Brad King wrote: > On 08/02/2013 05:20 PM, Rolf Eike Beer wrote: > > Brad, please don't merge that > > Do all of the remaining test failures need the cxx-flags topic to fix?
No, since both are in next already everything that still shows up as a bug is a bug. > http://open.cdash.org/testSummary.php?project=1&name=Module.FindCXXFeatures > &date=2013-08-13 > > I'd like to get these resolved or revert the topic because the noise > on the dashboard is very distracting while trying to get 2.8.12-rc1 out. I'm working on it. > On Win64-vs10-Tv90 it is building with VS 10 but the compiler is VS 9. > The logic that sets MSVC_VERSION should be correct in this case so I'm > not sure what is causing the expected results to mismatch. The expected features seen here indicate that MSVC_VERSION is 1600 (i.e. VC10). Which is correct IMHO, as it gives the version of the suite, not the compiler. I'll switch that over to use CMAKE_CXX_COMPILER_VERSION as suggested below, the error will probably go away then. > For Win32-mingw-gcc-4.5 and Win32-msys-gcc-4.5, does the MinGW port of > the compiler support all the same features? It is GNU 4.5.0. I would have expected it so. > For the Linux*-LSB4.0 builds with GNU 4.3.2: > > CMake Error at CMakeLists.txt:203 (message): > C++ feature 'CXXFeatures_cstdint_header_FOUND' was detected, but not > expected > > the LSB-enforcing compilers may provide that header. > > For MacOSX10.7-Xcode43-ub, it is the Apple port of the GNU 4.2.1 so > there are more features than expected. Because of the Clang version mixup already discussed I plan to make the "more features found than expected" message a WARNING instead of an error. I can do this now and sort the other expected features out next week or so when you have the RC ready. This is only part of how we verify that the module works, not the module itself. > For Windows-icl-11.1-*, why do the expected results not just match > for this version? That is a question I wanted to bring up, too. Can you log into that machine and get me the intermediate results of the try_compile calls? > Also, IMO "if (CMAKE_COMPILER_IS_GNUCXX)" and "elseif (MSVC)" should > use CMAKE_CXX_COMPILER_ID instead just like the rest of the cases. Agreed, as above. Eike --
signature.asc
Description: This is a digitally signed message part.
-- Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://public.kitware.com/cgi-bin/mailman/listinfo/cmake-developers
