CI (Gitea Actions, .github/workflows/ci.yml):
- CodeQL off the PR critical path: runs on push-to-main + a weekly cron only
(`if: github.event_name != 'pull_request'`, `schedule: Mondays 03:00 UTC`).
The 2-language 20-min matrix was the slowest thing on every PR; code is still
scanned on main + weekly.
- Cache Playwright browsers (~/.cache/ms-playwright) in the storybook-a11y + e2e
jobs — skips the chromium download on a hit; `install --with-deps` then only does
the fast apt deps check.
- Cache NuGet (~/.nuget/packages, keyed on **/*.csproj — no packages.lock.json) in
the backend / e2e / api-client-drift / codeql-csharp jobs.
- `npm ci --prefer-offline --no-audit --no-fund` in the 4 npm jobs.
Demo (docker-compose.yml, local only — NOT used by CI): web image node:24 →
node:24-slim (~1.1GB → 232MB verified). The container only runs `npm ci && ng serve`
and the native deps ship prebuilt glibc binaries, so slim needs no toolchain —
verified: npm ci clean, ng serve boots, app returns 200.
Note: CI timing/behaviour can only be confirmed on the Gitea runner (not observable
locally). Every change here is independently revertable.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The e2e job backgrounded `dotnet run &` and `ng serve &` in separate Actions
steps, then `npx wait-on` (no timeout) in a later step. A process started with
`&` in one step is killed when that step's shell exits, so wait-on waited forever
on servers that were already gone — the job hung until the runner's hard limit
(~2h; Gitea's act_runner doesn't reliably enforce timeout-minutes).
Move both servers into Playwright's `webServer` (an array: backend + `npm start`),
so Playwright starts them, waits for readiness, runs the suite, and tears them
down in the one `npm run e2e` process — CI and local alike. The CI e2e job is now
just npm ci / playwright install / npm run e2e. `reuseExistingServer` is on locally
(reuses a running app, incl. the docker stack) and off in CI (fresh start).
Verified locally via `CI=1 npm run e2e` against free ports + a clean db: both
smoke tests pass in ~15s (no hang). The earlier local failure was leftover SQLite
state (WP-22) from a dirty run resuming a Concept — CI checks out fresh, so its db
is always empty.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Applications, documents (+ audit log) and the brief move off static in-memory
Dictionaries onto a real SQLite file via EF Core, so demo data survives a
process restart or `docker compose restart api` for the first time. The three
stores (ApplicationStore/DocumentStore/BriefStore) keep their exact public
signatures and static-class shape — no DI, no async ripple into Program.cs's
minimal-API handlers — each method just opens a short-lived AppDbContext via
Db.Create() under the same lock it already had. Opaque nested shapes (a
wizard's draft snapshot, a brief's sections/placeholders/status) are stored as
JSON text columns rather than redesigned into relational tables, matching the
existing "don't interpret it" posture.
Found two things the WP's own text got wrong, corrected in
docs/backlog/WP-22-durable-persistence.md's Deviations section: SeedData never
seeded these three stores (only the read-only BRP/DUO-mimicking GETs, which
stay in-memory) so there's no seed step; and no new docker-compose volume is
needed since the existing bind mount already covers the SQLite file — verified
against this environment's real podman-backed compose stack, not just by
reading the file.
Also: pinned SQLitePCLRaw.bundle_e_sqlite3 to 3.0.3 (EF Core Sqlite's own
transitive default bundles a pre-3.50.2 SQLite with a known high-severity
memory-corruption advisory); found and fixed a real xUnit test race where
concurrent test-class hosts stomped a shared static connection-string field,
fixed by disabling cross-class test parallelization rather than adding DI the
stores don't otherwise need.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
angular.json gains an i18n block (sourceLocale nl, en translation file) and
an `en` build/serve configuration with i18nMissingTranslation: "error" so a
new $localize string without an English unit fails the build, not silently
falls back. CI now runs `ng build --localize` to build both locales every
run. Verified end-to-end, not just "the build succeeded": the nl bundle
ships "Inloggen met DigiD", the en bundle ships "Log in with DigiD".
Incidental: prettier/compodoc regen noise in docs/wcag-checklist.md,
src/docs/a11y.mdx, documentation.json from the same working session.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds a happy-path spec (login → dashboard → registratie wizard, including
a real identity-document upload → real submit) and a degraded-path spec
(?scenario=error → <app-async> error slot → retry), both driving the real
app against the real .NET backend, plus a CI job that boots both.
Writing the retry spec surfaced a real bug: AsyncComponent's retry() only
reloads a [resource]-fed instance, so every real page (all [data]-fed via
a store's RemoteData) had a silently no-op retry button. Added a
retryClicked output and wired it on the dashboard's two async blocks.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Turn the interactive Storybook a11y addon into a build gate:
- @storybook/test-runner + axe-playwright over the static build
(.storybook/test-runner.ts reads the a11y tags from story context)
- test-storybook / test-storybook:ci scripts; storybook-a11y CI job
- triage: escape-hatch a11y.disable on stories whose display:contents
wrapper splits <ul>/<li> or <dl>/<dt>/<dd> (structural, deferred to
WP-11/WP-12, each with justification + cross-ref)
- fix trivial violations: footer/wizard-shell contrast, text-input label,
wizard stories missing provideApiClient
Verified: broken story fails the gate; 133 stories pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ehoandClaude Opus 4.8