Compare commits

...

213 commits

Author SHA1 Message Date
7b3a4145a7 fix(channels): back active-workspace persistence with a jotai atom so the native pager sees switcher picks instead of a stale memoized localStorage read 2026-05-22 01:18:15 +03:00
765445c091 feat(discord-widget): render Open-in-Channels card after login via VOJO-LOGIN-SPACE-V1 sentinel and generic open-matrix-to widget action 2026-05-21 14:08:50 +03:00
408b9eefc3 chore(eslint): give widget apps their own Preact-aware root config so host airbnb and react rules stop flagging valid Preact code in pre-commit 2026-05-21 14:08:21 +03:00
8d8b39e897 fix(telegram-widget): stack password row column on mobile so the show toggle does not overflow off the right edge 2026-05-21 13:43:11 +03:00
055b7d3692 fix(workspace-switcher): switch space rows to SurfaceVariant so inactive bg blends with the sheet silhouette instead of reading as dark cards 2026-05-21 01:05:34 +03:00
78504262d3 feat(stream-header): flatten web curtain to a tabsRow-border divider pixel-aligned with PageHeader at WEB_TABS_ROW_PX=54 and gate keyboard probe to native 2026-05-20 23:20:17 +03:00
cde50cff0f fix(bots): pad BotStatePage wrapper with safe-top so the mobile back-arrow header clears the Android status bar on the connect-bot empty state 2026-05-20 02:39:21 +03:00
6ca6b69d48 feat(stream-header): contextual Plus on Channels opens create-channel inside workspace and create-community on landing via StreamHeader.primaryAction 2026-05-20 01:59:04 +03:00
240bb54c29 refactor(stream-header): reset live drag on gesture teardown, drop dead pinned-local fallback, narrow commit() to peek|closed, add exhaustive transition guards and align stale comments 2026-05-20 00:59:17 +03:00
8fb885df1b feat(stream-header): free-range curtain drag through full pin↔closed↔peek range with bottomPinned-aware body bail and native-only handle 2026-05-20 00:26:10 +03:00
ab283e9788 refactor(stream-header): unify curtain gestures onto dual handle+body surfaces with 1:1 handle and rubber-band body, scroll-aware bail 2026-05-19 23:27:19 +03:00
e866cd3830 feat(stream-header): move pin/unpin gesture onto dedicated 32px drag-handle with 1:1 finger tracking and desktop-style grabber animation 2026-05-19 18:26:37 +03:00
7c5a1f2ee7 fix(mobile-tabs-pager): restore segment-button taps in pager mode via opacity-0 per-pane row and route their commits through an instant no-transition strip jump 2026-05-19 14:28:56 +03:00
0422a9832f feat(stream-header): pin chats curtain over static pager header on drag-up with per-tab atom, native-only rubber-band gesture and pinned-aware horseshoe sheet coordination 2026-05-19 11:50:31 +03:00
4a9d5f6384 fix(mobile-tabs-pager): paint Background.OnContainer on pagerRoot so panes mounted outside PageRoot inherit readable text on native 2026-05-18 22:58:53 +03:00
870e13d895 feat(mobile-tabs-pager): swipe between Direct, Channels and Bots on Capacitor native with static header, 24px gap, atom-bridged action icons and inert offscreen panes 2026-05-18 22:00:53 +03:00
727a53a776 fix(horseshoe): extend mobile DM and Channels wrappers up over the safe-top zone so the StreamHeader curtain paints the status-bar strip on drag-up 2026-05-18 15:14:58 +03:00
af97549e48 tweak(stream-header): require curtain drag past 90% of full peek travel to commit so short drags snap back as accidental 2026-05-18 02:12:55 +03:00
0c704aac38 tweak(bubbles): paint peer bg and horseshoe void pure black and flip own/peer flat-corner to bottom-left/top-left at 16px radius 2026-05-18 02:04:39 +03:00
b26340fa7d feat(electron): add desktop wrapper packaging Vojo as Windows zip with privileged vojo:// scheme, HashRouter override and native chrome 2026-05-18 01:50:16 +03:00
5dbe83aa9d feat(push): real sender+room avatars via MXC bridge with adaptive shortcut icons, plus review fixes (GROUP_ALERT_ALL, eventId dedup, isEncrypted privacy default, structured roomname parse, mark-as-read optimistic docs) 2026-05-17 17:39:19 +03:00
4b4454fa1d docs(android): document MessagingStyle pipeline, channel split, callId-session dedup and edit-collapse defer rationale 2026-05-17 02:42:45 +03:00
de348eb4fc fix(push): close E2EE-flip race by re-checking room encryption at reply send time, pre-flight credentials before optimistic echo, share callId-dedup helpers between FCM and Worker 2026-05-17 02:41:27 +03:00
06778702b2 feat(push): inline RemoteInput reply on per-room notifications for cleartext rooms with optimistic local echo and encryption-state re-dump 2026-05-17 02:29:20 +03:00
8a80194fe5 fix(push): dedup re-rings within one call session via composite (roomId, callSessionId) key so a participant rejoin doesn't re-alert 2026-05-17 02:22:20 +03:00
38d24e5527 feat(push): group room messages into a per-room MessagingStyle conversation with DM/group channels, mark-as-read action and receipt-driven dismiss 2026-05-17 02:06:21 +03:00
408f165f60 feat(push): add WorkManager polling fallback that delivers notifications via /_matrix/client/v3/notifications when FCM is blocked 2026-05-17 01:27:55 +03:00
b9aad691b5 update gitignore with Vite config 2026-05-16 22:06:35 +03:00
770609b964 feat(media): make the desktop right-side media pane resizable with a smart max that subtracts chat-list width and a chat-column reservation 2026-05-16 20:45:54 +03:00
ebf2cfe07b fix(sidebar): raise resizable page-nav min width from 320 to 384px so StreamHeader segments and action icons stop clipping at the floor 2026-05-16 20:31:11 +03:00
2d101a40fc fix(channels): collapse /channels/ index into one mobile pane and add Create-community CTA next to Find-community on empty state 2026-05-16 20:25:01 +03:00
f2ecca64da feat(share): receive Android system share intents and drop the payload into the next chat the user opens via a top banner cue 2026-05-16 19:33:06 +03:00
6982ec374e chore(lint): close all typecheck and eslint tech debt to enable husky pre-commit hook with --max-warnings 0 2026-05-16 17:22:53 +03:00
45c69317ff feat(message): paint non-own bubbles via --vojo-peer-bubble-bg in Stream + Channel layouts and Stream rail/day-divider via --vojo-timeline-rail 2026-05-16 13:13:28 +03:00
c78984a6d8 fix(direct): exclude bridged portal rooms from Direct tab so Telegram chats live exclusively in their per-bridge personal filtering space 2026-05-16 00:31:38 +03:00
bfd72dc1ff tweak(stream-header): collapse peek1/peek2 into single peek snap so one drag reveals both action chips at once 2026-05-15 23:47:40 +03:00
0eb2e056c0 feat(stream-header): rebuild Direct/Channels/Bots header as a curtain layered above tabs with peek chips, inline forms, and VisualViewport keyboard compensation 2026-05-15 23:05:34 +03:00
81d23be61f feat(theme): ship Settings picker with system/light/dark and Vojo light palette reshading sidebar, chat, bubbles, horseshoe void and PWA chrome 2026-05-15 01:06:49 +03:00
8e2db986b4 feat(composer): tighten action-row padding, extract button JSX, and rotate placeholder across 12 hour-keyed variants 2026-05-15 00:13:51 +03:00
646cb7b124 feat(members): carve rounded TL/BL on members drawer with 12px void seam to chat and extract VoidGap helper consolidating four per-pane seams 2026-05-14 21:11:36 +03:00
8400ef54ee feat(syslines): render membership and room-state events as sender-anchored chat bubbles via StreamLayout instead of thin rail syslines 2026-05-14 01:39:51 +03:00
a893e86d92 tweak(direct): dial DM list avatars down to 48px with 68px row height for a less heavy density 2026-05-14 01:12:17 +03:00
2d74848509 feat(direct): enlarge DM list avatars to size 500 with 80px row height and matched virtualizer estimate for a denser two-line layout 2026-05-14 01:05:32 +03:00
c3a384b651 fix(page-nav): hold min-height:0 on scrolling middle and shrink-no on Direct bottom rows so flex pressure no longer squashes them below their NavItem floor 2026-05-14 01:05:21 +03:00
3c7c79fb6c fix(calls): split per-session bubbles by joined-count boundary with expiry-aware ongoing, post-ring duration, and same-caller retry merging 2026-05-14 00:27:38 +03:00
f5e992daad fix(time): use Intl numeric day-month-year everywhere so chat day dividers follow the system locale instead of hardcoded English full-month 2026-05-13 23:30:52 +03:00
2dac76f9af chore(i18n): rename DM direct-stream segment label to Direct 2026-05-13 23:30:42 +03:00
1ee1d50c41 fix(dm-name): drop matrix-js-sdk mxid disambiguation suffix from DM room names by using peer rawDisplayName 2026-05-13 22:54:26 +03:00
d3e69e042f fix(i18n): track system language on every launch instead of caching first-seen value in localStorage 2026-05-13 22:54:06 +03:00
5d4a50f593 feat(legal): publish Privacy Policy and account-deletion pages with About-screen link and Play Store feature graphic 2026-05-13 22:53:58 +03:00
5c16649e0c chore(versioning): derive Android versionName from git describe to match __APP_VERSION__ from vite 2026-05-13 22:53:48 +03:00
30e477d2cd feat(calls): render m.call.member events as one aggregate chat bubble per call aligned to initiator side 2026-05-13 15:57:23 +03:00
6a6b7acf15 feat(channels): rebuild thread drawer and channel rows as chat-style bubble cards that merge thread-summary into the bubble footer 2026-05-13 15:21:13 +03:00
3dee9f099f feat(thread-drawer): wrap in horseshoe seam with rounded TL/BL and add pointer/keyboard resize clamped to viewport/3 2026-05-13 14:42:15 +03:00
11c46d9250 feat(channels): replace workspace switcher popout with sliding horseshoe sheet and inline create-channel row, retire sidebar CreateTab 2026-05-13 14:21:39 +03:00
c27f8a7cc2 chore(mascot): re-encode as 2s VP9 cycle and center auth drop-shadow on character 2026-05-13 04:17:14 +03:00
4e7ea405ab chore(android): enable R8 shrinker with keep rules, set up release signing, and strip sourcemaps from APK 2026-05-13 04:17:09 +03:00
ffb80bff88 chore(bundle): trim web bundle by dropping unused Twemoji TTF, Prism languages, and MediaPipe blur wasm 2026-05-13 04:17:03 +03:00
663aece487 feat(media-viewer): atom-driven horseshoe shell over chat replacing Overlay modal for image+video with anchor-aware pinch/wheel zoom and swipe prev/next 2026-05-13 02:47:04 +03:00
4654836092 feat(chat): hide composer on scroll-up past 200px and replace jump-to-latest chip with circular FAB that pulses on incoming live messages 2026-05-13 01:36:29 +03:00
635fb91022 feat(settings): replace Modal500 with /settings route plus mobile bottom-up horseshoe sheet overlaying DM list via clip-path mask 2026-05-13 00:01:26 +03:00
c6bb66958d feat(profile-rail): size mobile user-card rail to measured content height with 85vh cap and inner scroll only on overflow 2026-05-12 02:06:10 +03:00
149382299a feat(safe-area): extend Android edge-to-edge top inset via --vojo-safe-top var and collapse profile horseshoe header with measured height 2026-05-12 01:54:30 +03:00
ce82d66883 feat(safe-area): paint Android edge-to-edge top/bottom strips with the active surface tone instead of a fixed body bg 2026-05-11 23:13:56 +03:00
f38cb42344 feat(direct-tabs): spread DM/Channels/Bots tabs edge-to-edge and underline active segment with purple bar on the header's grey rule 2026-05-11 21:04:07 +03:00
785b679b61 feat(horseshoe): add 12px void gap between chat and profile pane with rounded TL/BL on profile, drop page-nav right rounding 2026-05-11 20:33:51 +03:00
de2354f1da fix(composer): re-anchor timeline scrollTop when overlay composer height changes via prop-driven layout effect 2026-05-11 18:10:56 +03:00
41a9af19e3 feat(composer): floating overlay above timeline with Gemini-style two-row layout and Android-WebView stuck-hover gate 2026-05-11 17:47:37 +03:00
2337b05140 feat(chat): invert canvas/bubble tones and darken timeline rail and media frames 2026-05-11 15:16:53 +03:00
ab6c65a4e0 fix(profile): keep desktop user-room-profile pane open on outside clicks, close only via × or Esc 2026-05-11 14:09:07 +03:00
e3e61afd4c feat(profile): merge 1:1 header avatar and title into single identity button so name taps open the profile sheet 2026-05-11 14:09:07 +03:00
4d0b508ebb feat(horseshoe): split web page-nav and chat panel with 12px void gap and rounded inner corners on both sides 2026-05-11 13:54:40 +03:00
023a6a439c feat(sidebar): add resizable left page-nav via pointer/keyboard with localStorage-persisted width, clamped [320, viewport/3] and tactile min/max indicator 2026-05-11 02:49:39 +03:00
c992e910ee feat(profile): add mouse drag for mobile horseshoe via pointer events with origin-tagged drag state and userSelect:none on viewports 2026-05-11 02:22:28 +03:00
2bbaf4dfcf fix(profile): override avatar circle force via && self-doubled selector instead of source-order-dependent !important double-up 2026-05-11 02:00:12 +03:00
117bb9fba4 refactor(profile): rebuild mobile horseshoe as single silhouette wrapper with Vaul ease curves replacing the two-radius emerge handoff 2026-05-11 02:00:04 +03:00
626a7c2d1d feat(profile): inline-expand hero avatar in desktop side pane instead of swapping the whole card to full-view 2026-05-11 01:11:09 +03:00
9c204c1af6 feat(profile): drop pill background on user-card kebab and enlarge the dots icon 2026-05-11 00:58:39 +03:00
4b39046c09 feat(horseshoe): bump top profile and bottom call horseshoe radius to 32px with 12px void gap 2026-05-11 00:54:58 +03:00
3fed5ff873 chore(layout): disable SidebarNav 66px rail rendering; component preserved for later entry-point redistribution 2026-05-10 22:12:31 +03:00
d4b05619a8 feat(channels): ship M6 workspace switcher dropdown with rename-reactive trigger and rows for multi-space users 2026-05-10 20:56:14 +03:00
0b31e6b930 feat(channels): route room-in-space navigations through /channels/ with eventId-anchored permalinks and channels-aware cold-push redirect 2026-05-10 20:13:50 +03:00
896a2e2083 feat(channels): M4a drawer rich chrome with edit menu reactions and reply affordances 2026-05-10 18:48:55 +03:00
e80453785e feat(channels): ship M4 per-thread unread with thread receipts mute-aware atom and architectural cleanup 2026-05-10 14:35:51 +03:00
307af24d1e feat(channels): ship M3 channel timeline avatar-name layout with thread summary cards and drawer header counter polish 2026-05-10 01:06:29 +03:00
4632be30f7 feat(channels): ship M2 thread drawer + composer + shareable thread URL with cold-load relations repair 2026-05-09 22:49:53 +03:00
851f3d30a3 feat(channels): drop NEW badge from Каналы segment per product call 2026-05-09 15:09:24 +03:00
075b6cf69c fix(android): redirect to root on logout to dodge Capacitor SPA-fallback failure on URL-encoded path segments 2026-05-09 15:07:35 +03:00
efe58dc2e2 feat(channels): ship M1 — Каналы segment with /channels/ routes and channels-mode RoomTimeline filter 2026-05-09 15:06:13 +03:00
c6eba1e935 chore(deps): bump matrix-js-sdk 40.2.0 → 41.4.0 with SessionMembershipData barrel and EC sticky-event capability extension 2026-05-09 13:00:45 +03:00
3ea01a9c3f chore(deps): bump matrix-js-sdk 39.4.0 → 40.2.0 adapting sessionMembershipsForRoom removal and dedup-key trichotomy 2026-05-09 12:18:02 +03:00
0d93a223d0 chore(deps): bump matrix-js-sdk 38.4.0 → 39.4.0 adapting MatrixRTCSession.slotDescription rename 2026-05-09 11:33:55 +03:00
6560f0b424 chore(deps): bump matrix-js-sdk 38.2.0 → 38.4.0 2026-05-09 02:45:59 +03:00
b30704dd96 fix(boot): drop 3-dots splash menu, add 10s fetch timeouts, surface logout on init/start/sync-error per Element Web pattern 2026-05-09 02:26:16 +03:00
cd824e0c90 feat(profile): mobile top-horseshoe rail and desktop right pane host a Dawn-style user card with hero, info rows, and floating 3-dot actions menu 2026-05-08 19:04:12 +03:00
f4292611cf feat(calls): split-horseshoe call surface with redesigned ring/active pill, orbit border, custom outline icons, tap-to-room 2026-05-08 01:30:36 +03:00
d58e69d49f refactor(timeline): rebuild stream row as 3-track CSS grid with auto-sized time column so 24h and AM/PM render with consistent gaps without JS measurement 2026-05-07 23:36:18 +03:00
997375b307 feat(timeline): square image+video bubbles with username overlay, reactions outside bubble, edge-anchored mobile rail, horizontal day divider 2026-05-07 21:24:50 +03:00
ce308776cd update docs 2026-05-06 23:12:57 +03:00
97a50e29f9 feat(bots): add bot-widget loading bar with cycle-complete hide and sync-state deference, plus matching cycle-complete polish on SyncIndicator 2026-05-06 22:11:42 +03:00
17ba496b7e fix(bots-widgets): default data-input to mouse and drop dead danger:hover rose override so hover works from frame zero on hybrid devices 2026-05-06 17:29:11 +03:00
998813eff4 refactor(direct): drop the bottom DM-nav status strip with the vojo.chat label and e2ee chip 2026-05-06 16:37:22 +03:00
ece9e922e3 fix(user-profile): round the avatar surface plate and replace its outline with a box-shadow ring so corners no longer poke past the circle 2026-05-06 16:19:47 +03:00
7af69574f4 feat(bots): render bot-shell command cards as a fixed 2-column grid with equal-height rows 2026-05-06 16:19:41 +03:00
e8865cec5f fix(direct): lock DmStreamRow title-block height so the room name stops jumping when row 2 collapses on hover 2026-05-06 16:19:34 +03:00
1d64275bae feat(bots-discord): land hCaptcha challenge handling for QR-login with sentinel-prefixed bridge protocol and Dawn-themed widget UI 2026-05-06 14:19:45 +03:00
295dfbb796 feat(bots-discord): drop the obsolete Devices step from the QR-scan navigation copy 2026-05-05 22:59:39 +03:00
526515dcde feat(bots-whatsapp): drop the unpredictable-bans paragraph from the AboutCard risk callout 2026-05-05 22:53:27 +03:00
42d9ccfbf3 feat(bots): unify command-card chrome with left-side semantic icons and fold WA Meta-ToS warning into the AboutCard modal 2026-05-05 19:59:24 +03:00
bc360e84cc feat(bots-whatsapp): land Preact widget for mautrix-whatsapp QR + pairing-code login, Meta-ToS warning card, and cross-iframe external-link relay 2026-05-05 15:25:16 +03:00
5eb12f888b feat(bots-discord): land Preact widget for mautrix-discord QR-login with ping-based status, reconnect recovery, and discordapp.com URL parser 2026-05-05 02:17:30 +03:00
156570826a feat(bots-telegram): land QR-code login flow rendered client-side from m.image body via qrcode-generator with bridge-race-tolerant state machine 2026-05-05 01:02:36 +03:00
e46bba2f7d feat(bots): polish the Telegram bot widget UI and fix Android WebView sticky-hover via pointerType-based input-mode detection 2026-05-04 18:34:51 +03:00
b2f3b668c5 feat(splash): hold Android system splash on screen until web mascot paints via custom LaunchSplash plugin 2026-05-04 18:31:10 +03:00
817dad383c remove stale plan 2026-05-03 23:56:28 +03:00
949860bc1a feat(connection): replace 30s 'Connecting...' banner with bottom-edge sync indicator and resume-grace window for phone-unlock UX 2026-05-03 22:35:22 +03:00
f102593081 Revert "feat(connection): replace 30s 'Connecting...' banner with bottom-edge sync indicator and homeserver footer dot, drop mascot loading splash"
This reverts commit a1ff5db724.
2026-05-03 20:15:26 +03:00
e547c466a8 feat(settings): drop user-facing time/date format toggle and derive everything from system locale via Intl.DateTimeFormat 2026-05-03 19:27:54 +03:00
a1ff5db724 feat(connection): replace 30s 'Connecting...' banner with bottom-edge sync indicator and homeserver footer dot, drop mascot loading splash 2026-05-03 18:21:00 +03:00
8f49124043 feat(ui): force every user and room avatar to render as a circle via globalStyle override on the folds Avatar wrapper 2026-05-03 14:48:27 +03:00
ed1544dd5e feat(direct): land inline invite cards with spam-filter toggle and retire the /inbox/ tree along with its sidebar tab 2026-05-03 13:49:33 +03:00
bae6761683 feat(bots): render Matrix-native bot avatar in BotCard sidebar row and BotShellHero so server-side avatar_url propagates without client patches 2026-05-03 13:22:10 +03:00
316c3eb9fd feat(bots-telegram): land M12.5 timeline-resume hydrate to recover pending login forms after widget reload via read_events scan 2026-05-03 02:36:17 +03:00
35ade7e941 feat(bots-telegram): ship M12 login flow with BotShell host hero and Go bridgev2 dialect parser 2026-05-02 22:12:37 +03:00
e43b0fb597 feat(bots-telegram): land Phase 3 widget scaffold with Dawn UI, dev config overlay, and prod origin allowlist 2026-05-02 13:22:25 +03:00
d961dddfbc feat(bots): land Phase 2 widget host/driver with retry UX and route-aware notifications 2026-05-02 00:44:52 +03:00
83e246da1f Add runtime-configured bot tab 2026-05-01 20:21:55 +03:00
357a2024f4 feat(bots): M1 wire Direct segment to /bots/ placeholder and rename label to Роботы 2026-05-01 14:42:00 +03:00
96085ba6a1 Update auth footer branding 2026-04-29 23:12:08 +03:00
b5ea37d57a feat(direct): use compact native phone layout for new chat to keep form fitted under the on-screen keyboard 2026-04-29 22:30:39 +03:00
212d3e3482 feat(invite): split user id into username and server fields and close prompt on successful invite 2026-04-29 21:45:33 +03:00
3cd1611ee2 fix(call): restore Android CallStyle banner for DM voice calls in encrypted rooms 2026-04-29 13:45:13 +03:00
d2c77496a7 redesign(p4): land Dawn RoomViewHeader for all rooms with peer chrome, presence, member-count subline, and reactive bridge gate. 2026-04-29 01:03:12 +03:00
103d6ad8a1 redesign(p3c): collapse Home into universal Direct, drop legacy layouts, gate room flavour on member-count, and clear orphan settings. 2026-04-28 21:52:31 +03:00
5bf0aeb00b redesign(p3a): land Stream message layout for DMs with rail, author dots, asymmetric bubbles, Stream day-divider, and sysline state events 2026-04-28 00:54:53 +03:00
e230e688de chore: upgrade TypeScript to 5.4 with bundler module resolution and reformat repo against tightened ESLint 2026-04-27 13:07:49 +03:00
ed3e5c0640 redesign(p2): rebuild DM list panel with stream header, segmented tabs, self-row, new-chat row, footer status, and live timeline rerender 2026-04-26 23:34:20 +03:00
0c89e9fda0 redesign(p0): land Dawn dark-theme foundation with one-shot migration and edge-to-edge polish 2026-04-26 21:30:22 +03:00
52bf23cf38 update docs 2026-04-26 19:48:13 +03:00
1c079ddca2 fix(nav): collapse push-tap, tab and back-arrow navigations into the back stack via replace 2026-04-26 00:35:26 +03:00
dbda8728a8 fix(auth): keep form within visible band on small viewports 2026-04-25 23:43:11 +03:00
6ab905a7c3 fix(auth): stabilize compact auth form scrolling 2026-04-25 20:18:38 +03:00
6c65dee82e Replace 'following' banner with WhatsApp-style delivery status checkmarks on own messages. 2026-04-25 17:49:51 +03:00
3652320b0f Split DM create form into separate username and server fields with smart defaults. 2026-04-25 15:41:29 +03:00
0466e78233 Simplify theme settings to a single dropdown with System, Light, and Dark options. 2026-04-25 13:36:40 +03:00
6fadff26c2 Use commit count as patch version instead of raw git describe output. 2026-04-25 12:53:18 +03:00
c164dacf18 Stop tracking docs/ai/desired_features.md, already in .gitignore. 2026-04-25 12:38:26 +03:00
ff9d15f930 Auto-scroll chat timeline to bottom when Android keyboard opens. 2026-04-25 12:35:02 +03:00
adb0012834 Fix DM rooms showing as regular rooms for invited users by syncing m.direct on join and routing push navigation correctly. 2026-04-25 12:04:30 +03:00
f96c80f829 Localize room intro, membership events, call controls, and leave-room dialog across en/ru via i18next. 2026-04-25 01:28:07 +03:00
8463180c04 Move push_strings.xml generation into Gradle AGP task so XMLs are never committed and direct gradle builds work without npm run android:sync. 2026-04-25 00:36:07 +03:00
15f86bfcae Localize invite push notifications across web SW and Android FCM via shared JSON Push namespace. 2026-04-24 23:51:07 +03:00
d15a3b336b Harden DM call teardown with io.element.close listener, 8s hangup timeout, and same-room zombie escape so stale embeds stop blocking retries. 2026-04-24 21:20:05 +03:00
cf0bf56541 Replace FCM foreground-skip cache with Java ring registry to fix silent ring on mid-ring backgrounding and eliminate dual dismiss plane 2026-04-24 01:47:03 +03:00
a35dfb1a5b Use Capacitor pause/resume events on Android to gate appActive at the same lifecycle edge as MainActivity.isInForeground. 2026-04-23 22:00:26 +03:00
aaebdffc4d Track declined notification IDs and re-check them after resolveCallId await to block stale rings when a decline lands during the async yield. 2026-04-23 19:57:27 +03:00
fb6f7bd982 Gate native CallStyle dismiss on App.isActive so background incoming rings keep ringing instead of being prematurely cleared by /sync atom ADD. 2026-04-23 01:27:51 +03:00
0c30e37b70 Gate incoming-call ring audio on App.isActive to stop double-ring in the brief window after app backgrounding 2026-04-23 00:50:55 +03:00
0692c05408 Add microphone foreground service to keep Android DM calls alive under lock screen 2026-04-23 00:04:16 +03:00
ff1bf9c377 Enforce DM call switching and foreground native call ownership 2026-04-22 21:27:54 +03:00
6b8228cdca user dm links 2026-04-21 21:11:32 +03:00
13ec1e0054 dm calls mvp: phase 5.35 polish: cancel notif on null session, redact token-leak in JSON parse log, guard drain from resume race 2026-04-20 23:41:29 +03:00
91e3dd95ec add automatic app version and park git init tag from vojo/dev 2026-04-20 22:58:38 +03:00
cb16edbf37 dm calls mvp: phase 2.5.3: route call-notif FSI launch and body-tap to Direct tab instead of Home 2026-04-20 22:24:07 +03:00
63e9c65dd2 dm calls mvp: phase 5.35: decline via BroadcastReceiver with session bridge, pending-declines flusher, allowBackup=false 2026-04-20 22:23:44 +03:00
53a8bda18f dm calls mvp: phase 2.5.3: lockscreen CallStyle with FSI, Answer/Decline, ringtone channel, permission prompt 2026-04-20 22:22:43 +03:00
aa958b6e76 dm calls mvp: phase 2.5.2: drop dead code after SW-only pivot (invite.ogg, usePreviousValue, usePermission, notificationPermission) 2026-04-19 21:34:04 +03:00
6a690d4ecc dm calls mvp: phase 2.5.2: SW push is the only OS notification channel; in-app banners removed, calls bypass the visible-tab gate 2026-04-19 21:32:41 +03:00
6ced8246e6 dm calls mvp: phase 2: handle ring/decline + timeline render in encrypted DMs 2026-04-19 15:26:37 +03:00
79bd0ccc4d dm calls mvp: phase 2: incoming call strip + A-side auto-hangup on decline, peer-leave, no-answer 2026-04-19 13:59:33 +03:00
f862f19f09 dm calls mvp: phase 2: mvp of accept/decline calls (contain bugs) 2026-04-19 10:11:09 +03:00
dfcd69f1df dm calls mvp: phase 1: outgoing DM voice call via element-call voiceOnly intent 2026-04-18 23:31:42 +03:00
30c59e199d dm calls mvp: phase 0: fix CallEmbed listener leak by caching bound refs in class fields 2026-04-18 21:32:07 +03:00
380a4d7d70 update vive code tools with plans dir 2026-04-18 20:02:13 +03:00
20ac745ce9 workaournd about boxes stable size when stretching up auth page 2026-04-18 15:01:57 +03:00
03267cfc35 fix fonts/white boxes when autofll auth page 2026-04-18 14:05:13 +03:00
02d9c8cc76 remove bugs md from git index & update gitignore 2026-04-18 01:43:11 +03:00
5947cd5b82 update background colors 2026-04-18 01:39:28 +03:00
3717adb52e feature with back button on native app 2026-04-18 01:30:48 +03:00
46981791a3 update gitignore with desired features 2026-04-18 01:00:33 +03:00
ce6afe1de8 add server side doc & some info about config files 2026-04-17 23:46:24 +03:00
c7c47ec23a move all ai docs to folder 2026-04-17 23:42:04 +03:00
6fbcf94cd9 push notifications permission on start app 2026-04-17 23:31:21 +03:00
46659625cd add notifications support for android (mostly) 2026-04-17 22:54:44 +03:00
4c64b52827 update claude 2026-04-16 02:22:12 +03:00
9fe5bf3480 disable by default chat encryption 2026-04-16 02:04:47 +03:00
2b479bb7c6 default messages box in direct messages: bubble style 2026-04-16 02:02:03 +03:00
85829b2066 update claude with project specs 2026-04-16 01:57:06 +03:00
5e9560570f fix scrolling login page on native app 2026-04-16 01:55:03 +03:00
da9ead7db7 fix jumping login window layout 2026-04-16 01:47:13 +03:00
3c613d9960 icons for native capacitor webview app 2026-04-16 01:25:37 +03:00
3f6e0e0d45 add webview capacitor support 2026-04-16 01:14:09 +03:00
ac452aaf1d update icons with evil mascot 2026-04-15 00:57:56 +03:00
fdf447c9cb rework login page 2026-04-14 23:19:44 +03:00
18e09b81a2 change mascot background 2026-04-14 22:32:30 +03:00
9ac226002c localize logout 2026-04-14 22:27:23 +03:00
d933587d09 update colors for dark theme 2026-04-14 22:11:31 +03:00
c7a2d723ec update claude md with project archicture 2026-04-14 22:01:17 +03:00
becdbc3d3a localize room settings 2026-04-14 21:27:03 +03:00
d120a9a933 localize Add Space 2026-04-14 02:18:51 +03:00
f4e94a48cc localize Explore Community 2026-04-14 01:59:41 +03:00
ef0fa23642 localize Inbox/Notifications 2026-04-14 01:41:22 +03:00
77fda29820 localize direct messages 2026-04-14 01:20:26 +03:00
fc8346e50d localize home 2026-04-14 00:30:08 +03:00
7f4a013eb5 search localization 2026-04-14 00:09:24 +03:00
d4ea8b525f localize settings 2026-04-13 23:36:03 +03:00
96613c7a76 locale for register form 2026-04-13 22:39:59 +03:00
4539ab5114 rebrending contination: add icons 2026-04-13 22:22:01 +03:00
ff133b22b5 rebranding from cinny to vojo: change basic icons and naming 2026-04-13 02:22:44 +03:00
ae650b94b8 fix layout for login page with styles css 2026-04-13 00:47:20 +03:00
303d3c74c1 feat(auth): rebrand auth pages to Vojo with mascot, glassmorphism, and i18n
Rework the entire authentication UI to match the Vojo brand:

- Add mascot video with purple gradient halo behind the auth form
- Glassmorphism card (backdrop-blur, semi-transparent bg) with JS-driven
  layout ported from element-web (ResizeObserver + requestAnimationFrame)
- Custom folds theme overrides via color token references (not hardcoded
  CSS variable hashes) for transparent inputs and white primary button
- Server edit modal dialog replacing browser prompt, with proper
  role="dialog", aria-modal, and Escape key support
- Footer: "Powered by Matrix · Hosted on Yandex Cloud"

Localization:
- Add ru.json, update en.json and de.json with all auth page keys
- Wire up react-i18next t() across all auth components
- Set fallbackLng to 'ru' while preserving LanguageDetector for en/de

Cleanup:
- Remove SSO login flow (SSOLogin, getSSOFlow, SSO rendering)
- Remove token login flow (TokenLogin.tsx, getTokenFlow, loginToken param)
- Strip unused imports (useState, usePathWithOrigin, useClientConfig)
- Fix ESLint: nested ternary → if-return, consistent-return, a11y

Config: vojo.chat as default and only homeserver.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-13 00:27:34 +03:00
06a94c3f54 update gitignore 2026-04-13 00:14:04 +03:00
389a396507 update git ignore with vscode 2026-04-12 21:59:32 +03:00
677 changed files with 73528 additions and 17221 deletions

View file

@ -1,2 +1,3 @@
experiment experiment
node_modules node_modules
*.css

View file

@ -4,15 +4,15 @@ module.exports = {
es2021: true, es2021: true,
}, },
extends: [ extends: [
"eslint:recommended", 'eslint:recommended',
"plugin:react/recommended", 'plugin:react/recommended',
"plugin:react-hooks/recommended", 'plugin:react-hooks/recommended',
"plugin:@typescript-eslint/eslint-recommended", 'plugin:@typescript-eslint/eslint-recommended',
"plugin:@typescript-eslint/recommended", 'plugin:@typescript-eslint/recommended',
'airbnb', 'airbnb',
'prettier', 'prettier',
], ],
parser: "@typescript-eslint/parser", parser: '@typescript-eslint/parser',
parserOptions: { parserOptions: {
ecmaFeatures: { ecmaFeatures: {
jsx: true, jsx: true,
@ -20,53 +20,80 @@ module.exports = {
ecmaVersion: 'latest', ecmaVersion: 'latest',
sourceType: 'module', sourceType: 'module',
}, },
"globals": { globals: {
JSX: "readonly" JSX: 'readonly',
__APP_VERSION__: 'readonly',
}, },
plugins: [ plugins: ['react', '@typescript-eslint'],
'react',
'@typescript-eslint'
],
rules: { rules: {
'linebreak-style': 0, 'linebreak-style': 0,
'no-underscore-dangle': 0, 'no-underscore-dangle': 0,
"no-shadow": "off", 'no-shadow': 'off',
"import/prefer-default-export": "off", 'import/prefer-default-export': 'off',
"import/extensions": "off", 'import/extensions': 'off',
"import/no-unresolved": "off", 'import/no-unresolved': 'off',
"import/no-extraneous-dependencies": [ 'import/no-extraneous-dependencies': [
"error", 'error',
{ {
devDependencies: true, devDependencies: true,
}, },
], ],
'react/no-unstable-nested-components': [ 'react/no-unstable-nested-components': ['error', { allowAsProps: true }],
'react/jsx-filename-extension': [
'error', 'error',
{ allowAsProps: true },
],
"react/jsx-filename-extension": [
"error",
{ {
extensions: [".tsx", ".jsx"], extensions: ['.tsx', '.jsx'],
}, },
], ],
"react/require-default-props": "off", 'react/require-default-props': 'off',
"react/jsx-props-no-spreading": "off", 'react/jsx-props-no-spreading': 'off',
"react-hooks/rules-of-hooks": "error", 'react-hooks/rules-of-hooks': 'error',
"react-hooks/exhaustive-deps": "error", 'react-hooks/exhaustive-deps': 'error',
"@typescript-eslint/no-unused-vars": "error", // Disable base rules in favour of their @typescript-eslint counterparts —
"@typescript-eslint/no-shadow": "error" // the base rules can't see TS-specific constructs (interface members, type
// imports, etc.) and double-fire alongside the TS versions.
'no-unused-vars': 'off',
'@typescript-eslint/no-unused-vars': ['error', { argsIgnorePattern: '^_' }],
'@typescript-eslint/no-shadow': 'error',
// Policy: kept as `warn` at the rule level so editors / `eslint --fix` /
// ad-hoc runs surface them as warnings, but `npm run check:eslint` and
// `lint-staged` BOTH pass `--max-warnings 0`, so new occurrences block
// commit. When unavoidable (matrix-js-sdk boundary, generic helpers,
// third-party callback shapes), suppress on the line with
// `// eslint-disable-next-line` and a one-line justification.
'@typescript-eslint/no-explicit-any': 'warn',
'@typescript-eslint/no-non-null-assertion': 'warn',
}, },
overrides: [ overrides: [
{ {
files: ['*.ts'], files: ['*.ts', '*.tsx'],
rules: { rules: {
'no-undef': 'off', 'no-undef': 'off',
}, },
}, },
{
// Upstream-vendored binary parsing copied verbatim from matrix-react-sdk
// (src/util/cryptE2ERoomKeys.js header link). Bitwise ops, post-increment
// and string concatenation are correct for the domain — clean-up risks
// breaking E2E room-key import/export. Keep the body byte-identical to
// upstream and disable only the rules that fire on those idioms.
files: ['src/util/cryptE2ERoomKeys.js'],
rules: {
'no-bitwise': 'off',
'no-plusplus': 'off',
'prefer-template': 'off',
'no-param-reassign': 'off',
// `for (;;)` form upstream uses for the iter-loops trips eslint
// even though it's intentional — keep upstream control flow.
'no-constant-condition': 'off',
// Diagnostic `console.log` left as-is in vendor copy.
'no-console': 'off',
},
},
], ],
}; };

View file

@ -4,7 +4,7 @@ body:
attributes: attributes:
value: | value: |
> [!IMPORTANT] > [!IMPORTANT]
> Please read through [the Discussion rules](https://github.com/cinnyapp/cinny/discussions/2653) and check for both existing [Discussions](https://github.com/cinnyapp/cinny/discussions?discussions_q=) and [Issues](https://github.com/cinnyapp/cinny/issues?q=sort%3Areactions-desc) prior to opening a new Discussion. > Please check for both existing Discussions and Issues prior to opening a new Discussion.
- type: markdown - type: markdown
attributes: attributes:
value: "# Issue Details" value: "# Issue Details"
@ -26,7 +26,7 @@ body:
attributes: attributes:
label: Expected Behavior label: Expected Behavior
description: | description: |
Describe how you expect Cinny to behave in this situation. Describe how you expect Vojo to behave in this situation.
placeholder: | placeholder: |
I expected the message to appear in the room timeline immediately after sending. I expected the message to appear in the room timeline immediately after sending.
OR OR
@ -37,7 +37,7 @@ body:
attributes: attributes:
label: Actual Behavior label: Actual Behavior
description: | description: |
Describe how Cinny actually behaves in this situation. If it is not immediately obvious how the actual behavior differs from the expected behavior described above, please be sure to mention the deviation specifically. Describe how Vojo actually behaves in this situation. If it is not immediately obvious how the actual behavior differs from the expected behavior described above, please be sure to mention the deviation specifically.
placeholder: | placeholder: |
The application freezes for 3 seconds and then shows a white screen. The application freezes for 3 seconds and then shows a white screen.
validations: validations:
@ -48,7 +48,7 @@ body:
description: | description: |
Provide a detailed set of step-by-step instructions for reproducing this issue. Provide a detailed set of step-by-step instructions for reproducing this issue.
placeholder: | placeholder: |
1. Open Cinny and log in to my account 1. Open Vojo and log in to my account
2. Navigate to the #general room 2. Navigate to the #general room
3. Type a message in the message box 3. Type a message in the message box
4. Press Enter to send 4. Press Enter to send
@ -62,14 +62,12 @@ body:
Please provide information about your environment. Include the following: Please provide information about your environment. Include the following:
- OS: - OS:
- Browser: - Browser:
- Cinny Web Version: (app.cinny.in or self hosted) - Vojo Web Version: (vojo.chat or self hosted)
- Cinny desktop Version: (appimage or deb or flatpak)
- Matrix Homeserver: - Matrix Homeserver:
placeholder: | placeholder: |
- OS: Windows 11 - OS: Windows 11
- Browser: Chrome 120.0.6099.109 - Browser: Chrome 120.0.6099.109
- Cinny Web Version: 3.2.0 (app.cinny.in or self hosted) - Vojo Web Version: (vojo.chat or self hosted)
- Cinny desktop Version: 3.2.0 (appimage or deb or flatpak)
- Matrix Homeserver: matrix.org (Synapse 1.97.0) - Matrix Homeserver: matrix.org (Synapse 1.97.0)
render: text render: text
validations: validations:
@ -116,12 +114,12 @@ body:
value: | value: |
# User Acknowledgements # User Acknowledgements
> [!TIP] > [!TIP]
> Use these links to review the existing Cinny [Discussions](https://github.com/cinnyapp/cinny/discussions?discussions_q=) and [Issues](https://github.com/cinnyapp/cinny/issues?q=sort%3Areactions-desc). > Use the search function to review existing Discussions and Issues.
- type: checkboxes #add faqs in future - type: checkboxes #add faqs in future
attributes: attributes:
label: "I acknowledge that:" label: "I acknowledge that:"
options: options:
- label: I have searched the Cinny repository (both open and closed Discussions and Issues) and confirm this is not a duplicate of an existing issue or discussion. - label: I have searched the Vojo repository (both open and closed Discussions and Issues) and confirm this is not a duplicate of an existing issue or discussion.
required: true required: true
- label: I have checked the "Preview" tab on all text fields to ensure that everything looks right, and have wrapped all configuration and code in code blocks with a group of three backticks (` ``` `) on separate lines. - label: I have checked the "Preview" tab on all text fields to ensure that everything looks right, and have wrapped all configuration and code in code blocks with a group of three backticks (` ``` `) on separate lines.
required: true required: true

4
.github/FUNDING.yml vendored
View file

@ -1,3 +1 @@
github: ajbura # Vojo project funding
liberapay: ajbura
open_collective: cinny

View file

@ -1,5 +1,5 @@
blank_issues_enabled: false blank_issues_enabled: false
contact_links: contact_links:
- name: Features, Bug Reports, Questions - name: Features, Bug Reports, Questions
url: https://github.com/cinnyapp/cinny/discussions/new/choose url: https://github.com/nicejuice-cc/vojo/discussions/new/choose
about: Our preferred starting point if you have any questions or suggestions about features or behavior. about: Our preferred starting point if you have any questions or suggestions about features or behavior.

View file

@ -1,36 +0,0 @@
name: 'CLA Assistant'
on:
issue_comment:
types: [created]
pull_request_target:
types: [opened, closed, synchronize]
jobs:
CLAssistant:
runs-on: ubuntu-latest
steps:
- name: 'CLA Assistant'
if: (github.event.comment.body == 'recheck' || github.event.comment.body == 'I have read the CLA Document and I hereby sign the CLA') || github.event_name == 'pull_request_target'
# Beta Release
uses: cla-assistant/github-action@ca4a40a7d1004f18d9960b404b97e5f30a505a08 # v2.6.1
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
# the below token should have repo scope and must be manually added by you in the repository's secret
PERSONAL_ACCESS_TOKEN: ${{ secrets.CLA_PAT }}
with:
path-to-signatures: 'signatures.json'
path-to-document: 'https://github.com/cinnyapp/cla/blob/main/cla.md' # e.g. a CLA or a DCO document
# branch should not be protected
branch: 'main'
allowlist: ajbura,bot*
#below are the optional inputs - If the optional inputs are not given, then default values will be taken
remote-organization-name: cinnyapp
remote-repository-name: cla
#create-file-commit-message: 'For example: Creating file for storing CLA Signatures'
#signed-commit-message: 'For example: $contributorName has signed the CLA in #$pullRequestNo'
#custom-notsigned-prcomment: 'pull request comment with Introductory message to ask new contributors to sign'
#custom-pr-sign-comment: 'The signature to be committed in order to sign the CLA'
#custom-allsigned-prcomment: 'pull request comment when all contributors has signed, defaults to **CLA Assistant Lite bot** All Contributors have signed the CLA.'
#lock-pullrequest-aftermerge: false - if you don't want this bot to automatically lock the pull request after merging (default - true)
#use-dco-flag: true - If you are using DCO instead of CLA

View file

@ -43,7 +43,7 @@ jobs:
enable-commit-comment: false enable-commit-comment: false
env: env:
NETLIFY_AUTH_TOKEN: ${{ secrets.NETLIFY_AUTH_TOKEN }} NETLIFY_AUTH_TOKEN: ${{ secrets.NETLIFY_AUTH_TOKEN }}
NETLIFY_SITE_ID: ${{ secrets.NETLIFY_SITE_ID_PR_CINNY }} NETLIFY_SITE_ID: ${{ secrets.NETLIFY_SITE_ID_PR_VOJO }}
timeout-minutes: 1 timeout-minutes: 1
- name: Comment preview on PR - name: Comment preview on PR
uses: thollander/actions-comment-pull-request@24bffb9b452ba05a4f3f77933840a6a841d1b32b #v3.0.1 uses: thollander/actions-comment-pull-request@24bffb9b452ba05a4f3f77933840a6a841d1b32b #v3.0.1

View file

@ -46,7 +46,6 @@ jobs:
uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # v5.10.0 uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # v5.10.0
with: with:
images: | images: |
ajbura/cinny
ghcr.io/${{ github.repository }} ghcr.io/${{ github.repository }}
- name: Build Docker image (no push) - name: Build Docker image (no push)

View file

@ -21,14 +21,6 @@ jobs:
package-manager-cache: false package-manager-cache: false
- name: Install dependencies - name: Install dependencies
run: npm ci run: npm ci
- name: Run semantic release
run: npm run semantic-release
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
GIT_AUTHOR_NAME: ${{ secrets.GIT_AUTHOR_NAME }}
GIT_AUTHOR_EMAIL: ${{ secrets.GIT_AUTHOR_EMAIL }}
GIT_COMMITTER_NAME: ${{ secrets.GIT_AUTHOR_NAME }}
GIT_COMMITTER_EMAIL: ${{ secrets.GIT_AUTHOR_EMAIL }}
- name: Get version from tag - name: Get version from tag
id: vars id: vars
run: | run: |
@ -53,7 +45,7 @@ jobs:
NETLIFY_SITE_ID: ${{ secrets.NETLIFY_SITE_ID_APP }} NETLIFY_SITE_ID: ${{ secrets.NETLIFY_SITE_ID_APP }}
timeout-minutes: 1 timeout-minutes: 1
- name: Create tar.gz - name: Create tar.gz
run: tar -czvf cinny-${{ steps.vars.outputs.tag }}.tar.gz dist run: tar -czvf vojo-${{ steps.vars.outputs.tag }}.tar.gz dist
- name: Sign tar.gz - name: Sign tar.gz
run: | run: |
echo '${{ secrets.GNUPG_KEY }}' | gpg --batch --import echo '${{ secrets.GNUPG_KEY }}' | gpg --batch --import
@ -63,14 +55,14 @@ jobs:
# non-armored and hex-encode it so that its printable. # non-armored and hex-encode it so that its printable.
echo "PGP Signing key, in raw PGP format in hex. Import with cat ... | xxd -r -p - | gpg --import" echo "PGP Signing key, in raw PGP format in hex. Import with cat ... | xxd -r -p - | gpg --import"
gpg --export | xxd -p gpg --export | xxd -p
echo '${{ secrets.GNUPG_PASSPHRASE }}' | gpg --batch --yes --pinentry-mode loopback --passphrase-fd 0 --armor --detach-sign cinny-${{ steps.vars.outputs.tag }}.tar.gz echo '${{ secrets.GNUPG_PASSPHRASE }}' | gpg --batch --yes --pinentry-mode loopback --passphrase-fd 0 --armor --detach-sign vojo-${{ steps.vars.outputs.tag }}.tar.gz
- name: Upload tagged release - name: Upload tagged release
uses: softprops/action-gh-release@6cbd405e2c4e67a21c47fa9e383d020e4e28b836 # v2.3.3 uses: softprops/action-gh-release@6cbd405e2c4e67a21c47fa9e383d020e4e28b836 # v2.3.3
with: with:
tag_name: ${{ steps.vars.outputs.tag }} tag_name: ${{ steps.vars.outputs.tag }}
files: | files: |
cinny-${{ steps.vars.outputs.tag }}.tar.gz vojo-${{ steps.vars.outputs.tag }}.tar.gz
cinny-${{ steps.vars.outputs.tag }}.tar.gz.asc vojo-${{ steps.vars.outputs.tag }}.tar.gz.asc
publish-image: publish-image:
name: Push Docker image to Docker Hub, GHCR name: Push Docker image to Docker Hub, GHCR
@ -106,7 +98,7 @@ jobs:
uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # v5.10.0 uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # v5.10.0
with: with:
images: | images: |
${{ secrets.DOCKER_USERNAME }}/cinny ${{ secrets.DOCKER_USERNAME }}/vojo
ghcr.io/${{ github.repository }} ghcr.io/${{ github.repository }}
tags: | tags: |
type=raw,value=${{ env.VERSION }} type=raw,value=${{ env.VERSION }}

22
.gitignore vendored
View file

@ -2,6 +2,26 @@ experiment
dist dist
node_modules node_modules
devAssets devAssets
config.local.json
electron/dist-electron
release
.DS_Store .DS_Store
.idea .idea
.vscode/*
!.vscode/tasks.json
.codex
.claude
docs/plans
docs/design
docs/ai/*
!docs/ai/README.md
!docs/ai/android.md
!docs/ai/architecture.md
!docs/ai/electron.md
!docs/ai/i18n.md
!docs/ai/overview.md
!docs/ai/server-side.md
vite.config.*.timestamp-*.mjs

5
.husky/pre-commit Normal file → Executable file
View file

@ -1,3 +1,2 @@
# These are commented until we enable lint and typecheck npx tsc -p tsconfig.json --noEmit
# npx tsc -p tsconfig.json --noEmit npx lint-staged
# npx lint-staged

View file

@ -3,4 +3,40 @@ node_modules
package.json package.json
package-lock.json package-lock.json
LICENSE LICENSE
README.md README.md
# Generated by Capacitor / Gradle / AGP — never format these.
android/app/build/
android/build/
android/capacitor-cordova-android-plugins/build/
android/app/src/main/assets/public/
android/app/src/main/assets/capacitor.config.json
android/app/src/main/assets/capacitor.plugins.json
android/app/google-services.json
# Internal docs — hand-formatted markdown. Prettier reflows tables and
# fenced code blocks (e.g. YAML inside fences in server-side.md, tables in
# architecture.md) in ways that change document structure, not whitespace.
# Most paths under docs/ are gitignored anyway via top-level .gitignore.
docs/
# Upstream Cinny GitHub Actions / templates — leave as-is, format drift here
# is unrelated to our work.
.github/
# Minified third-party assets.
*.min.js
# Top-level docs / HTML inherited from upstream Cinny — not part of this
# infra cleanup's scope. They have minor pre-existing format drift; touching
# them would just add review noise.
CLAUDE.md
CODE_OF_CONDUCT.md
CONTRIBUTING.md
index.html
# Upstream-vendored files copied verbatim from external projects (links in
# their headers). Keep byte-identical to upstream to make future re-syncs
# trivially diffable. Same intent as the per-file ESLint override.
src/util/cryptE2ERoomKeys.js
src/util/colorMXID.js

104
.vscode/tasks.json vendored Normal file
View file

@ -0,0 +1,104 @@
{
"version": "2.0.0",
"tasks": [
{
"label": "Deploy to vojo.chat",
"type": "shell",
"command": "npm run build && rsync -avz --delete dist/ vojo-superuser@187.127.77.124:~/vojo/cinny/",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Deploy widgets",
"type": "shell",
"command": "(cd apps/widget-telegram && npm run build && rsync -avz --delete dist/ vojo-superuser@187.127.77.124:~/vojo/widgets/telegram/) & PID1=$!; (cd apps/widget-discord && npm run build && rsync -avz --delete dist/ vojo-superuser@187.127.77.124:~/vojo/widgets/discord/) & PID2=$!; (cd apps/widget-whatsapp && npm run build && rsync -avz --delete dist/ vojo-superuser@187.127.77.124:~/vojo/widgets/whatsapp/) & PID3=$!; FAIL=0; wait $PID1 || FAIL=1; wait $PID2 || FAIL=1; wait $PID3 || FAIL=1; exit $FAIL",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Build Android APK",
"type": "shell",
"command": "npm run build:android:debug",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Deploy to Android (ADB)",
"type": "shell",
"command": "npm run build:android:debug && adb install -r android/app/build/outputs/apk/debug/app-debug.apk",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Connect to Android device (ADB)",
"type": "shell",
"command": "adb connect 192.168.1.204:5555",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Start Electron (dev)",
"type": "shell",
"command": "npm run electron:dev",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Build Electron Windows",
"type": "shell",
"command": "npm run build:electron:win",
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
},
{
"label": "Deploy Discord bridge",
"type": "shell",
"command": "docker build -t vojo-mautrix-discord:custom . && docker save vojo-mautrix-discord:custom | gzip | ssh vojo-superuser@187.127.77.124 'gunzip | docker load'",
"options": {
"cwd": "${workspaceFolder}/../vojo-mautrix-discord"
},
"group": "none",
"presentation": {
"reveal": "always",
"panel": "shared",
"showReuseMessage": false
},
"problemMatcher": []
}
]
}

9
CLAUDE.md Normal file
View file

@ -0,0 +1,9 @@
# Directive for AI agents
**All project context for Vojo lives in [`docs/ai/`](docs/ai/README.md). Read it before making any non-trivial change.**
**All plans that you create should be always created in docs/plans
This file exists only as a pointer. Do not add project knowledge here — put it in `docs/ai/`. Same rule for `.cursorrules`, `.windsurfrules`, `AGENTS.md`, `.codex`, home-directory memory, or any other agent-specific context file: if you're tempted to write project knowledge there, write it in `docs/ai/` instead and keep those files as thin pointers.
Start here: [docs/ai/README.md](docs/ai/README.md).

View file

@ -60,7 +60,7 @@ representative at an online or offline event.
Instances of abusive, harassing, or otherwise unacceptable behavior may be Instances of abusive, harassing, or otherwise unacceptable behavior may be
reported to the community leaders responsible for enforcement at reported to the community leaders responsible for enforcement at
cinnyapp@gmail.com. vojo@vojo.chat.
All complaints will be reviewed and investigated promptly and fairly. All complaints will be reviewed and investigated promptly and fairly.
All community leaders are obligated to respect the privacy and security of the All community leaders are obligated to respect the privacy and security of the

View file

@ -1,26 +1,21 @@
# Contributing to Cinny # Contributing to Vojo
First off, thanks for taking the time to contribute! ❤️ First off, thanks for taking the time to contribute!
All types of contributions are encouraged and valued. Please make sure to read the relevant section before making your contribution. It will make it a lot easier for us maintainers and smooth out the experience for all involved. The community looks forward to your contributions. 🎉 All types of contributions are encouraged and valued. Please make sure to read the relevant section before making your contribution.
> And if you like the project, but just don't have time to contribute, that's fine. There are other easy ways to support the project and show your appreciation, which we would also be very happy about: > And if you like the project, but just don't have time to contribute, that's fine. There are other easy ways to support the project and show your appreciation:
> - Star the project > - Star the project
> - Tweet about it (tag @cinnyapp)
> - Refer this project in your project's readme > - Refer this project in your project's readme
> - Mention the project at local meetups and tell your friends/colleagues > - Mention the project at local meetups and tell your friends/colleagues
> - [Donate to us](https://cinny.in/#sponsor)
## Bug reports ## Bug reports
Bug reports and feature suggestions must use descriptive and concise titles and be submitted to [GitHub Issues](https://github.com/ajbura/cinny/issues). Please use the search function to make sure that you are not submitting duplicates, and that a similar report or request has not already been resolved or rejected. Bug reports and feature suggestions must use descriptive and concise titles and be submitted to GitHub Issues. Please use the search function to make sure that you are not submitting duplicates, and that a similar report or request has not already been resolved or rejected.
## Pull requests ## Pull requests
> ### Legal Notice **NOTE: If you want to add new features, please discuss with maintainers before coding or opening a pull request.** This is to ensure that we are on the same track.
> When contributing to this project, you must agree that you have authored 100% of the content, that you have the necessary rights to the content and that the content you contribute may be provided under the project license. You will also be asked to [sign the CLA](https://github.com/cinnyapp/cla) upon submiting your pull request.
**NOTE: If you want to add new features, please discuss with maintainers before coding or opening a pull request.** This is to ensure that we are on same track and following our roadmap.
**Please use clean, concise titles for your pull requests.** We use commit squashing, so the final commit in the dev branch will carry the title of the pull request. For easier sorting in changelog, start your pull request titles using one of the verbs "Add", "Change", "Remove", or "Fix" (present tense). **Please use clean, concise titles for your pull requests.** We use commit squashing, so the final commit in the dev branch will carry the title of the pull request. For easier sorting in changelog, start your pull request titles using one of the verbs "Add", "Change", "Remove", or "Fix" (present tense).
@ -34,11 +29,7 @@ It is not always possible to phrase every change in such a manner, but it is des
**The smaller the set of changes in the pull request is, the quicker it can be reviewed and merged.** Splitting tasks into multiple smaller pull requests is often preferable. **The smaller the set of changes in the pull request is, the quicker it can be reviewed and merged.** Splitting tasks into multiple smaller pull requests is often preferable.
Also, we use [ESLint](https://eslint.org/) for clean and stylistically consistent code syntax, so make sure your pull request follow it. Also, we use [ESLint](https://eslint.org/) for clean and stylistically consistent code syntax, so make sure your pull request follows it.
**For any query or design discussion, join our [Matrix room](https://matrix.to/#/#cinny:matrix.org).**
## Helpful links ## Helpful links
- [BEM methodology](http://getbem.com/introduction/)
- [Atomic design](https://bradfrost.com/blog/post/atomic-web-design/)
- [Matrix JavaScript SDK documentation](https://matrix-org.github.io/matrix-js-sdk/index.html) - [Matrix JavaScript SDK documentation](https://matrix-org.github.io/matrix-js-sdk/index.html)

View file

@ -1,52 +0,0 @@
# Vojo — Matrix-based messenger for the Russian market
## What is this
This is a fork of Cinny (https://cinny.in, MIT license) rebranded as Vojo.
Vojo is an independent messenger built on the Matrix protocol, targeting
Russian users affected by Telegram blocking.
Key value proposition: "Telegram works without VPN" — the server maintains
Telegram bridge connections, users just open vojo.chat and chat.
## Architecture
- **Domain**: vojo.chat
- **Server**: Yandex Cloud VPS, Ubuntu 24.04
- **Backend**: Synapse (Matrix homeserver) + PostgreSQL + Caddy (reverse proxy)
- **Bridges**: mautrix-telegram (Python v0.15.3, SOCKS5 proxy for Telegram)
- **Client**: This repo — Cinny fork, deployed as static files via Caddy
The client is served from `~/vojo/cinny/` on the server.
Caddy serves it at `https://vojo.chat`.
## Branding
- **Name**: Vojo (everywhere in UI, page titles, meta tags)
## Default homeserver
All configs must point to `vojo.chat` as the default and only homeserver.
Hide homeserver selection from the user — they don't need to know about Matrix.
## Build & Deploy
```bash
npm ci
npm run build
# Output in dist/
scp -r dist/* vojo-superuser@111.88.146.156:~/vojo/cinny/
```
Caddy serves from `/var/www/cinny` (symlinked to ~/vojo/cinny).
## Key API Reference
- Matrix Client-Server API: https://spec.matrix.org/latest/client-server-api/
- Synapse Admin API: https://element-hq.github.io/synapse/latest/usage/administration/admin_api/
## Developer
Julian — C++ backend engineer. Limited frontend experience.
Using Claude Code agents for frontend work.

View file

@ -1,93 +1,29 @@
# Cinny # Vojo
<p>
<a href="https://github.com/ajbura/cinny/releases">
<img alt="GitHub release downloads" src="https://img.shields.io/github/downloads/ajbura/cinny/total?logo=github&style=social"></a>
<a href="https://hub.docker.com/r/ajbura/cinny">
<img alt="DockerHub downloads" src="https://img.shields.io/docker/pulls/ajbura/cinny?logo=docker&style=social"></a>
<a href="https://fosstodon.org/@cinnyapp">
<img alt="Follow on Mastodon" src="https://img.shields.io/mastodon/follow/106845779685925461?domain=https%3A%2F%2Ffosstodon.org&logo=mastodon&style=social"></a>
<a href="https://twitter.com/intent/follow?screen_name=cinnyapp">
<img alt="Follow on Twitter" src="https://img.shields.io/twitter/follow/cinnyapp?logo=twitter&style=social"></a>
<a href="https://cinny.in/#sponsor">
<img alt="Sponsor Cinny" src="https://img.shields.io/opencollective/all/cinny?logo=opencollective&style=social"></a>
</p>
A Matrix client focusing primarily on simple, elegant and secure interface. The main goal is to have an instant messaging application that is easy on people and has a modern touch. A Matrix client focusing primarily on simple, elegant and secure interface. The main goal is to have an instant messaging application that is easy on people and has a modern touch.
- [Roadmap](https://github.com/orgs/cinnyapp/projects/1)
Based on [Cinny](https://github.com/cinnyapp/cinny) (MIT license).
- [Contributing](./CONTRIBUTING.md) - [Contributing](./CONTRIBUTING.md)
> [!IMPORTANT]
We are currently in the [process of replacing the matrix-js-sdk](https://github.com/cinnyapp/cinny/issues/257#issuecomment-3714406704) with our own SDK. As a result, we will not be accepting any pull requests until further notice.
Thank you for your understanding.
<img align="center" src="https://raw.githubusercontent.com/cinnyapp/cinny-site/main/assets/preview2-light.png" height="380">
## Getting started ## Getting started
The web app is available at [app.cinny.in](https://app.cinny.in/) and gets updated on each new release. The `dev` branch is continuously deployed at [dev.cinny.in](https://dev.cinny.in) but keep in mind that it could have things broken.
You can also download our desktop app from the [cinny-desktop repository](https://github.com/cinnyapp/cinny-desktop). The web app is available at [vojo.chat](https://vojo.chat).
## Self-hosting ## Self-hosting
To host Cinny on your own, simply download the tarball from [GitHub releases](https://github.com/cinnyapp/cinny/releases/latest), and serve the files from `dist/` using your preferred webserver. Alternatively, you can just pull the docker image from [DockerHub](https://hub.docker.com/r/ajbura/cinny) or [GitHub Container Registry](https://github.com/cinnyapp/cinny/pkgs/container/cinny).
* The default homeservers and explore pages are defined in [`config.json`](config.json). To host Vojo on your own, build from source and serve the files from `dist/` using your preferred webserver.
* You need to set up redirects to serve the assests. Example configurations; [netlify](netlify.toml), [nginx](contrib/nginx/cinny.domain.tld.conf), [caddy](contrib/caddy/caddyfile). * The default homeserver is defined in [`config.json`](config.json).
* If you have trouble configuring redirects you can [enable hash routing](config.json#L35) — the url in the browser will have a `/#/` between the domain and open channel (ie. `app.cinny.in/#/home/` instead of `app.cinny.in/home/`) but you won't have to configure your webserver.
* To deploy on subdirectory, you need to rebuild the app youself after updating the `base` path in [`build.config.ts`](build.config.ts). * You need to set up redirects to serve the assets. Example configurations: [nginx](contrib/nginx/vojo.domain.tld.conf), [caddy](contrib/caddy/caddyfile).
* For example, if you want to deploy on `https://cinny.in/app`, then set `base: '/app'`.
<details><summary><b>PGP Public Key to verify tarball</b></summary> * To deploy on a subdirectory, rebuild the app after updating the `base` path in [`build.config.ts`](build.config.ts).
```
-----BEGIN PGP PUBLIC KEY BLOCK-----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=WtqZ
-----END PGP PUBLIC KEY BLOCK-----
```
</details>
## Local development ## Local development
> [!TIP] > [!TIP]
> We recommend using a version manager as versions change very quickly. You will likely need to switch between multiple Node.js versions based on the needs of different projects you're working on. [NVM on windows](https://github.com/coreybutler/nvm-windows#installation--upgrades) on Windows and [nvm](https://github.com/nvm-sh/nvm) on Linux/macOS are pretty good choices. Recommended nodejs version is Krypton LTS (v24.13.1). > We recommend using a version manager as versions change very quickly. [NVM on Windows](https://github.com/coreybutler/nvm-windows#installation--upgrades) or [nvm](https://github.com/nvm-sh/nvm) on Linux/macOS are good choices.
Execute the following commands to start a development server: Execute the following commands to start a development server:
```sh ```sh
@ -101,15 +37,15 @@ npm run build # Compiles the app into the dist/ directory
``` ```
### Running with Docker ### Running with Docker
This repository includes a Dockerfile, which builds the application from source and serves it with Nginx on port 80. To
use this locally, you can build the container like so: This repository includes a Dockerfile, which builds the application from source and serves it with Nginx on port 80. To use this locally, you can build the container like so:
``` ```
docker build -t cinny:latest . docker build -t vojo:latest .
``` ```
You can then run the container you've built with a command similar to this: You can then run the container you've built with a command similar to this:
``` ```
docker run -p 8080:80 cinny:latest docker run -p 8080:80 vojo:latest
``` ```
This will forward your `localhost` port 8080 to the container's port 80. You can visit the app in your browser by navigating to `http://localhost:8080`. This will forward your `localhost` port 8080 to the container's port 80. You can visit the app in your browser by navigating to `http://localhost:8080`.

101
android/.gitignore vendored Normal file
View file

@ -0,0 +1,101 @@
# Using Android gitignore template: https://github.com/github/gitignore/blob/HEAD/Android.gitignore
# Built application files
*.apk
*.aar
*.ap_
*.aab
# Files for the ART/Dalvik VM
*.dex
# Java class files
*.class
# Generated files
bin/
gen/
out/
# Uncomment the following line in case you need and you don't have the release build type files in your app
# release/
# Gradle files
.gradle/
build/
# Local configuration file (sdk path, etc)
local.properties
# Proguard folder generated by Eclipse
proguard/
# Log Files
*.log
# Android Studio Navigation editor temp files
.navigation/
# Android Studio captures folder
captures/
# IntelliJ
*.iml
.idea/workspace.xml
.idea/tasks.xml
.idea/gradle.xml
.idea/assetWizardSettings.xml
.idea/dictionaries
.idea/libraries
# Android Studio 3 in .gitignore file.
.idea/caches
.idea/modules.xml
# Comment next line if keeping position of elements in Navigation Editor is relevant for you
.idea/navEditor.xml
# Keystore files
# Uncomment the following lines if you do not want to check your keystore files in.
#*.jks
#*.keystore
# External native build folder generated in Android Studio 2.2 and later
.externalNativeBuild
.cxx/
# Google Services (e.g. APIs or Firebase)
# google-services.json
# Freeline
freeline.py
freeline/
freeline_project_description.json
# fastlane
fastlane/report.xml
fastlane/Preview.html
fastlane/screenshots
fastlane/test_output
fastlane/readme.md
# Version control
vcs.xml
# lint
lint/intermediates/
lint/generated/
lint/outputs/
lint/tmp/
# lint/reports/
# Android Profiling
*.hprof
# Cordova plugins for Capacitor
capacitor-cordova-android-plugins
# Copied web assets
app/src/main/assets/public
# Generated Config files
app/src/main/assets/capacitor.config.json
app/src/main/assets/capacitor.plugins.json
app/src/main/res/xml/config.xml

2
android/app/.gitignore vendored Normal file
View file

@ -0,0 +1,2 @@
/build/*
!/build/.npmkeep

148
android/app/build.gradle Normal file
View file

@ -0,0 +1,148 @@
apply plugin: 'com.android.application'
// Mirror of resolveAppVersion() in ../../vite.config.js so the APK's
// versionName matches __APP_VERSION__ rendered in the About screen.
// `git describe --tags --match 'v*'` against tag v0.2.0 yields
// `v0.2.0-<commits>-g<hash>`; patch = commit count since the tag.
// Falls back to package.json only when git is unavailable.
def gitDescribe = providers.exec {
it.commandLine 'git', 'describe', '--tags', '--match', 'v*', '--always'
it.workingDir rootDir.parentFile
it.ignoreExitValue = true
}
def appVersion = {
def fromGit = gitDescribe.result.get().exitValue == 0 ? gitDescribe.standardOutput.asText.get().trim() : null
def m = fromGit =~ /^v?(\d+)\.(\d+)\.(\d+)(?:-(\d+)-g[0-9a-f]+)?$/
if (fromGit && m.matches()) {
def major = m[0][1].toInteger()
def minor = m[0][2].toInteger()
def patch = (m[0][4] ?: m[0][3]).toInteger()
return [name: "${major}.${minor}.${patch}", major: major, minor: minor, patch: patch]
}
def pkg = new groovy.json.JsonSlurper().parseText(file('../../package.json').text)
def parts = pkg.version.split('\\.')
return [name: pkg.version, major: parts[0].toInteger(), minor: parts[1].toInteger(), patch: parts[2].toInteger()]
}()
def computedVersionCode = appVersion.major * 1000000 + appVersion.minor * 1000 + appVersion.patch
android {
namespace = "chat.vojo.app"
compileSdk = rootProject.ext.compileSdkVersion
defaultConfig {
applicationId "chat.vojo.app"
minSdkVersion rootProject.ext.minSdkVersion
targetSdkVersion rootProject.ext.targetSdkVersion
versionCode computedVersionCode
versionName appVersion.name
testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
aaptOptions {
// Files and dirs to omit from the packaged assets dir, modified to accommodate modern web apps.
// Default: https://android.googlesource.com/platform/frameworks/base/+/282e181b58cf72b6ca770dc7ca5f91f135444502/tools/aapt/AaptAssets.cpp#61
ignoreAssetsPattern = '!.svn:!.git:!.ds_store:!*.scc:.*:!CVS:!thumbs.db:!picasa.ini:!*~'
}
}
// AGP 8+ requires explicit opt-in for BuildConfig generation. We rely on
// BuildConfig.DEBUG to gate Log.d calls that dump privacy-sensitive
// identifiers (roomId, eventId) so release builds don't leak them through
// logcat / crash-reporter buffers. See dlog() in VojoFirebaseMessagingService.
buildFeatures {
buildConfig = true
}
signingConfigs {
release {
if (project.hasProperty('VOJO_RELEASE_STORE_FILE')) {
storeFile file(VOJO_RELEASE_STORE_FILE)
storePassword VOJO_RELEASE_STORE_PASSWORD
keyAlias VOJO_RELEASE_KEY_ALIAS
keyPassword VOJO_RELEASE_KEY_PASSWORD
}
}
}
buildTypes {
release {
minifyEnabled true
shrinkResources true
proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
signingConfig signingConfigs.release
}
}
}
repositories {
flatDir{
dirs '../capacitor-cordova-android-plugins/src/main/libs', 'libs'
}
}
dependencies {
implementation fileTree(include: ['*.jar'], dir: 'libs')
implementation "androidx.activity:activity:$androidxActivityVersion"
implementation "androidx.appcompat:appcompat:$androidxAppCompatVersion"
implementation "androidx.coordinatorlayout:coordinatorlayout:$androidxCoordinatorLayoutVersion"
implementation "androidx.core:core-splashscreen:$coreSplashScreenVersion"
implementation project(':capacitor-android')
// Needed for VojoFirebaseMessagingService. @capacitor/push-notifications
// already depends on firebase-messaging but declares it `implementation`
// so classes aren't exposed at app-module compile time.
implementation "com.google.firebase:firebase-messaging:25.0.1"
// WorkManager hosts VojoPollWorker periodic /notifications poll that
// delivers messages and missed-call surfaces on networks where FCM
// (mtalk.google.com:5228) is blocked. Library self-registers its scheduler
// in the merged manifest; we declare no permission for it.
implementation "androidx.work:work-runtime:2.10.0"
testImplementation "junit:junit:$junitVersion"
androidTestImplementation "androidx.test.ext:junit:$androidxJunitVersion"
androidTestImplementation "androidx.test.espresso:espresso-core:$androidxEspressoCoreVersion"
implementation project(':capacitor-cordova-android-plugins')
}
apply from: 'capacitor.build.gradle'
abstract class GeneratePushStringsTask extends DefaultTask {
@InputFiles
abstract ConfigurableFileCollection getInputFiles()
@OutputDirectory
abstract DirectoryProperty getOutputDir()
@TaskAction
void generate() {
def nodeBin = project.findProperty('NODE_BIN') ?: 'node'
project.exec {
commandLine nodeBin,
new File(project.rootProject.projectDir.parentFile, 'scripts/gen-push-strings.mjs').absolutePath,
'--out', outputDir.get().asFile.absolutePath
}
}
}
androidComponents {
onVariants(selector().all()) { variant ->
def repoRoot = rootProject.projectDir.parentFile
def taskProvider = tasks.register(
"generatePushStrings${variant.name.capitalize()}",
GeneratePushStringsTask
) {
inputFiles.from(
new File(repoRoot, 'public/locales/en.json'),
new File(repoRoot, 'public/locales/ru.json'),
new File(repoRoot, 'scripts/gen-push-strings.mjs')
)
outputDir.set(layout.buildDirectory.dir("generated/res/push/${variant.name}"))
}
variant.sources.res.addGeneratedSourceDirectory(
taskProvider, GeneratePushStringsTask::getOutputDir
)
}
}
try {
def servicesJSON = file('google-services.json')
if (servicesJSON.text) {
apply plugin: 'com.google.gms.google-services'
}
} catch(Exception e) {
logger.info("google-services.json not found, google-services plugin not applied. Push Notifications won't work")
}

View file

@ -0,0 +1,23 @@
// DO NOT EDIT THIS FILE! IT IS GENERATED EACH TIME "capacitor update" IS RUN
android {
compileOptions {
sourceCompatibility JavaVersion.VERSION_21
targetCompatibility JavaVersion.VERSION_21
}
}
apply from: "../capacitor-cordova-android-plugins/cordova.variables.gradle"
dependencies {
implementation project(':capacitor-app')
implementation project(':capacitor-browser')
implementation project(':capacitor-preferences')
implementation project(':capacitor-push-notifications')
implementation project(':capacitor-toast')
}
if (hasProperty('postBuildExtras')) {
postBuildExtras()
}

View file

@ -0,0 +1,29 @@
{
"project_info": {
"project_number": "51806967595",
"project_id": "chat-vojo-app",
"storage_bucket": "chat-vojo-app.firebasestorage.app"
},
"client": [
{
"client_info": {
"mobilesdk_app_id": "1:51806967595:android:93921bf62aa9713a79576e",
"android_client_info": {
"package_name": "chat.vojo.app"
}
},
"oauth_client": [],
"api_key": [
{
"current_key": "AIzaSyBroeOOHxg-tEyU-O-zjSWF7mEejedRWsM"
}
],
"services": {
"appinvite_service": {
"other_platform_oauth_client": []
}
}
}
],
"configuration_version": "1"
}

45
android/app/proguard-rules.pro vendored Normal file
View file

@ -0,0 +1,45 @@
# Add project specific ProGuard rules here.
# You can control the set of applied configuration files using the
# proguardFiles setting in build.gradle.
#
# For more details, see
# http://developer.android.com/guide/developing/tools/proguard.html
# If your project uses WebView with JS, uncomment the following
# and specify the fully qualified class name to the JavaScript interface
# class:
#-keepclassmembers class fqcn.of.javascript.interface.for.webview {
# public *;
#}
# Uncomment this to preserve the line number information for
# debugging stack traces.
#-keepattributes SourceFile,LineNumberTable
# If you keep the line number information, uncomment this to
# hide the original source file name.
#-renamesourcefileattribute SourceFile
# Keep custom app classes entry points invoked by Android system (Intents,
# FCM, AndroidManifest references) or by JS bridge via reflection.
-keep class chat.vojo.app.MainActivity { *; }
-keep class chat.vojo.app.VojoFirebaseMessagingService { *; }
-keep class chat.vojo.app.CallForegroundPlugin { *; }
-keep class chat.vojo.app.CallForegroundService { *; }
-keep class chat.vojo.app.CallDeclineReceiver { *; }
-keep class chat.vojo.app.CallCancelReceiver { *; }
-keep class chat.vojo.app.FullScreenIntentPlugin { *; }
-keep class chat.vojo.app.LaunchSplashPlugin { *; }
# Firebase Messaging receivers/services resolved by Android via manifest.
-keep public class * extends com.google.firebase.messaging.FirebaseMessagingService
-keep class com.google.firebase.iid.** { *; }
-keep class com.google.firebase.messaging.** { *; }
# Capacitor plugins discovered by annotation/reflection.
-keep @com.getcapacitor.annotation.CapacitorPlugin class * { *; }
-keep class com.getcapacitor.** { *; }
-keep class com.getcapacitor.plugin.** { *; }
# AndroidX splashscreen reflection paths.
-keep class androidx.core.splashscreen.** { *; }

View file

@ -0,0 +1,26 @@
package com.getcapacitor.myapp;
import static org.junit.Assert.*;
import android.content.Context;
import androidx.test.ext.junit.runners.AndroidJUnit4;
import androidx.test.platform.app.InstrumentationRegistry;
import org.junit.Test;
import org.junit.runner.RunWith;
/**
* Instrumented test, which will execute on an Android device.
*
* @see <a href="http://d.android.com/tools/testing">Testing documentation</a>
*/
@RunWith(AndroidJUnit4.class)
public class ExampleInstrumentedTest {
@Test
public void useAppContext() throws Exception {
// Context of the app under test.
Context appContext = InstrumentationRegistry.getInstrumentation().getTargetContext();
assertEquals("com.getcapacitor.app", appContext.getPackageName());
}
}

View file

@ -0,0 +1,145 @@
<?xml version="1.0" encoding="utf-8"?>
<manifest xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:tools="http://schemas.android.com/tools">
<!-- allowBackup=false: CallDeclineReceiver reads the Matrix access_token
from shared_prefs/CapacitorStorage.xml (written by sessionBridge).
With allowBackup=true a rooted device or adb-backup-enabled user
could exfiltrate the cleartext token. Session data is cheap to
recreate via re-login, so excluding ourselves from Auto Backup
is the simpler control vs. fine-grained backup_rules.xml exclusions. -->
<application
android:allowBackup="false"
android:icon="@mipmap/ic_launcher"
android:label="@string/app_name"
android:roundIcon="@mipmap/ic_launcher_round"
android:supportsRtl="true"
android:theme="@style/AppTheme">
<activity
android:configChanges="orientation|keyboardHidden|keyboard|screenSize|locale|smallestScreenSize|screenLayout|uiMode|navigation|density"
android:name=".MainActivity"
android:label="@string/title_activity_main"
android:theme="@style/AppTheme.NoActionBarLaunch"
android:launchMode="singleTask"
android:exported="true">
<intent-filter>
<action android:name="android.intent.action.MAIN" />
<category android:name="android.intent.category.LAUNCHER" />
</intent-filter>
<!-- App Links for https://vojo.chat/u/<user>. autoVerify=true lets
Chrome/Gmail/SMS hand the tap straight to this activity; the
server must publish a matching /.well-known/assetlinks.json
over HTTPS with the installed APK's signing SHA-256. Telegram
and other in-app browsers ignore verification and load the
URL in their own webview — the intent-URL redirect injected
in index.html covers that case. -->
<intent-filter android:autoVerify="true">
<action android:name="android.intent.action.VIEW" />
<category android:name="android.intent.category.DEFAULT" />
<category android:name="android.intent.category.BROWSABLE" />
<data
android:scheme="https"
android:host="vojo.chat"
android:pathPrefix="/u/" />
</intent-filter>
<!-- System share-sheet target. Three filters because Android's
sheet UI dedupes by activity but resolves by MIME match:
text/* gets its own filter so the Vojo icon shows up
alongside WhatsApp/Telegram for «share link/selection»; */*
covers single-file (image/video/audio/pdf/…) and
SEND_MULTIPLE picks up gallery multi-select.
Payload extraction lives in ShareTargetPlugin — MainActivity
only routes the Intent to the plugin via onNewIntent. -->
<intent-filter>
<action android:name="android.intent.action.SEND" />
<category android:name="android.intent.category.DEFAULT" />
<data android:mimeType="text/*" />
</intent-filter>
<intent-filter>
<action android:name="android.intent.action.SEND" />
<category android:name="android.intent.category.DEFAULT" />
<data android:mimeType="*/*" />
</intent-filter>
<intent-filter>
<action android:name="android.intent.action.SEND_MULTIPLE" />
<category android:name="android.intent.category.DEFAULT" />
<data android:mimeType="*/*" />
</intent-filter>
</activity>
<provider
android:name="androidx.core.content.FileProvider"
android:authorities="${applicationId}.fileprovider"
android:exported="false"
android:grantUriPermissions="true">
<meta-data
android:name="android.support.FILE_PROVIDER_PATHS"
android:resource="@xml/file_paths"></meta-data>
</provider>
<!-- Replace Capacitor's default FCM service. VojoFirebaseMessagingService
extends MessagingService so super.onMessageReceived() still forwards
to the JS bridge; we add cold-start notification display on top. -->
<service
android:name="com.capacitorjs.plugins.pushnotifications.MessagingService"
tools:node="remove" />
<service
android:name=".VojoFirebaseMessagingService"
android:exported="false">
<intent-filter>
<action android:name="com.google.firebase.MESSAGING_EVENT" />
</intent-filter>
</service>
<service
android:name=".CallForegroundService"
android:exported="false"
android:foregroundServiceType="microphone" />
<receiver
android:name=".CallCancelReceiver"
android:exported="false" />
<receiver
android:name=".CallDeclineReceiver"
android:exported="false" />
<receiver
android:name=".MarkAsReadReceiver"
android:exported="false" />
<receiver
android:name=".NotificationDismissReceiver"
android:exported="false" />
<receiver
android:name=".ReplyReceiver"
android:exported="false" />
</application>
<!-- Permissions -->
<uses-permission android:name="android.permission.INTERNET" />
<uses-permission android:name="android.permission.POST_NOTIFICATIONS" />
<!-- DM voice calls: mic + audio routing. Capacitor auto-requests at getUserMedia time. -->
<uses-permission android:name="android.permission.RECORD_AUDIO" />
<uses-permission android:name="android.permission.MODIFY_AUDIO_SETTINGS" />
<!-- Required to unblock NotificationCompat.CallStyle on API 31+: NMS's
checkDisqualifyingFeatures rejects CallStyle notifications without
FSI/FGS/UIJ, throwing IllegalArgumentException on its own handler
thread (silent to the app). Declaring the permission flips
FLAG_FSI_REQUESTED_BUT_DENIED so the gate passes, even though we
never call setFullScreenIntent(). See ADR 2.5-heads-up. -->
<uses-permission android:name="android.permission.USE_FULL_SCREEN_INTENT" />
<!-- DM call lock-screen retention: CallForegroundService keeps the call
process foregrounded under lock so AppOps doesn't revoke RECORD_AUDIO
and netd doesn't block background network. -->
<uses-permission android:name="android.permission.FOREGROUND_SERVICE" />
<uses-permission android:name="android.permission.FOREGROUND_SERVICE_MICROPHONE" />
</manifest>

View file

@ -0,0 +1,65 @@
package chat.vojo.app;
import android.graphics.Bitmap;
import android.util.LruCache;
/**
* In-memory LRU cache of decoded avatar bitmaps keyed by MXC URL string.
*
* Sized as a process-singleton (~4 MB) so the FCM service, polling Worker
* and ReplyReceiver all share one pool. 96×96 ARGB_8888 bitmap is about
* 36 KB, so a 4 MB cache holds ~110 avatars enough for the active
* conversation set on a typical user. LruCache evicts the least-recently-
* read entry when full; this is the right shape for "rooms the user is
* actively talking in stay warm, dormant rooms reload on demand".
*
* Thread-safety: LruCache itself is synchronized internally on every
* get/put/remove. We don't need an outer lock for normal operation. The
* AvatarLoader funnels all puts through this class.
*
* Process death: cache is in-memory only. After a kill, the first push
* to any room cold-renders without avatars and re-renders once the
* loader populates the cache (see AvatarLoader.loadAllWithTimeout).
*/
final class AvatarBitmapCache {
// Heap budget: bytes. 4 MB is generous against ARGB_8888 96×96 bitmaps
// (~36 KB each) and stays comfortably under the 1/8-of-heap Android
// recommendation on every device we ship to (minSdk 24 at least
// 96 MB heap on a low-end phone).
private static final int MAX_SIZE_BYTES = 4 * 1024 * 1024;
private static final LruCache<String, Bitmap> CACHE =
new LruCache<String, Bitmap>(MAX_SIZE_BYTES) {
@Override
protected int sizeOf(String key, Bitmap value) {
return value.getByteCount();
}
};
private AvatarBitmapCache() {}
/**
* Returns the cached bitmap for an MXC URL, or null on miss.
*
* Bitmap references are NOT defensively copied the cache hands out
* the same reference to every caller. This is safe because no code
* path in the app calls Bitmap.recycle() on a cached bitmap (the
* intermediate square / source bitmaps inside AvatarLoader.
* toCircularBitmap ARE recycled, but the circular output that lands
* here is held until LRU evicts it). LRU eviction simply drops the
* cache's reference, and the GC reclaims memory only after every
* Notification that referenced the bitmap is also released by the
* system. Adding a defensive copy here would halve the effective
* cache size for no real-world benefit.
*/
static Bitmap get(String mxc) {
if (mxc == null || mxc.isEmpty()) return null;
return CACHE.get(mxc);
}
static void put(String mxc, Bitmap bitmap) {
if (mxc == null || mxc.isEmpty() || bitmap == null) return;
CACHE.put(mxc, bitmap);
}
}

View file

@ -0,0 +1,368 @@
package chat.vojo.app;
import android.content.Context;
import android.content.SharedPreferences;
import android.graphics.Bitmap;
import android.graphics.BitmapFactory;
import android.graphics.BitmapShader;
import android.graphics.Canvas;
import android.graphics.Paint;
import android.graphics.Shader;
import android.util.Log;
import java.io.IOException;
import java.io.InputStream;
import java.net.HttpURLConnection;
import java.net.URL;
import java.util.Collection;
import java.util.LinkedHashSet;
import java.util.Set;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;
/**
* Fetches and decodes avatar bitmaps from MXC URLs, populating
* {@link AvatarBitmapCache}.
*
* URL resolution mirrors matrix-js-sdk's auth-media v1.11+ pattern:
* mxc://server/mediaId
* <homeserver>/_matrix/client/v1/media/thumbnail/<server>/<mediaId>
* ?width=96&height=96&method=crop
* + Authorization: Bearer <accessToken>
*
* The legacy unauthenticated `/_matrix/media/v3/thumbnail/...` endpoint is
* NOT used every Synapse the Vojo audience runs against (vanilla, v1.11+
* by deployment policy, see docs/ai/server-side.md) speaks auth media.
* Removing the legacy fallback keeps the loader off the deprecated path
* and avoids leaking the access token to a server route that doesn't
* require it.
*
* Concurrency: each MXC URL is fetched at most once concurrently the
* `inFlight` set short-circuits duplicate requests from rapid
* append-rebuild cycles on the same conversation. Loads happen on a
* shared 4-thread pool; bigger than 1 so 5 senders in a group chat can
* load in parallel, capped to keep socket pressure under the typical
* mobile network budget.
*
* Two entry points:
* - {@link #loadAllWithTimeout}: synchronous wait, used by the render
* path to populate the cache before building the MessagingStyle so the
* first post already has avatars. Timeout-bounded to keep FCM thread
* responsive (Android budgets ~10s; we use 800 ms).
* - {@link #prefetch}: fire-and-forget, used for warm-up scenarios.
* Not currently called but kept for the room-metadata bridge to
* eventually warm the cache on visibility resume.
*/
final class AvatarLoader {
private static final String TAG = "AvatarLoader";
private static final int AVATAR_SIZE_PX = 96;
private static final int CONNECT_TIMEOUT_MS = 5_000;
private static final int READ_TIMEOUT_MS = 5_000;
private static final int RENDER_BLOCK_TIMEOUT_MS = 800;
// Cap decoded bitmap byte count a malicious / huge avatar shouldn't
// OOM the FCM service. 96×96 ARGB_8888 is ~36 KB; we accept up to
// 4× that (~140 KB) to allow some downscaling slack on servers that
// return slightly oversized thumbnails.
private static final int MAX_DECODED_BYTES = 144 * 1024;
private static final ExecutorService EXECUTOR = Executors.newFixedThreadPool(4);
// MXC URL CountDownLatch that fires when the in-flight download
// completes (success or failure). A second caller observing an
// already-pending mxc waits on the SAME latch instead of either
// returning empty-handed or kicking off a duplicate fetch. Latches
// are removed by the worker task in its finally block; the same task
// that put the entry is the only one allowed to remove it, so a slow
// remove() race is harmless.
private static final ConcurrentHashMap<String, CountDownLatch> inFlight =
new ConcurrentHashMap<>();
private AvatarLoader() {}
/**
* Block the caller for up to {@link #RENDER_BLOCK_TIMEOUT_MS} while
* fetching any of the given MXC URLs that are not yet in
* {@link AvatarBitmapCache}. Cache hits are no-ops. Already-in-flight
* URLs are awaited via the shared latch duplicate concurrent
* fetches do not happen.
*
* Designed to be called inline from the render path: after this
* returns, {@link AvatarBitmapCache#get} will be non-null for every
* MXC that loaded successfully within the budget. Failures are
* silent the render then falls back to a Person without icon
* (Android renders initials/blank).
*
* Returns the count of avatars that landed in the cache during this
* call (purely informational useful for logs).
*/
static int loadAllWithTimeout(Context ctx, Collection<String> mxcs) {
if (mxcs == null || mxcs.isEmpty()) {
Log.i(TAG, "loadAll: empty input, skip");
return 0;
}
SharedPreferences prefs = ctx.getSharedPreferences(
VojoPollWorker.PREFS, Context.MODE_PRIVATE);
String token = prefs.getString(VojoPollWorker.KEY_ACCESS_TOKEN, null);
String homeserver = prefs.getString(VojoPollWorker.KEY_HOMESERVER_URL, null);
if (token == null || token.isEmpty() || homeserver == null || homeserver.isEmpty()) {
// No credentials yet (fresh install + first push). We can't
// resolve MXC URLs without an access token. Falling back to
// no-icon Person renderer is the correct behaviour here.
Log.i(TAG, "loadAll: no credentials in prefs, skip"
+ " hasToken=" + (token != null && !token.isEmpty())
+ " hasHs=" + (homeserver != null && !homeserver.isEmpty()));
return 0;
}
// De-duplicate and filter to misses only; if the cache already has
// an entry, no work is needed.
Set<String> toLoad = new LinkedHashSet<>();
for (String mxc : mxcs) {
if (mxc == null || mxc.isEmpty()) continue;
if (!mxc.startsWith("mxc://")) continue;
if (AvatarBitmapCache.get(mxc) != null) continue;
toLoad.add(mxc);
}
if (toLoad.isEmpty()) return 0;
// Per-mxc latches shared across concurrent callers a second
// caller arriving while we're already mid-fetch waits on the
// SAME latch instead of forcing a duplicate HTTP or returning
// immediately empty-handed (which was the previous bug see
// git blame for the race description).
java.util.List<CountDownLatch> waits = new java.util.ArrayList<>(toLoad.size());
for (String mxc : toLoad) {
CountDownLatch myLatch = new CountDownLatch(1);
CountDownLatch existing = inFlight.putIfAbsent(mxc, myLatch);
if (existing != null) {
// Already in flight share the original latch.
waits.add(existing);
continue;
}
// We own this fetch; kick off the worker that will fire
// myLatch when done.
waits.add(myLatch);
final String capturedMxc = mxc;
final String capturedHomeserver = homeserver;
final String capturedToken = token;
EXECUTOR.execute(() -> {
try {
Bitmap bmp = fetchAndDecode(capturedMxc, capturedHomeserver, capturedToken);
if (bmp != null) AvatarBitmapCache.put(capturedMxc, bmp);
} catch (Throwable t) {
Log.w(TAG, "fetch threw mxc=" + capturedMxc, t);
} finally {
// Remove BEFORE countDown so a freshly-arriving caller
// doesn't observe a stale latch for an already-loaded
// mxc (would block until the next call with no fetch
// actually pending). Cache.get() on the post-await
// side covers the race where remove+put-cache happens
// between two latch waits.
inFlight.remove(capturedMxc);
myLatch.countDown();
}
});
}
// Single budget for the whole batch wait for all latches OR
// hit the timeout. Latches that fire early just return await()
// immediately; the slowest one consumes the remainder of the
// budget.
long deadline = System.nanoTime() + TimeUnit.MILLISECONDS.toNanos(RENDER_BLOCK_TIMEOUT_MS);
try {
for (CountDownLatch latch : waits) {
long remaining = deadline - System.nanoTime();
if (remaining <= 0) break;
latch.await(remaining, TimeUnit.NANOSECONDS);
}
} catch (InterruptedException ie) {
Thread.currentThread().interrupt();
}
// Count how many actually landed in the cache during this call
// includes both items we fetched and items that finished after our
// timeout (which won't be reflected in this count but are still
// usable on the next render).
int hits = 0;
for (String mxc : toLoad) {
if (AvatarBitmapCache.get(mxc) != null) hits += 1;
}
Log.i(TAG, "loadAll: requested=" + mxcs.size()
+ " toLoad=" + toLoad.size() + " hits=" + hits);
return hits;
}
/**
* Resolve an `mxc://server/mediaId` URL to a 96×96 thumbnail via the
* authenticated v1.11+ media endpoint and decode the response into a
* Bitmap. Returns null on any non-2xx, decode failure, or oversized
* payload (see {@link #MAX_DECODED_BYTES}).
*/
private static Bitmap fetchAndDecode(String mxc, String homeserver, String token)
throws IOException {
Parsed parsed = parseMxc(mxc);
if (parsed == null) {
Log.w(TAG, "fetch: malformed mxc=" + mxc);
return null;
}
// Server + mediaId are NOT URL-encoded matches matrix-js-sdk's
// content-repo.ts (it concatenates verbatim via `new URL()`).
// URLEncoder would turn `example.com:8448` into `example.com%3A8448`,
// which Synapse's media router rejects as an unknown server.
// mediaId is base64-ish per spec (URL-safe alphabet) so no
// encoding is needed there either.
StringBuilder url = new StringBuilder(homeserver);
if (!homeserver.endsWith("/")) url.append('/');
url.append("_matrix/client/v1/media/thumbnail/")
.append(parsed.server)
.append('/')
.append(parsed.mediaId)
.append("?width=").append(AVATAR_SIZE_PX)
.append("&height=").append(AVATAR_SIZE_PX)
.append("&method=crop");
HttpURLConnection conn = (HttpURLConnection) new URL(url.toString()).openConnection();
try {
conn.setRequestMethod("GET");
conn.setRequestProperty("Authorization", "Bearer " + token);
conn.setRequestProperty("Accept", "image/*");
conn.setConnectTimeout(CONNECT_TIMEOUT_MS);
conn.setReadTimeout(READ_TIMEOUT_MS);
int code = conn.getResponseCode();
Log.i(TAG, "fetch: mxc=" + mxc + " status=" + code);
if (code < 200 || code >= 300) return null;
int contentLength = conn.getContentLength();
if (contentLength > MAX_DECODED_BYTES) {
Log.w(TAG, "fetch: oversized contentLength=" + contentLength + " mxc=" + mxc);
return null;
}
try (InputStream in = conn.getInputStream()) {
BitmapFactory.Options opts = new BitmapFactory.Options();
// Stick with ARGB_8888 even on low-mem devices RGB_565
// would lose alpha (group avatars often have a
// transparent corner) and the cache cap (4 MB) already
// bounds total memory. inJustDecodeBounds + sample-size
// dance is overkill at 96×96.
opts.inPreferredConfig = Bitmap.Config.ARGB_8888;
Bitmap bmp = BitmapFactory.decodeStream(in, null, opts);
if (bmp == null) {
Log.w(TAG, "fetch: decodeStream returned null mxc=" + mxc);
return null;
}
if (bmp.getByteCount() > MAX_DECODED_BYTES) {
Log.w(TAG, "fetch: decoded oversized "
+ bmp.getByteCount() + " bytes mxc=" + mxc);
bmp.recycle();
return null;
}
// Crop into a circle BEFORE caching IconCompat.createWithBitmap
// renders the bitmap verbatim, with no shape mask, so a
// square thumbnail from the homeserver lands as a square
// tile in the shade (visible on Android 12+ where
// conversation Person icons used to be auto-rounded by the
// OS this changed). Pre-cropping guarantees a round
// visual on every API level instead of relying on the
// SystemUI of the day. The original square bitmap is
// recycled once the circular copy is in hand.
return toCircularBitmap(bmp);
}
} finally {
conn.disconnect();
}
}
/**
* Re-encode a circular avatar as an adaptive-icon-shaped bitmap:
* embeds the avatar inside a transparent canvas whose total size is
* 1.5× the avatar so Android's adaptive-icon safe zone (66% of total)
* covers the entire avatar without clipping.
*
* Required for conversation-shortcut icons per docs at
* developer.android.com/develop/ui/views/notifications/conversations:
* *"To avoid unintentional clipping of your shortcut avatar, provide
* an AdaptiveIconDrawable for the shortcut's icon."*
*
* Without this padding, IconCompat.createWithAdaptiveBitmap would
* crop ~17% off every edge of the avatar to fit the safe zone a
* visible mutilation. With it, the shortcut icon renders pixel-
* identical to the circular avatar inside the system shade's
* conversation slot.
*/
static Bitmap toAdaptivePaddedBitmap(Bitmap circularAvatar) {
int avatarSize = Math.min(circularAvatar.getWidth(), circularAvatar.getHeight());
// Pad to 150% so the adaptive safe-zone (66% of canvas = avatarSize)
// covers the full avatar. Rounded up to keep the canvas even.
int canvasSize = (int) Math.ceil(avatarSize / 0.66f);
if (canvasSize % 2 != 0) canvasSize += 1;
Bitmap output = Bitmap.createBitmap(canvasSize, canvasSize, Bitmap.Config.ARGB_8888);
Canvas canvas = new Canvas(output);
int offset = (canvasSize - avatarSize) / 2;
canvas.drawBitmap(circularAvatar, offset, offset, null);
return output;
}
/**
* Return a circular ARGB_8888 bitmap of the source centre-cropped to
* a square if non-square, then masked with a circular path so the
* corners are transparent. The source bitmap is recycled.
*
* Anti-aliased edges via Paint.setAntiAlias on the circle draw the
* BitmapShader copies the source's pixels into the circular region in
* a single drawCircle call, which keeps allocation to one output
* bitmap (vs the naive "decode → square crop → mask compose" path
* that touches three intermediate bitmaps).
*/
private static Bitmap toCircularBitmap(Bitmap source) {
int size = Math.min(source.getWidth(), source.getHeight());
Bitmap squareSource;
if (source.getWidth() == size && source.getHeight() == size) {
squareSource = source;
} else {
int x = (source.getWidth() - size) / 2;
int y = (source.getHeight() - size) / 2;
squareSource = Bitmap.createBitmap(source, x, y, size, size);
source.recycle();
}
Bitmap output = Bitmap.createBitmap(size, size, Bitmap.Config.ARGB_8888);
Canvas canvas = new Canvas(output);
Paint paint = new Paint();
paint.setAntiAlias(true);
paint.setShader(new BitmapShader(
squareSource, Shader.TileMode.CLAMP, Shader.TileMode.CLAMP));
float radius = size / 2f;
canvas.drawCircle(radius, radius, radius, paint);
if (squareSource != source) {
squareSource.recycle();
}
return output;
}
private static final class Parsed {
final String server;
final String mediaId;
Parsed(String server, String mediaId) {
this.server = server;
this.mediaId = mediaId;
}
}
/**
* Split an `mxc://server/mediaId` URL into its two components. Returns
* null on any malformed input caller drops the avatar silently.
*/
private static Parsed parseMxc(String mxc) {
if (mxc == null) return null;
final String prefix = "mxc://";
if (!mxc.startsWith(prefix)) return null;
int slash = mxc.indexOf('/', prefix.length());
if (slash < 0 || slash == prefix.length()) return null;
String server = mxc.substring(prefix.length(), slash);
String mediaId = mxc.substring(slash + 1);
if (server.isEmpty() || mediaId.isEmpty()) return null;
return new Parsed(server, mediaId);
}
}

View file

@ -0,0 +1,44 @@
package chat.vojo.app;
import android.app.NotificationManager;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
/**
* Dismisses the incoming-call notification when its RTC lifetime expires.
*
* Scheduled by {@link VojoFirebaseMessagingService} via AlarmManager at
* sender_ts + lifetime. The extras carry tag/id so we can target the exact
* notification even if multiple ring pushes overlap (rare but possible across
* rapid re-dials).
*
* Also invoked directly (same intent shape) when the app receives a decline /
* other-device-answer via the live Matrix sync and wants the system notification
* cleared see Capacitor removeDeliveredNotifications in the JS layer.
*/
public class CallCancelReceiver extends BroadcastReceiver {
public static final String ACTION_CANCEL_CALL = "chat.vojo.app.CANCEL_CALL";
public static final String EXTRA_NOTIF_TAG = "notif_tag";
public static final String EXTRA_NOTIF_ID = "notif_id";
// Carried so the receiver can tombstone and drop the matching registry
// entry otherwise a same-eventId re-delivery after expiry would seed the
// registry again and render a stale ring on next backgrounding.
public static final String EXTRA_NOTIF_EVENT_ID = "notif_event_id";
@Override
public void onReceive(Context context, Intent intent) {
if (intent == null || !ACTION_CANCEL_CALL.equals(intent.getAction())) return;
String tag = intent.getStringExtra(EXTRA_NOTIF_TAG);
int id = intent.getIntExtra(EXTRA_NOTIF_ID, -1);
String notifEventId = intent.getStringExtra(EXTRA_NOTIF_EVENT_ID);
if (tag == null || id == -1) return;
NotificationManager nm =
(NotificationManager) context.getSystemService(Context.NOTIFICATION_SERVICE);
if (nm != null) nm.cancel(tag, id);
if (notifEventId != null) {
VojoFirebaseMessagingService.removeIncomingRing(context, notifEventId);
}
}
}

View file

@ -0,0 +1,219 @@
package chat.vojo.app;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
import android.content.SharedPreferences;
import android.util.Log;
import androidx.core.app.NotificationManagerCompat;
import org.json.JSONObject;
import java.io.OutputStream;
import java.net.HttpURLConnection;
import java.net.URL;
import java.net.URLEncoder;
import java.util.UUID;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
/**
* Sends {@code m.call.decline} directly from the notification action tap,
* bypassing the WebView boot entirely. Fires via {@code PendingIntent.getBroadcast}
* from the CallStyle Decline button MainActivity never starts, so the
* lockscreen-unlock-and-app-flash UX from Phase 2.5.3 is gone.
*
* Session data ({@code accessToken}, {@code baseUrl}, {@code userId}) is
* mirrored from JS into {@code shared_prefs/CapacitorStorage.xml} by
* {@code writeSessionBridge()} on client mount.
*
* Recovery contract:
* - Before HTTP: write {@code vojo.pendingDeclines.{notifEventId}} tombstone
* so {@code usePendingDeclinesFlusher} can retry on app resume if our
* HTTP fails (network drop, token invalidated, process killed mid-PUT).
* - On 2xx: remove the tombstone receiver path succeeded, no flusher work.
* - On non-2xx or exception: leave tombstone; flusher drains on next resume.
*
* Null-session edge case (fresh reinstall + first push before first login):
* we cannot send the decline at all there's no access token, and the
* JS-path can't cover us either (a logged-out client has no Matrix session
* to call sendRtcDecline against). Cancelling the notification is the only
* feedback we can give; leaving the ring would trap the user on a call they
* can't accept or decline until the A-side times out.
*
* Note on idempotency: the flusher's retry generates a new txnId, so on a
* split-success-fail sequence (receiver HTTP timed out, flusher succeeds)
* we may land two {@code m.call.decline} events in the timeline with the
* same {@code rel.event_id}. This is cosmetic the caller's auto-hangup
* hook is idempotent and fires on the first decline.
*/
public class CallDeclineReceiver extends BroadcastReceiver {
public static final String ACTION_DECLINE_CALL = "chat.vojo.app.DECLINE_CALL";
public static final String EXTRA_ROOM_ID = "room_id";
public static final String EXTRA_NOTIF_EVENT_ID = "notif_event_id";
public static final String EXTRA_NOTIF_TAG = "notif_tag";
public static final String EXTRA_NOTIF_ID = "notif_id";
private static final String PREFS_FILE = "CapacitorStorage";
private static final String SESSION_KEY = "vojo.matrixSession";
private static final String PENDING_DECLINES_PREFIX = "vojo.pendingDeclines.";
private static final int CONNECT_TIMEOUT_MS = 8_000;
private static final int READ_TIMEOUT_MS = 8_000;
private static final String TAG = "CallDeclineReceiver";
// Single reusable executor keeps us off the main thread without spawning
// a fresh one per broadcast declines come rarely enough that a pool of 1
// is fine; a short-lived Thread would also work but is noisier on tracing.
private static final ExecutorService EXECUTOR = Executors.newSingleThreadExecutor();
@Override
public void onReceive(Context context, Intent intent) {
if (intent == null) return;
final String roomId = intent.getStringExtra(EXTRA_ROOM_ID);
final String notifEventId = intent.getStringExtra(EXTRA_NOTIF_EVENT_ID);
final String notifTag = intent.getStringExtra(EXTRA_NOTIF_TAG);
final int notifId = intent.getIntExtra(EXTRA_NOTIF_ID, -1);
if (roomId == null || notifEventId == null) {
Log.w(TAG, "onReceive: missing extras, abort");
return;
}
final Context appContext = context.getApplicationContext();
final SharedPreferences prefs =
appContext.getSharedPreferences(PREFS_FILE, Context.MODE_PRIVATE);
final String sessionJson = prefs.getString(SESSION_KEY, null);
if (sessionJson == null) {
// Fresh reinstall / logged-out: no access token, no Matrix client.
// We can't send m.call.decline and neither can the JS-path (no
// session to decline from). Cancel the notif so the user isn't
// stuck on a ring they can't action. Skip the tombstone a
// retry without a session would be equally impotent.
Log.w(TAG, "onReceive: no session in prefs, cancelling notif without HTTP");
if (notifTag != null && notifId != -1) {
NotificationManagerCompat.from(appContext).cancel(notifTag, notifId);
}
VojoFirebaseMessagingService.removeIncomingRing(appContext, notifEventId);
return;
}
final String accessToken;
final String baseUrl;
try {
JSONObject session = new JSONObject(sessionJson);
accessToken = session.optString("accessToken", null);
baseUrl = session.optString("baseUrl", null);
} catch (Throwable t) {
// Do NOT pass the Throwable JSONException.getMessage() embeds
// the malformed input, which here contains the access token.
Log.e(TAG, "onReceive: prefs JSON parse failed: " + t.getClass().getSimpleName());
// Still drop the native and the registry entry user tapped Decline,
// the ring should not re-surface on next backgrounding just because
// we can't send the decline over HTTP.
if (notifTag != null && notifId != -1) {
NotificationManagerCompat.from(appContext).cancel(notifTag, notifId);
}
VojoFirebaseMessagingService.removeIncomingRing(appContext, notifEventId);
return;
}
if (accessToken == null || accessToken.isEmpty() || baseUrl == null || baseUrl.isEmpty()) {
Log.w(TAG, "onReceive: empty accessToken/baseUrl in session, cancelling ring locally");
if (notifTag != null && notifId != -1) {
NotificationManagerCompat.from(appContext).cancel(notifTag, notifId);
}
VojoFirebaseMessagingService.removeIncomingRing(appContext, notifEventId);
return;
}
// 1. Cancel first for instant user feedback. Ringtone stops within
// NotificationManagerCompat's Binder latency (~tens of ms) HTTP
// completion is irrelevant to the perceived UX.
if (notifTag != null && notifId != -1) {
NotificationManagerCompat.from(appContext).cancel(notifTag, notifId);
}
// Drop the registry entry and tombstone the eventId so the next
// onPause renderRegistry can't resurrect the declined ring.
VojoFirebaseMessagingService.removeIncomingRing(appContext, notifEventId);
// 2. Write tombstone BEFORE HTTP so the flusher sees work to do
// if we fail/die. Remove only on confirmed 2xx.
try {
JSONObject tombstone = new JSONObject();
tombstone.put("roomId", roomId);
tombstone.put("ts", System.currentTimeMillis());
prefs.edit()
.putString(PENDING_DECLINES_PREFIX + notifEventId, tombstone.toString())
.apply();
} catch (Throwable t) {
Log.w(TAG, "onReceive: tombstone write failed (non-fatal)", t);
}
// 3. HTTP PUT off-main on goAsync. Receiver stays alive ~10s for us
// to finish; after that Android reclaims the process and the
// pending request dies flusher covers recovery on next resume.
final PendingResult pendingResult = goAsync();
final String txnId = UUID.randomUUID().toString();
EXECUTOR.execute(() -> {
try {
int status = sendDecline(baseUrl, accessToken, roomId, notifEventId, txnId);
if (status >= 200 && status < 300) {
prefs.edit().remove(PENDING_DECLINES_PREFIX + notifEventId).apply();
Log.d(TAG, "decline PUT ok status=" + status + " room=" + roomId);
} else {
Log.w(TAG, "decline PUT non-2xx status=" + status + " room=" + roomId);
}
} catch (Throwable t) {
Log.e(TAG, "decline PUT threw", t);
} finally {
pendingResult.finish();
}
});
}
private int sendDecline(
String baseUrl,
String accessToken,
String roomId,
String notifEventId,
String txnId
) throws Exception {
String url = trimTrailingSlash(baseUrl)
+ "/_matrix/client/v3/rooms/"
+ URLEncoder.encode(roomId, "UTF-8")
+ "/send/org.matrix.msc4310.rtc.decline/"
+ URLEncoder.encode(txnId, "UTF-8");
JSONObject relates = new JSONObject();
relates.put("rel_type", "m.reference");
relates.put("event_id", notifEventId);
JSONObject body = new JSONObject();
body.put("m.relates_to", relates);
byte[] payload = body.toString().getBytes("UTF-8");
HttpURLConnection conn = null;
try {
conn = (HttpURLConnection) new URL(url).openConnection();
conn.setRequestMethod("PUT");
conn.setConnectTimeout(CONNECT_TIMEOUT_MS);
conn.setReadTimeout(READ_TIMEOUT_MS);
conn.setDoOutput(true);
conn.setRequestProperty("Authorization", "Bearer " + accessToken);
conn.setRequestProperty("Content-Type", "application/json");
conn.setFixedLengthStreamingMode(payload.length);
try (OutputStream os = conn.getOutputStream()) {
os.write(payload);
}
return conn.getResponseCode();
} finally {
if (conn != null) conn.disconnect();
}
}
private static String trimTrailingSlash(String s) {
return (s != null && s.endsWith("/")) ? s.substring(0, s.length() - 1) : s;
}
}

View file

@ -0,0 +1,148 @@
package chat.vojo.app;
import android.Manifest;
import android.content.Context;
import android.content.Intent;
import android.content.pm.PackageManager;
import android.os.Build;
import android.util.Log;
import androidx.core.content.ContextCompat;
import com.getcapacitor.Plugin;
import com.getcapacitor.PluginCall;
import com.getcapacitor.PluginMethod;
import com.getcapacitor.annotation.CapacitorPlugin;
import java.util.HashMap;
import java.util.Map;
/**
* JS Android bridge for CallForegroundService lifecycle.
*
* start / stop map onto startForegroundService / stopService.
*
* RECORD_AUDIO permission is re-verified here before dispatch: the JS caller
* (useAndroidCallForegroundSync) gates on the widget's JoinCall signal, which
* implies getUserMedia has run and the grant is in place but the plugin
* checks defensively so the service never attempts startForeground with
* TYPE_MICROPHONE without the permission. The manifest declares the service
* as foregroundServiceType="microphone" only (no fallback type), so on API
* 34+ starting without RECORD_AUDIO would throw ForegroundServiceTypeException.
*/
@CapacitorPlugin(name = "CallForegroundService")
public class CallForegroundPlugin extends Plugin {
private static final String TAG = "CallFgsPlugin";
@PluginMethod
public void start(PluginCall call) {
String title = call.getString("title");
String body = call.getString("body");
Context ctx = getContext();
// Defense-in-depth: starting the microphone-typed FGS without
// RECORD_AUDIO granted is invalid on API 34+. JS side already gates
// on JoinCall (see useAndroidCallForegroundSync) so this should never
// fire in practice. If it does, resolve cleanly without starting
// the call will run without retention, which is the same fate as
// the first-ever-call window before getUserMedia prompt answered.
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.R) {
int micPerm = ContextCompat.checkSelfPermission(ctx, Manifest.permission.RECORD_AUDIO);
if (micPerm != PackageManager.PERMISSION_GRANTED) {
Log.w(TAG, "start: RECORD_AUDIO not granted, skipping FGS (would fail on TYPE_MICROPHONE)");
call.resolve();
return;
}
}
Intent intent = new Intent(ctx, CallForegroundService.class);
if (title != null) intent.putExtra(CallForegroundService.EXTRA_TITLE, title);
if (body != null) intent.putExtra(CallForegroundService.EXTRA_BODY, body);
try {
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.O) {
ctx.startForegroundService(intent);
} else {
ctx.startService(intent);
}
Log.d(TAG, "start: service started");
call.resolve();
} catch (Throwable t) {
Log.e(TAG, "start: failed to start service", t);
call.reject("start_failed: " + t.getClass().getSimpleName() + ": " + t.getMessage());
}
}
@PluginMethod
public void stop(PluginCall call) {
Context ctx = getContext();
try {
ctx.stopService(new Intent(ctx, CallForegroundService.class));
Log.d(TAG, "stop: stopService dispatched");
call.resolve();
} catch (Throwable t) {
Log.w(TAG, "stop: stopService threw", t);
call.reject("stop_failed: " + t.getClass().getSimpleName() + ": " + t.getMessage());
}
}
// JS upserts a live incoming ring into the native registry (atom ADD
// happy-path). Idempotent with any prior FCM seed for the same eventId
// Java merges metadata fields append-only. See VojoFirebaseMessagingService
// for the registry operations contract.
@PluginMethod
public void upsertIncomingRing(PluginCall call) {
String eventId = call.getString("eventId");
String roomId = call.getString("roomId");
if (eventId == null || eventId.isEmpty() || roomId == null || roomId.isEmpty()) {
call.reject("missing_eventId_or_roomId");
return;
}
Map<String, String> data = new HashMap<>();
data.put("event_id", eventId);
data.put("room_id", roomId);
String callerName = call.getString("callerName");
if (callerName != null && !callerName.isEmpty()) {
data.put("sender_display_name", callerName);
}
// Pass through senderTs/lifetime as strings registry stores the same
// Map<String,String> shape that FCM delivers, and downstream consumers
// (scheduleCallNotificationExpiry, isExpired) parseLong them.
Long senderTs = call.getLong("senderTs");
if (senderTs != null && senderTs > 0) {
data.put("content_sender_ts", Long.toString(senderTs));
}
Long lifetime = call.getLong("lifetime");
if (lifetime != null && lifetime > 0) {
data.put("content_lifetime", Long.toString(lifetime));
}
String messageId = call.getString("messageId");
// messageId is used as google.message_id in the Answer/Launch PendingIntent
// extras Capacitor PushNotificationsPlugin gates pushNotificationActionPerformed
// on containsKey. Empty string also satisfies the gate; we pass the
// caller's value through verbatim.
boolean seeded = VojoFirebaseMessagingService.upsertIncomingRing(data, messageId);
// Mark in NotificationDedup so a polling fire 15 minutes later
// doesn't post a "Missed call" notification for a ring the user
// already saw live via the in-app strip. Mirrors the FCM-arrival
// path in VojoFirebaseMessagingService.onMessageReceived.
if (seeded) {
NotificationDedup.markNotified(getContext(), eventId);
}
call.resolve();
}
// JS removes a ring from the native registry (atom REMOVE / suppress path /
// native action receiver path). Tombstones the eventId to reject late
// FCM or /sync re-seeds within the ring lifetime.
@PluginMethod
public void removeIncomingRing(PluginCall call) {
String eventId = call.getString("eventId");
if (eventId == null || eventId.isEmpty()) {
call.reject("missing_event_id");
return;
}
VojoFirebaseMessagingService.removeIncomingRing(getContext(), eventId);
call.resolve();
}
}

View file

@ -0,0 +1,143 @@
package chat.vojo.app;
import android.app.NotificationChannel;
import android.app.NotificationManager;
import android.app.PendingIntent;
import android.app.Service;
import android.content.Context;
import android.content.Intent;
import android.content.pm.ServiceInfo;
import android.os.Build;
import android.os.IBinder;
import android.util.Log;
import androidx.core.app.NotificationCompat;
/**
* Foreground service kept alive for the duration of an active DM call. Its
* sole job is to promote the process to PROCESS_STATE_FOREGROUND_SERVICE so
* Android doesn't:
* - revoke RECORD_AUDIO via AppOps (API 31+ while-in-use gating);
* - apply background network firewall via netd.
*
* Both revocations were observed in Phase 0 capture on Samsung OneUI API 36:
* mic went Active=false ~5s after screen-off, netd isBlocked=true ~13s after,
* causing Element Call inside the hidden WebView to tear down the LiveKit
* session and the call to drop.
*
* Preconditions enforced by callers:
* - RECORD_AUDIO runtime permission granted (plugin-side check in
* CallForegroundPlugin.start). The manifest declares
* foregroundServiceType="microphone" only, so TYPE_NONE is not a valid
* fallback on API 34+ we never attempt one.
* - JS side gates on useCallJoined so the widget's getUserMedia has already
* prompted for and received the grant by the time we start.
*/
public class CallForegroundService extends Service {
public static final String EXTRA_TITLE = "title";
public static final String EXTRA_BODY = "body";
private static final String CHANNEL_ID = "vojo_calls_ongoing";
// Stable id, distinct from VojoFirebaseMessagingService.SUMMARY_NOTIFICATION_ID
// (Integer.MIN_VALUE) and from per-room call ids (String.hashCode of "call_<roomId>").
private static final int NOTIFICATION_ID = 0x766F6A6F; // "vojo" as ASCII bytes
private static final String TAG = "CallFgs";
@Override
public int onStartCommand(Intent intent, int flags, int startId) {
String title = intent != null ? intent.getStringExtra(EXTRA_TITLE) : null;
String body = intent != null ? intent.getStringExtra(EXTRA_BODY) : null;
if (title == null || title.isEmpty()) title = "Активный звонок";
if (body == null) body = "";
NotificationManager nm = (NotificationManager) getSystemService(Context.NOTIFICATION_SERVICE);
if (nm == null) {
Log.w(TAG, "onStartCommand: NotificationManager null, cannot start FGS");
stopSelf(startId);
return START_NOT_STICKY;
}
ensureOngoingChannel(nm);
Intent launchIntent = new Intent(this, MainActivity.class)
.setAction(Intent.ACTION_MAIN)
.addCategory(Intent.CATEGORY_LAUNCHER)
.addFlags(Intent.FLAG_ACTIVITY_NEW_TASK | Intent.FLAG_ACTIVITY_SINGLE_TOP);
int piFlags = PendingIntent.FLAG_UPDATE_CURRENT
| (Build.VERSION.SDK_INT >= Build.VERSION_CODES.M ? PendingIntent.FLAG_IMMUTABLE : 0);
PendingIntent launchPI = PendingIntent.getActivity(this, 0, launchIntent, piFlags);
NotificationCompat.Builder builder = new NotificationCompat.Builder(this, CHANNEL_ID)
.setSmallIcon(R.mipmap.ic_launcher)
.setContentTitle(title)
.setContentText(body)
.setCategory(NotificationCompat.CATEGORY_CALL)
.setOngoing(true)
.setAutoCancel(false)
.setOnlyAlertOnce(true)
.setPriority(NotificationCompat.PRIORITY_LOW)
.setContentIntent(launchPI);
try {
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.R) {
// API 30+: FOREGROUND_SERVICE_TYPE_MICROPHONE constant exists
// and 3-arg startForeground is available. API 34+ REQUIRES
// the type to match the manifest we declared `microphone`
// and always pass it. RECORD_AUDIO grant is ensured by
// CallForegroundPlugin before this code runs.
startForeground(NOTIFICATION_ID, builder.build(),
ServiceInfo.FOREGROUND_SERVICE_TYPE_MICROPHONE);
Log.d(TAG, "startForeground ok type=microphone");
} else {
// API 24-29: 2-arg form; manifest foregroundServiceType attribute
// is enough for the OS to classify the service correctly.
startForeground(NOTIFICATION_ID, builder.build());
Log.d(TAG, "startForeground ok (pre-R, manifest-driven type)");
}
} catch (Throwable t) {
// If startForeground with TYPE_MICROPHONE throws despite our
// precondition checks (unexpected OEM behavior, race, manifest
// drift), we intentionally do NOT retry with TYPE_NONE that is
// invalid on API 34+ when the manifest declares `microphone`.
// Better to surface the failure and let the call proceed without
// retention than to silently crash with ForegroundServiceTypeException.
Log.e(TAG, "startForeground threw, stopping service without retry", t);
stopSelf(startId);
}
return START_NOT_STICKY;
}
@Override
public IBinder onBind(Intent intent) {
return null;
}
@Override
public void onDestroy() {
Log.d(TAG, "onDestroy");
// Belt-and-suspenders: if the service is being stopped via stopService
// and the FGS flag is still up, make sure the notification goes away.
// Idempotent if stopForeground was already called elsewhere.
stopForeground(STOP_FOREGROUND_REMOVE);
super.onDestroy();
}
private void ensureOngoingChannel(NotificationManager nm) {
if (Build.VERSION.SDK_INT < Build.VERSION_CODES.O) return;
if (nm.getNotificationChannel(CHANNEL_ID) != null) return;
NotificationChannel channel = new NotificationChannel(
CHANNEL_ID,
"Активные звонки",
NotificationManager.IMPORTANCE_LOW
);
channel.setDescription("Уведомление во время активного звонка Vojo");
channel.setShowBadge(false);
channel.enableLights(false);
channel.enableVibration(false);
channel.setSound(null, null);
nm.createNotificationChannel(channel);
}
}

View file

@ -0,0 +1,163 @@
package chat.vojo.app;
import android.content.Context;
import android.content.Intent;
import android.graphics.Bitmap;
import android.os.Build;
import android.util.Log;
import androidx.core.content.LocusIdCompat;
import androidx.core.content.pm.ShortcutInfoCompat;
import androidx.core.content.pm.ShortcutManagerCompat;
import androidx.core.graphics.drawable.IconCompat;
import java.util.Collections;
import java.util.Set;
/**
* Publish a long-lived sharing shortcut for a Matrix room so the system
* treats per-room MessagingStyle notifications as conversations on
* Android 11+ (API 30+).
*
* Without a published shortcut whose id matches the notification's
* setShortcutId(), Android falls back to the app icon for the collapsed-
* preview avatar regardless of Person.setIcon / Builder.setLargeIcon
* Person icons are only consulted by the Conversation styling layer,
* which activates exclusively for notifications backed by a real
* ShortcutInfoCompat marked Long Lived + the SHORTCUT_CATEGORY_CONVERSATION
* sharing category.
*
* Idempotent: republishing the same shortcut id is the documented "update"
* path; ShortcutManagerCompat handles dedup internally. Cheap to call
* from the render hot path (~ms on warm system, indistinguishable from a
* SharedPreferences write at our scale).
*/
final class ConversationShortcuts {
private static final String TAG = "ConvShortcuts";
private ConversationShortcuts() {}
/**
* Publish or refresh the shortcut backing a room's conversation
* notification. No-op on API < 30 Conversation styling is an
* Android 11+ feature; older OS versions render the notification
* fine without the shortcut, and the largeIcon/Person.setIcon
* pipeline is the primary avatar source on them.
*
* @param ctx Context for the shortcut manager binding.
* @param roomId Matrix room id, used as the shortcut id so it
* matches NotificationCompat.Builder.setShortcutId.
* @param isDirect Whether the room is a DM; flips the shortcut
* category so launchers can group DMs separately.
* @param label Short visible label, typically the room name (or
* the peer's display name for a DM).
* @param avatar Optional cached avatar bitmap. Null falls through
* to the app launcher icon still publishes the
* shortcut so the conversation styling activates.
*/
/**
* Returns the published ShortcutInfoCompat so the caller can attach
* it directly to the notification via setShortcutInfo() this is
* the documented "atomic publish + bind" path that avoids the race
* where the notification posts before the shortcut publish has
* settled and Android sees an orphan shortcut id. Null on API < 30,
* null on failure (notification still posts cleanly).
*/
static ShortcutInfoCompat publishForRoom(
Context ctx,
String roomId,
boolean isDirect,
String label,
Bitmap avatar
) {
if (Build.VERSION.SDK_INT < Build.VERSION_CODES.R) {
return null;
}
if (roomId == null || roomId.isEmpty()) return null;
try {
// Conversation shortcut icon MUST be adaptive official docs:
// "To avoid unintentional clipping of your shortcut avatar,
// provide an AdaptiveIconDrawable for the shortcut's icon."
// Without this, Android silently falls back to the app's
// launcher icon for the collapsed-shade conversation avatar
// slot, even though shortcut publish + bind succeed.
// Resource icons (mipmap.ic_launcher) already ship with
// adaptive layers in the manifest; bitmap avatars need padding
// so the safe zone doesn't crop them.
IconCompat icon;
if (avatar != null) {
Bitmap padded = AvatarLoader.toAdaptivePaddedBitmap(avatar);
icon = IconCompat.createWithAdaptiveBitmap(padded);
} else {
icon = IconCompat.createWithResource(ctx, R.mipmap.ic_launcher);
}
// Intent the shortcut launches when tapped from the launcher
// long-press menu or share sheet opens MainActivity and
// delivers the same `room_id` extra the notification tap
// path uses, so the existing pushNotificationActionPerformed
// listener navigates correctly.
Intent launchIntent = new Intent(ctx, MainActivity.class)
.setAction(Intent.ACTION_VIEW)
.addFlags(Intent.FLAG_ACTIVITY_SINGLE_TOP | Intent.FLAG_ACTIVITY_CLEAR_TOP)
.putExtra("room_id", roomId)
// Capacitor PushNotificationsPlugin gates its action
// delivery on bundle.containsKey("google.message_id"); we
// attach an empty value so a launcher-initiated open
// takes the same path as a push-tap.
.putExtra("google.message_id", "");
// Constant value of androidx.core's
// ShortcutInfoCompat.SHORTCUT_CATEGORY_CONVERSATION. Hardcoded
// verbatim because older androidx.core in our dependency
// graph doesn't export the constant; the string itself is
// platform-stable per the Android shortcut category contract.
Set<String> categories =
Collections.singleton("android.shortcut.conversation");
ShortcutInfoCompat.Builder b = new ShortcutInfoCompat.Builder(ctx, roomId)
.setShortLabel(label != null && !label.isEmpty() ? label : "Vojo")
.setLongLabel(label != null && !label.isEmpty() ? label : "Vojo")
.setIntent(launchIntent)
.setIcon(icon)
.setLongLived(true)
.setCategories(categories)
// LocusId mirrors the shortcut id; the OS uses it to
// attribute the notification to a specific conversation
// for digital-wellbeing dashboards and bubble grouping.
.setLocusId(new LocusIdCompat(roomId))
// Marks isDirect so launchers / share sheet can present
// person-style affordances on DMs.
.setIsConversation();
// setPerson is only needed for one-on-one conversations to
// unlock direct-share suggestions, but for a DM we also want
// it to anchor the shortcut on the peer's identity. Skipped
// for groups (single Person doesn't represent the room).
if (isDirect) {
b.setPerson(new androidx.core.app.Person.Builder()
// setKey must match the Person.key used in the
// MessagingStyle so Android's conversation
// attribution matches the shortcut to the
// notification on the same identity.
.setKey(roomId)
.setName(label != null ? label : "")
.setIcon(icon)
.build());
}
ShortcutInfoCompat shortcut = b.build();
boolean ok = ShortcutManagerCompat.pushDynamicShortcut(ctx, shortcut);
Log.i(TAG, "publish room=" + roomId + " label=" + label
+ " hasAvatar=" + (avatar != null) + " ok=" + ok);
return shortcut;
} catch (Throwable t) {
// Shortcut publish is best-effort UX a failure must not
// sink the notification. Worst case: collapsed preview
// falls back to app icon (same as before the shortcut path
// existed at all).
Log.w(TAG, "publish failed room=" + roomId, t);
return null;
}
}
}

View file

@ -0,0 +1,75 @@
package chat.vojo.app;
import android.app.NotificationManager;
import android.content.Context;
import android.content.Intent;
import android.net.Uri;
import android.os.Build;
import android.provider.Settings;
import com.getcapacitor.JSObject;
import com.getcapacitor.Plugin;
import com.getcapacitor.PluginCall;
import com.getcapacitor.PluginMethod;
import com.getcapacitor.annotation.CapacitorPlugin;
/**
* Bridges Android 14+ (API 34) full-screen-intent opt-in into JS.
*
* On API 34 `USE_FULL_SCREEN_INTENT` was reclassified from "normal" to
* "special appop" declaring it in the manifest is no longer enough to
* actually display a full-screen notification over the lockscreen. The user
* must opt in via Settings Apps Vojo Full-screen notifications. There's
* no runtime grant API, only a deep-link.
*
* Without the opt-in, `setFullScreenIntent(launchPI, true)` still satisfies
* the AOSP NotificationManagerService gate (so CallStyle doesn't get silently
* dropped), but the notification renders as a regular heads-up and the screen
* doesn't wake over the lockscreen which was the "why is this just a banner
* on the lockscreen?" symptom we saw on the Samsung OneUI test device.
*
* See docs/plans/dm_calls.md ADR 2.5-fsi for the full history.
*/
@CapacitorPlugin(name = "FullScreenIntent")
public class FullScreenIntentPlugin extends Plugin {
@PluginMethod
public void canUseFullScreenIntent(PluginCall call) {
JSObject ret = new JSObject();
ret.put("value", canUseFullScreenIntentInternal());
call.resolve(ret);
}
@PluginMethod
public void openSettings(PluginCall call) {
Context ctx = getContext();
Intent intent;
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.UPSIDE_DOWN_CAKE) {
// API 34+ has a dedicated Settings screen for the full-screen notification opt-in.
intent = new Intent(Settings.ACTION_MANAGE_APP_USE_FULL_SCREEN_INTENT);
intent.setData(Uri.parse("package:" + ctx.getPackageName()));
} else {
// Fallback for API 33: the per-app notification Settings page is the closest
// equivalent and also covers channel-level toggles (mute, DND bypass, etc).
intent = new Intent(Settings.ACTION_APP_NOTIFICATION_SETTINGS);
intent.putExtra(Settings.EXTRA_APP_PACKAGE, ctx.getPackageName());
}
intent.addFlags(Intent.FLAG_ACTIVITY_NEW_TASK);
try {
ctx.startActivity(intent);
call.resolve();
} catch (Throwable t) {
call.reject("Failed to open FSI settings: " + t.getMessage());
}
}
private boolean canUseFullScreenIntentInternal() {
// On API 33 `USE_FULL_SCREEN_INTENT` is a normal permission if it's
// declared in the manifest, the app already has it. Skip the runtime check.
if (Build.VERSION.SDK_INT < Build.VERSION_CODES.UPSIDE_DOWN_CAKE) return true;
NotificationManager nm = (NotificationManager)
getContext().getSystemService(Context.NOTIFICATION_SERVICE);
if (nm == null) return false;
return nm.canUseFullScreenIntent();
}
}

View file

@ -0,0 +1,16 @@
package chat.vojo.app;
import com.getcapacitor.Plugin;
import com.getcapacitor.PluginCall;
import com.getcapacitor.PluginMethod;
import com.getcapacitor.annotation.CapacitorPlugin;
@CapacitorPlugin(name = "LaunchSplash")
public class LaunchSplashPlugin extends Plugin {
@PluginMethod
public void ready(PluginCall call) {
MainActivity.releaseLaunchSplash();
call.resolve();
}
}

View file

@ -0,0 +1,128 @@
package chat.vojo.app;
import android.os.Bundle;
import android.os.Handler;
import android.os.Looper;
import androidx.activity.EdgeToEdge;
import androidx.core.splashscreen.SplashScreen;
import androidx.core.view.WindowCompat;
import androidx.core.view.WindowInsetsControllerCompat;
import com.getcapacitor.BridgeActivity;
public class MainActivity extends BridgeActivity {
public static volatile boolean isInForeground = false;
private static volatile boolean launchSplashReady = false;
// Safety net for setKeepOnScreenCondition: if JS never calls
// launchSplash.ready() (boot crash, exception during config load before
// AuthMascot mounts, network hang in useClientConfig, deep-link straight
// into AuthLayout where the centered AuthMascot variant doesn't render,
// ) the splash would otherwise hang indefinitely and the user can't
// interact with anything. 5s covers normal cold boots on mid-range
// Android (config + bundle parse + first paint typically lands inside
// 1-2s) with comfortable headroom; past it we drop the splash and let
// whatever the web side has rendered take over including blank
// AuthLayout, which is at least recoverable.
private static final long SPLASH_SAFETY_TIMEOUT_MS = 5000L;
// Short debounce on the onPauserenderRegistry edge so an in-flight JS
// removeIncomingRing bridge call (e.g. user accepted/declined, then
// immediately pressed Home) has a chance to land before we post native
// CallStyle for a ring that's about to be removed. 150ms covers the
// strip-accept chain (Capacitor roundtrip + switchOrStartDmCall
// resolve + sync-effect bridge) on mid-range Android; imperceptible
// as a silent-ring delay.
private static final long RENDER_DEBOUNCE_MS = 150L;
// Modest debounce on the onResumecancelRenderedIncomingRings edge so JS
// has a moment to hydrate incomingCallsAtom before the native surface
// goes away. Covers warm resume (hook alive, /sync delivers an ADD
// within ~100-200ms) cleanly. Cold resume (killed process, Matrix
// client rehydration takes 1-3s) still has a "no surface" window
// acceptable tradeoff since tap-native flows carry call_action and
// are handled by pendingCallActionConsumer regardless of atom state.
private static final long CANCEL_DEBOUNCE_MS = 300L;
private final Handler lifecycleHandler = new Handler(Looper.getMainLooper());
private final Runnable renderRunnable = () ->
VojoFirebaseMessagingService.renderRegistry(this);
private final Runnable cancelRunnable = () ->
VojoFirebaseMessagingService.cancelRenderedIncomingRings(this);
public static void releaseLaunchSplash() {
launchSplashReady = true;
}
@Override
protected void onCreate(Bundle savedInstanceState) {
if (savedInstanceState == null) {
launchSplashReady = false;
}
// Custom plugins must be registered before super.onCreate so BridgeActivity
// can wire them into the WebView bridge on load. Registering after
// super.onCreate would make the plugin invisible to JS until the next relaunch.
registerPlugin(FullScreenIntentPlugin.class);
registerPlugin(CallForegroundPlugin.class);
registerPlugin(LaunchSplashPlugin.class);
registerPlugin(ShareTargetPlugin.class);
registerPlugin(PollingPlugin.class);
// AndroidX SplashScreen must be installed before super.onCreate().
// Keep it until the web splash confirms its first visible frame is
// ready, OR the safety timeout elapses (see SPLASH_SAFETY_TIMEOUT_MS).
final long splashStartMs = System.currentTimeMillis();
SplashScreen splashScreen = SplashScreen.installSplashScreen(this);
splashScreen.setKeepOnScreenCondition(() -> {
if (launchSplashReady) return false;
return System.currentTimeMillis() - splashStartMs < SPLASH_SAFETY_TIMEOUT_MS;
});
EdgeToEdge.enable(this);
super.onCreate(savedInstanceState);
// Force light icons on both system bars: our CSS is permanently dark
// (Dawn redesign), but EdgeToEdge.enable auto-detects icon tint from
// the device uiMode on a light-mode device that gives dark icons
// over our dark bars and they vanish.
WindowInsetsControllerCompat controller =
WindowCompat.getInsetsController(getWindow(), getWindow().getDecorView());
controller.setAppearanceLightStatusBars(false);
controller.setAppearanceLightNavigationBars(false);
}
@Override
public void onResume() {
super.onResume();
isInForeground = true;
// Cancel any pending render: user came back before the debounce fired,
// JS strip will own UX, no need to surface native.
lifecycleHandler.removeCallbacks(renderRunnable);
// Defer the native cancel so JS strip has a moment to hydrate from
// incomingCallsAtom. Registry entries persist they still represent
// live rings, just the native surfaces go.
lifecycleHandler.removeCallbacks(cancelRunnable);
lifecycleHandler.postDelayed(cancelRunnable, CANCEL_DEBOUNCE_MS);
}
@Override
public void onPause() {
super.onPause();
isInForeground = false;
// Re-backgrounding: don't cancel a native the user will still need
// visible. The render runnable below will re-render if needed.
lifecycleHandler.removeCallbacks(cancelRunnable);
// Schedule render user is backgrounding, JS audio gate is about to
// close, native CallStyle must surface or the ring goes silent. Debounce
// absorbs the bridge-call race: if onResume fires within RENDER_DEBOUNCE_MS
// (user bounce), the render is cancelled.
lifecycleHandler.removeCallbacks(renderRunnable);
lifecycleHandler.postDelayed(renderRunnable, RENDER_DEBOUNCE_MS);
}
@Override
public void onDestroy() {
// Drop any pending render/cancel so runnables which capture `this`
// can't fire post-destroy and land an nm.notify / nm.cancel against a
// dead Activity context on config change (rotation) or process teardown.
lifecycleHandler.removeCallbacksAndMessages(null);
super.onDestroy();
}
}

View file

@ -0,0 +1,147 @@
package chat.vojo.app;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
import android.content.SharedPreferences;
import android.util.Log;
import java.io.IOException;
import java.net.HttpURLConnection;
import java.net.URL;
import java.net.URLEncoder;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
/**
* Handles the per-notification "Mark as read" action.
*
* Posts {@code POST /_matrix/client/v3/rooms/{roomId}/receipt/m.read/{eventId}}
* using the access token saved by the polling lifecycle in
* {@code vojo_poll_state} SharedPreferences (same storage VojoPollWorker uses;
* keeps the credential lifecycle single-sourced). After a successful 2xx the
* per-room MessagingStyle notification is dismissed and the
* {@link RoomMessageCache} is cleared so the next push to that room starts a
* fresh conversation rather than re-appending to the prior history.
*
* Dismiss policy: OPTIMISTIC. The per-room notification is dismissed
* synchronously in onReceive before the HTTP receipt PUT is even
* attempted so the user sees instant feedback. The async receipt POST
* happens on a worker thread afterwards. This mirrors element-android's
* NotificationBroadcastReceiver pattern and matches the user's mental
* model ("I tapped, it should disappear immediately").
*
* Failure mode: on any non-2xx or thrown exception we accept that the
* server-side read receipt did not land. We do NOT re-post the
* notification or implement a flusher because:
* - the next room open from the JS app issues a fresh read-receipt
* for the latest visible event, catching up the server state
* - the in-app read-marker logic is the authoritative path; this
* receiver is a convenience for the shade-tap shortcut
* - accumulating tombstones in prefs (the CallDeclineReceiver pattern)
* would risk leaking historical eventIds the JS side would re-issue
* on app resume anyway
*
* Null-credential edge case (fresh install + first push before any
* saveSession bridge): no token to use, we still dismiss the notification
* locally so the user isn't stuck looking at a "stuck" Mark-as-read
* button. The next room open from JS covers the server view.
*/
public class MarkAsReadReceiver extends BroadcastReceiver {
public static final String ACTION_MARK_AS_READ = "chat.vojo.app.MARK_AS_READ";
public static final String EXTRA_ROOM_ID = "room_id";
public static final String EXTRA_EVENT_ID = "event_id";
private static final int CONNECT_TIMEOUT_MS = 8_000;
private static final int READ_TIMEOUT_MS = 8_000;
private static final String TAG = "MarkAsReadRcvr";
private static final ExecutorService EXECUTOR = Executors.newSingleThreadExecutor();
@Override
public void onReceive(Context context, Intent intent) {
if (intent == null) return;
final String roomId = intent.getStringExtra(EXTRA_ROOM_ID);
final String eventId = intent.getStringExtra(EXTRA_EVENT_ID);
if (roomId == null || roomId.isEmpty()) {
Log.w(TAG, "onReceive: missing room_id, abort");
return;
}
final Context appContext = context.getApplicationContext();
// Dismiss first for instant UX feedback HTTP latency is irrelevant
// to the perceived "marked as read" action.
VojoFirebaseMessagingService.dismissRoomNotification(appContext, roomId);
final SharedPreferences prefs = appContext.getSharedPreferences(
VojoPollWorker.PREFS, Context.MODE_PRIVATE);
final String token = prefs.getString(VojoPollWorker.KEY_ACCESS_TOKEN, null);
final String homeserver = prefs.getString(VojoPollWorker.KEY_HOMESERVER_URL, null);
if (token == null || token.isEmpty() || homeserver == null || homeserver.isEmpty()) {
Log.w(TAG, "onReceive: no credentials in prefs, local dismiss only");
return;
}
if (eventId == null || eventId.isEmpty()) {
// Without an eventId we cannot issue a receipt PUT the JS-side
// read-marker handler will catch this up on the next room open.
Log.w(TAG, "onReceive: no event_id, local dismiss only");
return;
}
final PendingResult pendingResult = goAsync();
EXECUTOR.execute(() -> {
try {
int status = sendReceipt(homeserver, token, roomId, eventId);
if (status >= 200 && status < 300) {
if (BuildConfig.DEBUG) {
Log.d(TAG, "receipt ok status=" + status + " room=" + roomId);
}
} else {
Log.w(TAG, "receipt non-2xx status=" + status + " room=" + roomId);
}
} catch (Throwable t) {
Log.w(TAG, "receipt threw room=" + roomId, t);
} finally {
pendingResult.finish();
}
});
}
private int sendReceipt(
String baseUrl,
String accessToken,
String roomId,
String eventId
) throws IOException {
String url = trimTrailingSlash(baseUrl)
+ "/_matrix/client/v3/rooms/"
+ URLEncoder.encode(roomId, "UTF-8")
+ "/receipt/m.read/"
+ URLEncoder.encode(eventId, "UTF-8");
HttpURLConnection conn = (HttpURLConnection) new URL(url).openConnection();
try {
conn.setRequestMethod("POST");
conn.setRequestProperty("Authorization", "Bearer " + accessToken);
conn.setRequestProperty("Content-Type", "application/json");
conn.setConnectTimeout(CONNECT_TIMEOUT_MS);
conn.setReadTimeout(READ_TIMEOUT_MS);
conn.setDoOutput(true);
// Empty JSON body per spec; setFixedLengthStreamingMode keeps the
// connection on the cached path instead of chunked-transfer fallback.
byte[] payload = "{}".getBytes("UTF-8");
conn.setFixedLengthStreamingMode(payload.length);
try (java.io.OutputStream os = conn.getOutputStream()) {
os.write(payload);
}
return conn.getResponseCode();
} finally {
conn.disconnect();
}
}
private static String trimTrailingSlash(String s) {
return (s != null && s.endsWith("/")) ? s.substring(0, s.length() - 1) : s;
}
}

View file

@ -0,0 +1,104 @@
package chat.vojo.app;
import android.content.Context;
import android.content.SharedPreferences;
import java.util.Iterator;
import java.util.LinkedHashSet;
import java.util.Set;
/**
* Cross-source LRU dedup for rendered push event_ids.
*
* Both the FCM service (after a successful nm.notify) and the polling Worker
* write into the same bounded SharedPreferences-backed set. The Worker reads
* it to skip events FCM already delivered which fixes the regression where
* a user who dismissed an FCM notification before polling fired would see
* the same event resurface up to 15 minutes later via the polling fallback.
*
* The native `eventId.hashCode()` notification-id slot is still the primary
* dedup for *concurrent* render (Android NotificationManager replace), but
* that only collapses surfaces while both notifications are still visible;
* once the user dismisses, the slot is empty and the second render would
* post fresh. This shared set covers that gap.
*
* Synchronisation: SharedPreferences read-modify-write is not atomic across
* threads/processes, and FCM service runs on a Firebase-managed background
* thread while the Worker runs on WorkManager's executor. We serialise all
* mutations through a static lock. Critical sections are short (string split
* + LinkedHashSet trim + putString) no Binder calls.
*/
final class NotificationDedup {
// Capacity is intentionally larger than VojoPollWorker's worst-case per-run
// event count (MAX_PAGES_PER_RUN × PAGE_LIMIT = 250). If a single fire
// marks 250 events and the cap were 200, the 50 oldest of those would
// already be evicted by the time we finish writing so a sibling poll
// resuming the same window would re-render them. 500 gives 2× headroom
// while staying ~12 KB in SharedPreferences (negligible).
private static final int MAX_TRACKED = 500;
private static final Object lock = new Object();
private NotificationDedup() {}
/** Returns true iff the given event_id has been notified in a recent cycle. */
static boolean wasNotified(Context ctx, String eventId) {
if (eventId == null || eventId.isEmpty()) return false;
synchronized (lock) {
return readSet(ctx).contains(eventId);
}
}
/** Append the event_id to the LRU set, trimming the oldest when full. */
static void markNotified(Context ctx, String eventId) {
if (eventId == null || eventId.isEmpty()) return;
synchronized (lock) {
Set<String> set = readSet(ctx);
// LinkedHashSet preserves insertion order re-adding moves to tail
// only if we remove-then-add. The Set#add no-op on a present entry
// does NOT refresh position, but the simple "drop oldest" trim
// below is adequate for our scale and matches the Worker's
// existing semantics. Skip the disk write entirely when add()
// returned false the event was already in the set, persistence
// would just churn SharedPreferences for no state change.
if (!set.add(eventId)) return;
if (set.size() > MAX_TRACKED) {
Iterator<String> it = set.iterator();
int drop = set.size() - MAX_TRACKED;
while (it.hasNext() && drop > 0) {
it.next();
it.remove();
drop -= 1;
}
}
writeSet(ctx, set);
}
}
/** Caller must hold {@link #lock}. */
private static Set<String> readSet(Context ctx) {
SharedPreferences prefs = ctx.getSharedPreferences(
VojoPollWorker.PREFS, Context.MODE_PRIVATE);
String raw = prefs.getString(VojoPollWorker.KEY_NOTIFIED_IDS, "");
Set<String> out = new LinkedHashSet<>();
if (raw.isEmpty()) return out;
for (String id : raw.split(",")) {
if (!id.isEmpty()) out.add(id);
}
return out;
}
/** Caller must hold {@link #lock}. */
private static void writeSet(Context ctx, Set<String> set) {
SharedPreferences prefs = ctx.getSharedPreferences(
VojoPollWorker.PREFS, Context.MODE_PRIVATE);
StringBuilder sb = new StringBuilder(set.size() * 25);
boolean first = true;
for (String id : set) {
if (!first) sb.append(',');
sb.append(id);
first = false;
}
prefs.edit().putString(VojoPollWorker.KEY_NOTIFIED_IDS, sb.toString()).apply();
}
}

View file

@ -0,0 +1,37 @@
package chat.vojo.app;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
import android.util.Log;
/**
* Fires when the user swipes a per-room MessagingStyle notification away.
*
* Without this hook, RoomMessageCache would still hold the prior messages
* for that room and the next push would append onto that history and
* re-surface the messages the user just dismissed. With it, swipe clears
* the cache so the next push starts a fresh conversation for the room.
*
* NOTE: this only fires for user-driven dismissals programmatic
* nm.cancel calls (mark-as-read, receipt-driven dismiss, channel migration)
* already call RoomMessageCache.clear themselves and do NOT fire the
* delete intent. There's no double-clear risk.
*/
public class NotificationDismissReceiver extends BroadcastReceiver {
public static final String ACTION_NOTIFICATION_DISMISSED =
"chat.vojo.app.NOTIFICATION_DISMISSED";
public static final String EXTRA_ROOM_ID = "room_id";
private static final String TAG = "DismissRcvr";
@Override
public void onReceive(Context context, Intent intent) {
if (intent == null) return;
String roomId = intent.getStringExtra(EXTRA_ROOM_ID);
if (roomId == null || roomId.isEmpty()) return;
if (BuildConfig.DEBUG) Log.d(TAG, "swipe clear cache room=" + roomId);
RoomMessageCache.clear(roomId);
}
}

View file

@ -0,0 +1,236 @@
package chat.vojo.app;
import android.content.Context;
import android.content.SharedPreferences;
import android.util.Log;
import androidx.work.Constraints;
import androidx.work.ExistingPeriodicWorkPolicy;
import androidx.work.NetworkType;
import androidx.work.PeriodicWorkRequest;
import androidx.work.WorkManager;
import com.getcapacitor.JSObject;
import com.getcapacitor.Plugin;
import com.getcapacitor.PluginCall;
import com.getcapacitor.PluginMethod;
import com.getcapacitor.annotation.CapacitorPlugin;
import java.util.concurrent.TimeUnit;
/**
* JS Android bridge for the WorkManager-based polling fallback.
*
* Lifecycle:
* - JS calls saveSession({accessToken, homeserverUrl, userId}) on login,
* on push (re)enable, and on visibilitychange visible (to recover a
* 401-cleared credentials slot without a full remount).
* - JS calls schedule({intervalMinutes}) once push is enabled. Idempotent:
* KEEP policy means a second schedule() call against an already-enqueued
* worker is a no-op (the running period continues unchanged).
* - JS calls saveRoomNames({names}) on mount + visibilitychange visible
* so VojoPollWorker has a local cache to resolve room_id display name
* without making N extra GET /rooms/{id}/state/m.room.name requests.
* Brand-new rooms created between visibility events fall back to
* sender_display_name in the renderer.
* - JS calls cancel() + clearSession() on logout / push disable.
*
* Worker tag: a single unique periodic worker named UNIQUE_WORK_NAME KEEP
* policy prevents schedule churn from re-creating it. Cancel() removes it
* by the same name.
*/
@CapacitorPlugin(name = "Polling")
public class PollingPlugin extends Plugin {
private static final String TAG = "PollingPlugin";
private static final String UNIQUE_WORK_NAME = "vojo_push_poll";
// Android's hard floor for PeriodicWorkRequest. Requests with shorter
// intervals are silently clamped to 15 minutes. We accept the requested
// value from JS but enforce the floor here so misuse from JS doesn't
// produce a silently-different behavior.
private static final long MIN_INTERVAL_MINUTES = 15;
@PluginMethod
public void saveSession(PluginCall call) {
String accessToken = call.getString("accessToken");
String homeserverUrl = call.getString("homeserverUrl");
if (accessToken == null || accessToken.isEmpty()
|| homeserverUrl == null || homeserverUrl.isEmpty()) {
call.reject("missing_accessToken_or_homeserverUrl");
return;
}
String userId = call.getString("userId");
SharedPreferences prefs = getContext()
.getSharedPreferences(VojoPollWorker.PREFS, Context.MODE_PRIVATE);
SharedPreferences.Editor editor = prefs.edit()
.putString(VojoPollWorker.KEY_ACCESS_TOKEN, accessToken)
.putString(VojoPollWorker.KEY_HOMESERVER_URL, homeserverUrl);
if (userId != null && !userId.isEmpty()) {
editor.putString(VojoPollWorker.KEY_USER_ID, userId);
}
// Seed the watermark to "now minus a small clock-skew buffer" on the
// first saveSession after install / logout. Without seeding the
// Worker's first fire sees watermark=0 and renders every historical
// unread /notifications entry as a fresh push. The buffer covers the
// case where the device clock runs ahead of the homeserver's clock
// event ts is server-side, so a too-fresh local seed would silently
// skip recently-arrived events as "older than watermark" forever.
// 60s tolerates typical NTP drift while still suppressing days-old
// backlog on first enable. We seed only when the key is absent so
// subsequent saveSession calls (token rotation, visibilitychange
// re-bridge) don't reset live state.
if (!prefs.contains(VojoPollWorker.KEY_LAST_SEEN_TS)) {
editor.putLong(
VojoPollWorker.KEY_LAST_SEEN_TS,
System.currentTimeMillis() - SEED_CLOCK_SKEW_BUFFER_MS
);
}
editor.apply();
call.resolve();
}
private static final long SEED_CLOCK_SKEW_BUFFER_MS = 60_000L;
@PluginMethod
public void clearSession(PluginCall call) {
getContext()
.getSharedPreferences(VojoPollWorker.PREFS, Context.MODE_PRIVATE)
.edit()
.remove(VojoPollWorker.KEY_ACCESS_TOKEN)
.remove(VojoPollWorker.KEY_HOMESERVER_URL)
.remove(VojoPollWorker.KEY_USER_ID)
.remove(VojoPollWorker.KEY_LAST_SEEN_TS)
.remove(VojoPollWorker.KEY_DRAIN_CURSOR)
.remove(VojoPollWorker.KEY_DRAIN_TARGET_TS)
.remove(VojoPollWorker.KEY_NOTIFIED_IDS)
.remove(VojoPollWorker.KEY_ROOM_NAMES)
.remove(VojoPollWorker.KEY_USER_AVATARS)
.apply();
call.resolve();
}
/**
* user_id MXC avatar URL snapshot. Mirrors {@link #saveRoomNames}
* stored as a JSON blob in vojo_poll_state for the FCM service /
* polling Worker / ReplyReceiver to consult via
* VojoFirebaseMessagingService.lookupUserAvatarMxc. JS dumps on the
* same lifecycle triggers as room names (mount, visibility resume,
* m.direct change, m.room.encryption flip).
*/
@PluginMethod
public void saveUserAvatars(PluginCall call) {
JSObject avatars = call.getObject("avatars");
if (avatars == null) {
call.reject("missing_avatars");
return;
}
String serialized = avatars.toString();
getContext()
.getSharedPreferences(VojoPollWorker.PREFS, Context.MODE_PRIVATE)
.edit()
.putString(VojoPollWorker.KEY_USER_AVATARS, serialized)
.apply();
Log.i(TAG, "saveUserAvatars: " + avatars.length() + " entries, "
+ serialized.length() + " bytes");
call.resolve();
}
@PluginMethod
public void saveRoomNames(PluginCall call) {
JSObject names = call.getObject("names");
if (names == null) {
// Empty map is also valid (user cleared all rooms) JS passes
// {} explicitly in that case; missing key is a contract bug.
call.reject("missing_names");
return;
}
// `JSObject extends JSONObject`, so names.toString() is already a
// valid JSON serialisation of validated values no need to re-parse
// it through `new JSONObject(...)` just to re-serialise. Persist
// verbatim.
String serialized = names.toString();
getContext()
.getSharedPreferences(VojoPollWorker.PREFS, Context.MODE_PRIVATE)
.edit()
.putString(VojoPollWorker.KEY_ROOM_NAMES, serialized)
.apply();
Log.i(TAG, "saveRoomNames: " + names.length() + " entries, "
+ serialized.length() + " bytes");
call.resolve();
}
@PluginMethod
public void schedule(PluginCall call) {
Integer intervalMinutes = call.getInt("intervalMinutes", 15);
long interval = Math.max(MIN_INTERVAL_MINUTES, intervalMinutes != null ? intervalMinutes : 15);
Constraints constraints = new Constraints.Builder()
.setRequiredNetworkType(NetworkType.CONNECTED)
.build();
PeriodicWorkRequest req = new PeriodicWorkRequest.Builder(
VojoPollWorker.class, interval, TimeUnit.MINUTES
)
.setConstraints(constraints)
.addTag("vojo_push_poll")
.build();
try {
WorkManager.getInstance(getContext())
.enqueueUniquePeriodicWork(
UNIQUE_WORK_NAME,
ExistingPeriodicWorkPolicy.KEEP,
req
);
Log.d(TAG, "scheduled periodic poll every " + interval + " minutes");
call.resolve();
} catch (Throwable t) {
Log.w(TAG, "schedule failed", t);
call.reject("schedule_failed: " + t.getMessage());
}
}
/**
* Dismiss the per-room MessagingStyle notification + clear the in-memory
* RoomMessageCache for the room. Called from the JS receipt listener when
* a server-side read receipt zeroes the unread count (the user read on
* another device / tab). No-op if the notification was never posted or
* has already been swiped away.
*/
@PluginMethod
public void dismissRoom(PluginCall call) {
String roomId = call.getString("roomId");
if (roomId == null || roomId.isEmpty()) {
call.reject("missing_roomId");
return;
}
VojoFirebaseMessagingService.dismissRoomNotification(getContext(), roomId);
call.resolve();
}
@PluginMethod
public void cancel(PluginCall call) {
try {
// Block on the Operation so callers awaiting cancel() see the
// cancel committed to WorkManager's database before we resolve.
// (NOTE: this does NOT interrupt a Worker that's already mid
// doWork(); cooperative cancellation via isStopped() is owned
// by VojoPollWorker itself.) Without this wait a fast
// disablereenable sequence races with ExistingPeriodicWorkPolicy.KEEP
// the second enqueueUniquePeriodicWork can land before the
// cancel is committed and become a no-op. We're already off
// the main thread (Capacitor dispatches plugin calls on its
// own executor), so the blocking get() is safe here.
WorkManager.getInstance(getContext())
.cancelUniqueWork(UNIQUE_WORK_NAME)
.getResult()
.get();
Log.d(TAG, "cancelled periodic poll");
call.resolve();
} catch (Throwable t) {
Log.w(TAG, "cancel failed", t);
call.reject("cancel_failed: " + t.getMessage());
}
}
}

View file

@ -0,0 +1,170 @@
package chat.vojo.app;
import android.content.Context;
import android.content.SharedPreferences;
import android.content.res.Configuration;
import java.util.Locale;
/**
* Locale-aware push-notification strings. Resources (values{,-ru}/
* push_strings.xml) are auto-generated at build time by
* scripts/gen-push-strings.mjs from public/locales/{en,ru}.json so
* the Push namespace has a single source of truth shared with the
* web Service Worker. Do not edit push_strings.xml by hand it will
* be overwritten on the next `npm run android:sync`.
*
* Locale selection: Vojo ships an explicit in-app language picker that
* does NOT have to match the device locale. pushLanguageBridge.ts
* mirrors i18next's current language into Capacitor Preferences
* (shared_prefs/CapacitorStorage.xml, key "vojo.appLanguage") on every
* `languageChanged` event; we read it here and force it onto a
* Configuration-scoped Context before the getString call.
*
* Killed-process pushes may arrive before JS has ever booted no pref
* to read. In that case we fall back to the device locale, normalised
* to {en, ru}; anything else maps to en, matching i18n.ts
* fallbackLng: 'en' on the main thread.
*/
final class PushStrings {
private static final String PREFS_GROUP = "CapacitorStorage";
private static final String LANG_KEY = "vojo.appLanguage";
private PushStrings() {}
static String messageFallback(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_new_message);
}
static String messagesFallback(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_new_messages);
}
static String inviteTitle(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_invitation);
}
static String missedCallTitle(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_missed_call);
}
static String missedCallBody(Context ctx, String caller) {
String safeCaller = caller == null ? "" : caller;
return forAppLocale(ctx).getString(R.string.push_missed_call_body, safeCaller);
}
static String channelGroup(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_channel_group);
}
static String channelDm(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_channel_dm);
}
static String channelDmDescription(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_channel_dm_description);
}
static String channelGroupRoom(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_channel_group_room);
}
static String channelGroupRoomDescription(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_channel_group_room_description);
}
static String selfName(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_self_name);
}
static String markAsReadAction(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_action_mark_as_read);
}
static String replyAction(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_action_reply);
}
static String replyHint(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_reply_hint);
}
static String replyFailed(Context ctx) {
return forAppLocale(ctx).getString(R.string.push_reply_failed);
}
/**
* Build the invite-notification body from inviter + room name, falling
* back through four variants when one or both are absent. The res IDs
* all declare positional formatters (%1$s = inviter, %2$s = roomName)
* to keep the order stable across locales; we always pass both args
* even when only one is used, because Android's formatter requires
* every referenced position to be supplied.
*/
static String inviteBody(Context ctx, String inviter, String roomName) {
boolean hasInviter = inviter != null && !inviter.isEmpty();
boolean hasRoom = roomName != null && !roomName.isEmpty();
int resId;
if (hasInviter && hasRoom) {
resId = R.string.push_invite_body;
} else if (hasInviter) {
resId = R.string.push_invite_body_no_room;
} else if (hasRoom) {
resId = R.string.push_invite_body_no_inviter;
} else {
resId = R.string.push_invite_body_generic;
}
String safeInviter = hasInviter ? inviter : "";
String safeRoom = hasRoom ? roomName : "";
return forAppLocale(ctx).getString(resId, safeInviter, safeRoom);
}
private static Context forAppLocale(Context ctx) {
String lang = chooseLang(ctx);
try {
Configuration cfg = new Configuration(ctx.getResources().getConfiguration());
cfg.setLocale(Locale.forLanguageTag(lang));
return ctx.createConfigurationContext(cfg);
} catch (Throwable t) {
return ctx;
}
}
private static String chooseLang(Context ctx) {
String fromPref = readAppLanguage(ctx);
if (fromPref != null) return fromPref;
try {
Locale loc = Locale.getDefault();
if (loc != null) return normalize(loc.getLanguage());
} catch (Throwable ignored) {
// Locale.getDefault is unusually robust but defensively fall through.
}
return "en";
}
private static String readAppLanguage(Context ctx) {
try {
SharedPreferences prefs =
ctx.getSharedPreferences(PREFS_GROUP, Context.MODE_PRIVATE);
return normalize(prefs.getString(LANG_KEY, null));
} catch (Throwable t) {
return null;
}
}
// Match i18next config: `fallbackLng: 'en'`, `load: 'languageOnly'`.
// Vojo only ships en.json + ru.json; any other Configuration locale
// (e.g. "fr") would bypass both values-ru/ and our bundled resources
// and land on values/ anyway we collapse to "en" explicitly so the
// web and native surfaces render the same lingua-franca default and
// readers of this method don't have to reason about fall-through.
// Clamp on both write (pushLanguageBridge.ts) and read (here) so a
// stale or tampered pref value can't leak through.
private static String normalize(String raw) {
if (raw == null) return null;
String lower = raw.toLowerCase(Locale.ROOT);
if (lower.startsWith("ru")) return "ru";
return "en";
}
}

View file

@ -0,0 +1,248 @@
package chat.vojo.app;
import android.app.NotificationManager;
import android.content.BroadcastReceiver;
import android.content.Context;
import android.content.Intent;
import android.content.SharedPreferences;
import android.os.Bundle;
import android.util.Log;
import androidx.core.app.NotificationCompat;
import androidx.core.app.RemoteInput;
import org.json.JSONObject;
import org.json.JSONException;
import java.io.IOException;
import java.io.OutputStream;
import java.net.HttpURLConnection;
import java.net.URL;
import java.net.URLEncoder;
import java.util.UUID;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
/**
* Handles the inline-reply RemoteInput action on a per-room MessagingStyle
* notification.
*
* Flow:
* 1. User taps reply, types text, presses send broadcast fires here.
* 2. We immediately append the outgoing message to RoomMessageCache and
* re-post the notification (instant UX feedback the message appears
* as a self-Person bubble in the conversation while the HTTP is in
* flight).
* 3. PUT /_matrix/client/v3/rooms/{roomId}/send/m.room.message/{txnId}
* with {msgtype: "m.text", body}. Uses the vojo_poll_state token (same
* storage as Worker / MarkAsReadReceiver single credential lifecycle).
* 4. On 2xx: nothing further; the JS sync echo will eventually replace
* the local-echo bubble in-app.
* 5. On non-2xx or thrown: post a small error notification "Could not
* send your reply" so the user knows to retry from in-app — better
* than silently swallowing the message.
*
* E2EE rooms are guarded UP-STREAM in VojoFirebaseMessagingService.
* renderMessageNotification: we don't even attach the reply action when
* RoomMetadata.isEncrypted is true. So this receiver never has to encrypt.
* Defense in depth: if a stale notification with the action ever survives
* an encryption flip we still detect the failure as a non-2xx HTTP and
* surface the error notification rather than sending cleartext (which
* Synapse would in any case reject for an encrypted room).
*
* Null-credential edge case: post the error notification so the user
* notices and retries in-app. Same logic as a network failure.
*/
public class ReplyReceiver extends BroadcastReceiver {
public static final String ACTION_REPLY = "chat.vojo.app.REPLY";
public static final String EXTRA_ROOM_ID = "room_id";
public static final String KEY_TEXT_REPLY = "vojo.text_reply";
private static final int CONNECT_TIMEOUT_MS = 8_000;
private static final int READ_TIMEOUT_MS = 8_000;
private static final String TAG = "ReplyRcvr";
private static final ExecutorService EXECUTOR = Executors.newSingleThreadExecutor();
@Override
public void onReceive(Context context, Intent intent) {
if (intent == null) return;
final String roomId = intent.getStringExtra(EXTRA_ROOM_ID);
if (roomId == null || roomId.isEmpty()) {
Log.w(TAG, "onReceive: missing room_id, abort");
return;
}
Bundle remote = RemoteInput.getResultsFromIntent(intent);
if (remote == null) {
Log.w(TAG, "onReceive: no RemoteInput results");
return;
}
CharSequence reply = remote.getCharSequence(KEY_TEXT_REPLY);
if (reply == null) {
Log.w(TAG, "onReceive: RemoteInput missing text");
return;
}
final String text = reply.toString().trim();
if (text.isEmpty()) return;
final Context appContext = context.getApplicationContext();
// Pre-flight validation BEFORE the optimistic echo. Posting a self
// bubble first and then immediately stacking an error notif on top
// is jarring UX; for predictable failures (logged out, freshly
// encrypted room) we'd rather skip the echo and only surface the
// error.
final SharedPreferences prefs = appContext.getSharedPreferences(
VojoPollWorker.PREFS, Context.MODE_PRIVATE);
final String token = prefs.getString(VojoPollWorker.KEY_ACCESS_TOKEN, null);
final String homeserver = prefs.getString(VojoPollWorker.KEY_HOMESERVER_URL, null);
if (token == null || token.isEmpty() || homeserver == null || homeserver.isEmpty()) {
Log.w(TAG, "onReceive: no credentials in prefs, surfacing error notif");
postReplyError(appContext, roomId);
return;
}
// Race guard for E2EE flip: the per-room metadata snapshot is
// refreshed by JS on m.room.encryption Timeline events, but a push
// delivered in the narrow window between the encryption state
// landing and the dump completing could still expose the reply
// action on a freshly-encrypted room. Re-read the snapshot
// synchronously here Synapse does NOT enforce "no cleartext in
// encrypted rooms" at the spec level, so without this guard we'd
// leak the user's reply into an E2EE timeline as plaintext.
if (isRoomEncryptedAtSendTime(prefs, roomId)) {
Log.w(TAG, "onReceive: room flipped to encrypted between render and send, abort");
postReplyError(appContext, roomId);
return;
}
// Optimistic local echo appends a self-Person message to the
// conversation and re-posts, so the user sees their reply in the
// shade before the HTTP completes. Only happens after pre-flight
// checks pass so the user doesn't see an echo for a reply we know
// will fail.
long now = System.currentTimeMillis();
VojoFirebaseMessagingService.appendOutgoingMessage(appContext, roomId, text, now);
final PendingResult pendingResult = goAsync();
final String txnId = "vojo-reply-" + UUID.randomUUID();
EXECUTOR.execute(() -> {
try {
int status = sendReply(homeserver, token, roomId, txnId, text);
if (status >= 200 && status < 300) {
if (BuildConfig.DEBUG) {
Log.d(TAG, "reply ok status=" + status + " room=" + roomId);
}
} else {
Log.w(TAG, "reply non-2xx status=" + status + " room=" + roomId);
postReplyError(appContext, roomId);
}
} catch (Throwable t) {
Log.w(TAG, "reply threw room=" + roomId, t);
postReplyError(appContext, roomId);
} finally {
pendingResult.finish();
}
});
}
private int sendReply(
String baseUrl,
String accessToken,
String roomId,
String txnId,
String text
) throws IOException {
String url = trimTrailingSlash(baseUrl)
+ "/_matrix/client/v3/rooms/"
+ URLEncoder.encode(roomId, "UTF-8")
+ "/send/m.room.message/"
+ URLEncoder.encode(txnId, "UTF-8");
JSONObject body;
try {
body = new JSONObject();
body.put("msgtype", "m.text");
body.put("body", text);
} catch (org.json.JSONException je) {
// JSONObject.put only throws on NaN/Inf doubles, neither of
// which we use but keep the type contract honest.
throw new IOException("payload encode failed", je);
}
byte[] payload = body.toString().getBytes("UTF-8");
HttpURLConnection conn = (HttpURLConnection) new URL(url).openConnection();
try {
conn.setRequestMethod("PUT");
conn.setRequestProperty("Authorization", "Bearer " + accessToken);
conn.setRequestProperty("Content-Type", "application/json");
conn.setConnectTimeout(CONNECT_TIMEOUT_MS);
conn.setReadTimeout(READ_TIMEOUT_MS);
conn.setDoOutput(true);
conn.setFixedLengthStreamingMode(payload.length);
try (OutputStream os = conn.getOutputStream()) {
os.write(payload);
}
return conn.getResponseCode();
} finally {
conn.disconnect();
}
}
/**
* Surface a short error notification when the reply HTTP fails so the
* user knows the message did NOT land server-side and can retry from
* within the app. Posted on the DM channel as a one-shot. Unique notif
* id per room so it can't clobber the room's conversation slot.
*/
private static void postReplyError(Context ctx, String roomId) {
NotificationManager nm = (NotificationManager)
ctx.getSystemService(Context.NOTIFICATION_SERVICE);
if (nm == null) return;
try {
String channel = VojoFirebaseMessagingService.CHANNEL_ID_DM;
NotificationCompat.Builder b = new NotificationCompat.Builder(ctx, channel)
.setSmallIcon(R.mipmap.ic_launcher)
.setContentTitle(PushStrings.replyFailed(ctx))
.setContentText(PushStrings.replyFailed(ctx))
.setAutoCancel(true)
.setPriority(NotificationCompat.PRIORITY_DEFAULT);
int errId = ("replyErr_" + roomId).hashCode();
nm.notify(errId, b.build());
} catch (Throwable t) {
Log.w(TAG, "reply error notif failed", t);
}
}
private static String trimTrailingSlash(String s) {
return (s != null && s.endsWith("/")) ? s.substring(0, s.length() - 1) : s;
}
/**
* Synchronous re-check of the room's encryption flag at send time.
* Mirrors VojoFirebaseMessagingService.loadRoomMetadata's tolerant
* parse: legacy string-shape entries and missing flags both default
* to encrypted=true (privacy-first refusing a reply on a falsely-
* flagged room is harmless; sending cleartext into a truly encrypted
* room is a privacy leak).
*/
private static boolean isRoomEncryptedAtSendTime(SharedPreferences prefs, String roomId) {
String raw = prefs.getString(VojoPollWorker.KEY_ROOM_NAMES, null);
if (raw == null || raw.isEmpty()) return true;
try {
JSONObject map = new JSONObject(raw);
if (!map.has(roomId) || map.isNull(roomId)) return true;
JSONObject obj = map.optJSONObject(roomId);
if (obj == null) {
// Legacy string-shape predates the encryption flag
// assume encrypted to err on the side of privacy.
return true;
}
return obj.optBoolean("isEncrypted", true);
} catch (JSONException je) {
return true;
}
}
}

View file

@ -0,0 +1,176 @@
package chat.vojo.app;
import androidx.core.app.NotificationCompat;
import androidx.core.app.Person;
import java.util.ArrayDeque;
import java.util.ArrayList;
import java.util.Deque;
import java.util.List;
import java.util.concurrent.ConcurrentHashMap;
/**
* Per-room MessagingStyle history cache.
*
* Stores the last N messages observed for each room so renderMessageNotification
* can rebuild a NotificationCompat.MessagingStyle with conversation context on
* every new event instead of posting a fresh single-message notification per
* event. Without this every 5-message DM produced 5 distinct entries in the
* shade; with it the user sees one expandable conversation per room the
* WhatsApp/Telegram convention.
*
* Thread-safety: ConcurrentHashMap + per-key synchronized mutation via the
* compute() / get() pattern. Both VojoFirebaseMessagingService.onMessageReceived
* (Firebase-managed thread) and VojoPollWorker.doWork (WorkManager executor)
* mutate the cache; without serialization a same-room FCM + polling race could
* lose a message. Mutations are short only deque append + bounded trim.
*
* Persistence: in-memory only. After process kill the cache is empty, and
* renderMessageNotification falls back to extractMessagingStyleFromNotification
* to recover history from the live system shade. If the user dismissed the
* notification too, the conversation legitimately starts fresh no signal we
* could recover from there anyway.
*
* Eviction: bounded at MAX_MESSAGES_PER_ROOM per room, with FIFO eviction
* (oldest message at the head of the deque is dropped via pollFirst when the
* append would exceed the cap). Map itself is unbounded; in practice the
* dump from dismissRoom (when a server-side read receipt clears unread) keeps
* the room count proportional to active conversations. For safety against
* runaway growth from rooms the user never reads, we cap the map at MAX_ROOMS.
*/
final class RoomMessageCache {
// Element-android keeps a similar in-memory queue (NotificationEventQueue);
// 20 messages per room is generous enough for an active group chat while
// staying well under Android's MessagingStyle render budget Android only
// shows the last ~7 messages in the shade anyway.
private static final int MAX_MESSAGES_PER_ROOM = 20;
// Hard cap on the map size so a long-running session that touches many
// rooms without ever clearing receipts can't slowly leak memory.
// Eviction is approximate (oldest-touched first via insertion order from
// ConcurrentHashMap is NOT guaranteed, so we just clear the oldest by
// arbitrary entry on overflow acceptable for an LRU at this scale).
private static final int MAX_ROOMS = 200;
private static final ConcurrentHashMap<String, Deque<Entry>> store =
new ConcurrentHashMap<>();
private RoomMessageCache() {}
/**
* Snapshot of a single rendered message. We can't store
* NotificationCompat.MessagingStyle.Message directly because Person's
* Icon field is not safely shareable across threads / not cheap to
* rebuild on every poll. Building the Message at render time from this
* record matches element-android's RoomGroupMessageCreator pattern.
*/
static final class Entry {
// Matrix event_id when known (incoming pushes always carry one;
// outgoing optimistic-echo entries pass null). Used by append() to
// suppress duplicate appends when FCM retries / cross-source
// delivery hands the same event in twice without this the
// MessagingStyle conversation would render the same message N
// times in the shade.
final String eventId;
final String body;
final long timestamp;
final String senderKey;
final String senderName;
final boolean fromSelf;
Entry(
String eventId,
String body,
long timestamp,
String senderKey,
String senderName,
boolean fromSelf
) {
this.eventId = eventId;
this.body = body;
this.timestamp = timestamp;
this.senderKey = senderKey;
this.senderName = senderName;
this.fromSelf = fromSelf;
}
}
/**
* Append a message to the room's history and return an ordered snapshot
* including the newly-added entry. Snapshot is taken INSIDE the atomic
* compute() so a concurrent append for the same room can't mutate the
* deque between our addLast and our copy. Returning the deque reference
* and copying outside is unsafe ConcurrentHashMap.compute serialises
* only the lambda body per key, not subsequent reads of the value.
*/
static List<Entry> append(String roomId, Entry entry) {
if (roomId == null || roomId.isEmpty() || entry == null) {
return java.util.Collections.emptyList();
}
final List<Entry> snapshot = new ArrayList<>();
store.compute(roomId, (key, existing) -> {
Deque<Entry> d = (existing != null) ? existing : new ArrayDeque<>();
// Dedup by eventId protects against FCM retry / cross-source
// (FCM + polling Worker) double-delivery that would otherwise
// append the same event twice. Only applies when both the new
// entry and a prior one carry a non-empty eventId; outgoing
// self-echo entries have null eventId by design and never
// collide.
boolean isDup = false;
if (entry.eventId != null && !entry.eventId.isEmpty()) {
for (Entry prior : d) {
if (entry.eventId.equals(prior.eventId)) {
isDup = true;
break;
}
}
}
if (!isDup) {
d.addLast(entry);
while (d.size() > MAX_MESSAGES_PER_ROOM) {
d.pollFirst();
}
}
snapshot.addAll(d);
return d;
});
// Bound the map. Iteration order of ConcurrentHashMap is unspecified
// and the size() check is racy with concurrent puts; we accept ±1
// eviction precision at the 200-room cap as an acceptable approximation
// of LRU (the alternative is a global lock on every append which is
// far more expensive than letting the cache drift by one).
if (store.size() > MAX_ROOMS) {
java.util.Iterator<String> it = store.keySet().iterator();
while (it.hasNext() && store.size() > MAX_ROOMS) {
String key = it.next();
if (!key.equals(roomId)) it.remove();
}
}
return snapshot;
}
/**
* Seed the room's history from an already-posted MessagingStyle (recovered
* via NotificationCompat.MessagingStyle.extractMessagingStyleFromNotification
* after process kill). Idempotent: if the room already has cached entries
* we leave them alone they are by construction at least as recent.
*/
static void seedIfAbsent(String roomId, List<Entry> entries) {
if (roomId == null || roomId.isEmpty() || entries == null || entries.isEmpty()) return;
store.computeIfAbsent(roomId, key -> {
Deque<Entry> d = new ArrayDeque<>();
for (Entry e : entries) {
d.addLast(e);
while (d.size() > MAX_MESSAGES_PER_ROOM) d.pollFirst();
}
return d;
});
}
/** Drop all cached messages for a room (e.g. on receipt-driven dismiss). */
static void clear(String roomId) {
if (roomId == null || roomId.isEmpty()) return;
store.remove(roomId);
}
}

View file

@ -0,0 +1,273 @@
package chat.vojo.app;
import android.content.ContentResolver;
import android.content.Intent;
import android.database.Cursor;
import android.net.Uri;
import android.os.Build;
import android.provider.OpenableColumns;
import android.util.Log;
import android.webkit.MimeTypeMap;
import com.getcapacitor.JSArray;
import com.getcapacitor.JSObject;
import com.getcapacitor.Plugin;
import com.getcapacitor.PluginCall;
import com.getcapacitor.PluginMethod;
import com.getcapacitor.annotation.CapacitorPlugin;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.ArrayList;
import java.util.List;
import java.util.UUID;
/**
* Receives ACTION_SEND / ACTION_SEND_MULTIPLE intents from the system share-
* sheet and surfaces them to the WebView as a pending share that JS consumes
* via {@code pickPendingShare()} (or reacts to via the {@code shareReceived}
* event when the app was already in the foreground).
*
* Cold-start flow:
* 1. Share-sheet Vojo MainActivity.onCreate super.onCreate runs
* BridgeActivity.load(), which itself calls bridge.onNewIntent(getIntent())
* and fans the intent out to every plugin's handleOnNewIntent. So
* cold-start and warm-start share the SAME entry point we don't
* double-process via handleOnStart.
* 2. captureFromIntent copies payload bytes into the app cache and stashes
* the result in {@link #pendingShare}.
* 3. JS booting up (Matrix client ready, user logged in) calls
* pickPendingShare(); receives the JSON; opens the room-picker UI. The
* shareReceived event fired here is dropped silently because no JS
* listener is attached yet that's fine, pickPendingShare drains the
* slot regardless.
*
* Warm flow (app already running):
* 1. Share-sheet MainActivity.onNewIntent BridgeActivity forwards to
* plugin.handleOnNewIntent(intent).
* 2. We re-capture the payload AND emit {@code shareReceived} so JS can
* open the picker without polling.
*
* Why we copy to cache instead of handing JS a content:// URI:
* - WebView fetch() rejects content:// schemes outright, and
* `Capacitor.convertFileSrc()` only works on file paths.
* - The originating app holds the read-grant only for the lifetime of the
* launching task; routing the URI through JS+picker+RoomInput would race
* that grant on Android 14+.
* - Copying into our own cache means the share is self-contained: even if
* the user backgrounds Vojo for hours before picking a chat, the bytes
* are still there. We schedule no cleanup of our own Android's cache
* eviction handles long-tail garbage.
*/
@CapacitorPlugin(name = "ShareTarget")
public class ShareTargetPlugin extends Plugin {
private static final String TAG = "ShareTargetPlugin";
private static final String SHARE_CACHE_SUBDIR = "shared";
// Single-slot pending share. Multiple share-sheet invocations before JS
// drains the slot collapse the latest wins. JS contract is "consume
// once, then it's gone" via pickPendingShare(consume=true). This matches
// user intent: tapping share twice on different photos clearly means
// "share THIS one now".
private volatile JSObject pendingShare = null;
@Override
public void handleOnNewIntent(Intent intent) {
super.handleOnNewIntent(intent);
captureFromIntent(intent, /* notifyJs */ true);
}
@PluginMethod
public void pickPendingShare(PluginCall call) {
JSObject ret = new JSObject();
JSObject snapshot = pendingShare;
if (snapshot == null) {
ret.put("empty", true);
} else {
// Default: consume on read. Lets us treat the slot like a one-shot
// mailbox without an extra round-trip. Caller can pass consume=false
// to peek (not used today, but cheap to keep).
Boolean consume = call.getBoolean("consume", Boolean.TRUE);
ret = snapshot;
if (Boolean.TRUE.equals(consume)) {
pendingShare = null;
}
}
call.resolve(ret);
}
private void captureFromIntent(Intent intent, boolean notifyJs) {
if (intent == null) return;
String action = intent.getAction();
if (action == null) return;
// Capacitor's JSObject.put() silently swallows JSONException internally
// (it wraps org.json.JSONObject and returns `this` on failure) so no
// checked exception is thrown here unlike the raw org.json API.
JSObject share = new JSObject();
share.put("empty", false);
String text = intent.getStringExtra(Intent.EXTRA_TEXT);
String subject = intent.getStringExtra(Intent.EXTRA_SUBJECT);
if (text != null && !text.isEmpty()) share.put("text", text);
if (subject != null && !subject.isEmpty()) share.put("subject", subject);
JSArray items = new JSArray();
List<Uri> uris = new ArrayList<>();
if (Intent.ACTION_SEND.equals(action)) {
Uri uri;
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.TIRAMISU) {
uri = intent.getParcelableExtra(Intent.EXTRA_STREAM, Uri.class);
} else {
// Deprecated overload required to read EXTRA_STREAM on
// API 32, where the typed variant doesn't exist.
//noinspection deprecation
uri = intent.getParcelableExtra(Intent.EXTRA_STREAM);
}
if (uri != null) uris.add(uri);
} else if (Intent.ACTION_SEND_MULTIPLE.equals(action)) {
List<Uri> multi;
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.TIRAMISU) {
multi = intent.getParcelableArrayListExtra(Intent.EXTRA_STREAM, Uri.class);
} else {
//noinspection deprecation
multi = intent.getParcelableArrayListExtra(Intent.EXTRA_STREAM);
}
if (multi != null) uris.addAll(multi);
}
String intentMime = intent.getType();
for (Uri uri : uris) {
JSObject item = copyUriToCache(uri, intentMime);
if (item != null) items.put(item);
}
share.put("items", items);
// Drop pure-noise intents neither text nor a successfully
// copied file. Possible if a sender app handed us only a content://
// URI we can't read (permission revoked) or an EXTRA_STREAM with a
// null Uri. Keeps JS from showing an empty picker.
if (text == null && subject == null && items.length() == 0) {
Log.w(TAG, "Dropping share intent with no usable payload");
return;
}
pendingShare = share;
if (notifyJs) {
notifyListeners("shareReceived", new JSObject());
}
}
/**
* Stream the content of {@code uri} into a fresh file under
* cacheDir/shared/, then return {name, mimeType, size, path}. The path is
* an absolute filesystem path JS wraps it with
* {@code Capacitor.convertFileSrc} before fetch().
*/
private JSObject copyUriToCache(Uri uri, String fallbackMime) {
if (uri == null) return null;
ContentResolver resolver = getContext().getContentResolver();
String name = queryDisplayName(resolver, uri);
String mimeType = resolver.getType(uri);
if (mimeType == null) mimeType = fallbackMime;
if (mimeType == null) mimeType = "application/octet-stream";
if (name == null || name.isEmpty()) {
String ext = MimeTypeMap.getSingleton().getExtensionFromMimeType(mimeType);
name = "share-" + UUID.randomUUID() + (ext != null ? "." + ext : "");
}
File dir = new File(getContext().getCacheDir(), SHARE_CACHE_SUBDIR);
// mkdirs returns false if the directory already exists not an error.
// The real failure mode is the I/O exception below on FileOutputStream
// construction, which we surface.
if (!dir.exists() && !dir.mkdirs()) {
Log.e(TAG, "Could not create share cache dir: " + dir);
return null;
}
// Prefix with UUID so a repeated share of "IMG_1234.jpg" doesn't
// overwrite the previous payload while the user is still picking a
// chat for the older one (e.g. Gallery Vojo, see room-picker open,
// background Gallery re-share same file foreground Vojo). Both
// payloads stay independently addressable.
File out = new File(dir, UUID.randomUUID() + "_" + safeFileName(name));
// Open the input first; if the sender's provider hands us back
// null (revoked grant, gone-away ContentProvider, ) bail before
// creating any on-disk file otherwise the FileOutputStream
// initializer below would create a zero-byte orphan we'd never
// clean up (catch arm doesn't fire when we early-return).
long size;
try (InputStream in = resolver.openInputStream(uri)) {
if (in == null) {
Log.w(TAG, "openInputStream returned null for " + uri);
return null;
}
try (FileOutputStream fos = new FileOutputStream(out)) {
byte[] buf = new byte[64 * 1024];
int n;
long total = 0;
while ((n = in.read(buf)) > 0) {
fos.write(buf, 0, n);
total += n;
}
size = total;
}
} catch (IOException e) {
Log.e(TAG, "Failed to copy " + uri, e);
// Drop the partial file so we don't surface a truncated
// payload to JS as if it were valid.
//noinspection ResultOfMethodCallIgnored
out.delete();
return null;
}
JSObject item = new JSObject();
item.put("name", name);
item.put("mimeType", mimeType);
item.put("size", size);
item.put("path", out.getAbsolutePath());
return item;
}
private String queryDisplayName(ContentResolver resolver, Uri uri) {
// ContentResolver.query throws if the provider rejects the URI scheme
// (e.g. some senders pass a file:// directly no provider involved).
// Wrap in try/catch and fall back to the URI's last path segment.
try (Cursor c = resolver.query(uri, new String[]{ OpenableColumns.DISPLAY_NAME }, null, null, null)) {
if (c != null && c.moveToFirst()) {
int idx = c.getColumnIndex(OpenableColumns.DISPLAY_NAME);
if (idx >= 0) {
String name = c.getString(idx);
if (name != null && !name.isEmpty()) return name;
}
}
} catch (Throwable t) {
Log.d(TAG, "queryDisplayName failed for " + uri + ": " + t.getMessage());
}
String last = uri.getLastPathSegment();
if (last != null && !last.isEmpty()) {
// Strip any directory traversal a malicious sender might encode.
int slash = last.lastIndexOf('/');
return slash >= 0 ? last.substring(slash + 1) : last;
}
return null;
}
private static String safeFileName(String name) {
// Strip path separators and trim length the on-disk name is just an
// identifier; the display name we return to JS preserves the user's
// original filename verbatim. Trim from the tail so the recognisable
// head ("IMG_2025_05_16…") survives and the extension is the part
// that gets clipped on absurdly long names; the on-disk extension
// doesn't matter because nothing inside Vojo dispatches on it (the
// display name carries the real extension into JS).
String stripped = name.replaceAll("[/\\\\]", "_");
if (stripped.length() > 120) stripped = stripped.substring(0, 120);
return stripped;
}
}

File diff suppressed because it is too large Load diff

View file

@ -0,0 +1,675 @@
package chat.vojo.app;
import android.content.Context;
import android.content.SharedPreferences;
import android.util.Log;
import androidx.annotation.NonNull;
import androidx.core.app.NotificationManagerCompat;
import androidx.work.Worker;
import androidx.work.WorkerParameters;
import org.json.JSONArray;
import org.json.JSONObject;
import java.io.IOException;
import java.io.InputStream;
import java.net.HttpURLConnection;
import java.net.URL;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;
/**
* Periodic poll of `/_matrix/client/v3/notifications` as a fallback delivery
* channel for users whose network blocks FCM (mtalk.google.com:5228) the
* ~5% slice on whitelist intranets (corporate / school / government) that
* otherwise receive zero pushes.
*
* Scheduling: enqueued from PollingPlugin.schedule() with a 15-minute period
* (Android's minimum for PeriodicWorkRequest) and CONNECTED network constraint.
* Cancelled via PollingPlugin.cancel() on logout / push disable.
*
* Credentials: read from SharedPreferences (saved by the JS side through
* PollingPlugin.saveSession). Vanilla Synapse (no MAS/OIDC) issues
* non-expiring access tokens; we do not implement refresh-token flow here.
* If a 401 ever occurs, doWork returns Result.success() the next foreground
* launch re-saves the credentials and polling resumes. Retrying with a stale
* token would just waste battery and amplify rate limits.
*
* Output: messages and invites route through VojoFirebaseMessagingService
* .renderMessageNotification (shared with FCM, same notif-id slots
* Android dedupes by replace). RTC ring events route through
* .renderMissedCallNotification (always stale by the time we poll 15-min
* cadence vs 30-second ring lifetime), so the user sees "Missed call" instead
* of a phantom incoming-call CallStyle for a long-dead ring.
*
* E2EE caveat: Synapse cannot decrypt event content, so for end-to-end
* encrypted rooms the response carries `content.algorithm`+`ciphertext`
* with no `body`. The renderer falls through to PushStrings.messageFallback
* (i18n "New message") with the room name as title same UX as the web
* Service Worker on encrypted pushes. By design no key access from the
* Worker.
*
* Dedup is two complementary mechanisms:
* 1) A per-poll high-watermark on the latest event ts we've notified.
* Stored as KEY_LAST_SEEN_TS; advances only after a successful render
* (or a foreground-skipped event the user already saw in-app). Worker
* stops walking within a run as soon as it hits ts strictly less than
* watermark newest-first ordering guarantees the rest are also
* older. Same-ts events fall through to the secondary filters because
* multiple events can share a millisecond.
* 2) NotificationDedup a shared cross-source bounded LRU written by
* every renderer (FCM service after successful nm.notify, this Worker
* after successful render, and the ring-upsert paths at seed time).
* Lets the Worker skip events FCM already delivered even after the
* user dismissed the FCM notification.
*
* Each fire starts from the HEAD of /notifications (no persistent
* pagination cursor the spec's `next_token` walks BACKWARDS into
* history, so a persisted cursor silently drifts off the new events the
* next poll should see; see matrix-js-sdk client.ts:5040 for the
* reference traversal pattern). When a single fire's backlog exceeds
* MAX_PAGES_PER_RUN pages the leftover next_token is saved as
* KEY_DRAIN_CURSOR (with the head ts snapshotted in KEY_DRAIN_TARGET_TS)
* and resumed on the next run, so big backlogs (>250 events) drain over
* consecutive polls without being clipped.
*/
public class VojoPollWorker extends Worker {
private static final String TAG = "VojoPoll";
static final String PREFS = "vojo_poll_state";
static final String KEY_ACCESS_TOKEN = "access_token";
static final String KEY_HOMESERVER_URL = "homeserver_url";
static final String KEY_USER_ID = "user_id";
// High-watermark on the latest event ts we've already notified about.
// Stored as a long-millis string. Replaces an earlier `last_from` cursor
// experiment that misunderstood /notifications pagination direction.
static final String KEY_LAST_SEEN_TS = "last_seen_ts";
// Continuation cursor used when a single run hits MAX_PAGES_PER_RUN before
// reaching the watermark. Persists the next_token across runs so a >250
// event backlog drains over consecutive polls instead of being clipped
// forever by the page cap. Cleared once we either reach the watermark or
// exhaust pagination on a single run.
static final String KEY_DRAIN_CURSOR = "drain_cursor";
// The "head ts" we recorded when entering drain mode. After drain
// completes the watermark is jumped to THIS value rather than the
// (older) max ts seen during drain otherwise the bounded LRU could
// evict events from the original head and let the next normal run
// re-render them. Set once on entering drain mode, untouched while
// draining, cleared when drain completes.
static final String KEY_DRAIN_TARGET_TS = "drain_target_ts";
static final String KEY_NOTIFIED_IDS = "notified_ids";
static final String KEY_ROOM_NAMES = "room_names";
// user_id MXC avatar URL, JSON-encoded, bridged from JS via
// PollingPlugin.saveUserAvatars. Consumed by
// VojoFirebaseMessagingService.lookupUserAvatarMxc for per-sender
// Person.setIcon in MessagingStyle conversations. Bounded at 500
// entries on the JS side; read tolerantly here.
static final String KEY_USER_AVATARS = "user_avatars";
private static final int HTTP_TIMEOUT_MS = 30_000;
// Cap pages-per-fire so an unexpectedly large backlog (server-side bug,
// first run after a long offline window) cannot loop until Android's
// 10-minute Worker kill timer fires. 5 pages × 50 events = up to 250
// events per cycle well above realistic 15-minute backlog for a single
// user. We also break as soon as we hit ts watermark, so most polls
// touch only a single page.
private static final int MAX_PAGES_PER_RUN = 5;
private static final int PAGE_LIMIT = 50;
private static final String RTC_NOTIFICATION_TYPE = "org.matrix.msc4075.rtc.notification";
private static final String RTC_NOTIFICATION_TYPE_STABLE = "m.rtc.notification";
public VojoPollWorker(@NonNull Context context, @NonNull WorkerParameters params) {
super(context, params);
}
@NonNull
@Override
public Result doWork() {
Context ctx = getApplicationContext();
SharedPreferences prefs = ctx.getSharedPreferences(PREFS, Context.MODE_PRIVATE);
String token = prefs.getString(KEY_ACCESS_TOKEN, null);
String homeserver = prefs.getString(KEY_HOMESERVER_URL, null);
if (token == null || homeserver == null) {
// Not logged in (or JS hasn't bridged credentials yet). Return
// success so WorkManager keeps the periodic schedule alive
// we'll pick up the credentials on the next fire.
Log.i(TAG, "poll: no credentials, bail");
return Result.success();
}
// If POST_NOTIFICATIONS was revoked we'd fetch + parse + try to
// render and then watch every nm.notify fail with SecurityException
// which leaves the LRU/watermark unadvanced (correctly so for a
// transient failure) and re-runs the same loop every 15 minutes
// forever. Bail early to avoid burning battery on a permanent
// user choice. The next visibility re-bridge inside the JS app
// will pick up a re-granted permission.
if (!NotificationManagerCompat.from(ctx).areNotificationsEnabled()) {
Log.i(TAG, "poll: notifications disabled, bail");
return Result.success();
}
long watermark = prefs.getLong(KEY_LAST_SEEN_TS, 0L);
String drainCursor = prefs.getString(KEY_DRAIN_CURSOR, null);
long drainTargetTs = prefs.getLong(KEY_DRAIN_TARGET_TS, 0L);
boolean wasDraining = drainCursor != null;
Map<String, String> roomNames = loadRoomNamesMap(prefs);
// Mirror the FCM service's foreground gate: if the user is actively in
// the app, the live timeline owns the UX and a system notification for
// a backlog event would be both stale and visually noisy. We still
// consume state (LRU, watermark) so the same event doesn't surface
// when the user later backgrounds the app.
boolean inForeground = MainActivity.isInForeground;
Log.i(TAG, "poll: start fg=" + inForeground
+ " watermark=" + watermark
+ " draining=" + wasDraining);
int pagesFetched = 0;
int renderedCount = 0;
int skippedDedupCount = 0;
long highestTsSeen = watermark;
boolean reachedWatermark = false;
// The continuation cursor we'd save if this run is capped. Starts as
// the resumed drain cursor; advances with each successful page fetch
// so a transient mid-pagination error still preserves drain progress.
String pendingCursor = drainCursor;
boolean paginationExhausted = false;
try {
// Cursor strategy: drain cursor resumes from where a previous capped
// run stopped; otherwise we start from the HEAD. next_token from
// /notifications paginates BACKWARDS into history, so a stored
// cursor must be used as a drain-only continuation, NOT as an
// ongoing "since" mark (the latter would silently drift off new
// events). Within a single fire we stop as soon as ts < watermark
// (newest-first ordering means everything past that is covered).
String nextFrom = drainCursor;
for (int page = 0; page < MAX_PAGES_PER_RUN && !reachedWatermark; page += 1) {
// Cooperative cancellation. WorkManager.cancelUniqueWork (called
// from PollingPlugin.cancel during logout / push disable) only
// marks future scheduling it does NOT interrupt this thread.
// Without these checks the Worker keeps fetching pages, posting
// notifications, and (worst of all) running the final
// editor.apply() with stale state written AFTER clearSession
// wiped prefs leaking watermark / drain cursor from the
// logged-out account into the next login.
if (isStopped()) return Result.success();
JSONObject body = fetchNotifications(homeserver, token, nextFrom);
// fetchNotifications throws on every failure path; a null
// return is unreachable in current code. The early-break here
// is a defensive belt-and-suspenders keep paginationExhausted
// consistent so the drain-bookkeeping below clears the cursor
// instead of replaying the same empty page forever.
if (body == null) {
paginationExhausted = true;
pendingCursor = null;
break;
}
JSONArray notifications = body.optJSONArray("notifications");
if (notifications == null || notifications.length() == 0) {
// Server returned no entries for this page. Treat as
// end-of-pagination so a drain in progress can complete
// (otherwise pendingCursor would keep its old value and
// we'd re-fetch the same empty page next cycle forever).
paginationExhausted = true;
pendingCursor = null;
break;
}
for (int i = 0; i < notifications.length(); i += 1) {
if (isStopped()) return Result.success();
JSONObject entry = notifications.optJSONObject(i);
if (entry == null) continue;
String eventId = extractEventId(entry);
if (eventId == null) continue;
// ts gate: server returns newest-first, so once we hit
// ts STRICTLY less than the watermark we know the rest of
// the page (and every subsequent page) is already covered.
// Same-ts events fall through to the LRU/read filters
// below multiple events can share a millisecond, and
// collapsing them at the ts boundary would silently drop
// a fresh sibling of a previously-rendered one.
long ts = entry.optLong("ts", 0L);
if (ts > 0 && ts < watermark) {
reachedWatermark = true;
break;
}
// Skip notifications the user already read on another
// client (web tab, Element, second device). Spec marks
// `read` as a required boolean on each entry.
if (entry.optBoolean("read", false)) {
if (ts > highestTsSeen) highestTsSeen = ts;
continue;
}
// Skip events the push rules said don't notify (muted
// rooms, dont_notify overrides). Without this gate
// polling would re-surface events Sygnal already
// suppressed for the FCM path the mute toggle
// wouldn't actually mute on whitelist networks.
if (!notifyAllowed(entry)) {
if (ts > highestTsSeen) highestTsSeen = ts;
continue;
}
// Cross-source dedup via NotificationDedup: FCM writes
// into this set after every successful render, so the
// Worker correctly skips events the FCM service already
// delivered even if the user dismissed the FCM
// notification before this cycle fired.
if (NotificationDedup.wasNotified(ctx, eventId)) {
skippedDedupCount += 1;
if (ts > highestTsSeen) highestTsSeen = ts;
continue;
}
// Three outcomes for marking + watermark advance:
// foreground mark + advance (skip render
// but consume state, otherwise
// next bg poll would replay)
// background + posted mark + advance
// background + !posted DON'T mark, DON'T advance
// (transient render failure
// should be retried next poll)
boolean posted = false;
boolean treatAsNotRenderable = false;
if (!inForeground) {
Map<String, String> flattened = flattenNotification(entry, roomNames);
String type = flattened.get("type");
boolean isRtcType = RTC_NOTIFICATION_TYPE.equals(type)
|| RTC_NOTIFICATION_TYPE_STABLE.equals(type);
boolean isRing = "ring".equals(flattened.get("content_notification_type"));
if (isRtcType && isRing) {
// Composite session dedup: if FCM already alerted
// for this call session (different ring event,
// same parent), skip posting a duplicate
// missed-call. Without this, a session with one
// FCM live-alert ring + one re-ring through
// polling would surface as both a CallStyle and
// a missed-call card. Helpers live in
// VojoFirebaseMessagingService so the key shape
// stays in lock-step across FCM and polling.
String roomIdField = flattened.get("room_id");
String sessionId = VojoFirebaseMessagingService
.extractCallSessionId(flattened);
String composite = null;
if (roomIdField != null && sessionId != null) {
composite = VojoFirebaseMessagingService
.compositeCallDedupKey(roomIdField, sessionId);
if (NotificationDedup.wasNotified(ctx, composite)) {
if (ts > highestTsSeen) highestTsSeen = ts;
treatAsNotRenderable = true;
}
}
if (!treatAsNotRenderable) {
// Stale ring (call lifetime is 30 seconds; we
// poll every 15 minutes). Show "Missed call"
// so the user knows somebody tried, without
// phantom-ringing a long-dead call via
// CallStyle.
posted = VojoFirebaseMessagingService
.renderMissedCallNotification(ctx, flattened);
if (posted && composite != null) {
// Mark the composite so the next polling
// cycle observing a re-ring for the same
// session doesn't double-post.
NotificationDedup.markNotified(ctx, composite);
}
}
} else if (isRtcType) {
// Non-ring RTC sub-type. MSC4075 defines at least
// "ring" and "notification" the latter is the
// chat-style alert variant which doesn't make
// sense to surface as a stale "missed" entry from
// a 15-minute poll. Falling through to
// renderMessageNotification would post a generic
// "New message" with no body (no content.body on
// RTC events). Skip rendering but still mark seen
// so we don't re-walk it next poll.
treatAsNotRenderable = true;
} else {
posted = VojoFirebaseMessagingService
.renderMessageNotification(ctx, flattened, null);
}
}
// Mark + advance ts whenever we've consumed the event
// (foreground-skipped, non-ring-RTC skipped, or
// successfully rendered). Render-failure (bg branch where
// posted==false) is intentionally excluded so the next
// poll retries it.
if (inForeground || posted || treatAsNotRenderable) {
NotificationDedup.markNotified(ctx, eventId);
if (ts > highestTsSeen) highestTsSeen = ts;
if (posted) renderedCount += 1;
}
}
pagesFetched += 1;
// optString returns the fallback only when the key is absent;
// a literal JSON `null` becomes the string "null" guard
// against the rare server quirk so we don't loop on it.
String rawNext = body.optString("next_token", null);
if (rawNext == null || rawNext.isEmpty() || "null".equals(rawNext)) {
nextFrom = null;
} else {
nextFrom = rawNext;
}
pendingCursor = nextFrom;
if (nextFrom == null) {
paginationExhausted = true;
break;
}
}
} catch (UnauthorizedException e) {
Log.w(TAG, "poll: 401 — clearing credentials, awaiting next foreground re-bridge");
prefs.edit()
.remove(KEY_ACCESS_TOKEN)
.apply();
return Result.success();
} catch (ForbiddenException e) {
// 403 from Synapse is usually rate-limit or a transient server
// policy reject, not a dead token. Don't clear credentials
// just let the next periodic fire retry. Avoid Result.retry()
// because we don't want an immediate accelerated retry that
// amplifies the rate-limit cause.
Log.w(TAG, "poll: 403/429 — skipping this cycle, will retry on next scheduled fire");
return Result.success();
} catch (Throwable t) {
Log.w(TAG, "poll: failed at page " + pagesFetched, t);
return Result.retry();
}
// Final stopped-check before persisting state. If cancellation landed
// between the last in-loop check and here, do NOT apply: the
// accumulated editor writes would otherwise overwrite KEY_LAST_SEEN_TS
// and KEY_DRAIN_CURSOR AFTER JS clearSession wiped them, leaking
// stale state from the just-logged-out account into the next login.
if (isStopped()) return Result.success();
SharedPreferences.Editor editor = prefs.edit();
// Drain-mode bookkeeping. Three transitions:
// - normal normal (cap not hit): advance watermark to highestTsSeen.
// - normal drain (cap hit, no prior drain): save continuation
// cursor AND snapshot drainTargetTs = highestTsSeen. The current
// run's highest ts becomes the "fast-forward" target for when
// drain eventually completes without this, the bounded LRU
// could evict the original head events and let the post-drain
// normal run re-render them.
// - drain drain (still capped): keep cursor + target unchanged.
// Don't overwrite drainTargetTs with this run's highestTsSeen,
// because drain pages are always OLDER than the original head.
// - drain normal (drain complete): clear cursor + target. Advance
// watermark to drainTargetTs drain pages always walk backwards
// (older than the snapshotted head), so highestTsSeen accumulated
// during drain is by construction drainTargetTs.
boolean cappedWithMore = !reachedWatermark && !paginationExhausted && pendingCursor != null;
long newWatermark = watermark;
String drainState;
if (cappedWithMore) {
editor.putString(KEY_DRAIN_CURSOR, pendingCursor);
if (!wasDraining) {
// First run entering drain mode snapshot the head ts.
editor.putLong(KEY_DRAIN_TARGET_TS, highestTsSeen);
drainState = "drain-entered";
} else {
drainState = "drain-continued";
}
} else {
editor.remove(KEY_DRAIN_CURSOR);
editor.remove(KEY_DRAIN_TARGET_TS);
long advanceTo = wasDraining ? drainTargetTs : highestTsSeen;
if (advanceTo > watermark) {
editor.putLong(KEY_LAST_SEEN_TS, advanceTo);
newWatermark = advanceTo;
}
drainState = wasDraining ? "drain-exited" : "normal";
}
editor.apply();
Log.i(TAG, "poll: done pages=" + pagesFetched
+ " rendered=" + renderedCount
+ " dedupSkipped=" + skippedDedupCount
+ " watermark=" + newWatermark
+ " state=" + drainState);
return Result.success();
}
// Returns true iff at least one element of entry.actions is the literal
// string "notify". Per Matrix spec §13.13.1, tweak objects
// (`{set_tweak: ...}`) only MODIFY a notification produced by a separate
// `"notify"` action they do not by themselves imply notify. "dont_notify"
// or an empty actions array means the push rule explicitly suppressed
// this event (most commonly: a muted room).
private static boolean notifyAllowed(JSONObject entry) {
JSONArray actions = entry.optJSONArray("actions");
if (actions == null || actions.length() == 0) return false;
for (int i = 0; i < actions.length(); i += 1) {
Object a = actions.opt(i);
if ((a instanceof String) && "notify".equals(a)) return true;
}
return false;
}
//
// HTTP
//
private static final class UnauthorizedException extends IOException {
UnauthorizedException() {
super("401 Unauthorized");
}
}
// 403 from Synapse is most commonly a rate-limit or a transient policy
// reject (M_LIMIT_EXCEEDED, M_FORBIDDEN). It is NOT "token died" we
// surface it as a distinct exception so doWork can skip this cycle
// without clearing credentials and without an accelerated Result.retry()
// that would amplify the rate-limit cause.
private static final class ForbiddenException extends IOException {
ForbiddenException() {
super("403 Forbidden");
}
}
private JSONObject fetchNotifications(String homeserverUrl, String token, String fromCursor)
throws IOException {
StringBuilder url = new StringBuilder(homeserverUrl);
if (!homeserverUrl.endsWith("/")) url.append('/');
url.append("_matrix/client/v3/notifications?limit=").append(PAGE_LIMIT);
if (fromCursor != null && !fromCursor.isEmpty()) {
url.append("&from=").append(java.net.URLEncoder.encode(fromCursor, "UTF-8"));
}
HttpURLConnection conn = (HttpURLConnection) new URL(url.toString()).openConnection();
try {
conn.setRequestMethod("GET");
conn.setRequestProperty("Authorization", "Bearer " + token);
conn.setRequestProperty("Accept", "application/json");
// Identifiable UA so server logs can attribute polling traffic
// (some WAFs also flag bare "Java/<version>" as suspicious).
conn.setRequestProperty("User-Agent", "Vojo-Android-Poll/" + BuildConfig.VERSION_NAME);
conn.setConnectTimeout(HTTP_TIMEOUT_MS);
conn.setReadTimeout(HTTP_TIMEOUT_MS);
int code = conn.getResponseCode();
if (code == 401) throw new UnauthorizedException();
// Treat 429 (rate limited) and 403 (Synapse policy reject) the
// same: skip this cycle, don't retry-storm. Result.retry()'s 30s
// backoff would amplify the rate-limit cause; the next periodic
// fire in 15 minutes is well past any realistic Retry-After
// window from a Matrix homeserver.
if (code == 403 || code == 429) throw new ForbiddenException();
if (code < 200 || code >= 300) {
throw new IOException("HTTP " + code);
}
try (InputStream in = conn.getInputStream()) {
return new JSONObject(readAll(in));
} catch (org.json.JSONException je) {
throw new IOException("malformed JSON", je);
}
} finally {
conn.disconnect();
}
}
private static String readAll(InputStream in) throws IOException {
// Accumulate raw bytes, then decode the whole buffer as a single UTF-8
// string. Decoding each 8 KB chunk separately would corrupt multi-byte
// sequences that straddle a chunk boundary for a Russian-content
// notification body that crosses ~8 KB, the result is U+FFFD in place
// of a Cyrillic character. Also use != -1 rather than > 0 for the
// read loop: InputStream.read(byte[]) is contractually allowed to
// return 0 without indicating EOF.
java.io.ByteArrayOutputStream out = new java.io.ByteArrayOutputStream();
byte[] buf = new byte[8 * 1024];
int n;
while ((n = in.read(buf)) != -1) {
if (n > 0) out.write(buf, 0, n);
}
return out.toString("UTF-8");
}
//
// Payload shaping
//
// The /notifications response shape is structured (event{type,sender,
// content{}}, room_id, ts, read, actions) different from Sygnal's
// flattened FCM payload. We flatten into the Sygnal-shape Map<String,
// String> so the shared renderer in VojoFirebaseMessagingService can
// stay source-agnostic. Keys we set: event_id, room_id, sender, type,
// content_membership, content_body, content_notification_type,
// content_sender_ts, content_lifetime, room_name (from local cache).
//
// NOTE: sender_display_name is NOT set here /notifications returns the
// raw event without the Sygnal-side profile resolution that gives FCM
// its `sender_display_name`. The renderer's title-fallback chain
// (room_name sender_display_name sender "Vojo") therefore lands
// on `sender` (a raw MXID) when the room name isn't cached. The renderer
// strips the MXID to its local-part as a final cosmetic guard so users
// see "alice" instead of "@alice:hs.tld".
//
private static Map<String, String> flattenNotification(
JSONObject entry, Map<String, String> roomNames
) {
Map<String, String> out = new HashMap<>();
String roomId = entry.optString("room_id", null);
if (roomId != null) out.put("room_id", roomId);
JSONObject event = entry.optJSONObject("event");
if (event != null) {
putIfPresent(out, event, "event_id", "event_id");
putIfPresent(out, event, "sender", "sender");
putIfPresent(out, event, "type", "type");
JSONObject content = event.optJSONObject("content");
if (content != null) {
putIfPresent(out, content, "membership", "content_membership");
putIfPresent(out, content, "body", "content_body");
putIfPresent(out, content, "notification_type", "content_notification_type");
if (content.has("sender_ts")) {
out.put("content_sender_ts", String.valueOf(content.optLong("sender_ts")));
}
if (content.has("lifetime")) {
out.put("content_lifetime", String.valueOf(content.optLong("lifetime")));
}
// Parent call event_id for session-level dedup. The shared
// FCM renderer reads this from the flattened key
// `content_m.relates_to_event_id` (mirroring one of Sygnal's
// flatten shapes); writing the literal-dot variant here keeps
// FCM and polling on the same key.
JSONObject relates = content.optJSONObject("m.relates_to");
if (relates != null) {
String parentEventId = relates.optString("event_id", null);
if (parentEventId != null && !parentEventId.isEmpty()) {
out.put("content_m.relates_to_event_id", parentEventId);
}
}
// Legacy MSC2746 call_id fallback. Modern MSC4075 sessions
// surface via m.relates_to above; this branch is a no-op for
// them but keeps the shape symmetric for older deployments.
if (content.has("call_id")) {
String callId = content.optString("call_id", null);
if (callId != null && !callId.isEmpty()) {
out.put("content_call_id", callId);
}
}
}
}
// Room name from the snapshot the JS side pushes through
// PollingPlugin.saveRoomNames, parsed once at the start of doWork().
// Brand-new rooms (not yet observed by JS at last bridge time) miss
// the cache the renderer falls back to sender / "Vojo".
if (roomId != null) {
String roomName = roomNames.get(roomId);
if (roomName != null && !roomName.isEmpty()) out.put("room_name", roomName);
}
return out;
}
// Parse the SharedPreferences-stored room-name JSON snapshot once per
// doWork() so we don't redo the parse for every event in the page (up to
// PAGE_LIMIT × MAX_PAGES_PER_RUN = 250 events).
//
// The snapshot shape evolved: legacy was {roomId: "Display name"}, current
// is {roomId: {name, isDirect, isEncrypted, avatarMxc?}}. We parse both
// tolerantly for the structured shape we extract `name`, for the legacy
// shape we use the string verbatim. A naive optString on the structured
// entry serialises the whole object as JSON ("{name:Alice,...}") and that
// string leaked into the missed-call / message title on the polling
// path visible bug.
private static Map<String, String> loadRoomNamesMap(SharedPreferences prefs) {
Map<String, String> out = new HashMap<>();
String raw = prefs.getString(KEY_ROOM_NAMES, null);
if (raw == null || raw.isEmpty()) return out;
try {
JSONObject map = new JSONObject(raw);
for (Iterator<String> it = map.keys(); it.hasNext(); ) {
String roomId = it.next();
if (map.isNull(roomId)) continue;
JSONObject obj = map.optJSONObject(roomId);
String name = obj != null
? obj.optString("name", null)
: map.optString(roomId, null);
if (name != null && !name.isEmpty()) out.put(roomId, name);
}
} catch (org.json.JSONException je) {
// Corrupt blob return empty map. Renderer falls back to sender.
}
return out;
}
private static void putIfPresent(
Map<String, String> out, JSONObject src, String srcKey, String dstKey
) {
// Guard against a literal JSON null at the key: JSONObject.optString
// returns the *fallback* only when the key is absent, but on a
// present-but-null key it coerces JSONObject.NULL to the four-char
// string "null", which would leak as "null" into a notification body.
if (!src.has(srcKey) || src.isNull(srcKey)) return;
String v = src.optString(srcKey, null);
if (v != null && !v.isEmpty()) out.put(dstKey, v);
}
private static String extractEventId(JSONObject entry) {
JSONObject event = entry.optJSONObject("event");
if (event == null) return null;
if (!event.has("event_id") || event.isNull("event_id")) return null;
String eventId = event.optString("event_id", null);
if (eventId == null || eventId.isEmpty()) return null;
return eventId;
}
}

View file

@ -0,0 +1,34 @@
<vector xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:aapt="http://schemas.android.com/aapt"
android:width="108dp"
android:height="108dp"
android:viewportHeight="108"
android:viewportWidth="108">
<path
android:fillType="evenOdd"
android:pathData="M32,64C32,64 38.39,52.99 44.13,50.95C51.37,48.37 70.14,49.57 70.14,49.57L108.26,87.69L108,109.01L75.97,107.97L32,64Z"
android:strokeColor="#00000000"
android:strokeWidth="1">
<aapt:attr name="android:fillColor">
<gradient
android:endX="78.5885"
android:endY="90.9159"
android:startX="48.7653"
android:startY="61.0927"
android:type="linear">
<item
android:color="#44000000"
android:offset="0.0" />
<item
android:color="#00000000"
android:offset="1.0" />
</gradient>
</aapt:attr>
</path>
<path
android:fillColor="#FFFFFF"
android:fillType="nonZero"
android:pathData="M66.94,46.02L66.94,46.02C72.44,50.07 76,56.61 76,64L32,64C32,56.61 35.56,50.11 40.98,46.06L36.18,41.19C35.45,40.45 35.45,39.3 36.18,38.56C36.91,37.81 38.05,37.81 38.78,38.56L44.25,44.05C47.18,42.57 50.48,41.71 54,41.71C57.48,41.71 60.78,42.57 63.68,44.05L69.11,38.56C69.84,37.81 70.98,37.81 71.71,38.56C72.44,39.3 72.44,40.45 71.71,41.19L66.94,46.02ZM62.94,56.92C64.08,56.92 65,56.01 65,54.88C65,53.76 64.08,52.85 62.94,52.85C61.8,52.85 60.88,53.76 60.88,54.88C60.88,56.01 61.8,56.92 62.94,56.92ZM45.06,56.92C46.2,56.92 47.13,56.01 47.13,54.88C47.13,53.76 46.2,52.85 45.06,52.85C43.92,52.85 43,53.76 43,54.88C43,56.01 43.92,56.92 45.06,56.92Z"
android:strokeColor="#00000000"
android:strokeWidth="1" />
</vector>

View file

@ -0,0 +1,170 @@
<?xml version="1.0" encoding="utf-8"?>
<vector xmlns:android="http://schemas.android.com/apk/res/android"
android:width="108dp"
android:height="108dp"
android:viewportHeight="108"
android:viewportWidth="108">
<path
android:fillColor="#26A69A"
android:pathData="M0,0h108v108h-108z" />
<path
android:fillColor="#00000000"
android:pathData="M9,0L9,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,0L19,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M29,0L29,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M39,0L39,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M49,0L49,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M59,0L59,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M69,0L69,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M79,0L79,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M89,0L89,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M99,0L99,108"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,9L108,9"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,19L108,19"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,29L108,29"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,39L108,39"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,49L108,49"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,59L108,59"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,69L108,69"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,79L108,79"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,89L108,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M0,99L108,99"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,29L89,29"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,39L89,39"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,49L89,49"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,59L89,59"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,69L89,69"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M19,79L89,79"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M29,19L29,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M39,19L39,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M49,19L49,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M59,19L59,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M69,19L69,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
<path
android:fillColor="#00000000"
android:pathData="M79,19L79,89"
android:strokeColor="#33FFFFFF"
android:strokeWidth="0.8" />
</vector>

Binary file not shown.

After

Width:  |  Height:  |  Size: 164 KiB

View file

@ -0,0 +1,12 @@
<?xml version="1.0" encoding="utf-8"?>
<androidx.coordinatorlayout.widget.CoordinatorLayout xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:app="http://schemas.android.com/apk/res-auto"
xmlns:tools="http://schemas.android.com/tools"
android:layout_width="match_parent"
android:layout_height="match_parent"
tools:context=".MainActivity">
<WebView
android:layout_width="match_parent"
android:layout_height="match_parent" />
</androidx.coordinatorlayout.widget.CoordinatorLayout>

View file

@ -0,0 +1,5 @@
<?xml version="1.0" encoding="utf-8"?>
<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
<background android:drawable="@color/ic_launcher_background"/>
<foreground android:drawable="@mipmap/ic_launcher_foreground"/>
</adaptive-icon>

View file

@ -0,0 +1,5 @@
<?xml version="1.0" encoding="utf-8"?>
<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
<background android:drawable="@color/ic_launcher_background"/>
<foreground android:drawable="@mipmap/ic_launcher_foreground"/>
</adaptive-icon>

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 18 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 5.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.7 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 5.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 19 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 26 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 19 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 36 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 45 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 36 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 66 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

View file

@ -0,0 +1,7 @@
<?xml version="1.0" encoding="utf-8"?>
<resources>
<!-- Matches web safe-area / DM 1:1 chat background (DAWN.bg2) so the
native splash, the WebView body, and the in-app AuthSplashScreen all
share a single backdrop and read as one continuous splash. -->
<color name="splash_bg">#0d0e11</color>
</resources>

View file

@ -0,0 +1,4 @@
<?xml version="1.0" encoding="utf-8"?>
<resources>
<color name="ic_launcher_background">#121314</color>
</resources>

View file

@ -0,0 +1,7 @@
<?xml version='1.0' encoding='utf-8'?>
<resources>
<string name="app_name">Vojo</string>
<string name="title_activity_main">Vojo</string>
<string name="package_name">chat.vojo.app</string>
<string name="custom_url_scheme">chat.vojo.app</string>
</resources>

View file

@ -0,0 +1,53 @@
<?xml version="1.0" encoding="utf-8"?>
<resources>
<!-- Base application theme. -->
<style name="AppTheme" parent="Theme.AppCompat.DayNight.NoActionBar">
<item name="colorPrimary">@color/colorPrimary</item>
<item name="colorPrimaryDark">@color/colorPrimaryDark</item>
<item name="colorAccent">@color/colorAccent</item>
<item name="windowActionBar">false</item>
<item name="windowNoTitle">true</item>
</style>
<style name="AppTheme.NoActionBar" parent="Theme.AppCompat.DayNight.NoActionBar">
<item name="windowActionBar">false</item>
<item name="windowNoTitle">true</item>
<item name="android:background">@null</item>
<item name="android:windowLayoutInDisplayCutoutMode">shortEdges</item>
<!-- Bridges the gap between native splash exit and the WebView's first
body paint: without this the window paints transparent/black for
~200ms while the bundle hydrates, producing a visible black flash
between the native and the in-app splash. Matches splash_bg so
cold start reads as one continuous backdrop. -->
<item name="android:windowBackground">@color/splash_bg</item>
</style>
<!-- Launch theme: Android 12+ system splash (Theme.SplashScreen via
androidx.core.splashscreen). Renders the mascot centered on the same
#0d0e11 backdrop the web AuthSplashScreen uses, so cold start reads
as one continuous splash (native → WebView mount → web splash) instead
of three visual jumps. MainActivity installs AndroidX SplashScreen
before super.onCreate() and keeps it visible until Capacitor's local
WebView has loaded the app shell. -->
<style name="AppTheme.NoActionBarLaunch" parent="Theme.SplashScreen">
<!-- Theme.SplashScreen only sets the native android:windowActionBar /
android:windowNoTitle attrs. Capacitor's BridgeActivity extends
AppCompatActivity, whose ActionBar delegate reads the un-prefixed
AppCompat attrs — without these two overrides, AppCompat keeps
its ActionBar enabled, paints the activity label ("Vojo" from
strings.xml/title_activity_main) at the top of the WebView, and
persists past the splash exit. -->
<item name="windowActionBar">false</item>
<item name="windowNoTitle">true</item>
<item name="windowSplashScreenBackground">@color/splash_bg</item>
<item name="windowSplashScreenAnimatedIcon">@drawable/vojo_mascot_splash</item>
<!-- Intentionally NO windowSplashScreenIconBackgroundColor: setting it
switches the system to the "with-background" canvas, which is
actually 240dp (vs 288dp without) — the colored ring would just
shrink the visible icon zone. Background already matches via
windowSplashScreenBackground above. -->
<item name="postSplashScreenTheme">@style/AppTheme.NoActionBar</item>
<item name="android:windowLayoutInDisplayCutoutMode">shortEdges</item>
</style>
</resources>

View file

@ -0,0 +1,5 @@
<?xml version="1.0" encoding="utf-8"?>
<paths xmlns:android="http://schemas.android.com/apk/res/android">
<external-path name="my_images" path="." />
<cache-path name="my_cache_images" path="." />
</paths>

View file

@ -0,0 +1,18 @@
package com.getcapacitor.myapp;
import static org.junit.Assert.*;
import org.junit.Test;
/**
* Example local unit test, which will execute on the development machine (host).
*
* @see <a href="http://d.android.com/tools/testing">Testing documentation</a>
*/
public class ExampleUnitTest {
@Test
public void addition_isCorrect() throws Exception {
assertEquals(4, 2 + 2);
}
}

29
android/build.gradle Normal file
View file

@ -0,0 +1,29 @@
// Top-level build file where you can add configuration options common to all sub-projects/modules.
buildscript {
repositories {
google()
mavenCentral()
}
dependencies {
classpath 'com.android.tools.build:gradle:8.13.0'
classpath 'com.google.gms:google-services:4.4.4'
// NOTE: Do not place your application dependencies here; they belong
// in the individual module build.gradle files
}
}
apply from: "variables.gradle"
allprojects {
repositories {
google()
mavenCentral()
}
}
task clean(type: Delete) {
delete rootProject.buildDir
}

View file

@ -0,0 +1,18 @@
// DO NOT EDIT THIS FILE! IT IS GENERATED EACH TIME "capacitor update" IS RUN
include ':capacitor-android'
project(':capacitor-android').projectDir = new File('../node_modules/@capacitor/android/capacitor')
include ':capacitor-app'
project(':capacitor-app').projectDir = new File('../node_modules/@capacitor/app/android')
include ':capacitor-browser'
project(':capacitor-browser').projectDir = new File('../node_modules/@capacitor/browser/android')
include ':capacitor-preferences'
project(':capacitor-preferences').projectDir = new File('../node_modules/@capacitor/preferences/android')
include ':capacitor-push-notifications'
project(':capacitor-push-notifications').projectDir = new File('../node_modules/@capacitor/push-notifications/android')
include ':capacitor-toast'
project(':capacitor-toast').projectDir = new File('../node_modules/@capacitor/toast/android')

22
android/gradle.properties Normal file
View file

@ -0,0 +1,22 @@
# Project-wide Gradle settings.
# IDE (e.g. Android Studio) users:
# Gradle settings configured through the IDE *will override*
# any settings specified in this file.
# For more details on how to configure your build environment visit
# http://www.gradle.org/docs/current/userguide/build_environment.html
# Specifies the JVM arguments used for the daemon process.
# The setting is particularly useful for tweaking memory settings.
org.gradle.jvmargs=-Xmx1536m
# When configured, Gradle will run in incubating parallel mode.
# This option should only be used with decoupled projects. More details, visit
# http://www.gradle.org/docs/current/userguide/multi_project_builds.html#sec:decoupled_projects
# org.gradle.parallel=true
# AndroidX package structure to make it clearer which packages are bundled with the
# Android operating system, and which are packaged with your app's APK
# https://developer.android.com/topic/libraries/support-library/androidx-rn
android.useAndroidX=true

Binary file not shown.

View file

@ -0,0 +1,7 @@
distributionBase=GRADLE_USER_HOME
distributionPath=wrapper/dists
distributionUrl=https\://services.gradle.org/distributions/gradle-8.14.3-all.zip
networkTimeout=10000
validateDistributionUrl=true
zipStoreBase=GRADLE_USER_HOME
zipStorePath=wrapper/dists

251
android/gradlew vendored Executable file
View file

@ -0,0 +1,251 @@
#!/bin/sh
#
# Copyright © 2015-2021 the original authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
# SPDX-License-Identifier: Apache-2.0
#
##############################################################################
#
# Gradle start up script for POSIX generated by Gradle.
#
# Important for running:
#
# (1) You need a POSIX-compliant shell to run this script. If your /bin/sh is
# noncompliant, but you have some other compliant shell such as ksh or
# bash, then to run this script, type that shell name before the whole
# command line, like:
#
# ksh Gradle
#
# Busybox and similar reduced shells will NOT work, because this script
# requires all of these POSIX shell features:
# * functions;
# * expansions «$var», «${var}», «${var:-default}», «${var+SET}»,
# «${var#prefix}», «${var%suffix}», and «$( cmd )»;
# * compound commands having a testable exit status, especially «case»;
# * various built-in commands including «command», «set», and «ulimit».
#
# Important for patching:
#
# (2) This script targets any POSIX shell, so it avoids extensions provided
# by Bash, Ksh, etc; in particular arrays are avoided.
#
# The "traditional" practice of packing multiple parameters into a
# space-separated string is a well documented source of bugs and security
# problems, so this is (mostly) avoided, by progressively accumulating
# options in "$@", and eventually passing that to Java.
#
# Where the inherited environment variables (DEFAULT_JVM_OPTS, JAVA_OPTS,
# and GRADLE_OPTS) rely on word-splitting, this is performed explicitly;
# see the in-line comments for details.
#
# There are tweaks for specific operating systems such as AIX, CygWin,
# Darwin, MinGW, and NonStop.
#
# (3) This script is generated from the Groovy template
# https://github.com/gradle/gradle/blob/HEAD/platforms/jvm/plugins-application/src/main/resources/org/gradle/api/internal/plugins/unixStartScript.txt
# within the Gradle project.
#
# You can find Gradle at https://github.com/gradle/gradle/.
#
##############################################################################
# Attempt to set APP_HOME
# Resolve links: $0 may be a link
app_path=$0
# Need this for daisy-chained symlinks.
while
APP_HOME=${app_path%"${app_path##*/}"} # leaves a trailing /; empty if no leading path
[ -h "$app_path" ]
do
ls=$( ls -ld "$app_path" )
link=${ls#*' -> '}
case $link in #(
/*) app_path=$link ;; #(
*) app_path=$APP_HOME$link ;;
esac
done
# This is normally unused
# shellcheck disable=SC2034
APP_BASE_NAME=${0##*/}
# Discard cd standard output in case $CDPATH is set (https://github.com/gradle/gradle/issues/25036)
APP_HOME=$( cd -P "${APP_HOME:-./}" > /dev/null && printf '%s\n' "$PWD" ) || exit
# Use the maximum available, or set MAX_FD != -1 to use that value.
MAX_FD=maximum
warn () {
echo "$*"
} >&2
die () {
echo
echo "$*"
echo
exit 1
} >&2
# OS specific support (must be 'true' or 'false').
cygwin=false
msys=false
darwin=false
nonstop=false
case "$( uname )" in #(
CYGWIN* ) cygwin=true ;; #(
Darwin* ) darwin=true ;; #(
MSYS* | MINGW* ) msys=true ;; #(
NONSTOP* ) nonstop=true ;;
esac
CLASSPATH="\\\"\\\""
# Determine the Java command to use to start the JVM.
if [ -n "$JAVA_HOME" ] ; then
if [ -x "$JAVA_HOME/jre/sh/java" ] ; then
# IBM's JDK on AIX uses strange locations for the executables
JAVACMD=$JAVA_HOME/jre/sh/java
else
JAVACMD=$JAVA_HOME/bin/java
fi
if [ ! -x "$JAVACMD" ] ; then
die "ERROR: JAVA_HOME is set to an invalid directory: $JAVA_HOME
Please set the JAVA_HOME variable in your environment to match the
location of your Java installation."
fi
else
JAVACMD=java
if ! command -v java >/dev/null 2>&1
then
die "ERROR: JAVA_HOME is not set and no 'java' command could be found in your PATH.
Please set the JAVA_HOME variable in your environment to match the
location of your Java installation."
fi
fi
# Increase the maximum file descriptors if we can.
if ! "$cygwin" && ! "$darwin" && ! "$nonstop" ; then
case $MAX_FD in #(
max*)
# In POSIX sh, ulimit -H is undefined. That's why the result is checked to see if it worked.
# shellcheck disable=SC2039,SC3045
MAX_FD=$( ulimit -H -n ) ||
warn "Could not query maximum file descriptor limit"
esac
case $MAX_FD in #(
'' | soft) :;; #(
*)
# In POSIX sh, ulimit -n is undefined. That's why the result is checked to see if it worked.
# shellcheck disable=SC2039,SC3045
ulimit -n "$MAX_FD" ||
warn "Could not set maximum file descriptor limit to $MAX_FD"
esac
fi
# Collect all arguments for the java command, stacking in reverse order:
# * args from the command line
# * the main class name
# * -classpath
# * -D...appname settings
# * --module-path (only if needed)
# * DEFAULT_JVM_OPTS, JAVA_OPTS, and GRADLE_OPTS environment variables.
# For Cygwin or MSYS, switch paths to Windows format before running java
if "$cygwin" || "$msys" ; then
APP_HOME=$( cygpath --path --mixed "$APP_HOME" )
CLASSPATH=$( cygpath --path --mixed "$CLASSPATH" )
JAVACMD=$( cygpath --unix "$JAVACMD" )
# Now convert the arguments - kludge to limit ourselves to /bin/sh
for arg do
if
case $arg in #(
-*) false ;; # don't mess with options #(
/?*) t=${arg#/} t=/${t%%/*} # looks like a POSIX filepath
[ -e "$t" ] ;; #(
*) false ;;
esac
then
arg=$( cygpath --path --ignore --mixed "$arg" )
fi
# Roll the args list around exactly as many times as the number of
# args, so each arg winds up back in the position where it started, but
# possibly modified.
#
# NB: a `for` loop captures its iteration list before it begins, so
# changing the positional parameters here affects neither the number of
# iterations, nor the values presented in `arg`.
shift # remove old arg
set -- "$@" "$arg" # push replacement arg
done
fi
# Add default JVM options here. You can also use JAVA_OPTS and GRADLE_OPTS to pass JVM options to this script.
DEFAULT_JVM_OPTS='"-Xmx64m" "-Xms64m"'
# Collect all arguments for the java command:
# * DEFAULT_JVM_OPTS, JAVA_OPTS, and optsEnvironmentVar are not allowed to contain shell fragments,
# and any embedded shellness will be escaped.
# * For example: A user cannot expect ${Hostname} to be expanded, as it is an environment variable and will be
# treated as '${Hostname}' itself on the command line.
set -- \
"-Dorg.gradle.appname=$APP_BASE_NAME" \
-classpath "$CLASSPATH" \
-jar "$APP_HOME/gradle/wrapper/gradle-wrapper.jar" \
"$@"
# Stop when "xargs" is not available.
if ! command -v xargs >/dev/null 2>&1
then
die "xargs is not available"
fi
# Use "xargs" to parse quoted args.
#
# With -n1 it outputs one arg per line, with the quotes and backslashes removed.
#
# In Bash we could simply go:
#
# readarray ARGS < <( xargs -n1 <<<"$var" ) &&
# set -- "${ARGS[@]}" "$@"
#
# but POSIX shell has neither arrays nor command substitution, so instead we
# post-process each arg (as a line of input to sed) to backslash-escape any
# character that might be a shell metacharacter, then use eval to reverse
# that process (while maintaining the separation between arguments), and wrap
# the whole thing up as a single "set" statement.
#
# This will of course break if any of these variables contains a newline or
# an unmatched quote.
#
eval "set -- $(
printf '%s\n' "$DEFAULT_JVM_OPTS $JAVA_OPTS $GRADLE_OPTS" |
xargs -n1 |
sed ' s~[^-[:alnum:]+,./:=@_]~\\&~g; ' |
tr '\n' ' '
)" '"$@"'
exec "$JAVACMD" "$@"

94
android/gradlew.bat vendored Normal file
View file

@ -0,0 +1,94 @@
@rem
@rem Copyright 2015 the original author or authors.
@rem
@rem Licensed under the Apache License, Version 2.0 (the "License");
@rem you may not use this file except in compliance with the License.
@rem You may obtain a copy of the License at
@rem
@rem https://www.apache.org/licenses/LICENSE-2.0
@rem
@rem Unless required by applicable law or agreed to in writing, software
@rem distributed under the License is distributed on an "AS IS" BASIS,
@rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
@rem See the License for the specific language governing permissions and
@rem limitations under the License.
@rem
@rem SPDX-License-Identifier: Apache-2.0
@rem
@if "%DEBUG%"=="" @echo off
@rem ##########################################################################
@rem
@rem Gradle startup script for Windows
@rem
@rem ##########################################################################
@rem Set local scope for the variables with windows NT shell
if "%OS%"=="Windows_NT" setlocal
set DIRNAME=%~dp0
if "%DIRNAME%"=="" set DIRNAME=.
@rem This is normally unused
set APP_BASE_NAME=%~n0
set APP_HOME=%DIRNAME%
@rem Resolve any "." and ".." in APP_HOME to make it shorter.
for %%i in ("%APP_HOME%") do set APP_HOME=%%~fi
@rem Add default JVM options here. You can also use JAVA_OPTS and GRADLE_OPTS to pass JVM options to this script.
set DEFAULT_JVM_OPTS="-Xmx64m" "-Xms64m"
@rem Find java.exe
if defined JAVA_HOME goto findJavaFromJavaHome
set JAVA_EXE=java.exe
%JAVA_EXE% -version >NUL 2>&1
if %ERRORLEVEL% equ 0 goto execute
echo. 1>&2
echo ERROR: JAVA_HOME is not set and no 'java' command could be found in your PATH. 1>&2
echo. 1>&2
echo Please set the JAVA_HOME variable in your environment to match the 1>&2
echo location of your Java installation. 1>&2
goto fail
:findJavaFromJavaHome
set JAVA_HOME=%JAVA_HOME:"=%
set JAVA_EXE=%JAVA_HOME%/bin/java.exe
if exist "%JAVA_EXE%" goto execute
echo. 1>&2
echo ERROR: JAVA_HOME is set to an invalid directory: %JAVA_HOME% 1>&2
echo. 1>&2
echo Please set the JAVA_HOME variable in your environment to match the 1>&2
echo location of your Java installation. 1>&2
goto fail
:execute
@rem Setup the command line
set CLASSPATH=
@rem Execute Gradle
"%JAVA_EXE%" %DEFAULT_JVM_OPTS% %JAVA_OPTS% %GRADLE_OPTS% "-Dorg.gradle.appname=%APP_BASE_NAME%" -classpath "%CLASSPATH%" -jar "%APP_HOME%\gradle\wrapper\gradle-wrapper.jar" %*
:end
@rem End local scope for the variables with windows NT shell
if %ERRORLEVEL% equ 0 goto mainEnd
:fail
rem Set variable GRADLE_EXIT_CONSOLE if you need the _script_ return code instead of
rem the _cmd.exe /c_ return code!
set EXIT_CODE=%ERRORLEVEL%
if %EXIT_CODE% equ 0 set EXIT_CODE=1
if not ""=="%GRADLE_EXIT_CONSOLE%" exit %EXIT_CODE%
exit /b %EXIT_CODE%
:mainEnd
if "%OS%"=="Windows_NT" endlocal
:omega

5
android/settings.gradle Normal file
View file

@ -0,0 +1,5 @@
include ':app'
include ':capacitor-cordova-android-plugins'
project(':capacitor-cordova-android-plugins').projectDir = new File('./capacitor-cordova-android-plugins/')
apply from: 'capacitor.settings.gradle'

16
android/variables.gradle Normal file
View file

@ -0,0 +1,16 @@
ext {
minSdkVersion = 24
compileSdkVersion = 36
targetSdkVersion = 36
androidxActivityVersion = '1.11.0'
androidxAppCompatVersion = '1.7.1'
androidxCoordinatorLayoutVersion = '1.3.0'
androidxCoreVersion = '1.17.0'
androidxFragmentVersion = '1.8.9'
coreSplashScreenVersion = '1.2.0'
androidxWebkitVersion = '1.14.0'
junitVersion = '4.13.2'
androidxJunitVersion = '1.3.0'
androidxEspressoCoreVersion = '3.7.0'
cordovaAndroidVersion = '14.0.1'
}

60
apps/.eslintrc.cjs Normal file
View file

@ -0,0 +1,60 @@
// Per-package ESLint config for the Preact widget apps under `apps/`.
//
// `root: true` stops ESLint from walking up to the host's
// `cinny/.eslintrc.cjs`, which extends airbnb + the React plugin. Those
// rule sets are tuned for the React host and flag legitimate Preact /
// small-widget patterns as errors (`class=` attributes, arrow-fn
// components, inline icon sub-components, for-of loops, etc.). Keeping
// the hierarchy open would force every widget file to fight host style
// for no real win.
//
// Widgets keep a minimal but real lint pass via the rule sets below:
//
// * `eslint:recommended` — catches genuine bugs (no-undef, no-dupe-*,
// no-redeclare, no-unused-vars, …) without enforcing style.
// * `@typescript-eslint/recommended` — TS-aware variants of the above
// plus type-level checks the recommended set ships.
//
// We deliberately DON'T extend `plugin:react/recommended` —
// `react/react-in-jsx-scope` and `react/no-unknown-property` both flag
// Preact-correct code as errors, and disabling them one by one creates
// a long suppression list. Widget JSX is type-checked by each app's
// `tsc --noEmit` (run by `vite build`), which is the better signal for
// JSX correctness anyway.
module.exports = {
root: true,
// `node` covers `module.exports` in this very file (CommonJS config);
// `browser` is the runtime widget code itself sees.
env: { browser: true, es2021: true, node: true },
extends: [
'eslint:recommended',
'plugin:@typescript-eslint/recommended',
// preact/hooks has the same dep-array semantics as react/hooks, and
// the widget code already carries `// eslint-disable-next-line
// react-hooks/exhaustive-deps` directives at the relevant sites;
// loading the plugin (a) keeps those directives meaningful (without
// it ESLint errors on the «unknown rule» referenced by the comment)
// and (b) catches the real exhaustive-deps mistakes in widget hooks
// for free.
'plugin:react-hooks/recommended',
],
parser: '@typescript-eslint/parser',
parserOptions: {
ecmaVersion: 'latest',
sourceType: 'module',
ecmaFeatures: { jsx: true },
},
plugins: ['@typescript-eslint', 'react-hooks'],
rules: {
// Underscore-prefixed args are intentionally unused (Preact event
// handlers receive args the body doesn't need); match the host's
// convention so lint reads consistently across both trees.
'@typescript-eslint/no-unused-vars': ['error', { argsIgnorePattern: '^_' }],
// Widget bridge-protocol regexes occasionally escape `-` inside
// character classes for visual clarity (e.g. `[0-9\-]`). The escape
// is harmless and pre-existing across all three widgets — keeping
// the rule on would force a churn-y diff in code that's been stable
// since the v0.7.6 bridge dialect work.
'no-useless-escape': 'off',
},
};

3
apps/widget-discord/.gitignore vendored Normal file
View file

@ -0,0 +1,3 @@
node_modules
dist
*.log

View file

@ -0,0 +1,193 @@
# @vojo/widget-discord
Vojo Discord bridge management widget — mounts inside `/bots/discord`
in the Vojo client. Mirrors the Telegram widget contract; protocol
specifics differ because mautrix-discord runs on the **legacy** mautrix
command framework, not bridgev2 (the Discord bridge had not yet been
ported to v2 as of January 2026 — see
https://mau.fi/blog/2026-01-mautrix-release/).
This is **not** a Discord client. It's a small panel that drives the
mautrix-discord bridge bot (`@discordbot:vojo.chat`) by sending text
commands in the control DM and rendering the bot's text replies. It
ships QR-only login (the Discord token-login flow stays accessible via
chat-fallback for power users).
## Layout
```
src/
├── bootstrap.ts Parse URL params (matches BotWidgetEmbed.ts)
├── widget-api.ts Inline matrix-widget-api postMessage transport
├── App.tsx UI: status pill, QR panel, logout / reconnect cards, transcript
├── main.tsx Entry: bootstrap + render
├── state.ts LoginState reducer + hydrate-from-timeline
├── styles.css Theme-aware CSS variables (Dawn palette)
├── i18n/ Tiny RU/EN dictionary harness
└── bridge-protocol/
├── types.ts LoginEvent + ParsableEvent types
├── parser.ts Dialect dispatch shim
└── dialects/
└── legacy_v076.ts mautrix-discord v0.7.6 wording
```
## Login flow (QR only)
1. Widget sends `!discord login-qr`.
2. Bridge replies with an `m.image` event whose `body` is a Discord
remoteauth URL (`https://discord.com/ra/<token>`). The host driver
strips `url`/`file`/`info` so the widget never touches the uploaded
PNG bytes — it re-encodes the URL into an SVG QR matrix client-side
via `qrcode-generator`.
3. The user scans the QR with the **Discord mobile app** (Settings →
Devices → Scan QR Code). Discord's remoteauth gateway requires the
mobile app — desktop Discord and the browser cannot scan.
4. Bridge redacts the `m.image` event after a successful scan and sends
`Successfully logged in as @<username>`.
5. Widget fires `!discord ping` to pick up the discord snowflake for
the connected pill.
If Discord asks for a CAPTCHA, the bridge replies with the standard
error line plus a hint about token-login. The widget surfaces an amber
warning suggesting the user retry later or use chat-fallback.
## Status probe
Discord's legacy command system has no `list-logins` API; status is
queried via `!discord ping`. The four reply variants map to four UI
states:
- `You're not logged in` → disconnected
- `You're logged in as @x (\`<id>\`)` → connected
- `You have a Discord token stored, but are not connected for some reason 🤔` → connected_dead (token_stored)
- `You're logged in, but the Discord connection seems to be dead 💥` → connected_dead (connection_dead)
`connected_dead` exposes a «Переподключиться» card that sends
`!discord reconnect`. `disconnect` is recognised for chat-fallback
typists but never sent by the widget.
## Local development
Same overlay mechanism as the Telegram widget — create
`config.local.json` at the project root (gitignored) with a `bots[]`
entry overriding the discord widget's `experience.url` to your local
dev server:
```bash
# one-time: install widget deps
cd apps/widget-discord && npm install
# config.local.json (gitignored) at the project root
cat > /home/ubuntu/projects/vojo/cinny/config.local.json <<'JSON'
{
"bots": [
{
"id": "discord",
"experience": {
"type": "matrix-widget",
"url": "http://localhost:8082/",
"commandPrefix": "!discord"
}
}
]
}
JSON
```
`http://localhost:*` URLs pass the host's URL validator only in dev
builds — see `src/app/features/bots/catalog.ts` `import.meta.env.DEV`
branch. Production builds drop the branch via Vite's dead-code
elimination AND enforce an origin allowlist (`PROD_WIDGET_ORIGINS`).
Run both servers:
```bash
# terminal 1 — widget on :8082 with HMR
cd apps/widget-discord && npm run dev
# terminal 2 — host SPA on :8080
cd /home/ubuntu/projects/vojo/cinny && npm start
```
Open `http://localhost:8080/bots/discord`. The Telegram widget on :8081
can run in parallel with no port conflict.
## Build
```bash
npm run build
```
Outputs to `apps/widget-discord/dist/`. Deploy by rsyncing `dist/*` into
`~/vojo/widgets/discord/` on the production host (Caddy serves this via
the `widgets.vojo.chat` block).
## Hosting (server-side, runbook)
Pre-requisite: `widgets.vojo.chat` already exists for the Telegram
widget — only the Caddy `widgets.vojo.chat` block needs a new
`handle_path` and the docker host needs a new directory.
1. `~/vojo/caddy/Caddyfile` — append to the existing
`widgets.vojo.chat { … }` block, beside the Telegram `handle_path`:
```
handle_path /discord/* {
root * /var/www/widgets/discord
try_files {path} /index.html
file_server
}
```
2. `mkdir -p ~/vojo/widgets/discord` (placeholder so the bind-mount has
something to serve), then `docker compose up -d caddy` (or `reload`).
3. Verify directly:
`curl -I https://widgets.vojo.chat/discord/index.html` should
return 200 and the `Content-Security-Policy` header.
## Adding the discord bridge to docker-compose
```yaml
discord-bridge:
image: dock.mau.dev/mautrix/discord:v0.7.6
restart: unless-stopped
volumes:
- ./mautrix-discord:/data
```
Then `~/vojo/synapse/homeserver.yaml` needs the discord registration
file added to `app_service_config_files`:
```yaml
app_service_config_files:
- /data/telegram-registration.yaml
- /data/discord-registration.yaml
```
The bridge's `command_prefix` defaults to `!discord` — keep it that
way so it matches the widget's `experience.commandPrefix`. If you
override it in `mautrix-discord/config.yaml`, mirror the override in
`/config.json`.
## Capacitor (Android)
`capacitor.config.ts` already allow-navigates `widgets.vojo.chat` for
the Telegram widget; no further change needed.
## Capability contract
The widget requests EXACTLY this set (matches the host's
`BotWidgetDriver.getBotWidgetCapabilities`):
```
org.matrix.msc2762.timeline:<roomId>
org.matrix.msc2762.send.event:m.room.message#m.text
org.matrix.msc2762.receive.event:m.room.message#m.text
org.matrix.msc2762.receive.event:m.room.message#m.notice
org.matrix.msc2762.receive.event:m.room.message#m.image
org.matrix.msc2762.receive.event:m.room.redaction
org.matrix.msc2762.receive.state_event:m.room.member
```
`m.image` is the QR carrier; `m.room.redaction` signals the bridge
consumed the QR after a successful scan. The host sanitizer strips
`url`/`file`/`info` from `m.image` content, so only the QR URL string
inside `body` survives the boundary.

View file

@ -0,0 +1,12 @@
<!doctype html>
<html lang="ru">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, viewport-fit=cover" />
<title>Discord bridge — Vojo</title>
</head>
<body>
<div id="app"></div>
<script type="module" src="/src/main.tsx"></script>
</body>
</html>

1999
apps/widget-discord/package-lock.json generated Normal file

File diff suppressed because it is too large Load diff

View file

@ -0,0 +1,21 @@
{
"name": "@vojo/widget-discord",
"version": "0.0.1",
"private": true,
"description": "Vojo Discord bridge management widget — mounts inside /bots/discord",
"type": "module",
"scripts": {
"dev": "vite",
"build": "tsc --noEmit && vite build",
"preview": "vite preview"
},
"dependencies": {
"preact": "10.22.1",
"qrcode-generator": "1.4.4"
},
"devDependencies": {
"@preact/preset-vite": "2.9.0",
"typescript": "5.4.5",
"vite": "5.4.19"
}
}

File diff suppressed because it is too large Load diff

View file

@ -0,0 +1,69 @@
// Parse the URL params the Phase 2 bot widget host appends when loading
// experience.url. Source of truth on the host side:
// src/app/features/bots/BotWidgetEmbed.ts (getBotWidgetUrl).
// Keep this in sync if the host adds params.
//
// Identical shape to apps/widget-telegram/src/bootstrap.ts on purpose —
// the host emits the same param set for every bot. Differences between
// telegram and discord live in the bridge protocol, not in bootstrap.
export type WidgetBootstrap = {
widgetId: string;
parentUrl: string;
parentOrigin: string;
roomId: string;
userId: string;
botId: string;
botMxid: string;
/** Bridge command prefix (e.g. `!discord`). Always non-empty the host
* validator (catalog.ts) defaults missing values to `!tg` and rejects
* malformed overrides, so the discord bot's /config.json entry MUST set
* `experience.commandPrefix: "!discord"` to override the default. The
* widget prepends `<commandPrefix> ` to every outbound command. */
commandPrefix: string;
theme: 'light' | 'dark';
clientLanguage: string;
};
export type BootstrapResult =
| { ok: true; bootstrap: WidgetBootstrap }
| { ok: false; missing: string[] };
const REQUIRED = ['widgetId', 'parentUrl', 'roomId', 'userId', 'botMxid', 'commandPrefix'] as const;
export const readBootstrap = (search: string): BootstrapResult => {
const params = new URLSearchParams(search);
const get = (k: string) => params.get(k) ?? '';
const missing = REQUIRED.filter((k) => !params.get(k));
if (missing.length > 0) return { ok: false, missing: [...missing] };
// Origin is what the widget validates against on incoming postMessage —
// see widget-api.ts. Falling back to '*' would defeat the security
// boundary, so a malformed parentUrl bails out as a missing-param error.
let parentOrigin: string;
try {
parentOrigin = new URL(get('parentUrl')).origin;
} catch {
return { ok: false, missing: ['parentUrl'] };
}
const themeRaw = get('theme');
const theme: 'light' | 'dark' = themeRaw === 'dark' ? 'dark' : 'light';
return {
ok: true,
bootstrap: {
widgetId: get('widgetId'),
parentUrl: get('parentUrl'),
parentOrigin,
roomId: get('roomId'),
userId: get('userId'),
botId: get('botId'),
botMxid: get('botMxid'),
commandPrefix: get('commandPrefix'),
theme,
clientLanguage: get('clientLanguage'),
},
};
};

View file

@ -0,0 +1,554 @@
// Dialect: mautrix-discord v0.7.6 (16 Feb 2026). The bridge runs on the
// LEGACY mautrix command framework — `maunium.net/go/mautrix/bridge/commands`,
// NOT bridgev2. As of January 2026 the mautrix maintainers flagged Discord
// as «not yet migrated to bridgev2» (mau.fi/blog/2026-01-mautrix-release/),
// so this dialect is the canonical one until the v2 migration lands.
//
// Each regex below is paired with its upstream source line in
// github.com/mautrix/discord/blob/v0.7.6/commands.go. If wording drifts in
// a future patch, replace this file with a sibling `legacy_v077.ts`
// (or whatever) and switch the import in ../parser.ts.
//
// Body encoding note: legacy mautrix commands use `ce.Reply(...)` which
// renders through `format.RenderMarkdown` in the bridge framework. Our
// host driver strips `formatted_body` (Phase 2 contract), so the widget
// only sees the markdown source — backticks, asterisks, escaped angle-
// brackets stay literal.
import type { LoginEvent, ParsableEvent } from '../types';
// --- Regex table ----------------------------------------------------------
// Ping replies — commands.go:fnPing (l.297-310 in v0.7.6). All four are
// distinct phrasings; we capture each separately so the state machine can
// route them to different status pills.
//
// «You're logged in as @<username> (`<id>`)» — the trailing parens hold the
// numeric Discord snowflake wrapped in markdown backticks. Both are useful
// for surfacing in the UI.
const PING_LOGGED_IN_RE = /^you'?re logged in as\s+@?(.+?)\s+\(`?(\d+)`?\)\.?$/i;
// «You're not logged in» — exact match, no period. The legacy framework
// doesn't append punctuation here.
const PING_NOT_LOGGED_IN_RE = /^you'?re not logged in\.?$/i;
// «You have a Discord token stored, but are not connected for some reason 🤔»
// — the emoji is part of the literal upstream string and we tolerate optional
// trailing whitespace / period.
const PING_TOKEN_STORED_RE = /^you have a discord token stored, but are not connected/i;
// «You're logged in, but the Discord connection seems to be dead 💥»
const PING_CONNECTION_DEAD_RE = /^you'?re logged in, but the discord connection seems to be dead/i;
// login-token / login-qr success — commands.go:fnLoginToken (l.156) and
// fnLoginQR (l.220). Format: `Successfully logged in as @<username>`. The
// QR-login path doesn't include the snowflake; the token-login path has
// «Connecting to Discord as user ID %d» BEFORE the success line, but we
// only need the success terminator. Capturing the handle is enough — App
// fires `ping` after to pick up the snowflake.
const LOGIN_SUCCESS_RE = /^successfully logged in as\s+@?(.+?)\.?$/i;
// login-qr CAPTCHA path — Vojo-patched bridge sends a single-line
// `VOJO-CAPTCHA-CHALLENGE-V1 {json}` m.notice carrying the hCaptcha
// sitekey, session_id, rqdata, rqtoken. The sentinel is markdown-inert by
// design (no `_`, `*`, `` ` ``, `[`, `<` characters): even if a future
// patch routes the bridge reply through goldmark again, the prefix
// survives intact. The bridge currently sends the notice via
// SendMessageEvent directly to bypass the framework's markdown round-trip.
// See bridge `commands_captcha.go` for the producer side.
const CAPTCHA_CHALLENGE_PREFIX = 'VOJO-CAPTCHA-CHALLENGE-V1';
const CAPTCHA_CHALLENGE_RE = /^VOJO-CAPTCHA-CHALLENGE-V1\s+(\{[\s\S]*\})\s*$/;
// Vojo-patched bridge emits this sentinel right after «Successfully logged
// in as @user» (commands_login_space.go::sendLoginSpaceNotice). Carries the
// matrix.to URL of the user's personal Discord space so the widget can
// render a CTA. Same markdown-inert + structured-JSON discipline as the
// captcha sentinel above; the bridge sends this via SendMessageEvent to
// bypass goldmark round-trip.
const LOGIN_SPACE_SENTINEL_PREFIX = 'VOJO-LOGIN-SPACE-V1';
const LOGIN_SPACE_SENTINEL_RE = /^VOJO-LOGIN-SPACE-V1\s+(\{[\s\S]*\})\s*$/;
// Legacy CAPTCHA fallback — commands.go:fnLoginQR (l.207-209) on UNPATCHED
// upstream v0.7.6: «CAPTCHAs are currently not supported - use token login
// instead». Kept so a deployment running unpatched bridge still produces a
// useful hint rather than a generic Go-error tail.
const CAPTCHA_REQUIRED_RE = /captchas? are currently not supported/i;
// Generic «Error logging in: %v» — fnLoginQR l.205 / 211 (different
// branches funnel through the same Reply call). Capture the Go-error tail
// as `reason`. Order matters: CAPTCHA_REQUIRED must be checked BEFORE this
// trap because the captcha case is a more specific subset.
const LOGIN_FAILED_RE = /^error logging in:\s*(.*)$/i;
// «Error connecting to login websocket: %v» — fnLoginQR l.184. Pre-QR
// failure (couldn't reach Discord's remoteauth gateway). Distinguish from
// LOGIN_FAILED so the App can surface a more accurate message.
const LOGIN_WEBSOCKET_FAILED_RE = /^error connecting to login websocket:\s*(.*)$/i;
// «Error connecting after login: %v» — fnLoginQR l.213. Post-QR rare path:
// remoteauth handed us a token but the immediate Discord connect failed.
const CONNECT_AFTER_LOGIN_FAILED_RE = /^error connecting after login:\s*(.*)$/i;
// «Failed to prepare login: %v» — fnLoginQR l.176. Pre-QR initialisation
// failure (remoteauth couldn't even start). Routes back to disconnected.
const PREPARE_LOGIN_FAILED_RE = /^failed to prepare login:\s*(.*)$/i;
// «You're already logged in» — both fnLoginToken (l.117) and fnLoginQR
// (l.171). Replied when the user clicks login but ping would have shown
// connected. We dispatch a re-ping to reconcile.
const ALREADY_LOGGED_IN_RE = /^you'?re already logged in\.?$/i;
// Logout — commands.go:fnLogout (l.275-280).
const LOGOUT_OK_RE = /^logged out successfully\.?$/i;
const LOGOUT_NO_OP_RE = /^you weren'?t logged in, but data was re-cleared/i;
// Disconnect — commands.go:fnDisconnect (l.318-326). User-typed-only path
// (the widget never sends `disconnect`), but recognising the replies keeps
// chat-fallback typists from confusing the state machine.
const DISCONNECT_OK_RE = /^successfully disconnected\.?$/i;
const DISCONNECT_NO_OP_RE = /^you'?re already not connected\.?$/i;
const DISCONNECT_FAILED_RE = /^error while disconnecting:\s*(.*)$/i;
// Reconnect — commands.go:fnReconnect (l.339-347). Used as recovery from
// `connection_dead` / `token_stored_not_connected` ping replies.
const RECONNECT_OK_RE = /^successfully reconnected\.?$/i;
const RECONNECT_NO_OP_RE = /^you'?re already connected\.?$/i;
const RECONNECT_FAILED_RE = /^error while reconnecting:\s*(.*)$/i;
// Unknown command — bridge/commands/processor.go (legacy framework). The
// exact wording differs between framework versions; this regex tolerates
// the canonical «Unknown command. Try `help`.» phrasing.
const UNKNOWN_COMMAND_RE = /^unknown command\.?\s*(?:try\s+)?`?help`?/i;
// --- Body parser ----------------------------------------------------------
const trimReplyBody = (raw: string): string => raw.trim();
export const parseLegacyV076Body = (rawBody: string): LoginEvent => {
const body = trimReplyBody(rawBody);
if (body.length === 0) return { kind: 'unknown' };
// ORDER MATTERS:
// 1. CAPTCHA must be checked before the generic LOGIN_FAILED — captcha
// bodies match LOGIN_FAILED_RE but carry the more-specific suffix.
// 2. LOGIN_SUCCESS_RE has a permissive `(.+?)` capture; we keep it AFTER
// explicit ping replies so a future ping wording drift can't swallow
// a success line.
// Ping replies (most common) — try first.
if (PING_NOT_LOGGED_IN_RE.test(body)) return { kind: 'not_logged_in' };
if (PING_TOKEN_STORED_RE.test(body)) return { kind: 'token_stored_not_connected' };
if (PING_CONNECTION_DEAD_RE.test(body)) return { kind: 'connection_dead' };
const pingLoggedInMatch = PING_LOGGED_IN_RE.exec(body);
if (pingLoggedInMatch) {
return {
kind: 'logged_in',
handle: pingLoggedInMatch[1].trim(),
discordId: pingLoggedInMatch[2],
};
}
// Login lifecycle.
// Vojo-patched bridge: structured hCaptcha challenge wins over every
// legacy regex — checked first so the JSON payload survives.
if (body.startsWith(CAPTCHA_CHALLENGE_PREFIX)) {
const match = CAPTCHA_CHALLENGE_RE.exec(body);
if (match) {
try {
const payload = JSON.parse(match[1]) as Record<string, unknown>;
const service = typeof payload.service === 'string' ? payload.service : '';
const sitekey = typeof payload.sitekey === 'string' ? payload.sitekey : '';
const sessionId = typeof payload.session_id === 'string' ? payload.session_id : '';
const rqdata = typeof payload.rqdata === 'string' ? payload.rqdata : '';
const rqtoken = typeof payload.rqtoken === 'string' ? payload.rqtoken : '';
if (sitekey && rqtoken) {
return { kind: 'captcha_challenge', service, sitekey, sessionId, rqdata, rqtoken };
}
} catch {
// fall through — malformed payload is treated as unknown
}
}
return { kind: 'unknown' };
}
if (CAPTCHA_REQUIRED_RE.test(body)) return { kind: 'captcha_required' };
// Vojo login-space sentinel: structured JSON with the personal Discord
// space's matrix.to URL. Checked alongside the captcha sentinel —
// markdown-inert prefix means it lands verbatim from the bridge, parsed
// into a `space_ready` event for the reducer to attach to connected state.
// Malformed payload (missing/empty `matrix_to_url`, JSON parse failure) is
// silently dropped as `unknown` rather than surfacing a stale CTA.
if (body.startsWith(LOGIN_SPACE_SENTINEL_PREFIX)) {
const match = LOGIN_SPACE_SENTINEL_RE.exec(body);
if (match) {
try {
const payload = JSON.parse(match[1]) as Record<string, unknown>;
const matrixToUrl = typeof payload.matrix_to_url === 'string' ? payload.matrix_to_url : '';
if (matrixToUrl) {
return { kind: 'space_ready', matrixToUrl };
}
} catch {
// fall through — malformed payload is treated as unknown
}
}
return { kind: 'unknown' };
}
const loginWsMatch = LOGIN_WEBSOCKET_FAILED_RE.exec(body);
if (loginWsMatch) return { kind: 'login_websocket_failed', reason: loginWsMatch[1].trim() };
const connectAfterMatch = CONNECT_AFTER_LOGIN_FAILED_RE.exec(body);
if (connectAfterMatch)
return { kind: 'connect_after_login_failed', reason: connectAfterMatch[1].trim() };
const prepareMatch = PREPARE_LOGIN_FAILED_RE.exec(body);
if (prepareMatch) return { kind: 'prepare_login_failed', reason: prepareMatch[1].trim() };
const loginFailedMatch = LOGIN_FAILED_RE.exec(body);
if (loginFailedMatch) return { kind: 'login_failed', reason: loginFailedMatch[1].trim() };
if (ALREADY_LOGGED_IN_RE.test(body)) return { kind: 'already_logged_in' };
// Login success — capture the handle. Discord usernames may include `.`
// and other ASCII punctuation; the regex's `(.+?)` is greedy-enough.
const successMatch = LOGIN_SUCCESS_RE.exec(body);
if (successMatch) {
const handleRaw = successMatch[1].trim();
return { kind: 'login_success', handle: handleRaw };
}
// Logout / disconnect / reconnect lifecycle.
if (LOGOUT_OK_RE.test(body)) return { kind: 'logout_ok' };
if (LOGOUT_NO_OP_RE.test(body)) return { kind: 'logout_no_op' };
if (DISCONNECT_OK_RE.test(body)) return { kind: 'disconnect_ok' };
if (DISCONNECT_NO_OP_RE.test(body)) return { kind: 'disconnect_no_op' };
const disconnectFailedMatch = DISCONNECT_FAILED_RE.exec(body);
if (disconnectFailedMatch)
return { kind: 'disconnect_failed', reason: disconnectFailedMatch[1].trim() };
if (RECONNECT_OK_RE.test(body)) return { kind: 'reconnect_ok' };
if (RECONNECT_NO_OP_RE.test(body)) return { kind: 'reconnect_no_op' };
const reconnectFailedMatch = RECONNECT_FAILED_RE.exec(body);
if (reconnectFailedMatch)
return { kind: 'reconnect_failed', reason: reconnectFailedMatch[1].trim() };
if (UNKNOWN_COMMAND_RE.test(body)) return { kind: 'unknown_command' };
return { kind: 'unknown' };
};
// --- Full-event parser ----------------------------------------------------
//
// `parseEventLegacyV076` dispatches on `event.type`:
//
// * `m.room.redaction` → `qr_redacted`. The state machine pairs the
// redaction's `redacts` against the active QR event id; an unrelated
// redaction is dropped silently.
//
// * `m.room.message` + `msgtype=m.image` → `qr_displayed` when the body
// contains a Discord remoteauth URL. Discord doesn't rotate the QR (no
// m.replace edits), but we still honour `m.relates_to.rel_type=m.replace`
// for forward-compat with a hypothetical future bridge that does.
//
// * `m.room.message` + `msgtype=m.text|m.notice` → existing
// `parseLegacyV076Body(body)` path.
// Discord remoteauth URLs encode the auth handshake in a path on
// `discordapp.com` (the OLD Discord domain — Discord still uses it as
// the canonical remoteauth host because the URL is consumed by the
// mobile app's deep-link handler, not by browser routing).
//
// Verified upstream: mautrix/discord/remoteauth/serverpackets.go at v0.7.6
// builds the QR string as `"https://discordapp.com/ra/" + Fingerprint` —
// see https://github.com/mautrix/discord/blob/v0.7.6/remoteauth/serverpackets.go.
//
// We accept both `discordapp.com` (canonical) AND `discord.com` because
// Discord has been gradually consolidating onto discord.com over years
// and a future bridge release could flip — keeping both means the
// widget survives the transition without a co-ordinated push.
// Subdomains (`canary.`, `ptb.`) aren't expected here (bridge talks to
// production remoteauth) but we tolerate them as belt-and-suspenders.
const DISCORD_REMOTEAUTH_URL_RE =
/https:\/\/(?:[a-z0-9-]+\.)?(?:discordapp|discord)\.com\/[A-Za-z0-9/_\-+=.~?&]+/i;
const isObject = (value: unknown): value is Record<string, unknown> =>
typeof value === 'object' && value !== null && !Array.isArray(value);
export const parseEventLegacyV076 = (event: ParsableEvent): LoginEvent => {
if (event.type === 'm.room.redaction') {
const target =
typeof event.redacts === 'string'
? event.redacts
: isObject(event.content) && typeof event.content.redacts === 'string'
? event.content.redacts
: undefined;
if (!target) return { kind: 'unknown' };
return { kind: 'qr_redacted', redactsEventId: target };
}
if (event.type !== 'm.room.message') return { kind: 'unknown' };
const msgtype = event.content?.msgtype;
if (msgtype === 'm.image') {
// Edits replace `body` by spec; bridges typically also mirror the new
// body into `m.new_content.body`. Discord's bridge doesn't edit QRs in
// the v0.7.6 timeline, but we read both spots so a future change
// doesn't quietly break the parser.
const newContent = isObject(event.content['m.new_content'])
? (event.content['m.new_content'] as { body?: unknown })
: undefined;
const editedBody = typeof newContent?.body === 'string' ? newContent.body : undefined;
const directBody = typeof event.content.body === 'string' ? event.content.body : '';
const body = editedBody ?? directBody;
const match = body.match(DISCORD_REMOTEAUTH_URL_RE);
if (!match) return { kind: 'unknown' };
const relatesTo = isObject(event.content['m.relates_to'])
? (event.content['m.relates_to'] as { rel_type?: unknown; event_id?: unknown })
: undefined;
const replacesEventId =
relatesTo?.rel_type === 'm.replace' && typeof relatesTo.event_id === 'string'
? relatesTo.event_id
: undefined;
return {
kind: 'qr_displayed',
discordUrl: match[0],
eventId: event.event_id,
replacesEventId,
};
}
if (msgtype !== 'm.text' && msgtype !== 'm.notice') return { kind: 'unknown' };
const body = typeof event.content.body === 'string' ? event.content.body : '';
return parseLegacyV076Body(body);
};
// --- DEV sanity assertions ------------------------------------------------
// Vite tree-shakes this branch in production builds: `import.meta.env.DEV`
// is replaced with the literal `false` and the call site collapses, so the
// fixture array never ships. Failure throws — HMR/dev-overlay surfaces the
// first regression on reload.
if (import.meta.env.DEV) {
runSanityChecks();
}
function runSanityChecks(): void {
// Body-only cases (`parseLegacyV076Body`).
const cases: Array<[string, LoginEvent]> = [
// Ping replies.
["You're not logged in", { kind: 'not_logged_in' }],
["You're not logged in.", { kind: 'not_logged_in' }],
[
'You have a Discord token stored, but are not connected for some reason 🤔',
{ kind: 'token_stored_not_connected' },
],
[
"You're logged in, but the Discord connection seems to be dead 💥",
{ kind: 'connection_dead' },
],
[
"You're logged in as @example (`123456789`)",
{ kind: 'logged_in', handle: 'example', discordId: '123456789' },
],
// Discord usernames support `.` since the 2026 username migration.
[
"You're logged in as @user.name (`987654321`)",
{ kind: 'logged_in', handle: 'user.name', discordId: '987654321' },
],
// Login success (post-QR scan). No snowflake in this line; App fires
// `ping` afterwards to pick up the discordId.
['Successfully logged in as @example', { kind: 'login_success', handle: 'example' }],
['Successfully logged in as @user.name', { kind: 'login_success', handle: 'user.name' }],
// Login failure paths.
['Error logging in: rate limited 429', { kind: 'login_failed', reason: 'rate limited 429' }],
// CAPTCHA legacy fallback — pre-empts LOGIN_FAILED_RE. Fires only on
// unpatched upstream v0.7.6.
[
'Error logging in: captcha-required 400\n\nCAPTCHAs are currently not supported - use token login instead',
{ kind: 'captcha_required' },
],
// Vojo-patched bridge — structured hCaptcha challenge. Sentinel prefix
// is checked before any regex so the JSON body is never misclassified.
// Use a realistic-shape rqtoken (JWT-style segments separated by `.`,
// base64url payload with `_`/`-`/`=`) so a regression where the regex
// accidentally trips on those characters is caught in CI.
[
'VOJO-CAPTCHA-CHALLENGE-V1 {"service":"hcaptcha","sitekey":"a9b5fb07-92ff-493f-86fe-352a2803b3df","session_id":"e971514e-4a6e-4a45-a869-01e61421327c","rqdata":"fgpS6hRTe96TX5qXD7QZgLQwgbQal50jmYsSPyZHqdY+UdfpECcH9gAZESHuGwi0k3n2aCUDs/32bITzKBYGSYjRUbKJqsN0gSo3JHr7SzyPB4bMLcwkOT15yro6f2ax","rqtoken":"IlRGQ3BWQVdGLzJhZUxHMDUrWkV3OE9TNjF6MjNCOS9zOWx2Nk1idzBOdlVvTy9abmZqUnZoZDNnZ2lBUm80Ull1NVRPL0E9PXhFTVRpOW5QeXFmaGF1MFEi.afpL4w.vi8MtSRKgUhesyHDNy4uWwpft1A"}',
{
kind: 'captcha_challenge',
service: 'hcaptcha',
sitekey: 'a9b5fb07-92ff-493f-86fe-352a2803b3df',
sessionId: 'e971514e-4a6e-4a45-a869-01e61421327c',
rqdata:
'fgpS6hRTe96TX5qXD7QZgLQwgbQal50jmYsSPyZHqdY+UdfpECcH9gAZESHuGwi0k3n2aCUDs/32bITzKBYGSYjRUbKJqsN0gSo3JHr7SzyPB4bMLcwkOT15yro6f2ax',
rqtoken:
'IlRGQ3BWQVdGLzJhZUxHMDUrWkV3OE9TNjF6MjNCOS9zOWx2Nk1idzBOdlVvTy9abmZqUnZoZDNnZ2lBUm80Ull1NVRPL0E9PXhFTVRpOW5QeXFmaGF1MFEi.afpL4w.vi8MtSRKgUhesyHDNy4uWwpft1A',
},
],
// Malformed challenge JSON falls through to `unknown` so a corrupted
// bridge build doesn't crash the parser.
['VOJO-CAPTCHA-CHALLENGE-V1 {not-json', { kind: 'unknown' }],
[
'Error connecting to login websocket: dial tcp i/o timeout',
{ kind: 'login_websocket_failed', reason: 'dial tcp i/o timeout' },
],
[
'Error connecting after login: gateway timeout',
{ kind: 'connect_after_login_failed', reason: 'gateway timeout' },
],
[
'Failed to prepare login: remoteauth init failed',
{ kind: 'prepare_login_failed', reason: 'remoteauth init failed' },
],
["You're already logged in", { kind: 'already_logged_in' }],
// Logout.
['Logged out successfully.', { kind: 'logout_ok' }],
["You weren't logged in, but data was re-cleared just to be safe.", { kind: 'logout_no_op' }],
// Disconnect / reconnect.
['Successfully disconnected', { kind: 'disconnect_ok' }],
["You're already not connected", { kind: 'disconnect_no_op' }],
[
'Error while disconnecting: connection already closed',
{ kind: 'disconnect_failed', reason: 'connection already closed' },
],
['Successfully reconnected', { kind: 'reconnect_ok' }],
["You're already connected", { kind: 'reconnect_no_op' }],
[
'Error while reconnecting: dial tcp connection refused',
{ kind: 'reconnect_failed', reason: 'dial tcp connection refused' },
],
// Unknown command — the bridge framework's wording.
['Unknown command. Try `help`.', { kind: 'unknown_command' }],
// Catch-all.
['Some completely unknown bridge reply that does not match any anchor', { kind: 'unknown' }],
];
for (const [body, expected] of cases) {
const actual = parseLegacyV076Body(body);
if (!sameEvent(actual, expected)) {
// eslint-disable-next-line no-console
console.error('[legacy_v076 sanity] mismatch', { body, actual, expected });
throw new Error(
`legacy_v076 parser sanity failed for body ${JSON.stringify(body)} — see console for diff`
);
}
}
// Full-event cases — m.image / m.room.redaction / m.notice fall-through.
const eventCases: Array<[ParsableEvent, LoginEvent]> = [
[
// Canonical upstream form — `discordapp.com` (verified at v0.7.6
// serverpackets.go). The legacy domain is what the Discord mobile
// app's deep-link handler accepts.
{
type: 'm.room.message',
event_id: '$qr1',
sender: '@discordbot:vojo.chat',
content: { msgtype: 'm.image', body: 'https://discordapp.com/ra/ABCDEF' },
},
{ kind: 'qr_displayed', discordUrl: 'https://discordapp.com/ra/ABCDEF', eventId: '$qr1' },
],
[
// Forward-compat: a future bridge release could flip to
// `discord.com`. The regex tolerates both.
{
type: 'm.room.message',
event_id: '$qr1b',
sender: '@discordbot:vojo.chat',
content: { msgtype: 'm.image', body: 'https://discord.com/ra/ABCDEF' },
},
{ kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/ABCDEF', eventId: '$qr1b' },
],
[
// Bare m.image without a discord URL — bridge has no business sending
// these here, but the parser declines to invent state.
{
type: 'm.room.message',
event_id: '$rand',
sender: '@discordbot:vojo.chat',
content: { msgtype: 'm.image', body: 'random non-discord image caption' },
},
{ kind: 'unknown' },
],
[
// Forward-compat: hypothetical future edit. Verifies the rotation
// path works even though Discord doesn't currently rotate.
{
type: 'm.room.message',
event_id: '$qr2',
sender: '@discordbot:vojo.chat',
content: {
msgtype: 'm.image',
body: 'https://discordapp.com/ra/OLD',
'm.relates_to': { rel_type: 'm.replace', event_id: '$qr1' },
'm.new_content': { msgtype: 'm.image', body: 'https://discordapp.com/ra/ROTATED' },
},
},
{
kind: 'qr_displayed',
discordUrl: 'https://discordapp.com/ra/ROTATED',
eventId: '$qr2',
replacesEventId: '$qr1',
},
],
[
// Redaction — top-level `redacts` (host sanitizer mirrors there).
{
type: 'm.room.redaction',
event_id: '$red1',
sender: '@discordbot:vojo.chat',
content: { redacts: '$qr1' },
redacts: '$qr1',
},
{ kind: 'qr_redacted', redactsEventId: '$qr1' },
],
[
// Redaction missing target — sanitizer should already reject; defence
// in depth.
{
type: 'm.room.redaction',
event_id: '$red2',
sender: '@discordbot:vojo.chat',
content: {},
},
{ kind: 'unknown' },
],
[
// m.notice fall-through — preserves the body-side parser path.
{
type: 'm.room.message',
event_id: '$n1',
sender: '@discordbot:vojo.chat',
content: { msgtype: 'm.notice', body: "You're not logged in" },
},
{ kind: 'not_logged_in' },
],
];
for (const [event, expected] of eventCases) {
const actual = parseEventLegacyV076(event);
if (!sameEvent(actual, expected)) {
// eslint-disable-next-line no-console
console.error('[legacy_v076 event sanity] mismatch', { event, actual, expected });
throw new Error(
`legacy_v076 event-parser sanity failed for type=${event.type} msgtype=${
event.content?.msgtype ?? '<none>'
}`
);
}
}
}
function sameEvent(a: LoginEvent, b: LoginEvent): boolean {
if (a.kind !== b.kind) return false;
return JSON.stringify(a) === JSON.stringify(b);
}

View file

@ -0,0 +1,18 @@
// Parser shim. The widget consumes a single `parseEvent(rawEvent)` and
// the dialect handles the full event surface — m.text, m.notice, m.image
// (QR broadcasts), m.room.redaction (post-scan cleanup). M-discord ships
// one dialect, `legacy_v076`, for the operator's current bridge image.
// When mautrix-discord eventually migrates to bridgev2 (the team flagged
// this as «not yet» as of 2026-01), add a sibling dialect file and
// switch the import below.
//
// The dialects/ subdirectory is kept as a seam for that swap; we don't
// implement runtime autodetect (the operator owns one bridge image at a
// time and a parser pin is honest about that).
import type { LoginEvent, ParsableEvent } from './types';
import { parseEventLegacyV076 } from './dialects/legacy_v076';
export type { ParsableEvent };
export const parseEvent = (event: ParsableEvent): LoginEvent => parseEventLegacyV076(event);

View file

@ -0,0 +1,130 @@
// LoginEvent — discriminated union the parser emits and the state machine
// consumes. One LoginEvent per inbound m.room.message / m.room.redaction
// from the bridge bot.
//
// Source-of-truth for every kind below is mautrix/discord legacy command
// system (commands.go), tag v0.7.6 — see ./dialects/legacy_v076.ts for the
// per-string upstream pointers. Discord uses the OLDER mautrix command
// processor (`maunium.net/go/mautrix/bridge/commands`), NOT bridgev2 — so
// the wording differs from mautrix-telegram and there's no list-logins
// API; status is queried via `ping`, and there's no per-account login id.
// `ping` reply variants — the bridge's only status-probe surface for the
// legacy command system. Each variant maps to a different LoginEvent so
// the state machine can render distinct status pills.
export type PingResult =
| { kind: 'not_logged_in' }
| { kind: 'token_stored_not_connected' }
| { kind: 'connection_dead' }
| { kind: 'logged_in'; handle: string; discordId?: string };
// Shape of an inbound event the dialect parser needs to look at. Matches
// the wire shape produced by the host's BotWidgetDriver sanitizer; declared
// here (not in widget-api.ts) so the dialect doesn't import from the
// transport layer.
export type ParsableEvent = {
type: string;
event_id: string;
sender: string;
origin_server_ts?: number;
content: { msgtype?: string; body?: string; [k: string]: unknown };
redacts?: string;
};
export type LoginEvent =
// --- ping reply ----------------------------------------------------------
| { kind: 'not_logged_in' }
| { kind: 'token_stored_not_connected' }
| { kind: 'connection_dead' }
// `ping` says we're live; handle parsed from `You're logged in as @x (`id`)`.
// Same shape as `login_success` — App routes both into the connected state.
| { kind: 'logged_in'; handle: string; discordId?: string }
// --- login-qr lifecycle --------------------------------------------------
// `m.image` carrying the remoteauth URL inside `content.body`. The widget
// renders the QR client-side from that URL and never touches the uploaded
// PNG. Discord's remoteauth does NOT rotate the URL (unlike Telegram
// MTProto): the bridge sends one m.image per login attempt and either
// redacts it on success or leaves it (and replies with an error) on
// failure. `replacesEventId` is here for forward-compat / paranoia — if a
// future bridge ever does an edit, the state machine handles it gracefully.
| { kind: 'qr_displayed'; discordUrl: string; eventId: string; replacesEventId?: string }
// Bridge redacted the QR event after a successful scan. NOT terminal — the
// success line («Successfully logged in as @x») typically lands in the same
// breath; the state machine moves us into a `qr_verifying` interstitial
// until it does.
| { kind: 'qr_redacted'; redactsEventId: string }
// Successful login (after QR scan). Captures handle and optional snowflake.
| { kind: 'login_success'; handle: string; discordId?: string }
// Generic login failure (wraps gotd / remoteauth Go errors). Most common
// bodies: «Error logging in: ...» — we surface the verbatim Go-error tail
// as a yellow warning on the QR panel.
| { kind: 'login_failed'; reason?: string }
// Vojo-patched bridge replies with a sentinel-prefixed JSON line
// («VOJO-CAPTCHA-CHALLENGE-V1 {...}») when Discord demands an hCaptcha
// before completing remote-auth. The widget renders a hCaptcha iframe
// with the supplied sitekey + rqdata, then sends the solved token back
// via `<commandPrefix> login-captcha <token>` to retry the login.
| {
kind: 'captcha_challenge';
service: string;
sitekey: string;
sessionId: string;
rqdata: string;
rqtoken: string;
}
// Legacy «CAPTCHAs are currently not supported - use token login instead»
// path — only fires against an UNPATCHED upstream v0.7.6 bridge. Kept so
// a deployment that forgets to ship the Vojo bridge image still surfaces
// a useful hint instead of a generic Go-error tail.
| { kind: 'captcha_required' }
// bridge sets up a websocket against Discord's remoteauth gateway; this is
// the «we couldn't even reach Discord» error — different from
// login_failed, which lands AFTER the websocket is up.
| { kind: 'login_websocket_failed'; reason?: string }
// Surfaces when QR-login starts but the bridge is already logged in.
// Race against ping/status — the App fires `ping` to reconcile.
| { kind: 'already_logged_in' }
// bridge couldn't initialise remoteauth at all (rare, indicates bridge-
// image misconfiguration). Routes back to disconnected with a warn line.
| { kind: 'prepare_login_failed'; reason?: string }
// bridge received the token but couldn't connect to Discord with it (rare
// post-scan failure; remoteauth can return a stale token if the gateway
// race trips). Surfaces as «Signed in, but couldn't connect: <reason>»
// and routes back to disconnected.
| { kind: 'connect_after_login_failed'; reason?: string }
// --- logout --------------------------------------------------------------
| { kind: 'logout_ok' }
// Bridge says the user wasn't logged in but cleared state defensively.
// Idempotent confirmation that we're now disconnected.
| { kind: 'logout_no_op' }
// --- disconnect / reconnect ---------------------------------------------
// Used as recovery from `connection_dead` / `token_stored_not_connected`.
// The widget never SENDS `disconnect` — that's an admin-only state op —
// but if the user typed it manually in chat-fallback, the parser still
// recognises the reply.
| { kind: 'disconnect_ok' }
| { kind: 'disconnect_no_op' }
| { kind: 'disconnect_failed'; reason?: string }
| { kind: 'reconnect_ok' }
| { kind: 'reconnect_no_op' }
| { kind: 'reconnect_failed'; reason?: string }
// --- Vojo: bridge-managed personal space ---------------------------------
// Vojo-patched bridge emits the sentinel `VOJO-LOGIN-SPACE-V1 {...}` as a
// separate m.notice right after the «Successfully logged in» line. Carries
// a `matrix.to` URL pointing at the user's auto-created Discord space
// (user.go::GetSpaceRoom on the bridge side). The widget surfaces this as
// an «Open in Channels» card; click → host navigates cinny to the space.
// See vojo-mautrix-discord/commands_login_space.go for the wire format.
| { kind: 'space_ready'; matrixToUrl: string }
// --- bridge-side errors --------------------------------------------------
// Generic «I don't know that command» — should not happen since we only
// ship known commands, but visible if the bridge image is misconfigured
// or the prefix in /config.json drifted from the bridge's command_prefix.
| { kind: 'unknown_command' }
| { kind: 'unknown' };

View file

@ -0,0 +1,89 @@
// English fallback. Mirror the RU key set; `Record<StringKey, string>` enforces
// that every RU key has an EN counterpart at compile time.
import type { StringKey } from './ru';
export const EN: Record<StringKey, string> = {
'status.unknown': 'Checking status…',
'status.disconnected': 'Discord not linked',
'status.connected': 'Discord linked',
'status.connected-as': 'Discord linked as {handle}',
'status.connection-dead': 'Discord connection lost',
'status.token-stored': 'Discord session is not active',
'status.qr-verifying': 'Verifying sign-in…',
'status.logging-out': 'Signing out…',
'status.reconnecting': 'Reconnecting to Discord…',
'card.login-qr.name': 'Sign in with QR code',
'card.login-qr.desc': 'Scan a QR code from the Discord mobile app',
'card.refresh.aria': 'Refresh status',
'card.refresh.label': 'Refresh status',
'card.refresh.name': 'Refresh status',
'card.refresh.desc': 'Re-check whether Discord is linked',
'card.refresh.in-flight': 'Checking…',
'card.about.name': 'How the Discord bot works',
'card.about.desc': 'Sign-in, safety, and source code',
'about.title': 'About the Discord bot',
'about.body-1':
'This bot connects Discord to Vojo. After sign-in, your DMs and servers from Discord will appear in Vojos chat list, and replies from the Vojo app will be sent to your contacts as normal Discord messages.',
'about.body-2':
'Sign-in requires the Discord mobile app — scan the QR code via Settings → Scan QR Code. Desktop Discord and the browser cannot be used: the bridge uses Discords “remoteauth” mechanism, available only in the mobile app.',
'about.body-3':
'The connection runs through the open-source mautrix-discord bridge. It creates a Discord session on the Vojo server and uses it to connect Discord with your Vojo account: receive messages from Discord and send your replies back.',
'about.github-label': 'The bridge source code is public on GitHub:',
'about.github-url': 'https://github.com/mautrix/discord',
'about.body-4':
'You can revoke access at any time — either with the “Sign out of Discord” button here, or inside Discord itself under Settings → Devices → Log out of Vojo.',
'about.close': 'Close',
'about.aria-close': 'Close “About this bot”',
'auth-card.qr.title': 'QR code sign-in',
'auth-card.qr.hint': 'Open the Discord mobile app and scan this QR code.',
'auth-card.qr.preparing': 'Preparing QR code…',
'auth-card.qr.aria': 'QR code for Discord sign-in. Scan it with your phone.',
'auth-card.qr.countdown': 'Time left to scan: {minutes}:{seconds}',
'auth-card.qr.expired': 'Sign-in window expired. Tap Cancel and try again.',
'auth-card.qr.step-1': 'Open the Discord mobile app.',
'auth-card.qr.step-2': 'Open Settings → Scan QR Code.',
'auth-card.qr.step-3': 'Scan the QR and confirm sign-in on your phone.',
'auth-card.captcha.title': 'Confirm youre not a robot',
'auth-card.captcha.hint':
'Discord asked for a CAPTCHA. Solve it below — sign-in will continue automatically once youre done.',
'auth-card.captcha.load-error':
'Could not load the CAPTCHA. Check your network, tap Cancel and try signing in again.',
'auth-card.cancel': 'Cancel',
'auth-card.waiting-hint': 'The bot is still thinking… replies may take up to 30 seconds.',
'auth-error.captcha-required':
'Discord requested a CAPTCHA — QR sign-in is temporarily unavailable. Try again later, or sign in with a token via the bots chat.',
'auth-error.captcha-send-failed':
'Could not deliver your CAPTCHA solution. Check your network and try signing in again.',
'auth-error.captcha-expired': 'CAPTCHA expired — tap «Sign in with QR code» and solve it again.',
'auth-error.login-failed': 'Sign-in failed: {reason}',
'auth-error.prepare-failed': 'Failed to prepare sign-in: {reason}',
'auth-error.websocket-failed': 'Could not connect to the sign-in server: {reason}',
'auth-error.connect-after-login-failed': 'Signed in, but could not connect to Discord: {reason}',
'auth-error.already-logged-in': 'You are already signed in to Discord — refresh status.',
'auth-error.unknown-command':
'The bot does not recognise this command — check the prefix in config.json.',
'auth-error.disconnect-failed': 'Disconnect failed: {reason}',
'card.reconnect.name': 'Reconnect',
'card.reconnect.desc': 'Restore the Discord connection without signing in again',
'card.logout.name': 'Sign out of Discord',
'card.logout.desc': 'End the session for this account',
'card.logout.confirm-prompt': 'Sign out for real?',
'card.logout.confirm-yes': 'Sign out',
'card.logout.confirm-no': 'Cancel',
'card.open-space.name': 'Open in Channels',
'card.open-space.desc': 'Jump to your Discord space with all chats and servers',
'diag.space-ready': 'Discord space ready to open.',
'diag.connecting': 'Connecting to Vojo… awaiting capability handshake.',
'diag.ready': 'Ready to send commands.',
'diag.checking-status': 'Checking connection status…',
'diag.send-failed': 'send failed: {message}',
'diag.history-marker': '─── history ───',
'diag.history-unavailable': 'Could not read history — re-checking status.',
'diag.qr-issued': 'QR code issued.',
'diag.qr-consumed': 'QR code consumed — bridge confirmed the scan.',
'diag.captcha-issued': 'Discord requested a CAPTCHA — solve it in the form above.',
'bootstrap.failed': 'Widget failed to start',
'bootstrap.missing-params': 'Missing required URL params: {names}.',
'bootstrap.embedded-only': 'This page is meant to be embedded by Vojo at {route}.',
};

View file

@ -0,0 +1,34 @@
// Tiny i18n harness — Russian primary, English fallback (BCP-47 prefix
// match — any `en` variant). Bootstrap forwards `clientLanguage` from the
// host; main.tsx can also call `createT()` without args before bootstrap
// completes (falls back to navigator.language, then RU).
//
// Identical mechanics to apps/widget-telegram/src/i18n/index.ts; the
// Discord widget keeps its own dictionary file because the copy differs —
// QR-only flow, no SMS, no 2FA password form.
import { RU, type StringKey } from './ru';
import { EN } from './en';
const interpolate = (s: string, vars?: Record<string, string>): string => {
if (!vars) return s;
return s.replace(/\{(\w+)\}/g, (_, k) => vars[k] ?? `{${k}}`);
};
const pickDict = (clientLanguage: string | undefined): Record<StringKey, string> => {
const lang = (
clientLanguage ||
(typeof navigator !== 'undefined' ? navigator.language : '') ||
'ru'
).toLowerCase();
return lang.startsWith('en') ? EN : RU;
};
export type T = (key: StringKey, vars?: Record<string, string>) => string;
export const createT = (clientLanguage?: string): T => {
const dict = pickDict(clientLanguage);
return (key, vars) => interpolate(dict[key], vars);
};
export type { StringKey };

View file

@ -0,0 +1,132 @@
// Russian primary copy. To add a string:
// 1. add the key + RU value here (this file is the canonical key list — `en.ts`
// and the `StringKey` type derive from it),
// 2. add the same key + EN value in `en.ts`,
// 3. consume via `t('key', { var: 'x' })` in components.
// Interpolation uses `{name}` placeholders resolved against the second arg.
//
// The widget no longer renders a hero (avatar/name/handle/description) —
// that block lives in the host's BotShellHero. Status is surfaced inline
// inside the relevant section.
export const RU = {
// --- Inline section status ---------------------------------------------
'status.unknown': 'Проверка статуса…',
'status.disconnected': 'Discord не привязан',
'status.connected': 'Discord привязан',
'status.connected-as': 'Discord привязан как {handle}',
'status.connection-dead': 'Соединение с Discord потеряно',
'status.token-stored': 'Сессия Discord не активна',
'status.qr-verifying': 'Проверяем вход…',
'status.logging-out': 'Завершение сеанса…',
'status.reconnecting': 'Переподключаюсь к Discord…',
// --- Section headers ---------------------------------------------------
'card.login-qr.name': 'Войти по QR-коду',
// Discord QR требует МОБИЛЬНОЕ приложение Discord (legacy remoteauth
// не работает с десктопным клиентом) — это важная подсказка, чтобы у
// пользователя без мобильного клиента не возникло тупика «попробовал
// и не работает».
'card.login-qr.desc': 'Отсканировать QR из мобильного приложения Discord',
'card.refresh.aria': 'Обновить статус',
'card.refresh.label': 'Обновить статус',
'card.refresh.name': 'Обновить статус',
'card.refresh.desc': 'Перепроверить, привязан ли Discord',
'card.refresh.in-flight': 'Проверяю…',
// --- About panel -------------------------------------------------------
'card.about.name': 'Как работает Discord-бот',
'card.about.desc': 'Вход, безопасность и исходный код',
'about.title': 'О боте Discord',
'about.body-1':
'Этот бот подключает Discord к Vojo. После входа личные чаты и серверы Discord появятся в списке чатов Vojo, а ответы из приложения Vojo будут отправляться собеседникам как обычные сообщения в Discord.',
'about.body-2':
'Для входа нужен мобильный клиент Discord — отсканируйте QR-код через «Настройки → Сканировать QR-код». Десктопный Discord или браузер для входа не подходят: используется механизм remoteauth, который доступен только в мобильном приложении.',
'about.body-3':
'Подключение работает через open-source мост mautrix-discord. Он создаёт Discord-сессию на сервере Vojo и использует её для связи Discord с вашим аккаунтом Vojo: получает сообщения из Discord и отправляет ваши ответы обратно.',
'about.github-label': 'Исходный код моста открыт на GitHub:',
'about.github-url': 'https://github.com/mautrix/discord',
'about.body-4':
'Отозвать доступ можно в любой момент — кнопкой «Выйти из Discord» здесь, либо в самом Discord через «Настройки → Устройства → Выйти из Vojo».',
'about.close': 'Закрыть',
'about.aria-close': 'Закрыть «О боте»',
// --- QR form -----------------------------------------------------------
// Discord QR не ротируется в отличие от Telegram MTProto — мост держит
// одну сессию remoteauth до успеха, ошибки или таймаута. Поэтому в
// тексте говорим про «отсканируйте этот QR-код», без указаний на
// обновление, и таймаут показываем «всего окна» одной строкой.
'auth-card.qr.title': 'Вход по QR-коду',
'auth-card.qr.hint': 'Откройте мобильный Discord и отсканируйте этот QR-код.',
'auth-card.qr.preparing': 'Готовим QR-код…',
'auth-card.qr.aria': 'QR-код для входа в Discord. Отсканируйте его телефоном.',
'auth-card.qr.countdown': 'На сканирование осталось {minutes}:{seconds}',
'auth-card.qr.expired': 'Окно входа истекло. Нажмите «Отмена» и попробуйте снова.',
'auth-card.qr.step-1': 'Откройте мобильное приложение Discord.',
'auth-card.qr.step-2': 'Откройте «Настройки → Сканировать QR-код».',
'auth-card.qr.step-3': 'Отсканируйте QR-код и подтвердите вход на телефоне.',
// --- hCaptcha challenge -----------------------------------------------
// Discord иногда требует решить hCaptcha перед завершением remoteauth —
// anti-abuse-сигнал, не зависящий от конкретного юзера. Vojo-патч
// отдаёт челлендж сюда в виджет; пользователь решает, токен уходит
// обратно на мост и логин завершается. Текст не упоминает «бан»: это
// обычный механизм Discord, а не санкция.
'auth-card.captcha.title': 'Подтвердите, что вы не робот',
'auth-card.captcha.hint':
'Discord попросил решить капчу. Решите её ниже — после этого вход продолжится автоматически.',
'auth-card.captcha.load-error':
'Не удалось загрузить капчу. Проверьте сеть и нажмите «Отмена», затем войдите снова.',
// --- Shared form chrome ------------------------------------------------
// Cancel в Discord-flow ЛОКАЛЬНЫЙ: legacy-мост не имеет команды отмены
// активного login-qr, поэтому кнопка просто возвращает виджет в
// disconnected, а серверная сторона сама истекает по таймауту remoteauth
// (~2 минуты по умолчанию). Это написано в about.body-2, и пользователь,
// увидев «Окно входа истекло», понимает, что стало с QR.
'auth-card.cancel': 'Отмена',
'auth-card.waiting-hint': 'Бот ещё думает… ответ может идти до 30 секунд.',
// --- Inline errors -----------------------------------------------------
'auth-error.captcha-required':
'Discord потребовал CAPTCHA — вход через QR временно недоступен. Попробуйте позже или войдите через токен в чате с ботом.',
'auth-error.captcha-send-failed':
'Не удалось отправить ответ на CAPTCHA. Проверьте сеть и попробуйте войти заново.',
'auth-error.captcha-expired': 'CAPTCHA устарела — нажмите «Войти по QR-коду» и решите её заново.',
'auth-error.login-failed': 'Не удалось войти: {reason}',
'auth-error.prepare-failed': 'Не удалось подготовить вход: {reason}',
'auth-error.websocket-failed': 'Не удалось подключиться к серверу входа: {reason}',
'auth-error.connect-after-login-failed':
'Вход прошёл, но соединиться с Discord не получилось: {reason}',
'auth-error.already-logged-in': 'Вы уже вошли в Discord — обновите статус.',
'auth-error.unknown-command': 'Бот не знает эту команду — проверьте префикс в config.json.',
'auth-error.disconnect-failed': 'Не удалось отключиться: {reason}',
// --- Logout / Reconnect ------------------------------------------------
// Reconnect-action нужен только в connection_dead / token_stored —
// здоровая сессия не показывает кнопку. Текст глагольный, без префиксов.
'card.reconnect.name': 'Переподключиться',
'card.reconnect.desc': 'Восстановить соединение с Discord без повторного входа',
'card.logout.name': 'Выйти из Discord',
'card.logout.desc': 'Завершить сеанс на этом аккаунте',
'card.logout.confirm-prompt': 'Точно выйти?',
'card.logout.confirm-yes': 'Выйти',
'card.logout.confirm-no': 'Отмена',
// --- Open Discord space (Vojo bridge sentinel) ------------------------
'card.open-space.name': 'Открыть в Каналах',
'card.open-space.desc': 'Перейти в спейс Discord со списком чатов и серверов',
// --- Diagnostics in transcript ----------------------------------------
'diag.space-ready': 'Discord-спейс готов к открытию.',
'diag.connecting': 'Соединение с Vojo… ожидаем capability handshake.',
'diag.ready': 'Готов отправлять команды.',
'diag.checking-status': 'Проверяю статус подключения…',
'diag.send-failed': 'ошибка отправки: {message}',
'diag.history-marker': '─── история ───',
'diag.history-unavailable': 'Не удалось прочитать историю — проверяю статус заново.',
// QR-сообщения никогда не выводятся целиком в transcript — body содержит
// токен `https://discord.com/ra/…`, который мост стирает после скана;
// сохранять его в DOM-логе виджета означало бы пережить эту защиту.
// Поэтому в логе только нейтральные диагностические строки.
'diag.qr-issued': 'QR-код выдан.',
'diag.qr-consumed': 'QR-код использован — мост подтверждает скан.',
'diag.captcha-issued': 'Discord прислал CAPTCHA — решите её на форме выше.',
// --- Bootstrap failure -------------------------------------------------
'bootstrap.failed': 'Widget не запустился',
'bootstrap.missing-params': 'Отсутствуют обязательные параметры URL: {names}.',
'bootstrap.embedded-only': 'Эта страница предназначена для встраивания Vojo по маршруту {route}.',
} as const;
export type StringKey = keyof typeof RU;

View file

@ -0,0 +1,62 @@
import { render } from 'preact';
import { readBootstrap } from './bootstrap';
import { App } from './App';
import { createT } from './i18n';
import { WidgetApi, buildCapabilities } from './widget-api';
import './styles.css';
// Input-mode detector — see apps/widget-telegram/src/main.tsx for the
// full rationale. Default to 'mouse'; the capture-phase pointerdown
// listener flips to 'touch' on the first non-mouse pointerType.
// matchMedia guessing was dropped — every variant
// (`any-pointer: coarse|fine`, `hover: hover`, `pointer: fine|coarse`)
// is mis-reported on at least one shipping device.
const setInputMode = (mode: 'touch' | 'mouse'): void => {
document.documentElement.dataset.input = mode;
};
setInputMode('mouse');
window.addEventListener(
'pointerdown',
(event) => {
setInputMode(event.pointerType === 'mouse' ? 'mouse' : 'touch');
},
{ passive: true, capture: true }
);
const root = document.getElementById('app');
if (!root) {
throw new Error('#app root element missing — index.html out of sync');
}
const result = readBootstrap(window.location.search);
if (!result.ok) {
// Either someone opened the widget URL directly (no host params), or a
// host bug failed to provide them. Render a self-contained diagnostic
// instead of going silent. Bootstrap failed before we could read
// clientLanguage, so let createT fall back to navigator.language.
const t = createT();
render(
<div class="app">
<div class="error-banner">
<strong>{t('bootstrap.failed')}</strong>
{t('bootstrap.missing-params', { names: result.missing.join(', ') })}{' '}
{t('bootstrap.embedded-only', { route: '/bots/discord' })}
</div>
</div>,
root
);
} else {
// Apply initial theme synchronously so the first paint isn't flashed
// through the wrong palette.
document.documentElement.dataset.theme = result.bootstrap.theme;
// Instantiate WidgetApi BEFORE React render. The constructor attaches
// the message listener synchronously, so by the time the host's
// ClientWidgetApi fires its capabilities request on iframe `load`,
// we're already listening. Constructing inside a useEffect would race
// with the cached-bundle remount path. See widget-telegram for full
// rationale.
const api = new WidgetApi(result.bootstrap, buildCapabilities(result.bootstrap.roomId));
render(<App bootstrap={result.bootstrap} api={api} />, root);
}

View file

@ -0,0 +1,939 @@
// Login state machine — consumes LoginEvent (one per inbound bridge bot
// reply) and emits a typed UI state. The widget renders the QR panel and
// the status pill from this state, never from raw reply strings.
//
// Discord vs Telegram differences:
// - QR-only: there's no phone/code/password ladder, so the state space
// is much smaller than the Telegram reducer.
// - status comes from `ping` (legacy mautrix command system), not
// `list-logins` (bridgev2). The four ping replies map to four states:
// disconnected / connected / connection_dead / token_stored.
// - no list-logins-derived `loginId`; logout is the bare `logout` verb,
// so the connected state doesn't need to gate on a login id.
// - the QR is NOT rotated by Discord remoteauth (single image per
// login attempt). The state machine still tracks `qrEventId` so the
// redaction handler can match against it and ignore unrelated cleanup.
//
// State-gating policy: late-arriving replies from cancelled flows must
// not resurrect dead state. The `cancel_pending` action ALWAYS lands us
// in `disconnected` immediately; later bridge events arriving after
// cancel are filtered by the live reducer.
import type { LoginEvent } from './bridge-protocol/types';
export type LoginErrorFlag =
| { kind: 'login_failed'; reason?: string }
| { kind: 'captcha_required' }
// Local rollback flag — fired when the widget couldn't deliver
// `login-captcha <token>` to the bridge (transport / Matrix-API failure
// before the bridge even saw the command). Distinct from
// `login_failed` (which IS a bridge reply) so the UX can read «sign-in
// didn't reach the bot, retry» instead of hinting at a Discord-side
// problem the user can't act on.
| { kind: 'captcha_send_failed' }
// The captcha challenge timed out (rqtoken expiry on Discord's side)
// before the user solved it. Surfaced as a soft warning to retry
// login-qr from scratch.
| { kind: 'captcha_expired' }
| { kind: 'login_websocket_failed'; reason?: string }
| { kind: 'connect_after_login_failed'; reason?: string }
| { kind: 'prepare_login_failed'; reason?: string }
| { kind: 'already_logged_in' }
| { kind: 'unknown_command' };
// `reconnect_failed` is intentionally NOT a LoginErrorFlag arm: the live
// reducer routes that event back to `connected_dead` (no error surface
// there — the connected-dead pill IS the error indicator) without
// staging a reason for `localizeError`. If a future UI change wants to
// surface the reason, add `lastError?: ...` to the connected_dead state
// shape and route `reconnect_failed` through it.
// A live form is open and waiting for user action. M-discord ships with
// only one: the QR panel. Hydrate's restorable shape collapses to this
// single variant + the `qr_verifying` interstitial.
export type PendingFormState = {
kind: 'awaiting_qr_scan';
discordUrl: string;
qrEventId: string;
firstShownAt: number;
lastError?: LoginErrorFlag;
};
// hCaptcha challenge surfaced by the Vojo-patched bridge after Discord
// returned 400+captcha-required. The widget renders the hCaptcha iframe
// from `sitekey` + `rqdata`; on solve, the App sends `login-captcha
// <token>` and we transition to `qr_verifying` until the bridge replies.
export type CaptchaSolveState = {
kind: 'awaiting_captcha_solve';
service: string;
sitekey: string;
sessionId: string;
rqdata: string;
rqtoken: string;
firstShownAt: number;
};
export type LoginState =
// Pre-handshake / pre-ping. Status pill: --faint.
| { kind: 'unknown' }
// ping returned `not_logged_in`, OR logout completed. Status pill:
// --rose. The card grid offers the QR-login affordance.
| { kind: 'disconnected'; lastError?: LoginErrorFlag }
// QR-login in progress. Optimistically transitioned by `start_qr_login`;
// overwritten with real discordUrl/qrEventId by the live `qr_displayed`
// event. Status pill: --amber.
| PendingFormState
// hCaptcha challenge from Discord — Vojo-patched bridge surfaced the
// sitekey + rqdata via `VOJO-CAPTCHA-CHALLENGE-V1` notice. The widget
// renders the hCaptcha iframe; on solve we send `login-captcha <token>`
// and transition to `qr_verifying`. Status pill: --amber.
| CaptchaSolveState
// QR was redacted (i.e. the bridge accepted a scan), but we don't yet
// know whether login succeeded. Held as an intermediate spinner until
// the next bridge signal arrives. Status pill: --amber.
| { kind: 'qr_verifying' }
// logout in flight — waiting for `Logged out successfully.`. Status
// pill: --amber.
| { kind: 'logging_out' }
// reconnect in flight (recovery from connection_dead / token_stored).
// Waiting for `Successfully reconnected` or `You're already connected`.
// Status pill: --amber. `handle` is carried through from the
// connected_dead state so a successful reconnect can flip directly to
// `connected{handle}` without bouncing through a transient `unknown`
// (which would briefly paint a faint «Проверка статуса…» pill — bad
// UX immediately after the user took an action).
| { kind: 'reconnecting'; handle?: string }
// Live session — ping or login_success confirmed. Discord legacy bridge
// doesn't have a per-account loginId concept (single Discord account
// per Matrix user), so logout doesn't need an id. `spaceMatrixToUrl`
// is populated from the Vojo `VOJO-LOGIN-SPACE-V1` sentinel that lands
// right after login_success; it survives the post-login re-ping and the
// reconnect-ok transitions so the «Open in Channels» card stays visible
// until logout. Absent until the sentinel arrives (and absent forever
// against an UNPATCHED bridge — the card simply never appears).
| { kind: 'connected'; handle: string; discordId?: string; spaceMatrixToUrl?: string }
// ping says we have a token but the connection's down. Status pill:
// green-ish but with a Reconnect recovery action exposed. The reducer
// distinguishes `connection_dead` (Discord WS dropped) from `token_stored`
// (we have the token but never got far enough to connect), but the UI
// collapses both into the same shape — they share the recovery path.
| { kind: 'connected_dead'; reason: 'connection_dead' | 'token_stored'; handle?: string };
// States that the hydrate path can restore after a reload. The QR panel
// (`awaiting_qr_scan`) survives reloads via the m.image / m.room.redaction
// timeline; `qr_verifying` covers the post-scan pre-success interstitial;
// `awaiting_captcha_solve` covers the case where the user reloads while
// staring at an hCaptcha challenge (rqdata/rqtoken are short-lived but
// often valid for a couple of minutes — fresh enough to reuse). Other
// transient states (logging_out, reconnecting) deliberately don't survive.
export type HydrateRestoredState = PendingFormState | CaptchaSolveState | { kind: 'qr_verifying' };
// Outbound user actions the App dispatches. Form-submit actions clear any
// pending lastError; structural transitions optimistically advance state —
// the App rolls them back on send-failure where the bot would otherwise
// leave us stuck.
export type LoginAction =
| { kind: 'event'; event: LoginEvent }
| { kind: 'start_qr_login' } // user clicked «Войти по QR»
| { kind: 'request_logout' } // user clicked «Выйти из Discord»
// user clicked «Переподключиться» — App passes the current handle
// (from `connected_dead.handle` or `connected.handle`) so the
// transient `reconnecting` state carries it forward; without this the
// post-reconnect_ok branch can't paint the connected pill until the
// follow-up ping resolves.
| { kind: 'request_reconnect'; handle?: string }
// Discord legacy mautrix has no `cancel` command. Cancel is LOCAL —
// returns the widget to disconnected immediately; the bridge's
// remoteauth websocket eventually times out on its own. The action is
// kept symmetrical with TG's reducer for shape consistency, but
// dispatching it doesn't trigger any send.
| { kind: 'cancel_pending' }
// User finished an hCaptcha challenge — token is non-empty. Optimistic
// transition to `qr_verifying`; the App fires `login-captcha <token>`
// and the bridge's reply (`login_success` / chained `captcha_challenge`
// / `login_failed`) lands via the live event stream.
| { kind: 'submit_captcha_token' }
// Rollback for `submit_captcha_token` — the App couldn't deliver the
// command to the bridge. Routes back to disconnected with a localized
// error so the user sees what happened.
| { kind: 'captcha_send_failed' }
// hCaptcha challenge expired (server rqtoken TTL or local 90s timer).
// Routes to disconnected with a localized warn.
| { kind: 'captcha_expired' }
| { kind: 'hydrate'; state: HydrateRestoredState };
export const initialLoginState: LoginState = { kind: 'unknown' };
const isFormState = (s: LoginState): s is PendingFormState => s.kind === 'awaiting_qr_scan';
// States that a fresh captcha challenge can clobber: the QR scan has
// landed (or is mid-flight), or we're already showing a previous
// challenge that Discord chained on top.
const isCaptchaAcceptingState = (
s: LoginState
): s is PendingFormState | { kind: 'qr_verifying' } | CaptchaSolveState =>
s.kind === 'awaiting_qr_scan' || s.kind === 'qr_verifying' || s.kind === 'awaiting_captcha_solve';
export const loginReducer = (state: LoginState, action: LoginAction): LoginState => {
if (action.kind === 'hydrate') {
// hydrate is a one-shot mount-time seed. If a live event already
// moved us off `unknown`, the live truth wins; the cached timeline
// snapshot is by definition older than what the live event just told
// us. Without this gate, a stale `awaiting_qr_scan` from a previous
// session could overwrite a legitimate `connected` that arrived
// during the readTimeline await.
if (state.kind !== 'unknown') return state;
return action.state;
}
if (action.kind === 'start_qr_login') {
// Optimistic placeholder; the live `qr_displayed` event overwrites
// discordUrl + qrEventId + firstShownAt. If the `!discord login-qr`
// send fails, the App rolls back to `disconnected`.
return {
kind: 'awaiting_qr_scan',
discordUrl: '',
qrEventId: '',
firstShownAt: Date.now(),
};
}
if (action.kind === 'request_logout') {
return { kind: 'logging_out' };
}
if (action.kind === 'request_reconnect') {
return { kind: 'reconnecting', handle: action.handle };
}
if (action.kind === 'cancel_pending') {
// Optimistic: drop straight back to disconnected. Discord legacy mautrix
// has no `cancel` command — the bridge's remoteauth websocket continues
// until it succeeds or times out internally. From the user's POV the
// widget returns to disconnected, and any later QR redaction / login
// success / login failure event from the abandoned flow is filtered
// by the per-event gates below (qr_redacted gated on awaiting_qr_scan,
// login_success / login_failed gated on awaiting_qr_scan|qr_verifying).
return { kind: 'disconnected' };
}
if (action.kind === 'submit_captcha_token') {
// User solved hCaptcha and we're about to fire login-captcha. Hold
// `qr_verifying` until the bridge replies (success / chained challenge
// / generic login_failed). Only honour from the captcha state — this
// action is App-emitted right after the hCaptcha callback, so it
// shouldn't ever fire from anywhere else, but defensive: a stale send
// shouldn't suddenly paint a verifying spinner over a connected pill.
if (state.kind !== 'awaiting_captcha_solve') return state;
return { kind: 'qr_verifying' };
}
if (action.kind === 'captcha_send_failed') {
// App couldn't ship the `login-captcha` command (transport / Matrix
// API failure before the bridge saw it). Roll the optimistic
// `qr_verifying` back to disconnected with a localized error.
//
// Honour ONLY from `qr_verifying` — narrowed from also accepting
// `awaiting_captcha_solve` to avoid clobbering a fresh chained
// captcha that may have arrived between the optimistic dispatch and
// the failed-send rollback. If the live state already moved to a
// newer challenge, the stale send-failure should be silently dropped.
if (state.kind !== 'qr_verifying') return state;
return { kind: 'disconnected', lastError: { kind: 'captcha_send_failed' } };
}
if (action.kind === 'captcha_expired') {
// Local 90s timer or hCaptcha's expired-callback fired — the rqtoken
// is dead, the user has nothing to solve. Route to disconnected with
// a localized warn so they retry login-qr from scratch.
if (state.kind !== 'awaiting_captcha_solve') return state;
return { kind: 'disconnected', lastError: { kind: 'captcha_expired' } };
}
const event = action.event;
switch (event.kind) {
// --- ping replies ----------------------------------------------------
case 'not_logged_in':
// Accept from states where flipping to disconnected is correct.
// Late-arriving `not_logged_in` MUST NOT clobber an active QR-scan
// (which was started after the ping was fired but before the reply
// landed) — that's the same race the TG reducer guards against.
if (
state.kind === 'unknown' ||
state.kind === 'disconnected' ||
state.kind === 'logging_out' ||
state.kind === 'qr_verifying' ||
state.kind === 'reconnecting' ||
state.kind === 'connected_dead'
) {
return { kind: 'disconnected' };
}
return state;
case 'logged_in':
// Authoritative source — accept from any state. Used by both the
// initial ping AND the post-`login_success` re-ping that picks up
// the discordId snowflake. Preserve `spaceMatrixToUrl` from a prior
// `connected` so the post-login_success re-ping doesn't blank the
// CTA before the user gets a chance to click it.
return {
kind: 'connected',
handle: event.handle,
discordId: event.discordId,
spaceMatrixToUrl: state.kind === 'connected' ? state.spaceMatrixToUrl : undefined,
};
case 'connection_dead':
// ping says token's good but the WS is down. Show the connected
// chrome with a Reconnect recovery action.
return {
kind: 'connected_dead',
reason: 'connection_dead',
handle: state.kind === 'connected' ? state.handle : undefined,
};
case 'token_stored_not_connected':
return {
kind: 'connected_dead',
reason: 'token_stored',
handle: state.kind === 'connected' ? state.handle : undefined,
};
// --- QR lifecycle ----------------------------------------------------
case 'qr_displayed': {
// Defence-in-depth: an inbound qr_displayed MUST carry a non-empty
// event id (the host driver rejects empty event_id at the sanitizer;
// this is a redundant guard).
if (event.eventId.length === 0) return state;
// Initial QR from a fresh login attempt — accept from:
// * `unknown` — cold-start before ping resolves;
// * placeholder `awaiting_qr_scan{qrEventId=''}` from start_qr_login.
//
// We DO NOT accept from `disconnected`. Discord legacy mautrix has
// no cancel command, so when the user clicks Cancel locally the
// bridge's remoteauth goroutine continues until success / failure
// / internal timeout. The widget transitions to `disconnected`
// immediately, but the bridge eventually emits the m.image. If we
// accepted that here, the user would see a QR they didn't ask for
// — the bridge has no way to know the user moved on. Drop it
// silently; the user has to click «Войти по QR» again to express
// intent (which resets the placeholder and lets the next m.image
// land).
if (
state.kind === 'unknown' ||
(state.kind === 'awaiting_qr_scan' && state.qrEventId === '')
) {
return {
kind: 'awaiting_qr_scan',
discordUrl: event.discordUrl,
qrEventId: event.eventId,
firstShownAt:
state.kind === 'awaiting_qr_scan' && state.firstShownAt
? state.firstShownAt
: Date.now(),
};
}
if (state.kind !== 'awaiting_qr_scan') return state;
// Hypothetical edit pointing at our anchor — repaint URL, keep id.
// Discord doesn't currently edit QRs but the path stays for
// forward-compat (cheaper to keep than to reconstruct).
if (event.replacesEventId === state.qrEventId) {
return { ...state, discordUrl: event.discordUrl };
}
// Fresh non-edit qr_displayed while we're already tracking one —
// could be a bridge-side restart (rare). Adopt as new anchor.
if (!event.replacesEventId) {
return {
kind: 'awaiting_qr_scan',
discordUrl: event.discordUrl,
qrEventId: event.eventId,
firstShownAt: Date.now(),
};
}
// Edit pointing at something we don't track — ignore.
return state;
}
case 'qr_redacted': {
// Bridge cleaned up the QR after a successful scan (commands.go
// l.197: `_, _ = ce.MainIntent().RedactEvent(ce.RoomID, qrCodeEvent)`
// — only fires on the success path). Held as `qr_verifying` until
// the success line lands. Only honour from awaiting_qr_scan with a
// matching event id.
if (state.kind !== 'awaiting_qr_scan') return state;
if (state.qrEventId !== event.redactsEventId) return state;
return { kind: 'qr_verifying' };
}
case 'login_success':
// Honour from any non-terminal state. The bridge's success line
// doesn't include the discordId; the App fires `ping` afterwards
// to upgrade to the full `connected{handle, discordId}` shape.
return { kind: 'connected', handle: event.handle };
case 'login_failed':
// Generic Discord-side login failure — bridge replies «Error logging
// in: <go-error>». Routes back to disconnected with the verbatim
// reason as a warn line. Only honour when a QR flow is in flight,
// OR while the user is solving a captcha (the Vojo-patched bridge
// can also reply «Error logging in: …» AFTER `login-captcha` if
// Discord rejects the post-solve replay). Otherwise it's stale
// (e.g. an old failure replaying after page reload while the user
// is already connected).
if (!isCaptchaAcceptingState(state)) return state;
return {
kind: 'disconnected',
lastError: { kind: 'login_failed', reason: event.reason },
};
case 'captcha_required':
// UNPATCHED bridge fallback: «CAPTCHAs are currently not supported».
// Surface as a hint suggesting token-login (chat-fallback only). On
// a Vojo-patched bridge this branch never fires — see captcha_challenge.
if (!isCaptchaAcceptingState(state)) return state;
return { kind: 'disconnected', lastError: { kind: 'captcha_required' } };
case 'captcha_challenge':
// Vojo-patched bridge surfaced an hCaptcha challenge — pivot the
// widget to the captcha screen. Accept from awaiting_qr_scan
// (challenge landed before the QR was redacted), qr_verifying
// (challenge landed while we were still in the post-redact spinner)
// or awaiting_captcha_solve (Discord chained another challenge after
// the previous solve). Other states drop the event silently — a
// stale challenge from an abandoned flow shouldn't repaint UI.
if (!isCaptchaAcceptingState(state)) return state;
return {
kind: 'awaiting_captcha_solve',
service: event.service,
sitekey: event.sitekey,
sessionId: event.sessionId,
rqdata: event.rqdata,
rqtoken: event.rqtoken,
firstShownAt: Date.now(),
};
case 'login_websocket_failed':
// Pre-QR failure: couldn't reach Discord remoteauth. The QR was
// never displayed in the first place. State `awaiting_qr_scan` with
// empty discordUrl is the placeholder set by `start_qr_login`;
// this fires before the first qr_displayed lands.
if (!isCaptchaAcceptingState(state)) return state;
return {
kind: 'disconnected',
lastError: { kind: 'login_websocket_failed', reason: event.reason },
};
case 'connect_after_login_failed':
// Post-scan rare: remoteauth gave us a token, but the bridge couldn't
// connect to Discord with it. The bridge has the token cached and
// might recover on next ping; we still route to disconnected so the
// user can retry.
if (!isCaptchaAcceptingState(state)) return state;
return {
kind: 'disconnected',
lastError: { kind: 'connect_after_login_failed', reason: event.reason },
};
case 'prepare_login_failed':
if (!isCaptchaAcceptingState(state)) return state;
return {
kind: 'disconnected',
lastError: { kind: 'prepare_login_failed', reason: event.reason },
};
case 'already_logged_in':
// The user clicked «Войти по QR» but the bridge is already logged
// in — race against ping. Surface a soft warning and let the App's
// re-ping reconcile to the connected state.
if (isFormState(state)) {
return { ...state, lastError: { kind: 'already_logged_in' } };
}
return state;
// --- logout ----------------------------------------------------------
case 'logout_ok':
case 'logout_no_op':
// Late `Logged out` from a previous session can arrive while the
// user is mid-new-flow. Only honour from logging_out; other states
// keep their flow.
if (state.kind !== 'logging_out') return state;
return { kind: 'disconnected' };
// --- disconnect (read-only, never sent by widget) -------------------
case 'disconnect_ok':
case 'disconnect_no_op':
// User typed `disconnect` manually in chat-fallback while the widget
// was open. Reflect the bridge's truth: no token-loss, but no live
// connection either — same shape as `token_stored`. Both
// `connected` (string handle) and `connected_dead` (handle?:
// string) expose `handle` on the same key, so a single read works.
if (state.kind === 'connected' || state.kind === 'connected_dead') {
return {
kind: 'connected_dead',
reason: 'token_stored',
handle: state.handle,
};
}
return state;
case 'disconnect_failed':
// Manual disconnect attempt failed — keep current state, the widget
// doesn't surface a UI for this since it never sent the command.
return state;
// --- reconnect -------------------------------------------------------
case 'reconnect_ok':
case 'reconnect_no_op':
// After a successful reconnect, ping is the source of truth for the
// handle. The App fires `ping` after this event lands to refresh.
// We flip to `connected` immediately so the user sees an immediate
// green pill confirming their click; the post-event ping refreshes
// the handle / discordId within ~100ms. Both `reconnecting` and
// `connected_dead` carry `handle?` — a missing handle still flips
// green with an empty handle, which the UI's
// `state.handle ? connected-as : connected` ternary tolerates.
// This avoids the `unknown` flap that the previous draft would
// produce when no handle was stashed. spaceMatrixToUrl is not
// restorable from connected_dead (the dead state never carried it),
// so the CTA stays hidden until a fresh sentinel arrives — bridge
// does NOT re-emit on reconnect, but the card returns once the user
// explicitly re-logs in.
if (state.kind === 'reconnecting' || state.kind === 'connected_dead') {
return { kind: 'connected', handle: state.handle ?? '' };
}
return state;
case 'space_ready':
// Vojo-patched bridge surfaced the personal Discord space — attach
// its matrix.to URL to the connected state so the «Open in Channels»
// card renders. Late-arriving sentinels from an abandoned flow drop
// here silently (e.g. a sentinel that lands during `logging_out`
// mustn't resurrect a connected state). Honour only from the
// canonical alive states.
if (state.kind === 'connected') {
return { ...state, spaceMatrixToUrl: event.matrixToUrl };
}
return state;
case 'reconnect_failed':
if (state.kind !== 'reconnecting') return state;
// Roll back to connected_dead carrying the previous handle. The
// user can hit Reconnect again or refresh. We don't surface the
// error reason here — the connected_dead pill itself reads as
// «something is wrong, try Reconnect» — adding a transient red
// banner adjacent to a recovery affordance is overkill.
return {
kind: 'connected_dead',
reason: 'connection_dead',
handle: state.handle,
};
// --- bridge-side errors ---------------------------------------------
case 'unknown_command':
// Shouldn't happen — we only send commands the bridge knows. Visible
// when /config.json's commandPrefix drifts from the bridge's actual
// command_prefix. Surface loudly on disconnected.
return { kind: 'disconnected', lastError: { kind: 'unknown_command' } };
case 'unknown':
return state;
default: {
// Exhaustiveness check — TS flags this if a new LoginEvent kind is
// added without a case here.
const exhaustive: never = event;
return exhaustive;
}
}
};
// --- hydrate-from-timeline -----------------------------------------------
//
// Discord's hydrate is simpler than Telegram's because the QR flow has
// fewer states. We walk past→present and let each event freely transition
// the state — like the TG hydrate, this is permissive (no out-of-thin-air
// rejection) because we trust the bridge's durable timeline.
// 3 minutes — Discord remoteauth's server-side timeout sits around 2
// minutes (verified empirically against v0.7.6's remoteauth/client.go;
// no explicit constant in the lib, the server-side gateway closes the
// websocket on inactivity). We use 3 min as a slight safety margin so
// reload-after-success grace still works while the panel is still
// fresh enough to scan. Telegram's QR rotates internally and lives ~10
// min, which is why the TG widget uses 10 min — Discord's single-shot
// remoteauth needs the tighter window.
const HYDRATE_FRESHNESS_MS = 3 * 60 * 1000;
// hCaptcha rqtoken is even more short-lived than the QR ticket — Discord
// invalidates after ~90s in practice. If the user reloads while staring
// at a captcha challenge older than this, restoring the captcha screen
// only sets them up for a server-side rejection on solve. Drop the state
// instead and let live ping reconcile.
const CAPTCHA_HYDRATE_FRESHNESS_MS = 90 * 1000;
export type HydrateInput = {
ev: LoginEvent;
ts: number;
};
type HydrateAccumulator = {
state: LoginState;
pendingTs: number | null;
terminated: boolean;
};
const stepHydrate = (prevAcc: HydrateAccumulator, input: HydrateInput): HydrateAccumulator => {
const { ev, ts } = input;
// After a terminal event we normally stop — except if a fresh
// `qr_displayed` shows up, that's the bridge signature of a NEW login
// flow. The user cancelled (or finished) and is now logging in again;
// the chain should resume tracking from the new start. Without this
// re-entry, `[qr_displayed, login_success, qr_displayed]` (logout-then-
// re-login-mid-QR) would return null.
if (prevAcc.terminated && ev.kind !== 'qr_displayed') {
return prevAcc;
}
// Restart-on-re-entry: clear the terminated bit AND any prior tracked
// state so the new flow's first event becomes the new anchor without
// inheriting the old QR's eventId.
const acc: HydrateAccumulator = prevAcc.terminated
? { state: { kind: 'unknown' }, pendingTs: null, terminated: false }
: prevAcc;
switch (ev.kind) {
case 'qr_displayed': {
// Same anchor logic as the live reducer.
if (acc.state.kind !== 'awaiting_qr_scan') {
return {
state: {
kind: 'awaiting_qr_scan',
discordUrl: ev.discordUrl,
qrEventId: ev.eventId,
firstShownAt: ts,
},
pendingTs: ts,
terminated: false,
};
}
if (ev.replacesEventId === acc.state.qrEventId) {
return {
state: { ...acc.state, discordUrl: ev.discordUrl },
pendingTs: ts,
terminated: false,
};
}
if (!ev.replacesEventId) {
return {
state: {
kind: 'awaiting_qr_scan',
discordUrl: ev.discordUrl,
qrEventId: ev.eventId,
firstShownAt: ts,
},
pendingTs: ts,
terminated: false,
};
}
return acc;
}
case 'qr_redacted': {
if (acc.state.kind !== 'awaiting_qr_scan') return acc;
if (acc.state.qrEventId !== ev.redactsEventId) return acc;
// Move into qr_verifying and keep the chain open — the success line
// typically follows in the same scan window.
return { state: { kind: 'qr_verifying' }, pendingTs: ts, terminated: false };
}
case 'captcha_challenge': {
// SECURITY: only accept a captcha challenge if the same chain has
// already seen a `qr_displayed` (i.e. WE initiated the login).
// Otherwise a malicious / compromised homeserver could craft an
// m.notice with the sentinel JSON pointing at an attacker-controlled
// sitekey, the user solves it on reload, and the resulting hCaptcha
// token is sent verbatim to the bridge — useful free captcha-solving
// labour for the attacker, and a phishing surface. The live reducer
// already gates on `isCaptchaAcceptingState` (which requires we're
// mid-flow), but the hydrate path replays raw timeline events
// without the live state — drop unsolicited challenges here.
if (acc.state.kind !== 'awaiting_qr_scan' && acc.state.kind !== 'qr_verifying') {
return acc;
}
// Vojo-patched bridge surfaced an hCaptcha — keep the chain open so a
// later `login_success` / `login_failed` still lands as terminal.
// The rqdata/rqtoken are short-lived on Discord's side (~2 min);
// the captcha-specific freshness gate in `hydrateFromTimeline`
// (CAPTCHA_HYDRATE_FRESHNESS_MS) drops stale states before they
// surface to the user.
return {
state: {
kind: 'awaiting_captcha_solve',
service: ev.service,
sitekey: ev.sitekey,
sessionId: ev.sessionId,
rqdata: ev.rqdata,
rqtoken: ev.rqtoken,
firstShownAt: ts,
},
pendingTs: ts,
terminated: false,
};
}
// Terminal events — collapse the chain. State becomes whatever the
// bot confirmed last; the caller returns null and lets live `ping`
// reconcile.
case 'login_success':
case 'logged_in':
case 'logout_ok':
case 'logout_no_op':
case 'not_logged_in':
case 'connection_dead':
case 'token_stored_not_connected':
case 'reconnect_ok':
case 'reconnect_no_op':
case 'reconnect_failed':
case 'disconnect_ok':
case 'disconnect_no_op':
case 'disconnect_failed':
case 'login_failed':
case 'captcha_required':
case 'login_websocket_failed':
case 'connect_after_login_failed':
case 'prepare_login_failed':
case 'unknown_command':
return { state: acc.state, pendingTs: null, terminated: true };
case 'already_logged_in':
case 'unknown':
case 'space_ready':
// Soft no-op for hydrate. already_logged_in is a live-flow warning
// that doesn't reflect persistent state; unknown is a wording-drift
// catch-all; space_ready is a post-terminal sentinel — hydrate
// terminates on login_success and lets live ping reconcile, so
// the URL gets attached on the live path, not here.
return acc;
default: {
const exhaustive: never = ev;
return exhaustive;
}
}
};
export const hydrateFromTimeline = (
inputs: ReadonlyArray<HydrateInput>,
now: number = Date.now()
): HydrateRestoredState | null => {
const acc = inputs.reduce<HydrateAccumulator>(stepHydrate, {
state: { kind: 'unknown' },
pendingTs: null,
terminated: false,
});
if (acc.terminated) return null;
if (acc.pendingTs === null) return null;
// Tighter freshness gate for captcha state — rqtoken expires faster
// than the QR ticket. This protects the user from a "solved captcha,
// bridge rejects, user confused" UX after a slow reload.
const freshness =
acc.state.kind === 'awaiting_captcha_solve'
? CAPTCHA_HYDRATE_FRESHNESS_MS
: HYDRATE_FRESHNESS_MS;
if (now - acc.pendingTs > freshness) return null;
if (acc.state.kind === 'qr_verifying') return acc.state;
if (acc.state.kind === 'awaiting_captcha_solve') return acc.state;
if (!isFormState(acc.state)) return null;
return acc.state;
};
// --- DEV sanity assertions ------------------------------------------------
if (import.meta.env.DEV) {
runHydrateSanity();
}
function runHydrateSanity(): void {
const t0 = 1_700_000_000_000;
const recent = (offset: number) => t0 + offset;
const now = t0 + 60 * 1000;
const cases: Array<{
name: string;
inputs: HydrateInput[];
expected: LoginState | null;
nowOverride?: number;
}> = [
{ name: 'empty timeline → null', inputs: [], expected: null },
{
name: 'lone qr_displayed → awaiting_qr_scan',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
],
expected: {
kind: 'awaiting_qr_scan',
discordUrl: 'https://discord.com/ra/A',
qrEventId: '$qrA',
firstShownAt: recent(0),
},
},
{
name: 'qr_redacted with mismatched target → ignored',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
{ ev: { kind: 'qr_redacted', redactsEventId: '$other' }, ts: recent(30000) },
],
expected: {
kind: 'awaiting_qr_scan',
discordUrl: 'https://discord.com/ra/A',
qrEventId: '$qrA',
firstShownAt: recent(0),
},
},
{
name: 'qr scan → no follow-up → qr_verifying (reload during the gap)',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
{ ev: { kind: 'qr_redacted', redactsEventId: '$qrA' }, ts: recent(30000) },
],
expected: { kind: 'qr_verifying' },
},
{
name: 'qr scan → login_success → null (terminal — let ping reconcile)',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
{ ev: { kind: 'qr_redacted', redactsEventId: '$qrA' }, ts: recent(30000) },
{ ev: { kind: 'login_success', handle: 'example' }, ts: recent(31000) },
],
expected: null,
},
{
name: 'login_failed after qr → null (terminal)',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
{ ev: { kind: 'login_failed', reason: 'rate limited' }, ts: recent(15000) },
],
expected: null,
},
{
name: 'captcha_required after qr → null (terminal)',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/A', eventId: '$qrA' },
ts: recent(0),
},
{ ev: { kind: 'captcha_required' }, ts: recent(10000) },
],
expected: null,
},
{
name: 'logout-then-relogin-mid-qr → awaiting_qr_scan (resume tracking)',
inputs: [
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/OLD', eventId: '$qrOld' },
ts: recent(0),
},
{ ev: { kind: 'qr_redacted', redactsEventId: '$qrOld' }, ts: recent(15000) },
{ ev: { kind: 'login_success', handle: 'old' }, ts: recent(16000) },
{ ev: { kind: 'logout_ok' }, ts: recent(20000) },
{
ev: { kind: 'qr_displayed', discordUrl: 'https://discord.com/ra/NEW', eventId: '$qrNew' },
ts: recent(25000),
},
],
expected: {
kind: 'awaiting_qr_scan',
discordUrl: 'https://discord.com/ra/NEW',
qrEventId: '$qrNew',
firstShownAt: recent(25000),
},
},
{
name: 'pending too old (5 min) → null (freshness guard, 3-min window)',
inputs: [
{
ev: {
kind: 'qr_displayed',
discordUrl: 'https://discordapp.com/ra/A',
eventId: '$qrA',
},
ts: t0 - 5 * 60 * 1000,
},
],
expected: null,
nowOverride: t0,
},
{
name: 'pending just inside window (2 min) → state',
inputs: [
{
ev: {
kind: 'qr_displayed',
discordUrl: 'https://discordapp.com/ra/A',
eventId: '$qrA',
},
ts: t0 - 2 * 60 * 1000,
},
],
expected: {
kind: 'awaiting_qr_scan',
discordUrl: 'https://discordapp.com/ra/A',
qrEventId: '$qrA',
firstShownAt: t0 - 2 * 60 * 1000,
},
nowOverride: t0,
},
{
name: 'connection_dead alone → null (terminal — let live ping reconcile)',
inputs: [{ ev: { kind: 'connection_dead' }, ts: recent(0) }],
expected: null,
},
{
name: 'token_stored_not_connected alone → null (terminal — let live ping reconcile)',
inputs: [{ ev: { kind: 'token_stored_not_connected' }, ts: recent(0) }],
expected: null,
},
{
name: 'logged_in alone → null (terminal — let live ping reconcile)',
inputs: [{ ev: { kind: 'logged_in', handle: 'x' }, ts: recent(0) }],
expected: null,
},
{
name: 'unknown alone → null',
inputs: [{ ev: { kind: 'unknown' }, ts: recent(0) }],
expected: null,
},
];
for (const c of cases) {
const actual = hydrateFromTimeline(c.inputs, c.nowOverride ?? now);
if (!sameLoginState(actual, c.expected)) {
// eslint-disable-next-line no-console
console.error('[hydrate sanity] mismatch', { case: c.name, actual, expected: c.expected });
throw new Error(`hydrate sanity failed: ${c.name}`);
}
}
}
function sameLoginState(a: LoginState | null, b: LoginState | null): boolean {
if (a === null || b === null) return a === b;
return JSON.stringify(a) === JSON.stringify(b);
}

View file

@ -0,0 +1,794 @@
/* Dawn palette must stay in sync with
* docs/design/new-direct-messages-design/project/stream-v2-dawn.jsx
* (DAWN const, lines 4-23). The widget renders inside the Vojo chat slot
* which is itself a Dawn surface; the iframe inherits the same visual
* canon to feel like a continuation of the host.
*
* Identical visual vocabulary to apps/widget-telegram/src/styles.css
* the Discord widget keeps fleet-violet (Vojo accent) rather than
* adopting Discord blurple, per product decision: «used Vojo style». */
:root {
--bg: #181a20;
--bg2: #0d0e11;
--surface: #21232b;
--surface2: #2a2d36;
--divider: rgba(255, 255, 255, 0.06);
--hairline: rgba(255, 255, 255, 0.08);
--text: #e6e6e9;
--muted: rgba(230, 230, 233, 0.55);
--faint: rgba(230, 230, 233, 0.32);
--fleet: #9580ff;
--fleet-soft: #a59cff;
--green: #7dd3a8;
--amber: #d4b88a;
--rose: #c08e7b;
--section-pad-x: 40px;
}
[data-theme='light'] {
/* Light theme is intentionally a thin remap. Vojo is dark-default; the
* theme param exists so we don't fight an explicit user/host setting,
* not because we expect daily light-mode use. */
--bg: #f5f5f7;
--bg2: #ffffff;
--surface: #f0f0f2;
--surface2: #e8e8ec;
--divider: rgba(0, 0, 0, 0.08);
--hairline: rgba(0, 0, 0, 0.1);
--text: #1a1a1d;
--muted: rgba(26, 26, 29, 0.62);
--faint: rgba(26, 26, 29, 0.4);
}
@media (max-width: 600px) {
:root {
--section-pad-x: 20px;
}
}
* {
box-sizing: border-box;
/* Kills the translucent grey overlay iOS/Android WebViews paint on top
* of any tapped element. Web browsers ignore this. */
-webkit-tap-highlight-color: transparent;
}
html,
body,
#app {
height: 100%;
}
body {
margin: 0;
padding: 0;
background: var(--bg);
color: var(--text);
font: 14px/1.45 -apple-system, 'Segoe UI', 'Inter', system-ui, sans-serif;
-webkit-font-smoothing: antialiased;
text-rendering: optimizeLegibility;
}
.app {
display: flex;
flex-direction: column;
min-height: 100%;
max-width: 960px;
margin: 0 auto;
}
/* The hero is OWNED BY THE HOST (BotShellHero). The widget body starts
* with the active-state section directly. */
/* ── Section ──────────────────────────────────────────────────────── */
.section {
padding: 24px var(--section-pad-x) 20px;
}
.section + .section {
padding-top: 4px;
}
/* Status pill non-interactive (no cursor:pointer, no hover). The pill
* carries the section's identity for stateful sections. */
.section-status {
display: inline-flex;
align-items: center;
gap: 10px;
font-size: 13px;
line-height: 20px;
color: var(--muted);
background: var(--bg2);
border: 1px solid var(--divider);
border-radius: 8px;
padding: 8px 14px;
margin: 0 0 14px;
user-select: none;
white-space: nowrap;
}
.section-status .dot {
width: 8px;
height: 8px;
border-radius: 50%;
background: var(--faint);
flex-shrink: 0;
}
.section-status.connected {
color: var(--green);
}
.section-status.connected .dot {
background: var(--green);
box-shadow: 0 0 0 3px rgba(125, 211, 168, 0.16);
}
.section-status.disconnected {
color: var(--rose);
}
.section-status.disconnected .dot {
background: var(--rose);
}
.section-status.checking {
color: var(--amber);
}
.section-status.checking .dot {
background: var(--amber);
}
/* Section row: status pill + recovery button (refresh / reconnect /
* cancel) when the state has no other affordance. Without this row, the
* user can stare at a «Проверка статуса» pill forever if the first
* ping reply dropped on the wire. */
.section-recovery-row {
display: flex;
align-items: center;
gap: 10px;
flex-wrap: wrap;
margin-bottom: 14px;
}
.section-recovery-row > .section-status {
margin-bottom: 0;
}
.recovery-action {
display: inline-flex;
align-items: center;
gap: 8px;
background: var(--bg2);
border: 1px solid var(--divider);
border-radius: 8px;
padding: 8px 14px;
font: inherit;
font-size: 13px;
line-height: 20px;
color: var(--muted);
cursor: pointer;
transition: background 0.12s, color 0.12s, border-color 0.12s;
}
:root[data-input='mouse'] .recovery-action:hover:not(:disabled) {
background: var(--surface);
color: var(--text);
border-color: var(--hairline);
}
.recovery-action:disabled {
opacity: 0.5;
cursor: not-allowed;
}
.recovery-action svg {
width: 16px;
height: 16px;
}
/* ── Command card (action card with name + desc + chevron) ──────── */
.command-card {
/* `appearance:none` strips native WebView focus paint that otherwise
* sits ON TOP of our explicit background see telegram widget for
* the full debugging trail (Capacitor Android WebView holds native
* focus paint until focus moves elsewhere). */
-webkit-appearance: none;
appearance: none;
background: var(--bg2);
border: 1px solid var(--divider);
border-radius: 10px;
padding: 14px 16px;
display: flex;
align-items: center;
gap: 12px;
cursor: pointer;
text-align: left;
font: inherit;
color: inherit;
transition: border-color 0.12s, background 0.12s;
}
/* Hover scoped to mouse-mode sessions only Capacitor Android WebView
* reports `(hover: hover)` as TRUE on a pure-touch device, so a media-
* query gate doesn't work. `[data-input]` is set in main.tsx from the
* actual `pointerdown.pointerType`. */
:root[data-input='mouse'] .command-card:hover:not(:disabled) {
background: var(--surface);
border-color: var(--hairline);
}
.command-card:focus {
outline: none;
}
:root[data-input='mouse'] .command-card:focus-visible {
outline: 2px solid var(--fleet);
outline-offset: 2px;
}
.command-card:disabled {
opacity: 0.5;
cursor: not-allowed;
}
.command-card-body {
flex: 1;
min-width: 0;
}
.command-card-name {
font-size: 15px;
color: var(--text);
font-weight: 600;
margin-bottom: 3px;
}
.command-card-desc {
font-size: 14px;
color: var(--muted);
line-height: 19px;
}
.command-card-chevron {
color: var(--muted);
font-size: 18px;
flex-shrink: 0;
line-height: 1;
display: inline-flex;
align-items: center;
justify-content: center;
}
.command-card-chevron svg {
width: 18px;
height: 18px;
display: block;
}
/* Generic leading-icon slot every command-card carries a semantic
* left-side glyph (mirror of the right-side chevron). Picks up
* `currentColor` from the parent and stays muted by default;
* `.danger` deliberately does NOT colour the lead icon so the rose
* accent stays reserved for the title. */
.command-card-lead-icon {
flex-shrink: 0;
display: inline-flex;
align-items: center;
justify-content: center;
color: var(--muted);
}
.command-card-lead-icon svg {
width: 20px;
height: 20px;
display: block;
}
/* Spin the leading refresh icon while the card is in its `refreshing`
* in-flight state. The selector targets the lead slot since the
* refresh card moved its glyph from the chevron (right) to the lead
* slot (left) for parity with every other card. */
.command-card.refreshing .command-card-lead-icon svg {
animation: command-card-spin 0.8s linear infinite;
}
@keyframes command-card-spin {
to {
transform: rotate(360deg);
}
}
/* ── Transcript ──────────────────────────────────────────────────── */
.transcript {
background: var(--bg2);
border: 1px solid var(--divider);
border-radius: 10px;
padding: 12px 14px;
font-family: ui-monospace, 'JetBrains Mono', 'SF Mono', monospace;
font-size: 12.5px;
line-height: 1.55;
max-height: 360px;
overflow-y: auto;
scrollbar-width: thin;
scrollbar-color: var(--surface2) transparent;
}
.transcript::-webkit-scrollbar {
width: 8px;
}
.transcript::-webkit-scrollbar-track {
background: transparent;
}
.transcript::-webkit-scrollbar-thumb {
background: var(--surface2);
border-radius: 4px;
border: 2px solid var(--bg2);
background-clip: padding-box;
}
.transcript::-webkit-scrollbar-thumb:hover {
background: var(--surface);
border: 2px solid var(--bg2);
background-clip: padding-box;
}
.transcript-line {
padding: 4px 0;
display: flex;
gap: 10px;
align-items: flex-start;
white-space: pre-wrap;
word-break: break-word;
}
.transcript-line + .transcript-line {
border-top: 1px dashed var(--divider);
}
.transcript-line .ts {
color: var(--faint);
flex-shrink: 0;
font-variant-numeric: tabular-nums;
}
.transcript-line .body {
flex: 1;
min-width: 0;
}
.transcript-line.from-bot .body {
color: var(--text);
}
.transcript-line.from-user .body {
color: var(--fleet-soft);
}
.transcript-line.diag .body {
color: var(--muted);
}
.transcript-line.error .body {
color: var(--rose);
}
.transcript-empty {
color: var(--faint);
text-align: center;
padding: 16px 0;
font-style: italic;
}
/* Destructive card red name marks logout as destructive vs the primary
* login card. The hover border stays on the generic
* `:root[data-input='mouse'] .command-card:hover` rule (hairline) so the
* accent is consistent across touch and mouse modes a previous
* `.command-card.danger:hover { border-color: var(--rose) }` override
* was dead in mouse mode (lower specificity than the input-gated rule)
* and only fired in the pre-first-pointerdown touch stub, leaking a
* rose tint that looked amber on the dark surface. */
.command-card.danger .command-card-name {
color: var(--rose);
}
/* Inline confirm-in-place body for the destructive logout card. */
.command-card-confirm {
display: flex;
align-items: center;
gap: 12px;
flex-wrap: wrap;
flex: 1;
min-width: 0;
}
.command-card-confirm-prompt {
font-size: 14px;
color: var(--text);
flex: 1;
min-width: 0;
}
.command-card-confirm-yes,
.command-card-confirm-no,
.btn-primary,
.btn-text {
font: inherit;
cursor: pointer;
}
.command-card-confirm-yes {
background: var(--rose);
color: #0c0c0e;
border: none;
border-radius: 7px;
padding: 7px 14px;
font-size: 13px;
font-weight: 600;
}
.command-card-confirm-no {
background: transparent;
color: var(--muted);
border: 1px solid var(--divider);
border-radius: 7px;
padding: 7px 14px;
font-size: 13px;
}
.command-card-confirm-yes:disabled,
.command-card-confirm-no:disabled {
opacity: 0.5;
cursor: not-allowed;
}
.command-grid {
display: grid;
grid-template-columns: repeat(2, minmax(0, 1fr));
grid-auto-rows: 1fr;
gap: 10px;
}
/* ── Auth card (QR panel chrome) ─────────────────────────────────── */
.auth-card {
background: var(--bg2);
border: 1px solid var(--divider);
border-radius: 10px;
padding: 16px 18px;
display: flex;
flex-direction: column;
gap: 12px;
}
.auth-card.error {
border-color: var(--rose);
}
.auth-card-title {
font-size: 13px;
color: var(--muted);
text-transform: uppercase;
letter-spacing: 1.4px;
font-weight: 600;
}
.auth-card-hint {
font-size: 14px;
color: var(--muted);
line-height: 19px;
}
.auth-card-row {
display: flex;
align-items: stretch;
gap: 10px;
flex-wrap: wrap;
}
.btn-primary {
background: var(--fleet);
color: #0c0c0e;
border: none;
border-radius: 8px;
padding: 10px 18px;
font-size: 13px;
font-weight: 600;
}
.btn-primary:disabled {
opacity: 0.5;
cursor: not-allowed;
}
.btn-text {
background: transparent;
border: none;
color: var(--muted);
padding: 10px 12px;
font-size: 13px;
}
.btn-text:hover:not(:disabled) {
color: var(--text);
}
.auth-card-error {
font-size: 13px;
line-height: 18px;
color: var(--rose);
}
.auth-card-warn {
font-size: 13px;
line-height: 18px;
color: var(--amber);
}
.auth-card-countdown {
font-size: 13px;
color: var(--muted);
line-height: 18px;
font-variant-numeric: tabular-nums;
transition: color 0.2s ease-out;
}
.auth-card-countdown.expired {
color: var(--amber);
}
/* ── QR-login panel ─────────────────────────────────────────────── */
/* Override the auth-card row layout QR panel stacks vertically with the
* matrix as the visual anchor. */
.auth-card-qr {
align-items: stretch;
}
/* The QR matrix sits on a hard #fff plate regardless of theme phone
* camera scanners need maximum contrast. */
.auth-card-qr-frame {
align-self: center;
background: #fff;
border-radius: 12px;
padding: 14px;
display: inline-flex;
align-items: center;
justify-content: center;
/* Lock the inner box to the SVG's rendered size so the placeholder
* variant doesn't collapse to zero height while the matrix is being
* computed. */
min-width: 260px;
min-height: 260px;
box-shadow: 0 1px 0 rgba(255, 255, 255, 0.06), 0 12px 24px rgba(0, 0, 0, 0.32);
}
.auth-card-qr-placeholder {
display: inline-flex;
align-items: center;
gap: 8px;
color: rgba(26, 26, 29, 0.62);
font-size: 13px;
line-height: 20px;
padding: 96px 16px;
}
.auth-card-qr-placeholder .dot {
width: 8px;
height: 8px;
border-radius: 50%;
background: var(--amber);
flex-shrink: 0;
}
.auth-card-qr-steps {
margin: 0;
padding-left: 1.4em;
display: flex;
flex-direction: column;
gap: 6px;
font-size: 13px;
line-height: 19px;
color: var(--muted);
}
.auth-card-qr-steps li::marker {
color: var(--faint);
}
/* ── hCaptcha challenge panel ───────────────────────────────────── */
.auth-card-captcha {
align-items: stretch;
}
/* hCaptcha renders its own iframe inside this host. We just provide a
* minimum frame height so the layout doesn't collapse during script load,
* and centre the iframe horizontally. The hCaptcha widget is a fixed-
* width 304px element by default; on narrow viewports we let it overflow
* its container's flex line rather than try to scale the iframe (the
* upstream SDK handles its own responsive variants). */
.auth-card-captcha-frame {
align-self: center;
display: flex;
flex-direction: column;
align-items: center;
gap: 10px;
min-height: 88px;
}
.auth-card-captcha-host {
/* hCaptcha injects a 304x78 (normal) or 156x144 (compact) iframe. The
* default rendering is 'normal' we size accordingly and let the SDK
* place its own iframe inside. */
display: flex;
justify-content: center;
}
@media (max-width: 600px) {
.auth-card-row {
flex-direction: column;
}
.btn-primary,
.btn-text {
width: 100%;
}
.command-card {
padding: 12px 14px;
border-radius: 8px;
}
.command-card-name {
font-size: 14px;
margin-bottom: 2px;
}
.command-card-desc {
font-size: 13px;
line-height: 17px;
}
.command-grid {
grid-template-columns: minmax(0, 1fr);
}
.auth-card-qr-frame {
min-width: 232px;
min-height: 232px;
padding: 10px;
}
.auth-card-qr-placeholder {
padding: 80px 12px;
}
}
/* ── Linkified transcript bodies ─────────────────────────────────── */
.transcript-line a {
color: var(--fleet-soft);
text-decoration: underline;
}
.transcript-line a:hover {
color: var(--text);
}
/* ── Diagnostic banner (pre-bootstrap failure) ───────────────────── */
.error-banner {
margin: var(--section-pad-x);
padding: 14px 16px;
background: rgba(192, 142, 123, 0.08);
border: 1px solid var(--rose);
border-radius: 10px;
color: var(--rose);
font-size: 13px;
line-height: 19px;
}
.error-banner strong {
display: block;
margin-bottom: 4px;
color: var(--rose);
font-weight: 600;
}
.error-banner code {
background: var(--bg2);
padding: 1px 6px;
border-radius: 4px;
font-family: ui-monospace, 'JetBrains Mono', monospace;
font-size: 12px;
color: var(--text);
}
/* ── About modal ─────────────────────────────────────────────────── */
.about-overlay {
position: fixed;
inset: 0;
background: rgba(13, 14, 17, 0.72);
display: flex;
align-items: center;
justify-content: center;
z-index: 1000;
padding: 20px;
animation: about-fade 0.15s ease-out;
}
@keyframes about-fade {
from {
opacity: 0;
}
to {
opacity: 1;
}
}
.about-panel {
background: var(--bg);
border: 1px solid var(--hairline);
border-radius: 14px;
width: 100%;
max-width: 520px;
max-height: 90vh;
display: flex;
flex-direction: column;
overflow: hidden;
box-shadow: 0 20px 60px rgba(0, 0, 0, 0.5);
}
.about-header {
display: flex;
align-items: center;
gap: 12px;
padding: 16px 18px;
border-bottom: 1px solid var(--divider);
}
.about-title {
flex: 1;
font-size: 17px;
font-weight: 600;
color: var(--text);
margin: 0;
line-height: 1.3;
}
.about-close-x {
background: transparent;
border: none;
color: var(--muted);
width: 32px;
height: 32px;
border-radius: 8px;
cursor: pointer;
display: inline-flex;
align-items: center;
justify-content: center;
flex-shrink: 0;
font: inherit;
font-size: 24px;
line-height: 1;
transition: background 0.12s, color 0.12s;
}
.about-close-x:hover {
background: var(--surface);
color: var(--text);
}
.about-body {
padding: 16px 18px;
overflow-y: auto;
display: flex;
flex-direction: column;
gap: 12px;
}
.about-body p {
margin: 0;
font-size: 14px;
line-height: 1.55;
color: var(--text);
}
.about-body a {
color: var(--fleet-soft);
text-decoration: underline;
overflow-wrap: anywhere;
}
.about-body a:hover {
color: var(--text);
}
.about-footer {
padding: 12px 18px 16px;
display: flex;
justify-content: flex-end;
border-top: 1px solid var(--divider);
}

1
apps/widget-discord/src/vite-env.d.ts vendored Normal file
View file

@ -0,0 +1 @@
/// <reference types="vite/client" />

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