random: Rely on free(NULL) being safe for random status freeing

[TimWolla]

No description provided.

[Girgias]

Can status ever be NULL or should an assertion be added?

[TimWolla]

See https://chat.stackoverflow.com/transcript/message/55792527#55792527 for context. I'm used to free()-style functions being a noop if passed NULL and consider it a good thing, because is avoids checking for NULL in every place that frees something that might be absent. In this case there might be no such place, but still this should just work in case there is.

[TimWolla]

I've just restructured the method a little, though. Instead of returning early, it checks for NULL before dereferencing status and just passes a possible-NULL status to pefree(). This pattern is a little more common for more complex structs.

[cmb69]

I think we can assume that free() and friends handle NULL as nop, even though custom allocators (USE_ZEND_ALLOC=0) may not. It might be a good idea to document that assumption (and others, like the infallability) in the phpinternalsbook.

[TimWolla]

It might be a good idea to document that assumption (and others, like the infallability) in the phpinternalsbook.

I've added an issue for that in the internals book repository.