patchmake shallow repository deepening more network efficient

> First of all, I can't find any reason why thin pack generation is 
> explicitly disabled when dealing with a shallow repository.  The 
> possible delta base objects are collected from the edge commits which 
> are always obtained through history walking with the same shallow refs 
> as the client, Therefore the client is always going to have those base 
> objects available. So let's remove that restriction.
>
> Then we can make shallow repository deepening much more efficient by 
> using the remote's unshallowed commits as edge commits to get preferred 
> base objects for thin pack generation.  On git.git, this makes the data 
> transfer for the deepening of a shallow repository from depth 1 to depth 2
> around 134 KB instead of 3.68 MB.
>
> Signed-off-by: Nicolas Pitre <nico@cam.org>

> diff --git a/upload-pack.c b/upload-pack.c
> index f73e3c9..c77ab71 100644
> --- a/upload-pack.c
> +++ b/upload-pack.c
> @@ -32,6 +32,7 @@ static int no_progress, daemon_mode;
>  static int shallow_nr;
>  static struct object_array have_obj;
>  static struct object_array want_obj;
> +static struct object_array extra_edge_obj;
>  static unsigned int timeout;
>  /* 0 for no sideband,
>   * otherwise maximum packet size (up to 65520 bytes).
> @@ -135,6 +136,10 @@ static int do_rev_list(int fd, void *create_full_pack)
>  	if (prepare_revision_walk(&revs))
>  		die("revision walk setup failed");
>  	mark_edges_uninteresting(revs.commits, &revs, show_edge);
> +	if (use_thin_pack)
> +		for (i = 0; i < extra_edge_obj.nr; i++)
> +			fprintf(pack_pipe, "-%s\n", sha1_to_hex(
> +					extra_edge_obj.objects[i].item->sha1));
>  	traverse_commit_list(&revs, show_commit, show_object, NULL);
>  	fflush(pack_pipe);
>  	fclose(pack_pipe);
> @@ -562,7 +567,6 @@ static void receive_needs(void)
>  		if (!prefixcmp(line, "shallow ")) {
>  			unsigned char sha1[20];
>  			struct object *object;
> -			use_thin_pack = 0;
>  			if (get_sha1(line + 8, sha1))
>  				die("invalid shallow line: %s", line);
>  			object = parse_object(sha1);
> @@ -574,7 +578,6 @@ static void receive_needs(void)
>  		}
>  		if (!prefixcmp(line, "deepen ")) {
>  			char *end;
> -			use_thin_pack = 0;
>  			depth = strtol(line + 7, &end, 0);
>  			if (end == line + 7 || depth <= 0)
>  				die("Invalid deepen: %s", line);
> @@ -657,6 +660,7 @@ static void receive_needs(void)
>  							NULL, &want_obj);
>  					parents = parents->next;
>  				}
> +				add_object_array(object, NULL, &extra_edge_obj);
>  			}
>  			/* make sure commit traversal conforms to client */
>  			register_shallow(object->sha1);

> Nicolas Pitre <nico@cam.org> writes:
> 
> > First of all, I can't find any reason why thin pack generation is 
> > explicitly disabled when dealing with a shallow repository.  The 
> > possible delta base objects are collected from the edge commits which 
> > are always obtained through history walking with the same shallow refs 
> > as the client, Therefore the client is always going to have those base 
> > objects available. So let's remove that restriction.
> >
> > Then we can make shallow repository deepening much more efficient by 
> > using the remote's unshallowed commits as edge commits to get preferred 
> > base objects for thin pack generation.  On git.git, this makes the data 
> > transfer for the deepening of a shallow repository from depth 1 to depth 2
> > around 134 KB instead of 3.68 MB.
> >
> > Signed-off-by: Nicolas Pitre <nico@cam.org>
> 
> Dscho, this is your code from around ed09aef (support fetching into a
> shallow repository, 2006-10-30) and f53514b (allow deepening of a shallow
> repository, 2006-10-30).  The latter disables thin pack transfer but the
> log does not attempt to justify the change.
> 
> Have any comments?