Re: [PATCH] clone: teach --single-branch and --branch during --recurse

From: Emily Shaffer <emilyshaffer@google.com>
Date: Jan 27, 2020, 22:49 UTC
Message-ID: <20200127224914.GB233139@google.com>
In-Reply-To: <20200127222019.GA233139@google.com>

On Mon, Jan 27, 2020 at 02:20:19PM -0800, Emily Shaffer wrote:

Show 13 quoted lines

> On Fri, Jan 17, 2020 at 04:03:19PM -0500, Jeff King wrote:
> 
> > (like trying to replace the use of "repo" in Android)
> 
> Oops, you saw right through us ;)
> 
> > It would make more sense to me to either (or both):
> > 
> >   - make sure that .gitmodules has enough information about which branch
> >     to use for each submodule
> 
> Hum. I don't work with them day to day, but aren't we already in that
> state? Is that not what the 'branch' option for each submodule means?

I've been corrected off-list that the 'branch' in .gitmodules is used during 'git submodule update --remote', but not during 'git submodule init' or 'git clone --recurse-submodules'. Then, for the problem in discussion for this thread, it seems like a better choice is something like 'git clone --recurse-submdoules --use-gitmodules' or whatever we want to call it - e.g., rather than fetching the branch where the server knows HEAD, ask the .gitmodules to figure out which branch?

It seems like that ought to live separately from --single-branch. In the case where you very strictly only want to fetch one branch (not two branches) I suppose you'd want something like 'git clone --recurse-submodules --single-branch --branch=mysuperprojectbranch --use-gitmodules' to make sure that only one branch per repo comes down.

With n submodules of various naming schemas, provenance, etc., I don't think there's a good case for recursing --branch one way or another; it seems like filling out some config is the way to go.

I guess we could also teach it to take some input like --submodule-branch-spec=foo.txt, and/or a multiply provided --submodule-branch foo=foobranch --submodule-branch bar/baz=bazbranch.

  [foo.txt]
  foo=foobranch
  bar/baz=bazbranch

With that approach, then someone gets a little more flexibility than relying on what the .gitmodules has set up.

Show 15 quoted lines

> >   - offer an extra option for the default branch to use for any
> >     submodules. This is still not general enough to cover all situations
> >     (e.g., the bar/baz you showed above), but it at least makes it
> >     relatively easy to cover the simple cases, without breaking any
> >     existing ones.
> 
> Yeah, this is sort of the direction my mind went too - "not
> --branch recursively, but --submodule-branch". But that breaks down when you've
> got a nontrivial number of submodules, at which point you're gonna have
> a hard time unless you've got the .gitmodules configured correctly.
> 
> 
> Well, as for this patch, let me try it with just --single-branch and see
> whether that works for the case the user reported. I can head back to
> the drawing board if not.

With only half the rework of my patch done, I'm starting to convince myself it's not actually going to work :) Well, I'll still try and see.

 - Emily