2 * Copyright (C) 2005 Junio C Hamano
11 #include "xdiff-interface.h"
13 static int use_size_cache;
15 int diff_rename_limit_default = -1;
17 int git_diff_config(const char *var, const char *value)
19 if (!strcmp(var, "diff.renamelimit")) {
20 diff_rename_limit_default = git_config_int(var, value);
24 return git_default_config(var, value);
27 static char *quote_one(const char *str)
34 needlen = quote_c_style(str, NULL, NULL, 0);
37 xp = xmalloc(needlen + 1);
38 quote_c_style(str, xp, NULL, 0);
42 static char *quote_two(const char *one, const char *two)
44 int need_one = quote_c_style(one, NULL, NULL, 1);
45 int need_two = quote_c_style(two, NULL, NULL, 1);
48 if (need_one + need_two) {
49 if (!need_one) need_one = strlen(one);
50 if (!need_two) need_one = strlen(two);
52 xp = xmalloc(need_one + need_two + 3);
54 quote_c_style(one, xp + 1, NULL, 1);
55 quote_c_style(two, xp + need_one + 1, NULL, 1);
56 strcpy(xp + need_one + need_two + 1, "\"");
59 need_one = strlen(one);
60 need_two = strlen(two);
61 xp = xmalloc(need_one + need_two + 1);
63 strcpy(xp + need_one, two);
67 static const char *external_diff(void)
69 static const char *external_diff_cmd = NULL;
70 static int done_preparing = 0;
73 return external_diff_cmd;
74 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
76 return external_diff_cmd;
79 #define TEMPFILE_PATH_LEN 50
81 static struct diff_tempfile {
82 const char *name; /* filename external diff should read from */
85 char tmp_path[TEMPFILE_PATH_LEN];
88 static int count_lines(const char *data, int size)
90 int count, ch, completely_empty = 1, nl_just_seen = 0;
101 completely_empty = 0;
104 if (completely_empty)
107 count++; /* no trailing newline */
111 static void print_line_count(int count)
121 printf("1,%d", count);
126 static void copy_file(int prefix, const char *data, int size)
128 int ch, nl_just_seen = 1;
140 printf("\n\\ No newline at end of file\n");
143 static void emit_rewrite_diff(const char *name_a,
145 struct diff_filespec *one,
146 struct diff_filespec *two)
149 diff_populate_filespec(one, 0);
150 diff_populate_filespec(two, 0);
151 lc_a = count_lines(one->data, one->size);
152 lc_b = count_lines(two->data, two->size);
153 printf("--- %s\n+++ %s\n@@ -", name_a, name_b);
154 print_line_count(lc_a);
156 print_line_count(lc_b);
159 copy_file('-', one->data, one->size);
161 copy_file('+', two->data, two->size);
164 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
166 if (!DIFF_FILE_VALID(one)) {
167 mf->ptr = ""; /* does not matter */
171 else if (diff_populate_filespec(one, 0))
174 mf->size = one->size;
178 struct emit_callback {
179 const char **label_path;
182 static int fn_out(void *priv, mmbuffer_t *mb, int nbuf)
185 struct emit_callback *ecbdata = priv;
187 if (ecbdata->label_path[0]) {
188 printf("--- %s\n", ecbdata->label_path[0]);
189 printf("+++ %s\n", ecbdata->label_path[1]);
190 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
192 for (i = 0; i < nbuf; i++)
193 if (!fwrite(mb[i].ptr, mb[i].size, 1, stdout))
199 struct xdiff_emit_state xm;
203 struct diffstat_file {
205 unsigned is_unmerged:1;
206 unsigned is_binary:1;
207 unsigned int added, deleted;
211 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
214 struct diffstat_file *x;
215 x = xcalloc(sizeof (*x), 1);
216 if (diffstat->nr == diffstat->alloc) {
217 diffstat->alloc = alloc_nr(diffstat->alloc);
218 diffstat->files = xrealloc(diffstat->files,
219 diffstat->alloc * sizeof(x));
221 diffstat->files[diffstat->nr++] = x;
222 x->name = strdup(name);
226 static void diffstat_consume(void *priv, char *line, unsigned long len)
228 struct diffstat_t *diffstat = priv;
229 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
233 else if (line[0] == '-')
237 static const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
238 static const char minuses[]= "----------------------------------------------------------------------";
240 static void show_stats(struct diffstat_t* data)
243 int i, len, add, del, total, adds = 0, dels = 0;
244 int max, max_change = 0, max_len = 0;
245 int total_files = data->nr;
250 for (i = 0; i < data->nr; i++) {
251 struct diffstat_file *file = data->files[i];
253 if (file->is_binary || file->is_unmerged)
255 if (max_change < file->added + file->deleted)
256 max_change = file->added + file->deleted;
257 len = strlen(file->name);
262 for (i = 0; i < data->nr; i++) {
263 char *name = data->files[i]->name;
264 int added = data->files[i]->added;
265 int deleted = data->files[i]->deleted;
267 if (0 < (len = quote_c_style(name, NULL, NULL, 0))) {
268 char *qname = xmalloc(len + 1);
269 quote_c_style(name, qname, NULL, 0);
271 data->files[i]->name = name = qname;
275 * "scale" the filename
286 slash = strchr(name, '/');
293 * scale the add/delete
299 if (data->files[i]->is_binary) {
300 printf(" %s%-*s | Bin\n", prefix, len, name);
301 goto free_diffstat_file;
303 else if (data->files[i]->is_unmerged) {
304 printf(" %s%-*s | Unmerged\n", prefix, len, name);
305 goto free_diffstat_file;
307 else if (added + deleted == 0) {
309 goto free_diffstat_file;
318 if (max_change > 0) {
319 total = (total * max + max_change / 2) / max_change;
320 add = (add * max + max_change / 2) / max_change;
323 printf(" %s%-*s |%5d %.*s%.*s\n", prefix,
324 len, name, added + deleted,
325 add, pluses, del, minuses);
327 free(data->files[i]->name);
328 free(data->files[i]);
331 printf(" %d files changed, %d insertions(+), %d deletions(-)\n",
332 total_files, adds, dels);
335 #define FIRST_FEW_BYTES 8000
336 static int mmfile_is_binary(mmfile_t *mf)
339 if (FIRST_FEW_BYTES < sz)
340 sz = FIRST_FEW_BYTES;
341 if (memchr(mf->ptr, 0, sz))
346 static void builtin_diff(const char *name_a,
348 struct diff_filespec *one,
349 struct diff_filespec *two,
350 const char *xfrm_msg,
351 int complete_rewrite)
357 a_one = quote_two("a/", name_a);
358 b_two = quote_two("b/", name_b);
359 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
360 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
361 printf("diff --git %s %s\n", a_one, b_two);
362 if (lbl[0][0] == '/') {
364 printf("new file mode %06o\n", two->mode);
365 if (xfrm_msg && xfrm_msg[0])
368 else if (lbl[1][0] == '/') {
369 printf("deleted file mode %06o\n", one->mode);
370 if (xfrm_msg && xfrm_msg[0])
374 if (one->mode != two->mode) {
375 printf("old mode %06o\n", one->mode);
376 printf("new mode %06o\n", two->mode);
378 if (xfrm_msg && xfrm_msg[0])
381 * we do not run diff between different kind
384 if ((one->mode ^ two->mode) & S_IFMT)
385 goto free_ab_and_return;
386 if (complete_rewrite) {
387 emit_rewrite_diff(name_a, name_b, one, two);
388 goto free_ab_and_return;
392 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
393 die("unable to read files to diff");
395 if (mmfile_is_binary(&mf1) || mmfile_is_binary(&mf2))
396 printf("Binary files %s and %s differ\n", lbl[0], lbl[1]);
398 /* Crazy xdl interfaces.. */
399 const char *diffopts = getenv("GIT_DIFF_OPTS");
403 struct emit_callback ecbdata;
405 ecbdata.label_path = lbl;
406 xpp.flags = XDF_NEED_MINIMAL;
408 xecfg.flags = XDL_EMIT_FUNCNAMES;
411 else if (!strncmp(diffopts, "--unified=", 10))
412 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
413 else if (!strncmp(diffopts, "-u", 2))
414 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
417 xdl_diff(&mf1, &mf2, &xpp, &xecfg, &ecb);
426 static void builtin_diffstat(const char *name_a, const char *name_b,
427 struct diff_filespec *one, struct diff_filespec *two,
428 struct diffstat_t *diffstat)
431 struct diffstat_file *data;
433 data = diffstat_add(diffstat, name_a ? name_a : name_b);
436 data->is_unmerged = 1;
440 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
441 die("unable to read files to diff");
443 if (mmfile_is_binary(&mf1) || mmfile_is_binary(&mf2))
446 /* Crazy xdl interfaces.. */
451 xpp.flags = XDF_NEED_MINIMAL;
454 ecb.outf = xdiff_outf;
456 xdl_diff(&mf1, &mf2, &xpp, &xecfg, &ecb);
460 struct diff_filespec *alloc_filespec(const char *path)
462 int namelen = strlen(path);
463 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
465 memset(spec, 0, sizeof(*spec));
466 spec->path = (char *)(spec + 1);
467 memcpy(spec->path, path, namelen+1);
471 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
475 spec->mode = canon_mode(mode);
476 memcpy(spec->sha1, sha1, 20);
477 spec->sha1_valid = !!memcmp(sha1, null_sha1, 20);
482 * Given a name and sha1 pair, if the dircache tells us the file in
483 * the work tree has that object contents, return true, so that
484 * prepare_temp_file() does not have to inflate and extract.
486 static int work_tree_matches(const char *name, const unsigned char *sha1)
488 struct cache_entry *ce;
492 /* We do not read the cache ourselves here, because the
493 * benchmark with my previous version that always reads cache
494 * shows that it makes things worse for diff-tree comparing
495 * two linux-2.6 kernel trees in an already checked out work
496 * tree. This is because most diff-tree comparisons deal with
497 * only a small number of files, while reading the cache is
498 * expensive for a large project, and its cost outweighs the
499 * savings we get by not inflating the object to a temporary
500 * file. Practically, this code only helps when we are used
501 * by diff-cache --cached, which does read the cache before
508 pos = cache_name_pos(name, len);
511 ce = active_cache[pos];
512 if ((lstat(name, &st) < 0) ||
513 !S_ISREG(st.st_mode) || /* careful! */
514 ce_match_stat(ce, &st, 0) ||
515 memcmp(sha1, ce->sha1, 20))
517 /* we return 1 only when we can stat, it is a regular file,
518 * stat information matches, and sha1 recorded in the cache
519 * matches. I.e. we know the file in the work tree really is
520 * the same as the <name, sha1> pair.
525 static struct sha1_size_cache {
526 unsigned char sha1[20];
529 static int sha1_size_cache_nr, sha1_size_cache_alloc;
531 static struct sha1_size_cache *locate_size_cache(unsigned char *sha1,
536 struct sha1_size_cache *e;
539 last = sha1_size_cache_nr;
540 while (last > first) {
541 int cmp, next = (last + first) >> 1;
542 e = sha1_size_cache[next];
543 cmp = memcmp(e->sha1, sha1, 20);
555 /* insert to make it at "first" */
556 if (sha1_size_cache_alloc <= sha1_size_cache_nr) {
557 sha1_size_cache_alloc = alloc_nr(sha1_size_cache_alloc);
558 sha1_size_cache = xrealloc(sha1_size_cache,
559 sha1_size_cache_alloc *
560 sizeof(*sha1_size_cache));
562 sha1_size_cache_nr++;
563 if (first < sha1_size_cache_nr)
564 memmove(sha1_size_cache + first + 1, sha1_size_cache + first,
565 (sha1_size_cache_nr - first - 1) *
566 sizeof(*sha1_size_cache));
567 e = xmalloc(sizeof(struct sha1_size_cache));
568 sha1_size_cache[first] = e;
569 memcpy(e->sha1, sha1, 20);
575 * While doing rename detection and pickaxe operation, we may need to
576 * grab the data for the blob (or file) for our own in-core comparison.
577 * diff_filespec has data and size fields for this purpose.
579 int diff_populate_filespec(struct diff_filespec *s, int size_only)
582 if (!DIFF_FILE_VALID(s))
583 die("internal error: asking to populate invalid file.");
584 if (S_ISDIR(s->mode))
592 if (!s->sha1_valid ||
593 work_tree_matches(s->path, s->sha1)) {
596 if (lstat(s->path, &st) < 0) {
597 if (errno == ENOENT) {
606 s->size = st.st_size;
611 if (S_ISLNK(st.st_mode)) {
613 s->data = xmalloc(s->size);
615 ret = readlink(s->path, s->data, s->size);
622 fd = open(s->path, O_RDONLY);
625 s->data = mmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
627 if (s->data == MAP_FAILED)
629 s->should_munmap = 1;
633 struct sha1_size_cache *e;
636 e = locate_size_cache(s->sha1, 1, 0);
641 if (!sha1_object_info(s->sha1, type, &s->size))
642 locate_size_cache(s->sha1, 0, s->size);
645 s->data = read_sha1_file(s->sha1, type, &s->size);
652 void diff_free_filespec_data(struct diff_filespec *s)
656 else if (s->should_munmap)
657 munmap(s->data, s->size);
658 s->should_free = s->should_munmap = 0;
664 static void prep_temp_blob(struct diff_tempfile *temp,
667 const unsigned char *sha1,
672 fd = git_mkstemp(temp->tmp_path, TEMPFILE_PATH_LEN, ".diff_XXXXXX");
674 die("unable to create temp-file");
675 if (write(fd, blob, size) != size)
676 die("unable to write temp-file");
678 temp->name = temp->tmp_path;
679 strcpy(temp->hex, sha1_to_hex(sha1));
681 sprintf(temp->mode, "%06o", mode);
684 static void prepare_temp_file(const char *name,
685 struct diff_tempfile *temp,
686 struct diff_filespec *one)
688 if (!DIFF_FILE_VALID(one)) {
690 /* A '-' entry produces this for file-2, and
691 * a '+' entry produces this for file-1.
693 temp->name = "/dev/null";
694 strcpy(temp->hex, ".");
695 strcpy(temp->mode, ".");
699 if (!one->sha1_valid ||
700 work_tree_matches(name, one->sha1)) {
702 if (lstat(name, &st) < 0) {
704 goto not_a_valid_file;
705 die("stat(%s): %s", name, strerror(errno));
707 if (S_ISLNK(st.st_mode)) {
709 char buf[PATH_MAX + 1]; /* ought to be SYMLINK_MAX */
710 if (sizeof(buf) <= st.st_size)
711 die("symlink too long: %s", name);
712 ret = readlink(name, buf, st.st_size);
714 die("readlink(%s)", name);
715 prep_temp_blob(temp, buf, st.st_size,
717 one->sha1 : null_sha1),
719 one->mode : S_IFLNK));
722 /* we can borrow from the file in the work tree */
724 if (!one->sha1_valid)
725 strcpy(temp->hex, sha1_to_hex(null_sha1));
727 strcpy(temp->hex, sha1_to_hex(one->sha1));
728 /* Even though we may sometimes borrow the
729 * contents from the work tree, we always want
730 * one->mode. mode is trustworthy even when
731 * !(one->sha1_valid), as long as
732 * DIFF_FILE_VALID(one).
734 sprintf(temp->mode, "%06o", one->mode);
739 if (diff_populate_filespec(one, 0))
740 die("cannot read data blob for %s", one->path);
741 prep_temp_blob(temp, one->data, one->size,
742 one->sha1, one->mode);
746 static void remove_tempfile(void)
750 for (i = 0; i < 2; i++)
751 if (diff_temp[i].name == diff_temp[i].tmp_path) {
752 unlink(diff_temp[i].name);
753 diff_temp[i].name = NULL;
757 static void remove_tempfile_on_signal(int signo)
760 signal(SIGINT, SIG_DFL);
764 static int spawn_prog(const char *pgm, const char **arg)
772 die("unable to fork");
774 execvp(pgm, (char *const*) arg);
778 while (waitpid(pid, &status, 0) < 0) {
784 /* Earlier we did not check the exit status because
785 * diff exits non-zero if files are different, and
786 * we are not interested in knowing that. It was a
787 * mistake which made it harder to quit a diff-*
788 * session that uses the git-apply-patch-script as
789 * the GIT_EXTERNAL_DIFF. A custom GIT_EXTERNAL_DIFF
790 * should also exit non-zero only when it wants to
791 * abort the entire diff-* session.
793 if (WIFEXITED(status) && !WEXITSTATUS(status))
798 /* An external diff command takes:
800 * diff-cmd name infile1 infile1-sha1 infile1-mode \
801 * infile2 infile2-sha1 infile2-mode [ rename-to ]
804 static void run_external_diff(const char *pgm,
807 struct diff_filespec *one,
808 struct diff_filespec *two,
809 const char *xfrm_msg,
810 int complete_rewrite)
812 const char *spawn_arg[10];
813 struct diff_tempfile *temp = diff_temp;
815 static int atexit_asked = 0;
816 const char *othername;
817 const char **arg = &spawn_arg[0];
819 othername = (other? other : name);
821 prepare_temp_file(name, &temp[0], one);
822 prepare_temp_file(othername, &temp[1], two);
823 if (! atexit_asked &&
824 (temp[0].name == temp[0].tmp_path ||
825 temp[1].name == temp[1].tmp_path)) {
827 atexit(remove_tempfile);
829 signal(SIGINT, remove_tempfile_on_signal);
835 *arg++ = temp[0].name;
836 *arg++ = temp[0].hex;
837 *arg++ = temp[0].mode;
838 *arg++ = temp[1].name;
839 *arg++ = temp[1].hex;
840 *arg++ = temp[1].mode;
850 retval = spawn_prog(pgm, spawn_arg);
853 fprintf(stderr, "external diff died, stopping at %s.\n", name);
858 static void run_diff_cmd(const char *pgm,
861 struct diff_filespec *one,
862 struct diff_filespec *two,
863 const char *xfrm_msg,
864 int complete_rewrite)
867 run_external_diff(pgm, name, other, one, two, xfrm_msg,
872 builtin_diff(name, other ? other : name,
873 one, two, xfrm_msg, complete_rewrite);
875 printf("* Unmerged path %s\n", name);
878 static void diff_fill_sha1_info(struct diff_filespec *one)
880 if (DIFF_FILE_VALID(one)) {
881 if (!one->sha1_valid) {
883 if (lstat(one->path, &st) < 0)
884 die("stat %s", one->path);
885 if (index_path(one->sha1, one->path, &st, 0))
886 die("cannot hash %s\n", one->path);
890 memset(one->sha1, 0, 20);
893 static void run_diff(struct diff_filepair *p, struct diff_options *o)
895 const char *pgm = external_diff();
896 char msg[PATH_MAX*2+300], *xfrm_msg;
897 struct diff_filespec *one;
898 struct diff_filespec *two;
901 char *name_munged, *other_munged;
902 int complete_rewrite = 0;
905 if (DIFF_PAIR_UNMERGED(p)) {
907 run_diff_cmd(pgm, p->one->path, NULL, NULL, NULL, NULL, 0);
912 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
913 name_munged = quote_one(name);
914 other_munged = quote_one(other);
915 one = p->one; two = p->two;
917 diff_fill_sha1_info(one);
918 diff_fill_sha1_info(two);
922 case DIFF_STATUS_COPIED:
923 len += snprintf(msg + len, sizeof(msg) - len,
924 "similarity index %d%%\n"
927 (int)(0.5 + p->score * 100.0/MAX_SCORE),
928 name_munged, other_munged);
930 case DIFF_STATUS_RENAMED:
931 len += snprintf(msg + len, sizeof(msg) - len,
932 "similarity index %d%%\n"
935 (int)(0.5 + p->score * 100.0/MAX_SCORE),
936 name_munged, other_munged);
938 case DIFF_STATUS_MODIFIED:
940 len += snprintf(msg + len, sizeof(msg) - len,
941 "dissimilarity index %d%%\n",
942 (int)(0.5 + p->score *
944 complete_rewrite = 1;
953 if (memcmp(one->sha1, two->sha1, 20)) {
955 int abbrev = o->full_index ? 40 : DEFAULT_ABBREV;
956 memcpy(one_sha1, sha1_to_hex(one->sha1), 41);
958 len += snprintf(msg + len, sizeof(msg) - len,
960 abbrev, one_sha1, abbrev,
961 sha1_to_hex(two->sha1));
962 if (one->mode == two->mode)
963 len += snprintf(msg + len, sizeof(msg) - len,
965 len += snprintf(msg + len, sizeof(msg) - len, "\n");
970 xfrm_msg = len ? msg : NULL;
973 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
974 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
975 /* a filepair that changes between file and symlink
976 * needs to be split into deletion and creation.
978 struct diff_filespec *null = alloc_filespec(two->path);
979 run_diff_cmd(NULL, name, other, one, null, xfrm_msg, 0);
981 null = alloc_filespec(one->path);
982 run_diff_cmd(NULL, name, other, null, two, xfrm_msg, 0);
986 run_diff_cmd(pgm, name, other, one, two, xfrm_msg,
993 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
994 struct diffstat_t *diffstat)
999 if (DIFF_PAIR_UNMERGED(p)) {
1001 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat);
1005 name = p->one->path;
1006 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
1008 diff_fill_sha1_info(p->one);
1009 diff_fill_sha1_info(p->two);
1011 builtin_diffstat(name, other, p->one, p->two, diffstat);
1014 void diff_setup(struct diff_options *options)
1016 memset(options, 0, sizeof(*options));
1017 options->output_format = DIFF_FORMAT_RAW;
1018 options->line_termination = '\n';
1019 options->break_opt = -1;
1020 options->rename_limit = -1;
1022 options->change = diff_change;
1023 options->add_remove = diff_addremove;
1026 int diff_setup_done(struct diff_options *options)
1028 if ((options->find_copies_harder &&
1029 options->detect_rename != DIFF_DETECT_COPY) ||
1030 (0 <= options->rename_limit && !options->detect_rename))
1032 if (options->detect_rename && options->rename_limit < 0)
1033 options->rename_limit = diff_rename_limit_default;
1034 if (options->setup & DIFF_SETUP_USE_CACHE) {
1036 /* read-cache does not die even when it fails
1037 * so it is safe for us to do this here. Also
1038 * it does not smudge active_cache or active_nr
1039 * when it fails, so we do not have to worry about
1040 * cleaning it up ourselves either.
1044 if (options->setup & DIFF_SETUP_USE_SIZE_CACHE)
1046 if (options->abbrev <= 0 || 40 < options->abbrev)
1047 options->abbrev = 40; /* full */
1052 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
1054 const char *arg = av[0];
1055 if (!strcmp(arg, "-p") || !strcmp(arg, "-u"))
1056 options->output_format = DIFF_FORMAT_PATCH;
1057 else if (!strcmp(arg, "--patch-with-raw")) {
1058 options->output_format = DIFF_FORMAT_PATCH;
1059 options->with_raw = 1;
1061 else if (!strcmp(arg, "--stat"))
1062 options->output_format = DIFF_FORMAT_DIFFSTAT;
1063 else if (!strcmp(arg, "--patch-with-stat")) {
1064 options->output_format = DIFF_FORMAT_PATCH;
1065 options->with_stat = 1;
1067 else if (!strcmp(arg, "-z"))
1068 options->line_termination = 0;
1069 else if (!strncmp(arg, "-l", 2))
1070 options->rename_limit = strtoul(arg+2, NULL, 10);
1071 else if (!strcmp(arg, "--full-index"))
1072 options->full_index = 1;
1073 else if (!strcmp(arg, "--name-only"))
1074 options->output_format = DIFF_FORMAT_NAME;
1075 else if (!strcmp(arg, "--name-status"))
1076 options->output_format = DIFF_FORMAT_NAME_STATUS;
1077 else if (!strcmp(arg, "-R"))
1078 options->reverse_diff = 1;
1079 else if (!strncmp(arg, "-S", 2))
1080 options->pickaxe = arg + 2;
1081 else if (!strcmp(arg, "-s"))
1082 options->output_format = DIFF_FORMAT_NO_OUTPUT;
1083 else if (!strncmp(arg, "-O", 2))
1084 options->orderfile = arg + 2;
1085 else if (!strncmp(arg, "--diff-filter=", 14))
1086 options->filter = arg + 14;
1087 else if (!strcmp(arg, "--pickaxe-all"))
1088 options->pickaxe_opts = DIFF_PICKAXE_ALL;
1089 else if (!strcmp(arg, "--pickaxe-regex"))
1090 options->pickaxe_opts = DIFF_PICKAXE_REGEX;
1091 else if (!strncmp(arg, "-B", 2)) {
1092 if ((options->break_opt =
1093 diff_scoreopt_parse(arg)) == -1)
1096 else if (!strncmp(arg, "-M", 2)) {
1097 if ((options->rename_score =
1098 diff_scoreopt_parse(arg)) == -1)
1100 options->detect_rename = DIFF_DETECT_RENAME;
1102 else if (!strncmp(arg, "-C", 2)) {
1103 if ((options->rename_score =
1104 diff_scoreopt_parse(arg)) == -1)
1106 options->detect_rename = DIFF_DETECT_COPY;
1108 else if (!strcmp(arg, "--find-copies-harder"))
1109 options->find_copies_harder = 1;
1110 else if (!strcmp(arg, "--abbrev"))
1111 options->abbrev = DEFAULT_ABBREV;
1112 else if (!strncmp(arg, "--abbrev=", 9)) {
1113 options->abbrev = strtoul(arg + 9, NULL, 10);
1114 if (options->abbrev < MINIMUM_ABBREV)
1115 options->abbrev = MINIMUM_ABBREV;
1116 else if (40 < options->abbrev)
1117 options->abbrev = 40;
1124 static int parse_num(const char **cp_p)
1126 unsigned long num, scale;
1128 const char *cp = *cp_p;
1135 if ( !dot && ch == '.' ) {
1138 } else if ( ch == '%' ) {
1139 scale = dot ? scale*100 : 100;
1140 cp++; /* % is always at the end */
1142 } else if ( ch >= '0' && ch <= '9' ) {
1143 if ( scale < 100000 ) {
1145 num = (num*10) + (ch-'0');
1154 /* user says num divided by scale and we say internally that
1155 * is MAX_SCORE * num / scale.
1157 return (num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale);
1160 int diff_scoreopt_parse(const char *opt)
1162 int opt1, opt2, cmd;
1167 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
1168 return -1; /* that is not a -M, -C nor -B option */
1170 opt1 = parse_num(&opt);
1176 else if (*opt != '/')
1177 return -1; /* we expect -B80/99 or -B80 */
1180 opt2 = parse_num(&opt);
1185 return opt1 | (opt2 << 16);
1188 struct diff_queue_struct diff_queued_diff;
1190 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
1192 if (queue->alloc <= queue->nr) {
1193 queue->alloc = alloc_nr(queue->alloc);
1194 queue->queue = xrealloc(queue->queue,
1195 sizeof(dp) * queue->alloc);
1197 queue->queue[queue->nr++] = dp;
1200 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
1201 struct diff_filespec *one,
1202 struct diff_filespec *two)
1204 struct diff_filepair *dp = xmalloc(sizeof(*dp));
1209 dp->source_stays = 0;
1210 dp->broken_pair = 0;
1216 void diff_free_filepair(struct diff_filepair *p)
1218 diff_free_filespec_data(p->one);
1219 diff_free_filespec_data(p->two);
1225 /* This is different from find_unique_abbrev() in that
1226 * it stuffs the result with dots for alignment.
1228 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
1233 return sha1_to_hex(sha1);
1235 abbrev = find_unique_abbrev(sha1, len);
1237 return sha1_to_hex(sha1);
1238 abblen = strlen(abbrev);
1240 static char hex[41];
1241 if (len < abblen && abblen <= len + 2)
1242 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
1244 sprintf(hex, "%s...", abbrev);
1247 return sha1_to_hex(sha1);
1250 static void diff_flush_raw(struct diff_filepair *p,
1251 int line_termination,
1252 int inter_name_termination,
1253 struct diff_options *options,
1258 int abbrev = options->abbrev;
1259 const char *path_one, *path_two;
1261 path_one = p->one->path;
1262 path_two = p->two->path;
1263 if (line_termination) {
1264 path_one = quote_one(path_one);
1265 path_two = quote_one(path_two);
1269 sprintf(status, "%c%03d", p->status,
1270 (int)(0.5 + p->score * 100.0/MAX_SCORE));
1272 status[0] = p->status;
1275 switch (p->status) {
1276 case DIFF_STATUS_COPIED:
1277 case DIFF_STATUS_RENAMED:
1280 case DIFF_STATUS_ADDED:
1281 case DIFF_STATUS_DELETED:
1288 if (output_format != DIFF_FORMAT_NAME_STATUS) {
1289 printf(":%06o %06o %s ",
1290 p->one->mode, p->two->mode,
1291 diff_unique_abbrev(p->one->sha1, abbrev));
1293 diff_unique_abbrev(p->two->sha1, abbrev));
1295 printf("%s%c%s", status, inter_name_termination, path_one);
1297 printf("%c%s", inter_name_termination, path_two);
1298 putchar(line_termination);
1299 if (path_one != p->one->path)
1300 free((void*)path_one);
1301 if (path_two != p->two->path)
1302 free((void*)path_two);
1305 static void diff_flush_name(struct diff_filepair *p,
1306 int inter_name_termination,
1307 int line_termination)
1309 char *path = p->two->path;
1311 if (line_termination)
1312 path = quote_one(p->two->path);
1314 path = p->two->path;
1315 printf("%s%c", path, line_termination);
1316 if (p->two->path != path)
1320 int diff_unmodified_pair(struct diff_filepair *p)
1322 /* This function is written stricter than necessary to support
1323 * the currently implemented transformers, but the idea is to
1324 * let transformers to produce diff_filepairs any way they want,
1325 * and filter and clean them up here before producing the output.
1327 struct diff_filespec *one, *two;
1329 if (DIFF_PAIR_UNMERGED(p))
1330 return 0; /* unmerged is interesting */
1335 /* deletion, addition, mode or type change
1336 * and rename are all interesting.
1338 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
1339 DIFF_PAIR_MODE_CHANGED(p) ||
1340 strcmp(one->path, two->path))
1343 /* both are valid and point at the same path. that is, we are
1344 * dealing with a change.
1346 if (one->sha1_valid && two->sha1_valid &&
1347 !memcmp(one->sha1, two->sha1, sizeof(one->sha1)))
1348 return 1; /* no change */
1349 if (!one->sha1_valid && !two->sha1_valid)
1350 return 1; /* both look at the same file on the filesystem. */
1354 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
1356 if (diff_unmodified_pair(p))
1359 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
1360 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
1361 return; /* no tree diffs in patch format */
1366 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
1367 struct diffstat_t *diffstat)
1369 if (diff_unmodified_pair(p))
1372 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
1373 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
1374 return; /* no tree diffs in patch format */
1376 run_diffstat(p, o, diffstat);
1379 int diff_queue_is_empty(void)
1381 struct diff_queue_struct *q = &diff_queued_diff;
1383 for (i = 0; i < q->nr; i++)
1384 if (!diff_unmodified_pair(q->queue[i]))
1390 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
1392 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
1395 DIFF_FILE_VALID(s) ? "valid" : "invalid",
1397 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
1398 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
1400 s->size, s->xfrm_flags);
1403 void diff_debug_filepair(const struct diff_filepair *p, int i)
1405 diff_debug_filespec(p->one, i, "one");
1406 diff_debug_filespec(p->two, i, "two");
1407 fprintf(stderr, "score %d, status %c stays %d broken %d\n",
1408 p->score, p->status ? p->status : '?',
1409 p->source_stays, p->broken_pair);
1412 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
1416 fprintf(stderr, "%s\n", msg);
1417 fprintf(stderr, "q->nr = %d\n", q->nr);
1418 for (i = 0; i < q->nr; i++) {
1419 struct diff_filepair *p = q->queue[i];
1420 diff_debug_filepair(p, i);
1425 static void diff_resolve_rename_copy(void)
1428 struct diff_filepair *p, *pp;
1429 struct diff_queue_struct *q = &diff_queued_diff;
1431 diff_debug_queue("resolve-rename-copy", q);
1433 for (i = 0; i < q->nr; i++) {
1435 p->status = 0; /* undecided */
1436 if (DIFF_PAIR_UNMERGED(p))
1437 p->status = DIFF_STATUS_UNMERGED;
1438 else if (!DIFF_FILE_VALID(p->one))
1439 p->status = DIFF_STATUS_ADDED;
1440 else if (!DIFF_FILE_VALID(p->two))
1441 p->status = DIFF_STATUS_DELETED;
1442 else if (DIFF_PAIR_TYPE_CHANGED(p))
1443 p->status = DIFF_STATUS_TYPE_CHANGED;
1445 /* from this point on, we are dealing with a pair
1446 * whose both sides are valid and of the same type, i.e.
1447 * either in-place edit or rename/copy edit.
1449 else if (DIFF_PAIR_RENAME(p)) {
1450 if (p->source_stays) {
1451 p->status = DIFF_STATUS_COPIED;
1454 /* See if there is some other filepair that
1455 * copies from the same source as us. If so
1456 * we are a copy. Otherwise we are either a
1457 * copy if the path stays, or a rename if it
1458 * does not, but we already handled "stays" case.
1460 for (j = i + 1; j < q->nr; j++) {
1462 if (strcmp(pp->one->path, p->one->path))
1463 continue; /* not us */
1464 if (!DIFF_PAIR_RENAME(pp))
1465 continue; /* not a rename/copy */
1466 /* pp is a rename/copy from the same source */
1467 p->status = DIFF_STATUS_COPIED;
1471 p->status = DIFF_STATUS_RENAMED;
1473 else if (memcmp(p->one->sha1, p->two->sha1, 20) ||
1474 p->one->mode != p->two->mode)
1475 p->status = DIFF_STATUS_MODIFIED;
1477 /* This is a "no-change" entry and should not
1478 * happen anymore, but prepare for broken callers.
1480 error("feeding unmodified %s to diffcore",
1482 p->status = DIFF_STATUS_UNKNOWN;
1485 diff_debug_queue("resolve-rename-copy done", q);
1488 static void flush_one_pair(struct diff_filepair *p,
1489 int diff_output_format,
1490 struct diff_options *options,
1491 struct diffstat_t *diffstat)
1493 int inter_name_termination = '\t';
1494 int line_termination = options->line_termination;
1495 if (!line_termination)
1496 inter_name_termination = 0;
1498 switch (p->status) {
1499 case DIFF_STATUS_UNKNOWN:
1502 die("internal error in diff-resolve-rename-copy");
1505 switch (diff_output_format) {
1506 case DIFF_FORMAT_DIFFSTAT:
1507 diff_flush_stat(p, options, diffstat);
1509 case DIFF_FORMAT_PATCH:
1510 diff_flush_patch(p, options);
1512 case DIFF_FORMAT_RAW:
1513 case DIFF_FORMAT_NAME_STATUS:
1514 diff_flush_raw(p, line_termination,
1515 inter_name_termination,
1516 options, diff_output_format);
1518 case DIFF_FORMAT_NAME:
1520 inter_name_termination,
1523 case DIFF_FORMAT_NO_OUTPUT:
1529 void diff_flush(struct diff_options *options)
1531 struct diff_queue_struct *q = &diff_queued_diff;
1533 int diff_output_format = options->output_format;
1534 struct diffstat_t *diffstat = NULL;
1536 if (diff_output_format == DIFF_FORMAT_DIFFSTAT || options->with_stat) {
1537 diffstat = xcalloc(sizeof (struct diffstat_t), 1);
1538 diffstat->xm.consume = diffstat_consume;
1541 if (options->with_raw) {
1542 for (i = 0; i < q->nr; i++) {
1543 struct diff_filepair *p = q->queue[i];
1544 flush_one_pair(p, DIFF_FORMAT_RAW, options, NULL);
1546 putchar(options->line_termination);
1548 if (options->with_stat) {
1549 for (i = 0; i < q->nr; i++) {
1550 struct diff_filepair *p = q->queue[i];
1551 flush_one_pair(p, DIFF_FORMAT_DIFFSTAT, options,
1554 show_stats(diffstat);
1557 putchar(options->line_termination);
1559 for (i = 0; i < q->nr; i++) {
1560 struct diff_filepair *p = q->queue[i];
1561 flush_one_pair(p, diff_output_format, options, diffstat);
1562 diff_free_filepair(p);
1566 show_stats(diffstat);
1572 q->nr = q->alloc = 0;
1575 static void diffcore_apply_filter(const char *filter)
1578 struct diff_queue_struct *q = &diff_queued_diff;
1579 struct diff_queue_struct outq;
1581 outq.nr = outq.alloc = 0;
1586 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
1588 for (i = found = 0; !found && i < q->nr; i++) {
1589 struct diff_filepair *p = q->queue[i];
1590 if (((p->status == DIFF_STATUS_MODIFIED) &&
1592 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
1594 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
1595 ((p->status != DIFF_STATUS_MODIFIED) &&
1596 strchr(filter, p->status)))
1602 /* otherwise we will clear the whole queue
1603 * by copying the empty outq at the end of this
1604 * function, but first clear the current entries
1607 for (i = 0; i < q->nr; i++)
1608 diff_free_filepair(q->queue[i]);
1611 /* Only the matching ones */
1612 for (i = 0; i < q->nr; i++) {
1613 struct diff_filepair *p = q->queue[i];
1615 if (((p->status == DIFF_STATUS_MODIFIED) &&
1617 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
1619 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
1620 ((p->status != DIFF_STATUS_MODIFIED) &&
1621 strchr(filter, p->status)))
1624 diff_free_filepair(p);
1631 void diffcore_std(struct diff_options *options)
1633 if (options->break_opt != -1)
1634 diffcore_break(options->break_opt);
1635 if (options->detect_rename)
1636 diffcore_rename(options);
1637 if (options->break_opt != -1)
1638 diffcore_merge_broken();
1639 if (options->pickaxe)
1640 diffcore_pickaxe(options->pickaxe, options->pickaxe_opts);
1641 if (options->orderfile)
1642 diffcore_order(options->orderfile);
1643 diff_resolve_rename_copy();
1644 diffcore_apply_filter(options->filter);
1648 void diffcore_std_no_resolve(struct diff_options *options)
1650 if (options->pickaxe)
1651 diffcore_pickaxe(options->pickaxe, options->pickaxe_opts);
1652 if (options->orderfile)
1653 diffcore_order(options->orderfile);
1654 diffcore_apply_filter(options->filter);
1657 void diff_addremove(struct diff_options *options,
1658 int addremove, unsigned mode,
1659 const unsigned char *sha1,
1660 const char *base, const char *path)
1662 char concatpath[PATH_MAX];
1663 struct diff_filespec *one, *two;
1665 /* This may look odd, but it is a preparation for
1666 * feeding "there are unchanged files which should
1667 * not produce diffs, but when you are doing copy
1668 * detection you would need them, so here they are"
1669 * entries to the diff-core. They will be prefixed
1670 * with something like '=' or '*' (I haven't decided
1671 * which but should not make any difference).
1672 * Feeding the same new and old to diff_change()
1673 * also has the same effect.
1674 * Before the final output happens, they are pruned after
1675 * merged into rename/copy pairs as appropriate.
1677 if (options->reverse_diff)
1678 addremove = (addremove == '+' ? '-' :
1679 addremove == '-' ? '+' : addremove);
1681 if (!path) path = "";
1682 sprintf(concatpath, "%s%s", base, path);
1683 one = alloc_filespec(concatpath);
1684 two = alloc_filespec(concatpath);
1686 if (addremove != '+')
1687 fill_filespec(one, sha1, mode);
1688 if (addremove != '-')
1689 fill_filespec(two, sha1, mode);
1691 diff_queue(&diff_queued_diff, one, two);
1694 void diff_change(struct diff_options *options,
1695 unsigned old_mode, unsigned new_mode,
1696 const unsigned char *old_sha1,
1697 const unsigned char *new_sha1,
1698 const char *base, const char *path)
1700 char concatpath[PATH_MAX];
1701 struct diff_filespec *one, *two;
1703 if (options->reverse_diff) {
1705 const unsigned char *tmp_c;
1706 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
1707 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
1709 if (!path) path = "";
1710 sprintf(concatpath, "%s%s", base, path);
1711 one = alloc_filespec(concatpath);
1712 two = alloc_filespec(concatpath);
1713 fill_filespec(one, old_sha1, old_mode);
1714 fill_filespec(two, new_sha1, new_mode);
1716 diff_queue(&diff_queued_diff, one, two);
1719 void diff_unmerge(struct diff_options *options,
1722 struct diff_filespec *one, *two;
1723 one = alloc_filespec(path);
1724 two = alloc_filespec(path);
1725 diff_queue(&diff_queued_diff, one, two);