https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87096
Martin Sebor <msebor at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |wrong-code
Status|UNCONFIRMED |NEW
Last reconfirmed| |2018-08-24
CC| |msebor at gcc dot gnu.org
Ever confirmed|0 |1
Known to fail| |4.7.4, 5.3.0, 6.4.0, 7.3.0,
| |8.1.0, 9.0
--- Comment #1 from Martin Sebor <msebor at gcc dot gnu.org> ---
Confirmed. This goes back to the original folding of snprintf to memcpy.
$ cat f.c && gcc -O2 -S -Wall -fdump-tree-optimized=/dev/stdout f.c
int f (char *d)
{
return __builtin_snprintf (d, __INT_MAX__ + 1LLU, "");
}
f.c: In function ‘f’:
f.c:3:53: warning: zero-length gnu_printf format string [-Wformat-zero-length]
return __builtin_snprintf (d, __INT_MAX__ + 1LLU, "");
^~
;; Function f (f, funcdef_no=0, decl_uid=1756, cgraph_uid=0, symbol_order=0)
f (char * d)
{
<bb 2>:
__builtin_memcpy (d_2(D), "", 1);
return 0;
}
GCC 8 warns on this case but still folds it into memcpy and return 0.
f.c:3:10: warning: specified bound 2147483648 exceeds ‘INT_MAX’
[-Wformat-truncation=]
return __builtin_snprintf (d, __INT_MAX__ + 1LLU, "");
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
