git-rev-list libification: rev-list walking
[git/git.git] / revision.c
1 #include "cache.h"
2 #include "tag.h"
3 #include "blob.h"
4 #include "tree.h"
5 #include "commit.h"
6 #include "diff.h"
7 #include "refs.h"
8 #include "revision.h"
9
10 static char *path_name(struct name_path *path, const char *name)
11 {
12 struct name_path *p;
13 char *n, *m;
14 int nlen = strlen(name);
15 int len = nlen + 1;
16
17 for (p = path; p; p = p->up) {
18 if (p->elem_len)
19 len += p->elem_len + 1;
20 }
21 n = xmalloc(len);
22 m = n + len - (nlen + 1);
23 strcpy(m, name);
24 for (p = path; p; p = p->up) {
25 if (p->elem_len) {
26 m -= p->elem_len + 1;
27 memcpy(m, p->elem, p->elem_len);
28 m[p->elem_len] = '/';
29 }
30 }
31 return n;
32 }
33
34 struct object_list **add_object(struct object *obj,
35 struct object_list **p,
36 struct name_path *path,
37 const char *name)
38 {
39 struct object_list *entry = xmalloc(sizeof(*entry));
40 entry->item = obj;
41 entry->next = *p;
42 entry->name = path_name(path, name);
43 *p = entry;
44 return &entry->next;
45 }
46
47 static void mark_blob_uninteresting(struct blob *blob)
48 {
49 if (blob->object.flags & UNINTERESTING)
50 return;
51 blob->object.flags |= UNINTERESTING;
52 }
53
54 void mark_tree_uninteresting(struct tree *tree)
55 {
56 struct object *obj = &tree->object;
57 struct tree_entry_list *entry;
58
59 if (obj->flags & UNINTERESTING)
60 return;
61 obj->flags |= UNINTERESTING;
62 if (!has_sha1_file(obj->sha1))
63 return;
64 if (parse_tree(tree) < 0)
65 die("bad tree %s", sha1_to_hex(obj->sha1));
66 entry = tree->entries;
67 tree->entries = NULL;
68 while (entry) {
69 struct tree_entry_list *next = entry->next;
70 if (entry->directory)
71 mark_tree_uninteresting(entry->item.tree);
72 else
73 mark_blob_uninteresting(entry->item.blob);
74 free(entry);
75 entry = next;
76 }
77 }
78
79 void mark_parents_uninteresting(struct commit *commit)
80 {
81 struct commit_list *parents = commit->parents;
82
83 while (parents) {
84 struct commit *commit = parents->item;
85 commit->object.flags |= UNINTERESTING;
86
87 /*
88 * Normally we haven't parsed the parent
89 * yet, so we won't have a parent of a parent
90 * here. However, it may turn out that we've
91 * reached this commit some other way (where it
92 * wasn't uninteresting), in which case we need
93 * to mark its parents recursively too..
94 */
95 if (commit->parents)
96 mark_parents_uninteresting(commit);
97
98 /*
99 * A missing commit is ok iff its parent is marked
100 * uninteresting.
101 *
102 * We just mark such a thing parsed, so that when
103 * it is popped next time around, we won't be trying
104 * to parse it and get an error.
105 */
106 if (!has_sha1_file(commit->object.sha1))
107 commit->object.parsed = 1;
108 parents = parents->next;
109 }
110 }
111
112 static void add_pending_object(struct rev_info *revs, struct object *obj, const char *name)
113 {
114 add_object(obj, &revs->pending_objects, NULL, name);
115 }
116
117 static struct commit *get_commit_reference(struct rev_info *revs, const char *name, const unsigned char *sha1, unsigned int flags)
118 {
119 struct object *object;
120
121 object = parse_object(sha1);
122 if (!object)
123 die("bad object %s", name);
124
125 /*
126 * Tag object? Look what it points to..
127 */
128 while (object->type == tag_type) {
129 struct tag *tag = (struct tag *) object;
130 object->flags |= flags;
131 if (revs->tag_objects && !(object->flags & UNINTERESTING))
132 add_pending_object(revs, object, tag->tag);
133 object = parse_object(tag->tagged->sha1);
134 if (!object)
135 die("bad object %s", sha1_to_hex(tag->tagged->sha1));
136 }
137
138 /*
139 * Commit object? Just return it, we'll do all the complex
140 * reachability crud.
141 */
142 if (object->type == commit_type) {
143 struct commit *commit = (struct commit *)object;
144 object->flags |= flags;
145 if (parse_commit(commit) < 0)
146 die("unable to parse commit %s", name);
147 if (flags & UNINTERESTING) {
148 mark_parents_uninteresting(commit);
149 revs->limited = 1;
150 }
151 return commit;
152 }
153
154 /*
155 * Tree object? Either mark it uniniteresting, or add it
156 * to the list of objects to look at later..
157 */
158 if (object->type == tree_type) {
159 struct tree *tree = (struct tree *)object;
160 if (!revs->tree_objects)
161 return NULL;
162 if (flags & UNINTERESTING) {
163 mark_tree_uninteresting(tree);
164 return NULL;
165 }
166 add_pending_object(revs, object, "");
167 return NULL;
168 }
169
170 /*
171 * Blob object? You know the drill by now..
172 */
173 if (object->type == blob_type) {
174 struct blob *blob = (struct blob *)object;
175 if (!revs->blob_objects)
176 return NULL;
177 if (flags & UNINTERESTING) {
178 mark_blob_uninteresting(blob);
179 return NULL;
180 }
181 add_pending_object(revs, object, "");
182 return NULL;
183 }
184 die("%s is unknown object", name);
185 }
186
187 static int everybody_uninteresting(struct commit_list *orig)
188 {
189 struct commit_list *list = orig;
190 while (list) {
191 struct commit *commit = list->item;
192 list = list->next;
193 if (commit->object.flags & UNINTERESTING)
194 continue;
195 return 0;
196 }
197 return 1;
198 }
199
200 #define TREE_SAME 0
201 #define TREE_NEW 1
202 #define TREE_DIFFERENT 2
203 static int tree_difference = TREE_SAME;
204
205 static void file_add_remove(struct diff_options *options,
206 int addremove, unsigned mode,
207 const unsigned char *sha1,
208 const char *base, const char *path)
209 {
210 int diff = TREE_DIFFERENT;
211
212 /*
213 * Is it an add of a new file? It means that
214 * the old tree didn't have it at all, so we
215 * will turn "TREE_SAME" -> "TREE_NEW", but
216 * leave any "TREE_DIFFERENT" alone (and if
217 * it already was "TREE_NEW", we'll keep it
218 * "TREE_NEW" of course).
219 */
220 if (addremove == '+') {
221 diff = tree_difference;
222 if (diff != TREE_SAME)
223 return;
224 diff = TREE_NEW;
225 }
226 tree_difference = diff;
227 }
228
229 static void file_change(struct diff_options *options,
230 unsigned old_mode, unsigned new_mode,
231 const unsigned char *old_sha1,
232 const unsigned char *new_sha1,
233 const char *base, const char *path)
234 {
235 tree_difference = TREE_DIFFERENT;
236 }
237
238 static struct diff_options diff_opt = {
239 .recursive = 1,
240 .add_remove = file_add_remove,
241 .change = file_change,
242 };
243
244 static int compare_tree(struct tree *t1, struct tree *t2)
245 {
246 if (!t1)
247 return TREE_NEW;
248 if (!t2)
249 return TREE_DIFFERENT;
250 tree_difference = TREE_SAME;
251 if (diff_tree_sha1(t1->object.sha1, t2->object.sha1, "", &diff_opt) < 0)
252 return TREE_DIFFERENT;
253 return tree_difference;
254 }
255
256 static int same_tree_as_empty(struct tree *t1)
257 {
258 int retval;
259 void *tree;
260 struct tree_desc empty, real;
261
262 if (!t1)
263 return 0;
264
265 tree = read_object_with_reference(t1->object.sha1, "tree", &real.size, NULL);
266 if (!tree)
267 return 0;
268 real.buf = tree;
269
270 empty.buf = "";
271 empty.size = 0;
272
273 tree_difference = 0;
274 retval = diff_tree(&empty, &real, "", &diff_opt);
275 free(tree);
276
277 return retval >= 0 && !tree_difference;
278 }
279
280 static void try_to_simplify_commit(struct rev_info *revs, struct commit *commit)
281 {
282 struct commit_list **pp, *parent;
283
284 if (!commit->tree)
285 return;
286
287 if (!commit->parents) {
288 if (!same_tree_as_empty(commit->tree))
289 commit->object.flags |= TREECHANGE;
290 return;
291 }
292
293 pp = &commit->parents;
294 while ((parent = *pp) != NULL) {
295 struct commit *p = parent->item;
296
297 if (p->object.flags & UNINTERESTING) {
298 pp = &parent->next;
299 continue;
300 }
301
302 parse_commit(p);
303 switch (compare_tree(p->tree, commit->tree)) {
304 case TREE_SAME:
305 parent->next = NULL;
306 commit->parents = parent;
307 return;
308
309 case TREE_NEW:
310 if (revs->remove_empty_trees && same_tree_as_empty(p->tree)) {
311 *pp = parent->next;
312 continue;
313 }
314 /* fallthrough */
315 case TREE_DIFFERENT:
316 pp = &parent->next;
317 continue;
318 }
319 die("bad tree compare for commit %s", sha1_to_hex(commit->object.sha1));
320 }
321 commit->object.flags |= TREECHANGE;
322 }
323
324 static void add_parents_to_list(struct rev_info *revs, struct commit *commit, struct commit_list **list)
325 {
326 struct commit_list *parent = commit->parents;
327
328 /*
329 * If the commit is uninteresting, don't try to
330 * prune parents - we want the maximal uninteresting
331 * set.
332 *
333 * Normally we haven't parsed the parent
334 * yet, so we won't have a parent of a parent
335 * here. However, it may turn out that we've
336 * reached this commit some other way (where it
337 * wasn't uninteresting), in which case we need
338 * to mark its parents recursively too..
339 */
340 if (commit->object.flags & UNINTERESTING) {
341 while (parent) {
342 struct commit *p = parent->item;
343 parent = parent->next;
344 parse_commit(p);
345 p->object.flags |= UNINTERESTING;
346 if (p->parents)
347 mark_parents_uninteresting(p);
348 if (p->object.flags & SEEN)
349 continue;
350 p->object.flags |= SEEN;
351 insert_by_date(p, list);
352 }
353 return;
354 }
355
356 /*
357 * Ok, the commit wasn't uninteresting. Try to
358 * simplify the commit history and find the parent
359 * that has no differences in the path set if one exists.
360 */
361 if (revs->paths)
362 try_to_simplify_commit(revs, commit);
363
364 parent = commit->parents;
365 while (parent) {
366 struct commit *p = parent->item;
367
368 parent = parent->next;
369
370 parse_commit(p);
371 if (p->object.flags & SEEN)
372 continue;
373 p->object.flags |= SEEN;
374 insert_by_date(p, list);
375 }
376 }
377
378 static void limit_list(struct rev_info *revs)
379 {
380 struct commit_list *list = revs->commits;
381 struct commit_list *newlist = NULL;
382 struct commit_list **p = &newlist;
383
384 while (list) {
385 struct commit_list *entry = list;
386 struct commit *commit = list->item;
387 struct object *obj = &commit->object;
388
389 list = list->next;
390 free(entry);
391
392 if (revs->max_age != -1 && (commit->date < revs->max_age))
393 obj->flags |= UNINTERESTING;
394 if (revs->unpacked && has_sha1_pack(obj->sha1))
395 obj->flags |= UNINTERESTING;
396 add_parents_to_list(revs, commit, &list);
397 if (obj->flags & UNINTERESTING) {
398 mark_parents_uninteresting(commit);
399 if (everybody_uninteresting(list))
400 break;
401 continue;
402 }
403 if (revs->min_age != -1 && (commit->date > revs->min_age))
404 continue;
405 p = &commit_list_insert(commit, p)->next;
406 }
407 revs->commits = newlist;
408 }
409
410 static void add_one_commit(struct commit *commit, struct rev_info *revs)
411 {
412 if (!commit || (commit->object.flags & SEEN))
413 return;
414 commit->object.flags |= SEEN;
415 commit_list_insert(commit, &revs->commits);
416 }
417
418 static int all_flags;
419 static struct rev_info *all_revs;
420
421 static int handle_one_ref(const char *path, const unsigned char *sha1)
422 {
423 struct commit *commit = get_commit_reference(all_revs, path, sha1, all_flags);
424 add_one_commit(commit, all_revs);
425 return 0;
426 }
427
428 static void handle_all(struct rev_info *revs, unsigned flags)
429 {
430 all_revs = revs;
431 all_flags = flags;
432 for_each_ref(handle_one_ref);
433 }
434
435 /*
436 * Parse revision information, filling in the "rev_info" structure,
437 * and removing the used arguments from the argument list.
438 *
439 * Returns the number of arguments left ("new argc").
440 */
441 int setup_revisions(int argc, const char **argv, struct rev_info *revs, const char *def)
442 {
443 int i, flags, seen_dashdash;
444 const char **unrecognized = argv+1;
445 int left = 1;
446
447 memset(revs, 0, sizeof(*revs));
448 revs->lifo = 1;
449 revs->dense = 1;
450 revs->prefix = setup_git_directory();
451 revs->max_age = -1;
452 revs->min_age = -1;
453 revs->max_count = -1;
454
455 /* First, search for "--" */
456 seen_dashdash = 0;
457 for (i = 1; i < argc; i++) {
458 const char *arg = argv[i];
459 if (strcmp(arg, "--"))
460 continue;
461 argv[i] = NULL;
462 argc = i;
463 revs->paths = get_pathspec(revs->prefix, argv + i + 1);
464 seen_dashdash = 1;
465 break;
466 }
467
468 flags = 0;
469 for (i = 1; i < argc; i++) {
470 struct commit *commit;
471 const char *arg = argv[i];
472 unsigned char sha1[20];
473 char *dotdot;
474 int local_flags;
475
476 if (*arg == '-') {
477 if (!strncmp(arg, "--max-count=", 12)) {
478 revs->max_count = atoi(arg + 12);
479 continue;
480 }
481 if (!strncmp(arg, "--max-age=", 10)) {
482 revs->max_age = atoi(arg + 10);
483 revs->limited = 1;
484 continue;
485 }
486 if (!strncmp(arg, "--min-age=", 10)) {
487 revs->min_age = atoi(arg + 10);
488 revs->limited = 1;
489 continue;
490 }
491 if (!strcmp(arg, "--all")) {
492 handle_all(revs, flags);
493 continue;
494 }
495 if (!strcmp(arg, "--not")) {
496 flags ^= UNINTERESTING;
497 continue;
498 }
499 if (!strcmp(arg, "--default")) {
500 if (++i >= argc)
501 die("bad --default argument");
502 def = argv[i];
503 continue;
504 }
505 if (!strcmp(arg, "--topo-order")) {
506 revs->topo_order = 1;
507 revs->limited = 1;
508 continue;
509 }
510 if (!strcmp(arg, "--date-order")) {
511 revs->lifo = 0;
512 revs->topo_order = 1;
513 revs->limited = 1;
514 continue;
515 }
516 if (!strcmp(arg, "--dense")) {
517 revs->dense = 1;
518 continue;
519 }
520 if (!strcmp(arg, "--sparse")) {
521 revs->dense = 0;
522 continue;
523 }
524 if (!strcmp(arg, "--remove-empty")) {
525 revs->remove_empty_trees = 1;
526 continue;
527 }
528 if (!strcmp(arg, "--objects")) {
529 revs->tag_objects = 1;
530 revs->tree_objects = 1;
531 revs->blob_objects = 1;
532 continue;
533 }
534 if (!strcmp(arg, "--objects-edge")) {
535 revs->tag_objects = 1;
536 revs->tree_objects = 1;
537 revs->blob_objects = 1;
538 revs->edge_hint = 1;
539 continue;
540 }
541 if (!strcmp(arg, "--unpacked")) {
542 revs->unpacked = 1;
543 revs->limited = 1;
544 continue;
545 }
546 *unrecognized++ = arg;
547 left++;
548 continue;
549 }
550 dotdot = strstr(arg, "..");
551 if (dotdot) {
552 unsigned char from_sha1[20];
553 char *next = dotdot + 2;
554 *dotdot = 0;
555 if (!*next)
556 next = "HEAD";
557 if (!get_sha1(arg, from_sha1) && !get_sha1(next, sha1)) {
558 struct commit *exclude;
559 struct commit *include;
560
561 exclude = get_commit_reference(revs, arg, from_sha1, flags ^ UNINTERESTING);
562 include = get_commit_reference(revs, next, sha1, flags);
563 if (!exclude || !include)
564 die("Invalid revision range %s..%s", arg, next);
565 add_one_commit(exclude, revs);
566 add_one_commit(include, revs);
567 continue;
568 }
569 *dotdot = '.';
570 }
571 local_flags = 0;
572 if (*arg == '^') {
573 local_flags = UNINTERESTING;
574 arg++;
575 }
576 if (get_sha1(arg, sha1) < 0) {
577 struct stat st;
578 int j;
579
580 if (seen_dashdash || local_flags)
581 die("bad revision '%s'", arg);
582
583 /* If we didn't have a "--", all filenames must exist */
584 for (j = i; j < argc; j++) {
585 if (lstat(argv[j], &st) < 0)
586 die("'%s': %s", arg, strerror(errno));
587 }
588 revs->paths = get_pathspec(revs->prefix, argv + i);
589 break;
590 }
591 commit = get_commit_reference(revs, arg, sha1, flags ^ local_flags);
592 add_one_commit(commit, revs);
593 }
594 if (def && !revs->commits) {
595 unsigned char sha1[20];
596 struct commit *commit;
597 if (get_sha1(def, sha1) < 0)
598 die("bad default revision '%s'", def);
599 commit = get_commit_reference(revs, def, sha1, 0);
600 add_one_commit(commit, revs);
601 }
602 if (revs->paths)
603 revs->limited = 1;
604 *unrecognized = NULL;
605 return left;
606 }
607
608 void prepare_revision_walk(struct rev_info *revs)
609 {
610 if (revs->paths)
611 diff_tree_setup_paths(revs->paths);
612 sort_by_date(&revs->commits);
613 if (revs->limited)
614 limit_list(revs);
615 if (revs->topo_order)
616 sort_in_topological_order(&revs->commits, revs->lifo);
617 }
618
619 struct commit *get_revision(struct rev_info *revs)
620 {
621 if (!revs->commits)
622 return NULL;
623 return pop_most_recent_commit(&revs->commits, SEEN);
624 }
625
626