Fix bug when using an private aliased factory service #17942
Conversation
| @@ -131,6 +131,6 @@ private function updateFactoryServiceReference($factoryService, $currentId, $new | |||
| return; | |||
| } | |||
|
|
|||
| return $currentId === $factoryService ? $newId : $currentId; | |||
| return $currentId === $factoryService ? $newId : $factoryService; | |||
There was a problem hiding this comment.
$currentId is the current ID of the removed alias, while $factoryService is the original ID of the factory service.
If the factory service equals the ID of the removed alias, it should be updated to become the actual service ID. However, if the factory service does not equal the ID of the removed alias, the original value should be used.
With the old code, it was always replacing the original code, either with the actual definition ID or the ID of the removed alias (which was causing the troubles people were reporting).
|
👍 Oh my God, what a silly mistake. Thanks Wouter for the fix! Status: Reviewed |
|
Was the test failing before? If not, a test would be nice to prevent this bug from happening again |
|
@iltar I've added a test now (although this is a very specific test, so I'm not 100% sure if we should include it). |
| @@ -45,6 +47,26 @@ public function testProcess() | |||
| } | |||
|
|
|||
| /** | |||
| * @group legacy | |||
There was a problem hiding this comment.
it's not really legacy in 2.3, but this makes things easier to merge to 2.7/2.8
|
Thank you @wouterj. |
…terJ) This PR was squashed before being merged into the 2.3 branch (closes #17942). Discussion ---------- Fix bug when using an private aliased factory service | Q | A | ------------- | --- | Bug fix? | yes | New feature? | no BC breaks? | no | Deprecations? | no | Tests pass? | yes | Fixed tickets | #17910, #17915 | License | MIT | Doc PR | - /cc @xabbuh Commits ------- de406c0 Fix bug when using an private aliased factory service
/cc @xabbuh