https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120526
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
>But clearly tuples[i].str is not NULL.
Not to the compiler since it looks like: ...
```
_1 = MEM[(const char * *)_24];
*_15 = 0;
size_16 = strtoimax (_1, &end, 10);
if (size_16 < 0)
goto <bb 4>; [0.73%]
else
goto <bb 8>; [99.27%]
if (_1 == 0B)
goto <bb 5>; [100.00%]
else
goto <bb 6>; [0.00%]
<bb 5> [local count: 42423513]:
stderr.1_3 = stderr;
fprintf (stderr.1_3, "oops %s\n", 0B);
```
strtoimax's first argument not being marked as nonnull.
Adding:
```
extern intmax_t strtoimax (const char *__restrict __nptr,
char **__restrict __endptr, int __base) __attribute__ ((__nothrow__ ,
__leaf__, nonnull(1)));
```
Fixes the warning and the missed optimization.
Let me see why glibc didn't mark strtoimax/strtoumax, etc. as nonnull.
