Currently macOS eats 2 CI tasks. This is fine, but not needed and duplicates compilation time and space requirements.
Slim down the tasks into one to reduce the ccache footprint, as well as total compilation time.
To give some numbers for the macOS task: It passed here in 29min (with an empty ccache), whereas on master, the times were 19min+14min=32min (https://github.com/bitcoin/bitcoin/actions/runs/14638899803/job/41076397151)
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/32339.
See the guideline for information on the review process. A summary of reviews will appear here.
Reviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
Debug, do we really want that?
Don’t these jobs run in parallel? i.e. how does this save time?
The tasks don’t run in parallel, when the GHA limit is reached. However this should be rare. The time and space savings come from re-using build artefacts that can be shared between the fuzz binary and other binaries.
It is true that the end-to-end duration of the combined task is now longer, albeit it being less than the sum of the times of the split tasks. I don’t think this is a concern, but no strong opinion.
This requires changing the build type to
Debug, do we really want that?
I’d say yes, because the optimization flags are unchanged here. Enabling the extra debug checks seems reasonable for CI, as the risk would otherwise be that no one else is running them on the platforms.