Re: [PATCH v2 09/12] clone: handle overridden main branch names

From: Jeff King <peff@peff.net>
Date: Jun 16, 2020, 13:22 UTC
Message-ID: <20200616132216.GH666057@coredump.intra.peff.net>
In-Reply-To: <0e59b6181699abe17eb46fe3ca5a48ce71889780.1592225416.git.gitgitgadget@gmail.com>

On Mon, Jun 15, 2020 at 12:50:13PM +0000, Johannes Schindelin via GitGitGadget wrote:

Show 12 quoted lines

> diff --git a/builtin/clone.c b/builtin/clone.c
> index 487b0a42d75..755fcaeb0ba 100644
> --- a/builtin/clone.c
> +++ b/builtin/clone.c
> @@ -718,6 +718,7 @@ static void update_head(const struct ref *our, const struct ref *remote,
>  		/* Local default branch link */
>  		if (create_symref("HEAD", our->name, NULL) < 0)
>  			die(_("unable to update HEAD"));
> +		git_config_set("core.mainbranch", head);
>  		if (!option_bare) {
>  			update_ref(msg, "HEAD", &our->old_oid, NULL, 0,
>  				   UPDATE_REFS_DIE_ON_ERR);

Just making sure I understand what's going on here...

This covers the case that we've run "clone -b foo" or similar, but there are two other case arms when "foo" is a tag, or the remote HEAD is unreachable. And there we don't set core.mainbranch at all.

But we would not want it to be missing, because that will likely need to stay a default for "master" indefinitely (to keep behavior for existing repositories). However, it won't be missing. We'll always have set it during the init_db() call, and this is just overriding that. So we'd end update_head() with either:

  - core.mainbranch set to the same branch we point HEAD to, whether we
    got it from the remote side or from "-b foo"

  - if we write a detached HEAD, then core.mainbranch remains at
    init.mainbranch (or defaulting to "master" now, and probably "main"
    later). We have no better option.

If so, then that makes sense to me.

-Peff