[gmx-developers] Difficult interpreting Jenkins output "Text Finder"

[Paul bauer]

Debugging this on a different machine, I think I found the error (not 
sure if it is icc related or double, but have not been able to reproduce 
it otherwise)

The bug only appears in release mode, and is not reproducible in a debug 
build

==29175==    at 0x56DD9A: __intel_sse2_strchr (in 
***/build-icc18-double-simdnone-mkl-release/bin/nbnxnkernelref-test)
==29175==    by 0xAF635E8: strchr (string.h:226)
==29175==    by 0xAF635E8: quoteIfNecessary (cmdlineprogramcontext.cpp:78)
==29175==    by 0xAF635E8: 
gmx::CommandLineProgramContext::Impl::Impl(int, char const* const*, 
std::unique_ptr<gmx::IExecutableEnvironment, 
std::default_delete<gmx::IExecutableEnvironment> >) 
(cmdlineprogramcontext.cpp:332)


On 03/09/2018 20:37, David van der Spoel wrote:
> Hi,
>
> for one of my groups projects, Jenkins produces this which is hard to 
> interpret for us:
> 09:57:18 The following tests FAILED:
> 09:57:18 Build step 'Jenkins Text Finder' changed build result to 
> UNSTABLE
>
> Any clue?
> The whole output is here:
> http://jenkins.gromacs.org/job/Matrix_PreSubmit_master/5623/OPTIONS=icc-18%20no-thread-mpi%20double%20openmp%20mkl%20cmake-3.9.6%20simd=none%20release%20hwloc%20host=bs_mic,label=bs_mic/console 
>