Files
ersatztv/docs/decisions/records/spa/app-shell-extraction.md
T
timothy fba5233caf
PR Gates / CI image pin matches docker/ci (pull_request) Successful in 11s
PR Gates / Docs update reminder (pull_request) Successful in 16s
PR Gates / decisions lifecycle (pull_request) Failing after 23s
Build ErsatzTV Image / Formatting (changed .cs conform to .editorconfig) (pull_request) Successful in 1m17s
Build ErsatzTV Image / API docs in sync (OpenAPI + endpoint index) (pull_request) Successful in 1m29s
Build ErsatzTV Image / Build & test (.NET) (pull_request) Successful in 8m5s
Build ErsatzTV Image / EF migration integrity (SQLite + MySql) (pull_request) Successful in 16m5s
Build ErsatzTV Image / Functional E2E (curl + UI contracts) (pull_request) Successful in 17m6s
Build ErsatzTV Image / Build & push image (amd64) (pull_request) Has been skipped
feat(610): split the decision corpus into one YAML-frontmatter file per record
168 records -> docs/decisions/records/<area>/<topic>.md (163 active, 23 dirs) and
docs/decisions/archive/<area>/<topic>.md (5 archived). The filename IS the key,
so one-active-record-per-key becomes a filesystem property rather than a
validator check, and supersession becomes a `git mv`.

WHY: the monolith was a concurrency problem before an aesthetic one. A
3,900-line append target made parallel sessions collide -- PR #605 and PR #614
both hit append-vs-append conflicts during routine rebases, and hand-resolving
those inside the corpus is exactly the operation the rationale-rewrite guard
exists to police.

HOW IT IS VERIFIED: a ~170-file diff cannot be meaningfully read, so correctness
does not rest on reading it. The parser was taught BOTH formats first, so the
body-diff guard parses the old form at the merge-base and the new form at head --
the migration validates itself, no bypass. The proof is a field-level equivalence
harness: 168 records before and after, zero lost, zero gained, zero field
mismatches, zero rationale bodies differing. Reviewers should scrutinise the
harness; it is the actual evidence.

What measuring caught that reading would not have:

- ~500 lines sit OUTSIDE any record -- decisions.md's lifecycle schema and each
  topic file's preamble, mostly the only copy. Source files are kept and
  stripped, never deleted. They also cannot be filed per-area: topic files hold
  several areas and 4 of 23 areas span several files.
- Archive discovery was a non-recursive glob; after the split it found ZERO
  archived records, surfacing as four bogus "supersedes points to unknown key"
  errors rather than an obvious failure.
- ~32 live docs point into the corpus BY DATE, which the split dangles. Each
  stripped file now ends with a generated "Records formerly in this file" index,
  which also rescues the identical breadcrumbs in old issue comments.
- decisions.md's "In this file:" list was 97 same-file anchor bullets that the
  split makes WRONG, not merely stale. Dropped; the generated index replaces
  them with links that resolve.

The equivalence harness now runs against a checked-in FIXTURE, not the live
corpus. The earlier version migrated the real tree, which made it a one-shot:
the moment the migration landed there was nothing left to move and the tests
failed for reasons unrelated to the code. A fixture keeps them testing the
SCRIPT rather than the repo's current state.

Keys preserved verbatim, warts included: `sched` (12) and `scheduling` (1) remain
two directories for one concept. Renaming a key is not a move -- it changes
identity, breaks the equivalence proof, and invalidates MemPalace's per-key
drawers. Taxonomy normalisation is separate work.

refs #610
2026-07-25 19:45:09 +02:00

3.5 KiB

key, title, status, since, supersedes, superseded-by, rule, signals, mechanics
key title status since supersedes superseded-by rule signals mechanics
spa.app-shell-extraction 2026-07-15 — App shell/routing extraction + explicit primary-action ownership (#247) active 2026-07-15 none none `App.tsx` is only the composition root over `web/src/app/routes.tsx` (stable route-object identity), `app/AppShell.tsx` (shell chrome), and `app/ScreenContent.tsx` (exhaustive screen dispatch); primary actions are one explicit `PrimaryActionProvider` registration per screen, replacing the old global `ctv:primary-action` window event. app shell extraction, route identity, primary-action ownership · paths: `web/src/app/routes.tsx`, `web/src/app/AppShell.tsx`, `web/src/app/ScreenContent.tsx` · issues: #247, #243, #238, #230 `docs/spa-conventions.md` §10; `App.test.tsx`

Final phase of the App.tsx modularization epic (#243). web/src/App.tsx is now only the composition root: it owns activeRoute, currentPathRef, the navigate/popstate pair that consults the dirty guard, the approved librariesSubPath, and the global theme/health inputs, then composes the shell with the matched screen. Route metadata/matching/sidebar groups moved to web/src/app/routes.tsx; shell chrome (Sidebar, TopBar, Connect menu, version and theme controls) moved to app/AppShell.tsx; exhaustive screen dispatch plus the Media/Libraries wrappers moved to app/ScreenContent.tsx. No route, styling, API, or dependency changed.

Route identity is deliberate infrastructure. routes.tsx contains the ONE stable module-level array of shared ScreenRoute objects. routeFromLocation() returns those same references for an allowSubPaths base path and every owned sub-path, so React's Object.is state bailout remains part of the contract rather than an accidental implementation detail. The three existing sub-path mechanisms stay distinct: Playouts and Media self-own pathname/popstate; guarded Libraries receives only the path App approved after its dirty guard; the remaining route screens retain their established self-owned or keyed remount behavior. The extraction does not unify them into a speculative router/framework. ScreenContent uses an exhaustive ScreenId switch, so a future route added without a dispatch branch fails typecheck instead of silently falling through to a placeholder.

Primary actions are one explicit screen-owned registration, not a global event. The old ctv:primary-action window CustomEvent and string-keyed dispatcher are removed. PrimaryActionProvider holds exactly one { routeId, handler, owner } registration shared by the active screen and TopBar; usePrimaryAction registers it, keeps the latest handler behind a ref without ownership churn, and clears it only when the same opaque owner unmounts. A stale cleanup therefore cannot erase a newer screen's action. TopBar renders the Plus action only when the route has a non-empty label AND that active route owns a matching registration, so metadata alone can no longer create a dead button. This is intentionally not a generic action bus: one declared action per screen is the whole contract.

Tests pin stable route-object identity (including sibling sub-paths and query-only URLs), composition-level sub-path rendering and dirty-popstate restoration, and primary-action matching/latest-handler/route-change/ cleanup/stale-owner behavior. Detailed Builder and Settings behavior is colocated with those screens while App.test.tsx remains shell/routing/composition coverage. Refs #247 #243 #238 #230.