A deep investigative review of the export concurrency (four parallel adversarial
reviewers) found that three rounds of race fixes had been chasing the wrong bug, and
that the model itself was the root cause. This replaces the model and fixes the real
data loss.
THE ACTUAL BUG (upload path, not the state machine)
`upload` checked `uploads_locked_at` BEFORE streaming the body — minutes, for a large
video — then committed the row without re-checking. A release landing mid-upload let the
row commit AFTER the export snapshot: the photo appeared in the live feed and was
PERMANENTLY missing from the keepsake, with nothing to regenerate it. release_gallery's
comment claims to prevent exactly this; it never did. Every prior round fixed the DB
state machine, and the leak was upstream of it.
Fix: re-assert the lock under `SELECT ... FOR SHARE` INSIDE the commit tx. That row lock
conflicts with release_gallery's `UPDATE event`, so either the upload commits first (and
the release — hence the snapshot — is ordered after it) or the release wins and the
upload is rejected as `uploads_locked` (reversible; the client keeps the blob).
Regression test verified NON-VACUOUS: with the fix reverted it fails, reporting accepted
uploads missing from the keepsake.
THE MODEL (migration 014)
"Which generation is current?" had no single answer: it was assembled from three
separately-written pieces across two tables (`export_released_at`, a per-job
`release_seq`, and cached `export_{zip,html}_ready` flags). Keeping them in agreement
took ~10 hand-written guards; each review round found one more path that slipped through.
Now: ONE monotonic `event.export_epoch`, bumped in the same UPDATE as any change to
`export_released_at`. Each job carries a copy. Downloadable IFF
`released AND job.epoch = event.export_epoch AND status = 'done'` — readiness DERIVED,
never stored. A retired-epoch worker is INERT BY CONSTRUCTION: losing a race can no
longer corrupt state, only waste work.
Deleted: both ready-flag columns, both ready-flip blocks, `if ready { continue }`, the
reopen seq-bump, open_event's transaction, and the cross-table claim guard.
That last one was also UNSOUND: under READ COMMITTED, a blocked UPDATE re-evaluates its
WHERE against the updated target row but answers subqueries on OTHER tables from the
original snapshot — so the previous `EXISTS (SELECT ... FROM event ...)` claim guard
could admit a claim against an already-reopened event while RETURNING the post-bump seq.
Every guard is now same-row (`epoch = $n`), which EPQ re-evaluates correctly.
OTHER REAL DEFECTS FIXED
- release_gallery committed the release and THEN awaited the enqueue inside the handler
future. Axum drops that future on client disconnect → event released, uploads locked,
ZERO export_job rows, host unable to retry ("bereits freigegeben"), restart-only
recovery. Now one tx; workers spawn (detached) after commit.
- No flush/fsync before the tmp→final rename. async_zip's close() never flushes the inner
file and tokio's File drop DISCARDS write errors — so a full disk silently produced a
truncated archive that was then marked done and published, after which the last good
generation was pruned. Now flush + sync_all, which also surfaces ENOSPC.
- Download read the ready flag and the file path in TWO queries; a reopen between them
served a keepsake that should 404. Now one read through the `export_current` view.
- `if !src.exists() { continue }` then `open()?` — a TOCTOU that turned a tolerated gap
into a hard failure of the ENTIRE keepsake (unretryable while released). Now open-first,
warn, and skip.
- Recovery was flag-driven and never checked the disk: a `done` row whose archive was gone
was a permanent dead end needing manual DB surgery. Now verifies the file and re-arms.
- Temp files/dirs leaked on every error path (the leak is what fills the disk that
corrupts the next archive). Now cleaned up.
- Export archives were not event-scoped (`viewer_tmp_` already was), so a second event
would collide on paths and one event's prune would delete another's live keepsake.
- Host deletes after release never regenerated the keepsake — a photo removed on request
lived on in the archive forever. Now bumps the epoch and rebuilds without it.
- claim_job swallowed DB errors as "someone else won", stranding a job pending at 0%.
- export_status reported retired-epoch rows (a progress bar that never moves).
- The startup sweep's single-instance assumption is now documented, not hidden.
Verified: backend 40 tests, svelte-check 0 errors, e2e typecheck clean,
e2e 160 passed / 1 skipped on chromium-desktop (incl. 2 new regressions; the
upload-acceptance one confirmed to fail without the fix).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adversarial re-review of the persona-audit + audit-followup rounds (6411747..)
found one HIGH and one MED regression plus LOW gaps. All fixed with coverage.
HIGH — export stale-keepsake resurrected by an open_event race
A reopen landing in the window between a *current* export worker's finalize_job
and its ready-flag flip cleared export_released_at + the ready flags but left the
export_job row `done` at the same release_seq. The seq-guarded flip then still
matched and re-set export_{zip,html}_ready=TRUE on a pre-reopen snapshot; the next
re-release read that stale TRUE and skipped regeneration (`if ready { continue }`),
serving a keepsake missing every upload from the reopen window — the exact data
loss migration 012 exists to prevent. Both ready-flip UPDATEs are now additionally
anchored on `export_released_at IS NOT NULL`, so a landed reopen makes the flip a
no-op and the re-release regenerates cleanly.
MED — queue dedup broke for reloaded items
loadQueue rebuilt QueueItems from IndexedDB without copying lastModified, which the
new addToQueue dedup keys on. A file re-selected after a page reload / PWA relaunch
missed the duplicate check and uploaded twice. Rehydration now carries lastModified
(extracted to a pure, tested entryToQueueItem helper).
LOW
- diashow: clear the upload-processed debounce timer in onDestroy (no stray
post-unmount /feed fetch).
- USER_JOURNEYS §9.5: document the reconnect-delta ban replay (hidden_user_ids /
uploads_hidden_at, migration 013), not just the live user-hidden SSE.
- e2e api-client: drop the misleading hide_uploads param from banUser — the backend
takes no body and always hides; strip the dead boolean at all call sites.
Tests
- Extract isReversibleLock (the terminal-403 KEEP-vs-PURGE-blob discriminator) into a
pure exported helper + unit tests, so the data-loss-critical branch is covered
without an XHR harness.
- entryToQueueItem unit tests lock the lastModified-carry regression.
- Document the export flip-race guard in the reopen/re-release spec (the sub-ms
finalize↔flip interleave isn't deterministically forceable with fast fixtures;
covered by the SQL guard + the end-to-end completeness test).
Verified: backend 40 tests, frontend 44 unit tests, svelte-check 0 errors,
e2e 156 passed / 1 skipped on chromium-desktop.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follows the perf + security + user-flow work with a role/persona audit (guest, host,
admin, projector) and fixes across three review rounds. Highlights:
HIGH
- Ban now replays on reconnect. A ban isn't a soft-delete, and the `user-hidden` SSE has
no replay, so a client that missed it (esp. the unattended diashow) kept cycling a
banned user's slides. New `uploads_hidden_at` (migration 013) + `hidden_user_ids` in
/feed/delta; feed + diashow evict those users. Applied even on a truncated delta.
MEDIUM
- Locked-upload data loss: a photo staged offline during a lock/release was purged as a
terminal 4xx and lost when the host reopened. New reversible `uploads_locked` error code;
the queue keeps the blob and auto-resumes on the `event-opened` SSE.
- Reopen after release now warns (ConfirmSheet) that it revokes the published keepsake.
- Host "forgotten-PIN" badge updates live (`pin-reset-requested` was broadcast but never
in KNOWN_EVENTS / subscribed); host page also refetches on `pin-reset` so a two-host
race can't hand out a conflicting PIN.
- Ban modal copy fixed (read-only ban, not "session ended"); Degradieren/Sperren/Entsperren
hidden on peer-host rows for non-admins (they always 403'd).
- Host dashboard shows live keepsake generation progress / ready state + link to /export.
- Admin JWT moved to sessionStorage (§11.1) to bound exposure on shared devices.
Export generation guard (H1 from the prior round, hardened): per-(event,type) `release_seq`
(migration 012) with seq-guarded claim/finalize/mark_failed/update_progress, per-generation
temp/final paths, download follows `file_path`, prune only strictly-older generations.
LOW: diashow coalesces upload-processed (avoids self-rate-limit); event-closed reconciles
galleryReleased; /recover gains a forgot-PIN request + drops a stale cached PIN on 401;
delta `>=` tie-break + 429 retry; misc copy/labels.
Adds e2e: ban-replay, upload-lock-code, and rewrites export-reopen-rerelease with a
data-completeness test. Reconciles USER_JOURNEYS §9/§11.
Verified: cargo build clean, 40 unit tests, svelte-check 0 errors, 33 frontend unit
tests, 155 e2e passing on chromium-desktop.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- compression_concurrency wired to boot config (state.rs) and removed as a
dead admin control.
- patch_config gains per-key integer/range validation so numerically-valid-but-
nonsensical values (-1/NaN/3.5) are rejected instead of silently reverting to
default at read time.
- clearAuth now also clears the display name (shared-device residual).
- e2e/Caddyfile.test mirrors prod security headers + the SSE encode-exclusion so
the test stack is representative and can catch header regressions.
- .gitignore: ignore root-level Playwright artifacts (test-results/, report).
- docs: USER_JOURNEYS §10 and SECURITY-BACKLOG updated to match the implemented
read-only-ban behavior and the export-path fix.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Re-review follow-ups (non-blocking quality items):
Tests:
- moderation H1 specs now assert exact `→ 403` instead of bare .rejects.toThrow(),
so a spurious 500 can no longer masquerade as "revocation worked".
- config: added case-insensitivity (upper/mixed-case placeholder) and the
len==32/31 boundary cases for validate_secrets.
- rate_limiter: added the trailing-comma empty-entry case for client_ip (must
fall back, not return "").
(Backend unit tests: 35 pass.)
Docs:
- README: note that with APP_ENV=production a placeholder .env makes the app
refuse to boot and Caddy wait unhealthy — reason is in `docker compose logs app`.
- SECURITY-BACKLOG + sse.rs comment: document that a mid-session ban does not
tear down an already-open SSE stream (new tickets are blocked; low blast radius).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
C1: reset_user_pin wrote to a non-existent column (pin_failed_attempts);
the real column is failed_pin_attempts, so every PIN reset 500'd. Fixed
the column name; new e2e (pin-reset.spec.ts) proves a reset returns a
usable PIN and the target can recover with it.
C2: config.rs::validate_secrets now rejects placeholder-ish secrets
(change_me/dev_secret/placeholder), enforces len>=32 in prod, and
requires a real ADMIN_PASSWORD_HASH. docker-compose.yml sets
APP_ENV=production so the guard actually runs. Corrected the false
"fixed" claim in SECURITY-BACKLOG.md. .env.example documents the rule.
Riders in these files (documented here since git can't split hunks):
- host.rs also carries the event-scoped ban_user fix and the
close_event/open_event no-op broadcast guard (medium).
- docker-compose.yml also adds ORIGIN (H6), app/frontend healthchecks with
Caddy waiting on health, and per-service memory limits (medium).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
From the 2026-06-27 audit branch (fix/audit-2026-06-27-critical-medium), whose
work was ~80% absorbed into main via the batch branches. This ports the pieces
that were NOT in main:
- compression.rs: cap image decode with image::Limits (12000x12000, 256 MiB
max_alloc) via ImageReader instead of image::open(). The upload body-size cap
bounds the file on disk but not the *decoded* dimensions, so a small
decompression-bomb image could OOM the box during decode/resize. Surgical port
of the audit's decode guard only (not its image/video semaphore split).
- docs/SECURITY-BACKLOG.md: the audit's triaged backlog, reconciled against main
(each item tagged done / open / contingent-on-signed-media). Records the still-
open items: host moderation UI gap, owner-delete SSE broadcast, quota
mount-detection (starts_with) + low-disk guard.
- docs/DECISION-media-auth.md: writes up the one real architectural divergence —
main serves media unauthenticated (ServeDir + UUID-capability) while the audit
built a signed/TTL'd gateway. Lays out the tradeoff (leaked URL = permanent vs
~24h access; banned-user access) and a recommendation, for a human decision.
Verified: cargo build clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A bundle of v0.16 platform work landing on top of the SvelteKit viewer.
Brings in 7 commits across docs, backend infra + features, frontend
plumbing, the live diashow, all UI surfaces (with dark mode), and a
viewer rebuild.
Headline features:
- Live diashow with two-queue policy (live drains first, shuffle as
fallback) and pluggable transitions. Design: docs/CONCEPT_DIASHOW.md.
- Dark mode: 'system' / 'light' / 'dark' preference, picked in the
onboarding step + the Account page, applied via Tailwind v4
class-based dark variant with FOUC guard.
- Host/Admin PIN reset with one-time PIN modal; admin may also reset
host PINs. Hosts may demote other hosts.
- Per-user dynamic storage quota enforced on upload + live widget in
My Account and on the upload screen. Toggleable per-area.
- All rate-limits + quotas individually toggleable from the admin
config UI (rendered as switches + a privacy_note textarea).
- Mobile-first gestures: long-press → context sheet, double-tap to
like with heart-burst. Buttons stay as desktop equivalents.
- Data mode (Saver vs Original) per device, applied across feed,
lightbox, and diashow.
- Per-event Datenschutzhinweis admin-editable, live-refreshed on all
clients via SSE event-updated.
- /api/v1/upload/{id}/original endpoint, /me/context + /me/quota.
Hardening (latent issues from the long-term review):
- Startup recovery for stuck compression / export jobs after a crash.
- Hourly cleanup of expired sessions + cold rate-limiter HashMap keys.
- ffmpeg 120s timeout with kill_on_drop (no more permit leaks).
- Per-user IndexedDB upload queue (no more cross-user leak on shared
devices); IDB schema bumped to v2.
- SSE reconnect uses exponential backoff (no more retry storm).
- PIN lockout no longer escalates — attempts reset when the cooldown
expires.
- soft_delete is now transactional and decrements total_upload_bytes
so quotas don't drift.
- pin-reset SSE handler filters by user_id so a host resetting Anna's
PIN doesn't clear Bob's cached PIN.
- Privacy note shown preformatted; admin-editable, ≤16 KiB cap.
Docs:
- New: FEATURES.md (role matrix), USER_JOURNEYS.md, IDEAS.md,
CONCEPT_DIASHOW.md, backend/migrations/README.md,
frontend/src/lib/README.md.
- Refresh: PROJECT.md, README.md, TEST_GUIDE.md, the two existing
CONCEPT_*.md banners.
- PROJECT.md, README.md, TEST_GUIDE.md: status line refreshed; rate-limiter
doc-vs-code drift fixed; HTML export section rewritten for the SvelteKit-
static viewer; SSE event names + new events documented; config seed block
extended with planned toggles + privacy_note; decision log entries added.
- docs/CONCEPT_HTML_VIEWER.md, docs/CONCEPT_MOBILE_UI.md: banner the design
intent as shipped; point at the source-of-truth code paths.
- docs/CONCEPT_DIASHOW.md: planned-then-shipped design for the live diashow
(two-queue policy, pluggable transitions, data-mode aware).
- docs/FEATURES.md: capability matrix by role (Guest / Host / Admin) plus
prose per area (auth, posting, feed, moderation, admin, export, gestures,
data mode, quotas, privacy note, extensibility).
- docs/USER_JOURNEYS.md: step-by-step flows for every supported scenario,
including PIN reset by host, data mode, privacy note, gestures, and the
admin toggles.
- docs/IDEAS.md: speculative extensions (global diashow, reactions,
multi-tenancy, animation pack, etc.) — explicitly out of v0.16 scope.
- backend/migrations/README.md, frontend/src/lib/README.md: codify the
"never edit a shipped migration" rule and the lib/ conventions
(one store per concern, gestures via actions, sheets via ContextSheet,
transitions as drop-in components).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Persistent bottom tab bar (Feed · FAB · Account) on all authenticated pages
- Upload FAB triggers bottom sheet (Galerie / Kamera) → navigates to composer
- Upload page redesigned as full-screen composer with thumbnail strip, textarea,
quick-tag chips, sticky submit button; bottom nav suppressed while composing
- Slim upload progress bar above bottom nav driven by queue state
- Feed: list/grid view toggle; list = chronological full-width FeedListCard;
grid = 3-col with search bar, autocomplete from loaded posts, filter chips
- Account page: role-gated dashboard links (Host / Admin); Konto section with
leave-confirm bottom sheet; no more per-page header nav icons
- Host dashboard: back arrow, collapsible sections, 2-col stats, user search
- Admin dashboard: back arrow, inner tab bar (Stats/Config/Export/Nutzer),
stacked config inputs with sticky save, new Nutzer tab
- BottomNav hidden on unauthenticated pages via isAuthenticated store
- FeedGrid: threeCol prop; OnboardingGuide upload step updated for FAB
- Concept docs added to docs/
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
