2 * Copyright (C) 2005 Junio C Hamano
12 static const char *diff_opts = "-pu";
13 static unsigned char null_sha1[20] = { 0, };
15 static const char *external_diff(void)
17 static const char *external_diff_cmd = NULL;
18 static int done_preparing = 0;
21 return external_diff_cmd;
24 * Default values above are meant to match the
25 * Linux kernel development style. Examples of
26 * alternative styles you can specify via environment
31 if (gitenv("GIT_EXTERNAL_DIFF"))
32 external_diff_cmd = gitenv("GIT_EXTERNAL_DIFF");
34 /* In case external diff fails... */
35 diff_opts = gitenv("GIT_DIFF_OPTS") ? : diff_opts;
38 return external_diff_cmd;
41 /* Help to copy the thing properly quoted for the shell safety.
42 * any single quote is replaced with '\'', and the caller is
43 * expected to enclose the result within a single quote pair.
46 * original sq_expand result
47 * name ==> name ==> 'name'
48 * a b ==> a b ==> 'a b'
49 * a'b ==> a'\''b ==> 'a'\''b'
51 static char *sq_expand(const char *src)
53 static char *buf = NULL;
58 /* count bytes needed to store the quoted string. */
59 for (cnt = 1, cp = src; *cp; cnt++, cp++)
65 while ((c = *src++)) {
69 bp = strcpy(bp, "'\\''");
77 static struct diff_tempfile {
85 unsigned char blob_sha1[20];
86 unsigned short mode; /* file mode */
87 unsigned sha1_valid : 1; /* if true, use blob_sha1 and trust mode;
88 * however with a NULL SHA1, read them
89 * from the file system.
90 * if false, use the name and read mode from
93 unsigned file_valid : 1; /* if false the file does not even exist */
96 static void builtin_diff(const char *name_a,
98 struct diff_tempfile *temp)
101 const char *diff_cmd = "diff -L'%s%s' -L'%s%s'";
102 const char *diff_arg = "'%s' '%s'||:"; /* "||:" is to return 0 */
103 const char *input_name_sq[2];
104 const char *path0[2];
105 const char *path1[2];
106 const char *name_sq[2];
109 name_sq[0] = sq_expand(name_a);
110 name_sq[1] = sq_expand(name_b);
112 /* diff_cmd and diff_arg have 6 %s in total which makes
113 * the sum of these strings 12 bytes larger than required.
114 * we use 2 spaces around diff-opts, and we need to count
115 * terminating NUL, so we subtract 9 here.
117 int cmd_size = (strlen(diff_cmd) + strlen(diff_opts) +
118 strlen(diff_arg) - 9);
119 for (i = 0; i < 2; i++) {
120 input_name_sq[i] = sq_expand(temp[i].name);
121 if (!strcmp(temp[i].name, "/dev/null")) {
122 path0[i] = "/dev/null";
125 path0[i] = i ? "b/" : "a/";
126 path1[i] = name_sq[i];
128 cmd_size += (strlen(path0[i]) + strlen(path1[i]) +
129 strlen(input_name_sq[i]));
132 cmd = xmalloc(cmd_size);
135 next_at += snprintf(cmd+next_at, cmd_size-next_at,
137 path0[0], path1[0], path0[1], path1[1]);
138 next_at += snprintf(cmd+next_at, cmd_size-next_at,
140 next_at += snprintf(cmd+next_at, cmd_size-next_at,
141 diff_arg, input_name_sq[0], input_name_sq[1]);
143 printf("diff --git a/%s b/%s\n", name_a, name_b);
145 printf("new file mode %s\n", temp[1].mode);
146 else if (!path1[1][0])
147 printf("deleted file mode %s\n", temp[0].mode);
149 if (strcmp(temp[0].mode, temp[1].mode)) {
150 printf("old mode %s\n", temp[0].mode);
151 printf("new mode %s\n", temp[1].mode);
153 if (strcmp(name_a, name_b)) {
154 printf("rename old %s\n", name_a);
155 printf("rename new %s\n", name_b);
157 if (strncmp(temp[0].mode, temp[1].mode, 3))
158 /* we do not run diff between different kind
164 execlp("/bin/sh","sh", "-c", cmd, NULL);
168 * Given a name and sha1 pair, if the dircache tells us the file in
169 * the work tree has that object contents, return true, so that
170 * prepare_temp_file() does not have to inflate and extract.
172 static int work_tree_matches(const char *name, const unsigned char *sha1)
174 struct cache_entry *ce;
178 /* We do not read the cache ourselves here, because the
179 * benchmark with my previous version that always reads cache
180 * shows that it makes things worse for diff-tree comparing
181 * two linux-2.6 kernel trees in an already checked out work
182 * tree. This is because most diff-tree comparisons deal with
183 * only a small number of files, while reading the cache is
184 * expensive for a large project, and its cost outweighs the
185 * savings we get by not inflating the object to a temporary
186 * file. Practically, this code only helps when we are used
187 * by diff-cache --cached, which does read the cache before
194 pos = cache_name_pos(name, len);
197 ce = active_cache[pos];
198 if ((lstat(name, &st) < 0) ||
199 !S_ISREG(st.st_mode) ||
200 ce_match_stat(ce, &st) ||
201 memcmp(sha1, ce->sha1, 20))
206 static void prep_temp_blob(struct diff_tempfile *temp,
214 strcpy(temp->tmp_path, ".diff_XXXXXX");
215 fd = mkstemp(temp->tmp_path);
217 die("unable to create temp-file");
218 if (write(fd, blob, size) != size)
219 die("unable to write temp-file");
221 temp->name = temp->tmp_path;
222 strcpy(temp->hex, sha1_to_hex(sha1));
224 sprintf(temp->mode, "%06o", mode);
227 static void prepare_temp_file(const char *name,
228 struct diff_tempfile *temp,
229 struct diff_spec *one)
231 if (!one->file_valid) {
233 /* A '-' entry produces this for file-2, and
234 * a '+' entry produces this for file-1.
236 temp->name = "/dev/null";
237 strcpy(temp->hex, ".");
238 strcpy(temp->mode, ".");
242 if (!one->sha1_valid ||
243 work_tree_matches(name, one->blob_sha1)) {
246 if (lstat(temp->name, &st) < 0) {
248 goto not_a_valid_file;
249 die("stat(%s): %s", temp->name, strerror(errno));
251 if (S_ISLNK(st.st_mode)) {
253 char *buf, buf_[1024];
254 buf = ((sizeof(buf_) < st.st_size) ?
255 xmalloc(st.st_size) : buf_);
256 ret = readlink(name, buf, st.st_size);
258 die("readlink(%s)", name);
259 prep_temp_blob(temp, buf, st.st_size,
261 one->blob_sha1 : null_sha1),
263 one->mode : S_IFLNK));
266 if (!one->sha1_valid)
267 strcpy(temp->hex, sha1_to_hex(null_sha1));
269 strcpy(temp->hex, sha1_to_hex(one->blob_sha1));
270 sprintf(temp->mode, "%06o",
271 S_IFREG |ce_permissions(st.st_mode));
280 blob = read_sha1_file(one->blob_sha1, type, &size);
281 if (!blob || strcmp(type, "blob"))
282 die("unable to read blob object for %s (%s)",
283 name, sha1_to_hex(one->blob_sha1));
284 prep_temp_blob(temp, blob, size, one->blob_sha1, one->mode);
289 static void remove_tempfile(void)
293 for (i = 0; i < 2; i++)
294 if (diff_temp[i].name == diff_temp[i].tmp_path) {
295 unlink(diff_temp[i].name);
296 diff_temp[i].name = NULL;
300 static void remove_tempfile_on_signal(int signo)
305 static int detect_rename;
306 static int reverse_diff;
307 static const char **pathspec;
309 static int diff_rename_minimum_score;
311 static int matches_pathspec(const char *name)
319 namelen = strlen(name);
320 for (i = 0; i < speccnt; i++) {
321 int speclen = strlen(pathspec[i]);
322 if (! strncmp(pathspec[i], name, speclen) &&
323 speclen <= namelen &&
324 (name[speclen] == 0 || name[speclen] == '/'))
330 /* An external diff command takes:
332 * diff-cmd name infile1 infile1-sha1 infile1-mode \
333 * infile2 infile2-sha1 infile2-mode [ rename-to ]
336 static void run_external_diff(const char *name,
338 struct diff_spec *one,
339 struct diff_spec *two)
341 struct diff_tempfile *temp = diff_temp;
344 static int atexit_asked = 0;
347 struct diff_spec *tmp_spec;
348 tmp_spec = one; one = two; two = tmp_spec;
351 tmp = name; name = other; other = tmp;
355 if (!matches_pathspec(name) && (!other || !matches_pathspec(other)))
359 prepare_temp_file(name, &temp[0], one);
360 prepare_temp_file(other ? : name, &temp[1], two);
361 if (! atexit_asked &&
362 (temp[0].name == temp[0].tmp_path ||
363 temp[1].name == temp[1].tmp_path)) {
365 atexit(remove_tempfile);
367 signal(SIGINT, remove_tempfile_on_signal);
373 die("unable to fork");
375 const char *pgm = external_diff();
378 const char *exec_arg[9];
379 const char **arg = &exec_arg[0];
382 *arg++ = temp[0].name;
383 *arg++ = temp[0].hex;
384 *arg++ = temp[0].mode;
385 *arg++ = temp[1].name;
386 *arg++ = temp[1].hex;
387 *arg++ = temp[1].mode;
391 execvp(pgm, (char *const*) exec_arg);
394 execlp(pgm, pgm, name, NULL);
397 * otherwise we use the built-in one.
400 builtin_diff(name, other ? : name, temp);
402 printf("* Unmerged path %s\n", name);
405 if (waitpid(pid, &status, 0) < 0 ||
406 !WIFEXITED(status) || WEXITSTATUS(status)) {
407 /* Earlier we did not check the exit status because
408 * diff exits non-zero if files are different, and
409 * we are not interested in knowing that. It was a
410 * mistake which made it harder to quit a diff-*
411 * session that uses the git-apply-patch-script as
412 * the GIT_EXTERNAL_DIFF. A custom GIT_EXTERNAL_DIFF
413 * should also exit non-zero only when it wants to
414 * abort the entire diff-* session.
417 fprintf(stderr, "external diff died, stopping at %s.\n", name);
424 * We do not detect circular renames. Just hold created and deleted
425 * entries and later attempt to match them up. If they do not match,
426 * then spit them out as deletes or creates as original.
429 static struct diff_spec_hold {
430 struct diff_spec_hold *next;
435 #define SHOULD_FREE 2
436 #define SHOULD_MUNMAP 4
439 } *createdfile, *deletedfile;
441 static void hold_diff(const char *name,
442 struct diff_spec *one,
443 struct diff_spec *two)
445 struct diff_spec_hold **list, *elem;
447 if (one->file_valid && two->file_valid)
448 die("internal error");
450 if (!detect_rename) {
451 run_external_diff(name, NULL, one, two);
454 elem = xmalloc(sizeof(*elem) + strlen(name));
455 strcpy(elem->path, name);
459 if (one->file_valid) {
471 static int populate_data(struct diff_spec_hold *s)
477 if (s->it.sha1_valid) {
478 s->data = read_sha1_file(s->it.blob_sha1, type, &s->size);
479 s->flags |= SHOULD_FREE;
484 fd = open(s->path, O_RDONLY);
487 if (fstat(fd, &st)) {
491 s->size = st.st_size;
492 s->data = mmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
497 s->flags |= SHOULD_MUNMAP;
502 static void free_data(struct diff_spec_hold *s)
504 if (s->flags & SHOULD_FREE)
506 else if (s->flags & SHOULD_MUNMAP)
507 munmap(s->data, s->size);
508 s->flags &= ~(SHOULD_FREE|SHOULD_MUNMAP);
511 static void flush_remaining_diff(struct diff_spec_hold *elem,
514 static struct diff_spec null_file_spec;
516 null_file_spec.file_valid = 0;
517 for ( ; elem ; elem = elem->next) {
519 if (elem->flags & MATCHED)
522 run_external_diff(elem->path, NULL,
523 &null_file_spec, &elem->it);
525 run_external_diff(elem->path, NULL,
526 &elem->it, &null_file_spec);
530 static int is_exact_match(struct diff_spec_hold *src,
531 struct diff_spec_hold *dst)
533 if (src->it.sha1_valid && dst->it.sha1_valid &&
534 !memcmp(src->it.blob_sha1, dst->it.blob_sha1, 20))
536 if (populate_data(src) || populate_data(dst))
537 /* this is an error but will be caught downstream */
539 if (src->size == dst->size &&
540 !memcmp(src->data, dst->data, src->size))
545 #define MINIMUM_SCORE 5000
546 int estimate_similarity(struct diff_spec_hold *src, struct diff_spec_hold *dst)
548 /* src points at a deleted file and dst points at a created
549 * file. They may be quite similar, in which case we want to
550 * say src is renamed to dst.
552 * Compare them and return how similar they are, representing
553 * the score as an integer between 0 and 10000. 10000 is
554 * reserved for the case where they match exactly.
557 unsigned long delta_size;
559 delta_size = ((src->size < dst->size) ?
560 (dst->size - src->size) : (src->size - dst->size));
562 /* We would not consider rename followed by more than
563 * 20% edits; that is, delta_size must be smaller than
564 * (src->size + dst->size)/2 * 0.2, which means...
566 if ((src->size + dst->size) < delta_size * 10)
569 delta = diff_delta(src->data, src->size,
570 dst->data, dst->size,
574 /* This "delta" is really xdiff with adler32 and all the
575 * overheads but it is a quick and dirty approximation.
577 * Now we will give some score to it. Let's say 20% edit gets
578 * 5000 points and 0% edit gets 9000 points. That is, every
579 * 1/20000 edit gets 1 point penalty. The amount of penalty is:
581 * (delta_size * 2 / (src->size + dst->size)) * 20000
584 return 9000 - (40000 * delta_size / (src->size+dst->size));
588 struct diff_spec_hold *src;
589 struct diff_spec_hold *dst;
593 static int score_compare(const void *a_, const void *b_)
595 const struct diff_score *a = a_, *b = b_;
596 return b->score - a->score;
599 static void flush_rename_pair(struct diff_spec_hold *src,
600 struct diff_spec_hold *dst)
602 src->flags |= MATCHED;
603 dst->flags |= MATCHED;
606 run_external_diff(src->path, dst->path,
610 static void free_held_diff(struct diff_spec_hold *list)
612 struct diff_spec_hold *h;
613 for (h = list; list; list = h) {
620 void diff_flush(void)
622 int num_create, num_delete, c, d;
623 struct diff_spec_hold *elem, *src, *dst;
624 struct diff_score *mx;
626 /* We really want to cull the candidates list early
627 * with cheap tests in order to avoid doing deltas.
629 for (dst = createdfile; dst; dst = dst->next) {
630 for (src = deletedfile; src; src = src->next) {
631 if (! is_exact_match(src, dst))
633 flush_rename_pair(src, dst);
638 /* Count surviving candidates */
639 for (num_create = 0, elem = createdfile; elem; elem = elem->next)
640 if (!(elem->flags & MATCHED))
643 for (num_delete = 0, elem = deletedfile; elem; elem = elem->next)
644 if (!(elem->flags & MATCHED))
647 if (num_create == 0 || num_delete == 0)
650 mx = xmalloc(sizeof(*mx) * num_create * num_delete);
651 for (c = 0, dst = createdfile; dst; dst = dst->next) {
652 int base = c * num_delete;
653 if (dst->flags & MATCHED)
655 for (d = 0, src = deletedfile; src; src = src->next) {
656 struct diff_score *m = &mx[base+d];
657 if (src->flags & MATCHED)
661 m->score = estimate_similarity(src, dst);
666 qsort(mx, num_create*num_delete, sizeof(*mx), score_compare);
668 for (c = 0; c < num_create * num_delete; c++) {
671 if ((src->flags & MATCHED) || (dst->flags & MATCHED))
675 for (c = 0; c < num_create * num_delete; c++) {
678 if ((src->flags & MATCHED) || (dst->flags & MATCHED))
680 if (mx[c].score < diff_rename_minimum_score)
682 flush_rename_pair(src, dst);
686 flush_remaining_diff(createdfile, 1);
687 flush_remaining_diff(deletedfile, 0);
688 free_held_diff(createdfile);
689 free_held_diff(deletedfile);
690 createdfile = deletedfile = NULL;
693 void diff_setup(int detect_rename_, int minimum_score_, int reverse_diff_,
694 const char **pathspec_, int speccnt_)
696 free_held_diff(createdfile);
697 free_held_diff(deletedfile);
698 createdfile = deletedfile = NULL;
700 detect_rename = detect_rename_;
701 reverse_diff = reverse_diff_;
702 pathspec = pathspec_;
704 diff_rename_minimum_score = minimum_score_ ? : MINIMUM_SCORE;
707 void diff_addremove(int addremove, unsigned mode,
708 const unsigned char *sha1,
709 const char *base, const char *path)
711 char concatpath[PATH_MAX];
712 struct diff_spec spec[2], *one, *two;
714 memcpy(spec[0].blob_sha1, sha1, 20);
716 spec[0].sha1_valid = !!memcmp(sha1, null_sha1, 20);
717 spec[0].file_valid = 1;
718 spec[1].file_valid = 0;
720 if (addremove == '+') {
721 one = spec + 1; two = spec;
723 one = spec; two = one + 1;
727 strcpy(concatpath, base);
728 strcat(concatpath, path);
730 hold_diff(path ? concatpath : base, one, two);
733 void diff_change(unsigned old_mode, unsigned new_mode,
734 const unsigned char *old_sha1,
735 const unsigned char *new_sha1,
736 const char *base, const char *path) {
737 char concatpath[PATH_MAX];
738 struct diff_spec spec[2];
741 strcpy(concatpath, base);
742 strcat(concatpath, path);
745 memcpy(spec[0].blob_sha1, old_sha1, 20);
746 spec[0].mode = old_mode;
747 memcpy(spec[1].blob_sha1, new_sha1, 20);
748 spec[1].mode = new_mode;
749 spec[0].sha1_valid = !!memcmp(old_sha1, null_sha1, 20);
750 spec[1].sha1_valid = !!memcmp(new_sha1, null_sha1, 20);
751 spec[1].file_valid = spec[0].file_valid = 1;
753 /* We do not look at changed files as candidate for
754 * rename detection ever.
756 run_external_diff(path ? concatpath : base, NULL, &spec[0], &spec[1]);
759 void diff_unmerge(const char *path)
761 run_external_diff(path, NULL, NULL, NULL);