 35-39 | Remove if ID matches (guards against removing replaced session) |
| `tryReadSessVar` | 41-42 | Non-blocking read of session var value |
Browser: `TMVar a``{value: T | undefined, resolve: (() => void) | null}`. `getSessVar` returns Left (new, empty) or Right (existing).
---
## Piece 1: RetryInterval (`retry.ts`)
Transpile from `Agent/RetryInterval.hs` (119 lines).
| Function | Haskell lines | Purpose |
|----------|--------------|---------|
| `RetryInterval` type | 27-31 | `{initialInterval, increaseAfter, maxInterval}` (microseconds) |
| `RetryInterval2` type | 33-36 | `{riSlow, riFast}` |
| `RI2State` type | 38-41 | `{slowInterval, fastInterval}` |
| `RetryIntervalMode` type | 51 | `RISlow \| RIFast` |
| `nextRetryDelay` | 114-118 | Pure: if elapsed < increaseAfter, keep delay; else min(delay*3/2, max) |
| `updateRetryInterval2` | 44-49 | Update RI2 from saved state |
| `withRetryInterval` | 54-55 | Wrapper around withRetryIntervalCount |
| `withRetryIntervalCount` | 57-66 | Loop: action(n, delay, loop); loop sleeps then recurses with updated delay |
| `withRetryLock2` | 90-112 | Two-mode retry with lock: action gets RI2State + loop function that takes mode |
Browser adaptation: `threadDelay'``setTimeout` wrapped in Promise. `TMVar` lock → Promise-based signal. Logic identical.
---
## Piece 2: TSessionSubs (`subscriptions.ts`)
Transpile from `Agent/TSessionSubs.hs` (202 lines). Every function, every branch.
Transport session key: `(UserId, SMPServer)` — serialized to string for Map key. One session per server, no per-entity multiplexing.
| Function | Haskell lines | Purpose |
|----------|--------------|---------|
| `TSessionSubs` type | 49-51 | `Map<string, SessSubs>` (string = serialized transport session) |
| `SessSubs` type | 53-57 | `{sessId: SessionId \| null, activeSubs: Map<string, RcvQueueSub>, pendingSubs: Map<string, RcvQueueSub>}` |
| `emptyIO` | 59-61 | Create empty TSessionSubs |
| `clear` | 63-65 | Clear all |
| `getSessSubs` | 71-77 | Get or create SessSubs for a transport session |
| `hasActiveSub` | 79-81 | Check if rcvId has active subscription |
| `hasPendingSub` | 83-85 | Check if rcvId has pending subscription |
| `addPendingSub` | 91-92 | Add to pendingSubs |
| `setSessionId` | 94-99 | Set session ID; if changed, move active→pending |
| `addActiveSub` | 101-110 | If sessId matches, add to active + remove from pending; else add to pending |
| `batchAddActiveSubs` | 112-121 | Batch version of addActiveSub |
| `batchAddPendingSubs` | 123-126 | Batch add to pending |
| `deletePendingSub` | 128-129 | Delete from pending |
| `batchDeletePendingSubs` | 131-134 | Batch delete from pending |
| `deleteSub` | 136-137 | Delete from both active and pending |
| `batchDeleteSubs` | 139-143 | Batch delete from both |
| `hasPendingSubs` | 145-146 | Check if any pending exist for session |
| `getPendingSubs` | 148-150 | Get all pending for session |
| `getActiveSubs` | 152-154 | Get all active for session |
| `setSubsPending` | 159-177 | Move active→pending on disconnect; handles session mode transitions |
| `setSubsPending_` | 179-187 | Internal: write new sessId, move active→pending |
| `updateClientNotices` | 189-192 | Update clientNoticeId on pending subs |
| `foldSessionSubs` | 194-195 | Fold over all sessions |
| `mapSubs` | 197-201 | Map over active and pending |
Critical: `setSubsPending` has mode-dependent logic (TSMEntity vs TSMUser/TSMServer). Must be transpiled exactly.
---
## Piece 3: AgentClient state + worker infrastructure (`client.ts`)
### AgentClient state
Transpile from `AgentClient` record (Client.hs:328-378) and `newAgentClient` (Client.hs:498-584).
| Field | Haskell type | TS type | Purpose |
|-------|-------------|---------|---------|
| `active` | `TVar Bool` | `boolean` | Is agent active |
| `subQ` | `TBQueue ATransmission` | `ABQueue` | Events to user |
| `msgQ` | `TBQueue (ServerTransmissionBatch ...)` | `ABQueue` | WebSocket onmessage enqueues, subscriber dequeues |
| `smpServers` | `TMap UserId (UserServers 'PSMP)` | `Map<UserId, UserServers>` | Server configs per user |
| `smpClients` | `TMap SMPTransportSession SMPClientVar` | `Map<string, SMPClient \| Promise<SMPClient>>` | Active SMP connections |
| `smpProxiedRelays` | `TMap SMPTransportSession SMPServerWithAuth` | `Map<string, SMPServerWithAuth>` | Proxy routing |
| `useNetworkConfig` | `TVar (NetworkConfig, NetworkConfig)` | `{slow: NetworkConfig, fast: NetworkConfig}` | Network config |
| `userNetworkInfo` | `TVar UserNetworkInfo` | `UserNetworkInfo` | Online/offline state |
| `subscrConns` | `TVar (Set ConnId)` | `Set<string>` | Connections being subscribed |
| `currentSubs` | `TSessionSubs` | `TSessionSubs` | Active/pending subscriptions |
| `removedSubs` | `TMap ...` | `Map<string, Map<string, SMPClientError>>` | Failed subscriptions |
| `workerSeq` | `TVar Int` | `number` | Worker ID sequence |
| `smpDeliveryWorkers` | `TMap SndQAddr (Worker, TMVar ())` | `Map<string, DeliveryWorker>` | Per-queue delivery workers |
| `asyncCmdWorkers` | `TMap (ConnId, Maybe SMPServer) Worker` | `Map<string, Worker>` | Async command workers |
| `rcvNetworkOp` | `TVar AgentOpState` | `AgentOpState` | Receive operation state |
| `msgDeliveryOp` | `TVar AgentOpState` | `AgentOpState` | Delivery operation state |
| `sndNetworkOp` | `TVar AgentOpState` | `AgentOpState` | Send operation state |
| `agentState` | `TVar AgentState` | `AgentState` | Foreground/suspended/suspending |
| `connLocks` | `TMap ConnId Lock` | `Map<string, Promise<void>>` | Connection locks |
| `invLocks` | `TMap ByteString Lock` | `Map<string, Promise<void>>` | Invitation locks |
| `agentEnv` | `Env` | closure | Config + store + RNG |
Fields NOT needed for MVP: `ntfServers`, `ntfClients`, `xftpServers`, `xftpClients`, `smpSubWorkers`, `clientNotices`, `clientNoticesLock`, `getMsgLocks`, `deleteLock`, `proxySessTs`, `*Stats`, `srvStatsStartedAt`, `acThread`, `presetDomains`, `presetServers`.
### Worker infrastructure
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `Worker` type | Env/SQLite.hs:317-322 | `{workerId, doWork: TMVar (), action: TMVar (Maybe ThreadId), restarts}` |
| `RestartCount` type | Env/SQLite.hs:324-327 | `{restartMinute, restartCount}` |
| `updateRestartCount` | Env/SQLite.hs:329-332 | Reset count if minute changed, else increment |
| `newWorker` | Client.hs:439-445 | Create worker with doWork TMVar |
| `getAgentWorker` | Client.hs:387-389 | Get-or-create worker for key |
| `getAgentWorker'` | Client.hs:391-437 | Full version with restart logic |
| `runWorkerAsync` | Client.hs:447-454 | Start worker if not running |
| `waitForWork` | Client.hs:2118-2119 | Block until doWork has value |
| `hasWorkToDo` / `hasWorkToDo'` | Client.hs:2171-2176 | Signal work available (tryPutTMVar) |
| `withWork` / `withWork_` | Client.hs:2122-2140 | Wait for work, get item from store, run action |
Browser adaptation: `TMVar ()` → boolean flag + resolver. `forkIO``setTimeout(0)`. Worker restart logic must be preserved exactly.
### Operation state management
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `AgentOpState` type | Client.hs:470 | `{opSuspended, opsInProgress}` |
| `AgentState` type | Client.hs:472-473 | `ASForeground \| ASSuspending \| ASSuspended` |
| `agentOperationBracket` | Client.hs:2232-2245 | Begin/end operation with suspend check |
| `beginAgentOperation` | Client.hs:2223-2230 | Increment opsInProgress |
| `endAgentOperation` | Client.hs:2179-2197 | Decrement opsInProgress, cascade suspend |
| `waitUntilActive` | Client.hs:956-957 | Block until agent is active |
| `throwWhenInactive` | Client.hs:959-962 | Throw if not active |
| `waitWhileSuspended` | Client.hs:2248-2253 | Block while suspended |
| `waitForUserNetwork` | Client.hs:924-928 | Block until network online |
| `noWorkToDo` | Client.hs:2167-2168 | Clear work flag (tryTakeTMVar) |
### Store wrappers
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `withStore` | Client.hs:2259-2270 | Run store action, convert StoreError to AgentErrorType |
| `withStore'` | Client.hs:2255-2257 | Simplified withStore (always Right) |
| `storeError` | Client.hs (exported) | StoreError → AgentErrorType |
### Server selection
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `userServers` | Client.hs:2312-2318 | Get user's server map |
| `pickServer` | Client.hs:2318-2325 | Pick server from NonEmpty list |
| `getNextServer` | Client.hs:2325-2350 | Get next server avoiding used hosts |
| `withNextSrv` | Client.hs:2375-2407 | Retry with next server on failure |
### Locking
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `withConnLock` | Client.hs:1003-1006 | Per-connection mutex |
| `withConnLocks` | Client.hs:1020-1022 | Multiple connection mutex |
| `withInvLock` | Client.hs:1012-1015 | Per-invitation mutex |
Browser: locks via Promise chains. Single-threaded JS means no actual contention, but the ordering semantics must be preserved for async operations.
---
## Piece 4: SMP session management
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `getSMPServerClient` | Client.hs:642-651 | Get or create SMP WebSocket client for transport session |
| `getSMPProxyClient` | Client.hs:653-702 | Get or create proxied relay session |
| `smpConnectClient` | Client.hs:704-718 | Actually connect SMP client via WebSocket |
| `smpClientDisconnected` | Client.hs:720-754 | Handle disconnect: move subs to pending, notify, resubscribe |
| `resubscribeSMPSession` | Client.hs:756-790 | Create resubscription worker |
| `mkTransportSession` | Client.hs:1345-1348 | Build transport session key |
| `mkSMPTransportSession` | Client.hs:1357-1360 | Build SMP transport session from queue |
| `getSessionMode` | Client.hs:1369-1370 | Get current session mode |
| `withClient_` | Client.hs:1037-1045 | Bracket: get client → run action → handle errors |
| `withClient` | Client.hs:1071-1073 | withClient_ + liftClient |
| `withSMPClient` | Client.hs:1079-1082 | withClient for SMP queues |
| `withLogClient_` | Client.hs:1064-1069 | withClient_ with logging |
| `withProxySession` | Client.hs:1047-1062 | Bracket for proxied operations |
| `sendOrProxySMPMessage` | Client.hs:1084-1094 | Decide direct vs proxy for SEND |
| `sendOrProxySMPCommand` | Client.hs:1096-1180 | Decide direct vs proxy for commands (SKEY etc) |
| `ipAddressProtected` | Client.hs:1181-1185 | Check if server is in protected domains |
| `liftClient` | Client.hs:1201-1203 | Convert protocol client error |
| `protocolClientError` | Client.hs:1205-1235 | Error conversion |
| `waitForProtocolClient` | Client.hs:847-868 | Wait for pending client connection |
| `newProtocolClient` | Client.hs:870-896 | Create protocol client with error handling |
| `activeClientSession` | Client.hs:1663-1666 | Check if client session is current (compares sessionId) |
| `removeSubscription` | Client.hs:1752-1755 | Remove single subscription from currentSubs + subscrConns |
| `removeSubscriptions` | Client.hs:1757-1763 | Remove multiple subscriptions |
| `hasActiveSubscription` | Client.hs:1736-1740 | Check if queue has active sub |
| `hasPendingSubscription` | Client.hs:1742-1747 | Check if queue has pending sub |
| `getClientConfig` | Client.hs:904-908 | Get protocol client config (slow/fast network) |
| `getNetworkConfig` | Client.hs:910-918 | Get current network config |
| `getFastNetworkConfig` | Client.hs:920-922 | Get fast network config |
| `slowNetworkConfig` | Client.hs:586-592 | Derive slow config from fast |
| `batchQueues` | Client.hs:1679-1684 | Group queues by transport session |
| `sendTSessionBatches` | Client.hs:1674-1678 | Send batched operations per session (mapConcurrently) |
| `sendClientBatch` | Client.hs:1686-1688 | Send batch to single client session (wrapper) |
| `sendClientBatch_` | Client.hs:1690-1722 | Send batch to single client: get client, run action, handle errors |
| `checkQueues` | Client.hs:1590-1595 | Filter out prohibited queues (GET lock check) |
| `subscribeSessQueues_` | Client.hs:1611-1651 | Send SUB batch via sendClientBatch_ + process results |
---
## Piece 5: Queue operations (use session management)
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `newRcvQueue` | Client.hs:1373-1377 | Generate keys, create queue on SMP server |
| `newRcvQueue_` | Client.hs:1394-1474 | Full queue creation: auth keys, DH, createSMPQueue, build RcvQueue record |
| `subscribeQueues` | Client.hs:1543-1556 | Batch subscribe rcv queues grouped by transport session |
| `subscribeQueues_` | Client.hs:1556-1720 | Subscribe batch for one session |
| `processSubResults` | Client.hs:1476-1510 | Process subscribe results: partition into failed/subscribed/notices |
| `addNewQueueSubscription` | Client.hs:1724-1728 | Add queue to active subs after creation |
| `sendConfirmation` | Client.hs:1788-1794 | Per-queue E2E encrypt confirmation → SEND via sendOrProxySMPMessage |
| `sendInvitation` | Client.hs:1796-1806 | Per-queue E2E encrypt invitation → SEND via sendOrProxySMPMessage |
| `sendAgentMessage` | Client.hs:1948-1952 | Per-queue E2E encrypt message → SEND via sendOrProxySMPMessage |
| `agentCbEncrypt` | Client.hs:2074-2082 | Per-queue E2E encrypt with stored DH secret |
| `agentCbEncryptOnce` | Client.hs:2085-2095 | Per-queue E2E encrypt with ephemeral DH (for invitations) |
| `agentCbDecrypt` | Client.hs:2099-2102 | Per-queue E2E decrypt |
| `secureQueue` | Client.hs:1830-1833 | Send KEY command |
| `secureSndQueue` | Client.hs:1835-1841 | Send SKEY command via sendOrProxySMPCommand |
| `sendAck` | Client.hs:1904-1907 | Send ACK command |
| `decryptSMPMessage` | Client.hs:1824-1828 | Decrypt received SMP message body |
| `suspendQueue` | Client.hs:1926-1929 | Send OFF command |
| `deleteQueue` | Client.hs:1931-1934 | Send DEL command |
| `deleteQueues` | Client.hs:1936-1945 | Batch DEL |
| `getQueueMessage` | Client.hs:1808-1822 | Send GET + decrypt (for polling, if needed) |
| `notifySub` / `notifySub'` | Client.hs:791-797 | Write event to subQ |
| `cryptoError` | Client.hs:2104-2115 | CryptoError → AgentErrorType |
---
## Piece 6: Agent API (top level)
| Function | Haskell location | Purpose |
|----------|-----------------|---------|
| `joinConn` | Agent.hs:1260-1263 | Top-level join: pick server, delegate |
| `joinConnSrv` (invitation) | Agent.hs:1342-1357 | Lock, startJoinInvitation, secureConfirmQueue |
| `joinConnSrv` (contact) | Agent.hs:1358-1388 | Lock, create rcv queue, sendInvitation |
| `startJoinInvitation` | Agent.hs:1270-1310 | Version check, ratchet params, snd queue, createRatchet_ |
| `compatibleInvitationUri` | Agent.hs:1321-1328 | Version range compatibility |
| `compatibleContactUri` | Agent.hs:1330-1336 | Version range compatibility |
| `secureConfirmQueue` | Agent.hs:3653-3671 | Secure + send confirmation synchronously |
| `secureConfirmQueueAsync` | Agent.hs:3645-3651 | Secure + store confirmation for async delivery |
| `agentSecureSndQueue` | Agent.hs:3673-3684 | SKEY decision logic |
| `mkAgentConfirmation` | Agent.hs:3686-3691 | Build AgentConnInfoReply with reply queue |
| `storeConfirmation` | Agent.hs:3698-3712 | Ratchet encrypt + store as SndMsg |
| `enqueueConfirmation` | Agent.hs:3693-3696 | Store + submit for delivery |
| `createReplyQueue` | Agent.hs:1415-1424 | Create rcv queue for reply |
| `newRcvConnSrv` | Agent.hs:1155-1220 | Full create-connection-with-rcv-queue |
| `createRcvQueue` | Agent.hs:972-983 | Wrapper: newRcvQueue_ + updateNewConnRcv + addNewQueueSubscription |
| `newConnToJoin` | Agent.hs:1237-1253 | Create ConnData for join, store connection |
| `newSndQueue` | Agent.hs:3769-3802 | Build SndQueue from SMPQueueInfo |
| `getNextSMPServer` | Agent.hs (via Client.hs) | Pick server for new queue, avoiding contact's server |
| `connReqQueue` | Agent.hs:1265-1268 | Extract first queue from ConnectionRequestUri |
| `versionPQSupport_` | Agent.hs:1338-1340 | PQ support based on agent+e2e versions |
| `sendMessage'` | Agent.hs:1705-1706 | Top-level send |
| `sendMessagesB_` | Agent.hs:1725-1760 | Get conn, prepare, delegate |
| `enqueueMessageB` | Agent.hs:1989-2048 | Core: updateSndIds, encode, ratchetEncryptHeader, createSndMsg, createSndMsgDelivery |
| `encodeAgentMsgStr` | Agent.hs:2050-2054 | Encode AgentMessage to bytes |
| `agentRatchetEncrypt` | Agent.hs:3742-3746 | Ratchet encrypt message body |
| `agentRatchetEncryptHeader` | Agent.hs:3748-3754 | Get encrypt key from ratchet |
| `agentRatchetDecrypt` | Agent.hs:3757-3767 | Ratchet decrypt with skipped keys |
| `submitPendingMsg` | Agent.hs:2087-2090 | Signal delivery worker |
| `getDeliveryWorker` | Agent.hs:2079-2085 | Get-or-create delivery worker for queue |
| `runSmpQueueMsgDelivery` | Agent.hs:2092-2270 | Delivery loop: getPendingQueueMsg, dispatch on msgType, send, handle errors |
| `ackMessage'` | Agent.hs:2285-2323 | ACK + delete + optional receipt |
| `ackQueueMessage` | Agent.hs:2410-2430 | Send ACK to server, handle response |
| `subscriber` | Agent.hs:2912-2919 | Read from msgQ, dispatch |
| `processSMPTransmissions` | Agent.hs:2997-3297 | Incoming message dispatcher |
| `decryptClientMessage` | Agent.hs:3282-3296 | Per-queue E2E decrypt + parse envelope |
| `agentClientMsg` | Agent.hs:3207-3225 | Ratchet decrypt, parse, store RcvMsg |
| `smpConfirmation` | Agent.hs:3298-3370 | Process received confirmation |
| `helloMsg` | Agent.hs:3372-3393 | Process HELLO |
| `smpInvitation` | Agent.hs:3515-3570 | Process received invitation |
| `allowConnection'` | Agent.hs:1427-1434 | Accept confirmation, enqueue ICAllowSecure |
| `connectReplyQueues` | Agent.hs:3630-3643 | Process reply queues from confirmation |
| `enqueueCommand` | Agent.hs:1764-1767 | Store command + start worker |
| `runCommandProcessing` | Agent.hs:1789-1902 | Async command worker loop |
| `enqueueMessage` / `enqueueMessages` | Agent.hs:~2370+ | Convenience wrappers |
| `resumeMsgDelivery` | Agent.hs:2072-2076 | Resume delivery worker for a snd queue |
| `resumeConnCmds` | Agent.hs:1773-1776 | Resume async command workers for connections |
| `resumeAllCommands` | Agent.hs:1778-1781 | Resume all pending async commands on startup |
| `enqueueSavedMessage` | Agent.hs:2056-2057 | Create delivery for additional snd queues |
| `checkMsgIntegrity` | Agent.hs:3603-3610 | Verify message sequence integrity (local fn in processSMPTransmissions) |
| `subscribeConnection'` | Agent.hs:1472-1474 | Subscribe single connection (delegates to subscribeConnections') |
| `subscribeConnections'` | Agent.hs:1488-1490 | Get conn subs from store, delegate to subscribeConnections_ |
| `subscribeConnections_` | Agent.hs:1492-1527 | Core: partition conns, resume delivery, subscribe rcv queues |
---
## Resolved decisions
1. **Store field naming**: No mapping. IDB returns snake_case (`row.conn_id`), agent code uses it directly.
2. **Error handling**: Throw custom `AgentError` exception with typed error data matching Haskell `AgentErrorType`. Catches process errors by type.
3. **subQ**: Keep ABQueue. Agent writes events, user reads. Queues prevent deadlock — without them, a callback within message processing that triggers a send could deadlock single-threaded JS.
4. **msgQ**: Keep ABQueue. WebSocket onmessage enqueues, subscriber loop dequeues. Prevents reentrancy.
5. **Structured commands in IDB**: Store as JS objects. New fields optional.
6. **Transport session key**: `(userId, server)` tuple. One WebSocket per server. No TSMEntity. All TSMEntity-specific branches dropped.
7. **Join flows**: Both invitation and contact needed. Contact for widget's primary flow (joining address). Invitation for internal group member connections.
8. **Async join**: `joinConnSrvAsync` / `secureConfirmQueueAsync` as primary path.
9. **Version ranges**: Match Haskell defaults.
10. **Config defaults**: Match Haskell defaults.
11. **`ep/conc-msgs` branch**: Ignore. Use queues.
12. **Connection type dispatch**: Compute from `conn_mode` field + queue presence. `conn_mode = "INV"` with rcv+snd queues = duplex, with only rcv = rcv, etc.
13. **`withAgentEnv`**: No-op in TS — env in closure.
14. **`getConnSubs` for subscribe**: Use `getConn` (returns conn + queues) in subscribe flow.
15. **Client notices**: Skip in `subscribeSessQueues_`.
---
## Store methods to add
These store methods are not yet in `store.ts` / `store-idb.ts` but are needed by the agent layer:
| Method | AgentStore.hs lines | Used by |
|--------|-------------------|---------|
| `createSndRatchet` | 1271-1287 | `startJoinInvitation` — stores ratchet + e2e pub keys for sending side |
| `getSndRatchet` | 1289-1300 | `startJoinInvitation` — retry path, get previously created snd ratchet |
| `updateNewConnSnd` | 424-431 | `startJoinInvitation` — add snd queue to new connection |
| `createSndConn` | 433-440 | May be needed for contact join flow |
| `setRcvQueueStatus` | already exists | — |
| `setSndQueueStatus` | already exists | — |
| `setRcvSwitchStatus` | skip (queue switching) | — |
| `setSndSwitchStatus` | skip (queue switching) | — |
---
## What to skip for MVP
| Feature | Functions to skip |
|---------|------------------|
| Queue switching | `switchConnection`, QADD/QKEY/QUSE/QTEST handlers, `switchDuplexConnection` |
| Ratchet sync | `synchronizeRatchet`, EREADY handler, `newRatchetKey` |
| Notifications | All NTF functions, `newQueueNtfSubscription` |
| File transfer | All XFTP functions |
| Remote control | All RC functions |
| Connection creation | `createConnection`, `newConn`, short links creation |
| Delivery receipts sending | `sendRcpt` in ackMessage (receiving A_RCVD is kept) |
| Multiple rcv queues | Queue replacement logic in processSMPTransmissions |
| Cleanup manager | `cleanupManager`, `deleteRcvMsgHashesExpired`, etc. |
| Server management | `setProtocolServers`, `testProtocolServer` |
| Client notices | `processClientNotices`, `subscribeClientService` |
| Statistics | All `inc*ServerStat` calls, `getAgentServersSummary` |
| Connection comparison | `compareConnections`, `syncConnections` |
## Testing strategy
### Principle: every step produces testable output
Cross-language tests from Haskell give the highest confidence because they verify wire compatibility. Pure TS tests verify internal logic. The goal is to have Haskell tests at every step where the TS code touches the network.
### Step 1 tests: pure TS (no server)
File: `tests/infra-test.ts` (run with `node`, like store-test.ts)
- **RetryInterval**: `nextRetryDelay` returns correct values for various elapsed/delay combinations. `withRetryIntervalCount` calls action with increasing delays.
- **TSessionSubs**: Full lifecycle: add pending → set session ID → add active (moves from pending) → disconnect (moves back to pending) → reconnect. Test `setSubsPending` mode logic (entity vs user session). Test batch operations.
- **SessionVar**: `getSessVar` returns Left for new, Right for existing. `removeSessVar` only removes matching ID.
### Step 2 tests: TS with store (no server)
File: `tests/worker-test.ts`
- Create AgentClient with real IndexedDB store (fake-indexeddb).
- Test worker lifecycle: create worker → signal work → worker runs → no work → worker waits → signal again.
- Test server selection: configure servers → `getNextServer` rotates avoiding used hosts.
- Test locking: `withConnLock` serializes async operations on same connId.
### Step 3 tests: agent-repl + Haskell (real SMP server)
File: `tests/agent-repl.ts` — new REPL with higher-level commands.
The agent-repl exposes mid-level operations that Haskell can drive:
```
AGENT_INIT <serverUrl> <userId>
→ creates AgentClient, connects to server
CREATE_RCV_QUEUE <connIdHex>
→ newRcvQueue on server, returns rcvId, sndId, sndQueueUri
→ Haskell can then SEND to this queue
SUBSCRIBE <connIdHex>
→ subscribeQueues for connection's rcv queue
SEND_AGENT_MSG <sndQueueHex> <msgBodyHex>
→ agentCbEncrypt + sendAgentMessage
RECV
→ wait for MSG from WebSocket, decryptSMPMessage, return parsed body
SECURE_SND <sndQueueHex>
→ secureSndQueue (SKEY)
SEND_CONFIRMATION <sndQueueHex> <confirmationHex>
→ sendConfirmation
ACK <rcvIdHex> <msgIdHex>
→ sendAck
```
Haskell test scenarios:
1. **Queue creation**: TS creates rcv queue → Haskell verifies by sending to it → TS receives
2. **Subscribe + receive**: TS subscribes → Haskell sends MSG → TS decrypts and returns
3. **Send**: TS sends agent message → Haskell receives and decrypts
4. **SKEY**: TS sends SKEY → Haskell verifies queue secured
5. **Proxy send**: TS sends via proxy → Haskell receives
These tests verify the entire session management + queue operations layer without needing the full agent handshake.
### Step 4 tests: agent-repl + Haskell (ratchet operations)
Extend agent-repl:
```
RATCHET_ENCRYPT <connIdHex> <plaintextHex>
→ agentRatchetEncrypt, return encrypted agent envelope
RATCHET_DECRYPT <connIdHex> <encryptedHex>
→ agentRatchetDecrypt, return plaintext
ENQUEUE_MSG <connIdHex> <msgBodyHex>
→ enqueueMessageB (encrypt + store + create delivery)
DELIVER
→ runSmpQueueMsgDelivery one iteration (getPendingQueueMsg + send)
```
Haskell test scenarios:
1. **Ratchet encrypt cross-language**: Initialize ratchet in both → TS encrypts → Haskell decrypts (and vice versa). Already have ratchet cross-language tests, extend to agent envelope level.
2. **Enqueue + deliver**: TS enqueues message → delivery worker sends → Haskell receives and decrypts entire agent message envelope.
3. **Receive + store**: Haskell sends agent message → TS receives, decrypts, stores RcvMsg → verify stored correctly.
### Step 5 tests: full handshake (end-to-end)
Extend agent-repl or create dedicated test:
```
JOIN <connectionRequestUri> <connInfo>
→ full joinConnection flow
ALLOW <confId> <connInfo>
→ allowConnection
SEND <connIdHex> <msgBody>
→ sendMessage
ACK_MSG <connIdHex> <msgId>
→ ackMessage
```
Haskell test scenarios:
1. **Join invitation**: Haskell creates invitation → TS joins → handshake completes (CONF, HELLO exchange) → messages flow both ways.
2. **Join contact**: Haskell creates contact address → TS joins → Haskell accepts → messages flow.
3. **Multiple connections**: TS joins two different connections on different servers simultaneously.
4. **Reconnect**: Connection established → WebSocket drops → resubscribe → messages resume.
### Test infrastructure
All cross-language tests go in `tests/SMPWebTests.hs`, extending the existing 99 tests. Each agent-repl command is a single stdin/stdout exchange (like client-repl). Haskell `callNode` drives the TS process.
Estimated test count per step:
- Step 1: ~15 TS tests (retry: 5, subs: 8, session: 2)
- Step 2: ~8 TS tests (worker: 4, server selection: 2, locking: 2)
- Step 3: ~8 Haskell tests (queue create, subscribe, send, receive, SKEY, proxy)
- Step 4: ~6 Haskell tests (ratchet encrypt/decrypt, enqueue+deliver, receive+store)
- Step 5: ~6 Haskell tests (join invitation, join contact, send/receive/ack, reconnect)
+150
View File
@@ -0,0 +1,150 @@
# SimpleX Network Protocol Specifications — Governance and Evolution (draft)
## Why this document exists
SimpleX Network protocol specifications must evolve as the network grows. This document defines how specifications change, who governs those changes, and how the history of changes is preserved.
### Lessons from the web: why ratcheted governance matters
The web's governance history demonstrates both the necessity of consortium governance and the dangers of getting the transition wrong.
[Tim Berners-Lee invented the web in 1991](https://home.cern/science/computing/birth-web/short-history-web). [Netscape took over in 1994](https://en.wikipedia.org/wiki/Netscape_Navigator), driving rapid innovation as a single company — SSL, cookies, JavaScript, and the features that made the web commercially viable. In 1994, [W3C was founded](https://www.w3.org/about/history/) as a consortium hosted across multiple independent institutions (MIT in the US, INRIA/ERCIM in Europe, Keio University in Japan, later Beihang University in China) to govern web standards.
The transition from company-led innovation to consortium governance was abrupt rather than gradual. Netscape's decline (accelerated by the [browser wars](https://en.wikipedia.org/wiki/Browser_wars) and [AOL acquisition](https://cybercultural.com/p/1999-the-fall-of-netscape-and-the-rise-of-mozilla/)) transferred control to a standards body that prioritized process over progress. The result was [a lost decade of web stagnation](https://eev.ee/blog/2020/02/01/old-css-new-css/): CSS 2.0 shipped in 1998; CSS 2.1 didn't reach Candidate Recommendation until 2004 and wasn't finalized until 2011. W3C pursued XHTML and rejected proposed enhancements to HTML, until frustrated engineers from Apple, Mozilla, and Opera formed [WHATWG in 2004](https://en.wikipedia.org/wiki/WHATWG) to build HTML5 outside W3C's process. The abrupt governance transition, without a mechanism to balance community guarantees against the imperative to continue evolving the product at pace, dramatically slowed web evolution at the time it was needed most.
Then in 2023, [W3C restructured from a multi-host consortium into a single 501(c)(3) nonprofit entity](https://www.w3.org/press-releases/2023/w3c-le-launched/) — W3C Inc, incorporated in the US. The previous structure distributed governance across four independent university hosts in different countries, making capture by any single entity structurally difficult. The new structure concentrates governance in a single legal entity with a board of directors. While presented as modernization, this effectively ended the decentralized consortium model that had protected web standards for nearly three decades.
### The governance double ratchet
SimpleX follows the same Netscape-to-consortium evolution path, but with two ratchets designed to prevent both failure modes — stagnation from premature governance transfer, and capture from governance centralization:
- **Licensing ratchet**: all contributed IP is licensed under AGPLv3 (software) and Creative Commons (documentation), perpetually and irrevocably. What is licensed cannot be unlicensed. If a Party transfers Licensed IP, the licensing obligations transfer with it.
- **Governance ratchet**: power can be given to the SimpleX Network Consortium, but never taken back. The Consortium Agreement requires majority decision of all Governing Parties for changes to the agreement itself, IP policy, and admission or removal of parties.
The ratcheted transition is historically proven to be necessary. It allows the company to continue driving rapid product innovation (as Netscape did for the web) while incrementally and irreversibly transferring governance to the consortium, without the abrupt handover that stalled web evolution or the centralization that later undermined it.
### Specification governance via the Consortium Agreement
The SimpleX Network Consortium Agreement (being deployed in 2026) establishes two levels of intellectual property governance: **Licensed IP** (all contributed protocol specifications, software, and documentation, licensed perpetually and irrevocably) and **Core IP** (the subset essential to the network, requiring consortium governance to change). The distinction between these levels and how they map to the RFC process is described in [Standard vs Core specifications](#standard-vs-core-specifications) below.
## Specification change process: protocol specifications and RFCs
Protocol knowledge lives in two places:
### `protocol/` — Consolidated specifications
Each file is a complete, self-contained description of a protocol as it exists today. Like consolidated legislation in the UK legal system: the full current law in one document, not a patchwork of amendments.
Consolidated specifications are maintained on every code change that affects protocol behavior. With LLMs, the cost of maintaining consolidated documents collapses — reworking prose to incorporate a new RFC is now inexpensive relative to the value of a single authoritative document per protocol.
Implementers read `protocol/`. They should never need to reconstruct current behavior from a base spec plus a chain of RFCs.
### `rfcs/` — Protocol evolution commits
Each RFC describes a single change to a protocol specification. RFCs are the atomic unit of protocol evolution — analogous to commits in version control, or amending acts in legislation.
An RFC is not part of the protocol specification. It becomes part of the specification only when embedded into the consolidated `protocol/` document. The RFC itself remains as a permanent historical record of what changed, when, and why.
## RFC lifecycle
```
┌——> done/ ——> standard/
draft (root) ——>──┤
└——> rejected/
```
### Draft — `rfcs/*.md`
A proposal for a protocol change. Not yet implemented. Active proposals live in the `rfcs/` root directory.
Named by proposal date: `YYYY-MM-DD-topic.md`.
A draft may be rejected if the proposal is considered but not accepted for implementation.
### Done — `rfcs/done/`
Implemented in code. The protocol change described by this RFC exists in the codebase, but the RFC has not yet been verified against the actual implementation (code may have diverged from the proposal during implementation).
### Standard — `rfcs/standard/`
Verified against the actual implementation and synchronized with code. The RFC accurately describes what was implemented. This is a permanent historical record — standard RFCs are never modified or removed.
On promotion to standard, the RFC is:
1. Renamed from proposal date to standardization date: `YYYY-MM-DD-topic.md` (new date, same topic slug)
2. Updated with a document history header capturing the full lifecycle
3. Embedded into the corresponding `protocol/` consolidated specification
The `protocol/` document references embedded RFCs by name (e.g., "Private message routing added by RFC 2023-09-12-second-relays, standardized 2026-XX-XX"), similar to UK legislation citing the amending act for each clause.
Protocol version numbers make it clear which RFCs are included in which protocol revision — no separate tracking is needed.
### Rejected — `rfcs/rejected/`
Draft proposals that were considered but not accepted for implementation. Only drafts move to rejected — once an RFC is implemented (done/), it proceeds to standard/ after verification. Preserved for historical record of design decisions.
### Document history header
Every RFC in `standard/` carries a history header:
```
---
Proposed: YYYY-MM-DD
Implemented: YYYY-MM-DD
Standardized: YYYY-MM-DD
Protocol: simplex-messaging v9 (or whichever protocol this amends)
---
```
## Governance
SimpleX Network follows the Netscape-to-W3C evolution path, with ratcheted rather than abrupt transitions:
| Phase | Period | Governance | Development process |
|-------|--------|-----------|-------------------|
| Protocol invented | 2020 | Two people | Prototype developed |
| SimpleX Chat Ltd | 2022 | One company | Product-first: code leads, specs follow |
| SimpleX Network Consortium | 2026 | Agreement of SimpleX Chat Ltd and non-profit entities | Product-first for standard; standards-first for core |
| Decentralized governance | Future | TBD (DAO research ongoing) | Standards-first |
### Current: product-first development
SimpleX protocols currently follow a product-first development process: requirements drive code, code drives specification. RFCs are written as design proposals before implementation, but implementation details are figured out in code. Consolidated protocol specifications in `protocol/` are then amended to match the implementation.
This process is governed by SimpleX Chat Ltd as the IP Holding Party under the Consortium Agreement.
Any Specification Author (as defined in the Consortium Agreement) may propose RFCs. Acceptance and standardization decisions are made by SimpleX Chat Ltd during the current product-first phase.
### Standard vs Core specifications
The distinction between standard and core maps directly to the two levels of IP governance in the Consortium Agreement, and reflects the difference between product-first and standards-first development:
**Standard** — Licensed IP, not yet under consortium governance. Governed by the company.
All contributed protocol specifications are Licensed IP under the Consortium Agreement. Standard specifications follow product-first development: the company can evolve them with product needs, and they must be maintained on every code change that affects protocol behavior.
Standard specifications live in `rfcs/standard/` and `protocol/`.
**Core** — Governed IP, governed by the consortium.
A subset of standard specifications will be designated as Core IP under the Consortium Agreement. Core specifications will follow standards-first development: specification changes must be agreed via Governing Decision before code changes.
This is a legally binding commitment. Once Licensed IP is included in Core IP, the company that owns the code cannot unilaterally change it — even though they own the code, the Consortium Agreement requires a Governing Decision for any change to Core IP. This protects the fundamental properties of the network (privacy, security, decentralization) from unilateral modification by any single party.
The designation of specific specifications as Core IP is itself a Governing Decision that requires Consortium vote. The transition will happen incrementally as protocols stabilize — the governance ratchet ensures that each designation is irreversible.
The exact mechanism for distinguishing core from standard within the RFC and protocol folder structure is TBD — it will be decided as the first protocols are designated as Core IP.
### Future: standards-first development
As more protocols are designated as Core IP, development naturally transitions to a standards-first process for a growing portion of the protocol suite. The governance ratchet ensures this transition is gradual and irreversible — each protocol that becomes core gains the protection of consortium governance permanently, while remaining standard protocols continue to evolve at product pace.
## Current state
| Location | Contents | Count |
|----------|----------|-------|
| `protocol/` | Consolidated specs (SMP v9, Agent v5, XFTP v2, XRCP v1, Push v2, PQDR v1) | 6 specs + overview |
| `rfcs/` root | Active draft proposals | 19 |
| `rfcs/done/` | Implemented, not yet verified | 25 |
| `rfcs/standard/` | Verified against implementation | (to be populated) |
| `rfcs/rejected/` | Draft proposals not accepted | 7 |
+6 -6
View File
@@ -67,14 +67,14 @@ if [ ! -f "${confd}/smp-server.ini" ]; then
# Fix path to certificates
if [ -n "${WEB_MANUAL}" ]; then
sed -i -e 's|^[^#]*https: |#&|' \
-e 's|^[^#]*cert: |#&|' \
-e 's|^[^#]*key: |#&|' \
-e 's|^port:.*|port: 5223|' \
sed -i -e 's|^[^#]*https = |#&|' \
-e 's|^[^#]*cert = |#&|' \
-e 's|^[^#]*key = |#&|' \
-e 's|^port = .*|port = 5223|' \
"${confd}/smp-server.ini"
else
sed -i -e "s|cert: /etc/opt/simplex/web.crt|cert: $cert_path/$ADDR.crt|" \
-e "s|key: /etc/opt/simplex/web.key|key: $cert_path/$ADDR.key|" \
sed -i -e "s|cert = /etc/opt/simplex/web.crt|cert = $cert_path/$ADDR.crt|" \
-e "s|key = /etc/opt/simplex/web.key|key = $cert_path/$ADDR.key|" \
"${confd}/smp-server.ini"
fi
fi
+1 -1
View File
@@ -76,7 +76,7 @@ if [ ! -f "${confd}/file-server.ini" ]; then
# Optionally, set password
if [ -n "${PASS}" ]; then
sed -i -e "/^# create_password:/a create_password: $PASS" \
sed -i -e "/^# create_password =/a create_password = $PASS" \
"${confd}/file-server.ini"
fi
fi
+512 -169
View File
@@ -1,13 +1,8 @@
#!/usr/bin/env sh
set -eu
# Links to scripts/configs
scripts="https://raw.githubusercontent.com/simplex-chat/simplexmq/stable/scripts/main"
scripts_systemd_smp="$scripts/smp-server.service"
scripts_systemd_xftp="$scripts/xftp-server.service"
scripts_update="$scripts/simplex-servers-update"
scripts_uninstall="$scripts/simplex-servers-uninstall"
scripts_stopscript="$scripts/simplex-servers-stopscript"
# Make sure that PATH variable contains /usr/local/bin
PATH="/usr/local/bin:$PATH"
# Default installation paths
path_bin="/usr/local/bin"
@@ -22,7 +17,9 @@ path_systemd_smp="$path_systemd/smp-server.service"
path_systemd_xftp="$path_systemd/xftp-server.service"
# Temporary paths
path_tmp_bin="$(mktemp -d)"
path_tmp_bin="/tmp/simplex-servers"
path_tmp_bin_smp="$path_tmp_bin/smp-server"
path_tmp_bin_xftp="$path_tmp_bin/xftp-server"
path_tmp_bin_update="$path_tmp_bin/simplex-servers-update"
path_tmp_bin_uninstall="$path_tmp_bin/simplex-servers-uninstall"
path_tmp_bin_stopscript="$path_tmp_bin/simplex-servers-stopscript"
@@ -37,203 +34,549 @@ BLU='\033[1;36m'
YLW='\033[1;33m'
RED='\033[0;31m'
NC='\033[0m'
BLD='\033[1m'
UNDRL='\033[4m'
NL='
'
# Set VER globally and only once
VER="${VER:-latest}"
# Currently, XFTP default to v0.1.0, so it doesn't make sense to check its version
os_test() {
. /etc/os-release
######################
### Misc functions ###
######################
case "$VERSION_ID" in
20.04|22.04) : ;;
24.04) VERSION_ID='22.04' ;;
*) printf "${RED}Unsupported Ubuntu version!${NC}\nPlease file Github issue with request to support Ubuntu %s: https://github.com/simplex-chat/simplexmq/issues/new\n" "$VERSION_ID" && exit 1 ;;
esac
# Checks "sanity" of downloaded thing, e.g. if it's really a script or binary
check_sanity() {
path="$1"
criteria="$2"
version="$(printf '%s' "$VERSION_ID" | tr '.' '_')"
arch="$(uname -p)"
case "$criteria" in
string:*)
pattern="$(printf '%s' "$criteria" | awk '{print $2}')"
if grep -q "$pattern" "$path"; then
sane=0
else
sane=1
fi
;;
file:*)
pattern="$(printf '%s' "$criteria" | awk '{print $2}')"
case "$arch" in
x86_64) arch="$(printf '%s' "$arch" | tr '_' '-')" ;;
*) printf "${RED}Unsupported architecture!${NC}\nPlease file Github issue with request to support %s architecture: https://github.com/simplex-chat/simplexmq/issues/new" "$arch" && exit 1 ;;
esac
if file "$path" | grep -q "$pattern"; then
sane=0
else
sane=1
fi
;;
*) printf 'Unknown criteria.\n'; sane=1 ;;
esac
bin_smp="$bin/smp-server-ubuntu-${version}-${arch}"
bin_xftp="$bin/xftp-server-ubuntu-${version}-${arch}"
unset path string
return "$sane"
}
installed_test() {
# Checks if old thing and new thing is different
change_check() {
old="$1"
new="$2"
if [ -x "$new" ] || [ -f "$new" ]; then
type="$(file $new)"
else
type='string'
fi
case "$type" in
*script*|*text*)
if diff -q "$old" "$new" > /dev/null 2>&1; then
changed=1
else
changed=0
fi
;;
string)
if [ "$old" = "$new" ]; then
changed=1
else
changed=0
fi
;;
esac
return "$changed"
}
##########################
### Misc functions END ###
##########################
#########################
### Support functions ###
#########################
# Sets local/remote versions and "apps" variables
check_versions() {
# Sets:
# - ver
# - bin_url
# - remote_version
# - local_version
# - apps
case "$VER" in
latest)
remote_version="$(curl --proto '=https' --tlsv1.2 -sSf -L https://api.github.com/repos/simplex-chat/simplexmq/releases/latest 2>/dev/null | grep -i "tag_name" | awk -F \" '{print $4}')"
if [ -z "$remote_version" ]; then
printf "${RED}Something went wrong when ${YLW}resolving the lastest version${NC}: either you don't have connection to Github or you're rate-limited.\n"
exit 1
fi
;;
*)
# Check if this version really exist
ver_check="https://github.com/simplex-chat/simplexmq/releases/tag/${VER}"
if curl -o /dev/null --proto '=https' --tlsv1.2 -sf -L "${ver_check}"; then
remote_version="${VER}"
else
printf "Provided version ${BLU}%s${NC} ${RED}doesn't exist${NC}! Switching to ${BLU}latest${NC}.\n" "${VER}"
VER='latest'
# Re-execute check
check_versions
# Everything has been done, so return from the function
return 0
fi
;;
esac
# Links to scripts/configs
bin_url="https://github.com/simplex-chat/simplexmq/releases/download/${remote_version}"
scripts_url="https://raw.githubusercontent.com/simplex-chat/simplexmq/refs/tags/${remote_version}/scripts/main"
scripts_url_systemd_smp="$scripts_url/smp-server.service"
scripts_url_systemd_xftp="$scripts_url/xftp-server.service"
scripts_url_update="$scripts_url/simplex-servers-update"
scripts_url_uninstall="$scripts_url/simplex-servers-uninstall"
scripts_url_stopscript="$scripts_url/simplex-servers-stopscript"
set +u
for i in $path_conf_etc/*; do
if [ -d "$i" ]; then
case "$i" in
*simplex) apps="smp $apps" ;;
*simplex-xftp) apps="xftp $apps" ;;
esac
for i in smp xftp; do
# Only check local directory where binaries are installed by the script
if command -v "/usr/local/bin/$i-server" >/dev/null; then
apps="$i $apps"
fi
done
set -u
}
set_version() {
ver="${VER:-latest}"
case "$ver" in
latest)
bin="https://github.com/simplex-chat/simplexmq/releases/latest/download"
remote_version="$(curl --proto '=https' --tlsv1.2 -sSf -L https://api.github.com/repos/simplex-chat/simplexmq/releases/latest | grep -i "tag_name" | awk -F \" '{print $4}')"
;;
*)
bin="https://github.com/simplex-chat/simplexmq/releases/download/${ver}"
remote_version="${ver}"
;;
esac
}
update_scripts() {
curl --proto '=https' --tlsv1.2 -sSf -L "$scripts_update" -o "$path_tmp_bin_update" && chmod +x "$path_tmp_bin_update"
curl --proto '=https' --tlsv1.2 -sSf -L "$scripts_uninstall" -o "$path_tmp_bin_uninstall" && chmod +x "$path_tmp_bin_uninstall"
curl --proto '=https' --tlsv1.2 -sSf -L "$scripts_stopscript" -o "$path_tmp_bin_stopscript" && chmod +x "$path_tmp_bin_stopscript"
if diff -q "$path_bin_uninstall" "$path_tmp_bin_uninstall" > /dev/null 2>&1; then
printf -- "- ${YLW}Uninstall script is up-to-date${NC}.\n"
rm "$path_tmp_bin_uninstall"
else
printf -- "- Updating uninstall script..."
mv "$path_tmp_bin_uninstall" "$path_bin_uninstall"
printf "${GRN}Done!${NC}\n"
if [ -z "$apps" ]; then
printf "${RED}No simplex servers installed! Aborting.${NC}\n"
exit 1
fi
if diff -q "$path_bin_stopscript" "$path_tmp_bin_stopscript" > /dev/null 2>&1; then
printf -- "- ${YLW}Stopscript script is up-to-date${NC}.\n"
rm "$path_tmp_bin_stopscript"
else
printf -- "- Updating stopscript script..."
mv "$path_tmp_bin_stopscript" "$path_bin_stopscript"
printf "${GRN}Done!${NC}\n"
fi
for server in $apps; do
# Check if info file is present
if [ -f "$path_conf_info/release" ]; then
# If present, source it
. "$path_conf_info/release" 2>/dev/null
if diff -q "$path_bin_update" "$path_tmp_bin_update" > /dev/null 2>&1; then
printf -- "- ${YLW}Update script is up-to-date${NC}.\n"
rm "$path_tmp_bin_update"
else
printf -- "- Updating update script..."
mv "$path_tmp_bin_update" "$path_bin_update"
printf "${GRN}Done!${NC}\n"
printf -- "- Re-executing Update script with latest updates..."
exec sh "$path_bin_update" "continue"
fi
}
update_systemd() {
service="${1}-server"
eval "scripts_systemd=\$scripts_systemd_${1}"
eval "path_systemd=\$path_systemd_${1}"
eval "path_tmp_systemd=\$path_tmp_systemd_${1}"
curl --proto '=https' --tlsv1.2 -sSf -L "$scripts_systemd" -o "$path_tmp_systemd"
if diff -q "$path_systemd" "$path_tmp_systemd" > /dev/null 2>&1; then
printf -- "- ${YLW}%s service is up-to-date${NC}.\n" "$service"
rm "$path_tmp_systemd"
else
printf -- "- Updating %s service..." "$service"
mv "$path_tmp_systemd" "$path_systemd"
systemctl daemon-reload
printf "${GRN}Done!${NC}\n"
fi
unset service scripts_systemd path_systemd path_tmp_systemd
}
update_bins() {
service="${1}-server"
eval "bin=\$bin_${1}"
eval "path_bin=\$path_bin_${1}"
set_ver() {
local_version='unset'
sed -i -- "s/local_version_${1}=.*/local_version_${1}='${remote_version}'/" "$path_conf_info/release"
}
if [ -f "$path_conf_info/release" ]; then
. "$path_conf_info/release" 2>/dev/null
set +u
eval "local_version=\$local_version_${1}"
set -u
if [ -z "${local_version}" ]; then
set_ver "$1"
fi
else
printf 'local_version_xftp=\nlocal_version_smp=\n' > "$path_conf_info/release"
set_ver "$1"
fi
if [ "$local_version" != "$remote_version" ]; then
if systemctl is-active --quiet "$service"; then
printf -- "- Stopping %s service..." "$service"
systemctl stop "$service"
printf "${GRN}Done!${NC}\n"
printf -- "- Updating %s to %s..." "$service" "$remote_version"
curl --proto '=https' --tlsv1.2 -sSf -L "$bin" -o "$path_bin" && chmod +x "$path_bin"
printf "${GRN}Done!${NC}\n"
printf -- "- Starting %s service..." "$service"
systemctl start "$service"
printf "${GRN}Done!${NC}\n"
# Check if line containing local version exists in file
if grep -q "local_version_${server}" "$path_conf_info/release"; then
# if exists, set the var
eval "local_version=\$local_version_${server}"
else
# If it doesn't, append it to file
printf "local_version_${server}=unset\n" >> "$path_conf_info/release"
# And set it in script (so we don't have to re-source the file)
eval "local_version_${server}=unset"
fi
else
printf -- "- Updating %s to %s..." "$service" "$remote_version"
curl --proto '=https' --tlsv1.2 -sSf -L "$bin" -o "$path_bin" && chmod +x "$path_bin"
printf "${GRN}Done!${NC}\n"
# If there isn't info file, populate it
printf "local_version_${server}=unset\n" >> "$path_conf_info/release"
fi
else
printf -- "- ${YLW}%s is up-to-date${NC}.\n" "$service"
fi
done
set_ver "$1"
unset service bin path_bin local_version
# Return
return 0
}
# Checks the distro and sets the urls variables
check_distro() {
. /etc/os-release
case "$VERSION_ID" in
20.04|22.04) : ;;
24.04) VERSION_ID='22.04' ;;
*) printf "${RED}Unsupported Ubuntu version!${NC}\nPlease file Github issue with request to support Ubuntu %s: https://github.com/simplex-chat/simplexmq/issues/new\n" "$VERSION_ID" && exit 1 ;;
esac
version="$(printf '%s' "$VERSION_ID" | tr '.' '_')"
arch="$(uname -p)"
case "$arch" in
x86_64) arch="$(printf '%s' "$arch" | tr '_' '-')" ;;
*) printf "${RED}Unsupported architecture!${NC}\nPlease file Github issue with request to support %s architecture: https://github.com/simplex-chat/simplexmq/issues/new" "$arch" && exit 1 ;;
esac
bin_url_smp="$bin_url/smp-server-ubuntu-${version}-${arch}"
bin_url_xftp="$bin_url/xftp-server-ubuntu-${version}-${arch}"
return 0
}
# General checks that must be performed on the initial execution of script
checks() {
if [ "$(id -u)" -ne 0 ]; then
printf "This script is intended to be run with root privileges. Please re-run script using sudo.\n"
exit 1
fi
set_version
os_test
installed_test
check_versions
check_distro
mkdir -p $path_conf_info
mkdir -p $path_conf_info $path_tmp_bin
return 0
}
main() {
checks
#############################
### Support functions END ###
#############################
set +u
if [ "$1" != "continue" ]; then
set -u
printf "Updating scripts...\n"
update_scripts
else
set -u
######################
### Main functions ###
######################
# Downloads thing to directory and checks its sanity
download_thing() {
thing="$1"
path="$2"
check_pattern="$3"
err_msg="$4"
if ! curl --proto '=https' --tlsv1.2 -sSf -L "$thing" -o "$path"; then
printf "${RED}Something went wrong when downloading ${YLW}%s${NC}: either you don't have connection to Github or you're rate-limited.\n" "$err_msg"
exit 1
fi
type="$(file "$path")"
case "$type" in
*script*|*executable*) chmod +x "$path" ;;
esac
if ! check_sanity "$path" "$check_pattern"; then
printf "${RED}Something went wrong with downloaded ${YLW}%s${NC}: file is corrupted.\n" "$err_msg"
exit 1
fi
return 0
}
# Downloads all necessary files to temp dir and set update messages for the menu
download_all() {
download_thing "$scripts_url_update" "$path_tmp_bin_update" 'string: /usr/bin/env' 'Update script'
if change_check "$path_tmp_bin_update" "$path_bin_update"; then
msg_scripts="${msg_scripts+$msg_scripts, }${YLW}simplex-servers-update${NC}"
msg_scripts_raw="${msg_scripts_raw+$msg_scripts_raw/}update"
fi
download_thing "$scripts_url_stopscript" "$path_tmp_bin_stopscript" 'string: /usr/bin/env' 'Stop script'
if change_check "$path_tmp_bin_stopscript" "$path_bin_stopscript"; then
msg_scripts="${msg_scripts+$msg_scripts, }${YLW}simplex-servers-stopscript${NC}"
msg_scripts_raw="${msg_scripts_raw+$msg_scripts_raw/}stop"
fi
download_thing "$scripts_url_uninstall" "$path_tmp_bin_uninstall" 'string: /usr/bin/env' 'Uninstall script'
if change_check "$path_tmp_bin_uninstall" "$path_bin_uninstall"; then
msg_scripts="${msg_scripts+$msg_scripts, }${YLW}simplex-servers-uninstall${NC}"
msg_scripts_raw="${msg_scripts_raw+$msg_scripts_raw/}uninstall"
fi
for i in $apps; do
service="${i}-server"
eval "scripts_url_systemd_final=\$scripts_url_systemd_${i}"
eval "path_tmp_systemd_final=\$path_tmp_systemd_${i}"
eval "path_systemd_final=\$path_systemd_${i}"
download_thing "$scripts_url_systemd_final" "$path_tmp_systemd_final" 'string: [Unit]' "$service systemd service"
if change_check "$path_tmp_systemd_final" "$path_systemd_final"; then
msg_services="${msg_services+$msg_services, }${YLW}$service.service${NC}"
msg_services_raw="${msg_services_raw+$msg_services_raw/}$service"
fi
done
for i in $apps; do
service="${i}-server"
eval "local_version=\$local_version_${i}"
if change_check "$local_version" "$remote_version"; then
msg_bins="${msg_bins+$msg_bins$NL} - ${YLW}$service${NC}: from ${BLU}$local_version${NC} to ${BLU}$remote_version${NC}"
msg_bins_alt="${msg_bins_alt+$msg_bins_alt, }${YLW}$service${NC}"
msg_bins_raw="${msg_bins_raw+$msg_bins_raw/}$service"
fi
done
return 0
}
# Updates systemd and scripts. This function depends om variables from "download_all"
update_misc() {
OLD_IFS="$IFS"
IFS='/'
for script in ${msg_scripts_raw:-}; do
case "$script" in
update)
printf -- "- Updating update script..."
mv "$path_tmp_bin_update" "$path_bin_update"
printf "${GRN}Done!${NC}\n"
printf -- "- Re-executing Update script..."
exec env UPDATE_SCRIPT_DONE=1 VER="$remote_version" "$path_bin_update" "${selection}"
;;
stop)
printf -- "- Updating stopscript script..."
mv "$path_tmp_bin_stopscript" "$path_bin_stopscript"
printf "${GRN}Done!${NC}\n"
;;
uninstall)
printf -- "- Updating uninstall script..."
mv "$path_tmp_bin_uninstall" "$path_bin_uninstall"
printf "${GRN}Done!${NC}\n"
;;
esac
done
for service in ${msg_services_raw:-}; do
app="${service%%-*}"
eval "path_systemd=\$path_systemd_${app}"
eval "path_tmp_systemd=\$path_tmp_systemd_${app}"
printf -- "- Updating %s service..." "$service"
mv "$path_tmp_systemd" "$path_systemd"
systemctl daemon-reload
printf "${GRN}Done!${NC}\n"
done
IFS="$OLD_IFS"
return 0
}
# Updates binaries. This function depends on variables from "download_all"
update_bins() {
OLD_IFS="$IFS"
IFS='/'
for service in ${msg_bins_raw:-}; do
app="${service%%-*}"
eval "local_version=\$local_version_${app}"
eval "bin_url_final=\$bin_url_${app}"
eval "path_tmp_bin_final=\$path_tmp_bin_${app}"
eval "path_bin_final=\$path_bin_${app}"
# If systemd service is active
if systemctl is-active --quiet "$service"; then
printf -- "- Stopping %s service..." "$service"
systemctl stop "$service"
printf "${GRN}Done!${NC}\n"
printf -- "- Updating ${YLW}%s${NC} from ${BLU}%s${NC} to ${BLU}%s${NC}..." "$service" "$local_version" "$remote_version"
download_thing "$bin_url_final" "$path_tmp_bin_final" 'file: ELF' "$service"
mv "$path_tmp_bin_final" "$path_bin_final"
printf "${GRN}Done!${NC}\n"
printf -- "- Starting %s service..." "$service"
systemctl start "$service"
printf "${GRN}Done!${NC}\n"
else
# If systemd service is NOT active
printf -- "- Updating ${YLW}%s${NC} from ${BLU}%s${NC} to ${BLU}%s${NC}..." "$service" "$local_version" "$remote_version"
download_thing "$bin_url_final" "$path_tmp_bin_final" 'file: ELF' "$service"
mv "$path_tmp_bin_final" "$path_bin_final"
printf "${GRN}Done!${NC}\n"
fi
# Don't forget to set version
sed -i -- "s|local_version_${app}=.*|local_version_${app}='${remote_version}'|" "$path_conf_info/release"
done
IFS="$OLD_IFS"
return 0
}
# Just download binaries
download_bins() {
OLD_IFS="$IFS"
IFS='/'
for service in ${msg_bins_raw:-}; do
app="${service%%-*}"
eval "local_version=\$local_version_${app}"
eval "bin_url_final=\$bin_url_${app}"
eval "path_tmp_bin_final=\$path_tmp_bin_${app}"
eval "path_bin_final=\$path_bin_${app}"
printf -- "- Downloading ${YLW}%s${NC} binary..." "$service"
download_thing "$bin_url_final" "$path_tmp_bin_final" 'file: ELF' "$service"
printf "${GRN}Done!${NC}\n"
done
IFS="$OLD_IFS"
return 0
}
menu_init_help() {
menu_help="Update script for SimpleX servers and scripts.${NL}${NL}"
menu_help="${menu_help}${BLD}${UNDRL}Usage:${NC} [<VARIABLE>] ${BLD}simplex-servers-update${NC}${NL} [<VARIABLE>] ${BLD}simplex-servers-update${NC} [<SUBCOMMAND>]${NL}${NL}"
menu_help="${menu_help}${BLD}${UNDRL}Subcommands:${NC}${NL}"
menu_help_sub=" ${BLD}[a]ll${NC} Update everything without confirmation${NL}"
menu_help_sub="${menu_help_sub} ${BLD}[b]inaries${NC} Update binaries only without confirmation${NL}"
menu_help_sub="${menu_help_sub} ${BLD}[d]ownload${NC} Download everything without updating${NL}"
menu_help_sub="${menu_help_sub} ${BLD}[h]elp${NC} Print this message${NL}${NL}"
menu_help="${menu_help}${menu_help_sub}"
menu_help="${menu_help}${BLD}${UNDRL}Variables:${NC}${NL}"
menu_help="${menu_help} ${BLD}VER=v3.2.1-beta.0${NC} Update binaries to specified version${NL}"
return 0
}
menu_init() {
menu_end="${RED}x${NC}) Exit${NL}${NL}Selection: "
menu_option_download="${GRN}d${NC}) Download files only${NL}"
if [ -n "${msg_scripts:-}" ]; then
menu_option_misc_raw="${menu_option_misc_raw+${menu_option_misc_raw}${NL}} - script(s): ${msg_scripts}"
fi
if [ -n "${msg_services:-}" ]; then
menu_option_misc_raw="${menu_option_misc_raw+${menu_option_misc_raw}${NL}} - systemd service file(s): ${msg_services}"
fi
menu_option_all="${GRN}a${NC}) Update all: ${BLU}(recommended)${NC}${NL}${menu_option_misc_raw+${menu_option_misc_raw}${NL}}${msg_bins+${msg_bins}${NL}}"
if [ -n "${msg_bins:-}" ]; then
menu_option_bins="${GRN}b${NC}) Update server binaries: ${msg_bins_alt}${NL}"
fi
# Abort early if there's neither update binaries, nor update scripts options
if [ -z "${menu_option_bins:-}" ] && [ -z "${menu_option_misc_raw:-}" ]; then
printf "${YLW}Everything is up-to-date${NC}.\n"
exit 0
fi
menu="${menu_option_all}${menu_option_bins:-}${menu_option_download}${menu_end}"
return 0
}
options_parse() {
selection="$1"
case "$selection" in
a|all)
check=0
if [ -z "${menu_option_misc_raw:-}" ] && [ -z "${menu_option_bins:-}" ]; then
printf "${YLW}Everything is up-to-date${NC}.\n"
else
if [ -n "${menu_option_misc_raw:-}" ]; then
update_misc
fi
if [ -n "${menu_option_bins:-}" ]; then
update_bins
fi
fi
;;
b|binaries)
check=0
if [ -n "${menu_option_bins:-}" ]; then
update_bins
else
printf "${YLW}Binaries is up-to-date${NC}.\n"
fi
;;
d|download)
check=0
if [ -n "${menu_option_bins:-}" ]; then
download_bins
fi
printf "\n${YLW}Scripts${NC}/${YLW}services${NC}/${YLW}binaries${NC} has been downloaded to ${BLU}%s${NC}\n" "$path_tmp_bin"
;;
x)
check=0
;;
*)
check=1
;;
esac
return "$check"
}
##########################
### Main functions END ###
##########################
############
### Init ###
############
main() {
# Early hook to print Done after script re-execution
if [ -n "${UPDATE_SCRIPT_DONE:-}" ]; then
printf "${GRN}Done!${NC}\n"
fi
printf "Updating systemd services...\n"
for i in $apps; do
update_systemd "$i"
done
printf "Updating simplex servers...\n"
for i in $apps; do
update_bins "$i"
done
# Early help menu
menu_init_help
rm -rf "$path_tmp_bin"
case "${1:-}" in
h|help)
printf '%b' "$menu_help"
exit 0
;;
esac
checks
download_all
menu_init
onetime=0
while true; do
if [ "$onetime" = 0 ]; then
onetime=1
if [ -n "${1:-}" ]; then
selection="$1"
else
printf '%b' "$menu"
read selection
fi
else
read selection
fi
if options_parse "$selection"; then
break
else
# Rerender whole menu if the first non-interactive option was bogus
if [ -n "${1:-}" ]; then
onetime=0
shift 1
else
# Erase last line
printf '\e[A\e[K'
# Only rerended selection
printf 'Selection: '
fi
fi
done
}
main "$@"
+139
View File
@@ -0,0 +1,139 @@
#!/usr/bin/env sh
set -eu
TAG="$1"
tempdir="$(mktemp -d)"
init_dir="$PWD"
repo_name="simplexmq"
repo="https://github.com/simplex-chat/${repo_name}"
export DOCKER_BUILDKIT=1
cleanup() {
docker exec -t builder sh -c 'rm -rf ./dist-newstyle' 2>/dev/null || :
rm -rf -- "$tempdir"
docker rm --force builder 2>/dev/null || :
docker image rm local 2>/dev/null || :
cd "$init_dir"
}
trap 'cleanup' EXIT INT
mkdir -p "$init_dir/$TAG-$repo_name/from-source" "$init_dir/$TAG-$repo_name/prebuilt"
git -C "$tempdir" clone "$repo.git" &&\
cd "$tempdir/${repo_name}" &&\
git checkout "$TAG"
for os in 22.04 24.04; do
os_url="$(printf '%s' "$os" | tr '.' '_')"
# Build image
docker build \
--no-cache \
--build-arg TAG=${os} \
--build-arg GHC=9.6.3 \
-f "$tempdir/simplexmq/Dockerfile.build" \
-t local \
.
# Run container in background
docker run -t -d \
--name builder \
-v "$tempdir/${repo_name}:/project" \
local
# PostgreSQL build (only smp-server)
docker exec \
-t \
builder \
sh -c 'git config --global --add safe.directory \*; cabal update && cabal build --jobs=$(nproc) --enable-tests -fserver_postgres && mkdir -p /out && for i in smp-server simplexmq-test; do bin=$(find /project/dist-newstyle -name "$i" -type f -executable) && chmod +x "$bin" && mv "$bin" /out/; done && strip /out/smp-server'
# Copy smp-server postgresql binary and prepare it
docker cp \
builder:/out/smp-server \
"$init_dir/$TAG-$repo_name/from-source/smp-server-postgres-ubuntu-${os_url}-x86-64"
# Download prebuilt postgresql binary
curl -L \
--output-dir "$init_dir/$TAG-$repo_name/prebuilt/" \
-O \
"$repo/releases/download/${TAG}/smp-server-postgres-ubuntu-${os_url}-x86-64"
# Regular build (all)
apps='smp-server xftp-server ntf-server xftp'
docker exec \
-t \
-e apps="$apps" \
builder \
sh -c 'cabal build --jobs=$(nproc) && mkdir -p /out && for i in $apps; do bin=$(find /project/dist-newstyle -name "$i" -type f -executable) && strip "$bin" && chmod +x "$bin" && mv "$bin" /out/; done'
# Copy regular binaries
docker cp \
builder:/out \
out-${os}
# Prepare regular binaries and download the prebuilt ones
for app in $apps; do
curl -L \
--output-dir "$init_dir/$TAG-$repo_name/prebuilt/" \
-O \
"$repo/releases/download/${TAG}/${app}-ubuntu-${os_url}-x86-64"
mv "./out-${os}/$app" "$init_dir/$TAG-$repo_name/from-source/${app}-ubuntu-${os_url}-x86-64"
done
# Important! Remove dist-newstyle for the next interation
docker exec \
-t \
builder \
sh -c 'rm -rf ./dist-newstyle'
# Also restore git to previous state
git reset --hard && git clean -dfx
# Stop containers, delete images
docker stop builder
docker rm --force builder
docker image rm local
done
# Cleanup
rm -rf -- "$tempdir"
cd "$init_dir"
# Final stage: compare hashes
# Path to binaries
path_bin="$init_dir/$TAG-$repo_name"
# Assume everything is okay for now
bad=0
# Check hashes for all binaries
for file in "$path_bin"/from-source/*; do
# Extract binary name
app="$(basename $file)"
# Compute hash for compiled binary
compiled=$(sha256sum "$path_bin/from-source/$app" | awk '{print $1}')
# Compute hash for prebuilt binary
prebuilt=$(sha256sum "$path_bin/prebuilt/$app" | awk '{print $1}')
# Compare
if [ "$compiled" != "$prebuilt" ]; then
# If hashes doesn't match, set bad...
bad=1
# ... and print affected binary
printf "%s - sha256sum hash doesn't match\n" "$app"
fi
done
# If everything is still okay, compute checksums file
if [ "$bad" = 0 ]; then
sha256sum "$path_bin"/from-source/* | sed -e "s|$PWD/||g" -e 's|from-source/||g' -e "s|-$repo_name||g" > "$path_bin/_sha256sums"
printf 'Checksums computed - %s\n' "$path_bin/_sha256sums"
fi
+170 -82
View File
@@ -1,7 +1,7 @@
cabal-version: 1.12
name: simplexmq
version: 6.3.0.3
version: 6.5.1.0
synopsis: SimpleXMQ message broker
description: This package includes <./docs/Simplex-Messaging-Server.html server>,
<./docs/Simplex-Messaging-Client.html client> and
@@ -24,33 +24,38 @@ extra-source-files:
CHANGELOG.md
cbits/sha512.h
cbits/sntrup761.h
apps/smp-server/static/index.html
apps/smp-server/static/link.html
apps/smp-server/static/media/apk_icon.png
apps/smp-server/static/media/apple_store.svg
apps/smp-server/static/media/contact.js
apps/smp-server/static/media/contact_page_mobile.png
apps/smp-server/static/media/f_droid.svg
apps/smp-server/static/media/favicon.ico
apps/smp-server/static/media/GilroyBold.woff2
apps/smp-server/static/media/GilroyLight.woff2
apps/smp-server/static/media/GilroyMedium.woff2
apps/smp-server/static/media/GilroyRegular.woff2
apps/smp-server/static/media/GilroyRegularItalic.woff2
apps/smp-server/static/media/google_play.svg
apps/smp-server/static/media/logo-dark.png
apps/smp-server/static/media/logo-light.png
apps/smp-server/static/media/logo-symbol-dark.svg
apps/smp-server/static/media/logo-symbol-light.svg
apps/smp-server/static/media/moon.svg
apps/smp-server/static/media/qrcode.js
apps/smp-server/static/media/script.js
apps/smp-server/static/media/style.css
apps/smp-server/static/media/sun.svg
apps/smp-server/static/media/swiper-bundle.min.css
apps/smp-server/static/media/swiper-bundle.min.js
apps/smp-server/static/media/tailwind.css
apps/smp-server/static/media/testflight.png
apps/common/Web/static/index.html
apps/common/Web/static/link.html
apps/common/Web/static/media/apk_icon.png
apps/common/Web/static/media/apple_store.svg
apps/common/Web/static/media/contact.js
apps/common/Web/static/media/contact_page_mobile.png
apps/common/Web/static/media/f_droid.svg
apps/common/Web/static/media/favicon.ico
apps/common/Web/static/media/GilroyBold.woff2
apps/common/Web/static/media/GilroyLight.woff2
apps/common/Web/static/media/GilroyMedium.woff2
apps/common/Web/static/media/GilroyRegular.woff2
apps/common/Web/static/media/GilroyRegularItalic.woff2
apps/common/Web/static/media/google_play.svg
apps/common/Web/static/media/logo-dark.png
apps/common/Web/static/media/logo-light.png
apps/common/Web/static/media/logo-symbol-dark.svg
apps/common/Web/static/media/logo-symbol-light.svg
apps/common/Web/static/media/moon.svg
apps/common/Web/static/media/qrcode.js
apps/common/Web/static/media/script.js
apps/common/Web/static/media/style.css
apps/common/Web/static/media/sun.svg
apps/common/Web/static/media/swiper-bundle.min.css
apps/common/Web/static/media/swiper-bundle.min.js
apps/common/Web/static/media/tailwind.css
apps/common/Web/static/media/testflight.png
apps/xftp-server/static/media/xftp-protocol.svg
apps/xftp-server/static/media/xftp-protocol-dark.svg
apps/xftp-server/static/xftp-web-bundle/crypto.worker.js
apps/xftp-server/static/xftp-web-bundle/index.css
apps/xftp-server/static/xftp-web-bundle/index.js
flag swift
description: Enable swift JSON format
@@ -72,6 +77,11 @@ flag client_postgres
manual: True
default: False
flag server_postgres
description: Build server with support of PostgreSQL.
manual: True
default: False
library
exposed-modules:
Simplex.FileTransfer.Agent
@@ -98,10 +108,13 @@ library
Simplex.Messaging.Agent.Store.AgentStore
Simplex.Messaging.Agent.Store.Common
Simplex.Messaging.Agent.Store.DB
Simplex.Messaging.Agent.Store.Entity
Simplex.Messaging.Agent.Store.Interface
Simplex.Messaging.Agent.Store.Migrations
Simplex.Messaging.Agent.Store.Migrations.App
Simplex.Messaging.Agent.Store.Postgres.Options
Simplex.Messaging.Agent.Store.Shared
Simplex.Messaging.Agent.TRcvQueues
Simplex.Messaging.Agent.TSessionSubs
Simplex.Messaging.Client
Simplex.Messaging.Client.Agent
Simplex.Messaging.Compression
@@ -114,6 +127,7 @@ library
Simplex.Messaging.Crypto.SNTRUP761.Bindings.Defines
Simplex.Messaging.Crypto.SNTRUP761.Bindings.FFI
Simplex.Messaging.Crypto.SNTRUP761.Bindings.RNG
Simplex.Messaging.Crypto.ShortLink
Simplex.Messaging.Encoding
Simplex.Messaging.Encoding.String
Simplex.Messaging.Notifications.Client
@@ -122,10 +136,13 @@ library
Simplex.Messaging.Notifications.Types
Simplex.Messaging.Parsers
Simplex.Messaging.Protocol
Simplex.Messaging.Protocol.Types
Simplex.Messaging.Server.Expiration
Simplex.Messaging.Server.QueueStore.Postgres.Config
Simplex.Messaging.Server.QueueStore.QueueInfo
Simplex.Messaging.ServiceScheme
Simplex.Messaging.Session
Simplex.Messaging.SystemTime
Simplex.Messaging.TMap
Simplex.Messaging.Transport
Simplex.Messaging.Transport.Buffer
@@ -137,6 +154,7 @@ library
Simplex.Messaging.Transport.HTTP2.Server
Simplex.Messaging.Transport.KeepAlive
Simplex.Messaging.Transport.Server
Simplex.Messaging.Transport.Shared
Simplex.Messaging.Util
Simplex.Messaging.Version
Simplex.Messaging.Version.Internal
@@ -147,20 +165,22 @@ library
Simplex.RemoteControl.Types
if flag(client_postgres)
exposed-modules:
Simplex.Messaging.Agent.Store.Postgres
Simplex.Messaging.Agent.Store.Postgres.Common
Simplex.Messaging.Agent.Store.Postgres.DB
Simplex.Messaging.Agent.Store.Postgres.Migrations
Simplex.Messaging.Agent.Store.Postgres.Migrations.App
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20241210_initial
if !flag(client_library)
exposed-modules:
Simplex.Messaging.Agent.Store.Postgres.Util
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20250203_msg_bodies
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20250322_short_links
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20250702_conn_invitations_remove_cascade_delete
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20251009_queue_to_subscribe
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20251010_client_notices
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20251230_strict_tables
Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260410_receive_attempts
else
exposed-modules:
Simplex.Messaging.Agent.Store.SQLite
Simplex.Messaging.Agent.Store.SQLite.Common
Simplex.Messaging.Agent.Store.SQLite.DB
Simplex.Messaging.Agent.Store.SQLite.Migrations
Simplex.Messaging.Agent.Store.SQLite.Migrations.App
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20220101_initial
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20220301_snd_queue_keys
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20220322_notifications
@@ -198,6 +218,21 @@ library
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20240930_ntf_tokens_to_delete
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20241007_rcv_queues_last_broker_ts
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20241224_ratchet_e2e_snd_params
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20250203_msg_bodies
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20250322_short_links
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20250702_conn_invitations_remove_cascade_delete
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20251009_queue_to_subscribe
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20251010_client_notices
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20251230_strict_tables
Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260410_receive_attempts
Simplex.Messaging.Agent.Store.SQLite.Util
if flag(client_postgres) || flag(server_postgres)
exposed-modules:
Simplex.Messaging.Agent.Store.Postgres
Simplex.Messaging.Agent.Store.Postgres.Common
Simplex.Messaging.Agent.Store.Postgres.DB
Simplex.Messaging.Agent.Store.Postgres.Migrations
Simplex.Messaging.Agent.Store.Postgres.Util
if !flag(client_library)
exposed-modules:
Simplex.FileTransfer.Client.Main
@@ -205,36 +240,55 @@ library
Simplex.FileTransfer.Server.Control
Simplex.FileTransfer.Server.Env
Simplex.FileTransfer.Server.Main
Simplex.FileTransfer.Server.Prometheus
Simplex.FileTransfer.Server.Stats
Simplex.FileTransfer.Server.Store
Simplex.FileTransfer.Server.StoreLog
Simplex.Messaging.Notifications.Server
Simplex.Messaging.Notifications.Server.Control
Simplex.Messaging.Notifications.Server.Env
Simplex.Messaging.Notifications.Server.Main
Simplex.Messaging.Notifications.Server.Push.APNS
Simplex.Messaging.Notifications.Server.Push.APNS.Internal
Simplex.Messaging.Notifications.Server.Stats
Simplex.Messaging.Notifications.Server.Store
Simplex.Messaging.Notifications.Server.StoreLog
Simplex.Messaging.Server
Simplex.Messaging.Server.CLI
Simplex.Messaging.Server.Control
Simplex.Messaging.Server.Env.STM
Simplex.Messaging.Server.Information
Simplex.Messaging.Server.Main
Simplex.Messaging.Server.Main.GitCommit
Simplex.Messaging.Server.Main.Init
Simplex.Messaging.Server.Web
Simplex.Messaging.Server.MsgStore
Simplex.Messaging.Server.MsgStore.Journal
Simplex.Messaging.Server.MsgStore.Journal.SharedLock
Simplex.Messaging.Server.MsgStore.STM
Simplex.Messaging.Server.MsgStore.Types
Simplex.Messaging.Server.NtfStore
Simplex.Messaging.Server.Prometheus
Simplex.Messaging.Server.QueueStore
Simplex.Messaging.Server.QueueStore.STM
Simplex.Messaging.Server.QueueStore.Types
Simplex.Messaging.Server.Stats
Simplex.Messaging.Server.StoreLog
Simplex.Messaging.Server.StoreLog.ReadWrite
Simplex.Messaging.Server.StoreLog.Types
Simplex.Messaging.Transport.WebSockets
if flag(server_postgres)
exposed-modules:
Simplex.Messaging.Notifications.Server
Simplex.Messaging.Notifications.Server.Control
Simplex.Messaging.Notifications.Server.Env
Simplex.Messaging.Notifications.Server.Main
Simplex.Messaging.Notifications.Server.Prometheus
Simplex.Messaging.Notifications.Server.Push.APNS
Simplex.Messaging.Notifications.Server.Push.APNS.Internal
Simplex.Messaging.Notifications.Server.Stats
Simplex.Messaging.Notifications.Server.Store
Simplex.Messaging.Notifications.Server.Store.Migrations
Simplex.Messaging.Notifications.Server.Store.Postgres
Simplex.Messaging.Notifications.Server.Store.Types
Simplex.Messaging.Notifications.Server.StoreLog
Simplex.FileTransfer.Server.Store.Postgres
Simplex.FileTransfer.Server.Store.Postgres.Config
Simplex.FileTransfer.Server.Store.Postgres.Migrations
Simplex.Messaging.Server.MsgStore.Postgres
Simplex.Messaging.Server.QueueStore.Postgres
Simplex.Messaging.Server.QueueStore.Postgres.Migrations
other-modules:
Paths_simplexmq
hs-source-dirs:
@@ -271,7 +325,6 @@ library
, hourglass ==0.2.*
, http-types ==0.12.*
, http2 >=4.2.2 && <4.3
, ini ==0.4.1
, iproute ==1.7.*
, iso8601-time ==0.1.*
, memory ==0.18.*
@@ -280,13 +333,11 @@ library
, network-info ==0.2.*
, network-transport ==0.5.6
, network-udp ==0.0.*
, optparse-applicative >=0.15 && <0.17
, process ==1.6.*
, random >=1.1 && <1.3
, scientific ==0.3.7.*
, simple-logger ==0.1.*
, socks ==0.6.*
, stm ==2.5.*
, temporary ==1.3.*
, time ==1.12.*
, time-manager ==0.0.*
, tls >=1.9.0 && <1.10
@@ -302,29 +353,50 @@ library
build-depends:
case-insensitive ==1.2.*
, hashable ==1.4.*
, ini ==0.4.1
, optparse-applicative >=0.15 && <0.17
, process ==1.6.*
, temporary ==1.3.*
, wai >=3.2 && <3.3
, wai-app-static >=3.1 && <3.2
, wai-websockets >=3.0.1 && <3.1
, warp ==3.3.30
, warp-tls ==3.4.7
, websockets ==0.12.*
if flag(client_postgres)
, zlib >=0.6 && <0.8
if flag(client_postgres) || flag(server_postgres)
build-depends:
postgresql-libpq >=0.10.0.0
, postgresql-simple ==0.7.*
, raw-strings-qq ==1.1.*
if flag(client_postgres)
cpp-options: -DdbPostgres
else
build-depends:
direct-sqlcipher ==2.3.*
, sqlcipher-simple ==0.4.*
if flag(server_postgres)
cpp-options: -DdbServerPostgres
build-depends:
hex-text ==0.1.*
if impl(ghc >= 9.6.2)
build-depends:
bytestring ==0.11.*
, template-haskell ==2.20.*
, text >=2.0.1 && <2.2
if impl(ghc < 9.6.2)
build-depends:
bytestring ==0.10.*
, template-haskell ==2.16.*
, text >=1.2.3.0 && <1.3
executable ntf-server
if flag(client_library)
buildable: False
if flag(server_postgres)
cpp-options: -DdbServerPostgres
else
buildable: False
main-is: Main.hs
other-modules:
Paths_simplexmq
@@ -342,32 +414,26 @@ executable ntf-server
executable smp-server
if flag(client_library)
buildable: False
if flag(server_postgres)
cpp-options: -DdbServerPostgres
main-is: Main.hs
other-modules:
Static
Static.Embedded
SMPWeb
Web.Embedded
Paths_simplexmq
hs-source-dirs:
apps/smp-server
apps/smp-server/web
apps/common
default-extensions:
StrictData
ghc-options: -Weverything -Wno-missing-exported-signatures -Wno-missing-import-lists -Wno-missed-specialisations -Wno-all-missed-specialisations -Wno-unsafe -Wno-safe -Wno-missing-local-signatures -Wno-missing-kind-signatures -Wno-missing-deriving-strategies -Wno-monomorphism-restriction -Wno-prepositive-qualified-module -Wno-implicit-prelude -Wno-missing-safe-haskell-mode -Wno-missing-export-lists -Wno-partial-fields -Wcompat -Werror=incomplete-record-updates -Werror=incomplete-patterns -Werror=incomplete-uni-patterns -Werror=missing-methods -Werror=tabs -Wredundant-constraints -Wincomplete-record-updates -Wunused-type-patterns -O2 -threaded -rtsopts
build-depends:
base
, bytestring
, directory
, file-embed
, filepath
, network
, file-embed >=0.0.10 && <0.1
, simple-logger
, simplexmq
, text
, unliftio
, wai
, wai-app-static
, warp ==3.3.30
, warp-tls ==3.4.7
default-language: Haskell2010
executable xftp
@@ -391,14 +457,21 @@ executable xftp-server
buildable: False
main-is: Main.hs
other-modules:
XFTPWeb
Web.Embedded
Paths_simplexmq
hs-source-dirs:
apps/xftp-server
apps/common
default-extensions:
StrictData
ghc-options: -Weverything -Wno-missing-exported-signatures -Wno-missing-import-lists -Wno-missed-specialisations -Wno-all-missed-specialisations -Wno-unsafe -Wno-safe -Wno-missing-local-signatures -Wno-missing-kind-signatures -Wno-missing-deriving-strategies -Wno-monomorphism-restriction -Wno-prepositive-qualified-module -Wno-implicit-prelude -Wno-missing-safe-haskell-mode -Wno-missing-export-lists -Wno-partial-fields -Wcompat -Werror=incomplete-record-updates -Werror=incomplete-patterns -Werror=incomplete-uni-patterns -Werror=missing-methods -Werror=tabs -Wredundant-constraints -Wincomplete-record-updates -Wunused-type-patterns -O2 -threaded -rtsopts
build-depends:
base
, bytestring
, directory
, file-embed >=0.0.10 && <0.1
, filepath
, simple-logger
, simplexmq
default-language: Haskell2010
@@ -415,8 +488,8 @@ test-suite simplexmq-test
AgentTests.EqInstances
AgentTests.FunctionalAPITests
AgentTests.MigrationTests
AgentTests.NotificationTests
AgentTests.ServerChoice
AgentTests.ShortLinkTests
CLITests
CoreTests.BatchingTests
CoreTests.CryptoFileTests
@@ -426,13 +499,10 @@ test-suite simplexmq-test
CoreTests.RetryIntervalTests
CoreTests.SOCKSSettings
CoreTests.StoreLogTests
CoreTests.TRcvQueuesTests
CoreTests.TSessionSubs
CoreTests.UtilTests
CoreTests.VersionRangeTests
FileDescriptionTests
Fixtures
NtfClient
NtfServerTests
RemoteControl
ServerTests
SMPAgentClient
@@ -443,18 +513,35 @@ test-suite simplexmq-test
XFTPCLI
XFTPClient
XFTPServerTests
Static
Static.Embedded
WebTests
XFTPWebTests
SMPWebTests
SMPWeb
XFTPWeb
Web.Embedded
Paths_simplexmq
if !flag(client_postgres)
if flag(client_postgres)
other-modules:
Fixtures
else
other-modules:
AgentTests.SchemaDump
AgentTests.SQLiteTests
if flag(server_postgres)
other-modules:
AgentTests.NotificationTests
CoreTests.XFTPStoreTests
NtfClient
NtfServerTests
PostgresSchemaDump
hs-source-dirs:
tests
apps/smp-server/web
apps/smp-server
apps/xftp-server
apps/common
default-extensions:
StrictData
-- add -fhpc to ghc-options to run tests with coverage
ghc-options: -Weverything -Wno-missing-exported-signatures -Wno-missing-import-lists -Wno-missed-specialisations -Wno-all-missed-specialisations -Wno-unsafe -Wno-safe -Wno-missing-local-signatures -Wno-missing-kind-signatures -Wno-missing-deriving-strategies -Wno-monomorphism-restriction -Wno-prepositive-qualified-module -Wno-implicit-prelude -Wno-missing-safe-haskell-mode -Wno-missing-export-lists -Wno-partial-fields -Wcompat -Werror=incomplete-record-updates -Werror=incomplete-patterns -Werror=incomplete-uni-patterns -Werror=missing-methods -Werror=tabs -Wredundant-constraints -Wincomplete-record-updates -Wunused-type-patterns -O2 -threaded -rtsopts -with-rtsopts=-A64M -with-rtsopts=-N1
build-depends:
base
@@ -462,18 +549,19 @@ test-suite simplexmq-test
, async
, base64-bytestring
, bytestring
, case-insensitive ==1.2.*
, containers
, crypton
, crypton-x509
, crypton-x509-store
, crypton-x509-validation
, deepseq ==1.4.*
, directory
, file-embed
, file-embed >=0.0.10 && <0.1
, filepath
, generic-random ==1.5.*
, hashable
, hspec ==2.11.*
, hspec-core ==2.11.*
, http-client
, http-types
, http2
@@ -484,7 +572,6 @@ test-suite simplexmq-test
, memory
, mtl
, network
, process
, QuickCheck ==2.14.*
, random
, silently ==1.2.*
@@ -498,18 +585,19 @@ test-suite simplexmq-test
, unliftio
, unliftio-core
, unordered-containers
, wai
, wai-app-static
, warp
, warp-tls
, yaml
default-language: Haskell2010
if flag(server_postgres)
cpp-options: -DdbServerPostgres
if flag(client_postgres)
build-depends:
postgresql-libpq >=0.10.0.0
, postgresql-simple ==0.7.*
, raw-strings-qq ==1.1.*
cpp-options: -DdbPostgres
else
build-depends:
sqlcipher-simple
if !flag(client_postgres) || flag(client_postgres) || flag(server_postgres)
build-depends:
deepseq ==1.4.*
, process
if flag(client_postgres) || flag(server_postgres)
build-depends:
postgresql-simple ==0.7.*
+4
View File
@@ -0,0 +1,4 @@
node_modules/
dist/
dist-test/
package-lock.json
+14
View File
@@ -0,0 +1,14 @@
addToLibrary({
js_random_bytes: function(buf, len) {
var bytes = new Uint8Array(len);
if (typeof crypto !== 'undefined' && crypto.getRandomValues) {
crypto.getRandomValues(bytes);
} else {
// Node.js fallback
var nodeCrypto = require('crypto');
var nodeBytes = nodeCrypto.randomBytes(len);
bytes.set(nodeBytes);
}
HEAPU8.set(bytes, buf);
}
});
+315
View File
@@ -0,0 +1,315 @@
/*
20080913
D. J. Bernstein
Public domain.
SHA-512 implementation from SUPERCOP/NaCl.
Source: https://bench.cr.yp.to/supercop.html
crypto_hashblocks/sha512/ref/blocks.c
crypto_hash/sha512/ref/hash.c
Combined into a single file for WASM compilation alongside sntrup761.
*/
#include "sha512.h"
typedef unsigned long long uint64;
/* -- crypto_hashblocks_sha512 (blocks.c) -- */
static uint64 load_bigendian(const unsigned char *x)
{
return
(uint64) (x[7]) \
| (((uint64) (x[6])) << 8) \
| (((uint64) (x[5])) << 16) \
| (((uint64) (x[4])) << 24) \
| (((uint64) (x[3])) << 32) \
| (((uint64) (x[2])) << 40) \
| (((uint64) (x[1])) << 48) \
| (((uint64) (x[0])) << 56)
;
}
static void store_bigendian(unsigned char *x,uint64 u)
{
x[7] = u; u >>= 8;
x[6] = u; u >>= 8;
x[5] = u; u >>= 8;
x[4] = u; u >>= 8;
x[3] = u; u >>= 8;
x[2] = u; u >>= 8;
x[1] = u; u >>= 8;
x[0] = u;
}
#define SHR(x,c) ((x) >> (c))
#define ROTR(x,c) (((x) >> (c)) | ((x) << (64 - (c))))
#define Ch(x,y,z) ((x & y) ^ (~x & z))
#define Maj(x,y,z) ((x & y) ^ (x & z) ^ (y & z))
#define Sigma0(x) (ROTR(x,28) ^ ROTR(x,34) ^ ROTR(x,39))
#define Sigma1(x) (ROTR(x,14) ^ ROTR(x,18) ^ ROTR(x,41))
#define sigma0(x) (ROTR(x, 1) ^ ROTR(x, 8) ^ SHR(x,7))
#define sigma1(x) (ROTR(x,19) ^ ROTR(x,61) ^ SHR(x,6))
#define M(w0,w14,w9,w1) w0 = sigma1(w14) + w9 + sigma0(w1) + w0;
#define EXPAND \
M(w0 ,w14,w9 ,w1 ) \
M(w1 ,w15,w10,w2 ) \
M(w2 ,w0 ,w11,w3 ) \
M(w3 ,w1 ,w12,w4 ) \
M(w4 ,w2 ,w13,w5 ) \
M(w5 ,w3 ,w14,w6 ) \
M(w6 ,w4 ,w15,w7 ) \
M(w7 ,w5 ,w0 ,w8 ) \
M(w8 ,w6 ,w1 ,w9 ) \
M(w9 ,w7 ,w2 ,w10) \
M(w10,w8 ,w3 ,w11) \
M(w11,w9 ,w4 ,w12) \
M(w12,w10,w5 ,w13) \
M(w13,w11,w6 ,w14) \
M(w14,w12,w7 ,w15) \
M(w15,w13,w8 ,w0 )
#define F(w,k) \
T1 = h + Sigma1(e) + Ch(e,f,g) + k + w; \
T2 = Sigma0(a) + Maj(a,b,c); \
h = g; \
g = f; \
f = e; \
e = d + T1; \
d = c; \
c = b; \
b = a; \
a = T1 + T2;
static int crypto_hashblocks_sha512(unsigned char *statebytes,const unsigned char *in,unsigned long long inlen)
{
uint64 state[8];
uint64 a;
uint64 b;
uint64 c;
uint64 d;
uint64 e;
uint64 f;
uint64 g;
uint64 h;
uint64 T1;
uint64 T2;
a = load_bigendian(statebytes + 0); state[0] = a;
b = load_bigendian(statebytes + 8); state[1] = b;
c = load_bigendian(statebytes + 16); state[2] = c;
d = load_bigendian(statebytes + 24); state[3] = d;
e = load_bigendian(statebytes + 32); state[4] = e;
f = load_bigendian(statebytes + 40); state[5] = f;
g = load_bigendian(statebytes + 48); state[6] = g;
h = load_bigendian(statebytes + 56); state[7] = h;
while (inlen >= 128) {
uint64 w0 = load_bigendian(in + 0);
uint64 w1 = load_bigendian(in + 8);
uint64 w2 = load_bigendian(in + 16);
uint64 w3 = load_bigendian(in + 24);
uint64 w4 = load_bigendian(in + 32);
uint64 w5 = load_bigendian(in + 40);
uint64 w6 = load_bigendian(in + 48);
uint64 w7 = load_bigendian(in + 56);
uint64 w8 = load_bigendian(in + 64);
uint64 w9 = load_bigendian(in + 72);
uint64 w10 = load_bigendian(in + 80);
uint64 w11 = load_bigendian(in + 88);
uint64 w12 = load_bigendian(in + 96);
uint64 w13 = load_bigendian(in + 104);
uint64 w14 = load_bigendian(in + 112);
uint64 w15 = load_bigendian(in + 120);
F(w0 ,0x428a2f98d728ae22ULL)
F(w1 ,0x7137449123ef65cdULL)
F(w2 ,0xb5c0fbcfec4d3b2fULL)
F(w3 ,0xe9b5dba58189dbbcULL)
F(w4 ,0x3956c25bf348b538ULL)
F(w5 ,0x59f111f1b605d019ULL)
F(w6 ,0x923f82a4af194f9bULL)
F(w7 ,0xab1c5ed5da6d8118ULL)
F(w8 ,0xd807aa98a3030242ULL)
F(w9 ,0x12835b0145706fbeULL)
F(w10,0x243185be4ee4b28cULL)
F(w11,0x550c7dc3d5ffb4e2ULL)
F(w12,0x72be5d74f27b896fULL)
F(w13,0x80deb1fe3b1696b1ULL)
F(w14,0x9bdc06a725c71235ULL)
F(w15,0xc19bf174cf692694ULL)
EXPAND
F(w0 ,0xe49b69c19ef14ad2ULL)
F(w1 ,0xefbe4786384f25e3ULL)
F(w2 ,0x0fc19dc68b8cd5b5ULL)
F(w3 ,0x240ca1cc77ac9c65ULL)
F(w4 ,0x2de92c6f592b0275ULL)
F(w5 ,0x4a7484aa6ea6e483ULL)
F(w6 ,0x5cb0a9dcbd41fbd4ULL)
F(w7 ,0x76f988da831153b5ULL)
F(w8 ,0x983e5152ee66dfabULL)
F(w9 ,0xa831c66d2db43210ULL)
F(w10,0xb00327c898fb213fULL)
F(w11,0xbf597fc7beef0ee4ULL)
F(w12,0xc6e00bf33da88fc2ULL)
F(w13,0xd5a79147930aa725ULL)
F(w14,0x06ca6351e003826fULL)
F(w15,0x142929670a0e6e70ULL)
EXPAND
F(w0 ,0x27b70a8546d22ffcULL)
F(w1 ,0x2e1b21385c26c926ULL)
F(w2 ,0x4d2c6dfc5ac42aedULL)
F(w3 ,0x53380d139d95b3dfULL)
F(w4 ,0x650a73548baf63deULL)
F(w5 ,0x766a0abb3c77b2a8ULL)
F(w6 ,0x81c2c92e47edaee6ULL)
F(w7 ,0x92722c851482353bULL)
F(w8 ,0xa2bfe8a14cf10364ULL)
F(w9 ,0xa81a664bbc423001ULL)
F(w10,0xc24b8b70d0f89791ULL)
F(w11,0xc76c51a30654be30ULL)
F(w12,0xd192e819d6ef5218ULL)
F(w13,0xd69906245565a910ULL)
F(w14,0xf40e35855771202aULL)
F(w15,0x106aa07032bbd1b8ULL)
EXPAND
F(w0 ,0x19a4c116b8d2d0c8ULL)
F(w1 ,0x1e376c085141ab53ULL)
F(w2 ,0x2748774cdf8eeb99ULL)
F(w3 ,0x34b0bcb5e19b48a8ULL)
F(w4 ,0x391c0cb3c5c95a63ULL)
F(w5 ,0x4ed8aa4ae3418acbULL)
F(w6 ,0x5b9cca4f7763e373ULL)
F(w7 ,0x682e6ff3d6b2b8a3ULL)
F(w8 ,0x748f82ee5defb2fcULL)
F(w9 ,0x78a5636f43172f60ULL)
F(w10,0x84c87814a1f0ab72ULL)
F(w11,0x8cc702081a6439ecULL)
F(w12,0x90befffa23631e28ULL)
F(w13,0xa4506cebde82bde9ULL)
F(w14,0xbef9a3f7b2c67915ULL)
F(w15,0xc67178f2e372532bULL)
EXPAND
F(w0 ,0xca273eceea26619cULL)
F(w1 ,0xd186b8c721c0c207ULL)
F(w2 ,0xeada7dd6cde0eb1eULL)
F(w3 ,0xf57d4f7fee6ed178ULL)
F(w4 ,0x06f067aa72176fbaULL)
F(w5 ,0x0a637dc5a2c898a6ULL)
F(w6 ,0x113f9804bef90daeULL)
F(w7 ,0x1b710b35131c471bULL)
F(w8 ,0x28db77f523047d84ULL)
F(w9 ,0x32caab7b40c72493ULL)
F(w10,0x3c9ebe0a15c9bebcULL)
F(w11,0x431d67c49c100d4cULL)
F(w12,0x4cc5d4becb3e42b6ULL)
F(w13,0x597f299cfc657e2aULL)
F(w14,0x5fcb6fab3ad6faecULL)
F(w15,0x6c44198c4a475817ULL)
a += state[0];
b += state[1];
c += state[2];
d += state[3];
e += state[4];
f += state[5];
g += state[6];
h += state[7];
state[0] = a;
state[1] = b;
state[2] = c;
state[3] = d;
state[4] = e;
state[5] = f;
state[6] = g;
state[7] = h;
in += 128;
inlen -= 128;
}
store_bigendian(statebytes + 0,state[0]);
store_bigendian(statebytes + 8,state[1]);
store_bigendian(statebytes + 16,state[2]);
store_bigendian(statebytes + 24,state[3]);
store_bigendian(statebytes + 32,state[4]);
store_bigendian(statebytes + 40,state[5]);
store_bigendian(statebytes + 48,state[6]);
store_bigendian(statebytes + 56,state[7]);
return inlen;
}
/* -- crypto_hash_sha512 (hash.c) -- */
static const unsigned char iv[64] = {
0x6a,0x09,0xe6,0x67,0xf3,0xbc,0xc9,0x08,
0xbb,0x67,0xae,0x85,0x84,0xca,0xa7,0x3b,
0x3c,0x6e,0xf3,0x72,0xfe,0x94,0xf8,0x2b,
0xa5,0x4f,0xf5,0x3a,0x5f,0x1d,0x36,0xf1,
0x51,0x0e,0x52,0x7f,0xad,0xe6,0x82,0xd1,
0x9b,0x05,0x68,0x8c,0x2b,0x3e,0x6c,0x1f,
0x1f,0x83,0xd9,0xab,0xfb,0x41,0xbd,0x6b,
0x5b,0xe0,0xcd,0x19,0x13,0x7e,0x21,0x79
};
void crypto_hash_sha512(unsigned char *out,
const unsigned char *in,
unsigned long long inlen)
{
unsigned char h[64];
unsigned char padded[256];
int i;
unsigned long long bytes = inlen;
for (i = 0;i < 64;++i) h[i] = iv[i];
crypto_hashblocks_sha512(h,in,inlen);
in += inlen;
inlen &= 127;
in -= inlen;
for (i = 0;i < (int)inlen;++i) padded[i] = in[i];
padded[inlen] = 0x80;
if (inlen < 112) {
for (i = inlen + 1;i < 119;++i) padded[i] = 0;
padded[119] = bytes >> 61;
padded[120] = bytes >> 53;
padded[121] = bytes >> 45;
padded[122] = bytes >> 37;
padded[123] = bytes >> 29;
padded[124] = bytes >> 21;
padded[125] = bytes >> 13;
padded[126] = bytes >> 5;
padded[127] = bytes << 3;
crypto_hashblocks_sha512(h,padded,128);
} else {
for (i = inlen + 1;i < 247;++i) padded[i] = 0;
padded[247] = bytes >> 61;
padded[248] = bytes >> 53;
padded[249] = bytes >> 45;
padded[250] = bytes >> 37;
padded[251] = bytes >> 29;
padded[252] = bytes >> 21;
padded[253] = bytes >> 13;
padded[254] = bytes >> 5;
padded[255] = bytes << 3;
crypto_hashblocks_sha512(h,padded,256);
}
for (i = 0;i < 64;++i) out[i] = h[i];
}
+11
View File
@@ -0,0 +1,11 @@
interface Sntrup761Module {
_sntrup761_wasm_keypair(pk: number, sk: number): void
_sntrup761_wasm_enc(c: number, k: number, pk: number): void
_sntrup761_wasm_dec(k: number, c: number, sk: number): void
_malloc(size: number): number
_free(ptr: number): void
HEAPU8: Uint8Array
}
declare function createSntrup761(): Promise<Sntrup761Module>
export default createSntrup761
+34
View File
@@ -0,0 +1,34 @@
/*
* WASM wrapper for sntrup761.
* Provides JS-callable functions with RNG from JS imports.
*
* Build: emcc sntrup761_wasm.c sntrup761.c sha512.c -O2 -o sntrup761.js \
* -s EXPORTED_FUNCTIONS='["_sntrup761_wasm_keypair","_sntrup761_wasm_enc","_sntrup761_wasm_dec","_malloc","_free"]' \
* -s EXPORTED_RUNTIME_METHODS='["ccall","cwrap"]'
*/
#include "sntrup761.h"
#include <stdlib.h>
/* Import RNG from JS environment */
extern void js_random_bytes(unsigned char *buf, int len);
/* RNG callback adapter for sntrup761 */
static void wasm_random(void *ctx, size_t length, uint8_t *dst) {
(void)ctx;
js_random_bytes(dst, (int)length);
}
/* JS-callable wrappers */
void sntrup761_wasm_keypair(unsigned char *pk, unsigned char *sk) {
sntrup761_keypair(pk, sk, NULL, wasm_random);
}
void sntrup761_wasm_enc(unsigned char *c, unsigned char *k, const unsigned char *pk) {
sntrup761_enc(c, k, pk, NULL, wasm_random);
}
void sntrup761_wasm_dec(unsigned char *k, const unsigned char *c, const unsigned char *sk) {
sntrup761_dec(k, c, sk);
}
+36
View File
@@ -0,0 +1,36 @@
{
"name": "@simplex-chat/smp-web",
"version": "0.1.0",
"description": "SMP protocol client for web/browser environments",
"license": "AGPL-3.0-only",
"repository": {
"type": "git",
"url": "git+https://github.com/simplex-chat/simplexmq.git",
"directory": "smp-web"
},
"type": "module",
"files": [
"src",
"dist"
],
"scripts": {
"build:wasm": "mkdir -p dist/wasm && npx emcc cbits/sntrup761_wasm.c ../cbits/sntrup761.c cbits/sha512.c -I../cbits -O2 -o dist/wasm/sntrup761.mjs -s EXPORTED_FUNCTIONS='[\"_sntrup761_wasm_keypair\",\"_sntrup761_wasm_enc\",\"_sntrup761_wasm_dec\",\"_malloc\",\"_free\"]' -s EXPORTED_RUNTIME_METHODS='[\"ccall\",\"cwrap\",\"HEAPU8\"]' -s MODULARIZE=1 -s EXPORT_NAME='createSntrup761' -s ALLOW_MEMORY_GROWTH=1 -s ENVIRONMENT='web,node' --js-library cbits/js_random.js && cp cbits/sntrup761.d.mts dist/wasm/",
"build:ts": "tsc",
"build:test": "tsc -p tsconfig.test.json",
"build": "npm run build:wasm && npm run build:ts && npm run build:test"
},
"dependencies": {
"@noble/ciphers": "^2.2.0",
"@noble/curves": "^2.2.0",
"@noble/hashes": "^1.5.0",
"@simplex-chat/xftp-web": "file:../xftp-web",
"emsdk": "^0.4.0"
},
"devDependencies": {
"@types/node": "^25.5.0",
"@types/ws": "^8.18.1",
"fake-indexeddb": "^6.2.5",
"typescript": "^5.4.0",
"ws": "^8.0.0"
}
}

Some files were not shown because too many files have changed in this diff Show More