[Bug tree-optimization/69359] Warn about constant comparisons between pointers and arrays

["fw at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69359

--- Comment #2 from Florian Weimer <fw at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #1)
> I'd find warning about p <= weird, warning about p < a makes sense, as well
> as e.g.
> int g (void)
> {
>   int a[3], b;
>   int *p = f (a, &b);
>   return (p < a)
>          + (p < &a[0])
>          + (p > &a[3])
>          + (p < &b)
>          + (p > &b + 1);
> }
> Not a stage4 material though.

Agreed.

>From my perspective, p <= a is suspicious for the same reason why u <= 0 for an
unsigned variable u is suspicious.  If the range of the variable is limited and
the condition can only be true if there is equality, why not use == instead of
<=?  <= suggests that the programmer expected something to happen in the code
which is impossible, which could point to a logic bug or illegal program, and I
think it is fine to warn in such cases.