fix(uploads): prevent cross-photo contamination from filename collisions and non-atomic writes (#931) (stable) (#934)

* fix(uploads): prevent cross-photo contamination from filename collisions and non-atomic writes (#931)

* test: pin the suffixed photo filename format in the NFD pipeline suite (#931)

* chore(deps): promote p-limit to a direct dependency for the watermark limiter (#931)

* test: make the suffix-uniqueness check deterministic-in-practice (#931)

* fix(uploads): widen the anti-collision suffix to 48 bits (#931)

* fix(uploads): hide staging files from list() + share one watermark limiter process-wide (#931)

* fix(uploads): reclaim orphaned staging files + revalidate watermark settings in queued jobs (#931)

---------

Co-authored-by: Paul Nothaft <[email protected]>
This commit is contained in:
Paul Nothaft
2026-08-01 12:29:23 +02:00
committed by GitHub
co-authored by Paul Nothaft
parent 7974b9c6d7
commit fc99e2b233
10 changed files with 246 additions and 35 deletions
+4 -3
View File
@@ -315,9 +315,10 @@ router.post('/:eventId/upload', adminAuth, requirePermission('photos.upload'), r
const crypto = require('crypto');
const uploadId = crypto.randomBytes(16).toString('hex');
// Counter base — same approximation as before. Strict uniqueness is
// already enforced by the filename template + DB unique index, so a
// small race here just retries a counter on conflict (rare).
// Counter base — a per-request approximation (concurrent upload
// requests can compute the same base; there is NO unique index on
// photos.filename). Uniqueness of the final path comes from the
// random suffix inside generatePhotoFilename (#931).
const existingCount = await db('photos')
.where({ event_id: eventId, type: photoType })
.count('id as count')