Add failing test with arguably the correct behaviour.

[Danack]

Adds a failing test for discussion.

The behaviour seems consistent with the principle of least surprise.

https://bugs.php.net/bug.php?id=64346 for reference.

[Girgias]

I'm not sure how to go about fixing it. This seems to be pretty orthogonal to function autoloading, but will spend some brain cells on this.

[Danack]

Well, I'm not even sure that it's the right behaviour.

Semi-sane question, is there any infrastructure inside the engine for 'mapping' a closure to a function? So something like:

<?php

namespace Foo;

function custom_autoloader($name)
{
	if ($name ===  'Foo\count') {
		return count(...);
	}
}

autoload_register_function("custom_autoloader");

Or would that be a huge chunk of work that needs to be done separately?

Obviously this would be nice for nefarious purposes. But it might also be useful for the fallback to root problem..

[Girgias]

Well, I'm not even sure that it's the right behaviour.

Semi-sane question, is there any infrastructure inside the engine for 'mapping' a closure to a function? So something like:

<?php

namespace Foo;

function custom_autoloader($name)
{
	if ($name ===  'Foo\count') {
		return count(...);
	}
}

autoload_register_function("custom_autoloader");

Or would that be a huge chunk of work that needs to be done separately?

Obviously this would be nice for nefarious purposes. But it might also be useful for the fallback to root problem..

I don't think there is, but doing $closure = func(...) emits the ZEND_CALLABLE_CONVERT opcode so it might be possible to add a way to track it?

I do wonder if it is possible to erase the RUNTIME cache via a definition (and/or throw) by doing a check in ZEND_DECLARE_FUNCTION, this might need to provide additional info at compile time. But it may be doable?