midx: implement verification support for incremental MIDXs

Teach the verification implementation used by `git multi-pack-index verify` to perform verification for incremental MIDX chains by independently validating each layer within the chain. Signed-off-by: Taylor Blau <me@ttaylorr.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Taylor Blau committed Aug 6, 2024 at 11:37 UTC 3592796d0a176f51bb257b31dd9afe12a5af7932
2 files changed +32 -17
midx.c
+30 -17
@@ -470,6 +470,13 @@ int prepare_midx_pack(struct repository *r, struct multi_pack_index *m,
470 return 0;
471 }
472
473 +struct packed_git *nth_midxed_pack(struct multi_pack_index *m,
474 + uint32_t pack_int_id)
475 +{
476 + uint32_t local_pack_int_id = midx_for_pack(&m, pack_int_id);
477 + return m->packs[local_pack_int_id];
478 +}
479 +
480 #define MIDX_CHUNK_BITMAPPED_PACKS_WIDTH (2 * sizeof(uint32_t))
481
482 int nth_bitmapped_pack(struct repository *r, struct multi_pack_index *m,
@@ -818,6 +825,7 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
825 uint32_t i;
826 struct progress *progress = NULL;
827 struct multi_pack_index *m = load_multi_pack_index(object_dir, 1);
828 + struct multi_pack_index *curr;
829 verify_midx_error = 0;
830
831 if (!m) {
@@ -840,8 +848,8 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
848
849 if (flags & MIDX_PROGRESS)
850 progress = start_delayed_progress(_("Looking for referenced packfiles"),
843 - m->num_packs);
844 - for (i = 0; i < m->num_packs; i++) {
851 + m->num_packs + m->num_packs_in_base);
852 + for (i = 0; i < m->num_packs + m->num_packs_in_base; i++) {
853 if (prepare_midx_pack(r, m, i))
854 midx_report("failed to load pack in position %d", i);
855
@@ -861,17 +869,20 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
869 if (flags & MIDX_PROGRESS)
870 progress = start_sparse_progress(_("Verifying OID order in multi-pack-index"),
871 m->num_objects - 1);
864 - for (i = 0; i < m->num_objects - 1; i++) {
865 - struct object_id oid1, oid2;
872
867 - nth_midxed_object_oid(&oid1, m, i);
868 - nth_midxed_object_oid(&oid2, m, i + 1);
873 + for (curr = m; curr; curr = curr->base_midx) {
874 + for (i = 0; i < m->num_objects - 1; i++) {
875 + struct object_id oid1, oid2;
876
870 - if (oidcmp(&oid1, &oid2) >= 0)
871 - midx_report(_("oid lookup out of order: oid[%d] = %s >= %s = oid[%d]"),
872 - i, oid_to_hex(&oid1), oid_to_hex(&oid2), i + 1);
877 + nth_midxed_object_oid(&oid1, m, m->num_objects_in_base + i);
878 + nth_midxed_object_oid(&oid2, m, m->num_objects_in_base + i + 1);
879
874 - midx_display_sparse_progress(progress, i + 1);
880 + if (oidcmp(&oid1, &oid2) >= 0)
881 + midx_report(_("oid lookup out of order: oid[%d] = %s >= %s = oid[%d]"),
882 + i, oid_to_hex(&oid1), oid_to_hex(&oid2), i + 1);
883 +
884 + midx_display_sparse_progress(progress, i + 1);
885 + }
886 }
887 stop_progress(&progress);
888
@@ -881,8 +892,8 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
892 * each of the objects and only require 1 packfile to be open at a
893 * time.
894 */
884 - ALLOC_ARRAY(pairs, m->num_objects);
885 - for (i = 0; i < m->num_objects; i++) {
895 + ALLOC_ARRAY(pairs, m->num_objects + m->num_objects_in_base);
896 + for (i = 0; i < m->num_objects + m->num_objects_in_base; i++) {
897 pairs[i].pos = i;
898 pairs[i].pack_int_id = nth_midxed_pack_int_id(m, i);
899 }
@@ -896,16 +907,18 @@ int verify_midx_file(struct repository *r, const char *object_dir, unsigned flag
907
908 if (flags & MIDX_PROGRESS)
909 progress = start_sparse_progress(_("Verifying object offsets"), m->num_objects);
899 - for (i = 0; i < m->num_objects; i++) {
910 + for (i = 0; i < m->num_objects + m->num_objects_in_base; i++) {
911 struct object_id oid;
912 struct pack_entry e;
913 off_t m_offset, p_offset;
914
915 if (i > 0 && pairs[i-1].pack_int_id != pairs[i].pack_int_id &&
905 - m->packs[pairs[i-1].pack_int_id])
906 - {
907 - close_pack_fd(m->packs[pairs[i-1].pack_int_id]);
908 - close_pack_index(m->packs[pairs[i-1].pack_int_id]);
916 + nth_midxed_pack(m, pairs[i-1].pack_int_id)) {
917 + uint32_t pack_int_id = pairs[i-1].pack_int_id;
918 + struct packed_git *p = nth_midxed_pack(m, pack_int_id);
919 +
920 + close_pack_fd(p);
921 + close_pack_index(p);
922 }
923
924 nth_midxed_object_oid(&oid, m, pairs[i].pos);
midx.h
+2
@@ -95,6 +95,8 @@ void get_split_midx_filename_ext(struct strbuf *buf, const char *object_dir,
95
96 struct multi_pack_index *load_multi_pack_index(const char *object_dir, int local);
97 int prepare_midx_pack(struct repository *r, struct multi_pack_index *m, uint32_t pack_int_id);
98 +struct packed_git *nth_midxed_pack(struct multi_pack_index *m,
99 + uint32_t pack_int_id);
100 int nth_bitmapped_pack(struct repository *r, struct multi_pack_index *m,
101 struct bitmapped_pack *bp, uint32_t pack_int_id);
102 int bsearch_one_midx(const struct object_id *oid, struct multi_pack_index *m,