9 Commits (dev)

Author	SHA1	Message	Date
Shlee	07afdc8f4f	Merge pull request #7839 from pixelfed/fix/edit-mentions-hashtags ... Re-derive mentions and hashtags when a post is edited	1 week ago
Ashley	463608e283	Re-derive mentions and hashtags when a post is edited ... Editing a local post saved the new caption but never re-lexed it, so Mention/StatusHashtag rows (written only at creation by StatusEntityLexer) went stale. UpdateNote.object.tag and scopeToAudience('cc') kept emitting the old entity set while content HTML regenerated live: new remote mentions were never delivered/notified, removed mentions stayed cc'd, and edits never entered newly added hashtag feeds remotely. Add StatusEntityReconciler::reconcile(), which re-runs the Extractor over the edited caption and reconciles rows using the same primitives as the lexer: create/restore Mention rows (dispatching MentionPipeline for newly added ones only), soft-delete dropped mentions, firstOrCreate StatusHashtag rows and hard-delete dropped ones (never touching the shared Hashtag row). Wired into UpdateStatusService::call() only when the caption changed, so captions whose entities are unchanged keep identical behavior and dispatch nothing. Edit history and media handling are untouched. Scoped to local edits. StatusRemoteUpdatePipeline (incoming remote edits) has the same staleness but its own media/NSFW handling and tests; left as a follow-up. Live-deriving tag entries in UpdateNote is unnecessary once rows are reconciled; left as an optional defensive follow-up. Verification (run from the worktree via ddev, sqlite :memory:): - tests/Feature/Status/EditReLexEntitiesTest.php: 2 passed. - Red/green: with the reconcile call commented out, the swap/payload case fails (new mention row absent, tag/cc stale) and the control case passes; with it restored both pass. - tests/Feature/Status + tests/Feature/StatusPipeline: 29 passed. - pint --test on the changed files: clean.	1 week ago
Ashley	1df355b82a	Return 404 for reply permalinks whose parent was deleted ... Status::parent() returns null once a parent is soft- or force-deleted, so the reply permalink page dereferenced null and threw a 500 for guests. StatusController::show() now aborts 404 when a reply's parent is gone, and reply.blade.php guards the parent/grandparent blocks so rendering can never null-deref. Verified red->green: before the fix the soft-delete and force-delete cases returned 500 (Call to a member function parent() on null); after the fix all three cases pass (parent alive 200, soft-deleted 404, force-deleted 404).	1 week ago
Ashley	dfa054f60a	Fix test collisions that broke the full suite ... Resolve name clashes that only surfaced when running the whole suite together rather than files in isolation: - Drop the redundant uses(TestCase::class) from two Feature tests; the TestCase binding already comes from tests/Pest.php, and binding it twice throws TestCaseAlreadyInUse. - Rename the per-file seedFilter() helpers to seedCustomFilter() and seedUserFilter() so the two files no longer redeclare the same global function. - Configure a captcha driver in two CaptchaManagerTest cases so they assert per-surface behaviour instead of the unconfigured-driver path.	1 week ago
Shlee	f121317748	Merge pull request #7810 from pixelfed/fix/nsfw-feed-reconcile-gates ... Apply public feed eligibility gates when restoring edited status	1 week ago
Ashley	3dcd5bce37	Apply public feed eligibility gates when restoring edited status	1 week ago
Ashley	1b3f39a177	Fix private child reply visibility leak on reply permalink page	1 week ago
Your Name	c32642e686	Clear place status caches after save to close invalidation race ... handleImmediateAttributes cleared the old and new per-place status-id caches while place_id was only mutated in memory, before $status->save() committed the change. A concurrent PlaceController::show read landing in that window re-warmed the old place cache from the pre-move DB state, pinning a stale membership list for the 4-day TTL. Move both PlaceService::clearStatusesByPlaceId calls to after save() so the cache can only be re-warmed from post-move DB state. Add a regression test that re-warms the old place cache during the updating event (fired inside save, before the DB write) and asserts it does not survive.	1 month ago
Your Name	acc5c930d6	Invalidate place status caches on API publish and location edits	1 month ago