Reorder the Small now-playing bar's right control set to
Connect / Cast / Audio settings / Now-Playing view / Lyrics / Volume /
Queue, matching PlayerBar. Pure reorder; the leading AudioStamp stays.
cc
Port Tauri's AwardView/AwardAlbumsView to Slint: the award landing
(hero + acclaimed-releases preview grid + other-awards rail + follow
heart), the paginated albums listing with client-side search, and
clickable album-sidebar laurels that open the award page (resolving the
award id by name when /album/get omits it).
- AwardState/AwardActions globals + AwardEntry; AlbumState.awards now
carries award ids
- award controller (src/award.rs): /award/page hero (never page.releases),
/award/getAlbums grid, /award/explore other-awards + name->id resolver;
follow via the shared favorite_awards cache (plural-read/singular-write
favType, skip-if-remote, login seed)
- nav registration: NavEntry::Award/AwardAlbums, ContentView arms,
navigate_award/navigate_award_albums, AppShell mounts, app.slint exports
- ArtworkTarget::AwardAlbum/AwardOther
cc
The 36px accent-blue mic-vocal tile at the top of the lyrics sidebar was a
Tauri-era treatment that no longer matches the current QBZ header style.
Removed per owner request; the LYRICS overline + track meta stay.
cc
A wrapped lyric line (one logical line over two visual rows) lit BOTH rows at
once because the active line was a single word-wrap Text masked by one clip
Rectangle of full block height + scalar width. Now it progresses per visual
line — row 1 fills to 100% before row 2 starts — matching the final Tauri
behavior. NOT a Slint limitation (proven in the investigation doc).
- lyrics_measure.rs (new): swash advance pass over the active line at the same
font (by font-index) + ppem + wght=700 + 0.2px letter-spacing the view
renders with, greedy word-wrap (CJK per-char) -> per-visual-line segments
with pixel widths. Embeds the same TTFs the view registers. swash is the
shaper femtovg uses, so widths match the render.
- lyrics_sync.rs: on active-index / content-width / font / size / uppercase
change, segment the active line and push [LyricsSegment{text, start-weight,
width-weight}] (Tauri's width-weighted partition); pushed atomically with the
active index. Per-surface cache keys (main + mini). line-progress unchanged.
- state.slint: LyricsSegment struct + active-segments + content-width on
LyricsState.
- LyricsLinesView.slint: active row renders a VerticalLayout of per-segment
rows (no-wrap), each the existing dim-base + clipped-bright-overlay with clip
width = segwidth * clamp((line-progress - start-weight)/width-weight, 0, 1).
Root publishes content-width. Inactive rows unchanged.
- miniplayer.rs: mini lyrics surface segmentation fan-out.
Investigation + plan: qbz-nix-docs/lyrics/2026-06-18-karaoke-highlight-investigation/.
Owner-verified working. Known tuning point: wrap-boundary precision per font tier.
cc
Pre-existing dead code unrelated to cast; restores the zero-warnings policy.
Deleted (abandoned refactor leftovers): ModeTextToggle (LocalLibraryView.slint,
unreferenced), section_artwork_jobs + ArtworkTarget::Section variant + its match
arm (Discover-slice leftover; live render uses DiscoverSectionAlbum),
rename_folder (orphan DB wrapper).
Marked #[allow(dead_code)] with a "reserved, pending wiring" note (ported work
from other epics, not discarded): qconnect_service set_autoplay_mode_if_remote /
autoplay_load_tracks_if_remote / stop_if_remote (QConnect autoplay/stop API);
myqbz_play skip_to_next_item / skip_to_previous_item (Mixtape item-skip, spec 40).
cargo check -p qbz-slint: 0 warnings.
cc
- Lyrics auto-follow while casting: the local poll's cast-skip now runs BEFORE
clear_remote_anchor() (it was wiping the anchor every 450ms), and the cast
poll publishes its position via lyrics_sync::publish_remote_anchor each tick
(the 30Hz engine extrapolates between ticks, like the QConnect path). Anchor
cleared on disconnect so local lyrics resume.
- DLNA seekbar: many renderers report TrackDuration=0 in GetPositionInfo, which
left the bar permanently full (position/0 -> clamp 1.0) with no total time.
Fall back to the track's catalog duration when the renderer reports 0 (the
renderer's position is still authoritative). Chromecast reports a real
duration, so it's unaffected.
cc
The cast was slow because the resolver always did a full CMAF download
(~10s for a 190MB hi-res FLAC) and never read the cache — unlike the fast last
Tauri build, which resolved bytes cache-first.
- P1 cache-first: Player::fetch_for_external_stream now checks L1 (in-memory) ->
L2 (on-disk playback cache) before any download. The gapless prefetch already
warms the next 2 tracks at UltraHiRes while casting, so next/advance/replay are
now instant; only a cold track pays the download.
- Offline tier: CoreBridge::fetch_for_external_stream_resolved gains offline +
sink params and tries resolve_offline_bytes (local CMAF decrypt, no network)
before the network path — a downloaded track casts fast with no connection
(the same local-segments+decrypt path offline playback uses). Replaces the
hard "Qobuz cast needs a connection" gate (offline-cached tracks now work).
- P2 quality label: derived from the track's catalog metadata
(bit_depth/sample_rate/hires) in cast_track — no extra API call, consistent
whether bytes came from cache, offline, or network.
cargo check -p qbz-slint: clean.
cc
Smoke-test fixes after the first real Chromecast session:
- Local 450ms poll now skips the bar update while casting (it was pushing the
stopped local player's position=0/playing=false/local-volume every tick,
fighting the cast poll → seekbar flicker, wrong play icon, dead volume bar).
The cast poll owns the bar, mirroring the existing QConnect-controller skip.
- Cast poll + connection-state push now set NowPlayingState.playing (the
play/pause icon was stuck on play), and set_volume_if_cast pushes the slider
position (the local set_volume that normally moves it is skipped while casting).
- next/previous no longer use a cast-only path: that advanced the renderer +
core cursor but never refreshed the card/queue, and left the cursor desynced
so queue-click resolved the wrong track. They now run the local
next()/previous() flow (move cursor -> refresh card+queue -> play_audible
casts the new track), keeping UI + renderer in sync. Auto-advance on track end
uses the same local next() flow.
cc
rust_cast opens a rustls TLS channel to the Chromecast. In the full binary both
the aws-lc-rs and ring providers are enabled (reqwest pulls ring, rust_cast pulls
aws-lc-rs), so rustls can't auto-select one and the worker thread panicked with
"Could not automatically determine the process-level CryptoProvider" — surfacing
as "Connection error: Thread response error".
Add qbz_cast::ensure_crypto_provider() (installs aws-lc-rs once, idempotent) and
call it at the top of ChromecastHandle::connect and DlnaConnection::connect.
cc
- play_audible now gates Cast-first (before the QConnect peer route): when a
renderer is connected, a newly-played track is cast to it instead of starting
the local audio backend (no double audio). Mirrors the existing QConnect
play_on_peer_if_active chokepoint.
- Offline gate: casting a Qobuz track while offline is refused with a clear
message (the external-stream resolver is network/CMAF only; local files still
cast offline). Offline-cached Qobuz cast is a tracked follow-up.
Deferred (documented, non-blocking): continue-locally prompt on disconnect
(needs a small Slint modal); skipIfRemote UI gate (cast while this instance is
a QConnect remote — rare edge).
cargo check -p qbz-slint: clean.
cc
- CastPicker.slint: real modal (scrim + 400px card) replacing the static
PlayerBar/PlayerBarSmall cast-menu PopupWindow stubs. Two rectangular tabs
(Chromecast/DLNA, no pills, plain count) + per-tab device list from
CastState.devices, connect-on-click (DLNA rows gate on can-play), connected
banner with Disconnect + delivered-quality line (D3), scanning/empty states,
error line. Mounted last in AppShell (on top).
- PlayerBar + PlayerBarSmall cast buttons now open the modal
(CastState.picker-open = true; CastActions.open()) instead of the dead popup.
- Cast teardown wired: shutdown() on logout (fixes the Tauri connection + poll
leak #33) and best-effort on app exit (renderer Stop + media server, #32).
cargo check -p qbz-slint: clean (only the 6 pre-existing warnings).
cc
The Slint-side seam over the qbz-cast crate, modeled on qconnect_service.rs.
cast_service.rs (singleton):
- discovery (mDNS + SSDP) with a 2s device-refresh loop and 15s scan window;
- connect/disconnect: snapshot local playback, stop the local backend, suspend
QConnect (remember + restore), connect, then re-cast the current track
passing its REAL source (fixes the Tauri resume-source bug) + deferred seek;
- cast_track: source-routed (qobuz via the Phase-0 shared resolver at
UltraHiRes; local via register_file streaming + rich MIME; plex = TODO),
one shared lazy MediaServer, per-device source-IP URL, Chromecast auto-play
vs DLNA two-step load->play;
- 1s position poll mirroring position/duration/elapsed to the bar + CastState,
with the verbatim track-end heuristic (Chromecast ->IDLE, DLNA ->STOPPED/
NO_MEDIA_PRESENT) driving auto-advance;
- delivered-quality (D3) pushed to CastState; shutdown() for logout/exit.
state.slint: CastState global + CastActions + CastDevice (tagged record with
DLNA capability flags); re-exported in app.slint.
main.rs: init + CastActions wiring (open/close/refresh/connect/disconnect) and
transport gating — play/pause/next/previous/seek/volume now route Cast >
QConnect > local.
Compiles clean (only stop_if_cast/shutdown unused pending the stop callback +
logout/exit wiring, next).
cc
Additive, backward-compatible enhancements to the shared cast crate (the
Tauri shims are pure re-exports, so existing callers are unaffected).
- DLNA timeouts: pause/stop/seek/set_volume now run through a shared
run_action() with a 10s timeout (was none) so a hung renderer can't block
the caller forever. New DlnaError::Timeout / CastError::Timeout variants.
- DLNA mute: add set_mute / get_mute (RenderingControl SetMute/GetMute,
Master channel) — was missing alongside set_volume.
- DLNA capability detection is now version-agnostic (matches AVTransport:2/:3,
not only :1) so newer renderers report has_av_transport correctly.
- Media server endpoint auth: served path is now /audio/<token>/<id> with a
128-bit per-session token (RandomState-seeded, no new dep). Token lives in
the PATH, not a query string, so DLNA renderers that strip queries still
authenticate. Mismatch -> 403.
- register_file content-type parity: content_type_for_path now mirrors the
rich local-cast MIME map (audio/x-ape, mp3/ogg/opus/aac, m4a/alac->audio/mp4)
instead of the poorer table; add register_file_with_content_type so the
caller can supply an exact MIME. Local files stream from disk (no full-RAM
read) per decision D4.
Unit-tested: tokened-path parsing, rich MIME map, token randomness.
cc
Add a frontend-agnostic API to resolve a track's bytes + MIME + actually-
delivered quality for an external renderer (Chromecast / DLNA), so the Slint
Cast port does not need an improvised downloader (ADR-006).
- qbz-models: ExternalStreamAsset { bytes, content_type, quality, duration,
origin }, StreamQualityInfo (kHz-normalized via from_raw, handles the
kHz-vs-Hz endpoint ambiguity), AssetOrigin.
- qbz-qobuz/cmaf: download_full_with_quality(_progress) returns the resolved
format_id/sampling_rate/bit_depth alongside the FLAC bytes; download_full is
now a thin wrapper (no duplicated segment assembly).
- qbz-player: Player::fetch_for_external_stream — CMAF-first (FLAC, accurate
quality) with legacy stream-URL fallback; never reads the local playback
cache (casting attempts the best tier and must report the real delivered
quality); never serves an empty content type (external_content_type helper).
- qbz-core: CoreBridge::fetch_for_external_stream_resolved wrapper.
Decision D3 (attempt UltraHiRes first + surface delivered quality) is satisfied
by passing Quality::UltraHiRes and returning StreamQualityInfo. Unit-tested:
content-type fallback by format id, sampling-rate unit normalization.
cc
Port the Tauri MiniPlayer to a second Slint MiniPlayerWindow (not an
AppWindow morph). Borderless via the native no-frame property (Slint's
winit adapter re-applies decorations from no-frame on every realization,
so the attributes hook alone was undone); the vendored renderer gives the
transparent rounded floating card. First QBZ surface with no system
titlebar.
State is mirror-driven with no second poll loop: the existing 450ms
playback poll and the 30Hz lyrics tick each fan out one mirror call into
the mini's per-window globals (REQ-1), gating the expensive copies
(artwork / queue / lyric lines) on track-id change. The queue surface owns
a self-contained navigable model (current + full upcoming, not the 40-row
sidebar page), and per-row play routes remote-first and bails before
mutating local (REQ-2). Transport callbacks delegate to the main window's
already-wired remote-first handlers (zero duplication).
Surfaces: micro / compact / artwork / queue / lyrics (reuses
LyricsLinesView + LyricsState). Footer with pending-seek state, transport,
volume popover, and the hover-expand controls capsule (surface tabs +
background-blur toggle + move-drag + close). Optional static
artwork-derived background-blur backdrop. Lifecycle in src/miniplayer.rs
(create-once / enter / exit / close / per-surface geometry persistence /
winit hooks). Menu entry enabled in PlayerBar + PlayerBarSmall.
cc
Slint 1.16's femtovg-wgpu renderer configures the wgpu surface via
get_default_config() and leaves the composite alpha mode at Auto, which
resolves to an opaque surface on KWin/Wayland — a borderless transparent
window then shows a black rectangular backing behind the rounded card.
WGPUConfiguration exposes no surface/alpha hook, so the only fix is a
renderer patch.
Vendor i-slint-renderer-femtovg-1.16.1 with a preferred_alpha_mode() that
picks a supported non-opaque mode (PreMultiplied / PostMultiplied /
Inherit, guarded by the surface's advertised alpha_modes, else Auto) and
wire it into set_surface(). Pulled in via [patch.crates-io]; Slint stays
pinned =1.16.1. The main window is unaffected (it paints opaque).
cx
ListScrollbar gains an auto-hide mode: track and thumb fade in on list hover
or active scroll (900ms idle timer) and out otherwise. Discover and Home derive
hover from the global ShellState pointer over the flickable bounds; the
Blacklist manager swaps std ScrollView for Flickable + ListScrollbar driven by
a per-row hover counter. Resolves the hover-only scrollbar over album cards that
a TouchArea overlay would have broken.
cx
Drop to the winit backend (on_winit_window_event) to (1) route the browser
Back/Forward mouse buttons to navigation before Slint hands them to the
topmost TouchArea, so cards/sidebar rows no longer swallow them, and (2)
mirror the cursor position into ShellState (pointer-x/y/in-window) so passive
hover chrome can react without installing a TouchArea over interactive
content. This is the seam that unblocks the auto-hide scrollbars.
cx
- Checkboxes are now whole-row clickable (new CheckRow primitive, HTML <label>
behavior): clicking the text toggles it, not just the box. Applied to the
welcome accept + the three Configure progress checks.
- "Use my current queue" guards an empty queue: it checks get_all_queue_tracks
and shows "Your queue is empty — add some tracks first, or press Play test."
instead of sitting on a dead read-back.
- Welcome step gains a "What to test with" notice recommending one track of each
bit-depth/sample-rate (16/44.1, 24/44.1, 24/96, 24/192) so users who bring
their own queue exercise every rate.
cc
"Use my own music" -> "Use my current queue" (the button resumes the existing
queue, it doesn't open a picker). Adds explicit Prev/Next icon buttons that call
the proven NowPlayingState.previous()/next() transport, so they step through
whatever is playing — the 4 curated test tracks or the user's own queue.
cc
From owner feedback:
- The body Flickable now has a visible ListScrollbar (wrapped in a Rectangle so
it can overlay) — the review step with multiple DAC accordions can be tall.
- Configure step gains Tauri-style progress checkboxes (backup / config files /
restart) so the user can track what they ran; Next is gated on all three.
- The trigger moves to the end of the BIT-PERFECT group (the wizard's whole
point is bit-perfect) and now wears the FLACDalf (gandalf) icon, which also
replaces the modal header icon. gandalf.svg copied into the Slint icon assets.
cc
The review step generates the copy-paste config per selected DAC (read-only —
QBZ never writes these files, Decision 2): an honest info banner, a backup
CommandBlock first, then one collapsible accordion per DAC (Decision 5 multi-DAC)
holding the three snippets ported 1:1 from the Tauri wizard — pipewire
allowed-rates, per-app bit-perfect client rules, and the WirePlumber rule that
pins node.name + rates + disables resample/channelmix — and finally an
init-aware restart CommandBlock. Rates are re-probed at generation time
(query_dac_capabilities); the filename slug is the description (fixes the old
dacShortName regex that only kept two underscore tokens). The done step lists
the config paths.
Controller: checked_dacs -> gen_configs_blocking (off-thread re-probe + template
build via line-join, no format!-brace escaping) -> apply_configs; toggle_config
flips an accordion. gen-configs fires on the select-dacs -> review transition.
2 unit tests (slugify, wireplumber pins node+rates). DacConfigRow state +
gen-configs/toggle-config actions.
cc
Fixes from owner testing of the Slice 9 playback test:
- No audio: removed the incomplete reinit_device(node) that routed output to the
selected DAC (the proper switch is set_output_device -> reload_settings ->
reinit; doing it half-way broke playback). The test now plays through the
current output; route QBZ to the DAC under test in Settings > Audio.
- Bit-perfect read-back: dropped the pw-top panel (batch snapshot never reflected
the live nodes) and enriched the N6 read-back to show the DAC's real
/proc/asound hw_params — rate + ALSA container format (S32_LE etc.) + channels.
That container format is the bit-perfect proof. Probe scans for the actively
playing card, so it's DAC-agnostic (answers the multi-DAC case: one output at
a time).
- Track jump: clicking a test row now re-sets the queue at that index via the
working play_tracks path (play_index alone only moved the pointer). Resolved
tracks are stashed for this.
- "Use my own music" button: starts the read-back + resumes the user's own queue
for those who don't want the curated tracks.
- Row hover no longer changes the row height (fixed-size glyph slot; the ♪->▶
swap can't reflow what's below).
cc
player.get_sample_rate/get_bit_depth don't exist on Arc<Player>; the accessors
live on player.state (as the QConnect engine already does). Fixes the Slice 9
build.
cc
Replaces the impossible "watch pw-top" verify step with an in-wizard test. The
overlay modal doesn't grab the window, so playback runs underneath while the
test plays.
start-test resolves the 4 owner-curated tracks by id-hint first, then by
"artist title" search if the id 404s (a pulled license — never raw-id-only),
routes QBZ output to the selected DAC (reinit_device), and plays them via the
existing play_tracks path. A 1.5s Slint Timer drives poll-test while playing:
it reads the rate QBZ REQUESTED (player.get_sample_rate — honest) and the DAC's
REAL negotiated rate (N6 / negotiated_stream_rate over /proc/asound), showing a
green check only when they actually match. Read-only; protected audio path
untouched.
State: reuses the test-* fields from Slice 5; adds a poll-test callback and a
TEST_NODE thread_local (no extra model). 3 unit tests already cover the pure
seed-matching/formatting helpers.
cc
Replaces the manual node.name paste with auto-enumeration. Controller
detect_blocking() enumerates via the Slice-0 pw-dump path, projects each sink to
a DacCandidate (looks_like_dac = is_hardware && bus in {usb,pci}), and probes
rates (Slice-2 query_dac_capabilities) for the likely DACs; apply_candidates
pre-selects them and flips has-enumeration off when empty. toggle_dac flips a
VecModel row + recomputes the Next gate; validate_manual is a 1:1 port of the
Tauri validateNodeName/detectDacType (+3 unit tests).
UI step 2: circular SelectionCheckbox rows with a bordered bus label (no pill),
a default star and the probed rate list; a "Can't see your DAC?" disclosure
reveals the two wpctl CommandBlocks + a validated node.name input (shown
unconditionally when nothing enumerated). Next gates on any-dac-selected ||
manual-valid; enumeration fires on the check->dacs transition. Read-only.
cc
Wires the check step on top of the Slice-0/2/6 backend: a health verdict
(success/warning WarningBanner — new `success` variant), a per-missing-piece
CommandBlock with copy, and TWO overridable dropdowns — distro (package
manager) and init system (service commands), which are orthogonal so a Gentoo
box on systemd vs OpenRC gets the right restart command.
New `src/dac_wizard.rs` controller: open() runs the probe off-thread
(spawn_blocking → upgrade_in_event_loop), caches it, and recompute() rebuilds
remediations from the cached probe + the current distro/init selections.
Init-aware restart_cmd (systemd/OpenRC/runit/s6/dinit). NixOS special-cased to
a declarative configuration.nix block.
Sandbox honesty: in Flatpak/Snap the host probes are blind, so the check step
shows an info banner + reference setup commands (full stack incl. pipewire-alsa
+ restart) for the chosen distro/init instead of a false health verdict.
main.rs wires on_open/on_set_distro/on_set_init/on_copy_command; the Audio
settings trigger calls DacWizardActions.open().
cc
InitSystem gains ALL/index/label so the wizard can offer a manual init override
(for sandboxes and for generating commands for another machine).
Sandbox awareness: detect_sandbox() (/.flatpak-info, $SNAP). Inside a sandbox
/etc/os-release is the RUNTIME's (Flatpak freedesktop-sdk) or the snap base's,
not the user's distro, and /run + /proc/1 reflect the sandbox — so detect_distro
now reads the host-exposed os-release first (Flatpak /run/host/os-release, Snap
/var/lib/snapd/hostfs/etc/os-release, then /etc), and detect_init returns
Unknown in a sandbox (the host init is invisible; the wizard's override decides).
cc
Decouple service commands from the distro: the init system is detected at
RUNTIME (detect_init: /run/systemd/system, /run/openrc, /run/runit, /run/s6-rc,
/run/dinitctl, then /proc/1/comm), because it is orthogonal to the distro —
Gentoo runs OpenRC *or* systemd, Debian runs systemd or sysVinit/runit (antiX).
So install commands key off the distro (package manager) and service/restart
commands key off the detected init.
Adds the systemd-free derivatives Antix (Debian) and Artix (Arch) — matched
BEFORE their parent family since their ID_LIKE would otherwise misclassify them
and emit systemd commands — plus NixOS (declarative). InitSystem enum +
parse_init_from_comm. 6 health tests (incl. systemd-free-beats-parent + init
classification).
cc
`qbz-audio::health` for the wizard's check step: audio_stack_health() probes
wireplumber (systemctl OR `pgrep` fallback so non-systemd inits like
Gentoo/OpenRC and Void/runit don't read as down), pw-dump, the ALSA->PipeWire
bridge PCM (`aplay -L` `^pipewire$` — needed for CPAL PLAYBACK, not just
enumeration), pactl, and any-sink presence. detect_distro() classifies
/etc/os-release (ID then ID_LIKE; strips single AND double quotes — Gentoo
single-quotes its value). Distro enum carries the dropdown order + labels.
Pure parse_distro + 4 unit tests (Ubuntu, Pop/EndeavourOS via ID_LIKE,
fedora/arch/suse/gentoo+single-quote/void/other, index round-trip). Read-only.
cc
DacWizardState/DacWizardActions/DacCandidateRow in state.slint (re-exported in
app.slint). DacWizardModal: scrim + 620px card + dot/label stepper rail +
Flickable body with if-step blocks + footer Back/Close/primary. Navigation is
inline state mutation (no Rust yet); the Actions callbacks (copy-command/
run-detect/query-caps/start-test) are declared for Slices 6-9. welcome step is
real (disclaimer + accept checkbox gating Start); steps 1-5 are titled
placeholders. Mounted in AppShell; "Open Wizard" 160x34 trigger added to Audio
settings, gated on PipeWire.
cc
CommandBlock: surface-elevated copy-paste card (Radius.sm, no pills) with an
optional caption, a monospace wrapping command body and a copy button that
flips to a green check for 2s (Timer). Read-only — the copy() callback is wired
to the clipboard by the consumer; nothing is executed.
WarningBanner: info/warning/error variants (tone drives icon + 10%/30% alpha
bg/border, matching the existing inline banners). Adds lucide copy.svg.
cc
Extract the DAC capability assembly out of the Tauri-only
`v2_query_dac_capabilities` into `qbz-audio::dac_capabilities` so the Slint
wizard can probe a DAC's real supported sample rates + pretty description
without a Tauri command. The detection itself already lived in this crate
(`PipeWireBackend::get_sink_supported_rates` -> /proc/asound, with the
`get_device_supported_rates` ALSA-direct fallback); this just assembles the DTO
frontend-agnostically and benefits from the Slice-0 pw-dump-robust enumeration
for the description.
Formats stay nominal (S16/S24/F32) and are documented as NOT probed — honest
port of the prior behavior, not a fake measurement. Pure `assemble` helper +
3 unit tests. Read-only.
cc
The playback test needs the rate the DAC is ACTUALLY clocked at, not the rate
QBZ requested (`player.state.get_sample_rate()` only proves "QBZ asked for X").
Ground truth on Linux is `/proc/asound/cardN/pcm*p/sub0/hw_params` (`rate:` line
while a stream is open, `closed` when idle). The card number is resolved from
the PipeWire node.name via `pw-dump` (robust — no pactl/pipewire-pulse).
New `qbz-audio::dac_probe`: `NegotiatedRate { sample_rate, format, channels }`,
pure `parse_hw_params` + `parse_alsa_card_for_node` helpers, and the composed
`negotiated_stream_rate(node_name)`. Read-only — never opens/reconfigures a
stream, protected bit-perfect path untouched. 4 unit tests (active 24/192
fixture, closed/empty/no-rate, card resolution from int and string props).
cc