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 struct got_commit_graph {
59 	/* The set of all commits we have traversed. */
60 	struct got_object_idset *node_ids;
61 
62 	/* The commit at which traversal began (youngest commit in node_ids). */
63 	struct got_commit_graph_node *head_node;
64 
65 	/*
66 	 * A set of object IDs of known parent commits which we have not yet
67 	 * traversed. Each commit ID in this set represents a branch in commit
68 	 * history: Either the first-parent branch of the head node, or another
69 	 * branch corresponding to a traversed merge commit for which we have
70 	 * not traversed a branch point commit yet.
71 	 *
72 	 * Whenever we add a commit with a matching ID to the graph, we remove
73 	 * its corresponding element from this set, and add new elements for
74 	 * each of that commit's parent commits which were not traversed yet.
75 	 *
76 	 * When API users ask us to fetch more commits, we fetch commits from
77 	 * all currently open branches. This allows API users to process
78 	 * commits in linear order even though the history contains branches.
79 	 */
80 	struct got_object_idset *open_branches;
81 
82 	/* The next commit to return when the API user asks for one. */
83 	struct got_commit_graph_node *iter_node;
84 
85 	TAILQ_HEAD(, got_commit_graph_node) iter_candidates;
86 };
87 
88 static struct got_commit_graph *
89 alloc_graph(void)
90 {
91 	struct got_commit_graph *graph;
92 
93 	graph = calloc(1, sizeof(*graph));
94 	if (graph == NULL)
95 		return NULL;
96 
97 	graph->node_ids = got_object_idset_alloc();
98 	if (graph->node_ids == NULL) {
99 		free(graph);
100 		return NULL;
101 	}
102 
103 	graph->open_branches = got_object_idset_alloc();
104 	if (graph->open_branches == NULL) {
105 		got_object_idset_free(graph->node_ids);
106 		free(graph);
107 		return NULL;
108 	}
109 
110 	TAILQ_INIT(&graph->iter_candidates);
111 	return graph;
112 }
113 
114 #if 0
115 static int
116 is_head_node(struct got_commit_graph_node *node)
117 {
118 	return node->nchildren == 0;
119 }
120 
121 static int
122 is_merge_point(struct got_commit_graph_node *node)
123 {
124 	return node->nparents > 1;
125 }
126 
127 int
128 is_branch_point(struct got_commit_graph_node *node)
129 {
130 	return node->nchildren > 1;
131 }
132 #endif
133 
134 static int
135 is_root_node(struct got_commit_graph_node *node)
136 {
137 	return node->nparents == 0;
138 }
139 
140 static void
141 add_iteration_candidate(struct got_commit_graph *graph,
142     struct got_commit_graph_node *node)
143 {
144 	struct got_commit_graph_node *n, *next;
145 	
146 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
147 		TAILQ_INSERT_TAIL(&graph->iter_candidates, node, entry);
148 		return;
149 	}
150 
151 	TAILQ_FOREACH(n, &graph->iter_candidates, entry) {
152 		if (node->commit_timestamp < n->commit_timestamp) {
153 			next = TAILQ_NEXT(n, entry);
154 			if (next == NULL) {
155 				TAILQ_INSERT_AFTER(&graph->iter_candidates, n,
156 				    node, entry);
157 				break;
158 			}
159 			if (node->commit_timestamp >= next->commit_timestamp) {
160 				TAILQ_INSERT_BEFORE(next, node, entry);
161 				break;
162 			}
163 		} else {
164 			TAILQ_INSERT_BEFORE(n, node, entry);
165 			break;
166 		}
167 	}
168 }
169 
170 static const struct got_error *
171 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
172 {
173 	struct got_object_qid *qid;
174 
175 	qid = calloc(1, sizeof(*qid));
176 	if (qid == NULL)
177 		return got_error_from_errno();
178 
179 	qid->id = got_object_id_dup(id);
180 	if (qid->id == NULL) {
181 		const struct got_error *err = got_error_from_errno();
182 		free(qid);
183 		return err;
184 	}
185 
186 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
187 	return NULL;
188 }
189 
190 static const struct got_error *
191 add_node(struct got_commit_graph_node **new_node,
192     struct got_commit_graph *graph, struct got_object_id *commit_id,
193     struct got_commit_object *commit, struct got_object_id *child_commit_id)
194 {
195 	const struct got_error *err = NULL;
196 	struct got_commit_graph_node *node, *existing_node;
197 	struct got_object_qid *qid;
198 
199 	*new_node = NULL;
200 
201 	node = calloc(1, sizeof(*node));
202 	if (node == NULL)
203 		return got_error_from_errno();
204 
205 	memcpy(&node->id, commit_id, sizeof(node->id));
206 	SIMPLEQ_INIT(&node->parent_ids);
207 	SIMPLEQ_INIT(&node->child_ids);
208 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
209 		err = add_vertex(&node->parent_ids, qid->id);
210 		if (err)
211 			return err;
212 		node->nparents++;
213 	}
214 	node->commit_timestamp = commit->committer_time; /* XXX not UTC! */
215 
216 	err = got_object_idset_add((void **)(&existing_node),
217 	    graph->node_ids, &node->id, node);
218 	if (err == NULL) {
219 		struct got_object_qid *qid;
220 
221 		add_iteration_candidate(graph, node);
222 		err = got_object_idset_remove(graph->open_branches, commit_id);
223 		if (err && err->code != GOT_ERR_NO_OBJ)
224 			return err;
225 		SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
226 			if (got_object_idset_get(graph->node_ids, qid->id))
227 				continue; /* parent already traversed */
228 			err = got_object_idset_add(NULL, graph->open_branches,
229 			    qid->id, node);
230 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
231 				return err;
232 		}
233 		*new_node = node;
234 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
235 		err = NULL;
236 		free(node);
237 		node = existing_node;
238 	} else {
239 		free(node);
240 		return err;
241 	}
242 
243 	if (child_commit_id) {
244 		struct got_object_qid *cid;
245 
246 		/* Prevent linking to self. */
247 		if (got_object_id_cmp(commit_id, child_commit_id) == 0)
248 			return got_error(GOT_ERR_BAD_OBJ_ID);
249 
250 		/* Prevent double-linking to the same child. */
251 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
252 			if (got_object_id_cmp(cid->id, child_commit_id) == 0)
253 				return got_error(GOT_ERR_BAD_OBJ_ID);
254 		}
255 
256 		err = add_vertex(&node->child_ids, child_commit_id);
257 		if (err)
258 			return err;
259 		node->nchildren++;
260 
261 	}
262 
263 	return err;
264 }
265 
266 const struct got_error *
267 got_commit_graph_open(struct got_commit_graph **graph,
268     struct got_object_id *commit_id, struct got_repository *repo)
269 {
270 	const struct got_error *err = NULL;
271 	struct got_commit_object *commit;
272 
273 	*graph = NULL;
274 
275 	err = got_object_open_as_commit(&commit, repo, commit_id);
276 	if (err)
277 		return err;
278 
279 	*graph = alloc_graph();
280 	if (*graph == NULL) {
281 		got_object_commit_close(commit);
282 		return got_error_from_errno();
283 	}
284 
285 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
286 	got_object_commit_close(commit);
287 	if (err) {
288 		got_commit_graph_close(*graph);
289 		*graph = NULL;
290 		return err;
291 	}
292 	
293 	return NULL;
294 }
295 
296 struct got_commit_graph_branch {
297 	struct got_object_id parent_id;
298 	struct got_commit_graph_node *node;
299 };
300 
301 struct gather_branches_arg {
302 	struct got_commit_graph_branch *branches;
303 	int nbranches;
304 };
305 
306 static void
307 gather_branches(struct got_object_id *id, void *data, void *arg)
308 {
309 	struct gather_branches_arg *a = arg;
310 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
311 	a->branches[a->nbranches].node = data;
312 	a->nbranches++;
313 }
314 
315 static const struct got_error *
316 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
317     struct got_commit_graph *graph, struct got_repository *repo,
318     struct got_object_id *wanted_id)
319 {
320 	const struct got_error *err;
321 	struct got_commit_graph_branch *branches;
322 	struct gather_branches_arg arg;
323 	int i;
324 
325 	*ncommits = 0;
326 	if (wanted_id_added)
327 		*wanted_id_added = 0;
328 
329 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
330 	if (arg.nbranches == 0)
331 		return NULL;
332 
333 	/*
334 	 * Adding nodes to the graph might change the graph's open
335 	 * branches state. Create a local copy of the current state.
336 	 */
337 	branches = calloc(arg.nbranches, sizeof(*branches));
338 	if (branches == NULL)
339 		return got_error_from_errno();
340 	arg.nbranches = 0; /* reset; gather_branches() will increment */
341 	arg.branches = branches;
342 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
343 
344 	for (i = 0; i < arg.nbranches; i++) {
345 		struct got_object_id *commit_id;
346 		struct got_commit_graph_node *child_node, *new_node;
347 		struct got_commit_object *commit;
348 
349 		commit_id = &branches[i].parent_id;
350 		child_node = branches[i].node;
351 
352 		err = got_object_open_as_commit(&commit, repo, commit_id);
353 		if (err)
354 			break;
355 
356 		err = add_node(&new_node, graph, commit_id, commit,
357 		    &child_node->id);
358 		got_object_commit_close(commit);
359 		if (err) {
360 			if (err->code != GOT_ERR_OBJ_EXISTS)
361 				break;
362 			err = NULL;
363 		}
364 		if (new_node)
365 			(*ncommits)++;
366 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
367 			*wanted_id_added = 1;
368 	}
369 
370 	free(branches);
371 	return err;
372 }
373 
374 const struct got_error *
375 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
376     int limit, struct got_repository *repo)
377 {
378 	const struct got_error *err;
379 	int ncommits;
380 
381 	*nfetched = 0;
382 
383 	while (*nfetched < limit) {
384 		err = fetch_commits_from_open_branches(&ncommits, NULL,
385 		    graph, repo, NULL);
386 		if (err)
387 			return err;
388 		if (ncommits == 0)
389 			break;
390 		*nfetched += ncommits;
391 	}
392 
393 	return NULL;
394 }
395 
396 const struct got_error *
397 got_commit_graph_fetch_commits_up_to(int *nfetched,
398     struct got_commit_graph *graph, struct got_object_id *wanted_id,
399     struct got_repository *repo)
400 {
401 	const struct got_error *err;
402 	int ncommits, wanted_id_added = 0;
403 
404 	*nfetched = 0;
405 	while (!wanted_id_added) {
406 		err = fetch_commits_from_open_branches(&ncommits,
407 		    &wanted_id_added, graph, repo, wanted_id);
408 		if (err)
409 			return err;
410 		if (ncommits == 0)
411 			return NULL;
412 		*nfetched += ncommits;
413 	}
414 
415 	return NULL;
416 }
417 
418 static void
419 free_graph_node(struct got_object_id *id, void *data, void *arg)
420 {
421 	struct got_commit_graph_node *node = data;
422 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
423 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
424 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
425 		free(child);
426 	}
427 	free(node);
428 }
429 
430 void
431 got_commit_graph_close(struct got_commit_graph *graph)
432 {
433 	got_object_idset_free(graph->open_branches);
434 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
435 	got_object_idset_free(graph->node_ids);
436 	free(graph);
437 }
438 
439 const struct got_error *
440 got_commit_graph_iter_start(struct got_commit_graph *graph,
441     struct got_object_id *id)
442 {
443 	struct got_commit_graph_node *start_node, *node;
444 	struct got_object_qid *qid;
445 
446 	start_node = got_object_idset_get(graph->node_ids, id);
447 	if (start_node == NULL)
448 		return got_error(GOT_ERR_NO_OBJ);
449 
450 	graph->iter_node = start_node;
451 
452 	while (!TAILQ_EMPTY(&graph->iter_candidates)) {
453 		node = TAILQ_FIRST(&graph->iter_candidates);
454 		TAILQ_REMOVE(&graph->iter_candidates, node, entry);
455 	}
456 
457 	/* Put all known parents of this commit on the candidate list. */
458 	SIMPLEQ_FOREACH(qid, &start_node->parent_ids, entry) {
459 		node = got_object_idset_get(graph->node_ids, qid->id);
460 		if (node)
461 			add_iteration_candidate(graph, node);
462 	}
463 
464 	return NULL;
465 }
466 
467 const struct got_error *
468 got_commit_graph_iter_next(struct got_object_id **id,
469     struct got_commit_graph *graph)
470 {
471 	struct got_commit_graph_node *node;
472 
473 	if (graph->iter_node == NULL) {
474 		/* We are done interating, or iteration was not started. */
475 		*id = NULL;
476 		return NULL;
477 	}
478 
479 	if (TAILQ_EMPTY(&graph->iter_candidates)) {
480 		if (is_root_node(graph->iter_node) &&
481 		    got_object_idset_num_elements(graph->open_branches) == 0) {
482 			*id = &graph->iter_node->id;
483 			/* We are done interating. */
484 			graph->iter_node = NULL;
485 			return NULL;
486 		}
487 		return got_error(GOT_ERR_ITER_NEED_MORE);
488 	}
489 
490 	*id = &graph->iter_node->id;
491 	node = TAILQ_FIRST(&graph->iter_candidates);
492 	TAILQ_REMOVE(&graph->iter_candidates, node, entry);
493 	graph->iter_node = node;
494 	return NULL;
495 }
496 
497 int
498 got_commit_graph_contains_object(struct got_commit_graph *graph,
499     struct got_object_id *id)
500 {
501 	return (got_object_idset_get(graph->node_ids, id) != NULL);
502 }