On Tue, Nov 17, 2015 at 4:14 PM, Levi Morrison <morrison.l...@gmail.com> wrote:
> On Tue, Nov 17, 2015 at 3:09 PM, Levi Morrison <morrison.l...@gmail.com>
> wrote:
>
>> On Tue, Nov 17, 2015 at 2:38 PM, Brad King <brad.k...@kitware.com> wrote:
>>
>>> On 11/17/2015 04:34 PM, Levi Morrison wrote:
>>> > On Tue, Nov 17, 2015 at 2:07 PM, Levi Morrison wrote:
>>> >>> icpc: command line warning #10159: invalid argument for option '-std'
>>> >> Hmm. Can you tell me what the value of -std was for that invocation?
>>> >
>>> > Hazarding a guess I think it might be gnu++14 – can you confirm this?
>>>
>>> It was `-std=c++14`. I'm using:
>>>
>>> $ icpc --version
>>> icpc (ICC) 15.0.0 20140723
>>>
>>> According to the `man icpc` page there is no -std=c++14 option.
>>>
>>> -Brad
>>>
>>>
>> It seems it was added in either 15.0.1 or 15.0.2. I can't find it in the
>> documentation. For now I will tell it 15.0.2 which I can confirm works.
>>
>
> Version 15.0.0 seems to use `-std=c++1y`.
>
> After I have fixed some more issues I have encountered two tests that I am
> stumped on:
>
> - cxx_right_angle_brackets
> - LinkImplementationFeatureCycle
>
> The first the test has this error:
>
> Tests/CompileFeatures/cxx_right_angle_brackets.cpp(10): error: space
> required between adjacent ">" delimiters of nested template argument lists
> (">>" is the right shift operator)
>
> In some experimentation with icpc this test passes if -std=c++0x/11/1y/14
> is used. When no standard is defined it has this error. How can I see
> exactly what flags were used? I am using -VV and do not see the exact
> invocation flags.
>
> The second failing test has this for the expect-err>:
>
> 286: expect-err> CMake Error in CMakeLists.txt:
> 286: expect-err> The COMPILE_FEATURES property of target "empty1" was
> evaluated when
> 286: expect-err> computing the link implementation, and the
> "CXX_STANDARD" was "98" for that
> 286: expect-err> computation. Computing the COMPILE_FEATURES based
> on the link
> 286: expect-err> implementation resulted in a higher "CXX_STANDARD"
> "11". This is not
> 286: expect-err> permitted. The COMPILE_FEATURES may not both depend
> on and be depended on
> 286: expect-err> by the link implementation.
>
> It will actually build successfully instead of error. This indicates to me
> I have set something up incorrectly. Any suggestions here?
>
> I have attached my current patch.
>
I have fixed a smattering of issues and was able to get Intel versions
14.0.2 , 15.0.2 and 15.0.3 to pass the CompilerFeature tests! Hooray!
However, I am stuck on Intel 15.0.0. In Intel-CXX-FeatureTests the entry
for cxx_return_type_deduction is giving me issues. If I put it into the
section just above it, the test will fail when it should pass. If I try to
exclude it by only doing stuff not less than 15.0.2 then it now suddenly
passes but the test suite expects it to fail. Any ideas on this one?
I'm going to install Intel 16.0 in the meantime and try to get its support
working.
0001-Add-CXX_STANDARD-support-to-Intel-C-compilers.patch
Description: Binary data
-- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
