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 	/*
68 	 * A set of object IDs of known parent commits which we have not yet
69 	 * traversed. Each commit ID in this set represents a branch in commit
70 	 * history: Either the first-parent branch of the head node, or another
71 	 * branch corresponding to a traversed merge commit for which we have
72 	 * not traversed a branch point commit yet.
73 	 *
74 	 * Whenever we add a commit with a matching ID to the graph, we remove
75 	 * its corresponding element from this set, and add new elements for
76 	 * each of that commit's parent commits which were not traversed yet.
77 	 *
78 	 * When API users ask us to fetch more commits, we fetch commits from
79 	 * all currently open branches. This allows API users to process
80 	 * commits in linear order even though the history contains branches.
81 	 */
82 	struct got_object_idset *open_branches;
83 
84 	/* The next commit to return when the API user asks for one. */
85 	struct got_commit_graph_node *iter_node;
86 
87 	/* The graph iteration list contains all nodes in sorted order. */
88 	struct got_commit_graph_iter_list iter_list;
89 };
90 
91 static struct got_commit_graph *
92 alloc_graph(void)
93 {
94 	struct got_commit_graph *graph;
95 
96 	graph = calloc(1, sizeof(*graph));
97 	if (graph == NULL)
98 		return NULL;
99 
100 	graph->node_ids = got_object_idset_alloc();
101 	if (graph->node_ids == NULL) {
102 		free(graph);
103 		return NULL;
104 	}
105 
106 	graph->open_branches = got_object_idset_alloc();
107 	if (graph->open_branches == NULL) {
108 		got_object_idset_free(graph->node_ids);
109 		free(graph);
110 		return NULL;
111 	}
112 
113 	TAILQ_INIT(&graph->iter_list);
114 	return graph;
115 }
116 
117 #if 0
118 static int
119 is_head_node(struct got_commit_graph_node *node)
120 {
121 	return node->nchildren == 0;
122 }
123 
124 static int
125 is_merge_point(struct got_commit_graph_node *node)
126 {
127 	return node->nparents > 1;
128 }
129 
130 int
131 is_branch_point(struct got_commit_graph_node *node)
132 {
133 	return node->nchildren > 1;
134 }
135 
136 static int
137 is_root_node(struct got_commit_graph_node *node)
138 {
139 	return node->nparents == 0;
140 }
141 #endif
142 
143 static void
144 add_node_to_iter_list(struct got_commit_graph *graph,
145     struct got_commit_graph_node *node,
146     struct got_commit_graph_node *child_node)
147 {
148 	struct got_commit_graph_node *n, *next;
149 	
150 	if (TAILQ_EMPTY(&graph->iter_list)) {
151 		TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
152 		return;
153 	}
154 
155 	/*
156 	 * If a child node is known, begin looping over the list there
157 	 * instead of beginning from the list head. Parent commits are
158 	 * traversed before their children.
159 	 */
160 	n = child_node ? child_node : TAILQ_FIRST(&graph->iter_list);
161 
162 	/* Ensure that an iteration in progress will see this new commit. */
163 	if (graph->iter_node) {
164 		n = graph->iter_node;
165 		do {
166 			next = TAILQ_NEXT(n, entry);
167 			if (next &&
168 			    node->commit_timestamp >= next->commit_timestamp) {
169 				TAILQ_INSERT_BEFORE(next, node, entry);
170 				return;
171 			}
172 			n = next;
173 		} while (next);
174 
175 		TAILQ_INSERT_AFTER(&graph->iter_list, graph->iter_node,
176 		    node, entry);
177 		return;
178 	}
179 
180 	/* Insert into list based on committer timestamp. */
181 	do {
182 		if (node->commit_timestamp == n->commit_timestamp) {
183 			TAILQ_INSERT_AFTER(&graph->iter_list, n, node, entry);
184 			break;
185 		} else if (node->commit_timestamp < n->commit_timestamp) {
186 			next = TAILQ_NEXT(n, entry);
187 			if (next == NULL) {
188 				TAILQ_INSERT_AFTER(&graph->iter_list, n,
189 				    node, entry);
190 				break;
191 			}
192 			if (node->commit_timestamp >= next->commit_timestamp) {
193 				TAILQ_INSERT_BEFORE(next, node, entry);
194 				break;
195 			}
196 		} else {
197 			TAILQ_INSERT_BEFORE(n, node, entry);
198 			break;
199 		}
200 		n = TAILQ_NEXT(n, entry);
201 	} while (n);
202 }
203 
204 static const struct got_error *
205 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
206 {
207 	struct got_object_qid *qid;
208 
209 	qid = calloc(1, sizeof(*qid));
210 	if (qid == NULL)
211 		return got_error_from_errno();
212 
213 	qid->id = got_object_id_dup(id);
214 	if (qid->id == NULL) {
215 		const struct got_error *err = got_error_from_errno();
216 		free(qid);
217 		return err;
218 	}
219 
220 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
221 	return NULL;
222 }
223 
224 static const struct got_error *
225 add_node(struct got_commit_graph_node **new_node,
226     struct got_commit_graph *graph, struct got_object_id *commit_id,
227     struct got_commit_object *commit, struct got_commit_graph_node *child_node)
228 {
229 	const struct got_error *err = NULL;
230 	struct got_commit_graph_node *node, *existing_node;
231 	struct got_object_qid *qid;
232 
233 	*new_node = NULL;
234 
235 	node = calloc(1, sizeof(*node));
236 	if (node == NULL)
237 		return got_error_from_errno();
238 
239 	memcpy(&node->id, commit_id, sizeof(node->id));
240 	SIMPLEQ_INIT(&node->parent_ids);
241 	SIMPLEQ_INIT(&node->child_ids);
242 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
243 		err = add_vertex(&node->parent_ids, qid->id);
244 		if (err)
245 			return err;
246 		node->nparents++;
247 	}
248 	node->commit_timestamp = commit->committer_time; /* XXX not UTC! */
249 
250 	err = got_object_idset_add((void **)(&existing_node),
251 	    graph->node_ids, &node->id, node);
252 	if (err == NULL) {
253 		struct got_object_qid *qid;
254 
255 		add_node_to_iter_list(graph, node, child_node);
256 		err = got_object_idset_remove(graph->open_branches, commit_id);
257 		if (err && err->code != GOT_ERR_NO_OBJ)
258 			return err;
259 		SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
260 			if (got_object_idset_get(graph->node_ids, qid->id))
261 				continue; /* parent already traversed */
262 			err = got_object_idset_add(NULL, graph->open_branches,
263 			    qid->id, node);
264 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
265 				return err;
266 		}
267 		*new_node = node;
268 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
269 		err = NULL;
270 		free(node);
271 		node = existing_node;
272 	} else {
273 		free(node);
274 		return err;
275 	}
276 
277 	if (child_node) {
278 		struct got_object_qid *cid;
279 
280 		/* Prevent linking to self. */
281 		if (got_object_id_cmp(commit_id, &child_node->id) == 0)
282 			return got_error(GOT_ERR_BAD_OBJ_ID);
283 
284 		/* Prevent double-linking to the same child. */
285 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
286 			if (got_object_id_cmp(cid->id, &child_node->id) == 0)
287 				return got_error(GOT_ERR_BAD_OBJ_ID);
288 		}
289 
290 		err = add_vertex(&node->child_ids, &child_node->id);
291 		if (err)
292 			return err;
293 		node->nchildren++;
294 	}
295 
296 	return err;
297 }
298 
299 const struct got_error *
300 got_commit_graph_open(struct got_commit_graph **graph,
301     struct got_object_id *commit_id, struct got_repository *repo)
302 {
303 	const struct got_error *err = NULL;
304 	struct got_commit_object *commit;
305 
306 	*graph = NULL;
307 
308 	err = got_object_open_as_commit(&commit, repo, commit_id);
309 	if (err)
310 		return err;
311 
312 	*graph = alloc_graph();
313 	if (*graph == NULL) {
314 		got_object_commit_close(commit);
315 		return got_error_from_errno();
316 	}
317 
318 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
319 	got_object_commit_close(commit);
320 	if (err) {
321 		got_commit_graph_close(*graph);
322 		*graph = NULL;
323 		return err;
324 	}
325 	
326 	return NULL;
327 }
328 
329 struct got_commit_graph_branch {
330 	struct got_object_id parent_id;
331 	struct got_commit_graph_node *node;
332 };
333 
334 struct gather_branches_arg {
335 	struct got_commit_graph_branch *branches;
336 	int nbranches;
337 };
338 
339 static void
340 gather_branches(struct got_object_id *id, void *data, void *arg)
341 {
342 	struct gather_branches_arg *a = arg;
343 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
344 	a->branches[a->nbranches].node = data;
345 	a->nbranches++;
346 }
347 
348 static const struct got_error *
349 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
350     struct got_commit_graph *graph, struct got_repository *repo,
351     struct got_object_id *wanted_id)
352 {
353 	const struct got_error *err;
354 	struct got_commit_graph_branch *branches;
355 	struct gather_branches_arg arg;
356 	int i;
357 
358 	*ncommits = 0;
359 	if (wanted_id_added)
360 		*wanted_id_added = 0;
361 
362 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
363 	if (arg.nbranches == 0)
364 		return NULL;
365 
366 	/*
367 	 * Adding nodes to the graph might change the graph's open
368 	 * branches state. Create a local copy of the current state.
369 	 */
370 	branches = calloc(arg.nbranches, sizeof(*branches));
371 	if (branches == NULL)
372 		return got_error_from_errno();
373 	arg.nbranches = 0; /* reset; gather_branches() will increment */
374 	arg.branches = branches;
375 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
376 
377 	for (i = 0; i < arg.nbranches; i++) {
378 		struct got_object_id *commit_id;
379 		struct got_commit_graph_node *child_node, *new_node;
380 		struct got_commit_object *commit;
381 
382 		commit_id = &branches[i].parent_id;
383 		child_node = branches[i].node;
384 
385 		err = got_object_open_as_commit(&commit, repo, commit_id);
386 		if (err)
387 			break;
388 
389 		err = add_node(&new_node, graph, commit_id, commit, child_node);
390 		got_object_commit_close(commit);
391 		if (err) {
392 			if (err->code != GOT_ERR_OBJ_EXISTS)
393 				break;
394 			err = NULL;
395 		}
396 		if (new_node)
397 			(*ncommits)++;
398 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
399 			*wanted_id_added = 1;
400 	}
401 
402 	free(branches);
403 	return err;
404 }
405 
406 const struct got_error *
407 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
408     int limit, struct got_repository *repo)
409 {
410 	const struct got_error *err;
411 	int ncommits;
412 
413 	*nfetched = 0;
414 
415 	while (*nfetched < limit) {
416 		err = fetch_commits_from_open_branches(&ncommits, NULL,
417 		    graph, repo, NULL);
418 		if (err)
419 			return err;
420 		if (ncommits == 0)
421 			break;
422 		*nfetched += ncommits;
423 	}
424 
425 	return NULL;
426 }
427 
428 const struct got_error *
429 got_commit_graph_fetch_commits_up_to(int *nfetched,
430     struct got_commit_graph *graph, struct got_object_id *wanted_id,
431     struct got_repository *repo)
432 {
433 	const struct got_error *err;
434 	int ncommits, wanted_id_added = 0;
435 
436 	*nfetched = 0;
437 
438 	if (got_object_idset_get(graph->node_ids, wanted_id) != NULL)
439 		return NULL;
440 
441 	while (!wanted_id_added) {
442 		err = fetch_commits_from_open_branches(&ncommits,
443 		    &wanted_id_added, graph, repo, wanted_id);
444 		if (err)
445 			return err;
446 		if (ncommits == 0)
447 			return NULL;
448 		*nfetched += ncommits;
449 	}
450 
451 	return NULL;
452 }
453 
454 static void
455 free_graph_node(struct got_object_id *id, void *data, void *arg)
456 {
457 	struct got_commit_graph_node *node = data;
458 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
459 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
460 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
461 		free(child);
462 	}
463 	free(node);
464 }
465 
466 void
467 got_commit_graph_close(struct got_commit_graph *graph)
468 {
469 	got_object_idset_free(graph->open_branches);
470 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
471 	got_object_idset_free(graph->node_ids);
472 	free(graph);
473 }
474 
475 const struct got_error *
476 got_commit_graph_iter_start(struct got_commit_graph *graph,
477     struct got_object_id *id)
478 {
479 	struct got_commit_graph_node *start_node;
480 
481 	start_node = got_object_idset_get(graph->node_ids, id);
482 	if (start_node == NULL)
483 		return got_error(GOT_ERR_NO_OBJ);
484 
485 	graph->iter_node = start_node;
486 	return NULL;
487 }
488 
489 const struct got_error *
490 got_commit_graph_iter_next(struct got_object_id **id,
491     struct got_commit_graph *graph)
492 {
493 	*id = NULL;
494 
495 	if (graph->iter_node == NULL) {
496 		/* We are done interating, or iteration was not started. */
497 		return got_error(GOT_ERR_ITER_COMPLETED);
498 	}
499 
500 	if (graph->iter_node ==
501 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
502 	    got_object_idset_num_elements(graph->open_branches) == 0) {
503 		*id = &graph->iter_node->id;
504 		/* We are done interating. */
505 		graph->iter_node = NULL;
506 		return NULL;
507 	}
508 
509 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
510 		return got_error(GOT_ERR_ITER_NEED_MORE);
511 
512 	*id = &graph->iter_node->id;
513 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
514 	return NULL;
515 }