414 Commits

Include renames

Author	SHA1	Message	Date
timothyandClaude Opus 4.8	d2d678aae8	test/docs(404): add unrelated-edit round-trip test; correct MultiCollection-only rationale ... Review + live-E2E follow-ups (no behavior change): - Add a test proving the canonical silent-reset trap directly: a rename-only save round-trips both weights untouched (the prior test only edited the weight it asserted). Cold review nit. - Correct the rationale in itemRules.ts + decisions.md: WeightedShuffle is MultiCollection-only in the SPA for *meaningfulness* (per-source weights need 2+ sources), NOT because the classic write path rejects it — live-E2E confirmed the classic engine ACCEPTS it on a plain Collection (200) and degrades to fair-share. The rejection is on the separate playlist/block write paths, whose editors keep their own order lists and already omit it. Live-E2E (real API): weighted multi-collection create + read round-trips weights; rename-only PUT preserves them (no silent reset); WeightedShuffle persists on a classic MultiCollection schedule item. Ratio itself is pinned by the existing PlayoutBuildGoldenTests.Classic_weighted (3:1), untouched by this SPA change. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-19 11:39:29 +02:00
timothyandClaude Opus 4.8	d639946b5c	feat(404): weighted-distribution SPA — per-source weight inputs + WeightedShuffle order ... The UI half of #70 (backend + API shipped in PR #402). Pure SPA + docs — no new endpoint/DTO (`weight` was already on MultiCollectionItemRequest/Response, and `WeightedShuffle` already in the PlaybackOrder enum). - Multi-collection editor (`/app/multi-collections`): per-source weight input (1..1000, mirroring the API validator), a computed % share (3:1 shows 75/25), and a "Reset to fair share" action (fair-share = all weights 1, decisions.md 2026-07-17 — a reset, not a separate mode). Weight round-trips through the draft (read in itemsFromMultiCollection, written in toItemRequest) so the replace-all PUT never silently resets it. - Classic schedule editor: `WeightedShuffle` offered as a Playback Order ONLY for MultiCollection sources (itemRules `MULTI_COLLECTION_ORDERS`) — it needs per-source weights and the write path rejects it elsewhere. Excluded from fillWithGroup like ShuffleInOrder (PlayoutBuilder schedules fill-groups per-group, incompatible with whole-collection weighted share). - `Input` gained min/max/inputMode/onBlur passthroughs for bounded numeric fields (reusable by #425's weight UI); weight held as a string for smooth editing, clamped on blur and at save so an out-of-range value never 400s. - Docs: domain-model + spa-conventions (replace-all round-trip trap, bounded numeric input pattern) + decisions.md entry. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-19 11:28:12 +02:00
timothyandClaude Opus 4.8	708e602197	feat(queue): deterministic scripts/select-queue.sh — stop re-deriving the selector by hand ... The kickoff dispatches a cheap model to rank the backlog, and the lore then made the orchestrator re-derive its mechanical claims (deps, milestone tiering, priority order, in-progress state) because a small model kept getting them wrong. This pays that tax down: the mechanical checks are now CODE — GET /dependencies exclusion, LOCAL .milestone.state/review/priority tiering (never the no-op ?milestones= filter), (tier,priority,issue#) ordering, in-progress/parked/PR exclusion — leaving only two JUDGMENT flags (CLAIM?, UMBRELLA?) for a human/model to resolve. - scripts/select-queue.sh: fail-open on no-creds/unreachable; ~1.3s; shellcheck clean; UMBRELLA? computed in the jq pass (no per-candidate body fetch). - handoff kickoff: run the script FIRST; trust its deps/tiering/ordering, recheck only flags. - handoff lore: the three 'cheap selector unreliable → re-derive by hand' bullets kept as the EVIDENCE for why the script exists; the prescription is redirected to 'run the script'. Operator-requested this session: 'rather than have the lore make us redo the selector's work, improve the selector.' Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-19 01:09:57 +02:00
timothyandtimothy	df8c5202d6	fix(409): report never-scanned LastScan as null for API/MCP parity (migration + read coercion) (#459) ... Co-authored-by: Timothy <timothy.look@gmail.com> Co-committed-by: Timothy <timothy.look@gmail.com>	2026-07-18 19:39:25 +00:00
timothyandClaude Opus 4.8	dd55f00ed4	docs(420): sharpen the skip's honest hit-rate (rebase-before-merge) + cite live combined-status check ... Review nits from the scoped #420 review. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 20:53:23 +02:00
timothy	922b1ef53c	docs(398,420): document the #420 skip + record build-once measured-and-rejected	2026-07-18 20:53:23 +02:00
timothy	8152afead1	Merge pull request 'feat(403): make unsupported PlaybackOrder loud at build time + tripwire' (#457) from feat/403-loud-playbackorder into main	2026-07-18 17:10:35 +00:00
timothyandClaude Opus 4.8	0f34c86afa	feat(403): make unsupported PlaybackOrder loud at build time + tripwire ... Adding a new PlaybackOrder was unsafe by construction: three build-time dispatch sites turned an unknown value into an enumerator silently. Classic substituted RandomizedMediaCollectionEnumerator (the // TODO default arm), PlaylistEnumerator had no default arm so the item was dropped, and BlockPlayoutBuilder's allow-list continue skipped it. (#70 already made YAML/Scripted log a warning and MultiCollectionGroup throws.) - each silent site now logs a Warning naming the order + engine + the fallback taken; the fallback itself is preserved so a live channel never goes dark on one misconfigured item and scheduler goldens do not move. - PlaylistEnumerator.Create gained an optional Option<ILogger> (it was static with no logger -- why the drop was unreportable); loggered callers pass it. - BlockPlayoutBuilder gained an explicit Random arm (it previously reached an enumerator only via the coincidental _ => fallback) and a loud defensive fallback. - new PlaybackOrderSupport matrix (per SchedulingEngineKind) + tripwire PlaybackOrderSupportTests: Supported ∪ Unsupported must partition the enum for every engine, so a new order fails the test until classified. BlockPlayoutBuilder consumes the matrix for its allow-list. - write-path rejection left unchanged (#70 closed the persistence hole; the perimeter has been wrong three times per decisions.md); reverse _ => None mappings reviewed and deferred (different axis; making them loud would warn on legit enumerator types). docs/decisions.md updated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 18:23:21 +02:00
timothyandtimothy	ab8e5d7a91	ci(338): distinguish ZAP warning (exit 2) from failure (exit 1) in security-scan (#452) ... Co-authored-by: Timothy <timothy.look@gmail.com> Co-committed-by: Timothy <timothy.look@gmail.com>	2026-07-18 16:16:41 +00:00
timothyandClaude Opus 4.8	eb47aed767	feat(396): collapsible sidebar + nav-group accordions ... Add two independent, persisted collapse states to the shell sidebar: - a header toggle that collapses it to a 60px icon rail - collapsible accordions per labeled nav group (Media, System); the unlabeled Primary group is always open, default-collapsed groups State + persistence live in web/src/app/sidebarState.ts (useSidebarState); AppShell stamps ctv-app-shell-collapsed on the shell root and the rail look is CSS-driven. Two namespaced localStorage keys (ctv-sidebar-collapsed, ctv-sidebar-groups) per the persisted-UI-state convention. In the rail, accordions are ignored (all items shown icon-only, label kept in the a11y tree + surfaced as a title tooltip, badges as a corner dot); active-route indicator works in both states; grid-column transition respects prefers-reduced-motion. Tests: colocated sidebarState.test.ts + a new describe in App.test.tsx (default-collapsed, accordion toggle+persist, rail, reload persistence). Docs: spa-conventions §13 + decisions.md 2026-07-18. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 16:43:33 +02:00
timothyandClaude Opus 4.8	a1b75c1f2c	docs(363): sweep the scan-flow drift the re-review caught ... The fix commit softened the script's scan-409 wording + dropped ?deep=true but left the docs describing the old behavior. Match them: - docs/e2e-local.md: POST .../scan (no ?deep; note local scans always ForceScan) + "409, deterministic bar a tiny residual TOCTOU gap" (was "guaranteed 409") - docs/ci-cd.md: same "guaranteed 409" -> "409 (deterministic bar ...)" Docs-only; harness behavior unchanged (still 38/38 green, deterministic). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 15:44:47 +02:00
timothyandClaude Opus 4.8	4345180a56	review(363): robustness + wording fixes from cold review ... - check seed_library_path succeeded (print lastrowid) so a silent seeding failure surfaces as a FAIL instead of degrading Flow A to an advisory skip with no diagnostic (Medium) - anchor the scan-status match to '"libraryId":2,' so it can't substring-match "libraryId":20/23 if the suite ever creates more libraries (Low) - drop the no-op ?deep=true (local scans always ForceScan; deep only affects Plex/Jellyfin/Emby) + note why (Low) - soften "guaranteed 409" for the scan flow to note the tiny residual TOCTOU gap the multi-second scan covers; Flow B stays race-free by construction (Low) - correct the "WAL tolerates a second writer" wording to the real reason (the busy-timeout retry serializes the writer) in the script + both docs (Nit) - use TEST-NET-1 192.0.2.1 (RFC 5737) instead of RFC1918 10.255.255.1 for the non-routable Jellyfin address (Nit) Re-verified: fresh-instance harness runs green (38/38), lock section deterministic. Functional E2E CI job already green on the prior head. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 15:33:39 +02:00
timothyandClaude Opus 4.8	8a85f9ddb5	test(363): functional-E2E harness — add deterministic scan-lock + collections-lock 409 flows ... Extends scripts/e2e-functional.sh with the two IEntityLocker 409 contracts the first cut (ersatztv#299) deferred as "racy", made DETERMINISTIC by firing the racing request only once the lock is provably held (no sleep-and-hope): - library-scan "already scanning" 409: seed ~60 tiny ffmpeg clips into the built-in Shows library so the scanner subprocess runs a few seconds, poll GET /libraries/scan-status until the library is active (that window is a strict subset of the scan lock's held window — StartScan after LockLibrary, EndScan before UnlockLibrary), then a second POST .../scan is a guaranteed 409. Self-skips (advisory) when ffmpeg is absent. - external-collections "already scanning" 409: seed a Jellyfin media-source row pointing at a non-routable address so the background sync hangs and the per-family lock stays held; the lock is taken synchronously before the 202, so the 202 proves it held. collections-scan-status corroborates; unknown source 404. Seeding uses python3's stdlib sqlite3 (already a harness dep) to insert rows the API can't create (LibraryPath, media-source); WAL mode tolerates the second writer. No new CI step/dependency — ffmpeg ships in the toolchain image. Verified: 4/4 fresh-instance runs green (38/38), lock section deterministic. Still deferred to #363 follow-ups: the playout-build lock 409 + isLocked projection (#215) and the UI-interactive Playwright flows. Docs updated same PR: docs/e2e-local.md, docs/ci-cd.md, the functional-e2e job comment in .gitea/workflows/docker-build.yml. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 15:17:21 +02:00
timothy	551366aa72	Merge pull request 'feat(293): paginate GET /api/v1/search/all-items to cap DoS exposure' (#442) from feat/293-search-allitems-cap into main	2026-07-18 12:17:20 +00:00
timothyandClaude Opus 4.8	c57fbf9826	docs(293): note the pageNum upper clamp (MaxAllItemsPageNum) in the decision + api-conventions ... Fix delta re-review flagged the decision entry + §5 note still described the pre-fix pageNum = Math.Max(0, pageNum); the shipped code clamps the upper bound too (0..2_000_000) to stop pageNum*pageSize overflowing int to a 500. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 13:40:20 +02:00
timothyandClaude Opus 4.8	ac7965dee4	feat(293): paginate GET /api/v1/search/all-items to cap DoS exposure ... The all-items endpoint fired ten index searches with limit:0 (every hit), so a broad authenticated query materialized the whole index into one response. Add optional pageNum/pageSize (clamped 1..1000; pageNum 0..2_000_000 so skip can't overflow int) and an additive per-kind Totals on the response; the SPA add-all flow now pages to completeness instead of a single unbounded fetch. - SearchController.SearchAllItems: clamp params (Logs §1 precedent), map Totals - QuerySearchIndexAllItemsHandler: skip=pageNum*pageSize, limit=pageSize, read SearchResult.TotalCount per kind - SearchResultAllItemsResponseModel: additive Totals (frozen-v1-safe) - web/src/api/search.ts: getSearchAllItems paging params + getAllSearchItemIds (pages until each kind hits its total; empty-page safety break) - tests: controller clamp/thread/totals, handler skip/limit/totals, SPA paging - docs: decisions.md 2026-07-18 (#293), api-conventions.md §5; regenerated OpenAPI Design: issue option (a) full pagination, operator-confirmed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 13:09:06 +02:00
timothyandClaude Opus 4.8	2281f2e764	fix(308): idempotent concurrent Add*ToCollection instead of a composite-PK 500 ... Two concurrent adds of the same item both membership-check it absent, both insert the CollectionItem composite key, and the loser's SaveChangesForcingVersion threw an uncaught DbUpdateException (SQLite 19 / MySQL 1062) -> 500. Now the loser is an idempotent no-op. - ConcurrencyExtensions.TrySaveChangesForcingVersion: bool-returning sibling that catches only a classified unique/PK violation and returns false. - 10 single-item Add*ToCollection handlers: return Unit.Default (no-op, skip fan-out) on false — the racing winner already inserted + rotated + rebuilt. - Bulk AddItemsToCollection: retry on a fresh context against recomputed membership so a partial-overlap collision doesn't drop the non-colliding items (bounded loop; common no-collision path runs once). - Provider detection via a TvContext.IsUniqueConstraintViolation static delegate (matches the existing IsSqlite/LastInsertedRowId provider seam), wired from Startup to SqliteErrorClassifier / MySqlErrorClassifier. - Add*ToPlaylist is NOT affected (PlaylistItem has its own identity PK; a playlist may legitimately contain the same item more than once). Tests: a negative-control anchor proves the race genuinely throws a classified exception; end-to-end handler tests reproduce a real cross-connection race via a shared-cache SQLite harness + a SavingChanges interceptor (the single-conn in-memory fixture cannot). Every fix-dependent test verified to fail with the catch disabled. Docs: api-conventions.md §7a (idempotent insert under concurrency) + decisions/optimistic-concurrency.md. fixes #308 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 13:02:14 +02:00
timothyandClaude Opus 4.8	e364b338e6	feat(425): per-source rotation weights + query corrections for auto-tune channels ... Auto-tune channels can now carry per-content-source rotation weights (weighted round-robin, e.g. 3x Show A / 1x Show B) and query corrections (exclude / add-untagged), supplied at bulk-create time via an optional `sources: [{sourceId, weight, excluded}]` on each AutoTunedChannelRequest. Design (Option A, reuse #70): when a source is customized the channel is backed by a system-owned MultiCollection of per-source SmartCollections carrying the weights, with PlaybackOrder.WeightedShuffle -- the exact path WeightedShuffleCollectionEnumerator already consumes. All-default weights keep the #69 single-SmartCollection fair-share shape. - Discriminators: TV -> live show_title:"X" (episodes carry no parent-show id in the index); movies -> stable id:{mediaItemId}. - Materialization is axis-dependent: TV materializes every base show individually (un-weighted shows keep per-show fair-share) + a live remainder at weight 1; MovieGenre materializes only touched movies + one count-weighted remainder. - Remainder = (base) AND NOT (materialized union excluded) -- a partition. - New nullable OwnedByChannelId on SmartCollection + MultiCollection (dual-provider migration); owned rows are hidden from the collection lists and cascade-cleaned on channel delete. Tests: AutoTuneAxisMap query/partition units; DB-backed weighted-path handler tests (TV materialize-all, movie count-remainder, exclusion, no-customization fallback); delete-cleanup. Docs: decisions.md, domain-model.md, api-conventions.md; OpenAPI trio regenerated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 03:52:06 +02:00
timothy	c91c34f113	docs(176): decisions + spa-conventions for the SmartCollection rule builder	2026-07-18 02:43:54 +02:00
timothy	8483af41fc	docs(176): plan Task 8 — guard Builder toggle + resync query (matches shipped fix)	2026-07-18 02:41:50 +02:00
timothy	279c0d6ce5	docs(176): plan — hook exposes clean non-null RuleField (Core DTOs are nullable-disabled)	2026-07-18 02:41:50 +02:00
timothy	5186bed910	docs(176): plan — broaden round-trip generator to special chars; escaping now total	2026-07-18 02:41:50 +02:00
timothy	218e1810cb	fix(176): total escaping so wildcard/quoted values round-trip (parse review)	2026-07-18 02:41:50 +02:00
timothy	7f7062729d	docs(176): fix plan operator model — split fulltext matches/notMatches from text contains (Task 3 review Critical)	2026-07-18 02:41:50 +02:00
timothy	ae22107d98	feat(176): GET /api/v1/search/fields endpoint + regenerated api artifacts	2026-07-18 02:41:50 +02:00
timothy	aa4127fcef	docs(176): fix plan test-project path + parse.ts self-review nits; add SDD ledger	2026-07-18 02:41:50 +02:00
timothyandClaude Opus 4.8	8c7af0e202	docs(176): implementation plan — SmartCollection visual rule builder ... Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 02:41:50 +02:00
timothyandClaude Opus 4.8	71159c1a8a	docs(176): design spec — SmartCollection visual rule builder (sub-project #1) ... Scopes #176 down after recon (Autotune #69 + composite create-channel #63 already shipped): the remaining leg is the visual WYSIWYG rule builder. Kodi one-level-nested model, compile-only to a closed Lucene subset (no schema change), backed by a new read-only GET /api/v1/search/fields catalog. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 02:41:50 +02:00
timothy	f44eee85c5	feat(386): Auto-Tune per-channel DetailPanel slide-over (SPA) ... Adds a right-hand "Configure" slide-over to each Auto-Tune Preview row, making a proposed channel editable before bulk-create — against the shipped #384/#385 backend only, so no control lacks a wire target. - New reusable SlideOver primitive (components/overlay.tsx), sharing a useOverlayBehavior hook (focus/scroll-lock/Escape/scrim) with Dialog. - Extract the Channel Builder's advanced-options model to builder/advancedOptions.tsx (enum catalogs, ADVANCED_KEYS, effectiveValue, INHERIT/omit useAdvancedOverrides hook); ChannelBuilder imports it unchanged (its tests pass byte-for-byte). The DetailPanel writes its own field JSX over the same hook — shared logic, per-screen layout. - Panes: identity (name/number + logo upload), Playback (Shuffle/Always-playing → advanced.playbackOrder/playoutMode), per-channel template picker, Advanced disclosure, lean read-only Query&size, read-only Content-sources via GET /members. - getAutoTuneChannelMembers API client (#384 read endpoint) + tests. - Screen-scoped §8 unsaved-changes guard + "Edited" row badge. - Dropped as backend-less decoration: MiniEpg, bug-initials generator, query text. Deferred to #425 with an in-pane hint: per-source weight steppers + corrections. - Docs: spa-conventions §11 (SlideOver + shared advanced-options), decisions.md. Refs #386	2026-07-18 01:31:40 +02:00
timothyandClaude Opus 4.8	ed6c43065f	feat(164): guided remediation for health checks (server-declared {Kind, Target}) ... Make the ~14 health checks actionable: each check that has a fix now declares where to go, and the SPA acts on it. Backend: - Widen domain HealthCheckLink (string Link) -> (string Target, HealthCheckLinkKind Kind) with ExternalDoc|AppRoute + factories; only the 4 link-building checks and the API mapper touched .Link. - Evolve HealthCheckResponseModel additively (/api/v1 frozen-additive): keep deprecated string? Link (still populated), add Brief (the BriefMessage the mapper was silently dropping) and nested Remediation {Kind, Target}. Kind is a mapped string, not a wire enum. - Make Mapper.GetStatus total: NotApplicable no longer throws (defensive; handler still filters it). InternalsVisibleTo(ErsatzTV.Tests) added to unit-test totality. - Fix 2 stale Blazor route links (media/trash -> /app/trash, search?query -> /app/search); add AppRoute remediation to actionable checks that had none (libraries / schedules / ffmpeg-profiles / settings). SPA: - DashboardScreen health panel renders remediation: AppRoute -> client-side nav button, ExternalDoc -> new-tab anchor; detail text truncates with title-hover. - Remove the dead "Open Classic UI" -> /system/health row from SettingsScreen (a #91b leftover that just 302'd to /app); update its regression test. Docs: decisions.md (#164), api-conventions.md (deprecate-in-place DTO evolution), blazor-route-parity.md (Section 4 correction); v1.json/v1.d.ts/endpoint-index regenerated. fixes #164 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-18 00:51:40 +02:00
timothyandClaude Opus 4.8	8f61ad6530	feat(385): per-channel overrides in auto-tune bulk-create ... Auto-Tune DetailPanel backend (#385), additive half. The create request `AutoTunedChannelRequest` gains three optional per-channel fields, all backward-compatible (omit = PR1 behavior): - `templateId` — overrides the batch template per channel - `advanced` — reuses the manual Channel Builder's `CreateChannelFromLineupAdvancedOptionsRequest` verbatim (24-field override set, `advanced.X ?? template.X` stamp contract). Axis default fills `PlaybackOrder` only when the caller leaves it null. - `logo` — uploaded channel image, `Sanitized()` at the request boundary (#283 stored-XSS defense), forwarded to `CreateChannelFromLineup.Logo` Resolved per channel inside `CreateAutoTunedChannelsHandler.CreateOne`, so one channel's bad override still yields a per-channel Failed/Skipped without aborting the batch. Per-source rotation weights + query corrections are split out to #425 (they need a MultiCollection-of-per-source-SmartCollections redesign — #70's WeightedShuffle reads weights only off MultiCollection join rows, and an auto-tuned channel is one SmartCollection). Bug-initials/colour generated logo also deferred (needs persisted Channel state + FFmpeg-pipeline wiring). Tests: handler override-threading (per-channel wins, axis default preserved, no-override baseline) + request `ToCommand()` logo sanitization. OpenAPI trio regenerated. Docs: decisions.md, api-conventions.md, domain-model.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 23:33:10 +02:00
timothy	1df08e86bb	ci: pre-push guard against pushing an uncommitted working-tree change (H13) ... A file in the pushed diff that still has uncommitted working-tree/index changes means the pushed commit doesn't match what you built/reviewed — the #416 index/worktree trap, where a --no-renames review fix lived only in the working tree and shipped without being committed (CI, the push, and the reviewer each saw a different tree). New fail-open pre-push hook blocks that precisely (only files in the branch's diff vs origin/main), after the H11 rebase check. Escape: ETV_ALLOW_DIRTY_PUSH=1. Documented in decisions.md. Refs #416	2026-07-17 23:05:30 +02:00
timothy	eb1aba99c1	docs(lore): run local gate + cold review BEFORE pushing to Gitea CI ... The handoff PR-routine and HARD-CONSTRAINTS placed the cold adversarial review AFTER push/PR-open. New standing rule (Timothy, 2026-07-17): local gate + cold review run BEFORE the push; fold pre-push fixes into the branch; CI only confirms. Runners should see only already-validated work — a Gitea run can't be cancelled, so a run spent on a lint/review-caught issue is pure waste. Also notes docs-only PRs are seconds now (ersatztv#416).	2026-07-17 22:39:42 +02:00
timothy	082124f706	docs(416): note shallow-checkout FETCH_HEAD/two-dot fix (ci-cd + decisions) ... Refs #416	2026-07-17 22:02:06 +02:00
timothy	92bb63b6da	docs(416): document docs-only CI skip ... ci-cd.md gains a 'Docs-only skip' section + triggers-table note; decisions.md records the decision. Cross-refs the separate PR-vs-main rerun redundancy (#420) and the within-run triple build (#398). Refs #416	2026-07-17 19:44:25 +02:00
3 people	caaae4cd00	docs(70): re-derive the stale-claim fix by grep instead of working the review's list ... Round-3 review returned BLOCKED: must-fix (b) was not closed. It was right, and the root cause it named is the point of this commit — the previous correction "was scoped to the four sites the reviewer listed rather than re-derived by grep". Fixing the list is not fixing the class. That is the same failure as B1, where the gate covered the two writers already in hand and missed CreateChannelFromLineup. Re-grepped the behavior class instead. Three survivors, two of them missed and one freshly introduced by the correction itself: - CreateMultiCollectionHandler.cs — the create twin of a comment whose UPDATE twin I corrected and whose create twin I never opened. Present tense, and contradicted by two tests in this same PR. - decisions.md — corrected one line in that file and left its sibling. - MultiCollectionItemWeight.cs (and its decisions.md mirror) — the ceiling rationale still claimed unbounded weights overflow the sum. They cannot: EffectiveWeight clamps before every sum and CycleLength widens to long. The earlier pass pattern-matched on the word "filtered" and left the identical defect on the ceiling. The ceiling's real job is the floor's argument — a billion is not a share of airtime any more than 0 is — so it now says that, and credits the clamp with the arithmetic safety it actually provides. Also corrected the writer claim to the right predicate: not "two persisting writers" (Add*ToPlaylist and Trakt persist it too, hardcoded) but two writers that persist a CALLER-SUPPLIED order. The full set is now classified persists-caller-value / persists-hardcoded / in-memory, including Engine/PlaylistHelper, which the previous "two Preview handlers" phrasing missed. That bullet has been wrong three times in the same shape; it now says so, since a lesson that keeps being re-learned is worth recording as a pattern rather than a fact. The BOM check caught this commit re-adding a BOM to the one file patched with utf-8-sig — the same trap, an hour after writing it down. Stripped; the mechanical pre-push check is what makes that survivable. Core.Tests 566, ErsatzTV.Tests 1673, 0 failed. Format verify exit 0. decisions.md +90/-0 (append-only guard green). Refs #70 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 17:00:46 +00:00
3 people	ccef0ca88a	fix(70): test the weight gate; correct rationale my own clamp made false ... Re-review of the fix commit returned MERGEABLE-WITH-NITS. It verified the gate is now complete by enumerating the writers itself (no fourth persisting writer) and proved B2's fix works by writing throwaway handler tests — which was also its point: the fix shipped with none. B2 was create and update silently DISAGREEING on the same input, and the fix re-established agreement with nothing pinning it. Both paths are now driven from one shared case list, plus an explicit test that create and update agree on every case — the per-path tests would both have passed while the two diverged, which is how the bug existed in the first place. Non-vacuity proven: inverting only the update path's validation fails 10 of 20 on a clean build (0 errors, so not a stale-dll pass), and the agreement test is among the failures. The rest is my own prose contradicting my own code. The commit that added EffectiveWeight removed the weight filter, then left four statements asserting a 0-weight source "is filtered out" — two of them authored by that same commit, including the stated justification for Minimum=1 in MultiCollectionItemWeight. A future agent could have read that and deleted the clamp or the floor as redundant; they are belt-and-braces and neither is. Corrected to describe what the code now does: the gate refuses input that means nothing on a share-of-airtime scale, the clamp protects rows predating the gate. Also corrected the writer count in the very bullet whose lesson is "grep every writer of the field": ReplaceBlockItems writes BlockItem.PlaybackOrder, not PlaylistItem.PlaybackOrder. There are TWO persisting writers of PlaylistItem's, and the correction itself had miscounted by conflating the two fields — so the lesson now says to grep each field separately. Core.Tests 565 passed, ErsatzTV.Tests 1673 passed, 0 failed. Refs #70 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 17:00:46 +00:00
3 people	c0da414a4c	fix(70): close the review blockers — third playlist writer, weight bounds, overflow ... Adversarial review of PR #402 returned BLOCKED. It could not break the WRR math or the stateless-restore claim (it probed restore across wraps at indices 12/13/20/37 — all held, and the clamp preserves a 1000:1 ratio exactly). What it broke was the perimeter. B1 — the validation gate had a hole, so the silent-drop bug shipped. CreateChannelFromLineup is a THIRD writer of PlaylistItem.PlaybackOrder; its own guard only covered MultiCollection entries, so a 2+ entry lineup of plain collections persisted WeightedShuffle straight through to PlaylistEnumerator's null-drop. My decisions.md claim that "the silent sites never see it" was false as written — corrected in place, with the lesson recorded: grep every writer of the field, the non-obvious composite handler is the one that gets missed. The Add*ToPlaylist handlers are safe only because they hardcode their order. B2 — Weight had no validation at all, and create/update disagreed on the same input. EF's HasDefaultValue(1) substitutes 1 for a 0 on INSERT (0 reads as "not set") but an UPDATE writes the 0 through — and a 0-weight source was filtered out of the rotation, deleting it from the channel silently. Exactly the failure this order is careful to avoid everywhere else. Now bounded 1..1000 by a shared MultiCollectionItemWeight used by both paths so they cannot drift, and clamped again in the enumerator for rows that predate the gate. B3 — Sum(weights) is checked arithmetic, so two int.MaxValue weights threw OverflowException from inside a playout build. Reachable through the API precisely because of B2. The ceiling fixes both; the sum also widens to long. M1 the lineup mirror now allows WeightedShuffle for multi collections, matching the PlayoutModeMustBeValid change it claims to mirror. M3 ScheduleAsGroup is documented as deliberately unread by this order. L1 MinimumDuration is computed over every source instead of the current rotation — under the clamp a rotation is a strict subset and is rebuilt each wrap, so caching over it went stale. L2 the retry guard keys off the rotation, not the raw collection count. N1 the tautological default test is gone: it built entities in C#, so it asserted the property initializer, not the migration — it could not have failed. Replaced with clamp, overflow, and cross-wrap restore cases (the property the review proved but found unpinned). H1 the two follow-ups the PR body claimed were "filed" did not exist. Now filed: #403 (silent dispatch-fallback hardening) and #404 (SPA weight UI, blocked-by #388). Core.Tests 565 passed, ErsatzTV.Tests 1643 passed, 0 failed. Refs #70 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 17:00:46 +00:00
3 people	953481c177	docs(70): record the WeightedShuffle design and correct the ShuffleInOrder misreading ... decisions.md entry covers why one enum value rather than two or a separate setting, why the weight lives on both multi-collection join entities with a DB default of 1, why it can only be applied on the ShuffleInOrder-shaped path (source identity is destroyed on the Shuffle path), why it is stateless, and why cross-engine exposure is closed by write-path validation instead of by changing five shipped fallbacks. The load-bearing part is the ShuffleInOrder clarification. Its balanced-shuffle padding reads as equalization, but the spacers emit nothing — one cycle plays every item once, so airtime stays proportional to collection size. It is anti-clumping, not fair-share. That was misread during this issue's own design pass and would have collapsed #70 to "already shipped"; the distinction is the whole justification for the feature, so it is now written down in both decisions.md and the domain-model glossary rather than left to be rediscovered. domain-model.md gains a Playback order row contrasting Shuffle / ShuffleInOrder / WeightedShuffle, and notes on the multi-collection row that the two join entities are mirrors. Refs #70 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 17:00:46 +00:00
timothy	fe6e2722f8	docs(lore): batching, no "main checkout", trust the queue, killed≠failed ... Operator-requested after PR #405 pushed 5 times, orphaning live runs the operator had to cancel by hand. Batches every standing-lore correction this session produced into one commit (per the batching rule it adds). New HARD CONSTRAINTS: - BATCH PUSHES. Cancellation is impossible from the agent side on Gitea 1.25.4 — REST .../runs/{id}/cancel and MCP cancel_run both 404, and the web-UI route needs a session+CSRF that doesn't script. Only the operator can cancel, so an orphaned run holds a runner slot until it finishes. If you must supersede a live run, SAY SO. - TRUST THE GITEA BUILD QUEUE. Do not gate/throttle a push on host health; the runners were retuned for stability. Batch because you can't cancel what you orphan, not to protect the host. - BOM-CHECK touched .cs before pushing. The #311 gate is fix-as-you-touch, and it bit two sessions the same day (PR #405 ×6; #70/PR #402 ×19 via Python utf-8-sig writing BOMs back). Verify your detector — an od-based grep reported all-clean while 19 files were dirty. Corrects a claim I nearly published: `dotnet format --include` DOES work here. The apparent no-op was the SHELL — CI's mapfile is bash-only, zsh has no mapfile → empty array → zero files → exit 0. Run it under bash -c. THERE IS NO "main checkout" — the biggest correction here. /Users/timothy/ersatztv is a shared mutable working tree whose HEAD is whatever the last session left there. Its name lies, and it bit TWO sessions on 2026-07-17, both doing the obvious thing: one assumed main and committed onto the #604/CI session's branch 24s after that session's own commit; another read git log there and concluded main was "4 behind origin" — a phantom. Framed as a design flaw, not a discipline failure: "check git status first" appears to confirm the false assumption and then goes stale (it WAS on main at 12:46 and wasn't by 14:17). Worktree discipline itself is healthy — 10 feature worktrees. Diagnosing CI reds — three ways to misread one, all hit this session: - A KILLED job reports conclusion=failure, not cancelled. The tell is a log that stops mid-step with NO error and NO failure marker. A runner retune killed run 1006's migration + E2E on a BOM-only diff that couldn't break them. Log timestamps are UTC, host is UTC+2 — convert before correlating. - cancelled ≠ failure: a cancel is NO verdict, and a run marked failure may hold a genuine job failure from BEFORE the cancel. Monitors must count FAILED and CANCELLED separately. - "Unable to pull refs/heads/v4" is act refreshing its action cache and is followed by "Cloned …" — noise, not a cause. Grep the failure marker, not the word "error". An infra-shaped red (setup/cache step, before your code compiles) is not a code failure; don't file a CI bug off one sample. Also: the cheap selector's failure modes are wider than deps+priority — it also misses in-progress claim state and umbrella-vs-child. Docs-only.	2026-07-17 18:15:32 +02:00
timothy	d92e717ebb	docs(72): fix the stale "Problems" filter name in domain-model (re-review) ... Re-review, Low: domain-model.md still documented a "Problems" filter — a line added by 50005b12 earlier in this same PR, so the PR contradicted itself once 38483681 renamed the filter to "No playout". Ironic in exactly the way the rename is about: the next person implementing the deferred fault classes would read the doc, see "Problems" documented, and restore the over-promising label that ChannelsScreen's own comment asks them not to. Stale docs re-introduce the bug the code comment defends against. decisions.md needed no change — it only ever named the badge, never the filter. Refs #72	2026-07-17 17:51:43 +02:00
timothy	9b3121f0c2	docs(72): record the channel-health API shape + mirror the lineup fault state ... Same-PR docs obligations for #72: - decisions.md — why health is the raw `PlayoutCount` fact on the lean list DTO rather than a derived status enum or a new endpoint, and, more importantly, why empty-schedule / broken-source / auto-tuned-origin are deliberately NOT computed. The empty-schedule one is the trap worth writing down: it only understands Classic ProgramSchedule playouts, so a badge driven off it would silently lie on the other four schedule kinds. - domain-model.md — a glossary row for channel health, drawing the line against /channels/state's OnAir (runtime liveness, not "would play if tuned"). - design-system — mirror the "No playout" badge + Problems filter into the Channels prototype so design-system/ stays a faithful mirror of prod (design-sync #388). Nature Docs carries playouts: 0 so the fault state is visible in the canvas. Refs #72	2026-07-17 17:51:43 +02:00
3 people	6052626316	chore: widen the BOM-guard remedy to staged + dirty files (review nit) ... Cold review (MERGEABLE-WITH-NITS) flagged that the printed fix snippet and the docs by-hand check used `git diff origin/main...HEAD`, which excludes staged-but-uncommitted files. A brand-new staged .cs is exactly what fires the pre-commit deny yet is absent from that diff, so the copy-paste remedy would strip nothing → deny again → loop. (The deny still names the file, so manual stripping always worked; this just makes the printed remedy match the guard's own detection set: branch diff ∪ staged ∪ dirty.) Verified on a fresh fixture: a brand-new staged BOM file is denied, and the widened snippet strips it. Left as accepted (fail-open, Medium, CI backstops; shared with sibling hooks): `git -C <path>` isn't parsed for the target tree (only `cd` is — matches the `cd <wt> && git` usage here), and the heredoc-body false-deny vector that bash-guard/worktree-guard also carry. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 15:50:01 +00:00
3 people	139bdaff5a	chore: PreToolUse guard that blocks a commit/push with a BOM'd touched .cs ... The #311 fix-as-you-touch rule is already enforced by a pre-commit hook and a blocking CI job, and both missed the same mistake twice on 2026-07-17: PR #405 shipped 6 BOM'd files, PR #402 shipped 19. The gap is that the pre-commit hook is routinely skipped — worktree hook friction makes `git commit --no-verify` the normal path here — which leaves CI, a ~10-minute round trip, as the first thing that tells you. This guard fires on Bash before git runs, so --no-verify can't skip it. It denies `git commit`/`git push` when a .cs the branch touches still carries a BOM, names the files, and gives the strip command. Why a hook rather than a note: a memory describing this exact trap did not prevent either failure — the PR #402 session re-added a BOM an hour after writing that memory down, because the usual ways of touching a legacy file re-add it silently (Python io.open(..., encoding='utf-8-sig') WRITES a BOM back; sed/perl round-trips keep it). A check that runs beats one you have to remember. Follows the existing gate pattern: a fail-open script in .claude/hooks/ registered on the PreToolUse/Bash matcher alongside bash-guard and worktree-guard. Any parse or lookup trouble → allow; this must never be the reason a commit can't happen, and CI remains the backstop. Generated *.Designer.cs / TvContextModelSnapshot.cs are exempt, matching what dotnet format itself skips. Scoped to this repo, and it resolves the target tree from an in-command `cd` because commits here run as `cd <worktree> && git ...` and the harness resets the shell cwd between calls. Verified against a throwaway fixture: denies commit and push for a BOM'd touched file (naming it); allows a clean tree, a BOM in generated files only, `git status` with a BOM present, and `echo 'run git push later'` (no false-trip on the words). Proved live via a sentinel — it fires on the Bash matcher — sentinel removed. docs/contributing.md updated: three enforcement layers, why the third exists, the utf-8-sig re-add trap, a by-hand check, and the bash-not-zsh caveat for `dotnet format --include` (mapfile is bash-only; under zsh the file list is empty and the tool looks like it silently did nothing). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 15:50:01 +00:00
timothyandClaude Opus 4.8	469d725559	ci(406): apply the memory-swap rule to our own two sites; stop leaning on the peak reading [decisions-edit] ... Re-review of the fix commits came back MERGEABLE with three findings worth acting on. MEDIUM -- the PR documented a standing rule ("--memory without --memory-swap silently grants 2x in swap") and then didn't apply it to the two sites this repo owns: docker-build.yml's smoke container and scripts/migration-smoke.sh, both `--memory 2g` with no --memory-swap. Pre-existing rather than a regression, but a rule you don't follow in your own repo isn't a rule. The migration-smoke one matters most: it runs on the PROD host in the release path, so a runaway migration should die against its cap rather than quietly swap out the box serving media. LOW -- and this is the important one: the docs leaned "peak 8305 MiB is probably mostly reclaimable cache". An independent probe (full solution build, same CI image, shared compilation off) measured peak 9457 MiB / anon 7134 MiB / file 421 MiB. ANON DOMINATED. Having verified the *mechanism* (peak overstates because it counts page cache), I guessed the *magnitude* in the direction I preferred -- the exact failure this entry criticises, committed inside the entry criticising it. Corrected in ci-cd.md, decisions.md and on server-management#604 (where the previous comment could have led to an unsafe 6g cap). Consequences now recorded honestly: a 6g cap looks UNSAFE, #570's "6g proved too tight" is the rule not an outlier, and #406's premise ("if this brings peak RSS well under 6 GiB the whole budget loosens") is looking DEAD -- the 7134 MiB anon was measured with shared compilation already off. The switches remain right; the looser budget they were meant to buy does not follow. NIT -- dropped the unverified claim that this also disables the Razor build server. The UseRazorBuildServer -> UseSharedCompilation fallback is .NET 5-era; Razor has been an in-process source generator since .NET 6, so there is likely no separate server to disable on .NET 10. Unverified, zero impact, so it has no business in a doc arguing for measurement over assumption. [decisions-edit]: the touched docs/decisions.md lines were added by this PR's own earlier commits, not settled entries on main -- net vs origin/main remains a pure insertion (0 deletions, verified). Also the sanctioned reason: the entry was factually wrong (see LOW). Verified: both workflows parse; migration-smoke.sh passes bash -n; the parsed mysql option string is `--memory=2g --memory-swap=2g --cpus=2`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 16:13:39 +02:00
timothyandClaude Opus 4.8	7859510c78	ci(406): report anon/file split — memory.peak counts page cache, not just RSS [decisions-edit] ... The peak-memory step from the previous commit worked on its first real CI run and immediately produced a number that would have led #604 to the WRONG decision: peak container memory: 8305 MiB (8708448256 bytes, from /sys/fs/cgroup/memory.peak) Read naively that says "the test job needs 8.1 GiB, so the 10g cap must stay and #406's 'the budget loosens' premise is dead". That reading is probably wrong, because `memory.peak` is the high-water mark of `memory.current`, which charges PAGE CACHE to the cgroup as well as anonymous memory. It is not peak RSS, and a build job does heavy NuGet/npm/obj/bin/coverage I/O. Proven on bumblebee -- a container with ZERO anonymous memory that merely reads an 800MB file: memory.peak : 826 MiB anon : 0 MiB file : 800 MiB This inverts the decision: page cache is RECLAIMED under a tighter cap, not OOM-killed, so a large peak that is mostly `file` is NOT evidence that the cap must stay high. `anon` is what forces an OOM. So the step now prints the anon/file split and says, in the log and the docs, to size caps on anon rather than peak. Honest about what this still isn't: the split is read at end-of-job, so it is the composition then, not at the peak instant -- a job peaking mid-`dotnet test` and freeing after would under-report anon. Sampling a true peak-anon is ersatztv#412, filed with the options. Also recorded there and in ci-cd.md: there is NO pre-change baseline from this instrument (7.8 GB VBCSCompiler was measured host-wide across concurrent jobs, not inside one job container), so #406's premise is neither confirmed nor killed yet. What IS established: no persistent compiler server survives a build, and migrations is green with mysql capped at 2g with swap disabled. The irony is the point, and it is now written into decisions.md: this repo's CI perf work keeps stating numbers from plausibility rather than measurement (#390's "2-4min" apt-ffmpeg estimate; real 110s). Measuring the wrong quantity precisely is the same failure wearing a lab coat -- and it bit this very change, whose whole purpose was to replace an inherited number with a measured one. Verified: shellcheck clean; the real step body run inside a --memory=2g container under `bash -e -o pipefail` reads memory.peak, prints the split, writes the step summary, exits 0; and it degrades to a skip + exit 0 where no cgroup file exists. [decisions-edit]: the modified docs/decisions.md lines are ones THIS PR added in its own earlier commits, not a settled entry on main -- net vs origin/main is still a pure insertion (0 deletions, verified). It is also the sanctioned reason: the entry as written was factually wrong about what memory.peak measures. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 15:53:24 +02:00
timothyandClaude Opus 4.8	04e3a5249d	ci(406): disable swap on the capped mysql; fix peak-step placement and failure mode [decisions-edit] ... Review fixes from the cold adversarial pass on PR #411. Verdict was MERGEABLE, but one HIGH is a genuine defect worth the round-trip. [decisions-edit] rationale: the 12 "modified" lines in docs/decisions.md are lines THIS PR added in its own previous commit (e33205e8), not a settled entry on main. The hook diffs against HEAD; net against origin/main this file is still a pure insertion (65 added, 0 deleted -- verified, and CI's decisions-guard diffs against the PR base). It is also the sanctioned reason regardless: the entry as first written contained a factual error (see HIGH below). HIGH -- `--memory=2g` alone silently grants 2g of SWAP as well. Docker defaults an unset --memory-swap to 2x --memory. Verified live on bumblebee: `--memory=2g` alone gives memory.max=2147483648 AND memory.swap.max=2147483648; adding `--memory-swap=2g` gives memory.swap.max=0. Capping RAM while permitting swap is close to the worst outcome on the one host whose swap thrash is the entire reason for the cap -- and a swapping mysqld mid-DDL is exactly the pathology behind the known `Command Timeout expired` migrations flake, so the naive cap could have made that flake worse rather than better. The comment promising "if this ever OOM-kills the service" described a failure mode that would not have happened first: it would have swapped, silently. Now --memory-swap=2g (swap off); prefer a loud OOM over silent degradation. The same 2x applies to the runners' 10g job slots (10 GiB RAM + 10 GiB swap each), which is a plausible direct mechanism for the incident's 21 GiB swapped. That is #604's boundary -- reporting it there. MEDIUM -- `if: always()` does not make a step advisory. It controls whether the step RUNS, not whether its failure fails the job, and defaults.run.shell: bash means -e is on, so a failed cat/redirect in an informational step could redden a green test job. Use continue-on-error: true, matching the neighbouring Coverage summary step. MEDIUM -- the peak step must run LAST. memory.peak read at step N reports the peak only up to N, so sitting before Coverage summary excluded reportgenerator's memory -- under- reporting the very number #604 sizes caps on. Moved after it. LOW -- dependency-scan.yml gets the env too; workflow env does not cross workflow files, and that one runs unattended on a cron against the prod media host. NIT -- Dockerfile ENV moved below the openapi-generator wget: an ENV invalidates every layer under it, and that ~30MB download is deliberately ordered early to stay cached (#190). Nothing between it and restore compiles, so this placement is free. NIT -- docs no longer imply this shrinks `format`. dotnet format loads Roslyn in-process via MSBuildWorkspace and never spawns csc, so its 3.95 GiB is untouched; sizing the small lane on the assumption it shrank would be a trap. Also record honestly that 543 MiB is init+idle rather than the 787-migration replay (so 2g is a measured floor + headroom, not a measured ceiling -- the migrations job going green is what validates it) and that --cpus=2 is judgement with no measurement behind it. The entry criticises dressing estimates as measurements; it should not do that itself. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 15:36:18 +02:00
timothyandClaude Opus 4.8	e33205e8eb	ci(406): disable persistent compiler servers, cap the mysql service, report peak RSS ... Refs #406, server-management#604. Items 1 and 2 of #406 (item 3 landed in PR #407). Root cause of the 2026-07-17 bumblebee incident (load 340, 21 GiB swapped, prod media down until reboot) is CI-induced memory pressure, and the largest single consumer was Roslyn's persistent VBCSCompiler at 7.8 GB RSS. It is a compiler *server*: it outlives the build that started it to keep its heap warm for the next one. In CI there is no next build -- each job container is torn down at the end of the run -- so it is pure cost, and it is the actual reason each job needed a 10 GiB cap. 1. Disable the compiler servers via the workflow's top-level env (UseSharedCompilation, DOTNET_CLI_USE_MSBUILD_SERVER, MSBUILDDISABLENODEREUSE). MSBuild surfaces env vars as properties and only defaults UseSharedCompilation when empty, so the env var wins and every dotnet call is covered without touching each call site. Also set as ENV in the Dockerfile's SDK stage: the workflow env reaches runner-side jobs only, and the `build` job compiles inside `docker build` where it does not propagate -- that is the job server-management#570 measured pegging 5.999/6 GiB, so it is the one that most needs this. Build stage only; the final image is FROM runtime-base, so nothing lands in the shipped image. Verified locally: a default build leaves 1 VBCSCompiler alive, the same build under these vars leaves 0, and ErsatzTV.sln still builds clean (0 errors). Trade-off: csc runs per project instead of via the server, costing some build time. 2. Cap the mysql service (--memory=2g --cpus=2). A runner's container.options applies to the job container ONLY: a live migrations job reported HostConfig.Memory=10737418240 for the job and mem=0 nanocpus=0 for its mysql service, i.e. unbounded. 2g is sized on measurement -- mysql:8.4 with this exact env peaked at 543 MiB during init, 481 MiB idle (probed on bumblebee) -- not on inheritance. Also add a `Report peak container memory` step to the test job (cgroup memory.peak, always(), never fails a build). server-management#604 sizes both runners' caps on that number and it was inherited rather than measured; now every run reports it. Docs: ci-cd.md gains the compiler-server + services-cap sections and the lane table is corrected (PR #407 reverted the #390 lane move without updating it). decisions.md records all three, incl. the standing rule that a services: container never inherits a cap. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	2026-07-17 15:22:31 +02:00
timothy	71708633f5	test(264): pin the path-scan-error log; correct testing.md accuracy ... Re-review of the previous fix commit (986ccfaf, verdict MERGEABLE) raised one fair hit and one claim that did not survive checking. Fair hit — that commit's message asserted "LogError is an extension method that NSubstitute can't cleanly verify" as the reason for adding no test. That is FALSE, and this repo disproves it: ErsatzTV.Tests' ReleaseMemoryHandlerTests.ShouldHaveLogged asserts on ILogger via ReceivedCalls(), which works precisely because the LogError extension dispatches to the substituted ILogger.Log. Reusing that idiom here costs 4 lines, so pin the log the previous commit added. Verified non-vacuous: asserting a message the handler never logs fails the test. Not applied — the same review called the row's "1471 tests" misleading on the grounds that TranscodingTests is [Explicit] and contributes most of that count. TranscodingTests is indeed [Explicit], but filtering it out yields exactly 1471, so 1471 is already the runnable count and the row was accurate. Kept the number; documented TranscodingTests as [Explicit]/opt-in instead, since it was a genuine omission from a table that claims to be authoritative. Also rewrap the verification-gate paragraph the previous commit left over-long, and name the two always-run projects instead of "both". Deferred (filed separately): a canceled local scan now logs at ERROR per path via this log. It mirrors the remote handlers exactly, so diverging here would be the inconsistency, not the fix.	2026-07-17 14:35:09 +02:00
timothy	986ccfaf6c	fix(264): log local path scan errors (review finding) + map Scanner.Tests ... Adversarial review (MERGEABLE, no blockers) raised two items worth folding in rather than deferring: Medium — ScanLocalLibraryHandler silently swallowed path scan errors, while the three remote scanners it mirrors all log result.LeftToSeq(). That mattered less when a failed path only skipped the path-level LastScan, but the previous commit makes a failed path suppress the library-level scan time too — so the user would see exactly the #264 symptom ("Never scanned") with nothing in the log explaining why. That is a diagnosis dead-end of the same class as the bug being fixed, so log it here rather than file a follow-up. No test: the sibling Synchronize*LibraryByIdHandlerTests don't assert on logging either, and LogError is an extension method that NSubstitute can't cleanly verify. Low — docs/testing.md bills itself as the authoritative map of what each test project covers but omitted ErsatzTV.Scanner.Tests entirely (1471 pre-existing tests). This PR adds a file to that project, so add the row and include it in the per-PR verification gate.	2026-07-17 14:27:50 +02:00