Skip to content

Conversation

Flamefire
Copy link
Contributor

@Flamefire Flamefire commented Oct 7, 2025

(created using eb --new-pr)

When an exception aborted the build the error reporting in the test report is badly formatted, especially as the traceback is printed as normal text which breaks markdown formatting

This fixes that by putting the traceback into code tags and adjusting the list items:

  • FAIL cuDNN-9.5.1.17-CUDA-12.6.0.eb (unhandled exception: name 'doesnt_exist' is not defined)
traceback

Prior: https://gist.github.com/Flamefire/a1e367edcdf25758c5fbc4be39dc710d
After: https://gist.github.com/Flamefire/24526ae1becbae76e1b31b9347d48ec7

Related: Should we improve the stdout for test report runs? Currently it is totally missing the cause for the failure:

== processing EasyBuild easyconfig /tmp/easybuild-tmp/eb-8uru6x75/files_pr23923/c/cuDNN/cuDNN-9.5.1.17-CUDA-12.6.0.eb
== building and installing Core/cuDNN/9.5.1.17-CUDA-12.6.0...
== fetching files and verifying checksums...
[snip]
== restore after iterating...
== postprocessing...
== sanity checking...
== ... (took 8 secs)
== processing EasyBuild easyconfig /tmp/easybuild-tmp/eb-8uru6x75/files_pr23923/c/cuSPARSELt/cuSPARSELt-0.6.3.2-CUDA-12.6.0.eb
== building and installing Core/cuSPARSELt/0.6.3.2-CUDA-12.6.0...
== fetching files and verifying checksums...
[snip]
== sanity checking...
== ... (took 5 secs)


Adding comment [snip]
**FAILED**
Build succeeded for 0 out of 7 (7 easyconfigs in total)
c144 - [snip machine...]
See https://gist.github.com/[...] for a full test report.'
== Test report uploaded [...]
== Build succeeded for 0 out of 7
== Summary:
   * [FAILED]  Core/cuDNN/9.5.1.17-CUDA-12.6.0
   * [FAILED]  Core/cuSPARSELt/0.6.3.2-CUDA-12.6.0
   * [FAILED]  Compiler/GCCcore/13.3.0/NCCL/2.26.2-CUDA-12.6.0
   * [FAILED]  Compiler/GCC/13.3.0/nvidia-cutlass/3.8.0.0-CUDA-12.6.0
   * [FAILED]  Compiler/GCCcore/13.3.0/pyzstd/0.16.2
   * [FAILED]  MPI/GCC/13.3.0/OpenMPI/5.0.3/Triton/3.3.1-CUDA-12.6.0
   * [FAILED]  MPI/GCC/13.3.0/OpenMPI/5.0.3/PyTorch/2.7.1-CUDA-12.6.0

@boegel boegel added this to the next release (5.2.0?) milestone Oct 8, 2025
@boegel boegel added the change label Oct 8, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants