2 * Copyright (C) 2005 Junio C Hamano
11 static const char *diff_opts = "-pu";
12 static unsigned char null_sha1[20] = { 0, };
14 static int reverse_diff;
16 static const char *external_diff(void)
18 static const char *external_diff_cmd = NULL;
19 static int done_preparing = 0;
22 return external_diff_cmd;
25 * Default values above are meant to match the
26 * Linux kernel development style. Examples of
27 * alternative styles you can specify via environment
32 if (gitenv("GIT_EXTERNAL_DIFF"))
33 external_diff_cmd = gitenv("GIT_EXTERNAL_DIFF");
35 /* In case external diff fails... */
36 diff_opts = gitenv("GIT_DIFF_OPTS") ? : diff_opts;
39 return external_diff_cmd;
42 /* Help to copy the thing properly quoted for the shell safety.
43 * any single quote is replaced with '\'', and the caller is
44 * expected to enclose the result within a single quote pair.
47 * original sq_expand result
48 * name ==> name ==> 'name'
49 * a b ==> a b ==> 'a b'
50 * a'b ==> a'\''b ==> 'a'\''b'
52 static char *sq_expand(const char *src)
54 static char *buf = NULL;
59 /* count bytes needed to store the quoted string. */
60 for (cnt = 1, cp = src; *cp; cnt++, cp++)
66 while ((c = *src++)) {
70 bp = strcpy(bp, "'\\''");
78 static struct diff_tempfile {
79 const char *name; /* filename external diff should read from */
85 static void builtin_diff(const char *name_a,
87 struct diff_tempfile *temp,
90 int i, next_at, cmd_size;
91 const char *diff_cmd = "diff -L'%s%s' -L'%s%s'";
92 const char *diff_arg = "'%s' '%s'||:"; /* "||:" is to return 0 */
93 const char *input_name_sq[2];
96 const char *name_sq[2];
99 name_sq[0] = sq_expand(name_a);
100 name_sq[1] = sq_expand(name_b);
102 /* diff_cmd and diff_arg have 6 %s in total which makes
103 * the sum of these strings 12 bytes larger than required.
104 * we use 2 spaces around diff-opts, and we need to count
105 * terminating NUL, so we subtract 9 here.
107 cmd_size = (strlen(diff_cmd) + strlen(diff_opts) +
108 strlen(diff_arg) - 9);
109 for (i = 0; i < 2; i++) {
110 input_name_sq[i] = sq_expand(temp[i].name);
111 if (!strcmp(temp[i].name, "/dev/null")) {
112 path0[i] = "/dev/null";
115 path0[i] = i ? "b/" : "a/";
116 path1[i] = name_sq[i];
118 cmd_size += (strlen(path0[i]) + strlen(path1[i]) +
119 strlen(input_name_sq[i]));
122 cmd = xmalloc(cmd_size);
125 next_at += snprintf(cmd+next_at, cmd_size-next_at,
127 path0[0], path1[0], path0[1], path1[1]);
128 next_at += snprintf(cmd+next_at, cmd_size-next_at,
130 next_at += snprintf(cmd+next_at, cmd_size-next_at,
131 diff_arg, input_name_sq[0], input_name_sq[1]);
133 printf("diff --git a/%s b/%s\n", name_a, name_b);
135 printf("new file mode %s\n", temp[1].mode);
136 else if (!path1[1][0])
137 printf("deleted file mode %s\n", temp[0].mode);
139 if (strcmp(temp[0].mode, temp[1].mode)) {
140 printf("old mode %s\n", temp[0].mode);
141 printf("new mode %s\n", temp[1].mode);
143 if (xfrm_msg && xfrm_msg[0])
146 if (strncmp(temp[0].mode, temp[1].mode, 3))
147 /* we do not run diff between different kind
153 execlp("/bin/sh","sh", "-c", cmd, NULL);
156 struct diff_filespec *alloc_filespec(const char *path)
158 int namelen = strlen(path);
159 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
160 spec->path = (char *)(spec + 1);
161 strcpy(spec->path, path);
162 spec->should_free = spec->should_munmap = 0;
163 spec->xfrm_flags = 0;
167 memset(spec->sha1, 0, 20);
171 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
175 spec->mode = DIFF_FILE_CANON_MODE(mode);
176 memcpy(spec->sha1, sha1, 20);
177 spec->sha1_valid = !!memcmp(sha1, null_sha1, 20);
182 * Given a name and sha1 pair, if the dircache tells us the file in
183 * the work tree has that object contents, return true, so that
184 * prepare_temp_file() does not have to inflate and extract.
186 static int work_tree_matches(const char *name, const unsigned char *sha1)
188 struct cache_entry *ce;
192 /* We do not read the cache ourselves here, because the
193 * benchmark with my previous version that always reads cache
194 * shows that it makes things worse for diff-tree comparing
195 * two linux-2.6 kernel trees in an already checked out work
196 * tree. This is because most diff-tree comparisons deal with
197 * only a small number of files, while reading the cache is
198 * expensive for a large project, and its cost outweighs the
199 * savings we get by not inflating the object to a temporary
200 * file. Practically, this code only helps when we are used
201 * by diff-cache --cached, which does read the cache before
208 pos = cache_name_pos(name, len);
211 ce = active_cache[pos];
212 if ((lstat(name, &st) < 0) ||
213 !S_ISREG(st.st_mode) || /* careful! */
214 ce_match_stat(ce, &st) ||
215 memcmp(sha1, ce->sha1, 20))
217 /* we return 1 only when we can stat, it is a regular file,
218 * stat information matches, and sha1 recorded in the cache
219 * matches. I.e. we know the file in the work tree really is
220 * the same as the <name, sha1> pair.
226 * While doing rename detection and pickaxe operation, we may need to
227 * grab the data for the blob (or file) for our own in-core comparison.
228 * diff_filespec has data and size fields for this purpose.
230 int diff_populate_filespec(struct diff_filespec *s)
233 if (!DIFF_FILE_VALID(s))
234 die("internal error: asking to populate invalid file.");
235 if (S_ISDIR(s->mode))
240 if (!s->sha1_valid ||
241 work_tree_matches(s->path, s->sha1)) {
244 if (lstat(s->path, &st) < 0) {
245 if (errno == ENOENT) {
254 s->size = st.st_size;
257 if (S_ISLNK(st.st_mode)) {
259 s->data = xmalloc(s->size);
261 ret = readlink(s->path, s->data, s->size);
268 fd = open(s->path, O_RDONLY);
271 s->data = mmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
272 s->should_munmap = 1;
277 s->data = read_sha1_file(s->sha1, type, &s->size);
283 void diff_free_filespec_data(struct diff_filespec *s)
287 else if (s->should_munmap)
288 munmap(s->data, s->size);
289 s->should_free = s->should_munmap = 0;
293 static void prep_temp_blob(struct diff_tempfile *temp,
301 strcpy(temp->tmp_path, ".diff_XXXXXX");
302 fd = mkstemp(temp->tmp_path);
304 die("unable to create temp-file");
305 if (write(fd, blob, size) != size)
306 die("unable to write temp-file");
308 temp->name = temp->tmp_path;
309 strcpy(temp->hex, sha1_to_hex(sha1));
311 sprintf(temp->mode, "%06o", mode);
314 static void prepare_temp_file(const char *name,
315 struct diff_tempfile *temp,
316 struct diff_filespec *one)
318 if (!DIFF_FILE_VALID(one)) {
320 /* A '-' entry produces this for file-2, and
321 * a '+' entry produces this for file-1.
323 temp->name = "/dev/null";
324 strcpy(temp->hex, ".");
325 strcpy(temp->mode, ".");
329 if (!one->sha1_valid ||
330 work_tree_matches(name, one->sha1)) {
332 if (lstat(name, &st) < 0) {
334 goto not_a_valid_file;
335 die("stat(%s): %s", name, strerror(errno));
337 if (S_ISLNK(st.st_mode)) {
339 char *buf, buf_[1024];
340 buf = ((sizeof(buf_) < st.st_size) ?
341 xmalloc(st.st_size) : buf_);
342 ret = readlink(name, buf, st.st_size);
344 die("readlink(%s)", name);
345 prep_temp_blob(temp, buf, st.st_size,
347 one->sha1 : null_sha1),
349 one->mode : S_IFLNK));
352 /* we can borrow from the file in the work tree */
354 if (!one->sha1_valid)
355 strcpy(temp->hex, sha1_to_hex(null_sha1));
357 strcpy(temp->hex, sha1_to_hex(one->sha1));
358 sprintf(temp->mode, "%06o",
359 S_IFREG |ce_permissions(st.st_mode));
364 if (diff_populate_filespec(one))
365 die("cannot read data blob for %s", one->path);
366 prep_temp_blob(temp, one->data, one->size,
367 one->sha1, one->mode);
371 static void remove_tempfile(void)
375 for (i = 0; i < 2; i++)
376 if (diff_temp[i].name == diff_temp[i].tmp_path) {
377 unlink(diff_temp[i].name);
378 diff_temp[i].name = NULL;
382 static void remove_tempfile_on_signal(int signo)
387 /* An external diff command takes:
389 * diff-cmd name infile1 infile1-sha1 infile1-mode \
390 * infile2 infile2-sha1 infile2-mode [ rename-to ]
393 static void run_external_diff(const char *pgm,
396 struct diff_filespec *one,
397 struct diff_filespec *two,
398 const char *xfrm_msg)
400 struct diff_tempfile *temp = diff_temp;
403 static int atexit_asked = 0;
406 prepare_temp_file(name, &temp[0], one);
407 prepare_temp_file(other ? : name, &temp[1], two);
408 if (! atexit_asked &&
409 (temp[0].name == temp[0].tmp_path ||
410 temp[1].name == temp[1].tmp_path)) {
412 atexit(remove_tempfile);
414 signal(SIGINT, remove_tempfile_on_signal);
420 die("unable to fork");
424 const char *exec_arg[10];
425 const char **arg = &exec_arg[0];
428 *arg++ = temp[0].name;
429 *arg++ = temp[0].hex;
430 *arg++ = temp[0].mode;
431 *arg++ = temp[1].name;
432 *arg++ = temp[1].hex;
433 *arg++ = temp[1].mode;
439 execvp(pgm, (char *const*) exec_arg);
442 execlp(pgm, pgm, name, NULL);
445 * otherwise we use the built-in one.
448 builtin_diff(name, other ? : name, temp, xfrm_msg);
450 printf("* Unmerged path %s\n", name);
453 if (waitpid(pid, &status, 0) < 0 ||
454 !WIFEXITED(status) || WEXITSTATUS(status)) {
455 /* Earlier we did not check the exit status because
456 * diff exits non-zero if files are different, and
457 * we are not interested in knowing that. It was a
458 * mistake which made it harder to quit a diff-*
459 * session that uses the git-apply-patch-script as
460 * the GIT_EXTERNAL_DIFF. A custom GIT_EXTERNAL_DIFF
461 * should also exit non-zero only when it wants to
462 * abort the entire diff-* session.
465 fprintf(stderr, "external diff died, stopping at %s.\n", name);
471 static void run_diff(const char *name,
473 struct diff_filespec *one,
474 struct diff_filespec *two,
475 const char *xfrm_msg)
477 const char *pgm = external_diff();
479 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
480 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
481 /* a filepair that changes between file and symlink
482 * needs to be split into deletion and creation.
484 struct diff_filespec *null = alloc_filespec(two->path);
485 run_external_diff(NULL, name, other, one, null, xfrm_msg);
487 null = alloc_filespec(one->path);
488 run_external_diff(NULL, name, other, null, two, xfrm_msg);
492 run_external_diff(pgm, name, other, one, two, xfrm_msg);
495 void diff_setup(int flags)
497 if (flags & DIFF_SETUP_REVERSE)
501 struct diff_queue_struct diff_queued_diff;
503 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
505 if (queue->alloc <= queue->nr) {
506 queue->alloc = alloc_nr(queue->alloc);
507 queue->queue = xrealloc(queue->queue,
508 sizeof(dp) * queue->alloc);
510 queue->queue[queue->nr++] = dp;
513 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
514 struct diff_filespec *one,
515 struct diff_filespec *two)
517 struct diff_filepair *dp = xmalloc(sizeof(*dp));
521 dp->source_stays = 0;
526 void diff_free_filepair(struct diff_filepair *p)
528 diff_free_filespec_data(p->one);
529 diff_free_filespec_data(p->two);
533 static void diff_flush_raw(struct diff_filepair *p,
534 int line_termination,
535 int inter_name_termination)
540 if (line_termination) {
541 const char *err = "path %s cannot be expressed without -z";
542 if (strchr(p->one->path, line_termination) ||
543 strchr(p->one->path, inter_name_termination))
544 die(err, p->one->path);
545 if (strchr(p->two->path, line_termination) ||
546 strchr(p->two->path, inter_name_termination))
547 die(err, p->two->path);
553 sprintf(status, "%c%03d", p->status,
554 (int)(0.5 + p->score * 100.0/MAX_SCORE));
558 status[0] = p->status;
562 printf(":%06o %06o %s ",
563 p->one->mode, p->two->mode, sha1_to_hex(p->one->sha1));
565 sha1_to_hex(p->two->sha1),
567 inter_name_termination,
570 printf("%c%s", inter_name_termination, p->two->path);
571 putchar(line_termination);
574 int diff_unmodified_pair(struct diff_filepair *p)
576 /* This function is written stricter than necessary to support
577 * the currently implemented transformers, but the idea is to
578 * let transformers to produce diff_filepairs any way they want,
579 * and filter and clean them up here before producing the output.
581 struct diff_filespec *one, *two;
583 if (DIFF_PAIR_UNMERGED(p))
584 return 0; /* unmerged is interesting */
589 /* deletion, addition, mode or type change
590 * and rename are all interesting.
592 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
593 DIFF_PAIR_MODE_CHANGED(p) ||
594 strcmp(one->path, two->path))
597 /* both are valid and point at the same path. that is, we are
598 * dealing with a change.
600 if (one->sha1_valid && two->sha1_valid &&
601 !memcmp(one->sha1, two->sha1, sizeof(one->sha1)))
602 return 1; /* no change */
603 if (!one->sha1_valid && !two->sha1_valid)
604 return 1; /* both look at the same file on the filesystem. */
608 static void diff_flush_patch(struct diff_filepair *p)
610 const char *name, *other;
611 char msg_[PATH_MAX*2+200], *msg;
613 if (diff_unmodified_pair(p))
617 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
618 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
619 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
620 return; /* no tree diffs in patch format */
625 "similarity index %d%%\n"
628 (int)(0.5 + p->score * 100.0/MAX_SCORE),
629 p->one->path, p->two->path);
634 "similarity index %d%%\n"
637 (int)(0.5 + p->score * 100.0/MAX_SCORE),
638 p->one->path, p->two->path);
645 if (DIFF_PAIR_UNMERGED(p))
646 run_diff(name, NULL, NULL, NULL, NULL);
648 run_diff(name, other, p->one, p->two, msg);
651 int diff_queue_is_empty(void)
653 struct diff_queue_struct *q = &diff_queued_diff;
655 for (i = 0; i < q->nr; i++)
656 if (!diff_unmodified_pair(q->queue[i]))
662 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
664 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
667 DIFF_FILE_VALID(s) ? "valid" : "invalid",
669 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
670 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
672 s->size, s->xfrm_flags);
675 void diff_debug_filepair(const struct diff_filepair *p, int i)
677 diff_debug_filespec(p->one, i, "one");
678 diff_debug_filespec(p->two, i, "two");
679 fprintf(stderr, "score %d, status %c source_stays %d\n",
680 p->score, p->status ? : '?', p->source_stays);
683 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
687 fprintf(stderr, "%s\n", msg);
688 fprintf(stderr, "q->nr = %d\n", q->nr);
689 for (i = 0; i < q->nr; i++) {
690 struct diff_filepair *p = q->queue[i];
691 diff_debug_filepair(p, i);
696 static void diff_resolve_rename_copy(void)
699 struct diff_filepair *p, *pp;
700 struct diff_queue_struct *q = &diff_queued_diff;
702 diff_debug_queue("resolve-rename-copy", q);
704 for (i = 0; i < q->nr; i++) {
706 p->status = 0; /* undecided */
707 if (DIFF_PAIR_UNMERGED(p))
709 else if (!DIFF_FILE_VALID(p->one))
711 else if (!DIFF_FILE_VALID(p->two)) {
712 /* Deleted entry may have been picked up by
713 * another rename-copy entry. So we scan the
714 * queue and if we find one that uses us as the
715 * source we do not say delete for this entry.
717 for (j = 0; j < q->nr; j++) {
719 if (!strcmp(p->one->path, pp->one->path) &&
721 /* rename/copy are always valid
722 * so we do not say DIFF_FILE_VALID()
723 * on pp->one and pp->two.
732 else if (DIFF_PAIR_TYPE_CHANGED(p))
735 /* from this point on, we are dealing with a pair
736 * whose both sides are valid and of the same type, i.e.
737 * either in-place edit or rename/copy edit.
740 if (p->source_stays) {
744 /* See if there is some other filepair that
745 * copies from the same source as us. If so
746 * we are a copy. Otherwise we are a rename.
748 for (j = i + 1; j < q->nr; j++) {
750 if (strcmp(pp->one->path, p->one->path))
751 continue; /* not us */
753 continue; /* not a rename/copy */
754 /* pp is a rename/copy from the same source */
761 else if (memcmp(p->one->sha1, p->two->sha1, 20) ||
762 p->one->mode != p->two->mode)
765 /* this is a "no-change" entry.
766 * should not happen anymore.
769 die("internal error in diffcore: unmodified entry remains");
771 diff_debug_queue("resolve-rename-copy done", q);
774 void diff_flush(int diff_output_style, int resolve_rename_copy)
776 struct diff_queue_struct *q = &diff_queued_diff;
778 int line_termination = '\n';
779 int inter_name_termination = '\t';
781 if (diff_output_style == DIFF_FORMAT_MACHINE)
782 line_termination = inter_name_termination = 0;
783 if (resolve_rename_copy)
784 diff_resolve_rename_copy();
786 for (i = 0; i < q->nr; i++) {
787 struct diff_filepair *p = q->queue[i];
788 if ((diff_output_style == DIFF_FORMAT_NO_OUTPUT) ||
792 die("internal error in diff-resolve-rename-copy");
793 switch (diff_output_style) {
794 case DIFF_FORMAT_PATCH:
797 case DIFF_FORMAT_HUMAN:
798 case DIFF_FORMAT_MACHINE:
799 diff_flush_raw(p, line_termination,
800 inter_name_termination);
804 for (i = 0; i < q->nr; i++)
805 diff_free_filepair(q->queue[i]);
808 q->nr = q->alloc = 0;
811 void diff_addremove(int addremove, unsigned mode,
812 const unsigned char *sha1,
813 const char *base, const char *path)
815 char concatpath[PATH_MAX];
816 struct diff_filespec *one, *two;
818 /* This may look odd, but it is a preparation for
819 * feeding "there are unchanged files which should
820 * not produce diffs, but when you are doing copy
821 * detection you would need them, so here they are"
822 * entries to the diff-core. They will be prefixed
823 * with something like '=' or '*' (I haven't decided
824 * which but should not make any difference).
825 * Feeding the same new and old to diff_change()
826 * also has the same effect.
827 * Before the final output happens, they are pruned after
828 * merged into rename/copy pairs as appropriate.
831 addremove = (addremove == '+' ? '-' :
832 addremove == '-' ? '+' : addremove);
834 if (!path) path = "";
835 sprintf(concatpath, "%s%s", base, path);
836 one = alloc_filespec(concatpath);
837 two = alloc_filespec(concatpath);
839 if (addremove != '+')
840 fill_filespec(one, sha1, mode);
841 if (addremove != '-')
842 fill_filespec(two, sha1, mode);
844 diff_queue(&diff_queued_diff, one, two);
847 void diff_helper_input(unsigned old_mode,
849 const unsigned char *old_sha1,
850 const unsigned char *new_sha1,
851 const char *old_path,
854 const char *new_path)
856 struct diff_filespec *one, *two;
857 struct diff_filepair *dp;
859 one = alloc_filespec(old_path);
860 two = alloc_filespec(new_path);
862 fill_filespec(one, old_sha1, old_mode);
864 fill_filespec(two, new_sha1, new_mode);
865 dp = diff_queue(&diff_queued_diff, one, two);
866 dp->score = score * MAX_SCORE / 100;
870 void diff_change(unsigned old_mode, unsigned new_mode,
871 const unsigned char *old_sha1,
872 const unsigned char *new_sha1,
873 const char *base, const char *path)
875 char concatpath[PATH_MAX];
876 struct diff_filespec *one, *two;
880 const unsigned char *tmp_c;
881 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
882 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
884 if (!path) path = "";
885 sprintf(concatpath, "%s%s", base, path);
886 one = alloc_filespec(concatpath);
887 two = alloc_filespec(concatpath);
888 fill_filespec(one, old_sha1, old_mode);
889 fill_filespec(two, new_sha1, new_mode);
891 diff_queue(&diff_queued_diff, one, two);
894 void diff_unmerge(const char *path)
896 struct diff_filespec *one, *two;
897 one = alloc_filespec(path);
898 two = alloc_filespec(path);
899 diff_queue(&diff_queued_diff, one, two);