Static callable

[LiamKearn]

Is there a syntax for a non-capturing closure?

eg. static Closure(int): void?

[ondrejmirtes]

No, and I'm not sure what you'd like to achieve with it.

[mad-briller]

code that calls Closure::bind on a static closure will fail, so it could be useful to specify "this doesn't rebind the closure" as a guarantee that static closures are allowed

https://3v4l.org/2Vhif
https://phpstan.org/r/cc935152-4334-4d95-a5fa-b36f33eec06f

it is a bit of an esoteric use case though

[LiamKearn]

Using (new ReflectionFunction($userSuppliedCallback))->__toString() for signature based hashing, binding of anything pretty much throws that and the use-case for $userSuppliedCallback out the door. It ain't pretty, it ain't stable but it's likely less code and faster than collecting all the different signature via reflection methods (e.g. ReflectionFunctionAbstract::getReturnType)

More than open to simple signature based hashing solutions.

[peterfox]

Had this question myself. Currently working with Pest testing framework and it would be useful to tell static tools like Rector to avoid converting \Closures to static for the tests e.g.

test('it works, static function () {

});

Breaks the testing command because under the hood the test function causes the \Closure to be bound to an object which isn't possible as it's a static method. Something like:

/**
 * @param \Closure<TestCase> $closure
 */

To signify what the closure will be bound to would be useful.

Equally in Laravel world we have things like this come up with PHPStan and macros as well.

[ondrejmirtes]

This is simply a Pest bug. Closure that doesn't use $this can simply always be made static. If Pest tries to bind it then it's a Pest issue.

Pest can ask ReflectionFunctionAbstract::isStatic() before trying to bind the closure: https://3v4l.org/dHfU6

[peterfox]

@ondrejmirtes I agree and I've highlighted it as a bug in Pest.

But then a tool like Rector has a rule to convert functions to static https://github.com/rectorphp/rector/blob/main/docs/rector_rules_overview.md#staticclosurerector. But it can't be determined by the parameters of the test() function that only non-static closures are allowed.

It feels like it would be worthwhile having a way to know the expected context for a closure.

Would avoid people having to define $this in closures where $this isn't the class the closure is defined in. e.g.

/** @var SomeObject $this */

This is a common problem I see in even level 0 PHPStan when it comes to closures.

[ondrejmirtes]

So in PHPStan 1.11 you'll be able to use @param-closure-this to tell PHPStan what $this is going to be inside the passed closure: https://twitter.com/OndrejMirtes/status/1770803825982497186

Pest is going to take advantage of this: https://twitter.com/enunomaduro/status/1770828317039497630

I also added a rule to check if a static closure is passed into such a parameter and report it as an error: phpstan/phpstan-src@2b8bbde

[peterfox]

Thank you @ondrejmirtes that's an amazing update 🎉