Step 1 of moving the desktop app's login out of the webview's plaintext
localStorage: keep a verified copy of the tokens in the OS keychain
(Windows Credential Manager, via cinny-desktop's new secure_session_*
commands). The session is still read from localStorage exactly as before,
so nothing about login changes and a keychain problem can't log anyone out.
Step 2 (a later release, once this has run on real installs) switches reads
to the keychain and drops the tokens from localStorage.
- sessions.ts: onSessionPersisted — listeners told about every session
write (login, token rotation) and removal (logout); a throwing listener
can't break the write.
- keychainMirror.ts: desktop only. Mirrors userId/deviceId/accessToken/
refreshToken (not the rest of the session); reads first and writes only
when the copy differs, then verifies by reading back; serialized, 5 s
timeouts; no session → clear (also covers a logout whose reload beat the
clear). Every failure is a status, never an exception. A desktop build
without the commands reads as "unsupported", so this can ship before the
desktop side.
- Settings → General (desktop): "Login in the system keychain" status.
Tested: unit tests (fake keychain: store, no rewrite when current, rotation,
clear, unsupported, missing commands, denied write, read-back mismatch,
timeout); a simulated desktop app with a fake keychain, 14/14 (login
mirrors only the secrets, Settings status, reload verifies without
rewriting, logout clears, an existing session is mirrored after upgrade,
Linux/denied show an honest status and stay logged in); the real Linux
desktop binary (commands answer "unsupported", login unaffected, Settings
says so). Unit 1295 pass, Playwright 20 passed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPmy3tPq869XDW4njjVaKA
register() rejects on non-http(s) origins, and the rejection was
unhandled. The desktop app's debug build loads from tauri://localhost,
which surfaced as a Sentry "serviceWorker.register() must be called with
a script URL whose protocol is either HTTP or HTTPS". Skip registration
there, and catch any other failure (e.g. SWs disabled) with a warning.
The app works without a SW.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPmy3tPq869XDW4njjVaKA
vite-plugin-pwa's dev-sw.js imports workbox as a module; registering it
as a classic script failed with 'script evaluation failed', leaving the
dev client with no SW — authenticated media 401'd (broken images in every
dev screenshot) and SW notification routing was untestable. Production
sw.js is unaffected.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPmy3tPq869XDW4njjVaKA
deleteSearchCacheDatabase() resolved after a 3 s "blocked" timeout while
another tab still held the DB, so decrypted rows could survive logout.
It now broadcasts lotus-logout first; every tab closes its handle and
refuses to reopen, then the delete proceeds. A boot with no session
re-runs the wipe once in case a race was still lost. Unit-tested.
Fixes#45
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PPmy3tPq869XDW4njjVaKA
Two console-noise / glitch fixes surfaced while a room with link
previews was open:
- URL preview og:image thumbnails 400 when Synapse can't thumbnail a
cached preview image (SVG/animated), leaving a broken image that the
browser keeps re-requesting. GenericCard now falls back to the full
image on error, then hides the image (and shows the link icon) if that
also fails, so no broken image and no repeated failing requests.
- Extend the existing console.warn filter to drop matrix-js-sdk's
high-volume, benign timeline bookkeeping warnings ("EventTimelineSet…"
and "Decrypted event … is not in room …"), which fire constantly in
E2EE rooms with threads. Real warnings still log.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- index.tsx: request navigator.storage.persist() for logged-in sessions so the
browser can't evict the IndexedDB rust-crypto store (eviction while the
localStorage session survives resurrects the device with a blank store → the
KE-1 "one time key already exists" upload storm). Guarded, checks persisted()
first, best-effort.
- Docs: remove HANDOFF_ELEMENT_CALL_FORK.md, LOTUS_E2EE_INVESTIGATION.md, and
LOTUS_BUGS.md. Port their live content into the three kept docs — verification
backlog → LOTUS_TESTING; open bugs + E2EE (KE-1..4) + an Element Call fork
operational reference (publish steps + io.lotus action catalog) → LOTUS_TODO.
Fix all dangling references (README, code comments, cross-doc links). Full
history of the removed docs remains in git.
Gates: tsc/eslint/prettier clean, build OK, 665 tests.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
We no longer use Sentry. Removed:
- @sentry/react + @sentry/vite-plugin (package.json + lockfile)
- Sentry.init in index.tsx and the VITE_SENTRY_DSN env (.env.production)
- @sentry/vite-plugin + the SENTRY_AUTH_TOKEN sourcemap-upload path in
vite.config.js (sourcemap now always false) and the CI env var
- Sentry.ErrorBoundary in App.tsx -> react-error-boundary's ErrorBoundary with a
folds-native fallback (Box/Text/Button + config tokens), which also resolves
the native-cinny audit's raw-#hex/#5865f2 fallback finding.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- CallEmbed: inject :root { color-scheme } into iframe so EC respects Cinny
theme regardless of OS preference (fixes white background in dark mode)
- CallEmbed: store themeKind, update color-scheme CSS on live setTheme() calls
- CallEmbed: catch transport.send() rejection in setTheme() to prevent
unhandled promise rejection when widget not ready yet (fixes REACT-8)
- CallEmbed: html + body both set to background:none so wallpaper shows through
- CallEmbedProvider: apply chatBackground wallpaper style to call embed
container in full-view mode (not PiP) -- wallpapers carry over to calls
- useCallEmbed: pass themeKind through to CallEmbed constructor
- index.tsx: ignoreErrors: [Request timed out] to suppress matrixRTC
heartbeat timeouts (REACT-9) from Sentry noise
- README: document 0.19.4, positioning fix, dark mode fix, wallpaper,
millify Rolldown interop fix, Sentry noise filter
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Synapse does not yet ship MSC3786/MSC3914 as server-default push rules.
matrix-js-sdk patches them client-side every login and warns. Filter these
at console.warn level -- functionality is unaffected.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When a new deploy lands while a tab is open, lazy-loaded chunks (like
GifPicker) disappear because their content-hash filename changes. Vite
dispatches a vite:preloadError event in this case. We reload once and
clear the flag on successful load so future deploys can trigger again.
browserTracingIntegration injects sentry-trace and baggage headers into all
outgoing fetch calls. Synapse does not list these in Access-Control-Allow-Headers,
so every Matrix API call was blocked by the browser CORS preflight check.
Removed browserTracingIntegration, set tracePropagationTargets:[] and
tracesSampleRate:0. Error capture (the useful part) is unaffected.
CSP fix (Sentry ingest domain) is applied via nginx — no code change needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Initialize Sentry SDK in index.tsx when VITE_SENTRY_DSN env var is set
- Wrap entire App with Sentry.ErrorBoundary (replaces the hard crash with a retry UI)
- 5% trace sample rate, sendDefaultPii disabled, strip events containing accessToken
- Add .env.production template with VITE_SENTRY_DSN placeholder
- Get your DSN from sentry.io -> Project Settings -> Client Keys
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* request session info from sw if missing
* fix async session request in fetch
* respond fetch synchronously and add early check for non media requests (#2670)
* make sure we call respondWith synchronously
* simplify isMediaRequest in sw
* improve naming in sw
* get back baseUrl check into validMediaRequest
* pass original request into fetch in sw
* extract mediaPath util and performs checks properly
---------
Co-authored-by: mmmykhailo <35040944+mmmykhailo@users.noreply.github.com>
* fix set power level broken after sdk update
* add media authentication hook
* fix service worker types
* fix service worker not working in dev mode
* fix env mode check when registering sw
* chore: Bump matrix-js-sdk to 34.4.0
* feat: Authenticated media support
* chore: Use Vite PWA for service worker support
* fix: Fix Vite PWA SW entry point
Forget this. :P
* fix: Also add Nginx rewrite for sw.js
* fix: Correct Nginx rewrite
* fix: Add Netlify redirect for sw.js
Otherwise the generic SPA rewrite to index.html would take effect, breaking Service Worker.
* fix: Account for subpath when regisering service worker
* chore: Correct types
Lotus CIandClaude Opus 5.5
jaredandClaude Opus 5
Lotus BotandClaude Sonnet 4.6
Ajay Buraandmmmykhailo
夜坂雅
aceArt-GmbHandAjay Bura