Docs rewrite: current-state docs, HANDOFF retired

README for the repo page; HOW-IT-WORKS overview; ARCHITECTURE,
PROTOCOL, DEVELOPMENT checked against the code; ROADMAP is backlog
only (isolate/dedicated server, sync fixes, rendering+shaders,
Windows build, Jellyfin epic). HANDOFF.md and mpv-http-contract.md
folded in and deleted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
foxyzandClaude Opus 5.5 committed 2026-10-07 00:42:28 +03:00
1 parent aeff087c7d
commit d8a0729029
9 files changed
+680 -687

No files matched your search

+216 -68
View File
@@ -1,84 +1,232 @@
# ARCHITECTURE
# Architecture
One binary, three modes. `packages/server` (Relic) is embedded, not a separate product.
Plain overview first: [HOW-IT-WORKS.md](HOW-IT-WORKS.md). Wire format:
[PROTOCOL.md](PROTOCOL.md). Plans: [ROADMAP.md](ROADMAP.md).
- **Client**: GUI Join, sync WS client (`web_socket_channel` in
`apps/desktop`, NOT `packages/core` — core stays Flutter-free)
+ mpv. No server code runs.
- **Server+Client (Host)**: GUI Host starts `packages/server` in-process, binds port, self-connects via loopback. Host opens ports, runs traffic. Clients open nothing outbound-only.
- **Headless**: `--headless`, same server package, no window/player init.
## Packages
Server side: **Relic** (HTTP + built-in WS). Client side WS: **`web_socket_channel`**
(Relic has no client; server WS handling stays in Relic).
| Path | What | Rule |
|---|---|---|
| `packages/core` | protocol messages, sync engine (`RoomEngine`), constants, config, Range parsing, logging | pure Dart, no Flutter |
| `packages/server` | Relic HTTP + WebSocket server (`RoomServer`, `RoomHub`, `Session`, media endpoint), `bin/server.dart` | pure Dart, no Flutter |
| `apps/desktop` | Flutter app: UI, `SyncClient`, `SessionCoordinator`, `PlaybackController` (media_kit/mpv), updater | the only shipped binary |
| `tool` | harnesses: `range_logger`, `shaper`, `fake_viewer`, `e2e_check` | dev only |
| `deploy` | systemd unit for the standalone server (stale until the dedicated server exists) | |
Media order (locked): links first (shared http(s) URL, Syncplay-style sync),
file relay after (server-proxied bytes for the whitelist case).
Guard: `packages/core/test/flutter_free_test.dart` fails if anything under
`core/lib` imports Flutter, `dart:ui` or `signals_flutter`; CI runs
analyze + tests for core and server on every push. Server has no such test but
the same rule; it is checked by CI building without Flutter. The client
WebSocket (`web_socket_channel`) lives in `apps/desktop`; Relic has no client.
## Media paths (links vs file relay)
## Server lifecycle
Two ways to watch, same sync engine, same Range endpoint:
Today (in-process): `SessionState.host()` creates a `RoomServer`, starts it
on the chosen port (bound to all IPv4 interfaces), then the app joins it over
`127.0.0.1` like any other client. Leaving or closing the window stops it.
`RoomServer` runs a 1 s heartbeat: stale sweep, closed-socket sweep, silent
socket drop, room + feed broadcast. There is one session, id `default`.
1. **Links (primary, shipped):** host shares an http(s) media URL
(e.g. a Jellyfin Download URL). Viewers stream it directly; the
server is sync-authority only. The URL travels as `media.url`
in room broadcasts and as welcome `mediaUrl` (set to the shared
URL when the server holds no file). This
covers most sessions — proxying is usually unnecessary.
2. **File relay (secondary, local-file half shipped):** host serves
bytes from their own machine through the server's token-gated
Range endpoint `GET /m/<sid>/<token>/<name>`. Viewers (and the
host via loopback) play the server URL. This is the fallback for
the whitelist case: viewers cannot reach the source, but the
server/host can (e.g. Jellyfin allowlists the host, not the
viewers). No server storage — bytes stream file→socket, config +
logs only, never to disk.
`--headless` does not work: the Flutter binary needs a display and exits with
"cannot open display" on a headless box. `packages/server/bin/server.dart` is
the only working standalone server (flags `--port`, `--secret`, `--name`,
`--media`, `--advertise-host`, `--config`; default port 8080).
Backends behind `MediaSource` (`packages/core/lib/src/media_source.dart`):
`FileMediaSource` (local file — SHIPPED as an abstraction, but the
live route serves `SessionMedia.file` directly via `file.openRead`;
nothing in `packages/` references `FileMediaSource` today);
`HttpSource` (server fetches a remote URL and re-serves it —
DEFERRED, Phase 8). `resolveMediaRequest` stays
framework-agnostic so both ride the same tested Range table.
Planned: hosting from the app runs the server on a Dart isolate (ports and
sockets stay in the same process, UI stays responsive); the dedicated server
is a separate `dart compile exe` of `packages/server`. See [ROADMAP.md](ROADMAP.md).
## Deferred: remote-host upload tunnel (NOT active protocol)
Passwords: the default secret is empty, which means an open server (a
warning is logged). Otherwise one shared secret, compared at hello.
The host is the member that connected from a loopback address; it cannot be
kicked. Kick = revoke token, send `kicked`, close; the
person may rejoin at once with a new token.
If a VPS tier ever needs a *remote* host to push bytes to a server
it cannot serve from disk, the reserved design is the brief's WSS
tunnel: host dials out, JSON control frames
(`hello`/`welcome`+mediaUrl, `open`{id,start,end}, `ack`, `cancel`,
`eof`/`err`) plus binary `[u32 BE reqId][payload]` chunks —
chunk 256 KiB, credit window 8 (≈2 MiB in flight), max 16 concurrent
`open`, relay acks only after socket flush (real backpressure).
`RelayConstants.chunkBytes/creditWindowChunks/maxConcurrentOpen`
are reserved for this. No code exists; direct-serve covers v1
(host IS the server, opens ports itself).
## Sync model
On file relay, host playback goes through loopback like everyone
else's (`http://127.0.0.1:<port>/…`) — same code path as remote
viewers, so server bugs surface on the host screen first. (On
links the host plays the direct URL.) Deployment ladder,
cheapest first: host's own PC (LAN, or internet with port-forward) →
same binary `--headless` on a friend's always-on box → cheap VPS
(CGNAT escape; TLS/Caddy deferred with this tier).
Loosely modeled on Syncplay, with a different readiness design.
Sync: heavily Syncplay-inspired + original readiness design, NOT 1:1
(audit done; deviations: clock never moves from reports, no
pause-on-leave, force-for-all, quorum gate — see HANDOFF §2).
Additive namespaced extensions only, `proto` bump + `PROTOCOL.md` same-commit.
- **Transport** is `playing` or `held`. The server owns the clock:
`positionAt(now) = position + (playing ? now - serverTime : 0)`.
Reports from clients never move it. Only seek, swap and gate transitions
write `position`/`serverTime` and bump `eventId`.
- **Reports**: clients send `position`, `buffered`, `eventId` every 1 s. The
server stores `buffered` and liveness; a report with `eventId` older than
current has its position ignored.
- **Ready gate** (HELD to PLAYING): the in-set is non-empty, every in-set
member is ready, fresh and has `buffered >= 4 s` (`minBufferSecs`), and
quorum holds: in-set size >= 2, unless at most one member is fresh. So a
lone viewer plays alone; with a second viewer present one Space waits for the
other. Joining members start with `inSet=false, ready=false` and never block.
Disconnect removes from the set; swap clears everyone.
- **Pause**: while PLAYING, any in-set member who becomes not ready, under
4 s buffered or stale flips the server to HELD (feed action `pause` by `server`).
Recovery resumes by itself when the gate passes.
- **Seek**: allowed from any member in any transport; the server writes the
position and bumps `eventId`; flags are untouched. Clients diff each `room`
broadcast by `eventId` (`RoomDiff`): new event means seek to the computed
room position, transport change means play/pause. Between events: drift
over 5 s for more than 2 s seeks, ahead by more than 1.5 s sets rate 0.95
until level, otherwise nothing. Local seeks are detected by the coordinator
polling mpv position once per second (`_detectUserSeek`, jump over 2 s and
more than 2 s from the room) and sent as a `seek` intent; a retry
(`_checkSeekLanded`) re-applies a room seek that has not landed after 1.5 s.
Known roughness: see ROADMAP.
- **Liveness**: WS ping every 5 s; silent over 15 s means stale (out of the
gate, `stale` feed); over 30 s the socket is closed and the member removed.
Closed sockets are polled with `isClosed` every second: Relic's server-side
`onDone` does not fire for dead peers, so presence never relies on it.
- **Client policies**: no room broadcast for 5 s means connection lost, clear
the server state and reconnect every 1.5 s until cancelled (an epoch counter makes
a cancelled attempt stale) or kicked. After a reconnect the ready flag is
restored from the server's state. Cancel during setup aborts the join.
Closing the window leaves gracefully (2 s timeout).
- **Idle** (`IdleTracker`): measured from the later of the last input and the
moment playback stopped. While HELD and self ready: warning at 8 min,
auto-unready at 10 min, and never less than 2 min after the warning.
- **Deviations from Syncplay**: clock never follows the slowest watcher; no
pause-on-leave; anyone can force ready/unready (all or one); quorum gate;
the clock is server time, and remote clock offset is not corrected (clients
assume their clock matches the server's).
No server storage: config + logs only. Media never touches disk.
Additive wire changes only; bump `proto` and PROTOCOL.md in the same commit.
## Health reporting (standing rule)
## Media
Every failure must display. Nothing stalls silently.
Two sources, same sync.
- Server `/health` returns machine-readable status: ok, room, viewer
count, uptime, media, bytes served, version. The app reads it.
- App surfaces all failure states in UI with cause + what-to-do:
connecting/reconnecting, wrong secret (401), kicked (403), session
gone (410), host offline / buffering (503), starvation
("waiting for X — N.N s buffered"). Backed by a structured status
signal (severity + message), not bare strings.
- Errors log via `debugPrint` + stack trace at the boundary, then
convert to a displayable state. No swallowed exceptions.
- **Link**: the media field holds an http(s) URL. It travels as `media.url`
in `swap` and `room`, and as `mediaUrl` in `welcome`. Each app plays it
directly; the server is only the clock.
- **Host local file**: `RoomServer.setLocalFile` / `mediaPath` makes the file
the session media. `GET /m/<sid>/<token>/<name>` serves it with a Range
table, gated by the viewer's token. The server builds the URL per viewer
(`http://<host>:<port>/m/default/<token>/<name>`), so tokens never reach
other viewers. `<host>` is the authority the client dialled; `advertiseHost`
(default empty) only overrides it when set explicitly. When a `room`
broadcast has no `media.url`, the client builds the URL itself from the host
it connected to. Bytes stream file to socket and are counted as streamed
(`bytesServed`); nothing is stored. The host plays through loopback like
everyone else.
Range table (`resolveMediaRequest`, tested byte-exact):
| Request | Response |
|---|---|
| no Range | 200, `Content-Length` = size |
| `bytes=a-b` | 206, `Content-Range: bytes a-b/size` |
| `bytes=a-` | 206 to EOF |
| `bytes=-n` | 206 last n bytes |
| unsatisfiable | 416, `Content-Range: bytes */size` |
| host has no file | 503 + `Retry-After: 2` |
| bad / unknown token, kicked, unknown session | 401, 403, 410 |
Always `Accept-Ranges: bytes`; `Content-Type` from the extension
(`contentTypeForFile`). Relic hands the Range over as a typed `RangeHeader`;
the endpoint takes the first range and keeps the unit, so non-`bytes` units
still get 416.
What mpv does against it (recorded with `tool/range_logger.dart`): mp4 opens
once with no Range then streams `bytes=0-` to the end; mkv opens `bytes=0-`
twice (probe, playback), reads a small tail at `size-n` (cues, expressed
open-ended), then every seek is a fresh `bytes=N-` connection. A server
outage of a few seconds is ridden out from mpv's cache with no error. Longer
outages exhaust the cache and rely on reconnect (not chaos-tested yet).
mpv settings (`PlaybackController`): `Player` with `bufferSize` 64 MiB,
`libass` on; pinned via runtime property: `demuxer-readahead-secs=10`,
`cache-secs=10`, `stream-lavf-o=reconnect_streamed=1,reconnect_on_http_error=1,reconnect_delay_max=4`
(comma form; this mpv build rejects colon-joined). media_kit has no public
`setProperty`; the code calls it through a guarded dynamic call on
`player.platform`. Hardware decoding comes from Settings (auto, off, custom)
and is passed to the `VideoController`; env `SYNC_RELAY_HWDEC` (e.g. `no`)
overrides it. It is read per player, so a change applies on the next join or
Reload video.
Player lifecycle: the `Player` and its `VideoController` live for one
session: created on join, disposed on leave, replaced by `recreate()` for
"Reload video" (button or Ctrl+R; volume kept). `openUrl` opens with `play: false`; play/pause/seek/rate come only from
`RoomDiff` actions or the play gate (`localPlayAllowed`: self in-set and
ready). A room seek within 0.5 s of the current position is skipped. `leave()`
stops the player (otherwise audio keeps playing). On media change
(`mediaId` bump) the coordinator opens the new URL and `RoomDiff` seeks to 0,
holds and resets rate.
Frozen picture: `FrameWatchdog` samples about once a second. It hashes a
small capture of the compositor's centre (`ShownFrameProbe`) and suspects a
freeze when position advances while that hash stays constant for 3 s. Two
differing mpv `screenshot-raw` frames show the decoder is still producing
pictures, which confirms the freeze (a still scene is dismissed). A confirmed
freeze reloads the video with a toast. It fires once until the picture moves
again for 10 s, so it cannot loop. Seeks, pause, buffering and a hidden window
disarm it.
Never unmount GPU-textured widgets (avatar images) on a transport-gated
branch: teardown races mpv's frame upload and the window goes black. Keep them
mounted and skip painting (`Visibility` with `maintainState`).
## Desktop UI
One `HomeScreen` (`screens/home/`) with a nav rail: Home (profile chip, Join
card, Host card, up to 6 Recent server chips for one-click rejoin), Settings
(profile, hardware decoding, updates, shortcuts, shaders "coming soon"),
Integrations (Jellyfin "coming soon") and About. The update indicator sits on
Home. Settings persist in two Hive boxes: `connect_form` (last form and
recents; saved even when a join fails) and `settings`. Link media names are
derived from the URL path and never contain a query string.
Window chrome is frameless (`window_manager`): every screen brings its own
drag area and min/max/close buttons (`widgets/window_chrome_bar.dart`,
`window_buttons.dart`); the session drag layer is `window_drag_layer.dart`.
## Health reporting
Every failure must show in the UI with a cause and a next step; nothing
stalls silently. The app has a `HealthStatus` signal (severity + message)
for connecting, reconnecting, wrong password, version mismatch, kicked,
host offline, and starvation ("waiting for X - N.N s buffered").
Errors are logged with a stack trace at the boundary, then turned into a
displayable state. The server's `GET /health` returns JSON (`ok`, `room`,
`proto`, `viewers`, `bytesServed`, `uptimeSecs`); it is unauthenticated and
the app does not read it.
## Updater
`apps/desktop/lib/src/update/updater.dart` (no Flutter imports, unit tested).
Reads a manifest (`version`, `protocol`, `url`, optional `sha256`, `notes`)
from `updateManifestUrl`, compares versions, downloads the tarball, checks
sha256 when present, then extracts into a staging dir inside the install
folder (same filesystem, so moves are renames; the folder must be
user-writable). Each top-level entry is moved aside and the new one moved in,
with rollback on failure (writing the running executable fails with ETXTBSY;
rename works). A packaged install is detected by `data/` and `lib/` next to
the binary; anything else exits with "update manually". A manifest whose
protocol is newer than the local one is a required update. Windows apply
(`*.old` rename, swept on next start) exists but is untested. The version is
stamped with `--dart-define=APP_VERSION` (dev runs report `0.0.0-dev` and skip
checks). `watch_together --upgrade` (or the shipped `update.sh`) runs it from
the CLI; `UpdateController` runs the same code for Home's update indicator,
which opens the update card, and restarts the app ("Update and restart").
Settings has Check now and an auto-check toggle. The manifest is `updates/manifest.json` in
this repo, written by `tool/release.sh`.
## Decisions
- Desktop + mpv only, no web clients: exists because Jellyfin SyncPlay is bad.
- No transcoding: bandwidth is plentiful and the Jellyfin box should not be loaded.
- No roles, friends trust model: simplest thing that fits the use.
- Blank password = open server: the owner chooses between open and private.
- The host cannot be kicked: the server dies with the host anyway.
- Server ships inside the app (and, planned, as a separate compiled server):
a Flutter binary cannot run headless, so a dedicated server needs its own
build of `packages/server`.
- Server owns the clock; reports never move it: avoids ping-pong between clients.
- Quorum gate: a lone viewer must be able to watch; two or more must start together.
- Relic for the server, `web_socket_channel` for the client; packages are
chosen on merit, not locked.
- Plain `ws://`, no TLS: friends use LAN, VPN or a forwarded port; TLS
(e.g. a reverse proxy) is a documentation item.
- One session per server, id `default`: the map is keyed for later but unused.
- No server storage: config and logs only; media never touches the server disk.
- Avatars ride in the 1 Hz `room` broadcast (cap 256 KiB base64 per avatar):
simple, costs host uplink; hash-based serving is on the roadmap.