From: Patrick Steinhardt <ps@pks.im>
Date: Tue, 06 Oct 2026 12:18:14 GMT
Subject: Re: [PATCH 01/13] commit-graph: require resolved packfile paths for `stdin_packs`
Message-ID: <asTnBsWxEvgphOV7@pks.im>
In-Reply-To: <CAOLa=ZS_S3bYXEufor7AgXpX76NwtV4NJdGRO57-bpSkYZrzvw@mail.gmail.com>

On Mon, Oct 05, 2026 at 03:27:27PM -0400, Karthik Nayak wrote:
> Patrick Steinhardt <ps@pks.im> writes:
> 
> > Users can ask git-commit-graph(1) to write a commit graph specifically
> > for a set of packfiles via the "--stdin-packs" option. Those users are
> > expected to pass in relative paths, and those eventually get resolved in
> > `fill_oids_from_packs()`. This ties the logic in "commit-graph.c" to the
> > specific object database source.
> >
> > Refactor the logic to instead require the caller to pass in resolved
> > packfiles to untangle that dependency.
> >
> 
> Nit: the changes look good, what I'm missing is 'why' are we doing this
> change.

It's basically this sentence:

  This ties the logic in "commit-graph.c" to the specific object
  database source.

With the current logic we have assumptions in "commit-graph.c" about
where a specific packfile lives relative to an object database source.
But with the next commit that becomes a bit harder to realize So
resolving packfile paths early on gets rid of parts of these
assumptions, so that we don't have to resolve those paths in
"commit-graph.c" anymore.

I'll rephrase this a bit.

Patrick