[r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[haochen.jiang]

On Linux/x86_64,

03cebd304955a6b9c5607e09312d77f1307cc98e is the first bad commit
commit 03cebd304955a6b9c5607e09312d77f1307cc98e
Author: Jason Merrill <jason@redhat.com>
Date:   Tue Apr 18 21:32:07 2023 -0400

    c++: fix 'unsigned typedef-name' extension [PR108099]

caused

FAIL: std/ranges/iota/max_size_type.cc execution test

with GCC configured with

../../gcc/configure --prefix=/export/users/haochenj/src/gcc-bisect/master/master/r14-159/usr --enable-clocale=gnu --with-system-zlib --with-demangler-in-ld --with-fpmath=sse --enable-languages=c,c++,fortran --enable-cet --without-isl --enable-libmpx x86_64-linux --disable-bootstrap

To reproduce:

$ cd {build_dir}/x86_64-linux/libstdc++-v3/testsuite && make check RUNTESTFLAGS="conformance.exp=std/ranges/iota/max_size_type.cc --target_board='unix{-m32}'"
$ cd {build_dir}/x86_64-linux/libstdc++-v3/testsuite && make check RUNTESTFLAGS="conformance.exp=std/ranges/iota/max_size_type.cc --target_board='unix{-m32\ -march=cascadelake}'"
$ cd {build_dir}/x86_64-linux/libstdc++-v3/testsuite && make check RUNTESTFLAGS="conformance.exp=std/ranges/iota/max_size_type.cc --target_board='unix{-m64}'"
$ cd {build_dir}/x86_64-linux/libstdc++-v3/testsuite && make check RUNTESTFLAGS="conformance.exp=std/ranges/iota/max_size_type.cc --target_board='unix{-m64\ -march=cascadelake}'"

(Please do not reply to this email, for question about this report, contact me at haochen dot jiang at intel.com)

Re: [r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[Jakub Jelinek]

On Sat, Apr 22, 2023 at 11:30:24AM +0800, haochen.jiang via Gcc-patches wrote:
> On Linux/x86_64,
> 
> 03cebd304955a6b9c5607e09312d77f1307cc98e is the first bad commit
> commit 03cebd304955a6b9c5607e09312d77f1307cc98e
> Author: Jason Merrill <jason@redhat.com>
> Date:   Tue Apr 18 21:32:07 2023 -0400
> 
>     c++: fix 'unsigned typedef-name' extension [PR108099]
> 
> caused
> 
> FAIL: std/ranges/iota/max_size_type.cc execution test

That is mentioned in
https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616439.html
as known problem.  I think the test should use std::make_signed<rep_t>
instead of signed rep_t which it uses, but as Jason mentioned,
it reveals further problems that need to be looked at.

	Jakub

Re: [r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[Jonathan Wakely]

[-- Attachment #1: Type: text/plain, Size: 966 bytes --]

On Saturday, April 22, 2023, Jakub Jelinek <jakub@redhat.com> wrote:
> On Sat, Apr 22, 2023 at 11:30:24AM +0800, haochen.jiang via Gcc-patches
wrote:
>> On Linux/x86_64,
>>
>> 03cebd304955a6b9c5607e09312d77f1307cc98e is the first bad commit
>> commit 03cebd304955a6b9c5607e09312d77f1307cc98e
>> Author: Jason Merrill <jason@redhat.com>
>> Date:   Tue Apr 18 21:32:07 2023 -0400
>>
>>     c++: fix 'unsigned typedef-name' extension [PR108099]
>>
>> caused
>>
>> FAIL: std/ranges/iota/max_size_type.cc execution test
>
> That is mentioned in
> https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616439.html
> as known problem.  I think the test should use std::make_signed<rep_t>
> instead of signed rep_t

It's already testing non-public implementation details, it could just use
#if and use signed __int128 or signed long long as appropriate.

> which it uses, but as Jason mentioned,
> it reveals further problems that need to be looked at.
>
>         Jakub
>
>

Re: [r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[Patrick Palka]

[-- Attachment #1: Type: text/plain, Size: 1171 bytes --]

On Sun, 23 Apr 2023, Jonathan Wakely wrote:

> 
> 
> On Saturday, April 22, 2023, Jakub Jelinek <jakub@redhat.com> wrote:
> > On Sat, Apr 22, 2023 at 11:30:24AM +0800, haochen.jiang via Gcc-patches wrote:
> >> On Linux/x86_64,
> >>
> >> 03cebd304955a6b9c5607e09312d77f1307cc98e is the first bad commit
> >> commit 03cebd304955a6b9c5607e09312d77f1307cc98e
> >> Author: Jason Merrill <jason@redhat.com>
> >> Date:   Tue Apr 18 21:32:07 2023 -0400
> >>
> >>     c++: fix 'unsigned typedef-name' extension [PR108099]
> >>
> >> caused
> >>
> >> FAIL: std/ranges/iota/max_size_type.cc execution test
> >
> > That is mentioned in
> > https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616439.html
> > as known problem.  I think the test should use std::make_signed<rep_t>
> > instead of signed rep_t

IIUC we can't portably use make_signed here since __int128 is considered
an integral type only in GNU mode and not in strict mode.

> 
> It's already testing non-public implementation details, it could just use #if and use signed __int128 or signed long long as appropriate.

Sounds good, patch posted at https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616598.html

Re: [r14-159 Regression] FAIL: std/ranges/iota/max_size_type.cc execution test on Linux/x86_64

[Jonathan Wakely]

On Mon, 24 Apr 2023 at 17:27, Patrick Palka <ppalka@redhat.com> wrote:
>
> On Sun, 23 Apr 2023, Jonathan Wakely wrote:
>
> >
> >
> > On Saturday, April 22, 2023, Jakub Jelinek <jakub@redhat.com> wrote:
> > > On Sat, Apr 22, 2023 at 11:30:24AM +0800, haochen.jiang via Gcc-patches wrote:
> > >> On Linux/x86_64,
> > >>
> > >> 03cebd304955a6b9c5607e09312d77f1307cc98e is the first bad commit
> > >> commit 03cebd304955a6b9c5607e09312d77f1307cc98e
> > >> Author: Jason Merrill <jason@redhat.com>
> > >> Date:   Tue Apr 18 21:32:07 2023 -0400
> > >>
> > >>     c++: fix 'unsigned typedef-name' extension [PR108099]
> > >>
> > >> caused
> > >>
> > >> FAIL: std/ranges/iota/max_size_type.cc execution test
> > >
> > > That is mentioned in
> > > https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616439.html
> > > as known problem.  I think the test should use std::make_signed<rep_t>
> > > instead of signed rep_t
>
> IIUC we can't portably use make_signed here since __int128 is considered
> an integral type only in GNU mode and not in strict mode.

Right (for now ... C++23 means we can make it an integer type, and I
think we might as well apply that to all -std modes, but that work
isn't done yet).

>
> >
> > It's already testing non-public implementation details, it could just use #if and use signed __int128 or signed long long as appropriate.
>
> Sounds good, patch posted at https://gcc.gnu.org/pipermail/gcc-patches/2023-April/616598.html