A zero download rate is not a stalled torrent: one between peers reads
zero for a poll or two and finishes fine, and at §9.8's 15s cadence that
flicker would raise the one chip reserved for a download that never
finishes. Transmission already decides this with its own stalled window;
carry isStalled and use it.
Four tests force a filesystem failure by freezing a directory to 0o555.
The CI container runs as root, mode bits do not constrain root, and the
rename those tests expect to fail succeeds — main has been red on
a_failed_rename_leaves_the_row_alone since the move-on-root-change work
landed, with nextest's fail-fast hiding the other three.
The guard probes the filesystem rather than the uid: what the tests
depend on is the refusal, and a container can hold CAP_DAC_OVERRIDE
without being uid 0.
DESIGN.md §15 end to end: provider search and download, embedded-track
extraction, translation behind four feature-gated backends, `alass` sync,
sidecar naming and the `.mt` machine-made segment, per-provider and
per-translator daily budgets, and the subtitle lane in /settings.
One sidecar per language, enforced in the schema (#222), so a forced track
never consumes the language's only slot. The OpenAI-compatible backend's
base URL and model are database rows, not bootstrap config (#220) — that
backend is any endpoint speaking the shape, `llama.cpp` included, and an
endpoint that needs no key is a valid configuration.
41 issues, plus #255 landing the provider reorder control on the shared
icon and control vocabulary.
`just ci` green: 778 tests. Migrations 0024, 0025 and 0027-0029 apply out
of order against a database that already has 0030-0032 — verified against
a database built from main's migration set, which is the shape production
is in.
CI's `rust` job is red on main independently of this merge; see #256.
`visibility: hidden` on the disabled reorder button left a control in the
DOM that rendered as nothing. `disabled` already blocks the click and drops
it out of the tab order, so the hiding was decoration; the shared
`.control:disabled` colour says the same thing on screen, and the row does
not reflow because the hidden button reserved its box anyway.
Exposing it surfaced a second problem: `.control-quiet:hover` sits after
`.control:disabled` at the same specificity, so a disabled quiet control
lit accent-bright under the cursor and invited a click that does nothing.
Guarded with `:not(:disabled)` in the shared rule rather than on this one
control.
`refreshProviderEdges` found the buttons by `button:nth-of-type(1|2)`.
Position in the row is not a contract — a third button would have quietly
moved the edge logic onto the wrong control. They carry `data-move` now.
Closes#255
The subtitle chip needs the settings arm-then-confirm at chip scale: a
chip fits neither the control plate nor the word "confirm". The shared
control grows a compact variant that swaps the glyph instead, rather than
each view keeping a private copy of the idiom.
relocate.rs gathered only media_files rows for a title/root move,
leaving subtitle_files rows pointing at the old folder. Sidecars
already ride along in the folder rename; only their rows were stale.
Tags each rewrite with its owning table (media_files or
subtitle_files) and updates both in the same transaction.
Feedback pass 2 and the size-band work landed on main while this branch
was finishing. Brings them in ahead of the merge back.
# Conflicts:
# crates/arr-api/src/movies.rs
# crates/arr-api/src/state.rs
# crates/arr-daemon/src/main.rs
# web/src/main.ts
17 issues: the operator's UI feedback (back button, icon-only controls,
one-line settings rows, plainer words), the root-move and root-path
relocation work, the attention queue's threshold, window, anchor and
liveness rule, reclassify on every action that changes an effective
policy, and the waived rule reaching the deck.
Gate green at 524 tests. Migrations 0030, 0031 and 0032 verified against
a database built through the production upgrade path; 0032's rebuild
preserves rows and children with foreign_key_check clean.
A pack that hard-failed at import blacklisted its release, put every
episode back to missing and left the season reading 0/10, with nothing
on screen joining the two. Every fact was already recorded.
The blacklist now carries its reason out of the database: deck rows read
`blacklisted · size` instead of a bare `blacklisted`, and say whether the
policy turned the file down — relaxable for this title — or the release
itself failed, which a retry only repeats. A season whose pack was
abandoned says so on its row and above its deck, with the release name,
when it failed, and what it failed on. A row the blacklist no longer
answers for keeps rendering and claims no reason.
Two defects from the integration review of #211 sit in the same code and
are fixed here: a waived row threw away the rule it now carries and read
a bare `below policy`, and the empty-eligible count called every waived
row force-grabbable, since #211 gave those rows the rule `overridable`
reads.
Verified against a real browser: series detail, both season decks and
their buckets, at 1280 and 390 px.
Refs #227, #211
#239 moved §5.7's attention window to `failed_at` and left §6.2's pack
ladder on `grabbed_at`. A torrent that stalls for weeks before ffprobe
condemns it at import has elapsed the whole ladder the moment it fails,
so the pack lane retried a source that had just failed — the one thing
the backoff exists to prevent.
The ladder now measures from the failure, the same anchor and the same
column §5.7 reads, with `grabbed_at` as the fallback for rows written
before the column existed. All three sites read
`max(coalesce(failed_at, grabbed_at))`, so the `last_failed_at` alias
holds what its name says — including the one the season deck feeds into
`reopens_at` and `pack_retry_at`, which was showing a grab time under a
name §5.7 had redefined.
DESIGN.md §6.2 states the anchor the way §5.7 states its own.
Tests cover a pack grabbed 35 days ago and failed 10 minutes ago on the
targeted lane, the RSS lane and the season deck.
`just ci` through the gate: 519/519 tests pass, web checks clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
§9.5 restated the pre-#226 decision-bar rule without the 30-day window,
giving the document two versions of the same rule. Make it defer to
§5.7 instead.
§5.7 used two phrasings for the season-queuing predicate in one
paragraph (existence-of-file vs. state check) though both readers
implement the state check; picked the state phrasing throughout. Also
fixed the arithmetically confusing "seen twice ... third face"
sentence, and rewrapped the single unwrapped ~450-character line in
§7.4 to the document's ~78-column width.
Ref #240. Ran full `just ci` through the gate (517 tests, exit 0).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
`releases` forbade a rule name on anything but a rejection, so §9.3's
deck showed a bare `waived` beside rejections that each named their own,
and §5.7's "watchable but not what was asked" lost the half that says
what was not asked for. Since #210 that is the ordinary outcome of
waiving a size rejection, not a rare one.
0032 rebuilds the table with `CHECK (verdict != 'rejected' OR
rejected_rule IS NOT NULL)`, and the daemon and arr-api's
reclassification both store the waived rule. Existing rows keep NULL and
read as they do today.
`releases` is a parent — `grabs`, `movie_releases`, `episode_releases`
and `season_releases` point at it, three ON DELETE CASCADE — so the
rebuild runs `-- no-transaction` with foreign keys off around one
explicit transaction, per SQLite's own procedure. Verified against a
real database: the pre-0032 binary created and populated it, this build
migrated a copy, and every release row, child row and created_at came
through byte-identical with `PRAGMA foreign_key_check` clean.
Refs #211
Three words the operator reads change; nothing underneath does. Waivers
are still written, recorded and served under their existing names, the
`blocked` column and flag keep theirs, and the release deck keeps its
name in DESIGN.md §9.3 and in the code.
- `waive + grab` reads `force grab`, and its accessible name says which
rule the click relaxes. The `waived` bucket reads `below policy`, and
so does the verdict chip on its rows — an operator can act on "below
policy" and cannot act on the name the record keeps.
- A waived import already said what was relaxed for two rules; `size`
joins them and the fallback names the rule rather than badging it
`waived`.
- `blocked` reads `no targeted search`, everywhere the flag surfaces.
§6.3 gives it one effect and a bare toggle hid it: the accessible name
carries the RSS half at rest, and a note under the controls spells it
out while the flag is on.
- The season and episode controls are already the #230 search icon; only
their labels still said "deck". They now say what the click does.
`just ci` passes through the gate: 509 tests, biome, tsc, tokens.
Verified in a real browser (agent-browser) on the library, movie detail,
series detail, the episode release view, queues and settings, at 1440
and 390 wide — no horizontal scroll at either.
Refs #232
PUT /api/policies/{id} changed the rule every title under every root
pointing at the policy is judged by, and re-derived nothing, so §9.3's
deck and the daemon's grab gate kept reading verdicts computed under
rules that no longer existed.
Drives #241's walker from a policy id: root by root through
reclassify::root, so the skip rules and the leave-unchanged-rows-alone
rule stay in one place. A rename touches no rule and walks nothing.
Inline still holds at this width. Measured on a release build over 2000
titles and 10 000 stored releases across two roots sharing one policy:
0.36 s when no verdict moves, 2.7 s when all 10 000 do. DESIGN.md §5.1
now names four actions and carries those numbers.
just ci passed through the gate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
#243 normalised the incoming path but compared it against the value read
raw from the database, so a root stored with a trailing separator never
compared equal. Every edit of it -- a policy change included -- took the
relocation branch, where each planned destination is its own source and
the pre-check refuses. That root could not be edited at all.
`update` now normalises both sides, and hands `relocate_root` the
normalised stored value. `path_is_free` normalises the stored side in SQL
and `create` goes through it too, so `/mnt/x` and `/mnt/x/` cannot be two
roots for one directory -- the unique index compares raw strings and
cannot see that.
Migration 0031 strips the separator from rows already written. It skips
any row whose stripped form another row would also hold, rather than
tripping the unique index: a migration that cannot apply stops the daemon
booting, which is worse than two roots naming one directory.
Also from the same review: `undo` recorded only the leaf directory, so a
failed move into `/mnt/media-v2/tv/kids` left `tv` behind. It now records
every level `create_dir_all` materialised, deepest first, and still never
touches one that was already on disk.
Refs #244.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Miguel Palhas
naps62-yolo