Skip to content
This repository was archived by the owner on Feb 2, 2026. It is now read-only.

Add archspec packages for x86_64#72

Merged
beckermr merged 6 commits intoregro:mainfrom
chrisburr:fake-archspc
Feb 7, 2025
Merged

Add archspec packages for x86_64#72
beckermr merged 6 commits intoregro:mainfrom
chrisburr:fake-archspc

Conversation

@chrisburr
Copy link
Contributor

When dependent packages require archspec the bot fails to solve despite the situation being okay. For example: https://conda-forge.org/status/migration/?name=root_base6328

This fixes this by always including the x86_64_v3 virtual package when solving for x86_64. We could think of including other virtual packages but I'm not sure if it's worth it.

Copy link
Contributor

@beckermr beckermr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The virtual packages are written to disk as a separate channel here: https://github.com/chrisburr/conda-forge-feedstock-check-solvable/blob/fake-archspc/conda_forge_feedstock_check_solvable/virtual_packages.py#L114

We need to add the archspec version / names we support as globals in utils.py and then write them in the same way.

Also we'll need to add tests.

@chrisburr
Copy link
Contributor Author

I thought I had tried that and it didn't work I'll try again

@chrisburr
Copy link
Contributor Author

What do you think about doing the opposite and setting them all properly instead of making a fake channel?

@beckermr
Copy link
Contributor

beckermr commented Feb 7, 2025

The fake channel is there so that we can use mamba as the solver backend in addition to rattler. We are approaching a time where mamba support can be removed (and it is currently broken in the tests), but we have not reached that point yet IMHO.

@chrisburr
Copy link
Contributor Author

Given the mamba support will likely be removed how about I port it over and make the fake channel mamba specific?

@beckermr
Copy link
Contributor

beckermr commented Feb 7, 2025

That'd be fine as long as we have a test that ensure the virtual packages end up the same for each solver.

@beckermr
Copy link
Contributor

beckermr commented Feb 7, 2025

I pushed changes to at least get the mamba tests to run even if they fail.

@chrisburr
Copy link
Contributor Author

Having thought about it more I think the virtual_package_repodata approach is better as it allows the solver to have multiple of the same virtual package with different versions.

@chrisburr chrisburr changed the title Assume x86_64_v3 when solving for x86_64 platforms Add archspec packages for x86_64 Feb 7, 2025
Copy link
Contributor

@beckermr beckermr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM and thank you!

@beckermr beckermr enabled auto-merge February 7, 2025 22:44
@beckermr beckermr merged commit d37b1bd into regro:main Feb 7, 2025
3 checks passed
@chrisburr chrisburr deleted the fake-archspc branch February 8, 2025 06:47
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants