[Bug libstdc++/61532] New: [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[Bug libstdc++/61532] New: [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[ramana at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

            Bug ID: 61532
           Summary: [4.10 regression]  make_signed and make_unsigned
                    wchar_t have started failing in the libstdc++
                    testsuite.
           Product: gcc
           Version: 4.10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ramana at gcc dot gnu.org

Somewhere between 210455 and 210475 these tests started failing on
arm-none-linux-gnueabihf. I see the same failures on aarch64-none-elf too.


FAIL: 20_util/make_signed/requirements/typedefs-1.cc (test for excess errors)
FAIL: 20_util/make_signed/requirements/typedefs-2.cc (test for excess errors)
FAIL: 20_util/make_unsigned/requirements/typedefs-1.cc (test for excess errors)
FAIL: 20_util/make_unsigned/requirements/typedefs-2.cc (test for excess errors)
FAIL: tr1/2_general_utilities/shared_ptr/modifiers/reset_neg.cc  (test for
errors, line 36)

All with 

   static_assert( is_same<test23_type, volatile signed wchar_t>::value,


or their equivalent.

In C on ARM the wchar_t type is defined to unsigned int on AAPCS configurations
which is true for Linux, I'm not sure what we do in C++ but I expect this to be
the same. I suspect there is something underlying here for all targets that
define wchar_t to be unsigned int by default.

Looking at other testresults in June it looks like atleast the first 2 tests
fail on powerpc-linux and powerpc-aix 




regards
Ramana

[Bug libstdc++/61532] [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[ramana at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

Ramana Radhakrishnan <ramana at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-06-17
   Target Milestone|---                         |4.10.0
     Ever confirmed|0                           |1

[Bug libstdc++/61532] [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Targets that define wchar_t as unsigned int are not conforming to the C++
standard, which requires wchar_t to be a distinct type, not a typedef.

I can look into a hack to detect non-conforming targets, but probably not until
next week.

[Bug libstdc++/61532] [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> ---
This is likely to also be failing on the 4.9 branch, since the weekend

[Bug libstdc++/61532] [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |redi at gcc dot gnu.org

[Bug libstdc++/61532] [4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[ramana at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #3 from Ramana Radhakrishnan <ramana at gcc dot gnu.org> ---

(In reply to Jonathan Wakely from comment #2)
> This is likely to also be failing on the 4.9 branch, since the weekend

I was about to add 4.9 but then couldn't find a link to post. 

(In reply to Jonathan Wakely from comment #1)
> Targets that define wchar_t as unsigned int are not conforming to the C++
> standard, which requires wchar_t to be a distinct type, not a typedef.

The standard appears to say in Section 3.9.1 (basic.fundamental) #5 that the 

"Type wchar_t shall have the same size, signedness, and alignment requirements
(3.11) as one of the other integral types, called its underlying type." 

That's my interpretation from n3337.

> 
> I can look into a hack to detect non-conforming targets, but probably not
> until next week.

Please do take a look. 

Ramana

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
      Known to work|                            |4.9.0
   Target Milestone|4.10.0                      |4.9.1
            Summary|[4.10 regression]           |[4.9/4.10 regression]
                   |make_signed and             |make_signed and
                   |make_unsigned wchar_t have  |make_unsigned wchar_t have
                   |started failing in the      |started failing in the
                   |libstdc++ testsuite.        |libstdc++ testsuite.
      Known to fail|                            |4.10.0, 4.9.1

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #4 from Marc Glisse <glisse at gcc dot gnu.org> ---
Writing "signed wchar_t" doesn't seem legal to me, and clang and intel warn or
reject such code.

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Marc Glisse from comment #4)
> Writing "signed wchar_t" doesn't seem legal to me, and clang and intel warn
> or reject such code.

Yeah, it's a GCC extension, and I thought I'd replaced the uses of it in the
tests with something portable, but apparently not!

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Ramana, are you actually testing on trunk? r210517 should have fixed three of
those failures

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Author: redi
Date: Mon Jun 23 18:17:03 2014
New Revision: 211909

URL: https://gcc.gnu.org/viewcvs?rev=211909&root=gcc&view=rev
Log:
    PR libstdc++/61532
    * testsuite/20_util/make_signed/requirements/typedefs-1.cc: Do not
    apply the signed specifier to wchar_t.
    * testsuite/20_util/make_signed/requirements/typedefs-2.cc: Likewise.
    * testsuite/20_util/make_unsigned/requirements/typedefs-1.cc: Check
    cv-qualifier and size.
    * testsuite/20_util/make_unsigned/requirements/typedefs-2.cc:
    Likewise.

Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/testsuite/20_util/make_signed/requirements/typedefs-1.cc
    trunk/libstdc++-v3/testsuite/20_util/make_signed/requirements/typedefs-2.cc
   
trunk/libstdc++-v3/testsuite/20_util/make_unsigned/requirements/typedefs-1.cc
   
trunk/libstdc++-v3/testsuite/20_util/make_unsigned/requirements/typedefs-2.cc

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Author: redi
Date: Mon Jun 23 20:29:57 2014
New Revision: 211914

URL: https://gcc.gnu.org/viewcvs?rev=211914&root=gcc&view=rev
Log:
    PR libstdc++/61532
    * testsuite/20_util/make_signed/requirements/typedefs-1.cc: Do not
    apply the signed specifier to wchar_t.
    * testsuite/20_util/make_signed/requirements/typedefs-2.cc: Likewise.
    * testsuite/20_util/make_unsigned/requirements/typedefs-1.cc: Do not
    apply the unsigned specifier to wchar_t.
    * testsuite/20_util/make_unsigned/requirements/typedefs-2.cc:
    Likewise.

Modified:
    branches/gcc-4_9-branch/libstdc++-v3/ChangeLog
   
branches/gcc-4_9-branch/libstdc++-v3/testsuite/20_util/make_signed/requirements/typedefs-1.cc
   
branches/gcc-4_9-branch/libstdc++-v3/testsuite/20_util/make_signed/requirements/typedefs-2.cc
   
branches/gcc-4_9-branch/libstdc++-v3/testsuite/20_util/make_unsigned/requirements/typedefs-1.cc
   
branches/gcc-4_9-branch/libstdc++-v3/testsuite/20_util/make_unsigned/requirements/typedefs-2.cc

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The tests should be fixed now - please check.

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[ramana.radhakrishnan at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

--- Comment #10 from ramana.radhakrishnan at arm dot com <ramana.radhakrishnan at arm dot com> ---
>
> On 23/06/14 21:31, redi at gcc dot gnu.org wrote:> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532
>>
>> --- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> ---
>> The tests should be fixed now - please check.
>>

This commit has fixed the issues with the 2 tests that were still 
remaining .

https://gcc.gnu.org/ml/gcc-testresults/2014-06/msg02109.html

libstdc++ test results are now clear on armhf.

Thanks for fixing this up.

regards
Ramana

[Bug libstdc++/61532] [4.9/4.10 regression] make_signed and make_unsigned wchar_t have started failing in the libstdc++ testsuite.

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61532

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #11 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Thanks for checking - fixed then.