Resolved a bug where a soft-deleted object isn't remove from the ObjectManager #2930
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2930 +/- ##
==========================================
+ Coverage 78.51% 78.53% +0.01%
==========================================
Files 168 168
Lines 8782 8790 +8
==========================================
+ Hits 6895 6903 +8
Misses 1887 1887 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
@phansys it'd be nice if the tests could be run on new commits. Also, I believe the CHANGELOG will need to be something you commit, as I don't have the target version. |
| @@ -129,10 +137,27 @@ public function onFlush(EventArgs $args) | |||
| $postSoftDeleteEventArgs | |||
| ); | |||
| } | |||
|
|
|||
| $this->softDeletedObjects[] = $object; | |||
There was a problem hiding this comment.
More than a suggestion, this is a question or a request for analysis.
Given storing the list of deleted objects in this array could lead to high memory consumption, do you evaluated a way to mitigate this by using the object references (like the ones produced by spl_object_id(), spl_object_hash(), or a combination of the model class and its id)?
I don' t know if this can produce better results, but I think this is a consideration we should take into account.
There was a problem hiding this comment.
Being that these objects are unset (cleared from memory) postFlush, I really don't see it as much of a concern - certainly not enough to introduce spl_object_id complexity and the potential to expose bugs through that.
Co-authored-by: Javier Spagnoletti <phansys@gmail.com>
Co-authored-by: Javier Spagnoletti <phansys@gmail.com>
Co-authored-by: Javier Spagnoletti <phansys@gmail.com>
Currently, after you remove an entity, it's still queryable by the pk, as it's cached in the
ObjectManager. This PR fixes this bug by removing itpostFlush. See test for more details.