Two screens, a controller and a bridge. `QbzPurchases` publishes one JSON
document per screen plus a revision counter, so QML re-parses even when the text
is byte-identical.
- The list: two tabs with per-type counters, a server-side search on a 300 ms
debounce, and a toolbar that DIFFERS by tab — Albums gets group, sort,
direction, two filters and grid/list; Tracks gets group and the filters only.
Building the Albums toolbar on both would have invented two controls.
- The detail: a format dropdown whose selection re-scopes every downloaded mark,
the progress section and Add-to-Library together; a determinate bar counted in
TRACKS, never bytes; goodies as an album-level action counted separately.
- The album download is the frontend loop, not the Rust command the reference
registered and never called. Sequential, concurrency 1, cancellable, and a
failed track does not abort the album.
- Both prefs are read at runtime, not merely written. `show_purchases` is the
master gate and `nav_tb_purchases` only relocates the entry to the title bar;
both default false, so Purchases still ships hidden.
- Scanned local rows keep their raw `qobuz_purchase` source so the gold badge
survives in list mode as well as grid, while still classifying under the
Offline source filter. Remote feed rows deliberately stay `qobuz`: stamping
them would badge every purchase the user merely OWNS as locally downloaded.
Defects fixed after the screens were written, none of which any test could have
caught:
- `QbzPurchases.boot()` was called nowhere. Without the Qt-thread hop every
publish is a silent no-op, so both screens would have stayed permanently empty
with nothing logged. The cause was an ownership gap rather than the code:
`Main.qml` belonged to no lane because the plan listed the files each lane
CREATES rather than the files it TOUCHES, and "belongs to nobody" is
indistinguishable from "done" until something runs.
- A second download of the same album nested inside the first. The store rewrites
its destination to the album subfolder after the first successful track so that
Add-to-Library adds the album rather than the whole music root, and the download
planner read that value first — producing `.../Album [q]/Artist/Album [q]/...`,
one level deeper each time. The picked ROOT and the resolved ALBUM FOLDER are
now separate fields with names that cannot be confused, with a regression test.
- A tab counter could stick at zero forever with its list plainly full: the two
ids calls each swallow their error, and the reference's backfill from the list
response was missing.
- Cancelling made the whole progress block vanish instead of showing the
cancelled state, whose msgid was already translated in all eight catalogs.
- Add-to-Library re-walked the entire library when the folder id it needed was
already in hand.
- A live language switch and a logout left the previous state behind.
Deliberate divergences from the contract, each recorded with its reason: the page
header is a plain `fontSection` title with no icon tile, because nothing else in
this application announces itself over a large coloured square; purchase download
status travels inside the album document rather than on the shared cache channel,
which is keyed by track id alone and would let the two features overwrite each
other on the same row.
Gate: release build clean, QML audits 278 files / 37 bridges, `cargo test -p
qbz-qt` 398 passing, offscreen smoke green with zero QML complaints.
Not verified, and cannot be from here: every populated screen. This account owns
nothing, so only the empty states have ever rendered.
cc
Accumulated qt-frontend work from the 2026-08-11/12 sessions (cc),
uncommitted until now: the app-wide dynamic background in both modes
(AmbientField with GPU/Canvas arms, AppShell layering, ambient palette
pipeline), the wave-1 clip/batch sweep across the tree (alpha batches
449 -> 327, reproducible), the rounded-cover and icon-tint arms, and
scripts/qt-batches.sh (batches + presents/s sampler).
The GPU behaviour of these pieces was finalized in the follow-up pulse
commits; see qbz-nix-docs/qt-frontend/2026-08-11-scenegraph-batches/
GPU-COST-INVESTIGATION.md for the full campaign.
km
Block B6 of the miniplayer/tray contract. B5 landed a tray whose clicks
reached Rust and stopped there; this is the other half.
qml/Main.qml — hideToTray(), showFromTray() and the ONE close choreography
every exit now calls, plus the handlers for the tray bridge's four signals
and the mini's close-app signal. The close predicate is both conditions,
trayLive && closeToTray: with the tray off a close QUITS, because gating on
the setting alone strands the process with no window on a desktop that has no
SNI host. close.accepted = false happens before the hide — hiding never arms
Qt's quit check, an accepted close does.
The miniplayer's enter and exit are rewired onto those two verbs, which is
what makes the tray toggle honest: the reference hides the main window behind
the tray's back and its second tray click brings the main window up beside a
live mini.
THE EXITS ARE FIVE, NOT FOUR. The contract enumerated four and gave the grep
that finds them; re-run today it returns a fifth, because the kiosk shell
grew its own drawn close button after the contract was written. It goes
through the same choreography. Four POC-NOTEs die here.
showFromTray also un-iconifies. Qt's hide() preserves windowStates where the
reference destroys the surface, so a window minimized before it was hidden
would otherwise come back minimized — present in the taskbar, invisible on
screen, which reads as the tray having done nothing.
qml/settings/AppearanceSettings.qml + src/settings_qt.rs — the three macOS
rows: the MENU BAR header swap, the macOS wording on Close-to-tray, and the
Hide-Dock row, all absent on Linux. minimize_to_tray gets storage and no row,
which is the owner's standing decision. And the tray-close-to-tray write arm
now pushes to the bridge beside its persist, so toggling it changes the very
next close with no restart.
src/mini_bridge.rs — close_app and the signal it routes through, so the
mini's red X will honour close-to-tray like every other exit. Its button is
B3's, so the chain has no caller yet.
Driven, not assumed: with the tray live and close-to-tray just pushed false,
the app-menu Close row QUIT the process — which can only happen if the push
reached the bridge, since the seed had set it true. And Shift+M then Escape
returns the main window pixel-identical, zero of 1.6 M pixels changed, with
the mini alone on a black framebuffer in between: the main window really is
hidden, and Qt's hide/show preserves its geometry with no restore call.
cc
Contract sections 4, 5, 6 and the QML half of 7.
ContentRouter.qml is the D3 extraction: ONE view-mount chain, shared by both
shells. The Slint kiosk's content block is a verbatim copy of AppShell's and
KioskShell.slint:10-17 names this extraction as the follow-up it was waiting
for. 23 desktop arms carried over with their comments, 8 kiosk overrides, and
the kiosk-only nowplaying route. AppShell's four viewLoader.item sites now go
through contentRouter.currentItem; a tree-wide grep confirmed no other
reference (LESSONS L4 -- grep, do not trust memory about "both places").
KioskShell.qml carries the back bar with the full search focus lifecycle, the
content-height arithmetic including the now-playing exception, the offline
gate, the forced small transport, the player-zone focus ring, the root key
handler and the 30ms initial focus grab. Two findings worth keeping:
- the NavRail tiles fire header-menu-navigate("favorites-tracks") in the
reference, which switches the view AND selects a landing tab; navigateTo
carries only the route, so the tiles reuse the port's existing pending-tab
handshake rather than landing on the wrong tab.
- the visualizer tile's action is a sanctioned no-op (R2). Wiring
QbzImmersive would be worse than nothing: `open` would go true over an
empty mount and the key guard would then reject every key against a
surface that renders nothing.
Three fractional font sizes had to be rounded. Qt's font.pixelSize is an INT
property, so the reference's 11.5px (NavRail.slint:47) and 12.5px
(KioskNowPlaying.slint:59,67) are not expressible; font.pointSize is the only
real-valued alternative and is DPI-relative, which would make those three
labels scale differently from every other in the tree.
The NavRail one was caught by the offscreen smoke run IN THE KIOSK PROFILE --
it made the whole rail unavailable, i.e. a kiosk with no footer. The build,
both QML audits and qmllint were all green while that was true, and the
desktop smoke was clean because it never mounts KioskShell. The other two
were still latent: KioskNowPlaying is not on the smoke's route.
Gate: build 9 warnings = baseline, 291 tests, both audits OK over 229 QML
files (+15), qmllint filtered = the same 20 pre-existing ArtistView lines and
nothing else, kiosk-profile offscreen smoke clean.
cc
vicrodh