- Install nyc for Istanbul instrumentation
- Add scripts/instrument-frontend.sh to instrument public/*.js
- Add scripts/collect-frontend-coverage.js to extract window.__coverage__
- Add scripts/combined-coverage.sh for combined server+frontend coverage
- Make server.js serve public-instrumented/ when COVERAGE=1 is set
- Add test:full-coverage npm script
- Add public-instrumented/ and .nyc_output/ to .gitignore
Tests now run in the test job, not after deploy. Spins up server.js
on port 13581, runs Playwright against it, kills it after.
If E2E fails, deploy is blocked — broken code never reaches prod.
BASE_URL env var makes the test configurable.
test-server-routes.js destructures { cache, pktStore, db } but these
weren't in module.exports. Also adds require.main guard so server
doesn't listen when imported by tests.
8 smoke tests against prod after deployment completes.
Uses Playwright bundled Chromium on x86 runner.
Falls back to CHROMIUM_PATH env var for other architectures.
Proof of concept: bare Playwright (not @playwright/test) running 8 critical
flow tests against analyzer.00id.net:
- Home page, nodes, map, packets, node detail, theme customizer, dark mode, analytics
- Uses system Chromium on ARM (Playwright bundled binary doesn't work on musl)
- Not added to test-all.sh or CI yet — POC only
- Run with: node test-e2e-playwright.js
Badges show: 'tests: 844/844 passed' and 'coverage: 76%'
Updated automatically by CI after each run via .badges/ JSON files.
Color: green >80%, yellow >60%, red <60%.
decodeEncryptedPayload: dest(1)+src(1)+MAC(2) per PAYLOAD_VER_1
decodeAck: dest(1)+src(1)+ack_hash(4)
decodeAnonReq: dest(1)+pubkey(32)+MAC(2)
decodePath: dest(1)+src(1)+MAC(2)+data
Source: firmware/src/Mesh.cpp lines 129-130, MeshCore.h CIPHER_MAC_SIZE=2
Golden fixture tests need updating to match correct output.
Replace duplicated function definitions in server.js with imports from
server-helpers.js. Functions replaced: loadConfigFile, loadThemeFile,
buildHealthConfig, getHealthMs, isHashSizeFlipFlop, computeContentHash,
geoDist, deriveHashtagChannelKey, buildBreakdown, updateHashSizeForPacket,
rebuildHashSizeMap, requireApiKey, CONFIG_PATHS, THEME_PATHS.
disambiguateHops kept in server.js due to behavioral differences in the
distance sanity check (server version nulls lat/lon on unreliable hops
and adds ambiguous field in output mapping).
server.js: 3201 → 3001 lines (-200 lines, -224 deletions/+24 insertions)
All tests pass (unit, e2e, frontend).
Added test job that runs unit tests + integration tests + coverage
before deploy. Deploy job depends on test job passing.
If any test fails, deploy is blocked.
npm test: all tests + coverage summary
npm run test:unit: fast unit tests only
npm run test:coverage: full suite + HTML report in coverage/
Baseline: 37% statements, 42% branches, 54% functions
Fixed e2e channels crash (undefined .length on null)
Cloned meshcore-dev/MeshCore to firmware/ (gitignored).
AGENTS.md now mandates reading firmware source before implementing
anything protocol-related. Lists key files to check.
Derived from git history analysis: 4.3x fix ratio, 12 reverts, 7 cache
buster regressions, 21 commits for hash size, 6 for QR overlay.
Rules: test before push, bump cache busters, verify API shape, plan
before implementing, one commit per change, understand before fixing.
renderRows() in nodes.js and three places in map.js were using only
n.last_seen to compute active/stale status, ignoring the more recent
n.last_heard from in-memory packets. This caused nodes that were recently
heard but had an old DB last_seen to incorrectly show as stale.
Also adds 29 unit tests for the aging system (getNodeStatus,
getStatusInfo, getStatusTooltip, threshold values).
Keep the 📍map link in the Location metadata row (goes to app map).
Remove the redundant 📍 Map pill in the hex breakdown (went to Google Maps).
One link, one style.
Was making N API calls per observer for ambiguous hops on every page load,
plus another per packet detail view. All hop resolution now uses the
client-side HopResolver which already handles ambiguous prefixes.
Eliminates the main perf regression.
ReferenceError: node is not defined — was using wrong variable name.
Verified with 37 tests covering: firmware type names, aliases, route,
numeric ops, string ops, payload dot notation, hops, size, observations,
AND/OR/NOT, parentheses, and error handling.
Was using display names like 'Channel Msg' which aren't standard.
Now resolves to firmware names: GRP_TXT, TXT_MSG, REQ, ADVERT, etc.
Also accepts aliases: 'channel', 'dm', 'Channel Msg' all map to the
correct firmware name for convenience.
queryGrouped was missing route_type, snr, rssi fields. The packet filter
language couldn't filter by route/snr/rssi since grouped packets didn't
have those fields.
- Add getStatusTooltip() helper with role-aware explanations
- Tooltips on status labels in: node badges, status explanation, detail table
- Tooltips on map legend active/stale counts per role
- Native title attributes (long-press on mobile)
- Bump cache busters
Server now computes last_heard from in-memory packet store (all traffic
types) and includes it in /api/nodes response. Client prefers last_heard
over DB last_seen for display, sort, filter, and status calculation.
Fixes inconsistency where list showed '5d ago' but side pane showed
'26m ago' for the same node.
you