Bug in <complex> header in trunk from malformed "requires" statement, causes compilation errors if included

Bug in <complex> header in trunk from malformed "requires" statement, causes compilation errors if included

[Gavin Ray]

Hello all,

Hit a very confusing bug this morning. I'm using Clang with GCC
upstream libstdc++ from jwakely's repo.
The following program fails to compile:

#include <complex>
int main() { return 0; }

It appears there was an initial commit which added "requires" clauses,
and then a fixup commit later:
- libstdc++: Fix up <complex> for extended floating point types
[PR107649] · gcc-mirror/gcc@0e2c551 (github.com)

But there seems to be one last malformed "requires" statement in trunk:
https://github.com/gcc-mirror/gcc/blob/59cc4da605e5cb8e31e9f1d54ef2b5ba47fc8f88/libstdc%2B%2B-v3/include/std/complex#L2531

(Big thanks to Luke D'Alessandro for figuring this all out)
Thought I'd let you folks know.

Thank you!

Re: Bug in <complex> header in trunk from malformed "requires" statement, causes compilation errors if included

[Jonathan Wakely]

On Fri, 18 Nov 2022 at 18:59, Gavin Ray via Libstdc++
<libstdc++@gcc.gnu.org> wrote:
>
> Hello all,
>
> Hit a very confusing bug this morning. I'm using Clang with GCC
> upstream libstdc++ from jwakely's repo.
> The following program fails to compile:
>
> #include <complex>
> int main() { return 0; }
>
> It appears there was an initial commit which added "requires" clauses,
> and then a fixup commit later:
> - libstdc++: Fix up <complex> for extended floating point types
> [PR107649] · gcc-mirror/gcc@0e2c551 (github.com)

Yes, that was https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107649 but
isn't included in my gcc-latest builds yet.

>
> But there seems to be one last malformed "requires" statement in trunk:
> https://github.com/gcc-mirror/gcc/blob/59cc4da605e5cb8e31e9f1d54ef2b5ba47fc8f88/libstdc%2B%2B-v3/include/std/complex#L2531

Thanks.

Re: Bug in <complex> header in trunk from malformed "requires" statement, causes compilation errors if included

[Jonathan Wakely]

On Fri, 18 Nov 2022 at 21:14, Jonathan Wakely wrote:
>
> On Fri, 18 Nov 2022 at 18:59, Gavin Ray via Libstdc++
> <libstdc++@gcc.gnu.org> wrote:
> >
> > Hello all,
> >
> > Hit a very confusing bug this morning. I'm using Clang with GCC
> > upstream libstdc++ from jwakely's repo.
> > The following program fails to compile:
> >
> > #include <complex>
> > int main() { return 0; }
> >
> > It appears there was an initial commit which added "requires" clauses,
> > and then a fixup commit later:
> > - libstdc++: Fix up <complex> for extended floating point types
> > [PR107649] · gcc-mirror/gcc@0e2c551 (github.com)
>
> Yes, that was https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107649 but
> isn't included in my gcc-latest builds yet.
>
> >
> > But there seems to be one last malformed "requires" statement in trunk:
> > https://github.com/gcc-mirror/gcc/blob/59cc4da605e5cb8e31e9f1d54ef2b5ba47fc8f88/libstdc%2B%2B-v3/include/std/complex#L2531
>
> Thanks.

That's fixed now, thanks for reporting it.

Re: Bug in <complex> header in trunk from malformed "requires" statement, causes compilation errors if included

[Gavin Ray]

No problem, thank you for fixing and for maintaining the gcc-latest repo!

On Sat, Nov 19, 2022 at 10:09 AM Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
>
> On Fri, 18 Nov 2022 at 21:14, Jonathan Wakely wrote:
> >
> > On Fri, 18 Nov 2022 at 18:59, Gavin Ray via Libstdc++
> > <libstdc++@gcc.gnu.org> wrote:
> > >
> > > Hello all,
> > >
> > > Hit a very confusing bug this morning. I'm using Clang with GCC
> > > upstream libstdc++ from jwakely's repo.
> > > The following program fails to compile:
> > >
> > > #include <complex>
> > > int main() { return 0; }
> > >
> > > It appears there was an initial commit which added "requires" clauses,
> > > and then a fixup commit later:
> > > - libstdc++: Fix up <complex> for extended floating point types
> > > [PR107649] · gcc-mirror/gcc@0e2c551 (github.com)
> >
> > Yes, that was https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107649 but
> > isn't included in my gcc-latest builds yet.
> >
> > >
> > > But there seems to be one last malformed "requires" statement in trunk:
> > > https://github.com/gcc-mirror/gcc/blob/59cc4da605e5cb8e31e9f1d54ef2b5ba47fc8f88/libstdc%2B%2B-v3/include/std/complex#L2531
> >
> > Thanks.
>
> That's fixed now, thanks for reporting it.