Files
qbz/crates/Cargo.toml
T
vicrodh f1b4241342 feat(source): add qbz-source, the source-agnostic media seam
Stage 1 of the seam the owner asked for: "son vistas súper híbridas donde se
mezcla Qobuz/Local(Folders/Plex) ... hay que hacer una solución agnóstica
porque no tardamos en meter a Jellyfin y quizás Navidrome". Nothing is wired
yet — this lands the crate, its three source implementations and its tests.
`qbz-qt` and `qbz` are untouched, so this commit changes no behaviour.

The problem it exists to end, measured rather than assumed: 129 source-branching
sites across 30 files in qbz-qt alone, SEVEN distinct item->tracks resolvers
none of which can reach the others, FIVE non-interchangeable artwork pipelines
over one correct taxonomy, and six incompatible source vocabularies — three of
which cannot express Plex at all, including the PERSISTED one (`AlbumSource` is
`Qobuz|Local`). Every hybrid view re-wires the per-source handling by hand,
which is why something is always missing.

One `Source` trait covering the four concerns together — tracks, metadata and
origin, artwork, playback — because splitting them is exactly what lets a view
forget one. A `SourceRegistry` with claim-priority ordering (Qobuz, Plex, then
Local; Plex-before-Local is load-bearing). `RegistryResolver` implements
`qbz_mixtape::ItemResolver`, so that crate and the Slint glue stay byte-
untouched and keep compiling.

The three measured bugs die structurally, not by discipline:
- The local DB was reopened PER CALL, so a My QBZ collection logged one
  "Opening library database" per item (34 measured live). `LocalSource` owns a
  pooled handle, and it hands out no guard — only `with(|db| ...)` — so a
  `!Send` `&LibraryDatabase` provably cannot cross an await. `qbz-mixtape`'s
  sync-resolver constraint becomes an API property instead of a comment.
- A `plex:<hash>` album key reached an API that wants a rating key. `MediaRef`
  has no public constructor: only `registry.claim` mints one, and
  `PlexSource::claim` rejects that shape in track position. The survey also
  found a second, undiscovered face of the same bug — "Go to album" on a Plex
  track row opens an empty page.
- Local rows in a hybrid list rendered placeholder discs. Traced to its real
  mechanism: those items are stored with `artwork_url: None` and the only
  backfill is behind a Qobuz-client gate, so offline or pre-login the art is on
  disk and never shown. `Source::artwork` is one call per row whatever the
  source, and keeps the cheap/cold split the local path already had.

Adding Jellyfin is one new file, four diff lines inside this crate, and one
line in `auth_qt.rs` — and the design states this only holds BECAUSE artwork
sits on the trait.

36 tests. The id tests derive their fixtures from `PLEX_TRACK_ID_FLOOR` instead
of transcribing them, which is not pedantry: the survey's own id table — the
artefact meant to be the definitive reference — paired a namespaced id with the
wrong rating key, and the transcribed test caught it on the first run.

Design and adversarial review (6 BLOCKER / 10 MAJOR / 7 MINOR, all corrected):
qbz-nix-docs/qt-frontend/2026-07-30-source-seam/.

cc
2026-07-30 16:14:30 -06:00

160 lines
6.0 KiB
TOML

[workspace]
resolver = "2"
members = [
"qbz-app",
"qbz-theme",
"qbz-models",
"qbz-audio",
"qbz-player",
"qbz-qobuz",
"qbz-radio",
"qbz-core",
"qbz-integrations",
"qbz-cache",
"qbz-library",
"qbz-cast",
"qbz-cmaf",
"qbz-dsd",
"qbz-secrets",
"qbz-offline-cache",
"qbz-plex",
"qconnect-core",
"qconnect-protocol",
"qconnect-transport-ws",
"qconnect-app",
"qbz-ui",
"qbz-dac-wizard",
"qbz-text-utils",
"qbz-slint-common",
"qbz",
"qbz-credentials",
"qbz-media-controls",
"qbz-mixtape",
"qbz-playlist-import",
"qbz-lyrics",
"qbz-i18n",
"qbz-music-link",
"qbz-reco",
"qbz-external-reco",
"qbz-log",
"qbzd",
"qbz-qt",
# The source-agnostic seam (one trait + one registry over Qobuz / local /
# Plex). Purely additive: nothing in `qbz` or `qbz-mixtape` changes, and it
# deliberately does NOT depend on qbz-audio / qbz-player.
"qbz-source",
]
# Qt-frontend POC worktree: plain `cargo build`/`cargo check` in this workspace
# produces ONLY the Qt binary (plus its Slint-free dependency closure). The
# Slint crates remain members but are never built by default here.
default-members = ["qbz-qt"]
[workspace.package]
version = "2.0.2"
edition = "2021"
license = "MIT"
repository = "https://github.com/vicrodh/qbz"
[workspace.dependencies]
# Async runtime
tokio = { version = "1", features = ["full"] }
async-trait = "0.1"
futures-util = "0.3"
tokio-stream = { version = "0.1", features = ["sync"] }
# Serialization
serde = { version = "1", features = ["derive"] }
serde_json = "1"
# HTTP
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls-webpki-roots-no-provider", "cookies", "stream"] }
# Logging
log = "0.4"
tracing = "0.1"
# Error handling
thiserror = "2"
anyhow = "1"
# Crypto (Qobuz auth)
md-5 = "0.10"
base64 = "0.22"
# Time
chrono = "0.4"
# Regex
regex = "1"
# Audio
symphonia = { version = "0.5", features = ["all"] }
# Random number generation (shuffle / DJ-mix) — mirrors src-tauri pin
rand = "0.10"
# Fuzzy string similarity (track-mix dedup) — mirrors src-tauri pin
strsim = "0.11"
# Note: Linux-specific dependencies (alsa) go in individual crate Cargo.toml files
# because workspace manifests cannot have [target] sections
[profile.release]
# Strip symbol tables from release binaries. Measured on 2026-07-02:
# 400 MB -> 218 MB (-45%). There are no .debug_* sections in a default
# release build — the weight is .strtab (151 MB) + .symtab (22 MB), inflated
# by Slint's generated generic symbol names. Trade-off: panic backtraces show
# addresses instead of names; the release workflow archives the UNSTRIPPED
# binary as an Actions artifact per tag for offline symbolication.
# Codegen knobs (opt-level / codegen-units / jobs) intentionally stay in the
# workflows' env (CARGO_PROFILE_RELEASE_*) — they are CI-memory tuning, not
# a property of the package.
strip = "symbols"
[patch.crates-io]
# Slint 1.16's femtovg-wgpu renderer leaves the wgpu surface alpha mode at
# `Auto`, which resolves as an opaque compositor surface on KWin/Wayland and
# makes borderless transparent windows show a black rectangular backing. The
# patch keeps Slint pinned but chooses a supported non-opaque composite alpha
# mode when the surface advertises one, so the miniplayer card can have real
# transparent rounded corners. The non-opaque mode is PER-WINDOW opt-in
# (`wgpu::set_surface_prefers_transparent`, flipped from qbz's winit
# window-attributes hook): only the miniplayer pays the compositor-blend cost,
# the main window keeps an Opaque swapchain.
i-slint-renderer-femtovg = { path = "vendor/i-slint-renderer-femtovg-1.16.1" }
# On modern macOS, fontique's CoreText per-script fallback can fail to resolve
# a Han-script family (the suggested PingFang is not always in the scanned
# family map), so Japanese kanji rendered blank while kana worked. The patch
# pre-registers the known system CJK families (Hiragino/PingFang/Apple SD
# Gothic Neo/Arial Unicode MS, Japanese-biased default) in the fallback map at
# collection creation, which takes priority over the CoreText query. macOS-only
# (#[cfg]); Linux keeps its working fontconfig fallback.
i-slint-common = { path = "vendor/i-slint-common-1.16.1" }
# With `overflow: elide`, sharedparley culled EVERY line whose bottom edge
# exceeded max_height — including the FIRST one, eliding the text to nothing.
# CJK fallback fonts (Hiragino on macOS) report line metrics taller than
# snug fixed-height boxes (TrackRow's 20px title line at 14px), so every
# CJK track title rendered blank on macOS while Latin fit (#543; verified
# with a headless repro matrix on the Mac Mini). The patch always draws the
# first line — it still gets the ellipsis treatment as the last visible one.
i-slint-core = { path = "vendor/i-slint-core-1.16.1" }
# Slint compiles femtovg with the swash raster path. Swash fails to scale
# some variable fonts at a non-default variation instance (the CFF2-based
# CJK system fonts on modern macOS), and femtovg cached those glyphs as
# permanently missing — so CJK text at any non-regular weight rendered
# blank (#543; regular weight = empty coords = renders fine). The patch
# retries the glyph at the font's DEFAULT instance before giving up: worst
# case the glyph draws at regular weight instead of not at all. No effect
# on fonts/instances that already render.
femtovg = { path = "vendor/femtovg-0.23.2" }
# macOS Skia renderer: skia bug 310510989 makes new_from_data fail for TTC
# faces at index > 0, and for some system collections (AppleSDGothicNeo.ttc,
# where EVERY weight sits at index > 0) the upstream read/write-fonts rebuild
# workaround ALSO fails — every Korean glyph run at non-regular weights drew
# nothing. The patch adds a second-chance workaround: resolve the same face
# through CoreText by family name + OS/2 weight, guarded by cmap glyph-id
# spot checks so a wrong match is rejected instead of drawing wrong glyphs.
i-slint-renderer-skia = { path = "vendor/i-slint-renderer-skia-1.16.1" }