Re: js/early-config-with-onbranch, was Re: What's cooking in git.git (Aug 2019, #01; Thu, 1)

From: Jeff King <peff@peff.net>
Date: Aug 2, 2019, 18:38 UTC
Message-ID: <20190802183840.GB26918@sigill.intra.peff.net>
In-Reply-To: <nycvar.QRO.7.76.6.1908021410540.46@tvgsbejvaqbjf.bet>

On Fri, Aug 02, 2019 at 02:12:34PM +0200, Johannes Schindelin wrote:

Show 21 quoted lines

> > > * js/early-config-with-onbranch (2019-07-31) 1 commit
> > >   (merged to 'next' on 2019-08-01 at 26b713c824)
> > >  + config: work around bug with includeif:onbranch and early config
> > >
> > >  The recently added [includeif "onbranch:branch"] feature does not
> > >  work well with an early config mechanism, as it attempts to find
> > >  out what branch we are on before we even haven't located the git
> > >  repository.  The inclusion during early config scan is ignored to
> > >  work around this issue.
> > >
> > >  Will merge to 'master'.
> >
> > I had some open comments here on how the "do we have a repo" check is
> > done, but I think what is committed here is functionally equivalent. I
> > can pursue the NULL the_repository cleanups separately.
> 
> Right, I read that as "maybe later" comments, as that project is so
> different from trying to fix the `onbranch` feature with regards to the
> early config machinery.
> 
> Or did I misread and you want me to do anything about it before v2.23.0?

I think when I wrote my original replies in the thread that I wasn't sure _what_ I wanted. :) I had originally hoped to convince you to switch to have_git_dir() now, but I do agree after investigating that it would require an extra patch to common-main.

Mostly I was just trying to clarify here that my comments were not blockers for merging this; I couldn't find any case where your patch as-is would behave incorrectly.

-Peff