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 len = strlen(file->name);
257 if (file->is_binary || file->is_unmerged)
259 if (max_change < file->added + file->deleted)
260 max_change = file->added + file->deleted;
263 for (i = 0; i < data->nr; i++) {
264 char *name = data->files[i]->name;
265 int added = data->files[i]->added;
266 int deleted = data->files[i]->deleted;
268 if (0 < (len = quote_c_style(name, NULL, NULL, 0))) {
269 char *qname = xmalloc(len + 1);
270 quote_c_style(name, qname, NULL, 0);
272 data->files[i]->name = name = qname;
276 * "scale" the filename
287 slash = strchr(name, '/');
294 * scale the add/delete
300 if (data->files[i]->is_binary) {
301 printf(" %s%-*s | Bin\n", prefix, len, name);
302 goto free_diffstat_file;
304 else if (data->files[i]->is_unmerged) {
305 printf(" %s%-*s | Unmerged\n", prefix, len, name);
306 goto free_diffstat_file;
308 else if (added + deleted == 0) {
310 goto free_diffstat_file;
319 if (max_change > 0) {
320 total = (total * max + max_change / 2) / max_change;
321 add = (add * max + max_change / 2) / max_change;
324 printf(" %s%-*s |%5d %.*s%.*s\n", prefix,
325 len, name, added + deleted,
326 add, pluses, del, minuses);
328 free(data->files[i]->name);
329 free(data->files[i]);
332 printf(" %d files changed, %d insertions(+), %d deletions(-)\n",
333 total_files, adds, dels);
336 #define FIRST_FEW_BYTES 8000
337 static int mmfile_is_binary(mmfile_t *mf)
340 if (FIRST_FEW_BYTES < sz)
341 sz = FIRST_FEW_BYTES;
342 if (memchr(mf->ptr, 0, sz))
347 static void builtin_diff(const char *name_a,
349 struct diff_filespec *one,
350 struct diff_filespec *two,
351 const char *xfrm_msg,
352 int complete_rewrite)
358 a_one = quote_two("a/", name_a);
359 b_two = quote_two("b/", name_b);
360 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
361 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
362 printf("diff --git %s %s\n", a_one, b_two);
363 if (lbl[0][0] == '/') {
365 printf("new file mode %06o\n", two->mode);
366 if (xfrm_msg && xfrm_msg[0])
369 else if (lbl[1][0] == '/') {
370 printf("deleted file mode %06o\n", one->mode);
371 if (xfrm_msg && xfrm_msg[0])
375 if (one->mode != two->mode) {
376 printf("old mode %06o\n", one->mode);
377 printf("new mode %06o\n", two->mode);
379 if (xfrm_msg && xfrm_msg[0])
382 * we do not run diff between different kind
385 if ((one->mode ^ two->mode) & S_IFMT)
386 goto free_ab_and_return;
387 if (complete_rewrite) {
388 emit_rewrite_diff(name_a, name_b, one, two);
389 goto free_ab_and_return;
393 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
394 die("unable to read files to diff");
396 if (mmfile_is_binary(&mf1) || mmfile_is_binary(&mf2))
397 printf("Binary files %s and %s differ\n", lbl[0], lbl[1]);
399 /* Crazy xdl interfaces.. */
400 const char *diffopts = getenv("GIT_DIFF_OPTS");
404 struct emit_callback ecbdata;
406 ecbdata.label_path = lbl;
407 xpp.flags = XDF_NEED_MINIMAL;
409 xecfg.flags = XDL_EMIT_FUNCNAMES;
412 else if (!strncmp(diffopts, "--unified=", 10))
413 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
414 else if (!strncmp(diffopts, "-u", 2))
415 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
418 xdl_diff(&mf1, &mf2, &xpp, &xecfg, &ecb);
427 static void builtin_diffstat(const char *name_a, const char *name_b,
428 struct diff_filespec *one, struct diff_filespec *two,
429 struct diffstat_t *diffstat)
432 struct diffstat_file *data;
434 data = diffstat_add(diffstat, name_a ? name_a : name_b);
437 data->is_unmerged = 1;
441 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
442 die("unable to read files to diff");
444 if (mmfile_is_binary(&mf1) || mmfile_is_binary(&mf2))
447 /* Crazy xdl interfaces.. */
452 xpp.flags = XDF_NEED_MINIMAL;
455 ecb.outf = xdiff_outf;
457 xdl_diff(&mf1, &mf2, &xpp, &xecfg, &ecb);
461 struct diff_filespec *alloc_filespec(const char *path)
463 int namelen = strlen(path);
464 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
466 memset(spec, 0, sizeof(*spec));
467 spec->path = (char *)(spec + 1);
468 memcpy(spec->path, path, namelen+1);
472 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
476 spec->mode = canon_mode(mode);
477 memcpy(spec->sha1, sha1, 20);
478 spec->sha1_valid = !!memcmp(sha1, null_sha1, 20);
483 * Given a name and sha1 pair, if the dircache tells us the file in
484 * the work tree has that object contents, return true, so that
485 * prepare_temp_file() does not have to inflate and extract.
487 static int work_tree_matches(const char *name, const unsigned char *sha1)
489 struct cache_entry *ce;
493 /* We do not read the cache ourselves here, because the
494 * benchmark with my previous version that always reads cache
495 * shows that it makes things worse for diff-tree comparing
496 * two linux-2.6 kernel trees in an already checked out work
497 * tree. This is because most diff-tree comparisons deal with
498 * only a small number of files, while reading the cache is
499 * expensive for a large project, and its cost outweighs the
500 * savings we get by not inflating the object to a temporary
501 * file. Practically, this code only helps when we are used
502 * by diff-cache --cached, which does read the cache before
509 pos = cache_name_pos(name, len);
512 ce = active_cache[pos];
513 if ((lstat(name, &st) < 0) ||
514 !S_ISREG(st.st_mode) || /* careful! */
515 ce_match_stat(ce, &st, 0) ||
516 memcmp(sha1, ce->sha1, 20))
518 /* we return 1 only when we can stat, it is a regular file,
519 * stat information matches, and sha1 recorded in the cache
520 * matches. I.e. we know the file in the work tree really is
521 * the same as the <name, sha1> pair.
526 static struct sha1_size_cache {
527 unsigned char sha1[20];
530 static int sha1_size_cache_nr, sha1_size_cache_alloc;
532 static struct sha1_size_cache *locate_size_cache(unsigned char *sha1,
537 struct sha1_size_cache *e;
540 last = sha1_size_cache_nr;
541 while (last > first) {
542 int cmp, next = (last + first) >> 1;
543 e = sha1_size_cache[next];
544 cmp = memcmp(e->sha1, sha1, 20);
556 /* insert to make it at "first" */
557 if (sha1_size_cache_alloc <= sha1_size_cache_nr) {
558 sha1_size_cache_alloc = alloc_nr(sha1_size_cache_alloc);
559 sha1_size_cache = xrealloc(sha1_size_cache,
560 sha1_size_cache_alloc *
561 sizeof(*sha1_size_cache));
563 sha1_size_cache_nr++;
564 if (first < sha1_size_cache_nr)
565 memmove(sha1_size_cache + first + 1, sha1_size_cache + first,
566 (sha1_size_cache_nr - first - 1) *
567 sizeof(*sha1_size_cache));
568 e = xmalloc(sizeof(struct sha1_size_cache));
569 sha1_size_cache[first] = e;
570 memcpy(e->sha1, sha1, 20);
576 * While doing rename detection and pickaxe operation, we may need to
577 * grab the data for the blob (or file) for our own in-core comparison.
578 * diff_filespec has data and size fields for this purpose.
580 int diff_populate_filespec(struct diff_filespec *s, int size_only)
583 if (!DIFF_FILE_VALID(s))
584 die("internal error: asking to populate invalid file.");
585 if (S_ISDIR(s->mode))
593 if (!s->sha1_valid ||
594 work_tree_matches(s->path, s->sha1)) {
597 if (lstat(s->path, &st) < 0) {
598 if (errno == ENOENT) {
607 s->size = st.st_size;
612 if (S_ISLNK(st.st_mode)) {
614 s->data = xmalloc(s->size);
616 ret = readlink(s->path, s->data, s->size);
623 fd = open(s->path, O_RDONLY);
626 s->data = mmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
628 if (s->data == MAP_FAILED)
630 s->should_munmap = 1;
634 struct sha1_size_cache *e;
637 e = locate_size_cache(s->sha1, 1, 0);
642 if (!sha1_object_info(s->sha1, type, &s->size))
643 locate_size_cache(s->sha1, 0, s->size);
646 s->data = read_sha1_file(s->sha1, type, &s->size);
653 void diff_free_filespec_data(struct diff_filespec *s)
657 else if (s->should_munmap)
658 munmap(s->data, s->size);
659 s->should_free = s->should_munmap = 0;
665 static void prep_temp_blob(struct diff_tempfile *temp,
668 const unsigned char *sha1,
673 fd = git_mkstemp(temp->tmp_path, TEMPFILE_PATH_LEN, ".diff_XXXXXX");
675 die("unable to create temp-file");
676 if (write(fd, blob, size) != size)
677 die("unable to write temp-file");
679 temp->name = temp->tmp_path;
680 strcpy(temp->hex, sha1_to_hex(sha1));
682 sprintf(temp->mode, "%06o", mode);
685 static void prepare_temp_file(const char *name,
686 struct diff_tempfile *temp,
687 struct diff_filespec *one)
689 if (!DIFF_FILE_VALID(one)) {
691 /* A '-' entry produces this for file-2, and
692 * a '+' entry produces this for file-1.
694 temp->name = "/dev/null";
695 strcpy(temp->hex, ".");
696 strcpy(temp->mode, ".");
700 if (!one->sha1_valid ||
701 work_tree_matches(name, one->sha1)) {
703 if (lstat(name, &st) < 0) {
705 goto not_a_valid_file;
706 die("stat(%s): %s", name, strerror(errno));
708 if (S_ISLNK(st.st_mode)) {
710 char buf[PATH_MAX + 1]; /* ought to be SYMLINK_MAX */
711 if (sizeof(buf) <= st.st_size)
712 die("symlink too long: %s", name);
713 ret = readlink(name, buf, st.st_size);
715 die("readlink(%s)", name);
716 prep_temp_blob(temp, buf, st.st_size,
718 one->sha1 : null_sha1),
720 one->mode : S_IFLNK));
723 /* we can borrow from the file in the work tree */
725 if (!one->sha1_valid)
726 strcpy(temp->hex, sha1_to_hex(null_sha1));
728 strcpy(temp->hex, sha1_to_hex(one->sha1));
729 /* Even though we may sometimes borrow the
730 * contents from the work tree, we always want
731 * one->mode. mode is trustworthy even when
732 * !(one->sha1_valid), as long as
733 * DIFF_FILE_VALID(one).
735 sprintf(temp->mode, "%06o", one->mode);
740 if (diff_populate_filespec(one, 0))
741 die("cannot read data blob for %s", one->path);
742 prep_temp_blob(temp, one->data, one->size,
743 one->sha1, one->mode);
747 static void remove_tempfile(void)
751 for (i = 0; i < 2; i++)
752 if (diff_temp[i].name == diff_temp[i].tmp_path) {
753 unlink(diff_temp[i].name);
754 diff_temp[i].name = NULL;
758 static void remove_tempfile_on_signal(int signo)
761 signal(SIGINT, SIG_DFL);
765 static int spawn_prog(const char *pgm, const char **arg)
773 die("unable to fork");
775 execvp(pgm, (char *const*) arg);
779 while (waitpid(pid, &status, 0) < 0) {
785 /* Earlier we did not check the exit status because
786 * diff exits non-zero if files are different, and
787 * we are not interested in knowing that. It was a
788 * mistake which made it harder to quit a diff-*
789 * session that uses the git-apply-patch-script as
790 * the GIT_EXTERNAL_DIFF. A custom GIT_EXTERNAL_DIFF
791 * should also exit non-zero only when it wants to
792 * abort the entire diff-* session.
794 if (WIFEXITED(status) && !WEXITSTATUS(status))
799 /* An external diff command takes:
801 * diff-cmd name infile1 infile1-sha1 infile1-mode \
802 * infile2 infile2-sha1 infile2-mode [ rename-to ]
805 static void run_external_diff(const char *pgm,
808 struct diff_filespec *one,
809 struct diff_filespec *two,
810 const char *xfrm_msg,
811 int complete_rewrite)
813 const char *spawn_arg[10];
814 struct diff_tempfile *temp = diff_temp;
816 static int atexit_asked = 0;
817 const char *othername;
818 const char **arg = &spawn_arg[0];
820 othername = (other? other : name);
822 prepare_temp_file(name, &temp[0], one);
823 prepare_temp_file(othername, &temp[1], two);
824 if (! atexit_asked &&
825 (temp[0].name == temp[0].tmp_path ||
826 temp[1].name == temp[1].tmp_path)) {
828 atexit(remove_tempfile);
830 signal(SIGINT, remove_tempfile_on_signal);
836 *arg++ = temp[0].name;
837 *arg++ = temp[0].hex;
838 *arg++ = temp[0].mode;
839 *arg++ = temp[1].name;
840 *arg++ = temp[1].hex;
841 *arg++ = temp[1].mode;
851 retval = spawn_prog(pgm, spawn_arg);
854 fprintf(stderr, "external diff died, stopping at %s.\n", name);
859 static void run_diff_cmd(const char *pgm,
862 struct diff_filespec *one,
863 struct diff_filespec *two,
864 const char *xfrm_msg,
865 int complete_rewrite)
868 run_external_diff(pgm, name, other, one, two, xfrm_msg,
873 builtin_diff(name, other ? other : name,
874 one, two, xfrm_msg, complete_rewrite);
876 printf("* Unmerged path %s\n", name);
879 static void diff_fill_sha1_info(struct diff_filespec *one)
881 if (DIFF_FILE_VALID(one)) {
882 if (!one->sha1_valid) {
884 if (lstat(one->path, &st) < 0)
885 die("stat %s", one->path);
886 if (index_path(one->sha1, one->path, &st, 0))
887 die("cannot hash %s\n", one->path);
891 memset(one->sha1, 0, 20);
894 static void run_diff(struct diff_filepair *p, struct diff_options *o)
896 const char *pgm = external_diff();
897 char msg[PATH_MAX*2+300], *xfrm_msg;
898 struct diff_filespec *one;
899 struct diff_filespec *two;
902 char *name_munged, *other_munged;
903 int complete_rewrite = 0;
906 if (DIFF_PAIR_UNMERGED(p)) {
908 run_diff_cmd(pgm, p->one->path, NULL, NULL, NULL, NULL, 0);
913 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
914 name_munged = quote_one(name);
915 other_munged = quote_one(other);
916 one = p->one; two = p->two;
918 diff_fill_sha1_info(one);
919 diff_fill_sha1_info(two);
923 case DIFF_STATUS_COPIED:
924 len += snprintf(msg + len, sizeof(msg) - len,
925 "similarity index %d%%\n"
928 (int)(0.5 + p->score * 100.0/MAX_SCORE),
929 name_munged, other_munged);
931 case DIFF_STATUS_RENAMED:
932 len += snprintf(msg + len, sizeof(msg) - len,
933 "similarity index %d%%\n"
936 (int)(0.5 + p->score * 100.0/MAX_SCORE),
937 name_munged, other_munged);
939 case DIFF_STATUS_MODIFIED:
941 len += snprintf(msg + len, sizeof(msg) - len,
942 "dissimilarity index %d%%\n",
943 (int)(0.5 + p->score *
945 complete_rewrite = 1;
954 if (memcmp(one->sha1, two->sha1, 20)) {
956 int abbrev = o->full_index ? 40 : DEFAULT_ABBREV;
957 memcpy(one_sha1, sha1_to_hex(one->sha1), 41);
959 len += snprintf(msg + len, sizeof(msg) - len,
961 abbrev, one_sha1, abbrev,
962 sha1_to_hex(two->sha1));
963 if (one->mode == two->mode)
964 len += snprintf(msg + len, sizeof(msg) - len,
966 len += snprintf(msg + len, sizeof(msg) - len, "\n");
971 xfrm_msg = len ? msg : NULL;
974 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
975 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
976 /* a filepair that changes between file and symlink
977 * needs to be split into deletion and creation.
979 struct diff_filespec *null = alloc_filespec(two->path);
980 run_diff_cmd(NULL, name, other, one, null, xfrm_msg, 0);
982 null = alloc_filespec(one->path);
983 run_diff_cmd(NULL, name, other, null, two, xfrm_msg, 0);
987 run_diff_cmd(pgm, name, other, one, two, xfrm_msg,
994 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
995 struct diffstat_t *diffstat)
1000 if (DIFF_PAIR_UNMERGED(p)) {
1002 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat);
1006 name = p->one->path;
1007 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
1009 diff_fill_sha1_info(p->one);
1010 diff_fill_sha1_info(p->two);
1012 builtin_diffstat(name, other, p->one, p->two, diffstat);
1015 void diff_setup(struct diff_options *options)
1017 memset(options, 0, sizeof(*options));
1018 options->output_format = DIFF_FORMAT_RAW;
1019 options->line_termination = '\n';
1020 options->break_opt = -1;
1021 options->rename_limit = -1;
1023 options->change = diff_change;
1024 options->add_remove = diff_addremove;
1027 int diff_setup_done(struct diff_options *options)
1029 if ((options->find_copies_harder &&
1030 options->detect_rename != DIFF_DETECT_COPY) ||
1031 (0 <= options->rename_limit && !options->detect_rename))
1035 * These cases always need recursive; we do not drop caller-supplied
1036 * recursive bits for other formats here.
1038 if ((options->output_format == DIFF_FORMAT_PATCH) ||
1039 (options->output_format == DIFF_FORMAT_DIFFSTAT) ||
1040 (options->with_stat))
1041 options->recursive = 1;
1043 if (options->detect_rename && options->rename_limit < 0)
1044 options->rename_limit = diff_rename_limit_default;
1045 if (options->setup & DIFF_SETUP_USE_CACHE) {
1047 /* read-cache does not die even when it fails
1048 * so it is safe for us to do this here. Also
1049 * it does not smudge active_cache or active_nr
1050 * when it fails, so we do not have to worry about
1051 * cleaning it up ourselves either.
1055 if (options->setup & DIFF_SETUP_USE_SIZE_CACHE)
1057 if (options->abbrev <= 0 || 40 < options->abbrev)
1058 options->abbrev = 40; /* full */
1063 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
1065 const char *arg = av[0];
1066 if (!strcmp(arg, "-p") || !strcmp(arg, "-u"))
1067 options->output_format = DIFF_FORMAT_PATCH;
1068 else if (!strcmp(arg, "--patch-with-raw")) {
1069 options->output_format = DIFF_FORMAT_PATCH;
1070 options->with_raw = 1;
1072 else if (!strcmp(arg, "--stat"))
1073 options->output_format = DIFF_FORMAT_DIFFSTAT;
1074 else if (!strcmp(arg, "--patch-with-stat")) {
1075 options->output_format = DIFF_FORMAT_PATCH;
1076 options->with_stat = 1;
1078 else if (!strcmp(arg, "-z"))
1079 options->line_termination = 0;
1080 else if (!strncmp(arg, "-l", 2))
1081 options->rename_limit = strtoul(arg+2, NULL, 10);
1082 else if (!strcmp(arg, "--full-index"))
1083 options->full_index = 1;
1084 else if (!strcmp(arg, "--name-only"))
1085 options->output_format = DIFF_FORMAT_NAME;
1086 else if (!strcmp(arg, "--name-status"))
1087 options->output_format = DIFF_FORMAT_NAME_STATUS;
1088 else if (!strcmp(arg, "-R"))
1089 options->reverse_diff = 1;
1090 else if (!strncmp(arg, "-S", 2))
1091 options->pickaxe = arg + 2;
1092 else if (!strcmp(arg, "-s"))
1093 options->output_format = DIFF_FORMAT_NO_OUTPUT;
1094 else if (!strncmp(arg, "-O", 2))
1095 options->orderfile = arg + 2;
1096 else if (!strncmp(arg, "--diff-filter=", 14))
1097 options->filter = arg + 14;
1098 else if (!strcmp(arg, "--pickaxe-all"))
1099 options->pickaxe_opts = DIFF_PICKAXE_ALL;
1100 else if (!strcmp(arg, "--pickaxe-regex"))
1101 options->pickaxe_opts = DIFF_PICKAXE_REGEX;
1102 else if (!strncmp(arg, "-B", 2)) {
1103 if ((options->break_opt =
1104 diff_scoreopt_parse(arg)) == -1)
1107 else if (!strncmp(arg, "-M", 2)) {
1108 if ((options->rename_score =
1109 diff_scoreopt_parse(arg)) == -1)
1111 options->detect_rename = DIFF_DETECT_RENAME;
1113 else if (!strncmp(arg, "-C", 2)) {
1114 if ((options->rename_score =
1115 diff_scoreopt_parse(arg)) == -1)
1117 options->detect_rename = DIFF_DETECT_COPY;
1119 else if (!strcmp(arg, "--find-copies-harder"))
1120 options->find_copies_harder = 1;
1121 else if (!strcmp(arg, "--abbrev"))
1122 options->abbrev = DEFAULT_ABBREV;
1123 else if (!strncmp(arg, "--abbrev=", 9)) {
1124 options->abbrev = strtoul(arg + 9, NULL, 10);
1125 if (options->abbrev < MINIMUM_ABBREV)
1126 options->abbrev = MINIMUM_ABBREV;
1127 else if (40 < options->abbrev)
1128 options->abbrev = 40;
1135 static int parse_num(const char **cp_p)
1137 unsigned long num, scale;
1139 const char *cp = *cp_p;
1146 if ( !dot && ch == '.' ) {
1149 } else if ( ch == '%' ) {
1150 scale = dot ? scale*100 : 100;
1151 cp++; /* % is always at the end */
1153 } else if ( ch >= '0' && ch <= '9' ) {
1154 if ( scale < 100000 ) {
1156 num = (num*10) + (ch-'0');
1165 /* user says num divided by scale and we say internally that
1166 * is MAX_SCORE * num / scale.
1168 return (num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale);
1171 int diff_scoreopt_parse(const char *opt)
1173 int opt1, opt2, cmd;
1178 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
1179 return -1; /* that is not a -M, -C nor -B option */
1181 opt1 = parse_num(&opt);
1187 else if (*opt != '/')
1188 return -1; /* we expect -B80/99 or -B80 */
1191 opt2 = parse_num(&opt);
1196 return opt1 | (opt2 << 16);
1199 struct diff_queue_struct diff_queued_diff;
1201 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
1203 if (queue->alloc <= queue->nr) {
1204 queue->alloc = alloc_nr(queue->alloc);
1205 queue->queue = xrealloc(queue->queue,
1206 sizeof(dp) * queue->alloc);
1208 queue->queue[queue->nr++] = dp;
1211 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
1212 struct diff_filespec *one,
1213 struct diff_filespec *two)
1215 struct diff_filepair *dp = xmalloc(sizeof(*dp));
1220 dp->source_stays = 0;
1221 dp->broken_pair = 0;
1227 void diff_free_filepair(struct diff_filepair *p)
1229 diff_free_filespec_data(p->one);
1230 diff_free_filespec_data(p->two);
1236 /* This is different from find_unique_abbrev() in that
1237 * it stuffs the result with dots for alignment.
1239 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
1244 return sha1_to_hex(sha1);
1246 abbrev = find_unique_abbrev(sha1, len);
1248 return sha1_to_hex(sha1);
1249 abblen = strlen(abbrev);
1251 static char hex[41];
1252 if (len < abblen && abblen <= len + 2)
1253 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
1255 sprintf(hex, "%s...", abbrev);
1258 return sha1_to_hex(sha1);
1261 static void diff_flush_raw(struct diff_filepair *p,
1262 int line_termination,
1263 int inter_name_termination,
1264 struct diff_options *options,
1269 int abbrev = options->abbrev;
1270 const char *path_one, *path_two;
1272 path_one = p->one->path;
1273 path_two = p->two->path;
1274 if (line_termination) {
1275 path_one = quote_one(path_one);
1276 path_two = quote_one(path_two);
1280 sprintf(status, "%c%03d", p->status,
1281 (int)(0.5 + p->score * 100.0/MAX_SCORE));
1283 status[0] = p->status;
1286 switch (p->status) {
1287 case DIFF_STATUS_COPIED:
1288 case DIFF_STATUS_RENAMED:
1291 case DIFF_STATUS_ADDED:
1292 case DIFF_STATUS_DELETED:
1299 if (output_format != DIFF_FORMAT_NAME_STATUS) {
1300 printf(":%06o %06o %s ",
1301 p->one->mode, p->two->mode,
1302 diff_unique_abbrev(p->one->sha1, abbrev));
1304 diff_unique_abbrev(p->two->sha1, abbrev));
1306 printf("%s%c%s", status, inter_name_termination, path_one);
1308 printf("%c%s", inter_name_termination, path_two);
1309 putchar(line_termination);
1310 if (path_one != p->one->path)
1311 free((void*)path_one);
1312 if (path_two != p->two->path)
1313 free((void*)path_two);
1316 static void diff_flush_name(struct diff_filepair *p,
1317 int inter_name_termination,
1318 int line_termination)
1320 char *path = p->two->path;
1322 if (line_termination)
1323 path = quote_one(p->two->path);
1325 path = p->two->path;
1326 printf("%s%c", path, line_termination);
1327 if (p->two->path != path)
1331 int diff_unmodified_pair(struct diff_filepair *p)
1333 /* This function is written stricter than necessary to support
1334 * the currently implemented transformers, but the idea is to
1335 * let transformers to produce diff_filepairs any way they want,
1336 * and filter and clean them up here before producing the output.
1338 struct diff_filespec *one, *two;
1340 if (DIFF_PAIR_UNMERGED(p))
1341 return 0; /* unmerged is interesting */
1346 /* deletion, addition, mode or type change
1347 * and rename are all interesting.
1349 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
1350 DIFF_PAIR_MODE_CHANGED(p) ||
1351 strcmp(one->path, two->path))
1354 /* both are valid and point at the same path. that is, we are
1355 * dealing with a change.
1357 if (one->sha1_valid && two->sha1_valid &&
1358 !memcmp(one->sha1, two->sha1, sizeof(one->sha1)))
1359 return 1; /* no change */
1360 if (!one->sha1_valid && !two->sha1_valid)
1361 return 1; /* both look at the same file on the filesystem. */
1365 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
1367 if (diff_unmodified_pair(p))
1370 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
1371 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
1372 return; /* no tree diffs in patch format */
1377 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
1378 struct diffstat_t *diffstat)
1380 if (diff_unmodified_pair(p))
1383 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
1384 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
1385 return; /* no tree diffs in patch format */
1387 run_diffstat(p, o, diffstat);
1390 int diff_queue_is_empty(void)
1392 struct diff_queue_struct *q = &diff_queued_diff;
1394 for (i = 0; i < q->nr; i++)
1395 if (!diff_unmodified_pair(q->queue[i]))
1401 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
1403 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
1406 DIFF_FILE_VALID(s) ? "valid" : "invalid",
1408 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
1409 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
1411 s->size, s->xfrm_flags);
1414 void diff_debug_filepair(const struct diff_filepair *p, int i)
1416 diff_debug_filespec(p->one, i, "one");
1417 diff_debug_filespec(p->two, i, "two");
1418 fprintf(stderr, "score %d, status %c stays %d broken %d\n",
1419 p->score, p->status ? p->status : '?',
1420 p->source_stays, p->broken_pair);
1423 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
1427 fprintf(stderr, "%s\n", msg);
1428 fprintf(stderr, "q->nr = %d\n", q->nr);
1429 for (i = 0; i < q->nr; i++) {
1430 struct diff_filepair *p = q->queue[i];
1431 diff_debug_filepair(p, i);
1436 static void diff_resolve_rename_copy(void)
1439 struct diff_filepair *p, *pp;
1440 struct diff_queue_struct *q = &diff_queued_diff;
1442 diff_debug_queue("resolve-rename-copy", q);
1444 for (i = 0; i < q->nr; i++) {
1446 p->status = 0; /* undecided */
1447 if (DIFF_PAIR_UNMERGED(p))
1448 p->status = DIFF_STATUS_UNMERGED;
1449 else if (!DIFF_FILE_VALID(p->one))
1450 p->status = DIFF_STATUS_ADDED;
1451 else if (!DIFF_FILE_VALID(p->two))
1452 p->status = DIFF_STATUS_DELETED;
1453 else if (DIFF_PAIR_TYPE_CHANGED(p))
1454 p->status = DIFF_STATUS_TYPE_CHANGED;
1456 /* from this point on, we are dealing with a pair
1457 * whose both sides are valid and of the same type, i.e.
1458 * either in-place edit or rename/copy edit.
1460 else if (DIFF_PAIR_RENAME(p)) {
1461 if (p->source_stays) {
1462 p->status = DIFF_STATUS_COPIED;
1465 /* See if there is some other filepair that
1466 * copies from the same source as us. If so
1467 * we are a copy. Otherwise we are either a
1468 * copy if the path stays, or a rename if it
1469 * does not, but we already handled "stays" case.
1471 for (j = i + 1; j < q->nr; j++) {
1473 if (strcmp(pp->one->path, p->one->path))
1474 continue; /* not us */
1475 if (!DIFF_PAIR_RENAME(pp))
1476 continue; /* not a rename/copy */
1477 /* pp is a rename/copy from the same source */
1478 p->status = DIFF_STATUS_COPIED;
1482 p->status = DIFF_STATUS_RENAMED;
1484 else if (memcmp(p->one->sha1, p->two->sha1, 20) ||
1485 p->one->mode != p->two->mode)
1486 p->status = DIFF_STATUS_MODIFIED;
1488 /* This is a "no-change" entry and should not
1489 * happen anymore, but prepare for broken callers.
1491 error("feeding unmodified %s to diffcore",
1493 p->status = DIFF_STATUS_UNKNOWN;
1496 diff_debug_queue("resolve-rename-copy done", q);
1499 static void flush_one_pair(struct diff_filepair *p,
1500 int diff_output_format,
1501 struct diff_options *options,
1502 struct diffstat_t *diffstat)
1504 int inter_name_termination = '\t';
1505 int line_termination = options->line_termination;
1506 if (!line_termination)
1507 inter_name_termination = 0;
1509 switch (p->status) {
1510 case DIFF_STATUS_UNKNOWN:
1513 die("internal error in diff-resolve-rename-copy");
1516 switch (diff_output_format) {
1517 case DIFF_FORMAT_DIFFSTAT:
1518 diff_flush_stat(p, options, diffstat);
1520 case DIFF_FORMAT_PATCH:
1521 diff_flush_patch(p, options);
1523 case DIFF_FORMAT_RAW:
1524 case DIFF_FORMAT_NAME_STATUS:
1525 diff_flush_raw(p, line_termination,
1526 inter_name_termination,
1527 options, diff_output_format);
1529 case DIFF_FORMAT_NAME:
1531 inter_name_termination,
1534 case DIFF_FORMAT_NO_OUTPUT:
1540 void diff_flush(struct diff_options *options)
1542 struct diff_queue_struct *q = &diff_queued_diff;
1544 int diff_output_format = options->output_format;
1545 struct diffstat_t *diffstat = NULL;
1547 if (diff_output_format == DIFF_FORMAT_DIFFSTAT || options->with_stat) {
1548 diffstat = xcalloc(sizeof (struct diffstat_t), 1);
1549 diffstat->xm.consume = diffstat_consume;
1552 if (options->with_raw) {
1553 for (i = 0; i < q->nr; i++) {
1554 struct diff_filepair *p = q->queue[i];
1555 flush_one_pair(p, DIFF_FORMAT_RAW, options, NULL);
1557 putchar(options->line_termination);
1559 if (options->with_stat) {
1560 for (i = 0; i < q->nr; i++) {
1561 struct diff_filepair *p = q->queue[i];
1562 flush_one_pair(p, DIFF_FORMAT_DIFFSTAT, options,
1565 show_stats(diffstat);
1568 putchar(options->line_termination);
1570 for (i = 0; i < q->nr; i++) {
1571 struct diff_filepair *p = q->queue[i];
1572 flush_one_pair(p, diff_output_format, options, diffstat);
1573 diff_free_filepair(p);
1577 show_stats(diffstat);
1583 q->nr = q->alloc = 0;
1586 static void diffcore_apply_filter(const char *filter)
1589 struct diff_queue_struct *q = &diff_queued_diff;
1590 struct diff_queue_struct outq;
1592 outq.nr = outq.alloc = 0;
1597 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
1599 for (i = found = 0; !found && i < q->nr; i++) {
1600 struct diff_filepair *p = q->queue[i];
1601 if (((p->status == DIFF_STATUS_MODIFIED) &&
1603 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
1605 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
1606 ((p->status != DIFF_STATUS_MODIFIED) &&
1607 strchr(filter, p->status)))
1613 /* otherwise we will clear the whole queue
1614 * by copying the empty outq at the end of this
1615 * function, but first clear the current entries
1618 for (i = 0; i < q->nr; i++)
1619 diff_free_filepair(q->queue[i]);
1622 /* Only the matching ones */
1623 for (i = 0; i < q->nr; i++) {
1624 struct diff_filepair *p = q->queue[i];
1626 if (((p->status == DIFF_STATUS_MODIFIED) &&
1628 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
1630 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
1631 ((p->status != DIFF_STATUS_MODIFIED) &&
1632 strchr(filter, p->status)))
1635 diff_free_filepair(p);
1642 void diffcore_std(struct diff_options *options)
1644 if (options->break_opt != -1)
1645 diffcore_break(options->break_opt);
1646 if (options->detect_rename)
1647 diffcore_rename(options);
1648 if (options->break_opt != -1)
1649 diffcore_merge_broken();
1650 if (options->pickaxe)
1651 diffcore_pickaxe(options->pickaxe, options->pickaxe_opts);
1652 if (options->orderfile)
1653 diffcore_order(options->orderfile);
1654 diff_resolve_rename_copy();
1655 diffcore_apply_filter(options->filter);
1659 void diffcore_std_no_resolve(struct diff_options *options)
1661 if (options->pickaxe)
1662 diffcore_pickaxe(options->pickaxe, options->pickaxe_opts);
1663 if (options->orderfile)
1664 diffcore_order(options->orderfile);
1665 diffcore_apply_filter(options->filter);
1668 void diff_addremove(struct diff_options *options,
1669 int addremove, unsigned mode,
1670 const unsigned char *sha1,
1671 const char *base, const char *path)
1673 char concatpath[PATH_MAX];
1674 struct diff_filespec *one, *two;
1676 /* This may look odd, but it is a preparation for
1677 * feeding "there are unchanged files which should
1678 * not produce diffs, but when you are doing copy
1679 * detection you would need them, so here they are"
1680 * entries to the diff-core. They will be prefixed
1681 * with something like '=' or '*' (I haven't decided
1682 * which but should not make any difference).
1683 * Feeding the same new and old to diff_change()
1684 * also has the same effect.
1685 * Before the final output happens, they are pruned after
1686 * merged into rename/copy pairs as appropriate.
1688 if (options->reverse_diff)
1689 addremove = (addremove == '+' ? '-' :
1690 addremove == '-' ? '+' : addremove);
1692 if (!path) path = "";
1693 sprintf(concatpath, "%s%s", base, path);
1694 one = alloc_filespec(concatpath);
1695 two = alloc_filespec(concatpath);
1697 if (addremove != '+')
1698 fill_filespec(one, sha1, mode);
1699 if (addremove != '-')
1700 fill_filespec(two, sha1, mode);
1702 diff_queue(&diff_queued_diff, one, two);
1705 void diff_change(struct diff_options *options,
1706 unsigned old_mode, unsigned new_mode,
1707 const unsigned char *old_sha1,
1708 const unsigned char *new_sha1,
1709 const char *base, const char *path)
1711 char concatpath[PATH_MAX];
1712 struct diff_filespec *one, *two;
1714 if (options->reverse_diff) {
1716 const unsigned char *tmp_c;
1717 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
1718 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
1720 if (!path) path = "";
1721 sprintf(concatpath, "%s%s", base, path);
1722 one = alloc_filespec(concatpath);
1723 two = alloc_filespec(concatpath);
1724 fill_filespec(one, old_sha1, old_mode);
1725 fill_filespec(two, new_sha1, new_mode);
1727 diff_queue(&diff_queued_diff, one, two);
1730 void diff_unmerge(struct diff_options *options,
1733 struct diff_filespec *one, *two;
1734 one = alloc_filespec(path);
1735 two = alloc_filespec(path);
1736 diff_queue(&diff_queued_diff, one, two);