1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
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.
7  *
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.
15  */
16 
17 #include <sys/types.h>
18 #include <sys/stat.h>
19 #include <sys/queue.h>
20 #include <sys/stdint.h>
21 
22 #include <stdio.h>
23 #include <stdlib.h>
24 #include <string.h>
25 #include <sha1.h>
26 #include <zlib.h>
27 #include <ctype.h>
28 
29 #include "got_error.h"
30 #include "got_object.h"
31 #include "got_commit_graph.h"
32 
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"
37 
38 struct got_commit_graph_node {
39 	struct got_object_id id;
40 
41 	/*
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.
46 	 */
47 	struct got_commit_object *commit; /* contains list of parents */
48 	int nchildren;
49 	SIMPLEQ_HEAD(, got_parent_id) child_ids;
50 
51 	/* Used during graph iteration. */
52 	TAILQ_ENTRY(got_commit_graph_node) entry;
53 };
54 
55 struct got_commit_graph {
56 	/* The set of all commits we have traversed. */
57 	struct got_object_idset *node_ids;
58 
59 	/* The commit at which traversal began (youngest commit in node_ids). */
60 	struct got_commit_graph_node *head_node;
61 
62 	/*
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.
68 	 *
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.
72 	 *
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.
76 	 */
77 	struct got_object_idset *open_branches;
78 
79 	/* The next commit to return when the API user asks for one. */
80 	struct got_commit_graph_node *iter_node;
81 
82 	TAILQ_HEAD(, got_commit_graph_node) iter_candidates;
83 };
84 
85 static struct got_commit_graph *
86 alloc_graph(void)
87 {
88 	struct got_commit_graph *graph;
89 
90 	graph = calloc(1, sizeof(*graph));
91 	if (graph == NULL)
92 		return NULL;
93 
94 	graph->node_ids = got_object_idset_alloc();
95 	if (graph->node_ids == NULL) {
96 		free(graph);
97 		return NULL;
98 	}
99 
100 	graph->open_branches = got_object_idset_alloc();
101 	if (graph->open_branches == NULL) {
102 		got_object_idset_free(graph->node_ids);
103 		free(graph);
104 		return NULL;
105 	}
106 
107 	TAILQ_INIT(&graph->iter_candidates);
108 	return graph;
109 }
110 
111 #if 0
112 static int
113 is_head_node(struct got_commit_graph_node *node)
114 {
115 	return node->nchildren == 0;
116 }
117 
118 static int
119 is_merge_point(struct got_commit_graph_node *node)
120 {
121 	return node->commit->nparents > 1;
122 }
123 
124 int
125 is_branch_point(struct got_commit_graph_node *node)
126 {
127 	return node->nchildren > 1;
128 }
129 #endif
130 
131 static int
132 is_root_node(struct got_commit_graph_node *node)
133 {
134 	return node->commit->nparents == 0;
135 }
136 
137 static const struct got_error *
138 parse_commit_time(int64_t *time, struct got_commit_object *commit)
139 {
140 	const struct got_error *err = NULL;
141 	const char *errstr;
142 	char *committer, *space;
143 
144 	*time = 0;
145 
146 	committer = strdup(commit->committer);
147 	if (committer == NULL)
148 		return got_error_from_errno();
149 
150 	/* Strip off trailing timezone indicator. */
151 	space = strrchr(committer, ' ');
152 	if (space == NULL) {
153 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
154 		goto done;
155 	}
156 	*space = '\0';
157 
158 	/* Timestamp is separated from committer name + email by space. */
159 	space = strrchr(committer, ' ');
160 	if (space == NULL) {
161 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
162 		goto done;
163 	}
164 
165 	*time = strtonum(space + 1, 0, INT64_MAX, &errstr);
166 	if (errstr)
167 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
168 
169 done:
170 	free(committer);
171 	return err;
172 }
173 
174 static const struct got_error *
175 compare_commits(int *cmp, struct got_commit_object *c1,
176     struct got_commit_object *c2)
177 {
178 	const struct got_error *err;
179 	int64_t t1, t2;
180 
181 	err = parse_commit_time(&t1, c1);
182 	if (err)
183 		return err;
184 	err = parse_commit_time(&t2, c2);
185 	if (err)
186 		return err;
187 
188 	if (t1 == t2)	
189 		*cmp = 0;
190 	else if (t1 < t2)
191 		*cmp = -1;
192 	else
193 		*cmp = 1;
194 
195 	return NULL;
196 }
197 
198 static const struct got_error *
199 add_iteration_candidate(struct got_commit_graph *graph,
200     struct got_commit_graph_node *node)
201 {
202 	struct got_commit_graph_node *n, *next;
203 	
204 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
205 		TAILQ_INSERT_TAIL(&graph->iter_candidates, node, entry);
206 		return NULL;
207 	}
208 
209 	TAILQ_FOREACH(n, &graph->iter_candidates, entry) {
210 		const struct got_error *err;
211 		int cmp;
212 		err = compare_commits(&cmp, node->commit, n->commit);
213 		if (err)
214 			return err;
215 		if (cmp < 0) {
216 			next = TAILQ_NEXT(n, entry);
217 			if (next == NULL) {
218 				TAILQ_INSERT_AFTER(&graph->iter_candidates, n,
219 				    node, entry);
220 				break;
221 			}
222 			err = compare_commits(&cmp, node->commit, next->commit);
223 			if (err)
224 				return err;
225 			if (cmp >= 0) {
226 				TAILQ_INSERT_BEFORE(next, node, entry);
227 				break;
228 			}
229 		} else {
230 			TAILQ_INSERT_BEFORE(n, node, entry);
231 			break;
232 		}
233 	}
234 
235 	return NULL;
236 }
237 
238 static const struct got_error *
239 add_node(struct got_commit_graph_node **new_node,
240     struct got_commit_graph *graph, struct got_object_id *commit_id,
241     struct got_commit_object *commit, struct got_object_id *child_commit_id)
242 {
243 	const struct got_error *err = NULL;
244 	struct got_commit_graph_node *node, *existing_node;
245 
246 	*new_node = NULL;
247 
248 	node = calloc(1, sizeof(*node));
249 	if (node == NULL)
250 		return got_error_from_errno();
251 
252 	memcpy(&node->id, commit_id, sizeof(node->id));
253 	node->commit = commit;
254 	SIMPLEQ_INIT(&node->child_ids);
255 
256 	err = got_object_idset_add((void **)(&existing_node),
257 	    graph->node_ids, &node->id, node);
258 	if (err == NULL) {
259 		struct got_parent_id *pid;
260 
261 		err = add_iteration_candidate(graph, node);
262 		if (err)
263 			return err;
264 
265 		err = got_object_idset_remove(graph->open_branches, commit_id);
266 		if (err && err->code != GOT_ERR_NO_OBJ)
267 			return err;
268 		SIMPLEQ_FOREACH(pid, &commit->parent_ids, entry) {
269 			if (got_object_idset_get(graph->node_ids, pid->id))
270 				continue; /* parent already traversed */
271 			err = got_object_idset_add(NULL, graph->open_branches,
272 			    pid->id, node);
273 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
274 				return err;
275 		}
276 		*new_node = node;
277 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
278 		err = NULL;
279 		free(node);
280 		node = existing_node;
281 	} else {
282 		free(node);
283 		return err;
284 	}
285 
286 	if (child_commit_id) {
287 		struct got_parent_id *child, *cid;
288 
289 		/* Prevent linking to self. */
290 		if (got_object_id_cmp(commit_id, child_commit_id) == 0)
291 			return got_error(GOT_ERR_BAD_OBJ_ID);
292 
293 		/* Prevent double-linking to the same child. */
294 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
295 			if (got_object_id_cmp(cid->id, child_commit_id) == 0)
296 				return got_error(GOT_ERR_BAD_OBJ_ID);
297 		}
298 
299 		child = calloc(1, sizeof(*child));
300 		if (child == NULL)
301 			return got_error_from_errno();
302 		child->id = got_object_id_dup(child_commit_id);
303 		if (child->id == NULL) {
304 			err = got_error_from_errno();
305 			free(child);
306 			return err;
307 		}
308 		SIMPLEQ_INSERT_TAIL(&node->child_ids, child, entry);
309 		node->nchildren++;
310 	}
311 
312 	return err;
313 }
314 
315 const struct got_error *
316 got_commit_graph_open(struct got_commit_graph **graph,
317     struct got_object_id *commit_id, struct got_repository *repo)
318 {
319 	const struct got_error *err = NULL;
320 	struct got_object *obj;
321 	struct got_commit_object *commit;
322 
323 	*graph = NULL;
324 
325 	err = got_object_open(&obj, repo, commit_id);
326 	if (err)
327 		return err;
328 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
329 		err = got_error(GOT_ERR_OBJ_TYPE);
330 		got_object_close(obj);
331 		return err;
332 	}
333 
334 	err = got_object_commit_open(&commit, repo, obj);
335 	got_object_close(obj);
336 	if (err)
337 		return err;
338 
339 	*graph = alloc_graph();
340 	if (*graph == NULL) {
341 		got_object_commit_close(commit);
342 		return got_error_from_errno();
343 	}
344 
345 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
346 	if (err) {
347 		got_commit_graph_close(*graph);
348 		*graph = NULL;
349 		return err;
350 	}
351 	
352 	return NULL;
353 }
354 
355 static const struct got_error *
356 open_commit(struct got_commit_object **commit, struct got_object_id *id,
357     struct got_repository *repo)
358 {
359 	const struct got_error *err;
360 	struct got_object *obj;
361 
362 	err = got_object_open(&obj, repo, id);
363 	if (err)
364 		return err;
365 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
366 		err = got_error(GOT_ERR_OBJ_TYPE);
367 		goto done;
368 	}
369 
370 	err = got_object_commit_open(commit, repo, obj);
371 done:
372 	got_object_close(obj);
373 	return err;
374 }
375 
376 struct got_commit_graph_branch {
377 	struct got_object_id parent_id;
378 	struct got_commit_graph_node *node;
379 };
380 
381 struct gather_branches_arg {
382 	struct got_commit_graph_branch *branches;
383 	int nbranches;
384 };
385 
386 static void
387 gather_branches(struct got_object_id *id, void *data, void *arg)
388 {
389 	struct gather_branches_arg *a = arg;
390 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
391 	a->branches[a->nbranches].node = data;
392 	a->nbranches++;
393 }
394 
395 const struct got_error *
396 fetch_commits_from_open_branches(int *ncommits,
397     struct got_commit_graph *graph, struct got_repository *repo)
398 {
399 	const struct got_error *err;
400 	struct got_commit_graph_branch *branches;
401 	struct gather_branches_arg arg;
402 	int i;
403 
404 	*ncommits = 0;
405 
406 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
407 	if (arg.nbranches == 0)
408 		return NULL;
409 
410 	/*
411 	 * Adding nodes to the graph might change the graph's open
412 	 * branches state. Create a local copy of the current state.
413 	 */
414 	branches = calloc(arg.nbranches, sizeof(*branches));
415 	if (branches == NULL)
416 		return got_error_from_errno();
417 	arg.nbranches = 0; /* reset; gather_branches() will increment */
418 	arg.branches = branches;
419 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
420 
421 	for (i = 0; i < arg.nbranches; i++) {
422 		struct got_object_id *commit_id;
423 		struct got_commit_graph_node *child_node, *new_node;
424 		struct got_commit_object *commit;
425 
426 		commit_id = &branches[i].parent_id;
427 		child_node = branches[i].node;
428 
429 		err = open_commit(&commit, commit_id, repo);
430 		if (err)
431 			break;
432 
433 		err = add_node(&new_node, graph, commit_id, commit,
434 		    &child_node->id);
435 		if (err) {
436 			if (err->code != GOT_ERR_OBJ_EXISTS)
437 				break;
438 			err = NULL;
439 		}
440 		if (new_node)
441 			(*ncommits)++;
442 		else
443 			got_object_commit_close(commit);
444 	}
445 
446 	free(branches);
447 	return err;
448 }
449 
450 const struct got_error *
451 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
452     int limit, struct got_repository *repo)
453 {
454 	const struct got_error *err;
455 	int total = 0, ncommits;
456 
457 	*nfetched = 0;
458 
459 	while (total < limit) {
460 		err = fetch_commits_from_open_branches(&ncommits, graph, repo);
461 		if (err)
462 			return err;
463 		if (ncommits == 0)
464 			break;
465 		total += ncommits;
466 	}
467 
468 	*nfetched = total;
469 	return NULL;
470 }
471 
472 static void
473 free_graph_node(struct got_object_id *id, void *data, void *arg)
474 {
475 	struct got_commit_graph_node *node = data;
476 	got_object_commit_close(node->commit);
477 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
478 		struct got_parent_id *child = SIMPLEQ_FIRST(&node->child_ids);
479 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
480 		free(child);
481 	}
482 	free(node);
483 }
484 
485 void
486 got_commit_graph_close(struct got_commit_graph *graph)
487 {
488 	got_object_idset_free(graph->open_branches);
489 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
490 	got_object_idset_free(graph->node_ids);
491 	free(graph);
492 }
493 
494 const struct got_error *
495 got_commit_graph_iter_start(struct got_commit_graph *graph,
496     struct got_object_id *id)
497 {
498 	struct got_commit_graph_node *start_node, *node;
499 	struct got_parent_id *pid;
500 
501 	start_node = got_object_idset_get(graph->node_ids, id);
502 	if (start_node == NULL)
503 		return got_error(GOT_ERR_NO_OBJ);
504 
505 	graph->iter_node = start_node;
506 
507 	while (!TAILQ_EMPTY(&graph->iter_candidates)) {
508 		node = TAILQ_FIRST(&graph->iter_candidates);
509 		TAILQ_REMOVE(&graph->iter_candidates, node, entry);
510 	}
511 
512 	/* Put all known parents of this commit on the candidate list. */
513 	SIMPLEQ_FOREACH(pid, &start_node->commit->parent_ids, entry) {
514 		node = got_object_idset_get(graph->node_ids, pid->id);
515 		if (node) {
516 			const struct got_error *err;
517 			err = add_iteration_candidate(graph, node);
518 			if (err)
519 				return err;
520 		}
521 	}
522 
523 	return NULL;
524 }
525 
526 const struct got_error *
527 got_commit_graph_iter_next(struct got_commit_object **commit,
528     struct got_object_id **id, struct got_commit_graph *graph)
529 {
530 	struct got_commit_graph_node *node;
531 
532 	if (graph->iter_node == NULL) {
533 		/* We are done interating, or iteration was not started. */
534 		*commit = NULL;
535 		*id = NULL;
536 		return NULL;
537 	}
538 
539 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
540 		if (is_root_node(graph->iter_node) &&
541 		    got_object_idset_num_elements(graph->open_branches) == 0) {
542 			*commit = graph->iter_node->commit;
543 			*id = &graph->iter_node->id;
544 			/* We are done interating. */
545 			graph->iter_node = NULL;
546 			return NULL;
547 		}
548 		return got_error(GOT_ERR_ITER_NEED_MORE);
549 	}
550 
551 	*commit = graph->iter_node->commit;
552 	*id = &graph->iter_node->id;
553 	node = TAILQ_FIRST(&graph->iter_candidates);
554 	TAILQ_REMOVE(&graph->iter_candidates, node, entry);
555 	graph->iter_node = node;
556 	return NULL;
557 }
558 
559 const struct got_commit_object *
560 got_commit_graph_get_commit(struct got_commit_graph *graph,
561     struct got_object_id *id)
562 {
563 	struct got_commit_graph_node *node;
564 	node = got_object_idset_get(graph->node_ids, id);
565 	return node ? node->commit : NULL;
566 }