Re: [RFC] Cache negative delta pairs

[Junio C Hamano <junkio@cox.net>]

Nicolas Pitre <nico@cam.org> writes:

> The negative delta cache concept is certainly attractive even for normal 
> repositories, especially for public servers, since when used in 
> conjonction with delta reuse it makes the creation of a pack basically 
> free.  So I think this idea really has merits, as long as the cache 
> remains small.

Yes, I agree it is very attractive.

One thing to watch out for is that we probably would not want to
let git-daemon write into public repositories.  Which means that
use of negative cache should be strict "opt-in".

 - "$GIT_DIR/delta-cache" is read but not necessarily is written
   back when it exists; git-daemon uses it that way.

 - The owner of the repository shouldn't have to tell the tool
   to update the negative cache every time repack happens.

Which suggests that pack-objects.c can learn an option that
tells it to call delta_cache_save(), and we use it in
git-repack, perhaps like this:

diff --git a/git-repack.sh b/git-repack.sh
index 640ad8d..b07ed9b 100755
--- a/git-repack.sh
+++ b/git-repack.sh
@@ -44,7 +44,7 @@ case ",$all_into_one," in
 esac
 pack_objects="$pack_objects $local $quiet $no_reuse_delta$extra"
 name=$(git-rev-list --objects --all $rev_list 2>&1 |
-	git-pack-objects --non-empty $pack_objects .tmp-pack) ||
+	git-pack-objects --update-delta-cache --non-empty $pack_objects .tmp-pack) ||
 	exit 1
 if [ -z "$name" ]; then
 	echo Nothing new to pack.
 
diff --git a/pack-objects.c b/pack-objects.c
index bed2497..46b9775 100644
--- a/pack-objects.c
+++ b/pack-objects.c
...
@@ -1342,5 +1350,7 @@ int main(int argc, char **argv)
 	if (progress)
 		fprintf(stderr, "Total %d, written %d (delta %d), reused %d (delta %d)\n",
 			nr_result, written, written_delta, reused, reused_delta);
+	if (update_delta_cache)
+		delta_cache_save();
 	return 0;
 }