It happens that master fixes some issues present in the released version. Even if I would not say the new code for parameters is well tested, it may make sense to plan a release.
For example @Christian Doczkal just stumbled on a bug in
HB.instance that is just not there in master and the code is too different to make backporting the fix worth it. We are currently piling up fixes, issues and todos coming from the porting of mathcomp, but it may take a while before we reach modules, which would kind of be the testing we are looking for, ditto for the todos which may take time to implement. Also, I'm not aware of regressions w.r.t. the code that worked in 0.9.
What do you think? Shall we open a 0.10 milestone and keep 1.0 for the one to be reached after mathcomp is ported/portable to HB?
Last updated: Jan 29 2023 at 15:02 UTC