Add a pure evaluateQconnectReconnectAdvisory helper that, on a
reconnecting->connected transition, replays a recently-issued idempotent intent
(play/pause/set_track) verbatim but only advises for next/previous to avoid a
double-skip; stale or non-reconnect transitions are no-ops. No invoke, fully
unit-testable.
Add the qconnect.lastActionMayNotHaveApplied i18n key to all five locales
(en/es/de/fr/pt) and emit a qconnect:resync_complete diagnostic from the backend
after the P1-8 post-reconnect resync so the frontend can run the advisory.
cc
On reconnect the bootstrap only re-asked for queue state, never for the active
renderer's full state, and a Lagged broadcast drop could swallow the
SESSION_STATE that carries the session_uuid / active_renderer_id.
Add free helpers ask_for_active_renderer_state + ask_for_queue_state and a pure
should_reask_queue_state(session_uuid_known, attempts, max) predicate. Wire:
- AskForRendererState after the reconnect Subscribed bootstrap and after the
deferred-join's final AskForQueueState (when an active renderer is known).
- Idempotent deferred-join: store last_joined_session_uuid; a repeat of the same
session_uuid skips re-announcing us but still re-asks renderer state.
- On a Lagged recv drop, re-AskForQueueState until the session_uuid is confirmed
or the attempt budget (5) is spent; the budget resets on disconnect.
No reconnect counters are reset (#358 latch untouched).
cc
Add a pure compute_connection_state(was_active, was_playing, ServerActiveState,
queue_equal) -> ConnectionDecision in session.rs implementing the plan's reduced
takeover matrix, plus a guarded CtrlSrvrSetActiveRenderer sender that only fires
for a valid id differing from the current active renderer.
Wire it on SESSION_STATE: the event loop captures was_active/was_playing and
classifies the server's active renderer from the incoming payload BEFORE
event_sink applies the new state, then runs the decision after the apply and
claims the active slot when warranted.
The exact web truth-table cell values were not dumped verbatim; the OtherPaused
vs OtherPlaying divergence and repeat-mode reconciliation are pending a
decompiled-bundle cross-check. Repeat-mode reconciliation is out of scope here.
queue_equal is defaulted false pending a controller/renderer queue diff. No
reconnect counters are reset (#358 latch untouched).
cc
The deferred renderer JoinSession always sent reason=CONTROLLER_REQUEST,
even on a rejoin after a transport drop. The event loop already tracked
has_disconnected but never threaded it into the deferred join.
Add JOIN_SESSION_REASON_{CONTROLLER_REQUEST,RECONNECTION} consts to mod.rs,
a pure deferred_join_reason(has_disconnected) selector, and thread the flag
through deferred_renderer_join so a post-drop rejoin carries RECONNECTION=2
while a first join keeps CONTROLLER_REQUEST=1. Transient-drop continuity
already works via the server-echoed session_uuid; restart-persistence of
session_uuid is intentionally deferred.
cc
Add `reconnect_idle_retry_ms` (default 0 = legacy terminate). When > 0, the
transport loop no longer breaks permanently on Exhausted: the new
`idle_retry_after_exhausted` helper idles (shutdown-cancellable), resets the
attempt counter / backoff to base, and retries (gap #7). The real config
sets 60s. The reset happens ONLY after Exhausted plus a deliberate idle
wait, never on connect, so the issue #358 latch is untouched.
Critical cross-file coupling: `service.rs`'s `MaxReconnectAttemptsExceeded`
handler previously dropped the runtime and broke the event loop. With
idle-retry active the transport stays alive and re-emits events, so the
handler now surfaces lifecycle Exhausted to the UI but keeps the runtime and
continues the loop (only the legacy path still drops + breaks). When the
re-armed transport later reaches SessionEstablished, the existing handler
moves the lifecycle back to Connected.
Startup auto-connect now retries on a bounded, monotonic schedule
(`startup_retry_schedule` = 2s/5s/15s/30s) instead of giving up after one
failed attempt (gap #8), re-resolving the transport config each iteration.
cc
Add `require_jwt` (default false) to `WsTransportConfig` and a
`MissingAuthenticateJwt` error variant. When `require_jwt` is set but no
`jwt_qws` is present, the connect attempt now emits a `TransportError`,
closes the socket, and routes through the existing bounded reconnect path
instead of silently skipping the AUTHENTICATE frame (gap #12).
The default stays false so the InMemory / test transport path keeps
working without a JWT; `resolve_transport_config` flips it on for the real
Qobuz endpoint. Reuses the increment-only `handle_reconnect_delay`, so the
issue #358 latch is untouched.
cc
A playing active-peer renderer that goes silent left the controller UI
stuck on stale 'playing' (no equivalent of the webplayer's 12s
rendererLost timer). Add a generation-guarded 12s watchdog keyed on the
active peer renderer in the RENDERER_STATE_UPDATED handler: reset on every
update, arm only while playing_state==PLAYING, disarm on
pause/stop/active-change/disconnect. On expiry freeze the projection to
UNKNOWN and emit qconnect:renderer_unreachable; Svelte shows an error
toast (rendererUnreachable, all 5 locales). Reuses the
freeze-projection-to-unknown helper from the ACTIVE_DISCONNECTED path.
cc
The wire 'status' field on RENDERER_STATE_UPDATED was decoded into the
adapter payload (decoder.rs:801) then dropped — a graceful peer
disconnect was invisible and the controller kept showing stale 'playing'.
Add a RendererStatus enum (UNKNOWN/UNRECOGNIZED -> INACTIVE collapse) and,
on ACTIVE_DISCONNECTED for the active remote renderer (id != -1), freeze
the projection to UNKNOWN and emit qconnect:renderer_disconnected. Svelte
surfaces it as an error toast (rendererDisconnected, all 5 locales). Pure
consumer-side: no protocol change.
cc
resolve_qconnect_device_uuid() generated a fresh Uuid::new_v4() per
process via a OnceLock, so every launch registered a new renderer
identity cloud-side (zombie renderers) and broke reconnect continuity.
Persist it beside device_name in qconnect_settings.db: generate once on
first run, reuse thereafter, keeping the QBZ_QCONNECT_DEVICE_UUID env
override for tests. Mirrors load_persisted_device_name/persist_device_name.
cc
v2_qconnect_send_command_with_admission trusted a single frontend origin and
never inspected track_ids. Add track_origins[] to the request and
validate_track_origins_for_admission: re-evaluate every track via the core
gate, block if any is non-Qobuz or the list is empty. The per-track check is
only enforced when the caller actually ships per-track origins (bulk queue
loads); single-track play-next/play-later carry no track_origins and stay
gated by the command-level origin alone, so an empty list never over-blocks
them. The existing command-level check and admission_blocked emit are kept.
Frontend loadQconnectQueue now computes per-track origins from its tracks arg
via resolveQconnectTrackOrigin and ships them to the sender.
Defers the optional DB cross-check (LibraryState / has_local_track_by_qobuz_id
defense-in-depth) as a follow-up; this commit is the per-track-origins
validator + wiring + test only.
cc
The Jack backend variant was added to qbz-audio but the backend-name
match in commands_v2/settings.rs was not updated, leaving the lib crate
non-exhaustive (E0004) and uncompilable on this branch. Add the missing
arm and sync Cargo.lock with the jack/jack-sys deps. Pre-existing breakage
unrelated to qconnect; fixed here only to unblock the build and tests.
cc
Move the Plex LAN client, hand-rolled XML metadata parsing, and the
SQLite plex_cache.db logic out of src-tauri/src/plex/mod.rs into a new
qbz-plex core crate so it can be shared by every frontend (Tauri, Slint,
TUI) per ADR-006. Plex previously lived entirely in src-tauri (~1715
lines), which blocked the Slint Local Library port from shipping Plex.
The extraction is mechanical: the module was already almost entirely
plain Rust (reqwest + rusqlite + serde + hand-rolled XML). The only
Tauri coupling was a dead plex_play_track command bound to AppState
(unreferenced; the live path is commands_v2::v2_plex_play_track, which
calls plex_resolve_track_media + AppState.player itself). That dead fn
and the vestigial #[tauri::command] attributes are dropped.
src-tauri/src/plex/mod.rs becomes a `pub use qbz_plex::*;` re-export so
every existing crate::plex::* call site compiles unchanged. Full Tauri
workspace cargo check stays green; qbz-plex tests pass.
cc
The Cloudflare proxy funnelled every user's MusicBrainz traffic onto a few
egress IPs, and MusicBrainz rate-limits 1 req/s per IP, so the budget was
shared across the whole userbase and produced 503s that worsen with growth.
MB read access needs no key and authentication does not raise the limit, so
the proxy added only the funnel; going per-IP makes the budget per-user.
- qbz-integrations MusicBrainzConfig default is now use_proxy=false (direct).
- client: handle HTTP 503 + Retry-After (not just 429, which the proxy used
to translate); add parse_retry_after; route search_recording_by_isrc and
search_artist through handle_response_status.
- src-tauri session_lifecycle: MusicBrainz V2 init/teardown pass use_proxy=false.
Slint inherits this via CoreBridge's default-constructed client.
cc
Completes Slice 0. src-tauri now depends on the qbz-offline-cache crate; the
old src-tauri/src/offline_cache/ implementation files are deleted. What remains
in that module is a thin shim: a re-export of the crate plus the Tauri glue.
- offline_cache/mod.rs: pub use qbz_offline_cache::* + tauri_cache_sink (maps
CacheEvent -> the exact legacy offline:caching_*/offline:unlock_* app.emit
shapes, so the Svelte frontend is unchanged) + core_client (resolves the core
qbz_qobuz::QobuzClient from the CoreBridge for the download pipeline).
- Cache/redownload/migration commands: drop the legacy crate::api::QobuzClient
(state.client) + the CoreBridge double-plumb; pass one core client + an
app.emit-backed CacheEventSink. Unused `state: State<AppState>` params dropped.
- Playback + library cached-resolution call sites: pass &tauri_cache_sink(..)
to load_cmaf_bundle_with_ui_events instead of &app_handle.
- session_lifecycle: apply_persisted_limit now reads the persisted value here;
purge takes &library.db directly.
- migration: qobuz_client is the Option-wrapped core client (consistent with the
downloader); None mid-loop is recorded as a per-track failure.
Full src-tauri cargo check passes; behaviour-preserving for the Svelte frontend.
Backend extraction (ADR-006) done — next is Slice 1 (shared tier-walk in core).
cc
qbzd was a premature daemon experiment that never shipped (gated off
behind QBZD_ENABLED, no user access). Its ratatui 0.29 dependency pinned
unicode-width =0.2.0, which conflicts with slint's annotate-snippets in a
shared workspace. Remove it entirely; it will be reimplemented from
scratch later on the clean, UI-agnostic shell.
- Delete crates/qbzd daemon crate + workspace member (drops ratatui,
crossterm and the other qbzd-only deps from the lock).
- Delete src-tauri qbzd_discovery (mDNS) + its commands/state wiring in
lib.rs; drop the direct mdns-sd dep (Chromecast keeps its own via
qbz-cast).
- Delete remoteApi/remoteEvents/playbackTargetStore; neutralize
commandRouter to local-only dispatch (isRemote/skipIfRemote are now
always-false no-ops so the ~100 guard call-sites keep compiling).
- Strip qbzd branches from playbackService and the remote banner from
NowPlayingBar; excise qbzd from CastPicker (Chromecast/DLNA kept).
- Remove qbzd i18n keys (5 locales), README daemon node, gitignore note.
cargo check (crates + src-tauri) and svelte-check all green.
cc
OOTB playback froze (seekbar advancing, no audio) on hosts where the
configured backend was not usable — e.g. PipeWire as the Linux default but
`pactl` absent (minimal Debian/Ubuntu). "Auto" (a null backend_type) fell
through to a legacy CPAL path that forced the track rate onto the shared
default device, producing a suspended/silent stream and leaving a
process-wide stuck audio handle that survived Reset.
- Add a cross-platform "System" backend (CpalDefaultBackend now on Linux
too): opens the OS default output at its negotiated rate with rodio
resampling and an explicit ~100ms CPAL buffer (MixerDeviceSink has zero
internal buffering, so the bare default underran and stayed suspended).
Shared, no exclusivity, no `pactl` — mixes with other apps like any
normal player.
- Make System the OOTB default and what Reset lands on; seed it on first
run. Drop the per-startup #375 NULL->PipeWire backfill migration.
- Resolve a null backend_type to System in the player so the legacy CPAL
path is never reached (covers legacy installs and headless callers).
- "Auto" in Settings is now a resolve-and-set action: it picks the best
available backend (PipeWire if available, else System), stores that
concrete value and updates the dropdown + device picker, so backend_type
is never null.
- Explicit PipeWire (DAC passthrough) and ALSA Direct (exclusive) configs
are untouched and remain bit-perfect.
cc
Add log_sanitize::redact_url, which preserves a URL's scheme, host, path
and query-parameter keys while masking every value and any fragment.
Apply it to the OAuth WebView navigation log lines (main on_navigation,
popup-window request, and popup on_navigation) so navigation traces stay
useful for debugging the flow without carrying parameter values into log
files or user-uploaded reports.
Covered by unit tests for code/token params, base64-padded values,
fragments, and query-less URLs.
cc
The offline-skip recursion walked the entire remaining queue synchronously,
desyncing the logical queue from the still-playing engine. Fixes:
- The gapless-transition target is the track the engine is already playing;
never run the availability-skip branch for it.
- isTrackAvailable now counts the in-memory/disk cache (new v2_is_track_cached
command), not just downloaded copies.
- Cap consecutive offline skips and play-failure auto-skips at 5 in both
+page.svelte and playbackService.ts.
cc
Long sessions died when a single transient blip flipped the app to offline and
the queue then skipped every cached-but-not-downloaded track. Harden the
detector:
- Consecutive-failure hysteresis: declare offline only after 2 failed polls,
not a single blip.
- Replace the HEAD https://1.1.1.1 probe with a concurrent race of cheap
generate_204 endpoints, so a saturated link (long Hi-Res tracks) no longer
times the probe out.
- Positive liveness override: if a segment downloaded within the last 45s we
are online by definition, regardless of the probe.
cc
Recovered from 63ecc324 (authored on the native-port branch, never merged to
pre-release). Backend-only, no Slint. Scales the prefetch cap down on slow
links and during audio underruns, and exposes the per-segment download signal
the offline detector reuses as a positive liveness override (#467).
cc
Pure-logic move of two helper modules out of src-tauri/src/musicbrainz/
into the qbz-integrations crate, where the MusicBrainzClient and DTOs
already live:
src-tauri/src/musicbrainz/location_discovery.rs
-> crates/qbz-integrations/src/musicbrainz/location.rs
src-tauri/src/musicbrainz/genre_normalization.rs
-> crates/qbz-integrations/src/musicbrainz/genre.rs
The functions exposed are unchanged: extract_metadata,
compute_affinity_score, build_scene_cache_key, format_life_span_date
(from location); normalize_genre, is_broad_genre, extract_affinity_seeds,
genre_summary (from genre). Imports in src-tauri/src/commands_v2/
integrations.rs are repointed to the new home; the legacy file
declarations are removed from src-tauri/src/musicbrainz/mod.rs. No
behavior change.
Unblocks the Slint side from reaching the same helpers via qbz-core
without depending on src-tauri.
cc
SearchAllResults and MostPopularItem now live in qbz-models; the Tauri
helpers re-export them under the legacy V2 aliases so command code and
its response serialization are unchanged.
cc
Extracts two self-contained, framework-agnostic modules out of
src-tauri so they can be reused by other clients without duplicating
logic:
- ImageCacheService / ImageCacheStats move into the qbz-cache crate
(LRU disk cache, MD5-keyed, SQLite index).
- The credential store (keyring + AES-256-GCM encrypted file for Qobuz
credentials and the OAuth token) moves into a new qbz-credentials
crate.
src-tauri/image_cache.rs and src-tauri/credentials/mod.rs become thin
re-exports, keeping their Tauri managed-state wrappers. The Tauri app
is behavior-identical; both modules were already self-contained (no
Tauri and no crate:: references).
cc
Moves the credentials module (keyring + AES-256-GCM encrypted file
store for Qobuz credentials and the OAuth token) out of src-tauri into
a new qbz-credentials crate. src-tauri/credentials re-exports it, so the
Tauri app keeps identical behavior while the Slint shell can share the
exact same implementation, encrypted file, and keyring entries.
The module was already fully self-contained (no Tauri, no crate::
references); only the dependency declarations move with it.
cc
Cover art now goes through the QBZ image cache instead of being
downloaded directly every time, matching the Tauri behavior.
- Moves ImageCacheService / ImageCacheStats into the qbz-cache crate
(LRU disk cache, MD5-keyed, SQLite index). src-tauri/image_cache.rs
re-exports it and keeps the Tauri managed-state wrapper, so the Tauri
app stays behavior-identical and both shells share one implementation
and one on-disk cache (~/.cache/qbz/images).
- qbz-slint artwork.rs checks the cache first, downloads and stores on
a miss, and trims the cache to 200MB on startup.
cc
Switch the session store SQLite connection from synchronous=NORMAL to
synchronous=FULL. With WAL + NORMAL the last writes can sit unsynced in
the WAL and be lost on a hard reboot / power loss, dropping the restored
session.
cc
The graphics recommendation banner no longer writes settings: the Apply button and the v2_apply_graphics_recommendation command are removed, so the banner only detects and explains. The user applies the change with the real toggles, matching the HiFi Wizard's advisory model. write_recommendation stays for the qbz --autoconfig-graphics CLI.
The Force DMA-BUF help text is now GPU-aware: the instability warning only shows on NVIDIA-only systems; non-NVIDIA and hybrid hosts are told it is safe to enable. On hybrid hosts the rendering GPU dropdown marks the integrated GPU as recommended.
cc
`/favorite/getNewReleases` includes albums whose stream availability is
still in the future, so they appear in the Release Watch grid even
though `get_album` 404s as soon as the user clicks them. Same shape for
regional restrictions and label takedowns surfaced via `streamable=false`.
Add `release_date_stream` and `streamable` to the Album struct (both
optional with serde defaults so existing call sites stay happy) and
filter the response inside `v2_get_release_watch` after the artist
blacklist pass: skip any album marked `streamable=false` or whose
`release_date_stream` is strictly greater than today (UTC). The total
count gets adjusted alongside so pagination math stays consistent.
cc
Two-pronged improvement to the graphics safety net for users hitting
the NVIDIA + Wayland + WebKit GBM issue:
1) New opt-in toggle 'NVIDIA Wayland compatibility (experimental)' in
Settings > Graphics. When ON:
- Sets __NV_DISABLE_EXPLICIT_SYNC=1 to bypass libnvidia-egl-gbm's
buggy implicit-sync paths that produce "Failed to create GBM
buffer of size NxN: Invalid argument" on driver 550+.
- Lifts the defensive compositing-mode shutdown that QBZ applies
to NVIDIA-only Wayland sessions — the user is signaling the real
fix is now in place, so the safety net is no longer needed.
- Persisted in graphics_settings.db column nvidia_compat_mode (new,
default off). Migrated via ALTER TABLE so existing installs don't
pick up a behaviour change.
2) Reset-graphics CLI affordance, surfaced only when the toggle is on:
inline copyable input + Copy button rendering `qbz --reset-graphics`
(or `flatpak run com.blitzfc.qbz --reset-graphics` based on detected
install method). Works around the global select-disable elsewhere
in SettingsView via user-select: all + onclick={e.select()} + the
explicit button. The boot watchdog still auto-reverts on crash, but
this is the manual escape hatch when the user wants to recover
without waiting for the watchdog to fire.
3) Watchdog recovery now lands in full CPU mode after ANY revert.
The single-knob revert (hw_accel / dmabuf / preferred_gpu) still
narrows the cause for the UI banner, but the safest post-crash
state is full CPU mode (HW accel off + DMA-BUF off + preferred_gpu
reset to auto). A partial revert can still hit the next crash if
the failure was stack-wide (driver bug, missing extension); going
all-the-way-CPU guarantees the user always reaches a paintable
window after a recovery — they can selectively re-enable from the
Settings banner.
4) preferred_gpu='software' now implies hardware_acceleration=off at
env-var application time. Forcing Mesa llvmpipe while keeping
WebKit GPU compositing on would just round-trip software pixels
through a "GPU" path that has no GPU — nonsensical configuration
that the watchdog had to clean up after. Now it's prevented at
source.
i18n keys (en/es/de/fr/pt): settings.graphics.nvidiaCompatMode,
nvidiaCompatModeDesc, resetCmdTitle, resetCmdDesc;
actions.copiedToClipboard.
cc
DiscordIpcClient::new became infallible in 1.0 — it now returns Self
directly instead of Result<Self>. Failures surface at connect() time
where the catch already lives. Removed the now-spurious `if let Ok()`
wrap around the constructor.
Closes#428 (applied manually because the bump needed a code change
that Dependabot can't author).
cc
vicrodh