DLNA renderers, app updates told like Librarian, acquisition without Lidarr, Collection and admin rework (M502–M505, M542) #156

Merged

bvandeusen merged 117 commits from dev into main 2026-10-26 12:32:33 -04:00

Conversation 0 Commits 117 Files Changed 1154

+53547 -24475

bvandeusen commented 2026-10-26 12:32:29 -04:00

Owner

Copy Link

Copy Source

Everything on dev since #155 (117 commits). Highlights:

Playback and the Android app

• Any UPnP/DLNA renderer plays, not only a Sonos: a one-track-at-a-time driver beside the Sonos native queue (#6451).
• Casting no longer sticks on "loading" when the same queue plays again, and a replaced Sonos queue no longer falls back to the phone (#6456).
• A newer app is announced in a snackbar on Home, installed from a sheet, and downloaded ahead on Wi-Fi (#6452). An update is offered only when the app itself changed (#6455).
• The audio cache keeps Liked songs and Recently played budgets, and liking a track downloads it on Wi-Fi (#6307–#6309).

Acquisition and the library

• Minstrel searches, grabs, imports and verifies downloads itself; it no longer talks to Lidarr (M502).
• Duplicates settle themselves, and AcoustID names recordings cut to the wrong slot (M498 follow-ups, #6167–#6169, #6184).
• Library folders are set in Admin › Library, and moving a folder keeps plays, likes and playlists (#6129).

Web

• Collection replaces Wanted. Admin and Settings are left-rail pages with an Overview verdict (M503, M504).
• Notifications say what changed and alert once (M542). Opening the inbox marks it as read (#6275).

Server

• Subsonic: getAlbumList2 recent/frequent, getUser fix (#6292, #6293).
• Shutdown ends the live-event streams, so a redeploy drains cleanly (#5293).

CI is green on dev head 7d5dcd45 (run 983).

🤖 Generated with Claude Code

Everything on `dev` since #155 (117 commits). Highlights: **Playback and the Android app** - Any UPnP/DLNA renderer plays, not only a Sonos: a one-track-at-a-time driver beside the Sonos native queue (#6451). - Casting no longer sticks on "loading" when the same queue plays again, and a replaced Sonos queue no longer falls back to the phone (#6456). - A newer app is announced in a snackbar on Home, installed from a sheet, and downloaded ahead on Wi-Fi (#6452). An update is offered only when the app itself changed (#6455). - The audio cache keeps Liked songs and Recently played budgets, and liking a track downloads it on Wi-Fi (#6307–#6309). **Acquisition and the library** - Minstrel searches, grabs, imports and verifies downloads itself; it no longer talks to Lidarr (M502). - Duplicates settle themselves, and AcoustID names recordings cut to the wrong slot (M498 follow-ups, #6167–#6169, #6184). - Library folders are set in Admin › Library, and moving a folder keeps plays, likes and playlists (#6129). **Web** - Collection replaces Wanted. Admin and Settings are left-rail pages with an Overview verdict (M503, M504). - Notifications say what changed and alert once (M542). Opening the inbox marks it as read (#6275). **Server** - Subsonic: getAlbumList2 recent/frequent, getUser fix (#6292, #6293). - Shutdown ends the live-event streams, so a redeploy drains cleanly (#5293). CI is green on `dev` head 7d5dcd45 (run 983). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

bvandeusen added 117 commits 2026-10-26 12:32:29 -04:00

feat: Lidarr searches for a better copy of a rip that has none (M498 #5447) ... e17556dce9

The operator chose option 2: an album holding a video rip that is the only
copy of its song is searched in Lidarr (AlbumSearch), at most once a week
and ten albums a pass. Lidarr grabs only when the quality profile allows an
upgrade, so nothing is removed and the rip stays held back either way.

- lidarr.Client.SearchAlbums posts the AlbumSearch command.
- ListAlbumsWithSoleCopyRips finds the albums: a held-back rip with no
  present clean copy sharing its song or its pending group, and not searched
  in the last week (audit action lidarr_rip_search).
- Suspect sources shows when Lidarr last searched the track's album.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a release change Lidarr applied but did not answer is recorded (M498) ... e1ad1c9ba2

On the first deploy pass Lidarr took longer than the client's 30s timeout to
answer the release change for Cracker Island: its album update unlinks the
files and queues the rescan before it replies. The change may well have
landed, but the resolver logged a failure and wrote no audit row, so the
album had no settle window and the report did not show it.

An unanswered PUT (ErrUnreachable) is now read back: when the album's
monitored release is the chosen one, the change is recorded as made. Any
other error, or a read-back that disagrees, stands as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a merge keeps the song link when the removed copy names the song (M498) ... 8bd48b4302

MergeInheritSongLink took "linked" to mean song_key <> id, but the copy
whose id became the song's key has no song_id of its own. Removing that
copy dropped the survivor out of the song. It now asks whether another
copy shares the key. The test covers both copies as the one removed, so
it no longer passes or fails on which random id sorts first.

The release-change test counted audit rows other tests left behind
(ResetDB keeps audit_log); it now counts what its own pass adds.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: duplicates resolve themselves without Lidarr (M502 #5517) ... 5306a261c1

Lidarr is off, and with it the resolver had stopped merging: it only
merged when Lidarr had said which copies it mapped. Minstrel is now the
only program writing the library, so the rule is about the copies:

- A song held twice on one album: the extra copies merge into the one
  whose tags fit best (then lossless, larger, older).
- Across releases or under another title: when every copy but one is a
  video rip (held back, or named like one and not marked fine), the rips
  merge into the one that is not.
- Clean copies on different releases are linked as one song, as before.
- Anything else waits for the operator.

Gone with Lidarr: release switches and their settle window, the
unmapped-file read, the rip search, the merge guard and its 409/503,
lidarr.Client's release methods, and duplicate_group_members.lidarr_state
(migration 0078). The duplicates page drops the per-copy Lidarr badge
and the blocked-merge notice; past release changes stay in its history.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: indexers and download clients you add and configure (M502 #5518) ... a2e1890d4c

Admin → Integrations gains "Indexers and download clients": two lists of
connections the operator adds, Fabled Librarian's model. Nothing is seeded
or copied from Lidarr.

- Indexers: Prowlarr (its native /api/v1/search, which asks every indexer
  at once; Prowlarr has no aggregate Torznab feed) and any Torznab feed.
- Download clients: qBittorrent (WebUI v2, with re-login on an expired
  session) and FabledExchange, one entry with a switch per job:
  downloads (/api/downloads), search (/api/search) and release details
  (/api/torrents/{infohash}). All three start on; its key is optional.
- Each entry: enable, name, address, a write-only key or password, Test.
  A download client's Test lists its torrents and checks the downloads
  folder is visible and writable from Minstrel; an Exchange's Test
  reports each job switched on.
- Download clients carry a path mapping (Minstrel's /downloads vs the
  client's own name for it); search sources carry searches per hour.

Migration 0079 adds acquisition_services with its CHECKs. Ported from
Fabled Librarian: internal/indexer (Torznab, 429 Retry-After, magnet
redirects), internal/prowlarr, internal/exchange, internal/downloader,
and internal/acquire's one builder per kind, which Test already uses
so a passing Test vouches for the path a grab will take.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: a release catalogue from MusicBrainz; request search asks it directly (M502 #5519) ... 77389f88bc

Minstrel keeps its own record of what artists have released, read from
MusicBrainz, so the wanted list, search and import no longer need Lidarr.

- internal/musicbrainz: searches (artists, release groups, recordings) and
  browses (an artist's release groups; a group's releases with track lists).
  One process-wide pacer at 1.1s, shared with tag enrichment, because
  MusicBrainz's limit is per address. A 503 throttle is retried.
- Migration 0080: mb_artists, mb_release_groups, mb_artist_release_groups
  (a collaboration belongs to each credited artist), mb_releases,
  mb_release_tracks.
- internal/catalogue: Discography and Releases serve the held copy and
  re-read it when older than 7 or 30 days. An unreachable MusicBrainz
  serves the stale copy, marked. An hourly pass re-reads due artists.
- Art for what isn't held comes through /api/catalogue/art from the
  enabled cover-art sources (Cover Art Archive by release group, then the
  name-based sources) and is kept on disk, so the browser never fetches a
  third-party host. A miss is a 404 and the card shows its icon.
- Request search moves from /api/lidarr/search to /api/catalogue/search,
  with the same response shape. Albums are release groups, and "in
  library" matches albums.release_group_mbid. A 10-minute cache absorbs
  search-as-you-type. Discover suggestions point at the art endpoint
  instead of looking up Lidarr.
- Android calls the new path and loads Discover art through ServerImage.
  Installed builds older than this one lose Discover search until they
  update (no compatibility path, rule 22).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the result card's state prop no longer shadows the $state rune (M502 #5519) ... 867f52358c

svelte-check read `$state(false)` as a subscription to the `state` prop.
The prop keeps its name for callers and is bound as cardState inside.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: a wanted list — follow artists on release-type profiles; requests and re-acquisition become wants (M502 #5520) ... 04515a977c

Minstrel keeps its own list of what it is to get, keyed by MusicBrainz
release group. A want comes from an approved request, a new release by a
followed artist, or an album whose files went missing, and is held once the
library has a full edition of it.

- Migration 0081: release_type_profiles (Standard seeded: Album, EP,
  Single, studio only), follows, follow_seen_groups and wants.
  lidarr_requests becomes requests with plain *_mbid columns; approved
  album and track requests move onto the wanted list, approved artist
  requests become follows, and the library's album artists start out
  followed. missing_reacquisitions and the retry settings go.
- internal/wanted: follows are checked against MusicBrainz discographies
  daily. The first check sets a baseline, and only releases after it are
  wanted unless the follow asked for what already exists. Profiles filter
  by primary and secondary type, so live recordings stay out unless a
  profile ticks Live.
- internal/requests (was lidarrrequests): approving puts the request on
  the wanted list, and an artist request can be approved on a chosen
  profile. A request completes when its want is held.
- The re-acquisition sweeper wants a missing album back instead of asking
  Lidarr; searching for it is the search schedule's (#5521).
- Web: an admin Wanted page (Wanted, Missing files, Held, Following), a
  follow control on artist pages, a release-type profiles card on
  Integrations, "Approve on a profile" on Requests, and want state on
  Missing files. The artist detail API returns the artist's MBID.
- Android: the request body sends artist_mbid, album_mbid and track_mbid.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the wanted-list tests set a track's MBID with its source, and find the missing-files link in its row (M502 #5520) ... 1a1646a920

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: search and grab — music release names, a quality ladder, a paced search schedule and one live grab per want (M502 #5521) ... 89b3b44779

- internal/release: reads a release name (format, bitrate, bit depth and
  sample rate, source, year, video, discography) and judges it against the
  album wanted: the artist and title must be in it; video rips,
  discographies, other versions (drumless, instrumental, karaoke, remix,
  live unless the album is live…) and singles or EPs for an album are
  refused. Quality tiers on Lidarr's ladder; a .torrent's file list is read
  for format, video files and a track count short of the album.
- internal/acquire: fan-out across every search source with a narrowed
  search kept, a process-wide pacer per source, every candidate judged and
  kept with its reasons, and the grab: one live grab per want, tie-broken
  (a person's grab beats the schedule's; a better release replaces a worse
  one unless it is nearly done). Each download lands in its own folder
  under the client's downloads folder.
- internal/wanted: the search schedule, every ten minutes, never-searched
  first, then requests, missing files and followed artists, taking the
  best release when auto-grab is on (the default).
- Migration 0082: acquisition_settings (auto-grab, interval, allowed
  qualities seeded from Lidarr's "Any"), want search columns,
  want_candidates, grabs, release_blocks.
- Web: the Wanted page shows each want's last search, a Search button and
  every release found with its verdict (Grab, Grab anyway, Never this one,
  Allow again). Integrations gains "Searching and grabbing".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the search test reads the last pass as JSON, since jsonb respaces it (M502 #5521) ... 7f1fe57935

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the search schedule logs its state when it changes, so an idle pass says why (M502 #5521) ... 77081f1cef

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: import with verification, and a Wanted view of its own (M502 #5522) ... ebbc49b902

A finished download is checked against MusicBrainz before anything is
filed: the files are paired with an edition's tracks by position or path,
and AcoustID, tag ids, lengths and titles say whether it is the album.
What is the album is copied into the library as
Artist/Album (Year)/Artist - Album - NN - Title, tagged with Picard's
MusicBrainz ids, and scanned; the download keeps seeding. What is not is
refused: set aside, removed from the client, and the album searched for
again. A download the check cannot settle waits for a person, who imports
it anyway or refuses it. Tracks the library already holds are skipped.

The wanted list leaves Admin for a top-level Wanted view, admin-only and
web-only, with tabs for Wanted, Activity (downloads and their imports,
the ones needing a person first), Missing files, Held, Following (with the
release-type profiles, moved off Integrations) and Blocked. The search
settings gain "File new music into" when the library has several folders.

The search pass logs why each failed search failed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: WantList's prop is list, not state, which Svelte reads as the $state store; the length test's 4-minute case fits within 3% (M502 #5522) ... f268b4df09

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the integrations test's library folder isn't /music, which a Lidarr root folder there is named (M502 #5522) ... d9e00d8bc2

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: download lifecycle — stalled and failed downloads are given up and searched again; imported ones seed, then go (M502 #5523) ... 8245f4406e

Ported from Fabled Librarian's System 104. A new lifecycle poller, every
ten minutes:
- a download the client fails is given up at once;
- one nobody shares, or still fetching its file list, past the stall
  setting (6 h) is asked for again three times, then given up;
- giving up removes the download with its files, holds the release back
  for a week (release_blocks.expires_at), ends the grab and searches the
  album again, with no limit per album;
- a running download the client stops listing for half an hour ends and is
  searched again, not held back; an empty list is not taken as gone;
- an imported download seeds to ratio 2.0 or 14 days, then is removed;
- a download the client would not remove is asked again every pass.

Every condition and act is one open row per (client, download, condition)
in download_interventions. Activity shows what is wrong with a download and
how often it was retried, why one ended, its seeding, and a banner for a
client that does not answer. A running download can be given up by hand
(set aside for good). The seed ratio, seed days and stall hours are on the
search settings card. The download client lookup moves to
acquire.DownloadClientFor, shared by the importer and the lifecycle.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the Activity test's page carries clients_down as a value, not a literal Page<T> would refuse (M502 #5523) ... bc198a9389

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the library's albums go on the wanted list as held, once; after that the list is the truth (M502 #5567) ... 2c677b3184

A one-time job in the wanted worker lists every album in the library as a
held want, one per release group, with a record that it happened
(acquisition_settings.library_listed_at). Albums whose tags name only a
release are named on MusicBrainz first, 120 a pass; what nothing names is
counted and shown on the Held tab.

Re-acquisition now wants back only albums on the list. Taking a held album
off the list leaves its files, and it is never fetched back; albums put in
the library folder by hand afterwards are not added.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: every library album is matched to MusicBrainz before it is listed, by its tags, its artist's discography or a search, and checked against its tracks (M502 #5569) ... b887ab759b

The one-time listing took only albums whose tags already named their
release group and artist, and left 340 off the list. The rest are now
matched: the release their tags name, else the artist's release groups
called what the album is called, else a MusicBrainz search by artist and
title. A group found by name is taken only when its tracks agree with the
album's (by disc and number, then title; length and title, or recording),
so a partial copy still matches and a same-titled other album does not.

library_album_matches records each album the listing decided, listed or
unmatched with why, so an album taken off the list is never re-listed. The
albums no match was found for are named on the Held tab. Album rows with
no tracks are no longer counted.

The length and title checks move from the importer to release, shared by
both.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: under the album's own name, matching track titles are enough to match it; a refusal says which track differed and how (M502 #5569) ... 80d2fdc3ab

The first run on the library refused singles whose one track has the
album's title ("0 of 1 agree") and near-complete albums (11 of 14), because
every track's length had to fit as well. Lengths now only decide between
groups whose names merely overlap. The reason for a refusal names the first
track that differed, with both titles and lengths, so what is left can be
judged. Migration 0087 clears the refusals so those albums are matched
again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a second copy of a track counts once, joining words don't part two titles, and a newer matcher retries an older one's refusals (M502 #5569) ... bf409d5f3f

The second run's refusals were mostly albums holding a track twice (a
two-track single with three files), which no edition can pair, and
release names that differ only in "+" against "/". Tracks are now read
once per disc, number and title, and "and", "&" and "+" count for nothing
when titles are compared; an edition's own title also counts as the
album's name.

library_album_matches records which matcher version decided each album.
Refusals by an older version are matched again, even after the listing is
done, so an improvement reaches them without a migration.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the matcher-retry test stamps the listing's cutoff after its albums exist, as a listing done before them would not see them (M502 #5569) ... 1936fe77c6

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: an album holding every track of an edition, and more besides, is that release group (M502 #5569) ... 7ef1fca90c

Savant's "Void" folder holds all 23 tracks of the VOID (DLC) edition plus
20 more, and was refused at 23 of 43. A whole edition of three or more
tracks now matches whatever else the folder holds; singles and two-track
releases still need most of the album's tracks, since many albums hold
their songs. Matcher version 4 retries the refusals.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a release group whose title names a live, remix or demo version is wanted only when the profile allows that version (M502 #5573) ... ca57133c8a

MusicBrainz types My Chemical Romance's "Common People (Recorded For BBC
Radio 1)" as a plain Single, so a studio-only follow wanted it. Passes now
reads the title's qualifier (brackets, or after a dash) for the version it
names and treats it as a secondary type; a version no type names, an
acoustic or instrumental one, no profile passes. The follow check takes off
the list an open follow want the rule now refuses. Judge counts bbc,
sessions and unplugged as live.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the Wanted view lists artists A–Z with their albums inside, each album's tracks a click away (M502 #5572) ... a4fbe44bd5

The list runs like the library: artists by the library's sort name, each
artist's albums oldest first, grouped under the artist (linked into the
library when it holds them). A filter narrows it to an artist or album by
name, on the server, so it finds what has not been scrolled to. Tracks
opens an album's track list: the library's own for an album it holds, a
lost file marked, otherwise MusicBrainz's standard edition, named.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the Wanted view becomes Collection, shaped like Lidarr: an artist catalogue first, the state sections down a left rail (M502 #5577) ... 358849de77

The section opens on a catalogue of every artist on the list, A–Z with a
letter rail, each card showing how many of their albums are held. An artist
opens on their page: every album of theirs, held or not, with its status,
search, releases and tracks. Wanted, Missing files, Activity, Following and
Blocked sit down a left rail as the sections that show state; Held is gone,
the catalogue shows it. The section moves from /wanted to /collection.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the player bar's heart is Lucide's, outlined then filled as on Android, and its title and artist read at the size of the rest of the page (#5580) ... 10be068dcc

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Collection › Unmatched lists the library albums no MusicBrainz album matched, each matched by hand from a search or a link, or left off (M502 #5579) ... a8786abace

Like Lidarr's Unmapped Files. A local band's album, which MusicBrainz will
never have, is left off: never matched again and not counted. A re-tagged
album is matched again on the next listing. The Collection's verdict now
says when library albums need matching.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the re-tag test sets its clock one statement at a time, as a prepared statement takes only one (M502 #5579) ... 2f2573dd07

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Now Playing's cover grows to fill the screen, and the page takes the cover's dominant colour as Android does (#5592) ... c0c0f8a71f

The web colour helper now picks Palette's dominant swatch (5-bit
quantise, Palette's default filter, median cut to 16, the most-populated
box) instead of the 1x1 mean, so the web and Android agree; PlayerBar's
accent strip shares it. Now Playing draws Android's gradient (55% at the
top, 18% at 45%, then the page) as a masked background colour that fades
over 600ms between tracks, holding the old colour until the new resolves.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Settings is one page per section down a left rail, Account first, every setting a row; the profile form no longer saves blanks over your name and email (M503 #5595 #5596 #5597) ... c94c09aeae

- /settings lands on Account: the Android app, profile, password, signed-in
  devices. Then Playback, Appearance, Notifications, Connections, About.
- Each setting is a row (name, one short line, control at the end) in a
  capped-width column; one-of-N choices are a shared Segmented control.
- Collection and Settings share one SectionRail.
- Connections leads ListenBrainz with a status line: not connected,
  connected and not sending, or sending with the last scrobble's age.
- Notifications prints its column labels once; the digest email's
  settings link points at /settings/notifications.
- #5596: the profile form now loads what is stored and sends only the
  fields that changed. It used to start blank, and the server reads a
  blank as "clear", so one Save wiped the display name and email.
- Recommendation metrics moves to Library › Your listening, and Hidden
  tracks gets a Library tab.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: admin bugs from the audit: a playback-error delete no longer reports failure, a failed Overview load no longer reads as all-clear, and email settings can't be saved over with defaults (M503 #5604) ... 7c89e59536

- Playback errors: Delete file stops resolving the error row the track
  deletion already took with it (ON DELETE CASCADE), which 404'd into
  "Delete failed".
- Overview: a queue that fails to load says so with Try again, and its tile
  shows a dash, not 0. Delete via Lidarr shows only for an album Lidarr
  holds while Lidarr is connected. An approve refreshes the Collection.
- Quarantine: Delete via Lidarr is in the row menu, so a phone reaches it;
  "1 report", not "1 reports".
- Duplicates: the Keep badge follows the copy picked.
- Integrations: provider save failures and load errors are said, not
  swallowed or shown as "No providers registered"; the Lidarr test catches
  a failed call; SMTP fills once and Save stays off until it has loaded.
  Lidarr's blurb says what it is still used for.
- Requests: release profiles point at Collection › Following.
- Library: an empty library sends an admin to Admin to run a scan, and
  tells anyone else the admin scans it, instead of opening Settings.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Admin is a left rail with how many wait on you in each queue, and the Overview opens with a verdict, the queues that need you and since when, then the clear ones on one line (M503 #5605) ... ace7e22e6a

- GET /api/admin/summary: every admin queue (requests, reported tracks,
  playback errors, duplicates, suspect sources, missing files) with its
  count and the age of its oldest item, in one query.
- The admin layout drops the 11-tab strip for SectionRail, as Settings and
  the Collection have, with each queue's count as its badge; the counts
  refresh on each admin navigation and every 30s.
- The Overview leads with "Nothing needs you." or "N things need you.",
  lists each waiting queue with "oldest 3d ago" and a link to it, counts a
  failed library scan as a problem, and names the empty queues on one line.
  The request and quarantine previews and the stat tiles are gone; those
  queues are worked on their own pages. A failed summary says so with Try
  again instead of reading as all-clear.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Settings uses the width of a wide screen: groups sit in two columns, each narrow enough that a row's control stays near its name (M503 #5595) ... 864eb45580

The content was capped at max-w-3xl, so a wide window showed one narrow
strip and a lot of empty space. SettingsColumns lays a page's groups out as
one column on a phone or narrow window and two side by side from xl up;
the page cap moves to 100rem so the columns don't stretch on a very wide
screen.

- Account: Android app, Profile and Password, with Signed-in devices beside.
- Playback: "Between tracks" (crossfade) beside "Volume leveling".
- Connections: ListenBrainz beside Subsonic apps.
- Appearance, Notifications, About: one column's width.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Admin › Reports is one queue of broken tracks, whether a listener flagged it or it failed to play, one row per track with one set of actions (M503 #5606) ... 89a2653995

- /admin/reports folds the quarantine queue and the unresolved playback
  errors into one row per track, newest first: who flagged it and why,
  what failed and for whom, the file, with Resolve (clears the reports and
  marks the errors fixed), Play, Copy details, Delete file and, for an album
  Lidarr holds, Delete via Lidarr.
- One delete path: Delete file removes the track, and its reports and
  errors go with the row.
- A Handled tab shows what was done: reports cleared, files deleted,
  errors marked fixed or ignored.
- The summary counts reports by track across both sources, so the rail,
  the Overview and the page agree.
- /admin/quarantine and /admin/playback-errors redirect to Reports: the
  notifications and the Android app link to them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a report's details keep the space before a note, and the Reports menu test opens the row it means (M503 #5606) ... 0cc4bc686a

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Integrations splits into Library, Acquisition, Metadata sources and Email & network, each a page on the admin rail; the Overview keeps only what needs you (M503 #5607) ... 7e1cfb8d08

- Library: the folders, Run scan and a row per scan job, cover art, fingerprints and loudness (off the Overview and Duplicates)
- Acquisition: indexers and clients, searching, re-acquisition (off Missing files), Lidarr
- Metadata sources: AcoustID, and one provider block for cover art and tags in place of the pasted pair
- Email & network: outgoing email, the notification schedule, public address, proxies
- Lidarr, outgoing email and the provider block are rebuilt as SettingsGroup rows; failures read as sentences, not codes
- Duplicates and Missing files say what their setting is and link to where it is set

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the admin settings cards read as menus: each a titled group of rows with one short line, status first, failures in words (M503 #5607) ... 7f913e3e70

Indexers and clients, searching, missing files, Lidarr's neighbours, AcoustID, fingerprinting, loudness, outgoing email's schedule, public address and proxies move from paragraph cards to SettingsGroup and SettingRow. Switches are role="switch". Also fixes the two type errors run 9023 found: the Lidarr test result is narrowed before its lists are read, and the tag-source save mock carries version_bumped.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the AcoustID and Library tests expect the new copy and one folder per line (M503 #5607) ... 0049b64ff1

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Users, Diagnostics and Tuning sit on the admin layout like the rest: debug mode is a row action on Users, every admin load error has Try again, and figures carry their scale (M503 #5608) ... a5ad301c94

- Users: a one-line count of accounts, admins and open invites; Turn on/off debug mode in each row's actions, and a "debug on" badge that opens that account's events
- Diagnostics: no title or paragraph of its own; devices named by account, OS, app version and last seen, not client id; opens on ?account=
- Tuning: each profile, the taste build and Discover are row groups with one-line hints; the sparkline has a 0–100% axis
- Requests, Duplicates, Missing files and Suspect sources: load errors offer Try again
- Library: cover art reads "N of M albums have art"

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: a held album below the quality you settle for, or holding a video rip, stays wanted for a better copy, and the import replaces the old tracks with their likes and plays carried over (M502 #5524) ... c54792c51e

- acquisition_settings.upgrade_until (default High quality lossy, Lidarr's
  cutoff, note #5527; "" is off), an "Upgrade until" row under Searching.
- The wanted worker judges held albums hourly: tier per format from size per
  second (the scanner records no bitrate), the album's worst. Below the
  cutoff it is wanted for quality; a video rip among its tracks always.
- The schedule searches upgrades weekly, after open wants; a release no
  better than the copy held is refused ("not_better").
- The import matches each file to the track it would replace (recording,
  then disc/track) and replaces it when it is a rip or the file is a better
  tier: over the same path in one rename, or filed beside it and the old
  track folded in (library.ReplaceTrack: file first, then foldTrackInto).
  Nothing better: the release is set aside.
- Collection › Upgrades (Lidarr's Cutoff Unmet); Suspect sources says when a
  clean copy was looked for instead of Lidarr's search.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Minstrel no longer talks to Lidarr: its connection, Delete via Lidarr and the unmonitor options are gone, and the track reports are plain quarantine (M502 #5525) ... 3a16732883

Migration 0091 drops lidarr_config and renames the quarantine tables, types,
indexes and the history column off the Lidarr name. The Integrations Lidarr
card, the Reports "Delete via Lidarr" action, the merge and remove-track
unmonitor options, the Lidarr audit actions and error codes go, on the
server, web and Android. The README describes acquisition as it is now.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the reports test name keeps its apostrophe inside double quotes, and the merge API test sends no unmonitor (M502 #5525) ... 12725e71b0

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: Reports, Duplicates and Suspect sources live under Collection, with their counts on its rail and a "To look at" line naming the ones waiting; notices link there (M504 #5625) ... 38b2325d2d

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: one Missing files page under Collection: folders gone, each with its album's want, last search or download, and Search, Releases and Remove; an album held before its files went is no longer shown as held again (M504 #5626) ... ead0abf0cf

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the admin Overview shows who is listening, downloads, the library and its disk, and the server's version, uptime and recent errors; a download waiting on you, a client not answering or a full disk count in its verdict (M504 #5627 #5628 #5629 #5630) ... c1305a547d

The version moves to internal/buildinfo, so /rest/ping reports the build
instead of "dev".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: FLAC, Ogg and Opus tracks carry their MusicBrainz recording id from the tag: one reader in library gives the scanner, the backfills and the importer the same ids, and the next scan re-reads the library (#5637) ... 9e7d50b38e

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

refactor: track lengths and file sizes come from the shared formatters on Missing files, Duplicates, a want's Releases and the app download; an hour-long track reads 1:01:05, not 61:05 (M505 #5639 #5640) ... 622f248ea2

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

refactor: a running download is worded the same on the admin Overview as on the wanted list, from one liveDownloadLine (M505 #5641) ... af2d4f9859

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

refactor: every count in the web app is said by one plural(), with the viewer's thousands separator; a source guard names any hand-written copy (M505 #5642) ... 9e24b2b4fc

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

refactor: the server says a count one way, words.Count, for notifications, a wanted match's reason and the duplicate resolver's activity line (M505 #5643) ... db15713798

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

refactor: the admin and user Requests pages draw a request with one RequestRow, and the kind chip is one KindPill there and on Hidden tracks (#5651) ... 453a977e48

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the library's folders are set in Admin › Library, and changing a folder moves its tracks with it, keeping their plays, likes and playlists (#6129) ... edcb3a57ef

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test: TestPlanRoots looks for the empty-list refusal as the message words it (#6129) ... 2365514e5b

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: notifications say what changed and alert once: a grouped notice names the tracks, grows without re-alerting, and a phone pile lists its titles (M542 #6146-#6151) ... 854c307834

- An update to a grouped notice keeps its arrival time and email state, so the
  phone and the email no longer go off again each time it grows.
- Notices carry the names of what changed (newest ten), shown as three and
  "and N more": tracks still missing after the scan's move/replacement fold,
  duplicate groups new to a sweep, songs merged automatically, and the track
  and listener a playback error was for.
- Duplicates found counts only the groups new to that sweep; playback errors
  count reports, not the unresolved total.
- A scan failure says what stopped and links to Admin > Library.
- Android: a pile of more than three alerts lists the newest five titles.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the playback-error test matches the prefixed username, and announcementFor keeps detekt's two returns (M542) ... 8538f0dc68

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: duplicates settle themselves: retitled copies across releases link, alike versions on one album are dismissed, and a file in another song's slot is removed so the album is fetched again (M498 #6159-#6162) ... 0f58ebcd6c

The operator's library had 536 groups waiting for review, 2026-10-19. A sample
showed most were one recording under two titles on two releases.

- The headline and the Duplicates count leave out cross-release groups, which
  the resolver keeps and links: the headline read 3,479 when 536 needed review.
- Any group spanning albums is cross_release and is linked as one song,
  whatever its titles say. Nothing is deleted.
- One album, two versions named as such ("Lovesick (Instrumental)") and not
  byte-identical: a new class, distinct, dismissed by the resolver.
- One album, one recording under different song titles (mismatch): the
  album's cached MusicBrainz track list decides. When exactly one slot's song
  is as long as the file and every other slot's song is clearly not, the other
  copies are removed and their tracks marked missing, so the reacquisition
  sweeper wants the album back. Otherwise the group stays for review.
- The duplicates_found notice moves from the sweep to the resolver: it names
  only groups a pass first left for review.
- Resolved automatically lists dismissals and wrong-slot removals beside
  merges.

Migration 0094 adds 'distinct' to the class CHECK (rule 36).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a wrong-slot removal reads without a counted noun (M498) ... 80c711ac9d

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: a version named after a dash counts as a version, and a song held twice beside a wrong slot no longer blocks it (M498) ... c5868c00cc

- songKey drops a dashed version suffix (" - Instrumental Version", " - Radio
  Edit", live, remix, mix, acoustic, demo, reprise, rework, VIP), so Andora's
  "Goodbye Yesterday - Instrumental Version" is distinct, not a wrong slot.
- wrongSlots accepts several fitting copies when they share one title
  (Killjoy twice beside Dreamland): the other slot's copy is removed, and the
  doubled song is left for the next sweep to find as a pair the resolver merges.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the duplicate resolver fetches a missing album track list from MusicBrainz to decide a wrong slot, five albums a pass (M498) ... c8e855ae4a

20 of the 34 groups left for review had no cached track list for their album:
the catalogue only reads one when the wanted list or the importer needs it.
The resolver now asks catalogue.RefreshReleaseGroup for an album it must
decide and the catalogue has never read, through the shared MusicBrainz pacer,
at most five a pass (hourly). An unreachable MusicBrainz stops the pass's
fetching; a group read and found empty is not asked about again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: when no slot an album gives one recording is its own, every copy is removed and the album fetched again (M498) ... a3d9effca3

Shnabubula's whole album, one 45-minute file, was filed as both Fading Light
and Looking for Answers. Neither slot fits, so wrongSlots had nothing to keep
and held the group. Every copy now goes when each is clearly another song's;
a copy near its slot still holds the group.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the AcoustID rate-limit test bounds arrivals by the service's 3 a second, not the client's send spacing ... 304caaee02

Run 9543 saw two arrivals 389ms apart though the client sent them 400ms apart.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a slot emptied of another song's file drops that file's recording MBID, so missing-pair folding cannot fold it into the copy kept (M498) ... 8daa3d07b5

On 2026-10-20 twelve of the 24 slots the resolver emptied were folded by MBID
into the kept copy, so their albums were never wanted back.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: clearing a wrong slot's MBID clears its source too, as tracks_mbid_source_pair requires (M498) ... 39029b5cd9

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: an album with missing tracks is wanted back by its release group, not only when it carries a release MBID (#6167) ... 460e78c955

The release id was only ever used to find the group. 238 albums matched to a
group carried none, so their losses, five of the resolver's wrong-slot
removals among them, were never fetched again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: the duplicates review says what each match rests on and plays every copy (#6168) ... e3eddab80f

Each group now states its evidence: identical audio data, or the share of
audio fingerprints that agree. Copies titled as different songs in one
acoustic group say so and ask the reviewer to listen, and each copy has a
play button so they can.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: one recording cut to several songs' lengths is named by AcoustID; the copy in its slot stays, the rest go, and the release is set aside (#6169) ... 18d31e9cdd

Savant's Mortals held one recording as 16 of its 50 tracks, each trimmed to
its own slot's MusicBrainz length, so the lengths could not say which slot
was its own and every group waited for the operator. When a one-album
mismatch is the same audio (identical, or fingerprints within 0.02) and
every copy fits its own slot, the resolver asks what the recording is: the
copies' own AcoustID lookups, or a fresh lookup of one copy (five a pass).
The track list names each slot's recording; the copy in that slot is kept
and the others removed, or every copy when the recording is in none of
their slots or AcoustID knows none. AcoustID unreachable or answers in
several copies' slots: the group waits.

Whenever wrong-slot copies are removed, the download the files were
imported from is set aside, so the album fetched again is not the same
files. Albums imported before Minstrel downloaded anything have none.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: AcoustID decides any one-album mismatch the lengths cannot, not only copies each cut to its slot (#6169) ... 3a482bd715

The first pass on the operator's library held 16 groups. In several, every
copy's own lookup names one recording (Wombats Record Shop/Tormentor,
Alchemist (2)/Prelude, Monday Chill 8/9, 404sessions), but one copy sits 3-10s
from its slot, so the every-copy-fits precondition kept AcoustID out. Now:
copies whose lookups all agree are that recording, whatever their
fingerprints read; copies of the same audio are asked about as before; and
the track list places the recording in one copy's slot, or none. Copies
filed as one song are not settled here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the duplicate resolver waits for a running scan, so a copy it removes is not marked present again (#6169) ... 42e48efb79

At boot the scan lists the library, the resolver removes wrong-slot copies
and marks their tracks missing, then the scan reconciles from its list and
marks them present: 7 of the 13 removed on 2026-10-20 15:00 read as
present with no file, and would not be wanted back until a later scan. A
pass now waits while a scan younger than StuckScanThreshold is in flight
and tries again a minute later; a stuck run does not hold it up.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: imported FLAC, Ogg and Opus files keep their track and disc numbers (#6171) ... 80ecc2f7e7

The importer wrote track=N/M for every format, which ffmpeg stores in a
Vorbis comment as TRACKNUMBER=3/20; dhowden/tag reads a Vorbis TRACKNUMBER
with a bare Atoi, so it came back 0. All 245 tracks imported since
2026-10-10 (Humanz, Plastic Beach, the Gorillaz super deluxe and twelve
more albums) have no track or disc number, and their album pages have no
order.

- The importer writes a Vorbis track and disc number alone, with
  TRACKTOTAL and DISCTOTAL, as Picard does. ID3 and MP4 keep N/M.
- ReadTrackAndDisc takes the number before a slash when the Vorbis reader
  returns 0, for the scanner and for the import's own check of a download.
- tagReadVersion 6 re-reads the files already held (tag reads only).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: an album holding a song twice beside its single keeps one copy (#5401) ... 09e74a57a2

A cross-release group was only ever linked, never merged, so an album that
held the song twice among the releases kept both: Humanz has Saturnz Barz,
Andromeda, We Got the Power and Garage Palace as its FLAC and a YouTube mp3,
each beside the single. The resolver now merges an album's extra copies
into its best (the one that is not a video rip, else ProposeSurvivor:
lossless first) with ReplaceTrack, history and all, then links what is
left as one song.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a slot the resolver emptied folds only into a copy of its own song (M498) ... 3ebd35e51d

On 2026-10-20 the emptied "Open the Gateway" slot of Auvic's Speed Demons
was folded by title into another file tagged "Open the Gateway" that holds
Onward (244 s against the song's 254 s), so the album was never wanted
back. A wrong-slot removal now stamps tracks.emptied_at (migration 0095)
when the album holds no copy of the song as long as its MusicBrainz slot,
and missing-pair folding pairs such a row by title only with a track added
since: a copy fetched again. A slot whose song the album does hold (VOID
DLC's Weeper beside Void's) folds into it as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test: the album-copy fold test makes each album once, as UpsertAlbum folds only by MBID (#5401) ... 16a152b862

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a duplicate fold never deletes a lossless copy for a lossy one, nor folds two versions (M498) ... f32abc974f

The first album-copy fold pass on 2026-10-20 removed five FLACs (Måneskin
Rush!, Weezer 2016) in favour of mp3s whose only edge was a recording MBID,
and an explicit "Suite-Pee" for its clean version. ProposeSurvivor now ranks
lossless before tag fit; tag fit still outranks file size, which is what the
Humanz rips needed. The fold keys copies by their version words (title and
file name), so a clean, live or instrumental copy holding the same title tag
is linked, not folded.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: same-titled albums of two release groups are two albums, and a missing track with no disc is on disc 1 ... e45561e282

The scanner found an album by artist and title alone, so Weezer's six
self-titled albums (1994–2026) were one 118-track row and their merges,
wants and upgrades counted them as one. A same-named row whose release group
is known and differs from the file's is now another album; tagReadVersion 7
re-reads the library so existing tracks move to their own rows.

WantHoldings compared discs with IS NOT DISTINCT FROM, so a missing track 6
with no disc was never covered by a present disc 1 track 6. Air's Love 2
stayed wanted and was downloaded again for nothing.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a file tagged with no release group joins the same-named album its folder holds ... 210f884b2d

Under e45561e2 a file with no group tag took the oldest same-named album row,
which left three of Weezer's 1994 bonus tracks on the 2016 album. The lookup
now ranks same-named rows by how many tracks they hold from the file's
folder; tagReadVersion 8 re-places the existing ones.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix(android): sync pulls every page of the change log, and a 404 drops the cached row (#5654) ... 42132810b0

The server returns at most 1000 changes per /api/library/sync call and
expects clients to loop with the returned cursor. SyncController took one
page per trigger (cold start, reconnect), so the cache fell behind a log
that a full re-read grows by ~26,000 rows and, past the 30-day compaction,
reset through 410 to cursor 0 without ever reaching the merge deletes of
2026-10-09/10. The freshness sweep then re-asked for those tracks every
30 minutes: the same ~100 404s on every pass.

- SyncController pages until 204, storing the cursor after each page, and
  stops on a cursor that does not advance.
- MetadataProvider drops a cached album, artist or track the server
  answers 404 for, so rows orphaned before this fix stop being refetched.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test(android): name the 204 body type so Response.success resolves (#5654) ... 77f06b60f7

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test(android): run the sync pass in the test body, not through init's background collector (#5654) ... 46e1dd6c9c

advanceUntilIdle does not count backgroundScope work as pending, so the
pass init would start never ran. The network state is Offline so
construction starts none, and each test calls syncSafe itself and asserts
lastSyncError stays null.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: ListenBrainz similarity is stored relative to the seed's best, and the candidate arms keep tier order (#6177) ... e44025435e

LB's similarity scores are counts (16 to 1445 for recordings, to 10,875
for artists), and they reached a scorer that reads similarity as 0 to 1.
One edge added 64 to 5,800 points under Songs-like, so likes, skips and
recency stopped mattering for any candidate with an edge.

- ResolveListenBrainzTrackEdges divides each edge by the seed's best
  score in its whole cached answer. The cache keeps the raw answer.
- The worker stores matched and unmatched artist rows over the seed's
  top artist.
- Migration 0096 rescales the rows already stored.
- The candidate arms hold sim_score bands so tier 1 stays above tier 2:
  lb_similar 0.5 + 0.5 x edge, tag_overlap halved, likes_overlap 0.3.
  Otherwise a seed's weaker LB edges fell below a shared genre tag.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a track shorter than the fingerprint span keeps its print (#6178) ... cb1f619a21

The image's fpcalc 1.5.1 exits 3 with "Error decoding audio frame (End
of file)" whenever it reaches the end of a file before -length, at any
length, so every track under two minutes failed both the AcoustID lookup
print and the duplicate-detection print: 471 lookups and all 1,979
present short tracks' chromaprints.

- runFingerprintTool returns the tool's stdout beside a toolExitError.
- fpcalcReadToEnd keeps a partial exit whose stderr is the end of file
  and whose declared DURATION is within the span asked for. A damaged
  file that stops early declares more than it read, and still fails.
- Migration 0097 clears the stored failures so the lookup worker and
  the fingerprint backfill take those tracks again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a release its download client will not let Minstrel follow is held back, not taken every pass (#6176) ... 484a81f779

The Black Keys' Ohio Players was grabbed 40 times in a day. FabledExchange
already held the torrent under another category from before Minstrel took
the downloads over: the add answered "duplicate", which counted as added,
and Exchange's listing, filtered to Minstrel's category, never showed it.
Half an hour later the grab ended gone, and the next pass took it again.

- A duplicate add that the client does not list for Minstrel is not a
  grab: the release is held back for a day with the client's own reason
  (already downloading X (category)), and the search moves on. Forced,
  it is refused with that reason.
- A release that ends gone a second time within a week is held back for
  a day, whatever the cause.
- release_blocks gains the reason 'gone' (migration 0098), shown as
  "Held back: its download client lost track of it".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the taste arm round-robins across taste artists, and tag ties follow the order seed (#3880) ... 5bcbc80555

Measured on the operator's library (26,005 tracks, pool scale 2.28): the
arms the library scale grows all have far more rows than it asks for, and
two arms collapsed whatever their limits were.

- taste_overlap ordered by weight then track id, so the strongest taste
  artist's catalogue filled it: 20 tracks of saib. (197 tracks) out of 174
  taste artists, the same lowest-id tracks every time. It now takes one
  track per taste artist, strongest first, md5-seeded, as similar_artists
  does since #5297.
- tag_overlap broke ties in no order. A one-tag seed ties 7,804 tracks,
  and the LIMIT took whichever Postgres reached first, which the order
  seed did not control. Ties now follow md5(id || seed).
- library_scale.go records the per-arm measurement beside the reference
  size it leaves as is.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: a group of one recording in several slots is settled by getting the album again, not merged (#6184) ... b9684e6d12

The duplicates review offered only Merge for a wrong-slot group. A merge keeps
one copy's title whether or not it is the audio's, and neither sets the release
aside nor wants the missing song back.

For class mismatch the page now asks which song the recording is. "Get it
again" (the operator can't tell) removes every copy; "Keep <title>" keeps that
copy and removes the ones under other titles. Either way it runs the resolver's
own path: the release the files came from is set aside, the removed copies'
tracks stay marked missing and emptied, and the album is wanted again.

- library.SettleWrongSlot, POST /api/admin/library/duplicates/{id}/refetch
- ListDuplicateGroupsForResolve takes an optional group_id
- removeWrongSlots takes the acting operator (manual: not automatic, actor in
  the audit row) and removes files through removeTrackFileOnDisk, so a
  read-only library answers library_not_writable

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: the admin requests list names who asked, not a cut-off user id (#6265) ... c98bd359f7

/admin/requests showed "by user f7e1ec86" — the first eight characters of
the requester's id, a placeholder from the first build. The admin list now
sends each requester's username, looked up for the whole page in one query,
and the row reads "by <username>". The id stays as the fallback for a list
that comes back without names.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test: the admin requests list test names the requester by their seeded username (#6265) ... 687a7cf3ee

seedUser stores users as test-<name>, so the assertion compares against
alice.Username rather than the literal "alice".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: Android's admin requests names who asked from the request itself (#6267) ... fefc51439e

The admin requests list now sends each requester's username (#6265), so
the Android queue reads it off the row instead of fetching the whole
admin users list alongside and mapping ids to names. The short id stays
as the fallback for a row that comes back without a name.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: shutdown ends the live-event streams so a redeploy drains instead of waiting out its deadline (#5293) ... 79b90bda03

http.Server.Shutdown waits for in-flight requests but never cancels them,
and an SSE stream only ends when its client leaves, so any open tab or
signed-in phone held every redeploy for the full 10s and the old task
exited 1. With the swarm's stop-first update that was ~10s of downtime per
deploy. The bus now closes on shutdown, ending each stream; clients
reconnect to the next instance.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: the shutdown comments say what happens, without the task number ... 4756e957ce

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: opening the notifications inbox marks what it shows as read (#6275) ... f381622d20

The web panel and the Android Notifications screen now mark every unread
notice they show as read the moment they appear, so the badge clears on a
look instead of a dismissal. Rows that were unread keep looking new until
the inbox closes. The read goes through read-all with a cutoff of the newest
notice shown plus 1 ms, so one that arrived unseen stays unread. 'Mark all
read' is gone; opening the inbox does its job.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test: the notifications view model is built inside each test, after the test Main dispatcher is set (#6275) ... eac320c8d7

Built as a property initializer it ran at class construction, before
MainDispatcherExtension's beforeEach, so viewModelScope.launch in init hit
the missing Main dispatcher. Its uncaught exceptions also failed the next
coroutine test in the JVM (RetryTransientIoTest).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: README in the standard order; setup files say what works today (#6277) ... 9f9609591d

README: logo header, short summary, a compose quick start with one-line
comments, first-run steps, a configuration table, setup pitfalls,
updating and backups, then features grouped by what you do. The Specs
section (it called the Android client the "Flutter companion app"), the
dated tag-history sentence and the Fable pointer are gone. Feature lines
were checked against the code: radio seeds from a track only, and the
Android app plays from its cache rather than offering downloads.

.env.example used the SMARTMUSIC_ prefix, which the config loader has not
read for a long time, so copying it set nothing. It now uses MINSTREL_.

config.example.yaml is also the config the image bakes in at
/etc/smartmusic/config.yaml. It listed five recommendation weights
(base_weight, like_boost, recency_weight, skip_penalty, jitter_magnitude)
that moved to Admin > Tuning; yaml.Unmarshal ignored them silently. They
are removed, and the file parses to exactly Default(), so the running
server is unchanged. coverart_from_mbcaa is not listed: nothing reads it.

docker-compose.yml, the Makefile and the spec banners lose their task and
milestone ids.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: server comments say what the code does, not how it came to be (#6277) ... d191a28217

Comment blocks in cmd/minstrel and 30 smaller internal packages, rewritten
to rule 32's standard with each citation read in context: task, milestone
and rule ids go, and so do the stories they were attached to ("used to",
"Before #N", measurements from one install, quotes, pointers to Fabled
Librarian's records and to docs/superpowers files that are not in the
repo). Where a story wrapped a real constraint, the constraint stays as a
present-tense fact.

Comments that had gone stale are corrected:
- taste.Config: the tuned values are overlaid by internal/recsettings;
  it is no longer "a later follow-up".
- StorageConfig: the data dir holds art, caches and the stream secret,
  not only playlist covers.
- the Subsonic album list and getUser comments describe what they do
  today rather than a milestone that would change it.

Checked: the non-comment token stream of every changed file matches
HEAD, and gofmt -l is empty.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: wanted, dbtest, recommendation and playlists comments say what the code does (#6277) ... fcf5b7e5a8

The same rewrite as d191a282, for four packages whose comments carried the
most history: tuning stories, quotes from the operator, single-install
measurements in library_scale.go, and 'used to' accounts of bugs. The
constraint each story protected stays as a present-tense reason (why tier 3
is reduced not removed, why the randomised arms take an order seed, why the
seed artist is eligible in Songs-like). Comments naming system_cron.go and
config.RecommendationConfig, neither of which holds what they claimed, now
name the scheduler and internal/recsettings.

Checked: non-comment tokens identical to HEAD; gofmt clean.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: api comments say what the endpoints do, not how they came to be (#6277) ... 08a2c06b85

Rewrites the internal/api comment blocks to rule 32: task, milestone and
rule ids gone, the stories behind them restated as the constraint they
left, stale references corrected (the stream secret is wired from config,
the timezone handler already refreshes the scheduler, the Android app is
the native client, the design docs cited are not in the repo). Also
rewraps comment lines over 80 columns left by the earlier batches.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: library comments say what the scanner and resolver do, not how they came to be (#6277) ... 7e6228c13f

Rewrites the internal/library comment blocks to rule 32. Task, milestone,
decision and rule ids are gone; the incidents behind them are restated as
the constraint they left (why a file is deleted before its row, why the
missing mark is a mark, why duplicates block on duration). Counts and
album names from one library are dropped from production comments; test
comments keep the names their fixtures use. The tagReadVersion log keeps
each version's before and after, without the measurements.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: query comments say what each query selects and why, not how it came to be (#6277) ... 1d4ef1a23f

Rewrites the internal/db/queries comments to rule 32 and regenerates
dbq, whose doc comments and embedded SQL carry them. The bugs behind the
tiered mixes, the md5 order seed, the score bands and the round-robin
arms are restated as the constraint each one enforces; counts from one
library and stale cross-references are gone.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: migration headers keep the before and after, without ids, quotes or one library's numbers (#6277) ... 6028d78844

Applies rule 32's migration exception: each header still says what the
schema was, what changes and what it prevents, but task, milestone,
decision and rule ids are replaced by the practice they stood for, the
operator quotes by the decision, single-install counts are dropped, and
pointers to a deleted package and an external design doc are gone.
golang-migrate keeps no checksum, so editing applied migrations'
comments is safe.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: web comments say what the app does, not how it came to be (#6277) ... 10d50dccbd

Rewrites the web app's comments to rule 32. Task, milestone, rule and
spec ids are gone; regression stories are restated as the trap they
guard (public auth routes, the menu's stopPropagation, universal loads
under adapter-static, metric deltas with their margin). Flutter and
Tauri mentions now name the Android app or a native shell, operator
quotes and dates are dropped, and one library's counts no longer
appear outside test fixtures. Verified with a TypeScript-scanner token
comparison (template literals and regex literals re-scanned): only
comment text changed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: the history phrases the id sweep missed, restated as present-tense facts or traps (#6277) ... 9f588b6d61

A wider pass for "previous", "before the fix", "we shipped" and the
like. The chi double-subtree, uppercase Vorbis key and stopPropagation
comments now describe the trap they guard; commit hashes, CI run
numbers and a deleted file name are gone; uuidString's comment says
what it returns.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: Android comments say what the app does, not how it came to be (#6277) ... ae7e05c579

Every flagged comment block in the Android app, its tests and its XML/TOML
config, rewritten with the cited issue in view: the constraint is kept, the
story behind it dropped, and issue, milestone, rule, phase and drift ids,
dates and operator quotes removed. "Mirrors the Flutter client's X"
pointers are gone; that client is no longer in the repo.

Comments that no longer matched the code, corrected:
- SyncController: a 410 resets the cursor and re-pulls; it never dropped
  the three local caches.
- PlayerFactory, PlayerModule, AudioCacheIndexEntity, CacheConfig,
  CacheSource: there is no eviction worker and no liked-track bucket;
  SimpleCache evicts by plain LRU over the rolling cap, and the liked cap
  is stored but unenforced.
- CacheSettings, StorageCard: the prefetcher exists (AudioPrefetcher);
  pin-on-like still does not.
- LibraryApi, LibraryRepository, BrowseViewModel: the cache now carries
  the missing-file mark; the browse indexes stay server-backed for one
  source of grouping.
- OutputRoute: UPnP routes exist; only CAST and SONOS are reserved.
- PlayerFactory, RemotePlayerState, OutputPickerController: a failed
  transport command does not drop the UPnP route.
- LibraryScreen: History, Liked and Hidden are real tabs, not placeholders.
- AVTransportClient: the extend loop lives in SonosQueueLoader.
- "the operator" meaning the listener now says "the user".

Verified with a Kotlin lexer comparing every file's non-comment tokens
against HEAD: 394 files, 0 differences.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: workflow, Dockerfile and brand-tool comments say what each does (#6277) ... 557887bb07

release.yml, the Dockerfile and tools/gen-brand-assets.py keep every
constraint their comments explain (the tag map, the gate, immutable tags,
the shared-daemon Postgres guard, the version/channel split) and drop the
history around them: rule and family-idea numbers, issue and run ids,
dates, an operator quote, the "this used to say the opposite" paragraphs
and one-runner timings.

Corrected: the Dockerfile said client/ is filled on tag pushes only; every
build fills it (dev and tag push bundle their own APK, main bundles the
latest release's).

Only full-line `#` comments changed: non-comment lines are identical to
HEAD in all three files, and the three run: scripts whose text changed
differ only in shell comment lines. The release guard tests match raw
workflow text; none of their required or forbidden strings appear in the
new comments.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: setup files and the client, shared and CI docs describe what is (#6277) ... c96231106f

.dockerignore, .gitignore, deploy/initdb, ci-requirements.md,
client/README.md, shared/README.md and two web/scripts test names: dates,
issue/rule/spike ids and the stories behind each exclusion removed.

Corrected:
- client/README.md: the dev lane bundles its own APK; :main no longer
  exists; main bundles the recorded sidecar instead of re-deriving the
  version name. Its smoke-test step wrote a bare "v0.2.0", which the
  server rejects as a sidecar that is not JSON; it now writes the JSON
  shape, which the README documents.
- ci-requirements.md: one workflow, not four; the android lane builds no
  debug APK; main reads the recorded sidecar.
- .gitignore: CI fills client/ on every image build, not tag releases only.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: plan-era "v1", "Task N" and "Phase N" wording restated as facts (#6277) ... 466912e202

The marker list behind the earlier server and web passes did not catch
"for v1", "Task 5", "Phase 2" or "today: Flutter". Those comments now say
what the code does.

Corrected: coverart said "both v1 providers" ship on. There are four
providers (MBCAA, TheAudioDB, Deezer, Last.fm), and Last.fm, which needs a
key, ships off.

Non-comment tokens unchanged against HEAD: 10 Go files, 7 TS/Svelte files.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix(subsonic): getUser answers an admin's lookup with the requested user (#6292) ... 09e78962b2

An admin asking for another username got their own name and roles back.
The handler now looks the user up by exact username and answers with
that user's payload; an unknown username is ErrDataNotFound. A
non-admin asking about someone else is still ErrNotAuthorized.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat(subsonic): getAlbumList2 recent and frequent read the caller's play history (#6293) ... f3b124f9b2

type=recent orders the caller's albums by their latest play (skips
included); type=frequent orders them by completed plays (skips
excluded). Plays of tracks whose file is gone do not count, and both
page by size/offset like the other types.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

test(subsonic): album-list fixture users carry distinct api token hashes (#6293) ... 0d0b04bff0

users.api_token_hash is unique, so the second seeded user needs its own.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix(tracks): removing a track writes a track_remove audit row (#6294) ... 1aba1a41e6

RemoveTrack took the admin's id and dropped it, so a merge left an
audit trail and the more destructive removal did not. It now reads the
track's title, artist, album and file path before the delete and writes
them, with whether the album and artist went too, under the new
track_remove action.

The audit action groups lose their plan-era U1/U2/U3 headings.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

chore(android): delete the unread deprecated FabledSwordTokens alias (#6295) ... b61474744b

UI code reads LocalFabledSwordTheme.current; nothing referenced the
dark-bound alias.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat(android): the audio cache keeps a Liked songs budget and a Recently played budget (#6307) ... 7ec8a82871

SimpleCache ran one LRU evictor at the rolling cap, so liked audio was
dropped like anything else and the offline Liked pool shrank as other
music played. TwoBucketCacheEvictor replaces it:

- a span counts against Liked when its track is liked, else Recently
  played; a liked track that is also played is charged to Liked only;
- each budget has its own cap (0 = unlimited) and drops its own
  least-recently-used audio first;
- CacheBudgetSync feeds it the signed-in user's liked ids and the Storage
  caps, so a like, an unlike or a cap change applies at once;
- nothing is evicted until the liked set is known, so startup never trims
  liked audio as Recently played.

The decisions live in a Media3-free BucketLedger with unit tests.
CacheConfig and PlayerModule, which only snapshotted the caps at launch,
are gone.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat(android): liking a track downloads its audio into the Liked budget on Wi-Fi (#6308) ... ec1e59921b

LikedAudioPinner downloads liked tracks' whole audio when "Keep liked
songs offline" is on: most recently liked first, one at a time, only
on an unmetered network while the server is reachable, and only while
the Liked budget has room for another track at CD bitrate, so it never
pushes out liked audio already on the phone. Tracks liked before this
ships are covered by the same pass. Any change (a like, an unlike, the
switch, the network) restarts the pass and cancels a download in flight.

Supporting changes:
- ConnectivityObserver.unmetered: the default network's NOT_METERED bit.
- Liked track ids are read newest first, and refreshIds stamps the
  server's newest-first order into likedAt.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat(android): Storage shows the Liked songs and Recently played budgets (#6309) ... 5dc6aa9c2f

Two cap selectors, each with its usage, and a "Keep liked songs
offline" switch replace the single cache limit and its next-launch
caveat; caps apply at once.

Docs that said the liked bucket did not exist are restated, and the
audio-cache index's unused per-source eviction queries are deleted:
eviction follows the liked set, not the index's source column.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat(android): any UPnP/DLNA renderer plays, not only a Sonos (#6451) ... b1e55adc14

Discovery already listed every MediaRenderer on the LAN, but playback only
spoke the Sonos native queue (AddURIToQueue, x-rincon-queue, Seek TRACK_NR,
Next/Previous), which a generic renderer faults on, so picking one fell
straight back to the phone.

A renderer that isn't a Sonos is now fed one track at a time by
SingleTrackLoader: SetAVTransportURI + Play for the current track, and
SetNextAVTransportURI for the one after it so the renderer moves on by
itself without a gap. A renderer without SetNext is advanced by the app
when the poll sees it stop (or report NO_MEDIA_PRESENT) at the end of a
track; a stop partway is still the stall watchdog's.

ActiveUpnp carries the queue mode, picked from the route at select. Skip,
previous and queue taps point a one-track renderer at the cursor's track,
the poll takes the queue position from the loader instead of the
renderer's Track=1, queue changes keep the playing track and re-arm the
next, and adopting a running session matches by track id. Sonos keeps its
native queue unchanged.

DIDL-Lite drops the Rincon metadata for non-Sonos renderers and declares
DLNA.ORG_OP=01 so renderers that need it will seek.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix(android): detekt — no four-way destructuring, single-track loader within the function cap (#6451) ... de41169ccb

The library-duration fallback rides on the poll outcome instead of being its
own member, and the adopt path reads the target's fields by name.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

feat: a newer app is told in a snackbar on Home and installed from a sheet holding Settings' card, fetched ahead on Wi-Fi (#6452) ... fa75bd08d7

Mirrors Librarian (#6446, #6448), after Gallery's update tray (canon #6367).

- Home shows "Version X of the app is on your server." / "…is downloaded
  and ready." with Update / Install; the button opens a bottom sheet
  holding the same About card Settings shows, so there is one install
  path. Each build is told at most once a day, kept on disk.
- The shell-wide update banner and its in-memory dismissals are gone.
- AppUpdates (singleton) holds the offer and the install, so Home and
  Settings show the same state. Checks run when either screen opens, on
  every return to the front, every 5 minutes while shown, and on Home's
  pull to refresh.
- UpdateFetchWorker downloads the newer APK ahead on an unmetered network
  (battery and storage not low, three tries); the download writes a .part
  file and keeps it under the build's name only once its size and sha256
  match, so Install is usually one tap.
- Back from allowing installs, the update the listener asked for carries
  on (kept in SavedStateHandle) instead of asking them to tap again.
- An offer needs a higher code AND a different name: the same app built
  again at a later minute is not offered.
- Server: /api/client/version reports the APK's sha256, worked out from
  the served file and cached by size and mtime.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

fix: an app update is offered only when the app changed; casting no longer sticks on "loading" when the same queue plays again (#6455, #6456) ... 3fa5d9fc3a

ci/version.sh --apk names the APK from android/ alone, and the APK lane
calls it. The app offers an update when the server's APK name differs from
its own, so a name derived from the whole repo offered every phone a rebuild
of the app it already had on each server-only deploy. The image keeps the
union set: it bundles the APK under client/.

Casting (#6456):
- setMediaItems with the tracks already queued (play the queued album from
  another track) cleared the cast and waited for a resync that the id-keyed
  collector never ran. It is now a move within the queue: Seek TRACK_NR on a
  Sonos, start that track on a one-track renderer; the play() goes through.
- A queue replaced through setMediaItems could take the Sonos incremental
  diff, which cleared target without restoring active, so the app thought it
  was back on the phone while the speaker played on, and the chosen start
  track was ignored. The diff now runs only while the route is still active;
  a replaced queue reloads in full at its start index.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

docs: two comments state the rule rather than its history (#6277) ... 7d5dcd454e

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

bvandeusen merged commit 7529266f5b into main 2026-10-26 12:32:33 -04:00

bvandeusen referenced this pull request 2026-10-26 13:22:17 -04:00

Main builds and bundles its own signed APK, so a merge ships the app without a release (#6465) #157

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

Clear labels

deps

Dependency update (Renovate)

deps:major

Major dependency bump — held for manual review

No labels

Milestone

No items

No Milestone

Projects

Clear projects

No projects

Assignees

Clear assignees

renovate-bot bvandeusen

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: bvandeusen/minstrel#156