Commit Graph

2 Commits

Author SHA1 Message Date
fabi
3c683247c0 chore(deploy): expose the export rebuild hatch in the UI; rehearse migration 014
Three deploy-readiness gaps, none of them in the export invariant itself.

1. The escape hatch was unreachable. `POST /host/export/rebuild` was added as the fix
   for "a failed export is terminal", but nothing in the frontend called it — so recovery
   required a terminal, the API docs and a valid JWT. The host page instead told the host
   to reopen uploads and re-release, which is the destructive act the endpoint exists to
   avoid (it unlocks the gallery to every guest and retracts the release). The failed
   state now offers "Erneut versuchen"; a ready keepsake can be rebuilt behind a confirm,
   since a rebuild makes it briefly undownloadable.

2. Migration 014 had never been run against anything. It is the repo's first destructive
   migration and its backfill has to carry the old notion of "downloadable" across exactly
   — get it wrong and a released event's keepsake silently 404s, on data that cannot be
   re-created. backend/scripts/rehearse-014.sh proves it on a throwaway postgres, against
   a real pg_dump or a synthetic DB covering every pre-014 state, asserting up, down and
   up-again all preserve downloadability. Verified non-vacuous: retiring nothing (ELSE -1
   -> ELSE e.export_epoch) fails it, naming the three keepsakes it would resurrect.

   It also surfaced that the down migration is an inverse UP TO DRIFT: a ready flag that
   disagreed with its job row comes back FALSE. That heals corruption rather than
   restoring it, and costs nothing (no file_path to serve), so the assertion checks what
   actually matters — no genuinely downloadable keepsake loses its flag — and reports the
   normalisation instead of failing on it.

3. No advisory scanning. cargo audit + npm audit, in .github/workflows/ because Gitea
   Actions scans it too and e2e.yml already resolves actions/* from GitHub. The runner is
   not assumed to have cargo. RUSTSEC-2023-0071 (rsa/Marvin) is ignored SPECIFICALLY, not
   globally: it is unreachable here (HS256 + bcrypt, Postgres only) and unpatched, so
   failing on it would mean a permanently red pipeline everyone learns to ignore.

Tests: e2e pins that a failed keepsake recovers via rebuild while the event stays
released and uploads stay locked — so a future refactor implementing rebuild as an
internal reopen+re-release fails — and that rebuild cannot publish an unreleased gallery.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-14 21:09:53 +02:00
MechaCat02
e42d8a92a1 feat(e2e): Playwright suite — 134 tests across 9 spec areas + UA matrix
Adds an end-to-end Playwright test suite under e2e/ that spins up an
isolated docker-compose stack (Postgres :55432, Caddy :3101, backend with
EVENTSNAP_TEST_MODE=1, SvelteKit adapter-node frontend) and exercises the
SvelteKit app against the real Rust backend.

Phase 1 — happy paths covering every documented USER_JOURNEYS.md flow:
  01-auth/      join, recover, admin login, leave event, PIN lockout
  02-upload/    gallery picker (API path), rate-limit + admin toggle
  03-feed/      like/comment SSE, filters, SSE reconnect on visibility
  04-host/      event lock API, ban/unban, promote
  05-admin/     config validation, foundational authz guards, stats
  06-export/    /export status + download stub
  __smoke/      cross-UA happy-path (runs on every UA project)

Phase 2 — adversarial + browser chaos:
  07-adversarial/  XSS payloads (6 × display name path), SQLi shapes,
                   length / encoding / RTL override / NUL byte;
                   file-upload boundaries (ELF body claimed as JPEG,
                   oversize vs max_image_size_mb, zero-byte, NUL
                   filename, path-traversal, SVG-with-script);
                   JWT alg:none, signature/payload tamper, expired
                   session, PIN brute-force (serial + parallel),
                   admin password brute-force; deep authz (cross-user
                   delete, banned user across like/comment/feed-read,
                   host→admin escalation); small-scale DDoS (20× /join,
                   10MB comment body, 10 concurrent SSE).
  08-browser-chaos/ localStorage / sessionStorage / cookie purge,
                    IndexedDB drop mid-session, offline → reconnect,
                    slow-3G, 503 flakes, 429 with no retry storm,
                    multi-tab same/different user, no-JS, hostile CSS,
                    clock skew ±1h / -2d, localStorage quota exhausted.

Phase 3 — mobile gestures (runs only on chromium-mobile / Pixel 7):
  09-mobile/    touch-target ≥44px audit, env(safe-area-inset-bottom)
                structural check, long-press (FeedListCard → ContextSheet,
                quick-tap negation, click-suppression), double-tap
                (feed card like + lightbox heart-burst, via synthetic
                pointer events to bypass the first-tap-fires-click trap),
                viewport reflow (portrait/landscape/narrow/phablet),
                plus fixme stubs documenting planned gestures (swipe
                lightbox L/R, swipe-down dismiss, pull-to-refresh,
                long-press-comment).

Cross-UA matrix (chromium-engine projects run @smoke only):
  chromium-pixel7, chromium-galaxy-s22, samsung-internet (Samsung UA
  emulation on Galaxy viewport), edge-android, plus webkit-iphone,
  chrome-ios, firefox-android, firefox-desktop — the latter four need
  libavif16 on the host (Playwright dep) but the configs are in place.

Infrastructure:
  - fixtures/test.ts central test.extend (api, db, adminToken, guest,
    host, signIn). Per-test DB truncate via the dev-only POST
    /admin/__truncate route, gated by EVENTSNAP_TEST_MODE=1.
  - helpers/sse-listener.ts, helpers/upload-client.ts (Node-side
    multipart for adversarial file-upload tests + JPEG/PNG/ELF magic
    constants), helpers/touch.ts (longPress / doubleTap / swipe /
    inlineStyle / computedStyle).
  - 10 page objects covering every route + UploadSheet/Lightbox.
  - global-setup waits for /health, logs in admin, disables every
    rate-limit and quota toggle.
  - .github/workflows/e2e.yml: PR check runs chromium-desktop + the
    smoke matrix in parallel, uploads playwright-report/ and traces on
    failure.

Findings the suite surfaces as live `[finding]` warnings (not silenced):
  1. /admin/login has no rate-limit or lockout (bcrypt cost only).
  2. PIN-attempt counter races under parallel /recover requests.
  3. Zero-byte uploads pass /api/v1/upload.
  4. SVG-with-script can pass the magic-byte check (consider CSP +
     X-Content-Type-Options on /media/*).

Stack-internal docs live in e2e/README.md (UA tier table, Samsung
Internet escalation tiers A/B/C, debugging tips, roadmap).

Final tally: 134 passed / 0 failed / 9 skipped (test.fixme stubs for
not-yet-shipped gestures and one UI-upload-flow investigation).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-16 19:37:11 +02:00