git-blame.el: use mapc instead of mapcar
[git.git] / diff.c
blob9038f190ece061696a3290acf8f1e9951e9d0a6c
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 400;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static int diff_dirstat_permille_default = 30;
35 static struct diff_options default_diff_options;
37 static char diff_colors[][COLOR_MAXLEN] = {
38 GIT_COLOR_RESET,
39 GIT_COLOR_NORMAL, /* PLAIN */
40 GIT_COLOR_BOLD, /* METAINFO */
41 GIT_COLOR_CYAN, /* FRAGINFO */
42 GIT_COLOR_RED, /* OLD */
43 GIT_COLOR_GREEN, /* NEW */
44 GIT_COLOR_YELLOW, /* COMMIT */
45 GIT_COLOR_BG_RED, /* WHITESPACE */
46 GIT_COLOR_NORMAL, /* FUNCINFO */
49 static int parse_diff_color_slot(const char *var, int ofs)
51 if (!strcasecmp(var+ofs, "plain"))
52 return DIFF_PLAIN;
53 if (!strcasecmp(var+ofs, "meta"))
54 return DIFF_METAINFO;
55 if (!strcasecmp(var+ofs, "frag"))
56 return DIFF_FRAGINFO;
57 if (!strcasecmp(var+ofs, "old"))
58 return DIFF_FILE_OLD;
59 if (!strcasecmp(var+ofs, "new"))
60 return DIFF_FILE_NEW;
61 if (!strcasecmp(var+ofs, "commit"))
62 return DIFF_COMMIT;
63 if (!strcasecmp(var+ofs, "whitespace"))
64 return DIFF_WHITESPACE;
65 if (!strcasecmp(var+ofs, "func"))
66 return DIFF_FUNCINFO;
67 return -1;
70 static int parse_dirstat_params(struct diff_options *options, const char *params,
71 struct strbuf *errmsg)
73 const char *p = params;
74 int p_len, ret = 0;
76 while (*p) {
77 p_len = strchrnul(p, ',') - p;
78 if (!memcmp(p, "changes", p_len)) {
79 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
80 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
81 } else if (!memcmp(p, "lines", p_len)) {
82 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
83 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
84 } else if (!memcmp(p, "files", p_len)) {
85 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
86 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
87 } else if (!memcmp(p, "noncumulative", p_len)) {
88 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
89 } else if (!memcmp(p, "cumulative", p_len)) {
90 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
91 } else if (isdigit(*p)) {
92 char *end;
93 int permille = strtoul(p, &end, 10) * 10;
94 if (*end == '.' && isdigit(*++end)) {
95 /* only use first digit */
96 permille += *end - '0';
97 /* .. and ignore any further digits */
98 while (isdigit(*++end))
99 ; /* nothing */
101 if (end - p == p_len)
102 options->dirstat_permille = permille;
103 else {
104 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%.*s'\n"),
105 p_len, p);
106 ret++;
108 } else {
109 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%.*s'\n"),
110 p_len, p);
111 ret++;
114 p += p_len;
116 if (*p)
117 p++; /* more parameters, swallow separator */
119 return ret;
122 static int git_config_rename(const char *var, const char *value)
124 if (!value)
125 return DIFF_DETECT_RENAME;
126 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
127 return DIFF_DETECT_COPY;
128 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
132 * These are to give UI layer defaults.
133 * The core-level commands such as git-diff-files should
134 * never be affected by the setting of diff.renames
135 * the user happens to have in the configuration file.
137 int git_diff_ui_config(const char *var, const char *value, void *cb)
139 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
140 diff_use_color_default = git_config_colorbool(var, value, -1);
141 return 0;
143 if (!strcmp(var, "diff.renames")) {
144 diff_detect_rename_default = git_config_rename(var, value);
145 return 0;
147 if (!strcmp(var, "diff.autorefreshindex")) {
148 diff_auto_refresh_index = git_config_bool(var, value);
149 return 0;
151 if (!strcmp(var, "diff.mnemonicprefix")) {
152 diff_mnemonic_prefix = git_config_bool(var, value);
153 return 0;
155 if (!strcmp(var, "diff.noprefix")) {
156 diff_no_prefix = git_config_bool(var, value);
157 return 0;
159 if (!strcmp(var, "diff.external"))
160 return git_config_string(&external_diff_cmd_cfg, var, value);
161 if (!strcmp(var, "diff.wordregex"))
162 return git_config_string(&diff_word_regex_cfg, var, value);
164 if (!strcmp(var, "diff.ignoresubmodules"))
165 handle_ignore_submodules_arg(&default_diff_options, value);
167 return git_diff_basic_config(var, value, cb);
170 int git_diff_basic_config(const char *var, const char *value, void *cb)
172 if (!strcmp(var, "diff.renamelimit")) {
173 diff_rename_limit_default = git_config_int(var, value);
174 return 0;
177 switch (userdiff_config(var, value)) {
178 case 0: break;
179 case -1: return -1;
180 default: return 0;
183 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
184 int slot = parse_diff_color_slot(var, 11);
185 if (slot < 0)
186 return 0;
187 if (!value)
188 return config_error_nonbool(var);
189 color_parse(value, var, diff_colors[slot]);
190 return 0;
193 /* like GNU diff's --suppress-blank-empty option */
194 if (!strcmp(var, "diff.suppressblankempty") ||
195 /* for backwards compatibility */
196 !strcmp(var, "diff.suppress-blank-empty")) {
197 diff_suppress_blank_empty = git_config_bool(var, value);
198 return 0;
201 if (!strcmp(var, "diff.dirstat")) {
202 struct strbuf errmsg = STRBUF_INIT;
203 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
204 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
205 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
206 errmsg.buf);
207 strbuf_release(&errmsg);
208 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
209 return 0;
212 if (!prefixcmp(var, "submodule."))
213 return parse_submodule_config_option(var, value);
215 return git_color_default_config(var, value, cb);
218 static char *quote_two(const char *one, const char *two)
220 int need_one = quote_c_style(one, NULL, NULL, 1);
221 int need_two = quote_c_style(two, NULL, NULL, 1);
222 struct strbuf res = STRBUF_INIT;
224 if (need_one + need_two) {
225 strbuf_addch(&res, '"');
226 quote_c_style(one, &res, NULL, 1);
227 quote_c_style(two, &res, NULL, 1);
228 strbuf_addch(&res, '"');
229 } else {
230 strbuf_addstr(&res, one);
231 strbuf_addstr(&res, two);
233 return strbuf_detach(&res, NULL);
236 static const char *external_diff(void)
238 static const char *external_diff_cmd = NULL;
239 static int done_preparing = 0;
241 if (done_preparing)
242 return external_diff_cmd;
243 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
244 if (!external_diff_cmd)
245 external_diff_cmd = external_diff_cmd_cfg;
246 done_preparing = 1;
247 return external_diff_cmd;
250 static struct diff_tempfile {
251 const char *name; /* filename external diff should read from */
252 char hex[41];
253 char mode[10];
254 char tmp_path[PATH_MAX];
255 } diff_temp[2];
257 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
259 struct emit_callback {
260 int color_diff;
261 unsigned ws_rule;
262 int blank_at_eof_in_preimage;
263 int blank_at_eof_in_postimage;
264 int lno_in_preimage;
265 int lno_in_postimage;
266 sane_truncate_fn truncate;
267 const char **label_path;
268 struct diff_words_data *diff_words;
269 struct diff_options *opt;
270 int *found_changesp;
271 struct strbuf *header;
274 static int count_lines(const char *data, int size)
276 int count, ch, completely_empty = 1, nl_just_seen = 0;
277 count = 0;
278 while (0 < size--) {
279 ch = *data++;
280 if (ch == '\n') {
281 count++;
282 nl_just_seen = 1;
283 completely_empty = 0;
285 else {
286 nl_just_seen = 0;
287 completely_empty = 0;
290 if (completely_empty)
291 return 0;
292 if (!nl_just_seen)
293 count++; /* no trailing newline */
294 return count;
297 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
299 if (!DIFF_FILE_VALID(one)) {
300 mf->ptr = (char *)""; /* does not matter */
301 mf->size = 0;
302 return 0;
304 else if (diff_populate_filespec(one, 0))
305 return -1;
307 mf->ptr = one->data;
308 mf->size = one->size;
309 return 0;
312 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
313 static unsigned long diff_filespec_size(struct diff_filespec *one)
315 if (!DIFF_FILE_VALID(one))
316 return 0;
317 diff_populate_filespec(one, 1);
318 return one->size;
321 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
323 char *ptr = mf->ptr;
324 long size = mf->size;
325 int cnt = 0;
327 if (!size)
328 return cnt;
329 ptr += size - 1; /* pointing at the very end */
330 if (*ptr != '\n')
331 ; /* incomplete line */
332 else
333 ptr--; /* skip the last LF */
334 while (mf->ptr < ptr) {
335 char *prev_eol;
336 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
337 if (*prev_eol == '\n')
338 break;
339 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
340 break;
341 cnt++;
342 ptr = prev_eol - 1;
344 return cnt;
347 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
348 struct emit_callback *ecbdata)
350 int l1, l2, at;
351 unsigned ws_rule = ecbdata->ws_rule;
352 l1 = count_trailing_blank(mf1, ws_rule);
353 l2 = count_trailing_blank(mf2, ws_rule);
354 if (l2 <= l1) {
355 ecbdata->blank_at_eof_in_preimage = 0;
356 ecbdata->blank_at_eof_in_postimage = 0;
357 return;
359 at = count_lines(mf1->ptr, mf1->size);
360 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
362 at = count_lines(mf2->ptr, mf2->size);
363 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
366 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
367 int first, const char *line, int len)
369 int has_trailing_newline, has_trailing_carriage_return;
370 int nofirst;
371 FILE *file = o->file;
373 if (o->output_prefix) {
374 struct strbuf *msg = NULL;
375 msg = o->output_prefix(o, o->output_prefix_data);
376 assert(msg);
377 fwrite(msg->buf, msg->len, 1, file);
380 if (len == 0) {
381 has_trailing_newline = (first == '\n');
382 has_trailing_carriage_return = (!has_trailing_newline &&
383 (first == '\r'));
384 nofirst = has_trailing_newline || has_trailing_carriage_return;
385 } else {
386 has_trailing_newline = (len > 0 && line[len-1] == '\n');
387 if (has_trailing_newline)
388 len--;
389 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
390 if (has_trailing_carriage_return)
391 len--;
392 nofirst = 0;
395 if (len || !nofirst) {
396 fputs(set, file);
397 if (!nofirst)
398 fputc(first, file);
399 fwrite(line, len, 1, file);
400 fputs(reset, file);
402 if (has_trailing_carriage_return)
403 fputc('\r', file);
404 if (has_trailing_newline)
405 fputc('\n', file);
408 static void emit_line(struct diff_options *o, const char *set, const char *reset,
409 const char *line, int len)
411 emit_line_0(o, set, reset, line[0], line+1, len-1);
414 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
416 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
417 ecbdata->blank_at_eof_in_preimage &&
418 ecbdata->blank_at_eof_in_postimage &&
419 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
420 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
421 return 0;
422 return ws_blank_line(line, len, ecbdata->ws_rule);
425 static void emit_add_line(const char *reset,
426 struct emit_callback *ecbdata,
427 const char *line, int len)
429 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
430 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
432 if (!*ws)
433 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
434 else if (new_blank_line_at_eof(ecbdata, line, len))
435 /* Blank line at EOF - paint '+' as well */
436 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
437 else {
438 /* Emit just the prefix, then the rest. */
439 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
440 ws_check_emit(line, len, ecbdata->ws_rule,
441 ecbdata->opt->file, set, reset, ws);
445 static void emit_hunk_header(struct emit_callback *ecbdata,
446 const char *line, int len)
448 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
449 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
450 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
451 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
452 static const char atat[2] = { '@', '@' };
453 const char *cp, *ep;
454 struct strbuf msgbuf = STRBUF_INIT;
455 int org_len = len;
456 int i = 1;
459 * As a hunk header must begin with "@@ -<old>, +<new> @@",
460 * it always is at least 10 bytes long.
462 if (len < 10 ||
463 memcmp(line, atat, 2) ||
464 !(ep = memmem(line + 2, len - 2, atat, 2))) {
465 emit_line(ecbdata->opt, plain, reset, line, len);
466 return;
468 ep += 2; /* skip over @@ */
470 /* The hunk header in fraginfo color */
471 strbuf_add(&msgbuf, frag, strlen(frag));
472 strbuf_add(&msgbuf, line, ep - line);
473 strbuf_add(&msgbuf, reset, strlen(reset));
476 * trailing "\r\n"
478 for ( ; i < 3; i++)
479 if (line[len - i] == '\r' || line[len - i] == '\n')
480 len--;
482 /* blank before the func header */
483 for (cp = ep; ep - line < len; ep++)
484 if (*ep != ' ' && *ep != '\t')
485 break;
486 if (ep != cp) {
487 strbuf_add(&msgbuf, plain, strlen(plain));
488 strbuf_add(&msgbuf, cp, ep - cp);
489 strbuf_add(&msgbuf, reset, strlen(reset));
492 if (ep < line + len) {
493 strbuf_add(&msgbuf, func, strlen(func));
494 strbuf_add(&msgbuf, ep, line + len - ep);
495 strbuf_add(&msgbuf, reset, strlen(reset));
498 strbuf_add(&msgbuf, line + len, org_len - len);
499 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
500 strbuf_release(&msgbuf);
503 static struct diff_tempfile *claim_diff_tempfile(void) {
504 int i;
505 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
506 if (!diff_temp[i].name)
507 return diff_temp + i;
508 die("BUG: diff is failing to clean up its tempfiles");
511 static int remove_tempfile_installed;
513 static void remove_tempfile(void)
515 int i;
516 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
517 if (diff_temp[i].name == diff_temp[i].tmp_path)
518 unlink_or_warn(diff_temp[i].name);
519 diff_temp[i].name = NULL;
523 static void remove_tempfile_on_signal(int signo)
525 remove_tempfile();
526 sigchain_pop(signo);
527 raise(signo);
530 static void print_line_count(FILE *file, int count)
532 switch (count) {
533 case 0:
534 fprintf(file, "0,0");
535 break;
536 case 1:
537 fprintf(file, "1");
538 break;
539 default:
540 fprintf(file, "1,%d", count);
541 break;
545 static void emit_rewrite_lines(struct emit_callback *ecb,
546 int prefix, const char *data, int size)
548 const char *endp = NULL;
549 static const char *nneof = " No newline at end of file\n";
550 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
551 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
553 while (0 < size) {
554 int len;
556 endp = memchr(data, '\n', size);
557 len = endp ? (endp - data + 1) : size;
558 if (prefix != '+') {
559 ecb->lno_in_preimage++;
560 emit_line_0(ecb->opt, old, reset, '-',
561 data, len);
562 } else {
563 ecb->lno_in_postimage++;
564 emit_add_line(reset, ecb, data, len);
566 size -= len;
567 data += len;
569 if (!endp) {
570 const char *plain = diff_get_color(ecb->color_diff,
571 DIFF_PLAIN);
572 emit_line_0(ecb->opt, plain, reset, '\\',
573 nneof, strlen(nneof));
577 static void emit_rewrite_diff(const char *name_a,
578 const char *name_b,
579 struct diff_filespec *one,
580 struct diff_filespec *two,
581 struct userdiff_driver *textconv_one,
582 struct userdiff_driver *textconv_two,
583 struct diff_options *o)
585 int lc_a, lc_b;
586 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
587 const char *name_a_tab, *name_b_tab;
588 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
589 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
590 const char *reset = diff_get_color(color_diff, DIFF_RESET);
591 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
592 const char *a_prefix, *b_prefix;
593 char *data_one, *data_two;
594 size_t size_one, size_two;
595 struct emit_callback ecbdata;
596 char *line_prefix = "";
597 struct strbuf *msgbuf;
599 if (o && o->output_prefix) {
600 msgbuf = o->output_prefix(o, o->output_prefix_data);
601 line_prefix = msgbuf->buf;
604 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
605 a_prefix = o->b_prefix;
606 b_prefix = o->a_prefix;
607 } else {
608 a_prefix = o->a_prefix;
609 b_prefix = o->b_prefix;
612 name_a += (*name_a == '/');
613 name_b += (*name_b == '/');
614 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
615 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
617 strbuf_reset(&a_name);
618 strbuf_reset(&b_name);
619 quote_two_c_style(&a_name, a_prefix, name_a, 0);
620 quote_two_c_style(&b_name, b_prefix, name_b, 0);
622 size_one = fill_textconv(textconv_one, one, &data_one);
623 size_two = fill_textconv(textconv_two, two, &data_two);
625 memset(&ecbdata, 0, sizeof(ecbdata));
626 ecbdata.color_diff = color_diff;
627 ecbdata.found_changesp = &o->found_changes;
628 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
629 ecbdata.opt = o;
630 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
631 mmfile_t mf1, mf2;
632 mf1.ptr = (char *)data_one;
633 mf2.ptr = (char *)data_two;
634 mf1.size = size_one;
635 mf2.size = size_two;
636 check_blank_at_eof(&mf1, &mf2, &ecbdata);
638 ecbdata.lno_in_preimage = 1;
639 ecbdata.lno_in_postimage = 1;
641 lc_a = count_lines(data_one, size_one);
642 lc_b = count_lines(data_two, size_two);
643 fprintf(o->file,
644 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
645 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
646 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
647 line_prefix, fraginfo);
648 if (!o->irreversible_delete)
649 print_line_count(o->file, lc_a);
650 else
651 fprintf(o->file, "?,?");
652 fprintf(o->file, " +");
653 print_line_count(o->file, lc_b);
654 fprintf(o->file, " @@%s\n", reset);
655 if (lc_a && !o->irreversible_delete)
656 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
657 if (lc_b)
658 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
659 if (textconv_one)
660 free((char *)data_one);
661 if (textconv_two)
662 free((char *)data_two);
665 struct diff_words_buffer {
666 mmfile_t text;
667 long alloc;
668 struct diff_words_orig {
669 const char *begin, *end;
670 } *orig;
671 int orig_nr, orig_alloc;
674 static void diff_words_append(char *line, unsigned long len,
675 struct diff_words_buffer *buffer)
677 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
678 line++;
679 len--;
680 memcpy(buffer->text.ptr + buffer->text.size, line, len);
681 buffer->text.size += len;
682 buffer->text.ptr[buffer->text.size] = '\0';
685 struct diff_words_style_elem {
686 const char *prefix;
687 const char *suffix;
688 const char *color; /* NULL; filled in by the setup code if
689 * color is enabled */
692 struct diff_words_style {
693 enum diff_words_type type;
694 struct diff_words_style_elem new, old, ctx;
695 const char *newline;
698 static struct diff_words_style diff_words_styles[] = {
699 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
700 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
701 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
704 struct diff_words_data {
705 struct diff_words_buffer minus, plus;
706 const char *current_plus;
707 int last_minus;
708 struct diff_options *opt;
709 regex_t *word_regex;
710 enum diff_words_type type;
711 struct diff_words_style *style;
714 static int fn_out_diff_words_write_helper(FILE *fp,
715 struct diff_words_style_elem *st_el,
716 const char *newline,
717 size_t count, const char *buf,
718 const char *line_prefix)
720 int print = 0;
722 while (count) {
723 char *p = memchr(buf, '\n', count);
724 if (print)
725 fputs(line_prefix, fp);
726 if (p != buf) {
727 if (st_el->color && fputs(st_el->color, fp) < 0)
728 return -1;
729 if (fputs(st_el->prefix, fp) < 0 ||
730 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
731 fputs(st_el->suffix, fp) < 0)
732 return -1;
733 if (st_el->color && *st_el->color
734 && fputs(GIT_COLOR_RESET, fp) < 0)
735 return -1;
737 if (!p)
738 return 0;
739 if (fputs(newline, fp) < 0)
740 return -1;
741 count -= p + 1 - buf;
742 buf = p + 1;
743 print = 1;
745 return 0;
749 * '--color-words' algorithm can be described as:
751 * 1. collect a the minus/plus lines of a diff hunk, divided into
752 * minus-lines and plus-lines;
754 * 2. break both minus-lines and plus-lines into words and
755 * place them into two mmfile_t with one word for each line;
757 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
759 * And for the common parts of the both file, we output the plus side text.
760 * diff_words->current_plus is used to trace the current position of the plus file
761 * which printed. diff_words->last_minus is used to trace the last minus word
762 * printed.
764 * For '--graph' to work with '--color-words', we need to output the graph prefix
765 * on each line of color words output. Generally, there are two conditions on
766 * which we should output the prefix.
768 * 1. diff_words->last_minus == 0 &&
769 * diff_words->current_plus == diff_words->plus.text.ptr
771 * that is: the plus text must start as a new line, and if there is no minus
772 * word printed, a graph prefix must be printed.
774 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
775 * *(diff_words->current_plus - 1) == '\n'
777 * that is: a graph prefix must be printed following a '\n'
779 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
781 if ((diff_words->last_minus == 0 &&
782 diff_words->current_plus == diff_words->plus.text.ptr) ||
783 (diff_words->current_plus > diff_words->plus.text.ptr &&
784 *(diff_words->current_plus - 1) == '\n')) {
785 return 1;
786 } else {
787 return 0;
791 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
793 struct diff_words_data *diff_words = priv;
794 struct diff_words_style *style = diff_words->style;
795 int minus_first, minus_len, plus_first, plus_len;
796 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
797 struct diff_options *opt = diff_words->opt;
798 struct strbuf *msgbuf;
799 char *line_prefix = "";
801 if (line[0] != '@' || parse_hunk_header(line, len,
802 &minus_first, &minus_len, &plus_first, &plus_len))
803 return;
805 assert(opt);
806 if (opt->output_prefix) {
807 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
808 line_prefix = msgbuf->buf;
811 /* POSIX requires that first be decremented by one if len == 0... */
812 if (minus_len) {
813 minus_begin = diff_words->minus.orig[minus_first].begin;
814 minus_end =
815 diff_words->minus.orig[minus_first + minus_len - 1].end;
816 } else
817 minus_begin = minus_end =
818 diff_words->minus.orig[minus_first].end;
820 if (plus_len) {
821 plus_begin = diff_words->plus.orig[plus_first].begin;
822 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
823 } else
824 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
826 if (color_words_output_graph_prefix(diff_words)) {
827 fputs(line_prefix, diff_words->opt->file);
829 if (diff_words->current_plus != plus_begin) {
830 fn_out_diff_words_write_helper(diff_words->opt->file,
831 &style->ctx, style->newline,
832 plus_begin - diff_words->current_plus,
833 diff_words->current_plus, line_prefix);
834 if (*(plus_begin - 1) == '\n')
835 fputs(line_prefix, diff_words->opt->file);
837 if (minus_begin != minus_end) {
838 fn_out_diff_words_write_helper(diff_words->opt->file,
839 &style->old, style->newline,
840 minus_end - minus_begin, minus_begin,
841 line_prefix);
843 if (plus_begin != plus_end) {
844 fn_out_diff_words_write_helper(diff_words->opt->file,
845 &style->new, style->newline,
846 plus_end - plus_begin, plus_begin,
847 line_prefix);
850 diff_words->current_plus = plus_end;
851 diff_words->last_minus = minus_first;
854 /* This function starts looking at *begin, and returns 0 iff a word was found. */
855 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
856 int *begin, int *end)
858 if (word_regex && *begin < buffer->size) {
859 regmatch_t match[1];
860 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
861 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
862 '\n', match[0].rm_eo - match[0].rm_so);
863 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
864 *begin += match[0].rm_so;
865 return *begin >= *end;
867 return -1;
870 /* find the next word */
871 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
872 (*begin)++;
873 if (*begin >= buffer->size)
874 return -1;
876 /* find the end of the word */
877 *end = *begin + 1;
878 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
879 (*end)++;
881 return 0;
885 * This function splits the words in buffer->text, stores the list with
886 * newline separator into out, and saves the offsets of the original words
887 * in buffer->orig.
889 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
890 regex_t *word_regex)
892 int i, j;
893 long alloc = 0;
895 out->size = 0;
896 out->ptr = NULL;
898 /* fake an empty "0th" word */
899 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
900 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
901 buffer->orig_nr = 1;
903 for (i = 0; i < buffer->text.size; i++) {
904 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
905 return;
907 /* store original boundaries */
908 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
909 buffer->orig_alloc);
910 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
911 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
912 buffer->orig_nr++;
914 /* store one word */
915 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
916 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
917 out->ptr[out->size + j - i] = '\n';
918 out->size += j - i + 1;
920 i = j - 1;
924 /* this executes the word diff on the accumulated buffers */
925 static void diff_words_show(struct diff_words_data *diff_words)
927 xpparam_t xpp;
928 xdemitconf_t xecfg;
929 mmfile_t minus, plus;
930 struct diff_words_style *style = diff_words->style;
932 struct diff_options *opt = diff_words->opt;
933 struct strbuf *msgbuf;
934 char *line_prefix = "";
936 assert(opt);
937 if (opt->output_prefix) {
938 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
939 line_prefix = msgbuf->buf;
942 /* special case: only removal */
943 if (!diff_words->plus.text.size) {
944 fputs(line_prefix, diff_words->opt->file);
945 fn_out_diff_words_write_helper(diff_words->opt->file,
946 &style->old, style->newline,
947 diff_words->minus.text.size,
948 diff_words->minus.text.ptr, line_prefix);
949 diff_words->minus.text.size = 0;
950 return;
953 diff_words->current_plus = diff_words->plus.text.ptr;
954 diff_words->last_minus = 0;
956 memset(&xpp, 0, sizeof(xpp));
957 memset(&xecfg, 0, sizeof(xecfg));
958 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
959 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
960 xpp.flags = 0;
961 /* as only the hunk header will be parsed, we need a 0-context */
962 xecfg.ctxlen = 0;
963 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
964 &xpp, &xecfg);
965 free(minus.ptr);
966 free(plus.ptr);
967 if (diff_words->current_plus != diff_words->plus.text.ptr +
968 diff_words->plus.text.size) {
969 if (color_words_output_graph_prefix(diff_words))
970 fputs(line_prefix, diff_words->opt->file);
971 fn_out_diff_words_write_helper(diff_words->opt->file,
972 &style->ctx, style->newline,
973 diff_words->plus.text.ptr + diff_words->plus.text.size
974 - diff_words->current_plus, diff_words->current_plus,
975 line_prefix);
977 diff_words->minus.text.size = diff_words->plus.text.size = 0;
980 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
981 static void diff_words_flush(struct emit_callback *ecbdata)
983 if (ecbdata->diff_words->minus.text.size ||
984 ecbdata->diff_words->plus.text.size)
985 diff_words_show(ecbdata->diff_words);
988 static void free_diff_words_data(struct emit_callback *ecbdata)
990 if (ecbdata->diff_words) {
991 diff_words_flush(ecbdata);
992 free (ecbdata->diff_words->minus.text.ptr);
993 free (ecbdata->diff_words->minus.orig);
994 free (ecbdata->diff_words->plus.text.ptr);
995 free (ecbdata->diff_words->plus.orig);
996 if (ecbdata->diff_words->word_regex) {
997 regfree(ecbdata->diff_words->word_regex);
998 free(ecbdata->diff_words->word_regex);
1000 free(ecbdata->diff_words);
1001 ecbdata->diff_words = NULL;
1005 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1007 if (diff_use_color)
1008 return diff_colors[ix];
1009 return "";
1012 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1014 const char *cp;
1015 unsigned long allot;
1016 size_t l = len;
1018 if (ecb->truncate)
1019 return ecb->truncate(line, len);
1020 cp = line;
1021 allot = l;
1022 while (0 < l) {
1023 (void) utf8_width(&cp, &l);
1024 if (!cp)
1025 break; /* truncated in the middle? */
1027 return allot - l;
1030 static void find_lno(const char *line, struct emit_callback *ecbdata)
1032 const char *p;
1033 ecbdata->lno_in_preimage = 0;
1034 ecbdata->lno_in_postimage = 0;
1035 p = strchr(line, '-');
1036 if (!p)
1037 return; /* cannot happen */
1038 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1039 p = strchr(p, '+');
1040 if (!p)
1041 return; /* cannot happen */
1042 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1045 static void fn_out_consume(void *priv, char *line, unsigned long len)
1047 struct emit_callback *ecbdata = priv;
1048 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1049 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1050 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1051 struct diff_options *o = ecbdata->opt;
1052 char *line_prefix = "";
1053 struct strbuf *msgbuf;
1055 if (o && o->output_prefix) {
1056 msgbuf = o->output_prefix(o, o->output_prefix_data);
1057 line_prefix = msgbuf->buf;
1060 if (ecbdata->header) {
1061 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1062 strbuf_reset(ecbdata->header);
1063 ecbdata->header = NULL;
1065 *(ecbdata->found_changesp) = 1;
1067 if (ecbdata->label_path[0]) {
1068 const char *name_a_tab, *name_b_tab;
1070 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1071 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1073 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1074 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1075 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1076 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1077 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1080 if (diff_suppress_blank_empty
1081 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1082 line[0] = '\n';
1083 len = 1;
1086 if (line[0] == '@') {
1087 if (ecbdata->diff_words)
1088 diff_words_flush(ecbdata);
1089 len = sane_truncate_line(ecbdata, line, len);
1090 find_lno(line, ecbdata);
1091 emit_hunk_header(ecbdata, line, len);
1092 if (line[len-1] != '\n')
1093 putc('\n', ecbdata->opt->file);
1094 return;
1097 if (len < 1) {
1098 emit_line(ecbdata->opt, reset, reset, line, len);
1099 if (ecbdata->diff_words
1100 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1101 fputs("~\n", ecbdata->opt->file);
1102 return;
1105 if (ecbdata->diff_words) {
1106 if (line[0] == '-') {
1107 diff_words_append(line, len,
1108 &ecbdata->diff_words->minus);
1109 return;
1110 } else if (line[0] == '+') {
1111 diff_words_append(line, len,
1112 &ecbdata->diff_words->plus);
1113 return;
1115 diff_words_flush(ecbdata);
1116 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1117 emit_line(ecbdata->opt, plain, reset, line, len);
1118 fputs("~\n", ecbdata->opt->file);
1119 } else {
1121 * Skip the prefix character, if any. With
1122 * diff_suppress_blank_empty, there may be
1123 * none.
1125 if (line[0] != '\n') {
1126 line++;
1127 len--;
1129 emit_line(ecbdata->opt, plain, reset, line, len);
1131 return;
1134 if (line[0] != '+') {
1135 const char *color =
1136 diff_get_color(ecbdata->color_diff,
1137 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1138 ecbdata->lno_in_preimage++;
1139 if (line[0] == ' ')
1140 ecbdata->lno_in_postimage++;
1141 emit_line(ecbdata->opt, color, reset, line, len);
1142 } else {
1143 ecbdata->lno_in_postimage++;
1144 emit_add_line(reset, ecbdata, line + 1, len - 1);
1148 static char *pprint_rename(const char *a, const char *b)
1150 const char *old = a;
1151 const char *new = b;
1152 struct strbuf name = STRBUF_INIT;
1153 int pfx_length, sfx_length;
1154 int len_a = strlen(a);
1155 int len_b = strlen(b);
1156 int a_midlen, b_midlen;
1157 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1158 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1160 if (qlen_a || qlen_b) {
1161 quote_c_style(a, &name, NULL, 0);
1162 strbuf_addstr(&name, " => ");
1163 quote_c_style(b, &name, NULL, 0);
1164 return strbuf_detach(&name, NULL);
1167 /* Find common prefix */
1168 pfx_length = 0;
1169 while (*old && *new && *old == *new) {
1170 if (*old == '/')
1171 pfx_length = old - a + 1;
1172 old++;
1173 new++;
1176 /* Find common suffix */
1177 old = a + len_a;
1178 new = b + len_b;
1179 sfx_length = 0;
1180 while (a <= old && b <= new && *old == *new) {
1181 if (*old == '/')
1182 sfx_length = len_a - (old - a);
1183 old--;
1184 new--;
1188 * pfx{mid-a => mid-b}sfx
1189 * {pfx-a => pfx-b}sfx
1190 * pfx{sfx-a => sfx-b}
1191 * name-a => name-b
1193 a_midlen = len_a - pfx_length - sfx_length;
1194 b_midlen = len_b - pfx_length - sfx_length;
1195 if (a_midlen < 0)
1196 a_midlen = 0;
1197 if (b_midlen < 0)
1198 b_midlen = 0;
1200 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1201 if (pfx_length + sfx_length) {
1202 strbuf_add(&name, a, pfx_length);
1203 strbuf_addch(&name, '{');
1205 strbuf_add(&name, a + pfx_length, a_midlen);
1206 strbuf_addstr(&name, " => ");
1207 strbuf_add(&name, b + pfx_length, b_midlen);
1208 if (pfx_length + sfx_length) {
1209 strbuf_addch(&name, '}');
1210 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1212 return strbuf_detach(&name, NULL);
1215 struct diffstat_t {
1216 int nr;
1217 int alloc;
1218 struct diffstat_file {
1219 char *from_name;
1220 char *name;
1221 char *print_name;
1222 unsigned is_unmerged:1;
1223 unsigned is_binary:1;
1224 unsigned is_renamed:1;
1225 uintmax_t added, deleted;
1226 } **files;
1229 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1230 const char *name_a,
1231 const char *name_b)
1233 struct diffstat_file *x;
1234 x = xcalloc(sizeof (*x), 1);
1235 if (diffstat->nr == diffstat->alloc) {
1236 diffstat->alloc = alloc_nr(diffstat->alloc);
1237 diffstat->files = xrealloc(diffstat->files,
1238 diffstat->alloc * sizeof(x));
1240 diffstat->files[diffstat->nr++] = x;
1241 if (name_b) {
1242 x->from_name = xstrdup(name_a);
1243 x->name = xstrdup(name_b);
1244 x->is_renamed = 1;
1246 else {
1247 x->from_name = NULL;
1248 x->name = xstrdup(name_a);
1250 return x;
1253 static void diffstat_consume(void *priv, char *line, unsigned long len)
1255 struct diffstat_t *diffstat = priv;
1256 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1258 if (line[0] == '+')
1259 x->added++;
1260 else if (line[0] == '-')
1261 x->deleted++;
1264 const char mime_boundary_leader[] = "------------";
1266 static int scale_linear(int it, int width, int max_change)
1269 * make sure that at least one '-' is printed if there were deletions,
1270 * and likewise for '+'.
1272 if (max_change < 2)
1273 return it;
1274 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1277 static void show_name(FILE *file,
1278 const char *prefix, const char *name, int len)
1280 fprintf(file, " %s%-*s |", prefix, len, name);
1283 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1285 if (cnt <= 0)
1286 return;
1287 fprintf(file, "%s", set);
1288 while (cnt--)
1289 putc(ch, file);
1290 fprintf(file, "%s", reset);
1293 static void fill_print_name(struct diffstat_file *file)
1295 char *pname;
1297 if (file->print_name)
1298 return;
1300 if (!file->is_renamed) {
1301 struct strbuf buf = STRBUF_INIT;
1302 if (quote_c_style(file->name, &buf, NULL, 0)) {
1303 pname = strbuf_detach(&buf, NULL);
1304 } else {
1305 pname = file->name;
1306 strbuf_release(&buf);
1308 } else {
1309 pname = pprint_rename(file->from_name, file->name);
1311 file->print_name = pname;
1314 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1316 int i, len, add, del, adds = 0, dels = 0;
1317 uintmax_t max_change = 0, max_len = 0;
1318 int total_files = data->nr;
1319 int width, name_width;
1320 const char *reset, *add_c, *del_c;
1321 const char *line_prefix = "";
1322 struct strbuf *msg = NULL;
1324 if (data->nr == 0)
1325 return;
1327 if (options->output_prefix) {
1328 msg = options->output_prefix(options, options->output_prefix_data);
1329 line_prefix = msg->buf;
1332 width = options->stat_width ? options->stat_width : 80;
1333 name_width = options->stat_name_width ? options->stat_name_width : 50;
1335 /* Sanity: give at least 5 columns to the graph,
1336 * but leave at least 10 columns for the name.
1338 if (width < 25)
1339 width = 25;
1340 if (name_width < 10)
1341 name_width = 10;
1342 else if (width < name_width + 15)
1343 name_width = width - 15;
1345 /* Find the longest filename and max number of changes */
1346 reset = diff_get_color_opt(options, DIFF_RESET);
1347 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1348 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1350 for (i = 0; i < data->nr; i++) {
1351 struct diffstat_file *file = data->files[i];
1352 uintmax_t change = file->added + file->deleted;
1353 fill_print_name(file);
1354 len = strlen(file->print_name);
1355 if (max_len < len)
1356 max_len = len;
1358 if (file->is_binary || file->is_unmerged)
1359 continue;
1360 if (max_change < change)
1361 max_change = change;
1364 /* Compute the width of the graph part;
1365 * 10 is for one blank at the beginning of the line plus
1366 * " | count " between the name and the graph.
1368 * From here on, name_width is the width of the name area,
1369 * and width is the width of the graph area.
1371 name_width = (name_width < max_len) ? name_width : max_len;
1372 if (width < (name_width + 10) + max_change)
1373 width = width - (name_width + 10);
1374 else
1375 width = max_change;
1377 for (i = 0; i < data->nr; i++) {
1378 const char *prefix = "";
1379 char *name = data->files[i]->print_name;
1380 uintmax_t added = data->files[i]->added;
1381 uintmax_t deleted = data->files[i]->deleted;
1382 int name_len;
1385 * "scale" the filename
1387 len = name_width;
1388 name_len = strlen(name);
1389 if (name_width < name_len) {
1390 char *slash;
1391 prefix = "...";
1392 len -= 3;
1393 name += name_len - len;
1394 slash = strchr(name, '/');
1395 if (slash)
1396 name = slash;
1399 if (data->files[i]->is_binary) {
1400 fprintf(options->file, "%s", line_prefix);
1401 show_name(options->file, prefix, name, len);
1402 fprintf(options->file, " Bin ");
1403 fprintf(options->file, "%s%"PRIuMAX"%s",
1404 del_c, deleted, reset);
1405 fprintf(options->file, " -> ");
1406 fprintf(options->file, "%s%"PRIuMAX"%s",
1407 add_c, added, reset);
1408 fprintf(options->file, " bytes");
1409 fprintf(options->file, "\n");
1410 continue;
1412 else if (data->files[i]->is_unmerged) {
1413 fprintf(options->file, "%s", line_prefix);
1414 show_name(options->file, prefix, name, len);
1415 fprintf(options->file, " Unmerged\n");
1416 continue;
1418 else if (!data->files[i]->is_renamed &&
1419 (added + deleted == 0)) {
1420 total_files--;
1421 continue;
1425 * scale the add/delete
1427 add = added;
1428 del = deleted;
1429 adds += add;
1430 dels += del;
1432 if (width <= max_change) {
1433 add = scale_linear(add, width, max_change);
1434 del = scale_linear(del, width, max_change);
1436 fprintf(options->file, "%s", line_prefix);
1437 show_name(options->file, prefix, name, len);
1438 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1439 added + deleted ? " " : "");
1440 show_graph(options->file, '+', add, add_c, reset);
1441 show_graph(options->file, '-', del, del_c, reset);
1442 fprintf(options->file, "\n");
1444 fprintf(options->file, "%s", line_prefix);
1445 fprintf(options->file,
1446 " %d files changed, %d insertions(+), %d deletions(-)\n",
1447 total_files, adds, dels);
1450 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1452 int i, adds = 0, dels = 0, total_files = data->nr;
1454 if (data->nr == 0)
1455 return;
1457 for (i = 0; i < data->nr; i++) {
1458 if (!data->files[i]->is_binary &&
1459 !data->files[i]->is_unmerged) {
1460 int added = data->files[i]->added;
1461 int deleted= data->files[i]->deleted;
1462 if (!data->files[i]->is_renamed &&
1463 (added + deleted == 0)) {
1464 total_files--;
1465 } else {
1466 adds += added;
1467 dels += deleted;
1471 if (options->output_prefix) {
1472 struct strbuf *msg = NULL;
1473 msg = options->output_prefix(options,
1474 options->output_prefix_data);
1475 fprintf(options->file, "%s", msg->buf);
1477 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1478 total_files, adds, dels);
1481 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1483 int i;
1485 if (data->nr == 0)
1486 return;
1488 for (i = 0; i < data->nr; i++) {
1489 struct diffstat_file *file = data->files[i];
1491 if (options->output_prefix) {
1492 struct strbuf *msg = NULL;
1493 msg = options->output_prefix(options,
1494 options->output_prefix_data);
1495 fprintf(options->file, "%s", msg->buf);
1498 if (file->is_binary)
1499 fprintf(options->file, "-\t-\t");
1500 else
1501 fprintf(options->file,
1502 "%"PRIuMAX"\t%"PRIuMAX"\t",
1503 file->added, file->deleted);
1504 if (options->line_termination) {
1505 fill_print_name(file);
1506 if (!file->is_renamed)
1507 write_name_quoted(file->name, options->file,
1508 options->line_termination);
1509 else {
1510 fputs(file->print_name, options->file);
1511 putc(options->line_termination, options->file);
1513 } else {
1514 if (file->is_renamed) {
1515 putc('\0', options->file);
1516 write_name_quoted(file->from_name, options->file, '\0');
1518 write_name_quoted(file->name, options->file, '\0');
1523 struct dirstat_file {
1524 const char *name;
1525 unsigned long changed;
1528 struct dirstat_dir {
1529 struct dirstat_file *files;
1530 int alloc, nr, permille, cumulative;
1533 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1534 unsigned long changed, const char *base, int baselen)
1536 unsigned long this_dir = 0;
1537 unsigned int sources = 0;
1538 const char *line_prefix = "";
1539 struct strbuf *msg = NULL;
1541 if (opt->output_prefix) {
1542 msg = opt->output_prefix(opt, opt->output_prefix_data);
1543 line_prefix = msg->buf;
1546 while (dir->nr) {
1547 struct dirstat_file *f = dir->files;
1548 int namelen = strlen(f->name);
1549 unsigned long this;
1550 char *slash;
1552 if (namelen < baselen)
1553 break;
1554 if (memcmp(f->name, base, baselen))
1555 break;
1556 slash = strchr(f->name + baselen, '/');
1557 if (slash) {
1558 int newbaselen = slash + 1 - f->name;
1559 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1560 sources++;
1561 } else {
1562 this = f->changed;
1563 dir->files++;
1564 dir->nr--;
1565 sources += 2;
1567 this_dir += this;
1571 * We don't report dirstat's for
1572 * - the top level
1573 * - or cases where everything came from a single directory
1574 * under this directory (sources == 1).
1576 if (baselen && sources != 1) {
1577 if (this_dir) {
1578 int permille = this_dir * 1000 / changed;
1579 if (permille >= dir->permille) {
1580 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1581 permille / 10, permille % 10, baselen, base);
1582 if (!dir->cumulative)
1583 return 0;
1587 return this_dir;
1590 static int dirstat_compare(const void *_a, const void *_b)
1592 const struct dirstat_file *a = _a;
1593 const struct dirstat_file *b = _b;
1594 return strcmp(a->name, b->name);
1597 static void show_dirstat(struct diff_options *options)
1599 int i;
1600 unsigned long changed;
1601 struct dirstat_dir dir;
1602 struct diff_queue_struct *q = &diff_queued_diff;
1604 dir.files = NULL;
1605 dir.alloc = 0;
1606 dir.nr = 0;
1607 dir.permille = options->dirstat_permille;
1608 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1610 changed = 0;
1611 for (i = 0; i < q->nr; i++) {
1612 struct diff_filepair *p = q->queue[i];
1613 const char *name;
1614 unsigned long copied, added, damage;
1615 int content_changed;
1617 name = p->two->path ? p->two->path : p->one->path;
1619 if (p->one->sha1_valid && p->two->sha1_valid)
1620 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1621 else
1622 content_changed = 1;
1624 if (!content_changed) {
1626 * The SHA1 has not changed, so pre-/post-content is
1627 * identical. We can therefore skip looking at the
1628 * file contents altogether.
1630 damage = 0;
1631 goto found_damage;
1634 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1636 * In --dirstat-by-file mode, we don't really need to
1637 * look at the actual file contents at all.
1638 * The fact that the SHA1 changed is enough for us to
1639 * add this file to the list of results
1640 * (with each file contributing equal damage).
1642 damage = 1;
1643 goto found_damage;
1646 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1647 diff_populate_filespec(p->one, 0);
1648 diff_populate_filespec(p->two, 0);
1649 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1650 &copied, &added);
1651 diff_free_filespec_data(p->one);
1652 diff_free_filespec_data(p->two);
1653 } else if (DIFF_FILE_VALID(p->one)) {
1654 diff_populate_filespec(p->one, 1);
1655 copied = added = 0;
1656 diff_free_filespec_data(p->one);
1657 } else if (DIFF_FILE_VALID(p->two)) {
1658 diff_populate_filespec(p->two, 1);
1659 copied = 0;
1660 added = p->two->size;
1661 diff_free_filespec_data(p->two);
1662 } else
1663 continue;
1666 * Original minus copied is the removed material,
1667 * added is the new material. They are both damages
1668 * made to the preimage.
1669 * If the resulting damage is zero, we know that
1670 * diffcore_count_changes() considers the two entries to
1671 * be identical, but since content_changed is true, we
1672 * know that there must have been _some_ kind of change,
1673 * so we force all entries to have damage > 0.
1675 damage = (p->one->size - copied) + added;
1676 if (!damage)
1677 damage = 1;
1679 found_damage:
1680 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1681 dir.files[dir.nr].name = name;
1682 dir.files[dir.nr].changed = damage;
1683 changed += damage;
1684 dir.nr++;
1687 /* This can happen even with many files, if everything was renames */
1688 if (!changed)
1689 return;
1691 /* Show all directories with more than x% of the changes */
1692 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1693 gather_dirstat(options, &dir, changed, "", 0);
1696 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1698 int i;
1699 unsigned long changed;
1700 struct dirstat_dir dir;
1702 if (data->nr == 0)
1703 return;
1705 dir.files = NULL;
1706 dir.alloc = 0;
1707 dir.nr = 0;
1708 dir.permille = options->dirstat_permille;
1709 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1711 changed = 0;
1712 for (i = 0; i < data->nr; i++) {
1713 struct diffstat_file *file = data->files[i];
1714 unsigned long damage = file->added + file->deleted;
1715 if (file->is_binary)
1717 * binary files counts bytes, not lines. Must find some
1718 * way to normalize binary bytes vs. textual lines.
1719 * The following heuristic assumes that there are 64
1720 * bytes per "line".
1721 * This is stupid and ugly, but very cheap...
1723 damage = (damage + 63) / 64;
1724 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1725 dir.files[dir.nr].name = file->name;
1726 dir.files[dir.nr].changed = damage;
1727 changed += damage;
1728 dir.nr++;
1731 /* This can happen even with many files, if everything was renames */
1732 if (!changed)
1733 return;
1735 /* Show all directories with more than x% of the changes */
1736 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1737 gather_dirstat(options, &dir, changed, "", 0);
1740 static void free_diffstat_info(struct diffstat_t *diffstat)
1742 int i;
1743 for (i = 0; i < diffstat->nr; i++) {
1744 struct diffstat_file *f = diffstat->files[i];
1745 if (f->name != f->print_name)
1746 free(f->print_name);
1747 free(f->name);
1748 free(f->from_name);
1749 free(f);
1751 free(diffstat->files);
1754 struct checkdiff_t {
1755 const char *filename;
1756 int lineno;
1757 int conflict_marker_size;
1758 struct diff_options *o;
1759 unsigned ws_rule;
1760 unsigned status;
1763 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1765 char firstchar;
1766 int cnt;
1768 if (len < marker_size + 1)
1769 return 0;
1770 firstchar = line[0];
1771 switch (firstchar) {
1772 case '=': case '>': case '<': case '|':
1773 break;
1774 default:
1775 return 0;
1777 for (cnt = 1; cnt < marker_size; cnt++)
1778 if (line[cnt] != firstchar)
1779 return 0;
1780 /* line[1] thru line[marker_size-1] are same as firstchar */
1781 if (len < marker_size + 1 || !isspace(line[marker_size]))
1782 return 0;
1783 return 1;
1786 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1788 struct checkdiff_t *data = priv;
1789 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1790 int marker_size = data->conflict_marker_size;
1791 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1792 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1793 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1794 char *err;
1795 char *line_prefix = "";
1796 struct strbuf *msgbuf;
1798 assert(data->o);
1799 if (data->o->output_prefix) {
1800 msgbuf = data->o->output_prefix(data->o,
1801 data->o->output_prefix_data);
1802 line_prefix = msgbuf->buf;
1805 if (line[0] == '+') {
1806 unsigned bad;
1807 data->lineno++;
1808 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1809 data->status |= 1;
1810 fprintf(data->o->file,
1811 "%s%s:%d: leftover conflict marker\n",
1812 line_prefix, data->filename, data->lineno);
1814 bad = ws_check(line + 1, len - 1, data->ws_rule);
1815 if (!bad)
1816 return;
1817 data->status |= bad;
1818 err = whitespace_error_string(bad);
1819 fprintf(data->o->file, "%s%s:%d: %s.\n",
1820 line_prefix, data->filename, data->lineno, err);
1821 free(err);
1822 emit_line(data->o, set, reset, line, 1);
1823 ws_check_emit(line + 1, len - 1, data->ws_rule,
1824 data->o->file, set, reset, ws);
1825 } else if (line[0] == ' ') {
1826 data->lineno++;
1827 } else if (line[0] == '@') {
1828 char *plus = strchr(line, '+');
1829 if (plus)
1830 data->lineno = strtol(plus, NULL, 10) - 1;
1831 else
1832 die("invalid diff");
1836 static unsigned char *deflate_it(char *data,
1837 unsigned long size,
1838 unsigned long *result_size)
1840 int bound;
1841 unsigned char *deflated;
1842 git_zstream stream;
1844 memset(&stream, 0, sizeof(stream));
1845 git_deflate_init(&stream, zlib_compression_level);
1846 bound = git_deflate_bound(&stream, size);
1847 deflated = xmalloc(bound);
1848 stream.next_out = deflated;
1849 stream.avail_out = bound;
1851 stream.next_in = (unsigned char *)data;
1852 stream.avail_in = size;
1853 while (git_deflate(&stream, Z_FINISH) == Z_OK)
1854 ; /* nothing */
1855 git_deflate_end(&stream);
1856 *result_size = stream.total_out;
1857 return deflated;
1860 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1862 void *cp;
1863 void *delta;
1864 void *deflated;
1865 void *data;
1866 unsigned long orig_size;
1867 unsigned long delta_size;
1868 unsigned long deflate_size;
1869 unsigned long data_size;
1871 /* We could do deflated delta, or we could do just deflated two,
1872 * whichever is smaller.
1874 delta = NULL;
1875 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1876 if (one->size && two->size) {
1877 delta = diff_delta(one->ptr, one->size,
1878 two->ptr, two->size,
1879 &delta_size, deflate_size);
1880 if (delta) {
1881 void *to_free = delta;
1882 orig_size = delta_size;
1883 delta = deflate_it(delta, delta_size, &delta_size);
1884 free(to_free);
1888 if (delta && delta_size < deflate_size) {
1889 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1890 free(deflated);
1891 data = delta;
1892 data_size = delta_size;
1894 else {
1895 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1896 free(delta);
1897 data = deflated;
1898 data_size = deflate_size;
1901 /* emit data encoded in base85 */
1902 cp = data;
1903 while (data_size) {
1904 int bytes = (52 < data_size) ? 52 : data_size;
1905 char line[70];
1906 data_size -= bytes;
1907 if (bytes <= 26)
1908 line[0] = bytes + 'A' - 1;
1909 else
1910 line[0] = bytes - 26 + 'a' - 1;
1911 encode_85(line + 1, cp, bytes);
1912 cp = (char *) cp + bytes;
1913 fprintf(file, "%s", prefix);
1914 fputs(line, file);
1915 fputc('\n', file);
1917 fprintf(file, "%s\n", prefix);
1918 free(data);
1921 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1923 fprintf(file, "%sGIT binary patch\n", prefix);
1924 emit_binary_diff_body(file, one, two, prefix);
1925 emit_binary_diff_body(file, two, one, prefix);
1928 static void diff_filespec_load_driver(struct diff_filespec *one)
1930 /* Use already-loaded driver */
1931 if (one->driver)
1932 return;
1934 if (S_ISREG(one->mode))
1935 one->driver = userdiff_find_by_path(one->path);
1937 /* Fallback to default settings */
1938 if (!one->driver)
1939 one->driver = userdiff_find_by_name("default");
1942 int diff_filespec_is_binary(struct diff_filespec *one)
1944 if (one->is_binary == -1) {
1945 diff_filespec_load_driver(one);
1946 if (one->driver->binary != -1)
1947 one->is_binary = one->driver->binary;
1948 else {
1949 if (!one->data && DIFF_FILE_VALID(one))
1950 diff_populate_filespec(one, 0);
1951 if (one->data)
1952 one->is_binary = buffer_is_binary(one->data,
1953 one->size);
1954 if (one->is_binary == -1)
1955 one->is_binary = 0;
1958 return one->is_binary;
1961 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1963 diff_filespec_load_driver(one);
1964 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1967 static const char *userdiff_word_regex(struct diff_filespec *one)
1969 diff_filespec_load_driver(one);
1970 return one->driver->word_regex;
1973 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1975 if (!options->a_prefix)
1976 options->a_prefix = a;
1977 if (!options->b_prefix)
1978 options->b_prefix = b;
1981 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1983 if (!DIFF_FILE_VALID(one))
1984 return NULL;
1986 diff_filespec_load_driver(one);
1987 return userdiff_get_textconv(one->driver);
1990 static void builtin_diff(const char *name_a,
1991 const char *name_b,
1992 struct diff_filespec *one,
1993 struct diff_filespec *two,
1994 const char *xfrm_msg,
1995 int must_show_header,
1996 struct diff_options *o,
1997 int complete_rewrite)
1999 mmfile_t mf1, mf2;
2000 const char *lbl[2];
2001 char *a_one, *b_two;
2002 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
2003 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2004 const char *a_prefix, *b_prefix;
2005 struct userdiff_driver *textconv_one = NULL;
2006 struct userdiff_driver *textconv_two = NULL;
2007 struct strbuf header = STRBUF_INIT;
2008 struct strbuf *msgbuf;
2009 char *line_prefix = "";
2011 if (o->output_prefix) {
2012 msgbuf = o->output_prefix(o, o->output_prefix_data);
2013 line_prefix = msgbuf->buf;
2016 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2017 (!one->mode || S_ISGITLINK(one->mode)) &&
2018 (!two->mode || S_ISGITLINK(two->mode))) {
2019 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2020 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2021 show_submodule_summary(o->file, one ? one->path : two->path,
2022 one->sha1, two->sha1, two->dirty_submodule,
2023 del, add, reset);
2024 return;
2027 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2028 textconv_one = get_textconv(one);
2029 textconv_two = get_textconv(two);
2032 diff_set_mnemonic_prefix(o, "a/", "b/");
2033 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2034 a_prefix = o->b_prefix;
2035 b_prefix = o->a_prefix;
2036 } else {
2037 a_prefix = o->a_prefix;
2038 b_prefix = o->b_prefix;
2041 /* Never use a non-valid filename anywhere if at all possible */
2042 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2043 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2045 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2046 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2047 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2048 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2049 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2050 if (lbl[0][0] == '/') {
2051 /* /dev/null */
2052 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2053 if (xfrm_msg)
2054 strbuf_addstr(&header, xfrm_msg);
2055 must_show_header = 1;
2057 else if (lbl[1][0] == '/') {
2058 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2059 if (xfrm_msg)
2060 strbuf_addstr(&header, xfrm_msg);
2061 must_show_header = 1;
2063 else {
2064 if (one->mode != two->mode) {
2065 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2066 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2067 must_show_header = 1;
2069 if (xfrm_msg)
2070 strbuf_addstr(&header, xfrm_msg);
2073 * we do not run diff between different kind
2074 * of objects.
2076 if ((one->mode ^ two->mode) & S_IFMT)
2077 goto free_ab_and_return;
2078 if (complete_rewrite &&
2079 (textconv_one || !diff_filespec_is_binary(one)) &&
2080 (textconv_two || !diff_filespec_is_binary(two))) {
2081 fprintf(o->file, "%s", header.buf);
2082 strbuf_reset(&header);
2083 emit_rewrite_diff(name_a, name_b, one, two,
2084 textconv_one, textconv_two, o);
2085 o->found_changes = 1;
2086 goto free_ab_and_return;
2090 if (o->irreversible_delete && lbl[1][0] == '/') {
2091 fprintf(o->file, "%s", header.buf);
2092 strbuf_reset(&header);
2093 goto free_ab_and_return;
2094 } else if (!DIFF_OPT_TST(o, TEXT) &&
2095 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2096 (!textconv_two && diff_filespec_is_binary(two)) )) {
2097 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2098 die("unable to read files to diff");
2099 /* Quite common confusing case */
2100 if (mf1.size == mf2.size &&
2101 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2102 if (must_show_header)
2103 fprintf(o->file, "%s", header.buf);
2104 goto free_ab_and_return;
2106 fprintf(o->file, "%s", header.buf);
2107 strbuf_reset(&header);
2108 if (DIFF_OPT_TST(o, BINARY))
2109 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2110 else
2111 fprintf(o->file, "%sBinary files %s and %s differ\n",
2112 line_prefix, lbl[0], lbl[1]);
2113 o->found_changes = 1;
2114 } else {
2115 /* Crazy xdl interfaces.. */
2116 const char *diffopts = getenv("GIT_DIFF_OPTS");
2117 xpparam_t xpp;
2118 xdemitconf_t xecfg;
2119 struct emit_callback ecbdata;
2120 const struct userdiff_funcname *pe;
2122 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2123 fprintf(o->file, "%s", header.buf);
2124 strbuf_reset(&header);
2127 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2128 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2130 pe = diff_funcname_pattern(one);
2131 if (!pe)
2132 pe = diff_funcname_pattern(two);
2134 memset(&xpp, 0, sizeof(xpp));
2135 memset(&xecfg, 0, sizeof(xecfg));
2136 memset(&ecbdata, 0, sizeof(ecbdata));
2137 ecbdata.label_path = lbl;
2138 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2139 ecbdata.found_changesp = &o->found_changes;
2140 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2141 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2142 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2143 ecbdata.opt = o;
2144 ecbdata.header = header.len ? &header : NULL;
2145 xpp.flags = o->xdl_opts;
2146 xecfg.ctxlen = o->context;
2147 xecfg.interhunkctxlen = o->interhunkcontext;
2148 xecfg.flags = XDL_EMIT_FUNCNAMES;
2149 if (pe)
2150 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2151 if (!diffopts)
2153 else if (!prefixcmp(diffopts, "--unified="))
2154 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2155 else if (!prefixcmp(diffopts, "-u"))
2156 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2157 if (o->word_diff) {
2158 int i;
2160 ecbdata.diff_words =
2161 xcalloc(1, sizeof(struct diff_words_data));
2162 ecbdata.diff_words->type = o->word_diff;
2163 ecbdata.diff_words->opt = o;
2164 if (!o->word_regex)
2165 o->word_regex = userdiff_word_regex(one);
2166 if (!o->word_regex)
2167 o->word_regex = userdiff_word_regex(two);
2168 if (!o->word_regex)
2169 o->word_regex = diff_word_regex_cfg;
2170 if (o->word_regex) {
2171 ecbdata.diff_words->word_regex = (regex_t *)
2172 xmalloc(sizeof(regex_t));
2173 if (regcomp(ecbdata.diff_words->word_regex,
2174 o->word_regex,
2175 REG_EXTENDED | REG_NEWLINE))
2176 die ("Invalid regular expression: %s",
2177 o->word_regex);
2179 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2180 if (o->word_diff == diff_words_styles[i].type) {
2181 ecbdata.diff_words->style =
2182 &diff_words_styles[i];
2183 break;
2186 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2187 struct diff_words_style *st = ecbdata.diff_words->style;
2188 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2189 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2190 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2193 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2194 &xpp, &xecfg);
2195 if (o->word_diff)
2196 free_diff_words_data(&ecbdata);
2197 if (textconv_one)
2198 free(mf1.ptr);
2199 if (textconv_two)
2200 free(mf2.ptr);
2201 xdiff_clear_find_func(&xecfg);
2204 free_ab_and_return:
2205 strbuf_release(&header);
2206 diff_free_filespec_data(one);
2207 diff_free_filespec_data(two);
2208 free(a_one);
2209 free(b_two);
2210 return;
2213 static void builtin_diffstat(const char *name_a, const char *name_b,
2214 struct diff_filespec *one,
2215 struct diff_filespec *two,
2216 struct diffstat_t *diffstat,
2217 struct diff_options *o,
2218 int complete_rewrite)
2220 mmfile_t mf1, mf2;
2221 struct diffstat_file *data;
2223 data = diffstat_add(diffstat, name_a, name_b);
2225 if (!one || !two) {
2226 data->is_unmerged = 1;
2227 return;
2230 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2231 data->is_binary = 1;
2232 data->added = diff_filespec_size(two);
2233 data->deleted = diff_filespec_size(one);
2236 else if (complete_rewrite) {
2237 diff_populate_filespec(one, 0);
2238 diff_populate_filespec(two, 0);
2239 data->deleted = count_lines(one->data, one->size);
2240 data->added = count_lines(two->data, two->size);
2243 else {
2244 /* Crazy xdl interfaces.. */
2245 xpparam_t xpp;
2246 xdemitconf_t xecfg;
2248 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2249 die("unable to read files to diff");
2251 memset(&xpp, 0, sizeof(xpp));
2252 memset(&xecfg, 0, sizeof(xecfg));
2253 xpp.flags = o->xdl_opts;
2254 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2255 &xpp, &xecfg);
2258 diff_free_filespec_data(one);
2259 diff_free_filespec_data(two);
2262 static void builtin_checkdiff(const char *name_a, const char *name_b,
2263 const char *attr_path,
2264 struct diff_filespec *one,
2265 struct diff_filespec *two,
2266 struct diff_options *o)
2268 mmfile_t mf1, mf2;
2269 struct checkdiff_t data;
2271 if (!two)
2272 return;
2274 memset(&data, 0, sizeof(data));
2275 data.filename = name_b ? name_b : name_a;
2276 data.lineno = 0;
2277 data.o = o;
2278 data.ws_rule = whitespace_rule(attr_path);
2279 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2281 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2282 die("unable to read files to diff");
2285 * All the other codepaths check both sides, but not checking
2286 * the "old" side here is deliberate. We are checking the newly
2287 * introduced changes, and as long as the "new" side is text, we
2288 * can and should check what it introduces.
2290 if (diff_filespec_is_binary(two))
2291 goto free_and_return;
2292 else {
2293 /* Crazy xdl interfaces.. */
2294 xpparam_t xpp;
2295 xdemitconf_t xecfg;
2297 memset(&xpp, 0, sizeof(xpp));
2298 memset(&xecfg, 0, sizeof(xecfg));
2299 xecfg.ctxlen = 1; /* at least one context line */
2300 xpp.flags = 0;
2301 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2302 &xpp, &xecfg);
2304 if (data.ws_rule & WS_BLANK_AT_EOF) {
2305 struct emit_callback ecbdata;
2306 int blank_at_eof;
2308 ecbdata.ws_rule = data.ws_rule;
2309 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2310 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2312 if (blank_at_eof) {
2313 static char *err;
2314 if (!err)
2315 err = whitespace_error_string(WS_BLANK_AT_EOF);
2316 fprintf(o->file, "%s:%d: %s.\n",
2317 data.filename, blank_at_eof, err);
2318 data.status = 1; /* report errors */
2322 free_and_return:
2323 diff_free_filespec_data(one);
2324 diff_free_filespec_data(two);
2325 if (data.status)
2326 DIFF_OPT_SET(o, CHECK_FAILED);
2329 struct diff_filespec *alloc_filespec(const char *path)
2331 int namelen = strlen(path);
2332 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2334 memset(spec, 0, sizeof(*spec));
2335 spec->path = (char *)(spec + 1);
2336 memcpy(spec->path, path, namelen+1);
2337 spec->count = 1;
2338 spec->is_binary = -1;
2339 return spec;
2342 void free_filespec(struct diff_filespec *spec)
2344 if (!--spec->count) {
2345 diff_free_filespec_data(spec);
2346 free(spec);
2350 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2351 unsigned short mode)
2353 if (mode) {
2354 spec->mode = canon_mode(mode);
2355 hashcpy(spec->sha1, sha1);
2356 spec->sha1_valid = !is_null_sha1(sha1);
2361 * Given a name and sha1 pair, if the index tells us the file in
2362 * the work tree has that object contents, return true, so that
2363 * prepare_temp_file() does not have to inflate and extract.
2365 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2367 struct cache_entry *ce;
2368 struct stat st;
2369 int pos, len;
2372 * We do not read the cache ourselves here, because the
2373 * benchmark with my previous version that always reads cache
2374 * shows that it makes things worse for diff-tree comparing
2375 * two linux-2.6 kernel trees in an already checked out work
2376 * tree. This is because most diff-tree comparisons deal with
2377 * only a small number of files, while reading the cache is
2378 * expensive for a large project, and its cost outweighs the
2379 * savings we get by not inflating the object to a temporary
2380 * file. Practically, this code only helps when we are used
2381 * by diff-cache --cached, which does read the cache before
2382 * calling us.
2384 if (!active_cache)
2385 return 0;
2387 /* We want to avoid the working directory if our caller
2388 * doesn't need the data in a normal file, this system
2389 * is rather slow with its stat/open/mmap/close syscalls,
2390 * and the object is contained in a pack file. The pack
2391 * is probably already open and will be faster to obtain
2392 * the data through than the working directory. Loose
2393 * objects however would tend to be slower as they need
2394 * to be individually opened and inflated.
2396 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2397 return 0;
2399 len = strlen(name);
2400 pos = cache_name_pos(name, len);
2401 if (pos < 0)
2402 return 0;
2403 ce = active_cache[pos];
2406 * This is not the sha1 we are looking for, or
2407 * unreusable because it is not a regular file.
2409 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2410 return 0;
2413 * If ce is marked as "assume unchanged", there is no
2414 * guarantee that work tree matches what we are looking for.
2416 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2417 return 0;
2420 * If ce matches the file in the work tree, we can reuse it.
2422 if (ce_uptodate(ce) ||
2423 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2424 return 1;
2426 return 0;
2429 static int populate_from_stdin(struct diff_filespec *s)
2431 struct strbuf buf = STRBUF_INIT;
2432 size_t size = 0;
2434 if (strbuf_read(&buf, 0, 0) < 0)
2435 return error("error while reading from stdin %s",
2436 strerror(errno));
2438 s->should_munmap = 0;
2439 s->data = strbuf_detach(&buf, &size);
2440 s->size = size;
2441 s->should_free = 1;
2442 return 0;
2445 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2447 int len;
2448 char *data = xmalloc(100), *dirty = "";
2450 /* Are we looking at the work tree? */
2451 if (s->dirty_submodule)
2452 dirty = "-dirty";
2454 len = snprintf(data, 100,
2455 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2456 s->data = data;
2457 s->size = len;
2458 s->should_free = 1;
2459 if (size_only) {
2460 s->data = NULL;
2461 free(data);
2463 return 0;
2467 * While doing rename detection and pickaxe operation, we may need to
2468 * grab the data for the blob (or file) for our own in-core comparison.
2469 * diff_filespec has data and size fields for this purpose.
2471 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2473 int err = 0;
2474 if (!DIFF_FILE_VALID(s))
2475 die("internal error: asking to populate invalid file.");
2476 if (S_ISDIR(s->mode))
2477 return -1;
2479 if (s->data)
2480 return 0;
2482 if (size_only && 0 < s->size)
2483 return 0;
2485 if (S_ISGITLINK(s->mode))
2486 return diff_populate_gitlink(s, size_only);
2488 if (!s->sha1_valid ||
2489 reuse_worktree_file(s->path, s->sha1, 0)) {
2490 struct strbuf buf = STRBUF_INIT;
2491 struct stat st;
2492 int fd;
2494 if (!strcmp(s->path, "-"))
2495 return populate_from_stdin(s);
2497 if (lstat(s->path, &st) < 0) {
2498 if (errno == ENOENT) {
2499 err_empty:
2500 err = -1;
2501 empty:
2502 s->data = (char *)"";
2503 s->size = 0;
2504 return err;
2507 s->size = xsize_t(st.st_size);
2508 if (!s->size)
2509 goto empty;
2510 if (S_ISLNK(st.st_mode)) {
2511 struct strbuf sb = STRBUF_INIT;
2513 if (strbuf_readlink(&sb, s->path, s->size))
2514 goto err_empty;
2515 s->size = sb.len;
2516 s->data = strbuf_detach(&sb, NULL);
2517 s->should_free = 1;
2518 return 0;
2520 if (size_only)
2521 return 0;
2522 fd = open(s->path, O_RDONLY);
2523 if (fd < 0)
2524 goto err_empty;
2525 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2526 close(fd);
2527 s->should_munmap = 1;
2530 * Convert from working tree format to canonical git format
2532 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2533 size_t size = 0;
2534 munmap(s->data, s->size);
2535 s->should_munmap = 0;
2536 s->data = strbuf_detach(&buf, &size);
2537 s->size = size;
2538 s->should_free = 1;
2541 else {
2542 enum object_type type;
2543 if (size_only) {
2544 type = sha1_object_info(s->sha1, &s->size);
2545 if (type < 0)
2546 die("unable to read %s", sha1_to_hex(s->sha1));
2547 } else {
2548 s->data = read_sha1_file(s->sha1, &type, &s->size);
2549 if (!s->data)
2550 die("unable to read %s", sha1_to_hex(s->sha1));
2551 s->should_free = 1;
2554 return 0;
2557 void diff_free_filespec_blob(struct diff_filespec *s)
2559 if (s->should_free)
2560 free(s->data);
2561 else if (s->should_munmap)
2562 munmap(s->data, s->size);
2564 if (s->should_free || s->should_munmap) {
2565 s->should_free = s->should_munmap = 0;
2566 s->data = NULL;
2570 void diff_free_filespec_data(struct diff_filespec *s)
2572 diff_free_filespec_blob(s);
2573 free(s->cnt_data);
2574 s->cnt_data = NULL;
2577 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2578 void *blob,
2579 unsigned long size,
2580 const unsigned char *sha1,
2581 int mode)
2583 int fd;
2584 struct strbuf buf = STRBUF_INIT;
2585 struct strbuf template = STRBUF_INIT;
2586 char *path_dup = xstrdup(path);
2587 const char *base = basename(path_dup);
2589 /* Generate "XXXXXX_basename.ext" */
2590 strbuf_addstr(&template, "XXXXXX_");
2591 strbuf_addstr(&template, base);
2593 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2594 strlen(base) + 1);
2595 if (fd < 0)
2596 die_errno("unable to create temp-file");
2597 if (convert_to_working_tree(path,
2598 (const char *)blob, (size_t)size, &buf)) {
2599 blob = buf.buf;
2600 size = buf.len;
2602 if (write_in_full(fd, blob, size) != size)
2603 die_errno("unable to write temp-file");
2604 close(fd);
2605 temp->name = temp->tmp_path;
2606 strcpy(temp->hex, sha1_to_hex(sha1));
2607 temp->hex[40] = 0;
2608 sprintf(temp->mode, "%06o", mode);
2609 strbuf_release(&buf);
2610 strbuf_release(&template);
2611 free(path_dup);
2614 static struct diff_tempfile *prepare_temp_file(const char *name,
2615 struct diff_filespec *one)
2617 struct diff_tempfile *temp = claim_diff_tempfile();
2619 if (!DIFF_FILE_VALID(one)) {
2620 not_a_valid_file:
2621 /* A '-' entry produces this for file-2, and
2622 * a '+' entry produces this for file-1.
2624 temp->name = "/dev/null";
2625 strcpy(temp->hex, ".");
2626 strcpy(temp->mode, ".");
2627 return temp;
2630 if (!remove_tempfile_installed) {
2631 atexit(remove_tempfile);
2632 sigchain_push_common(remove_tempfile_on_signal);
2633 remove_tempfile_installed = 1;
2636 if (!one->sha1_valid ||
2637 reuse_worktree_file(name, one->sha1, 1)) {
2638 struct stat st;
2639 if (lstat(name, &st) < 0) {
2640 if (errno == ENOENT)
2641 goto not_a_valid_file;
2642 die_errno("stat(%s)", name);
2644 if (S_ISLNK(st.st_mode)) {
2645 struct strbuf sb = STRBUF_INIT;
2646 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2647 die_errno("readlink(%s)", name);
2648 prep_temp_blob(name, temp, sb.buf, sb.len,
2649 (one->sha1_valid ?
2650 one->sha1 : null_sha1),
2651 (one->sha1_valid ?
2652 one->mode : S_IFLNK));
2653 strbuf_release(&sb);
2655 else {
2656 /* we can borrow from the file in the work tree */
2657 temp->name = name;
2658 if (!one->sha1_valid)
2659 strcpy(temp->hex, sha1_to_hex(null_sha1));
2660 else
2661 strcpy(temp->hex, sha1_to_hex(one->sha1));
2662 /* Even though we may sometimes borrow the
2663 * contents from the work tree, we always want
2664 * one->mode. mode is trustworthy even when
2665 * !(one->sha1_valid), as long as
2666 * DIFF_FILE_VALID(one).
2668 sprintf(temp->mode, "%06o", one->mode);
2670 return temp;
2672 else {
2673 if (diff_populate_filespec(one, 0))
2674 die("cannot read data blob for %s", one->path);
2675 prep_temp_blob(name, temp, one->data, one->size,
2676 one->sha1, one->mode);
2678 return temp;
2681 /* An external diff command takes:
2683 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2684 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2687 static void run_external_diff(const char *pgm,
2688 const char *name,
2689 const char *other,
2690 struct diff_filespec *one,
2691 struct diff_filespec *two,
2692 const char *xfrm_msg,
2693 int complete_rewrite)
2695 const char *spawn_arg[10];
2696 int retval;
2697 const char **arg = &spawn_arg[0];
2699 if (one && two) {
2700 struct diff_tempfile *temp_one, *temp_two;
2701 const char *othername = (other ? other : name);
2702 temp_one = prepare_temp_file(name, one);
2703 temp_two = prepare_temp_file(othername, two);
2704 *arg++ = pgm;
2705 *arg++ = name;
2706 *arg++ = temp_one->name;
2707 *arg++ = temp_one->hex;
2708 *arg++ = temp_one->mode;
2709 *arg++ = temp_two->name;
2710 *arg++ = temp_two->hex;
2711 *arg++ = temp_two->mode;
2712 if (other) {
2713 *arg++ = other;
2714 *arg++ = xfrm_msg;
2716 } else {
2717 *arg++ = pgm;
2718 *arg++ = name;
2720 *arg = NULL;
2721 fflush(NULL);
2722 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2723 remove_tempfile();
2724 if (retval) {
2725 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2726 exit(1);
2730 static int similarity_index(struct diff_filepair *p)
2732 return p->score * 100 / MAX_SCORE;
2735 static void fill_metainfo(struct strbuf *msg,
2736 const char *name,
2737 const char *other,
2738 struct diff_filespec *one,
2739 struct diff_filespec *two,
2740 struct diff_options *o,
2741 struct diff_filepair *p,
2742 int *must_show_header,
2743 int use_color)
2745 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2746 const char *reset = diff_get_color(use_color, DIFF_RESET);
2747 struct strbuf *msgbuf;
2748 char *line_prefix = "";
2750 *must_show_header = 1;
2751 if (o->output_prefix) {
2752 msgbuf = o->output_prefix(o, o->output_prefix_data);
2753 line_prefix = msgbuf->buf;
2755 strbuf_init(msg, PATH_MAX * 2 + 300);
2756 switch (p->status) {
2757 case DIFF_STATUS_COPIED:
2758 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2759 line_prefix, set, similarity_index(p));
2760 strbuf_addf(msg, "%s\n%s%scopy from ",
2761 reset, line_prefix, set);
2762 quote_c_style(name, msg, NULL, 0);
2763 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2764 quote_c_style(other, msg, NULL, 0);
2765 strbuf_addf(msg, "%s\n", reset);
2766 break;
2767 case DIFF_STATUS_RENAMED:
2768 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2769 line_prefix, set, similarity_index(p));
2770 strbuf_addf(msg, "%s\n%s%srename from ",
2771 reset, line_prefix, set);
2772 quote_c_style(name, msg, NULL, 0);
2773 strbuf_addf(msg, "%s\n%s%srename to ",
2774 reset, line_prefix, set);
2775 quote_c_style(other, msg, NULL, 0);
2776 strbuf_addf(msg, "%s\n", reset);
2777 break;
2778 case DIFF_STATUS_MODIFIED:
2779 if (p->score) {
2780 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2781 line_prefix,
2782 set, similarity_index(p), reset);
2783 break;
2785 /* fallthru */
2786 default:
2787 *must_show_header = 0;
2789 if (one && two && hashcmp(one->sha1, two->sha1)) {
2790 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2792 if (DIFF_OPT_TST(o, BINARY)) {
2793 mmfile_t mf;
2794 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2795 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2796 abbrev = 40;
2798 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2799 find_unique_abbrev(one->sha1, abbrev));
2800 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2801 if (one->mode == two->mode)
2802 strbuf_addf(msg, " %06o", one->mode);
2803 strbuf_addf(msg, "%s\n", reset);
2807 static void run_diff_cmd(const char *pgm,
2808 const char *name,
2809 const char *other,
2810 const char *attr_path,
2811 struct diff_filespec *one,
2812 struct diff_filespec *two,
2813 struct strbuf *msg,
2814 struct diff_options *o,
2815 struct diff_filepair *p)
2817 const char *xfrm_msg = NULL;
2818 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2819 int must_show_header = 0;
2821 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2822 pgm = NULL;
2823 else {
2824 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2825 if (drv && drv->external)
2826 pgm = drv->external;
2829 if (msg) {
2831 * don't use colors when the header is intended for an
2832 * external diff driver
2834 fill_metainfo(msg, name, other, one, two, o, p,
2835 &must_show_header,
2836 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2837 xfrm_msg = msg->len ? msg->buf : NULL;
2840 if (pgm) {
2841 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2842 complete_rewrite);
2843 return;
2845 if (one && two)
2846 builtin_diff(name, other ? other : name,
2847 one, two, xfrm_msg, must_show_header,
2848 o, complete_rewrite);
2849 else
2850 fprintf(o->file, "* Unmerged path %s\n", name);
2853 static void diff_fill_sha1_info(struct diff_filespec *one)
2855 if (DIFF_FILE_VALID(one)) {
2856 if (!one->sha1_valid) {
2857 struct stat st;
2858 if (!strcmp(one->path, "-")) {
2859 hashcpy(one->sha1, null_sha1);
2860 return;
2862 if (lstat(one->path, &st) < 0)
2863 die_errno("stat '%s'", one->path);
2864 if (index_path(one->sha1, one->path, &st, 0))
2865 die("cannot hash %s", one->path);
2868 else
2869 hashclr(one->sha1);
2872 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2874 /* Strip the prefix but do not molest /dev/null and absolute paths */
2875 if (*namep && **namep != '/') {
2876 *namep += prefix_length;
2877 if (**namep == '/')
2878 ++*namep;
2880 if (*otherp && **otherp != '/') {
2881 *otherp += prefix_length;
2882 if (**otherp == '/')
2883 ++*otherp;
2887 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2889 const char *pgm = external_diff();
2890 struct strbuf msg;
2891 struct diff_filespec *one = p->one;
2892 struct diff_filespec *two = p->two;
2893 const char *name;
2894 const char *other;
2895 const char *attr_path;
2897 name = p->one->path;
2898 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2899 attr_path = name;
2900 if (o->prefix_length)
2901 strip_prefix(o->prefix_length, &name, &other);
2903 if (DIFF_PAIR_UNMERGED(p)) {
2904 run_diff_cmd(pgm, name, NULL, attr_path,
2905 NULL, NULL, NULL, o, p);
2906 return;
2909 diff_fill_sha1_info(one);
2910 diff_fill_sha1_info(two);
2912 if (!pgm &&
2913 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2914 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2916 * a filepair that changes between file and symlink
2917 * needs to be split into deletion and creation.
2919 struct diff_filespec *null = alloc_filespec(two->path);
2920 run_diff_cmd(NULL, name, other, attr_path,
2921 one, null, &msg, o, p);
2922 free(null);
2923 strbuf_release(&msg);
2925 null = alloc_filespec(one->path);
2926 run_diff_cmd(NULL, name, other, attr_path,
2927 null, two, &msg, o, p);
2928 free(null);
2930 else
2931 run_diff_cmd(pgm, name, other, attr_path,
2932 one, two, &msg, o, p);
2934 strbuf_release(&msg);
2937 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2938 struct diffstat_t *diffstat)
2940 const char *name;
2941 const char *other;
2942 int complete_rewrite = 0;
2944 if (DIFF_PAIR_UNMERGED(p)) {
2945 /* unmerged */
2946 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2947 return;
2950 name = p->one->path;
2951 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2953 if (o->prefix_length)
2954 strip_prefix(o->prefix_length, &name, &other);
2956 diff_fill_sha1_info(p->one);
2957 diff_fill_sha1_info(p->two);
2959 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2960 complete_rewrite = 1;
2961 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2964 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2966 const char *name;
2967 const char *other;
2968 const char *attr_path;
2970 if (DIFF_PAIR_UNMERGED(p)) {
2971 /* unmerged */
2972 return;
2975 name = p->one->path;
2976 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2977 attr_path = other ? other : name;
2979 if (o->prefix_length)
2980 strip_prefix(o->prefix_length, &name, &other);
2982 diff_fill_sha1_info(p->one);
2983 diff_fill_sha1_info(p->two);
2985 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2988 void diff_setup(struct diff_options *options)
2990 memcpy(options, &default_diff_options, sizeof(*options));
2992 options->file = stdout;
2994 options->line_termination = '\n';
2995 options->break_opt = -1;
2996 options->rename_limit = -1;
2997 options->dirstat_permille = diff_dirstat_permille_default;
2998 options->context = 3;
3000 options->change = diff_change;
3001 options->add_remove = diff_addremove;
3002 if (diff_use_color_default > 0)
3003 DIFF_OPT_SET(options, COLOR_DIFF);
3004 options->detect_rename = diff_detect_rename_default;
3006 if (diff_no_prefix) {
3007 options->a_prefix = options->b_prefix = "";
3008 } else if (!diff_mnemonic_prefix) {
3009 options->a_prefix = "a/";
3010 options->b_prefix = "b/";
3014 int diff_setup_done(struct diff_options *options)
3016 int count = 0;
3018 if (options->output_format & DIFF_FORMAT_NAME)
3019 count++;
3020 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3021 count++;
3022 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3023 count++;
3024 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3025 count++;
3026 if (count > 1)
3027 die("--name-only, --name-status, --check and -s are mutually exclusive");
3030 * Most of the time we can say "there are changes"
3031 * only by checking if there are changed paths, but
3032 * --ignore-whitespace* options force us to look
3033 * inside contents.
3036 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3037 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3038 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3039 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3040 else
3041 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3043 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3044 options->detect_rename = DIFF_DETECT_COPY;
3046 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3047 options->prefix = NULL;
3048 if (options->prefix)
3049 options->prefix_length = strlen(options->prefix);
3050 else
3051 options->prefix_length = 0;
3053 if (options->output_format & (DIFF_FORMAT_NAME |
3054 DIFF_FORMAT_NAME_STATUS |
3055 DIFF_FORMAT_CHECKDIFF |
3056 DIFF_FORMAT_NO_OUTPUT))
3057 options->output_format &= ~(DIFF_FORMAT_RAW |
3058 DIFF_FORMAT_NUMSTAT |
3059 DIFF_FORMAT_DIFFSTAT |
3060 DIFF_FORMAT_SHORTSTAT |
3061 DIFF_FORMAT_DIRSTAT |
3062 DIFF_FORMAT_SUMMARY |
3063 DIFF_FORMAT_PATCH);
3066 * These cases always need recursive; we do not drop caller-supplied
3067 * recursive bits for other formats here.
3069 if (options->output_format & (DIFF_FORMAT_PATCH |
3070 DIFF_FORMAT_NUMSTAT |
3071 DIFF_FORMAT_DIFFSTAT |
3072 DIFF_FORMAT_SHORTSTAT |
3073 DIFF_FORMAT_DIRSTAT |
3074 DIFF_FORMAT_SUMMARY |
3075 DIFF_FORMAT_CHECKDIFF))
3076 DIFF_OPT_SET(options, RECURSIVE);
3078 * Also pickaxe would not work very well if you do not say recursive
3080 if (options->pickaxe)
3081 DIFF_OPT_SET(options, RECURSIVE);
3083 * When patches are generated, submodules diffed against the work tree
3084 * must be checked for dirtiness too so it can be shown in the output
3086 if (options->output_format & DIFF_FORMAT_PATCH)
3087 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3089 if (options->detect_rename && options->rename_limit < 0)
3090 options->rename_limit = diff_rename_limit_default;
3091 if (options->setup & DIFF_SETUP_USE_CACHE) {
3092 if (!active_cache)
3093 /* read-cache does not die even when it fails
3094 * so it is safe for us to do this here. Also
3095 * it does not smudge active_cache or active_nr
3096 * when it fails, so we do not have to worry about
3097 * cleaning it up ourselves either.
3099 read_cache();
3101 if (options->abbrev <= 0 || 40 < options->abbrev)
3102 options->abbrev = 40; /* full */
3105 * It does not make sense to show the first hit we happened
3106 * to have found. It does not make sense not to return with
3107 * exit code in such a case either.
3109 if (DIFF_OPT_TST(options, QUICK)) {
3110 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3111 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3114 return 0;
3117 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3119 char c, *eq;
3120 int len;
3122 if (*arg != '-')
3123 return 0;
3124 c = *++arg;
3125 if (!c)
3126 return 0;
3127 if (c == arg_short) {
3128 c = *++arg;
3129 if (!c)
3130 return 1;
3131 if (val && isdigit(c)) {
3132 char *end;
3133 int n = strtoul(arg, &end, 10);
3134 if (*end)
3135 return 0;
3136 *val = n;
3137 return 1;
3139 return 0;
3141 if (c != '-')
3142 return 0;
3143 arg++;
3144 eq = strchr(arg, '=');
3145 if (eq)
3146 len = eq - arg;
3147 else
3148 len = strlen(arg);
3149 if (!len || strncmp(arg, arg_long, len))
3150 return 0;
3151 if (eq) {
3152 int n;
3153 char *end;
3154 if (!isdigit(*++eq))
3155 return 0;
3156 n = strtoul(eq, &end, 10);
3157 if (*end)
3158 return 0;
3159 *val = n;
3161 return 1;
3164 static int diff_scoreopt_parse(const char *opt);
3166 static inline int short_opt(char opt, const char **argv,
3167 const char **optarg)
3169 const char *arg = argv[0];
3170 if (arg[0] != '-' || arg[1] != opt)
3171 return 0;
3172 if (arg[2] != '\0') {
3173 *optarg = arg + 2;
3174 return 1;
3176 if (!argv[1])
3177 die("Option '%c' requires a value", opt);
3178 *optarg = argv[1];
3179 return 2;
3182 int parse_long_opt(const char *opt, const char **argv,
3183 const char **optarg)
3185 const char *arg = argv[0];
3186 if (arg[0] != '-' || arg[1] != '-')
3187 return 0;
3188 arg += strlen("--");
3189 if (prefixcmp(arg, opt))
3190 return 0;
3191 arg += strlen(opt);
3192 if (*arg == '=') { /* sticked form: --option=value */
3193 *optarg = arg + 1;
3194 return 1;
3196 if (*arg != '\0')
3197 return 0;
3198 /* separate form: --option value */
3199 if (!argv[1])
3200 die("Option '--%s' requires a value", opt);
3201 *optarg = argv[1];
3202 return 2;
3205 static int stat_opt(struct diff_options *options, const char **av)
3207 const char *arg = av[0];
3208 char *end;
3209 int width = options->stat_width;
3210 int name_width = options->stat_name_width;
3211 int argcount = 1;
3213 arg += strlen("--stat");
3214 end = (char *)arg;
3216 switch (*arg) {
3217 case '-':
3218 if (!prefixcmp(arg, "-width")) {
3219 arg += strlen("-width");
3220 if (*arg == '=')
3221 width = strtoul(arg + 1, &end, 10);
3222 else if (!*arg && !av[1])
3223 die("Option '--stat-width' requires a value");
3224 else if (!*arg) {
3225 width = strtoul(av[1], &end, 10);
3226 argcount = 2;
3228 } else if (!prefixcmp(arg, "-name-width")) {
3229 arg += strlen("-name-width");
3230 if (*arg == '=')
3231 name_width = strtoul(arg + 1, &end, 10);
3232 else if (!*arg && !av[1])
3233 die("Option '--stat-name-width' requires a value");
3234 else if (!*arg) {
3235 name_width = strtoul(av[1], &end, 10);
3236 argcount = 2;
3239 break;
3240 case '=':
3241 width = strtoul(arg+1, &end, 10);
3242 if (*end == ',')
3243 name_width = strtoul(end+1, &end, 10);
3246 /* Important! This checks all the error cases! */
3247 if (*end)
3248 return 0;
3249 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3250 options->stat_name_width = name_width;
3251 options->stat_width = width;
3252 return argcount;
3255 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3257 struct strbuf errmsg = STRBUF_INIT;
3258 if (parse_dirstat_params(options, params, &errmsg))
3259 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3260 errmsg.buf);
3261 strbuf_release(&errmsg);
3263 * The caller knows a dirstat-related option is given from the command
3264 * line; allow it to say "return this_function();"
3266 options->output_format |= DIFF_FORMAT_DIRSTAT;
3267 return 1;
3270 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3272 const char *arg = av[0];
3273 const char *optarg;
3274 int argcount;
3276 /* Output format options */
3277 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3278 options->output_format |= DIFF_FORMAT_PATCH;
3279 else if (opt_arg(arg, 'U', "unified", &options->context))
3280 options->output_format |= DIFF_FORMAT_PATCH;
3281 else if (!strcmp(arg, "--raw"))
3282 options->output_format |= DIFF_FORMAT_RAW;
3283 else if (!strcmp(arg, "--patch-with-raw"))
3284 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3285 else if (!strcmp(arg, "--numstat"))
3286 options->output_format |= DIFF_FORMAT_NUMSTAT;
3287 else if (!strcmp(arg, "--shortstat"))
3288 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3289 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3290 return parse_dirstat_opt(options, "");
3291 else if (!prefixcmp(arg, "-X"))
3292 return parse_dirstat_opt(options, arg + 2);
3293 else if (!prefixcmp(arg, "--dirstat="))
3294 return parse_dirstat_opt(options, arg + 10);
3295 else if (!strcmp(arg, "--cumulative"))
3296 return parse_dirstat_opt(options, "cumulative");
3297 else if (!strcmp(arg, "--dirstat-by-file"))
3298 return parse_dirstat_opt(options, "files");
3299 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3300 parse_dirstat_opt(options, "files");
3301 return parse_dirstat_opt(options, arg + 18);
3303 else if (!strcmp(arg, "--check"))
3304 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3305 else if (!strcmp(arg, "--summary"))
3306 options->output_format |= DIFF_FORMAT_SUMMARY;
3307 else if (!strcmp(arg, "--patch-with-stat"))
3308 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3309 else if (!strcmp(arg, "--name-only"))
3310 options->output_format |= DIFF_FORMAT_NAME;
3311 else if (!strcmp(arg, "--name-status"))
3312 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3313 else if (!strcmp(arg, "-s"))
3314 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3315 else if (!prefixcmp(arg, "--stat"))
3316 /* --stat, --stat-width, or --stat-name-width */
3317 return stat_opt(options, av);
3319 /* renames options */
3320 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3321 !strcmp(arg, "--break-rewrites")) {
3322 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3323 return error("invalid argument to -B: %s", arg+2);
3325 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3326 !strcmp(arg, "--find-renames")) {
3327 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3328 return error("invalid argument to -M: %s", arg+2);
3329 options->detect_rename = DIFF_DETECT_RENAME;
3331 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3332 options->irreversible_delete = 1;
3334 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3335 !strcmp(arg, "--find-copies")) {
3336 if (options->detect_rename == DIFF_DETECT_COPY)
3337 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3338 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3339 return error("invalid argument to -C: %s", arg+2);
3340 options->detect_rename = DIFF_DETECT_COPY;
3342 else if (!strcmp(arg, "--no-renames"))
3343 options->detect_rename = 0;
3344 else if (!strcmp(arg, "--relative"))
3345 DIFF_OPT_SET(options, RELATIVE_NAME);
3346 else if (!prefixcmp(arg, "--relative=")) {
3347 DIFF_OPT_SET(options, RELATIVE_NAME);
3348 options->prefix = arg + 11;
3351 /* xdiff options */
3352 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3353 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3354 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3355 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3356 else if (!strcmp(arg, "--ignore-space-at-eol"))
3357 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3358 else if (!strcmp(arg, "--patience"))
3359 DIFF_XDL_SET(options, PATIENCE_DIFF);
3361 /* flags options */
3362 else if (!strcmp(arg, "--binary")) {
3363 options->output_format |= DIFF_FORMAT_PATCH;
3364 DIFF_OPT_SET(options, BINARY);
3366 else if (!strcmp(arg, "--full-index"))
3367 DIFF_OPT_SET(options, FULL_INDEX);
3368 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3369 DIFF_OPT_SET(options, TEXT);
3370 else if (!strcmp(arg, "-R"))
3371 DIFF_OPT_SET(options, REVERSE_DIFF);
3372 else if (!strcmp(arg, "--find-copies-harder"))
3373 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3374 else if (!strcmp(arg, "--follow"))
3375 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3376 else if (!strcmp(arg, "--color"))
3377 DIFF_OPT_SET(options, COLOR_DIFF);
3378 else if (!prefixcmp(arg, "--color=")) {
3379 int value = git_config_colorbool(NULL, arg+8, -1);
3380 if (value == 0)
3381 DIFF_OPT_CLR(options, COLOR_DIFF);
3382 else if (value > 0)
3383 DIFF_OPT_SET(options, COLOR_DIFF);
3384 else
3385 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3387 else if (!strcmp(arg, "--no-color"))
3388 DIFF_OPT_CLR(options, COLOR_DIFF);
3389 else if (!strcmp(arg, "--color-words")) {
3390 DIFF_OPT_SET(options, COLOR_DIFF);
3391 options->word_diff = DIFF_WORDS_COLOR;
3393 else if (!prefixcmp(arg, "--color-words=")) {
3394 DIFF_OPT_SET(options, COLOR_DIFF);
3395 options->word_diff = DIFF_WORDS_COLOR;
3396 options->word_regex = arg + 14;
3398 else if (!strcmp(arg, "--word-diff")) {
3399 if (options->word_diff == DIFF_WORDS_NONE)
3400 options->word_diff = DIFF_WORDS_PLAIN;
3402 else if (!prefixcmp(arg, "--word-diff=")) {
3403 const char *type = arg + 12;
3404 if (!strcmp(type, "plain"))
3405 options->word_diff = DIFF_WORDS_PLAIN;
3406 else if (!strcmp(type, "color")) {
3407 DIFF_OPT_SET(options, COLOR_DIFF);
3408 options->word_diff = DIFF_WORDS_COLOR;
3410 else if (!strcmp(type, "porcelain"))
3411 options->word_diff = DIFF_WORDS_PORCELAIN;
3412 else if (!strcmp(type, "none"))
3413 options->word_diff = DIFF_WORDS_NONE;
3414 else
3415 die("bad --word-diff argument: %s", type);
3417 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3418 if (options->word_diff == DIFF_WORDS_NONE)
3419 options->word_diff = DIFF_WORDS_PLAIN;
3420 options->word_regex = optarg;
3421 return argcount;
3423 else if (!strcmp(arg, "--exit-code"))
3424 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3425 else if (!strcmp(arg, "--quiet"))
3426 DIFF_OPT_SET(options, QUICK);
3427 else if (!strcmp(arg, "--ext-diff"))
3428 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3429 else if (!strcmp(arg, "--no-ext-diff"))
3430 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3431 else if (!strcmp(arg, "--textconv"))
3432 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3433 else if (!strcmp(arg, "--no-textconv"))
3434 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3435 else if (!strcmp(arg, "--ignore-submodules")) {
3436 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3437 handle_ignore_submodules_arg(options, "all");
3438 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3439 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3440 handle_ignore_submodules_arg(options, arg + 20);
3441 } else if (!strcmp(arg, "--submodule"))
3442 DIFF_OPT_SET(options, SUBMODULE_LOG);
3443 else if (!prefixcmp(arg, "--submodule=")) {
3444 if (!strcmp(arg + 12, "log"))
3445 DIFF_OPT_SET(options, SUBMODULE_LOG);
3448 /* misc options */
3449 else if (!strcmp(arg, "-z"))
3450 options->line_termination = 0;
3451 else if ((argcount = short_opt('l', av, &optarg))) {
3452 options->rename_limit = strtoul(optarg, NULL, 10);
3453 return argcount;
3455 else if ((argcount = short_opt('S', av, &optarg))) {
3456 options->pickaxe = optarg;
3457 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3458 return argcount;
3459 } else if ((argcount = short_opt('G', av, &optarg))) {
3460 options->pickaxe = optarg;
3461 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3462 return argcount;
3464 else if (!strcmp(arg, "--pickaxe-all"))
3465 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3466 else if (!strcmp(arg, "--pickaxe-regex"))
3467 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3468 else if ((argcount = short_opt('O', av, &optarg))) {
3469 options->orderfile = optarg;
3470 return argcount;
3472 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3473 options->filter = optarg;
3474 return argcount;
3476 else if (!strcmp(arg, "--abbrev"))
3477 options->abbrev = DEFAULT_ABBREV;
3478 else if (!prefixcmp(arg, "--abbrev=")) {
3479 options->abbrev = strtoul(arg + 9, NULL, 10);
3480 if (options->abbrev < MINIMUM_ABBREV)
3481 options->abbrev = MINIMUM_ABBREV;
3482 else if (40 < options->abbrev)
3483 options->abbrev = 40;
3485 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3486 options->a_prefix = optarg;
3487 return argcount;
3489 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3490 options->b_prefix = optarg;
3491 return argcount;
3493 else if (!strcmp(arg, "--no-prefix"))
3494 options->a_prefix = options->b_prefix = "";
3495 else if (opt_arg(arg, '\0', "inter-hunk-context",
3496 &options->interhunkcontext))
3498 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3499 options->file = fopen(optarg, "w");
3500 if (!options->file)
3501 die_errno("Could not open '%s'", optarg);
3502 options->close_file = 1;
3503 return argcount;
3504 } else
3505 return 0;
3506 return 1;
3509 int parse_rename_score(const char **cp_p)
3511 unsigned long num, scale;
3512 int ch, dot;
3513 const char *cp = *cp_p;
3515 num = 0;
3516 scale = 1;
3517 dot = 0;
3518 for (;;) {
3519 ch = *cp;
3520 if ( !dot && ch == '.' ) {
3521 scale = 1;
3522 dot = 1;
3523 } else if ( ch == '%' ) {
3524 scale = dot ? scale*100 : 100;
3525 cp++; /* % is always at the end */
3526 break;
3527 } else if ( ch >= '0' && ch <= '9' ) {
3528 if ( scale < 100000 ) {
3529 scale *= 10;
3530 num = (num*10) + (ch-'0');
3532 } else {
3533 break;
3535 cp++;
3537 *cp_p = cp;
3539 /* user says num divided by scale and we say internally that
3540 * is MAX_SCORE * num / scale.
3542 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3545 static int diff_scoreopt_parse(const char *opt)
3547 int opt1, opt2, cmd;
3549 if (*opt++ != '-')
3550 return -1;
3551 cmd = *opt++;
3552 if (cmd == '-') {
3553 /* convert the long-form arguments into short-form versions */
3554 if (!prefixcmp(opt, "break-rewrites")) {
3555 opt += strlen("break-rewrites");
3556 if (*opt == 0 || *opt++ == '=')
3557 cmd = 'B';
3558 } else if (!prefixcmp(opt, "find-copies")) {
3559 opt += strlen("find-copies");
3560 if (*opt == 0 || *opt++ == '=')
3561 cmd = 'C';
3562 } else if (!prefixcmp(opt, "find-renames")) {
3563 opt += strlen("find-renames");
3564 if (*opt == 0 || *opt++ == '=')
3565 cmd = 'M';
3568 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3569 return -1; /* that is not a -M, -C nor -B option */
3571 opt1 = parse_rename_score(&opt);
3572 if (cmd != 'B')
3573 opt2 = 0;
3574 else {
3575 if (*opt == 0)
3576 opt2 = 0;
3577 else if (*opt != '/')
3578 return -1; /* we expect -B80/99 or -B80 */
3579 else {
3580 opt++;
3581 opt2 = parse_rename_score(&opt);
3584 if (*opt != 0)
3585 return -1;
3586 return opt1 | (opt2 << 16);
3589 struct diff_queue_struct diff_queued_diff;
3591 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3593 if (queue->alloc <= queue->nr) {
3594 queue->alloc = alloc_nr(queue->alloc);
3595 queue->queue = xrealloc(queue->queue,
3596 sizeof(dp) * queue->alloc);
3598 queue->queue[queue->nr++] = dp;
3601 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3602 struct diff_filespec *one,
3603 struct diff_filespec *two)
3605 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3606 dp->one = one;
3607 dp->two = two;
3608 if (queue)
3609 diff_q(queue, dp);
3610 return dp;
3613 void diff_free_filepair(struct diff_filepair *p)
3615 free_filespec(p->one);
3616 free_filespec(p->two);
3617 free(p);
3620 /* This is different from find_unique_abbrev() in that
3621 * it stuffs the result with dots for alignment.
3623 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3625 int abblen;
3626 const char *abbrev;
3627 if (len == 40)
3628 return sha1_to_hex(sha1);
3630 abbrev = find_unique_abbrev(sha1, len);
3631 abblen = strlen(abbrev);
3632 if (abblen < 37) {
3633 static char hex[41];
3634 if (len < abblen && abblen <= len + 2)
3635 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3636 else
3637 sprintf(hex, "%s...", abbrev);
3638 return hex;
3640 return sha1_to_hex(sha1);
3643 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3645 int line_termination = opt->line_termination;
3646 int inter_name_termination = line_termination ? '\t' : '\0';
3647 if (opt->output_prefix) {
3648 struct strbuf *msg = NULL;
3649 msg = opt->output_prefix(opt, opt->output_prefix_data);
3650 fprintf(opt->file, "%s", msg->buf);
3653 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3654 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3655 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3656 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3658 if (p->score) {
3659 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3660 inter_name_termination);
3661 } else {
3662 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3665 if (p->status == DIFF_STATUS_COPIED ||
3666 p->status == DIFF_STATUS_RENAMED) {
3667 const char *name_a, *name_b;
3668 name_a = p->one->path;
3669 name_b = p->two->path;
3670 strip_prefix(opt->prefix_length, &name_a, &name_b);
3671 write_name_quoted(name_a, opt->file, inter_name_termination);
3672 write_name_quoted(name_b, opt->file, line_termination);
3673 } else {
3674 const char *name_a, *name_b;
3675 name_a = p->one->mode ? p->one->path : p->two->path;
3676 name_b = NULL;
3677 strip_prefix(opt->prefix_length, &name_a, &name_b);
3678 write_name_quoted(name_a, opt->file, line_termination);
3682 int diff_unmodified_pair(struct diff_filepair *p)
3684 /* This function is written stricter than necessary to support
3685 * the currently implemented transformers, but the idea is to
3686 * let transformers to produce diff_filepairs any way they want,
3687 * and filter and clean them up here before producing the output.
3689 struct diff_filespec *one = p->one, *two = p->two;
3691 if (DIFF_PAIR_UNMERGED(p))
3692 return 0; /* unmerged is interesting */
3694 /* deletion, addition, mode or type change
3695 * and rename are all interesting.
3697 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3698 DIFF_PAIR_MODE_CHANGED(p) ||
3699 strcmp(one->path, two->path))
3700 return 0;
3702 /* both are valid and point at the same path. that is, we are
3703 * dealing with a change.
3705 if (one->sha1_valid && two->sha1_valid &&
3706 !hashcmp(one->sha1, two->sha1) &&
3707 !one->dirty_submodule && !two->dirty_submodule)
3708 return 1; /* no change */
3709 if (!one->sha1_valid && !two->sha1_valid)
3710 return 1; /* both look at the same file on the filesystem. */
3711 return 0;
3714 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3716 if (diff_unmodified_pair(p))
3717 return;
3719 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3720 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3721 return; /* no tree diffs in patch format */
3723 run_diff(p, o);
3726 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3727 struct diffstat_t *diffstat)
3729 if (diff_unmodified_pair(p))
3730 return;
3732 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3733 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3734 return; /* no useful stat for tree diffs */
3736 run_diffstat(p, o, diffstat);
3739 static void diff_flush_checkdiff(struct diff_filepair *p,
3740 struct diff_options *o)
3742 if (diff_unmodified_pair(p))
3743 return;
3745 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3746 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3747 return; /* nothing to check in tree diffs */
3749 run_checkdiff(p, o);
3752 int diff_queue_is_empty(void)
3754 struct diff_queue_struct *q = &diff_queued_diff;
3755 int i;
3756 for (i = 0; i < q->nr; i++)
3757 if (!diff_unmodified_pair(q->queue[i]))
3758 return 0;
3759 return 1;
3762 #if DIFF_DEBUG
3763 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3765 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3766 x, one ? one : "",
3767 s->path,
3768 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3769 s->mode,
3770 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3771 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3772 x, one ? one : "",
3773 s->size, s->xfrm_flags);
3776 void diff_debug_filepair(const struct diff_filepair *p, int i)
3778 diff_debug_filespec(p->one, i, "one");
3779 diff_debug_filespec(p->two, i, "two");
3780 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3781 p->score, p->status ? p->status : '?',
3782 p->one->rename_used, p->broken_pair);
3785 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3787 int i;
3788 if (msg)
3789 fprintf(stderr, "%s\n", msg);
3790 fprintf(stderr, "q->nr = %d\n", q->nr);
3791 for (i = 0; i < q->nr; i++) {
3792 struct diff_filepair *p = q->queue[i];
3793 diff_debug_filepair(p, i);
3796 #endif
3798 static void diff_resolve_rename_copy(void)
3800 int i;
3801 struct diff_filepair *p;
3802 struct diff_queue_struct *q = &diff_queued_diff;
3804 diff_debug_queue("resolve-rename-copy", q);
3806 for (i = 0; i < q->nr; i++) {
3807 p = q->queue[i];
3808 p->status = 0; /* undecided */
3809 if (DIFF_PAIR_UNMERGED(p))
3810 p->status = DIFF_STATUS_UNMERGED;
3811 else if (!DIFF_FILE_VALID(p->one))
3812 p->status = DIFF_STATUS_ADDED;
3813 else if (!DIFF_FILE_VALID(p->two))
3814 p->status = DIFF_STATUS_DELETED;
3815 else if (DIFF_PAIR_TYPE_CHANGED(p))
3816 p->status = DIFF_STATUS_TYPE_CHANGED;
3818 /* from this point on, we are dealing with a pair
3819 * whose both sides are valid and of the same type, i.e.
3820 * either in-place edit or rename/copy edit.
3822 else if (DIFF_PAIR_RENAME(p)) {
3824 * A rename might have re-connected a broken
3825 * pair up, causing the pathnames to be the
3826 * same again. If so, that's not a rename at
3827 * all, just a modification..
3829 * Otherwise, see if this source was used for
3830 * multiple renames, in which case we decrement
3831 * the count, and call it a copy.
3833 if (!strcmp(p->one->path, p->two->path))
3834 p->status = DIFF_STATUS_MODIFIED;
3835 else if (--p->one->rename_used > 0)
3836 p->status = DIFF_STATUS_COPIED;
3837 else
3838 p->status = DIFF_STATUS_RENAMED;
3840 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3841 p->one->mode != p->two->mode ||
3842 p->one->dirty_submodule ||
3843 p->two->dirty_submodule ||
3844 is_null_sha1(p->one->sha1))
3845 p->status = DIFF_STATUS_MODIFIED;
3846 else {
3847 /* This is a "no-change" entry and should not
3848 * happen anymore, but prepare for broken callers.
3850 error("feeding unmodified %s to diffcore",
3851 p->one->path);
3852 p->status = DIFF_STATUS_UNKNOWN;
3855 diff_debug_queue("resolve-rename-copy done", q);
3858 static int check_pair_status(struct diff_filepair *p)
3860 switch (p->status) {
3861 case DIFF_STATUS_UNKNOWN:
3862 return 0;
3863 case 0:
3864 die("internal error in diff-resolve-rename-copy");
3865 default:
3866 return 1;
3870 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3872 int fmt = opt->output_format;
3874 if (fmt & DIFF_FORMAT_CHECKDIFF)
3875 diff_flush_checkdiff(p, opt);
3876 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3877 diff_flush_raw(p, opt);
3878 else if (fmt & DIFF_FORMAT_NAME) {
3879 const char *name_a, *name_b;
3880 name_a = p->two->path;
3881 name_b = NULL;
3882 strip_prefix(opt->prefix_length, &name_a, &name_b);
3883 write_name_quoted(name_a, opt->file, opt->line_termination);
3887 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3889 if (fs->mode)
3890 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3891 else
3892 fprintf(file, " %s ", newdelete);
3893 write_name_quoted(fs->path, file, '\n');
3897 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3898 const char *line_prefix)
3900 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3901 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3902 p->two->mode, show_name ? ' ' : '\n');
3903 if (show_name) {
3904 write_name_quoted(p->two->path, file, '\n');
3909 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3910 const char *line_prefix)
3912 char *names = pprint_rename(p->one->path, p->two->path);
3914 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3915 free(names);
3916 show_mode_change(file, p, 0, line_prefix);
3919 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3921 FILE *file = opt->file;
3922 char *line_prefix = "";
3924 if (opt->output_prefix) {
3925 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3926 line_prefix = buf->buf;
3929 switch(p->status) {
3930 case DIFF_STATUS_DELETED:
3931 fputs(line_prefix, file);
3932 show_file_mode_name(file, "delete", p->one);
3933 break;
3934 case DIFF_STATUS_ADDED:
3935 fputs(line_prefix, file);
3936 show_file_mode_name(file, "create", p->two);
3937 break;
3938 case DIFF_STATUS_COPIED:
3939 fputs(line_prefix, file);
3940 show_rename_copy(file, "copy", p, line_prefix);
3941 break;
3942 case DIFF_STATUS_RENAMED:
3943 fputs(line_prefix, file);
3944 show_rename_copy(file, "rename", p, line_prefix);
3945 break;
3946 default:
3947 if (p->score) {
3948 fprintf(file, "%s rewrite ", line_prefix);
3949 write_name_quoted(p->two->path, file, ' ');
3950 fprintf(file, "(%d%%)\n", similarity_index(p));
3952 show_mode_change(file, p, !p->score, line_prefix);
3953 break;
3957 struct patch_id_t {
3958 git_SHA_CTX *ctx;
3959 int patchlen;
3962 static int remove_space(char *line, int len)
3964 int i;
3965 char *dst = line;
3966 unsigned char c;
3968 for (i = 0; i < len; i++)
3969 if (!isspace((c = line[i])))
3970 *dst++ = c;
3972 return dst - line;
3975 static void patch_id_consume(void *priv, char *line, unsigned long len)
3977 struct patch_id_t *data = priv;
3978 int new_len;
3980 /* Ignore line numbers when computing the SHA1 of the patch */
3981 if (!prefixcmp(line, "@@ -"))
3982 return;
3984 new_len = remove_space(line, len);
3986 git_SHA1_Update(data->ctx, line, new_len);
3987 data->patchlen += new_len;
3990 /* returns 0 upon success, and writes result into sha1 */
3991 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3993 struct diff_queue_struct *q = &diff_queued_diff;
3994 int i;
3995 git_SHA_CTX ctx;
3996 struct patch_id_t data;
3997 char buffer[PATH_MAX * 4 + 20];
3999 git_SHA1_Init(&ctx);
4000 memset(&data, 0, sizeof(struct patch_id_t));
4001 data.ctx = &ctx;
4003 for (i = 0; i < q->nr; i++) {
4004 xpparam_t xpp;
4005 xdemitconf_t xecfg;
4006 mmfile_t mf1, mf2;
4007 struct diff_filepair *p = q->queue[i];
4008 int len1, len2;
4010 memset(&xpp, 0, sizeof(xpp));
4011 memset(&xecfg, 0, sizeof(xecfg));
4012 if (p->status == 0)
4013 return error("internal diff status error");
4014 if (p->status == DIFF_STATUS_UNKNOWN)
4015 continue;
4016 if (diff_unmodified_pair(p))
4017 continue;
4018 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4019 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4020 continue;
4021 if (DIFF_PAIR_UNMERGED(p))
4022 continue;
4024 diff_fill_sha1_info(p->one);
4025 diff_fill_sha1_info(p->two);
4026 if (fill_mmfile(&mf1, p->one) < 0 ||
4027 fill_mmfile(&mf2, p->two) < 0)
4028 return error("unable to read files to diff");
4030 len1 = remove_space(p->one->path, strlen(p->one->path));
4031 len2 = remove_space(p->two->path, strlen(p->two->path));
4032 if (p->one->mode == 0)
4033 len1 = snprintf(buffer, sizeof(buffer),
4034 "diff--gita/%.*sb/%.*s"
4035 "newfilemode%06o"
4036 "---/dev/null"
4037 "+++b/%.*s",
4038 len1, p->one->path,
4039 len2, p->two->path,
4040 p->two->mode,
4041 len2, p->two->path);
4042 else if (p->two->mode == 0)
4043 len1 = snprintf(buffer, sizeof(buffer),
4044 "diff--gita/%.*sb/%.*s"
4045 "deletedfilemode%06o"
4046 "---a/%.*s"
4047 "+++/dev/null",
4048 len1, p->one->path,
4049 len2, p->two->path,
4050 p->one->mode,
4051 len1, p->one->path);
4052 else
4053 len1 = snprintf(buffer, sizeof(buffer),
4054 "diff--gita/%.*sb/%.*s"
4055 "---a/%.*s"
4056 "+++b/%.*s",
4057 len1, p->one->path,
4058 len2, p->two->path,
4059 len1, p->one->path,
4060 len2, p->two->path);
4061 git_SHA1_Update(&ctx, buffer, len1);
4063 if (diff_filespec_is_binary(p->one) ||
4064 diff_filespec_is_binary(p->two)) {
4065 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4066 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4067 continue;
4070 xpp.flags = 0;
4071 xecfg.ctxlen = 3;
4072 xecfg.flags = 0;
4073 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4074 &xpp, &xecfg);
4077 git_SHA1_Final(sha1, &ctx);
4078 return 0;
4081 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4083 struct diff_queue_struct *q = &diff_queued_diff;
4084 int i;
4085 int result = diff_get_patch_id(options, sha1);
4087 for (i = 0; i < q->nr; i++)
4088 diff_free_filepair(q->queue[i]);
4090 free(q->queue);
4091 DIFF_QUEUE_CLEAR(q);
4093 return result;
4096 static int is_summary_empty(const struct diff_queue_struct *q)
4098 int i;
4100 for (i = 0; i < q->nr; i++) {
4101 const struct diff_filepair *p = q->queue[i];
4103 switch (p->status) {
4104 case DIFF_STATUS_DELETED:
4105 case DIFF_STATUS_ADDED:
4106 case DIFF_STATUS_COPIED:
4107 case DIFF_STATUS_RENAMED:
4108 return 0;
4109 default:
4110 if (p->score)
4111 return 0;
4112 if (p->one->mode && p->two->mode &&
4113 p->one->mode != p->two->mode)
4114 return 0;
4115 break;
4118 return 1;
4121 static const char rename_limit_warning[] =
4122 "inexact rename detection was skipped due to too many files.";
4124 static const char degrade_cc_to_c_warning[] =
4125 "only found copies from modified paths due to too many files.";
4127 static const char rename_limit_advice[] =
4128 "you may want to set your %s variable to at least "
4129 "%d and retry the command.";
4131 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4133 if (degraded_cc)
4134 warning(degrade_cc_to_c_warning);
4135 else if (needed)
4136 warning(rename_limit_warning);
4137 else
4138 return;
4139 if (0 < needed && needed < 32767)
4140 warning(rename_limit_advice, varname, needed);
4143 void diff_flush(struct diff_options *options)
4145 struct diff_queue_struct *q = &diff_queued_diff;
4146 int i, output_format = options->output_format;
4147 int separator = 0;
4148 int dirstat_by_line = 0;
4151 * Order: raw, stat, summary, patch
4152 * or: name/name-status/checkdiff (other bits clear)
4154 if (!q->nr)
4155 goto free_queue;
4157 if (output_format & (DIFF_FORMAT_RAW |
4158 DIFF_FORMAT_NAME |
4159 DIFF_FORMAT_NAME_STATUS |
4160 DIFF_FORMAT_CHECKDIFF)) {
4161 for (i = 0; i < q->nr; i++) {
4162 struct diff_filepair *p = q->queue[i];
4163 if (check_pair_status(p))
4164 flush_one_pair(p, options);
4166 separator++;
4169 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4170 dirstat_by_line = 1;
4172 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4173 dirstat_by_line) {
4174 struct diffstat_t diffstat;
4176 memset(&diffstat, 0, sizeof(struct diffstat_t));
4177 for (i = 0; i < q->nr; i++) {
4178 struct diff_filepair *p = q->queue[i];
4179 if (check_pair_status(p))
4180 diff_flush_stat(p, options, &diffstat);
4182 if (output_format & DIFF_FORMAT_NUMSTAT)
4183 show_numstat(&diffstat, options);
4184 if (output_format & DIFF_FORMAT_DIFFSTAT)
4185 show_stats(&diffstat, options);
4186 if (output_format & DIFF_FORMAT_SHORTSTAT)
4187 show_shortstats(&diffstat, options);
4188 if (output_format & DIFF_FORMAT_DIRSTAT)
4189 show_dirstat_by_line(&diffstat, options);
4190 free_diffstat_info(&diffstat);
4191 separator++;
4193 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4194 show_dirstat(options);
4196 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4197 for (i = 0; i < q->nr; i++) {
4198 diff_summary(options, q->queue[i]);
4200 separator++;
4203 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4204 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4205 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4207 * run diff_flush_patch for the exit status. setting
4208 * options->file to /dev/null should be safe, becaue we
4209 * aren't supposed to produce any output anyway.
4211 if (options->close_file)
4212 fclose(options->file);
4213 options->file = fopen("/dev/null", "w");
4214 if (!options->file)
4215 die_errno("Could not open /dev/null");
4216 options->close_file = 1;
4217 for (i = 0; i < q->nr; i++) {
4218 struct diff_filepair *p = q->queue[i];
4219 if (check_pair_status(p))
4220 diff_flush_patch(p, options);
4221 if (options->found_changes)
4222 break;
4226 if (output_format & DIFF_FORMAT_PATCH) {
4227 if (separator) {
4228 putc(options->line_termination, options->file);
4229 if (options->stat_sep) {
4230 /* attach patch instead of inline */
4231 fputs(options->stat_sep, options->file);
4235 for (i = 0; i < q->nr; i++) {
4236 struct diff_filepair *p = q->queue[i];
4237 if (check_pair_status(p))
4238 diff_flush_patch(p, options);
4242 if (output_format & DIFF_FORMAT_CALLBACK)
4243 options->format_callback(q, options, options->format_callback_data);
4245 for (i = 0; i < q->nr; i++)
4246 diff_free_filepair(q->queue[i]);
4247 free_queue:
4248 free(q->queue);
4249 DIFF_QUEUE_CLEAR(q);
4250 if (options->close_file)
4251 fclose(options->file);
4254 * Report the content-level differences with HAS_CHANGES;
4255 * diff_addremove/diff_change does not set the bit when
4256 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4258 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4259 if (options->found_changes)
4260 DIFF_OPT_SET(options, HAS_CHANGES);
4261 else
4262 DIFF_OPT_CLR(options, HAS_CHANGES);
4266 static void diffcore_apply_filter(const char *filter)
4268 int i;
4269 struct diff_queue_struct *q = &diff_queued_diff;
4270 struct diff_queue_struct outq;
4271 DIFF_QUEUE_CLEAR(&outq);
4273 if (!filter)
4274 return;
4276 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4277 int found;
4278 for (i = found = 0; !found && i < q->nr; i++) {
4279 struct diff_filepair *p = q->queue[i];
4280 if (((p->status == DIFF_STATUS_MODIFIED) &&
4281 ((p->score &&
4282 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4283 (!p->score &&
4284 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4285 ((p->status != DIFF_STATUS_MODIFIED) &&
4286 strchr(filter, p->status)))
4287 found++;
4289 if (found)
4290 return;
4292 /* otherwise we will clear the whole queue
4293 * by copying the empty outq at the end of this
4294 * function, but first clear the current entries
4295 * in the queue.
4297 for (i = 0; i < q->nr; i++)
4298 diff_free_filepair(q->queue[i]);
4300 else {
4301 /* Only the matching ones */
4302 for (i = 0; i < q->nr; i++) {
4303 struct diff_filepair *p = q->queue[i];
4305 if (((p->status == DIFF_STATUS_MODIFIED) &&
4306 ((p->score &&
4307 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4308 (!p->score &&
4309 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4310 ((p->status != DIFF_STATUS_MODIFIED) &&
4311 strchr(filter, p->status)))
4312 diff_q(&outq, p);
4313 else
4314 diff_free_filepair(p);
4317 free(q->queue);
4318 *q = outq;
4321 /* Check whether two filespecs with the same mode and size are identical */
4322 static int diff_filespec_is_identical(struct diff_filespec *one,
4323 struct diff_filespec *two)
4325 if (S_ISGITLINK(one->mode))
4326 return 0;
4327 if (diff_populate_filespec(one, 0))
4328 return 0;
4329 if (diff_populate_filespec(two, 0))
4330 return 0;
4331 return !memcmp(one->data, two->data, one->size);
4334 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4336 int i;
4337 struct diff_queue_struct *q = &diff_queued_diff;
4338 struct diff_queue_struct outq;
4339 DIFF_QUEUE_CLEAR(&outq);
4341 for (i = 0; i < q->nr; i++) {
4342 struct diff_filepair *p = q->queue[i];
4345 * 1. Entries that come from stat info dirtiness
4346 * always have both sides (iow, not create/delete),
4347 * one side of the object name is unknown, with
4348 * the same mode and size. Keep the ones that
4349 * do not match these criteria. They have real
4350 * differences.
4352 * 2. At this point, the file is known to be modified,
4353 * with the same mode and size, and the object
4354 * name of one side is unknown. Need to inspect
4355 * the identical contents.
4357 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4358 !DIFF_FILE_VALID(p->two) ||
4359 (p->one->sha1_valid && p->two->sha1_valid) ||
4360 (p->one->mode != p->two->mode) ||
4361 diff_populate_filespec(p->one, 1) ||
4362 diff_populate_filespec(p->two, 1) ||
4363 (p->one->size != p->two->size) ||
4364 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4365 diff_q(&outq, p);
4366 else {
4368 * The caller can subtract 1 from skip_stat_unmatch
4369 * to determine how many paths were dirty only
4370 * due to stat info mismatch.
4372 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4373 diffopt->skip_stat_unmatch++;
4374 diff_free_filepair(p);
4377 free(q->queue);
4378 *q = outq;
4381 static int diffnamecmp(const void *a_, const void *b_)
4383 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4384 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4385 const char *name_a, *name_b;
4387 name_a = a->one ? a->one->path : a->two->path;
4388 name_b = b->one ? b->one->path : b->two->path;
4389 return strcmp(name_a, name_b);
4392 void diffcore_fix_diff_index(struct diff_options *options)
4394 struct diff_queue_struct *q = &diff_queued_diff;
4395 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4398 void diffcore_std(struct diff_options *options)
4400 if (options->skip_stat_unmatch)
4401 diffcore_skip_stat_unmatch(options);
4402 if (!options->found_follow) {
4403 /* See try_to_follow_renames() in tree-diff.c */
4404 if (options->break_opt != -1)
4405 diffcore_break(options->break_opt);
4406 if (options->detect_rename)
4407 diffcore_rename(options);
4408 if (options->break_opt != -1)
4409 diffcore_merge_broken();
4411 if (options->pickaxe)
4412 diffcore_pickaxe(options);
4413 if (options->orderfile)
4414 diffcore_order(options->orderfile);
4415 if (!options->found_follow)
4416 /* See try_to_follow_renames() in tree-diff.c */
4417 diff_resolve_rename_copy();
4418 diffcore_apply_filter(options->filter);
4420 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4421 DIFF_OPT_SET(options, HAS_CHANGES);
4422 else
4423 DIFF_OPT_CLR(options, HAS_CHANGES);
4425 options->found_follow = 0;
4428 int diff_result_code(struct diff_options *opt, int status)
4430 int result = 0;
4432 diff_warn_rename_limit("diff.renamelimit",
4433 opt->needed_rename_limit,
4434 opt->degraded_cc_to_c);
4435 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4436 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4437 return status;
4438 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4439 DIFF_OPT_TST(opt, HAS_CHANGES))
4440 result |= 01;
4441 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4442 DIFF_OPT_TST(opt, CHECK_FAILED))
4443 result |= 02;
4444 return result;
4447 int diff_can_quit_early(struct diff_options *opt)
4449 return (DIFF_OPT_TST(opt, QUICK) &&
4450 !opt->filter &&
4451 DIFF_OPT_TST(opt, HAS_CHANGES));
4455 * Shall changes to this submodule be ignored?
4457 * Submodule changes can be configured to be ignored separately for each path,
4458 * but that configuration can be overridden from the command line.
4460 static int is_submodule_ignored(const char *path, struct diff_options *options)
4462 int ignored = 0;
4463 unsigned orig_flags = options->flags;
4464 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4465 set_diffopt_flags_from_submodule_config(options, path);
4466 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4467 ignored = 1;
4468 options->flags = orig_flags;
4469 return ignored;
4472 void diff_addremove(struct diff_options *options,
4473 int addremove, unsigned mode,
4474 const unsigned char *sha1,
4475 const char *concatpath, unsigned dirty_submodule)
4477 struct diff_filespec *one, *two;
4479 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4480 return;
4482 /* This may look odd, but it is a preparation for
4483 * feeding "there are unchanged files which should
4484 * not produce diffs, but when you are doing copy
4485 * detection you would need them, so here they are"
4486 * entries to the diff-core. They will be prefixed
4487 * with something like '=' or '*' (I haven't decided
4488 * which but should not make any difference).
4489 * Feeding the same new and old to diff_change()
4490 * also has the same effect.
4491 * Before the final output happens, they are pruned after
4492 * merged into rename/copy pairs as appropriate.
4494 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4495 addremove = (addremove == '+' ? '-' :
4496 addremove == '-' ? '+' : addremove);
4498 if (options->prefix &&
4499 strncmp(concatpath, options->prefix, options->prefix_length))
4500 return;
4502 one = alloc_filespec(concatpath);
4503 two = alloc_filespec(concatpath);
4505 if (addremove != '+')
4506 fill_filespec(one, sha1, mode);
4507 if (addremove != '-') {
4508 fill_filespec(two, sha1, mode);
4509 two->dirty_submodule = dirty_submodule;
4512 diff_queue(&diff_queued_diff, one, two);
4513 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4514 DIFF_OPT_SET(options, HAS_CHANGES);
4517 void diff_change(struct diff_options *options,
4518 unsigned old_mode, unsigned new_mode,
4519 const unsigned char *old_sha1,
4520 const unsigned char *new_sha1,
4521 const char *concatpath,
4522 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4524 struct diff_filespec *one, *two;
4526 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4527 is_submodule_ignored(concatpath, options))
4528 return;
4530 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4531 unsigned tmp;
4532 const unsigned char *tmp_c;
4533 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4534 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4535 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4536 new_dirty_submodule = tmp;
4539 if (options->prefix &&
4540 strncmp(concatpath, options->prefix, options->prefix_length))
4541 return;
4543 one = alloc_filespec(concatpath);
4544 two = alloc_filespec(concatpath);
4545 fill_filespec(one, old_sha1, old_mode);
4546 fill_filespec(two, new_sha1, new_mode);
4547 one->dirty_submodule = old_dirty_submodule;
4548 two->dirty_submodule = new_dirty_submodule;
4550 diff_queue(&diff_queued_diff, one, two);
4551 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4552 DIFF_OPT_SET(options, HAS_CHANGES);
4555 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4557 struct diff_filepair *pair;
4558 struct diff_filespec *one, *two;
4560 if (options->prefix &&
4561 strncmp(path, options->prefix, options->prefix_length))
4562 return NULL;
4564 one = alloc_filespec(path);
4565 two = alloc_filespec(path);
4566 pair = diff_queue(&diff_queued_diff, one, two);
4567 pair->is_unmerged = 1;
4568 return pair;
4571 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4572 size_t *outsize)
4574 struct diff_tempfile *temp;
4575 const char *argv[3];
4576 const char **arg = argv;
4577 struct child_process child;
4578 struct strbuf buf = STRBUF_INIT;
4579 int err = 0;
4581 temp = prepare_temp_file(spec->path, spec);
4582 *arg++ = pgm;
4583 *arg++ = temp->name;
4584 *arg = NULL;
4586 memset(&child, 0, sizeof(child));
4587 child.use_shell = 1;
4588 child.argv = argv;
4589 child.out = -1;
4590 if (start_command(&child)) {
4591 remove_tempfile();
4592 return NULL;
4595 if (strbuf_read(&buf, child.out, 0) < 0)
4596 err = error("error reading from textconv command '%s'", pgm);
4597 close(child.out);
4599 if (finish_command(&child) || err) {
4600 strbuf_release(&buf);
4601 remove_tempfile();
4602 return NULL;
4604 remove_tempfile();
4606 return strbuf_detach(&buf, outsize);
4609 size_t fill_textconv(struct userdiff_driver *driver,
4610 struct diff_filespec *df,
4611 char **outbuf)
4613 size_t size;
4615 if (!driver || !driver->textconv) {
4616 if (!DIFF_FILE_VALID(df)) {
4617 *outbuf = "";
4618 return 0;
4620 if (diff_populate_filespec(df, 0))
4621 die("unable to read files to diff");
4622 *outbuf = df->data;
4623 return df->size;
4626 if (driver->textconv_cache && df->sha1_valid) {
4627 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4628 &size);
4629 if (*outbuf)
4630 return size;
4633 *outbuf = run_textconv(driver->textconv, df, &size);
4634 if (!*outbuf)
4635 die("unable to read files to diff");
4637 if (driver->textconv_cache && df->sha1_valid) {
4638 /* ignore errors, as we might be in a readonly repository */
4639 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4640 size);
4642 * we could save up changes and flush them all at the end,
4643 * but we would need an extra call after all diffing is done.
4644 * Since generating a cache entry is the slow path anyway,
4645 * this extra overhead probably isn't a big deal.
4647 notes_cache_write(driver->textconv_cache);
4650 return size;