[Php81] ArrayToFirstClassCallableRector should not skip non-public methods from owning scope #7760
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
If a class contains array callbacks with a non-public method, they are not converted into first-class callables. This is due to a fix for a bug that was specific to skipping non-public methods referenced from *outside* the owning class - there was no previous coverage for what should happen when a callable is referenced within a valid scope.
If an array callable is explicitly referencing `$this`, then it should always be safe to convert it to a first-class callable even if the method is private / protected. This was the original behaviour of the 8.1 fixers, but it changed in v1.1.1 following a fix for cases where a private / protected method was being referenced from outside the owning scope.
acoulton
commented
Dec 18, 2025
samsonasik
reviewed
Dec 18, 2025
...81/Rector/Array_/ArrayToFirstClassCallableRector/Fixture/some_class_with_own_private.php.inc
Outdated
Show resolved
Hide resolved
samsonasik
reviewed
Dec 18, 2025
...81/Rector/Array_/ArrayToFirstClassCallableRector/Fixture/some_class_with_own_private.php.inc
Outdated
Show resolved
Hide resolved
acoulton
added a commit
to Behat/Behat
that referenced
this pull request
Dec 18, 2025
We still had a few uses of array callables within our internal code, which I found when starting to look at adding parameter types to internal methods. Rector has a rule to fix these as part of the PHP 8.1 set that we are running, but it had skipped over them because they were not public methods, even though they are defined and used in their own scope. I think this may actually be a Rector bug - see rectorphp/rector-src#7760 for the patch which I have temporarily used locally to convert these. Refactoring them to first-class callables will help Rector (and possibly phpstan) to make better decisions about parameter & return types in future.
Co-authored-by: Abdul Malik Ikhsan <[email protected]>
samsonasik
approved these changes
Dec 18, 2025
Member
|
Thankl you 🥳 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
ArrayToFirstClassCallableRector currently skips over all callables that reference non-public methods, even if the callable is defined in the same scope as the method (e.g.
[$this, 'somePrivateMethod']).There are no existing test fixtures that demonstrate this behaviour either as an expected refactor or as an expected skip (the examples are all using public methods, or references from outside the owning scope).
It looks like this was the behaviour before #5929 was merged into v1.1.1. That was a bugfix for rectorphp/rector#8659 which was specific to cases where a private method is referenced from outside the owning scope.
I'm not sure if it was intentional to also stop refactoring callables that are defined and used within the same class? As far as I can tell, those should always be safe to convert to first-class callables.
I've added a test fixture to prove that these nodes are not currently refactored, and will push a second commit with a fix.
If you're not convinced about changing the default behaviour, then I could instead make this a configurable option for the rule?