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