Hi,
On Mon, 3 Mar 2008, Kristian Høgsberg wrote:
> On Sun, 2008-03-02 at 06:25 +0000, Johannes Schindelin wrote:
> > Signed-off-by: Johannes Schindelin <johannes.schindelin@xxxxxx>
> > ---
> >
> > Sorry, my previous patch was broken in so many ways. This one
> > is better, promise.
> >
> > BTW this incidentally fixes the branch.<branch>.{remote,merge}
> > setup: it used to strip all up-to and including a slash from the
> > ref name. This just _happened_ to work, because commonly HEAD is
> > at "refs/heads/master". However, if it is at "refs/heads/a/b", it
> > would fail.
> >
> > builtin-clone.c | 35 ++++++++++++++++++++---------------
> > 1 files changed, 20 insertions(+), 15 deletions(-)
> >
> > diff --git a/builtin-clone.c b/builtin-clone.c
> > index 056e8a3..f27d205 100644
> > --- a/builtin-clone.c
> > +++ b/builtin-clone.c
> > @@ -523,33 +523,38 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
> > git_config_set_multivar(key, value, "^$", 0);
> > }
> >
> > + if (head_points_at)
> > + /* Local default branch */
> > + create_symref("HEAD", head_points_at->name, NULL);
> > +
> > if (option_bare) {
> > - if (head_points_at) {
> > - /* Local default branch */
> > - create_symref("HEAD", head_points_at->name, NULL);
> > - }
> > junk_work_tree = NULL;
> > junk_git_dir = NULL;
> > return 0;
> > }
> >
> > if (head_points_at) {
> > - if (strrchr(head_points_at->name, '/'))
> > - head = strrchr(head_points_at->name, '/') + 1;
> > - else
> > - head = head_points_at->name;
> > + struct strbuf head_ref, real_ref;
> >
> > - /* Local default branch */
> > - create_symref("HEAD", head_points_at->name, NULL);
> > + head = head_points_at->name;
> > + if (!prefixcmp(head, "refs/heads/"))
> > + head += 11;
> >
> > /* Tracking branch for the primary branch at the remote. */
> > update_ref(NULL, "HEAD", head_points_at->old_sha1,
> > NULL, 0, DIE_ON_ERR);
> > - /*
> > - rm -f "refs/remotes/$origin/HEAD"
> > - git symbolic-ref "refs/remotes/$origin/HEAD" \
> > - "refs/remotes/$origin/$head_points_at" &&
> > - */
> > +
> > + strbuf_init(&head_ref, 0);
> > + strbuf_addstr(&head_ref, branch_top);
> > + strbuf_addstr(&head_ref, "/HEAD");
> > + delete_ref(head_ref.buf, head_points_at->old_sha1);
> > + strbuf_init(&real_ref, 0);
> > + strbuf_addstr(&real_ref, branch_top);
> > + strbuf_addch(&real_ref, '/');
> > + strbuf_addstr(&real_ref, head);
>
> What about just using
>
> strbuf_addf(&real_ref, "%s/%s", branch_top, head);
>
> Are you worried about performance? :-p
You know, just after sending, I thought the same.
> Oh and I'm wondering if
>
> strbuf_initf(&real_ref, "%s/%s", branch_top, head);
>
> would be a worthwhile addition to the strbuf API...
And exactly this was crossing my mind, too, as
static inline int strbuf_initf(struct strbuf *buf, const char *format, ...)
{
strbuf_init(buf, strlen(format));
return strbuf_addf(format, ...);
}
(just a sketch, but you get the idea...)
Ciao,
Dscho
