2 * Copyright (C) 2005 Junio C Hamano
12 static const char *diff_opts = "-pu";
13 static unsigned char null_sha1[20] = { 0, };
15 static int reverse_diff;
16 static int generate_patch;
17 static int line_termination = '\n';
18 static int inter_name_termination = '\t';
19 static const char **pathspec;
22 static const char *external_diff(void)
24 static const char *external_diff_cmd = NULL;
25 static int done_preparing = 0;
28 return external_diff_cmd;
31 * Default values above are meant to match the
32 * Linux kernel development style. Examples of
33 * alternative styles you can specify via environment
38 if (gitenv("GIT_EXTERNAL_DIFF"))
39 external_diff_cmd = gitenv("GIT_EXTERNAL_DIFF");
41 /* In case external diff fails... */
42 diff_opts = gitenv("GIT_DIFF_OPTS") ? : diff_opts;
45 return external_diff_cmd;
48 /* Help to copy the thing properly quoted for the shell safety.
49 * any single quote is replaced with '\'', and the caller is
50 * expected to enclose the result within a single quote pair.
53 * original sq_expand result
54 * name ==> name ==> 'name'
55 * a b ==> a b ==> 'a b'
56 * a'b ==> a'\''b ==> 'a'\''b'
58 static char *sq_expand(const char *src)
60 static char *buf = NULL;
65 /* count bytes needed to store the quoted string. */
66 for (cnt = 1, cp = src; *cp; cnt++, cp++)
72 while ((c = *src++)) {
76 bp = strcpy(bp, "'\\''");
84 static struct diff_tempfile {
85 const char *name; /* filename external diff should read from */
91 static void builtin_diff(const char *name_a,
93 struct diff_tempfile *temp,
96 int i, next_at, cmd_size;
97 const char *diff_cmd = "diff -L'%s%s' -L'%s%s'";
98 const char *diff_arg = "'%s' '%s'||:"; /* "||:" is to return 0 */
99 const char *input_name_sq[2];
100 const char *path0[2];
101 const char *path1[2];
102 const char *name_sq[2];
105 name_sq[0] = sq_expand(name_a);
106 name_sq[1] = sq_expand(name_b);
108 /* diff_cmd and diff_arg have 6 %s in total which makes
109 * the sum of these strings 12 bytes larger than required.
110 * we use 2 spaces around diff-opts, and we need to count
111 * terminating NUL, so we subtract 9 here.
113 cmd_size = (strlen(diff_cmd) + strlen(diff_opts) +
114 strlen(diff_arg) - 9);
115 for (i = 0; i < 2; i++) {
116 input_name_sq[i] = sq_expand(temp[i].name);
117 if (!strcmp(temp[i].name, "/dev/null")) {
118 path0[i] = "/dev/null";
121 path0[i] = i ? "b/" : "a/";
122 path1[i] = name_sq[i];
124 cmd_size += (strlen(path0[i]) + strlen(path1[i]) +
125 strlen(input_name_sq[i]));
128 cmd = xmalloc(cmd_size);
131 next_at += snprintf(cmd+next_at, cmd_size-next_at,
133 path0[0], path1[0], path0[1], path1[1]);
134 next_at += snprintf(cmd+next_at, cmd_size-next_at,
136 next_at += snprintf(cmd+next_at, cmd_size-next_at,
137 diff_arg, input_name_sq[0], input_name_sq[1]);
139 printf("diff --git a/%s b/%s\n", name_a, name_b);
141 printf("new file mode %s\n", temp[1].mode);
142 else if (!path1[1][0])
143 printf("deleted file mode %s\n", temp[0].mode);
145 if (strcmp(temp[0].mode, temp[1].mode)) {
146 printf("old mode %s\n", temp[0].mode);
147 printf("new mode %s\n", temp[1].mode);
149 if (xfrm_msg && xfrm_msg[0])
150 fputs(xfrm_msg, stdout);
152 if (strncmp(temp[0].mode, temp[1].mode, 3))
153 /* we do not run diff between different kind
159 execlp("/bin/sh","sh", "-c", cmd, NULL);
162 struct diff_filespec *alloc_filespec(const char *path)
164 int namelen = strlen(path);
165 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
166 spec->path = (char *)(spec + 1);
167 strcpy(spec->path, path);
168 spec->should_free = spec->should_munmap = 0;
169 spec->xfrm_flags = 0;
173 memset(spec->sha1, 0, 20);
177 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
180 if (mode) { /* just playing defensive */
182 memcpy(spec->sha1, sha1, 20);
183 spec->sha1_valid = !!memcmp(sha1, null_sha1, 20);
188 * Given a name and sha1 pair, if the dircache tells us the file in
189 * the work tree has that object contents, return true, so that
190 * prepare_temp_file() does not have to inflate and extract.
192 static int work_tree_matches(const char *name, const unsigned char *sha1)
194 struct cache_entry *ce;
198 /* We do not read the cache ourselves here, because the
199 * benchmark with my previous version that always reads cache
200 * shows that it makes things worse for diff-tree comparing
201 * two linux-2.6 kernel trees in an already checked out work
202 * tree. This is because most diff-tree comparisons deal with
203 * only a small number of files, while reading the cache is
204 * expensive for a large project, and its cost outweighs the
205 * savings we get by not inflating the object to a temporary
206 * file. Practically, this code only helps when we are used
207 * by diff-cache --cached, which does read the cache before
214 pos = cache_name_pos(name, len);
217 ce = active_cache[pos];
218 if ((lstat(name, &st) < 0) ||
219 !S_ISREG(st.st_mode) || /* careful! */
220 ce_match_stat(ce, &st) ||
221 memcmp(sha1, ce->sha1, 20))
223 /* we return 1 only when we can stat, it is a regular file,
224 * stat information matches, and sha1 recorded in the cache
225 * matches. I.e. we know the file in the work tree really is
226 * the same as the <name, sha1> pair.
232 * While doing rename detection and pickaxe operation, we may need to
233 * grab the data for the blob (or file) for our own in-core comparison.
234 * diff_filespec has data and size fields for this purpose.
236 int diff_populate_filespec(struct diff_filespec *s)
239 if (!DIFF_FILE_VALID(s))
240 die("internal error: asking to populate invalid file.");
241 if (S_ISDIR(s->mode))
246 if (!s->sha1_valid ||
247 work_tree_matches(s->path, s->sha1)) {
250 if (lstat(s->path, &st) < 0) {
251 if (errno == ENOENT) {
260 s->size = st.st_size;
263 if (S_ISLNK(st.st_mode)) {
265 s->data = xmalloc(s->size);
267 ret = readlink(s->path, s->data, s->size);
274 fd = open(s->path, O_RDONLY);
277 s->data = mmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
278 s->should_munmap = 1;
283 s->data = read_sha1_file(s->sha1, type, &s->size);
289 void diff_free_filespec_data(struct diff_filespec *s)
293 else if (s->should_munmap)
294 munmap(s->data, s->size);
295 s->should_free = s->should_munmap = 0;
299 static void prep_temp_blob(struct diff_tempfile *temp,
307 strcpy(temp->tmp_path, ".diff_XXXXXX");
308 fd = mkstemp(temp->tmp_path);
310 die("unable to create temp-file");
311 if (write(fd, blob, size) != size)
312 die("unable to write temp-file");
314 temp->name = temp->tmp_path;
315 strcpy(temp->hex, sha1_to_hex(sha1));
317 sprintf(temp->mode, "%06o", mode);
320 static void prepare_temp_file(const char *name,
321 struct diff_tempfile *temp,
322 struct diff_filespec *one)
324 if (!DIFF_FILE_VALID(one)) {
326 /* A '-' entry produces this for file-2, and
327 * a '+' entry produces this for file-1.
329 temp->name = "/dev/null";
330 strcpy(temp->hex, ".");
331 strcpy(temp->mode, ".");
335 if (!one->sha1_valid ||
336 work_tree_matches(name, one->sha1)) {
338 if (lstat(name, &st) < 0) {
340 goto not_a_valid_file;
341 die("stat(%s): %s", name, strerror(errno));
343 if (S_ISLNK(st.st_mode)) {
345 char *buf, buf_[1024];
346 buf = ((sizeof(buf_) < st.st_size) ?
347 xmalloc(st.st_size) : buf_);
348 ret = readlink(name, buf, st.st_size);
350 die("readlink(%s)", name);
351 prep_temp_blob(temp, buf, st.st_size,
353 one->sha1 : null_sha1),
355 one->mode : S_IFLNK));
358 /* we can borrow from the file in the work tree */
360 if (!one->sha1_valid)
361 strcpy(temp->hex, sha1_to_hex(null_sha1));
363 strcpy(temp->hex, sha1_to_hex(one->sha1));
364 sprintf(temp->mode, "%06o",
365 S_IFREG |ce_permissions(st.st_mode));
370 if (diff_populate_filespec(one))
371 die("cannot read data blob for %s", one->path);
372 prep_temp_blob(temp, one->data, one->size,
373 one->sha1, one->mode);
377 static void remove_tempfile(void)
381 for (i = 0; i < 2; i++)
382 if (diff_temp[i].name == diff_temp[i].tmp_path) {
383 unlink(diff_temp[i].name);
384 diff_temp[i].name = NULL;
388 static void remove_tempfile_on_signal(int signo)
393 static int matches_pathspec(const char *name)
401 namelen = strlen(name);
402 for (i = 0; i < speccnt; i++) {
403 int speclen = strlen(pathspec[i]);
404 if (! strncmp(pathspec[i], name, speclen) &&
405 speclen <= namelen &&
406 (name[speclen] == 0 || name[speclen] == '/'))
412 /* An external diff command takes:
414 * diff-cmd name infile1 infile1-sha1 infile1-mode \
415 * infile2 infile2-sha1 infile2-mode [ rename-to ]
418 static void run_external_diff(const char *name,
420 struct diff_filespec *one,
421 struct diff_filespec *two,
422 const char *xfrm_msg)
424 struct diff_tempfile *temp = diff_temp;
427 static int atexit_asked = 0;
429 if (!matches_pathspec(name) && (!other || !matches_pathspec(other)))
433 prepare_temp_file(name, &temp[0], one);
434 prepare_temp_file(other ? : name, &temp[1], two);
435 if (! atexit_asked &&
436 (temp[0].name == temp[0].tmp_path ||
437 temp[1].name == temp[1].tmp_path)) {
439 atexit(remove_tempfile);
441 signal(SIGINT, remove_tempfile_on_signal);
447 die("unable to fork");
449 const char *pgm = external_diff();
452 const char *exec_arg[10];
453 const char **arg = &exec_arg[0];
456 *arg++ = temp[0].name;
457 *arg++ = temp[0].hex;
458 *arg++ = temp[0].mode;
459 *arg++ = temp[1].name;
460 *arg++ = temp[1].hex;
461 *arg++ = temp[1].mode;
467 execvp(pgm, (char *const*) exec_arg);
470 execlp(pgm, pgm, name, NULL);
473 * otherwise we use the built-in one.
476 builtin_diff(name, other ? : name, temp, xfrm_msg);
478 printf("* Unmerged path %s\n", name);
481 if (waitpid(pid, &status, 0) < 0 ||
482 !WIFEXITED(status) || WEXITSTATUS(status)) {
483 /* Earlier we did not check the exit status because
484 * diff exits non-zero if files are different, and
485 * we are not interested in knowing that. It was a
486 * mistake which made it harder to quit a diff-*
487 * session that uses the git-apply-patch-script as
488 * the GIT_EXTERNAL_DIFF. A custom GIT_EXTERNAL_DIFF
489 * should also exit non-zero only when it wants to
490 * abort the entire diff-* session.
493 fprintf(stderr, "external diff died, stopping at %s.\n", name);
499 int diff_scoreopt_parse(const char *opt)
501 int diglen, num, scale, i;
502 if (opt[0] != '-' || (opt[1] != 'M' && opt[1] != 'C'))
503 return -1; /* that is not a -M nor -C option */
504 diglen = strspn(opt+2, "0123456789");
505 if (diglen == 0 || strlen(opt+2) != diglen)
506 return 0; /* use default */
507 sscanf(opt+2, "%d", &num);
508 for (i = 0, scale = 1; i < diglen; i++)
511 /* user says num divided by scale and we say internally that
512 * is MAX_SCORE * num / scale.
514 return MAX_SCORE * num / scale;
517 void diff_setup(int reverse_diff_, int diff_output_style)
519 reverse_diff = reverse_diff_;
521 switch (diff_output_style) {
522 case DIFF_FORMAT_HUMAN:
523 line_termination = '\n';
524 inter_name_termination = '\t';
526 case DIFF_FORMAT_MACHINE:
527 line_termination = inter_name_termination = 0;
529 case DIFF_FORMAT_PATCH:
535 struct diff_queue_struct diff_queued_diff;
537 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
538 struct diff_filespec *one,
539 struct diff_filespec *two)
541 struct diff_filepair *dp = xmalloc(sizeof(*dp));
545 dp->orig_order = queue->nr;
547 if (queue->alloc <= queue->nr) {
548 queue->alloc = alloc_nr(queue->alloc);
549 queue->queue = xrealloc(queue->queue,
550 sizeof(dp) * queue->alloc);
552 queue->queue[queue->nr++] = dp;
556 static void diff_flush_raw(struct diff_filepair *p)
559 * We used to reject rename/copy but new diff-raw can express them.
561 printf(":%06o %06o %s ",
562 p->one->mode, p->two->mode, sha1_to_hex(p->one->sha1));
563 printf("%s%c%s%c%s%c",
564 sha1_to_hex(p->two->sha1), inter_name_termination,
565 p->one->path, inter_name_termination,
566 p->two->path, line_termination);
569 static void diff_flush_patch(struct diff_filepair *p)
571 const char *name, *other;
574 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
575 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
576 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
577 return; /* no tree diffs in patch format */
579 run_external_diff(name, other, p->one, p->two, p->xfrm_msg);
582 static int identical(struct diff_filespec *one, struct diff_filespec *two)
584 /* This function is written stricter than necessary to support
585 * the currently implemented transformers, but the idea is to
586 * let transformers to produce diff_filepairs any way they want,
587 * and filter and clean them up here before producing the output.
590 if (!DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two))
591 return 1; /* not interesting */
593 /* deletion, addition, mode change and renames are all interesting. */
594 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
595 (one->mode != two->mode) ||
596 strcmp(one->path, two->path))
599 /* both are valid and point at the same path. that is, we are
600 * dealing with a change.
602 if (one->sha1_valid && two->sha1_valid &&
603 !memcmp(one->sha1, two->sha1, sizeof(one->sha1)))
604 return 1; /* no change */
605 if (!one->sha1_valid && !two->sha1_valid)
606 return 1; /* both look at the same file on the filesystem. */
610 static void diff_flush_one(struct diff_filepair *p)
612 if (identical(p->one, p->two))
620 int diff_queue_is_empty(void)
622 struct diff_queue_struct *q = &diff_queued_diff;
625 for (i = 0; i < q->nr; i++) {
626 struct diff_filepair *p = q->queue[i];
627 if (!identical(p->one, p->two))
633 void diff_flush(const char **pathspec_, int speccnt_)
635 struct diff_queue_struct *q = &diff_queued_diff;
638 pathspec = pathspec_;
641 for (i = 0; i < q->nr; i++)
642 diff_flush_one(q->queue[i]);
644 for (i = 0; i < q->nr; i++) {
645 struct diff_filepair *p = q->queue[i];
646 diff_free_filespec_data(p->one);
647 diff_free_filespec_data(p->two);
653 q->nr = q->alloc = 0;
656 void diff_addremove(int addremove, unsigned mode,
657 const unsigned char *sha1,
658 const char *base, const char *path)
660 char concatpath[PATH_MAX];
661 struct diff_filespec *one, *two;
663 /* This may look odd, but it is a preparation for
664 * feeding "there are unchanged files which should
665 * not produce diffs, but when you are doing copy
666 * detection you would need them, so here they are"
667 * entries to the diff-core. They will be prefixed
668 * with something like '=' or '*' (I haven't decided
669 * which but should not make any difference).
670 * Feeding the same new and old to diff_change() should
671 * also have the same effect. diff_flush() should
672 * filter the identical ones out at the final output
676 addremove = (addremove == '+' ? '-' :
677 addremove == '-' ? '+' : addremove);
679 if (!path) path = "";
680 sprintf(concatpath, "%s%s", base, path);
681 one = alloc_filespec(concatpath);
682 two = alloc_filespec(concatpath);
684 if (addremove != '+')
685 fill_filespec(one, sha1, mode);
686 if (addremove != '-')
687 fill_filespec(two, sha1, mode);
689 diff_queue(&diff_queued_diff, one, two);
692 void diff_guif(unsigned old_mode,
694 const unsigned char *old_sha1,
695 const unsigned char *new_sha1,
696 const char *old_path,
697 const char *new_path)
699 struct diff_filespec *one, *two;
703 const unsigned char *tmp_c;
704 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
705 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
707 one = alloc_filespec(old_path);
708 two = alloc_filespec(new_path);
710 fill_filespec(one, old_sha1, old_mode);
712 fill_filespec(two, new_sha1, new_mode);
713 diff_queue(&diff_queued_diff, one, two);
716 void diff_change(unsigned old_mode, unsigned new_mode,
717 const unsigned char *old_sha1,
718 const unsigned char *new_sha1,
719 const char *base, const char *path)
721 char concatpath[PATH_MAX];
722 struct diff_filespec *one, *two;
726 const unsigned char *tmp_c;
727 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
728 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
730 if (!path) path = "";
731 sprintf(concatpath, "%s%s", base, path);
732 one = alloc_filespec(concatpath);
733 two = alloc_filespec(concatpath);
734 fill_filespec(one, old_sha1, old_mode);
735 fill_filespec(two, new_sha1, new_mode);
737 diff_queue(&diff_queued_diff, one, two);
740 void diff_unmerge(const char *path)
743 run_external_diff(path, NULL, NULL, NULL, NULL);
745 printf("U %s%c", path, line_termination);