Move 6 functions that bridge cloud-emitted QConnect events to the
local CoreBridge audio engine into a dedicated corebridge submodule:
- queue_state_needs_materialization (dedup comparator)
- apply_remote_loop_mode_to_corebridge
- apply_renderer_command_to_corebridge (SetState/SetVolume/MuteVolume/
SetLoopMode/SetActive handler — ~180 lines, the busiest function)
- sync_corebridge_remote_shuffle_projection (shuffle projection with
identity-shuffle deferral; remains private to corebridge)
- materialize_remote_queue_to_corebridge (queue replication, ~210 lines)
- align_corebridge_queue_cursor
mod.rs deps that corebridge.rs reaches via super:: are now pub(super):
- PLAYING_STATE_PLAYING/PAUSED/STOPPED/UNKNOWN constants
- normalize_volume_to_fraction, qconnect_repeat_mode_from_loop_mode,
model_track_to_core_queue_track
- ensure_remote_track_loaded (still in mod.rs, will move to
track_loading.rs in a follow-up step)
corebridge.rs imports queue_resolution::* directly for the cursor and
order helpers, and session::is_peer_renderer_active for the echo-reset
gate.
Pure relocation. mod.rs: 5304 → 4777 lines. corebridge.rs: 565 lines.
cc
Move 22 public types (QconnectConnectOptions, QconnectLifecycleState,
QconnectConnectionStatus, QconnectVisibleQueueProjection, the 20-variant
QconnectOutboundCommandType + 5 request payload structs, the 4 admission/
handoff types, the 8 device/session/queue payloads) into a dedicated
types.rs submodule. mod.rs re-exports via `pub use types::*` so external
callers (lib.rs invoke handlers, commands_v2/queue.rs) keep their import
paths.
Pure data and enum-mapping impls only. Two helper methods on
QconnectOutboundCommandType (to_queue_command_type,
requires_remote_queue_admission) become pub(super) since they are only
used from within the qconnect module tree.
mod.rs: 6371 → 6109 lines. types.rs: 273 lines.
cc
Pure rename to enable subsequent splits into focused submodules. No
behavior change: 34 invoke handler refs in lib.rs and 2 use statements
in commands_v2/queue.rs updated from qconnect_service:: to qconnect::.
Foundation for the qconnect_service.rs → qconnect/ module split: the
file is 6371 lines and mixes 12 distinct responsibilities (types,
session topology, event sink, queue resolution, corebridge bridging,
track loading, transport, Tauri commands, etc). Splitting into focused
submodules makes the area navigable for both authors and reviewers,
enables independent reasoning per concern, and creates a clean target
for the eventual qbz-qconnect-proto crate extraction.
cc
Bundle of related fixes addressing QConnect echo handling and shuffle
authority. All four converge on the same architectural principle:
qbz must not invent state when the cloud is the source of truth.
1. Remove should_force_remote_track_restart and its echo-driven reload
The function fired whenever the cloud echoed SetState with
target_position_ms=0 for the same track currently playing. The cloud
re-emits this routinely on secondary field changes (next_track
corrections, queue_item_id refresh) — every echo triggered a
load_remote_track_into_player that interrupted in-progress audio.
Track-change cases continue to load via align_corebridge_queue_cursor
plus ensure_remote_track_loaded.
2. Reject echo seeks in apply_renderer_command_to_corebridge
Authoritative seek-to-zero is only honored when the command targets
a different track, or when local position is already near zero. If
command.current_track matches playback_state.track_id and target<=1s
while local_pos>2s, the seek is treated as echo and skipped. The
prior 2s threshold was insufficient to absorb cloud round-trip during
queue changes.
3. Drop has_loaded_audio gate from should_reload_remote_track
playback_state.track_id is updated by the audio thread only after
engine.append(source) succeeds. During the 1-3s buffer/decode window
of an initial load, has_loaded_audio is false even though loading is
in flight — the prior gate caused ensure_remote_track_loaded to
redundantly re-trigger load on every echo SetState arriving in that
window. This was the residual first-track hiccup.
4. Add load-attempt dedup across V2 play and ensure_remote_track_loaded
New sync_state.last_load_attempt: Option<(track_id, Instant)> with a
5s window. Two write sites:
- play_remote_renderer_track_if_active when active_renderer_is_local
(V2 frontend just dispatched a local play)
- ensure_remote_track_loaded when it triggers its own load
Read site: ensure_remote_track_loaded skips if last attempt for the
same track was within the window. Eliminates redundant loads when an
echo SetState arrives while a local play is still buffering.
5. Defer engine shuffle activation until cloud sends a valid order
The cloud broadcasts shuffle toggles in two stages: first a
QueueUpdated with shuffle_mode=true and shuffle_order=null, then
~400ms later a QueueUpdated with the computed shuffle_order. The
prior code applied set_queue_with_order with shuffle_enabled=true on
the first event, falling through to set_identity_shuffle_order_internal
— qbz inventing a sequence (0,1,2,...) that diverges from the order
the cloud is about to authorize. If the second event was delayed,
dropped, or raced against the first, qbz drifted from peer clients
(the long-standing "doble shuffle" symptom).
Now: effective_shuffle_enabled = shuffle_mode && core_shuffle_order.
is_some(). When the order is absent, the engine stays unshuffled and
waits for the second QueueUpdated. sync_corebridge_remote_shuffle_
projection short-circuits identically. Updated diagnostic log
surfaces remote_shuffle_mode, shuffle_order_present, and engine_
shuffle_enabled separately for future triage.
Tests: removed should_force_remote_track_restart test, simplified
should_reload_remote_track test to reflect the track-id-only contract.
cc
The isGaplessTransition predicate fired on every external track change
(QConnect renderer command, MPRIS, etc.) when qbz was the local
renderer, mis-classifying remote-driven track changes as gapless
transitions. The handler then ran onGaplessTransition every backend
tick (~1Hz) without updating currentTrack, leaving the UI stale and
spamming v2_next_track.
Added gaplessAttemptedForCurrent gate captured before the existing
null-out: only treat the event as a gapless transition when
gaplessAttemptTrackId matches the track currentTrack actually believes
is playing. External track changes fall through to the
"track changed externally" block which updates currentTrack properly.
cc
The inner `pub mod commands` block in network/mod.rs hosted three
unregistered Tauri commands (get_network_mounts_cmd,
check_mount_accessible, check_network_paths_batch) and was uncalled.
The non-command helpers in the parent module (get_network_mounts,
check_mount_accessibility, find_mount_for_path, parse_mount_info)
are also no longer referenced now that the commands are gone, but
they remain pub for potential future use without warnings.
- Remove the inner `pub mod commands { ... }` block.
- Drop the now-unused std::collections::HashMap import.
cc
src-tauri/src/commands/user_session.rs hosted three unregistered
#[tauri::command] entries (activate_user_session, deactivate_user_session,
get_last_user_id) plus two non-command helpers (init_type_alias_state,
teardown_type_alias_state) consumed only by session_lifecycle.rs.
- Move init_type_alias_state and teardown_type_alias_state into
session_lifecycle.rs (their only caller) as plain pub helpers.
- Delete src-tauri/src/commands/user_session.rs and
src-tauri/src/commands/mod.rs.
- Remove pub mod commands; from lib.rs — the legacy commands
namespace is now empty.
- Refresh the activate_session doc comment to match the new layout.
cc
The 1,375-line library/commands.rs hosted only unregistered Tauri
commands (54 of them) plus a path-compat pub use of state types.
Every #[tauri::command] in the file is V2-replaced and uncalled.
The state-types pub use was the only external surface and is
already exported canonically from library/mod.rs.
- Delete src-tauri/src/library/commands.rs entirely
- Drop pub mod commands; from library/mod.rs
- Update session_lifecycle.rs to use crate::library::LibraryState
(the canonical path) instead of crate::library::commands::LibraryState
- Refresh the doc comment in library/state.rs to reflect the
post-cleanup layout
cc
All 17 #[tauri::command] entries in offline_cache/commands.rs are
unregistered (V2 owns the live cache commands) and uncalled. The only
surviving consumer was session_lifecycle.rs reaching for
purge_all_cached_files (a non-command helper).
- Move purge_all_cached_files to a new offline_cache/purge.rs module
(drops the LibraryState path through library::commands; uses the
canonical crate::library::LibraryState directly)
- Delete offline_cache/commands.rs along with its dead siblings:
cache_track_for_offline + post_process_cached_track,
is_track_cached, get_cached_track_path, get_cached_track,
get_cached_tracks, get_offline_cache_stats, remove_cached_track,
clear_offline_cache + cleanup_empty_folder, set_offline_cache_limit
+ evict_if_needed, open_offline_cache_folder, open_album_folder,
open_track_folder, check_offline_root_mounted, validate_offline_path,
move_offline_cache_to_path, detect_legacy_cached_files (the command
shim — the helper of the same name lives in migration.rs and is
still consumed by V2), start_legacy_migration,
sync_offline_cache_to_library + SyncResult.
- Update session_lifecycle.rs to import from offline_cache::purge.
cc
The reco_store/commands.rs module no longer hosts any registered Tauri
commands. Its surviving content is purely helper logic consumed by V2
home recommendations in commands_v2/library.rs.
- Rename reco_store/commands.rs -> reco_store/helpers.rs
- Drop the unregistered/uncalled #[tauri::command] entries
(reco_log_event, reco_train_scores, get_playlist_suggestions,
reco_backfill_genres, reco_needs_genre_backfill) and their dead
scoring helpers (build_scores, build_track_entries, decay_factor,
current_timestamp, etc.) along with RecoTrainStats / RecoScoreCounts
/ BackfillStats structs that only those commands referenced.
- Update commands_v2/library.rs to import from
crate::reco_store::helpers::* (still uses get_home_seeds_internal,
resolve_albums, resolve_tracks, resolve_artists, album_to_card_meta).
cc
The Tauri commands in cast/commands.rs and cast/dlna/commands.rs are
unregistered (handler list is V2-only) and uncalled. The CastState /
DlnaState path-compat re-exports are also unused (canonical paths are
crate::cast::CastState and crate::cast::dlna::DlnaState). Remove the
files entirely and drop pub mod declarations.
cc
The last consumer (api_server REST handlers) was migrated to
commands_v2::v2_play_track / v2_seek / v2_set_volume in the previous
commit, so the file's three remaining #[tauri::command] functions
(play_track, seek, set_volume) plus the unregistered prefetch_track,
get_playback_state, and all helper symbols (parse_quality,
limit_quality_for_device, PlayTrackResult, StreamInfo, download_audio,
get_stream_info, extract_audio_format_from_header, download_and_stream,
spawn_prefetch, prefetch constants, PREFETCH_SEMAPHORE) had zero
remaining callers.
Drops the `pub mod playback;` and `pub use playback::*;` lines from
commands/mod.rs. Only user_session.rs remains in the legacy commands
module.
cc
Switches the REST endpoints serving the PWA remote control from the
legacy commands::playback::* helpers to their V2 equivalents:
- play_queue_track / play_queue_index / play_album:
commands::playback::play_track -> commands_v2::v2_play_track
- seek: commands::playback::seek -> commands_v2::v2_seek
- set_volume: commands::playback::set_volume -> commands_v2::v2_set_volume
The HTTP contract (paths, methods, payload shapes, response codes,
emit_playback_update side-effects) is unchanged. Behaviour gains the
V2 features that legacy lacked (CMAF cache hits, ALSA hardware
sample-rate guard, manual-offline-mode awareness) because the PWA
remote now flows through the same playback path as the desktop UI.
Drops the `commands` import from this module (now only `commands_v2`
is used) and adds the State acquisitions needed by v2_play_track:
CoreBridgeState, OfflineState, RuntimeManagerState.
cc
The function had zero callers across src/. Today's H2 audio rewiring
work updated the invoke names inside cmdUpdateAudioSettings but the
function itself was already unreachable. Remove it.
The remaining commandRouter exports (cmdGetAudioSettings,
cmdGetAvailableBackends, cmdGetDevicesForBackend,
cmdGetHardwareAudioStatus, plus all Playback/Queue helpers) still
have callers, so the file stays.
cc
SettingsView read qbz-fetch-artist-images with === 'true' (OFF when key
absent or unrecognised) while LocalLibraryView read with !== 'false'
(ON when key absent or unrecognised). The two files disagreed on
first-run behaviour and on any stored value other than 'true'/'false'.
Switch SettingsView to the !== 'false' convention so the toggle UI
matches what the consumer actually does at fetch time.
cc
The Show-Partial-Playlists toggle was intentionally hidden in commit
2579c1990 ("Remove offline playlist integration") because offline
playlists are not yet functional. The handler and its
setShowPartialPlaylists store import were left behind as dead code.
The store function and OfflineSettings.showPartialPlaylists field
remain because Sidebar.svelte and PlaylistManagerView.svelte still
read the setting (defaulting to true).
When offline playlist integration ships, the toggle and handler can
be re-added.
cc
Toggling Manual Token Mode only updated local $state, so on every
remount of the Settings view the toggle reverted to OFF and hid the
manual-token field. Hydrate from getUserItem and persist via
setUserItem on change, matching the pattern used by plexEnabled,
plexUiCollapsed, and plexMetadataWriteEnabled.
Storage key: qbz-plex-poc-manual-token-mode (string 'true'/'false').
cc
The RemoteControlSetupGuide modal was rendered but unreachable: nothing
ever set showRemoteControlGuide to true. Add a Setup-Guide button next
to the existing Regenerate / QR controls in the Remote Control developer
subsection so the onboarding wizard can actually be opened.
The button intentionally is not gated by remoteControlEnabled because
the guide is intended for users who have not yet enabled remote control.
Reuses the pre-existing settings.integrations.remoteControlSetupGuide
i18n key (already present in en/es/de/fr/pt).
cc
Inlines bodies of get_offline_status, get_offline_settings,
set_manual_offline, check_network, and get_pending_playlists from
offline::commands directly into the v2 wrappers in commands_v2/library.rs.
Deletes the entire offline::commands submodule (the inlined commands and
the additional dead Tauri command shims that lived alongside them — none
were registered with Tauri or referenced outside the submodule). Prunes
the now-unused tauri AppHandle/Emitter imports from offline/mod.rs.
Step 8 of legacy cleanup, group H.
cc
Inlines bodies of reco_get_home, reco_get_home_ml, and reco_get_home_resolved
from reco_store/commands.rs directly into the v2 wrappers in
commands_v2/library.rs. Promotes their shared helpers
(get_home_seeds_internal, resolve_tracks, resolve_artists,
album_to_card_meta) from private to pub(crate) so the v2 wrappers can
call them without crossing back into a legacy command module.
Deletes the corresponding legacy #[tauri::command] definitions and
prunes now-unused imports from reco_store/commands.rs.
Step 8 of legacy cleanup, group G.
cc
Inlines bodies of playlist_get_custom_order, playlist_has_custom_order,
and playlist_get_tracks_with_local_copies from library/commands.rs
directly into the v2 wrappers in commands_v2/playlists.rs, mapping
LibraryError to RuntimeError::Internal as needed.
Deletes the corresponding legacy #[tauri::command] definitions and
removes their re-exports from library/mod.rs.
Step 8 of legacy cleanup, group F.
cc
Inlines bodies of library_get_album_tracks, library_get_tracks_by_ids,
and update_playlist_folder from library/commands.rs directly into the
v2 wrappers in commands_v2/library.rs.
Deletes the corresponding legacy #[tauri::command] definitions and
removes their re-exports from library/mod.rs.
Step 8 of legacy cleanup, group E.
cc
Inlines bodies of library_fetch_missing_artwork, library_get_thumbnail,
and library_get_thumbnails_cache_size from library/commands.rs directly
into the v2 wrappers in commands_v2/library.rs.
Deletes the corresponding legacy #[tauri::command] definitions and the
unused legacy library_clear_thumbnails command. Removes their re-exports
from library/mod.rs.
Step 8 of legacy cleanup, group D.
cc
Inlines bodies of library_get_scan_progress, library_get_artists,
library_search, and library_get_stats from library/commands.rs directly
into the v2 wrappers in commands_v2/library.rs.
Deletes the corresponding legacy #[tauri::command] definitions and
removes their re-exports from library/mod.rs.
Step 8 of legacy cleanup, group C.
cc
Inlines bodies of library_add_folder, library_cleanup_missing_files,
library_get_folders, and library_get_folders_with_metadata from
library/commands.rs directly into the v2 wrappers in commands_v2/library.rs.
Deletes the corresponding legacy #[tauri::command] definitions and
removes their re-exports from library/mod.rs.
Step 8 of legacy cleanup, group B.
cc
Inlines bodies of library_scan_impl, library_stop_scan_impl,
library_scan_folder_impl, library_clear_impl, library_update_album_metadata_impl,
library_write_album_metadata_to_files_impl, and library_refresh_album_metadata_from_files_impl
from library/commands.rs directly into the v2 wrappers in commands_v2/legacy_compat.rs.
Moves their helpers (process_cue_file, apply_sidecar_override_if_present,
compute_track_artist_match, normalize_library_path) alongside the inlined bodies.
Deletes the corresponding legacy #[tauri::command] _impl definitions and
the dead public command siblings (library_scan, library_stop_scan,
library_scan_folder, library_clear, library_update_album_metadata,
library_write_album_metadata_to_files, library_refresh_album_metadata_from_files)
that were only ever called from the deleted _impl variants.
Updates library/mod.rs to drop the corresponding re-exports.
Step 8 of legacy cleanup, group A.
cc
Inlines `check_network_path` from crate::network::commands into
commands_v2/legacy_compat.rs. The legacy command was a one-line wrapper
over crate::network::is_network_path; the inlined v2 form calls
is_network_path directly with no behavior change.
Deletes the legacy #[tauri::command] definition in network/mod.rs.
Other tauri commands inside crate::network::commands remain (they have
no v2 delegates and keep being exported as their own legacy commands;
they will be addressed in a separate cleanup step if needed).
No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group G (network path).
cc
Inlines the bodies of `library_update_folder_settings` and
`get_track_by_path` from crate::library::commands::* directly into
commands_v2/legacy_compat.rs. Maps the legacy String error type to
RuntimeError::Internal at every return path.
Deletes the legacy #[tauri::command] definitions in
library/commands.rs. The data-layer methods on LibraryDatabase
(update_folder_settings, get_folder_by_id, get_track_by_path) are
unchanged.
No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group F (folder settings + track
lookup).
cc
Inlines the body of `library_backfill_downloads` from
crate::library::commands::* directly into commands_v2/legacy_compat.rs,
mapping the legacy String error type to RuntimeError::Internal at every
return path. Adjusts the v2 return type to refer to
crate::library::BackfillReport (the canonical re-export from
qbz-library) instead of the deleted crate::library::commands path.
Deletes the legacy #[tauri::command] definition in
library/commands.rs. MetadataExtractor and BackfillReport are reused
unchanged from their crate::library re-exports.
No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group E (backfill downloads).
cc
Inlines the bodies of `check_album_fully_cached` and
`check_albums_fully_cached_batch` from
crate::offline_cache::commands::* directly into
commands_v2/legacy_compat.rs, mapping the legacy String error type to
RuntimeError::Internal at the v2 boundary.
Note: v2_cache_notification_artwork is NOT a delegate — it lives in
commands_v2/helpers.rs and is a true helper used by other v2 commands.
The investigation list flagged it as a possible delegate, but
verification showed it owns its implementation.
Deletes the legacy #[tauri::command] definitions in
offline_cache/commands.rs. Other functions in the offline_cache::commands
module remain (still called from commands_v2/library.rs delegates,
which are step 8 scope).
No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group D (offline cache).
cc
Inlines the bodies of `create_pending_playlist`,
`get_pending_playlist_count`, `queue_scrobble`, `get_queued_scrobbles`,
`get_queued_scrobble_count`, and `cleanup_sent_scrobbles` from
crate::offline::commands::* directly into commands_v2/legacy_compat.rs.
The legacy bodies were uniform: lock state.store, unwrap optional store,
delegate to a method on the store struct. Inlined verbatim with the
String error type mapped to RuntimeError::Internal at each return path.
The store-side OfflineStore methods (data layer) are untouched.
Deletes the legacy #[tauri::command] definitions in offline/mod.rs.
Other functions in the offline::commands module remain (still called
from commands_v2/library.rs delegates, which are step 8 scope).
No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group C (offline scrobble/playlist).
cc
Inlines the bodies of `discogs_has_credentials`,
`discogs_search_artwork`, and `discogs_download_artwork` from
crate::library::commands::* directly into commands_v2/legacy_compat.rs.
The discogs delegates were single-line bodies (or trivial proxy logic);
DiscogsClient and library::get_artwork_cache_dir are reused as-is.
The String error type from the underlying DiscogsClient methods is
mapped to RuntimeError::Internal at the v2 boundary, matching the
existing v2 contracts.
Deletes the legacy #[tauri::command] definitions from
library/commands.rs. No frontend invoke changes required.
Step 6 of the legacy cleanup plan, group B (discogs).
cc
Inlines the bodies of `lyrics_get` and `lyrics_get_cache_stats` from
crate::lyrics::commands::* directly into commands_v2/legacy_compat.rs,
mapping the legacy String error type to RuntimeError::Internal at the
return boundary for v2_lyrics_get. Renames LyricsCacheStats to
V2LyricsCacheStats, hosted next to the inlined v2_lyrics_get_cache_stats.
Deletes lyrics/commands.rs entirely (also drops the dead
lyrics_clear_cache helper which had no callers; v2_lyrics_clear_cache in
commands_v2/library.rs already inlines its own logic).
The v2 functions retain their existing names, parameter shapes, and
Tauri::State signatures; only the body changes. No frontend invoke
changes required.
Step 6 of the legacy cleanup plan, group A (lyrics).
cc
Relocates CastState (struct + impl) from cast/commands.rs to
cast/state.rs, and DlnaState (struct + impl) from
cast/dlna/commands.rs to cast/dlna/state.rs.
Both types are consumed by commands_v2/library.rs (and the legacy
Tauri commands themselves) via the existing re-exports in
cast/mod.rs and cast/dlna/mod.rs. The re-exports are updated to
source from the new state modules, and each commands.rs re-exports
its state type so historical
crate::cast::commands::CastState / crate::cast::dlna::commands::DlnaState
paths continue to compile.
This unblocks future deletion of cast/commands.rs and
cast/dlna/commands.rs in a later cleanup step (the
#[tauri::command] functions in those files are already
unregistered and have zero Rust call sites).
cc
Relocates LibraryState (struct + impl), CleanupResult, BackfillReport,
LibraryAlbumMetadataUpdateRequest, and LibraryAlbumTrackMetadataUpdate
from library/commands.rs to a new library/state.rs module.
These types are consumed by commands_v2/library.rs,
commands_v2/legacy_compat.rs, commands_v2/mixtapes.rs,
commands_v2/link_resolver.rs, cast/commands.rs, and
session_lifecycle.rs. The existing pub use re-exports in
library/mod.rs are updated to source from state, and commands.rs
re-exports them in turn so historical
crate::library::commands::* paths (still used by
commands_v2/legacy_compat.rs, which is out of scope for this task)
continue to compile.
This unblocks future deletion of library/commands.rs once the
remaining #[tauri::command] definitions there are inlined into
their commands_v2 callers (separate task).
No serde shape changes; struct fields, derives and serde
attributes are preserved verbatim, so the JSON the frontend
receives is identical.
cc
Removes 11 #[tauri::command] functions that are unregistered and have
zero Rust call sites:
- pause_playback, resume_playback, stop_playback
- play_next_gapless, set_media_metadata
- get_audio_devices, reinit_audio_device
- get_audio_output_status, get_pipewire_sinks
- set_pipewire_default_sink, debug_get_cpal_devices
play_track, seek, and set_volume are preserved - they are still
consumed by src-tauri/src/api_server/mod.rs (REST endpoints for
the PWA remote). They will be migrated to call commands_v2/playback
or core_bridge in a later step.
Helper functions used solely by the deleted commands are also
removed: AudioDevice, AudioOutputStatus, PipewireSink structs,
get_pipewire_device_descriptions (linux + non-linux), and
cpal_device_name. Helpers reachable from play_track / seek /
set_volume / prefetch_track / get_playback_state (parse_quality,
limit_quality_for_device, download_audio, get_stream_info,
extract_audio_format_from_header, download_and_stream,
spawn_prefetch, prefetch constants and PREFETCH_SEMAPHORE) are
preserved.
cc
All 20 #[tauri::command] functions in this file are unregistered
(no longer in invoke_handler) and have zero Rust call sites. V2
equivalents in commands_v2/search.rs and commands_v2/catalog.rs
cover all use cases.
Removed:
- search_albums, search_tracks, search_artists, search_all
- get_album, get_featured_albums, get_track, get_artist
- get_similar_artists, get_genres
- get_discover_index, get_discover_playlists, get_playlist_tags, get_discover_albums
- get_artist_page, get_artist_basic, get_artist_detail
- get_artist_tracks, get_artist_albums, get_releases_grid
The corresponding pub mod search; declaration in commands/mod.rs is
also removed.
The SearchAllResults / MostPopularItem types (used by api_server's
REST /search/all endpoint, the PWA remote control surface) are
relocated into api_server/mod.rs since that is the only remaining
consumer; this avoids leaving a near-empty file just to host two
serde DTOs.
cc
These commands had v2_* equivalents registered alongside them; the
frontend has been migrated and no longer invokes them. The legacy
function definitions remain in their respective modules - they will
be removed in subsequent steps once we verify nothing internally
calls them.
cc
Path C: legacy CPAL output-sink enumeration was inlined directly in
commands/playback.rs (cfg-gated Linux + non-Linux variants), with
no equivalent in any domain crate. Extract a domain function
`qbz_audio::output_sinks::list_output_sinks` that returns the same
{name, description, volume, is_default} shape (mirrored as the new
public `OutputSinkInfo` struct) for both Linux and non-Linux
platforms, preserving the original logging cadence and the rule
that name == CPAL device name (so saved IDs reopen correctly).
Add v2_get_pipewire_sinks in commands_v2/audio.rs that calls into
qbz_audio directly — no `crate::commands::*` delegation. Updates the
two frontend call sites (AudioOutputBadges.loadStatus and
SettingsView.loadAudioDevices). Legacy command in
commands/playback.rs is preserved untouched; the audio-protected
files (pipewire_backend.rs, audio_settings.rs, player init_device)
are not modified.
cc
Path B: legacy command read state.player.state.current_device() and
state.player.state.is_playing() — both methods already exist on the
qbz-player Player.state in the V2 architecture. No domain extraction
needed; the v2 command reads them via CoreBridge.player() like the
existing v2_get_hardware_audio_status does.
Adds v2_get_audio_output_status returning V2AudioOutputStatus
(serialized to the same {device_name, is_playing} shape so the
AudioOutputBadges frontend type is unchanged). Updates both
AudioOutputBadges call sites (initial loadStatus + the 1s polling
interval). The legacy command in commands/playback.rs is preserved
untouched.
cc
Path C: legacy command called client.get_releases_grid for the
artist detail view's release-type pagination grid. No existing
v2 equivalent, so add v2_get_releases_grid via QbzCore ->
QobuzClient::get_releases_grid directly.
cc
Path C: legacy command called client.get_artist_tracks (the
/artist/get?extra=tracks path) and returned a TracksContainer.
No existing v2 equivalent, so add v2_get_artist_tracks via
QbzCore -> QobuzClient::get_artist_tracks directly.
cc
Path C: legacy command called client.get_artist_detail with the
hard-coded mobile defaults of limit=1000, offset=0 to populate the
suggestions panel's playlists and tracks_appears_on. No close v2
equivalent existed (v2_get_artist_with_albums uses
extra=albums only), so add v2_get_artist_detail that goes through
QbzCore -> QobuzClient::get_artist_detail directly.
cc
Path C: legacy command projected `Artist.albums` from
client.get_artist_with_pagination(with_albums=true). The existing
v2_get_artist_with_albums returns the full Artist envelope, so
add a dedicated v2_get_artist_albums in QbzCore that performs the
same projection and exposes only ArtistAlbums to match what the
two callsites destructure.
cc
Path A: existing v2_get_artist already returns Artist via
QbzCore::get_artist which delegates to client.get_artist_basic;
shape matches the legacy command exactly so the frontend just
swaps the invoke name.
cc
Add v2_get_flatpak_help_text as a thin wrapper over the existing
flatpak::get_flatpak_guidance(). Update the frontend invoke in
SettingsView.svelte to use the v2 name. Legacy command remains for
now.
Audit: legacy_callsites 12 -> 11.
cc
Add v2_detect_desktop_theme as a thin wrapper over the existing
desktop_theme::detect_desktop_theme(). Update the frontend invoke in
windowControlsStore.ts to use the v2 name. Legacy command remains
for now.
Audit: legacy_callsites 13 -> 12.
cc
Add v2_get_qobuz_tos_accepted as a thin wrapper over the existing
config::legal_settings::get_qobuz_tos_accepted. Update the frontend
invoke in qobuzLegalStore.ts to use the v2 name. Legacy command
remains for now.
Audit: legacy_callsites 14 -> 13.
cc
Add v2_has_snap_welcome_been_shown as a thin wrapper over the
existing UpdatesStore::has_snap_welcome_been_shown(). Update the
frontend invoke in updatesService.ts to use the v2 name. Legacy
command remains for now.
Audit: legacy_callsites 15 -> 14.
cc
Add v2_has_flatpak_welcome_been_shown as a thin wrapper over the
existing UpdatesStore::has_flatpak_welcome_been_shown(). Update the
frontend invoke in updatesService.ts to use the v2 name. Legacy
command remains for now.
Audit: legacy_callsites 16 -> 15.
cc
The test was added in 145974c29 when the default device name was a
const "QBZ Desktop". A later refactor (bf91822b4) replaced the const
with hostname-derived "Qbz - {hostname}" plus env-var/persisted-DB
overrides. This left the renderer fixture asserting a friendly_name
that no longer matches what default_qconnect_device_info() returns,
so refresh_local_renderer_id correctly returned None and the test
panicked.
Read the local device info inside the test and use those fields for
the renderer that should match. The test now exercises the contract
("when device_uuid is missing, fall back to a unique fingerprint
match against the local device info") regardless of host environment.
cc
When QConnect is on AND the local app is the active renderer, every
local next/play_index round-trips back through the Qobuz cloud as a
SET_STATE echo. apply_renderer_command_to_corebridge unconditionally
called align_corebridge_queue_cursor, which races the user's already-
advanced local cursor and triggers a second play_index, pushing a
duplicate index onto QueueManager.history. That is the duplicates half
of issue #316.
Mirror the gating that sync_active_renderer_projection already uses:
align the CoreBridge cursor only when local is NOT the active renderer.
The existing peer-driven flow (peer is active renderer, echoes arrive
to drive our cursor + history) is unchanged.
cc
set_queue and set_queue_with_order unconditionally cleared history, which
wiped legitimate plays whenever QConnect echoed a queue version bump
(reorder, shuffle toggle, peer-driven snapshot) without changing track
identity. That is the missing-tracks half of issue #316.
Instead, remap history entries by track id: each pre-existing index is
resolved to its track id via the old tracks list, then looked up in the
new tracks list. Entries whose track id is no longer present are dropped.
When the new queue has zero overlap with the old, history clears
naturally; when the queue is a pure reorder, history survives intact.
Six unit tests cover pure reorder, additions, partial removals,
disjoint replacement, and verify the remapped indices (not just rendered
ids) point at the right new tracks.
cc
Previously the prop was named isQobuzConnectConnected, which suggested
transport-up semantics, but after 23658b05 it carries the user-intent
toggle state (true during Connecting and Reconnecting too). Rename to
match what the value actually means, removing the chance of a future
reader misreading the prop as a "safe to send commands" gate.
No behavior change.
Refs: #358
cc
Previously, an Err from app.connect() would propagate via ? while
lifecycle_state remained at Connecting and runtime was None, leaving
the frontend toggle stuck "on" (because isQconnectToggleOn treats
Connecting as on) with no live runtime. This recreated the exact
user-visible bug 23658b05 was trying to fix, just triggered from a
different upstream failure.
Convert the ? propagation into explicit error handling that resets
lifecycle_state = Off and records last_error before returning Err,
so the toggle correctly flips off and a subsequent connect() starts
fresh.
Refs: #358
cc
When the reconnect loop is stuck waiting for a SESSION_STATE that never
comes (the actual symptom of issue #358), the runtime is alive but
`transport_connected` is false. The UI used to read its on/off toggle
from `transport_connected`, so the badge appeared OFF while the backend
loop kept retrying. Worse: clicking the toggle in that state called
`connect()`, which rejected with "QConnect service is already running".
The user had no way to disable QConnect from the UI — only to "turn it
on more", which failed.
Backend (`qconnect_service.rs`):
- Add `QconnectLifecycleState` enum (Off / Connecting / Connected /
Reconnecting / Exhausted) and surface it on
`QconnectConnectionStatus::state`. Driven from the spawned transport
event loop via a new `update_lifecycle_state_if_running` helper which
refuses to overwrite terminal states (Off / Exhausted) with late
Reconnecting / Connected events from the broadcast queue.
- Make `connect()` idempotent: if a runtime already exists (alive but
reconnecting), return current status instead of erroring. The toggle
is the single user-visible control; the previous "already running"
error was unactionable from the UI.
- Auto-stop on `MaxReconnectAttemptsExceeded`: take the runtime out,
set state=Exhausted, set last_error to the cloud reason, and emit a
`qconnect:status_changed` event so the frontend reflects the new
state without waiting for its 5 s poll. The transport loop has
already exited at this point, so a fresh user-initiated connect()
cleanly starts a new runtime.
- `disconnect()` now forces lifecycle to Off unconditionally and tears
down the runtime (already worked, the transport's `shutdown_tx`
watch breaks `handle_reconnect_delay` mid-sleep), satisfying the
"disable QConnect must always succeed" requirement.
Frontend (`qconnectRuntime.ts`, `+page.svelte`):
- Add `QconnectLifecycleState` mirror type and `isQconnectToggleOn()`
helper.
- Split state in the page: `isQobuzConnectConnected` keeps its old
semantics (`transport_connected`) and gates send-command paths
(volume reports, position reports) so we don't fire commands while
the WS is down. New `isQobuzConnectToggleOn` reflects the
user-facing on/off — true even during Reconnecting/Connecting — and
is what the badge component and the toggle handler now read.
- Listen for the new `qconnect:status_changed` event so the toggle
reflects auto-stop (Exhausted) and other transitions immediately
instead of after the 5 s status poll.
Refs #358
cc
The MSG_TYPE_ERROR (cloud_type=9) arm previously logged only the payload
byte count, so when Qobuz cloud rejected a session (zombie session,
conflicting device, expired JWT, ...) we had no signal beyond an opaque
"bytes=13" TransportError. That hid the actual cause of the 2s flicker
loop seen after the app sat minimized for 1-2 hours (issue #358).
The qws.proto ErrorMessage (messageTypeId=9) carries:
uint32 msg_id = 1
uint64 msg_date = 2
uint32 code = 3
string descr = 4
Decode it via a local QwsErrorMessage prost struct, log a warn line with
the decoded code+descr, and emit a new TransportEvent::CloudError so
qconnect-app and qconnect_service can react. qconnect_service forwards
it as a "qconnect:cloud_error" diagnostic (level=warning) and logs it.
qconnect-app treats it as a no-op forwarder for now, matching the
existing pattern for other transport-only events. Decode failures fall
back to the previous TransportError path under a distinct stage name so
they are still visible.
Adds a tokio test that builds a real MSG_TYPE_ERROR frame and asserts
the transport emits CloudError (not the legacy TransportError).
Refs #358
cc
The native WS transport reset its reconnect counters as soon as the WS
upgrade succeeded. Qobuz cloud accepts the WS connection well before the
JOIN_SESSION exchange and rejects unhealthy sessions with MSG_TYPE_ERROR
right after, so the counter reset every cycle and the loop hammered the
endpoint at base backoff (~2s) forever (issue #358).
Move the counter reset to the moment the cloud emits a
MESSAGE_TYPE_SRVR_CTRL_SESSION_STATE event (the actual proof of a healthy
session-level join), and add a configurable maximum-attempts cap that
makes the transport stop and emit MaxReconnectAttemptsExceeded instead
of looping indefinitely.
- Add SessionEstablished and MaxReconnectAttemptsExceeded transport events.
- Add reconnect_max_attempts to WsTransportConfig (default Some(10)).
- handle_incoming_binary now returns whether SESSION_STATE was decoded.
- Reconnect helper returns a ReconnectOutcome enum (Continue/Shutdown/
Exhausted) instead of a bare bool.
- qconnect_service.rs logs the new events and emits a diagnostic when
the cap is exceeded so the higher layer can react (#358-fix-3 will
surface it in the UI).
- Unit tests cover the cap behavior and the unbounded fallback.
Refs #358
cc
Completes the legacy root list by adding config::favorites_cache,
which is registered in lib.rs (e.g., is_track_favorite) but was
missing from the audit's known-roots list. No frontend currently
invokes it, so this is preventive — guards against future silent
misclassification as 'unknown'.
Found in code review of commit 3ea96fb5.
cc
The frontend invoke audit was capturing the wrong segment of legacy
handler registrations. The pattern 'commands::([A-Za-z0-9_]+)' grabbed
the submodule name (e.g. 'search', 'playback') instead of the function
leaf, so commands like 'commands::search::get_artist_albums' fell into
the unknown bucket even though they were registered.
Switch to a per-root regex of the form
'<root_path>(::<segment>)*::(<leaf>),' which captures the leaf
identifier regardless of how many submodule levels sit between the root
and the function. Also add four legacy module roots that were missing:
- updates (has_flatpak_welcome_been_shown, has_snap_welcome_been_shown)
- config::legal_settings (get_qobuz_tos_accepted)
- desktop_theme (detect_desktop_theme)
- flatpak (get_flatpak_help_text)
After this change the unknown bucket is empty and the 16 callsites that
were previously misclassified as 'unknown' now correctly show up as
'legacy', which is the accurate status pending future V2 migration.
cc
cmdUpdateAudioSettings was invoking four v2 commands that do not exist:
- v2_set_audio_backend -> v2_set_audio_backend_type
- v2_set_audio_device -> v2_set_audio_output_device
- v2_set_exclusive_mode -> v2_set_audio_exclusive_mode
- v2_set_dac_passthrough -> v2_set_audio_dac_passthrough
Also corrects the parameter for set_audio_output_device, which expects
'device' (not 'deviceName'). The other three already use the correct
parameter names. The function currently has no callers, but keeping the
shape correct prevents a future caller from hitting silent failures.
cc
The frontend was invoking 'v2_qconnect_get_status', but the backend
command is registered as 'v2_qconnect_status' under qconnect_service.
This caused two callsites in the diagnostics path to fail at runtime:
- src/lib/services/diagnosticsSnapshot.ts
- src/lib/components/DiagnosticsPanel.svelte
cc
The previous regex only matched commands registered under commands_v2::,
missing 50+ legitimate v2_qconnect_* commands registered under
qconnect_service::. This produced 64 false-positive "missing_v2" callsites
that obscured real audit signal.
Generalize the registration regex to match any module prefix ending in
::v2_* or ::runtime_*, so future modules that follow the v2_ naming
convention are picked up automatically without script changes. The inner
alternation still excludes legacy commands, which never carry the v2_ or
runtime_ prefix on the function name.
After the fix, missing_v2_callsites drops from 64 to 6 (the remaining 6
are real frontend typos, not audit-script bugs); legacy_callsites stays
at 0; unknown_callsites stays at 16.
cc
Custom-sort reordering inside a playlist (arrow buttons + drag and
drop) silently no-op'd or shifted by the wrong amount. Investigation
on a 15-track playlist showed customOrderMap with duplicate positions
(two tracks at position 6) and the moved track at map-position 10
while displayed at index 11 — the in-place shift algorithm assumed
display-index and map-position lived in the same number space, which
is only true when positions are unique 0..N-1.
Once any duplicate sneaked in, every subsequent click hit the no-op
path: target newPosition equalled current map-position, the per-key
shift loop's `pos > fromIndex && pos <= toIndex` window collapsed to
empty, and the backend's move_playlist_track saw current==new and
returned without writing. The user had to toggle sort mode to force
loadOrInitCustomOrder to re-fetch backend state — which masked the
fact that the local algorithm itself was wedged.
This rewrites moveTrack to ignore the existing map values and instead
splice the visible displayTracks order, then renumber 0..N-1. The new
order is shipped to v2_playlist_set_custom_order, whose backend impl
is a DELETE + INSERT for the playlist, so any pre-existing corruption
self-heals on the first move.
Concurrent clicks are serialized through a Promise chain so backend
writes land in click order even though tokio's mutex isn't strictly
FIFO. The optimistic local update happens synchronously before the
chain dispatch, so subsequent clicks see the post-move displayTracks
state without waiting for the round-trip.
Note: the latent bug in database.rs::move_playlist_track that lets
duplicate positions exist in the first place (the INSERT branch on
line 3354 doesn't shift existing rows) is not addressed here — the
new flow simply doesn't use that code path. Backend cleanup is a
separate fix.
cc
Replaces the first-pass implementation that rolled its own backdrop and
pill-style size buttons. The pills had two problems: pills are forbidden
by the project design rules (qbz-nix-docs/design/FIGMA_PROMPT.md), and
the custom modal centered itself with `transform: translate(-50%, -50%)`
which created a containing block for `position: fixed` descendants —
that broke `Dropdown.svelte`'s portaled menu, which landed in the
bottom-right corner of the screen instead of below the trigger.
Now wraps `Modal.svelte` (portaled to body, flexbox-centered, ESC
built in), uses `Dropdown.svelte` for the size picker, and follows
the project convention of `<div class="modal-actions">` + the global
`.btn .btn-primary` / `.btn .btn-secondary` classes for the footer.
Copy is reworked to read like a sentence: title "Random queue", body
"Create a random queue with songs from this collection. How many songs
do you want?", confirm button "Add to queue". Synced across en/es/de/fr/pt.
Drops the now-unused `mixModal.subtitle` key.
Also removes the duplicated mixModalOpen entry from the view's
keydown ESC chain — `Modal.svelte` already handles ESC.
cc
Adds a third circular header action between Shuffle and More that
opens TrackMixModal, then delegates the resolved sampleSize to
v2_collection_shuffle_tracks. Replaces the queue, starts playback,
and surfaces a toast when actualCount < requestedCount (dedup or
per-album cap shrunk the pool).
ESC chain extended to dismiss the mix modal before closing other UI.
collection.play_mode is left untouched.
cc
- TrackMixModal.svelte renders a centered radiogroup of queue sizes
(50, 100, 150, …, "All (N)") backed by buildSizeOptions. Fetches the
unique track count via v2_collection_unique_track_count when first
opened for a collection and caches it for subsequent opens of the
same collection.
- buildSizeOptions is a pure helper extracted to a separate file so it
can be unit-tested cleanly. 8 vitest cases cover the empty/null,
small, exact-50, multiple-of-50, and irregular paths.
- Adds 8 new i18n keys (common.shuffleTracksMix, mixModal.title /
computing / subtitle / allOption / confirm / loadFailed,
toast.mixTrimmed) across all 5 locale files.
- z-index 3000/3001 per ADR-003.
cc
End-to-end DJ-mix command. Loads the collection, resolves its natural
in-order track list, dedupes near-duplicate versions, samples up to
sample_size tracks under the per-album cap, replaces the queue, and
starts playback from index 0. Stamps queue_source_collection and bumps
the play counter (mirroring v2_enqueue_collection's replace mode), but
does not modify the collection's persistent play_mode.
Returns ShuffleTracksResult { requestedCount, actualCount } so the
frontend can show the trimmed-mix toast when the cap or dedup shrinks
the pool.
ThreadRng is not Send, so dedup + sample run inside a sync scope that
ends before the bridge awaits.
cc
Resolves the collection's natural in-order track list and returns the
count of distinct songs after similarity-based deduplication. Used by
the upcoming track-mix modal to size the slider of pickable queue
sizes. Auth gated; deterministic output.
cc
Round-robin sampling that picks up to `requested` tracks from a
deduped track list. Each album bucket is shuffled in place and
truncated to cap = max(2, ceil(requested * 0.30)); rounds permute
the album order so cross-album distribution is randomized too. May
return fewer than `requested` when the per-album cap and bucket
sizes do not add up — callers report this gap to the user.
7 new tests (37 total) including a 200-seed statistical test
confirming the cap is never exceeded.
cc
- unique_track_count returns the count of distinct songs after grouping
near-duplicate versions; deterministic, no RNG.
- dedup_by_similarity collapses each duplicate group to one survivor
picked at random via the supplied RNG. Output preserves input order
of survivors.
Both share a single bucketed Union-Find pass: tracks are first grouped
by exact normalized artist, then unioned within each artist bucket
when their normalized titles score >= 0.80 on token_set_ratio. Cross-
artist comparisons are skipped to avoid false positives like "Yesterday"
by different artists.
12 new tests (30 total in this module) including a 200-seed statistical
test confirming every duplicate version can be selected as the winner.
cc
Pure-function helpers used by the upcoming track-mix sampler:
- normalize_title strips parens/brackets, dash suffixes, feat patterns,
punctuation, and diacritics.
- normalize_artist keeps parens (band suffixes carry meaning) and just
lowercases + strips diacritics.
- token_set_ratio implements the RapidFuzz-style token-set similarity in
[0.0, 1.0] using strsim's normalized Levenshtein.
Promotes strsim from transitive to direct dep. 18 unit tests cover the
core normalization and similarity behaviors.
cc
Inlined SVG component for the upcoming track-shuffle mix button.
Uses fill="currentColor" so it inherits theme color. The source
SVG (static/dj-1.svg) is kept as a reference asset.
cc
Memory mitigations for issue #331 (Pi 3B / 1 GB hosts):
- qbz-core::system_capabilities reads /proc/meminfo at startup and
derives a MemoryProfile (Normal vs LowMemory). Cached in OnceLock.
- V2 prefetch (commands_v2/playback.rs) caps prefetch_count and
max_concurrent_prefetch from the profile, downgrades HiRes to
Lossless when allow_hires_prefetch is false.
- StreamingConfig::from_speed_mbps clamps the dynamic initial
buffer to a process-wide cap configured from the profile (256 KB
on LowMemory vs 2 MB historical max).
- AudioCache L1 cap also scales (50 MB on LowMemory vs 400 MB).
- Tokio watchdog spawned on LowMemory hosts polls MemAvailable
every 10 s; flushes audio caches when below 5 % to break the
swap-thrash loop codehd7 reported.
Includes the rodio 0.22 NonZero test fix (also in the bugfix
merge above) so the branch validates standalone.
cc
Fixes the iec958: -> hw:CARD=...,DEV=... fallback gate that was
silently disabled for non-USB sound cards (HifiBerry, intel-hda,
etc.) because the previous gate required /proc/asound/cardN/stream0,
which only USB-audio drivers emit. The HifiBerry Digi2 Pro on
Raspberry Pi OS reported in #331 hits exactly this case.
Also bumps the visualizer test fixtures to rodio 0.22's
NonZero<u16>/NonZero<u32> SamplesBuffer::new signature so
cargo test -p qbz-audio compiles.
cc
rodio 0.22 tightened SamplesBuffer::new to require NonZero<u16> for
channels and NonZero<u32> for sample_rate, but the visualizer test
fixtures still passed bare integer literals — so the qbz-audio test
binary stopped building. Wrap both call sites in NonZero::new(...)
.unwrap() so cargo test -p qbz-audio compiles again.
Pure test-only change; runtime visualizer code was already migrated.
cc
Add audio_cache_l1_max_bytes to MemoryProfile and wire it into
AppState's AudioCache constructor. Normal hosts keep the historical
400 MB cap; LowMemory hosts drop to 50 MB.
Without this, a Pi 3B (1 GB total) starts with the L1 cache cap set
at ~40 % of physical RAM. Even with the Feature #3 prefetch caps
already in place, the runtime watchdog from the previous commit
ends up firing more often than necessary because nothing prevents
a single L1 fill cycle from claiming 400 MB. Capping at 50 MB caps
how much the cache can grow before either: (a) eviction picks up
naturally, or (b) the watchdog steps in if the rest of the system
is also under pressure.
Pure refactor in lib.rs — replaces a hard-coded 400 * 1024 * 1024
with a profile field read. The default-constructed AudioCache
fallback (when disk-spillover init fails) keeps its own crate
default; no need to thread the profile there since the failure
path is rare.
Two tests added: a presence-bound assertion in the existing Pi 3B
case, and an explicit comparison that LowMemory's cap is strictly
smaller than Normal's.
cc
Tauri setup() now spawns a tokio task that polls
qbz_core::system_capabilities::read_memory_pressure() every 10 s
when the resolved MemoryProfile classifies the host as LowMemory.
On Normal-class machines the task is never spawned, so polling cost
is zero where it isn't needed.
Behavior:
- pressure.is_critical (< 5 % MemAvailable) triggers
audio_cache.clear() once per critical episode and logs at warn.
Both L1 (in-memory tracks) and L2 (disk spillover) are flushed —
the disk hit on subsequent playback is far cheaper than letting
the OS swap-thrash through the audio thread.
- pressure.is_low (< 15 %) is logged once at info as an early
warning but does not evict.
- Recovery (back above 15 %) logs once and arms the next critical
episode.
The was_critical edge guard prevents log spam when pressure
oscillates near the threshold during heavy prefetch.
For codehd7's Pi 3B scenario (issue #331): when QBZ pushes
MemAvailable below ~47 MB (5 % of 938 MB), the watchdog flushes
the cache, frees memory back to the kernel, and the network stack
recovers from the swap thrash that was killing playback.
cc
Extend qbz-core's system_capabilities with a parse_meminfo_available_kb
helper plus a small MemoryPressure value object that classifies the
host's current available memory against thresholds derived from
MemoryProfile. The watchdog (src-tauri side, next commit) polls this
to decide when to evict caches.
Thresholds intentionally simple for v1: is_low triggers under 15 %
available, is_critical under 5 %. Both are strict-inequality so the
exact-percent boundaries don't flap.
The pure pressure_from_figures(avail_kb, total_kb) factors out the
math from the /proc read so threshold behavior is tested without
touching the filesystem. Refactors the existing
parse_meminfo_total_kb to share a parse_meminfo_field_kb helper with
the new available-memory parser.
Adds 9 tests covering the parser, the threshold boundaries, the zero-
total divide-safety guard, and the codehd7 issue #331 scenario
(Pi 3B, ~25 MB available out of 938 MB → critical).
cc
StreamingConfig::from_speed_mbps inflates the initial buffer to 2 MB
on slow connections (< 1 MB/s) to absorb network jitter. On a healthy
desktop that's fine, but on a memory-constrained host the slow link
is itself a symptom: when QBZ on a Pi 3B starts allocating ~300 MB
of Hi-Res prefetch the kernel swaps, networking stalls, the speed
detector reports < 1 MB/s, and the buffer code reacts by allocating
ANOTHER 2 MB — exactly the wrong direction.
Add a process-wide AtomicUsize cap (defaults to usize::MAX, i.e.
no-op) plus a small public set_max_initial_buffer_bytes setter.
Refactor from_speed_mbps to clamp through the cap; extract the
speed-to-bytes ladder into a pure raw_initial_buffer_for_speed
helper plus a from_speed_mbps_with_cap variant that exposes the
clamping for unit tests without touching global state.
src-tauri/lib.rs's run() wires the cap into qbz-player from
qbz-core's MemoryProfile right after logging the profile. Effective
on Pi 3B (issue #331): all StreamingConfig::from_speed_mbps results
pin to 256 KB regardless of detected speed.
Re-export max_initial_buffer_bytes / set_max_initial_buffer_bytes
from qbz_player::player so external callers don't need to reach
into the private streaming_source module.
Adds 5 unit tests: ladder behavior, pass-through when cap is loose,
clamp at the LowMemory cap (the issue #331 case), no-op at the
Normal cap, and that max_buffer_bytes stays unchanged.
cc
Wire qbz-core's detected MemoryProfile into the V2 prefetch loop so
memory-constrained hosts (Pi 3B and similar, < 2 GB RAM) cap their
in-memory footprint:
- V2_PREFETCH_COUNT becomes a function reading profile.prefetch_count
(5 on Normal, 1 on LowMemory). Reading it once per loop call is
cheap — OnceLock-cached behind the scenes.
- V2_MAX_CONCURRENT_PREFETCH likewise becomes a function; the
semaphore lazy_static initializes from it once at process start.
- Effective prefetch quality gets a low-memory cap before the
hardware-rate ladder runs. When profile.allow_hires_prefetch is
false we drop UltraHiRes/HiRes to Lossless so each cached track
stays under ~15 MB instead of ~60 MB. The hardware downgrade
still runs after, so a low-memory + 48 kHz-only DAC still gets
the right rate.
- lib.rs forces an early call to memory_profile() in run() so the
resolved profile is logged at boot rather than at first prefetch.
For codehd7's reported case (Pi 3B, 1 GB, HifiBerry Digi2 Pro,
issue #331): the active profile becomes prefetch_count=1,
max_concurrent_prefetch=1, allow_hires_prefetch=false. Cached
prefetch footprint drops from ~5 × 60 MB = 300 MB to a single
~15 MB Lossless track — well within the headroom that 1 GB minus
WebView (~150 MB) leaves.
cc
New module in qbz-core that probes /proc/meminfo at startup and
classifies the host into Normal (>= 2 GB RAM) or LowMemory (< 2 GB)
buckets, then derives runtime caps for memory-sensitive subsystems:
prefetch_count, max_initial_buffer_bytes, max_concurrent_prefetch,
allow_hires_prefetch.
Detection is split into a pure parse_meminfo_total_kb +
detect_profile_from_meminfo pair (testable with synthetic input)
and a thin detect_profile wrapper that reads the file. The
process-wide profile is cached in a OnceLock and logged once at
first access, so subsequent callers pay zero filesystem cost.
The Normal-fallback path covers macOS, Windows, and any Linux where
/proc/meminfo is missing or unparseable — we never accidentally
throttle a system whose memory we couldn't measure.
This commit only ships the module; wiring into the prefetch loop
and streaming buffer follows in the next commits.
Adds 10 unit tests covering the parser, the threshold boundary
(exactly 2 GB → Normal, just under → LowMemory), the Pi 3B and
Pi Zero 2W cases that motivated this work (#331), and the
unparseable-input fallback.
cc
The hw:CARD=…,DEV=… fallback added in 1.2.8 (commit 927fabbd7) only
fired when the helper chain
extract_card_name_from_device(device_id)
.and_then(|card| get_hw_supported_rates(&card))
.is_some()
returned Some — but get_hw_supported_rates parses
/proc/asound/cardN/stream0, which is only emitted by the USB-audio
driver. I2S, PCI and built-in cards (HifiBerry HATs, intel-hda, etc.)
don't have a stream0 file at all, so the gate evaluated to false for
the very devices that needed the fallback the most. The reporter's
HifiBerry Digi2 Pro on Raspberry Pi OS hit this exactly: 1.2.8's
ALSA Direct path logged "Looking for device: iec958:CARD=…", failed
the lookup, and dropped straight to the proc_found=false error
branch without ever attempting hw:CARD=sndrpihifiberry,DEV=0.
Extract a small is_card_present_in_proc helper that uses
find_card_number_by_name (which reads /proc/asound/cards — present
for every registered card regardless of bus) and use it for both
gate checks: the fallback trigger and the proc_found error message
selector. get_hw_supported_rates keeps its three legitimate uses
(actual rate validation in create_output_stream, and the two public
device_supports_sample_rate / get_device_supported_rates queries).
Adds 8 unit tests covering build_hw_fallback_id, the alias-prefix
branches of extract_card_name_from_device, the short-circuit cases
of is_card_present_in_proc, and the existing is_known_pcm_id filter.
The /proc-touching branches stay covered by hardware testing on the
reporter's Pi when v1.2.9-rc ships.
cc
rodio 0.22 tightened SamplesBuffer::new to require NonZero<u16> for
channels and NonZero<u32> for sample_rate, but the visualizer test
fixtures still passed bare integer literals — so the qbz-audio test
binary stopped building. Wrap both call sites in NonZero::new(...)
.unwrap() so cargo test -p qbz-audio compiles again.
Pure test-only change; runtime visualizer code was already migrated.
cc
Two $t('toast.playingAlbum', { values: { count } }) call sites in
+page.svelte (collection/mixtape detail play paths) were printing the
raw key because the translation was never added. The `|| fallback`
idiom after each call never fired either — $t() returns the key
itself on miss, which is truthy.
Add an ICU plural entry covering both callers: count=1 (detail-view
play button on an album) reads as "Playing album", count>1 (bulk
play-all) reads as "Playing {count} tracks". Mirrored across en, es,
de, fr, pt.
cc
Both PlaylistDetailView's .tracks-sticky-header and ArtistDetailView's
.jump-nav opened a visible gap between the titlebar and the sticky
bar once pinned — scrolled content bled through as a blurred strip.
Two causes, both hit here:
- Subpixel/compositing rounding in WebKitGTK 2.50+ left a few px
transparent at the pinned top edge. Promoting each bar to its own
compositing layer via `will-change: transform; transform: translateZ(0)`
mirrors the same workaround imageCacheService applies to <img>.
- Even with a compositing layer, the scroll container above the pin
can briefly expose bg before the bar's fill kicks in. Add a solid
`box-shadow: 0 -32px 0 0 var(--bg-primary)` so the bg extends
upward and stays covered regardless of any remaining offset. For
.jump-nav this composes with the existing downward drop shadow.
Also set `background: var(--bg-primary)` on .playlist-detail and
.artist-detail (the scroll containers) and bump .track-controls'
padding-top to -2px (CSS clamps to 0 but documents the intent), so
the toolbar sits flush against the bar border without the 8px
breathing-room the unpinned state had.
cc
vicrodh