Nguyễn Thái Ngọc DuyIn pack version 4, ref-delta technically could be used to compress any objects including commits and trees (both canonical and v4). But it does not make sense to do so. It can only lead to performance degradation. Catch those packers.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> --- I could now verify that pack-objects does not compress commits nor trees using ref-delta in v4. But perhaps these are a bit too strict? Maybe downgrade from die() to warning() and still accept the pack?
builtin/index-pack.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
Show changes to builtin/index-pack.c +16 −0
diff --git a/builtin/index-pack.c b/builtin/index-pack.c index fbf97f0..e4ecf69 100644 --- a/builtin/index-pack.c +++ b/builtin/index-pack.c @@ -788,6 +788,12 @@ static void *unpack_raw_entry(struct object_entry *obj, case OBJ_BLOB: case OBJ_TAG: break; + + /* + * OBJ_PV4_* are all greater than 7, which is the limit of + * field "type" in pack v2. So we do not really need 'if + * (!packv4) die("wrong type");' here. + */ case OBJ_PV4_COMMIT: obj->real_type = OBJ_COMMIT; break; @@ -1288,6 +1294,16 @@ static void resolve_delta(struct object_entry *delta_obj, hash_sha1_file(result->data, result->size, typename(delta_obj->real_type), delta_obj->idx.sha1); check_against_sha1table(delta_obj->idx.sha1); + if (packv4 && + delta_obj->type == OBJ_REF_DELTA && + delta_obj->real_type == OBJ_TREE) + bad_object(delta_obj->idx.offset, + _("ref-delta on a tree is not supported in pack version 4")); + if (packv4 && + delta_obj->type == OBJ_REF_DELTA && + delta_obj->real_type == OBJ_COMMIT) + bad_object(delta_obj->idx.offset, + _("ref-delta on a commit is not supported in pack version 4")); sha1_object(result->data, NULL, result->size, delta_obj->real_type, delta_obj->idx.sha1); counter_lock();
-- 1.8.2.83.gc99314b
