A mirrored library that answers "0 items" is not believed (its rows stay), but
that verdict stopped the whole reconcile: one library emptied on Plex with its
mirrored items kept made every run incomplete, and un-watches in every other
library never synced again. Each section now settles on its own walk, and only
the settled sections' rows are cleared; the unsettled ones are named in the
warning.
A library left out of the allowlist is never scanned, so its watch states were cleared on
every reconcile; clearing is now limited to the sections walked. A section we mirror items
from that says it holds 0 is not believed complete. The page size is no longer read as the
section total when totalSize is missing (measured: the server always sends totalSize on a
paged request, 0 for an empty one).
E32 review round 1 (B1, B2, B3, B8).
- Plex watch reconcile: a section the server SAYS is empty (totalSize 0, nothing listed) is a
complete listing. It read as "cut short" forever, so one empty library stopped un-watches from
ever syncing. A listing with no total at all stays incomplete (the client now returns None for
it instead of 0), and a whole scan that listed nothing still clears nothing. Tests now cover
completeness across two sections, both orders.
- Plex collections: no members from a collection whose listing counts some (childCount) keeps
the membership we hold, in the full sync and in the targeted resync after a write.
- YouTube playlists: no items for a playlist whose listing counts some (itemCount) keeps the
mirror in the bulk pull, and is refused as retryable in "reset to YouTube" and in the push
plan/merge, where it read as "YouTube deleted them all". The rule sits beside vanished().
- A playlist deleted on YouTube answered "reconnect your account"; it now says it is gone
(errors.youtube.gone, HU+EN).
- The pure 3-way merge moved to sync/playlist_merge.py; playlists.py is back under 800 lines and
off the size list.
The missing identifier became server:/// in the write URI. Client tests over a MockTransport
(errors, paging, cutoff, write URIs, image transcode) and admin-route tests (connection test,
enabled-section sampling, no server URL in a 422, the one-time watch import).
The daily watch reconcile reads "absent from Plex's scan" as "Plex un-watched it" and deletes the
row — but it walked the sections without asking _paginate whether the walk reached the server's
total, so a Plex hiccup mid-walk (an empty page early) un-watched every title it failed to reach
until the next complete run. The scan now reports completeness; an incomplete one still pulls in
what it saw, re-pushes dirty rows, and clears nothing (logged as plex.watch.reconcile_incomplete).
The HU overlay had the same shape: clear every row's HU text, then refill from the listing. It now
reads the listing first and clears only after a complete one. The overlay moved to plex/overlay.py
(plex/sync.py drops under the 800-line limit and off the size ratchet's list).
Tests: plex/paths (23% -> 100%: prefix boundary, longest prefix, dead mount), the reconcile's
clear/keep/re-push rules, the cut-short scan and listing (each seen red with the guard removed).
The 7 baselined Optional accesses in plex/stream.py were one fact the type did not carry: the
initial spawn (replace_only=None) always supersedes, so only the fallback respawn can get None.
An assert at the call site says so; the pyright baseline drops 195 -> 188.
test_abuse_limits: the public-URL and YouTube enqueue tests stopped at the profile lookup
(400 no_profile) and passed on that unrelated refusal. They now send an inline spec and assert
the job is queued. The same shape — a request whose response is discarded — sat at 53 more
places across the suite (all returned what they should, measured with a TestClient probe);
each now asserts its status, and test_response_checked keeps it that way (seen red).
ruff gains RUF100, PLW1510 and flake8-bugbear, every hit fixed rather than
ignored. B008 stays on with FastAPI's Depends/Query declared immutable.
basedpyright moves to standard mode, pytest to --strict-markers, and a hand
run of pytest now requires the database unless SIFTLODE_TEST_REQUIRE_DB=0.
PLW1510 found a real bug: the three ffprobe probes in plex/stream.py never
looked at the exit status, so a probe that failed was read as the file's
truth (an HDR source sent down the HW-safe path, a multi-audio file collapsed
to one track, a subtitle offset taken from garbage) and the warning each
except branch promises never fired. They now run with check=True.
The playlist rows, the search box and the collection picker still
fetched the 400px poster for a 32-80px picture. A contract test pins the
set of bare Plex pictures, and the srcset widths to imgcache.SIZES.
Cards fetched the full cuts: a 400x600 poster (~38 KB) for a ~144px picture and the 1280x720
backdrop (~162 KB) for a ~212px shelf picture. The image proxy now also cuts thumb_sm (240x360)
and art_sm (480x270), cached apart; card pictures (grid, shelves, collection tiles, Up next) carry
both in srcset with sizes=auto, so the browser picks by laid-out width and DPR. Localdev:
17 KB a poster, 56 KB a backdrop. The detail page keeps the full cuts.
- warm_portraits: only 404/410 mark a headshot dead (measured: Plex answers a dead CDN file 404);
a refused token or a 5xx is Plex failing, ends the run after three in a row and marks nothing.
'failed' counts every failure, not only the current streak.
- is_portrait_url: plain path segments only — no dot or empty segment an HTTP client folds onto
one file; person_thumb_url uses the same predicate (27 852 stored headshots all pass).
- Tests: Plex refusing (401/403/429/5xx), failures between successes (portraits and bios), the kept
original served from cache, the audit_clear / demo_reset / discovery_purge rows' counts as data,
and the 0094 backfill on real rows.
- people_enrich: a failed TMDB lookup or a dead headshot is skipped, not the end of the run (a dead
file is marked and left for a week; three failures in a row end it); portraits wait for the Plex
module switch; a no-Plex original keeps its own cache key, so the resize still replaces it; the job
entry lives in the plex domain (run_people_enrich)
- person-image takes only a headshot-shaped CDN url (no query variants of one file)
- every synced row gets Hungarian genres, and a HU genre filter matches the word as picked too;
migration 0094 fills the rows no overlay reached
- a job run's audit row carries its result as data; audit_clear / demo_reset store their counts
- one credit-column list for the prune and the enrichment order; _overlay_pass typed
- tests: batch skip + stop, HU bio stored, dead headshots, module switch, the person-image route,
the job body, genre filter + facet write + twin fallback, tagline/summary fallbacks, audit data,
the reader's alphabet at every remaining collated() call site
A person's TMDB bio was fetched only when their page was first opened, off the request path, so the first
open showed no bio and the page never asked again (2 of 18,202 linked people were ever enriched).
- people_enrich scheduler job (every 10 min, 300 per run): linked people's bios from TMDB and their
portraits into the image cache, the most-credited people first. A failed TMDB lookup ends the batch.
- Portraits are resized through Plex's photo transcoder to a 320 box (~15 KB instead of ~170 KB; the
person-image proxy uses the same cache key, and a full-size original is removed as its small twin lands).
- The person response says `enriching` while a bio is on its way (a person the job hasn't reached yet).
- "Known for" is ranked by vote count, without guest visits as themselves (talk shows had crowded out the
films: Tom Cruise had 14 of 15), with directed works, from a 40-deep pool; the page shows up to 15.
Migration 0093 un-stamps enriched people so the job refetches them.
- A phased job result (backfill, people_enrich) now counts as a change for the audit trail; before, an
automatic backfill run that added videos was never audited.
- HU: the watch reconcile job reads "egyeztetés" (a leftover from the S5 sweep).
The /person fetch appends translations and keeps TMDB's Hungarian biography beside the instance-language
one (new plex_people.biography_hu, migration 0092). A Hungarian reader is served it, the English one when
TMDB has none. Already-enriched people are un-stamped, so their next open refetches them once.
- lib/vocab + the `vocab` locale area: TMDB crew jobs and departments, auto-tags (D1 A: the data
stays English, the catalog is named at its one source, useTags), built-in download profiles,
YouTube topics; held key-for-key to the backend's own lists by backendIds.contract.
- Countries by ISO code: TMDB parse stores {code, name}; rows stored before are coded from TMDB's
own country list in one request (run_enrich); Intl names them, historic codes (YU, SU, XI…)
from the catalog.
- Player tracks named from the language tag, forced/SDH flags and codec tail, not Plex's
server-language displayTitle; no invented "Audio N" on the server.
- Download lifecycle notices (shared / expiring / expired / evicted) named from data; the share
notice carries the title.
- Scheduler: the job result rides along as data and is named per key (dashboard + run notice).
- Invite status named; a render crash's notice no longer shows the engine's English message.
- Fix: YouTube links …/wiki/Humour, the tagger looked for "Humor" — Comedy never came from a topic.
Found in the S1b UAT: 'Mert megnézted: Batman Forever' kept the English title, and
the info page's TMDB recommendations kept TMDB's English titles even for titles we
hold. A held recommendation is now named as the library names it, in the reader's
language; one we lack keeps TMDB's title.
The grid served Hungarian titles, but every other card surface (Discover rails, the
show page's episodes, search, collections, the person page, playlists, Home's resume
rail, hero and comfort shelf) served English. The card serializers now take a
required keyword-only lang (a forgotten call site fails instead of going English), and
each of those reads takes ?lang=hu, with the language in its query key. A Hungarian
search finds an episode by its Hungarian title too.
An episode on a poster rail showed its 16:9 still cropped into a poster, titled only
with the episode name; it now shows the show's poster and name, with S·E · title under
it (option A). Plex's generic season names (Season 1, Specials) are rebuilt from the
number in the reader's language; a season with a name of its own keeps it.
Home's comfort-shelf move wrote its optimistic order to the key without the language,
which nothing read any more; it writes the query's own key.
The HU TV section lists every episode with its Hungarian title and summary, but the
overlay read only movies and shows, so every episode stayed English. It now overlays
the episodes too, and the HU tagline (new tagline_hu column, served on ?lang=hu).
Plex's HU agent returns a mixed genre vocabulary (Action beside Dráma), and a title
with no HU twin kept its English list, so the HU genre list showed Drama and Dráma.
plex/genres.py maps every genre word to one Hungarian word; the overlay writes it
into genres_hu for every row, and a HU genre filter is read through the same map,
so a saved filter still holding Action keeps matching.
One invalid byte sent a whole UTF-8 sidecar to a legacy code page, an untagged
file fell to Windows-1252 (ő/ű became õ/û) and Serbian was read as Cyrillic.
The cache key moves to sub3_ so a track the old rules mis-decoded is read again.
A Hungarian .srt beside the video (Windows-1250) was decoded as UTF-8 with
errors=replace: one episode served 1391 U+FFFD and no accented letter.
decode_subtitle now lets a BOM decide, keeps valid UTF-8, and otherwise
reads the legacy code page of the stream's language (cp1250 for Central
European, cp1251 for Cyrillic, else cp1252) — on the ffmpeg path too.
The cache has no eviction, so its key is versioned (sub2_): a track cached
broken is read again once instead of being served for good.
HLS sessions were keyed by the item, so a second account opening the same
film killed the first one's session; the displaced player reclaimed at its
playhead and evicted the other in turn. The registry is now keyed by
(viewer, rating_key), segment requests resolve by both, and the session
directory carries the viewer (the orphan sweep still recognises the older
shapes).
Only transcodes counted toward a cap, so remuxes — each a whole segmented
copy of the film on the scratch volume — were bounded by the library size.
Three bounds now apply after every start, each dropping the least recently
used: two sessions per account, the transcode cap, and a new admin knob,
plex_max_sessions (default 8), over every session. A superseded session
request returns at once, since it waits on the killed ffmpeg, so an account
cannot park more request workers than it may hold sessions.
The subtitle extractor's full-pass slots were global and minutes long each;
one account opening subtitles on film after film could hold them all. An
account may now hold three of the six, and still gets windows past that.
A full pass that ran out of time kept what it reached, but its retry an hour
later started again at 0, so a file too big to read inside the extraction
timeout never completed. The pass now starts where every embedded track's
coverage from the head ends (less the window pre-roll, so a cue across the
join is read whole). The seek needs -avoid_negative_ts on this read as on a
window, measured on a real file: without it the cue running at the start
shifts the whole output.
A partial track whose full pass had failed never reached a settled state, so
the player polled it every 2 s for the rest of the film. The status now says
whether the track is still growing, and a stalled one is polled every 15 s —
still inside the 60 s lookahead that reads the next window before an edge.
The comfort shelf resolved each pinned channel with three queries (the
row, its unwatched count, its newest video) and each playlist with two.
channels_summary gets the count and the newest row per channel as window
functions over one filtered set, playlists_summary reads every pinned list
in one query, and the rows come in one IN() each — so the shelf's cost no
longer grows with its length. A test counts the statements.
RAIL_LIMIT had two definitions held in step by a comment; the Plex Discover
route now imports the domain's. And rollup's "continue the latest-started
episode" was false: it takes the last in-progress one in broadcast order,
which is also what Home's next_episode picks.
The feed's channel-name match and the live search's term fold built their
LIKE pattern from the typed text as it was, so `_` and `%` acted as
wildcards: `build_log` also matched `buildXlog`. The term fold also tested
a substring, so a term already present inside a longer word (`cat` in
`category`) was never folded into search_terms. It now compares whole
terms, escaped.
`like_escape` moves from the Plex query module to `app.utils`, beside
`to_tsquery_str`, and every pattern built from typed text goes through it.
- yt-dlp's text is scrubbed wherever it is logged, not only in the adapter:
`formats.log_safe` takes out the URL, the `[extractor] <id>:` prefix and
file paths, and the job-failed, refused, edit-failed and backfill lines use
it. Done lines log the asset id, not the title-built path. The PII contract
now bans `rel`/`rel_path`/`title`.
- A traceback goes with a download/edit failure only when it is our own code
failing; for yt-dlp's DownloadError, our RuntimeErrors and OSErrors its foot
is the unscrubbed message.
- yt-dlp's `error` maps to WARNING: it fires on attempts the fallback client
set recovers from; the job's verdict is `download.job.failed`.
- The sink groups a bare `%s` pass-through by its text (numbers folded), and
a process's last batch closes the repeat window, so a short-lived download
child still reports what it dropped.
- Worker-side failures (stall, wall-clock, no asset, cached asset error) log
`download.job.failed` with a cause; they left no line at all before. The
crash, supervisor-error and non-zero-exit lines carry the code too.
- An edit failure is one coded line in `_fail_edit`, for every path.
- `logs.get_logger("__main__")` resolves to the `-m` module, so the worker
and the download child keep their names.
- TRACE lines carrying user content are gated on `logs.consented_trace()`,
not on the process level; a contract test pins it.
- The OAuth token refresh goes through `_upstream_failure`; transport errors
carry reason `transport`. Subscribe recognises a duplicate by reason code
only (the prose fallback was dead).
- Once-only guards for a stale sysconfig override and a refusing HLS session
directory; item_detail splits Plex failures from our own parse bugs.
- The playlist breaker line names the playlist and the action; comments that
were false (runner noqa, `_merge_vtt`, sink grouping) are corrected.
The gap this sprint was scoped against turned out to be smaller than the survey
said, and in a way worth writing down: `log_requests` already records every HTTP
call with its route template, status, duration, request id and user id, and the
audit log already records every admin act on someone else's account. Putting a
line in each of routes/downloads.py's twenty mutating endpoints would have added
a second copy of the first and nothing else. So this instruments what neither of
them can see.
yt-dlp now explains itself to our log. `quiet: True` was doing more than quieting:
with no `logger` its whole account of a download went to the child's stdout, so
"Requested format is not available", a signature-extraction failure or a POT
refusal reached us only as whatever text ended up in the raised exception — the
real reason a download failed was routinely one line the app never kept. Measured
against the pinned yt-dlp 2026.07.04 before wiring it: with `logger` set it calls
`debug` for extraction and progress lines and `warning`/`error` for the rest, and
neither `quiet` nor `no_warnings` gates those calls. `debug` lands at TRACE, being
several lines a second, so the noisy half is the half the user's own consent
gates. The full source URL is trimmed to its host on the way in — it is a row of
someone's download history, and `logs.set_job` already puts the job id on every
one of these lines. The option is set AFTER the adapter merge, like `cookiefile`
and the size cap: a source adapter must not be able to take away the one account
of what went wrong that a failed job leaves behind.
An HLS session now says which pipeline it got (transcode/remux, hw, tonemap,
variants, offset) — everything downstream is about that choice, and it was only
inferable from an ffmpeg command line nobody keeps. A session culled by the
concurrent-transcode cap says so: somebody's playback stops there, the route turns
it into a retryable 404, and "my film keeps rebuffering" had no other explanation
anywhere. The HW→SW fallback keeps its line and gains an event code.
A live search refused by a quota cap says WHICH cap: the 429 is in the request
line either way, but the per-user daily cap is answered by tomorrow and the shared
budget by the admin.
The download and edit job lifecycle lines gain event codes, which is what E22.6's
console filters on — the job id was already a structured field.
tests/test_ytdlp_logging.py; three of its four cases mutation-checked. A fourth
(a `%` in a progress line) survived its mutation — stdlib only applies `%`
formatting to a record that has args — so it is gone rather than kept as a case
nothing can break; the reasoning moved into the comment on `_emit`.
114 exception handlers neither logged nor re-raised. The defect was never that
they are silent — plenty of them should be — it is that their silence said
nothing, so nobody could tell a decision from an oversight. Each one has now been
read. 44 got a line; the other 70 carry `# silent-ok: <why>`.
What the silent ones had been hiding, in the cases that got a line:
- A failed ffprobe sent every HDR file down the 8-bit pipeline, so playback
degraded to a software transcode with nothing saying why; a failed audio probe
collapsed a multi-audio file to one track and the player showed no language
picker; a failed first-PTS probe left subtitles seconds out of sync.
- A subtitle sidecar that existed but would not parse made a PARTIAL track read as
complete, permanently — the viewer loses every cue past the cached window.
- An unreadable crop/trim/cut-list in an edit spec was dropped, so the clip came
out uncropped, untrimmed, or shorter than the one the user built.
- A dropped `max_height` downloaded at best quality against an explicit cap.
- A failed videos.list enrich made every candidate undroppable-from-judgement and
therefore dropped: a working search rendered as "no results".
- An unusable password hash answered as "wrong password", locking the owner out
and telling them they mistyped; a rotated TOKEN_ENCRYPTION_KEY turned every
stored token into "not connected".
- A stored admin setting that would not coerce fell back to the env default while
the settings page went on showing the saved value.
- A one-off InnerTube config failure pinned the whole process to the hard-coded
fallback key, and searches return nothing when that goes stale.
- The HLS sweep reported 0 swept — like a clean run — when it could not list its
own root, leaving every crashed session's segments on disk forever.
Four of them are counted rather than logged per item (dropped cues, dropped cut
list segments, unreclaimed GC files) and reported once per call, because the
alternative is one line per cue over a whole subtitle track.
ruff gains S110 and BLE001. Every blind `except Exception` that remains carries
`# noqa: BLE001 — <why>`, which is the point: catching everything is sometimes
right, and always worth a sentence. plex/sync.py's three best-effort backfill
catches now keep the traceback — a blind catch over a batch that fetches AND
writes could be Plex, the database or our own bug, and `%s` of the exception told
them apart badly. main.py's boot-time log-level read likewise.
logs.event's `code` is positional-only now: `event("x", code=...)` was a TypeError,
and `code` is exactly the word a caller reaches for when the thing that failed has
one of its own.
tests/test_swallowed_exceptions.py is the wider rule ruff cannot see: a handler
that neither logs nor re-raises must say why. Both cases mutation-checked — a
marker removed and a log line removed each fail it, naming the handler.
25 literal names covered 34 modules. plex/client.py, plex/stream.py,
plex/subtitles.py, plex/sync.py and two Plex route modules all called themselves
`siftlode.plex`; downloads/pipeline.py called itself `siftlode.worker`;
routes/channels.py called itself `siftlode.api`. The `logger` column is in every
stored row so E22.6's console can filter on it, and so an operator can pin a level
to one part of the system — neither works when six modules share a name.
`logs.get_logger(__name__)` derives it: app/plex/sync.py -> siftlode.plex.sync.
One name per module, and it cannot drift from the file it sits in.
tests/test_logger_names.py pins the derivation, that everything stays under the
one `siftlode` tree the handler and the tests attach to, and that no module
hand-rolls a name again unless the line says why (`# logger-name-ok: <why>`).
Two do: get_logger's own call, and `siftlode.http`, which is named for the traffic
it carries rather than the module it lives in — it replaced uvicorn's access log
and is what an operator pins when they want request lines and nothing else.
app/scheduler.py called its logger `logger`; every other module calls it `log`.
ruff gains LOG, G, T20, TRY400, TRY401 and S112. Four findings cleared with them:
worker.py logged `str(exc)[:120]` where the traceback was the whole question, and
three prints/f-strings that are deliberate now say so on the line. S110 and BLE001
stay off until S3 — 23 handlers are outstanding and each is a judgement about one
of them; enabling the rules first would just mean 23 blanket noqas.
log_level, search_source, download_layout and the transcode hw/preset
were typed by hand, and every consumer carried its own guess at what a
typo meant. The value set is now a Literal in config.py: pydantic
rejects a bad env value at boot (any letter case accepted), sysconfig
refuses a bad override and derives the admin picker's options from the
same Literal, and the consumers read the value as-is.
The Configuration page renders a choice key as a segmented bar, or a
dropdown past six options, with translated option names. Its tabs now
follow the registry's order instead of a hand-kept copy that had
already dropped the explore group.
A single-choice FilterSelect now opens on the option in use: in the
browser it could land on the first row instead.
Round 6 turned `artifact_may_arrive` into a positive rule, but keyed it to the HLS vocabulary in
general rather than to what THIS session produces: the playlist regex accepted master/index/stream_*
for every session and the `seq == 0` branch accepted any variant token unconditionally. The two argv
branches are disjoint — a single-audio session is given `seg_%d.ts` + `index.m3u8` and no `%v` at
all, so `master.m3u8`, `stream_*.m3u8` and every `seg_<variant>_<n>.ts` are names it can never write,
and vice versa — so each of those was read as "an HLS name, so wait and see" and held a threadpool
worker, plus its pooled DB connection, for the full timeout. That is verbatim the round-5 outage the
route's own comment claims it prevents.
Measured through the real route against a live single-audio session, before: master.m3u8 404 after
30.0s, stream_q7z.m3u8 404 after 30.0s, seg_a0_0.ts / seg_zzzz_0.ts / seg_0_0.ts 404 after 20.1s
each. After, on the running localdev API: 404 in 44-316ms for all of them, while index.m3u8,
seg_0.ts and seg_1.ts still answer 200 — nothing is over-refused.
The session now carries the rendition names its ffmpeg was told to write, and `entry` is derived from
them instead of passed. And "one place" now means the argv BUILDS its patterns from the same
functions the parser reads back (`segment_name("%d", "%v")`, `variant_playlist_name("%v")`): the
layout was still spelled out independently in six places, with `_VIDEO_VARIANT` read by no consumer,
which is the drift that caused round 6's HIGH. A test asserts the argv's own patterns are what those
functions produce, and that every name they expand to is one the guard accepts.
The load-bearing MEASURED comment cited ffmpeg 8.1.1 — the Windows host's. The binary that writes
these files is the API container's 7.1.5; re-measured there, the conclusion holds.
Round 5 added a guard so a request for a segment beyond the produced region
answers instead of polling out the full timeout — the route is a SYNC one, so
every such wait parks a threadpool worker and, through `current_user`, a pooled
DB connection. Measured against real ffmpeg, the guard could not fire where it
mattered.
`-hls_segment_filename seg_%v_%d.ts` substitutes the VARIANT first: a dual-audio
session writes seg_0_0.ts (video) and seg_a0_0.ts / seg_a1_0.ts (the `name:aN`
renditions). The regex read those as sequence-first, so every video segment
parsed as segment 0 and every audio segment did not parse at all — the guard was
inert on every multi-audio item, which is most of this library. `_seg()` carried
the same inversion: the warm-up waited 8 s for `seg_1_0.ts`, a name ffmpeg never
writes, on every multi-audio session start AND every seek-restart.
The layout is now one constant plus one parser (`segment_name` /
`_SEG_NAME_RE`), with a test pinning them to each other, instead of three
comments that could drift — and they had drifted, all three ways.
The guard is also a positive rule now. It was a blocklist, so anything that was
not a single-audio segment name was waved through: `GET …/anything.m3u8` still
parked a worker for 30 s, while the route's comment claimed a flood of hopeless
names was what it prevented. A session produces its playlists and its segments
and nothing else. `.vtt` leaves `_HLS_CT` with it — ffmpeg runs with `-sn` and
subtitles are a separate route, so that extension could only ever name a file
that will not arrive.
Also here:
- A 5000-digit segment index reached the client as a 500: `int()` on an
unbounded digit run raises past 4300 digits and only OSError was caught. The
name is bounded on the route and the index in the regex — the same bug class
the subtitle routes were bounded for.
- A session torn down between the wait and the read is a 404, not a 500. `_kill`
rmtrees the directory and no lock spans those two steps; the player's recovery
is keyed on the STATUS, so a 500 skipped the reclaim and ended playback with
"couldn't play" instead of restarting at the playhead. The trigger is ordinary
— `loadSession` awaits the new session before destroying the old hls instance.
- Two comments claimed an open player checkpoints every 10 s "whether or not it
is playing", which is the stated reason the 30 s finished-idle reap is safe.
`flush` sends nothing once the finish latch is up, or before a position is
known. The states it does cover — playing or paused mid-item — are the ones
the rule needs; the comments now say that instead of more.
- The route's tests leave the DB lane: the route reads nothing from the database
since the auto-start was removed, and there those four regressions SKIP on a
machine with no Postgres while pytest prints green.
14 tests, mutation-checked: the old parse, the old warm-up name, the blocklist,
both bounds off the absurd name, the unguarded read, and the route dropping
`proc` — each red on the test that names it.
Both found by UAT in a real browser, after the tests for them were already green — which is the
whole reason this repo settles focus and timing claims by running them.
- **The autoFocus fix worked, and then the dialog took the focus back.** React 18's StrictMode runs
every effect mount→cleanup→mount in development: the cleanup handed focus back to the trigger, and
the second mount, finding the focus outside the dialog, moved it to the first focusable — off the
`autoFocus` field a beat after the browser had put it there. Chrome measured it: INPUT (commitMount)
→ BODY (cleanup) → ✕ (re-mount). The effect is idempotent now — it remembers what the dialog was
focused on and restores THAT on a re-mount — and the close-time restore is conditional on the
dialog still holding the focus, which it tracks as it changes rather than computing at cleanup,
where the panel is already out of the document. A dialog no longer yanks focus back from wherever
the user has since moved it.
- **Passing ffmpeg's handle to the segment wait only shortened the FINISHED case.** Measured against
the running localdev API: a request for `seg_99999.ts` on a live session still took 20 195 ms,
because that ffmpeg has not exited and the wait has no other reason to stop. ffmpeg writes segments
in order, so a segment whose predecessor does not exist is not the next one and cannot arrive —
and that is knowable at once. Same request now: 45 ms.
Verified on localdev against the real library (Westworld S01E06, remux, multi-audio): Mod+K opens the
command palette with the focus in its search field and typing reaches it; playback runs on the new
session-scoped HLS URL; a request naming a stale session id is a 404; and a viewer paused for longer
than the finished-session rule keeps their 2.4 GB session across two reaper passes — the progress
checkpoint touches it — and resumes cleanly.
vitest 1009, pytest 1127.
Deep-review round 5, over round 4's own fixes.
- Sessions are keyed by rating_key alone while a session's DIRECTORY is per (key, offset, tag), and
the artifact URL carried no session identity — so once a seek-restart or a second viewer of the
same item took the registry slot, the old viewer's request for `seg_12.ts` was answered 200 with
the new session's twelfth segment: bytes from a different `-ss` offset under `-copyts`, on a
timeline the client is not on. Wrong content with a success status is the one failure a player
cannot notice, and round 4's reclaim-on-404 could never fire on it. The id is in the path now (so
the playlist's relative names resolve back into the same session), a mismatch is a 404, and the
player reclaims at the playhead. The entry playlist's auto-start is gone with it: it could only
ever begin at 0, silently discarding the viewer's position.
- Round 4 exempted paused viewers from the reaper by keeping EVERY finished session for the full
ten-minute idle window — one whole segmented copy of a film per item, on the volume the download
center shares. The signal it needed already existed: an open player checkpoints its position every
10 s whether or not it is playing, so that route now touches the session and the periodic pass is
back on the short rule.
- POST /session's opportunistic reap was never scoped to the key being started: it walked the whole
registry with the 30 s finished rule, so pressing play on ANY item — even a direct-play one, the
call ran before that early return — killed another viewer's paused session. It is gone; the
periodic pass owns reaping.
- A segment that can never arrive no longer polls out the full 20 s: the producer's handle is passed
to `wait_for`, so a finished ffmpeg answers at once. This is a sync route, so each such wait parked
a threadpool worker AND a pooled DB connection, and an ordinary seek past the produced region was
enough to do it.
- The DataError handler's re-raise branch — the half that keeps a genuine DataError a 500 — is
pinned, so simplifying it to an unconditional 422 cannot pass in silence.
Six regression tests (two route-level, in the DB lane), each mutation-checked red on the assertion
it is named after. pytest 1124, vitest 999.
Deep-review round 4, backend half.
- The periodic HLS reaper (round 3) applied the "finished and unfetched for 30 s" rule that only a
NEW session request may apply. A remux finishes long before the film, and a paused player with a
full buffer fetches nothing, so a one-minute pause killed the session and playback died ~30 s after
resuming. The periodic pass is `reap_abandoned` now (the long idle limit, finished or not).
- A subtitle `publish` that could not keep its output (full / read-only cache volume) returned
quietly, so every poll queued the window and the whole-container pass again. It reports failure
and the pass is backed off like a failed read; a cached VTT that cannot be READ is answered
`unavailable` (`cache_unreadable`) instead of scheduling reads that can never land.
- The subtitle routes bound `rating_key` / `sub_ord` (they become a cache file name: a 250-digit
one was ENAMETOOLONG, which pathlib's exists() re-raises -> a 500), and the cache helpers no longer
throw on such a name.
- One app-wide handler turns psycopg's "text fields cannot contain NUL" DataError into a 422: round 3
guarded only the by-rating_key lookup, and `contains=`, `contains_group=`, reorder/bulk key lists
and the collection search still 500'd.
- The hand-written `db.get(...) is None -> 404 "Unknown ..."` siblings use `get_or_404` with a
`what` (channel / video / user), so the Hungarian UI no longer shows English for a gone row.
Deep-review round 3 of bug hunt 2026-09.
- `failed` is now final only for a track that can never be text (bitmap, no item/track/file). A Plex
hiccup or a stretch ffmpeg could not read answers `unavailable`: said, polled again every 5 s, and
the <track> stays mounted when cues exist elsewhere (gen > 0). Re-picking the selected track asks
the server again at once.
- Within LOOKAHEAD_S of a covered stretch's end the next window is read ahead, instead of only once
the playhead has fallen off it (subtitles dropped out at every edge while the full pass was backed
off or slower than a window lasts).
- Windows and full passes run on separate pools with separate in-flight caps, so a window never
queues behind (or is refused by) minute-long full passes — the pool comment promised that and
nothing enforced it.
- A pass that times out is TERMinated and keeps what it reached (published as partial) instead of
being killed and discarded.
- Lapsed failure memos are swept on every write (they were only pruned when the same key was read
again, so polling fresh keys grew the process without bound).
- `publish` no longer merges into a sidecar with no VTT behind it, or claims a region whose VTT write
failed, or overwrites an unreadable track with one window.
- A NUL in a rating_key names no row (it was a DataError, a 500, on every Plex route) and no longer
500s the status poll.
Tests: 16 new, each mutation-checked (18 mutations, all red on the test that names them).
Deep-review round 2 of the 2026-09 bug hunt, over the round-1 fix diff.
- A window was LATE by the distance to the keyframe its input seek landed on, and then doubled for
good once the full pass merged in: a cue already running at the window's start came out at a
negative time, and the webvtt muxer's default shifted the whole output so it sat at 0. Measured on
the real Westworld S01E06 at 25:00: every window cue +5.005 s, 682 cues merged instead of 608.
`-avoid_negative_ts disabled` keeps the times exact (the parser reads ffmpeg's signed-field
spelling of a negative time); now 608/608 on that file, and a real-ffmpeg test pins it.
- A full pass that failed was re-armed every minute while the viewer watched what it did publish —
a whole-container read per minute. It now backs off for an hour; windows keep running meanwhile.
- A failed WINDOW set the track's failure memo, so any uncovered position read "failed" — a final
answer that ends the player's poll, even with the full pass about to deliver, and one any account
could plant with an absurd `at`. It now marks only the span it failed to read; "failed" is given
once the full pass AND a window there have both failed.
- Windows join by what they COVER, not a bucketed identity: a backward seek no longer waits behind a
window that cannot cover it, and a playhead crossing a bucket edge no longer queues a second read.
The queue is bounded per track and overall.
- shutdown() sets a flag a worker checks before it spawns ffmpeg (one still in its Plex lookup used to
start a read after the kill sweep); queued-job cancellation and the lifespan call are now tested.
- A window starts where ffmpeg is told to seek (rounded to the ms it gets), a rounded-up millisecond
carries into the minute (no more `00:00:60.000`), a key's first publish claims nothing before its
VTT lands, and `at`/`offset`/`start`/`aoff` refuse non-finite values (`at` also an absurd one).
Deep-review R2, plus one defect found while verifying it on localdev.
- The status poll re-armed every second after a FAILED request: `data` stays empty on an error,
so it read as "still preparing" for as long as the player was open — and the request was not
quiet, so a persistent 500 re-opened the global error dialog every second. It now backs off to
10 s (a blip still heals on its own) and is quiet; the player already shows the failure inline.
- A <track> whose own VTT fetch failed never fetched again: with the server's answer unchanged its
key was unchanged, so React kept the dead element — silently blank subtitles. The key (now built
in the hook, where it is tested) carries a failed-load count; after two remounts the track is
reported as unavailable instead.
- Found in UAT: until the first timeupdate `abs` is 0, so after a resume every poll said at=0 and
the first window was cut around the head of the film (a resume at 20:00 polled at=0 for as long
as it stayed paused). The poll reports the session's start until playback ticks. And the window's
dedup identity is the track plus a window-length bucket of its start — keyed on the track alone
(R1), the right window queued behind that wrong one.
The poll's start/stop rules finally have tests (every earlier one resolved a single fixed answer).
Verified on localdev against Westworld S1E6 resumed at 20:00: the polls carried at=1197, one
subtitle line on screen at 21:02, the poll stopped at `ready`; the real file's window (56 cues) +
full pass (608) merged to 608 — the old merge would have served 664. 11 mutations, all red.
Deep-review of the bug-hunt round (R1). The background extractor from ed4ab4e1 held, but its seams
did not:
- Window + full pass doubled every cue in the first hour. merge_vtt de-duplicated on the raw timing
text, and the two passes spell a timestamp differently (shift_vtt writes HH:MM:SS.mmm, ffmpeg's
webvtt muxer drops the hour below 1 h; shift_vtt(-0.0) never normalised the full pass). Timing is
now parsed and re-emitted canonically, keyed with the cue text so two cues sharing a time survive.
- A full pass that exited non-zero was published as complete: "ready" forever, subtitles ending
where ffmpeg died, and no eviction to ever retry. It now covers only up to its last cue, is never
complete, and records the failure memo.
- The window's dedup identity held the live playhead, so every 1 s poll queued another read of the
file. It is keyed on the track now, and a queued job re-checks coverage before reading anything.
- A sidecar track waited behind an unrelated embedded full pass (the full job is per item). Its
window job now fetches it whole — one small request.
- The <track> fetch ran the scheduler at position 0: an unrequested window over the head of the file,
and a refusal of a window that was on disk. It is a pure read of the cache now.
- publish() is serialised (a lost update became permanent once complete was set); the pool is shut
down and running ffmpeg children killed at app shutdown (they held interpreter exit for minutes);
?at=inf is a 422, not a 500; and both subtitle routes use current_user instead of the cookie-only
poster gate, which let a suspended or epoch-revoked session keep scheduling reads (pre-existing).
_extract_and_publish finally has tests of its own — every earlier test stubbed it out. 19 mutations,
each red on the assertion that names it.
Two halves of the same report ("Collectionben nem működik a rendezés?"). The pasted browse URL
sorts correctly — that part was working. What did not:
The Collection INFO page had no sort control at all and a hardcoded release-year-ascending order.
It now shows the browse grid's Sort control, with the browse grid's vocabulary, because a second
sort vocabulary on a neighbouring surface is exactly the kind of drift this repo keeps paying for:
`PLEX_SORT_KEYS` is now one tuple in `plex/query.py`, and a test pins BOTH implementations to it —
browse's key→column map (extracted from the handler to be reachable) and the collection page's
key→member map. A key added to the vocabulary without teaching one of them fails the gate instead
of putting a dead option in the menu.
Ordering a collection cannot be a single ORDER BY: its two legs are different tables, so the merged
member list is sorted in Python. Members with no value for the key sink to the END in both
directions (a movie with no rating is not "the worst rated" — the browse grid's `nullslast()`), and
ties always resolve title-ascending, so flipping the direction reverses the ranking rather than the
alphabet inside it. `lastwatched` needed the show leg's `resumed_at`, so the page now takes ONE
rollup pass that yields both the badge status and the recency instead of the status alone — without
it that menu entry would have ordered every show as never-watched. The default stays release-year
ascending: a franchise reads in release order, and that is what the page did before.
The other half: a shared Plex link carries the collection filter as bare rating_keys, so a link
opened cold had no name for them — the sidebar token rendered `17045` and the chip could only say
"Collection". `GET /api/plex/collections` takes a `keys=` CSV now, and a display-only hook resolves
just the missing ids. It deliberately does NOT write them into `collectionTitles`: that map is
hashed into the Plex query keys, so storing a resolved label there would split the cache from the
identical state and refetch the grid for a piece of text.
Verified in the browser on localdev against the real Plex library: the movie collection's menu
offers Length and the show collection's does not (the same scope rule as browse); rating-ascending
reorders 8.1/7.4/7.2/6.6/6.5 correctly; `lastwatched` on a show collection brings the in-progress
shows up; the choice survives a reload (it is a server pref, on its own key so opening a collection
never re-sorts the library behind it); and the reported URL now reads "Pirates of the Caribbean
Collection" in both the sidebar and the chip.
12 pytest + 8 vitest, mutation-checked against 14 mutations (direction ignored, nulls first, no
stable tiebreak, unknown key untouched, wrong default, show recency dropped, empty `keys=` falling
through to everything, a key missing from either sort map, every id refetched, ids unsorted,
fetched titles unmerged, the chip ignoring them, the hook's result never reaching the builder) —
each red on exactly the test that names it.
Bug hunt 2026-09, item 14 — "how does the multichannel downmix work, and is it
normalised?" It isn't: `plex/stream.py` has always encoded `-c:a aac -ac 2 -b:a
192k`, i.e. fixed stereo with ffmpeg's default downmix matrix and no loudness
handling at all, which is why a film mastered for a cinema arrives with whispered
dialogue and deafening action.
New admin knob `plex_audio_loudnorm` (Configuration → Plex), DEFAULT OFF: it makes
a film markedly louder and narrows its dynamic range, which is a taste call rather
than a fix. It sits at the end of the Plex group, after the software-transcode
knobs — it is an audio concern, not one of the libx264 ones.
-af aformat=channel_layouts=stereo,loudnorm=I=-16:TP=-1.5:LRA=11,aresample=48000
Three things about that chain are load-bearing and each has a test:
* the downmix is FORCED FIRST rather than left to `-ac 2` — measuring 5.1 and
folding it down afterwards would normalise something other than what is heard;
* the trailing resample is not decoration: `loudnorm` runs its own resampler at
192 kHz internally, which the AAC encoder will not take;
* one `-af`, no stream index, so EVERY audio rendition of a multi-audio file is
normalised — not just the first.
The knob is also part of the session directory name (`…_multi-ln_+0.00`), because
it changes the audio the segments contain, and a session left over from before the
admin flipped it must not be mistaken for one that honours it. The suffix is
appended with a DASH: `_SESSION_DIR_RE` treats the tag as one underscore-free
field, and that regex is the only thing keeping the orphan sweep off what may be a
shared scratch path. A test pins both halves — the name changes, and the sweep
still recognises it.
Measured end-to-end on the app's own HLS segments (a real 5.1 AC3 file, same item,
same 160 s offset, same ten segments): OFF -35.2 LUFS, ON -17.2 LUFS. On a 60 s
bench: I -27.6 → -16.6 LUFS with LRA 13.3 → 9.8 LU. A/V sync is untouched — under
this pipeline's own `-copyts` the audio start PTS is identical with and without the
filter, so the 3 s lookahead desyncs nothing.
Direct-played files are served untouched and stay so; a 5.1 mix is always `remux`,
so the reported case is fully covered.
Bug hunt 2026-09, items 8 and 12 — the two things the Plex player was measurably
lying about.
Item 8: picking a subtitle track showed nothing for a long time, and the reported
workaround was to stop/start until it suddenly worked. Measured on prod against a
17.06 GB MKV with two embedded SubRip tracks:
full pass, one track 154 s
full pass, both tracks in ONE run 111 s
-c:s copy (no WebVTT conversion) 143 s
from 1:00:00 to the end of file 66 s
a 630 s window at 1:00:00, both 8 s
So the conversion was never the cost: ffmpeg streams every byte of the container
between its seek point and its stop point, and the array delivers ~140 MB/s. The
only lever is reading less. Three changes follow from that:
* a window around the playhead is extracted and published FIRST (seconds), and
the full pass completes behind it;
* every embedded text track rides that same single read — picking Hungarian used
to cost a full pass and then switching to English cost another one;
* none of it happens inside a request any more. The endpoint answers 202 while a
background worker extracts, the player polls /status (carrying its position, so
the window follows the viewer), and one ffmpeg run finally has a timeout.
-copyts is deliberately absent from a windowed pass: with it ffmpeg ignores -t and
reads the whole container anyway (measured — a [0, 600 s] window still took the
full 145 s), so a window comes back on a zero-based clock and is shifted to
absolute time here. Cached tracks written by older builds have no coverage sidecar
and count as complete, so nothing is re-extracted after the upgrade.
The <track> is no longer mounted before the server has something to put in it; a
"Preparing subtitles: …" notice stands in, and a track that can never be shown says
why instead of leaving the viewer to conclude the picker does nothing.
Item 12: "Bitrate" read a constant 211 kbps on every remux, because it was the
master playlist's BANDWIDTH — which ffmpeg derives from the one stream whose rate
it knows, the -b:a 192k AAC track, plus 10%. It is a measurement now: bytes per
second of MEDIA from hls.js FRAG_LOADED, accumulated per fragment type and summed
(a multi-audio session's video and audio renditions cover the same media seconds,
so pooling them would report half the truth), and an honest dash until the first
fragment lands. The panel also finally names the audio and subtitle track in play.
Verified in the browser against the real Plex library, and every new test
mutation-checked.
The button deletes the local row and mirrors the un-mark to Plex, but it decided whether
to push from the previous STATUS alone (prev_status != "new") — and an in-progress item's
row is status="new" carrying a position, which is exactly how the importer maps a bare
viewOffset. So for the single case the button exists for, nothing was ever sent to Plex:
the offset stayed, and the next pull imported it straight back. The user cleared Protector
twice and it returned to Continue watching at 92s both times.
Push now when Plex holds anything to clear — a view count, or a resume point. "hidden"
stays Siftlode-only and pushes nothing, which is what this endpoint's docstring always
claimed and the status-only guard contradicted.
One unscrobble clears BOTH viewCount and viewOffset — measured against the live server from
a watched+offset state and an offset-only one, so the push itself needed no change. The
client docstring that claimed "clears viewCount" is what sent the first diagnosis down a
wrong path (a timeline write with time=0, which does not clear an offset at all); it now
records what the server actually does.
Verified end to end on the real Plex server: the button leaves viewCount=None viewOffset=None,
and a full 44,885-item re-import brings back no row for it.
`test_only_a_show_row_pays_for_a_resolution` seeded no watch state, so every rail query
inner-joined its way to zero rows and the loop body — with it the assertion — never ran. It
now seeds a video, a movie and an episode, and asserts all three kinds are present before
the loop, so it cannot go vacuous again.
The Plex half of the surprise rail had no test for either of its promises: every test that
reached `plex_surprise` passed budget="any" against a library with no state rows, so the
duration bounds and the already-started exclusion could both be deleted with the suite
green. The YouTube half has had tests for the same two rules since S1 — the asymmetric half
of one principle.
`plex_item` and `plex_show` — two of the five pinnable kinds — were pinned by no test at
all, leaving the rating_key → internal-id translation and both shelf arms uncovered,
including the self-reference cycle that once 500'd the shelf. And `GET /api/discover/pick`
was exercised only through its domain function, so the route's own query-string translation
could regress silently; the new test excludes all three candidates, which is what proves the
whole list parsed rather than one unsplit string.
`PIN_COLUMNS`/`PIN_KINDS` claimed in a comment to be the one place the pin vocabulary is
spoken and had no readers anywhere: `resolve_ref` re-spelled the column names and the route
re-typed the kind list. Both now drive off the constants.
The `_NO_TS` sentinel was introduced as a crash fix and is not one. The key it replaced,
`(ts is not None, ts)`, cannot raise — tuple comparison finds the first UNEQUAL element via
`==`, so two null keys compare equal and `None < None` is never reached; three independent
runs confirmed it over every permutation. The rationale now says what the sentinel actually
buys, lives once in `plex.rollup`, and the third rail that was still on the older spelling
has been swept to match. Its test is renamed and re-scoped to the mutation it does catch.
Frontend: the reroll's second press asserted `expect.any(Array)`, which cannot tell an
accumulating exclusion window from a replacing one — both that and SEEN_WINDOW = 1 (each
reinstating the measured nine-card cycle) left the suite green. And `onInfo`'s jump type was
untested while `onPlay`'s was pinned twice, so pointing the ⓘ affordance at the player
typechecked and shipped green.
All eight fixes mutation-checked; every mutant died on its intended test.
The browse scope is a pure DB query with no request state, but it lived in
routes/plex/_shared.py. E24's Home surface needs the same scope from a domain
package, and a domain package importing from routes/ inverts the one-way
dependency CLAUDE.md draws.
Moved to app/plex/query.py and renamed to content_libs (public — it is a domain
helper now). Swept all five call sites and both tests in one go rather than
leaving a re-export shim: the canonical-patterns table already names the
routes.admin re-export as the anti-pattern.
Deep-review round 2 on the round-1 fixes (all low):
- part_key churn (state): round 1's part-key reset re-probed EVERY sync for a
multi-version item (an original + a Plex 'Optimized' copy), because
_media_facts reads media[0]/part[0] whose order Plex doesn't guarantee stable,
oscillating transcode_hw between the two versions. Gate the part-key signal on
a single-version item (len(Media)<=1); codec-change reset still applies to all.
A multi-version same-codec re-encode not self-healing is an acceptable rare
edge; a per-sync re-probe loop is not.
- dead i18n key: this diff's playableClass() collapses to transcode_hw/sw, so the
bare plex.playable.transcode key it orphaned is now unused (knip can't see
dynamic keys). Dropped from both locales, parity kept.
- test: cross-facet narrowing by an active playable class (a co-facet's year
range collapses to the selected subset) — the flip side of self-exclusion,
previously unpinned.
The aria-label+title double-announcement on the card dot is a deliberate accepted
trade-off (reliable SR name + mouse tooltip). Backend + FE gates green.
Deep-review round 1 on feature/plex-playback-class-filter:
- Sticky watermark (medium): _apply_item reset the bit-depth probe watermark
only on a codec-string change, so a same-codec re-encode (8-bit->10-bit HEVC)
kept a stale video_bit_depth/transcode_hw forever (the backfill only re-probes
media_probed_at IS NULL rows). Also reset when the Plex Part key changes (the
file was replaced), not just the codec.
- i18n parity (medium): the HU playback-class labels were mixed-language (only
'direct' translated). Translated remux/transcode_hw/transcode_sw and the
tooltips consistently to Hungarian (Kozvetlen streameles / Atkodolas).
- Layering (low): _PLAYBACK_CLASSES was module-private yet imported across into
routes.plex.browse; renamed to public PLAYBACK_CLASSES (it is the shared
source of truth for facet order + zero-fill).
- a11y (low): the browse-card playback-class dot conveyed meaning only via a
mouse-hover title; added role=img + aria-label so SR/touch users get the class.
- tests: the NULL/unprobed transcode_hw -> SW fold (filter + facet), facet
self-exclusion under an active playable filter, the migration codec-set ==
HW_DECODABLE_CODECS drift guard, the probe-batch bound + oldest-first order,
the transient probe-failure retry path, and the same-codec re-encode re-probe.
Backend gate green (ruff + full pytest); FE gate green (tsc/lint/knip/prettier/
vitest). NOTE: the facet-vs-runtime disagreement for 10-bit HEVC (C1) is the
known merge-order constraint — this branch's runtime is codec-only; the facet
becomes truthful once BUG1/tonemap's bit-depth-aware routing lands with it. Left
as a sequencing constraint, not code-changed here.
Deep-review round 3: the only surviving finding — a comment-drift nit. The
fallback's cull-yield comment named the route's 'Not available yet' path, but a
culled-key segment request actually returns 'No active playback session' (the
substantive invariant — never a None that 404s as file-not-found — is unchanged).
Reworded to match the route. Concurrency + correctness came back clean.
Deep-review round 2 on the round-1 fix:
- Race (medium): the round-1 identity guard did the supersede check and the
respawn in TWO separate _lock sections. A newer same-key session installed in
the gap was still torn down by _spawn's unconditional pop+kill (the round-1
comment even had the direction backwards). Fold the check into _spawn via a
new replace_only= param: the identity test and the pop+replace now share ONE
lock, so the respawn is a true no-op when superseded/culled.
- Honest states (low): the superseded branch returned current_session(key),
which is None when the key was cap-culled during our out-of-lock wait; the
route turns None into a dishonest 'Media file not found on disk' 404 for a
file that exists. Yield 'current_session(key) or s' so a culled session
degrades to the route's retryable 'not available yet' path instead.
- Hygiene: _want_hw docstring corrected (it is a test-only shim now; production
decides via _select_encode) and the two stale _want_hw references refreshed
(stream.py header, config.py plex_transcode_hw).
- tests: the cap-cull no-respawn/no-None sub-case; the _opencl_present except
OSError branch (glob raising, not just empty); tightened a tautological -vf
assertion. Backend gate green (ruff + full pytest).
peter