Re: [PATCH 01/16] path: refactor `repo_common_path()` family of functions

From: Patrick Steinhardt <ps@pks.im>
Date: Feb 7, 2025, 06:16 UTC
Message-ID: <Z6WlNXbw6s4ojsiP@pks.im>
In-Reply-To: <CAOLa=ZRSSnqRO_=N-sVrYSRpCZkyW0KFh5=rwX0NY_VqyzKz3A@mail.gmail.com>

On Thu, Feb 06, 2025 at 03:17:12AM -0800, Karthik Nayak wrote:

Show 19 quoted lines

> Patrick Steinhardt <ps@pks.im> writes:
> 
> [snip]
> 
> > diff --git a/path.c b/path.c
> > index 07964f5d32..273b649e00 100644
> > --- a/path.c
> > +++ b/path.c
> > @@ -414,7 +414,7 @@ static void strbuf_worktree_gitdir(struct strbuf *buf,
> >  	else if (!wt->id)
> >  		strbuf_addstr(buf, repo->commondir);
> >  	else
> > -		strbuf_git_common_path(buf, repo, "worktrees/%s", wt->id);
> > +		repo_common_path_append(repo, buf, "worktrees/%s", wt->id);
> 
> This is the only place we replace `strbuf_git_common_path` with
> `repo_common_path_append`, whereas other places it is
> `repo_common_path_replace`. I assume that is because in those places the
> buf is newly initialized.

Yes, exactly. I think that it's easier to reason about to replace callsites with `_replace()` where we can. But I'm happy to adjust the approach when it's putting too much of a burden on the reviewer.

Patrick