Workspace flatten brought these operator-scoped state dirs (Claude
Code per-machine settings and remember-skill memory artifacts) into
the repo's working tree. Neither belongs in git — the canonical
project memory lives under ~/.claude/projects/ outside the repo.
Adds CoverageRollup type, getCoverageRollup() against
/api/admin/library/coverage, and createCoverageQuery() with the same
3s/30s pacing used by scan-status. New qk.coverage() key follows the
existing zero-arg tuple pattern. Vitest covers the GET path and the
empty-library zero-state plus the with_art + pending + settled = total
invariant.
Code-review polish on admin_coverage_test.go:
- Replace inline artist-seed SQL with the seedArtist helper used
across the rest of the api test suite (admin_covers_test.go,
likes_test.go, me_history_test.go, admin_quarantine_test.go).
- Rename the cover_art_source pointer locals from none/sidecar/mbcaa
to sourceNone/sourceSidecar/sourceMbcaa so they don't read like
zero-value identifiers.
New admin handler returns the library-wide cover-art coverage rollup
{total, with_art, pending, settled, pending_no_mbid} for the admin
dashboard gauge. Always 200; zeros on empty library. Same RequireAdmin
middleware as /api/admin/scan/status. Lives under a new /library/
admin sub-namespace, parallel to /scan/ (per-run state) and /covers/
(actions).
Tests gated on MINSTREL_TEST_DATABASE_URL: empty library returns all
zeros, mixed rows produce correct bucket math (verifying the
with_art + pending + settled = total invariant), non-admin returns 403.
Code-review polish on GetAlbumCoverageRollup's comment block:
- Records the invariant that with_art + pending + settled = total
(and that pending_no_mbid is a subset of pending, not a fourth
bucket).
- Notes that the IN ('sidecar','embedded','mbcaa') list must stay
in sync with the cover_art_source CHECK constraint in migration
0016 — without this note a future source addition could silently
undercount with_art.
FILTER-aggregate count over albums.cover_art_source and albums.mbid
returning total, with_art, pending, settled, pending_no_mbid in a
single round-trip. Sub-millisecond on realistic libraries; no index
needed for v1. pending_no_mbid is a subset of pending — the UI
tooltip uses it to show how many "pending" rows are blocked on
missing MBID and won't be moved by another scan.
Also bumps the Makefile sqlc pin from 1.27.0 to 1.31.1 to match the
version that produced all existing committed dbq/*.go files. The
prior pin was stale; running make generate against 1.27.0 silently
rolled the codegen back (cosmetic version headers on most files
plus a real codegen shape change in DeleteArtistIfEmpty).
Forward-fix from running the prior two commits in production. Backfill
on a real library exposed two further issues:
(1) gofmt -s flagged the column-aligned stubMeta methods in
mbids_test.go. Switched to the gofmt-canonical single-space form.
(2) Many albums failed to heal with SQLSTATE 23505 (unique constraint
albums_mbid_unique). Cause: the operator's library has duplicate
album rows in the DB — same MusicBrainz release, split into
multiple rows by the pre-MBID scanner because of subtle title or
artist disagreements between files. When backfill now correctly
extracts the MBID, two rows want to claim the same one and the
partial unique index rejects the second.
The conflict is correct DB behavior — we shouldn't have two rows
with the same MBID — but it's not a write failure to alarm the
operator about. Detect 23505 specifically:
- downgrade the log line from Warn to Info
- track these in a new BackfillMBIDsResult.Duplicates counter
(separate from Skipped, which retains its "no MBID in tag"
meaning)
- leave the duplicate row's mbid NULL; merging duplicates is a
separate (future) operator workflow
Same handling threaded through the scanner heal path so a regular
rescan doesn't generate the noise either.
JSON tally + admin UI gain a "Duplicates" line so the operator
can see how many duplicate rows their library carries.
Follow-up scope (separate task): a duplicate-album merge UX that
reparents the duplicate's tracks to the canonical row and deletes
the orphaned album row.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Follow-up to 947f944. With the dhowden/tag/mbz extractor now correctly
matching ID3v2 TXXX frames, mbid backfill started rejecting every row
with "invalid byte sequence for encoding UTF8: 0x00 (SQLSTATE 22021)".
Cause: dhowden's readTextWithDescrFrame (id3v2frames.go:454) — the path
that decodes TXXX text — does NOT strip the trailing/embedded NUL bytes
that ID3v2.4 uses as frame terminator and multi-value separator. Unlike
readTFrame, which does (line 313). So a TXXX:MusicBrainz Album Id of
"abc-123\x00" comes back to us verbatim, and Postgres refuses to store
NULs in text columns.
cleanMBID splits on NUL and returns the first non-empty segment, which
also handles the multi-value case (collaboration artist IDs are
NUL-separated; Minstrel uses the primary for MBCAA, so first wins).
Tests cover the trailing-NUL, multi-value, and all-NUL paths.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The hand-rolled extractor in internal/library/mbids.go looked up keys that
dhowden/tag never produces, for every supported audio format:
Vorbis (FLAC/OGG): dhowden lowercases keys at parse time (vorbis.go:59),
we looked up MUSICBRAINZ_ALBUMID. Always missed.
ID3v2 (MP3): dhowden stores TXXX frames under bare TXXX/TXXX_N keys
with *tag.Comm values; the Picard tag name is on the value's
Description field, not part of the key. We looked up
TXXX:MusicBrainz Album Id. Always missed.
MP4 (M4A): dhowden stores freeform iTunes atoms under their bare
sub-name. We looked up ----:com.apple.iTunes:MusicBrainz Album Id.
Always missed.
Net effect: every album in the m7-380 boot backfill reported
processed=N, healed=0, skipped=N. The m7-379 enricher gate then
short-circuited NULL on every album, so MBCAA was never queried.
Symptom: "all albums skipped" during cover enrichment.
The pre-existing unit tests passed because they hand-built Raw() maps
with the intended-but-incorrect keys, never round-tripping through
dhowden's parser.
Replace the hand-rolled extractor with dhowden's own mbz sub-package,
which knows the per-format Raw() conventions and uses lowercase
canonical keys (mbz.Album, mbz.Artist). Update both call sites to pass
the full tag.Metadata instead of just Raw(). Rewrite the test with a
stubMeta implementing tag.Metadata, asserting against the actual
per-format Raw() shape, plus a regression guard that pins "uppercase
Vorbis keys must not match" so the v0 bug can't sneak back.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds getScanStatus/triggerScan helpers, createScanStatusQuery factory
(3s poll), qk.scanStatus(), and a Library Scan section on the admin
overview page with per-stage tallies and inline Run scan button.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds GET /api/admin/scan/status and POST /api/admin/scan/run under the
existing RequireAdmin middleware block; plumbs *library.Scanner and
library.RunScanConfig through api.Mount, server.New, and main.go so the
manual trigger reuses the same RunScan orchestrator as startup scans.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds internal/library/scanrun.go with RunScan, which creates a scan_runs
row and sequences file-walk → MBID backfill → cover enrich, persisting
per-stage tallies as jsonb. Extends coverart.EnrichBatch to return
(processed, succeeded, failed, err) so the orchestrator can classify
outcomes. Replaces main.go's two separate boot goroutines with a single
RunScan call; passes nil scanner in the no-startup-scan branch.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a boot-time goroutine that walks albums with NULL mbid, re-reads
tags from one track per album via dhowden/tag, and persists album + artist
MBIDs. Healed albums also get their cover_art_source='none' cleared so the
enricher's next batch retries the MBCAA fetch. Caps at 5000 albums per
boot to avoid stalling startup on huge libraries.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- gofmt -s on system.go, system_cron.go, api.go
- rename unused r → _ in 3 fetcher_test.go HTTP handlers
- TrackRow +queue test uses /add .* to queue/i (track-aware aria-label from #377)
- /library/artists page test mocks likes + tanstack-query (ArtistCard now embeds LikeButton)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds cover_art_source to AlbumRef, two admin API helpers (refetchAlbumCover,
refetchMissingCovers), a per-album retry button gated on is_admin in the album
detail page, and a bulk-refetch section in the admin overview.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Production list-playlists handler reuses Service.List + in-memory kind
filter so other users' public playlists can surface alongside the
caller's filtered own. Comment clarifies the divergence so future
readers don't expect this query to be wired into the API path.
Push kind='system' guard into each of the five Service edit methods
(Update, Delete, AppendTracks, RemoveTrack, Reorder) immediately after
the ownership check; map the typed error to 403 in writePlaylistErr;
add table-driven test covering all five verbs against a seeded system
playlist.
Adds ?kind= filter (user|system|all, default user) to GET /api/playlists
so system mixes don't leak into slice-1 SPA views. Lazy background build
fires when kind=system|all and the caller's run row is stale (>24h).
Propagates Kind/SystemVariant/SeedArtistID through PlaylistRow and
playlistRowView wire types.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds internal/db/queries/system_playlists.sql with 12 named queries
covering active-user enumeration, per-user run tracking (claim/finish/fail),
seed-artist/track/cover selection, and playlist CRUD by kind.
Runs sqlc generate to emit dbq/system_playlists.sql.go; also updates
playlists.sql.go and models.go to reflect the new kind/system_variant/
seed_artist_id columns added in migration 0015.
Note: plan specified a.cover_path for PickTopAlbumCoverForArtistByUser
but albums uses cover_art_path — corrected in the query file.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Code-quality review flagged convention drift from sibling library
pages. Aligning so the history page matches the same shape:
- queryStore + $derived($queryStore) pattern (single subscription
point) instead of $query.X everywhere.
- h1 uses font-display text-2xl font-medium per FabledSword design
system (weights 400/500 only — font-semibold drifted to 600).
- h2 uses font-medium + z-10 for sticky-header layering.
- InfiniteScrollSentinel uses its `enabled` prop (which exists; the
earlier spec note claiming otherwise was wrong) so the observer
isn't recreated on every fetch cycle.
- Loading more… / End of history footers added to match albums.
- onRetry passes query.refetch by reference, not wrapped.
- Dropped redundant `as HistoryEvent[]` cast.
Also fixed test case 4 which trivially passed regardless of the
conditional gate's correctness — now queries the sentinel's actual
DOM root (div[aria-hidden="true"].h-px) and asserts presence/absence.
Added a positive twin test for hasNextPage=true.
Implements the listening-history endpoint for M7 #365: returns a user's
play events newest-first, excluding skipped events and quarantined tracks,
with offset/limit pagination and a has_more heuristic.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
CI test-web was blocking on:
- QueueDrawer.test close-button query: getByLabelText(/close queue/i)
matched BOTH the backdrop button (aria-label="Close queue backdrop")
and the close button (aria-label="Close queue"). Switched to exact
string match.
- QueueDrawer.test inert assertion: Svelte 5 + jsdom uses property
binding for `inert`, not attribute, so hasAttribute('inert') returns
false. Check the DOM property (with attribute fallback for
robustness) instead.
- queue-row-math.ts -0 normalization: Math.round(-0.5) returns -0,
and Object.is(-0, 0) is false — vitest's .toBe(0) fails. Added
`|| 0` to normalize -0 to +0 at the function boundary so consumers
never see -0.
- player/store.svelte.ts user import guard: persistence $effect.root
reads `user.value?.id` at module-init, but in test files where
auth/store.svelte.ts is imported transitively (auth/store.test,
playlists/playlists.test), the player module loads via auth's
import chain before auth's `user` binding is assigned. `user?.value`
guards against that init-order race. The proper fix is to invert
the auth↔player import dep (filed as I2 follow-up under #375).
CI svelte-check was blocking on:
- 1 ERROR I introduced in the previous cleanup: QueueDrawer.test.ts
passed { hidden: true } to getByLabelText, but that option only
exists on getByRole. Fixed by switching to a direct
document.querySelector for the aria-hidden assertion.
- 13 WARNINGS pre-existing in the codebase from M7 #352/#372/#349
era, never surfaced because earlier CI runs failed before reaching
type-check. Now that CI gets that far, they accumulate. Cleared:
- FlagPopover, RemoveTrackPopover, AddToPlaylistMenu, TrackMenu:
interactive role divs gain tabindex="-1" + onkeydown that stops
propagation and closes on Escape (functional, not just warning-
suppression).
- FlagPopover state-from-props: $state(untrack(() => prop ?? default))
for explicit initial-snapshot semantics; const isUpdate switched
to $derived so it reacts to prop changes.
- PlaylistTrackRow drag-div: role="listitem" added.
- playlists/+page.svelte: autofocus replaced with bind:this + $effect.
CI svelte-check failed on 5 type errors caused by the slice using
duration_ms/album_name fields that don't exist on TrackRef (the real
shape uses duration_sec + album_id/album_title). Fixed across
QueueDrawer, QueueDrawer.test, QueueTrackRow.test, persisted.test,
and the inline `state.current = null` in PlayerBar.test (TrackRef |
undefined, not | null).
Final whole-slice review concerns also rolled in:
- C1 (Critical): persisted position was restored to UI but never
seeked into the audio element — first timeupdate snapped _position
back to 0. Added a one-shot _pendingRestorePosition module ref +
consumePendingRestorePosition() export; layout's loadedmetadata
handler now seeks once on restore.
- I1 (Important): throttled-write effect now wraps _queue / _index
reads in untrack so the dependency tracking matches the design
intent. Position remains the only tracked dep.
- I4 (Important): drawer gets inert={!queueDrawerOpen} so its inner
buttons drop out of tab order when closed (aria-hidden alone
doesn't do that). Removed redundant role="complementary" from
<aside> (implied by the element). New test asserts inert binding.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Wrap the _position read inside untrack() in the immediate-write $effect
so it no longer registers as a reactive dependency, preventing the effect
from firing on every 4Hz position tick. Also adds three missing
restoreQueue unit tests with vi.mock for the auth store.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
M2 — Add TODO(#375) tripwire comments where the dark/light theme-color
hex pair is duplicated outside tokens.json (vite.config.ts and
applyMetaThemeColor.svelte.ts). Refactoring is out of scope for #363;
the comments are stop signs for the next person who edits these.
M5 — The SPA never reads window.__MINSTREL__.description. The OG meta
description is server-rendered and complete on its own. Drop the dead
inline-script field and the corresponding helpers in branding.ts.
Server-side BrandingConfig.Description stays — it's still used for
<meta name="description"> and og:description.
Copy nit — Playlist detail page title becomes "Playlist · Foo"
matching the singular form already used by Artist · / Album ·.
Nine golangci-lint failures accumulated across M7 #352, #372, and
6d1709c that were never surfaced because test-go never ran green long
enough to reach the lint step. The first push to land cleanly through
vet (the M7 #363 slice) ran lint and exposed them.
- errcheck: discard the error on defer Close() in collage.go,
collage_test.go, and server_test.go.
- gofmt -s: re-run on config.go and playlists_test.go.
- goimports: move the stray "time" import from the local-module group
to stdlib in server_test.go.
- revive unused-parameter: rename ctx to _ on stubScanner.Scan and
fakeLidarrUnmonitorer.UnmonitorTrack (test stubs); on
tracks.Service's adminID, add //nolint:revive directive rather than
renaming because the HTTP handler passes admin.ID (a real authed-user
UUID) and the existing comment already flags it as reserved for the
audit-log follow-up.
The legacy SMARTMUSIC_ prefix dates from when the project was specced
as "smart-music" before being renamed to Minstrel. Hard cutover to
MINSTREL_* across config.go, tests, config.example.yaml, README, and
docker-compose. Also renames SMARTMUSIC_CONFIG in cmd/minstrel/main.go
which was missed in the original audit.
Also seeds Auth.AdminBootstrap.Username = "admin" in Default() so
MINSTREL_DATABASE_URL is the only env var required for a fresh prod
deployment — the bootstrap auto-generates a password and prints it
to stderr. Drops the now-redundant MINSTREL_AUTH_ADMIN_USERNAME=admin
line from the README quickstart and docker-compose.
M1 follow-up from the #363 final review: guards
MINSTREL_BRANDING_APP_NAME and _DESCRIPTION against empty-string
overrides — empty would have shipped <title></title> to share-preview
crawlers.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
bvandeusen
