2 * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
4 * Permission to use, copy, modify, and distribute this software for any
5 * purpose with or without fee is hereby granted, provided that the above
6 * copyright notice and this permission notice appear in all copies.
8 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 #include <sys/types.h>
19 #include <sys/queue.h>
20 #include <sys/stdint.h>
29 #include "got_error.h"
30 #include "got_object.h"
31 #include "got_commit_graph.h"
33 #include "got_lib_delta.h"
34 #include "got_lib_zbuf.h"
35 #include "got_lib_object.h"
36 #include "got_lib_object_idset.h"
38 struct got_commit_graph_node {
39 struct got_object_id id;
42 * Each graph node corresponds to a commit object.
43 * Graph vertices are modelled with two separate adjacency lists:
44 * Adjacencies of a graph node are either parent (older) commits,
45 * and child (younger) commits.
47 struct got_commit_object *commit; /* contains list of parents */
49 SIMPLEQ_HEAD(, got_parent_id) child_ids;
51 /* Used during graph iteration. */
52 TAILQ_ENTRY(got_commit_graph_node) entry;
55 struct got_commit_graph {
56 /* The set of all commits we have traversed. */
57 struct got_object_idset *node_ids;
59 /* The commit at which traversal began (youngest commit in node_ids). */
60 struct got_commit_graph_node *head_node;
63 * A set of object IDs of known parent commits which we have not yet
64 * traversed. Each commit ID in this set represents a branch in commit
65 * history: Either the first-parent branch of the head node, or another
66 * branch corresponding to a traversed merge commit for which we have
67 * not traversed a branch point commit yet.
69 * Whenever we add a commit with a matching ID to the graph, we remove
70 * its corresponding element from this set, and add new elements for
71 * each of that commit's parent commits which were not traversed yet.
73 * When API users ask us to fetch more commits, we fetch commits from
74 * all currently open branches. This allows API users to process
75 * commits in linear order even though the history contains branches.
77 struct got_object_idset *open_branches;
79 /* The next commit to return when the API user asks for one. */
80 struct got_commit_graph_node *iter_node;
82 TAILQ_HEAD(, got_commit_graph_node) iter_candidates;
85 static struct got_commit_graph *
88 struct got_commit_graph *graph;
90 graph = calloc(1, sizeof(*graph));
94 graph->node_ids = got_object_idset_alloc();
95 if (graph->node_ids == NULL) {
100 graph->open_branches = got_object_idset_alloc();
101 if (graph->open_branches == NULL) {
102 got_object_idset_free(graph->node_ids);
107 TAILQ_INIT(&graph->iter_candidates);
113 is_head_node(struct got_commit_graph_node *node)
115 return node->nchildren == 0;
119 is_merge_point(struct got_commit_graph_node *node)
121 return node->commit->nparents > 1;
125 is_branch_point(struct got_commit_graph_node *node)
127 return node->nchildren > 1;
132 is_root_node(struct got_commit_graph_node *node)
134 return node->commit->nparents == 0;
137 static const struct got_error *
138 compare_commits(int *cmp, struct got_commit_object *c1,
139 struct got_commit_object *c2)
141 const struct got_error *err;
144 err = got_object_commit_get_committer_time(&t1, c1);
147 err = got_object_commit_get_committer_time(&t2, c2);
162 static const struct got_error *
163 add_iteration_candidate(struct got_commit_graph *graph,
164 struct got_commit_graph_node *node)
166 struct got_commit_graph_node *n, *next;
168 if (TAILQ_EMPTY(&graph->iter_candidates)) {
169 TAILQ_INSERT_TAIL(&graph->iter_candidates, node, entry);
173 TAILQ_FOREACH(n, &graph->iter_candidates, entry) {
174 const struct got_error *err;
176 err = compare_commits(&cmp, node->commit, n->commit);
180 next = TAILQ_NEXT(n, entry);
182 TAILQ_INSERT_AFTER(&graph->iter_candidates, n,
186 err = compare_commits(&cmp, node->commit, next->commit);
190 TAILQ_INSERT_BEFORE(next, node, entry);
194 TAILQ_INSERT_BEFORE(n, node, entry);
202 static const struct got_error *
203 add_node(struct got_commit_graph_node **new_node,
204 struct got_commit_graph *graph, struct got_object_id *commit_id,
205 struct got_commit_object *commit, struct got_object_id *child_commit_id)
207 const struct got_error *err = NULL;
208 struct got_commit_graph_node *node, *existing_node;
212 node = calloc(1, sizeof(*node));
214 return got_error_from_errno();
216 memcpy(&node->id, commit_id, sizeof(node->id));
217 node->commit = commit;
218 SIMPLEQ_INIT(&node->child_ids);
220 err = got_object_idset_add((void **)(&existing_node),
221 graph->node_ids, &node->id, node);
223 struct got_parent_id *pid;
225 err = add_iteration_candidate(graph, node);
229 err = got_object_idset_remove(graph->open_branches, commit_id);
230 if (err && err->code != GOT_ERR_NO_OBJ)
232 SIMPLEQ_FOREACH(pid, &commit->parent_ids, entry) {
233 if (got_object_idset_get(graph->node_ids, pid->id))
234 continue; /* parent already traversed */
235 err = got_object_idset_add(NULL, graph->open_branches,
237 if (err && err->code != GOT_ERR_OBJ_EXISTS)
241 } else if (err->code == GOT_ERR_OBJ_EXISTS) {
244 node = existing_node;
250 if (child_commit_id) {
251 struct got_parent_id *child, *cid;
253 /* Prevent linking to self. */
254 if (got_object_id_cmp(commit_id, child_commit_id) == 0)
255 return got_error(GOT_ERR_BAD_OBJ_ID);
257 /* Prevent double-linking to the same child. */
258 SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
259 if (got_object_id_cmp(cid->id, child_commit_id) == 0)
260 return got_error(GOT_ERR_BAD_OBJ_ID);
263 child = calloc(1, sizeof(*child));
265 return got_error_from_errno();
266 child->id = got_object_id_dup(child_commit_id);
267 if (child->id == NULL) {
268 err = got_error_from_errno();
272 SIMPLEQ_INSERT_TAIL(&node->child_ids, child, entry);
279 const struct got_error *
280 got_commit_graph_open(struct got_commit_graph **graph,
281 struct got_object_id *commit_id, struct got_repository *repo)
283 const struct got_error *err = NULL;
284 struct got_object *obj;
285 struct got_commit_object *commit;
289 err = got_object_open(&obj, repo, commit_id);
292 if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
293 err = got_error(GOT_ERR_OBJ_TYPE);
294 got_object_close(obj);
298 err = got_object_commit_open(&commit, repo, obj);
299 got_object_close(obj);
303 *graph = alloc_graph();
304 if (*graph == NULL) {
305 got_object_commit_close(commit);
306 return got_error_from_errno();
309 err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
311 got_commit_graph_close(*graph);
319 static const struct got_error *
320 open_commit(struct got_commit_object **commit, struct got_object_id *id,
321 struct got_repository *repo)
323 const struct got_error *err;
324 struct got_object *obj;
326 err = got_object_open(&obj, repo, id);
329 if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
330 err = got_error(GOT_ERR_OBJ_TYPE);
334 err = got_object_commit_open(commit, repo, obj);
336 got_object_close(obj);
340 struct got_commit_graph_branch {
341 struct got_object_id parent_id;
342 struct got_commit_graph_node *node;
345 struct gather_branches_arg {
346 struct got_commit_graph_branch *branches;
351 gather_branches(struct got_object_id *id, void *data, void *arg)
353 struct gather_branches_arg *a = arg;
354 memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
355 a->branches[a->nbranches].node = data;
359 const struct got_error *
360 fetch_commits_from_open_branches(int *ncommits,
361 struct got_commit_graph *graph, struct got_repository *repo)
363 const struct got_error *err;
364 struct got_commit_graph_branch *branches;
365 struct gather_branches_arg arg;
370 arg.nbranches = got_object_idset_num_elements(graph->open_branches);
371 if (arg.nbranches == 0)
375 * Adding nodes to the graph might change the graph's open
376 * branches state. Create a local copy of the current state.
378 branches = calloc(arg.nbranches, sizeof(*branches));
379 if (branches == NULL)
380 return got_error_from_errno();
381 arg.nbranches = 0; /* reset; gather_branches() will increment */
382 arg.branches = branches;
383 got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
385 for (i = 0; i < arg.nbranches; i++) {
386 struct got_object_id *commit_id;
387 struct got_commit_graph_node *child_node, *new_node;
388 struct got_commit_object *commit;
390 commit_id = &branches[i].parent_id;
391 child_node = branches[i].node;
393 err = open_commit(&commit, commit_id, repo);
397 err = add_node(&new_node, graph, commit_id, commit,
400 if (err->code != GOT_ERR_OBJ_EXISTS)
407 got_object_commit_close(commit);
414 const struct got_error *
415 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
416 int limit, struct got_repository *repo)
418 const struct got_error *err;
419 int total = 0, ncommits;
423 while (total < limit) {
424 err = fetch_commits_from_open_branches(&ncommits, graph, repo);
437 free_graph_node(struct got_object_id *id, void *data, void *arg)
439 struct got_commit_graph_node *node = data;
440 got_object_commit_close(node->commit);
441 while (!SIMPLEQ_EMPTY(&node->child_ids)) {
442 struct got_parent_id *child = SIMPLEQ_FIRST(&node->child_ids);
443 SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
450 got_commit_graph_close(struct got_commit_graph *graph)
452 got_object_idset_free(graph->open_branches);
453 got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
454 got_object_idset_free(graph->node_ids);
458 const struct got_error *
459 got_commit_graph_iter_start(struct got_commit_graph *graph,
460 struct got_object_id *id)
462 struct got_commit_graph_node *start_node, *node;
463 struct got_parent_id *pid;
465 start_node = got_object_idset_get(graph->node_ids, id);
466 if (start_node == NULL)
467 return got_error(GOT_ERR_NO_OBJ);
469 graph->iter_node = start_node;
471 while (!TAILQ_EMPTY(&graph->iter_candidates)) {
472 node = TAILQ_FIRST(&graph->iter_candidates);
473 TAILQ_REMOVE(&graph->iter_candidates, node, entry);
476 /* Put all known parents of this commit on the candidate list. */
477 SIMPLEQ_FOREACH(pid, &start_node->commit->parent_ids, entry) {
478 node = got_object_idset_get(graph->node_ids, pid->id);
480 const struct got_error *err;
481 err = add_iteration_candidate(graph, node);
490 const struct got_error *
491 got_commit_graph_iter_next(struct got_commit_object **commit,
492 struct got_object_id **id, struct got_commit_graph *graph)
494 struct got_commit_graph_node *node;
496 if (graph->iter_node == NULL) {
497 /* We are done interating, or iteration was not started. */
503 if (TAILQ_EMPTY(&graph->iter_candidates)) {
504 if (is_root_node(graph->iter_node) &&
505 got_object_idset_num_elements(graph->open_branches) == 0) {
506 *commit = graph->iter_node->commit;
507 *id = &graph->iter_node->id;
508 /* We are done interating. */
509 graph->iter_node = NULL;
512 return got_error(GOT_ERR_ITER_NEED_MORE);
515 *commit = graph->iter_node->commit;
516 *id = &graph->iter_node->id;
517 node = TAILQ_FIRST(&graph->iter_candidates);
518 TAILQ_REMOVE(&graph->iter_candidates, node, entry);
519 graph->iter_node = node;
523 const struct got_commit_object *
524 got_commit_graph_get_commit(struct got_commit_graph *graph,
525 struct got_object_id *id)
527 struct got_commit_graph_node *node;
528 node = got_object_idset_get(graph->node_ids, id);
529 return node ? node->commit : NULL;