Completely remove disabled functions from function table

[nikic]

Currently, disabling a function only replaces the internal function handler with one that throws a warning, and a few places in the engine special-case such functions, such as function_exists. This leaves us with a Schrödinger's function, which both does not exist (function_exists returns false) and does exist (you cannot define a function with the same name). In particular, this prevents the implementation of robust polyfills, as reported in https://bugs.php.net/bug.php?id=79382:

if (!function_exists('getallheaders')) {
    function getallheaders(...) { ... }
}

If getallheaders() is a disabled function, this code will break.

I don't think it is possible to reconcile this. Making function_exists() return true would fix this case, but would break other cases, that check function_exists() before calling a function.

I think the only way to make this fully consistent is to actually remove the function from the function table. I don't think this has any disadvantages, apart from throwing a more generic error message.

This will make the ReflectionFunction::isDisabled() method useless, as it will no longer be possible to create a ReflectionFunction of a disabled function. Similarly, the $exclude_disabled parameter of get_defined_functions() will be hardcoded to true.

[cmb69]

Thanks, I like that very much. Should disable_classes be treated in a similar manner (see also https://bugs.php.net/79516).

[nikic]

Thanks, I like that very much. Should disable_classes be treated in a similar manner (see also https://bugs.php.net/79516).

I agree, but the class case will probably have some complications, because we often reference class entries internally. We must make sure that the class entry is not destroyed.

[staabm]

should ini_get('disable_functions') return a empty value and emit a deprecation warning in case this change gets merged?

[nikic]

@staabm Why? ini_get() should return the value it has been set to, as usual.

[staabm]

oh, then I misread the PR. had the impression the disabled-function mechnics will be removed.

fine then.

[nikic]

@staabm Ooh, yes, I can see where that came from :) Renamed to be a bit more obvious.