On Sun, Sep 13, 2020 at 6:40 PM Gregory Nutt <spudan...@gmail.com> wrote:
> > > I've been assiging various PRs from both repos to 10.0 milestone. > > Please also see if you think not assigned ones are to be surely included > > and assign them. I've not done this with issues as many are old ones > > for which no PR was proposed. > > > > Regarding bumping release: I think we had many major changes > > and I was under the impression that we indeed should do a major > > bump (hence the 10.0 milestone name). > > > > But maybe I've read that wrong > > Well, I think that the issue is that we need to decide what the major > release number means. If it means that the code users have under 9.x > will not no longer work, it should be boosted to 10.0. Otherwise, under > what conditions to we boost the major revision number? > > Going from 8.x to 9.0 did not follow these rules. There was not > particular incompatibility at all. The major number changed only > because it was the first Apache release. Other people have suggested > boosting the major number to celebrate the NuttX2019 event. The only > point is that we need to decide what these numbers mean and be consistent. > > As a user, I expect to see issues if I upgraded from (n).(x) to (n+1).0. > I would not expect to see issues if I upgraded from (n).(x) to > (n).(n+1). What would be your expectation? > > > I think the best policy is described here: https://semver.org/ New major version = backwards incompatible API changes. New minor version = backwards compatible. New patch version = bugfixes (and backwards compatible obviously) Do we want to adopt that? Nathan
