On Tue, Aug 04, 2020 at 03:45:11PM -0700, Mike Stump wrote:
> On Aug 4, 2020, at 3:08 PM, Marek Polacek via Gcc-patches
> <gcc-patches@gcc.gnu.org> wrote:
> >
> > That works well if you know where to expect an error. But if you don't,
> > it's
> > worse. E.g.,
> >
> > // { dg-xfail-if "" { *-*-* } }
> > int i = nothere; // demonstrates something that errors out
> > // { dg-error "" } didn't know where to put this
> >
> > only prints unexpected failures, but no unexpected successes. I guess
> > that's
> > OK though, at least for now, so I'll drop dg-accepts-invalid.
>
> There are two cases, either you get an error message that is wrong, and you
> can use:
>
> strncpy (p, s, 60); /* { dg-bogus "-Wstringop-truncation" } */
>
> or, you don't get an error, but you should:
>
> A foo(void i = 0); // { dg-error "incomplete type|invalid use" }
>
> ? Do you have an example of a specific case that doesn't work? I'm not sure
> I'm following.
Sorry for being unclear. Say you have
// PR c++/55408
struct foo {
template<int*>
void bar();
};
template<int*...>
void foo::bar() {}
int main()
{
extern int i;
foo {}.bar<&i>();
}
which we wrongly accept. It might be unclear what line to put that dg-error
on. So you put it at the end of the file. Then when we start issuing an error
for the testcase, you will just get a FAIL, not an XPASS. That might be
confusing if that file isn't in unfixed/.
Maybe it's not a real problem though -- upon inspection you'll find the
dg-error line and just tweak the testcase as needed.
Marek
