Re: [PATCH 1/1] builtin/pack-objects.c: avoid iterating all refs

On Tue, Jan 19, 2021 at 06:08:31PM -0500, Taylor Blau wrote:

quoted

diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c
index 2a00358f34..2b32bc93bd 100644
--- a/builtin/pack-objects.c
+++ b/builtin/pack-objects.c

@@ -3740,7 +3740,7 @@ int cmd_pack_objects(int argc, const char **argv, const char *prefix)
 	}
 	cleanup_preferred_base();
 	if (include_tag && nr_result)
-		for_each_ref(add_ref_tag, NULL);
+		for_each_fullref_in("refs/tags/", add_ref_tag, NULL, 0);

Makes sense. I was curious why it wasn't this way from the beginning in
f0a24aa56e (git-pack-objects: Automatically pack annotated tags if
object was packed, 2008-03-03).

The patch doesn't say, but presumably it was because there wasn't any
speed-up to be had iterating only a subset of references back then if
they were packed (did packed refs even exist in 2008? unsure). In any
case, builtin/pack-objects.c:add_ref_tag() discards anything that
doesn't start with "refs/tags/", so I think your change is doing the
right thing here.

That said, you could shorten this to use 'for_each_tag_ref()' instead
(which compiles to the exact same thing).

You'd end up with "v1.2.3" in the refname field of the callback then,
rather than "refs/tags/v1.2.3". So we'd definitely need to drop the
prefix check in add_ref_tag() then (though I think it is a good idea
anyway). But I'm also not sure if it would interfere with peel_ref(),
and its magic for using packed-refs peeled information.

-Peff