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 	if (graph->node_ids == NULL) {
105 		free(graph);
106 		return NULL;
107 	}
108 
109 	graph->open_branches = got_object_idset_alloc();
110 	if (graph->open_branches == NULL) {
111 		got_object_idset_free(graph->node_ids);
112 		free(graph);
113 		return NULL;
114 	}
115 
116 	TAILQ_INIT(&graph->iter_list);
117 	return graph;
118 }
119 
120 #if 0
121 static int
122 is_head_node(struct got_commit_graph_node *node)
123 {
124 	return node->nchildren == 0;
125 }
126 
127 static int
128 is_merge_point(struct got_commit_graph_node *node)
129 {
130 	return node->nparents > 1;
131 }
132 
133 int
134 is_branch_point(struct got_commit_graph_node *node)
135 {
136 	return node->nchildren > 1;
137 }
138 
139 static int
140 is_root_node(struct got_commit_graph_node *node)
141 {
142 	return node->nparents == 0;
143 }
144 #endif
145 
146 static void
147 add_node_to_iter_list(struct got_commit_graph *graph,
148     struct got_commit_graph_node *node,
149     struct got_commit_graph_node *child_node)
150 {
151 	struct got_commit_graph_node *n, *next;
152 	
153 	if (TAILQ_EMPTY(&graph->iter_list)) {
154 		TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
155 		return;
156 	}
157 
158 	/* Ensure that an iteration in progress will see this new commit. */
159 	if (graph->iter_node) {
160 		n = graph->iter_node;
161 		while (n) {
162 			next = TAILQ_NEXT(n, entry);
163 			if (next &&
164 			    node->commit_timestamp >= next->commit_timestamp) {
165 				TAILQ_INSERT_BEFORE(next, node, entry);
166 				return;
167 			}
168 			n = next;
169 		}
170 		TAILQ_INSERT_AFTER(&graph->iter_list, graph->iter_node,
171 		    node, entry);
172 		return;
173 	}
174 
175 	/*
176 	 * If a child node is known, begin looping over the list there
177 	 * instead of beginning from the list head.
178 	 */
179 	n = child_node ? child_node : TAILQ_FIRST(&graph->iter_list);
180 
181 	/* Insert into list based on committer timestamp. */
182 	do {
183 		if (node->commit_timestamp == n->commit_timestamp) {
184 			TAILQ_INSERT_AFTER(&graph->iter_list, n, node, entry);
185 			break;
186 		} else if (node->commit_timestamp < n->commit_timestamp) {
187 			next = TAILQ_NEXT(n, entry);
188 			if (next == NULL) {
189 				TAILQ_INSERT_AFTER(&graph->iter_list, n,
190 				    node, entry);
191 				break;
192 			}
193 			if (node->commit_timestamp >= next->commit_timestamp) {
194 				TAILQ_INSERT_BEFORE(next, node, entry);
195 				break;
196 			}
197 		} else {
198 			TAILQ_INSERT_BEFORE(n, node, entry);
199 			break;
200 		}
201 		n = TAILQ_NEXT(n, entry);
202 	} while (n);
203 }
204 
205 static const struct got_error *
206 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
207 {
208 	struct got_object_qid *qid;
209 
210 	qid = calloc(1, sizeof(*qid));
211 	if (qid == NULL)
212 		return got_error_from_errno();
213 
214 	qid->id = got_object_id_dup(id);
215 	if (qid->id == NULL) {
216 		const struct got_error *err = got_error_from_errno();
217 		free(qid);
218 		return err;
219 	}
220 
221 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
222 	return NULL;
223 }
224 
225 static const struct got_error *
226 advance_open_branches(struct got_commit_graph *graph,
227     struct got_commit_graph_node *node,
228     struct got_object_id *commit_id, struct got_commit_object *commit)
229 {
230 	const struct got_error *err;
231 	struct got_object_qid *qid;
232 
233 	err = got_object_idset_remove(graph->open_branches, commit_id);
234 	if (err && err->code != GOT_ERR_NO_OBJ)
235 		return err;
236 
237 	if (graph->flags & GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL) {
238 		qid = SIMPLEQ_FIRST(&commit->parent_ids);
239 		if (qid == NULL)
240 			return NULL;
241 		err = got_object_idset_add(NULL, graph->open_branches, qid->id,
242 		    node);
243 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
244 			return err;
245 		return NULL;
246 	}
247 
248 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
249 		if (got_object_idset_get(graph->node_ids, qid->id))
250 			continue; /* parent already traversed */
251 		err = got_object_idset_add(NULL, graph->open_branches,
252 		    qid->id, node);
253 		if (err && err->code != GOT_ERR_OBJ_EXISTS)
254 			return err;
255 	}
256 
257 	return NULL;
258 }
259 
260 static const struct got_error *
261 add_node(struct got_commit_graph_node **new_node,
262     struct got_commit_graph *graph, struct got_object_id *commit_id,
263     struct got_commit_object *commit, struct got_commit_graph_node *child_node)
264 {
265 	const struct got_error *err = NULL;
266 	struct got_commit_graph_node *node, *existing_node;
267 	struct got_object_qid *qid;
268 
269 	*new_node = NULL;
270 
271 	node = calloc(1, sizeof(*node));
272 	if (node == NULL)
273 		return got_error_from_errno();
274 
275 	memcpy(&node->id, commit_id, sizeof(node->id));
276 	SIMPLEQ_INIT(&node->parent_ids);
277 	SIMPLEQ_INIT(&node->child_ids);
278 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
279 		err = add_vertex(&node->parent_ids, qid->id);
280 		if (err)
281 			return err;
282 		node->nparents++;
283 	}
284 	node->commit_timestamp = commit->committer_time; /* XXX not UTC! */
285 
286 	err = got_object_idset_add((void **)(&existing_node),
287 	    graph->node_ids, &node->id, node);
288 	if (err == NULL) {
289 		add_node_to_iter_list(graph, node, child_node);
290 		err = advance_open_branches(graph, node, commit_id, commit);
291 		*new_node = node;
292 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
293 		err = NULL;
294 		free(node);
295 		node = existing_node;
296 	} else {
297 		free(node);
298 		return err;
299 	}
300 
301 	if (child_node) {
302 		struct got_object_qid *cid;
303 
304 		/* Prevent linking to self. */
305 		if (got_object_id_cmp(commit_id, &child_node->id) == 0)
306 			return got_error(GOT_ERR_BAD_OBJ_ID);
307 
308 		/* Prevent double-linking to the same child. */
309 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
310 			if (got_object_id_cmp(cid->id, &child_node->id) == 0)
311 				return got_error(GOT_ERR_BAD_OBJ_ID);
312 		}
313 
314 		err = add_vertex(&node->child_ids, &child_node->id);
315 		if (err)
316 			return err;
317 		node->nchildren++;
318 	}
319 
320 	return err;
321 }
322 
323 const struct got_error *
324 got_commit_graph_open(struct got_commit_graph **graph,
325     struct got_object_id *commit_id, int first_parent_traversal,
326     struct got_repository *repo)
327 {
328 	const struct got_error *err = NULL;
329 	struct got_commit_object *commit;
330 
331 	*graph = NULL;
332 
333 	err = got_object_open_as_commit(&commit, repo, commit_id);
334 	if (err)
335 		return err;
336 
337 	*graph = alloc_graph();
338 	if (*graph == NULL) {
339 		got_object_commit_close(commit);
340 		return got_error_from_errno();
341 	}
342 
343 	if (first_parent_traversal)
344 		(*graph)->flags |= GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL;
345 
346 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
347 	got_object_commit_close(commit);
348 	if (err) {
349 		got_commit_graph_close(*graph);
350 		*graph = NULL;
351 		return err;
352 	}
353 	
354 	return NULL;
355 }
356 
357 struct got_commit_graph_branch {
358 	struct got_object_id parent_id;
359 	struct got_commit_graph_node *node;
360 };
361 
362 struct gather_branches_arg {
363 	struct got_commit_graph_branch *branches;
364 	int nbranches;
365 };
366 
367 static void
368 gather_branches(struct got_object_id *id, void *data, void *arg)
369 {
370 	struct gather_branches_arg *a = arg;
371 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
372 	a->branches[a->nbranches].node = data;
373 	a->nbranches++;
374 }
375 
376 static const struct got_error *
377 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
378     struct got_commit_graph *graph, struct got_repository *repo,
379     struct got_object_id *wanted_id)
380 {
381 	const struct got_error *err;
382 	struct got_commit_graph_branch *branches;
383 	struct gather_branches_arg arg;
384 	int i;
385 
386 	*ncommits = 0;
387 	if (wanted_id_added)
388 		*wanted_id_added = 0;
389 
390 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
391 	if (arg.nbranches == 0)
392 		return NULL;
393 
394 	/*
395 	 * Adding nodes to the graph might change the graph's open
396 	 * branches state. Create a local copy of the current state.
397 	 */
398 	branches = calloc(arg.nbranches, sizeof(*branches));
399 	if (branches == NULL)
400 		return got_error_from_errno();
401 	arg.nbranches = 0; /* reset; gather_branches() will increment */
402 	arg.branches = branches;
403 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
404 
405 	for (i = 0; i < arg.nbranches; i++) {
406 		struct got_object_id *commit_id;
407 		struct got_commit_graph_node *child_node, *new_node;
408 		struct got_commit_object *commit;
409 
410 		commit_id = &branches[i].parent_id;
411 		child_node = branches[i].node;
412 
413 		err = got_object_open_as_commit(&commit, repo, commit_id);
414 		if (err)
415 			break;
416 
417 		err = add_node(&new_node, graph, commit_id, commit, child_node);
418 		got_object_commit_close(commit);
419 		if (err) {
420 			if (err->code != GOT_ERR_OBJ_EXISTS)
421 				break;
422 			err = NULL;
423 		}
424 		if (new_node)
425 			(*ncommits)++;
426 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
427 			*wanted_id_added = 1;
428 	}
429 
430 	free(branches);
431 	return err;
432 }
433 
434 const struct got_error *
435 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
436     int limit, struct got_repository *repo)
437 {
438 	const struct got_error *err;
439 	int ncommits;
440 
441 	*nfetched = 0;
442 
443 	while (*nfetched < limit) {
444 		err = fetch_commits_from_open_branches(&ncommits, NULL,
445 		    graph, repo, NULL);
446 		if (err)
447 			return err;
448 		if (ncommits == 0)
449 			break;
450 		*nfetched += ncommits;
451 	}
452 
453 	return NULL;
454 }
455 
456 const struct got_error *
457 got_commit_graph_fetch_commits_up_to(int *nfetched,
458     struct got_commit_graph *graph, struct got_object_id *wanted_id,
459     struct got_repository *repo)
460 {
461 	const struct got_error *err;
462 	int ncommits, wanted_id_added = 0;
463 
464 	*nfetched = 0;
465 
466 	if (got_object_idset_get(graph->node_ids, wanted_id) != NULL)
467 		return NULL;
468 
469 	while (!wanted_id_added) {
470 		err = fetch_commits_from_open_branches(&ncommits,
471 		    &wanted_id_added, graph, repo, wanted_id);
472 		if (err)
473 			return err;
474 		if (ncommits == 0)
475 			return NULL;
476 		*nfetched += ncommits;
477 	}
478 
479 	return NULL;
480 }
481 
482 static void
483 free_graph_node(struct got_object_id *id, void *data, void *arg)
484 {
485 	struct got_commit_graph_node *node = data;
486 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
487 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
488 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
489 		free(child);
490 	}
491 	free(node);
492 }
493 
494 void
495 got_commit_graph_close(struct got_commit_graph *graph)
496 {
497 	got_object_idset_free(graph->open_branches);
498 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
499 	got_object_idset_free(graph->node_ids);
500 	free(graph);
501 }
502 
503 const struct got_error *
504 got_commit_graph_iter_start(struct got_commit_graph *graph,
505     struct got_object_id *id)
506 {
507 	struct got_commit_graph_node *start_node;
508 
509 	start_node = got_object_idset_get(graph->node_ids, id);
510 	if (start_node == NULL)
511 		return got_error(GOT_ERR_NO_OBJ);
512 
513 	graph->iter_node = start_node;
514 	return NULL;
515 }
516 
517 const struct got_error *
518 got_commit_graph_iter_next(struct got_object_id **id,
519     struct got_commit_graph *graph)
520 {
521 	*id = NULL;
522 
523 	if (graph->iter_node == NULL) {
524 		/* We are done interating, or iteration was not started. */
525 		return got_error(GOT_ERR_ITER_COMPLETED);
526 	}
527 
528 	if (graph->iter_node ==
529 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
530 	    got_object_idset_num_elements(graph->open_branches) == 0) {
531 		*id = &graph->iter_node->id;
532 		/* We are done interating. */
533 		graph->iter_node = NULL;
534 		return NULL;
535 	}
536 
537 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
538 		return got_error(GOT_ERR_ITER_NEED_MORE);
539 
540 	*id = &graph->iter_node->id;
541 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
542 	return NULL;
543 }