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>
30 #include "got_error.h"
31 #include "got_object.h"
32 #include "got_cancel.h"
33 #include "got_commit_graph.h"
36 #include "got_lib_delta.h"
37 #include "got_lib_inflate.h"
38 #include "got_lib_object.h"
39 #include "got_lib_object_idset.h"
41 struct got_commit_graph_node {
42 struct got_object_id id;
45 /* Used during graph iteration. */
46 TAILQ_ENTRY(got_commit_graph_node) entry;
49 TAILQ_HEAD(got_commit_graph_iter_list, got_commit_graph_node);
51 struct got_commit_graph_branch_tip {
52 struct got_object_id *commit_id;
53 struct got_commit_object *commit;
54 struct got_commit_graph_node *new_node;
57 struct got_commit_graph {
58 /* The set of all commits we have traversed. */
59 struct got_object_idset *node_ids;
62 #define GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL 0x01
65 * A set of object IDs of known parent commits which we have not yet
66 * traversed. Each commit ID in this set represents a branch in commit
67 * history: Either the first-parent branch of the head node, or another
68 * branch corresponding to a traversed merge commit for which we have
69 * not traversed a branch point commit yet.
71 * Whenever we add a commit with a matching ID to the graph, we remove
72 * its corresponding element from this set, and add new elements for
73 * each of that commit's parent commits which were not traversed yet.
75 * When API users ask us to fetch more commits, we fetch commits from
76 * all currently open branches. This allows API users to process
77 * commits in linear order even though the history contains branches.
79 struct got_object_idset *open_branches;
81 /* Array of branch tips for fetch_commits_from_open_branches(). */
82 struct got_commit_graph_branch_tip *tips;
85 /* Path of tree entry of interest to the API user. */
89 * Nodes which will be passed to the API user next, sorted by
92 struct got_commit_graph_iter_list iter_list;
95 static const struct got_error *
96 detect_changed_path(int *changed, struct got_commit_object *commit,
97 struct got_object_id *commit_id, const char *path,
98 struct got_repository *repo)
100 const struct got_error *err = NULL;
101 struct got_commit_object *pcommit = NULL;
102 struct got_tree_object *tree = NULL, *ptree = NULL;
103 struct got_object_qid *pid;
105 if (got_path_is_root_dir(path)) {
112 pid = SIMPLEQ_FIRST(&commit->parent_ids);
114 struct got_object_id *obj_id;
115 err = got_object_id_by_path(&obj_id, repo, commit_id, path);
117 if (err->code == GOT_ERR_NO_TREE_ENTRY)
120 *changed = 1; /* The path was created in this commit. */
125 err = got_object_open_as_tree(&tree, repo, commit->tree_id);
129 err = got_object_open_as_commit(&pcommit, repo, pid->id);
133 err = got_object_open_as_tree(&ptree, repo, pcommit->tree_id);
137 err = got_object_tree_path_changed(changed, tree, ptree, path, repo);
140 got_object_tree_close(tree);
142 got_object_tree_close(ptree);
144 got_object_commit_close(pcommit);
149 add_node_to_iter_list(struct got_commit_graph *graph,
150 struct got_commit_graph_node *node)
152 struct got_commit_graph_node *n, *next;
154 if (TAILQ_EMPTY(&graph->iter_list)) {
155 TAILQ_INSERT_HEAD(&graph->iter_list, node, entry);
159 n = TAILQ_FIRST(&graph->iter_list);
160 /* Ensure that an iteration in progress will see this new commit. */
162 next = TAILQ_NEXT(n, entry);
163 if (next && node->timestamp >= next->timestamp) {
164 TAILQ_INSERT_BEFORE(next, node, entry);
169 TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
172 static const struct got_error *
173 close_branch(struct got_commit_graph *graph, struct got_object_id *commit_id)
175 const struct got_error *err;
177 err = got_object_idset_remove(NULL, graph->open_branches, commit_id);
178 if (err && err->code != GOT_ERR_NO_OBJ)
183 static const struct got_error *
184 advance_branch(struct got_commit_graph *graph, struct got_object_id *commit_id,
185 struct got_commit_object *commit, struct got_repository *repo)
187 const struct got_error *err;
188 struct got_object_qid *qid;
190 err = close_branch(graph, commit_id);
194 if (graph->flags & GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL) {
195 qid = SIMPLEQ_FIRST(&commit->parent_ids);
197 got_object_idset_contains(graph->open_branches, qid->id))
199 return got_object_idset_add(graph->open_branches,
204 * If we are graphing commits for a specific path, skip branches
205 * which do not contribute any content to this path.
207 if (commit->nparents > 1 && !got_path_is_root_dir(graph->path)) {
208 struct got_object_id *merged_id, *prev_id = NULL;
209 int branches_differ = 0;
211 err = got_object_id_by_path(&merged_id, repo, commit_id,
216 SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
217 struct got_object_id *id;
219 if (got_object_idset_contains(graph->open_branches,
223 err = got_object_id_by_path(&id, repo, qid->id,
226 if (err->code == GOT_ERR_NO_TREE_ENTRY) {
236 if (!branches_differ &&
237 got_object_id_cmp(id, prev_id) != 0)
244 * If a branch has created the merged content we can
245 * skip any other branches.
247 if (got_object_id_cmp(merged_id, id) == 0) {
248 err = got_object_idset_add(graph->open_branches,
262 * If the path's content is the same on all branches,
263 * follow the first parent only.
265 if (!branches_differ) {
266 qid = SIMPLEQ_FIRST(&commit->parent_ids);
269 if (got_object_idset_contains(graph->open_branches,
272 if (got_object_idset_contains(graph->node_ids,
274 return NULL; /* parent already traversed */
275 return got_object_idset_add(graph->open_branches,
280 SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
281 if (got_object_idset_contains(graph->open_branches, qid->id))
283 if (got_object_idset_contains(graph->node_ids, qid->id))
284 continue; /* parent already traversed */
285 err = got_object_idset_add(graph->open_branches, qid->id, NULL);
293 static const struct got_error *
294 add_node(struct got_commit_graph_node **new_node,
295 struct got_commit_graph *graph,
296 struct got_object_id *commit_id,
297 struct got_commit_object *commit,
298 struct got_repository *repo)
300 const struct got_error *err = NULL;
301 struct got_commit_graph_node *node;
305 node = calloc(1, sizeof(*node));
307 return got_error_from_errno("calloc");
309 memcpy(&node->id, commit_id, sizeof(node->id));
310 node->timestamp = commit->committer_time;
312 err = got_object_idset_add(graph->node_ids, &node->id, NULL);
320 const struct got_error *
321 got_commit_graph_open(struct got_commit_graph **graph,
322 const char *path, int first_parent_traversal)
324 const struct got_error *err = NULL;
326 *graph = calloc(1, sizeof(**graph));
328 return got_error_from_errno("calloc");
330 TAILQ_INIT(&(*graph)->iter_list);
332 (*graph)->path = strdup(path);
333 if ((*graph)->path == NULL) {
334 err = got_error_from_errno("strdup");
338 (*graph)->node_ids = got_object_idset_alloc();
339 if ((*graph)->node_ids == NULL) {
340 err = got_error_from_errno("got_object_idset_alloc");
344 (*graph)->open_branches = got_object_idset_alloc();
345 if ((*graph)->open_branches == NULL) {
346 err = got_error_from_errno("got_object_idset_alloc");
350 if (first_parent_traversal)
351 (*graph)->flags |= GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL;
354 got_commit_graph_close(*graph);
360 struct add_branch_tip_arg {
361 struct got_commit_graph_branch_tip *tips;
363 struct got_repository *repo;
364 struct got_commit_graph *graph;
367 static const struct got_error *
368 add_branch_tip(struct got_object_id *commit_id, void *data, void *arg)
370 const struct got_error *err;
371 struct add_branch_tip_arg *a = arg;
372 struct got_commit_graph_node *new_node;
373 struct got_commit_object *commit;
375 err = got_object_open_as_commit(&commit, a->repo, commit_id);
379 err = add_node(&new_node, a->graph, commit_id, commit, a->repo);
383 a->tips[a->ntips].commit_id = new_node ? &new_node->id : NULL;
384 a->tips[a->ntips].commit = commit;
385 a->tips[a->ntips].new_node = new_node;
391 static const struct got_error *
392 fetch_commits_from_open_branches(struct got_commit_graph *graph,
393 struct got_repository *repo, got_cancel_cb cancel_cb, void *cancel_arg)
395 const struct got_error *err;
396 struct add_branch_tip_arg arg;
399 ntips = got_object_idset_num_elements(graph->open_branches);
403 /* (Re-)allocate branch tips array if necessary. */
404 if (graph->ntips < ntips) {
405 struct got_commit_graph_branch_tip *tips;
406 tips = recallocarray(graph->tips, graph->ntips, ntips,
409 return got_error_from_errno("recallocarray");
411 graph->ntips = ntips;
413 arg.tips = graph->tips;
414 arg.ntips = 0; /* add_branch_tip() will increment */
417 err = got_object_idset_for_each(graph->open_branches, add_branch_tip,
422 for (i = 0; i < arg.ntips; i++) {
423 struct got_object_id *commit_id;
424 struct got_commit_object *commit;
425 struct got_commit_graph_node *new_node;
429 err = (*cancel_cb)(cancel_arg);
434 commit_id = arg.tips[i].commit_id;
435 commit = arg.tips[i].commit;
436 new_node = arg.tips[i].new_node;
438 err = detect_changed_path(&changed, commit, commit_id,
441 if (err->code != GOT_ERR_NO_OBJ)
444 * History of the path stops here on the current
445 * branch. Keep going on other branches.
447 err = close_branch(graph, commit_id);
453 add_node_to_iter_list(graph, new_node);
454 err = advance_branch(graph, commit_id, commit, repo);
459 for (i = 0; i < arg.ntips; i++)
460 got_object_commit_close(arg.tips[i].commit);
465 got_commit_graph_close(struct got_commit_graph *graph)
467 if (graph->open_branches)
468 got_object_idset_free(graph->open_branches);
470 got_object_idset_free(graph->node_ids);
476 const struct got_error *
477 got_commit_graph_iter_start(struct got_commit_graph *graph,
478 struct got_object_id *id, struct got_repository *repo,
479 got_cancel_cb cancel_cb, void *cancel_arg)
481 const struct got_error *err = NULL;
483 if (!TAILQ_EMPTY(&graph->iter_list))
484 return got_error(GOT_ERR_ITER_BUSY);
486 err = got_object_idset_add(graph->open_branches, id, NULL);
490 /* Locate first commit which changed graph->path. */
491 while (TAILQ_EMPTY(&graph->iter_list) &&
492 got_object_idset_num_elements(graph->open_branches) > 0) {
493 err = fetch_commits_from_open_branches(graph, repo,
494 cancel_cb, cancel_arg);
499 if (TAILQ_EMPTY(&graph->iter_list)) {
501 if (got_path_is_root_dir(graph->path))
502 return got_error_no_obj(id);
504 while (path[0] == '/')
506 return got_error_path(path, GOT_ERR_NO_TREE_ENTRY);
512 const struct got_error *
513 got_commit_graph_iter_next(struct got_object_id **id,
514 struct got_commit_graph *graph, struct got_repository *repo,
515 got_cancel_cb cancel_cb, void *cancel_arg)
517 const struct got_error *err = NULL;
518 struct got_commit_graph_node *node;
522 if (TAILQ_EMPTY(&graph->iter_list)) {
523 /* We are done iterating, or iteration was not started. */
524 return got_error(GOT_ERR_ITER_COMPLETED);
527 node = TAILQ_FIRST(&graph->iter_list);
528 while (TAILQ_NEXT(node, entry) == NULL &&
529 got_object_idset_num_elements(graph->open_branches) > 0) {
530 err = fetch_commits_from_open_branches(graph, repo,
531 cancel_cb, cancel_arg);
537 TAILQ_REMOVE(&graph->iter_list, node, entry);
541 const struct got_error *
542 got_commit_graph_find_youngest_common_ancestor(struct got_object_id **yca_id,
543 struct got_object_id *commit_id, struct got_object_id *commit_id2,
544 struct got_repository *repo, got_cancel_cb cancel_cb, void *cancel_arg)
546 const struct got_error *err = NULL;
547 struct got_commit_graph *graph = NULL, *graph2 = NULL;
548 int completed = 0, completed2 = 0;
549 struct got_object_idset *commit_ids;
553 commit_ids = got_object_idset_alloc();
554 if (commit_ids == NULL)
555 return got_error_from_errno("got_object_idset_alloc");
557 err = got_commit_graph_open(&graph, "/", 1);
561 err = got_commit_graph_open(&graph2, "/", 1);
565 err = got_commit_graph_iter_start(graph, commit_id, repo,
566 cancel_cb, cancel_arg);
570 err = got_commit_graph_iter_start(graph2, commit_id2, repo,
571 cancel_cb, cancel_arg);
576 struct got_object_id *id = NULL, *id2 = NULL;
579 err = (*cancel_cb)(cancel_arg);
585 err = got_commit_graph_iter_next(&id, graph, repo,
586 cancel_cb, cancel_arg);
588 if (err->code != GOT_ERR_ITER_COMPLETED)
596 err = got_commit_graph_iter_next(&id2, graph2, repo,
597 cancel_cb, cancel_arg);
599 if (err->code != GOT_ERR_ITER_COMPLETED)
607 if (got_object_idset_contains(commit_ids, id)) {
608 *yca_id = got_object_id_dup(id);
611 err = got_error_from_errno("got_object_id_dup");
615 err = got_object_idset_add(commit_ids, id, NULL);
620 if (got_object_idset_contains(commit_ids, id2)) {
621 *yca_id = got_object_id_dup(id2);
624 err = got_error_from_errno("got_object_id_dup");
628 err = got_object_idset_add(commit_ids, id2, NULL);
633 if (completed && completed2) {
634 err = got_error(GOT_ERR_ANCESTRY);
640 got_object_idset_free(commit_ids);
642 got_commit_graph_close(graph);
644 got_commit_graph_close(graph2);