Surely the point of writing an if-block testing the equality of the pointers is precisely because the code isn’t assuming that they are the same - but maybe it has something it wants to do if it finds itself running somewhere where they are.
Writing code which assumed those pointers were equivalent would be bad, but this code doesn’t do that. Instead, it looks like the compiler assumes that the pointers won’t be the same and refuses to let you entertain the possibility at all.
> Instead, it looks like the compiler assumes that the pointers won’t be the same and refuses to let you entertain the possibility at all.
and indeed it should - it’s sane to use the address of a stack variable for the purposes of writing something into it, but assuming that it forms some sort of array in memory is sure to cause pain if you rely on that assumption.
I agree, when I can see the addresses match it would do my head in if the compiler says false, its unexpected, its surprising. If i write bad code, fine I'm used to that and I can debug it, if there are these new 'gotchas' thats less transparent. I've not used c11 yet, and don't likes the sound of it!
Writing code which assumed those pointers were equivalent would be bad, but this code doesn’t do that. Instead, it looks like the compiler assumes that the pointers won’t be the same and refuses to let you entertain the possibility at all.