An <img> can't send the X-Siftlode-Account header, so the poster.jpg request
resolved to the session-default account and _accessible_job 404'd it for any
download owned by a different (per-tab) account — every card then fell back to
the source's remote thumbnail, which for Facebook is an expiring fbcdn URL, so
the Library showed blank thumbnails in a multi-account session. The 0.69.1
local-poster-primary change made this the primary image, exposing it broadly.
Append ?account= to the poster URL via a new api.posterImageUrl(), the same
wallet-gated header-less selection downloadFileUrl / storyboardImageUrl already
use (resolved_user_id reads ?account= for exactly these contexts).
Three prod bug fixes: local-poster-primary Library thumbnails, viewport-scaled
feed read-ahead (fixes the fast-scroll "Load more" stall), and the channel About
popover fitting/blank-line cleanup.
Bump VERSION 0.67.0 -> 0.68.0 and add the user-facing release-notes entry for
the E2 epic (feed sort index, keyset/NOT NULL, search GIN + state inner-joins,
single /feed/summary pass, no card-action refetch storm, job partial indexes).
Review round on the E2 stack (S1-S5). No correctness bugs; the surviving
low/medium findings addressed:
- altitude (#1): extract the shared `tags_by_category` helper used by BOTH
filtered_query (SQL tag WHERE) and feed_summary (Python facet/count), removing
the category-grouping duplication, and add bidirectional "keep in lockstep —
equivalence-tested by test_feed_summary_matches_count_plus_facets" comments so
the SQL<->Python tag rule can't silently drift.
- test-coverage (#2): the /feed/summary equivalence test seeded only system tags,
leaving the `ChannelTag.user_id == user.id` visibility branch untested. Add a
user-owned Tag + link + two combos; the suite is now 346 (was 344) and the new
cases pass (feed_summary agrees with /feed/count + /facets on user-visible tags).
- dead-code (#3): annotate get_feed_count, get_facets, and the FE api.feedCount/
facets as "SUPERSEDED by /feed/summary — kept only as the equivalence oracle;
retire together" so the intentionally-retained surface reads as deliberate.
Not changed (confirmed no-op): the search channel-name resolve (#4) is
once-per-request on a 633-row table and a deliberate pre-resolve for PERF-12's
BitmapOr plan; the facet loop (#5) already computes passes() once per
(channel, category) with a distinct exclude each, so there is no recomputation.
Gate green: ruff + 346 pytest; frontend tsc/knip/prettier (FE change is a comment).
The FE fired /feed/count (1 pass) + /facets (3 passes) on every filter change =
4 full scans of the subscribed set. New /api/feed/summary does ONE grouped query
(the base filter with tags stripped, GROUP BY channel_id -> per-channel video
counts) and derives the total count + every tag's distinct-channel facet count in
Python on the ~few-hundred-row channel map, faithfully mirroring filtered_query's
channel-based tag logic + the /facets drill-down (own category dropped, unless a
conjunctive topic-AND). Returns {count, facets} — the same numbers the two old
endpoints returned.
FE: a single shared query (Feed reads .count, Sidebar reads .facets); same key +
filters, so React Query dedupes to ONE request. Invalidations repointed to
qk.feedSummary. The old /feed/count + /facets endpoints (and their now-unused FE
api/qk) are kept as the equivalence-test oracle + safety net; a later cleanup can
drop them.
New equivalence test (test_feed_views, 8 filter combos incl. cross-category,
disjunctive/conjunctive topic modes, search): /feed/summary count == /feed/count
and facets == /facets, byte-identical. Structural win: 2 requests / 4-6 queries
-> 1 request / 2-3 queries (the review measured 530 -> 144 ms). Gate green:
344 pytest, frontend tsc/eslint/knip/prettier/222 vitest.
Bump VERSION 0.66.0 -> 0.67.0 and add the user-facing release-notes entry for
the E1 Foundation epic (data-safety fixes, account-switch + search-error fixes,
graceful key-mismatch degrade; plus the security/test-net/CI/repro chores).
SEC-10 (account pre-emption / "pre-hijack"): a password set on an account whose
email was never verified is untrusted — whoever registered the address (to
pre-empt the real owner) could have chosen it. Both the Google-adoption callback
and admin activation (_activate_user_by_email) now drop that pre-set password and
bump the session epoch when the account isn't email_verified, so a later Google
login / admin approval can't hand the pre-set password a working login. No-op for
a verified account (the legit no-SMTP flow). +2 tests.
SEC-11 (link/upgrade bound to the session default): the OAuth link/upgrade GETs
resolved to the session DEFAULT account, so a tab acting as a secondary account
would link Google / the YouTube grant to the wrong account. The frontend now
threads the active account through the navigation URL (withActiveAccount ->
?account=<active>), and _complete_link accepts any wallet account, not only the
default, so the identity lands on the account that started the flow.
A vitest that fails the gate when the newest release-notes entry's version
diverges from the repo-root VERSION file (the two hand-edited version sources),
plus a no-duplicate-versions check. Catches the "bumped one, forgot the other"
release mistake.
Ships the E5 sprint stack (S1–S7): field-scoped Plex search + typeahead, cast &
crew pages, the HU-overlay Hungarian metadata + language-aware serving, the
IMDb/audience rating split, the toolbar relocation, and the narrow/mobile header
fix — plus the underlying Plex-module refactors (feedquery/plex.cards/plex.query
extraction, db.query→2.0 select).
- VERSION 0.65.2 → 0.66.0
- releaseNotes.ts: 0.66.0 user-facing entry
- prettier-format the E5-touched frontend files (release hygiene)
Merged main (0.65.2 download-gc hotfix) in first; no code conflicts (E5 is
Plex-only, the hotfix was notifications/downloads).
1. (correctness) Show hero rating regression: PlexBrowse's show hero rendered a
rating only when show.imdb_rating was set, so after the S6 IMDb-only change a
show with no imdb:// entry showed no rating. Now falls back to the audience
`rating` like PlexInfo.
2. (correctness) ratingMin column mismatch: the rating filter and the /facets
rating_max now key off coalesce(imdb_rating, rating) — the SAME "best score"
the badge displays — so a ratingMin picked from a shown IMDb score narrows the
right column instead of the raw audience `rating`.
3. (altitude) HU search gap: new query.vector_match() matches EITHER vector in HU
(or_(search_vector_hu, search_vector)) so a row with no HU twin stays findable
by its English title; ranking still uses the HU vector. field_recheck unions
the HU+EN per-field tsvectors in HU, staying a strict subset of the prefilter.
4. (test-coverage) Extracted the badge logic to lib/plexRating.pickRatingBadge
(shared by PlexInfo + the show hero) with a unit test (imdb-preferred, audience
fallback, 0-score, none). Backend tests added for the HU union search + the
IMDb-preferred rating filter.
5. (efficiency) Language switch no longer double-fetches the grid: browseQ is
gated on a committedLang state that lags `lang` by one render, so the
genre-clear effect settles into a single fetch instead of one throwaway fetch
with stale-language genres plus a second.
DB-lane suite 295 passed / 2 skip; FE tsc/knip/lint clean, 217 vitest pass.
E5 S6 frontend half.
Language threading (?lang=hu serves the HU overlay; en is the default):
- api/plex.ts: `lang` on plexLibrary / plexFacets / plexItem / plexShow (only
sent for hu, so EN URLs stay byte-stable); PlexItemDetail gains `rating`.
- queryKeys: plexItem / plexShow take an optional trailing lang so a switch caches
HU/EN separately; the ...rest keys (library/facets) already carry it. Prefix
invalidations (qk.plexShow(id) / qk.plexItem(id)) still match either.
- i18n: plexLang() collapses the UI language to en|hu for non-component code.
- PlexBrowse / PlexSidebar: thread lang into the grid, facets, item and show
queries (useTranslation re-renders on switch → the key changes → refetch).
- plexPlayerHooks: the optimistic resume setQueryData now keys by plexLang() so
it still lands on the language-keyed detail query.
- Clear the genre-filter selection on a language switch — genre chips are
language-specific strings, so a stale one silently empties the grid.
IMDb-only rating badge (folded-in S5b review #5):
- PlexInfo.tsx shows the IMDb badge + imdb.com link only when imdb_rating is
present; otherwise it renders the Plex audience `rating` (no IMDb link/branding),
so a non-IMDb score is never mislabelled. ratingMin filter keys off whichever
score is shown.
Tests: i18n parity.test.ts (en/hu key sets identical per namespace — currently
green, no drift). FE gate green: tsc / knip / lint clean, 213 vitest tests pass.
- PlexPersonView.dom.test.tsx: variadic vi.fn mock for plexSetState so the
spread call typechecks (TS2556 — zero-arg mock couldn't take ...args).
- plexSearch.ts: drop the unused `export` on PLEX_SEARCH_FIELDS (module-local
only; knip flagged it as an unused export).
tsc / knip / lint all green; ruff clean.
Backend: new GET /api/plex/person/{name} returns a cast/crew member's cached headshot plus
the movies (PlexItem, kind=movie) and shows (PlexShow) in the enabled libraries whose
cast_names OR directors include the name — the inverse of the cast → filter link. Introduce a
shared _content_libs(db) helper (enabled-library partition) and reuse it in people_cards;
S5 will extend it with the HU-overlay exclusion. Watch-state comes from PlexState / rollup
(reusing _movie_card / _show_card / state_map). DB-lane tests cover cast+director credits,
library scoping, the kind filter, unknown-name empties, and the partition helper.
Frontend: a new `person` history-subview. A cast-row click (PlexInfo cast row + the show
page's CastStrip) now opens the person's filmography page instead of filtering; the filter
gesture moves to a "Filter library by {name}" button on the page itself. PlexPersonView
mirrors PlexInfoView's fetch-wrapper + glass hero (headshot + name + credit counts) and
renders Movies / Shows as PlexPosterCard grids, with an optimistic per-card watch toggle on
the person cache. New plexPerson query key + api client + PlexPersonDetail type; HU/EN copy.
Move the two "what am I browsing" controls out of the filter sidebar to where
they're always in reach, matching the Feed's layout:
- Library scope (both/movie/show) → a compact SegmentedToggle in the header next
to the search box (new PlexScopeToggle, reads the same provider state).
- Sort → a SortControl in the row above the cards (into the PageToolbar band,
count on the left, sort on the right), like the Feed.
Sidebar cleanup: drop the `scope` and `sort` island cards; the panel badge now
counts only the filters still in the panel (expanded filters + watch-state);
remove the now-unused sort imports. panelLayout drops both from the default plex
order (stale saved-layout keys are auto-filtered, so existing layouts degrade
cleanly). usePlexActiveFilters no longer emits a sort chip — sort is an
always-visible control now, not a removable narrowing (same as the Feed).
FE-only, no backend/migration. typecheck + eslint clean; 171 vitest green;
self-E2E verified (header scope drives the grid, sort control reorders, sidebar
no longer shows the two cards, no sort chip).
Add a search scope-pill to the Plex header search box (GitHub-style, one
field-scope at a time) and an instant typeahead of the top matches below it.
Backend (unified_library):
- New `field` param: all | title | people | summary | episode. `all` is
byte-identical to the pre-scope behaviour (untouched weighted search_vector);
a specific field keeps the GIN `search_vector @@ tsq` prefilter AND ANDs a
per-column tsvector recheck, so scoped hits are a strict subset of Everywhere.
Per-field expressions mirror the search_vector Computed def (same TS config,
same left(summary,1000) truncation).
- `episode` scope empties the movie/show grid and title-only-matches the grouped
Episodes section; the title/people/summary scopes suppress that section.
- Unknown field falls back to `all`. New DB-lane characterization tests
(test_plex_field_scoped_search.py): each scope narrows correctly, field=all is
identical to today, episode empties the grid and needs shows in scope.
Frontend:
- PlexSearchBox wraps the shared SearchBar with the scope-pill (reusing
ActionMenu) and the typeahead; SearchBar gains generic leading/below/focus/
keydown slots so other modules' search boxes are untouched.
- Episode scope offered only when shows are in scope (movie-only hides it),
healed at the point of use like the sort state.
- Typeahead reuses /api/plex/library (limit=10), honours scope+field (not the
sidebar filters), shows poster+title+year, keyboard nav (arrows/Enter/Esc),
and jumps straight to a movie's info page or a show's drill-down via a new
provider jumpTo hand-off (mirrors the playlistOpen pattern).
- `field` persisted per-account; en/hu labels added.
Bumps VERSION to 0.65.1 and adds the release note for the backfill fix
(immediate "get full history" now completes channels over ~3000 videos in
one run) plus the per-page durability + centralized reserve-guard hardening.
Bump VERSION to 0.65.0 and add the user-facing release notes for R14
(zoom-fit manager table, ✓ charm, N/S/L moved to the channel page,
live-refresh, immediate "get full history").
- Make the immediate-backfill guard SYMMETRIC: run_deep_backfill now also
registers the channel it pages in _backfill_in_progress, so it and
backfill_channel_now mutually exclude per-channel and never advance the
same cursor concurrently (whichever started first). The interactive cap
moves to a separate _immediate_running counter so the scheduler's channels
don't false-trip it.
- Attribute the immediate backfill's spend: wrap it in
quota.attribute(user_id, VIDEOS_BACKFILL_FULL) (route passes user.id), so
it shows under the user's full-backfill action, not the OTHER/system bucket.
- Tests: the update_channel->background-task wiring + the demo guard
(scheduled for a real user, NOT for demo); distinct N/S/L counts (2 short
+ 1 live) so a short/live swap fails; extracted channelDetailPollMs with a
pure unit test of the poll-or-stop predicate.
- a11y: role="img" on the fully-synced charm so screen readers announce it.
- list_channels no longer computes the Normal/Shorts/Live aggregation the
frontend stopped reading when R14 dropped the N/S/L table column; the dead
ManagedChannel.count_* fields are gone too.
- Immediate "get full history" backfill now coordinates with the paced
scheduler via a small in-flight set: capped at 2 concurrent + per-channel
dedup, and run_deep_backfill skips a channel an immediate backfill owns —
killing the cursor race and bounding the quota-drain burst, without losing
the speed win.
- Add test_channel_backfill.py: the channelDetail N/S/L counts + the guard's
no-op / cap / scheduler-skip behaviour.
- Inline the syncedChannels useMemo (its dep is a fresh array each render, so
it never cached); fix the stale "~2-min ticks" comment (it's 10 minutes).
The channel detail page now polls (via useLiveQuery) while THIS user's
full-history backfill is running (deep_requested && !backfill_done), so
the stored count and the N/S/L breakdown climb on their own as the
scheduler adds videos — no manual F5 — and it stops polling once the
back-catalogue is in. Widen useLiveQuery's interval callback to allow
returning `false` (stop) so "poll while active, stop when done" is a
first-class, reusable cadence.
- DataTable: opt-in `fixedLayout` (colgroup + table-layout:fixed) so the
table fits its container at any zoom/text-size; only the channel manager
opts in — playlists/discovery/Plex keep auto-sizing.
- Remove the N/S/L and Sync columns from the manager table. Fully-synced
status becomes a ✓ charm on the Stored count; the sync filter moves to an
`applySyncFilter` pre-filter still driven by the top chips.
- Channel column is now fluid: ChannelLink truncates with the full name on
hover (title). Data columns get fixed widths so only Channel/About flex.
- N/S/L (normal/short/live) breakdown moves to the channel page: add the
three counts to the channelDetail endpoint (one grouped query, same
filters as the list) + the ChannelDetail type + a meta-line entry.
- Drop the now-dead sync/types i18n keys (en+hu).
Bump VERSION to 0.64.0 and add the user-facing release-notes entry for
the R13 sprint (unified sort/filter controls, Hidden channels tab,
channel-manager toolbar declutter, filter-rail jitter fix).
The deep-review flagged that the feed sidebar's new date single-select
derivation was the one bit of migrated logic without a test, while its sibling
primitives (FilterChipGroup/SegmentedToggle/SortControl) all gained one. Lift
the pure logic into lib/dateFilter (deriveDateKey + applyDateChip), same split
as columnSort, and cover it: dateKey for none/preset/custom/off-preset-value,
the preset toggle-off, the preset-not-active-while-custom-open case, and the
Custom open/close + maxAgeDays clearing. Sidebar now calls the helpers; the
onChange is behaviour-identical (self-E2E'd: preset on/off, Custom editor).
11 new tests (vitest 136 -> 147). gate green. Branch not merged.
Retire the hand-rolled channel-manager view pills for the shared <Tabs> component (given an
optional className so it embeds in the justify-between header without the default mb-4), and
back view + layout on usePersistedTab instead of manual getAccountRaw/setAccountRaw useState.
That subsumes the account-pin re-hydrate (usePersistedTab now re-reads on the pin), so
ChannelsProvider's bespoke markManagerDirty/useAccountRehydrate is gone; setView keeps the
history-tab logic (_chTab pushState + popstate) layered on the stable raw setter, and
useAccountPersistedState's setter is now memoized so that stable action holds.
Per the user's call, unify every channel-manager pill on the rounded-lg idiom: the view tabs
via <Tabs>, the sync chips and tag chips + the "manage tags" affordance from rounded-full to
rounded-lg — no more odd-one-out rounded-full.
gate-green: tsc + vitest 120 + eslint-0 + knip + prettier. Self-E2E'd: tab switch, browser
Back across tabs (Blocked->Hidden->Subscriptions with localStorage in sync), and a fresh tab
restoring view=hidden on the FIRST load; zero console errors.
Addresses the 2 findings from /code-review high on 726c382:
- NavigationProvider: openChannel/enterYtSearch no longer markDirty. A sub-view is an
overlay, not a page change — the page re-hydrate's replaceState already spreads the
_chan/_yt marker, so the overlay survives on top of the RESTORED page. Removing the
markDirty lets a fresh tab still restore the stored landing page under a pre-pin overlay
(previously it stuck on the default). setPage keeps markDirty (a real navigation).
- useFilterState.dom.test.ts: model a mutable pinned account and add 3 integration tests —
null->id re-hydrate adopts the stored map, the migrate seed carries a legacy value forward
on re-hydrate when the key is empty, and a pre-pin edit is preserved (markDirty guard).
gate-green: vite build + vitest 120 + eslint-0 + knip + prettier. Fresh-tab restore re-E2E'd
(lands on Plex with full state, zero console errors).
Fixes the 5 findings from /code-review high on the account-pin work:
- Extract the reactive-account primitives (ACTIVE_ACCOUNT_KEY, activeAccountId,
notifyActiveAccountChange, useActiveAccountId) into a new dependency-free
lib/activeAccount.ts, breaking the storage.ts <-> useAccountRehydrate.ts import cycle.
storage.ts / api/core.ts / useAccountRehydrate.ts now import from it.
- NavigationProvider: the page re-hydrate now PRESERVES existing history.state markers
(spread, not overwrite), and openChannel/enterYtSearch markDirty so a sub-view opened
before the account pins isn't clobbered by the page override.
- useAccountRehydrate: reset the dirty flag after each account transition, so a future
non-reloading account switch can't skip the re-hydrate and leak the prior account's state.
- Doc-link useAccountRehydrate to the pre-existing useEffect([me?.id]) server-adopt pattern.
- Tests: model a mutable pinned account and add useSortState integration tests for the
null->id re-hydrate + the markDirty no-clobber guard (vitest 115->117).
gate-green: vite build + vitest 117 + eslint-0 + knip + prettier. Self-E2E'd on a fresh
tab: still lands on Plex with Movies/In-progress/Year restored, zero console errors.
The fresh-tab restore bug wasn't Channels-specific: any provider that mounts above
<App> and reads per-account state at useState-init hit it, since <App> pins the tab's
account only after /api/me. Audited all of them — FeedFilters/FeedView/Prefs/App already
self-correct via their own me?.id server-prefs adopt effects, so the still-broken ones
were the pure-localStorage providers with no server adopt: NavigationProvider (landing
page), PlexProvider (scope/show/sort/filters/q), PlaylistsProvider (selected playlist).
Each now adopts the shared useAccountRehydrate hook (markDirty in the setters; page uses
replaceState so the restore doesn't add a phantom Back step). Also baked re-hydrate into
useAccountPersistedState/Object so their consumers (persisted tabs, player prefs) are
robust even if one ever mounts before the pin. Self-E2E'd on a genuinely fresh tab: it
lands on Plex with Movies/In-progress/Year/search all restored on the FIRST load.
The provider tree mounts above <App>, which pins this tab's account only after /api/me
resolves — so on a genuinely fresh tab (empty sessionStorage) the manager's sort, filters,
view and layout initialised from an unknown account and stayed at defaults until a second
reload. Make the active account reactive (a subscribable store + useActiveAccountId via
useSyncExternalStore; setActiveAccount/clearActiveAccount notify) and re-read the stored
value when the account transitions null→id, via a shared useAccountRehydrate hook. A
markDirty guard (called by each owning setter) prevents clobbering a value the user changed
in the brief pre-pin window; the effect is keyed on the account id only, so the user's own
writes can't loop it. useSortState, useFilterState and ChannelsProvider (view/layout) all
adopt it. Affects sort identically — this was a long-standing caveat, not new to R13.
Review fixes (user /code-review high):
- DataTable: hoist a stable EMPTY_FILTERS fallback so a controlled caller that omits
`filters` doesn't churn a fresh {} each render.
- DataTable: the persistKey blob now stores only the slices this table OWNS — a
controlled sort/filter is written as null/{} instead of a stale never-read duplicate.
- useFilterState: optional `migrate` seed (used only when storage is empty);
ChannelsProvider carries the retired LS.channelFilter sync chip forward via migrateSyncChip.
- Channels: guard the sync-chip read with typeof instead of an `as string` cast.
Also unify the tag filter, same as the sync one: the top "YOUR TAGS" chip row is now a
controlled view over the Tags COLUMN filter (`filters.tags`), replacing its own tagFilter
state — so the chips and the Tags funnel share one state bidirectionally, the card layout
applies it via filterRows, and it persists with the rest of the map.
The top sync chips and the Sync column's own select filter were two independent
states with different vocab (all/needs_full/fully_synced vs fully/needsFull) and
subtly different semantics. DataTable gains a controlled filters/onFiltersChange
pair (symmetric to the existing controlled sort); ChannelsProvider now owns the
filter map via a new useFilterState hook, so the top chips, the Sync column
funnel, the card layout, and the header "full history" deep-link all drive ONE
filters.sync. The card layout reuses the exported filterRows predicate instead of
its own status pre-filter. Vocab unified on the column's fully = recent_synced &&
backfill_done, which now matches the Sync cell's own "fully synced" badge.
- SortControl no longer renders a blank select when an always-sorted view
(no defaultLabel) is handed a key its options don't offer: it falls back to
the first option instead of value="" (which also tripped React's controlled
warning). Robustness now lives in the primitive, not each caller.
- PLEX_SORT_FALLBACK is Object.frozen: it's handed out by reference from
healSort/healPlexSort and stored as provider state, so freezing turns any
accidental in-place SortState mutation into a loud error instead of silently
corrupting the shared fallback.
- Test: SortControl falls back to the first option for an unavailable key.
Gate green: tsc/eslint(0 err)/prettier/knip clean, vitest 106.
- PlexSidebar activeCount now compares the HEALED sort, so an out-of-scope key
(movie "duration" after switching to shows) no longer inflates the
active-filter badge while the control + grid show the default.
- PlexProvider strips the legacy `sortDir` from the filters blob, so it stops
leaking into the facets query key and being re-persisted on every change.
- PlexProvider sort migration guards the fall-through: valid JSON that isn't a
SortState now falls back instead of using the raw JSON string as the key.
- New shared healPlexSort(sort, scope) is the single heal both PlexBrowse and
PlexSidebar use (removes the noop-label plexSortOptions call + the duplication).
- Tests: useSortState.dom.test.ts (hydrate/persist/heal round-trip via renderHook)
and SortControl.dom.test.tsx (options, pick, direction toggle, directionless,
trailing). vitest 94 -> 105.
Gate green: tsc/eslint(0 err)/prettier/knip clean, vitest 105.
ChannelsProvider and ChannelDiscovery each hand-rolled the same sort plumbing
(useState<SortState> + parseSortState + a persist key). Both now use the shared
useSortState hook. hydrateSort gains an empty-options rule: with no vocabulary
passed it parses without healing (these owners heal for display via
ColumnSortControl against the live columns), so the un-healed provider store is
preserved. plexSortKeys made module-private (knip).
Closes R13 S1: feed, Plex, and the two channel tables all sort through one
SortControl + one SortState model. Gate green: tsc/eslint(0 err)/prettier/knip
clean, vitest 94.
Plex sort was a bespoke chip-pill row plus a separate asc/desc chip pair, with
the key in PlexState.sort and the direction hidden in PlexFilters.sortDir. Now:
- sort is one SortState {key,dir}; sortDir is dropped from PlexFilters (and from
the facet-key destructure). The wire call still splits it into sort + sort_dir
at the server boundary, so the backend is untouched.
- The chip sort-row is replaced by the shared SortControl (select + arrow) — the
same idiom the feed and channel manager use. New lib/plexSort.ts owns the
scope-dependent vocabulary (movies add duration) + the title/asc fallback.
- PlexProvider migrates the old storage in place: a legacy bare-key plexSort
string plus the old filters.sortDir become the {key,dir} state, so an existing
sort preference survives the upgrade.
- PlexBrowse and PlexSidebar heal the stored sort against the current scope's
options, so a movie-only key (duration) left over after switching to shows
can't ask the server for an unavailable sort.
Gate green: tsc clean, eslint 0 errors, prettier clean, vitest 93.
Introduce the one sort control every content surface will share:
- columnSort.ts gains the SortOption vocabulary + pure pickSort/healSort/
hydrateSort helpers (direction lives in SortState, never in the key).
- SortControl.tsx: the generic dropdown + asc/desc toggle, with an optional
"no sort" choice (tables) and a trailing adornment slot (the feed's reshuffle).
- ColumnSortControl is now a thin wrapper over SortControl (API unchanged, so
Channels + ChannelDiscovery are untouched); optionsFromColumns moved to
lib/sortOptions.ts to keep the component file fast-refresh clean.
- useSortState.ts: the per-account persist/hydrate/heal plumbing the channel
manager and discovery each hand-rolled, ready to replace both.
Pure logic is fully unit-tested (columnSort.test.ts 10 -> 23). Gate green:
tsc/eslint/prettier clean, vitest 93 pass. Feed/Plex/channels migration follows.
- applyVolume/setVol now pass the element's CURRENT level as the normalize
fallback instead of undefined, so a corrupt persisted volume holds the
playing level rather than jumping to PLEX_SCALE.default=1 (full) — the "full
blast" failure mode playerVolume.ts's header explicitly warns against. (The
old bare clamp muted instead; both are wrong for corrupt data.) readMerged
shallow-merges stored prefs without per-field validation, so a corrupt
volume is reachable.
- setVol normalises the level ONCE and both applies + persists that same value,
so the <video> element and the stored pref can never disagree.
- Test documents the hold-last-known contract (corrupt raw + valid fallback ->
fallback; only an unusable fallback lands on the scale default).
Gate green (tsc / eslint-0 / prettier / knip / vitest 83 / pytest 234).
Volume-consolidation (conservative — no behaviour or prefs-format change):
- playerVolume.ts is now scale-generic (a VolumeScale parameterises range /
default / integer rounding). YT_SCALE (0-100 ints) and PLEX_SCALE (0..1
float) bind the same tested core; the YouTube wrappers keep the original API
so the existing tests are unchanged.
- The Plex usePlayerVolume hook drops its bespoke clamp01 arithmetic for the
shared normalizeVolumeIn / stepVolumeIn at PLEX_SCALE. Side effect (a latent
bug fix): a corrupt persisted volume now falls back to the default instead of
a bare clamp silently muting it. The separate .muted flag stays Plex-only.
- clamp01 stays for the seek-bar fraction (its non-volume use).
- Tests: the generic core is covered at both scales (float no-round, corrupt ->
default, step/clamp) — the Plex-side arithmetic finally has unit coverage.
PlexPlayerSettings extraction (verbatim JSX move, behaviour-equivalent):
- The ~240-line gear/tuning panel (sync / playback / subtitle / clock tabs)
moves out of PlexPlayer into PlexPlayerSettings.tsx (a pure prefs+patchPrefs
view owning no state). PlexPlayer shrinks 1402 -> 1173 lines; Slider/Toggle
imports it no longer uses are dropped.
Gate green (tsc / eslint-0 / prettier / knip / vitest 83 / pytest 234). Dev
server smoke: app builds + loads clean, no console errors. The interactive
settings panel needs a playing Plex video (unreachable on localdev) -> owed to
the user's UAT.
S4a-4a of the PlexPlayer decomposition. Move audio/subtitle track
selection + subtitle rendering into usePlexTracks: the selected-ordinal
state, change/cycle audio+subtitle, the debounced sub-offset, the audio
A/V-sync offset, and the two subtitle effects (force-showing the mounted
track + positioning cues at the user's subPos). The shared media refs
(video/hls/audio/aoff/abs/multiAudio) stay component-owned and are passed
in, since loadSession + the detail effect write audioRef too; absRef is
hoisted above the hook call for that. No behaviour change; verified by
self-E2E (subtitle on/off via 'S' mounts/removes the <track>, mode
'showing', VTT 804 cues all positioned to line=subPos, subLang persists;
playback uninterrupted). Audio switch is single-track here so cycleAudio
correctly no-ops.
S4a-3b of the PlexPlayer decomposition. Move the intro/credits auto-skip
machinery (activeMarker detection, the skipProgress countdown, doSkip,
cancelAutoSkip + the running-interval and reset-on-id effects) into
useAutoSkip. The hook returns skipProgressRef / cancelAutoSkipRef — the
forward-declaration bridge the earlier keyboard/back handlers read to
cancel a running countdown; those two stable refs are now listed in the
backIntent/keyboard dep arrays (they were provably-stable local refs
before, params now). No behaviour change; verified by self-E2E (credits
marker detected -> countdown fill renders + advances while playing ->
auto-skip fires -> item marked Watched + player closes).
peter