Ensure we're testing sorcha on multiple systems/operating systems

[mschwamb]

It would be good to ensure that were testing sorcha's unit tests on multiple systems/operating systems
linux, mac osx arm64 and also intel macs.

[drewoldag]

Currently our unit tests run on ubuntu-latest and macos-latest operating systems.

The details can be seen here: https://docs.github.com/en/actions/using-github-hosted-runners/about-github-hosted-runners#supported-runners-and-hardware-resources

Basically this documentation indicates that the GitHub hosted runners that are used to execute the CI currently doesn't include ARM processors. However, I found this in the GitHub product roadmap: github/roadmap#528

It's already been pushed out at least 1 quarter, but if it doesn't get pushed again, then some time in Q4, we may be able to specify mac M1 hardware for running tests.

[mschwamb]

Leaving this so when Apple Silicon becomes available that we add it for the unit tests to run on

[mschwamb]

Looks like this is now available in beta for the larger runner https://github.blog/2023-10-02-introducing-the-new-apple-silicon-powered-m1-macos-larger-runner-for-github-actions/

[drewoldag]

I'm sure that you're already aware, but these are not free runners. In the blog post it calls out: "With today’s launch, our macOS larger runners will be priced at $0.16/minute for XL and $0.12/minute for large."

[mschwamb]

Yep. No plan to implement. Just updating the ticket with the latest info.

[mjuric]

I tried it & hit that roadblock (@drewoldag ). Looking to upgrade us to Team plan (looks like we're eligible for a freebie, as an educational institution).

[mjuric]

Actually. we're already supposed to have this (I knew I set this up...). Looks like it's a GitHub issue -- I'll ping support.

[drewoldag]

Neato

[mschwamb]

I was secretly hoping this was the case....

[mjuric]

Ticket opened.

And the PR is here, for when it gets resolved.

[mschwamb]

macos-14 is available which is a runner for Mac M1 - we're getting HDF5 errors. It's a configuration thing, but I am not 100% sure how to fix it with pip. I fix it by conda installing the package and then the libraries and environmental variables are correct