Release candidate for v5.6.3#1182
Conversation
Global variables are zero-initialized anyway, so this makes no difference. Note that module state only matters for HPC-GAP. In the future, I'd like to hide the moduleStateSize and moduleStateOffsetPtr for "regular" GAP builds.
james-d-mitchell
added
the
release-candidate
|
@james-d-mitchell I tested this and to the best of my ability am convinced it will pass all PackageDistro tests and be fine. So let's say I am 99.9% confident. Anyway: even if I am wrong and there is some further obscure issue, then this is not necessarily a blocker for the release; we could agree to ignore it and postpone a package distro fix for later; I can make a 4.16.1 release quickly as needed. My main concern is that I'd like to avoid a release of GAP that breaks a major package like semigroups. In that sense, I guess even Semigroups 5.6.2 really is fine, because the "breakage" mostly is one test of the test suite that is not important for actual usage of the package. It's just that I'd rather avoid having "semigroups" always fail in package distro tests, because it means that actual failures can slip in without being noticed. tl;dr: releasing this as-is seems perfect, thank you so much, I'll take responsibility for dealing with any further issues |
|
Thanks @fingolfin happy to make a release, and don't want to be a blocker for releasing GAP 4.16! |
2 participants
@fingolfin can you please confirm that if this release is made and merged into
main, then this will allow GAP 4.16.0 to be released (at least as far as the Semigroups package is concerned of course). Thanks!