[CMake] Weired behavior with CTest & Fedora 10 & g++ 4.3.2

[Bill Lorensen]

Some more information. This is not a clean build each night. With the
compiler flags I use, it just takes too long. However, tonight, I'll
clean the build tree. Perhaps some system file has been updated. I let
Fedora automatically update my system.

Bill

On Wed, Dec 17, 2008 at 1:11 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> David,
>
> I have a Fedora 9, ctest  2.6-patch 1 RC-11 and gcc 4.3.0 that
> exhibits similar behavior. This is for the itk dashboard which has
> about 1400 tests on over 30 platforms. I have only noticed this on
> this one platform. Both tests segfault without any output. When I
> rerun them manually, they pass.
>
> Here's one example from last night where two tests fail:
> http://www.cdash.org/CDash/testDetails.php?test=14056769&build=232080
> http://www.cdash.org/CDash/testDetails.php?test=14056775&build=232080
>
> The previous night I had two different tests fail:
> http://www.cdash.org/CDash/testDetails.php?test=14002130&build=231405
> http://www.cdash.org/CDash/testDetails.php?test=14002264&build=231405
>
> Some of these build have tests that are truly failures, on all platforms.
>
> All of the bogus failures only occur on this one platform. I was
> assuming I had some memory errors or some other hardware thing going
> on.
>
> Maybe not...
>
> Until your post I had not noticed that it is always 2 bogus failures.
>
> Bill
>
>
>
> On Wed, Dec 17, 2008 at 7:00 AM, David Graf <david.graf at 28msec.com> wrote:
>> Hello
>>
>> We have a weired problem when we compile and test our software on Fedora 10
>> with g++ 4.3.2.
>> We have around 600 tests. When we execute these tests in one ctest run, two
>> random tests always fail (without any regularity). We tested it on two
>> different fedora installation with ctest 2.4 and 2.6. This strange behavior
>> occurs on both machines.
>> After some debugging, we observed that the failing tests do not produce any
>> log output. Even when adding the std::cout out statements into the code, we
>> do not get some output. We assume there is some bad interaction between
>> Fedora 10 and ctest.
>>
>> Has someone noticed the same strange behavior?
>>
>> David
>> _______________________________________________
>> CMake mailing list
>> CMake at cmake.org
>> http://www.cmake.org/mailman/listinfo/cmake
>>
>