continued discussion on git commit messages

[ng0 <ng0@infotropique.org>]

[-- Attachment #1: Type: text/plain, Size: 2348 bytes --]

It makes more sense to move this to guix-devel, the bug is finished.

Arun Isaac transcribed 1.6K bytes:
> 
> Tobias Geerinckx-Rice writes:
> 
> >> Also, having a staging branch to which all commits are made would help
> >> keep the master branch clean, and free of these mishaps. But, I guess
> >> this issue has been raised before, and the creation of such a staging
> >> branch has been postponed.
> >
> > Interesting. Do you have a link? We do have a ‘staging’ branch, but it
> > serves a different purpose (basically, ‘core-updates’, lite).
> 
> I was referring to this thread discussing the stability of master. It's
> not about any staging branch like I proposed, but there are some
> similarities.
> 
> https://lists.gnu.org/archive/html/guix-devel/2017-07/msg00021.html
> 
> > Since the commit message is already reviewed on the mailing list, such a
> > branch would only catch the — one hopes — very rare last-minute foul-up
> > by the committer themselves.
> >
> > ‘Someone’ would have to keep track of, review, and merge it. ‘People’
> > would then start following that branch instead of master (Why let a typo
> > keep me from the shiny?), making rebasing just as painful.
> 
> Agreed.
> 
> > I'm less opposed to an optional hook, but fear that it will discourage
> > committers from making a habit of checking all their commits before that
> > last and final push. Which they should definitely be doing.
> 
> Despite these concerns, I think it is still worth automating some commit
> checks. Given long enough time, even the most careful person is likely
> to mess up somewhere.

That's what I (although very obscure) meant to express with my grumpy
comments a while back. We could have a bot merging this after checks have
been run (mostly what I was concerned with is health of repository and
general build checks). Ludovic said we'd need to change large parts of our
workflow for this, and someone needs to work on getting this implemented.

> > (That's not a dig at you or anyone personally — you're talking to the
> > reason for the /current/ pre-push hook, after all :-)
> 
> No issues! :-)
> 
> 
> 
> 

-- 
ng0
GnuPG: A88C8ADD129828D7EAC02E52E22F9BBFEE348588
GnuPG: https://n0is.noblogs.org/my-keys
https://www.infotropique.org https://krosos.org

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]