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