- chunk download doesn't loop on permanent errors
- decryption errors are considered permanent - local worker doesn't retry
- update replica retries; to do - consider use for this field, or remove it
- rcv file Error status - to prevent repeat reads of chunks for download, files for decryption; also plan to use it for filtering on cleanup
- error string saved in separate field for debugging (not part of status type)
- agent event for rcv file errors
* send "quota exceeded" message from SMP server when sender gets ERR QUOTA (ignored in the agent for now)
* send msg quota to the recipient to indicate that sender got ERR QUOTA, test
* switch between slow/fast retry intervals (tests do not pass yet)
* send QCONT message, refactor RetryInterval, test
* refactor
* remove comment
* remove space
* unit test for withRetryLock2
* refactor
* add optional auxiliary data field to ConnectionRequestUri
* remove import
* fix, test
* fix StrEncoding Char
* data only in sync command, type
* fixing
* queryParamStr
* safeDecodeUtf8
Co-authored-by: Evgeny Poberezkin <2769109+epoberezkin@users.noreply.github.com>
* rfc: queue rotation
* update rfc
* messages for queue rotation
* allow multiple subscribed queues per connection in Agent/Client.hs
* refactor
* fix module name
* allow multiple queues in duplex connection type
* update commands
* add indices
* addConnectionRcvQueue
* switch connection to another queue (WIP)
* update schema/protocol
* switching queue works, but sending messages after the switch fails
* messages are delivered after rotation
* use connection-scoped queue ID
* rename queue records fields
* refactor using SMPQueue class/instances
* simplify queries
* QKEY: check queue is not secured, refactor
* update rfc
* mark queue as primary in QUSE
* queue rotation errors
* fix async ack
* fix async ACK to send OK
* correction
Co-authored-by: JRoberts <8711996+jr-simplex@users.noreply.github.com>
* use SWCH command
* rename
* take into account only active queue subscription when determining connection result if at least one queue is active
* remove comment
* only enable notifications for connections with enableNtfs = True
* async test (WIP)
* async queue rotation test
* simplify combining results
* test with 2 servers
* fix unused subscribeConnection
* switch to cabal build
* increase build timeout
* increase delay in async test
* skip queue rotation tests
* build matrix
* step name
* use ubuntu-18.04 in build matrix
* enable rotation tests
Co-authored-by: JRoberts <8711996+jr-simplex@users.noreply.github.com>
* store SMP client version per queue
* use versioned format for queue addresses (not compatible with v1)
* multiple server hosts in queue URI/data, versioned encoding (cross-version tests fail)
* choose server host based on socksProxy setting
* test works
* multiple server addresses mostly work
* add onion hosts for provided servers
* update SMP hosts
* SMP v3: encrypt message timestamp and flags together with the body between server and recipient
* v3 tests
* update protocol doc
* add test for max size message
* delay in notifications test
* simplify v3
* encrypt server message to the recipient when sent
* refactor
* exit on error restoring the messages
* refactor, increase test timeout
* style
* add prints to the test
* remove error from unsafeMaxLenBS
* update protocol
* lint, improve test, change func param
* support stopping and resuming agent (#385)
* export agentDbPath
* support fully closing and resuming agent
* whitespace
* clean up
* support message flags visible to SMP server to control notifications (and for any future extensions)
* simplify message flags encoding
* GET command
* primary tokens
* support repeat token registration (TODO fix tests)
* fix notifications tests
* fix/test repeat/new registrations of the same token
* re-register token when subsequent ntf command fails with AUTH error (e.g. when server is re-started)
* cancel periodic notifications when token is deleted on the server
* debug failing test on CI
* fix notification test in CI
* debug CI test
* add delay in notificaitons test after server restart
* process ntf server commands
* when subscription is re-created and it was ENDed, resubscribe to SMP
* SMPClientAgent draft
* SMPClientAgent: remove double tracking of subscriptions
* subscriber frame
* PING error now throws error to restart SMPClient for more reliable re-connection (#342)
* increase TCP timeout to 5 sec
* add pragmas and vacuum db (#343)
* vacuum in each connection to enable auto-vacuum (#344)
* update protocol, token verification
* refactor SMPClient to ProtocoClient, to use with notification server protocol
* notification server client, managing notification clients in the agent
* stub for push payload
Co-authored-by: JRoberts <8711996+jr-simplex@users.noreply.github.com>
* notifications: protocol
* update protocol to include subscription ID and DH public key
* update protocol, started server
* add notification server subscription DH key
* use the same command type in notifications protocol, protocol parsing, server frame
* remove empty files
* make sending confirmation asynchronous
* confirm first queue synchronously (on JOIN, and delete connection on failure), second queue asynchronously (from REPLY msg)
* test to retry JOIN when the first attempt fails
* process confirmation delivery errors
Evgeny Poberezkin