Hi Bert,
On Mon, 20 Jan 2020, Bert Wesarg wrote:
> On Mon, Jan 20, 2020 at 12:25 PM Johannes Schindelin
> <Johannes.Schindelin@xxxxxx> wrote:
> >
> > Hi Bert,
> >
> > On Fri, 17 Jan 2020, Bert Wesarg wrote:
> >
> > > On Fri, Jan 17, 2020 at 2:30 PM Johannes Schindelin
> > > <Johannes.Schindelin@xxxxxx> wrote:
> > > >
> > > > On Fri, 17 Jan 2020, Bert Wesarg wrote:
> > > >
> > > > > On Fri, Jan 17, 2020 at 12:50 PM Johannes Schindelin
> > > > > <Johannes.Schindelin@xxxxxx> wrote:
> > > > > AFAIU, the value of remote.pushDefault wont be renamed yet. So you
> > > > > suggest to issue a warning in case remote.pushDefault is X. But as X
> > > > > does not exists anymore after the rename, the value of
> > > > > remote.pushDefault is invalid. So why not rename it too?
> > > >
> > > > If this setting was usually a repository-specific one, I would suggest to
> > > > change its value, too. But it is my understanding that this might be set
> > > > in `~/.gitconfig` more often than not, so I recommend a warning instead.
> > >
> > > than why not rename it, if its a repository-specific setting and warn
> > > if it is a global one? If this is detectable at all.
> >
> > Sure, but you might need to re-parse the config to detect that (and you
> > have to use `git_config_from_file()` to make sure that you know that you
> > are looking at the repository config and not at anything else).
>
> I found current_config_scope() which serves the purpose for me.
> Anything wrong with this approach?
I guess you could go for that, even if it is not exactly elegant (and not
thread-safe, but who cares about that in `git remote`...). It would also
do way more work than you need.
If I were to implement this, I would definitely go for
`git_config_from_file(git_path("index"), ...)`.
Ciao,
Dscho
