-
Notifications
You must be signed in to change notification settings - Fork 36
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use MarkAllButFirstSubBags for T_BIPART #1003
Conversation
T_BIPART consists of a general pointer not managed by GAP's GC, plus two bag references (which may be NULL). Hence we can use MarkAllButFirstSubBags instead of a custom marking function.
(I based this on |
Thanks @fingolfin either branch is fine. Any idea why the ci is failing looks related to recent changes in GAP and not related to this pr at all |
The CI failure are due to the stable-4.12 tests using the latest package distro, which has a newer GAPDoc, which makes one of the tests in stable-4.12 fail. The ideal solution would be to test against GAP 4.12.2 (the release version) with its bundled packages, instead of against stable-4.12 plus "latest" packages. See gap-actions/setup-gap#24. This is something someone could look into during GAP Days next week (but realistically I think of the remaining "active" people only @james-d-mitchell, @ChrisJefferson and myself qualify... I can see if I can find a "young volunteer" to teach them, but I don't see it yet. Ah well) A quicker workaround for the interim would be to insert a step between building GAP and running its test suite where you just remove (or update) the offending .tst file. |
@james-d-mitchell do you think you might have time to get this into a new Semigroups release within the next couple days so we could release GAP 4.13.0 next week? Not trying to pressure you here at all, just trying to gauge what is still possible and what not :-) |
@fingolfin yes definitely, I can make a release, would like to see the CI passing first if possible. Happy to have the easy resolution of just deleting the offending files if we're using the wrong version version of GAP. |
Basically if running the tests in stable-4.12, delete GAP's |
T_BIPART consists of a general pointer not managed by GAP's GC, plus two bag references (which may be NULL). Hence we can use MarkAllButFirstSubBags instead of a custom marking function.
This prepares Semigroups for gap-system/gap#5662, but also works in previous GAP versions.