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(NULL, 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 = mktime(&commit->tm_committer); 
285 	if (node->commit_timestamp == -1)
286 		return got_error_from_errno();
287 
288 	err = got_object_idset_add((void **)(&existing_node),
289 	    graph->node_ids, &node->id, node);
290 	if (err == NULL) {
291 		add_node_to_iter_list(graph, node, child_node);
292 		err = advance_open_branches(graph, node, commit_id, commit);
293 		*new_node = node;
294 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
295 		err = NULL;
296 		free(node);
297 		node = existing_node;
298 	} else {
299 		free(node);
300 		return err;
301 	}
302 
303 	if (child_node) {
304 		struct got_object_qid *cid;
305 
306 		/* Prevent linking to self. */
307 		if (got_object_id_cmp(commit_id, &child_node->id) == 0)
308 			return got_error(GOT_ERR_BAD_OBJ_ID);
309 
310 		/* Prevent double-linking to the same child. */
311 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
312 			if (got_object_id_cmp(cid->id, &child_node->id) == 0)
313 				return got_error(GOT_ERR_BAD_OBJ_ID);
314 		}
315 
316 		err = add_vertex(&node->child_ids, &child_node->id);
317 		if (err)
318 			return err;
319 		node->nchildren++;
320 	}
321 
322 	return err;
323 }
324 
325 const struct got_error *
326 got_commit_graph_open(struct got_commit_graph **graph,
327     struct got_object_id *commit_id, int first_parent_traversal,
328     struct got_repository *repo)
329 {
330 	const struct got_error *err = NULL;
331 	struct got_commit_object *commit;
332 
333 	*graph = NULL;
334 
335 	err = got_object_open_as_commit(&commit, repo, commit_id);
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 	if (first_parent_traversal)
346 		(*graph)->flags |= GOT_COMMIT_GRAPH_FIRST_PARENT_TRAVERSAL;
347 
348 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
349 	got_object_commit_close(commit);
350 	if (err) {
351 		got_commit_graph_close(*graph);
352 		*graph = NULL;
353 		return err;
354 	}
355 	
356 	return NULL;
357 }
358 
359 struct got_commit_graph_branch {
360 	struct got_object_id parent_id;
361 	struct got_commit_graph_node *node;
362 };
363 
364 struct gather_branches_arg {
365 	struct got_commit_graph_branch *branches;
366 	int nbranches;
367 };
368 
369 static void
370 gather_branches(struct got_object_id *id, void *data, void *arg)
371 {
372 	struct gather_branches_arg *a = arg;
373 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
374 	a->branches[a->nbranches].node = data;
375 	a->nbranches++;
376 }
377 
378 static const struct got_error *
379 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
380     struct got_commit_graph *graph, struct got_repository *repo,
381     struct got_object_id *wanted_id)
382 {
383 	const struct got_error *err;
384 	struct got_commit_graph_branch *branches;
385 	struct gather_branches_arg arg;
386 	int i;
387 
388 	*ncommits = 0;
389 	if (wanted_id_added)
390 		*wanted_id_added = 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 = got_object_open_as_commit(&commit, repo, commit_id);
416 		if (err)
417 			break;
418 
419 		err = add_node(&new_node, graph, commit_id, commit, child_node);
420 		got_object_commit_close(commit);
421 		if (err) {
422 			if (err->code != GOT_ERR_OBJ_EXISTS)
423 				break;
424 			err = NULL;
425 		}
426 		if (new_node)
427 			(*ncommits)++;
428 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
429 			*wanted_id_added = 1;
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 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_graph_node(struct got_object_id *id, void *data, void *arg)
486 {
487 	struct got_commit_graph_node *node = data;
488 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
489 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
490 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
491 		free(child);
492 	}
493 	free(node);
494 }
495 
496 void
497 got_commit_graph_close(struct got_commit_graph *graph)
498 {
499 	got_object_idset_free(graph->open_branches);
500 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
501 	got_object_idset_free(graph->node_ids);
502 	free(graph);
503 }
504 
505 const struct got_error *
506 got_commit_graph_iter_start(struct got_commit_graph *graph,
507     struct got_object_id *id)
508 {
509 	struct got_commit_graph_node *start_node;
510 
511 	start_node = got_object_idset_get(graph->node_ids, id);
512 	if (start_node == NULL)
513 		return got_error(GOT_ERR_NO_OBJ);
514 
515 	graph->iter_node = start_node;
516 	return NULL;
517 }
518 
519 const struct got_error *
520 got_commit_graph_iter_next(struct got_object_id **id,
521     struct got_commit_graph *graph)
522 {
523 	*id = NULL;
524 
525 	if (graph->iter_node == NULL) {
526 		/* We are done interating, or iteration was not started. */
527 		return got_error(GOT_ERR_ITER_COMPLETED);
528 	}
529 
530 	if (graph->iter_node ==
531 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
532 	    got_object_idset_num_elements(graph->open_branches) == 0) {
533 		*id = &graph->iter_node->id;
534 		/* We are done interating. */
535 		graph->iter_node = NULL;
536 		return NULL;
537 	}
538 
539 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
540 		return got_error(GOT_ERR_ITER_NEED_MORE);
541 
542 	*id = &graph->iter_node->id;
543 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
544 	return NULL;
545 }