Raw
1 #define USE_THE_REPOSITORY_VARIABLE
2 #define DISABLE_SIGN_COMPARE_WARNINGS
3
4 #include "git-compat-util.h"
5 #include "config.h"
6 #include "commit.h"
7 #include "diff.h"
8 #include "environment.h"
9 #include "gettext.h"
10 #include "hex.h"
11 #include "revision.h"
12 #include "refs.h"
13 #include "list-objects.h"
14 #include "quote.h"
15 #include "run-command.h"
16 #include "log-tree.h"
17 #include "bisect.h"
18 #include "oid-array.h"
19 #include "strvec.h"
20 #include "commit-slab.h"
21 #include "commit-reach.h"
22 #include "object-name.h"
23 #include "odb.h"
24 #include "path.h"
25 #include "dir.h"
26
27 static struct oid_array good_revs;
28 static struct oid_array skipped_revs;
29
30 static struct object_id *current_bad_oid;
31
32 static char *term_bad;
33 static char *term_good;
34
35 /* Remember to update object flag allocation in object.h */
36 #define COUNTED (1u<<16)
37
38 /*
39 * This is a truly stupid algorithm, but it's only
40 * used for bisection, and we just don't care enough.
41 *
42 * We care just barely enough to avoid recursing for
43 * non-merge entries.
44 */
45 static int count_distance(struct commit_list *entry)
46 {
47 int nr = 0;
48
49 while (entry) {
50 struct commit *commit = entry->item;
51 struct commit_list *p;
52
53 if (commit->object.flags & (UNINTERESTING | COUNTED))
54 break;
55 if (!(commit->object.flags & TREESAME))
56 nr++;
57 commit->object.flags |= COUNTED;
58 p = commit->parents;
59 entry = p;
60 if (p) {
61 p = p->next;
62 while (p) {
63 nr += count_distance(p);
64 p = p->next;
65 }
66 }
67 }
68
69 return nr;
70 }
71
72 static void clear_distance(struct commit_list *list)
73 {
74 while (list) {
75 struct commit *commit = list->item;
76 commit->object.flags &= ~COUNTED;
77 list = list->next;
78 }
79 }
80
81 define_commit_slab(commit_weight, int *);
82 static struct commit_weight commit_weight;
83
84 #define DEBUG_BISECT 0
85
86 static inline int weight(struct commit_list *elem)
87 {
88 return **commit_weight_at(&commit_weight, elem->item);
89 }
90
91 static inline void weight_set(struct commit_list *elem, int weight)
92 {
93 **commit_weight_at(&commit_weight, elem->item) = weight;
94 }
95
96 static int count_interesting_parents(struct commit *commit, unsigned bisect_flags)
97 {
98 struct commit_list *p;
99 int count;
100
101 for (count = 0, p = commit->parents; p; p = p->next) {
102 if (!(p->item->object.flags & UNINTERESTING))
103 count++;
104 if (bisect_flags & FIND_BISECTION_FIRST_PARENT_ONLY)
105 break;
106 }
107 return count;
108 }
109
110 static inline int approx_halfway(struct commit_list *p, int nr)
111 {
112 int diff;
113
114 /*
115 * Don't short-cut something we are not going to return!
116 */
117 if (p->item->object.flags & TREESAME)
118 return 0;
119 if (DEBUG_BISECT)
120 return 0;
121 /*
122 * For small number of commits 2 and 3 are halfway of 5, and
123 * 3 is halfway of 6 but 2 and 4 are not.
124 */
125 diff = 2 * weight(p) - nr;
126 switch (diff) {
127 case -1: case 0: case 1:
128 return 1;
129 default:
130 /*
131 * For large number of commits we are not so strict, it's
132 * good enough if it's within ~0.1% of the halfway point,
133 * e.g. 5000 is exactly halfway of 10000, but we consider
134 * the values [4996, 5004] as halfway as well.
135 */
136 if (abs(diff) < nr / 1024)
137 return 1;
138 return 0;
139 }
140 }
141
142 static void show_list(const char *debug, int counted, int nr,
143 struct commit_list *list)
144 {
145 struct commit_list *p;
146
147 if (!DEBUG_BISECT)
148 return;
149
150 fprintf(stderr, "%s (%d/%d)\n", debug, counted, nr);
151
152 for (p = list; p; p = p->next) {
153 struct commit_list *pp;
154 struct commit *commit = p->item;
155 unsigned commit_flags = commit->object.flags;
156 enum object_type type;
157 size_t size;
158 char *buf = odb_read_object(the_repository->objects,
159 &commit->object.oid, &type,
160 &size);
161 const char *subject_start;
162 int subject_len;
163
164 if (!buf)
165 die(_("unable to read %s"), oid_to_hex(&commit->object.oid));
166
167 fprintf(stderr, "%c%c%c ",
168 (commit_flags & TREESAME) ? ' ' : 'T',
169 (commit_flags & UNINTERESTING) ? 'U' : ' ',
170 (commit_flags & COUNTED) ? 'C' : ' ');
171 if (*commit_weight_at(&commit_weight, p->item))
172 fprintf(stderr, "%3d", weight(p));
173 else
174 fprintf(stderr, "---");
175 fprintf(stderr, " %.*s", 8, oid_to_hex(&commit->object.oid));
176 for (pp = commit->parents; pp; pp = pp->next)
177 fprintf(stderr, " %.*s", 8,
178 oid_to_hex(&pp->item->object.oid));
179
180 subject_len = find_commit_subject(buf, &subject_start);
181 if (subject_len)
182 fprintf(stderr, " %.*s", subject_len, subject_start);
183 fprintf(stderr, "\n");
184 }
185 }
186
187 static struct commit_list *best_bisection(struct commit_list *list, int nr)
188 {
189 struct commit_list *p, *best;
190 int best_distance = -1;
191
192 best = list;
193 for (p = list; p; p = p->next) {
194 int distance;
195 unsigned commit_flags = p->item->object.flags;
196
197 if (commit_flags & TREESAME)
198 continue;
199 distance = weight(p);
200 if (nr - distance < distance)
201 distance = nr - distance;
202 if (distance > best_distance) {
203 best = p;
204 best_distance = distance;
205 }
206 }
207
208 return best;
209 }
210
211 struct commit_dist {
212 struct commit *commit;
213 int distance;
214 };
215
216 static int compare_commit_dist(const void *a_, const void *b_)
217 {
218 struct commit_dist *a, *b;
219
220 a = (struct commit_dist *)a_;
221 b = (struct commit_dist *)b_;
222 if (a->distance != b->distance)
223 return b->distance - a->distance; /* desc sort */
224 return oidcmp(&a->commit->object.oid, &b->commit->object.oid);
225 }
226
227 static struct commit_list *best_bisection_sorted(struct commit_list *list, int nr)
228 {
229 struct commit_list *p;
230 struct commit_dist *array = xcalloc(nr, sizeof(*array));
231 struct strbuf buf = STRBUF_INIT;
232 int cnt, i;
233
234 for (p = list, cnt = 0; p; p = p->next) {
235 int distance;
236 unsigned commit_flags = p->item->object.flags;
237
238 if (commit_flags & TREESAME)
239 continue;
240 distance = weight(p);
241 if (nr - distance < distance)
242 distance = nr - distance;
243 array[cnt].commit = p->item;
244 array[cnt].distance = distance;
245 cnt++;
246 }
247 QSORT(array, cnt, compare_commit_dist);
248 for (p = list, i = 0; i < cnt; i++) {
249 struct object *obj = &(array[i].commit->object);
250
251 strbuf_reset(&buf);
252 strbuf_addf(&buf, "dist=%d", array[i].distance);
253 add_name_decoration(DECORATION_NONE, buf.buf, obj);
254
255 p->item = array[i].commit;
256 if (i < cnt - 1)
257 p = p->next;
258 }
259 if (p) {
260 commit_list_free(p->next);
261 p->next = NULL;
262 }
263 strbuf_release(&buf);
264 free(array);
265 return list;
266 }
267
268 /*
269 * zero or positive weight is the number of interesting commits it can
270 * reach, including itself. Especially, weight = 0 means it does not
271 * reach any tree-changing commits (e.g. just above uninteresting one
272 * but traversal is with pathspec).
273 *
274 * weight = -1 means it has one parent and its distance is yet to
275 * be computed.
276 *
277 * weight = -2 means it has more than one parent and its distance is
278 * unknown. After running count_distance() first, they will get zero
279 * or positive distance.
280 */
281 static struct commit_list *do_find_bisection(struct commit_list *list,
282 int nr, int *weights,
283 unsigned bisect_flags)
284 {
285 int n, counted;
286 struct commit_list *p;
287
288 counted = 0;
289
290 for (n = 0, p = list; p; p = p->next) {
291 struct commit *commit = p->item;
292 unsigned commit_flags = commit->object.flags;
293
294 *commit_weight_at(&commit_weight, p->item) = &weights[n++];
295 switch (count_interesting_parents(commit, bisect_flags)) {
296 case 0:
297 if (!(commit_flags & TREESAME)) {
298 weight_set(p, 1);
299 counted++;
300 show_list("bisection 2 count one",
301 counted, nr, list);
302 }
303 /*
304 * otherwise, it is known not to reach any
305 * tree-changing commit and gets weight 0.
306 */
307 break;
308 case 1:
309 weight_set(p, -1);
310 break;
311 default:
312 weight_set(p, -2);
313 break;
314 }
315 }
316
317 show_list("bisection 2 initialize", counted, nr, list);
318
319 /*
320 * If you have only one parent in the resulting set
321 * then you can reach one commit more than that parent
322 * can reach. So we do not have to run the expensive
323 * count_distance() for single strand of pearls.
324 *
325 * However, if you have more than one parents, you cannot
326 * just add their distance and one for yourself, since
327 * they usually reach the same ancestor and you would
328 * end up counting them twice that way.
329 *
330 * So we will first count distance of merges the usual
331 * way, and then fill the blanks using cheaper algorithm.
332 */
333 for (p = list; p; p = p->next) {
334 if (p->item->object.flags & UNINTERESTING)
335 continue;
336 if (weight(p) != -2)
337 continue;
338 if (bisect_flags & FIND_BISECTION_FIRST_PARENT_ONLY)
339 BUG("shouldn't be calling count-distance in fp mode");
340 weight_set(p, count_distance(p));
341 clear_distance(list);
342
343 /* Does it happen to be at half-way? */
344 if (!(bisect_flags & FIND_BISECTION_ALL) &&
345 approx_halfway(p, nr))
346 return p;
347 counted++;
348 }
349
350 show_list("bisection 2 count_distance", counted, nr, list);
351
352 while (counted < nr) {
353 for (p = list; p; p = p->next) {
354 struct commit_list *q;
355 unsigned commit_flags = p->item->object.flags;
356
357 if (0 <= weight(p))
358 continue;
359
360 for (q = p->item->parents;
361 q;
362 q = bisect_flags & FIND_BISECTION_FIRST_PARENT_ONLY ? NULL : q->next) {
363 if (q->item->object.flags & UNINTERESTING)
364 continue;
365 if (0 <= weight(q))
366 break;
367 }
368 if (!q)
369 continue;
370
371 /*
372 * weight for p is unknown but q is known.
373 * add one for p itself if p is to be counted,
374 * otherwise inherit it from q directly.
375 */
376 if (!(commit_flags & TREESAME)) {
377 weight_set(p, weight(q)+1);
378 counted++;
379 show_list("bisection 2 count one",
380 counted, nr, list);
381 }
382 else
383 weight_set(p, weight(q));
384
385 /* Does it happen to be at half-way? */
386 if (!(bisect_flags & FIND_BISECTION_ALL) &&
387 approx_halfway(p, nr))
388 return p;
389 }
390 }
391
392 show_list("bisection 2 counted all", counted, nr, list);
393
394 if (!(bisect_flags & FIND_BISECTION_ALL))
395 return best_bisection(list, nr);
396 else
397 return best_bisection_sorted(list, nr);
398 }
399
400 void find_bisection(struct commit_list **commit_list, int *reaches,
401 int *all, unsigned bisect_flags)
402 {
403 int nr, on_list;
404 struct commit_list *list, *p, *best, *next, *last;
405 int *weights;
406
407 show_list("bisection 2 entry", 0, 0, *commit_list);
408 init_commit_weight(&commit_weight);
409
410 /*
411 * Count the number of total and tree-changing items on the
412 * list, while reversing the list.
413 */
414 for (nr = on_list = 0, last = NULL, p = *commit_list;
415 p;
416 p = next) {
417 unsigned commit_flags = p->item->object.flags;
418
419 next = p->next;
420 if (commit_flags & UNINTERESTING) {
421 free(p);
422 continue;
423 }
424 p->next = last;
425 last = p;
426 if (!(commit_flags & TREESAME))
427 nr++;
428 on_list++;
429 }
430 list = last;
431 show_list("bisection 2 sorted", 0, nr, list);
432
433 *all = nr;
434 CALLOC_ARRAY(weights, on_list);
435
436 /* Do the real work of finding bisection commit. */
437 best = do_find_bisection(list, nr, weights, bisect_flags);
438 if (best) {
439 if (!(bisect_flags & FIND_BISECTION_ALL)) {
440 list->item = best->item;
441 commit_list_free(list->next);
442 best = list;
443 best->next = NULL;
444 }
445 *reaches = weight(best);
446 }
447 *commit_list = best;
448
449 free(weights);
450 clear_commit_weight(&commit_weight);
451 }
452
453 static int register_ref(const struct reference *ref, void *cb_data UNUSED)
454 {
455 struct strbuf good_prefix = STRBUF_INIT;
456 strbuf_addstr(&good_prefix, term_good);
457 strbuf_addstr(&good_prefix, "-");
458
459 if (!strcmp(ref->name, term_bad)) {
460 free(current_bad_oid);
461 current_bad_oid = xmalloc(sizeof(*current_bad_oid));
462 oidcpy(current_bad_oid, ref->oid);
463 } else if (starts_with(ref->name, good_prefix.buf)) {
464 oid_array_append(&good_revs, ref->oid);
465 } else if (starts_with(ref->name, "skip-")) {
466 oid_array_append(&skipped_revs, ref->oid);
467 }
468
469 strbuf_release(&good_prefix);
470
471 return 0;
472 }
473
474 static int read_bisect_refs(void)
475 {
476 struct refs_for_each_ref_options opts = {
477 .prefix = "refs/bisect/",
478 .trim_prefix = strlen("refs/bisect/"),
479 };
480 return refs_for_each_ref_ext(get_main_ref_store(the_repository),
481 register_ref, NULL, &opts);
482 }
483
484 static GIT_PATH_FUNC(git_path_bisect_names, "BISECT_NAMES")
485 static GIT_PATH_FUNC(git_path_bisect_ancestors_ok, "BISECT_ANCESTORS_OK")
486 static GIT_PATH_FUNC(git_path_bisect_run, "BISECT_RUN")
487 static GIT_PATH_FUNC(git_path_bisect_start, "BISECT_START")
488 static GIT_PATH_FUNC(git_path_bisect_log, "BISECT_LOG")
489 static GIT_PATH_FUNC(git_path_bisect_terms, "BISECT_TERMS")
490 static GIT_PATH_FUNC(git_path_bisect_first_parent, "BISECT_FIRST_PARENT")
491 static GIT_PATH_FUNC(git_path_bisect_reset_when_found, "BISECT_RESET_WHEN_FOUND")
492
493 static void read_bisect_paths(struct strvec *array)
494 {
495 struct strbuf str = STRBUF_INIT;
496 const char *filename = git_path_bisect_names();
497 FILE *fp = xfopen(filename, "r");
498
499 while (strbuf_getline_lf(&str, fp) != EOF) {
500 strbuf_trim(&str);
501 if (sq_dequote_to_strvec(str.buf, array))
502 die(_("Badly quoted content in file '%s': %s"),
503 filename, str.buf);
504 }
505
506 strbuf_release(&str);
507 fclose(fp);
508 }
509
510 static char *join_oid_array_hex(struct oid_array *array, char delim)
511 {
512 struct strbuf joined_hexs = STRBUF_INIT;
513 int i;
514
515 for (i = 0; i < array->nr; i++) {
516 strbuf_add_oid_hex(&joined_hexs, array->oid + i);
517 if (i + 1 < array->nr)
518 strbuf_addch(&joined_hexs, delim);
519 }
520
521 return strbuf_detach(&joined_hexs, NULL);
522 }
523
524 /*
525 * In this function, passing a not NULL skipped_first is very special.
526 * It means that we want to know if the first commit in the list is
527 * skipped because we will want to test a commit away from it if it is
528 * indeed skipped.
529 * So if the first commit is skipped, we cannot take the shortcut to
530 * just "return list" when we find the first non skipped commit, we
531 * have to return a fully filtered list.
532 *
533 * We use (*skipped_first == -1) to mean "it has been found that the
534 * first commit is not skipped". In this case *skipped_first is set back
535 * to 0 just before the function returns.
536 */
537 struct commit_list *filter_skipped(struct commit_list *list,
538 struct commit_list **tried,
539 int show_all,
540 int *count,
541 int *skipped_first)
542 {
543 struct commit_list *filtered = NULL, **f = &filtered;
544
545 *tried = NULL;
546
547 if (skipped_first)
548 *skipped_first = 0;
549 if (count)
550 *count = 0;
551
552 if (!skipped_revs.nr)
553 return list;
554
555 while (list) {
556 struct commit_list *next = list->next;
557 list->next = NULL;
558 if (0 <= oid_array_lookup(&skipped_revs, &list->item->object.oid)) {
559 if (skipped_first && !*skipped_first)
560 *skipped_first = 1;
561 /* Move current to tried list */
562 *tried = list;
563 tried = &list->next;
564 } else {
565 if (!show_all) {
566 if (!skipped_first || !*skipped_first) {
567 commit_list_free(next);
568 commit_list_free(filtered);
569 return list;
570 }
571 } else if (skipped_first && !*skipped_first) {
572 /* This means we know it's not skipped */
573 *skipped_first = -1;
574 }
575 /* Move current to filtered list */
576 *f = list;
577 f = &list->next;
578 if (count)
579 (*count)++;
580 }
581 list = next;
582 }
583
584 if (skipped_first && *skipped_first == -1)
585 *skipped_first = 0;
586
587 return filtered;
588 }
589
590 #define PRN_MODULO 32768
591
592 /*
593 * This is a pseudo random number generator based on "man 3 rand".
594 * It is not used properly because the seed is the argument and it
595 * is increased by one between each call, but that should not matter
596 * for this application.
597 */
598 static unsigned get_prn(unsigned count)
599 {
600 count = count * 1103515245 + 12345;
601 return (count/65536) % PRN_MODULO;
602 }
603
604 /*
605 * Custom integer square root from
606 * https://en.wikipedia.org/wiki/Integer_square_root
607 */
608 static int sqrti(int val)
609 {
610 float d, x = val;
611
612 if (!val)
613 return 0;
614
615 do {
616 float y = (x + (float)val / x) / 2;
617 d = (y > x) ? y - x : x - y;
618 x = y;
619 } while (d >= 0.5);
620
621 return (int)x;
622 }
623
624 static struct commit_list *skip_away(struct commit_list *list, int count)
625 {
626 struct commit_list *cur, *previous, *result = list;
627 int prn, index, i;
628
629 prn = get_prn(count);
630 index = (count * prn / PRN_MODULO) * sqrti(prn) / sqrti(PRN_MODULO);
631
632 cur = list;
633 previous = NULL;
634
635 for (i = 0; cur; cur = cur->next, i++) {
636 if (i == index) {
637 if (!oideq(&cur->item->object.oid, current_bad_oid))
638 result = cur;
639 else if (previous)
640 result = previous;
641 else
642 result = list;
643 break;
644 }
645 previous = cur;
646 }
647
648 for (cur = list; cur != result; ) {
649 struct commit_list *next = cur->next;
650 free(cur);
651 cur = next;
652 }
653
654 return result;
655 }
656
657 static struct commit_list *managed_skipped(struct commit_list *list,
658 struct commit_list **tried)
659 {
660 int count, skipped_first;
661
662 *tried = NULL;
663
664 if (!skipped_revs.nr)
665 return list;
666
667 list = filter_skipped(list, tried, 0, &count, &skipped_first);
668
669 if (!skipped_first)
670 return list;
671
672 return skip_away(list, count);
673 }
674
675 static void bisect_rev_setup(struct repository *r, struct rev_info *revs,
676 struct strvec *rev_argv,
677 const char *prefix,
678 const char *bad_format, const char *good_format,
679 int read_paths)
680 {
681 int i;
682
683 repo_init_revisions(r, revs, prefix);
684 revs->abbrev = 0;
685 revs->commit_format = CMIT_FMT_UNSPECIFIED;
686
687 /* rev_argv.argv[0] will be ignored by setup_revisions */
688 strvec_push(rev_argv, "bisect_rev_setup");
689 strvec_pushf(rev_argv, bad_format, oid_to_hex(current_bad_oid));
690 for (i = 0; i < good_revs.nr; i++)
691 strvec_pushf(rev_argv, good_format,
692 oid_to_hex(good_revs.oid + i));
693 strvec_push(rev_argv, "--");
694 if (read_paths)
695 read_bisect_paths(rev_argv);
696
697 setup_revisions_from_strvec(rev_argv, revs, NULL);
698 }
699
700 static void bisect_common(struct rev_info *revs)
701 {
702 if (prepare_revision_walk(revs))
703 die("revision walk setup failed");
704 if (revs->tree_objects)
705 mark_edges_uninteresting(revs, NULL, 0);
706 }
707
708 static enum bisect_error error_if_skipped_commits(struct commit_list *tried,
709 const struct object_id *bad)
710 {
711 if (!tried)
712 return BISECT_OK;
713
714 printf("There are only 'skip'ped commits left to test.\n"
715 "The first '%s' commit could be any of:\n", term_bad);
716
717 for ( ; tried; tried = tried->next)
718 printf("%s\n", oid_to_hex(&tried->item->object.oid));
719
720 if (bad)
721 printf("%s\n", oid_to_hex(bad));
722 printf(_("We cannot bisect more!\n"));
723
724 return BISECT_ONLY_SKIPPED_LEFT;
725 }
726
727 static int is_expected_rev(const struct object_id *oid)
728 {
729 struct object_id expected_oid;
730 if (refs_read_ref(get_main_ref_store(the_repository), "BISECT_EXPECTED_REV", &expected_oid))
731 return 0;
732 return oideq(oid, &expected_oid);
733 }
734
735 enum bisect_error bisect_checkout(const struct object_id *bisect_rev,
736 int no_checkout)
737 {
738 struct commit *commit;
739 struct pretty_print_context pp = {0};
740 struct strbuf commit_msg = STRBUF_INIT;
741
742 refs_update_ref(get_main_ref_store(the_repository), NULL,
743 "BISECT_EXPECTED_REV", bisect_rev, NULL, 0,
744 UPDATE_REFS_DIE_ON_ERR);
745
746 if (no_checkout) {
747 refs_update_ref(get_main_ref_store(the_repository), NULL,
748 "BISECT_HEAD", bisect_rev, NULL, 0,
749 UPDATE_REFS_DIE_ON_ERR);
750 } else {
751 struct child_process cmd = CHILD_PROCESS_INIT;
752
753 cmd.git_cmd = 1;
754 strvec_pushl(&cmd.args, "checkout", "-q",
755 oid_to_hex(bisect_rev), "--", NULL);
756 if (run_command(&cmd))
757 /*
758 * Errors in `run_command()` itself, signaled by res < 0,
759 * and errors in the child process, signaled by res > 0
760 * can both be treated as regular BISECT_FAILED (-1).
761 */
762 return BISECT_FAILED;
763 }
764
765 commit = lookup_commit_reference(the_repository, bisect_rev);
766 repo_format_commit_message(the_repository, commit, "[%H] %s%n",
767 &commit_msg, &pp);
768 fputs(commit_msg.buf, stdout);
769 strbuf_release(&commit_msg);
770
771 return BISECT_OK;
772 }
773
774 static struct commit *get_commit_reference(struct repository *r,
775 const struct object_id *oid)
776 {
777 struct commit *c = lookup_commit_reference(r, oid);
778 if (!c)
779 die(_("Not a valid commit name %s"), oid_to_hex(oid));
780 return c;
781 }
782
783 static struct commit **get_bad_and_good_commits(struct repository *r,
784 size_t *rev_nr)
785 {
786 struct commit **rev;
787 size_t i, n = 0;
788
789 ALLOC_ARRAY(rev, 1 + good_revs.nr);
790 rev[n++] = get_commit_reference(r, current_bad_oid);
791 for (i = 0; i < good_revs.nr; i++)
792 rev[n++] = get_commit_reference(r, good_revs.oid + i);
793 *rev_nr = n;
794
795 return rev;
796 }
797
798 static enum bisect_error handle_bad_merge_base(void)
799 {
800 if (is_expected_rev(current_bad_oid)) {
801 char *bad_hex = oid_to_hex(current_bad_oid);
802 char *good_hex = join_oid_array_hex(&good_revs, ' ');
803 if (!strcmp(term_bad, "bad") && !strcmp(term_good, "good")) {
804 fprintf(stderr, _("The merge base %s is bad.\n"
805 "This means the bug has been fixed "
806 "between %s and [%s].\n"),
807 bad_hex, bad_hex, good_hex);
808 } else if (!strcmp(term_bad, "new") && !strcmp(term_good, "old")) {
809 fprintf(stderr, _("The merge base %s is new.\n"
810 "The property has changed "
811 "between %s and [%s].\n"),
812 bad_hex, bad_hex, good_hex);
813 } else {
814 fprintf(stderr, _("The merge base %s is '%s'.\n"
815 "This means the first '%s' commit is "
816 "between %s and [%s].\n"),
817 bad_hex, term_bad, term_good, bad_hex, good_hex);
818 }
819
820 free(good_hex);
821 return BISECT_MERGE_BASE_CHECK;
822 }
823
824 fprintf(stderr, _("Some '%s' revs are not ancestors of the '%s' rev.\n"
825 "git bisect cannot work properly in this case.\n"
826 "Maybe you mistook '%s' and '%s' revs?\n"),
827 term_good, term_bad, term_good, term_bad);
828 return BISECT_FAILED;
829 }
830
831 static void handle_skipped_merge_base(const struct object_id *mb)
832 {
833 char *mb_hex = oid_to_hex(mb);
834 char *bad_hex = oid_to_hex(current_bad_oid);
835 char *good_hex = join_oid_array_hex(&good_revs, ' ');
836
837 warning(_("the merge base between %s and [%s] "
838 "must be skipped.\n"
839 "So we cannot be sure the first '%s' commit is "
840 "between %s and %s.\n"
841 "We continue anyway."),
842 bad_hex, good_hex, term_bad, mb_hex, bad_hex);
843 free(good_hex);
844 }
845
846 /*
847 * "check_merge_bases" checks that merge bases are not "bad" (or "new").
848 *
849 * - If one is "bad" (or "new"), it means the user assumed something wrong
850 * and we must return error with a non 0 error code.
851 * - If one is "good" (or "old"), that's good, we have nothing to do.
852 * - If one is "skipped", we can't know but we should warn.
853 * - If we don't know, we should check it out and ask the user to test.
854 * - If a merge base must be tested, on success return
855 * BISECT_INTERNAL_SUCCESS_MERGE_BASE (-11) a special condition
856 * for early success, this will be converted back to 0 in
857 * check_good_are_ancestors_of_bad().
858 */
859 static enum bisect_error check_merge_bases(size_t rev_nr, struct commit **rev, int no_checkout)
860 {
861 enum bisect_error res = BISECT_OK;
862 struct commit_list *result = NULL;
863
864 if (repo_get_merge_bases_many(the_repository, rev[0], rev_nr - 1,
865 rev + 1, &result) < 0)
866 exit(128);
867
868 for (struct commit_list *l = result; l; l = l->next) {
869 const struct object_id *mb = &l->item->object.oid;
870 if (oideq(mb, current_bad_oid)) {
871 res = handle_bad_merge_base();
872 break;
873 } else if (0 <= oid_array_lookup(&good_revs, mb)) {
874 continue;
875 } else if (0 <= oid_array_lookup(&skipped_revs, mb)) {
876 handle_skipped_merge_base(mb);
877 } else {
878 printf(_("Bisecting: a merge base must be tested\n"));
879 res = bisect_checkout(mb, no_checkout);
880 if (!res)
881 /* indicate early success */
882 res = BISECT_INTERNAL_SUCCESS_MERGE_BASE;
883 break;
884 }
885 }
886
887 commit_list_free(result);
888 return res;
889 }
890
891 static int check_ancestors(struct repository *r, size_t rev_nr,
892 struct commit **rev, const char *prefix)
893 {
894 struct strvec rev_argv = STRVEC_INIT;
895 struct rev_info revs;
896 int res;
897
898 bisect_rev_setup(r, &revs, &rev_argv, prefix, "^%s", "%s", 0);
899
900 bisect_common(&revs);
901 res = (revs.commits != NULL);
902
903 /* Clean up objects used, as they will be reused. */
904 clear_commit_marks_many(rev_nr, rev, ALL_REV_FLAGS);
905
906 release_revisions(&revs);
907 strvec_clear(&rev_argv);
908 return res;
909 }
910
911 /*
912 * "check_good_are_ancestors_of_bad" checks that all "good" revs are
913 * ancestor of the "bad" rev.
914 *
915 * If that's not the case, we need to check the merge bases.
916 * If a merge base must be tested by the user, its source code will be
917 * checked out to be tested by the user and we will return.
918 */
919
920 static enum bisect_error check_good_are_ancestors_of_bad(struct repository *r,
921 const char *prefix,
922 int no_checkout)
923 {
924 char *filename;
925 struct stat st;
926 int fd;
927 size_t rev_nr;
928 enum bisect_error res = BISECT_OK;
929 struct commit **rev;
930
931 if (!current_bad_oid)
932 return error(_("a '%s' revision is needed"), term_bad);
933
934 filename = repo_git_path(the_repository, "BISECT_ANCESTORS_OK");
935
936 /* Check if file BISECT_ANCESTORS_OK exists. */
937 if (!stat(filename, &st) && S_ISREG(st.st_mode))
938 goto done;
939
940 /* Bisecting with no good rev is ok. */
941 if (!good_revs.nr)
942 goto done;
943
944 /* Check if all good revs are ancestor of the bad rev. */
945
946 rev = get_bad_and_good_commits(r, &rev_nr);
947 if (check_ancestors(r, rev_nr, rev, prefix))
948 res = check_merge_bases(rev_nr, rev, no_checkout);
949 free(rev);
950
951 if (!res) {
952 /* Create file BISECT_ANCESTORS_OK. */
953 fd = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0600);
954 if (fd < 0)
955 /*
956 * BISECT_ANCESTORS_OK file is not absolutely necessary,
957 * the bisection process will continue at the next
958 * bisection step.
959 * So, just signal with a warning that something
960 * might be wrong.
961 */
962 warning_errno(_("could not create file '%s'"),
963 filename);
964 else
965 close(fd);
966 }
967 done:
968 free(filename);
969 return res;
970 }
971
972 /*
973 * Display a commit summary to the user.
974 */
975 static void show_commit(struct commit *commit)
976 {
977 struct child_process show = CHILD_PROCESS_INIT;
978
979 /*
980 * Call git show with --no-pager, as it would otherwise
981 * paginate the "git show" output only, not the output
982 * from bisect_next_all(); this can be fixed by moving
983 * it into a --format parameter, but that would override
984 * the user's default options for "git show", which we
985 * are trying to honour.
986 */
987 strvec_pushl(&show.args,
988 "--no-pager",
989 "show",
990 "--stat",
991 "--summary",
992 "--no-abbrev-commit",
993 "--diff-merges=first-parent",
994 oid_to_hex(&commit->object.oid), NULL);
995 show.git_cmd = 1;
996 if (run_command(&show))
997 die(_("unable to start 'show' for object '%s'"),
998 oid_to_hex(&commit->object.oid));
999 }
1000
1001 /*
1002 * The terms used for this bisect session are stored in BISECT_TERMS.
1003 * We read them and store them to adapt the messages accordingly.
1004 * Default is bad/good.
1005 */
1006 void read_bisect_terms(char **read_bad, char **read_good)
1007 {
1008 struct strbuf str = STRBUF_INIT;
1009 const char *filename = git_path_bisect_terms();
1010 FILE *fp = fopen(filename, "r");
1011
1012 if (!fp) {
1013 if (errno == ENOENT) {
1014 free(*read_bad);
1015 *read_bad = xstrdup("bad");
1016 free(*read_good);
1017 *read_good = xstrdup("good");
1018 return;
1019 } else {
1020 die_errno(_("could not read file '%s'"), filename);
1021 }
1022 } else {
1023 if (strbuf_getline_lf(&str, fp) == EOF)
1024 die(_("could not read bad term from file '%s'"), filename);
1025 free(*read_bad);
1026 *read_bad = strbuf_detach(&str, NULL);
1027 if (strbuf_getline_lf(&str, fp) == EOF)
1028 die(_("could not read good term from file '%s'"), filename);
1029 free(*read_good);
1030 *read_good = strbuf_detach(&str, NULL);
1031 }
1032 strbuf_release(&str);
1033 fclose(fp);
1034 }
1035
1036 /*
1037 * We use the convention that return BISECT_INTERNAL_SUCCESS_1ST_BAD_FOUND (-10) means
1038 * the bisection process finished successfully.
1039 * In this case the calling function or command should not turn a
1040 * BISECT_INTERNAL_SUCCESS_1ST_BAD_FOUND return code into an error or a non zero exit code.
1041 *
1042 * Checking BISECT_INTERNAL_SUCCESS_1ST_BAD_FOUND
1043 * in bisect_helper::bisect_next() and only transforming it to 0 at
1044 * the end of bisect_helper::cmd_bisect__helper() helps bypassing
1045 * all the code related to finding a commit to test.
1046 */
1047 enum bisect_error bisect_next_all(struct repository *r, const char *prefix)
1048 {
1049 struct strvec rev_argv = STRVEC_INIT;
1050 struct rev_info revs = REV_INFO_INIT;
1051 struct commit_list *tried = NULL;
1052 int reaches = 0, all = 0, nr, steps;
1053 enum bisect_error res = BISECT_OK;
1054 struct object_id *bisect_rev;
1055 char *steps_msg;
1056 /*
1057 * If no_checkout is non-zero, the bisection process does not
1058 * checkout the trial commit but instead simply updates BISECT_HEAD.
1059 */
1060 int no_checkout = refs_ref_exists(get_main_ref_store(the_repository),
1061 "BISECT_HEAD");
1062 unsigned bisect_flags = 0;
1063
1064 read_bisect_terms(&term_bad, &term_good);
1065 if (read_bisect_refs())
1066 die(_("reading bisect refs failed"));
1067
1068 if (file_exists(git_path_bisect_first_parent()))
1069 bisect_flags |= FIND_BISECTION_FIRST_PARENT_ONLY;
1070
1071 if (skipped_revs.nr)
1072 bisect_flags |= FIND_BISECTION_ALL;
1073
1074 res = check_good_are_ancestors_of_bad(r, prefix, no_checkout);
1075 if (res)
1076 goto cleanup;
1077
1078 bisect_rev_setup(r, &revs, &rev_argv, prefix, "%s", "^%s", 1);
1079
1080 revs.first_parent_only = !!(bisect_flags & FIND_BISECTION_FIRST_PARENT_ONLY);
1081 revs.limited = 1;
1082
1083 bisect_common(&revs);
1084
1085 find_bisection(&revs.commits, &reaches, &all, bisect_flags);
1086 revs.commits = managed_skipped(revs.commits, &tried);
1087
1088 if (!revs.commits) {
1089 /*
1090 * We should return error here only if the "bad"
1091 * commit is also a "skip" commit.
1092 */
1093 res = error_if_skipped_commits(tried, NULL);
1094 if (res < 0)
1095 goto cleanup;
1096 printf(_("%s was both '%s' and '%s'\n"),
1097 oid_to_hex(current_bad_oid),
1098 term_good,
1099 term_bad);
1100
1101 res = BISECT_FAILED;
1102 goto cleanup;
1103 }
1104
1105 if (!all) {
1106 fprintf(stderr, _("No testable commit found.\n"
1107 "Maybe you started with bad path arguments?\n"));
1108
1109 res = BISECT_NO_TESTABLE_COMMIT;
1110 goto cleanup;
1111 }
1112
1113 bisect_rev = &revs.commits->item->object.oid;
1114
1115 if (oideq(bisect_rev, current_bad_oid)) {
1116 res = error_if_skipped_commits(tried, current_bad_oid);
1117 if (res)
1118 goto cleanup;
1119 printf("%s is the first '%s' commit\n", oid_to_hex(bisect_rev),
1120 term_bad);
1121
1122 show_commit(revs.commits->item);
1123 /*
1124 * This means the bisection process succeeded.
1125 * Using BISECT_INTERNAL_SUCCESS_1ST_BAD_FOUND (-10)
1126 * so that the call chain can simply check
1127 * for negative return values for early returns up
1128 * until the cmd_bisect__helper() caller.
1129 */
1130 res = BISECT_INTERNAL_SUCCESS_1ST_BAD_FOUND;
1131 goto cleanup;
1132 }
1133
1134 nr = all - reaches - 1;
1135 steps = estimate_bisect_steps(all);
1136
1137 steps_msg = xstrfmt(Q_("(roughly %d step)", "(roughly %d steps)",
1138 steps), steps);
1139 /*
1140 * TRANSLATORS: the last %s will be replaced with "(roughly %d
1141 * steps)" translation.
1142 */
1143 printf(Q_("Bisecting: %d revision left to test after this %s\n",
1144 "Bisecting: %d revisions left to test after this %s\n",
1145 nr), nr, steps_msg);
1146 free(steps_msg);
1147 /* Clean up objects used, as they will be reused. */
1148 repo_clear_commit_marks(r, ALL_REV_FLAGS);
1149
1150 res = bisect_checkout(bisect_rev, no_checkout);
1151 cleanup:
1152 commit_list_free(tried);
1153 release_revisions(&revs);
1154 strvec_clear(&rev_argv);
1155 return res;
1156 }
1157
1158 static inline int exp2i(int n)
1159 {
1160 return 1 << n;
1161 }
1162
1163 /*
1164 * Estimate the number of bisect steps left (after the current step)
1165 *
1166 * For any x between 0 included and 2^n excluded, the probability for
1167 * n - 1 steps left looks like:
1168 *
1169 * P(2^n + x) == (2^n - x) / (2^n + x)
1170 *
1171 * and P(2^n + x) < 0.5 means 2^n < 3x
1172 */
1173 int estimate_bisect_steps(int all)
1174 {
1175 int n, x, e;
1176
1177 if (all < 3)
1178 return 0;
1179
1180 n = log2u(all);
1181 e = exp2i(n);
1182 x = all - e;
1183
1184 return (e < 3 * x) ? n : n - 1;
1185 }
1186
1187 static int mark_for_removal(const struct reference *ref, void *cb_data)
1188 {
1189 struct string_list *refs = cb_data;
1190 string_list_append(refs, ref->name);
1191 return 0;
1192 }
1193
1194 int bisect_clean_state(void)
1195 {
1196 struct refs_for_each_ref_options opts = {
1197 .prefix = "refs/bisect/",
1198 };
1199 int result = 0;
1200
1201 /* There may be some refs packed during bisection */
1202 struct string_list refs_for_removal = STRING_LIST_INIT_DUP;
1203 refs_for_each_ref_ext(get_main_ref_store(the_repository),
1204 mark_for_removal, &refs_for_removal, &opts);
1205 string_list_append(&refs_for_removal, "BISECT_HEAD");
1206 string_list_append(&refs_for_removal, "BISECT_EXPECTED_REV");
1207 result = refs_delete_refs(get_main_ref_store(the_repository),
1208 "bisect: remove", &refs_for_removal,
1209 REF_NO_DEREF);
1210 string_list_clear(&refs_for_removal, 0);
1211 unlink_or_warn(git_path_bisect_ancestors_ok());
1212 unlink_or_warn(git_path_bisect_log());
1213 unlink_or_warn(git_path_bisect_names());
1214 unlink_or_warn(git_path_bisect_run());
1215 unlink_or_warn(git_path_bisect_terms());
1216 unlink_or_warn(git_path_bisect_first_parent());
1217 unlink_or_warn(git_path_bisect_reset_when_found());
1218 /*
1219 * Cleanup BISECT_START last to support the --no-checkout option
1220 * introduced in the commit 4796e823a.
1221 */
1222 unlink_or_warn(git_path_bisect_start());
1223
1224 return result;
1225 }