git: 5a20c351ea45 - main - [skip ci] add a CODEOWNERS file

Jessica Clarke jrtc27 at freebsd.org
Mon May 31 19:51:44 UTC 2021


On 31 May 2021, at 20:50, Alan Somers <asomers at freebsd.org> wrote:
> 
> On Mon, May 31, 2021 at 1:39 PM Jessica Clarke <jrtc27 at freebsd.org> wrote:
> On 31 May 2021, at 20:37, Warner Losh <imp at bsdimp.com> wrote:
> > 
> > 
> > 
> > On Mon, May 31, 2021 at 1:21 PM Jessica Clarke <jrtc27 at jrtc27.com> wrote:
> > On 31 May 2021, at 20:16, Warner Losh <imp at bsdimp.com> wrote:
> > > 
> > > 
> > > 
> > > On Mon, May 31, 2021, 1:11 PM Ed Maste <emaste at freebsd.org> wrote:
> > > On Mon, 31 May 2021 at 11:51, Alan Somers <asomers at freebsd.org> wrote:
> > > >
> > > >>
> > > >> I think this file isn't the right place for (another copy of) this
> > > >> text; perhaps the CODEOWNERS file should just reference the top-level
> > > >> MAINTAINERS?
> > > >
> > > > Except that CODEOWNERS is in a format that tools know how to parse.  If anything, MAINTAINERS should be a symlink to CODEOWNERS.
> > > 
> > > At least the file's location and user IDs suggest that it is specific
> > > to GitHub; in any case I don't really care which one points to which.
> > > I hope we can agree though that we don't really want two different
> > > files representing code ownership in different ways that both
> > > independently refer to a third mechanism for recording code ownership
> > > that's external to the source tree?
> > > 
> > > It also works on gitlab, FWIW. The format is standard. Bummer we can't generate it based on where it is published…
> > 
> > We could conceivably have a CODEOWNERS.master from which the others can be generated via `make codeowners` whenever someone edits it. Whether that’s worth the hassle of implementing though for a file that shouldn’t regularly be changing is unclear.
> > 
> > Doing that's trivial... However, it would mean that the automatic flagging that this file's presence gives wouldn't work... But then again, there's no issues or pull requests at the gitlab mirror, so at the moment it's a bit of a moot point.
> 
> A hypothetical .gitlab/CODEOWNERS would be checked in as a generated file, if that was unclear. Otherwise I’m not sure I understand your first point?
> 
> Jess
> 
> There's no need to generate anything.  Both github and gitlab support a CODEOWNERS file in the project's root.  Let's just move it there.  We can delete the old MAINTAINERS, too, once all of its entries have been converted.

Not if people have different usernames between FreeBSD, GitHub and GitLab? Or can it work off email addresses and we can rely on people adding their @FreeBSD.org to their account? Though that still doesn’t work for groups.

Jess



More information about the dev-commits-src-all mailing list