This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
flood of warnings about builtins with -traditional
- To: gcc-bugs at gcc dot gnu dot org
- Subject: flood of warnings about builtins with -traditional
- From: Zack Weinberg <zack at wolery dot cumb dot org>
- Date: Mon, 8 May 2000 12:40:11 -0700
$ cat test.c
int main(void) { return 0; }
$ ./cc1 -traditional -quiet test.c
<built-in>:0: warning: type of external `__builtin_constant_p' is not global
<built-in>:0: warning: type of external `__builtin_return_address' is not global
<built-in>:0: warning: type of external `__builtin_frame_address' is not global
<built-in>:0: warning: type of external `__builtin_alloca' is not global
<built-in>:0: warning: type of external `__builtin_ffs' is not global
<built-in>:0: warning: type of external `alloca' is not global
<built-in>:0: warning: type of external `ffs' is not global
<built-in>:0: warning: type of external `_exit' is not global
<built-in>:0: warning: type of external `bzero' is not global
<built-in>:0: warning: type of external `bcmp' is not global
<built-in>:0: warning: type of external `__builtin_abs' is not global
<built-in>:0: warning: type of external `__builtin_fabsf' is not global
[etc. etc. - 83 warnings total]
It also happens for implicit declarations in the code - e.g.
int main(void) { puts("hello world"); return 0; } would get the same
warning for puts.
We didn't do this in 2.95, I can't figure out what the warning means
even after reading the relevant chunk of c-decl.c, and it interferes
with testing -traditional. Also, warning about builtin declarations
is user-hostile, period.
zw