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 	int nparents;
48 	struct got_object_id_queue parent_ids;
49 	int nchildren;
50 	struct got_object_id_queue child_ids;
51 
52 	time_t commit_timestamp;
53 
54 	/* Used during graph iteration. */
55 	TAILQ_ENTRY(got_commit_graph_node) entry;
56 };
57 
58 TAILQ_HEAD(got_commit_graph_iter_list, got_commit_graph_node);
59 
60 struct got_commit_graph {
61 	/* The set of all commits we have traversed. */
62 	struct got_object_idset *node_ids;
63 
64 	/* The commit at which traversal began (youngest commit in node_ids). */
65 	struct got_commit_graph_node *head_node;
66 
67 	int flags;
68 #define GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL		0x01
69 
70 	/*
71 	 * A set of object IDs of known parent commits which we have not yet
72 	 * traversed. Each commit ID in this set represents a branch in commit
73 	 * history: Either the first-parent branch of the head node, or another
74 	 * branch corresponding to a traversed merge commit for which we have
75 	 * not traversed a branch point commit yet.
76 	 *
77 	 * Whenever we add a commit with a matching ID to the graph, we remove
78 	 * its corresponding element from this set, and add new elements for
79 	 * each of that commit's parent commits which were not traversed yet.
80 	 *
81 	 * When API users ask us to fetch more commits, we fetch commits from
82 	 * all currently open branches. This allows API users to process
83 	 * commits in linear order even though the history contains branches.
84 	 */
85 	struct got_object_idset *open_branches;
86 
87 	/* The next commit to return when the API user asks for one. */
88 	struct got_commit_graph_node *iter_node;
89 
90 	/* The graph iteration list contains all nodes in sorted order. */
91 	struct got_commit_graph_iter_list iter_list;
92 };
93 
94 static struct got_commit_graph *
95 alloc_graph(void)
96 {
97 	struct got_commit_graph *graph;
98 
99 	graph = calloc(1, sizeof(*graph));
100 	if (graph == NULL)
101 		return NULL;
102 
103 	graph->node_ids = got_object_idset_alloc(
104 	    GOT_OBJECT_IDSET_ITERATE_BY_OBJECT_ID);
105 	if (graph->node_ids == NULL) {
106 		free(graph);
107 		return NULL;
108 	}
109 
110 	graph->open_branches = got_object_idset_alloc(
111 	     GOT_OBJECT_IDSET_ITERATE_BY_OBJECT_ID);
112 	if (graph->open_branches == NULL) {
113 		got_object_idset_free(graph->node_ids);
114 		free(graph);
115 		return NULL;
116 	}
117 
118 	TAILQ_INIT(&graph->iter_list);
119 	return graph;
120 }
121 
122 #if 0
123 static int
124 is_head_node(struct got_commit_graph_node *node)
125 {
126 	return node->nchildren == 0;
127 }
128 
129 static int
130 is_merge_point(struct got_commit_graph_node *node)
131 {
132 	return node->nparents > 1;
133 }
134 
135 int
136 is_branch_point(struct got_commit_graph_node *node)
137 {
138 	return node->nchildren > 1;
139 }
140 
141 static int
142 is_root_node(struct got_commit_graph_node *node)
143 {
144 	return node->nparents == 0;
145 }
146 #endif
147 
148 static void
149 add_node_to_iter_list(struct got_commit_graph *graph,
150     struct got_commit_graph_node *node,
151     struct got_commit_graph_node *child_node)
152 {
153 	struct got_commit_graph_node *n, *next;
154 	
155 	if (TAILQ_EMPTY(&graph->iter_list)) {
156 		TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
157 		return;
158 	}
159 
160 	/* Ensure that an iteration in progress will see this new commit. */
161 	if (graph->iter_node) {
162 		n = graph->iter_node;
163 		while (n) {
164 			next = TAILQ_NEXT(n, entry);
165 			if (next &&
166 			    node->commit_timestamp >= next->commit_timestamp) {
167 				TAILQ_INSERT_BEFORE(next, node, entry);
168 				return;
169 			}
170 			n = next;
171 		}
172 		TAILQ_INSERT_AFTER(&graph->iter_list, graph->iter_node,
173 		    node, entry);
174 		return;
175 	}
176 
177 	/*
178 	 * If a child node is known, begin looping over the list there
179 	 * instead of beginning from the list head.
180 	 */
181 	n = child_node ? child_node : TAILQ_FIRST(&graph->iter_list);
182 
183 	/* Insert into list based on committer timestamp. */
184 	do {
185 		if (node->commit_timestamp == n->commit_timestamp) {
186 			TAILQ_INSERT_AFTER(&graph->iter_list, n, node, entry);
187 			break;
188 		} else if (node->commit_timestamp < n->commit_timestamp) {
189 			next = TAILQ_NEXT(n, entry);
190 			if (next == NULL) {
191 				TAILQ_INSERT_AFTER(&graph->iter_list, n,
192 				    node, entry);
193 				break;
194 			}
195 			if (node->commit_timestamp >= next->commit_timestamp) {
196 				TAILQ_INSERT_BEFORE(next, node, entry);
197 				break;
198 			}
199 		} else {
200 			TAILQ_INSERT_BEFORE(n, node, entry);
201 			break;
202 		}
203 		n = TAILQ_NEXT(n, entry);
204 	} while (n);
205 }
206 
207 static const struct got_error *
208 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
209 {
210 	struct got_object_qid *qid;
211 
212 	qid = calloc(1, sizeof(*qid));
213 	if (qid == NULL)
214 		return got_error_from_errno();
215 
216 	qid->id = got_object_id_dup(id);
217 	if (qid->id == NULL) {
218 		const struct got_error *err = got_error_from_errno();
219 		free(qid);
220 		return err;
221 	}
222 
223 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
224 	return NULL;
225 }
226 
227 static const struct got_error *
228 advance_open_branches(struct got_commit_graph *graph,
229     struct got_commit_graph_node *node,
230     struct got_object_id *commit_id, struct got_commit_object *commit)
231 {
232 	const struct got_error *err;
233 	struct got_object_qid *qid;
234 
235 	err = got_object_idset_remove(NULL, graph->open_branches, commit_id);
236 	if (err && err->code != GOT_ERR_NO_OBJ)
237 		return err;
238 
239 	if (graph->flags & GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL) {
240 		qid = SIMPLEQ_FIRST(&commit->parent_ids);
241 		if (qid == NULL)
242 			return NULL;
243 		err = got_object_idset_add(NULL, graph->open_branches, qid->id,
244 		    node);
245 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
246 			return err;
247 		return NULL;
248 	}
249 
250 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
251 		if (got_object_idset_get(graph->node_ids, qid->id))
252 			continue; /* parent already traversed */
253 		err = got_object_idset_add(NULL, graph->open_branches,
254 		    qid->id, node);
255 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
256 			return err;
257 	}
258 
259 	return NULL;
260 }
261 
262 static const struct got_error *
263 add_node(struct got_commit_graph_node **new_node,
264     struct got_commit_graph *graph, struct got_object_id *commit_id,
265     struct got_commit_object *commit, struct got_commit_graph_node *child_node)
266 {
267 	const struct got_error *err = NULL;
268 	struct got_commit_graph_node *node, *existing_node;
269 	struct got_object_qid *qid;
270 
271 	*new_node = NULL;
272 
273 	node = calloc(1, sizeof(*node));
274 	if (node == NULL)
275 		return got_error_from_errno();
276 
277 	memcpy(&node->id, commit_id, sizeof(node->id));
278 	SIMPLEQ_INIT(&node->parent_ids);
279 	SIMPLEQ_INIT(&node->child_ids);
280 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
281 		err = add_vertex(&node->parent_ids, qid->id);
282 		if (err)
283 			return err;
284 		node->nparents++;
285 	}
286 	node->commit_timestamp = mktime(&commit->tm_committer); 
287 	if (node->commit_timestamp == -1)
288 		return got_error_from_errno();
289 
290 	err = got_object_idset_add((void **)(&existing_node),
291 	    graph->node_ids, &node->id, node);
292 	if (err == NULL) {
293 		add_node_to_iter_list(graph, node, child_node);
294 		err = advance_open_branches(graph, node, commit_id, commit);
295 		*new_node = node;
296 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
297 		err = NULL;
298 		free(node);
299 		node = existing_node;
300 	} else {
301 		free(node);
302 		return err;
303 	}
304 
305 	if (child_node) {
306 		struct got_object_qid *cid;
307 
308 		/* Prevent linking to self. */
309 		if (got_object_id_cmp(commit_id, &child_node->id) == 0)
310 			return got_error(GOT_ERR_BAD_OBJ_ID);
311 
312 		/* Prevent double-linking to the same child. */
313 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
314 			if (got_object_id_cmp(cid->id, &child_node->id) == 0)
315 				return got_error(GOT_ERR_BAD_OBJ_ID);
316 		}
317 
318 		err = add_vertex(&node->child_ids, &child_node->id);
319 		if (err)
320 			return err;
321 		node->nchildren++;
322 	}
323 
324 	return err;
325 }
326 
327 const struct got_error *
328 got_commit_graph_open(struct got_commit_graph **graph,
329     struct got_object_id *commit_id, int first_parent_traversal,
330     struct got_repository *repo)
331 {
332 	const struct got_error *err = NULL;
333 	struct got_commit_object *commit;
334 
335 	*graph = NULL;
336 
337 	err = got_object_open_as_commit(&commit, repo, commit_id);
338 	if (err)
339 		return err;
340 
341 	*graph = alloc_graph();
342 	if (*graph == NULL) {
343 		got_object_commit_close(commit);
344 		return got_error_from_errno();
345 	}
346 
347 	if (first_parent_traversal)
348 		(*graph)->flags |= GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL;
349 
350 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
351 	got_object_commit_close(commit);
352 	if (err) {
353 		got_commit_graph_close(*graph);
354 		*graph = NULL;
355 		return err;
356 	}
357 	
358 	return NULL;
359 }
360 
361 struct got_commit_graph_branch {
362 	struct got_object_id parent_id;
363 	struct got_commit_graph_node *node;
364 };
365 
366 struct gather_branches_arg {
367 	struct got_commit_graph_branch *branches;
368 	int nbranches;
369 };
370 
371 static void
372 gather_branches(struct got_object_id *id, void *data, void *arg)
373 {
374 	struct gather_branches_arg *a = arg;
375 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
376 	a->branches[a->nbranches].node = data;
377 	a->nbranches++;
378 }
379 
380 static const struct got_error *
381 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
382     struct got_commit_graph *graph, struct got_repository *repo,
383     struct got_object_id *wanted_id)
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 	if (wanted_id_added)
392 		*wanted_id_added = 0;
393 
394 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
395 	if (arg.nbranches == 0)
396 		return NULL;
397 
398 	/*
399 	 * Adding nodes to the graph might change the graph's open
400 	 * branches state. Create a local copy of the current state.
401 	 */
402 	branches = calloc(arg.nbranches, sizeof(*branches));
403 	if (branches == NULL)
404 		return got_error_from_errno();
405 	arg.nbranches = 0; /* reset; gather_branches() will increment */
406 	arg.branches = branches;
407 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
408 
409 	for (i = 0; i < arg.nbranches; i++) {
410 		struct got_object_id *commit_id;
411 		struct got_commit_graph_node *child_node, *new_node;
412 		struct got_commit_object *commit;
413 
414 		commit_id = &branches[i].parent_id;
415 		child_node = branches[i].node;
416 
417 		err = got_object_open_as_commit(&commit, repo, commit_id);
418 		if (err)
419 			break;
420 
421 		err = add_node(&new_node, graph, commit_id, commit, child_node);
422 		got_object_commit_close(commit);
423 		if (err) {
424 			if (err->code != GOT_ERR_OBJ_EXISTS)
425 				break;
426 			err = NULL;
427 		}
428 		if (new_node)
429 			(*ncommits)++;
430 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
431 			*wanted_id_added = 1;
432 	}
433 
434 	free(branches);
435 	return err;
436 }
437 
438 const struct got_error *
439 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
440     int limit, struct got_repository *repo)
441 {
442 	const struct got_error *err;
443 	int ncommits;
444 
445 	*nfetched = 0;
446 
447 	while (*nfetched < limit) {
448 		err = fetch_commits_from_open_branches(&ncommits, NULL,
449 		    graph, repo, NULL);
450 		if (err)
451 			return err;
452 		if (ncommits == 0)
453 			break;
454 		*nfetched += ncommits;
455 	}
456 
457 	return NULL;
458 }
459 
460 const struct got_error *
461 got_commit_graph_fetch_commits_up_to(int *nfetched,
462     struct got_commit_graph *graph, struct got_object_id *wanted_id,
463     struct got_repository *repo)
464 {
465 	const struct got_error *err;
466 	int ncommits, wanted_id_added = 0;
467 
468 	*nfetched = 0;
469 
470 	if (got_object_idset_get(graph->node_ids, wanted_id) != NULL)
471 		return NULL;
472 
473 	while (!wanted_id_added) {
474 		err = fetch_commits_from_open_branches(&ncommits,
475 		    &wanted_id_added, graph, repo, wanted_id);
476 		if (err)
477 			return err;
478 		if (ncommits == 0)
479 			return NULL;
480 		*nfetched += ncommits;
481 	}
482 
483 	return NULL;
484 }
485 
486 static void
487 free_graph_node(struct got_object_id *id, void *data, void *arg)
488 {
489 	struct got_commit_graph_node *node = data;
490 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
491 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
492 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
493 		free(child);
494 	}
495 	free(node);
496 }
497 
498 void
499 got_commit_graph_close(struct got_commit_graph *graph)
500 {
501 	got_object_idset_free(graph->open_branches);
502 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
503 	got_object_idset_free(graph->node_ids);
504 	free(graph);
505 }
506 
507 const struct got_error *
508 got_commit_graph_iter_start(struct got_commit_graph *graph,
509     struct got_object_id *id)
510 {
511 	struct got_commit_graph_node *start_node;
512 
513 	start_node = got_object_idset_get(graph->node_ids, id);
514 	if (start_node == NULL)
515 		return got_error(GOT_ERR_NO_OBJ);
516 
517 	graph->iter_node = start_node;
518 	return NULL;
519 }
520 
521 const struct got_error *
522 got_commit_graph_iter_next(struct got_object_id **id,
523     struct got_commit_graph *graph)
524 {
525 	*id = NULL;
526 
527 	if (graph->iter_node == NULL) {
528 		/* We are done interating, or iteration was not started. */
529 		return got_error(GOT_ERR_ITER_COMPLETED);
530 	}
531 
532 	if (graph->iter_node ==
533 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
534 	    got_object_idset_num_elements(graph->open_branches) == 0) {
535 		*id = &graph->iter_node->id;
536 		/* We are done interating. */
537 		graph->iter_node = NULL;
538 		return NULL;
539 	}
540 
541 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
542 		return got_error(GOT_ERR_ITER_NEED_MORE);
543 
544 	*id = &graph->iter_node->id;
545 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
546 	return NULL;
547 }