https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80354
Alejandro Colomar <colomar.6.4.3 at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |colomar.6.4.3 at gmail dot com
--- Comment #10 from Alejandro Colomar <colomar.6.4.3 at gmail dot com> ---
In some cases, that warning saved me from writing dangerous code (using file
paths), and it's been very useful. But casting to void means one and only one
thing: _I DO know what I'm doing and I do NOT care, so STFU!_
And if the casting to void is dangerous (ie. casting to void the result of
storing a path), that's my fault.
I currently have a program where I want, for some reason, to truncate the
output to an 80-char line:
#include <stdio.h>
#define LINE_SIZE (80)
int main(int argc, char *argv[])
{
char truncated[LINE_SIZE];
(void)snprintf(truncated, LINE_SIZE, "I DO want to truncate this line "
"to a 80 char line, and discard any remaining text so "
"that it is not displayed or stored anywhere because I
"
"don't care about it. "
" Usually one would check for the return value of "
"snprintf, but in this precise situation, I do not "
"care about it, so the explicit cast to (void) should "
"silence the warning, just like it does in other "
"warnings such as -Wunused-variable.");
printf("%s\n", truncated);
return 0;
}
But in the same project I have paths and that kind of uses of snprintf where I
still want to have the warnings enabled (and even more, -Werror so that they
never compile):
if (snprintf(file_name, FILENAME_MAX, "%s/%s", file_path, saved_name))
{
goto err_path;
}
Many other warnings are supressed with (void), why is this one so special?
Anyways, I'm reporting a new bug, because this one is "RESOLVED INVALID".
