[1.55][mpl][atomic] Tests failing on MSVC-12

Hi,
I'm seeing Boost.Atomic test failures on MSVC-12 with the following errors:
C:\local\boost\boost/iterator/detail/facade_iterator_category.hpp(166)
: error C2039: 'assert_not_arg' : is not a member of 'boost::mpl'
C:\local\boost\boost/mpl/eval_if.hpp(41) : see reference to
class template instantiation
'boost::detail::facade_iterator_category_impl

Andrey Semashev
I'm seeing Boost.Atomic test failures on MSVC-12 with the following errors:
C:\local\boost\boost/iterator/detail/facade_iterator_category.hpp(166)
: error C2039: 'assert_not_arg' : is not a member of 'boost::mpl'
It doesn't seem to be a problem with Boost.Atomic but rather some
incompatibility of BOOST_MPL_ASSERT_NOT and MSVC-12. Does anyone know
what the problem is?
Yes, almost all MSVC12 release tests fail because of this error. It was fixed in trunk in r84443, but the fix hasn't been merged to the release branch yet. The problem has already been mentioned several times on the maling lists and the bug tracker: https://svn.boost.org/trac/boost/ticket/9205 https://svn.boost.org/trac/boost/ticket/9249 https://svn.boost.org/trac/boost/ticket/9278 http://thread.gmane.org/gmane.comp.lib.boost.user/79919 http://thread.gmane.org/gmane.comp.lib.boost.user/79932 http://thread.gmane.org/gmane.comp.lib.boost.devel/244986 http://thread.gmane.org/gmane.comp.lib.boost.devel/245596

On Tue, Oct 29, 2013 at 12:52 PM, Marcel Raad
Andrey Semashev
writes: I'm seeing Boost.Atomic test failures on MSVC-12 with the following errors:
C:\local\boost\boost/iterator/detail/facade_iterator_category.hpp(166)
: error C2039: 'assert_not_arg' : is not a member of 'boost::mpl'
It doesn't seem to be a problem with Boost.Atomic but rather some
incompatibility of BOOST_MPL_ASSERT_NOT and MSVC-12. Does anyone know
what the problem is?
Yes, almost all MSVC12 release tests fail because of this error. It was fixed in trunk in r84443, but the fix hasn't been merged to the release branch yet.
The problem has already been mentioned several times on the maling lists and the bug tracker:
https://svn.boost.org/trac/boost/ticket/9205
https://svn.boost.org/trac/boost/ticket/9249
https://svn.boost.org/trac/boost/ticket/9278
http://thread.gmane.org/gmane.comp.lib.boost.user/79919
http://thread.gmane.org/gmane.comp.lib.boost.user/79932
So that's an MSVC-12 RC bug? I was thinking the tester used the final MSVC-12. Ok then. Thanks for the references.

Andrey Semashev
So that's an MSVC-12 RC bug? I was thinking the tester used the final
MSVC-12. Ok then.
Thanks for the references.
No, it's a boost bug (part of an MSVC 11 bug workaround is active for only MSVC 11 while another part is active for MSVC 11+). RTM is also affected.

On Tue, Oct 29, 2013 at 4:09 PM, Marcel Raad
Andrey Semashev
writes: So that's an MSVC-12 RC bug? I was thinking the tester used the final
MSVC-12. Ok then.
Thanks for the references.
No, it's a boost bug (part of an MSVC 11 bug workaround is active for only MSVC 11 while another part is active for MSVC 11+). RTM is also affected.
I see. Thanks again for the clarification.

On Tue, Oct 29, 2013 at 8:09 AM, Marcel Raad
Andrey Semashev
writes: So that's an MSVC-12 RC bug? I was thinking the tester used the final
MSVC-12. Ok then.
Thanks for the references.
No, it's a boost bug (part of an MSVC 11 bug workaround is active for only MSVC 11 while another part is active for MSVC 11+). RTM is also affected.
I just finished merging and testing locally. It was causing both MSVC 2012 and 2013 to fail iterator_adaptor_test (and many other tests in many other libraries). After the merge, iterator_adaptor_test passes for both compilers. I'll request permission to merge.

On Tue, Oct 29, 2013 at 5:04 PM, Beman Dawes
On Tue, Oct 29, 2013 at 8:09 AM, Marcel Raad
wrote: Andrey Semashev
writes: So that's an MSVC-12 RC bug? I was thinking the tester used the final
MSVC-12. Ok then.
Thanks for the references.
No, it's a boost bug (part of an MSVC 11 bug workaround is active for only MSVC 11 while another part is active for MSVC 11+). RTM is also affected.
I just finished merging and testing locally. It was causing both MSVC 2012 and 2013 to fail iterator_adaptor_test (and many other tests in many other libraries). After the merge, iterator_adaptor_test passes for both compilers. I'll request permission to merge.
Thank you Beman for taking care of this.

On Tue, Oct 29, 2013 at 4:52 AM, Marcel Raad
Andrey Semashev
writes: I'm seeing Boost.Atomic test failures on MSVC-12 with the following errors:
C:\local\boost\boost/iterator/detail/facade_iterator_category.hpp(166)
: error C2039: 'assert_not_arg' : is not a member of 'boost::mpl'
It doesn't seem to be a problem with Boost.Atomic but rather some
incompatibility of BOOST_MPL_ASSERT_NOT and MSVC-12. Does anyone know
what the problem is?
Yes, almost all MSVC12 release tests fail because of this error. It was fixed in trunk in r84443, but the fix hasn't been merged to the release branch yet.
The problem has already been mentioned several times on the maling lists and the bug tracker:
I'm about to do the merge and test locally. Assuming it works I'll request permission to merge. --Beman
participants (3)
-
Andrey Semashev
-
Beman Dawes
-
Marcel Raad