Fix issue #280

[lohedges]

This PR closes #280.

• I confirm that I have merged the latest version of devel into this branch before issuing this pull request (e.g. by running git pull origin devel): [y]
• I confirm that I have permission to release this code under the GPL3 license: [y]

Suggested reviewers:

@chryswoods

Note that I expect the CI to fail on macOS due to this issue.

[lohedges]

Well, that was strange. The macOS build did indeed fail, but not for the reason that I was expecting. The error is:

 The reported errors are:
- Encountered problems while solving:
-   - nothing provides requested sire 2024.2.0.dev

The package does exist, so I don't know why it wasn't found. Will try re-running.

[lohedges]

Same issue with a re-run. Really strange.

[lohedges]

I see the issue, the runner is now osx-arm64. This must be the origin of the GROMACS issue too!

[xiki-tempula]

Hi @lohedges Is this PR ready? Do you mind merge this, please? Thanks.

[lohedges]

It's ready from a code perspective, but needs a few things first:

• We need to update the sire CI to use the old x86 macOS runner, of move to macOS ARM64. (This is causing the macOS CI failure above, i.e. there's no matching macOS sire package to build against.)
• Figure out the Windows CI failures that are causing the environment creation to fail. (Looks like a mamba issue on WIndows.) Looks like the environment failure issue has been fixed. The failure here was cause by one of the random IO errors that seems to happen when reading test files. Just re-running now.
• Needs a review. I can ask @mb2055 to take a look if @chryswoods is too busy this week.

[lohedges]

The Windows re-run passed. Everything should be good once the macOS ARM64 sire package is ready. If easier, we can always merge once reviewed and rebuild against the updated packages once available.

[xiki-tempula]

Do you mind merge this now, please? Thank you.

[lohedges]

@mb2055: Are you able to review?

@xiki-tempula: I've already mentioned twice that we need a review for this. We have an internal process to follow with a limited team who are busy working on other things. I see that you have already manually applied on of the fixes here to your own fork, i.e. in this commit. I know your fork is based from our main, but is there a reason that you can't apply patches from fix PRs for things that are blocking you? (It seems that you do manually add certain fixes, but not others.)

[mb2055]

Sure thing, I'll go through it now

[xiki-tempula]

Sorry, I was reading too fast and didn't noticed that it has not been reviewed. I thought that the only blocker is failed pipeline. I will manually apply the fix then.

[lohedges]

Thanks. I'm off for the school run in a mo, but should be able to check back later depending on the level of carnage.

[lohedges]

No problem. For fixes like this it is probably best (in future) that someone from your end reviews it, since you have a detailed understanding of the issue and can easily test if it works. I think you can add reviews (you have done before) but aren't available from the drop-down for some reason. (Presumably I need to adjust the permissions.)

[xiki-tempula]

Thanks for mentioning it. I always thought that only the OBS people can be asked to review these PRs.

[mb2055]

Everything working on my end 👍 sounds like everything is working for @xiki-tempula so it should be good to go