Upload wheels to PyPI from GitHub-hosted runner#142
Conversation
The gh-action-pypi-publish action does not support self-hosted runners. Use a GitHub-hosted runner instead. Resolves NVIDIA#75
|
/ok to test |
|
Seems like CI is not working at all now (and not just for this PR) - I will need to look into that. |
|
CI is not running on this PR because the workflow changes made the file invalid: https://github.com/NVIDIA/numba-cuda/actions/runs/13596857711 I still see CI runs from #143 and #145 right now: https://github.com/NVIDIA/numba-cuda/actions |
|
As you can see here, the fake publish step is picking up the wheels from the build job. I am going to revert the last two test commits and restore the PR to how I want it to be merged. |
vyasr
left a comment
vyasr
left a comment
There was a problem hiding this comment.
This looks good to me, but I don't think my review has any power here.
|
Once this is merged, would you guys be willing to do a patch release to check that it worked? |
|
@KyleFromNVIDIA @gmarkall FYI if you move it out of the reusable workflow into the top level, you could get rid of having to manage long-living API tokens + you'd be able to upload PEP 740 digital attestations to PyPI out of the box. This would mark more bits of metadata on the public PyPI page as verified. |
|
Thanks for the heads-up. So you would suggest moving the publish step out of wheels-build.yaml into the publish workflow after the build call? |
Move publish step out of `wheels-build.yaml` Follow-up to #142 (comment) Also `s/mambabuild/build/` in `build_conda.sh`
- Fix linking of external code from callees (NVIDIA#137) - Try using a newer branch workflow (NVIDIA#148) - Move publish step out of `wheels-build.yaml` (NVIDIA#147) - Upload wheels to PyPI from GitHub-hosted runner (NVIDIA#142) - Add paddle to interoperability chapter (NVIDIA#144) - Fix the debug info of GridGroup type (NVIDIA#131) - Remove dead `prepare_cuda_kernel()` (NVIDIA#130) - Add a CUDA DI Builder (NVIDIA#104) - dont launch extra kernels when stats counting is disabled (NVIDIA#127) - Fixup debug metadata in kernel fixup (NVIDIA#97) - Implement debuginfo bool name fix (numba/numba#9888) in numba-cuda (NVIDIA#106)
- Fix linking of external code from callees (#137) - Try using a newer branch workflow (#148) - Move publish step out of `wheels-build.yaml` (#147) - Upload wheels to PyPI from GitHub-hosted runner (#142) - Add paddle to interoperability chapter (#144) - Fix the debug info of GridGroup type (#131) - Remove dead `prepare_cuda_kernel()` (#130) - Add a CUDA DI Builder (#104) - dont launch extra kernels when stats counting is disabled (#127) - Fixup debug metadata in kernel fixup (#97) - Implement debuginfo bool name fix (numba/numba#9888) in numba-cuda (#106)
5 participants
The
gh-action-pypi-publishaction does not support self-hosted runners. Use a GitHub-hosted runner instead.Resolves #75