Require C & C++ compilers in cuda-nvcc

[jakirkham]

As the cuda-nvcc package is intended for usage by developers in their environments and having working C & C++ compilers is a prerequisite, simply make the C & C++ compilers requirements of cuda-nvcc.

---

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jakirkham]

@conda-forge-admin, please re-render

[github-actions]

Hi! This is the friendly automated conda-forge-webservice.

I tried to rerender for you, but it looks like I wasn't able to push to the cuda-nvcc_req_compilers branch of jakirkham-feedstocks/cuda-nvcc-feedstock. Did you check the "Allow edits from maintainers" box?

NOTE: PRs from organization accounts or PRs from forks made from organization forks cannot be rerendered because of GitHub permissions. Please fork the feedstock directly from conda-forge into your personal GitHub account.

This message was generated by GitHub actions workflow run https://github.com/conda-forge/cuda-nvcc-feedstock/actions/runs/5394898787.

[jakirkham]

Not seeing any changes when re-rendering locally 😕

Edit: Though wouldn't expect to necessarily. Unclear why the bot reached a different conclusion

[bdice]

Does it make sense to keep this warning, or can that be removed now that the compilers are hard dependencies?

cuda-nvcc-feedstock/recipe/activate.sh

Line 31 in f602ed9

echo 'cuda-nvcc: Please add the `compiler("c")` and `compiler("cxx")` packages to the environment.'

[jakirkham]

Think so because the activation scripts are part of cuda-nvcc_{{ cross_target_platform }}

cuda-nvcc-feedstock/recipe/meta.yaml

Lines 43 to 49 in f602ed9

	- name: cuda-nvcc_{{ cross_target_platform }}
	run_exports:
	strong:
	- cuda-version >={{ cuda_version }},<{{ cuda_version.split(".")[0]|int + 1 }}
	files:
	- etc/conda/activate.d/~cuda-nvcc_activate.sh # [linux]
	- etc/conda/deactivate.d/~cuda-nvcc_deactivate.sh # [linux]

[jakirkham]

Thanks all! 🙏

[isuruf]

You probably want to do {{ c_compiler }} instead to avoid pinning a version.

[jakirkham]

Thanks for the feedback! 🙏

Yeah that seems reasonable

Should we do the same thing in the compilers-feedstock?

[isuruf]

Nope. The purpose of compilers feedstock is to give the users the same compilers used by conda-build, whereas cuda-nvcc is a way to give users nvcc.

[jakirkham]

Though we would like one that plays well with the same compilers conda-build uses. So maybe it is not that far off

What if we did this?

Suggested change

	- {{ compiler('c') }}
	- {{ compiler('cxx') }}
	- c-compiler
	- cxx-compiler

[jakirkham]

Just to update this thread, as cuda-compiler fills a similar purpose to c-compiler & cxx-compiler (and includes them as dependencies), think Isuru's original suggestion is probably the best choice here