On Fri, Nov 25, 2011 at 05:40:51PM +0100, Joerg Sonnenberger wrote:
> > And why should the/any static analyzer have to figure that out when
> > coding it right makes it a nonissue?
>
> Because that's what is making warnings useful. To not complain about
> non-issues.
Like this?
const char *
foo(const char *s)
{
size_t len;
char *t;
len = strlen(s);
if (len >= 2 && s[0] == s[len-1]) {
t = s+1;
} else {
t = s;
}
return strchr(t, '"');
}
We require that this be written with a "const" on t, even though it's
a "non-issue". Part of the price of checking is the obligation to
provide code that has the details right.
--
David A. Holland
dholl...@netbsd.org
