On Wed, Jun 10, 2020 at 09:19:23PM +0000, Johannes Schindelin via GitGitGadget wrote:
> @@ -2099,7 +2100,10 @@ struct ref *guess_remote_head(const struct ref *head,
>
> /* If refs/heads/master could be right, it is. */
> if (!all) {
> - r = find_ref_by_name(refs, "refs/heads/master");
> + char *name = git_default_branch_name(0);
> +
> + r = find_ref_by_name(refs, name);
> + free(name);
> if (r && oideq(&r->old_oid, &head->old_oid))
> return copy_ref(r);
> }
You'd perhaps want to update the comment above, too.
However, I think we should be a bit more lenient on the "reading" side
default names. Just because "foo" is _my_ default branch name, does not
mean it is the default on the remote side. We cannot know what the other
side's default preference is, but in a world where we have 15 years of
repos that may have been created with "master", it is probably still a
good guess.
I.e., I think this probably ought to check the preferred name, and then
fall back to the existing behavior, like:
if (!all) {
char *name;
/* try the user's preferred default branch name */
name = git_default_branch_name(0);
r = find_ref_by_name(refs, name);
free(name);
if (r && oideq(&r->old_oid, &head->old_oid))
return copy_ref(r);
/* otherwise, try "master", which is the historical default */
r = find_ref_by_name(refs, "refs/heads/master");
if (r && oideq(&r->old_oid, &head->old_oid))
return copy_ref(r);
}
That will help minimize fallout when git_default_branch_name() changes,
either by user config or if we switch the baked-in default. In the
latter case, we might also consider hard-coding that as a guess between
the user's preferred name and the historical "master".
Hopefully this would not matter _too_ much either way, as most servers
would support the symref extension these days. But I still think we
should do our best to minimize spots where the user may see a
regression.
-Peff
