Update dependency phpstan/phpstan to v2 #1459
Conversation
| @@ -452,6 +455,7 @@ public function getValue(object|null $object = null): mixed | |||
|
|
|||
| $closure = Closure::bind(fn (string $implementingClassName, string $propertyName): mixed => $implementingClassName::${$propertyName}, null, $implementingClassName); | |||
|
|
|||
| /** @phpstan-ignore-next-line function.alreadyNarrowedType instanceof.alwaysTrue */ | |||
There was a problem hiding this comment.
Are the stubs wrong in phpstan upstream? Or in psalm? 🤔
There was a problem hiding this comment.
@ondrejmirtes Can you check please?
There was a problem hiding this comment.
/** @phpstan-ignore-next-line function.alreadyNarrowedType instanceof.alwaysTrue */
This ignores everything on the next line, not just those identifiers - you need @phpstan-ignore function.alreadyNarrowedType instanceof.alwaysTrue.
There was a problem hiding this comment.
Alright so PHPStan is complaining that assert($closure instanceof Closure);. $closure is what gets returned by Closure::bind. When it can return null?
There was a problem hiding this comment.
This is according to upstream types: https://www.php.net/manual/en/closure.bind.php
there's a clear : ?Closure return type there.
Happy to patch PHP upstream, if we manage to demonstrate that there's no nullability there :D
The actual return locations:
There was a problem hiding this comment.
Alright, I'm going to return a benevolent union there. It's not going to complain in your case but it's not going to require handling the null either.
There was a problem hiding this comment.
@kukulich BTW, happy to iterate over this in future changes: should we move forward with this patch as-is, meanwhile?
There was a problem hiding this comment.
@Ocramius Yes, we can :)
|
Thanks @kukulich! |
No description provided.