With video on the MediaCodec plane, real bitstream passthrough, media3
demuxing, and the GL fallback ladder in place, mpv measures at parity
or better than ExoPlayer on every device class tested (Tegra, Mali,
Amlogic armv7, Adreno; API 28-36; phone, TV, and foldable), including
the two historical Android mpv failure modes: low-end TV playback and
Hi10P.
Installs that never chose a backend move to mpv; a stored ExoPlayer
choice is preserved, and the Player Backend toggle stays user-visible
as the escape hatch. The automatic ExoPlayer-to-mpv runtime fallback
is unchanged.
The in-app FFmpeg container demuxer classified every DTS variant as
plain DTS (profile-blind MIME mapping at the demux boundary), so
DTS-HD MA lost its lossless identity downstream, and it ignored
container display-aspect-ratio overrides. media3's extractors get both
right.
Delete the FFmpeg demuxer (JNI, extractor, and its setting); ExoPlayer
direct play demuxes with media3's DefaultExtractorsFactory again, and
mpv - the Android default - reads container display dimensions itself.
close#2124close#2115
mpv's ao_audiotrack hardcoded stereo IEC61937 frames, which cannot
carry TrueHD MAT or DTS-HD MA 8-channel bursts, and the app fed
audio-spdif an empty list in self-defense because mpv
force-passthroughs every codec named there with no decode fallback.
The mpv path therefore decoded everything, and passthrough had to be
enabled by hand on every install.
The Kotlin audio route probe (AudioOutputPolicy) now feeds mpv a
per-route audio-spdif list, the fork's multichannel IEC61937 patch
(pinned libmpv v1.1.3) provides the burst geometry, and E-AC3, TrueHD
and DTS-HD MA bitstream on the mpv path. Android TV installs with no
stored preference default to passthrough on, on both backends; an
explicit user choice is preserved.
Verified by route probing and mpv logs on Shield and Box R; no AVR was
in the loop, so receiver-side decode is unconfirmed.
Android mpv drew every frame through vo=gpu: an ImageReader/GLES copy
pinned to 8-bit RGB0 with a 100 ms timed wait per frame, which
truncated 10-bit and HDR output, raced acquireLatestImage, and
performed badly on low-end TVs.
Video now scans out on a SurfaceFlinger video plane through the fork's
vo=mediacodec: decoder buffers are queued to the surface at their
target PTS (av_mediacodec_release_buffer_at_time), so 10-bit and HDR
dataspaces reach the display untouched and frame pacing no longer
depends on GL vsync. Subtitles and OSD render on a sibling transparent
surface presented for the same PTS, frame-locked by construction.
The plane takes decoder buffers only, so software-decoded video stays
on a GL vo, with a chain-failure watchdog re-initializing the output
when the plane refuses a stream mid-session. The GL fallback is vo=gpu
- gpu-next breaks the Tegra GLES linker (#2010) - and GpuVoPolicy
selects gpu-next only when Dolby Vision RPU reshaping needs
libplacebo. AV1 film grain is applied in the decoder
(vd-lavc-film-grain=cpu) because the GL raster grain fallback is not
available on this path.
Pins libmpv-android v1.1.3, which carries the fork patches this path
rides on: the vo itself, BT2020-PQ EGL window surfaces, GLES direct
rendering treated as slow (H.264 Hi10P software decode on Tegra went
from 0.3x realtime with wrong colors to 1.0x with correct 10-bit
output), and multichannel IEC61937 for ao_audiotrack.
Verified on Shield Pro (Tegra/GLES), Pixel 7 (Mali), Box R 4K Plus
(Amlogic armv7), and Galaxy Z Fold3 (Adreno): 520 play/teardown cycles
and 3 hours of continuous 4K HDR playback; HDR engagement confirmed
via SurfaceFlinger dataspace BT2020_ITU_PQ.
A device with no hardware HEVC decoder could still be handed an HEVC
transcode: the device profile advertised a fixed codec list that assumed
every device decodes everything. Prepending AV1 to reach the AV1 encoders
issue #2131 asks for would have made that worse - an Apple TV 4K and every
iPhone before the A17 Pro have no AV1 decoder at all.
Probe the platform instead. Android enumerates MediaCodecList for a
hardware decoder and iOS/tvOS ask VideoToolbox, both feeding one latched
VideoDecodeCapabilities that the device profile reads when it builds its
codec lists. Desktop deliberately answers nothing: a pre-Kaby-Lake Mac has
no hardware HEVC decoder and an M1 no AV1 one, yet both software-decode in
real time, so narrowing there would force transcodes for nothing. An
unanswered or failed probe advertises everything, so the list never
narrows on missing data.
The transcode target becomes av1,hevc,h264 filtered by the probe. Leading
with AV1 is safe because the server rotates codecs its admin has not
enabled ("Allow encoding in HEVC/AV1 format", both off by default) to the
back before picking one, so it costs nothing on a server that will not
emit AV1.
Audio now accepts everything the path can carry. The direct-play profile
drops its AudioCodec list entirely - an omitted list means "any codec" to
Jellyfin - so an audio stream can no longer be what blocks direct play.
The transcode target lists every codec Jellyfin can put in an fMP4
segment, so a video-only transcode copies DTS or TrueHD instead of
re-encoding it. Two limits bound that string: the server validates it
against ^[a-zA-Z0-9\-\._,|]{0,40}$ when it echoes the list into the
transcode URL, so alac does not fit and * is not a wildcard; and omitting
the key is not "accept everything" here the way it is for direct play,
because the server substitutes the source codec and then ships no audio at
all for a source fMP4 cannot carry.
close#2131
MPEG-TS segments cannot carry AV1, so a Jellyfin server with an AV1
hardware encoder (Intel Arc and similar) could never pick AV1 as the
transcode output no matter what the codec list advertised. Switch the
video TranscodingProfile container from ts to mp4, matching
jellyfin-web's fMP4 HLS behavior; mpv consumes fMP4 HLS on every
platform and the Plex VOD target has shipped it since issue #1859.
close#2131
Preferences that belong to the media-server account had nowhere to live in Plezy. Jellyfin keeps a user's audio/subtitle language, subtitle mode and several library options in `UserConfiguration`; plex.tv keeps the same language choices plus watched indicators and review visibility on the account. Plezy read four of those fields for auto-track selection and could never show or change any of them.
Adds an Account preferences section under Connections. One account edits in place; several show a picker first, scoped to the active profile's own connections so a managed Plex Home user never edits the owner's record. Every row is gated on what the backend can actually store.
It also moves "rewatching in Next Up" onto the account. There is no `UserConfiguration` field for it, but the per-user `DisplayPreferences` store is keyed `(userId, displayPreferencesId, client)` with no device component, so the switch goes there and follows the account.
`AccountRef` keys by account, not `clientScopeId`: MediaBrowser is `{machineId}/{userId}`, Plex is (account, Home user). Writes are patch-shaped because both backends replace whole objects — Jellyfin's `POST /Users/{id}/Configuration` and its `DisplayPreferences` row both reset omitted fields, so each write re-reads, merges only the patched keys, and posts back. Plex takes its changes as query parameters with an empty body, and its `experience` blob and the PMS `/accounts/1` mirror are deliberately untouched.
`AccountPreferencesController` owns a single repository above the profile session, so changing a language in settings reaches the next playback without a restart. Emby is gated out of rewatching through `MediaBrowserDialect.supportsNextUpRewatching`.
close#1910
Linking a Seerr instance had two rough edges, both on the Jellyfin path.
The URL step prepended https:// to anything without a scheme, so a
plain-HTTP instance on the LAN — 192.168.1.5:5055, or a bare host on
Seerr's default port — failed with "could not reach" unless the user knew
to type the scheme. And the only way to sign in with a Jellyfin account
was typing a username and password, which on a TV remote is misery.
Schemeless input now expands into candidates that are probed together,
but TLS wins by construction: a plaintext success is held while any https
candidate is still in flight and is accepted only once they have all
failed. The sign-in that follows posts a password to whichever URL wins
here, so a slow-but-working https endpoint has to beat a fast plaintext
one, and the wait is already bounded by the probe timeout. When nothing
answers, the failure that reached a server outranks a transport error, so
"this instance has not completed first-run setup" is not masked by "could
not reach https://..." from a candidate the user never typed.
Quick Connect goes through Seerr's own proxy routes (3.4+): initiate,
then check polled at Seerr's own 2s cadence through the shared
pollWithBackoff — a 404 mid-poll means the secret is gone and is terminal
— then authenticate, which mints the session cookie through the existing
sign-in path. The secret rides in a query string, so it is registered for
log redaction the way the Jellyfin flow registers its own. An instance
that predates the routes 404s the initiate and is told apart from a
generic rejection.
SeerrAuthMethod.quickConnect deliberately stores no secret: silent
re-auth is impossible by construction, so an expired cookie lands in the
existing "no stored credentials" arm, unlinks the session, and the
connect flow asks for a fresh code. The affordance is limited to the
Jellyfin credential form, since Seerr rejects Quick Connect for Emby, and
is not auto-started on TV the way the MediaBrowser form is:
/settings/public exposes no "Quick Connect enabled" flag to gate that on.
The waiting panel and its attempt/cancel bookkeeping moved out of
AddJellyfinScreen into QuickConnectCodePanel and QuickConnectFlowMixin so
both screens share one implementation, and the post-frame focus helper
both forms use moved to AsyncFormStateMixin.
Turning what a user typed into URLs worth probing lived inside
JellyfinEndpointDiscovery, next to a private copy of the scheme-detection
regex that url_utils already had. A second backend needs the same rules
with a different default port, and the existing copy carried two latent
defects: input with no host came back unchanged, so a candidate that can
never resolve went to the prober, and a query string survived into the
base URL that request paths are appended to.
expandBaseUrlCandidates takes the ordered guesses a backend wants — a
scheme plus the port to try when the user typed none — and owns the rest.
A typed port beats a guessed one, so host:8096 collapses to one candidate
per scheme and the same guess list covers both bare hosts and hosts with
ports. Candidates are built field by field rather than through
Uri.replace, whose null query means "keep mine", so query and fragment
are dropped for real. Blank input and input with no host expand to
nothing.
JellyfinEndpointDiscovery passes its existing order, so its probe and
persistence behaviour is unchanged, and hasUrlScheme replaces the regex
copy.
Playing a downloaded movie or episode failed outright once the storage
holding it was gone — an SD card ejected, a USB volume unplugged, the
files deleted from outside the app (issue #2101). The item still played
fine from the server, but Plezy never tried: it committed to the local
copy and surfaced mpv's open failure.
Resolution already refused a missing local copy and let the caller
stream instead, but only for filesystem paths. SAF `content://` URIs
skipped the check on the premise that they are always playable as
written. They are not: an unmounted volume answers
`FileNotFoundException: No root for <volume>`, and a revoked grant is
just as dead, while the row still reads `completed`.
Probe the stored document through SAF before preferring it, so an
unreachable copy resolves to nothing and every consumer degrades the
way a missing file already did — the player streams, and both
external-player entry points launch against the server instead of a
dead URI.
close#2101
Siri Remote navigation felt sluggish and then over-sensitive next to
native tvOS apps (issue #2006): swipes were priced at a fixed travel
per step, a single flick could glide into a second focus step, hard
lifts coasted several extrapolated steps, and rail scrolls snapped in
65-250ms where native glides.
Retuned the whole path against two hardware instrumentation passes on
an Apple TV 4K: committed-move telemetry through an experimental
UIFocusItem bridge (branch feat/tvos-native-focus-bridge), then a
dedicated native probe app logging every touch sample, pan velocity,
engine hint, focus step, and scroll tick across 101 swipe sessions on
160/230/300pt tiles. What the data showed, now encoded:
- step pricing follows geometry: one step costs the focused item's
extent along the swipe axis plus ~155pt (measured 314/391/410pt on
160/230/300pt tiles), not a fixed distance. Thresholds derive per
axis from the primary focus rect, normalized so a wide-flat control
steps vertically once the finger covers its height. Locked-focus
rows (hub rows, the TV browse rail) vend their selected card's rect
through the new LockedFocusRowNode so the row-wide focus node's
screen-sized rect never prices the step. Scopes, the player's
catch-all surfaces, and unbuilt cards fall back to a fixed 400pt.
- a lift never coasts more than one step: sessions with lift
velocities up to ~11400pt/s never produced a second coast step. The
glide is gated on a sustained drag (two consecutive same-direction
steps), cancelled by reversal pivots and new touches, so a discrete
flick moves exactly one item.
- the native 'inertia' feel is the scroll animation, not focus
physics: the engine's scrollable containers settle over ~450-900ms
of ease-out. TV rail and hub-row navigation scrolls now retarget a
500ms easeOutCubic animation per step, so drags and hold-repeats
chain into one continuous glide that catches up on release.
Non-English users saw English text in a dozen places and blank labels in
sixteen more.
The English came from sites that produce their copy away from the widget
that renders it, which is what the structural hardcoded-string check
cannot see: picture-in-picture refused with a raw literal instead of the
pipErrors.notSupported key that already existed; the two Jellyfin/Emby
auth throws missing display: rendered their developer message on the
add-server form; ServerParsingException.toString() fed its English into
the localized "Failed to load servers" wrapper; Watch Together
interpolated the whole PeerError, so a failed create read "Failed to
create session: PeerError(PeerErrorType.timeout): Timed out creating
session" and join printed its prefix twice; the hub and playlist
continuation footers rendered exception.toString(); shader rows showed an
English title over an already translated subtitle; the player queue fell
back to the raw Dart enum name; a Plex home user with no title showed
"Unknown"; a failed player start showed "Exception: Failed to initialize
player"; and the tvOS top-shelf header was hardcoded in an extension that
has no Flutter engine.
The blanks came from three recent features that added English keys
without translations. clean_translations.py filled all 21 siblings with
empty strings, so the Android TV resolution switch, every Jellyfin/Emby
recording-rule field, the demuxer row, the Companion Remote address
caption and the Seerr blocklist pill rendered nothing at all.
Two fixes are structural rather than key swaps. ContinuationStatusSliver
now takes an errorContext and calls a new non-logging
localizedLoadErrorText, so no future throw can leak through it. lib/mpv
stays free of user-facing copy: it raises a PlayerInitializationException
sentinel and a PlayerError.playerInitFailed cause tag that the player
screen resolves to localized text. The tvOS section title travels in the
shelf payload, additively, so an older cache still renders.
Sub-4K content on a 4K TV was always upscaled by the playback device, so the TV's own (usually better) upscaler never received the native signal and could not apply resolution-specific processing.
Adds a Match Content Resolution setting on Android TV: playback picks the smallest display mode that still contains the video — never downscaling it — and rate-matches within that resolution when frame rate matching is also enabled. A resolution-only switch keeps the refresh rate as close to the current one as possible, and the settle watchdog now verifies the requested mode id so such switches still trigger the Android MPV decoder refresh. Transcodes get their target from decoded dimensions on the post-first-frame path. Mode selection is extracted into a pure, JVM-tested DisplayModeSelector.
close#2073
In-player subtitle search always ended with "Subtitle downloaded, but it
could not be selected": the apply flow polls getVideoPlaybackData for the
new external stream, but the fresh-cache-first fast path (be9197ed) served
the pre-download /library/metadata row for every iteration — playback
start leaves a fresh row, and the first network poll re-stamps it, so the
new stream was never observed and the flow timed out.
Add forceRefresh to PlexClient.getVideoPlaybackData (mirroring
getPlaybackExtras) and pass it from the download poll loop. The forced
fetch still rewrites the cache row, so the subsequent source switch sees
the new track.
Jellyfin and Emby users had no way to record live TV from Plezy at all: no
record button in the EPG guide's program menu, and nothing in the recordings
UI, because every recording surface was gated behind a Plex-only DVR adapter.
The recording surfaces are now backed by a MediaBrowser DVR adapter that
synthesizes the neutral (Plex-shaped) payload models from the timer APIs:
- Template comes from `GET /LiveTv/Timers/Defaults?programId=`, offering a
"Record Episode" entry and, for series airings, "Record Series". The
defaults DTO travels JSON-encoded in the template's opaque `parameters`
and setting ids are DTO field names, so the existing template-driven
record-options sheet renders and round-trips them unchanged.
- Create POSTs the mutated defaults whole to `/LiveTv/Timers` or
`/LiveTv/SeriesTimers`; a duplicate one-off create answers 400, which the
adapter rethrows as `RecordingConflictException` so the sheet can show
"Already scheduled" without a backend check.
- Scheduled recordings read `/LiveTv/Timers` minus `Cancelled`/`Completed`
tombstones, rules read `/LiveTv/SeriesTimers` with their child timers
nested. Rule keys carry a `timer:`/`series:` prefix so cancel, edit and
delete dispatch to the right timer space inside the adapter.
Guide programs now carry their recording state: the program id doubles as the
record seed, and `TimerId`/`SeriesTimerId` become the rule keys that drive the
guide's scheduled dot and the Manage action. The series key is stamped only
when an airing actually records, so an episode a series rule skips does not
show a false indicator.
Three neutrality fixes on the shared UI: the record-options sheet only demands
a target library when the template declares one (MediaBrowser records into its
own configured folder), and "Re-evaluate rules" is hidden unless a connected
DVR supports it, since only Plex has that endpoint.
`fetchDvrs` deliberately stays empty so the synthesized per-server Live TV
identity that channel fetches, favorites and playback key off is preserved.
Verified end-to-end against a disposable jellyfin/jellyfin:10.11.11 container
with an M3U tuner and XMLTV guide, driven through the real client: template,
create, duplicate conflict, cancel, series create with child grabs, edit
round-trip and delete.
close#1645
Tapping a discovered desktop in the phone's companion remote did nothing:
the device list emptied to "No devices found on your network" and no
connection was ever made. Manual IP:port entry silently did nothing too.
A host advertises every local IPv4 address, so a PC with virtual adapters
(WSL, Hyper-V) broadcasts addresses the phone cannot route to. The client
races them all, and cleanup of the losing candidates awaited
`sink.close()` on channels that never connected. That future never
completes in web_socket_channel: the close future waits on the local
stream's listener, which is only attached on the connect-success path. So
the race never finished, the winner was never joined, and the flow just
stopped with no error.
Cleanup now tracks whether a candidate ever connected: unconnected probes
get a deferred close armed on `ready` settling instead of a blocking
await, and they skip the terminal drain since they hold no host admission
slot. The managed join gets a real connect timeout that surfaces as a
typed timeout error, and disconnect during a pending connect closes the
socket once the connect settles rather than blocking on it.
Two paper cuts around the same flow: a failed attempt now restarts
discovery instead of stranding the cleared list on "No devices found",
and the desktop's server card shows its listen addresses so manual entry
is not guesswork.
close#2077
A quoted value in the mpv.conf settings screen, copied verbatim from a
real mpv.conf (sub-font = 'NetflixSans-Bold'), reached mpv_set_property
with the quotes included: a string property silently selected a
nonexistent font family and fell back to the bundled font, and a
numeric property failed mpv's parse and was only logged. Either way the
line did nothing. mpv's own config-file parser strips one pair of
matching quotes around the whole value; parseMpvConfigText now does the
same before the startup pass applies the entries.
close#2025
On a 32-bit Amlogic TV box, 57% of UI frames while moving focus along a rail
missed the 16.68 ms budget, and the median frame was already over it at 20.7 ms.
The GPU was idle throughout (raster p90 9.1 ms, zero frames over budget) -- all
of it was main-isolate work, split roughly evenly between layout and semantics.
Five separate causes, all measured:
- The horizontal rail used `itemExtentBuilder`, so `RenderSliverVariedExtentList`
walked every preceding index on each realized child, in every layout pass, and
re-resolved the trailing slot inside the closure. Pinning the trailing cell to
the card extent makes the list uniform, which restores O(1) offset, index and
max-extent math.
- The sidebar-expand tween wrapped the whole shell with a `LayoutBuilder` inside
its builder, so ~15 ticks per focus flip rebuilt `SideNavigationRail` -- and its
non-virtualized child list -- during the layout phase. The tween now wraps only
the content `Positioned`, matching `SideNavigationBleedBuilder`.
- Each nav item crossfaded through two `Opacity` subtrees, i.e. 14-18 offscreen
save layers per frame for the whole 250 ms morph. It now fades colour alpha on
the leaf, the same substitution `AnimatedDimScrim` already documents.
- `HorizontalScrollWithArrows` kept a scroll listener on platforms where its arrow
chrome is compiled out, so crossing either scroll boundary `setState`d a whole
rail row of cards. Two full-row rebuild storms per traversal, for arrows that
cannot appear.
- The rail's semantic proxy rebuilt its label and six closures on every
`_RailFocusModel` notification, including vertical-scroll flips.
Also: one merged listener per `SettingsBuilder` instead of six, and the media card
semantic label is cached rather than rebuilt and then discarded by the rail's
`ExcludeSemantics`.
Measured on the device, median of three runs: frames over budget 120/212 -> 32/246,
total UI-thread work 4067 ms -> 1845 ms, layout p90 18.2 -> 8.1 ms, build p90
7.95 -> 2.42 ms. The sidebar scenario went from 121/148 frames over budget to
32/184.
Cold start on the target TV box reaches its first frame 21 ms after `dart_main`,
so nothing Dart-side gates the splash. Everything below is on the path to the
first *useful* frame, which is a strictly serial chain and where the viewer
actually waits.
- `monoTheme` is a pure function of two bools that builds a full `ColorScheme`, an
applied-and-copied 15-style `TextTheme`, ~14 sub-themes and then clones the whole
`ThemeData` again. It was rebuilt five times per cold start, two of them before
`runApp`, and twice more per app-shell rebuild. It is memoized now, keyed by
palette plus `TargetPlatform` -- the platform matters because `ThemeData()`
derives tap target size, visual density and typography from
`defaultTargetPlatform`, so a palette-only key would be wrong under a debug or
test platform override.
- `initializeDateFormatting` ignores its locale argument and builds CLDR symbols
and patterns for all 121 locales synchronously. It blocked the gate ahead of the
database open for data that only content screens use.
- `DownloadStorageService.initialize` ended in a `path_provider` round trip plus
mkdir at the tail of the gate, contradicting the comment above it that already
explained offline artwork is not a launch requirement.
- `recoverInterruptedDownloads()` and `TrackerCoordinator.initialize()` ran from
`initState` of the widget whose first build produces the first app frame, and
the RSS watchdog installed a periodic timer there whose first useful sample is
15 s away regardless.
- `CredentialVault` decrypted every token with pure-Dart AES-GCM on the main
isolate, uncached, on every registry read and on every Drift re-emit -- and the
binder writes tokens during the startup sweep, so writes re-triggered reads.
Decryption is memoized by ciphertext, with `invalidateCache()` wired into the
preference-store repair path so a repaired install cannot serve stale plaintext.
- `reloadFromStorage` now coalesces in-flight callers. The two serial awaits around
the legacy migration are deliberately not merged; only genuinely concurrent
callers share a snapshot.
- `_sameConnections` ran two `jsonEncode` calls per connection on every Drift emit
purely to compare, allocating two maps and two strings each time.
- The splash rendered one `CircularProgressIndicator` per pending server on top of
the aggregate one, so N+1 tickers scheduled a frame every vsync for the whole of
`awaitBindingSettle` -- competing with the startup work they were reporting on.
Measured on device in a settled dexopt state: time to `main_screen` 1495 -> 1400 ms,
`credentials_loaded` 1238 -> 1128 ms, `database_ready` 501 -> 443 ms. First frame is
unchanged at ~18 ms, as expected.
The image pipeline funnels through `MediaImageHelper.serverArtworkProvider`, which
sizes the request, bounds the decode and namespaces the disk key. Five places
escaped that funnel, and on the reduced tier they compete for a 64 MiB image cache.
- Jellyfin trickplay prefetch built a bare `CachedNetworkImageProvider` and called
`resolve()` on it, fully decoding a sprite sheet -- roughly 22 MiB of RGBA for a
10x10 grid of 320-wide tiles. None of it was ever painted: the render path wraps
the same provider in `ResizeImage`, so it decodes again under a different cache
key. The comment already said the intent was only to warm the disk cache, so it
now does exactly that through the cache manager and resolves no image stream.
- `CyclingMediaBackdrop` re-fetched and re-decoded a full-screen backdrop every
10 s forever, with no `DevicePerformance` term in `_canRotate` even though its
own fade duration is tier-gated. At the reduced tier's 1280x720 art cap that is
3.52 MiB per rotation, ~7 MiB live across the crossfade, churning while the
viewer sits still and evicting the rail posters behind it. The fade is already
zero-duration there, so the rotation bought variety and nothing else.
- Catalog detail passed `item.posterUrl`/`backdropUrl` with no client, which
`getOptimizedImageUrl` deliberately returns unmodified -- so a 40x60 slot pulled
a 600x900 TMDB asset and a 320-tall backdrop pulled 1920x800. `posterFor`/
`backdropFor` already existed for this; the sites now use them.
- The season strip, the Live TV programme sheet and the profile avatar built
`CachedNetworkImage` directly, keeping the sized request but losing the
`plex_optimized_<sha1>` disk key, so the same artwork was cached twice and
re-downloaded on detail open. The season strip also discarded its height bound.
Sources without published variants (Trakt, AniList, MAL, MDBList) still fall back
to the base URL.
The Auto demuxer mode routed only media3's weak container families
(AVI, ASF/WMV, MPEG-PS, Matroska/WebM) through FFmpeg and kept MP4/TS
on media3's extractors, with a second any-container FFmpeg instance
behind media3's list. With the goal of fully switching to the ffmpeg
demuxer, the split served no purpose and the catch-all could never add
coverage once the primary accepts everything.
Collapse the preference to two modes: FFmpeg (default), which demuxes
every progressive container ahead of media3's list, and media3 only as
the user-facing escape hatch in case a file misbehaves. media3's
extractors stay behind FFmpeg, so anything FFmpeg cannot sniff still
reaches them. The catch-all role and the Auto container list are
removed; unknown persisted wire values (including the retired "auto")
resolve to FFmpeg.
Verified on a Pixel 7: MP4, TS, and MKV all demux through FFmpeg under
the default ("sniff accepted mov,mp4 / mpegts / matroska,webm") with
forward and backward seeks landing; JVM suites and analyzer pass.
Direct play on Android kept accumulating container patches: AVI with
XviD packed-bitstream timestamps and missing VOL csd played broken,
ASF/WMV and MPEG-PS/VOB had no media3 extractor at all, and MKV needed
a custom extractor stack for zlib-compressed subtitles, LOAS/LATM
audio, cueless seeking, and font attachments.
Demux progressive containers with libavformat behind media3's
extractor API. An AVIO bridge serves libavformat from the
ExtractorInput: every position divergence defers into a RESULT_SEEK
round trip, header reads replay from a block cache while
avformat_open_input restarts, and avformat_seek_file executes seeks
with bounded loader round trips. Packets feed media3's TrackOutputs,
so decoders, passthrough carriers, Dolby Vision RPU/EL conversion
(dvh1 codecs string surfaced from DOVI side data), and the subtitle
pipeline are untouched: embedded fonts feed AssHandler over JNI, ASS
reaches libass as per-sample dialogue, SRT/VTT render as cues, VobSub
maps through media3's VobsubParser, and LOAS/LATM AAC unwraps via
LatmTrackOutput.
Under the default Auto preference FFmpeg demuxes AVI, ASF/WMV,
MPEG-PS, and Matroska/WebM and sits behind media3's list as an
any-container fallback; MP4/TS keep media3's extractors. A playback
setting exposes auto / FFmpeg first / media3 only, and the JNI
load-failure path falls back to stock media3. The custom Matroska
extractor stack and its reflection keeps are retired.
Verified on Pixel 7 (API 36), Box R 4K Plus TV (API 34), and SHIELD
TV (API 30): instrumentation playback suites, the minified R8
reachability gate, and an on-device container matrix with forward and
backward seek landings plus visual subtitle checks.
close#2052
Plex's tag.tag- removes comma separated tags and has no escape syntax, so
removing a tag containing a literal comma ('Action, Comedy') split it at the
separator and could over-remove siblings named 'Action' or 'Comedy'.
Fields whose removals contain a comma now defer to one removal per request
after the main edit; every request restates the kept tags and the lock, so
the sequence converges regardless of order. Removals without commas keep the
exact single-request wire shape.
The Latest Albums see-all sheet offers a Date Added sort, but its rows were
requested without DateCreated, so addedAt mapped null and the sort silently
compared nulls — the same gap the earlier DateCreated work closed for catalog
and hub rows. DateCreated is a direct dto property, not one of the per-row
COUNT fields that motivated this set's slimness (#1552), so the cost profile
of these folder-dto requests is unchanged.
Queued tracker writes still replayed through the wrong account when the session died by token expiry instead of explicit disconnect: the auth-failure teardown cleared the store and rebound null but never purged the service's retry queue, so rows queued under account A replayed through whichever account connected next. The invalidation callback now purges like the disconnect path, after the rebind so an in-flight failure is dropped by the account-binding check instead of re-queueing behind the purge.
Narrower second hole in the same invariant: removeService swept the in-memory fallback outside the queue lock, so an enqueue whose persist failed while a disconnect raced it re-buffered the row after the sweep and the next flush resurrected it. The fallback add and the fallback sweep both run inside the queue lock now, so lock-slot order covers the buffered store the same way it already covered the persisted one.
Two callbacks with the same state could both exchange the authorization code upstream (concurrently or by replay); the session is now atomically claimed under the proxy lock before the exchange.
GET /posters/ and the /auth/result long-poll had no per-IP limiting or concurrency bound, and every poster lookup serialized through an exclusive store lock; both endpoints now use the established limiter and non-expired hits take a read lock. The limiter also tracks active transfers per IP with caps below the global limits (4 for fetches, 2 for uploads), because one unauthenticated client could otherwise take all 16 global poster-fetch slots and hold them through slow ServeContent reads, starving everyone else with 429s; concurrency checks precede bucket charges so a capped denial consumes no admission tokens.
/auth/result gets its own per-IP budget instead of sharing /auth/start's burst-3 bucket, where two concurrent NAT'd sign-ins 429'd on the fourth request. It is charged only after the poll secret validates (bogus requests keep the generic 410 and cost nothing), denials carry an honest Retry-After, and the Dart poller - which treated 429 as terminal and abandoned a valid session - retries them until the session's 10-minute lifetime expires.
Jellyseerr shifts MediaStatus codes 6/7 relative to Overseerr (6=blocklisted, 7=deleted vs 6=deleted), and the shared unconditional mapping misread both. The product is now detected from the presence of mediaServerType in /settings/public, persisted on the session, and used to resolve raw wire codes; blocklisted titles are shown as non-requestable.
Request statuses failed(4) and completed(5) previously decoded as pending, blocking re-requests and mislabeling failures; both are now modeled and active requests are defined positively as pending or approved. Seerr also marks a request Failed on arr-push failure while leaving the media status Processing, and precedence checked processing first, so failed titles rendered as Processing and the request sheet kept blocking re-requests; a failed request with no live pending/approved request now wins over stale pipeline status in both the catalog state and the sheet's blocked labels.
Reauthentication completing from a stale session snapshot wholesale-replaced the session and could downgrade a concurrently detected product discriminator back to unknown permanently (the settings cache never reapplies it); adoption now merges the known product.
Post-restart completion recovery joined directory/filename, losing the base-directory component the Task constructor strips on custom roots. An unresolvable queue head (offline server) broke out of the drain loop, stalling every other queued download; the drain now excludes it and continues. The download tree also read all-cancelled or partial containers as completed.
The queued-download resume was a one-shot on the first online client, so a persisted row skipped because its server was offline stayed queued until restart. The resume now re-fires whenever a server comes online that the last resume did not cover (including reconnects), and a resume landing mid-drain re-drives the pass instead of being swallowed by the processing guard.
Deleting an episode fanned out one network playback-extras request per sibling row; the reference scan is now cache-only, retaining thumbnails when a sibling cannot be resolved. Season/show deletion removes rows sequentially, so that conservative retention would have counted siblings queued in the same batch - a sibling, or the container row whose extras cache is always a miss, made every episode retain its thumbnails and orphaned the files once the batch's rows were gone. The fan-out threads the batch's rating keys through so rows scheduled to disappear neither retain thumbnails nor contribute in-use paths; genuinely surviving rows still trigger retention.
Queued tracker writes carried no account identity, so rows created under account A replayed against account B after a reconnect; explicit disconnect now purges that service's queue. An in-flight markWatched/reconcile that failed after the purge was re-enqueued anyway (the write scope only tracked profile generation) and replayed into whichever account connected next: Tracker.accountBinding (client identity, rebound synchronously on every disconnect) is captured before each write and re-checked before queueing a retry, and because the check sits in the same synchronous segment as the enqueue, a row that passes it is claimed ahead of the purge in the queue mutex and is still removed by it.
The memoized anime-list snapshot survived writes, so sequential rewatch updates computed from pre-write state; successful writes now evict the entry.
A device-code dialog dismissed by system back never cancelled the poll (blocking new attempts until the deadline), and a user-cancelled sign-in showed the connection-failed snackbar; both now route through one cancel path.
Companion-remote reconnects derived their intent from a session status the peer overwrites mid-join, so a failed reconnect attempt often never rescheduled; the attempt now carries its own intent flag.
Library loads dropped retained state on failure. A delta load that failed for every requested server removed their libraries while adding nothing, and an in-place refresh where one server succeeded and another timed out replaced the whole list with only the successful response, wiping the failed server's sidebar entries even though it stayed out of the loaded set. Both paths now key retention to the servers that actually responded: unreachable servers keep their entries and are refetched on the next status emission, removed servers still drop.
Discord RPC kept the dead client after a disconnect, making every 30s reconnect a no-op for the rest of the session; disconnect now tears the client down so the timer builds a fresh one. Rapid disable/enable while a connect was still initializing also let the old client's failure dispose the fresh one, so each attempt binds its listeners and cleanup to its own client instance and stands down when it is no longer current.
A response cancelled before or during body consumption completed as a successful empty body, so cancellation-triggered aggregation could overwrite valid state and a cancelled download could commit an empty file. The managed client now delivers an abort error, and a request whose own abort has fired classifies secondary teardown errors as cancelled (timeouts keep their type). A Plex server whose /hubs leg failed was also counted as succeeded when only the optional music leg returned, letting Discover replace good hubs with a music-only result.
Removing a tag containing a space or symbol silently did nothing: the removed values were pre-encoded and the transport encoded them again, so Plex matched nothing while the UI showed the tag gone. Folder listings and the shared list wrapper also converted every transport failure (401, 500, timeout) into a successful empty list, so an unreachable server rendered 'no folders'/'no results' with no retry while Jellyfin surfaced an error. Failures now propagate to the existing error paths, and the match screen gained one.
Disconnecting a gamepad while holding a D-pad direction or a deflected stick left the synthetic key repeat running forever, since only window blur tore that state down. The disconnect event now runs the same teardown: stop repeats, release held keys, clear pressed/suppressed sets and stick latches.
Jellyfin gates DateCreated behind explicit Fields, so browse/hub rows mapped addedAt as null: the Date Added sort was a no-op and unplayed Next Up rows sorted dead last in merged Continue Watching. Both base field sets now request it.
Editing a release date re-posted the full DTO with the stale ProductionYear; the year now follows the edited date and clears with it.
The guide requested programmes by MinStartDate, dropping anything already airing; the lower bound is now MinEndDate so overlapping programmes stay in the window.
With two profiles bound to two Jellyfin/Emby users on one server, a legacy bare machine-id lookup (offline downloads, cached playback metadata) could resolve another user's cache row, serving their resume position and token-stamped image URLs. Mirror the write-path guard: when the bare scope matches more than one user, log and return nothing.
A 2-letter preferred language (es, ar, ru, ...) prefix-matched unrelated 3-letter track codes (est, arm, rum), so profile-based audio/subtitle selection could pick a wrong-language track whenever it preceded the wanted one in container order. Route the matcher through languageMatches, which compares exact codes, region variants, and ISO-639 variations.
Recording indicators never appeared in the EPG and the program sheet kept
offering "Record" for already-subscribed programmes, so rules were created
repeatedly (nine duplicate subscriptions in the linked report). Two defects,
both confirmed against the official web client bundle:
- The sheet's scheduled-state check called
/media/providers/<identifier>/media/subscriptions/mapping/<ratingKey>,
which 404s: PMS mounts that route under the numeric MediaProvider id from
/media/providers. The check failed instantly and unconditionally on every
platform, so the button always read "Record".
- The guide derived its dot solely from cross-matching grab metadata, and
the grab refresh issued right after scheduling raced the server
materializing the grab, wiping the optimistic key before one frame showed.
The grid response already tags subscribed airings with subscriptionID /
grandparentSubscriptionID - the signal the official client renders from.
LiveTvProgram now carries those attributes; the guide checks them first
(grab matching stays as the secondary signal) and keeps local
schedule/cancel actions authoritative until the next grid load. The sheet
resolves Record vs Manage from the tagged rule key without a network call,
falling back to the fixed numeric-id mapping route for untagged airings.
Grab parsing also accepts the nested airing under "Video", which PMS uses
for every non-scheduled grab status in JSON.
close#2009
The most frequent Dart crash in production: CompanionRemoteProvider's
async crypto rebuild can call stopListening() on a discovery service
that was already disposed, and _emitHosts() then threw 'Bad state:
Cannot add new events after calling close'. Emitting after dispose is
now a no-op.
Backgrounding the Android app or locking the screen near-always dumped the
companion remote back to the device-selection page, forcing a manual
re-connect. Three defects combined: the peer's trailing disconnected status
event knocked the session out of reconnecting and broke the retry chain, the
bounded backoff budget (~31s) burned out against restricted background
networking before the user returned, and nothing retried on resume.
The provider now ignores stale peer status/error events while a reconnect
cycle is active, defers retries while backgrounded instead of consuming the
budget, and on resume retries immediately with a fresh budget - or pings a
nominally-connected session so a dead socket fails into the reconnect path
right away. Authenticated socket errors in the peer now surface as
disconnects like clean closes, keeping both terminal signals on the
reconnect path.
close#2035
Plex "Other Videos" libraries returned items as type=movie with
subtype=clip, which mapped to MediaKind.movie and rendered 2:3 poster
cards that cropped the generated 16:9 video-frame thumbs.
Items keep the movie kind so downloads, add-to, delete-from-server, and
detail navigation stay available, but subtype=clip now renders wide with
the thumb-first artwork clips already use. Sections marked subtype=clip
in /media/providers map to the clip library kind, giving Plex home-video
libraries the same folder-first grouping and wide grid cells as
MediaBrowser homevideos views. Wide-only hubs also keep 16:9 cards in
the poster episode modes, matching the TV rail's gate.
close#2036
Two private formatters mapped the canonical server resolution label to
display text (media_quality_labels, media_version) with edge-case
drift, and several single-consumer helpers added indirection without
callers:
- add shared resolutionDisplayLabel in resolution_label.dart and
migrate MediaVersion.displayLabel and the quality-label builder to it
- drop the resolutionLabelFromHeight compat re-export from
jellyfin_mappers (no remaining consumers) and its stale doc note
- delete TraktCatalogSource.membershipKeysFor, byte-identical to the
CatalogWatchlistMachinery default
- derive the rating sheet backend label from
MediaBrowserDialect.productName instead of a hardcoded switch
- inline EndpointFailoverManager into failover_http_client.dart, its
only importer, and remove endpoint_failover_interceptor.dart
MAL token refresh waited on the general 20s request timeout while the
Trakt and MDBList refresh paths use the dedicated 15s refresh timeout,
so a hung refresh held the 401 retry longer on MAL than on its
siblings. Align MAL with the shared refreshTimeout.
Finishing an episode of a show set to rewatching stomped the entry back
to watching. The scrobble write now reads the entry's status and rewatch
count in the same request as the episode count, keeps REPEATING /
is_rewatching entries rewatching, starts a rewatch when progress lands on
a completed entry, and bumps the rewatch count when a rewatch completes.
close#2026
Swipe navigation felt sluggish next to native tvOS apps: Plezy priced a
focus step at 0.55x the focused item's extent, throttled repeats to 140ms,
and stopped dead on finger lift - the inverse of the native engine's feel.
Retuned AppleTvRemoteTouchService to on-device measurements of the native
focus engine (issue #2006):
- one focus step per fixed 400pt of touch travel; drop item-extent scaling
(measured step distance was identical across a 230pt and a 345pt axis)
- repeat cadence 140ms -> 60ms (native median)
- fast lifts glide one or two extra steps (>=2000 / >=8000 pt/s) within
~130ms of the lift; reversal pivots and new touches cancel the glide
Posters and episode thumbnails looked noticeably soft on Fire TV sticks and
similar boxes: the auto-detected reduced tier capped artwork DPR at 1.5 and
tightened tile decode caps, so every tile was fetched at 56% of its pixels
and upscaled twice on large 4K panels.
Drop the reduced-tier DPR cap and the poster/square/thumb decode caps so
tiles fetch and decode at full TV density on every tier; raise the reduced
image-cache budget to the 64MB TV baseline to absorb the larger tiles.
Backdrops keep their scrim-masked ~720p reduced caps, and the display budget
factor stays pinned to 1.0 there, so the low-RAM art ceiling is unchanged.
close#2020
Searching a name shared by a movie, a show, and audio drowns the wanted
item in episode and track rows. The search screen now offers kind chips
(All, Movies, TV Shows, ...) above the results; chips derive from the
pre-rank candidate pool, and selecting one re-ranks that kind's
candidates with the full display budget, so a kind crowded out of the
trimmed All view still surfaces everything the servers returned for it.
Chip switches never refetch, and the filter falls back to All when a
refined query no longer returns its kind.
close#2001
Removing an item from Continue Watching on Emby posts HideFromResume,
which only the dedicated /Users/{uid}/Items/Resume route honors: the
IsResumable /Items query and the per-series /Shows/NextUp lookups kept
returning hidden rows, so removed items blipped back on every refresh.
Emby's dedicated route conflates both shelves (in-progress items plus
one zero-position next episode per started series), so every playback
surface now reads one hide-aware window from it and splits the rows by
PlaybackPositionTicks: positive rows feed Continue Watching, zero-
position episode rows feed Next Up. The 1+N per-series Next Up
reconstruction is deleted; one recency scan still supplies the play
dates and the NextUpDateCutoff window. Jellyfin request strings are
unchanged.
close#2003
Browsing artwork showed PMS PhotoTranscoder log lines flagged
"upscaled: 1" on every request. The flag merely echoes the upscale=1
query param Plezy sent on every cover request (PMS prints it before
locating the source, including on cache hits and pure downscales), but
the param had a real cost: whenever the requested box exceeded the
source — hero art at the 4K display budget (3840x2160 requests against
the near-universal 1920x1080 agent art), portrait backdrops on >1080p
phones — PMS genuinely enlarged the image server-side. Verified against
PMS 1.43: ~2.5x the transcode time and transfer bytes for zero rendered
detail, since the client covers the slot on the GPU regardless and the
fit-policy decode bounds never enlarge.
Send upscale=0 on cover requests too: downscale output is byte-identical
and oversized requests now return the native image. Jellyfin needs no
change — MaxWidth/MaxHeight never enlarge.
close#1975
Closing the window now runs a graceful exit (so trackers get their
terminal scrobble), but the teardown kept the window on screen for
seconds while screens visibly emptied as servers disconnected.
Hide the window as the first step of exit teardown, bounded so a
stalled platform channel cannot hold an already-accepted exit open, and
cover every graceful-exit entry (window X, Cmd+Q, OS-initiated), not
just the close button. Add a terminal MultiServerManager.shutdown() that
closes the status streams before draining clients: the still-mounted
widget tree must never observe an "all servers gone" snapshot mid-exit,
which would flip the app into offline UI and dismantle screens (and
with it the teardown-driven AppLifecycleListener used-after-dispose
assert). A health result landing after shutdown is a silent no-op.
Supersedes the flag-based approach in #1999.
edde746