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;
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 			continue;
217 		TAILQ_INSERT_BEFORE(n, node, entry);
218 		break;
219 	}
220 
221 	return NULL;
222 }
223 
224 static const struct got_error *
225 add_node(struct got_commit_graph_node **new_node,
226     struct got_commit_graph *graph, struct got_object_id *commit_id,
227     struct got_commit_object *commit, struct got_object_id *child_commit_id)
228 {
229 	const struct got_error *err = NULL;
230 	struct got_commit_graph_node *node, *existing_node;
231 
232 	*new_node = NULL;
233 
234 	node = calloc(1, sizeof(*node));
235 	if (node == NULL)
236 		return got_error_from_errno();
237 
238 	memcpy(&node->id, commit_id, sizeof(node->id));
239 	node->commit = commit;
240 	SIMPLEQ_INIT(&node->child_ids);
241 
242 	err = got_object_idset_add((void **)(&existing_node),
243 	    graph->node_ids, &node->id, node);
244 	if (err == NULL) {
245 		struct got_parent_id *pid;
246 
247 		err = add_iteration_candidate(graph, node);
248 		if (err)
249 			return err;
250 
251 		err = got_object_idset_remove(graph->open_branches, commit_id);
252 		if (err && err->code != GOT_ERR_NO_OBJ)
253 			return err;
254 		SIMPLEQ_FOREACH(pid, &commit->parent_ids, entry) {
255 			if (got_object_idset_get(graph->node_ids, pid->id))
256 				continue; /* parent already traversed */
257 			err = got_object_idset_add(NULL, graph->open_branches,
258 			    pid->id, node);
259 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
260 				return err;
261 		}
262 		*new_node = node;
263 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
264 		err = NULL;
265 		free(node);
266 		node = existing_node;
267 	} else {
268 		free(node);
269 		return err;
270 	}
271 
272 	if (child_commit_id) {
273 		struct got_parent_id *child, *cid;
274 
275 		/* Prevent linking to self. */
276 		if (got_object_id_cmp(commit_id, child_commit_id) == 0)
277 			return got_error(GOT_ERR_BAD_OBJ_ID);
278 
279 		/* Prevent double-linking to the same child. */
280 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
281 			if (got_object_id_cmp(cid->id, child_commit_id) == 0)
282 				return got_error(GOT_ERR_BAD_OBJ_ID);
283 		}
284 
285 		child = calloc(1, sizeof(*child));
286 		if (child == NULL)
287 			return got_error_from_errno();
288 		child->id = got_object_id_dup(child_commit_id);
289 		if (child->id == NULL) {
290 			err = got_error_from_errno();
291 			free(child);
292 			return err;
293 		}
294 		SIMPLEQ_INSERT_TAIL(&node->child_ids, child, entry);
295 		node->nchildren++;
296 	}
297 
298 	return err;
299 }
300 
301 const struct got_error *
302 got_commit_graph_open(struct got_commit_graph **graph,
303     struct got_object_id *commit_id, struct got_repository *repo)
304 {
305 	const struct got_error *err = NULL;
306 	struct got_object *obj;
307 	struct got_commit_object *commit;
308 
309 	*graph = NULL;
310 
311 	err = got_object_open(&obj, repo, commit_id);
312 	if (err)
313 		return err;
314 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
315 		err = got_error(GOT_ERR_OBJ_TYPE);
316 		got_object_close(obj);
317 		return err;
318 	}
319 
320 	err = got_object_commit_open(&commit, repo, obj);
321 	got_object_close(obj);
322 	if (err)
323 		return err;
324 
325 	*graph = alloc_graph();
326 	if (*graph == NULL) {
327 		got_object_commit_close(commit);
328 		return got_error_from_errno();
329 	}
330 
331 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
332 	if (err) {
333 		got_commit_graph_close(*graph);
334 		*graph = NULL;
335 		return err;
336 	}
337 	
338 	return NULL;
339 }
340 
341 static const struct got_error *
342 open_commit(struct got_commit_object **commit, struct got_object_id *id,
343     struct got_repository *repo)
344 {
345 	const struct got_error *err;
346 	struct got_object *obj;
347 
348 	err = got_object_open(&obj, repo, id);
349 	if (err)
350 		return err;
351 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
352 		err = got_error(GOT_ERR_OBJ_TYPE);
353 		goto done;
354 	}
355 
356 	err = got_object_commit_open(commit, repo, obj);
357 done:
358 	got_object_close(obj);
359 	return err;
360 }
361 
362 struct got_commit_graph_branch {
363 	struct got_object_id parent_id;
364 	struct got_commit_graph_node *node;
365 };
366 
367 struct gather_branches_arg {
368 	struct got_commit_graph_branch *branches;
369 	int nbranches;
370 };
371 
372 static void
373 gather_branches(struct got_object_id *id, void *data, void *arg)
374 {
375 	struct gather_branches_arg *a = arg;
376 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
377 	a->branches[a->nbranches].node = data;
378 	a->nbranches++;
379 }
380 
381 const struct got_error *
382 fetch_commits_from_open_branches(int *ncommits,
383     struct got_commit_graph *graph, struct got_repository *repo)
384 {
385 	const struct got_error *err;
386 	struct got_commit_graph_branch *branches;
387 	struct gather_branches_arg arg;
388 	int i;
389 
390 	*ncommits = 0;
391 
392 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
393 	if (arg.nbranches == 0)
394 		return NULL;
395 
396 	/*
397 	 * Adding nodes to the graph might change the graph's open
398 	 * branches state. Create a local copy of the current state.
399 	 */
400 	branches = calloc(arg.nbranches, sizeof(*branches));
401 	if (branches == NULL)
402 		return got_error_from_errno();
403 	arg.nbranches = 0; /* reset; gather_branches() will increment */
404 	arg.branches = branches;
405 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
406 
407 	for (i = 0; i < arg.nbranches; i++) {
408 		struct got_object_id *commit_id;
409 		struct got_commit_graph_node *child_node, *new_node;
410 		struct got_commit_object *commit;
411 
412 		commit_id = &branches[i].parent_id;
413 		child_node = branches[i].node;
414 
415 		err = open_commit(&commit, commit_id, repo);
416 		if (err)
417 			break;
418 
419 		err = add_node(&new_node, graph, commit_id, commit,
420 		    &child_node->id);
421 		if (err) {
422 			if (err->code != GOT_ERR_OBJ_EXISTS)
423 				break;
424 			err = NULL;
425 		}
426 		if (new_node)
427 			(*ncommits)++;
428 		else
429 			got_object_commit_close(commit);
430 	}
431 
432 	free(branches);
433 	return err;
434 }
435 
436 const struct got_error *
437 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
438     int limit, struct got_repository *repo)
439 {
440 	const struct got_error *err;
441 	int total = 0, ncommits;
442 
443 	*nfetched = 0;
444 
445 	while (total < limit) {
446 		err = fetch_commits_from_open_branches(&ncommits, graph, repo);
447 		if (err)
448 			return err;
449 		if (ncommits == 0)
450 			break;
451 		total += ncommits;
452 	}
453 
454 	*nfetched = total;
455 	return NULL;
456 }
457 
458 static void
459 free_graph_node(struct got_object_id *id, void *data, void *arg)
460 {
461 	struct got_commit_graph_node *node = data;
462 	got_object_commit_close(node->commit);
463 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
464 		struct got_parent_id *child = SIMPLEQ_FIRST(&node->child_ids);
465 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
466 		free(child);
467 	}
468 	free(node);
469 }
470 
471 void
472 got_commit_graph_close(struct got_commit_graph *graph)
473 {
474 	got_object_idset_free(graph->open_branches);
475 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
476 	got_object_idset_free(graph->node_ids);
477 	free(graph);
478 }
479 
480 const struct got_error *
481 got_commit_graph_iter_start(struct got_commit_graph *graph,
482     struct got_object_id *id)
483 {
484 	struct got_commit_graph_node *start_node, *node;
485 	struct got_parent_id *pid;
486 
487 	start_node = got_object_idset_get(graph->node_ids, id);
488 	if (start_node == NULL)
489 		return got_error(GOT_ERR_NO_OBJ);
490 
491 	graph->iter_node = start_node;
492 
493 	while (!TAILQ_EMPTY(&graph->iter_candidates)) {
494 		node = TAILQ_FIRST(&graph->iter_candidates);
495 		TAILQ_REMOVE(&graph->iter_candidates, node, entry);
496 	}
497 
498 	/* Put all known parents of this commit on the candidate list. */
499 	SIMPLEQ_FOREACH(pid, &start_node->commit->parent_ids, entry) {
500 		node = got_object_idset_get(graph->node_ids, pid->id);
501 		if (node) {
502 			const struct got_error *err;
503 			err = add_iteration_candidate(graph, node);
504 			if (err)
505 				return err;
506 		}
507 	}
508 
509 	return NULL;
510 }
511 
512 const struct got_error *
513 got_commit_graph_iter_next(struct got_commit_object **commit,
514     struct got_object_id **id, struct got_commit_graph *graph)
515 {
516 	struct got_commit_graph_node *node;
517 
518 	if (graph->iter_node == NULL) {
519 		/* We are done interating, or iteration was not started. */
520 		*commit = NULL;
521 		*id = NULL;
522 		return NULL;
523 	}
524 
525 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
526 		if (is_root_node(graph->iter_node) &&
527 		    got_object_idset_num_elements(graph->open_branches) == 0) {
528 			*commit = graph->iter_node->commit;
529 			*id = &graph->iter_node->id;
530 			/* We are done interating. */
531 			graph->iter_node = NULL;
532 			return NULL;
533 		}
534 		return got_error(GOT_ERR_ITER_NEED_MORE);
535 	}
536 
537 	*commit = graph->iter_node->commit;
538 	*id = &graph->iter_node->id;
539 	node = TAILQ_FIRST(&graph->iter_candidates);
540 	TAILQ_REMOVE(&graph->iter_candidates, node, entry);
541 	graph->iter_node = node;
542 	return NULL;
543 }
544 
545 const struct got_commit_object *
546 got_commit_graph_get_commit(struct got_commit_graph *graph,
547     struct got_object_id *id)
548 {
549 	struct got_commit_graph_node *node;
550 	node = got_object_idset_get(graph->node_ids, id);
551 	return node ? node->commit : NULL;
552 }