Laravel 11 exposes MariaDB as a dedicated 'mariadb' driver, so
config('database.default') === 'mysql' checks silently misclassified
MariaDB as the non-mysql (postgres) branch.
- Add App\Util\Database\DatabaseDriver with isMysqlLike()/isPgsql()
plus db_is_mysql_like()/db_is_pgsql() global helpers.
- Route all database.default driver checks through the helpers so
MySQL and MariaDB are treated as one group.
- Use '' (not null) for share/compose caption+rendered, valid whether
the column is nullable or NOT NULL (it is NOT NULL on MySQL/MariaDB).
- Guard pgsql strtolower() in registration against missing fields.
- Scope CustomEmoji::duplicateShortcodes to the grouped column for
Postgres GROUP BY validity.
- Remove stale Postgres guard in status:dedup; use havingRaw for
cross-driver HAVING.
The status- and account-deletion jobs loaded whole collections with
->get() and then looped, running a per-row Notification lookup inside
each iteration.
- StatusDelete / RemoteStatusDelete: resolve associated DirectMessage and
MediaTag ids, fetch their notifications in a single whereIn query,
clear each via cursor (NotificationService::del must run per row for
cache/redis cleanup), then bulk delete the DMs and media tags.
- DeleteAccountPipeline / DeleteRemoteProfilePipeline: stream Story and
Collection deletions with cursor() instead of loading every row into
memory. Per-row file unlink and item deletes are preserved.
Adds StatusDeleteCleanupTest covering DM + notification cleanup, media
tag + notification cleanup, and the no-associations case.
MediaDeletePipeline skips deletion when media->status_id is set. status_id has
no FK/cascade, so deleting a status never clears it, and the delete jobs
dispatched by the status-delete paths were always skipped, leaking media files.
Detach media (status_id = null) before dispatching the delete in StatusDelete,
RemoteStatusDelete and DeleteRemoteStatusPipeline, so the row is genuinely
orphaned by the time the guard checks it and the deletion proceeds.
Also adds a status:media diagnostic command that dumps all metadata for a
media id (DB columns, computed URLs, attachment state, parent status including
the dangling status_id case, owner, metadata, and an optional live URL check).
1. Add FractalService with static item() and collection() helpers
replacing 22 call sites that repeated the 4-line Fractal Manager
+ ArraySerializer boilerplate.
2. Add AccountInterstitial::createFromStatus() factory method
consolidating 4 identical 15-line blocks that create interstitials
with status metadata.
3. Add NotificationService::createNotification() to handle the
repeated pattern of creating, caching, and registering a
notification in the recipient's feed.
4. Add NotificationService::firstOrCreateNotification() for
idempotent notifications (share/boost, mention) that should
only notify once per actor+action+item combination.
Add ActivityPubDeliveryService::pool() using Laravel's Http::pool() to
consolidate the duplicated delivery pattern found across 10 jobs.
Updated jobs:
- StatusActivityPubDeliver
- StatusDelete
- StatusLocalUpdateActivityPubDeliverPipeline
- FanoutDeletePipeline
- SharePipeline
- UndoSharePipeline
- StoryFanout
- StoryExpire
- StoryDelete
- ProfileMigrationDeliverMoveActivityPipeline
The shared method accepts a Profile (sender), audience (inbox URLs),
and activity (payload array), handling signing, user-agent, timeout,
and concurrency in one place. No direct Guzzle usage remains in
app/Jobs/.
Move all Eloquent models from the app/ root directory to app/Models/
for consistency with modern Laravel conventions. The project already had
54 models in App\Models; this migrates the remaining 52 legacy models.
Changes:
- Move 52 model files from app/ to app/Models/
- Update namespace declarations in each model
- Update all ~1000 import references across the codebase
- Add Relation::morphMap() in AppServiceProvider for backward
compatibility with existing polymorphic database records
- Add missing HasSnowflakePrimary imports for models that relied
on same-namespace resolution
Previous config (timeout=5, tries=1) was too aggressive — a single
transient failure would permanently lose the status publication.
New config:
- timeout: 5 → 30 (sufficient for DB check + job dispatch)
- tries: 1 → 3 (recover from transient Redis/DB issues)
- maxExceptions: 1 (don't retry actual bugs)
- backoff: [5, 10] (exponential delay between retries)
Replace direct GuzzleHttp\Client and Pool usage in fanoutDelete()
with Laravel's Http::pool() facade. This provides:
- Testability via Http::fake() in tests
- Consistent timeout/retry configuration
- No direct Guzzle dependency in application code
- Proper integration with Laravel's HTTP client features
Convert all 273 short facade alias imports (e.g. 'use Cache;') to their
fully-qualified class names (e.g. 'use Illuminate\Support\Facades\Cache;')
across 193 files.
This resolves 643 PHPStan 'class.notFound' errors caused by the static
analyzer being unable to resolve global aliases, and aligns with modern
Laravel conventions. It also unblocks removing the aliases array from
config/app.php in a future change.
All 107 tests pass.
Replace all deprecated helper function calls:
- str_slug() → Str::slug()
- starts_with() → str_starts_with()
- ends_with() → str_ends_with()
- array_first() → Arr::first()
- array_last() → Arr::last()
- array_flatten() → Arr::flatten()
Remove laravel/helpers package from composer.json as it is no longer
needed and will not be maintained for Laravel 13.
PHP 5.5.9 adds the new static `class` property which provides the fully qualified class name. This is preferred over using strings for class names since the `class` property references are checked by PHP.
Posting is done in two steps - first the media is uploaded and that begins a processing pipeline. When the media is finally
processed, the media pipeline sets the cdn_url to the final image (in the cloud if configured).
At any time, the user can click post. This begins a separate NewStatusPipeline to deliver the status over ActivityPub.
This causes a race. If the NewStatusPipeline runs before the media finishes processing, then the status is delivered with
the un-processed (and un-optimized images). If pixelfed is set to use cloud storage, then the ActivityPub message also incorrectly
uses the local media.
This commit fixes the race condition by waiting for all the media to be processed before sending the message over ActivityPub.
A new publish_delayed column is set per-post, indicating whether the
status should be delayed until all media is processed.
Then, the NewStatusPipeline is called twice per post -
When the media finishes processing and when the status is posted.
Both times, the pipeline will check to see if it's valid to post
If so, the pipeline will try to set publish_delayed to false
and only if it is the first time doing so, will it publish the post
Your Name
Shlee
Daniel Supernault
Shift
Anil Kulkarni