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 compare_commits(int *cmp, struct got_commit_object *c1,
139     struct got_commit_object *c2)
140 {
141 	const struct got_error *err;
142 	int64_t t1, t2;
143 
144 	err = got_object_commit_get_committer_time(&t1, c1);
145 	if (err)
146 		return err;
147 	err = got_object_commit_get_committer_time(&t2, c2);
148 
149 	if (err)
150 		return err;
151 
152 	if (t1 == t2)	
153 		*cmp = 0;
154 	else if (t1 < t2)
155 		*cmp = -1;
156 	else
157 		*cmp = 1;
158 
159 	return NULL;
160 }
161 
162 static const struct got_error *
163 add_iteration_candidate(struct got_commit_graph *graph,
164     struct got_commit_graph_node *node)
165 {
166 	struct got_commit_graph_node *n, *next;
167 	
168 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
169 		TAILQ_INSERT_TAIL(&graph->iter_candidates, node, entry);
170 		return NULL;
171 	}
172 
173 	TAILQ_FOREACH(n, &graph->iter_candidates, entry) {
174 		const struct got_error *err;
175 		int cmp;
176 		err = compare_commits(&cmp, node->commit, n->commit);
177 		if (err)
178 			return err;
179 		if (cmp < 0) {
180 			next = TAILQ_NEXT(n, entry);
181 			if (next == NULL) {
182 				TAILQ_INSERT_AFTER(&graph->iter_candidates, n,
183 				    node, entry);
184 				break;
185 			}
186 			err = compare_commits(&cmp, node->commit, next->commit);
187 			if (err)
188 				return err;
189 			if (cmp >= 0) {
190 				TAILQ_INSERT_BEFORE(next, node, entry);
191 				break;
192 			}
193 		} else {
194 			TAILQ_INSERT_BEFORE(n, node, entry);
195 			break;
196 		}
197 	}
198 
199 	return NULL;
200 }
201 
202 static const struct got_error *
203 add_node(struct got_commit_graph_node **new_node,
204     struct got_commit_graph *graph, struct got_object_id *commit_id,
205     struct got_commit_object *commit, struct got_object_id *child_commit_id)
206 {
207 	const struct got_error *err = NULL;
208 	struct got_commit_graph_node *node, *existing_node;
209 
210 	*new_node = NULL;
211 
212 	node = calloc(1, sizeof(*node));
213 	if (node == NULL)
214 		return got_error_from_errno();
215 
216 	memcpy(&node->id, commit_id, sizeof(node->id));
217 	node->commit = commit;
218 	SIMPLEQ_INIT(&node->child_ids);
219 
220 	err = got_object_idset_add((void **)(&existing_node),
221 	    graph->node_ids, &node->id, node);
222 	if (err == NULL) {
223 		struct got_parent_id *pid;
224 
225 		err = add_iteration_candidate(graph, node);
226 		if (err)
227 			return err;
228 
229 		err = got_object_idset_remove(graph->open_branches, commit_id);
230 		if (err && err->code != GOT_ERR_NO_OBJ)
231 			return err;
232 		SIMPLEQ_FOREACH(pid, &commit->parent_ids, entry) {
233 			if (got_object_idset_get(graph->node_ids, pid->id))
234 				continue; /* parent already traversed */
235 			err = got_object_idset_add(NULL, graph->open_branches,
236 			    pid->id, node);
237 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
238 				return err;
239 		}
240 		*new_node = node;
241 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
242 		err = NULL;
243 		free(node);
244 		node = existing_node;
245 	} else {
246 		free(node);
247 		return err;
248 	}
249 
250 	if (child_commit_id) {
251 		struct got_parent_id *child, *cid;
252 
253 		/* Prevent linking to self. */
254 		if (got_object_id_cmp(commit_id, child_commit_id) == 0)
255 			return got_error(GOT_ERR_BAD_OBJ_ID);
256 
257 		/* Prevent double-linking to the same child. */
258 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
259 			if (got_object_id_cmp(cid->id, child_commit_id) == 0)
260 				return got_error(GOT_ERR_BAD_OBJ_ID);
261 		}
262 
263 		child = calloc(1, sizeof(*child));
264 		if (child == NULL)
265 			return got_error_from_errno();
266 		child->id = got_object_id_dup(child_commit_id);
267 		if (child->id == NULL) {
268 			err = got_error_from_errno();
269 			free(child);
270 			return err;
271 		}
272 		SIMPLEQ_INSERT_TAIL(&node->child_ids, child, entry);
273 		node->nchildren++;
274 	}
275 
276 	return err;
277 }
278 
279 const struct got_error *
280 got_commit_graph_open(struct got_commit_graph **graph,
281     struct got_object_id *commit_id, struct got_repository *repo)
282 {
283 	const struct got_error *err = NULL;
284 	struct got_object *obj;
285 	struct got_commit_object *commit;
286 
287 	*graph = NULL;
288 
289 	err = got_object_open(&obj, repo, commit_id);
290 	if (err)
291 		return err;
292 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
293 		err = got_error(GOT_ERR_OBJ_TYPE);
294 		got_object_close(obj);
295 		return err;
296 	}
297 
298 	err = got_object_commit_open(&commit, repo, obj);
299 	got_object_close(obj);
300 	if (err)
301 		return err;
302 
303 	*graph = alloc_graph();
304 	if (*graph == NULL) {
305 		got_object_commit_close(commit);
306 		return got_error_from_errno();
307 	}
308 
309 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
310 	if (err) {
311 		got_commit_graph_close(*graph);
312 		*graph = NULL;
313 		return err;
314 	}
315 	
316 	return NULL;
317 }
318 
319 static const struct got_error *
320 open_commit(struct got_commit_object **commit, struct got_object_id *id,
321     struct got_repository *repo)
322 {
323 	const struct got_error *err;
324 	struct got_object *obj;
325 
326 	err = got_object_open(&obj, repo, id);
327 	if (err)
328 		return err;
329 	if (got_object_get_type(obj) != GOT_OBJ_TYPE_COMMIT) {
330 		err = got_error(GOT_ERR_OBJ_TYPE);
331 		goto done;
332 	}
333 
334 	err = got_object_commit_open(commit, repo, obj);
335 done:
336 	got_object_close(obj);
337 	return err;
338 }
339 
340 struct got_commit_graph_branch {
341 	struct got_object_id parent_id;
342 	struct got_commit_graph_node *node;
343 };
344 
345 struct gather_branches_arg {
346 	struct got_commit_graph_branch *branches;
347 	int nbranches;
348 };
349 
350 static void
351 gather_branches(struct got_object_id *id, void *data, void *arg)
352 {
353 	struct gather_branches_arg *a = arg;
354 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
355 	a->branches[a->nbranches].node = data;
356 	a->nbranches++;
357 }
358 
359 const struct got_error *
360 fetch_commits_from_open_branches(int *ncommits,
361     struct got_commit_graph *graph, struct got_repository *repo)
362 {
363 	const struct got_error *err;
364 	struct got_commit_graph_branch *branches;
365 	struct gather_branches_arg arg;
366 	int i;
367 
368 	*ncommits = 0;
369 
370 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
371 	if (arg.nbranches == 0)
372 		return NULL;
373 
374 	/*
375 	 * Adding nodes to the graph might change the graph's open
376 	 * branches state. Create a local copy of the current state.
377 	 */
378 	branches = calloc(arg.nbranches, sizeof(*branches));
379 	if (branches == NULL)
380 		return got_error_from_errno();
381 	arg.nbranches = 0; /* reset; gather_branches() will increment */
382 	arg.branches = branches;
383 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
384 
385 	for (i = 0; i < arg.nbranches; i++) {
386 		struct got_object_id *commit_id;
387 		struct got_commit_graph_node *child_node, *new_node;
388 		struct got_commit_object *commit;
389 
390 		commit_id = &branches[i].parent_id;
391 		child_node = branches[i].node;
392 
393 		err = open_commit(&commit, commit_id, repo);
394 		if (err)
395 			break;
396 
397 		err = add_node(&new_node, graph, commit_id, commit,
398 		    &child_node->id);
399 		if (err) {
400 			if (err->code != GOT_ERR_OBJ_EXISTS)
401 				break;
402 			err = NULL;
403 		}
404 		if (new_node)
405 			(*ncommits)++;
406 		else
407 			got_object_commit_close(commit);
408 	}
409 
410 	free(branches);
411 	return err;
412 }
413 
414 const struct got_error *
415 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
416     int limit, struct got_repository *repo)
417 {
418 	const struct got_error *err;
419 	int total = 0, ncommits;
420 
421 	*nfetched = 0;
422 
423 	while (total < limit) {
424 		err = fetch_commits_from_open_branches(&ncommits, graph, repo);
425 		if (err)
426 			return err;
427 		if (ncommits == 0)
428 			break;
429 		total += ncommits;
430 	}
431 
432 	*nfetched = total;
433 	return NULL;
434 }
435 
436 static void
437 free_graph_node(struct got_object_id *id, void *data, void *arg)
438 {
439 	struct got_commit_graph_node *node = data;
440 	got_object_commit_close(node->commit);
441 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
442 		struct got_parent_id *child = SIMPLEQ_FIRST(&node->child_ids);
443 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
444 		free(child);
445 	}
446 	free(node);
447 }
448 
449 void
450 got_commit_graph_close(struct got_commit_graph *graph)
451 {
452 	got_object_idset_free(graph->open_branches);
453 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
454 	got_object_idset_free(graph->node_ids);
455 	free(graph);
456 }
457 
458 const struct got_error *
459 got_commit_graph_iter_start(struct got_commit_graph *graph,
460     struct got_object_id *id)
461 {
462 	struct got_commit_graph_node *start_node, *node;
463 	struct got_parent_id *pid;
464 
465 	start_node = got_object_idset_get(graph->node_ids, id);
466 	if (start_node == NULL)
467 		return got_error(GOT_ERR_NO_OBJ);
468 
469 	graph->iter_node = start_node;
470 
471 	while (!TAILQ_EMPTY(&graph->iter_candidates)) {
472 		node = TAILQ_FIRST(&graph->iter_candidates);
473 		TAILQ_REMOVE(&graph->iter_candidates, node, entry);
474 	}
475 
476 	/* Put all known parents of this commit on the candidate list. */
477 	SIMPLEQ_FOREACH(pid, &start_node->commit->parent_ids, entry) {
478 		node = got_object_idset_get(graph->node_ids, pid->id);
479 		if (node) {
480 			const struct got_error *err;
481 			err = add_iteration_candidate(graph, node);
482 			if (err)
483 				return err;
484 		}
485 	}
486 
487 	return NULL;
488 }
489 
490 const struct got_error *
491 got_commit_graph_iter_next(struct got_commit_object **commit,
492     struct got_object_id **id, struct got_commit_graph *graph)
493 {
494 	struct got_commit_graph_node *node;
495 
496 	if (graph->iter_node == NULL) {
497 		/* We are done interating, or iteration was not started. */
498 		*commit = NULL;
499 		*id = NULL;
500 		return NULL;
501 	}
502 
503 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
504 		if (is_root_node(graph->iter_node) &&
505 		    got_object_idset_num_elements(graph->open_branches) == 0) {
506 			*commit = graph->iter_node->commit;
507 			*id = &graph->iter_node->id;
508 			/* We are done interating. */
509 			graph->iter_node = NULL;
510 			return NULL;
511 		}
512 		return got_error(GOT_ERR_ITER_NEED_MORE);
513 	}
514 
515 	*commit = graph->iter_node->commit;
516 	*id = &graph->iter_node->id;
517 	node = TAILQ_FIRST(&graph->iter_candidates);
518 	TAILQ_REMOVE(&graph->iter_candidates, node, entry);
519 	graph->iter_node = node;
520 	return NULL;
521 }
522 
523 const struct got_commit_object *
524 got_commit_graph_get_commit(struct got_commit_graph *graph,
525     struct got_object_id *id)
526 {
527 	struct got_commit_graph_node *node;
528 	node = got_object_idset_get(graph->node_ids, id);
529 	return node ? node->commit : NULL;
530 }