Derrick Stolee | 08fd81c | 2018-04-02 16:34:19 -0400 | [diff] [blame] | 1 | #ifndef COMMIT_GRAPH_H |
| 2 | #define COMMIT_GRAPH_H |
| 3 | |
Elijah Newren | a034e91 | 2023-05-16 06:34:06 +0000 | [diff] [blame] | 4 | #include "object-store-ll.h" |
Taylor Blau | 6830c36 | 2020-04-13 22:04:25 -0600 | [diff] [blame] | 5 | #include "oidset.h" |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 6 | |
Derrick Stolee | 859fdc0 | 2018-08-29 05:49:04 -0700 | [diff] [blame] | 7 | #define GIT_TEST_COMMIT_GRAPH "GIT_TEST_COMMIT_GRAPH" |
Derrick Stolee | 7b671f8 | 2020-06-23 17:47:01 +0000 | [diff] [blame] | 8 | #define GIT_TEST_COMMIT_GRAPH_DIE_ON_PARSE "GIT_TEST_COMMIT_GRAPH_DIE_ON_PARSE" |
Garima Singh | d5b873c | 2020-04-06 16:59:55 +0000 | [diff] [blame] | 9 | #define GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS "GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS" |
Derrick Stolee | 859fdc0 | 2018-08-29 05:49:04 -0700 | [diff] [blame] | 10 | |
Derrick Stolee | b23ea97 | 2020-04-16 20:14:03 +0000 | [diff] [blame] | 11 | /* |
Patrick Steinhardt | e04838e | 2023-10-31 08:16:13 +0100 | [diff] [blame] | 12 | * This environment variable controls whether commits looked up via the |
| 13 | * commit graph will be double checked to exist in the object database. |
| 14 | */ |
| 15 | #define GIT_COMMIT_GRAPH_PARANOIA "GIT_COMMIT_GRAPH_PARANOIA" |
| 16 | |
| 17 | /* |
Derrick Stolee | b23ea97 | 2020-04-16 20:14:03 +0000 | [diff] [blame] | 18 | * This method is only used to enhance coverage of the commit-graph |
| 19 | * feature in the test suite with the GIT_TEST_COMMIT_GRAPH and |
| 20 | * GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS environment variables. Do not |
| 21 | * call this method oustide of a builtin, and only if you know what |
| 22 | * you are doing! |
| 23 | */ |
| 24 | void git_test_write_commit_graph_or_die(void); |
| 25 | |
Jonathan Tan | e5c5ca2 | 2018-07-11 15:42:39 -0700 | [diff] [blame] | 26 | struct commit; |
Garima Singh | 76ffbca | 2020-04-06 16:59:49 +0000 | [diff] [blame] | 27 | struct bloom_filter_settings; |
SZEDER Gábor | fa79653 | 2020-06-05 13:00:28 +0000 | [diff] [blame] | 28 | struct repository; |
| 29 | struct raw_object_store; |
| 30 | struct string_list; |
Jonathan Tan | e5c5ca2 | 2018-07-11 15:42:39 -0700 | [diff] [blame] | 31 | |
Taylor Blau | ad2dd5b | 2020-02-03 13:18:02 -0800 | [diff] [blame] | 32 | char *get_commit_graph_filename(struct object_directory *odb); |
Derrick Stolee | 663b2b1 | 2020-09-17 18:11:46 +0000 | [diff] [blame] | 33 | char *get_commit_graph_chain_filename(struct object_directory *odb); |
Ævar Arnfjörð Bjarmason | 61df89c | 2019-03-25 13:08:30 +0100 | [diff] [blame] | 34 | int open_commit_graph(const char *graph_file, int *fd, struct stat *st); |
Jeff King | 7ed76b4 | 2023-09-28 00:38:07 -0400 | [diff] [blame] | 35 | int open_commit_graph_chain(const char *chain_file, int *fd, struct stat *st); |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 36 | |
Derrick Stolee | 177722b | 2018-04-10 08:56:05 -0400 | [diff] [blame] | 37 | /* |
| 38 | * Given a commit struct, try to fill the commit struct info, including: |
| 39 | * 1. tree object |
| 40 | * 2. date |
| 41 | * 3. parents. |
| 42 | * |
| 43 | * Returns 1 if and only if the commit was found in the packed graph. |
| 44 | * |
| 45 | * See parse_commit_buffer() for the fallback after this call. |
| 46 | */ |
Jonathan Tan | dade47c | 2018-07-11 15:42:42 -0700 | [diff] [blame] | 47 | int parse_commit_in_graph(struct repository *r, struct commit *item); |
Derrick Stolee | 177722b | 2018-04-10 08:56:05 -0400 | [diff] [blame] | 48 | |
Derrick Stolee | e2838d8 | 2018-05-01 12:47:13 +0000 | [diff] [blame] | 49 | /* |
Taylor Blau | 7805360 | 2022-07-12 19:10:31 -0400 | [diff] [blame] | 50 | * Fills `*pos` with the graph position of `c`, and returns 1 if `c` is |
| 51 | * found in the commit-graph belonging to `r`, or 0 otherwise. |
| 52 | * Initializes the commit-graph belonging to `r` if it hasn't been |
| 53 | * already. |
| 54 | * |
| 55 | * Note: this is a low-level helper that does not alter any slab data |
| 56 | * associated with `c`. Useful in circumstances where the slab data is |
| 57 | * already being modified (e.g., writing the commit-graph itself). |
| 58 | * |
| 59 | * In most cases, callers should use `parse_commit_in_graph()` instead. |
| 60 | */ |
| 61 | int repo_find_commit_pos_in_graph(struct repository *r, struct commit *c, |
| 62 | uint32_t *pos); |
| 63 | |
| 64 | /* |
Patrick Steinhardt | f559d6d | 2021-08-09 10:12:03 +0200 | [diff] [blame] | 65 | * Look up the given commit ID in the commit-graph. This will only return a |
| 66 | * commit if the ID exists both in the graph and in the object database such |
| 67 | * that we don't return commits whose object has been pruned. Otherwise, this |
| 68 | * function returns `NULL`. |
| 69 | */ |
| 70 | struct commit *lookup_commit_in_graph(struct repository *repo, const struct object_id *id); |
| 71 | |
| 72 | /* |
Derrick Stolee | e2838d8 | 2018-05-01 12:47:13 +0000 | [diff] [blame] | 73 | * It is possible that we loaded commit contents from the commit buffer, |
| 74 | * but we also want to ensure the commit-graph content is correctly |
| 75 | * checked and filled. Fill the graph_pos and generation members of |
| 76 | * the given commit. |
| 77 | */ |
Jonathan Tan | dade47c | 2018-07-11 15:42:42 -0700 | [diff] [blame] | 78 | void load_commit_graph_info(struct repository *r, struct commit *item); |
Derrick Stolee | e2838d8 | 2018-05-01 12:47:13 +0000 | [diff] [blame] | 79 | |
Jonathan Tan | dade47c | 2018-07-11 15:42:42 -0700 | [diff] [blame] | 80 | struct tree *get_commit_tree_in_graph(struct repository *r, |
| 81 | const struct commit *c); |
Derrick Stolee | 7b8a21d | 2018-04-06 19:09:46 +0000 | [diff] [blame] | 82 | |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 83 | struct commit_graph { |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 84 | const unsigned char *data; |
| 85 | size_t data_len; |
| 86 | |
| 87 | unsigned char hash_len; |
| 88 | unsigned char num_chunks; |
| 89 | uint32_t num_commits; |
| 90 | struct object_id oid; |
Derrick Stolee | 6c622f9 | 2019-06-18 11:14:27 -0700 | [diff] [blame] | 91 | char *filename; |
Taylor Blau | 13c2499 | 2020-02-03 13:18:00 -0800 | [diff] [blame] | 92 | struct object_directory *odb; |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 93 | |
Derrick Stolee | d4f4d60 | 2019-06-18 11:14:24 -0700 | [diff] [blame] | 94 | uint32_t num_commits_in_base; |
Abhishek Kumar | 1fdc383 | 2021-01-16 18:11:16 +0000 | [diff] [blame] | 95 | unsigned int read_generation_data; |
Derrick Stolee | d4f4d60 | 2019-06-18 11:14:24 -0700 | [diff] [blame] | 96 | struct commit_graph *base_graph; |
| 97 | |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 98 | const uint32_t *chunk_oid_fanout; |
| 99 | const unsigned char *chunk_oid_lookup; |
| 100 | const unsigned char *chunk_commit_data; |
Abhishek Kumar | e8b6300 | 2021-01-16 18:11:15 +0000 | [diff] [blame] | 101 | const unsigned char *chunk_generation_data; |
| 102 | const unsigned char *chunk_generation_data_overflow; |
Jeff King | ee6a792 | 2023-10-09 17:05:47 -0400 | [diff] [blame] | 103 | size_t chunk_generation_data_overflow_size; |
SZEDER Gábor | 5af7417 | 2019-01-19 21:21:13 +0100 | [diff] [blame] | 104 | const unsigned char *chunk_extra_edges; |
Jeff King | 9622610 | 2023-10-09 17:05:38 -0400 | [diff] [blame] | 105 | size_t chunk_extra_edges_size; |
Derrick Stolee | 118bd57 | 2019-06-18 11:14:26 -0700 | [diff] [blame] | 106 | const unsigned char *chunk_base_graphs; |
Jeff King | 6cf61d0 | 2023-10-09 17:05:41 -0400 | [diff] [blame] | 107 | size_t chunk_base_graphs_size; |
Garima Singh | 76ffbca | 2020-04-06 16:59:49 +0000 | [diff] [blame] | 108 | const unsigned char *chunk_bloom_indexes; |
| 109 | const unsigned char *chunk_bloom_data; |
Jeff King | 920f400 | 2023-10-09 17:05:50 -0400 | [diff] [blame] | 110 | size_t chunk_bloom_data_size; |
Garima Singh | 76ffbca | 2020-04-06 16:59:49 +0000 | [diff] [blame] | 111 | |
Abhishek Kumar | 72a2bfc | 2021-01-16 18:11:12 +0000 | [diff] [blame] | 112 | struct topo_level_slab *topo_levels; |
Garima Singh | 76ffbca | 2020-04-06 16:59:49 +0000 | [diff] [blame] | 113 | struct bloom_filter_settings *bloom_filter_settings; |
Derrick Stolee | 2a2e32b | 2018-04-10 08:56:02 -0400 | [diff] [blame] | 114 | }; |
| 115 | |
Taylor Blau | ab14d06 | 2020-09-09 11:22:56 -0400 | [diff] [blame] | 116 | struct commit_graph *load_commit_graph_one_fd_st(struct repository *r, |
| 117 | int fd, struct stat *st, |
Taylor Blau | a7df60c | 2020-02-03 13:18:04 -0800 | [diff] [blame] | 118 | struct object_directory *odb); |
Jeff King | 7ed76b4 | 2023-09-28 00:38:07 -0400 | [diff] [blame] | 119 | struct commit_graph *load_commit_graph_chain_fd_st(struct repository *r, |
Jeff King | 5f25919 | 2023-09-28 00:39:51 -0400 | [diff] [blame] | 120 | int fd, struct stat *st, |
| 121 | int *incomplete_chain); |
Taylor Blau | 13c2499 | 2020-02-03 13:18:00 -0800 | [diff] [blame] | 122 | struct commit_graph *read_commit_graph_one(struct repository *r, |
| 123 | struct object_directory *odb); |
Taylor Blau | a92d852 | 2022-07-14 14:43:06 -0700 | [diff] [blame] | 124 | |
| 125 | /* |
| 126 | * Callers should initialize the repo_settings with prepare_repo_settings() |
| 127 | * prior to calling parse_commit_graph(). |
| 128 | */ |
| 129 | struct commit_graph *parse_commit_graph(struct repo_settings *s, |
Taylor Blau | ab14d06 | 2020-09-09 11:22:56 -0400 | [diff] [blame] | 130 | void *graph_map, size_t graph_size); |
Josh Steadmon | aa65857 | 2019-01-15 14:25:50 -0800 | [diff] [blame] | 131 | |
Derrick Stolee | 6cc0174 | 2018-07-20 16:33:30 +0000 | [diff] [blame] | 132 | /* |
| 133 | * Return 1 if and only if the repository has a commit-graph |
| 134 | * file and generation numbers are computed in that file. |
| 135 | */ |
| 136 | int generation_numbers_enabled(struct repository *r); |
| 137 | |
Abhishek Kumar | 8d00d7c | 2021-01-16 18:11:17 +0000 | [diff] [blame] | 138 | /* |
| 139 | * Return 1 if and only if the repository has a commit-graph |
| 140 | * file and generation data chunk has been written for the file. |
| 141 | */ |
| 142 | int corrected_commit_dates_enabled(struct repository *r); |
| 143 | |
Taylor Blau | 4f36440 | 2020-09-09 11:22:44 -0400 | [diff] [blame] | 144 | struct bloom_filter_settings *get_bloom_filter_settings(struct repository *r); |
| 145 | |
SZEDER Gábor | 39d8831 | 2019-08-05 10:02:39 +0200 | [diff] [blame] | 146 | enum commit_graph_write_flags { |
| 147 | COMMIT_GRAPH_WRITE_APPEND = (1 << 0), |
| 148 | COMMIT_GRAPH_WRITE_PROGRESS = (1 << 1), |
SZEDER Gábor | 7c5c9b9 | 2019-08-05 10:02:40 +0200 | [diff] [blame] | 149 | COMMIT_GRAPH_WRITE_SPLIT = (1 << 2), |
Taylor Blau | 2f00c35 | 2020-05-13 15:59:55 -0600 | [diff] [blame] | 150 | COMMIT_GRAPH_WRITE_BLOOM_FILTERS = (1 << 3), |
Junio C Hamano | 70cdbbe | 2020-07-30 13:20:31 -0700 | [diff] [blame] | 151 | COMMIT_GRAPH_NO_WRITE_BLOOM_FILTERS = (1 << 4), |
SZEDER Gábor | 39d8831 | 2019-08-05 10:02:39 +0200 | [diff] [blame] | 152 | }; |
Derrick Stolee | 5af8039 | 2019-06-12 06:29:38 -0700 | [diff] [blame] | 153 | |
Taylor Blau | 4f02735 | 2020-04-13 22:04:08 -0600 | [diff] [blame] | 154 | enum commit_graph_split_flags { |
Taylor Blau | fdbde82 | 2020-04-13 22:04:12 -0600 | [diff] [blame] | 155 | COMMIT_GRAPH_SPLIT_UNSPECIFIED = 0, |
Taylor Blau | 8a6ac28 | 2020-04-13 22:04:17 -0600 | [diff] [blame] | 156 | COMMIT_GRAPH_SPLIT_MERGE_PROHIBITED = 1, |
| 157 | COMMIT_GRAPH_SPLIT_REPLACE = 2 |
Taylor Blau | 4f02735 | 2020-04-13 22:04:08 -0600 | [diff] [blame] | 158 | }; |
| 159 | |
Taylor Blau | 98bb796 | 2020-09-17 22:59:49 -0400 | [diff] [blame] | 160 | struct commit_graph_opts { |
Derrick Stolee | c2bc6e6 | 2019-06-18 11:14:32 -0700 | [diff] [blame] | 161 | int size_multiple; |
| 162 | int max_commits; |
| 163 | timestamp_t expire_time; |
Taylor Blau | 98bb796 | 2020-09-17 22:59:49 -0400 | [diff] [blame] | 164 | enum commit_graph_split_flags split_flags; |
Taylor Blau | 809e032 | 2020-09-18 09:27:27 -0400 | [diff] [blame] | 165 | int max_new_filters; |
Derrick Stolee | c2bc6e6 | 2019-06-18 11:14:32 -0700 | [diff] [blame] | 166 | }; |
| 167 | |
Derrick Stolee | e103f72 | 2019-06-12 06:29:37 -0700 | [diff] [blame] | 168 | /* |
| 169 | * The write_commit_graph* methods return zero on success |
| 170 | * and a negative value on failure. Note that if the repository |
| 171 | * is not compatible with the commit-graph feature, then the |
| 172 | * methods will return 0 without writing a commit-graph. |
| 173 | */ |
Taylor Blau | 0bd52e2 | 2020-02-03 21:51:50 -0800 | [diff] [blame] | 174 | int write_commit_graph_reachable(struct object_directory *odb, |
SZEDER Gábor | 39d8831 | 2019-08-05 10:02:39 +0200 | [diff] [blame] | 175 | enum commit_graph_write_flags flags, |
Taylor Blau | 98bb796 | 2020-09-17 22:59:49 -0400 | [diff] [blame] | 176 | const struct commit_graph_opts *opts); |
Taylor Blau | 0bd52e2 | 2020-02-03 21:51:50 -0800 | [diff] [blame] | 177 | int write_commit_graph(struct object_directory *odb, |
Ævar Arnfjörð Bjarmason | 4a04790 | 2022-03-04 19:32:12 +0100 | [diff] [blame] | 178 | const struct string_list *pack_indexes, |
Taylor Blau | 6830c36 | 2020-04-13 22:04:25 -0600 | [diff] [blame] | 179 | struct oidset *commits, |
SZEDER Gábor | 39d8831 | 2019-08-05 10:02:39 +0200 | [diff] [blame] | 180 | enum commit_graph_write_flags flags, |
Taylor Blau | 98bb796 | 2020-09-17 22:59:49 -0400 | [diff] [blame] | 181 | const struct commit_graph_opts *opts); |
Derrick Stolee | 08fd81c | 2018-04-02 16:34:19 -0400 | [diff] [blame] | 182 | |
Derrick Stolee | 3da4b60 | 2019-06-18 11:14:32 -0700 | [diff] [blame] | 183 | #define COMMIT_GRAPH_VERIFY_SHALLOW (1 << 0) |
| 184 | |
| 185 | int verify_commit_graph(struct repository *r, struct commit_graph *g, int flags); |
Derrick Stolee | 283e68c | 2018-06-27 09:24:32 -0400 | [diff] [blame] | 186 | |
Derrick Stolee | c3a3a96 | 2019-05-17 11:41:47 -0700 | [diff] [blame] | 187 | void close_commit_graph(struct raw_object_store *); |
Jonathan Tan | c3756d5 | 2018-07-11 15:42:40 -0700 | [diff] [blame] | 188 | void free_commit_graph(struct commit_graph *); |
| 189 | |
Jeff King | 6abada1 | 2019-09-12 10:44:45 -0400 | [diff] [blame] | 190 | /* |
| 191 | * Disable further use of the commit graph in this process when parsing a |
| 192 | * "struct commit". |
| 193 | */ |
| 194 | void disable_commit_graph(struct repository *r); |
| 195 | |
Abhishek Kumar | 4844812 | 2020-06-17 14:44:09 +0530 | [diff] [blame] | 196 | struct commit_graph_data { |
| 197 | uint32_t graph_pos; |
Abhishek Kumar | d7f9278 | 2021-01-16 18:11:13 +0000 | [diff] [blame] | 198 | timestamp_t generation; |
Abhishek Kumar | 4844812 | 2020-06-17 14:44:09 +0530 | [diff] [blame] | 199 | }; |
| 200 | |
| 201 | /* |
| 202 | * Commits should be parsed before accessing generation, graph positions. |
| 203 | */ |
Abhishek Kumar | d7f9278 | 2021-01-16 18:11:13 +0000 | [diff] [blame] | 204 | timestamp_t commit_graph_generation(const struct commit *); |
Abhishek Kumar | 4844812 | 2020-06-17 14:44:09 +0530 | [diff] [blame] | 205 | uint32_t commit_graph_position(const struct commit *); |
Taylor Blau | c08645b | 2023-03-20 11:26:52 +0000 | [diff] [blame] | 206 | |
| 207 | /* |
| 208 | * After this method, all commits reachable from those in the given |
| 209 | * list will have non-zero, non-infinite generation numbers. |
| 210 | */ |
| 211 | void ensure_generations_valid(struct repository *r, |
| 212 | struct commit **commits, size_t nr); |
| 213 | |
Derrick Stolee | 08fd81c | 2018-04-02 16:34:19 -0400 | [diff] [blame] | 214 | #endif |