Brad King wrote:
> On 08/14/2012 07:04 AM, Stephen Kelly wrote:
>> Why do you think that's not the correct fix?
>> Is there really no C compiler available?
>> Do you have a backtrace?
>> If you comment out the run_tests("C") in the GenerateExportHeader unit
>> test, does the problem go away?
>
> As a side effect of the 'next' rewrite the generate_export_header-fixes
> branch is no longer in 'next'. It is still on the stage. Please try
> to resolve this issue and rebase/rewrite the topic before merging to
> 'next' again.
Yes.
I have no way of debugging the problems interactively, however. I would tend
to skip tests which I can not debug and which are raising problems to which
I do not already know the solution, but that also did not meet with
satisfaction, so it's catch 22:
http://thread.gmane.org/gmane.comp.programming.tools.cmake.devel/4448/focus=4457
I know that skipping tests is not great either due to them not showing up at
all on the dashboard, but without a way to debug the problems that appear on
the dashboard, examine the content of generated files, try things, and have
enough knowledge about the affected platforms (such as the void issue you
write below), it's the best I can do I think.
One very inefficient way would be to try to debug it by adding lots of
message(), but that conflicts with the requirement to resolve the issues
before merging to next.
> P.S. Borland C wants functions with no arguments to be declared
> explicitly with "(void)" rather than just "()".
Thanks. This is something I probably would not have even known to try.
Bill, aside from the above requests I made previously, could you see if this
makes a difference?
Thanks,
Steve.
--
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
