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_inflate.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 an adjacency list.
44 	 * Adjacencies of a graph node are parent (older) commits.
45 	 */
46 	int nparents;
47 	struct got_object_id_queue parent_ids;
48 
49 	time_t commit_timestamp;
50 
51 	/* Used during graph iteration. */
52 	TAILQ_ENTRY(got_commit_graph_node) entry;
53 };
54 
55 TAILQ_HEAD(got_commit_graph_iter_list, got_commit_graph_node);
56 
57 struct got_commit_graph_branch_tip {
58 	struct got_object_id id;
59 	struct got_commit_graph_node *node;
60 };
61 
62 struct got_commit_graph {
63 	/* The set of all commits we have traversed. */
64 	struct got_object_idset *node_ids;
65 
66 	/* The commit at which traversal began (youngest commit in node_ids). */
67 	struct got_commit_graph_node *head_node;
68 
69 	int flags;
70 #define GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL		0x01
71 
72 	/*
73 	 * A set of object IDs of known parent commits which we have not yet
74 	 * traversed. Each commit ID in this set represents a branch in commit
75 	 * history: Either the first-parent branch of the head node, or another
76 	 * branch corresponding to a traversed merge commit for which we have
77 	 * not traversed a branch point commit yet.
78 	 *
79 	 * Whenever we add a commit with a matching ID to the graph, we remove
80 	 * its corresponding element from this set, and add new elements for
81 	 * each of that commit's parent commits which were not traversed yet.
82 	 *
83 	 * When API users ask us to fetch more commits, we fetch commits from
84 	 * all currently open branches. This allows API users to process
85 	 * commits in linear order even though the history contains branches.
86 	 */
87 	struct got_object_idset *open_branches;
88 
89 	/* Copy of known branch tips for fetch_commits_from_open_branches(). */
90 	struct got_commit_graph_branch_tip *tips;
91 	size_t ntips;
92 #define GOT_COMMIT_GRAPH_MIN_TIPS 100	/* minimum amount of tips to allocate */
93 
94 	/* The next commit to return when the API user asks for one. */
95 	struct got_commit_graph_node *iter_node;
96 
97 	/* The graph iteration list contains all nodes in sorted order. */
98 	struct got_commit_graph_iter_list iter_list;
99 };
100 
101 static struct got_commit_graph *
102 alloc_graph(void)
103 {
104 	struct got_commit_graph *graph;
105 
106 	graph = calloc(1, sizeof(*graph));
107 	if (graph == NULL)
108 		return NULL;
109 
110 	graph->node_ids = got_object_idset_alloc();
111 	if (graph->node_ids == NULL) {
112 		free(graph);
113 		return NULL;
114 	}
115 
116 	graph->open_branches = got_object_idset_alloc();
117 	if (graph->open_branches == NULL) {
118 		got_object_idset_free(graph->node_ids);
119 		free(graph);
120 		return NULL;
121 	}
122 
123 	TAILQ_INIT(&graph->iter_list);
124 	return graph;
125 }
126 
127 #if 0
128 static int
129 is_head_node(struct got_commit_graph_node *node)
130 {
131 	return node->nchildren == 0;
132 }
133 
134 static int
135 is_merge_point(struct got_commit_graph_node *node)
136 {
137 	return node->nparents > 1;
138 }
139 
140 int
141 is_branch_point(struct got_commit_graph_node *node)
142 {
143 	return node->nchildren > 1;
144 }
145 
146 static int
147 is_root_node(struct got_commit_graph_node *node)
148 {
149 	return node->nparents == 0;
150 }
151 #endif
152 
153 static void
154 add_node_to_iter_list(struct got_commit_graph *graph,
155     struct got_commit_graph_node *node,
156     struct got_commit_graph_node *child_node)
157 {
158 	struct got_commit_graph_node *n, *next;
159 	
160 	if (TAILQ_EMPTY(&graph->iter_list)) {
161 		TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
162 		return;
163 	}
164 
165 	/* Ensure that an iteration in progress will see this new commit. */
166 	if (graph->iter_node) {
167 		n = graph->iter_node;
168 		while (n) {
169 			next = TAILQ_NEXT(n, entry);
170 			if (next &&
171 			    node->commit_timestamp >= next->commit_timestamp) {
172 				TAILQ_INSERT_BEFORE(next, node, entry);
173 				return;
174 			}
175 			n = next;
176 		}
177 		TAILQ_INSERT_AFTER(&graph->iter_list, graph->iter_node,
178 		    node, entry);
179 		return;
180 	}
181 
182 	/*
183 	 * If a child node is known, begin looping over the list there
184 	 * instead of beginning from the list head.
185 	 */
186 	n = child_node ? child_node : TAILQ_FIRST(&graph->iter_list);
187 
188 	/* Insert into list based on committer timestamp. */
189 	do {
190 		if (node->commit_timestamp == n->commit_timestamp) {
191 			TAILQ_INSERT_AFTER(&graph->iter_list, n, node, entry);
192 			break;
193 		} else if (node->commit_timestamp < n->commit_timestamp) {
194 			next = TAILQ_NEXT(n, entry);
195 			if (next == NULL) {
196 				TAILQ_INSERT_AFTER(&graph->iter_list, n,
197 				    node, entry);
198 				break;
199 			}
200 			if (node->commit_timestamp >= next->commit_timestamp) {
201 				TAILQ_INSERT_BEFORE(next, node, entry);
202 				break;
203 			}
204 		} else {
205 			TAILQ_INSERT_BEFORE(n, node, entry);
206 			break;
207 		}
208 		n = TAILQ_NEXT(n, entry);
209 	} while (n);
210 }
211 
212 static const struct got_error *
213 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
214 {
215 	struct got_object_qid *qid;
216 
217 	qid = calloc(1, sizeof(*qid));
218 	if (qid == NULL)
219 		return got_error_from_errno();
220 
221 	qid->id = got_object_id_dup(id);
222 	if (qid->id == NULL) {
223 		const struct got_error *err = got_error_from_errno();
224 		got_object_qid_free(qid);
225 		return err;
226 	}
227 
228 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
229 	return NULL;
230 }
231 
232 static const struct got_error *
233 advance_open_branches(struct got_commit_graph *graph,
234     struct got_commit_graph_node *node,
235     struct got_object_id *commit_id, struct got_commit_object *commit)
236 {
237 	const struct got_error *err;
238 	struct got_object_qid *qid;
239 
240 	err = got_object_idset_remove(NULL, graph->open_branches, commit_id);
241 	if (err && err->code != GOT_ERR_NO_OBJ)
242 		return err;
243 
244 	if (graph->flags & GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL) {
245 		qid = SIMPLEQ_FIRST(&commit->parent_ids);
246 		if (qid == NULL)
247 			return NULL;
248 		err = got_object_idset_add(NULL, graph->open_branches, qid->id,
249 		    node);
250 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
251 			return err;
252 		return NULL;
253 	}
254 
255 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
256 		if (got_object_idset_get(graph->node_ids, qid->id))
257 			continue; /* parent already traversed */
258 		err = got_object_idset_add(NULL, graph->open_branches,
259 		    qid->id, node);
260 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
261 			return err;
262 	}
263 
264 	return NULL;
265 }
266 
267 static void
268 free_node(struct got_commit_graph_node *node)
269 {
270 	while (!SIMPLEQ_EMPTY(&node->parent_ids)) {
271 		struct got_object_qid *pid = SIMPLEQ_FIRST(&node->parent_ids);
272 		SIMPLEQ_REMOVE_HEAD(&node->parent_ids, entry);
273 		got_object_qid_free(pid);
274 	}
275 	free(node);
276 }
277 
278 static const struct got_error *
279 add_node(struct got_commit_graph_node **new_node,
280     struct got_commit_graph *graph, struct got_object_id *commit_id,
281     struct got_commit_object *commit, struct got_commit_graph_node *child_node)
282 {
283 	const struct got_error *err = NULL;
284 	struct got_commit_graph_node *node;
285 	struct got_object_qid *pid;
286 
287 	*new_node = NULL;
288 
289 	node = calloc(1, sizeof(*node));
290 	if (node == NULL)
291 		return got_error_from_errno();
292 
293 	memcpy(&node->id, commit_id, sizeof(node->id));
294 	SIMPLEQ_INIT(&node->parent_ids);
295 	SIMPLEQ_FOREACH(pid, &commit->parent_ids, entry) {
296 		err = add_vertex(&node->parent_ids, pid->id);
297 		if (err) {
298 			free_node(node);
299 			return err;
300 		}
301 		node->nparents++;
302 	}
303 	node->commit_timestamp = mktime(&commit->tm_committer); 
304 	if (node->commit_timestamp == -1) {
305 		free_node(node);
306 		return got_error_from_errno();
307 	}
308 
309 	err = got_object_idset_add(NULL, graph->node_ids, &node->id, node);
310 	if (err) {
311 		if (err->code == GOT_ERR_OBJ_EXISTS)
312 			err = NULL;
313 		free_node(node);
314 		return err;
315 	}
316 
317 	add_node_to_iter_list(graph, node, child_node);
318 	err = advance_open_branches(graph, node, commit_id, commit);
319 	if (err)
320 		free_node(node);
321 	else
322 		*new_node = node;
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 gather_branch_tips_arg {
362 	struct got_commit_graph_branch_tip *tips;
363 	int ntips;
364 };
365 
366 static void
367 gather_branch_tips(struct got_object_id *id, void *data, void *arg)
368 {
369 	struct gather_branch_tips_arg *a = arg;
370 	memcpy(&a->tips[a->ntips].id, id, sizeof(*id));
371 	a->tips[a->ntips].node = data;
372 	a->ntips++;
373 }
374 
375 static const struct got_error *
376 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
377     struct got_commit_graph *graph, struct got_repository *repo,
378     struct got_object_id *wanted_id)
379 {
380 	const struct got_error *err;
381 	struct gather_branch_tips_arg arg;
382 	int i;
383 
384 	*ncommits = 0;
385 	if (wanted_id_added)
386 		*wanted_id_added = 0;
387 
388 	arg.ntips = got_object_idset_num_elements(graph->open_branches);
389 	if (arg.ntips == 0)
390 		return NULL;
391 
392 	/*
393 	 * Adding nodes to the graph might change the graph's open
394 	 * branches state. Create a local copy of the current state.
395 	 */
396 	if (graph->ntips < arg.ntips) {
397 		struct got_commit_graph_branch_tip *tips;
398 		if (arg.ntips < GOT_COMMIT_GRAPH_MIN_TIPS)
399 			arg.ntips = GOT_COMMIT_GRAPH_MIN_TIPS;
400 		tips = reallocarray(graph->tips, arg.ntips, sizeof(*tips));
401 		if (tips == NULL)
402 			return got_error_from_errno();
403 		graph->tips = tips;
404 		graph->ntips = arg.ntips;
405 	}
406 	arg.ntips = 0; /* reset; gather_branch_tips() will increment */
407 	arg.tips = graph->tips;
408 	got_object_idset_for_each(graph->open_branches,
409 	    gather_branch_tips, &arg);
410 
411 	for (i = 0; i < arg.ntips; i++) {
412 		struct got_object_id *commit_id;
413 		struct got_commit_graph_node *child_node, *new_node;
414 		struct got_commit_object *commit;
415 
416 		commit_id = &graph->tips[i].id;
417 		child_node = graph->tips[i].node;
418 
419 		err = got_object_open_as_commit(&commit, repo, commit_id);
420 		if (err)
421 			break;
422 
423 		err = add_node(&new_node, graph, commit_id, commit, child_node);
424 		got_object_commit_close(commit);
425 		if (err)
426 			break;
427 		if (new_node)
428 			(*ncommits)++;
429 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
430 			*wanted_id_added = 1;
431 	}
432 
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 ncommits;
442 
443 	*nfetched = 0;
444 
445 	while (*nfetched < limit) {
446 		err = fetch_commits_from_open_branches(&ncommits, NULL,
447 		    graph, repo, NULL);
448 		if (err)
449 			return err;
450 		if (ncommits == 0)
451 			break;
452 		*nfetched += ncommits;
453 	}
454 
455 	return NULL;
456 }
457 
458 const struct got_error *
459 got_commit_graph_fetch_commits_up_to(int *nfetched,
460     struct got_commit_graph *graph, struct got_object_id *wanted_id,
461     struct got_repository *repo)
462 {
463 	const struct got_error *err;
464 	int ncommits, wanted_id_added = 0;
465 
466 	*nfetched = 0;
467 
468 	if (got_object_idset_get(graph->node_ids, wanted_id) != NULL)
469 		return NULL;
470 
471 	while (!wanted_id_added) {
472 		err = fetch_commits_from_open_branches(&ncommits,
473 		    &wanted_id_added, graph, repo, wanted_id);
474 		if (err)
475 			return err;
476 		if (ncommits == 0)
477 			return NULL;
478 		*nfetched += ncommits;
479 	}
480 
481 	return NULL;
482 }
483 
484 static void
485 free_node_iter(struct got_object_id *id, void *data, void *arg)
486 {
487 	struct got_commit_graph_node *node = data;
488 	free_node(node);
489 }
490 
491 void
492 got_commit_graph_close(struct got_commit_graph *graph)
493 {
494 	got_object_idset_free(graph->open_branches);
495 	got_object_idset_for_each(graph->node_ids, free_node_iter, NULL);
496 	got_object_idset_free(graph->node_ids);
497 	free(graph->tips);
498 	free(graph);
499 }
500 
501 const struct got_error *
502 got_commit_graph_iter_start(struct got_commit_graph *graph,
503     struct got_object_id *id)
504 {
505 	struct got_commit_graph_node *start_node;
506 
507 	start_node = got_object_idset_get(graph->node_ids, id);
508 	if (start_node == NULL)
509 		return got_error(GOT_ERR_NO_OBJ);
510 
511 	graph->iter_node = start_node;
512 	return NULL;
513 }
514 
515 const struct got_error *
516 got_commit_graph_iter_next(struct got_object_id **id,
517     struct got_commit_graph *graph)
518 {
519 	*id = NULL;
520 
521 	if (graph->iter_node == NULL) {
522 		/* We are done interating, or iteration was not started. */
523 		return got_error(GOT_ERR_ITER_COMPLETED);
524 	}
525 
526 	if (graph->iter_node ==
527 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
528 	    got_object_idset_num_elements(graph->open_branches) == 0) {
529 		*id = &graph->iter_node->id;
530 		/* We are done interating. */
531 		graph->iter_node = NULL;
532 		return NULL;
533 	}
534 
535 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
536 		return got_error(GOT_ERR_ITER_NEED_MORE);
537 
538 	*id = &graph->iter_node->id;
539 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
540 	return NULL;
541 }