Raw
1 #define USE_THE_REPOSITORY_VARIABLE
2 #define DISABLE_SIGN_COMPARE_WARNINGS
3
4 #include "git-compat-util.h"
5 #include "abspath.h"
6 #include "config.h"
7 #include "copy.h"
8 #include "environment.h"
9 #include "gettext.h"
10 #include "hex.h"
11 #include "lockfile.h"
12 #include "string-list.h"
13 #include "read-cache-ll.h"
14 #include "rerere.h"
15 #include "xdiff-interface.h"
16 #include "dir.h"
17 #include "resolve-undo.h"
18 #include "merge-ll.h"
19 #include "path.h"
20 #include "pathspec.h"
21 #include "object-file.h"
22 #include "odb.h"
23 #include "strmap.h"
24
25 #define RESOLVED 0
26 #define PUNTED 1
27 #define THREE_STAGED 2
28 void *RERERE_RESOLVED = &RERERE_RESOLVED;
29
30 /* if rerere_enabled == -1, fall back to detection of .git/rr-cache */
31 static int rerere_enabled = -1;
32
33 /* automatically update cleanly resolved paths to the index */
34 static int rerere_autoupdate;
35
36 #define RR_HAS_POSTIMAGE 1
37 #define RR_HAS_PREIMAGE 2
38 struct rerere_dir {
39 int status_alloc, status_nr;
40 unsigned char *status;
41 char name[FLEX_ARRAY];
42 };
43
44 static struct strmap rerere_dirs = STRMAP_INIT;
45
46 static void free_rerere_dirs(void)
47 {
48 struct hashmap_iter iter;
49 struct strmap_entry *ent;
50
51 strmap_for_each_entry(&rerere_dirs, &iter, ent) {
52 struct rerere_dir *rr_dir = ent->value;
53 free(rr_dir->status);
54 free(rr_dir);
55 }
56 strmap_clear(&rerere_dirs, 0);
57 }
58
59 static void free_rerere_id(struct string_list_item *item)
60 {
61 free(item->util);
62 }
63
64 static const char *rerere_id_hex(const struct rerere_id *id)
65 {
66 return id->collection->name;
67 }
68
69 static void fit_variant(struct rerere_dir *rr_dir, int variant)
70 {
71 variant++;
72 ALLOC_GROW(rr_dir->status, variant, rr_dir->status_alloc);
73 if (rr_dir->status_nr < variant) {
74 memset(rr_dir->status + rr_dir->status_nr,
75 '\0', variant - rr_dir->status_nr);
76 rr_dir->status_nr = variant;
77 }
78 }
79
80 static void assign_variant(struct rerere_id *id)
81 {
82 int variant;
83 struct rerere_dir *rr_dir = id->collection;
84
85 variant = id->variant;
86 if (variant < 0) {
87 for (variant = 0; variant < rr_dir->status_nr; variant++)
88 if (!rr_dir->status[variant])
89 break;
90 }
91 fit_variant(rr_dir, variant);
92 id->variant = variant;
93 }
94
95 const char *rerere_path(struct strbuf *buf, const struct rerere_id *id, const char *file)
96 {
97 if (!file)
98 return repo_git_path_replace(the_repository, buf, "rr-cache/%s",
99 rerere_id_hex(id));
100
101 if (id->variant <= 0)
102 return repo_git_path_replace(the_repository, buf, "rr-cache/%s/%s",
103 rerere_id_hex(id), file);
104
105 return repo_git_path_replace(the_repository, buf, "rr-cache/%s/%s.%d",
106 rerere_id_hex(id), file, id->variant);
107 }
108
109 static int is_rr_file(const char *name, const char *filename, int *variant)
110 {
111 const char *suffix;
112 char *ep;
113
114 if (!strcmp(name, filename)) {
115 *variant = 0;
116 return 1;
117 }
118 if (!skip_prefix(name, filename, &suffix) || *suffix != '.')
119 return 0;
120
121 errno = 0;
122 *variant = strtol(suffix + 1, &ep, 10);
123 if (errno || *ep)
124 return 0;
125 return 1;
126 }
127
128 static void scan_rerere_dir(struct rerere_dir *rr_dir)
129 {
130 struct dirent *de;
131 char *path;
132 DIR *dir;
133
134 path = repo_git_path(the_repository, "rr-cache/%s", rr_dir->name);
135 dir = opendir(path);
136 free(path);
137 if (!dir)
138 return;
139 while ((de = readdir(dir)) != NULL) {
140 int variant;
141
142 if (is_rr_file(de->d_name, "postimage", &variant)) {
143 fit_variant(rr_dir, variant);
144 rr_dir->status[variant] |= RR_HAS_POSTIMAGE;
145 } else if (is_rr_file(de->d_name, "preimage", &variant)) {
146 fit_variant(rr_dir, variant);
147 rr_dir->status[variant] |= RR_HAS_PREIMAGE;
148 }
149 }
150 closedir(dir);
151 }
152
153 static struct rerere_dir *find_rerere_dir(const char *hex)
154 {
155 struct rerere_dir *rr_dir;
156
157 rr_dir = strmap_get(&rerere_dirs, hex);
158 if (!rr_dir) {
159 FLEX_ALLOC_STR(rr_dir, name, hex);
160 rr_dir->status = NULL;
161 rr_dir->status_nr = 0;
162 rr_dir->status_alloc = 0;
163 strmap_put(&rerere_dirs, hex, rr_dir);
164
165 scan_rerere_dir(rr_dir);
166 }
167 return rr_dir;
168 }
169
170 static int has_rerere_resolution(const struct rerere_id *id)
171 {
172 const int both = RR_HAS_POSTIMAGE|RR_HAS_PREIMAGE;
173 int variant = id->variant;
174
175 if (variant < 0)
176 return 0;
177 return ((id->collection->status[variant] & both) == both);
178 }
179
180 static struct rerere_id *new_rerere_id_hex(char *hex)
181 {
182 struct rerere_id *id = xmalloc(sizeof(*id));
183 id->collection = find_rerere_dir(hex);
184 id->variant = -1; /* not known yet */
185 return id;
186 }
187
188 static struct rerere_id *new_rerere_id(unsigned char *hash)
189 {
190 return new_rerere_id_hex(hash_to_hex(hash));
191 }
192
193 /*
194 * $GIT_DIR/MERGE_RR file is a collection of records, each of which is
195 * "conflict ID", a HT and pathname, terminated with a NUL, and is
196 * used to keep track of the set of paths that "rerere" may need to
197 * work on (i.e. what is left by the previous invocation of "git
198 * rerere" during the current conflict resolution session).
199 */
200 static void read_rr(struct repository *r, struct string_list *rr)
201 {
202 struct strbuf buf = STRBUF_INIT;
203 FILE *in = fopen_or_warn(git_path_merge_rr(r), "r");
204
205 if (!in)
206 return;
207 while (!strbuf_getwholeline(&buf, in, '\0')) {
208 char *path;
209 unsigned char hash[GIT_MAX_RAWSZ];
210 struct rerere_id *id;
211 int variant;
212 const unsigned hexsz = the_hash_algo->hexsz;
213
214 /* There has to be the hash, tab, path and then NUL */
215 if (buf.len < hexsz + 2 || get_hash_hex(buf.buf, hash))
216 die(_("corrupt MERGE_RR"));
217
218 if (buf.buf[hexsz] != '.') {
219 variant = 0;
220 path = buf.buf + hexsz;
221 } else {
222 errno = 0;
223 variant = strtol(buf.buf + hexsz + 1, &path, 10);
224 if (errno)
225 die(_("corrupt MERGE_RR"));
226 }
227 if (*(path++) != '\t')
228 die(_("corrupt MERGE_RR"));
229 buf.buf[hexsz] = '\0';
230 id = new_rerere_id_hex(buf.buf);
231 id->variant = variant;
232 /*
233 * make sure id->collection->status has enough space
234 * for the variant we are interested in
235 */
236 fit_variant(id->collection, variant);
237 string_list_insert(rr, path)->util = id;
238 }
239 strbuf_release(&buf);
240 fclose(in);
241 }
242
243 static struct lock_file write_lock;
244
245 static int write_rr(struct string_list *rr, int out_fd)
246 {
247 int i;
248 for (i = 0; i < rr->nr; i++) {
249 struct strbuf buf = STRBUF_INIT;
250 struct rerere_id *id;
251
252 assert(rr->items[i].util != RERERE_RESOLVED);
253
254 id = rr->items[i].util;
255 if (!id)
256 continue;
257 assert(id->variant >= 0);
258 if (0 < id->variant)
259 strbuf_addf(&buf, "%s.%d\t%s%c",
260 rerere_id_hex(id), id->variant,
261 rr->items[i].string, 0);
262 else
263 strbuf_addf(&buf, "%s\t%s%c",
264 rerere_id_hex(id),
265 rr->items[i].string, 0);
266
267 if (write_in_full(out_fd, buf.buf, buf.len) < 0)
268 die(_("unable to write rerere record"));
269
270 strbuf_release(&buf);
271 }
272 if (commit_lock_file(&write_lock) != 0)
273 die(_("unable to write rerere record"));
274 return 0;
275 }
276
277 /*
278 * "rerere" interacts with conflicted file contents using this I/O
279 * abstraction. It reads a conflicted contents from one place via
280 * "getline()" method, and optionally can write it out after
281 * normalizing the conflicted hunks to the "output". Subclasses of
282 * rerere_io embed this structure at the beginning of their own
283 * rerere_io object.
284 */
285 struct rerere_io {
286 int (*getline)(struct strbuf *, struct rerere_io *);
287 FILE *output;
288 int wrerror;
289 /* some more stuff */
290 };
291
292 static void ferr_write(const void *p, size_t count, FILE *fp, int *err)
293 {
294 if (!count || *err)
295 return;
296 if (fwrite(p, count, 1, fp) != 1)
297 *err = errno;
298 }
299
300 static inline void ferr_puts(const char *s, FILE *fp, int *err)
301 {
302 ferr_write(s, strlen(s), fp, err);
303 }
304
305 static void rerere_io_putstr(const char *str, struct rerere_io *io)
306 {
307 if (io->output)
308 ferr_puts(str, io->output, &io->wrerror);
309 }
310
311 static void rerere_io_putmem(const char *mem, size_t sz, struct rerere_io *io)
312 {
313 if (io->output)
314 ferr_write(mem, sz, io->output, &io->wrerror);
315 }
316
317 /*
318 * Subclass of rerere_io that reads from an on-disk file
319 */
320 struct rerere_io_file {
321 struct rerere_io io;
322 FILE *input;
323 };
324
325 /*
326 * ... and its getline() method implementation
327 */
328 static int rerere_file_getline(struct strbuf *sb, struct rerere_io *io_)
329 {
330 struct rerere_io_file *io = (struct rerere_io_file *)io_;
331 return strbuf_getwholeline(sb, io->input, '\n');
332 }
333
334 /*
335 * Require the exact number of conflict marker letters, no more, no
336 * less, followed by SP or any whitespace
337 * (including LF).
338 */
339 static int is_cmarker(char *buf, int marker_char, int marker_size)
340 {
341 int want_sp;
342
343 /*
344 * The beginning of our version and the end of their version
345 * always are labeled like "<<<<< ours" or ">>>>> theirs",
346 * hence we set want_sp for them. Note that the version from
347 * the common ancestor in diff3-style output is not always
348 * labelled (e.g. "||||| common" is often seen but "|||||"
349 * alone is also valid), so we do not set want_sp.
350 */
351 want_sp = (marker_char == '<') || (marker_char == '>');
352
353 while (marker_size--)
354 if (*buf++ != marker_char)
355 return 0;
356 if (want_sp && *buf != ' ')
357 return 0;
358 return isspace(*buf);
359 }
360
361 static void rerere_strbuf_putconflict(struct strbuf *buf, int ch, size_t size)
362 {
363 strbuf_addchars(buf, ch, size);
364 strbuf_addch(buf, '\n');
365 }
366
367 static int handle_conflict(struct strbuf *out, struct rerere_io *io,
368 int marker_size, struct git_hash_ctx *ctx)
369 {
370 enum {
371 RR_SIDE_1 = 0, RR_SIDE_2, RR_ORIGINAL
372 } hunk = RR_SIDE_1;
373 struct strbuf one = STRBUF_INIT, two = STRBUF_INIT;
374 struct strbuf buf = STRBUF_INIT, conflict = STRBUF_INIT;
375 int has_conflicts = -1;
376
377 while (!io->getline(&buf, io)) {
378 if (is_cmarker(buf.buf, '<', marker_size)) {
379 if (handle_conflict(&conflict, io, marker_size, NULL) < 0)
380 break;
381 if (hunk == RR_SIDE_1)
382 strbuf_addbuf(&one, &conflict);
383 else
384 strbuf_addbuf(&two, &conflict);
385 strbuf_release(&conflict);
386 } else if (is_cmarker(buf.buf, '|', marker_size)) {
387 if (hunk != RR_SIDE_1)
388 break;
389 hunk = RR_ORIGINAL;
390 } else if (is_cmarker(buf.buf, '=', marker_size)) {
391 if (hunk != RR_SIDE_1 && hunk != RR_ORIGINAL)
392 break;
393 hunk = RR_SIDE_2;
394 } else if (is_cmarker(buf.buf, '>', marker_size)) {
395 if (hunk != RR_SIDE_2)
396 break;
397 if (strbuf_cmp(&one, &two) > 0)
398 strbuf_swap(&one, &two);
399 has_conflicts = 1;
400 rerere_strbuf_putconflict(out, '<', marker_size);
401 strbuf_addbuf(out, &one);
402 rerere_strbuf_putconflict(out, '=', marker_size);
403 strbuf_addbuf(out, &two);
404 rerere_strbuf_putconflict(out, '>', marker_size);
405 if (ctx) {
406 git_hash_update(ctx, one.buf, one.len + 1);
407 git_hash_update(ctx, two.buf, two.len + 1);
408 }
409 break;
410 } else if (hunk == RR_SIDE_1)
411 strbuf_addbuf(&one, &buf);
412 else if (hunk == RR_ORIGINAL)
413 ; /* discard */
414 else if (hunk == RR_SIDE_2)
415 strbuf_addbuf(&two, &buf);
416 }
417 strbuf_release(&one);
418 strbuf_release(&two);
419 strbuf_release(&buf);
420
421 return has_conflicts;
422 }
423
424 /*
425 * Read contents a file with conflicts, normalize the conflicts
426 * by (1) discarding the common ancestor version in diff3-style,
427 * (2) reordering our side and their side so that whichever sorts
428 * alphabetically earlier comes before the other one, while
429 * computing the "conflict ID", which is just an SHA-1 hash of
430 * one side of the conflict, NUL, the other side of the conflict,
431 * and NUL concatenated together.
432 *
433 * Return 1 if conflict hunks are found, 0 if there are no conflict
434 * hunks and -1 if an error occurred.
435 */
436 static int handle_path(unsigned char *hash, struct rerere_io *io, int marker_size)
437 {
438 struct git_hash_ctx ctx;
439 struct strbuf buf = STRBUF_INIT, out = STRBUF_INIT;
440 int has_conflicts = 0;
441 if (hash)
442 the_hash_algo->init_fn(&ctx);
443
444 while (!io->getline(&buf, io)) {
445 if (is_cmarker(buf.buf, '<', marker_size)) {
446 has_conflicts = handle_conflict(&out, io, marker_size,
447 hash ? &ctx : NULL);
448 if (has_conflicts < 0)
449 break;
450 rerere_io_putmem(out.buf, out.len, io);
451 strbuf_reset(&out);
452 } else
453 rerere_io_putstr(buf.buf, io);
454 }
455 strbuf_release(&buf);
456 strbuf_release(&out);
457
458 if (hash)
459 git_hash_final(hash, &ctx);
460
461 return has_conflicts;
462 }
463
464 /*
465 * Scan the path for conflicts, do the "handle_path()" thing above, and
466 * return the number of conflict hunks found.
467 */
468 static int handle_file(struct index_state *istate,
469 const char *path, unsigned char *hash, const char *output)
470 {
471 int has_conflicts = 0;
472 struct rerere_io_file io;
473 int marker_size = ll_merge_marker_size(istate, path);
474
475 memset(&io, 0, sizeof(io));
476 io.io.getline = rerere_file_getline;
477 io.input = fopen(path, "r");
478 io.io.wrerror = 0;
479 if (!io.input)
480 return error_errno(_("could not open '%s'"), path);
481
482 if (output) {
483 io.io.output = fopen(output, "w");
484 if (!io.io.output) {
485 error_errno(_("could not write '%s'"), output);
486 fclose(io.input);
487 return -1;
488 }
489 }
490
491 has_conflicts = handle_path(hash, (struct rerere_io *)&io, marker_size);
492
493 fclose(io.input);
494 if (io.io.wrerror)
495 error(_("there were errors while writing '%s' (%s)"),
496 path, strerror(io.io.wrerror));
497 if (io.io.output && fclose(io.io.output))
498 io.io.wrerror = error_errno(_("failed to flush '%s'"), path);
499
500 if (has_conflicts < 0) {
501 if (output)
502 unlink_or_warn(output);
503 return error(_("could not parse conflict hunks in '%s'"), path);
504 }
505 if (io.io.wrerror)
506 return -1;
507 return has_conflicts;
508 }
509
510 /*
511 * Look at a cache entry at "i" and see if it is not conflicting,
512 * conflicting and we are willing to handle, or conflicting and
513 * we are unable to handle, and return the determination in *type.
514 * Return the cache index to be looked at next, by skipping the
515 * stages we have already looked at in this invocation of this
516 * function.
517 */
518 static int check_one_conflict(struct index_state *istate, int i, int *type)
519 {
520 const struct cache_entry *e = istate->cache[i];
521
522 if (!ce_stage(e)) {
523 *type = RESOLVED;
524 return i + 1;
525 }
526
527 *type = PUNTED;
528 while (i < istate->cache_nr && ce_stage(istate->cache[i]) == 1)
529 i++;
530
531 /* Only handle regular files with both stages #2 and #3 */
532 if (i + 1 < istate->cache_nr) {
533 const struct cache_entry *e2 = istate->cache[i];
534 const struct cache_entry *e3 = istate->cache[i + 1];
535 if (ce_stage(e2) == 2 &&
536 ce_stage(e3) == 3 &&
537 ce_same_name(e, e3) &&
538 S_ISREG(e2->ce_mode) &&
539 S_ISREG(e3->ce_mode))
540 *type = THREE_STAGED;
541 }
542
543 /* Skip the entries with the same name */
544 while (i < istate->cache_nr && ce_same_name(e, istate->cache[i]))
545 i++;
546 return i;
547 }
548
549 /*
550 * Scan the index and find paths that have conflicts that rerere can
551 * handle, i.e. the ones that have both stages #2 and #3.
552 *
553 * NEEDSWORK: we do not record or replay a previous "resolve by
554 * deletion" for a delete-modify conflict, as that is inherently risky
555 * without knowing what modification is being discarded. The only
556 * safe case, i.e. both side doing the deletion and modification that
557 * are identical to the previous round, might want to be handled,
558 * though.
559 */
560 static int find_conflict(struct repository *r, struct string_list *conflict)
561 {
562 int i;
563
564 if (repo_read_index(r) < 0)
565 return error(_("index file corrupt"));
566
567 for (i = 0; i < r->index->cache_nr;) {
568 int conflict_type;
569 const struct cache_entry *e = r->index->cache[i];
570 i = check_one_conflict(r->index, i, &conflict_type);
571 if (conflict_type == THREE_STAGED)
572 string_list_insert(conflict, (const char *)e->name);
573 }
574 return 0;
575 }
576
577 /*
578 * The merge_rr list is meant to hold outstanding conflicted paths
579 * that rerere could handle. Abuse the list by adding other types of
580 * entries to allow the caller to show "rerere remaining".
581 *
582 * - Conflicted paths that rerere does not handle are added
583 * - Conflicted paths that have been resolved are marked as such
584 * by storing RERERE_RESOLVED to .util field (where conflict ID
585 * is expected to be stored).
586 *
587 * Do *not* write MERGE_RR file out after calling this function.
588 *
589 * NEEDSWORK: we may want to fix the caller that implements "rerere
590 * remaining" to do this without abusing merge_rr.
591 */
592 int rerere_remaining(struct repository *r, struct string_list *merge_rr)
593 {
594 int i;
595
596 if (setup_rerere(r, merge_rr, RERERE_READONLY))
597 return 0;
598 if (repo_read_index(r) < 0)
599 return error(_("index file corrupt"));
600
601 for (i = 0; i < r->index->cache_nr;) {
602 int conflict_type;
603 const struct cache_entry *e = r->index->cache[i];
604 i = check_one_conflict(r->index, i, &conflict_type);
605 if (conflict_type == PUNTED)
606 string_list_insert(merge_rr, (const char *)e->name);
607 else if (conflict_type == RESOLVED) {
608 struct string_list_item *it;
609 it = string_list_lookup(merge_rr, (const char *)e->name);
610 if (it) {
611 free_rerere_id(it);
612 it->util = RERERE_RESOLVED;
613 }
614 }
615 }
616 return 0;
617 }
618
619 /*
620 * Try using the given conflict resolution "ID" to see
621 * if that recorded conflict resolves cleanly what we
622 * got in the "cur".
623 */
624 static int try_merge(struct index_state *istate,
625 const struct rerere_id *id, const char *path,
626 mmfile_t *cur, mmbuffer_t *result)
627 {
628 enum ll_merge_result ret;
629 mmfile_t base = {NULL, 0}, other = {NULL, 0};
630 struct strbuf buf = STRBUF_INIT;
631
632 if (read_mmfile(&base, rerere_path(&buf, id, "preimage")) ||
633 read_mmfile(&other, rerere_path(&buf, id, "postimage"))) {
634 ret = LL_MERGE_CONFLICT;
635 } else {
636 /*
637 * A three-way merge. Note that this honors user-customizable
638 * low-level merge driver settings.
639 */
640 ret = ll_merge(result, path, &base, NULL, cur, "", &other, "",
641 istate, NULL);
642 }
643
644 strbuf_release(&buf);
645 free(base.ptr);
646 free(other.ptr);
647
648 return ret;
649 }
650
651 /*
652 * Find the conflict identified by "id"; the change between its
653 * "preimage" (i.e. a previous contents with conflict markers) and its
654 * "postimage" (i.e. the corresponding contents with conflicts
655 * resolved) may apply cleanly to the contents stored in "path", i.e.
656 * the conflict this time around.
657 *
658 * Returns 0 for successful replay of recorded resolution, or non-zero
659 * for failure.
660 */
661 static int merge(struct index_state *istate, const struct rerere_id *id, const char *path)
662 {
663 FILE *f;
664 int ret;
665 struct strbuf buf = STRBUF_INIT;
666 mmfile_t cur = {NULL, 0};
667 mmbuffer_t result = {NULL, 0};
668
669 /*
670 * Normalize the conflicts in path and write it out to
671 * "thisimage" temporary file.
672 */
673 if ((handle_file(istate, path, NULL, rerere_path(&buf, id, "thisimage")) < 0) ||
674 read_mmfile(&cur, rerere_path(&buf, id, "thisimage"))) {
675 ret = 1;
676 goto out;
677 }
678
679 ret = try_merge(istate, id, path, &cur, &result);
680 if (ret)
681 goto out;
682
683 /*
684 * A successful replay of recorded resolution.
685 * Mark that "postimage" was used to help gc.
686 */
687 if (utime(rerere_path(&buf, id, "postimage"), NULL) < 0)
688 warning_errno(_("failed utime() on '%s'"),
689 rerere_path(&buf, id, "postimage"));
690
691 /* Update "path" with the resolution */
692 f = fopen(path, "w");
693 if (!f)
694 return error_errno(_("could not open '%s'"), path);
695 if (fwrite(result.ptr, result.size, 1, f) != 1)
696 error_errno(_("could not write '%s'"), path);
697 if (fclose(f))
698 return error_errno(_("writing '%s' failed"), path);
699
700 out:
701 free(cur.ptr);
702 free(result.ptr);
703 strbuf_release(&buf);
704
705 return ret;
706 }
707
708 static void update_paths(struct repository *r, struct string_list *update)
709 {
710 struct lock_file index_lock = LOCK_INIT;
711 int i;
712
713 repo_hold_locked_index(r, &index_lock, LOCK_DIE_ON_ERROR);
714
715 for (i = 0; i < update->nr; i++) {
716 struct string_list_item *item = &update->items[i];
717 if (add_file_to_index(r->index, item->string, 0))
718 exit(128);
719 fprintf_ln(stderr, _("Staged '%s' using previous resolution."),
720 item->string);
721 }
722
723 if (write_locked_index(r->index, &index_lock,
724 COMMIT_LOCK | SKIP_IF_UNCHANGED))
725 die(_("unable to write new index file"));
726 }
727
728 static void remove_variant(struct rerere_id *id)
729 {
730 struct strbuf buf = STRBUF_INIT;
731 unlink_or_warn(rerere_path(&buf, id, "postimage"));
732 unlink_or_warn(rerere_path(&buf, id, "preimage"));
733 id->collection->status[id->variant] = 0;
734 strbuf_release(&buf);
735 }
736
737 /*
738 * The path indicated by rr_item may still have conflict for which we
739 * have a recorded resolution, in which case replay it and optionally
740 * update it. Or it may have been resolved by the user and we may
741 * only have the preimage for that conflict, in which case the result
742 * needs to be recorded as a resolution in a postimage file.
743 */
744 static void do_rerere_one_path(struct index_state *istate,
745 struct string_list_item *rr_item,
746 struct string_list *update)
747 {
748 const char *path = rr_item->string;
749 struct rerere_id *id = rr_item->util;
750 struct rerere_dir *rr_dir = id->collection;
751 struct strbuf buf = STRBUF_INIT;
752 int variant;
753
754 variant = id->variant;
755
756 /* Has the user resolved it already? */
757 if (variant >= 0) {
758 if (!handle_file(istate, path, NULL, NULL)) {
759 copy_file(rerere_path(&buf, id, "postimage"), path, 0666);
760 id->collection->status[variant] |= RR_HAS_POSTIMAGE;
761 fprintf_ln(stderr, _("Recorded resolution for '%s'."), path);
762 free_rerere_id(rr_item);
763 rr_item->util = NULL;
764 goto out;
765 }
766 /*
767 * There may be other variants that can cleanly
768 * replay. Try them and update the variant number for
769 * this one.
770 */
771 }
772
773 /* Does any existing resolution apply cleanly? */
774 for (variant = 0; variant < rr_dir->status_nr; variant++) {
775 const int both = RR_HAS_PREIMAGE | RR_HAS_POSTIMAGE;
776 struct rerere_id vid = *id;
777
778 if ((rr_dir->status[variant] & both) != both)
779 continue;
780
781 vid.variant = variant;
782 if (merge(istate, &vid, path))
783 continue; /* failed to replay */
784
785 /*
786 * If there already is a different variant that applies
787 * cleanly, there is no point maintaining our own variant.
788 */
789 if (0 <= id->variant && id->variant != variant)
790 remove_variant(id);
791
792 if (rerere_autoupdate)
793 string_list_insert(update, path);
794 else
795 fprintf_ln(stderr,
796 _("Resolved '%s' using previous resolution."),
797 path);
798 free_rerere_id(rr_item);
799 rr_item->util = NULL;
800 goto out;
801 }
802
803 /* None of the existing one applies; we need a new variant */
804 assign_variant(id);
805
806 variant = id->variant;
807 handle_file(istate, path, NULL, rerere_path(&buf, id, "preimage"));
808 if (id->collection->status[variant] & RR_HAS_POSTIMAGE) {
809 const char *path = rerere_path(&buf, id, "postimage");
810 if (unlink(path))
811 die_errno(_("cannot unlink stray '%s'"), path);
812 id->collection->status[variant] &= ~RR_HAS_POSTIMAGE;
813 }
814 id->collection->status[variant] |= RR_HAS_PREIMAGE;
815 fprintf_ln(stderr, _("Recorded preimage for '%s'"), path);
816
817 out:
818 strbuf_release(&buf);
819 }
820
821 static int do_plain_rerere(struct repository *r,
822 struct string_list *rr, int fd)
823 {
824 struct string_list conflict = STRING_LIST_INIT_DUP;
825 struct string_list update = STRING_LIST_INIT_DUP;
826 struct strbuf buf = STRBUF_INIT;
827 int i;
828
829 find_conflict(r, &conflict);
830
831 /*
832 * MERGE_RR records paths with conflicts immediately after
833 * merge failed. Some of the conflicted paths might have been
834 * hand resolved in the working tree since then, but the
835 * initial run would catch all and register their preimages.
836 */
837 for (i = 0; i < conflict.nr; i++) {
838 struct rerere_id *id;
839 unsigned char hash[GIT_MAX_RAWSZ];
840 const char *path = conflict.items[i].string;
841 int ret;
842
843 /*
844 * Ask handle_file() to scan and assign a
845 * conflict ID. No need to write anything out
846 * yet.
847 */
848 ret = handle_file(r->index, path, hash, NULL);
849 if (ret != 0 && string_list_has_string(rr, path)) {
850 remove_variant(string_list_lookup(rr, path)->util);
851 string_list_remove(rr, path, 1);
852 }
853 if (ret < 1)
854 continue;
855
856 id = new_rerere_id(hash);
857 string_list_insert(rr, path)->util = id;
858
859 /* Ensure that the directory exists. */
860 safe_create_dir_in_gitdir(the_repository, rerere_path(&buf, id, NULL));
861 }
862
863 for (i = 0; i < rr->nr; i++)
864 do_rerere_one_path(r->index, &rr->items[i], &update);
865
866 if (update.nr)
867 update_paths(r, &update);
868
869 string_list_clear(&conflict, 0);
870 string_list_clear(&update, 0);
871 strbuf_release(&buf);
872 return write_rr(rr, fd);
873 }
874
875 static void git_rerere_config(void)
876 {
877 repo_config_get_bool(the_repository, "rerere.enabled", &rerere_enabled);
878 repo_config_get_bool(the_repository, "rerere.autoupdate", &rerere_autoupdate);
879 repo_config(the_repository, git_default_config, NULL);
880 }
881
882 static GIT_PATH_FUNC(git_path_rr_cache, "rr-cache")
883
884 static int is_rerere_enabled(void)
885 {
886 int rr_cache_exists;
887
888 if (!rerere_enabled)
889 return 0;
890
891 rr_cache_exists = is_directory(git_path_rr_cache());
892 if (rerere_enabled < 0)
893 return rr_cache_exists;
894
895 if (!rr_cache_exists &&
896 safe_create_dir_in_gitdir(the_repository, git_path_rr_cache()))
897 die(_("could not create directory '%s'"), git_path_rr_cache());
898 return 1;
899 }
900
901 int setup_rerere(struct repository *r, struct string_list *merge_rr, int flags)
902 {
903 int fd;
904
905 git_rerere_config();
906 if (!is_rerere_enabled())
907 return -1;
908
909 if (flags & (RERERE_AUTOUPDATE|RERERE_NOAUTOUPDATE))
910 rerere_autoupdate = !!(flags & RERERE_AUTOUPDATE);
911 if (flags & RERERE_READONLY)
912 fd = 0;
913 else
914 fd = hold_lock_file_for_update(&write_lock,
915 git_path_merge_rr(r),
916 LOCK_DIE_ON_ERROR);
917 read_rr(r, merge_rr);
918 return fd;
919 }
920
921 /*
922 * The main entry point that is called internally from codepaths that
923 * perform mergy operations, possibly leaving conflicted index entries
924 * and working tree files.
925 */
926 int repo_rerere(struct repository *r, int flags)
927 {
928 struct string_list merge_rr = STRING_LIST_INIT_DUP;
929 int fd, status;
930
931 fd = setup_rerere(r, &merge_rr, flags);
932 if (fd < 0)
933 return 0;
934 status = do_plain_rerere(r, &merge_rr, fd);
935 free_rerere_dirs();
936 string_list_clear(&merge_rr, 1);
937 return status;
938 }
939
940 /*
941 * Subclass of rerere_io that reads from an in-core buffer that is a
942 * strbuf
943 */
944 struct rerere_io_mem {
945 struct rerere_io io;
946 struct strbuf input;
947 };
948
949 /*
950 * ... and its getline() method implementation
951 */
952 static int rerere_mem_getline(struct strbuf *sb, struct rerere_io *io_)
953 {
954 struct rerere_io_mem *io = (struct rerere_io_mem *)io_;
955 char *ep;
956 size_t len;
957
958 strbuf_release(sb);
959 if (!io->input.len)
960 return -1;
961 ep = memchr(io->input.buf, '\n', io->input.len);
962 if (!ep)
963 ep = io->input.buf + io->input.len;
964 else if (*ep == '\n')
965 ep++;
966 len = ep - io->input.buf;
967 strbuf_add(sb, io->input.buf, len);
968 strbuf_remove(&io->input, 0, len);
969 return 0;
970 }
971
972 static int handle_cache(struct index_state *istate,
973 const char *path, unsigned char *hash, const char *output)
974 {
975 mmfile_t mmfile[3] = {{NULL}};
976 mmbuffer_t result = {NULL, 0};
977 const struct cache_entry *ce;
978 int pos, len, i, has_conflicts;
979 struct rerere_io_mem io;
980 int marker_size = ll_merge_marker_size(istate, path);
981
982 /*
983 * Reproduce the conflicted merge in-core
984 */
985 len = strlen(path);
986 pos = index_name_pos(istate, path, len);
987 if (0 <= pos)
988 return -1;
989 pos = -pos - 1;
990
991 while (pos < istate->cache_nr) {
992 enum object_type type;
993 unsigned long size;
994
995 ce = istate->cache[pos++];
996 if (ce_namelen(ce) != len || memcmp(ce->name, path, len))
997 break;
998 i = ce_stage(ce) - 1;
999 if (!mmfile[i].ptr) {
1000 mmfile[i].ptr = odb_read_object(the_repository->objects,
1001 &ce->oid, &type, &size);
1002 if (!mmfile[i].ptr)
1003 die(_("unable to read %s"),
1004 oid_to_hex(&ce->oid));
1005 mmfile[i].size = size;
1006 }
1007 }
1008 for (i = 0; i < 3; i++)
1009 if (!mmfile[i].ptr && !mmfile[i].size)
1010 mmfile[i].ptr = xstrdup("");
1011
1012 /*
1013 * NEEDSWORK: handle conflicts from merges with
1014 * merge.renormalize set, too?
1015 */
1016 ll_merge(&result, path, &mmfile[0], NULL,
1017 &mmfile[1], "ours",
1018 &mmfile[2], "theirs",
1019 istate, NULL);
1020 for (i = 0; i < 3; i++)
1021 free(mmfile[i].ptr);
1022
1023 memset(&io, 0, sizeof(io));
1024 io.io.getline = rerere_mem_getline;
1025 if (output)
1026 io.io.output = fopen(output, "w");
1027 else
1028 io.io.output = NULL;
1029 strbuf_init(&io.input, 0);
1030 strbuf_attach(&io.input, result.ptr, result.size, result.size);
1031
1032 /*
1033 * Grab the conflict ID and optionally write the original
1034 * contents with conflict markers out.
1035 */
1036 has_conflicts = handle_path(hash, (struct rerere_io *)&io, marker_size);
1037 strbuf_release(&io.input);
1038 if (io.io.output)
1039 fclose(io.io.output);
1040 return has_conflicts;
1041 }
1042
1043 static int rerere_forget_one_path(struct index_state *istate,
1044 const char *path,
1045 struct string_list *rr)
1046 {
1047 const char *filename;
1048 struct rerere_id *id;
1049 unsigned char hash[GIT_MAX_RAWSZ];
1050 int ret;
1051 struct strbuf buf = STRBUF_INIT;
1052 struct string_list_item *item;
1053
1054 /*
1055 * Recreate the original conflict from the stages in the
1056 * index and compute the conflict ID
1057 */
1058 ret = handle_cache(istate, path, hash, NULL);
1059 if (ret < 1)
1060 return error(_("could not parse conflict hunks in '%s'"), path);
1061
1062 /* Nuke the recorded resolution for the conflict */
1063 id = new_rerere_id(hash);
1064
1065 for (id->variant = 0;
1066 id->variant < id->collection->status_nr;
1067 id->variant++) {
1068 mmfile_t cur = { NULL, 0 };
1069 mmbuffer_t result = {NULL, 0};
1070 int cleanly_resolved;
1071
1072 if (!has_rerere_resolution(id))
1073 continue;
1074
1075 handle_cache(istate, path, hash, rerere_path(&buf, id, "thisimage"));
1076 if (read_mmfile(&cur, rerere_path(&buf, id, "thisimage"))) {
1077 free(cur.ptr);
1078 error(_("failed to update conflicted state in '%s'"), path);
1079 goto fail_exit;
1080 }
1081 cleanly_resolved = !try_merge(istate, id, path, &cur, &result);
1082 free(result.ptr);
1083 free(cur.ptr);
1084 if (cleanly_resolved)
1085 break;
1086 }
1087
1088 if (id->collection->status_nr <= id->variant) {
1089 error(_("no remembered resolution for '%s'"), path);
1090 goto fail_exit;
1091 }
1092
1093 filename = rerere_path(&buf, id, "postimage");
1094 if (unlink(filename)) {
1095 if (errno == ENOENT)
1096 error(_("no remembered resolution for '%s'"), path);
1097 else
1098 error_errno(_("cannot unlink '%s'"), filename);
1099 goto fail_exit;
1100 }
1101
1102 /*
1103 * Update the preimage so that the user can resolve the
1104 * conflict in the working tree, run us again to record
1105 * the postimage.
1106 */
1107 handle_cache(istate, path, hash, rerere_path(&buf, id, "preimage"));
1108 fprintf_ln(stderr, _("Updated preimage for '%s'"), path);
1109
1110 /*
1111 * And remember that we can record resolution for this
1112 * conflict when the user is done.
1113 */
1114 item = string_list_insert(rr, path);
1115 free_rerere_id(item);
1116 item->util = id;
1117 fprintf(stderr, _("Forgot resolution for '%s'\n"), path);
1118 strbuf_release(&buf);
1119 return 0;
1120
1121 fail_exit:
1122 strbuf_release(&buf);
1123 free(id);
1124 return -1;
1125 }
1126
1127 int rerere_forget(struct repository *r, struct pathspec *pathspec)
1128 {
1129 int i, fd, ret;
1130 struct string_list conflict = STRING_LIST_INIT_DUP;
1131 struct string_list merge_rr = STRING_LIST_INIT_DUP;
1132
1133 if (repo_read_index(r) < 0)
1134 return error(_("index file corrupt"));
1135
1136 fd = setup_rerere(r, &merge_rr, RERERE_NOAUTOUPDATE);
1137 if (fd < 0)
1138 return 0;
1139
1140 /*
1141 * The paths may have been resolved (incorrectly);
1142 * recover the original conflicted state and then
1143 * find the conflicted paths.
1144 */
1145 unmerge_index(r->index, pathspec, 0);
1146 find_conflict(r, &conflict);
1147 for (i = 0; i < conflict.nr; i++) {
1148 struct string_list_item *it = &conflict.items[i];
1149 if (!match_pathspec(r->index, pathspec, it->string,
1150 strlen(it->string), 0, NULL, 0))
1151 continue;
1152 rerere_forget_one_path(r->index, it->string, &merge_rr);
1153 }
1154
1155 ret = write_rr(&merge_rr, fd);
1156
1157 string_list_clear(&conflict, 0);
1158 string_list_clear(&merge_rr, 1);
1159 return ret;
1160 }
1161
1162 /*
1163 * Garbage collection support
1164 */
1165
1166 static timestamp_t rerere_created_at(struct rerere_id *id)
1167 {
1168 struct strbuf buf = STRBUF_INIT;
1169 struct stat st;
1170 timestamp_t ret;
1171
1172 ret = stat(rerere_path(&buf, id, "preimage"), &st) ? (time_t) 0 : st.st_mtime;
1173
1174 strbuf_release(&buf);
1175 return ret;
1176 }
1177
1178 static timestamp_t rerere_last_used_at(struct rerere_id *id)
1179 {
1180 struct strbuf buf = STRBUF_INIT;
1181 struct stat st;
1182 timestamp_t ret;
1183
1184 ret = stat(rerere_path(&buf, id, "postimage"), &st) ? (time_t) 0 : st.st_mtime;
1185
1186 strbuf_release(&buf);
1187 return ret;
1188 }
1189
1190 /*
1191 * Remove the recorded resolution for a given conflict ID
1192 */
1193 static void unlink_rr_item(struct rerere_id *id)
1194 {
1195 struct strbuf buf = STRBUF_INIT;
1196 unlink_or_warn(rerere_path(&buf, id, "thisimage"));
1197 remove_variant(id);
1198 id->collection->status[id->variant] = 0;
1199 strbuf_release(&buf);
1200 }
1201
1202 static void prune_one(struct rerere_id *id,
1203 timestamp_t cutoff_resolve, timestamp_t cutoff_noresolve)
1204 {
1205 timestamp_t then;
1206 timestamp_t cutoff;
1207
1208 then = rerere_last_used_at(id);
1209 if (then)
1210 cutoff = cutoff_resolve;
1211 else {
1212 then = rerere_created_at(id);
1213 if (!then)
1214 return;
1215 cutoff = cutoff_noresolve;
1216 }
1217 if (then < cutoff)
1218 unlink_rr_item(id);
1219 }
1220
1221 /* Does the basename in "path" look plausibly like an rr-cache entry? */
1222 static int is_rr_cache_dirname(const char *path)
1223 {
1224 struct object_id oid;
1225 const char *end;
1226 return !parse_oid_hex(path, &oid, &end) && !*end;
1227 }
1228
1229 void rerere_gc(struct repository *r, struct string_list *rr)
1230 {
1231 struct string_list to_remove = STRING_LIST_INIT_DUP;
1232 DIR *dir;
1233 struct dirent *e;
1234 int i;
1235 timestamp_t now = time(NULL);
1236 timestamp_t cutoff_noresolve = now - 15 * 86400;
1237 timestamp_t cutoff_resolve = now - 60 * 86400;
1238 struct strbuf buf = STRBUF_INIT;
1239
1240 if (setup_rerere(r, rr, 0) < 0)
1241 return;
1242
1243 repo_config_get_expiry_in_days(the_repository, "gc.rerereresolved",
1244 &cutoff_resolve, now);
1245 repo_config_get_expiry_in_days(the_repository, "gc.rerereunresolved",
1246 &cutoff_noresolve, now);
1247 repo_config(the_repository, git_default_config, NULL);
1248 dir = opendir(repo_git_path_replace(the_repository, &buf, "rr-cache"));
1249 if (!dir)
1250 die_errno(_("unable to open rr-cache directory"));
1251 /* Collect stale conflict IDs ... */
1252 while ((e = readdir_skip_dot_and_dotdot(dir))) {
1253 struct rerere_dir *rr_dir;
1254 struct rerere_id id;
1255 int now_empty;
1256
1257 if (!is_rr_cache_dirname(e->d_name))
1258 continue; /* or should we remove e->d_name? */
1259
1260 rr_dir = find_rerere_dir(e->d_name);
1261
1262 now_empty = 1;
1263 for (id.variant = 0, id.collection = rr_dir;
1264 id.variant < id.collection->status_nr;
1265 id.variant++) {
1266 prune_one(&id, cutoff_resolve, cutoff_noresolve);
1267 if (id.collection->status[id.variant])
1268 now_empty = 0;
1269 }
1270 if (now_empty)
1271 string_list_append(&to_remove, e->d_name);
1272 }
1273 closedir(dir);
1274
1275 /* ... and then remove the empty directories */
1276 for (i = 0; i < to_remove.nr; i++)
1277 rmdir(repo_git_path_replace(the_repository, &buf,
1278 "rr-cache/%s", to_remove.items[i].string));
1279
1280 string_list_clear(&to_remove, 0);
1281 rollback_lock_file(&write_lock);
1282 strbuf_release(&buf);
1283 }
1284
1285 /*
1286 * During a conflict resolution, after "rerere" recorded the
1287 * preimages, abandon them if the user did not resolve them or
1288 * record their resolutions. And drop $GIT_DIR/MERGE_RR.
1289 *
1290 * NEEDSWORK: shouldn't we be calling this from "reset --hard"?
1291 */
1292 void rerere_clear(struct repository *r, struct string_list *merge_rr)
1293 {
1294 int i;
1295
1296 if (setup_rerere(r, merge_rr, 0) < 0)
1297 return;
1298
1299 for (i = 0; i < merge_rr->nr; i++) {
1300 struct rerere_id *id = merge_rr->items[i].util;
1301 struct strbuf buf = STRBUF_INIT;
1302
1303 if (!has_rerere_resolution(id)) {
1304 unlink_rr_item(id);
1305 rmdir(rerere_path(&buf, id, NULL));
1306 }
1307
1308 strbuf_release(&buf);
1309 }
1310 unlink_or_warn(git_path_merge_rr(r));
1311 rollback_lock_file(&write_lock);
1312 }