`emitNormalizationDebug` (debug-mode trace forwarder, 15+ internal call
sites) and `isInOrbitSession` (Orbit-active guard used by next() and the
async fallback paths to suppress local queue extensions, 5 call sites)
move into dedicated modules under src/store/. Both were file-private —
no caller-side changes outside playerStore's own imports.
playerStore 3434 → 3415 LOC.
`deriveNormalizationSnapshot` — the loudness / replaygain / off branch
that the runtime uses to compute the normalization fields on every track
switch and queue rewrite — moves into src/store/normalizationSnapshot.ts.
File-private, four internal call sites updated to import from the new
module. Adds 8 focused tests pinning each branch (off, loudness with
target LUFS, replaygain enabled with tag + pre-gain, replaygain enabled
with fallback) plus neighbour-track context for album-mode resolution.
playerStore 3470 → 3434 LOC.
Three file-private helpers + their shared module-scoped track id move
into src/store/playbackUrlRouting.ts: recordEnginePlayUrl,
playbackSourceHintForResolvedUrl, shouldRebindPlaybackToHotCache. The
`lastOpenedWithHttpTrackId` mutable goes with them — only those three
read it. No external callers, no re-exports needed.
Adds 12 focused tests covering the source-kind classifier
(stream / hot / offline), the rebind decision across stream:-prefix
forms, the empty-serverId / un-recorded edge cases, and the
test-only reset helper.
playerStore 3490 → 3470 LOC.
The high-frequency PlaybackProgressSnapshot channel — type, mutable
snapshot, listener Set, emit/get/subscribe — moves into a dedicated
src/store/playbackProgress.ts. playerStore re-exports the public 3
(getPlaybackProgressSnapshot, subscribePlaybackProgress, the type) so
the 5+ external callers (PlayerBar, FullscreenPlayer, WaveformSeek,
LyricsPane, MobilePlayerView, TauriEventBridge) keep their existing
imports.
Direct unit tests pin the delta short-circuit (`currentTime <0.005`,
`progress/buffered <0.0002`) that keeps idle CPU bounded. The
end-to-end Tauri-event drive path remains covered by the existing
playerStore.progress characterization test (unchanged).
playerStore 3516 → 3490 LOC.
Three small tranches out of playerStore.ts:
- src/utils/waveformParse.ts — waveformBlobLenOk + coerceWaveformBins (the
parser that handles number[] / Uint8Array / ArrayLike payloads Rust
serializes as).
- src/utils/normalizationCompare.ts — normalizationAlmostEqual (tolerant
null-aware dB comparison).
- src/utils/seekErrors.ts — isRecoverableSeekError (retry classifier for
the Rust seek pipeline).
Each gets focused unit tests. All four were file-private, no external
callers — pure code move. playerStore 3556 → 3516 LOC.
Move the bounded undo/redo stacks, snapshot factory, scroll-top reader
registry, and pending-scroll-top channel into a dedicated module under
src/store/. playerStore re-exports the three public APIs that QueuePanel
and the test harness already imported (registerQueueListScrollTopReader,
consumePendingQueueListScrollTop, _resetQueueUndoStacksForTest), so no
caller-side changes are needed.
The new module adds explicit push/pop accessors so the undo/redo store
actions stop reaching into module-scoped arrays directly. Inline
`pendingQueueListScrollTop = …` writes inside applyQueueHistorySnapshot
become `setPendingQueueListScrollTop(…)` calls — same effect, scoped
through the module.
PlayerState gets the `export` keyword so queueUndo.ts can type its
state-shaped parameters.
playerStore.ts 3618 → 3576 LOC. queueUndo gains a focused test file
covering snapshot deep-cloning, the redo-stack invalidation on a fresh
undo push, max-size enforcement, and the scroll-top channel.
Pre-PR check: PASS.
Four pure helpers move out of playerStore: normalizeAnalysisTrackId,
sameQueueTrackId, queuesStructuralEqual, shallowCloneQueueTracks.
Adds focused unit tests for the stream:-prefix normalization and the
no-op detection that prevents unnecessary queue rewrites.
Behaviour preserved verbatim. playerStore 3618 → 3598 LOC.
Final piece of the App.tsx slim-down. The `TauriEventBridge` component
— ZIP download progress, track-preview lifecycle, audio device
changed/reset, the full `cli:*` listener surface (audio-device-set,
instant-mix, library list/set, server list/set, search, player-command),
tray-icon visibility sync, in-app keybindings keydown handler, media
keys + tray actions, `shortcut:global-action` / `shortcut:run-action`,
seek-relative / seek-absolute / set-volume, the window:close-requested
+ app:force-quit flow (with the shared `performExit` Orbit teardown),
and the `psysonic --info` snapshot publisher — moves into
`src/app/TauriEventBridge.tsx`.
`MainApp` imports it from the new file. `App.tsx` is now 57 LOC: just
the `App()` default-export that branches between `MiniPlayerApp` and
`MainApp` after wiring the shared theme / font / track-preview
document attributes.
A.K.A. App.tsx 1453 → 57 LOC over Phase 2 (M0 + B.1 + B.2 + C.1 + C.2
+ D). Pure code move at every step — no behaviour change.
Pre-PR check: PASS (frontend tests, tsc, coverage gates, prod build,
backend tests, clippy, backend coverage gates).
Continues the App.tsx slim-down. Two pieces move out of the monolith:
- `src/app/AppRoutes.tsx` — the route table and its 32 lazy page imports.
AppShell now renders `<AppRoutes />` inside the existing `<Suspense>`;
the `perfFlags.disableMainRouteContentMount` placeholder stays in
AppShell because that branch is a layout concern, not a routing one.
`useIsMobile()` moves inside AppRoutes so the `/now-playing` mobile
swap stays self-contained.
- `src/app/RequireAuth.tsx` — the 4-line auth guard, with a focused test
that covers all three reject paths (no login, no active server id,
empty server list) plus the happy path. MainApp imports it from the
new file instead of routing through the App.tsx re-export.
Side-cleanup of imports that B.2 had already orphaned in App.tsx
(`version`, `initAudioListeners`, `lazy`, `Routes`, `Route`, `Navigate`,
`MobilePlayerView`).
`App.tsx` 1308 → 1232 LOC. `AppShell` stays in App.tsx for Phase C.2.
Pre-PR check: PASS (frontend tests, tsc, coverage gates, prod build,
backend tests, clippy, backend coverage gates).
The 186-LOC default export shrinks to a thin window-kind switch with
shared document-attribute hooks. The mini-player tree and the main-app
tree each move into their own module under src/app/.
- src/app/MiniPlayerApp.tsx (48 LOC):
DragDropProvider + MiniPlayer + cross-window storage sync
- src/app/MainApp.tsx (129 LOC):
BrowserRouter + Routes + main-only lifecycle hooks
(audio listeners, hot cache, global shortcuts, mini-player
bridge, easter egg, scrollbar auto-hide)
AppShell + RequireAuth + TauriEventBridge are now named exports from
App.tsx so MainApp can compose them; Phase C/D will extract those into
their own modules.
App.tsx: 1453 -> 1308 LOC. Behaviour-preserving.
Moves four self-contained helpers into src/utils/, each with co-located
characterization tests. playerStore re-exports them for the ~30 existing
call sites; Phase E will migrate those imports.
- shuffleArray (Fisher-Yates, generic)
- resolveReplayGainDb (track/album/auto mode resolution)
- songToTrack (Subsonic -> Track shape)
- buildInfiniteQueueCandidates (Instant-Mix top-up source)
playerStore.ts: 3732 -> 3618 LOC (-114).
24 new tests covering the smaller action surfaces F1 / 2a-c skipped:
- setStarredOverride / setUserRatingOverride (per-id maps)
- openContextMenu / closeContextMenu (state + isOpen flip)
- openSongInfo / closeSongInfo (modal state)
- toggleQueue / setQueueVisible (visibility flip with persisted side effect)
- toggleFullscreen (boolean flip)
- setLastfmLoved (writes verbatim + caches under title::artist when track is
set; does NOT cache without a track) + toggleLastfmLove (no-op without
track or session-key; flips state + cache otherwise)
- setLastfmLovedForSong (caches under title::artist key)
- setProgress (currentTime + derived progress)
- stop (invokes audio_stop, resets playback bookkeeping)
- shuffleQueue (no-op when queue < 2; deterministic with mocked RNG,
current track stays at queueIndex 0)
- shuffleUpcomingQueue (no-op when upcoming < 2; head + current untouched,
upcoming tail shuffled)
- pruneUpcomingToCurrent (drops everything after queueIndex; clears the
queue entirely when no current track; early return when queue is already
empty)
- setRadioArtistId (does-not-throw smoke; no public getter for the
module-level state it writes)
playerStore.ts coverage 40.48% -> 48.02% lines (functions 37.84% -> 50.34%).
Close to but not over the F1 50% line floor — the remaining ~2pp lives in
playTrack's async hot-cache/replay-gain body, which is the same surface PR
2c flagged for a separate follow-up. Not adding playerStore.ts to the gate
yet; staying out one more PR to see the number stabilise.
Frontend suite: 451 -> 475 tests (+24).
39 new tests targeting the response normalization paths the F3 PR (#544)
deferred. Mocks axios at the module boundary; pins:
- api() helper envelope: unwrap subsonic-response on status=ok, throw
"Invalid response" without envelope, throw the server message on
status=failed, throw a generic on failed-without-message, propagate
network failures.
- song-array vs single-object normalization paths:
- getMusicDirectory normalizes child (object -> [object]), empty -> [].
- getMusicIndexes flattens index -> artist arrays (object or array).
- getMusicFolders coerces numeric ids to strings + defaults name.
- getRandomSongs pass-through behaviour pinned (no normalization --
Navidrome always returns the array form).
- collection-shape contracts: getAlbum splits { album, songs } + empty
fallback when album.song is absent, getStarred returns empty arrays
on missing starred2 / missing fields / pass-through arrays.
- single endpoint behaviours: getSong null on failure, getTopSongs []
+ slice to 5, getArtists flatten + empty, search whitespace-query
short-circuit (no HTTP), getAlbumInfo2 null on error, ping
true/false based on status.
- pingWithCredentials (explicit-URL path): full response (type +
serverVersion + openSubsonic), http:// prepend when scheme missing,
trailing-slash strip before /rest/ping.view, ok=false on any
failure / status=failed, openSubsonic defaults to false when omitted.
subsonic.ts coverage 12.66% -> 31.87% lines (+19pp). The remaining
surface lives in niche endpoints (playlist mutations, statistics
overview/aggregates, internet radio CRUD, cover-art uploads, ratings
prefetch). Not gate-eligible yet; a follow-up could push to ~60% but
diminishing returns relative to other backlog items.
Frontend suite: 412 -> 451 tests (+39).
MiniPlayer.test.tsx (4): mounts without throwing, renders the
always-present titlebar controls (Pin + Open main window — Close is
Linux-only and lives on the manual smoke list per pick 4a), click on
Open main window / Close does not throw, clicking the Pin button
flips the alwaysOnTop label between "Unpin" and "Pin on top". Bridge
contract (mini:ready / mini:sync, geometry persistence) deferred to
B-tier phase B5 -- jsdom does not model two webviews.
FullscreenPlayer.test.tsx (9): renders the labelled Fullscreen Player
dialog + Close Fullscreen control. Control wiring: Close calls
onClose, Stop calls stop, Previous calls previous, Next calls next,
Repeat cycles via toggleRepeat.
§4.5 of the v2 plan -- useCachedUrl(coverUrl, coverKey, false)
regression. Mocks the CachedImage module so the call args are
observable. Pins:
- the 500 px cover-art call passes opt=false (no fetchUrl fallback;
prevents double crossfade fetchUrl -> blobUrl);
- the 300 px art-box call passes opt=true (default behaviour).
A refactor that "tidies up" the useCachedUrl call sites would silently
regress the FS player cover; this test makes it loud.
Harness fix: vi.mock for @tauri-apps/api/event now returns an async
emit that resolves -- components chain .catch() on emit which crashed
on the bare vi.fn() return value during first-render useEffect.
Benefits any future component test that mounts something using emit.
Frontend suite: 399 -> 412 tests (+13). F5 (and Phases F0-F6) complete.
QueuePanel.test.tsx (8): empty-queue affordance, one row per queue track
with matching data-queue-idx, track titles render. Toolbar exposes
Shuffle queue / Save Playlist / Load Playlist / Copy queue share link /
Clear via aria-label; Shuffle is disabled when the queue has fewer than
2 tracks.
§4.4 of the v2 plan -- DnD architecture pin. Queue rows do NOT declare
draggable=true (no HTML5 native drag); the source file has no
dataTransfer.setData / dataTransfer.getData / onDragStart / onDrop JSX
usage; no application/json MIME anywhere. The project's psy-drop custom
event system sidesteps WebView2's text/plain-only restriction by
avoiding HTML5 DnD entirely -- a refactor that re-introduces native DnD
on the queue would silently break Windows.
PlayerBar.test.tsx (9): renders the labelled "Music Player" region.
Surfaces Previous Track / Play / Next Track / Repeat / Stop when a
track is loaded. The middle control flips between "Play" and "Pause"
based on isPlaying. Control wiring: clicking Play/Pause calls
togglePlay, Previous calls previous, Next calls next, Repeat cycles
off -> all -> one, Stop calls stop. The region landmark is still
rendered when no track is loaded.
Adds a generic scrollIntoView no-op stub in src/test/mocks/browser.ts --
jsdom does not implement it and QueuePanel's queueAutoScroll calls it
on mount with auto-advance enabled. Benefits any future component test
that touches scroll affordances.
Frontend suite: 382 -> 399 tests (+17).
ContextMenu.test.tsx (11): renders nothing when closed, renders when
openContextMenu has run, closeContextMenu hides on next render. Song-type
shows Play Now / Play Next / Add to Queue; album-type shows Open Album
/ Play Next / Enqueue Album / Go to Artist; artist-type shows Start
Radio + share affordances; queue-item shows a Remove option the song
menu does not. Clicking an action calls the expected store method
(playNext, enqueue) and closes the menu. Escape on the menu closes it.
WaveformSeek.test.tsx (8): renders a canvas, cursor=default without
trackId (no-track-loaded affordance), cursor=pointer with a trackId.
Wheel guards: no seek without a trackId, no seek with duration=0. Wheel
commit wiring: 350 ms trailing debounce delays the seek call until
activity settles; rapid wheel events coalesce (fewer commits than
events). Mount + unmount completes without throwing.
Neither component joins the hot-path gate -- jsdom skips canvas drawing
and these files have large branch surfaces (ContextMenu has 13 menu
types, WaveformSeek has the animation loop + 11 seekbar styles) that
need behavioural smoke rather than line coverage. The tests cover the
contract the refactor must preserve; visual / interactive layers stay
on the manual smoke list.
Frontend suite: 363 -> 382 tests (+19).
Adds 22 new tests on top of the existing 7 _on* / stopPreview ones.
startPreview happy path: invokes audio_preview_play with the configured
args (id, url, durationSec, startSec, volume) and stores the previewing
track + duration + reset elapsed / audioStarted. Short tracks
(duration <= previewDuration * 1.5) start at 0; longer tracks seek to
duration * trackPreviewStartRatio. Camel-case IPC keys pinned
(startSec / durationSec, not snake_case -- CLAUDE.md gotcha).
Cross-store guard tests: no-op when previews globally disabled, no-op
when disabled at the calling location, no-op while a host or guest is
inside any Orbit phase (active / joining / starting), falls through to
play when role is null (no session).
Same-id re-click: treats it as a stop -- audio_preview_stop fires,
audio_preview_play does not.
Failure path: engine invoke rejects -> store state rolls back
(previewingId / previewingTrack / audioStarted) and the error propagates
to the caller.
Loudness pre-attenuation folding: with normalizationEngine=loudness +
loudnessPreAnalysisAttenuationDb=-6 dB, volume is multiplied by
10^(-6/20). normalizationEngine=off keeps volume verbatim. Positive
pre-attenuation values are pulled to 0 by the Math.min(0, ...) guard.
Main-player volume sync side-effect (module-level
usePlayerStore.subscribe): pings audio_preview_set_volume when volume
changes during a preview, skips when no preview is active, skips when
the new value equals the prior value (subscription guard).
previewStore.ts coverage 33% -> 100% lines. Added to the hot-path gate.
Plus the typed `OrbitRole` is `'host' | 'guest'` (null when no session),
not 'idle' as a string -- minor type-correctness alignment.
subsonic.contract.test.ts (21): parseSubsonicEntityStarRating (userRating
first then rating fallback, numeric-string coercion, undefined for null /
NaN / non-numeric), libraryFilterParams (empty without active server, empty
on "all" filter, returns musicFolderId on specific filter), getClient
(throws without a server, returns baseUrl + auth params, rotates token + salt
across calls), coverArtCacheKey (serverId:cover:id:size shape, "_" fallback
without active server, no ephemeral salt embedded -- stays cacheable),
buildStreamUrl (URL shape + Subsonic auth params: id u t s v=1.16.1
c=psysonic/* f=json, rotates t/s across calls so Rust matches by id, special
character ids encoded once not twice), buildCoverArtUrl (default size=256),
buildDownloadUrl (download.view path), trailing-slash + scheme handling on
base URL.
resolvePlaybackUrl.test.ts (15): precedence offline > hot-cache > stream
(first priority wins even when later sources also have the track), forwards
trackId + serverId to both stores. getPlaybackSourceKind for offline / hot
/ stream / engine-preload-hint cases. streamUrlTrackId parser (id from
stream.view query, null for non-stream URLs / no query / missing id, decodes
URL-encoded ids, manual-query fallback for relative paths).
copyEntityShareLink.test.ts (5): writes a psysonic2-prefixed payload that
round-trips, returns false without an active server, returns false on
empty / whitespace id, trims surrounding whitespace before encoding,
propagates clipboard-failure return.
Gate broadens with src/utils/resolvePlaybackUrl.ts (95.8 %) +
src/utils/copyEntityShareLink.ts (100 %). subsonic.ts at 12.7 % stays out
-- the URL-builder + parser surface this PR covers is the structural part;
the async API endpoints need axios mocking, deferred to a follow-up.
authStore.ts (79 %) and playerStore.ts (40 %) deferred-list comments
updated to reflect F2 + F1 actuals.
login.test.ts: composed addServer -> setActiveServer -> setLoggedIn flow,
failed-login leaves prior state intact, addServer assigns unique ids,
setConnecting / setConnectionError independence, logout clears
isLoggedIn + musicFolders but keeps server entry, Last.fm session
(setLastfm / connectLastfm / disconnectLastfm contracts).
servers.test.ts: addServer / updateServer (patch by id, no-op on unknown),
setActiveServer (clears musicFolders), removeServer (non-active no-effect,
active picks newServers[0] fallback, last server -> null + isLoggedIn
false, cleans per-server bookkeeping maps), getActiveServer / getBaseUrl
selectors. Includes the gapless / crossfade mutex regression test from
the v2 plan section 4.3 -- callers clear the other flag before setting,
the setters themselves do NOT auto-clear (contract pin).
persistence.test.ts: hydration loads existing localStorage shape, defaults
missing fields, preserves saved values verbatim. Robust to corrupt JSON
and missing top-level state. onRehydrate migrations: clears conflicting
hotCacheEnabled + preloadMode!=off legacy combo, keeps hotCache when
preload was already off, migrates legacy waveform seekbarStyle to truewave,
strips removed animationMode / reducedAnimations fields. partialize strips
musicFolders. Includes the synchronous-storage invariant regression from
the v2 plan section 2 (CLAUDE.md gotcha) -- getActiveServer is visible in
the same tick after addServer + setActiveServer with no await.
settings.test.ts: API-pin sweep across 22 trivial setters via it.each
(rename-resistant), focused tests for setters with logic (clamping in
setTrackPreviewStartRatio / setTrackPreviewDurationSec / setRandomMixSize,
boolean coercion in setTrackPreviewsEnabled, finite-number guard in
setLoudnessPreAnalysisAttenuationDb, default reset in
resetLoudnessPreAnalysisAttenuationDbDefault), per-server bookkeeping
contracts (setEntityRatingSupport / setAudiomuseNavidromeEnabled
positive-opt-in semantics), enum-value setters
(setPreloadMode / setDiscordCoverSource / setLoggingMode / setReplayGainMode /
setNormalizationEngine / setLyricsMode), genre blacklist + audio output
device replacement.
authStore.ts coverage 45.37% -> 79.29% lines (target was 60%). authStore.ts
not yet added to the hot-path gate -- want to see it stable across a few
real PRs first per the gate's curation rule.
progress.test.ts: getPlaybackProgressSnapshot shape + post-emit reflection,
subscribePlaybackProgress (notify, (next, prev) pair, near-duplicate epsilon
coalesce at 0.005 s, unsub stops notifications, multiple subscribers
independent), live-emit throttling guard (drops within 1500 ms + < 0.9 s
delta; large delta passes; time threshold passes).
persistence.test.ts: flushPlayQueuePosition forwards (ids, currentId, posMs)
to savePlayQueue, caps song-id list at 1000, no-ops on radio / no-track /
empty queue, swallows backend errors, floors position to whole ms.
playerStore.ts coverage 39.55% -> 40.48%. F1 50% floor not met -- remaining
~10pp lives in playTrack's async hot-cache/replay-gain body, shuffleQueue,
stop, enqueueRadio, initializeFromServerQueue and the orbit auto-merge
paths. Either a follow-up PR 2d or a revised F1 floor; flagged in the PR
body. Gate unchanged -- playerStore.ts stays out until the floor is met.
Discovered + fixed during this PR: the spread ...await vi.importActual()
mock pattern lets the real savePlayQueue leak through to playerStore.ts's
relative import (../api/subsonic) even when the alias-form is mocked.
Switched persistence.test.ts to an explicit non-spread mock map listing
every export the store uses.
trackShape.test.ts: songToTrack mapping (required, optional, replayGain
flattening, missing-block fallback, no invented flags); resolveReplayGainDb
precedence (disabled, track, album, auto with prev/next neighbour match,
missing albumId fallback, both gains missing); shuffleArray (non-mutating,
preserves multiset, copies, empty/single, deterministic under mocked RNG).
queue.test.ts: enqueue / enqueueAt (auto-added separator placement,
queueIndex shift, index clamping), playNext (playNextAdded tag, position,
empty input), clearQueue (state reset, audio_stop call), reorderQueue
(splice, currentTrack-id-following queueIndex), removeTrack (clamp on
shrink, keep when after cursor), undo / redo (empty stacks return false,
rollback, replay, new edit drops pending redo).
Adds a 5-line test-only export _resetQueueUndoStacksForTest in playerStore.ts
because the undo/redo arrays live at module scope outside the Zustand state
graph; storeReset.ts now calls it so resetPlayerStore is fully complete.
playerStore.ts coverage 4.76% -> 18.1% lines. PR 2b + 2c push further toward
the F1 floor of 50%.
* fix(orbit): event-driven host push on play/pause flips
Without this, the worst-case delay between "host hits pause" and "guest
stops" is two full polling windows (host's 2.5 s push tick + guest's
2.5 s read tick, plus network) — long enough for the guest to noticeably
run past the host. Subscribing to playerStore.isPlaying changes adds at
most one extra remote write per flip; non-flip state ticks still ride
the existing 2.5 s timer. The listener filters on isPlaying so the
per-second currentTime ticks don't trigger spurious pushes.
* fix(orbit): lock seekbar for guests — sync follows the host
Guests could drag/click/wheel the seekbar, which would jump the local
player and then snap back at the next host poll (2.5 s of inconsistent
UX) — or push the guest into a diverged state where Catch Up was the
only way back. The seekbar is host-controlled in Orbit; the guest input
path now reflects that.
- App.tsx exposes `data-orbit-role="host"|"guest"` on the root element
alongside the existing `data-orbit-active` marker.
- WaveformSeek's container gains a `.waveform-seek-container` class so
CSS can target it.
- Guest rule: `pointer-events: none` on children blocks click / drag /
wheel / hover; the parent keeps `cursor: not-allowed` + reduced opacity
so the disabled state is visually unambiguous.
Hosts and non-orbit users see no change.
* docs(changelog): credit PR #537 (orbit sync latency + guest seekbar)
* Add drag-and-drop reordering and visibility toggles for queue toolbar
* docs(changelog): credit PR #534 (queue toolbar customization)
Adds the v1.46.0 CHANGELOG entry and a new bullet on kveld9's
contributors block in Settings → System.
---------
Co-authored-by: Psychotoxical <171614930+Psychotoxical@users.noreply.github.com>
* fix(orbit): host single-track playTrack appends instead of replacing
Reported by cucadmuh: "When playing from offline, the Orbit queue
doesn't get appended to — it gets overwritten."
Root cause: the Orbit bulk-guard fires only when `queue.length > 1`.
A `playTrack(track, [track])` call (one example: OfflineLibrary's
"Play this album" on a single-track album, but any UI that passes an
explicit 1-track replacement queue triggers it) slips past the guard
and replaces the host's `playerStore.queue`. The host's queue *is* the
shared Orbit queue — replacing it wipes every guest suggestion + every
upcoming track in one click.
Re-route to append + jump when role is host: if the track is already
in the queue, jump to that slot; otherwise append it and jump to the
new tail. The guest path is intentionally left alone — guests opting
out of host-sync via a local Play is the existing "guest running
their own show" divergence behaviour. `useOrbitGuest`'s `syncToHost`
is the only guest-side caller of `playTrack(track, [track])`, and it
never matches `role === 'host'` so it's never intercepted.
* docs(changelog): add host single-track Orbit-queue protection bullet
* fix(orbit): make initial-sync seek visually stick on join
Reported: "I join the room, the waveform shows the host's live
position for a second or two, then snaps back to 0:00 and audio is
still playing from the start. Only after that snap can I press Catch
Up."
Two compounding causes in `useOrbitGuest`:
**1. Poll fires `applyMirror` before the engine is genuinely playing.**
`playTrack` flips `isPlaying` to `true` *synchronously* in its
optimistic store write, so the post-`playTrack` poll satisfied its
"engine ready" check before the Tauri `audio_play` had even
started producing sample. The `seek` inside `applyMirror` updates
the store position immediately (waveform jumps to host's live
position), but `audio_seek` is debounced and lands on a not-ready
engine — it silently no-ops. The engine then starts playing from
0, and its first progress events overwrite the optimistic seek
position, snapping the waveform back. Add `currentTime > 0.1` to
the poll's "engine genuinely playing" condition: once audio has
flowed past the cold-start barrier, the seek commits and the
seek-target guard correctly filters subsequent progress events.
**2. `applyMirror`'s play-state mirror raced its seek**, same shape
as the `onCatchUp` race fixed in #527. `player.seek` debounces
`audio_seek` via setTimeout(0) while `pause`/`resume` invoke
synchronously — pause arriving first leaves the engine paused at
the old position. Defer the play-state mirror by 200 ms so the
seek lands first.
* docs(changelog): add initial-sync seek-stickiness bullet
* fix(orbit): keep Catch Up button visible long enough to click
Reported during testing: the Catch Up button "appears for a moment,
then disappears too fast to click". The single-stage debounce only
filtered the show direction (require ≥ 3 s sustained over-threshold
before showing), but on a real-world high-latency session the genuine
drift fluctuates between ~1 s and ~8 s in lockstep with both sides'
chunked `currentTime` updates — so the button vanished as soon as
drift dipped briefly under 3 s, even though the drift baseline was
still 5–8 s.
Add a hide-side hysteresis: once visible, the button stays visible
until drift has been under a tighter 1 s threshold for ≥ 1 s.
Otherwise the noisy 1–3 s drift valleys keep the button up so the
user can actually click it.
Constants left as locals; if testing wants different floors we can
extract.
* docs(changelog): add catch-up hysteresis bullet
* fix(orbit): show Catch Up when host paused + serialise seek-then-pause
Two follow-ons after the hysteresis change:
**1. Show Catch Up even when the host is paused.** The visibility
gate required `state.isPlaying === true`, but a guest who joined
while the host was paused still benefits from Catch Up if their
sync to the host's paused position failed (engine drop, brief
network blip during initial-sync, manual local seek). Drop the
`state.isPlaying` gate — the only signal that should matter is
"drift between us and the host's reported state" — which
`computeOrbitDriftMs` already computes correctly in the paused
case (no time-extrapolation, just `guestPos - hostPos`).
**2. Defer the play-state mirror after `player.seek`.** Reported
symptom: "I press Catch Up, the player pauses; when I press play,
it resumes from the *old* position and the waveform jumps back."
Root cause: `player.seek` debounces `audio_seek` via
`setTimeout(0)`, while `player.pause` / `player.resume` fire their
invokes synchronously. Pause arrives at the engine first, leaving
it paused at the pre-Catch-Up position with the seek queued
behind. When resume happens later, the engine plays from the old
spot. Push the play-state mirror behind a 200 ms `setTimeout` so
the seek's invoke lands first.
* fix(orbit): kick a fresh playTrack when engine is stuck mid-load
Follow-up to #525. Symptom: occasionally on join, the guest gets no
audio until the next host-driven track change, at which point a
fresh `playTrack` runs and audio plays normally.
Root cause: when the initial `syncToHost` poll hits its 5 s deadline
without the engine reporting `isPlaying === true` (slow Navidrome
cold-start), the next pull tick takes the cheap "track already loaded"
shortcut and calls `applyMirror`. `applyMirror` fires `seek` + `resume`
on an engine that is stuck in a "loaded but never started" limbo —
seek silently no-ops and `resume` can't kick a track that never began.
Guest is silent until something else triggers a fresh `playTrack`.
Tighten the shortcut: take the cheap path only when the engine is
already in the state the host expects (or playing while host is
paused, which is fine to align via pause). Otherwise fall through to a
fresh `playTrack` and the existing 5 s ready-poll, which re-initialises
the engine and lets audio actually start.
* docs(changelog): add engine-limbo follow-up bullet
* fix(orbit): re-sync when engine silently fell back to paused
The optimistic `isPlaying: true` `playTrack` writes synchronously
masks a `audio_play` failure: the post-playTrack poll sees the
optimistic flag, fires `applyMirror`, and the outer tick records
`lastAppliedRef = { ..., isPlaying: true }` as a successful sync.
But if the underlying `invoke('audio_play')` rejects later (network
blip / cold-start exhaustion), the catch handler flips the flag back
to `false` and schedules `next()`, which short-circuits to `audio_stop`
in an Orbit guest — leaving the player silent while `lastAppliedRef`
still claims we're playing. None of the divergence-detection branches
(track-change / play-pause-flip) match, so the guest never re-syncs.
Add a recovery check before the if/else-if chain: when the captured
`last` says we applied playing, the engine is currently not playing,
and the host is still playing the same track, reset
`lastAppliedRef.current = null`. The next iteration of the chain
re-runs initial-sync (with the 500 ms fast-poll cadence) which fires
a fresh `playTrack`. If `audio_play` succeeds the second time, audio
finally starts; if it keeps failing, we loop with no extra harm
(audio was already silent).
Adds an `engine-recovery` event scope to the diagnostics buffer so
future captures can see when this re-sync fired.
* docs(changelog): expand engine-recovery bullet to cover both cases
* fix(orbit): guest short-circuits queue-exhaustion fallback paths
When a guest's local queue runs out (single-track queue from `syncToHost`
empties on `audio:ended`), the player walks the standard fallback chain
in `next()`: radio top-up → infinite-queue → stop. The infinite-queue
branch builds a 6-track queue and calls `playTrack`, which trips
`orbitBulkGuard` and pops a "Add 6 tracks to the Orbit queue?" modal.
Hitting Cancel leaves playback frozen; "Add them all" injects unrelated
tracks into the host's shared queue.
In an active Orbit guest session the host owns the queue. Skip the
fallback paths entirely and just stop — the next `useOrbitGuest` pull
tick will sync to whatever the host advanced to.
Bonus side-effect: kills the deferred-promise race where a
`buildInfiniteQueueCandidates().then(...)` from a guest's track end
could resolve *after* a Catch Up replaced the queue and pop the modal
a second time against the now-current 1-track queue.
* fix(orbit): treat natural track-end as not-diverged in guest sync
When a guest's track ended naturally before the host advanced, the
divergence-detection branch read `player.isPlaying === false` and
classified it as the user manually paused — so it refused to load the
host's next track. The guest sat silent until they clicked Catch Up.
`handleAudioEnded` keeps `currentTrack` pinned to the just-ended track
and resets `currentTime` to 0, while a real manual pause leaves
`currentTime` somewhere mid-track. Use the 0-position discriminator to
classify natural-end as not-diverged so the host's new track loads.
Confirmed via the captured guest log buffer:
18:43:08.598 [track-change] host: VJkV5… → 6i6RP… BUT guest diverged
(player.isPlaying=false ≠ last.isPlaying=true)
— guest stuck for ~33s until Catch Up was pressed.
* fix(orbit): Catch Up polls until engine is ready before seeking
The 400 ms blind setTimeout in `onCatchUp` was too short for an
HTTP-streamed cold-start on high-latency links. If the audio engine
wasn't ready by then, `seek(fraction)` silently no-oped and playback
started at 0:00, making Catch Up effectively useless on exactly the
slow links where it's needed. Captured log shows a Catch Up bringing
the guest to posSec=30, then 5 s later the guest was at posSec=6
(playback restarted from the head).
Replace with the same poll-until-ready pattern `syncToHost` already
uses: check every 100 ms, fire the seek as soon as the engine reports
playing, fall back to a blind apply at the 4 s deadline.
* docs(changelog): add orbit guest playback fixes entry
* fix(orbit): debounce Catch Up button + match bar item height
Two follow-on UX fixes after PR #525's three primary bugs landed:
1. **Debounce visibility.** Drift is computed from an asymmetric signal:
guest's `currentTime` updates in coarse ~5 s chunks, while host's
position is extrapolated linearly via `(nowMs - posAt)`. Even on a
perfectly-synced session the diff swings ±5 s every tick, so the
button flickered in and out continuously. Show only after drift has
stayed over the 3 s threshold for ≥ 3 s of wall clock — measurement
noise is filtered out, real sustained drift still surfaces in time.
2. **Match neighbour height.** The button was 32 px tall against 26 px
for the other action buttons (.orbit-bar__settings) so every flicker
shifted the entire bar height. Set `height: 26 px` and tighten the
padding/font so the layout is stable regardless of visibility.
* fix(orbit): tighten queue-extension lockout + reliable initial-sync seek
Two follow-on fixes after the 4-bug umbrella:
**1. Local queue-extension paths fully off during Orbit.**
Phase check broadened from `active` to cover `starting` / `joining` /
`active` so a fetch-then-join race can't pop the bulk-add modal *after*
the join. The proactive infinite-queue topper inside `next()` (which
fires when ≤ 2 auto-tracks remain ahead) is now also gated, plus each
async `.then()` callback in the radio + infinite-queue paths re-checks
at resolution time. A `playTrack(... 6-track queue ...)` after the user
joined Orbit was the path that re-triggered the "Add 5 tracks?" modal
on a freshly-joined guest.
**2. `syncToHost` only seeks once the engine reports playing.**
The previous 2 s deadline-fallback applied the seek even when the
engine hadn't started, where the seek silently no-ops and the track
plays from 0:00. Symptom: clicking Catch Up makes the song "jump 50 %
forward" — that's the seek finally landing because the engine is now
ready, the initial-sync seek had already failed silently. New deadline
is 5 s, and on timeout we return `false` so the outer pull tick keeps
`lastAppliedRef` null and the 500 ms fast-poll retries.
* fix(orbit): double-click play button + hide preview during session
Two cucadmuh-flagged gaps:
**1. Double-click on the inline play button now reaches the orbit-add
path.** The album-track row's onDoubleClick already routes to
`addTrackToOrbit` when in Orbit, but the inline play button stopped
propagation on click — so clicking it twice just fired the "double-
click to add" hint toast and never touched the orbit queue. Add an
onDoubleClick on the button itself that delegates to the parent's
`onDoubleClickSong`.
**2. Track preview is suppressed during an Orbit session.** Preview
shares the Rust audio engine with the shared playback, so starting
one as a guest yanks the host's track out from under everyone. A new
`[data-orbit-active]` attribute on `<html>` (set whenever role is
host/guest and phase is starting/joining/active) hides every
preview button via a single CSS rule, and `previewStore.startPreview`
short-circuits as a defensive guard for keyboard shortcuts and any
programmatic callers.
* feat(orbit): in-app diagnostics popover with copyable event log
Multiple users on Discord report Orbit guests stopping after the first
song with no errors anywhere — Settings → Debug → Export Logs is too
buried for non-technical reporters, and the relevant code branches
have no logging at all (silent fail). This adds a one-click "Copy log"
path right inside the Orbit session bar.
The new Activity-icon button next to Help opens a popover with:
- Live mini-display: role, host vs. guest track id + position, drift,
age of the host's last state write — all updating once a second.
- Scrolling event log textarea fed by an in-memory ring (200 events).
- Copy + Clear buttons. Copy formats `[ISO] [scope] body` lines and
drops them on the clipboard — paste straight into a Discord report.
Instrumentation lands at the previously-silent decision points:
- Guest pull tick: full snapshot of host vs. guest state on every read.
- Each branch of the divergence detection in `useOrbitGuest.ts` logs
which path it took and why (initial / track-change-followed /
track-change-diverged / play-pause-flip), making the
"stuck after first song" symptom diagnosable from the buffer alone.
- Host pushes log track id, isPlaying, queue length, guest count.
Events are also bridged to the existing `frontend_debug_log` Tauri
command when Settings → Logging is on Debug, so power users still get
the same data in `psysonic-logs-*.log` for offline triage.
i18n: full `orbit.diag.*` namespace in all eight locales. EN + DE are
native; ES / FR / NB / NL / RU / ZH are first-pass and may want a
polish from native speakers later.
* docs(changelog): add orbit diagnostics popover entry
* fix(context-menu): drop inline z-index that hid menus under floating player
The main context menu wrapper carried an inline `zIndex: 999` that
overrode the `.context-menu { z-index: 10000 }` stylesheet rule. The
floating player bar sits at z-index 1000, so when a menu opened near
the bottom of the screen — long enough or anchored low — the player
bar covered it.
Removing the inline override lets the stylesheet rule (10000) win, so
the menu always paints on top of the floating bar. Submenus inherit
the same stacking context so they follow the parent menu.
Closes#521.
* docs(changelog): add context-menu floating-player z-index fix entry
* fix(home): swap Because-you-listened rail to AlbumRow under 696 px
The hero-style BecauseCards are tuned for full-rail widths (3 cards at
1052 px+, 2 cards at 696-1051 px). Below that the cards stretched
full-width with a fixed 160 px cover stuck on the left and centred text
floating in a wide empty area — looked like three over-sized banners
stacked vertically instead of a compact recommendation rail.
A `ResizeObserver` on the rail wrapper now watches the container width
and below 696 px renders a standard `AlbumRow` (which is already
perf-tuned for narrow rails: artwork budget, viewport windowing, scroll
paging). Wide layouts keep the unchanged hero card layout, so the
mainstage view at full width is identical to before.
* docs(changelog): add Because-you-listened narrow-layout fix entry
* fix(i18n): OpenDyslexic font supports Cyrillic
The OpenDyslexic subtitle in the font picker stated "no RU/ZH support",
but the bundled `@fontsource/opendyslexic` 5.x ships Cyrillic glyphs and
Russian renders correctly — verified empirically. Only Chinese (CJK)
actually falls back to the system font. Updated the subtitle string in
all 8 locales accordingly; the Russian locale itself no longer claims
unsupported-self.
* docs(changelog): correct OpenDyslexic locale-coverage claim (#513)
The #507 entry in the unreleased v1.46.0 block claimed `Latin +
Latin-extended only` and listed Cyrillic among the system-font
fallbacks. Cyrillic glyphs ship in @fontsource/opendyslexic 5.x and
render correctly — only Chinese (CJK) actually falls back. Updated the
bullet in place to match the corrected subtitle hints from this PR.
* feat(player-bar): album context menu on song title right-click
Right-clicking the track title in the player bar now opens the same
album context menu that album cards use (open, play next, enqueue,
go to artist, favorite, rate, share, download, add to playlist).
Mirrors the existing left-click behavior on the title, which already
navigates to the album. Suppressed for radio and preview, matching
the click handler.
MarqueeText gains an optional onContextMenu prop; PlayerBar builds a
SubsonicAlbum shape from currentTrack on demand.
* docs(changelog): add entry for PR #512 (player-bar title context menu)
* refactor(settings): collapse all sections, drop font dropdown, surface OpenDyslexic
Settings opened on a tab where four or five sub-sections were expanded
on first render — audio device, theme list, lyrics sources, sidebar
customizer, random-mix copy, offline dir, language picker, keybindings
table. The page felt like a wall of controls before the user had even
looked for something specific. Removed every `defaultOpen` flag from
the SettingsSubSection call sites so each tab now boots with only the
section headers visible. Component default was already `false`.
ThemePicker auto-expanded the group containing the active theme on
mount. Same noise on a screen that already has the longest accordion
list in the app, and the blue dot in the group header already tells
the user which group holds the active theme. Initial open-group is
now `null` — all groups collapsed until the user clicks one.
Font picker had a dropdown-style button that toggled a list inside
the sub-section, which meant two clicks (open the section, then open
the dropdown) for what should be a one-click choice. Removed the
button + the `fontPickerOpen` state — opening the Font sub-section
now reveals the full list directly and a click sets the font without
collapsing anything. OpenDyslexic moved to the top of the list so
users with dyslexia don't scroll past 14 sans-serifs to find their
option; the rest stays in the original order.
* docs: changelog entry for PR #508
Logs the Settings collapse-by-default + font picker cleanup +
OpenDyslexic ordering in v1.46.0 "## Changed".
* feat(settings): OpenDyslexic font option for dyslexic readers
Next step on the accessibility track. The first pass was on the colour
side — WCAG contrast audits across every theme and dedicated colour-
vision-deficiency variants for the protanopia / deuteranopia / tritan-
opia palettes. Typography is the other axis: some users with dyslexia
find a font with a heavier weighted baseline and asymmetric glyph
shapes (b/d, p/q never mirror, italic forms differentiated rather
than slanted-regular) easier to track than a typical sans.
Adds OpenDyslexic to the existing Fontsource font picker. SIL OFL
licensed, freely redistributable, and the de-facto open-source
standard for this use case. Non-variable axis, ships as four discrete
weight/style files (regular, bold, italic, bold-italic) — the Settings
picker grew an optional `hint` field on font entries so this one row
can carry a "dyslexia-friendly · no RU/ZH support" subtitle without
bloating the other 14 entries.
Latin + Latin-extended only. Cyrillic and CJK locales (RU, ZH) fall
back to the system font when this is selected; the subtitle calls out
that limitation upfront. i18n: hint string in all 8 locales
(settings.fontHintOpenDyslexic).
Accessibility is intentional product positioning here — it's an
underserved corner of the Subsonic-client ecosystem.
* chore(nix): sync npmDepsHash with package-lock.json
* docs: changelog entry for PR #507
Logs the OpenDyslexic font option in v1.46.0 "## Added".
* docs(settings): contributor entry for PR #507
Adds the OpenDyslexic accessibility bullet to Psychotoxical's
contributions list.
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
* wip(home): Lossless rail + dedicated /lossless-albums page
Rail under Home > mostPlayed and a dedicated infinite-scroll page that
list albums whose tracks are tagged in lossless containers. Walks
Navidrome's native /api/song?_sort=bit_depth&_order=DESC, dedupes by
albumId on the way down, stops when the song stream crosses into lossy
(bitDepth==0) or the server runs out of rows. _filters has no operators
on quality columns, so a sort + walk is the only path; equality on
sample_rate / bit_depth probes returned empty (verified).
The page paginates through the song-cursor with an in-flight ref so
overlapping IntersectionObserver fires don't double-add albums, plus a
cancelled flag so React StrictMode's double-mount doesn't apply two
parallel result sets in dev.
Suffix allowlist excludes ambiguous wrappers — m4a/m4b can be ALAC
*or* AAC and Navidrome's response carries an empty codec field, so we
can't tell them apart; same story for wma. Allowlist is flac, wav,
aiff/aif, dsf/dff, ape, wv, shn, tta — containers that are only
lossless. ALAC-in-m4a setups will miss out, acceptable trade-off
without a reliable codec field.
Status: WIP. Settings home-customizer label and en.ts strings landed,
no other locales yet, no quality badge on AlbumCards across the rest of
the app, no CHANGELOG. Branch was 'feat/home-hires-rail' during the
earlier hi-res-only iteration before the lossless broadening.
* wip(home): Lossless page header parity + streaming + sidebar nav
Page header now mirrors All Albums: selection mode with three action
buttons (Enqueue Selected, Add Offline, Download ZIPs) wired to the
same handlers Albums.tsx uses, selection-counter title swap, and the
perfFlags.disableMainstageStickyHeader respect path. Filters were
intentionally skipped — the rail is sorted by bit_depth, mixing client-
side filters with server-driven pagination would produce gaps.
Loading feels noticeably faster: ndListLosslessAlbumsPage takes an
optional onProgress callback that fires once per internal fetch with
the entries discovered in that fetch, so the page can stream new
albums into state instead of waiting on the whole loadMore. Page-side
budget dropped from 5×200 to 2×100 songs per loadMore (~1 MB worst
case vs 5 MB before), since the rail's catch-em-all pass is wrong for
infinite-scroll UX. Subtitle under the page title primes the user that
this is slower than other album pages because Psysonic walks the song
catalog by quality (Navidrome ignores _fields, so per-song responses
ship with lyrics + tags + participants whether we want them or not).
Sidebar nav entry registered under 'losslessAlbums' with a Gem icon,
defaults to visible:false (matches composers / folderBrowser /
deviceSync — niche browsing modes). Existing users get the entry
appended at the end of their persisted sidebar list automatically via
the onRehydrateStorage merge that sidebarStore already runs for new
DEFAULT_SIDEBAR_ITEMS.
i18n: full coverage across all 8 locales for sidebar.losslessAlbums,
home.losslessAlbums, losslessAlbums.empty, losslessAlbums.unsupported
and the new losslessAlbums.slowFetchHint subtitle. ru/zh are
machine-translation quality, flagged for a polish pass.
* feat(home): default Lossless sidebar entry to visible
Flips the DEFAULT_SIDEBAR_ITEMS entry for `losslessAlbums` from false
to true. Existing installs keep whatever the user has in persisted
storage; fresh installs see the entry in the sidebar from the start.
Earlier wip commit defaulted it off (matched composers / folderBrowser
/ deviceSync as a niche browse mode), but the rail + page do show
something useful for any library with at least one FLAC/WAV/etc. album,
so off-by-default just hid the feature.
* docs: changelog entry for PR #506
Logs the Lossless Albums rail + page + sidebar entry in v1.46.0
"## Added".
* docs(settings): contributor entry for PR #506
Adds the Lossless Albums rail/page bullet to Psychotoxical's
contributions list.
* feat(song-info): show absolute file path on Navidrome via native API
Subsonic's `getSong.view` returns at most a relative path (or none on
Navidrome), so the Path row in the Song Info modal stayed empty for
most users. Feishin and the Navidrome web client surface the full
server-side path by hitting Navidrome's native `/api/song/{id}` instead.
Added an `nd_get_song_path` Tauri command that logs in to the native
API with the active server's credentials, fetches the song, and returns
the `path` string. Wired it into `SongInfoModal`: the Subsonic
`getSong` call still drives the rest of the dialog, and the native call
runs in parallel only when the active server's identity is
"navidrome". When it returns a path, that absolute path replaces the
relative Subsonic value; native-API failures are silent and the modal
falls back to whatever Subsonic provided.
No token cache yet — the modal is opened occasionally enough that one
fresh login per call is fine.
Closes discussion #479.
* docs: changelog entry for PR #504
Logs the Navidrome native-API absolute file path support for the
Song Info dialog in v1.46.0 "## Added".
* docs: settings contributor entry for PR #504, drop competitor mention
Adds the song-info absolute path bullet to Psychotoxical's contributors
list in Settings, and rewords the existing CHANGELOG entry so neither
text references competing clients.
* fix(radio): queue navigation, dedup, and similar-first variety (#500)
After a Radio session ran a while, three things broke:
Queue navigation through duplicates. playTrack re-resolved the active
queue index by `findIndex(t.id === track.id)`, returning the *first*
matching id, so reaching the second occurrence of a track snapped
queueIndex back to the earlier slot — highlight jumped and the next
advance played the wrong follow-up. Added an optional
`targetQueueIndex` to playTrack, threaded through next(), previous(),
the audio:ended repeat-one path, queue-row click, and the queue-item
context menu. findIndex stays as the fallback for callers that just
have a track and a fresh queue.
Queue accumulation. enqueueRadio didn't dedupe incoming tracks; the
next() top-up deduped against the live queue but trimmed the played
tail down to HISTORY_KEEP=5, so a song heard 8 ago was gone from
`existingIds` and a later Last.fm/topSongs response could re-add it;
and the `.filter(...)` pass admitted intra-batch repeats (top +
similar overlap is common) because it read the dedup set before
mutating it. A module-level radioSessionSeenIds set, fed by
enqueueRadio and both top-up paths and reset on artist change and
clearQueue, closes all three: trimmed ids stay in the set, ids about
to be replaced (fresh enqueueRadio wiping the pending radio block)
are removed first so callers can re-introduce them, and the dedup
pass mutates the set inline.
Variety. Starting Radio on a track stacked five top tracks of the
seed artist before any similar-artist material played. Switched the
seed path and both top-up paths to lead with similar songs (other
artists) and only fall back to top tracks when similar comes back
empty — preserves the "no Last.fm" graceful degradation but stops
the seed artist from monopolising the front of the queue.
Not affected: gapless audio:track_switched (already index-based, no
findIndex), AudioMuse Instant Mix / Lucky Mix (single-element queues
or enqueue-only paths), the artist-radio path (no seedTrack — already
picks just one top track and fills the rest from similar).
Reported by netherguy4.
* docs: changelog entry for PR #503
Logs the radio queue navigation/dedup/similar-first fix in v1.46.0
"## Fixed".
* fix(preview): keep preview sink volume in sync with player slider (#498)
The Rust preview sink had its volume set once at audio_preview_play and
then never updated. audio_set_volume only ramps the main sink, so slider
movements during a preview had zero effect on the preview level. With
the default loudness normalization (-4.5 dB pre-analysis attenuation)
applied at start, even a 100% slider gives 1.0 × 0.596 × MASTER_HEADROOM
≈ 53% — matching the user-visible "fixed at around 50%" symptom.
- Add audio_preview_set_volume Rust command that updates the preview
sink if one is active (clamp + master headroom mirror the path used
in audio_preview_play).
- Extract the preview-volume calculation in previewStore into
computePreviewVolume() so startPreview and the new sync path share
one formula (slider value, plus the LUFS pre-analysis attenuation
the engine already applies to the main sink).
- Subscribe to playerStore at module level: when volume changes and a
preview is active, push the recomputed value to Rust. Auth /
normalization tweaks during preview are intentionally not synced —
preview is short and that case is rare.
Reported by netherguy4.
* docs: changelog entry for PR #502
Logs the preview-volume-slider sync fix in v1.46.0 "## Fixed".
* refactor(settings): remove redundant Animations 3-state setting under Seekbar Style
The `animationMode` setting (Full / Reduced / Static) duplicated work
the perf-flag system and OS-level reduced-motion preference already
covered:
- `perfFlags.disableMarqueeScroll` already kills marquee scrolling on
demand, replacing what `static` mode used to gate.
- The `data-perf-disable-animations` html-level switch already strips
every `*` animation, replacing what `static` mode used to do globally.
- `@media (prefers-reduced-motion: reduce)` honours the OS setting for
every user that asked for it via system preferences.
- The 30 fps cap that `reduced` mode applied to the seekbar wave was
better served by per-feature perf toggles cucadmuh added later.
Removed:
- `AnimationMode` type, `animationMode` field + setter from auth store.
- Settings UI block (3 buttons + hint text) under Appearance > Seekbar
Style.
- `animationMode === 'static'` short-circuit in WaveformSeek's rAF
effect; `isReduced` skip-every-other-frame logic; `static`-checks in
`drawNow` / `needsDirectDraw`.
- `animationMode !== 'static'` guard and `data-anim-mode` attribute in
MarqueeText.
- `[data-anim-mode="static"]` and `[data-anim-mode="reduced"]` rules in
layout.css.
- Seven i18n keys (animationMode + 6 variants) across all eight
locales.
Migration: the persist layer strips `animationMode` (and the legacy
`reducedAnimations` boolean predecessor) so anyone who had `'reduced'`
or `'static'` selected silently lands on the former `'full'` path on
first launch after upgrade. No user-facing prompt — the missing setting
just stops existing.
cucadmuh's PR #472 (FPS overlay), #476 (preview-freeze main seekbar,
sleep-recovery hooks, card-hover removal) and #486 (interpolation
anchor reset on resume) are all preserved untouched — they live in
separate effects / files and were not driven by `animationMode`.
* docs(changelog): add Removed section for animationMode setting (PR #495)
* docs(changelog): refine animationMode removal rationale (drop prefers-reduced-motion overstatement)
Two reports rolled into one pass:
1. Rail occasionally rendered nothing on Home open. Cause: the rotated
anchor sometimes had no Last.fm similar-artists or no library matches
among the sampled set, and the old code gave up after one try and
stored that dud anchor as the rotation cursor — so the next mount
started from the same dud's neighbour and could fail again.
2. Recommendations felt repetitive. Cause: pool of 12 anchors walked
round-robin made each anchor recur every 12 mounts, similar-artist
sample of 6 from 12 had heavy overlap visit-to-visit, and per-artist
single-album random pick meant artists with one library album always
surfaced the same record.
Anchor selection: random pick from pool with a per-server cooldown
buffer (last 5 anchors excluded, capped at floor(pool/2) so small
libraries don't soft-lock). Up to 4 anchors are tried in a shuffled
candidates list before giving up; the localStorage cursor only advances
on a successful anchor so duds don't poison future mounts.
Picks variety: similar-artist fetch raised from 12 to 25 (same
getArtistInfo call, larger response — Last.fm typically returns up to
~50). Per-server ring buffer of the last 30 shown album ids; per-similar
-artist album choice prefers an album not in that buffer, falling back
to any album when the artist's whole catalogue is stale so the slot is
never lost.
Pool cap raised 12 -> 20 to give the cooldown buffer room to breathe in
libraries with varied listening history.
Storage: legacy `psysonic_because_anchor:` single-id keys from the
round-robin era are stripped on module load (one-shot localStorage
sweep, the new `..._anchor_history:` prefix has a different colon
position so no false matches).
API budget unchanged in the hot path: 1 getArtistInfo + 6 getArtist
per Home mount. Worst case (3 dud anchors, 4th succeeds) is 4
getArtistInfo + 6 getArtist.
* feat(home): mix recently-played + starred into Because-you-like anchor pool
Anchor pool was sourced only from getAlbumList(frequent), so the rotation
cursor walked the same eight top-played artists no matter how varied the
rest of the listening history was. Round-robin merge of mostPlayed,
recentlyPlayed and starred (dedup by artistId) means each mount can land
on a different listening *mode* — heavy rotation, current focus, or
explicit favorites — instead of stepping through the same top-played
sequence.
Pool size 8 -> 12 to let the cursor visit all three modes before
wrapping. Visibility guard widened so the rail still renders when the
server has no frequent-play data yet but starred or recent items exist.
Zero new API calls — all three lists are already in Home's initial
fetch.
* fix(home): drop orphan 3rd Because-card in 2-col range, keep all 3 stacked on mobile
auto-fit grid wraps to 2 cols between 696-1051px container width, which
left the third card alone on a second row at 1080p. Container query
hides the 3rd card only inside that 2-col band; on wider screens the
full 3-up row stays, on narrow viewports (single column) all three
cards stack vertically as expected.
* docs(changelog): Because-you-listened seed pool + 1080p layout polish (PR #493)
* docs(changelog): fold PR #493 refinements into the existing Because-you-listened entry
Drop the separate Changed section entry — the feature is in the same
1.46.0 release window as PR #489, so readers want a single description
of the final behaviour, not "added X, then changed X" for the same
release. PR reference becomes "PRs #489, #493".
Frank Stellmacher
Kveld.
Maxim Isaev