+1 for avoiding merge commits in favour of linear history. While undocumented, I learned it on the job by reading the commit history which is mostly linear except developing big feature branches like the major upgrade. Of course, it would be better if the policy is documented clearly somewhere.
A couple points that support linear commit history: * Camel accepts many community contributions via GitHub pull reqs. Not rebasing them would add non-necessary merge-point commits in the history and make it lengthier. * We have many regular committers who work in parallel. Allowing merges would make the history messy and hard to trace changes for troubleshooting and learning from history. On Thu, Apr 16, 2020 at 7:19 AM David Jencks <david.a.jen...@gmail.com> wrote: > I’m in favor of changing the GitHub camel repositories to enforce > rebase-and-merge: Rebasing and merging your commits < > https://help.github.com/en/github/administering-a-repository/about-merge-methods-on-github#rebasing-and-merging-your-commits>. > This will produce linear history. IIUC, the committer on each commit will > change to the person doing the rebase-and-merge, and the author will remain > unchanged. > > Is it possible to do something equivalent on the Apache git repo? > > AFAICT without this, GitHub merges will definitely have a merge commit and > may have non-linear history. > > Thanks > David Jencks > > > On Apr 15, 2020, at 2:08 PM, Andrea Cosentino <anco...@gmail.com> wrote: > > > > I don't think there is a consensus around this.. we can collect here the > > different views. Personally I don't think merge commits are evil, but if > a > > clean history is desired it's fine for me > > > > > > Il mer 15 apr 2020, 22:09 Pascal Schumacher <pascalschumac...@gmx.net> > ha > > scritto: > > > >> Hi, > >> > >> recently there were several merge commits (especially for merged pull > >> request). > >> > >> I thought the consensus was to avoid merge commits to keep the git > >> history as clean as possible. > >> > >> Should we keep this policy? > >> > >> What do you think? > >> > >> Cheers, > >> > >> Pascal > >> > >> > >> > > -- Tadayoshi Sato
