Greg Troxel <[email protected]> writes:
> I have long found it funny that the only related warning I've ever heard
> is due to gcc and the NetBSD ctype implementation. It's a macro (as is
> allowed):
>
> #define isspace(c) ((int)((_ctype_tab_ + 1)[(c)] & _CTYPE_S))
>
> which is very efficient when the input is EOF or 0..255. Well, it's
> efficient regardless, reading random memory. But it takes advantage of
> the UB leeway in the spec to avoid extra operations.
I meant to say: the warning is "array subscript of type char". Which is
either because char could be (usually is) "signed char", or because char
isn't a good kind of int for arrays. gcc has no built-in clue about
ctype functios; it's warning on the inlined code, correctly and
helpfully but somewhat inscrutably.
Coverity ought to do better. I have never been clear on whether
coverity aims to find UB/etc. in the sources, or if it is looking for UB
in the sources as built on a particular system.
I just ran on Debian 13 amd64:
#include <ctype.h>
int
main(){
int c = -1 * 1024 * 1024;
int r;
r = isalpha(c);
return r;
}
which got me a segfault. So it's not bounds checking. Which is allowed
and not a glibc bug!
