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 n = TAILQ_FIRST(&graph->iter_list);
156 next = TAILQ_NEXT(n, entry);
157 if (next && node->timestamp >= next->timestamp) {
158 TAILQ_INSERT_BEFORE(next, node, entry);
163 TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
166 static const struct got_error *
167 advance_branch(struct got_commit_graph *graph, struct got_object_id *commit_id,
168 struct got_commit_object *commit, struct got_repository *repo)
170 const struct got_error *err;
171 struct got_object_qid *qid;
173 if (graph->flags & GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL) {
174 qid = SIMPLEQ_FIRST(&commit->parent_ids);
176 got_object_idset_contains(graph->open_branches, qid->id))
178 return got_object_idset_add(graph->open_branches,
183 * If we are graphing commits for a specific path, skip branches
184 * which do not contribute any content to this path.
186 if (commit->nparents > 1 && !got_path_is_root_dir(graph->path)) {
187 struct got_object_id *merged_id, *prev_id = NULL;
188 int branches_differ = 0;
190 err = got_object_id_by_path(&merged_id, repo, commit_id,
195 SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
196 struct got_object_id *id;
198 if (got_object_idset_contains(graph->open_branches,
202 err = got_object_id_by_path(&id, repo, qid->id,
205 if (err->code == GOT_ERR_NO_TREE_ENTRY) {
215 if (!branches_differ &&
216 got_object_id_cmp(id, prev_id) != 0)
223 * If a branch has created the merged content we can
224 * skip any other branches.
226 if (got_object_id_cmp(merged_id, id) == 0) {
227 err = got_object_idset_add(graph->open_branches,
241 * If the path's content is the same on all branches,
242 * follow the first parent only.
244 if (!branches_differ) {
245 qid = SIMPLEQ_FIRST(&commit->parent_ids);
248 if (got_object_idset_contains(graph->open_branches,
251 if (got_object_idset_contains(graph->node_ids,
253 return NULL; /* parent already traversed */
254 return got_object_idset_add(graph->open_branches,
259 SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
260 if (got_object_idset_contains(graph->open_branches, qid->id))
262 if (got_object_idset_contains(graph->node_ids, qid->id))
263 continue; /* parent already traversed */
264 err = got_object_idset_add(graph->open_branches, qid->id, NULL);
272 static const struct got_error *
273 add_node(struct got_commit_graph_node **new_node,
274 struct got_commit_graph *graph,
275 struct got_object_id *commit_id,
276 struct got_commit_object *commit,
277 struct got_repository *repo)
279 const struct got_error *err = NULL;
280 struct got_commit_graph_node *node;
284 node = calloc(1, sizeof(*node));
286 return got_error_from_errno("calloc");
288 memcpy(&node->id, commit_id, sizeof(node->id));
289 node->timestamp = commit->committer_time;
291 err = got_object_idset_add(graph->node_ids, &node->id, NULL);
299 const struct got_error *
300 got_commit_graph_open(struct got_commit_graph **graph,
301 const char *path, int first_parent_traversal)
303 const struct got_error *err = NULL;
305 *graph = calloc(1, sizeof(**graph));
307 return got_error_from_errno("calloc");
309 TAILQ_INIT(&(*graph)->iter_list);
311 (*graph)->path = strdup(path);
312 if ((*graph)->path == NULL) {
313 err = got_error_from_errno("strdup");
317 (*graph)->node_ids = got_object_idset_alloc();
318 if ((*graph)->node_ids == NULL) {
319 err = got_error_from_errno("got_object_idset_alloc");
323 (*graph)->open_branches = got_object_idset_alloc();
324 if ((*graph)->open_branches == NULL) {
325 err = got_error_from_errno("got_object_idset_alloc");
329 if (first_parent_traversal)
330 (*graph)->flags |= GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL;
333 got_commit_graph_close(*graph);
339 struct add_branch_tip_arg {
340 struct got_commit_graph_branch_tip *tips;
342 struct got_repository *repo;
343 struct got_commit_graph *graph;
346 static const struct got_error *
347 add_branch_tip(struct got_object_id *commit_id, void *data, void *arg)
349 const struct got_error *err;
350 struct add_branch_tip_arg *a = arg;
351 struct got_commit_graph_node *new_node;
352 struct got_commit_object *commit;
354 err = got_object_open_as_commit(&commit, a->repo, commit_id);
358 err = add_node(&new_node, a->graph, commit_id, commit, a->repo);
362 a->tips[a->ntips].commit_id = new_node ? &new_node->id : NULL;
363 a->tips[a->ntips].commit = commit;
364 a->tips[a->ntips].new_node = new_node;
370 static const struct got_error *
371 fetch_commits_from_open_branches(struct got_commit_graph *graph,
372 struct got_repository *repo, got_cancel_cb cancel_cb, void *cancel_arg)
374 const struct got_error *err;
375 struct add_branch_tip_arg arg;
378 ntips = got_object_idset_num_elements(graph->open_branches);
382 /* (Re-)allocate branch tips array if necessary. */
383 if (graph->ntips < ntips) {
384 struct got_commit_graph_branch_tip *tips;
385 tips = recallocarray(graph->tips, graph->ntips, ntips,
388 return got_error_from_errno("recallocarray");
390 graph->ntips = ntips;
392 arg.tips = graph->tips;
393 arg.ntips = 0; /* add_branch_tip() will increment */
396 err = got_object_idset_for_each(graph->open_branches, add_branch_tip,
401 for (i = 0; i < arg.ntips; i++) {
402 struct got_object_id *commit_id;
403 struct got_commit_object *commit;
404 struct got_commit_graph_node *new_node;
408 err = (*cancel_cb)(cancel_arg);
413 commit_id = arg.tips[i].commit_id;
414 commit = arg.tips[i].commit;
415 new_node = arg.tips[i].new_node;
417 err = got_object_idset_remove(NULL, graph->open_branches,
419 if (err && err->code != GOT_ERR_NO_OBJ)
422 err = detect_changed_path(&changed, commit, commit_id,
425 if (err->code != GOT_ERR_NO_OBJ)
428 * History of the path stops here on the current
429 * branch. Keep going on other branches.
434 add_node_to_iter_list(graph, new_node);
435 err = advance_branch(graph, commit_id, commit, repo);
440 for (i = 0; i < arg.ntips; i++)
441 got_object_commit_close(arg.tips[i].commit);
446 got_commit_graph_close(struct got_commit_graph *graph)
448 if (graph->open_branches)
449 got_object_idset_free(graph->open_branches);
451 got_object_idset_free(graph->node_ids);
457 const struct got_error *
458 got_commit_graph_iter_start(struct got_commit_graph *graph,
459 struct got_object_id *id, struct got_repository *repo,
460 got_cancel_cb cancel_cb, void *cancel_arg)
462 const struct got_error *err = NULL;
464 if (!TAILQ_EMPTY(&graph->iter_list))
465 return got_error(GOT_ERR_ITER_BUSY);
467 err = got_object_idset_add(graph->open_branches, id, NULL);
471 /* Locate first commit which changed graph->path. */
472 while (TAILQ_EMPTY(&graph->iter_list) &&
473 got_object_idset_num_elements(graph->open_branches) > 0) {
474 err = fetch_commits_from_open_branches(graph, repo,
475 cancel_cb, cancel_arg);
480 if (TAILQ_EMPTY(&graph->iter_list)) {
482 if (got_path_is_root_dir(graph->path))
483 return got_error_no_obj(id);
485 while (path[0] == '/')
487 return got_error_path(path, GOT_ERR_NO_TREE_ENTRY);
493 const struct got_error *
494 got_commit_graph_iter_next(struct got_object_id **id,
495 struct got_commit_graph *graph, struct got_repository *repo,
496 got_cancel_cb cancel_cb, void *cancel_arg)
498 const struct got_error *err = NULL;
499 struct got_commit_graph_node *node;
503 node = TAILQ_FIRST(&graph->iter_list);
505 /* We are done iterating, or iteration was not started. */
506 return got_error(GOT_ERR_ITER_COMPLETED);
509 while (TAILQ_NEXT(node, entry) == NULL &&
510 got_object_idset_num_elements(graph->open_branches) > 0) {
511 err = fetch_commits_from_open_branches(graph, repo,
512 cancel_cb, cancel_arg);
518 TAILQ_REMOVE(&graph->iter_list, node, entry);
522 const struct got_error *
523 got_commit_graph_find_youngest_common_ancestor(struct got_object_id **yca_id,
524 struct got_object_id *commit_id, struct got_object_id *commit_id2,
525 struct got_repository *repo, got_cancel_cb cancel_cb, void *cancel_arg)
527 const struct got_error *err = NULL;
528 struct got_commit_graph *graph = NULL, *graph2 = NULL;
529 int completed = 0, completed2 = 0;
530 struct got_object_idset *commit_ids;
534 commit_ids = got_object_idset_alloc();
535 if (commit_ids == NULL)
536 return got_error_from_errno("got_object_idset_alloc");
538 err = got_commit_graph_open(&graph, "/", 1);
542 err = got_commit_graph_open(&graph2, "/", 1);
546 err = got_commit_graph_iter_start(graph, commit_id, repo,
547 cancel_cb, cancel_arg);
551 err = got_commit_graph_iter_start(graph2, commit_id2, repo,
552 cancel_cb, cancel_arg);
557 struct got_object_id *id = NULL, *id2 = NULL;
560 err = (*cancel_cb)(cancel_arg);
566 err = got_commit_graph_iter_next(&id, graph, repo,
567 cancel_cb, cancel_arg);
569 if (err->code != GOT_ERR_ITER_COMPLETED)
577 err = got_commit_graph_iter_next(&id2, graph2, repo,
578 cancel_cb, cancel_arg);
580 if (err->code != GOT_ERR_ITER_COMPLETED)
588 if (got_object_idset_contains(commit_ids, id)) {
589 *yca_id = got_object_id_dup(id);
592 err = got_error_from_errno("got_object_id_dup");
596 err = got_object_idset_add(commit_ids, id, NULL);
601 if (got_object_idset_contains(commit_ids, id2)) {
602 *yca_id = got_object_id_dup(id2);
605 err = got_error_from_errno("got_object_id_dup");
609 err = got_object_idset_add(commit_ids, id2, NULL);
614 if (completed && completed2) {
615 err = got_error(GOT_ERR_ANCESTRY);
621 got_object_idset_free(commit_ids);
623 got_commit_graph_close(graph);
625 got_commit_graph_close(graph2);