[CMake] CTest: Weird behaviour

[Marcel Loose]

On Tue, 2010-10-19 at 08:24 -0400, Bill Hoffman wrote:
> On 10/19/2010 4:21 AM, Marcel Loose wrote:
> > Hi all,
> >
> > I stumbled upon IMHO weird behaviour of CTest.
> > It seems that compilation errors are not picked up, somehow. Look at
the
> > output of a run of "ctest -V -D ExperimentalBuild" below.
> >
> >
/export/home/loose/work/LOFAR_trunk/LOFAR/LCS/pyparameterset/src/pyparameterset.cc:22:27:
> > error: lofar_config.hh: No such file or directory
> 
> New regular expression is needed for this error most likely.
> 
> The CTEST_USE_LAUNCHERS option would fix that:
> 
> http://www.kitware.com/blog/home/post/11
> 
> We should also add the regular expression for thie error.
> 
> > Linking CXX shared
> > module
> >
../../../lib64/python2.6/site-packages/lofar/parameterset/_pyparameterset.so
> > g++: CMakeFiles/_pyparameterset.dir/pyparameterset.cc.o: No such
file or
> > directory
> > [ 96%] Built target _pyparameterset
> > [ 98%] Built target lofar_pyparameterset
> > [100%] Built target versionpyparameterset
> >   Size of output: 1K
> > Command exited with the value: 0
> >     0 or more Compiler errors
> >     0 Compiler warnings
> > Elapsed: 0:12.30 - CPU: 9.944u+1.396s = 92.1%
> >
> > The compiler fails to compile pyparmeterset.cc, because it cannot
find
> > one of the files that are #include-d. CTest, however, reports 0
compiler
> > errors and exits with status 0.
> >
> > Could someone please explain what's going on here?
> >
> > Best regards,
> > Marcel Loose.
> >
> >
> > _______________________________________________
> > 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://www.cmake.org/mailman/listinfo/cmake
> >
> 
> 

Hi Bill,

It's not just this kind of error. This was just an example. I get the
same result when I strip off a semi-colon at the end of a C++ statement,
which definitely gives a different compiler error. It looks as if the
error count is not being tracked correctly, or being reset somewhere.

Best regards,
Marcel Loose.