Jan Patrick Lehr

Thanks for the details. I'm not sure I'm fully following along though. The error that you shared above was about missing includes in the MetaCG project, not about a failing...

Thanks for clarifying. No, there is no need to switch to a different OS. The error you are seeing is due to the CLI of MetaCG change from v0.4.0 (previously...

I'll close this for now, due to inactivity.

I don't think we want to test the sytem's rocgdb (by accident or on purpose).

Hi, This broke our OpenMP / Offload bots. It would be great if that can be easily fixed or reverted if more time is required to inspect the issue. Thanks!

> I just pushed a fix in [75252e2](https://github.com/llvm/llvm-project/commit/75252e29ea6a0959f3c1670e641a03fc18fc65fa). I see another error in one of our bots for this patch: https://lab.llvm.org/staging/#/builders/130/builds/7112 And another one in https://lab.llvm.org/buildbot/#/builders/140/builds/9495

This may have turned our HIP blender bot red: https://lab.llvm.org/buildbot/#/builders/123/builds/24793 ``` [34/59] Building CXX object External/HIP/CMakeFiles/TheNextWeek-hip-6.3.0.dir/workload/ray-tracing/TheNextWeek/main.cc.o FAILED: External/HIP/CMakeFiles/TheNextWeek-hip-6.3.0.dir/workload/ray-tracing/TheNextWeek/main.cc.o /home/botworker/bbot/clang-hip-vega20/botworker/clang-hip-vega20/llvm/bin/clang++ -DNDEBUG -O3 -DNDEBUG -w -Werror=date-time --rocm-path=/opt/botworker/llvm/External/hip/rocm-6.3.0 --offload-arch=gfx908 --offload-arch=gfx90a --offload-arch=gfx1030 --offload-arch=gfx1100 -xhip...

This broke https://lab.llvm.org/buildbot/#/builders/10/builds/9125/steps/8/logs/stdio The bot does a SHARED_LIBS=ON build. I can reproduce locally with the CMake cache file in `offload/cmake/caches/AMDGPUBot.cmake` as the build config.

Thanks for the report! Which release or branch is this happening? Is this running bare metal or in a container environment?

I looked at the results again, and I think the test is not doing what its name is conveying (to me) what it is intended to do.