Codex-review nits: drop the now-inert "MudBlazor" Serilog level override in
appsettings.json (package removed) and reword the PlaylistController comment
that referenced the deleted Blazor MultiSelectBase.AddItemsToPlaylist path. No
behavior change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Cold-review Low-1: ErsatzTV/Extensions/NavigationManagerExtensions.cs survived
the removal — the last non-deleted .cs still importing
Microsoft.AspNetCore.Components/JSInterop and calling the deleted
blazorHelpers.scrollToFragment JS. Fully unreferenced (compiled only via the
shared framework). Removing it completes the Blazor deletion.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The ChicoryTV React SPA (web/, served at /app) now has full parity for every
route the Blazor UI served, so the legacy Blazor Server / MudBlazor UI is
deleted. This is the milestone-capping removal of #91 phase (b).
Deleted: ErsatzTV/Pages/**, Shared/**, ViewModels/** (39 edit VMs),
Validators/** (10 edit-VM validators), App.razor, _Imports.razor,
Locals/{Shared,Pages}/** (Blazor loc resx; Locals/Resources.* kept),
wwwroot/css + wwwroot/lib, libman.json, and the orphaned MultiSelectBaseTests.
Startup.cs (surgical, not wholesale): removed AddRazorPages/AuthorizeFolder,
AddServerSideBlazor, AddMudServices, AddSortable, AddCourier, the HtmlSanitizer
registration, the Blazor-attached OIDC UseAuthentication/UseAuthorization
middleware (per the #206 auth-posture sign-off), MapBlazorHub, and
MapFallbackToPage("/_Host"). Renamed the branch blazor->legacy; it still
co-hosts MapControllers, /docs (Scalar), dev MapOpenApi and the redirect
middleware. Replaced the _Host fallback with a catch-all (MapFallback ->
302 /app) that excludes /api|/artwork|/docs|/openapi (genuine 404) per #204.
Kept all OIDC/JWT/API-key service wiring (inert unless configured; real auth
is #197), ConditionalIptvAuthorizeFilter, ApiKeyAuthorizationFilter.
Pruned 9 now-unused packages (all verified zero remaining consumers) from
Directory.Packages.props + ErsatzTV.csproj: MudBlazor, Heron.MudCalendar,
Blazored.FluentValidation, BlazorSortable, MediatR.Courier.DependencyInjection,
Markdig, HtmlSanitizer, Chronic.Core, NaturalSort.Extension. Also removed the
now-dead #25 razor-Sonar NoWarn.
LegacyUiRedirects: added the 14 /media/sources/* -> /app/libraries/* redirects
(SPA screens landed in #202) and lifted the #204-era /media/sources prefix ban.
Tests: Release build clean; full solution suite green. Updated Startup
source-text tests + added regression coverage that Blazor wiring is gone, the
catch-all is wired, and all 14 media-sources routes redirect.
Docs: blazor-route-parity.md (phase b COMPLETE), decisions.md (removal entry),
CLAUDE.md, contributing.md, README.md all updated in this PR.
Rollback: tag blazor-final is cut on pre-merge main as the first merge action.
Part of #91.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adversarial review caught a HIGH the fan-out introduced: activating Version as an
IsConcurrencyToken on Playout/Collection makes EF append `WHERE Version=@orig` to
EVERY root UPDATE, so a non-If-Match writer that saves via plain SaveChangesAsync
now throws DbUpdateConcurrencyException → 500 when a replace-all editor bumps the
row between its load and save. Realistic two-tab trigger (edit playout settings while
editing its alt-schedules; edit a collection's name while reordering) — a new crash,
previously silent last-write-wins.
Fix: shared ConcurrencyExtensions.SaveChangesForcingVersion — on a concurrency
failure it adopts the stored token as original+current (client-wins merge scoped to
the token, never reverting the concurrent bump) and retries, i.e. Phase-1 force-write
semantics for a missing If-Match. Applied to the exposed UPDATE writers:
UpdatePlayout, Update{Sequential,Scripted,ExternalJson}Playout, UpdateOnDemandCheckpoint,
UpdateCollection. Non-vacuous test proves the write lands and the bump survives.
Deletes + repo-mediated Add* writers (rarer / join-rows-only) re-scoped onto #269.
Refs #253#269
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Installs husky git hooks (via web/'s lint-staged + npm, since the JS/TS
project lives in web/ with no root package.json) to catch lint, format,
type, and generated-API-drift errors locally before they reach CI.
Hooks (committed at repo root under .husky/):
- pre-commit: (a) lint-staged runs eslint --fix on staged
web/src/**/*.{ts,tsx} + a project-wide typecheck; (b) if any *.cs are
staged, dotnet format --verify-no-changes on just those files (skipped
when no .cs staged, so web-only commits skip the sln load).
- pre-push: CI-parity gate — cd web && check:api && lint && typecheck &&
build. Blocks pushing drift or a change that breaks an unstaged file.
- commit-msg: requires a Co-Authored-By trailer (merge commits exempt).
Wiring: web/package.json gains husky + lint-staged devDeps, a lint-staged
config, and a `prepare` script (cd .. && husky) that points git's
core.hooksPath at the repo-root .husky dir on npm install. A fresh
`web/` npm install installs all four hooks automatically.
Monorepo/worktree gotchas handled:
- husky init hard-checks for .git in cwd, so `prepare` cd's to the repo
root before invoking husky (npm keeps web/node_modules/.bin on PATH).
- git exports GIT_DIR while running hooks; in a worktree/subdir that made
pre-push's `git diff` (check:api) mislocate the working tree and pass
silently on drift — pre-push now unsets GIT_DIR/GIT_WORK_TREE/GIT_INDEX_FILE.
docs/ci-cd.md: new "Pre-commit hooks (web/)" section covering all four.
Verified: eslint error blocks commit; clean commit passes; bad-format .cs
blocks (dotnet format ~6-7s scoped), good .cs passes; check:api drift and
a lint error each block `git push --dry-run`, clean state passes; missing
Co-Authored-By blocks commit-msg, present passes; non-web/.cs commits skip
lint/format. npm run lint clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Last SPA pre-work before deleting Blazor Libraries.razor (#91 phase b):
wire the shipped scanLibrary(id, deep) + scanCollections(family, id, deep)
clients (F9 API, #235) into LibrariesScreen so the SPA reaches parity with
Libraries.razor's four scan actions.
- Deep Scan Library button on each remote (Plex/Jellyfin/Emby) library row,
threading `deep` through the existing optimistic-pending/poll hook (quick +
deep share the per-library lock).
- External Collections section (quick + deep per remote source). Rows derive
client-side from getMediaSources(): the media-sources API handler already
filters each source's `libraries` to sync-enabled entries, so a remote
source with a non-empty libraries list is exactly GetExternalCollections's
Libraries.Any(ShouldSyncItems) filter — no new endpoint.
- useCollectionsScan hook: collections scans have no scan-status poll surface
(the endpoint is library-keyed; Blazor observed collections locks via
in-process IEntityLocker events), so pending is optimistic + timeout-bounded
(409 benign, 404/network surfaces the error). Follow-up #271 for a proper
collections status surface.
Pure SPA change (no backend/OpenAPI). Docs: blazor-route-parity.md §5 (SPA
affordance DONE), decisions.md (derive-vs-endpoint + optimistic-timeout).
Refs #91
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The prior fix set setItemsLoading/setItemsLoaded at the top of loadItems, but the
activeId effect calls loadItems synchronously → react-hooks/set-state-in-effect lint
error (Main Lint SPA, the real CI failure). Move the not-loaded/loading gating into
reloadAfterConflict (an event handler, lint-clean), which is exactly the 412 conflict-
reload path Codex's Medium-3 targeted; canEdit stays false through that reload window.
Behavior unchanged; the normal switch/initial-load paths (guarded separately by the
#242 dirty-guard) are left untouched. Local: lint clean, vitest 667, check:api no drift,
full dotnet solution test green.
Independent Codex review of #268 found two Blockers the fork missed + two Mediums:
- Blocker: replace PUTs returned the handler's item snapshot but re-queried the root
for the ETag separately, so a racing writer could pair stale items with a newer ETag
(silent overwrite). All four controllers now reload root-then-items (version-first,
fail-safe) and 404 when the root is gone between commit and reload — matching the
Block reference. Fixes the Blocker + the Medium '200 without ETag' case together.
- Blocker: PlaylistsScreen loaded items+root via Promise.all (concurrent), pairing a
stale name with the current ETag; now sequential (items-with-meta first, then root).
- Medium: SchedulesScreen loadItems now marks not-loaded/loading up front so canEdit is
false through the 412 conflict reload (no stale-draft edits lost).
Controller unit-test mocks updated to stub the new reload query. Full suite green
(ErsatzTV.Tests 1334, web 667, check:api no drift).
Codex-review Low: NotFound previously reached the catch-all by coincidence; a
future enum value would silently 404. Explicit arm + UnreachableException fallback
so an unmapped outcome fails loudly rather than mis-mapping to 404.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The new POST /api/media-sources/{plex|jellyfin|emby}/{id}/scan-collections
endpoints acquire a per-provider collections lock (409 if held) and hand the
single release to the ScannerService finally. But SchedulerService's periodic
collection scans were enqueued WITHOUT the lock, and ScannerService's finally
released the collections lock whenever held with no ownership check. A
scheduler-queued scan running while an API request held the lock cross-released
the API's lock (#250 bug class), letting a second API request get a spurious
202 instead of 409.
Fix (mirrors the SynchronizePlexLibraryByIdIfNeeded(Unlock: !networksFollow)
library-scan precedent):
- Add `bool Unlock = true` (4th positional param) to the three
Synchronize{Plex,Jellyfin,Emby}Collections records; default keeps the
controller + Libraries.razor call sites compiling and releasing on run.
- ScannerService: the three collection finallys now honor `request.Unlock`
(the concrete typed request is in scope in each method) so a batch member
with Unlock:false never releases a lock it doesn't own.
- SchedulerService: replace the unlocked per-source enqueue with a lock-once
per-provider batch — LockX Collections() once, enqueue each source with
Unlock:isLast (last message owns the release), compensating unlock in catch,
and SKIP the whole provider loop if the lock is already held. A naive
"lock-per-source, skip if held" would deterministically starve the 2nd+
source; lock-once-batch does not.
Tests (ErsatzTV.Tests/Services/): ScannerServiceCollectionLockTests drives the
real ScannerService read loop + real EntityLocker and asserts Unlock:false
leaves a held lock intact while Unlock:true releases (all three providers);
SchedulerServiceCollectionLockTests reflect-invokes ScanPlexMediaSources and
asserts it locks once + skips the enqueue when held, and hands the release to
the last message when acquired. Proven non-vacuous: reverting the Plex fix
fails exactly the three Plex tests.
No OpenAPI/v1.json change (internal channel-message record, not a DTO).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wire the frozen ETag/If-Match/412 recipe (Block reference implementation)
onto the Template and DecoTemplate aggregates:
- ReplaceTemplateItems / ReplaceDecoTemplateItems commands gain
Option<int> ExpectedVersion; ToCommand() on the request DTOs threads it
through from If-Match.
- Handlers introduce the version check as a standalone Either after
validation (never via Apply), bump Version unconditionally before
saving, and persist through SaveChangesWithConcurrencyGuard so a losing
writer maps to 412 instead of 500. DecoTemplate's post-commit playout
Reset enqueue now only runs after a successful save.
- TemplateViewModel / DecoTemplateViewModel carry Version (header-only,
not echoed in the response body), populated in Mapper.
- TemplateController / DecoTemplateController: GET items emits a strong
ETag of the root's version; PUT parses If-Match (400 on malformed),
threads the expected version into the command, and returns the new
ETag from the refreshed root on success. Both PUT actions now use the
handler's returned item list directly instead of re-querying items.
- SPA: templates.ts / decoTemplates.ts gain getXItemsWithMeta and an
If-Match-aware replaceX; TemplateEditor / DecoTemplateEditor hold the
ETag in a ref, read items-with-meta first on load, and open a
"changed elsewhere" ConfirmDialog on a 412 instead of navigating away.
Tests: new ReplaceTemplateItemsHandlerConcurrencyTests /
ReplaceDecoTemplateItemsHandlerConcurrencyTests mirror the Block
concurrency contract tests (stale/matching/absent If-Match, no-op bump,
racing-save 412, non-vacuous backstop). TemplateControllerTests /
DecoTemplateControllerTests gain ETag/If-Match/412 coverage.
TemplatesScreen.test.tsx / DecoTemplatesScreen.test.tsx gain a 412
conflict-dialog test mirroring BlocksScreen's.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fans the frozen ETag/If-Match/412 recipe (Block reference, #253) onto the
Playlist aggregate:
- ReplacePlaylistItems command carries ExpectedVersion; the handler runs
CheckVersion as a standalone Either after validation (so a stale write
survives as 412, not flattened to 422 by Apply/Join), bumps Version
unconditionally before saving, and persists via
SaveChangesWithConcurrencyGuard (EF concurrency-token backstop).
- PlaylistViewModel carries Version; the items GET sets a strong ETag and
the PUT parses If-Match, threads it into the command, and returns the
refreshed ETag on success (400 on a malformed If-Match).
- Sibling item-adding handlers (AddItemsToPlaylist, AddMovie/Episode/
Season/ShowToPlaylist) bump Version too, since they mutate the same
editor-visible item list.
- SPA: playlists.ts exposes getPlaylistItemsWithMeta and an
If-Match-aware updatePlaylist; PlaylistEditor holds the ETag in a ref,
round-trips it on save, and opens a "changed elsewhere" ConfirmDialog on
412 (mirrors BlockEditor).
Tests: new ReplacePlaylistItemsHandlerConcurrencyTests (stale/match/
force-write/no-op-bump/racing-save), new PlaylistController tests
(ETag on GET items, 400/412/thread-version/force-write on PUT), and a
vitest 412-conflict-dialog test for PlaylistsScreen. dotnet test:
1304/1304 green. web: npm run typecheck clean, npm run build clean,
vitest 664/664 green.
Ref #253 PR2.
Closes the two F9 Libraries.razor parity gaps and normalizes scan-show error
mapping to ProblemDetails.
TASK 1 — library-wide deep scan:
- QueueLibraryScanByLibraryId gains optional `bool DeepScan = false`; handler
threads it into ForceSynchronize{Plex,Jellyfin,Emby}LibraryById.
- POST /api/libraries/{id}/scan?deep=false binds it via [FromQuery].
TASK 2 — external-collections scan (new endpoints):
- POST /api/media-sources/{plex|jellyfin|emby}/{id}/scan-collections?deep=false
acquires the per-source collections lock (§3b: lock IS the running scan → 409),
enqueues Synchronize{X}Collections(id, ForceScan:true, deep) to the scanner
channel, returns 202; compensating-unlock on enqueue throw.
TASK 3 — scan-show normalization:
- New QueueShowScanResult enum; handler returns it instead of bool.
- POST /api/libraries/{id}/scan-show now maps 202/404/409/422 (all errors
ProblemDetails) instead of 200/404/400-anonymous-object.
- Updated the lone Blazor caller (TelevisionSeasonList.razor).
Tests: LibrariesController (scan deep=true, scan-show enum→status), the three
media-source controllers (scan-collections route/404/409/202/compensating-unlock),
and handler tests for both changed handlers (deep threading + show-scan outcomes).
Docs: api-conventions §3b exemplar + blazor-route-parity §5 F9 gate.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Slice C of the async-op contract normalization:
- channel reset (POST /api/channels/{channelNumber}/playout/reset) now
returns 202 Accepted (was 200 Ok) — it only queues a background rebuild
- reset-all (POST /api/playouts/reset-all) still 202 but now returns a
ResetAllPlayoutsResponseModel body reporting QueuedPlayoutIds /
SkippedLocked / SkippedUnsupported instead of silently swallowing skips;
handler returns a new ResetAllPlayoutsResult record
- single-playout GET (GET /api/playouts/{id}) now exposes IsLocked on
PlayoutResponseModel, set from IEntityLocker.IsPlayoutLocked mirroring
the list projection — gives a polling client the lock flag
Tests: channel reset asserts 202; reset-all asserts 202 + skipped-body
shape; single GET asserts IsLocked; new ResetAllPlayoutsHandlerTests
(in-memory SQLite) asserts locked/ExternalJson/None land in skipped lists
and eligible playouts in queued. docs/api-conventions.md §3a updated.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The global Trakt lock is acquired by SchedulerService.RefreshTraktLists /
MatchTraktLists (and TraktController) and released only when the *terminal*
message of a batch — the one carrying Unlock: true (list == traktLists.Last())
— is processed by WorkerService, whose handler (AddTraktListHandler /
MatchTraktListItemsHandler) calls IEntityLocker.UnlockTrakt() in a finally.
WorkerService.ExecuteAsync breaks out of the read loop on
stoppingToken.IsCancellationRequested (and exits on channel completion /
reader cancellation) BEFORE processing the next message. If shutdown lands
after a batch is enqueued but before its terminal Unlock: true message is
handled, UnlockTrakt() never runs and the in-memory Trakt lock leaks for the
rest of the process lifetime (subsequent Trakt operations 409 forever).
Fix (option a): make the batch-release loss-tolerant with a compensating
release in a finally around the read loop — if the Trakt lock is still held
when the worker stops, release it. Chosen over tracking pending ownership
(b) because the lock is a global singleton and WorkerService is its sole
batch-release site, so "held at shutdown" unambiguously means "the terminal
release was lost"; covers all three exit paths (break / channel completion /
cancellation) in one place. Same lock-lifecycle class as #231/#233/#234.
Regression test: WorkerServiceTests gates the first (non-terminal) batch
message on the stopping token, then StopAsync-cancels so the worker breaks
before the terminal Unlock: true message — asserts the lock is released and
the terminal message was never processed. Proven non-vacuous: inverting the
finally condition fails the test.
Backend-only; no controller/DTO/SPA/OpenAPI impact.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Slice A of the async-op API contract normalization.
MaintenanceController:
- EmptyTrash error path: was 500 text/plain (error.ToString()); now maps the
BaseError Left through ApiResults.ToErrorResult() -> 404 (NotFoundError) / 422
ProblemDetails. Success stays 200 OkResult. Added ProducesResponseType 200 + 422.
- CleanArtwork: fire-and-forget enqueue of DeleteOrphanedArtwork was a silent 200;
now returns 202 Accepted (AcceptedResult) since it queues background work.
Added ProducesResponseType 202. (Controller does not derive from ControllerBase,
so results are built directly as before.)
TroubleshootController.TroubleshootPlayback (GET|HEAD /api/troubleshoot/playback.m3u8):
- Two bare body-less NotFound() call sites conflated "not found" with "prepare/
playback failure". Both now return a ProblemDetails body:
* prepare-failure (result.IsLeft): mapped through error.ToErrorResult() -> 404 for
NotFoundError (unknown media item/channel) else 422 for a validation BaseError.
* terminal fall-through (prepare ok but no playable output): kept 404 with a
distinguishing ApiResults.NotFoundProblem(...) detail.
- Added ProducesResponseType 404 + 422 (409 already present).
Consumer check: the SPA (PlaybackTroubleshootingScreen) feeds the playback.m3u8 URL
straight to hls.js via HlsPlayer, which never inspects the HTTP status code — playback
state is surfaced via the separate /api/troubleshoot/playback/status poll. So the
404->422 split for the validation subcase is safe; no player code branches on the
status code.
Tests: MaintenanceControllerTests (200/422/202 + enqueue assertion),
TroubleshootControllerTests (prepare 404 NotFoundError, 422 validation). All green;
Api error-metadata/contract/security scans still pass.
Note: OpenAPI artifacts (v1.json / v1.d.ts) intentionally NOT regenerated here — the
orchestrator regenerates once after all #235 slices merge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Independent adversarial review of PR #266 found two sibling handlers with the
identical post-commit `_smartCollectionCache.Refresh(cancellationToken)` pattern
that the sweep missed (only UpdateSmartCollection was caught). Same audit#22 F4
class: a late client-disconnect after the commit lands would throw and leave the
in-memory smart-collection cache stale vs the committed DB. → CancellationToken.None.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Backend mutation-hardening cluster from the 2026-07-11 mutation-safety audit
sweep (adversarial-reviewer #22/#23), the parallel-safe backend-isolated slice.
audit#22 F4 — standardize post-commit enqueues on CancellationToken.None:
20 command handlers under MediaCollections/, ProgramSchedules/, Playouts/,
Channels/ threaded the request cancellationToken into work that runs AFTER
SaveChangesAsync commits (WriteAsync rebuild/refresh enqueues, mediator.Publish,
reindex, cache Refresh, and post-commit lookups that gate an enqueue). A late
client-disconnect then turns an already-durable commit into a thrown request AND
drops the side effect. Generalizes the #251 deco-handler fix. Excludes
BuildPlayoutHandler (worker/background token, not a client-disconnect token),
the config/FFmpeg multi-upsert handlers (partial-commit case, separate
follow-up), and response-projection reloads (correctly keep the request token).
audit#22 F2 — DeleteChannelHandler/DeletePlayoutHandler now delete the channel
guide {number}.xml through IFileSystem.File.Delete (observable under
MockFileSystem) and BEFORE the commit (a post-commit delete orphans the xml on a
crash; the xml is regenerable on demand, so pre-commit delete is the safe order).
audit#23 F4 — ReplacePlayoutAlternateScheduleItemsHandler rejects an empty item
list in the handler (not only the controller pre-guard) so a direct caller can't
trip the Max()-on-empty crash.
Docs: api-conventions.md §7a (post-commit token convention + boundaries),
decisions.md entry (rationale, sweep scope, #253 PR2-4 coordination note).
Tests: guide-cache-delete-through-FS for both delete handlers, empty-list guard.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Codex re-review of the fix commit confirmed both prior findings resolved and raised one
new Medium: RFC 7232 would 412 (not 400) a syntactically-valid but non-matching If-Match
(non-canonical "03", weak W/"3", tag lists, empty, overflow). Deferred to #197 (cold
contract pass) as #265 — fail-safe today (the mutation is rejected, never applied) and no
first-party client is affected. Records the deferral where the #253 fan-out will copy the
parser: a code comment in ConcurrencyHeaders + a note in api-conventions §7a.
Refs #253#265
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Codex independent review of #263 surfaced two defects the fork review missed:
- High — client load TOCTOU: BlockEditor read root metadata (getBlock) and items+ETag
(getBlockItemsWithMeta) concurrently, so a concurrent write landing between them (with
the items read resolving last) left a stale root paired with a current ETag → the save
silently overwrote the concurrent change with no 412. Fix: read items+ETag FIRST, then
the root metadata, so the captured ETag is never newer than the root version and any
inconsistency fails safe (save 412s → conflict dialog → reload).
- Medium — `ParseIfMatch` accepted non-canonical strong tags ("03", "+3", " 3 ") as
version 3. An ETag is opaque; only the exact emitted form is valid. Fix: canonical
decimal only (`NumberStyles.None` + no leading zeros) → else 400.
Tests: new `ConcurrencyHeadersTests` (canonical parse + padded/signed/whitespace/weak/
unquoted/list/overflow/empty → malformed); `ApiResultsTests` gains the 412 mapping case.
Existing BlocksScreen tests still green (load reordering is behavior-preserving for the
non-concurrent path).
Refs #253
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Resolve the two SHOULD-FIX gate findings from the #91 cold review by making the
removal plan address them explicitly instead of clearing the gate by omission.
Pages (verified in code, not assumed) — OIDC's AuthorizeFolder("/") gates only the
Blazor _Host Razor Page; /app (SPA) and /api/* were already unauthenticated since
phase (a); /iptv JWT + API-key filters are independent of Blazor and survive
removal. Sign-off: no capability lost, no NEW exposure beyond phase (a); real
SPA/API auth deferred to #197. Recorded in docs/decisions.md.
(cut at removal time on the pre-deletion main commit — not a v* tag, no release
build) + the restore path (checkout+build+pin test container, or revert the merge).
Recorded in docs/decisions.md.
Both fold into a new "Section 5 — Removal execution runbook" in blazor-route-parity.md
so the (gated) removal PR has an ordered checklist. Docs-only; no code change.
refs #205#206#91
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the shared optimistic-concurrency contract so a stale second tab can no longer
silently overwrite a fresher edit. PR1 lands the infra + the Block reference aggregate;
PRs 2–4 fan the same recipe across the other 8 roots (design: #253#issuecomment-8472).
Contract
- `IVersionedAggregate` (`int Version`) on all 9 replace-all roots (ProgramSchedule,
Block, Template, DecoTemplate, Playlist, Collection, Playout, MultiCollection,
RerunCollection), EF-mapped `.IsConcurrencyToken()`; one dual-provider migration
`AddAggregateVersions` (nullable:false, default 0).
- Strong `ETag` of `Version` on the aggregate GET; `If-Match` on the PUT; mismatch →
412 (distinct from the §3a 409 build-lock guard). Successful PUT returns the new ETag.
- `PreconditionFailedError : BaseError` → 412 in `ApiResults.ToErrorResult`;
`ConcurrencyHeaders.ParseIfMatch/SetETag`; malformed If-Match → 400; `*`/absent =
Phase-1 force-write.
Block reference wiring
- Handler: standalone `Either` via `CheckVersion` AFTER validation (never through
`Apply`, which Join()-flattens the subtype to 422), unconditional `Version++`,
`SaveChangesWithConcurrencyGuard` backstop (DbUpdateConcurrencyException → 412).
- `BlockViewModel.Version` (header-only, not echoed in the body); controller sets the
ETag on GET items and on the successful PUT.
- SPA: `client.requestWithMeta` seam; `blocks.getBlockItemsWithMeta` + `replaceBlock`
If-Match/ETag round-trip; `BlockEditor` holds the ETag, sends If-Match, and on 412
opens a blocking "changed elsewhere — reload" dialog.
Tests
- Handler contract tests: stale-If-Match → 412 (no mutation), matching/absent → success
+ bump, no-op save still bumps, and a two-context racing save → 412; proven
non-vacuous (drop `.IsConcurrencyToken()` → the race test fails).
- Controller tests: malformed If-Match → 400, If-Match threaded to the command, ETag on
GET/PUT, 412 passthrough. SPA: requestWithMeta ETag, replaceBlock If-Match, 412 dialog.
Docs: api-conventions §7a, spa-conventions §4a, domain-model glossary, decisions log.
Refs #253
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fork adversarial review nit: Map_Keys_Should_Not_Begin_With_Forbidden_Prefix
covered only Tier-1 Map keys, not the Tier-2 PatternRule templates. That guard
invariant is the load-bearing protection for the un-prefix-guarded /api|/artwork|
/docs|/openapi surface, so make it self-enforcing over ALL rules — a future
prefix-violating template now fails the test instead of slipping through.
Exposes internal LegacyUiRedirects.PatternTemplates (InternalsVisibleTo already
set for ErsatzTV.Tests); stores the raw template on PatternRule.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Extend LegacyUiRedirects from an exact-match dictionary to a two-tier matcher:
Tier 1 keeps the exact Map (now 52 entries incl. the ?kind= browse roots),
Tier 2 adds 36 ordered segment-template PatternRules for id-carrying routes.
{id} is a strict positive integer (non-int/0/neg/overflow falls through), which
also makes the rule set collision-free by construction. New AppendQueryString
helper merges the incoming query into ?kind= targets with '&' (kills the
double-'?' bug); one-line Startup change keeps the redirect GET/HEAD-only 302
before UseRouting.
Completes phase-(a) Step 1 for every PARITY-OK route (#91 phase b); the
catch-all fallback replacing MapFallbackToPage stays with the removal PR.
/media/sources/* (#202) and /system/health remain deliberately un-redirected.
fixes#204
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Build the Remote media-source SPA screens over the S5 foundation, replacing
the MediaSourceEditorPlaceholder for the plex/jellyfin/emby dispatch branches
only (Local branches left for S6a):
- PlexSourceScreen: pin-flow sign-in / fix-credentials / sign-out with the
§C1 poll state machine — polls GET /api/media-sources/plex every 2s up to
150s and keeps polling while authorized-but-locked ("finalizing"); the
terminal success is the lock releasing. Popup-blocked fallback link. Server
table (Refresh disabled while locked / Edit Libraries / Edit Path
Replacements) + sign-out content-removal confirm dialog.
- RemoteSourceScreen (shared Jellyfin/Emby): connect / edit-connection /
disconnect (warning dialog) + server table.
- RemoteConnectionEditScreen (shared): secure key affordance (§C3/finding 1)
— address prefilled, "leave blank to keep" when hasApiKey, required on first
connect; stored key never rendered or requested.
- RemoteLibrariesEditScreen (shared): client-side sortable Name + MediaKind
columns, per-library sync Switch, one Save; draft keyed by (name,mediaKind)
not id, refetch after save (ids change on disable, §C4a).
- PathReplacementsEditScreen (shared): row list + selected-row edit form,
add/remove, one Save; both fields required; family remote-path column label.
All editors use the ChannelEditScreen draft/save model + a shared useDirtyGuard
(registerNavigationGuard + beforeunload), Save gated !valid||!dirty||saving,
draft retained on 422/network, destructive actions gated on saving, 409 →
refetch. Colocated tests cover the poll (waiting→finalizing→success asserting
it does NOT stop at authorized&locked, timeout, budget-exhausted), the secure
key affordance, sortable columns, draft-retained-on-422, dirty-guard veto, and
the disconnect/sign-out dialogs.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the Local library create/edit editor (create at /app/libraries/local/new,
edit at /app/libraries/local/{id}) wired into the S5-built LibrariesRouteScreen
dispatch switch, replacing MediaSourceEditorPlaceholder for the local-new and
local-edit sub-routes only. Remote (Plex/Jellyfin/Emby) branches are untouched
(S6b).
- Name (required) + Media Kind (create-only, disabled+annotated on edit)
- Add Path: path-exists pre-check (L7) + in-draft duplicate detection
(mediaSources/paths.ts normalizePath)
- Delete path: draft-local removal with a media-item-count confirm dialog
- Move path: dialog filtered to same-MediaKind libraries excluding the source,
including "(New Library)" which composes createLocalLibrary + moveLocalLibraryPath
(surfaces the error and leaves the new empty library on a failed move, matching
Blazor); gated on !dirty to avoid clobbering unsaved edits with the post-move
refetch
- Draft/saved model with explicit Save (POST L3 / PUT L4), draft retained on
422/network error, dirty-guard (registerNavigationGuard + beforeunload)
- Delete library (L5) with a media-item-count confirm; 409 refetches detail
Extended the existing App.test.tsx App-owned-popstate regression test (design
§D.2) to exercise the real screen's dirty guard instead of a manually-armed
stand-in, now that S6a has landed the editor it was stubbing out for.
Verification (web/): vitest (632 passed), eslint clean, tsc -b + vite build
clean, check:api reports no drift.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
New JellyfinMediaSourcesController (/api/media-sources/jellyfin, J1-J9) and
EmbyMediaSourcesController (/api/media-sources/emby, E1-E9), wrapping the
existing Jellyfin/Emby MediatR commands per the #202 design doc §A.3/§A.4.
Secure connection contract (§C3/§B, finding 1): the connection GET returns
only { address, hasApiKey } — the API key never crosses the wire. The PUT
retains the existing key when the incoming key is blank, sets a new one when
non-blank, and 422s "API key is required" on a blank first connect.
Finding 7 (lock-release discipline): DisconnectJellyfinHandler and
DisconnectEmbyHandler now wrap their work in try/finally so a throw from any
awaited dependency (repo delete, search-index commit, secret store) still
releases the family lock instead of wedging every future disconnect at 409.
Findings 2c/8 (path-replacement cross-source guard): UpdateJellyfinPathReplacementsHandler
and UpdateEmbyPathReplacementsHandler now reject, before any write, an incoming
positive Id that isn't owned by the route's media source, a null item, or a
blank RemotePath/LocalPath — all 422 with no partial mutation. Defense-in-depth
repo fix: the Jellyfin/Emby path-replacement UPDATE SQL in MediaSourceRepository
now scopes by {Jellyfin,Emby}MediaSourceId (was previously unscoped by Id alone,
allowing a PUT to one source to silently overwrite another source's row). The
Plex path-replacement method (~line 397) is untouched — that's slice S2's file.
Library preferences (§C4a): the controller validates the incoming id set
against the source's known libraries (reject foreign ids, require full
coverage, no Id=0) before dispatch, then — for §C7 — LockLibrary + enqueues
the SynchronizeXLibraries/SynchronizeXLibraryByIdIfNeeded pair per enabled
library (compensating unlock if the enqueue throws), and returns the reloaded
list (ids are not stable across a disable).
404s on id-taking endpoints come from a controller pre-check (GetXMediaSourceById
is None), not a handler NotFoundError, since Either.Apply/ToEitherAsync join any
NotFoundError into a flat 422 (finding 9).
Tests: controller route/404/409/422 tests for both families; disconnect
fault-injection tests proving the lock releases even when a dependency throws;
path-replacement handler tests for cross-source-id/blank/null-item rejection
and correct add/update/delete merge; a repository-level test proving the SQL
fix stops a same-family cross-source path-replacement overwrite.
No new commands, no DB migration, no OpenAPI regen (gated until S1-S3 merge
per the design doc's build-slice plan).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds LocalLibrariesController (L1-L7: list/get/create/update/delete/move-path/
path-exists) wrapping the existing local-library MediatR commands, mapping to
the shared S0 response DTOs. Per design #202 §A.1/§C5/§C6:
- 404 for L4/L5/L6 comes from a controller pre-check (GetLocalLibraryById is
None), not the handler -- .Apply/.ToEitherAsync both .Join() a NotFoundError
into a plain 422, so relying on the handler would be dead code. This is
check-then-act; a delete racing the pre-check falls through to the handler's
422, documented in the controller.
- L4/L5 409 via IEntityLocker.IsLibraryLocked(id); L6 resolves the source
library from the path id (new ILibraryRepository.GetLibraryIdForPath) before
its own lock check.
- MoveLocalLibraryPathHandler gains same-MediaKind and different-library
validation (finding 3) -- Blazor only filtered these client-side in the move
dialog, so an API/MCP client could bypass them.
- CreateLocalLibraryHandler/UpdateLocalLibraryHandler gain a shared
NewPathsMustExist validation (LocalLibraryHandlerBase) that Directory.Exists-
checks only new paths (Id < 1); existing rows stay exempt so an unmounted
share doesn't block a rename. L7 (path-exists) is a controller-local
IFileSystem check with no command.
Tests: controller route/mediator-arg tests incl. 404-pre-check vs
fall-through-422 and 409-lock cases; handler tests for the move-path
cross-kind/same-library 422s, new-path 422 (missing/mixed), and a lossless
round-trip proving local paths are identified by normalized path string, not
id. Full solution test suite (Scanner/Core/Architecture/Tests/Infrastructure)
green, 0 regressions.
Deviations: none from the S1 slice description. Did not touch
MediaSourceRepository.cs or any Plex/Jellyfin/Emby file (S2/S3 scope). Did not
run the OpenAPI regen scripts (separate gate after S1-S3 merge per design §E).
New PlexMediaSourcesController (/api/media-sources/plex) P1-P8 wrapping
existing MediatR commands: state GET, pin-flow, sign-out, per-server
libraries/path-replacements GET+PUT, and refresh — VMs projected to the
S0 shared DTOs, ApiResults mapping, 404 controller pre-checks, #215-style
409 lock guards, [EndpointGroupName("general")].
Lock-lifecycle hardening (the tricky part):
- TryCompletePlexPinFlowHandler now releases the Plex lock ONLY on its
non-handoff exits (timeout-throw, poll exception, enqueue exception, the
dead return-false) via try/catch — NOT an unconditional finally. On
success the lock is handed off to SynchronizePlexMediaSources (the sole
releaser after discovery); a finally would double-release and release
before discovery, re-opening the finding-5 poll race. Fixes the latent
leak where an abandoned pin flow wedged Plex locked until restart.
- StartPlexPinFlow controller compensates UnlockPlex on the Left branch AND
any thrown dispatch/enqueue; only the Right/200 path holds the lock.
- SignOutOfPlexHandler wraps its work in try/finally { UnlockPlex() } — a
terminal handler with no handoff, so unconditional release is correct.
- Post-save library sync enqueues SynchronizePlexLibraryByIdIfNeeded
(Unlock:false) then SynchronizePlexNetworks (Unlock:true) — one lock, one
release on the last message, compensating-unlock if the 2nd enqueue throws
(corrects the Blazor Unlock-ordering bug, finding 6).
Data-integrity hardening:
- UpdatePlexPathReplacementsHandler rejects (422, no mutation) any positive
Id not owned by the route source, blank RemotePath/LocalPath, and null
list/items (findings 2c/8).
- MediaSourceRepository Plex path-replacement UPDATE gains
AND PlexMediaSourceId = @id (Jellyfin/Emby untouched — slice S3).
- ReplaceLibraryPreferences controller validates the id set against the
source's libraries (rejects unowned + Id=0), returns the reloaded list
(ids change on disable).
Tests (NUnit/Shouldly/NSubstitute), 34 new, all green: pin-flow lock
released on thrown-cancellation/poll-throw/enqueue-throw AND held on
success (no double-release); sign-out finally-release under a throwing
dependency; cross-source/nonblank/null path-replacement 422s; library-prefs
id-not-owned 422; post-save enqueue exact messages + Unlock flags; full
controller route/404/409/422 coverage.
Refs #202
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move the Libraries domain verbatim out of web/src/App.tsx into
web/src/screens/LibrariesScreen.tsx (zero-prop, self-sufficient), mirroring the
ChannelsScreen extraction (#244). Pure structural move: no API, route, CSS, or
visual change. App.tsx retains only the import + the <LibrariesScreen /> dispatch.
- 10 symbols moved (LibrariesLoadingState -> sourceLastScanLabel); App.tsx's
formatDateTime is inlined into the moved screen so it has no import back into
App.tsx (behavior-identical), matching the Channels precedent.
- Libraries behavior tests moved to a colocated LibrariesScreen.test.tsx with its
own scoped fetch mock (renders <LibrariesScreen /> directly); App.test.tsx keeps
one nav-smoke test for the route.
- Pruned now-dead App.tsx imports (Server, MonitorPlay, Music, FileImage, Folder,
HardDrive icons; useLibrariesScreenQuery, LibraryScanStatus/MediaSource/
MediaSourceLibrary types) and the now-dead runPollTick test helper (its doc
comment named it Libraries-specific).
- Disabled "Add Source"/gear/"Scan All" affordances are unchanged (wired in later
S5/S6 slices, not here).
Verified: web vitest 584 passed, eslint clean, tsc/vite build clean, check:api no
drift.
refs #202
Response DTOs (ErsatzTV.Core/Api/MediaSources) and request DTOs
(ErsatzTV/Controllers/Api/Requests) per the #202 design doc §B — the
shared shapes that backend slices S1 (local libraries), S2 (Plex), and
S3 (Jellyfin/Emby) will consume. No controllers or handler changes;
DTOs are unused so far.
Notable: RemoteConnectionResponseModel deliberately never carries the
raw API key (secure connection contract); SaveRemoteConnectionRequest's
To{Jellyfin,Emby}Command(existingApiKey) retains the existing key when
the incoming ApiKey is blank/omitted.
No conventions changed; nothing to update in docs/api-conventions.md.
Independent Codex review of the fix diff surfaced two real findings the fork pass missed:
- Medium: the #251 affected-playout QUERIES in ReplaceDecoTemplateItemsHandler and
UpdateDecoHandler still ran on the request `cancellationToken`, so a cancellation
landing after SaveChanges committed but before those queries executed would throw
before the CancellationToken.None enqueue — the edit committed but no playout Reset,
re-opening the stale-content bug in that window. Run the entire post-commit
invalidation (queries + enqueue) on CancellationToken.None so the side effect can't
be half-aborted once the data has changed.
- Low: UpdateDefaultDecoHandler enqueued a Reset for request.PlayoutId even when
ExecuteUpdateAsync matched 0 rows (nonexistent playout), creating a background build
request for an id that isn't there. Guard the enqueue on rows-updated > 0 so the
enqueued set equals the affected set. Added a regression test.
Also corrected the ReplaceProgramScheduleItemsHandler comments: the schedule-item
hierarchy is TPT (table-per-type), not TPH — the SetValues reconcile is safe either way
(same-runtime-type guard; no discriminator to corrupt), Codex confirmed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
#252: ReplaceProgramScheduleItems deleted and re-inserted every item on every
save (even a no-op PUT-back), and PlayoutScheduleItemFillGroupIndex.ProgramScheduleItemId
is OnDelete(Cascade) — so every schedule save silently wiped persisted
fill-group/shuffle enumerator progression for all playouts using the schedule.
Switch to a positional in-place reconcile: for a same-typed slot, copy scalars via
CurrentValues.SetValues (BuildItem stays the single source of item construction, so
no field is dropped) and rebuild the watermark/graphics join rows, keeping the item
id — and with it the fill-group index. Subtype change / surplus falls back to
delete+insert for that slot only. The request DTO carries no stable item id, so
position is the only key here; true content-aware stable identity is deferred to the
shared concurrency/round-trip contract in #253.
#251: deco / deco-template CONTENT edits (and default-deco assignment) only take
effect on a playout Reset build — deco/break/default-filler content is applied during
Reset, a Continue keeps the frozen filler items, and BlockKey change-detection has no
deco dimension to self-heal. The editors enqueued nothing (a commented-out TODO in
ReplaceDecoTemplateItemsHandler), so filler/break content stayed stale indefinitely
until a manual Reset. Enqueue BuildPlayout(Reset) for exactly the affected playouts:
- ReplaceDecoTemplateItemsHandler: playouts via PlayoutTemplate.DecoTemplateId
- UpdateDecoHandler: playouts via Playout.DecoId and via deco-template items
- UpdateDefaultDecoHandler: the reassigned playout (adjacent same-class fix)
Post-commit enqueues use CancellationToken.None (audit #22 policy).
Tests: DecoInvalidationTests + ReplaceProgramScheduleItemsReconcileTests, each proven
non-vacuous against a negative control (inverted the primitive, verified 0 CS errors so
the --no-build run used a fresh dll). Full ErsatzTV.Tests suite green (1067).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Independent Codex review of PR #250 caught a cross-release the fork missed: the
outer catch in PrepareTroubleshootingPlaybackHandler released the troubleshooting
lock unconditionally, so an exception BEFORE this caller acquired it (e.g. request
cancellation during validation, or a DB error) would release a lock held by another
session.
Fix: track ownership with a Handle-scoped `lockAcquired` flag and gate the catch on
it. Acquisition for the media-item path moves out of GetProcess up into Handle (after
validation succeeds), so one place owns the full lifecycle: acquire -> on Left release
-> on success hand off to StartTroubleshootingPlayback -> on any exception release only
if we own it. GetProcess is now lock-free.
Test: Handle_Should_Not_Release_Lock_It_Never_Acquired (exception before acquisition ->
no Unlock, no cross-release), verified non-vacuous against an inverted-condition
negative control. Strengthened the empty-path test to also assert acquisition happened.
Deferred (noted for close comment): worker-dispatch-failure lock leak and
BuildPlayout silent-skip observability are pre-existing / sanctioned -> #235.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Consume the EntityLocker ownership contract (#231/#241: Lock* returns true iff
this caller won the slot) at three lock-leak sites surfaced by adversarial-reviewer#20.
#233 (F3) — troubleshooting playback:
- PrepareTroubleshootingPlaybackHandler: both lock sites now acquire via
`if (!LockTroubleshootingPlayback())` (kills the check-then-set TOCTOU) and the
empty-media-path Left return releases the lock it acquired — previously it leaked,
wedging the status endpoint at "running" forever for a file gone from disk.
- TroubleshootController.TroubleshootPlayback: lock conflict is now 409 ProblemDetails
(was a bare 404, indistinguishable from a bad id); the Prepare-success -> enqueue
window releases the lock if we never hand off to StartTroubleshootingPlayback.
#234 (F4 + F5.2) — playout builds:
- ExtractEmbeddedSubtitlesHandler: try/finally releases exactly the playouts it
locked, on every terminal path (cancellation early-return, swallowed cancellation,
any exception) — no more permanent leaks after cancelled mid-extraction, and no
cross-release of playouts held by someone else.
- BuildPlayoutHandler: skips (logs, returns Right) when LockPlayout returns false
instead of building unlocked and cross-releasing the other owner's lock in finally.
Tests: handler-level release-discipline tests (Prepare empty-path, Extract
cancellation + no-cross-release, BuildPlayout skip + finally-release) via the
InMemoryTvContext harness, a TroubleshootController 409 test, and OpenApi contract
cases for the m3u8 endpoint's 409. OpenAPI regenerated. All non-vacuous (F3 verified
against a negative control).
Fixes#233, #234
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move the Channels domain verbatim out of web/src/App.tsx into
web/src/screens/ChannelsScreen.tsx (zero-prop, self-sufficient, mirroring the
SchedulesScreen extraction). Pure structural move: no API, route, CSS, or
visual change. App.tsx retains only the import + the <ChannelsScreen /> dispatch.
- 14 symbols moved (ChannelViewFilter → ChannelTableRow); the Dashboard-owned
progressFromNowPlaying is inlined into the moved progressFromChannelState so
the screen has no import back into App.tsx (behavior-identical).
- 12 Channels behavior tests moved to a colocated ChannelsScreen.test.tsx with
its own scoped fetch mock (renders <ChannelsScreen /> directly, no
mockDashboardApi); App.test.tsx keeps one nav-smoke test for the route.
- Pruned 12 now-dead App.tsx imports; shared symbols (ChannelState,
messageFromError, ApiError, useChannelsQuery) verified still used and kept.
- Docs: spa-conventions §6 (extracted-screen own-fetch-mock convention),
decisions.md (single-file rationale; no web/src/channels/ sibling dir, unlike
Schedules; inlined helper; #238 deferral).
Verified: web vitest 587 passed, eslint clean, tsc/vite build clean,
check:api no drift. #212 empty-lineup bare-create success+failure coverage
preserved. #238 TopBar dead-button left as-is (its owned bug; shell redesign
is epic phase 4 / #247).
refs #244#243
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Codex re-review of the prior fix commit found the Edit gate closed the exact
repro but two paths remained. One is reachable: Delete was the only
schedule-switch path that guardedSwitch's `saving` guard didn't cover — deleting
mid-save runs applySwitch to the next schedule while the in-flight items PUT is
still outstanding, and that PUT's completion handler then overwrites the next
schedule's draft with the deleted schedule's response. Delete is now
`disabled={saving}`, consistent with the Select, Edit, and guardedSwitch.
Regression: the deferred-PUT test now also asserts Delete is disabled in-flight
and re-enables after the save settles.
Also widens the test mock's onRequest return type to `Response | Promise<Response>
| null` (removes the `as unknown as Response` cast — a test-only type hole the
re-review flagged).
Deferred to #248: the other residual path (properties dialog not focus-trapped, so
keyboard focus can escape to underlying Add/Save mid-save) is a pre-existing,
cross-cutting overlay.tsx a11y gap affecting all dialogs — out of scope for this
targeted blocker fix.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two findings from the PR #242 adversarial review (Codex + Claude fork):
1. ChannelsEmptyState swallowed mutationError: on the fresh-install path #212
targets, a bare-create can 4xx (e.g. no default ffmpeg profile), but the empty
branch never rendered the error alert the non-empty screen shows — the user saw
only a spinner re-enable. The empty state now renders the same ctv-channels-error
alert. Regression: App.test.tsx asserts the error surfaces + no navigation on an
empty lineup.
2. SchedulesScreen Edit button was not gated on `saving` (Codex): during an items
save PUT the draft is still dirty, so a discard-to-open → shuffle-flip could let
the in-flight PUT resolve AFTER the shuffle reload and clobber the normalized
draft with the pre-shuffle body. Edit is now disabled while saving (consistent
with the schedule Select). Regression: a deferred PUT proves Edit is disabled
in-flight and re-enables once the save settles.
Deferred as nits (both reviews rate low): vetoed navigateToPath leaves two stray
history entries (cosmetic, rare — fixing means refactoring central nav); rapid
double-Back is best-effort (inherent popstate non-cancellability).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adversarial review (fork + Codex, both flagged) of PR #241:
- SPA (both reviewers): removing PENDING_GRACE_TICKS wholesale reintroduced a
stuck scan button. A 202'd scan that finishes between 10s polls (short/empty
library) is never observed active, so its optimistic pending flag wedged the
button disabled until reload. Restore a BOUNDED grace net (pruneGraceExpiredPending)
— re-scoped honestly: it absorbs the inherent queue->observed-active lag and the
fast-completion race, NOT the removed lying-200 compensation (the POST now returns
409/404/422 honestly). Bounds pending to PENDING_GRACE_TICKS * pollMs (~30s).
- SPA 409 (Codex): on "already scanning" the button was cleared+reconciled, but a
scan-status still lagging the in-progress scan re-enabled the button and let the
user fire repeated 409s. Keep the pending flag on 409 (no toast) so the button
stays disabled; polling promotes or expires it.
- Scheduler (Codex): the Plex-Shows tail-token batch and the local/Jellyfin/Emby
scan enqueues had no compensating unlock — a WriteAsync failure after LockLibrary
(cancellation on shutdown) stranded the library lock. Wrap each acquired-lock
enqueue in try/catch → UnlockLibrary → rethrow (the Plex catch covers both writes,
since the library message carries Unlock: false and the un-enqueued networks
message was the sole releaser).
Tests: two new App.test.tsx cases — grace-window expiry re-enables the button, and
409 keeps it disabled through the queue->active lag.
Ref #232.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Editing a schedule's properties to change shuffleScheduleItems for the active
schedule left the open items draft stale: hidden Fixed/Flood start values could
be saved back and the inspector kept offering start-type controls the schedule
no longer supports. onScheduleSaved now detects a shuffle flip on the active
schedule and reloads the items via GET so the server's EnforceProperties
re-normalizes the draft. The schedule-level flags already refresh from the save
response (setBoot maps `saved` into the list).
Decision: chose "block opening the edit dialog while the item draft is dirty"
(confirm-to-discard, guardedSwitch semantics) over confirm-at-reload — the
smaller fully-consistent change. It guarantees the properties editor only ever
opens over a clean baseline draft, so the post-save reload is lossless and
avoids the awkward state where a cancelled discard leaves a now-shuffled
schedule holding Fixed values.
Regression: SchedulesScreen.test.tsx — Fixed item on a non-shuffled schedule →
edit properties to shuffle=true → a fresh items GET fires, the Fixed option is
gone, and a subsequent Save's PUT carries no Fixed startType.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
canLeaveCurrentScreen() was only consulted in App's navigate() (sidebar/nav
clicks); browser Back/Forward switched screens unguarded. A popstate can't be
cancelled, so App's popstate handler now, on a vetoed guard, re-pushes the
pre-pop path (tracked in currentPathRef, updated on every approved navigation)
and leaves activeRoute untouched — undoing the browser's URL change. The same
handler covers the synthetic pop navigateToPath() dispatches. Re-pushing is
safe: only one screen is mounted at a time and the guard-registering screen
(schedules) owns no internal popstate listener, so no sub-path screen's
pathname state can desync. Effect cleanup keeps StrictMode double-mount from
double-registering.
Docs: spa-conventions §8 rewritten from navigate-only to describe popstate
coverage. Regression: App.test.tsx dirties the schedules draft, simulates
popstate → confirm called; cancel keeps route + re-pushes path; accept
switches route and unmounts the draft.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
ChannelsScreen returned ChannelsEmptyState before the action bar that owns
"New blank channel", so a fresh install could never create its first channel.
The empty state now offers both create paths (bare-create + ChannelBuilder),
reusing the exact createBlankChannel handler (number = max+1 → 1 on empty,
group "ErsatzTV", default ffmpeg profile).
Regression: App.test.tsx bare-creates from a [] lineup, asserts the POST
payload (number "1") + navigation to /app/edit-channel/{id}.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Scan queue handler now returns a QueueLibraryScanResult enum
(Queued|NotFound|SyncDisabled|AlreadyScanning) instead of a lying bool;
LibrariesController.ScanLibrary maps them to 202/404/422/409 with ProblemDetails.
Guard the lock->enqueue with the EnqueueWithTraktLock compensating-unlock pattern.
ScannerService now releases every library/collection lock in a finally so a handler
exception can't leak the lock. Plex "Shows" scheduler batch (one lock, two messages)
now has only the trailing SynchronizePlexNetworks carry the single release
(Unlock flag), mirroring the scheduler Trakt tail-token precedent.
Guard the other lock->enqueue producers (Create/UpdateLocalLibrary, UpdateTraktList)
with compensating unlock. SPA drops the PENDING_GRACE_TICKS heuristic now that the
POST reports 202/409/404/422 directly: 202 -> pending+poll, 409 -> reconcile (no
error toast), 404/422 -> surface error.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The six plain-bool lock flags (Plex, Trakt, Emby/Jellyfin/Plex collections,
troubleshooting playback) used a non-atomic check-then-set, so two concurrent
Lock* callers could both win. Convert them to int flags mutated only via
Interlocked.CompareExchange, so the caller that wins the 0->1 transition is the
sole owner and the only one that fires the change event. The three
ConcurrentDictionary-backed kinds (Library/Playout/RemoteMediaSource) were
already atomic; drop their redundant ContainsKey pre-checks.
Define the ownership contract (tokenless single-owner discipline, no interface
change) on IEntityLocker and in docs/decisions.md: a true from Lock* confers
ownership of exactly one release; Unlock* on an unlocked slot returns false,
fires no event, and logs a warning (the double-release / non-owner tripwire).
Adds EntityLockerTests (real locker, parallel-caller races) proving exactly one
winner per kind, one-releaser-per-slot, and event-fires-once-per-transition.
Ref #231. Scan-lifecycle call-site fixes that consume this contract land in the
same PR (#232); the BuildPlayout/subtitle finally-gating is #234.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Closes the reviewer-verification gap from adversarial-reviewer#18: the
locked-path 409 guard tests already existed for Delete, EraseItems,
EraseItemsAndHistory, Update, and UpdateDefaultDeco, but were missing for
PlayoutController.ReplaceAlternateSchedules and .ReplaceTemplates even
though their production guards (entityLocker.IsPlayoutLocked) were in
place. Added two tests mirroring the existing pattern exactly (409
ConflictObjectResult + ProblemDetails + DidNotReceive() on the mediator
command). ApiErrorResponseMetadataTests and OpenApiErrorResponseContractTests
already carried 409 rows for both PUT endpoints, so no changes were needed
there.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- build & docs-reminder jobs -> runs-on: small (dedicated small-jobs runner,
server-management#574). Gitea dispatches a job as a runner task even when
its 'if' skips it; the PR-run skip of 'Build & push image' waited up to
31 min for an ubuntu-latest slot, stalling every PR run.
- concurrency scoped per event+ref with cancel-in-progress for PRs: runs
parallelize across PRs, superseded syncs auto-cancel. Previously one global
group serialized ALL runs (single-runner relic). Main/tag builds still
serialize within their ref; don't push main + v* tag simultaneously
(shared :buildcache / smoke container) — tag after main is green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
timothy
