aboutsummaryrefslogtreecommitdiff
path: root/packfile.c
diff options
context:
space:
mode:
authorPatrick Steinhardt <ps@pks.im>2026-03-20 08:07:39 +0100
committerJunio C Hamano <gitster@pobox.com>2026-03-20 13:16:42 -0700
commit6c2ede6e4abed754bb5891c2904212c05efcfb11 (patch)
tree84ad390cef9248deda96d8fca0dd290c2a69d23c /packfile.c
parentab3ab1038dd38d2be62e3bacf39a3248929a7a98 (diff)
downloadgit-6c2ede6e4abed754bb5891c2904212c05efcfb11.tar.xz
object-file: move logic to compute packed abbreviation length
Same as the preceding commit, move the logic that computes the minimum required prefix length to make a given object ID unique for the packfile store into a new function `packfile_store_find_abbrev_len()` that is part of "packfile.c". This prepares for making the logic fully generic via pluggable object databases. Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Diffstat (limited to 'packfile.c')
-rw-r--r--packfile.c111
1 files changed, 111 insertions, 0 deletions
diff --git a/packfile.c b/packfile.c
index 2539a371c1..ee9c7ea1d1 100644
--- a/packfile.c
+++ b/packfile.c
@@ -2597,6 +2597,117 @@ out:
return ret;
}
+static int extend_abbrev_len(const struct object_id *a,
+ const struct object_id *b,
+ unsigned *out)
+{
+ unsigned len = oid_common_prefix_hexlen(a, b);
+ if (len != hash_algos[a->algo].hexsz && len >= *out)
+ *out = len + 1;
+ return 0;
+}
+
+static void find_abbrev_len_for_midx(struct multi_pack_index *m,
+ const struct object_id *oid,
+ unsigned min_len,
+ unsigned *out)
+{
+ unsigned len = min_len;
+
+ for (; m; m = m->base_midx) {
+ int match = 0;
+ uint32_t num, first = 0;
+ struct object_id found_oid;
+
+ if (!m->num_objects)
+ continue;
+
+ num = m->num_objects + m->num_objects_in_base;
+ match = bsearch_one_midx(oid, m, &first);
+
+ /*
+ * first is now the position in the packfile where we
+ * would insert the object ID if it does not exist (or the
+ * position of the object ID if it does exist). Hence, we
+ * consider a maximum of two objects nearby for the
+ * abbreviation length.
+ */
+
+ if (!match) {
+ if (nth_midxed_object_oid(&found_oid, m, first))
+ extend_abbrev_len(&found_oid, oid, &len);
+ } else if (first < num - 1) {
+ if (nth_midxed_object_oid(&found_oid, m, first + 1))
+ extend_abbrev_len(&found_oid, oid, &len);
+ }
+ if (first > 0) {
+ if (nth_midxed_object_oid(&found_oid, m, first - 1))
+ extend_abbrev_len(&found_oid, oid, &len);
+ }
+ }
+
+ *out = len;
+}
+
+static void find_abbrev_len_for_pack(struct packed_git *p,
+ const struct object_id *oid,
+ unsigned min_len,
+ unsigned *out)
+{
+ int match;
+ uint32_t num, first = 0;
+ struct object_id found_oid;
+ unsigned len = min_len;
+
+ num = p->num_objects;
+ match = bsearch_pack(oid, p, &first);
+
+ /*
+ * first is now the position in the packfile where we would insert
+ * the object ID if it does not exist (or the position of mad->hash if
+ * it does exist). Hence, we consider a maximum of two objects
+ * nearby for the abbreviation length.
+ */
+ if (!match) {
+ if (!nth_packed_object_id(&found_oid, p, first))
+ extend_abbrev_len(&found_oid, oid, &len);
+ } else if (first < num - 1) {
+ if (!nth_packed_object_id(&found_oid, p, first + 1))
+ extend_abbrev_len(&found_oid, oid, &len);
+ }
+ if (first > 0) {
+ if (!nth_packed_object_id(&found_oid, p, first - 1))
+ extend_abbrev_len(&found_oid, oid, &len);
+ }
+
+ *out = len;
+}
+
+int packfile_store_find_abbrev_len(struct packfile_store *store,
+ const struct object_id *oid,
+ unsigned min_len,
+ unsigned *out)
+{
+ struct packfile_list_entry *e;
+ struct multi_pack_index *m;
+
+ m = get_multi_pack_index(store->source);
+ if (m)
+ find_abbrev_len_for_midx(m, oid, min_len, &min_len);
+
+ for (e = packfile_store_get_packs(store); e; e = e->next) {
+ if (e->pack->multi_pack_index)
+ continue;
+ if (open_pack_index(e->pack) || !e->pack->num_objects)
+ continue;
+
+ find_abbrev_len_for_pack(e->pack, oid, min_len, &min_len);
+ }
+
+ *out = min_len;
+ return 0;
+}
+
struct add_promisor_object_data {
struct repository *repo;
struct oidset *set;