The proposal looks good but we should also consider if we want to support LTS maintenance releases containing bugfixed backported from mainline
Regards Alin On Fri, Dec 13, 2019, 09:27 David Sidrane <david.sidr...@nscdg.com> wrote: > Precisely! We cut a branch as a Release Candidate. nuttx-MM.mm.rr-rcnn. > During the release cycle it can have back ports from master if a new > feature > or bug fix is found it is added if deemed necessary to the release. > > Workflow Proposal > > I would ask that we adopt a workflow similar to PX4. [1] - see page 16. Or > parts of it that work for us. > > All development would be done on branches and only squashed atomic commits > would be done to master. (Think constantly working bisect on master) > > Branches can be rebased - to reduce noise until review [1] see page 17 > > All commit message would have context. [1] - see page 16 (Think emails > subject tells me what happed, Filter out noise) > > Regards, > > David > > [1] https://drive.google.com/open?id=1XHaNB3nTarjPL-CzO9CpZTl1NgX_kpqY > > > -----Original Message----- > From: spudaneco [mailto:spudan...@gmail.com] > Sent: Thursday, December 12, 2019 9:16 PM > To: dev@nuttx.apache.org > Subject: Re: Project Emails > > Sent from Samsung tablet. > I think we should use release branches, unless we want to lock downthe repo > against all changes for the duration of the release prep ->release > candidate -> testing -> debating -> voting cycle, which couldtake a non > trivial amount of time.You could carry on business as usual on a temporary > branch, but that feels awkward. >
