reset: the command takes committish
[git/jnareb-git.git] / diff.c
blob7e154265f778c645192cbf17c65b9bea2a507402
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);
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 if (git_color_config(var, value, cb) < 0)
168 return -1;
170 return git_diff_basic_config(var, value, cb);
173 int git_diff_basic_config(const char *var, const char *value, void *cb)
175 if (!strcmp(var, "diff.renamelimit")) {
176 diff_rename_limit_default = git_config_int(var, value);
177 return 0;
180 switch (userdiff_config(var, value)) {
181 case 0: break;
182 case -1: return -1;
183 default: return 0;
186 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
187 int slot = parse_diff_color_slot(var, 11);
188 if (slot < 0)
189 return 0;
190 if (!value)
191 return config_error_nonbool(var);
192 color_parse(value, var, diff_colors[slot]);
193 return 0;
196 /* like GNU diff's --suppress-blank-empty option */
197 if (!strcmp(var, "diff.suppressblankempty") ||
198 /* for backwards compatibility */
199 !strcmp(var, "diff.suppress-blank-empty")) {
200 diff_suppress_blank_empty = git_config_bool(var, value);
201 return 0;
204 if (!strcmp(var, "diff.dirstat")) {
205 struct strbuf errmsg = STRBUF_INIT;
206 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
207 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
208 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
209 errmsg.buf);
210 strbuf_release(&errmsg);
211 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
212 return 0;
215 if (!prefixcmp(var, "submodule."))
216 return parse_submodule_config_option(var, value);
218 return git_default_config(var, value, cb);
221 static char *quote_two(const char *one, const char *two)
223 int need_one = quote_c_style(one, NULL, NULL, 1);
224 int need_two = quote_c_style(two, NULL, NULL, 1);
225 struct strbuf res = STRBUF_INIT;
227 if (need_one + need_two) {
228 strbuf_addch(&res, '"');
229 quote_c_style(one, &res, NULL, 1);
230 quote_c_style(two, &res, NULL, 1);
231 strbuf_addch(&res, '"');
232 } else {
233 strbuf_addstr(&res, one);
234 strbuf_addstr(&res, two);
236 return strbuf_detach(&res, NULL);
239 static const char *external_diff(void)
241 static const char *external_diff_cmd = NULL;
242 static int done_preparing = 0;
244 if (done_preparing)
245 return external_diff_cmd;
246 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
247 if (!external_diff_cmd)
248 external_diff_cmd = external_diff_cmd_cfg;
249 done_preparing = 1;
250 return external_diff_cmd;
253 static struct diff_tempfile {
254 const char *name; /* filename external diff should read from */
255 char hex[41];
256 char mode[10];
257 char tmp_path[PATH_MAX];
258 } diff_temp[2];
260 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
262 struct emit_callback {
263 int color_diff;
264 unsigned ws_rule;
265 int blank_at_eof_in_preimage;
266 int blank_at_eof_in_postimage;
267 int lno_in_preimage;
268 int lno_in_postimage;
269 sane_truncate_fn truncate;
270 const char **label_path;
271 struct diff_words_data *diff_words;
272 struct diff_options *opt;
273 int *found_changesp;
274 struct strbuf *header;
277 static int count_lines(const char *data, int size)
279 int count, ch, completely_empty = 1, nl_just_seen = 0;
280 count = 0;
281 while (0 < size--) {
282 ch = *data++;
283 if (ch == '\n') {
284 count++;
285 nl_just_seen = 1;
286 completely_empty = 0;
288 else {
289 nl_just_seen = 0;
290 completely_empty = 0;
293 if (completely_empty)
294 return 0;
295 if (!nl_just_seen)
296 count++; /* no trailing newline */
297 return count;
300 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
302 if (!DIFF_FILE_VALID(one)) {
303 mf->ptr = (char *)""; /* does not matter */
304 mf->size = 0;
305 return 0;
307 else if (diff_populate_filespec(one, 0))
308 return -1;
310 mf->ptr = one->data;
311 mf->size = one->size;
312 return 0;
315 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
316 static unsigned long diff_filespec_size(struct diff_filespec *one)
318 if (!DIFF_FILE_VALID(one))
319 return 0;
320 diff_populate_filespec(one, 1);
321 return one->size;
324 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
326 char *ptr = mf->ptr;
327 long size = mf->size;
328 int cnt = 0;
330 if (!size)
331 return cnt;
332 ptr += size - 1; /* pointing at the very end */
333 if (*ptr != '\n')
334 ; /* incomplete line */
335 else
336 ptr--; /* skip the last LF */
337 while (mf->ptr < ptr) {
338 char *prev_eol;
339 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
340 if (*prev_eol == '\n')
341 break;
342 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
343 break;
344 cnt++;
345 ptr = prev_eol - 1;
347 return cnt;
350 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
351 struct emit_callback *ecbdata)
353 int l1, l2, at;
354 unsigned ws_rule = ecbdata->ws_rule;
355 l1 = count_trailing_blank(mf1, ws_rule);
356 l2 = count_trailing_blank(mf2, ws_rule);
357 if (l2 <= l1) {
358 ecbdata->blank_at_eof_in_preimage = 0;
359 ecbdata->blank_at_eof_in_postimage = 0;
360 return;
362 at = count_lines(mf1->ptr, mf1->size);
363 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
365 at = count_lines(mf2->ptr, mf2->size);
366 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
369 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
370 int first, const char *line, int len)
372 int has_trailing_newline, has_trailing_carriage_return;
373 int nofirst;
374 FILE *file = o->file;
376 if (o->output_prefix) {
377 struct strbuf *msg = NULL;
378 msg = o->output_prefix(o, o->output_prefix_data);
379 assert(msg);
380 fwrite(msg->buf, msg->len, 1, file);
383 if (len == 0) {
384 has_trailing_newline = (first == '\n');
385 has_trailing_carriage_return = (!has_trailing_newline &&
386 (first == '\r'));
387 nofirst = has_trailing_newline || has_trailing_carriage_return;
388 } else {
389 has_trailing_newline = (len > 0 && line[len-1] == '\n');
390 if (has_trailing_newline)
391 len--;
392 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
393 if (has_trailing_carriage_return)
394 len--;
395 nofirst = 0;
398 if (len || !nofirst) {
399 fputs(set, file);
400 if (!nofirst)
401 fputc(first, file);
402 fwrite(line, len, 1, file);
403 fputs(reset, file);
405 if (has_trailing_carriage_return)
406 fputc('\r', file);
407 if (has_trailing_newline)
408 fputc('\n', file);
411 static void emit_line(struct diff_options *o, const char *set, const char *reset,
412 const char *line, int len)
414 emit_line_0(o, set, reset, line[0], line+1, len-1);
417 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
419 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
420 ecbdata->blank_at_eof_in_preimage &&
421 ecbdata->blank_at_eof_in_postimage &&
422 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
423 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
424 return 0;
425 return ws_blank_line(line, len, ecbdata->ws_rule);
428 static void emit_add_line(const char *reset,
429 struct emit_callback *ecbdata,
430 const char *line, int len)
432 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
433 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
435 if (!*ws)
436 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
437 else if (new_blank_line_at_eof(ecbdata, line, len))
438 /* Blank line at EOF - paint '+' as well */
439 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
440 else {
441 /* Emit just the prefix, then the rest. */
442 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
443 ws_check_emit(line, len, ecbdata->ws_rule,
444 ecbdata->opt->file, set, reset, ws);
448 static void emit_hunk_header(struct emit_callback *ecbdata,
449 const char *line, int len)
451 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
452 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
453 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
454 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
455 static const char atat[2] = { '@', '@' };
456 const char *cp, *ep;
457 struct strbuf msgbuf = STRBUF_INIT;
458 int org_len = len;
459 int i = 1;
462 * As a hunk header must begin with "@@ -<old>, +<new> @@",
463 * it always is at least 10 bytes long.
465 if (len < 10 ||
466 memcmp(line, atat, 2) ||
467 !(ep = memmem(line + 2, len - 2, atat, 2))) {
468 emit_line(ecbdata->opt, plain, reset, line, len);
469 return;
471 ep += 2; /* skip over @@ */
473 /* The hunk header in fraginfo color */
474 strbuf_add(&msgbuf, frag, strlen(frag));
475 strbuf_add(&msgbuf, line, ep - line);
476 strbuf_add(&msgbuf, reset, strlen(reset));
479 * trailing "\r\n"
481 for ( ; i < 3; i++)
482 if (line[len - i] == '\r' || line[len - i] == '\n')
483 len--;
485 /* blank before the func header */
486 for (cp = ep; ep - line < len; ep++)
487 if (*ep != ' ' && *ep != '\t')
488 break;
489 if (ep != cp) {
490 strbuf_add(&msgbuf, plain, strlen(plain));
491 strbuf_add(&msgbuf, cp, ep - cp);
492 strbuf_add(&msgbuf, reset, strlen(reset));
495 if (ep < line + len) {
496 strbuf_add(&msgbuf, func, strlen(func));
497 strbuf_add(&msgbuf, ep, line + len - ep);
498 strbuf_add(&msgbuf, reset, strlen(reset));
501 strbuf_add(&msgbuf, line + len, org_len - len);
502 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
503 strbuf_release(&msgbuf);
506 static struct diff_tempfile *claim_diff_tempfile(void) {
507 int i;
508 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
509 if (!diff_temp[i].name)
510 return diff_temp + i;
511 die("BUG: diff is failing to clean up its tempfiles");
514 static int remove_tempfile_installed;
516 static void remove_tempfile(void)
518 int i;
519 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
520 if (diff_temp[i].name == diff_temp[i].tmp_path)
521 unlink_or_warn(diff_temp[i].name);
522 diff_temp[i].name = NULL;
526 static void remove_tempfile_on_signal(int signo)
528 remove_tempfile();
529 sigchain_pop(signo);
530 raise(signo);
533 static void print_line_count(FILE *file, int count)
535 switch (count) {
536 case 0:
537 fprintf(file, "0,0");
538 break;
539 case 1:
540 fprintf(file, "1");
541 break;
542 default:
543 fprintf(file, "1,%d", count);
544 break;
548 static void emit_rewrite_lines(struct emit_callback *ecb,
549 int prefix, const char *data, int size)
551 const char *endp = NULL;
552 static const char *nneof = " No newline at end of file\n";
553 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
554 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
556 while (0 < size) {
557 int len;
559 endp = memchr(data, '\n', size);
560 len = endp ? (endp - data + 1) : size;
561 if (prefix != '+') {
562 ecb->lno_in_preimage++;
563 emit_line_0(ecb->opt, old, reset, '-',
564 data, len);
565 } else {
566 ecb->lno_in_postimage++;
567 emit_add_line(reset, ecb, data, len);
569 size -= len;
570 data += len;
572 if (!endp) {
573 const char *plain = diff_get_color(ecb->color_diff,
574 DIFF_PLAIN);
575 emit_line_0(ecb->opt, plain, reset, '\\',
576 nneof, strlen(nneof));
580 static void emit_rewrite_diff(const char *name_a,
581 const char *name_b,
582 struct diff_filespec *one,
583 struct diff_filespec *two,
584 struct userdiff_driver *textconv_one,
585 struct userdiff_driver *textconv_two,
586 struct diff_options *o)
588 int lc_a, lc_b;
589 const char *name_a_tab, *name_b_tab;
590 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
591 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
592 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
593 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
594 const char *a_prefix, *b_prefix;
595 char *data_one, *data_two;
596 size_t size_one, size_two;
597 struct emit_callback ecbdata;
598 char *line_prefix = "";
599 struct strbuf *msgbuf;
601 if (o && o->output_prefix) {
602 msgbuf = o->output_prefix(o, o->output_prefix_data);
603 line_prefix = msgbuf->buf;
606 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
607 a_prefix = o->b_prefix;
608 b_prefix = o->a_prefix;
609 } else {
610 a_prefix = o->a_prefix;
611 b_prefix = o->b_prefix;
614 name_a += (*name_a == '/');
615 name_b += (*name_b == '/');
616 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
617 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
619 strbuf_reset(&a_name);
620 strbuf_reset(&b_name);
621 quote_two_c_style(&a_name, a_prefix, name_a, 0);
622 quote_two_c_style(&b_name, b_prefix, name_b, 0);
624 size_one = fill_textconv(textconv_one, one, &data_one);
625 size_two = fill_textconv(textconv_two, two, &data_two);
627 memset(&ecbdata, 0, sizeof(ecbdata));
628 ecbdata.color_diff = want_color(o->use_color);
629 ecbdata.found_changesp = &o->found_changes;
630 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
631 ecbdata.opt = o;
632 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
633 mmfile_t mf1, mf2;
634 mf1.ptr = (char *)data_one;
635 mf2.ptr = (char *)data_two;
636 mf1.size = size_one;
637 mf2.size = size_two;
638 check_blank_at_eof(&mf1, &mf2, &ecbdata);
640 ecbdata.lno_in_preimage = 1;
641 ecbdata.lno_in_postimage = 1;
643 lc_a = count_lines(data_one, size_one);
644 lc_b = count_lines(data_two, size_two);
645 fprintf(o->file,
646 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
647 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
648 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
649 line_prefix, fraginfo);
650 if (!o->irreversible_delete)
651 print_line_count(o->file, lc_a);
652 else
653 fprintf(o->file, "?,?");
654 fprintf(o->file, " +");
655 print_line_count(o->file, lc_b);
656 fprintf(o->file, " @@%s\n", reset);
657 if (lc_a && !o->irreversible_delete)
658 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
659 if (lc_b)
660 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
661 if (textconv_one)
662 free((char *)data_one);
663 if (textconv_two)
664 free((char *)data_two);
667 struct diff_words_buffer {
668 mmfile_t text;
669 long alloc;
670 struct diff_words_orig {
671 const char *begin, *end;
672 } *orig;
673 int orig_nr, orig_alloc;
676 static void diff_words_append(char *line, unsigned long len,
677 struct diff_words_buffer *buffer)
679 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
680 line++;
681 len--;
682 memcpy(buffer->text.ptr + buffer->text.size, line, len);
683 buffer->text.size += len;
684 buffer->text.ptr[buffer->text.size] = '\0';
687 struct diff_words_style_elem {
688 const char *prefix;
689 const char *suffix;
690 const char *color; /* NULL; filled in by the setup code if
691 * color is enabled */
694 struct diff_words_style {
695 enum diff_words_type type;
696 struct diff_words_style_elem new, old, ctx;
697 const char *newline;
700 static struct diff_words_style diff_words_styles[] = {
701 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
702 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
703 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
706 struct diff_words_data {
707 struct diff_words_buffer minus, plus;
708 const char *current_plus;
709 int last_minus;
710 struct diff_options *opt;
711 regex_t *word_regex;
712 enum diff_words_type type;
713 struct diff_words_style *style;
716 static int fn_out_diff_words_write_helper(FILE *fp,
717 struct diff_words_style_elem *st_el,
718 const char *newline,
719 size_t count, const char *buf,
720 const char *line_prefix)
722 int print = 0;
724 while (count) {
725 char *p = memchr(buf, '\n', count);
726 if (print)
727 fputs(line_prefix, fp);
728 if (p != buf) {
729 if (st_el->color && fputs(st_el->color, fp) < 0)
730 return -1;
731 if (fputs(st_el->prefix, fp) < 0 ||
732 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
733 fputs(st_el->suffix, fp) < 0)
734 return -1;
735 if (st_el->color && *st_el->color
736 && fputs(GIT_COLOR_RESET, fp) < 0)
737 return -1;
739 if (!p)
740 return 0;
741 if (fputs(newline, fp) < 0)
742 return -1;
743 count -= p + 1 - buf;
744 buf = p + 1;
745 print = 1;
747 return 0;
751 * '--color-words' algorithm can be described as:
753 * 1. collect a the minus/plus lines of a diff hunk, divided into
754 * minus-lines and plus-lines;
756 * 2. break both minus-lines and plus-lines into words and
757 * place them into two mmfile_t with one word for each line;
759 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
761 * And for the common parts of the both file, we output the plus side text.
762 * diff_words->current_plus is used to trace the current position of the plus file
763 * which printed. diff_words->last_minus is used to trace the last minus word
764 * printed.
766 * For '--graph' to work with '--color-words', we need to output the graph prefix
767 * on each line of color words output. Generally, there are two conditions on
768 * which we should output the prefix.
770 * 1. diff_words->last_minus == 0 &&
771 * diff_words->current_plus == diff_words->plus.text.ptr
773 * that is: the plus text must start as a new line, and if there is no minus
774 * word printed, a graph prefix must be printed.
776 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
777 * *(diff_words->current_plus - 1) == '\n'
779 * that is: a graph prefix must be printed following a '\n'
781 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
783 if ((diff_words->last_minus == 0 &&
784 diff_words->current_plus == diff_words->plus.text.ptr) ||
785 (diff_words->current_plus > diff_words->plus.text.ptr &&
786 *(diff_words->current_plus - 1) == '\n')) {
787 return 1;
788 } else {
789 return 0;
793 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
795 struct diff_words_data *diff_words = priv;
796 struct diff_words_style *style = diff_words->style;
797 int minus_first, minus_len, plus_first, plus_len;
798 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
799 struct diff_options *opt = diff_words->opt;
800 struct strbuf *msgbuf;
801 char *line_prefix = "";
803 if (line[0] != '@' || parse_hunk_header(line, len,
804 &minus_first, &minus_len, &plus_first, &plus_len))
805 return;
807 assert(opt);
808 if (opt->output_prefix) {
809 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
810 line_prefix = msgbuf->buf;
813 /* POSIX requires that first be decremented by one if len == 0... */
814 if (minus_len) {
815 minus_begin = diff_words->minus.orig[minus_first].begin;
816 minus_end =
817 diff_words->minus.orig[minus_first + minus_len - 1].end;
818 } else
819 minus_begin = minus_end =
820 diff_words->minus.orig[minus_first].end;
822 if (plus_len) {
823 plus_begin = diff_words->plus.orig[plus_first].begin;
824 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
825 } else
826 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
828 if (color_words_output_graph_prefix(diff_words)) {
829 fputs(line_prefix, diff_words->opt->file);
831 if (diff_words->current_plus != plus_begin) {
832 fn_out_diff_words_write_helper(diff_words->opt->file,
833 &style->ctx, style->newline,
834 plus_begin - diff_words->current_plus,
835 diff_words->current_plus, line_prefix);
836 if (*(plus_begin - 1) == '\n')
837 fputs(line_prefix, diff_words->opt->file);
839 if (minus_begin != minus_end) {
840 fn_out_diff_words_write_helper(diff_words->opt->file,
841 &style->old, style->newline,
842 minus_end - minus_begin, minus_begin,
843 line_prefix);
845 if (plus_begin != plus_end) {
846 fn_out_diff_words_write_helper(diff_words->opt->file,
847 &style->new, style->newline,
848 plus_end - plus_begin, plus_begin,
849 line_prefix);
852 diff_words->current_plus = plus_end;
853 diff_words->last_minus = minus_first;
856 /* This function starts looking at *begin, and returns 0 iff a word was found. */
857 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
858 int *begin, int *end)
860 if (word_regex && *begin < buffer->size) {
861 regmatch_t match[1];
862 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
863 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
864 '\n', match[0].rm_eo - match[0].rm_so);
865 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
866 *begin += match[0].rm_so;
867 return *begin >= *end;
869 return -1;
872 /* find the next word */
873 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
874 (*begin)++;
875 if (*begin >= buffer->size)
876 return -1;
878 /* find the end of the word */
879 *end = *begin + 1;
880 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
881 (*end)++;
883 return 0;
887 * This function splits the words in buffer->text, stores the list with
888 * newline separator into out, and saves the offsets of the original words
889 * in buffer->orig.
891 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
892 regex_t *word_regex)
894 int i, j;
895 long alloc = 0;
897 out->size = 0;
898 out->ptr = NULL;
900 /* fake an empty "0th" word */
901 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
902 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
903 buffer->orig_nr = 1;
905 for (i = 0; i < buffer->text.size; i++) {
906 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
907 return;
909 /* store original boundaries */
910 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
911 buffer->orig_alloc);
912 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
913 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
914 buffer->orig_nr++;
916 /* store one word */
917 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
918 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
919 out->ptr[out->size + j - i] = '\n';
920 out->size += j - i + 1;
922 i = j - 1;
926 /* this executes the word diff on the accumulated buffers */
927 static void diff_words_show(struct diff_words_data *diff_words)
929 xpparam_t xpp;
930 xdemitconf_t xecfg;
931 mmfile_t minus, plus;
932 struct diff_words_style *style = diff_words->style;
934 struct diff_options *opt = diff_words->opt;
935 struct strbuf *msgbuf;
936 char *line_prefix = "";
938 assert(opt);
939 if (opt->output_prefix) {
940 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
941 line_prefix = msgbuf->buf;
944 /* special case: only removal */
945 if (!diff_words->plus.text.size) {
946 fputs(line_prefix, diff_words->opt->file);
947 fn_out_diff_words_write_helper(diff_words->opt->file,
948 &style->old, style->newline,
949 diff_words->minus.text.size,
950 diff_words->minus.text.ptr, line_prefix);
951 diff_words->minus.text.size = 0;
952 return;
955 diff_words->current_plus = diff_words->plus.text.ptr;
956 diff_words->last_minus = 0;
958 memset(&xpp, 0, sizeof(xpp));
959 memset(&xecfg, 0, sizeof(xecfg));
960 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
961 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
962 xpp.flags = 0;
963 /* as only the hunk header will be parsed, we need a 0-context */
964 xecfg.ctxlen = 0;
965 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
966 &xpp, &xecfg);
967 free(minus.ptr);
968 free(plus.ptr);
969 if (diff_words->current_plus != diff_words->plus.text.ptr +
970 diff_words->plus.text.size) {
971 if (color_words_output_graph_prefix(diff_words))
972 fputs(line_prefix, diff_words->opt->file);
973 fn_out_diff_words_write_helper(diff_words->opt->file,
974 &style->ctx, style->newline,
975 diff_words->plus.text.ptr + diff_words->plus.text.size
976 - diff_words->current_plus, diff_words->current_plus,
977 line_prefix);
979 diff_words->minus.text.size = diff_words->plus.text.size = 0;
982 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
983 static void diff_words_flush(struct emit_callback *ecbdata)
985 if (ecbdata->diff_words->minus.text.size ||
986 ecbdata->diff_words->plus.text.size)
987 diff_words_show(ecbdata->diff_words);
990 static void free_diff_words_data(struct emit_callback *ecbdata)
992 if (ecbdata->diff_words) {
993 diff_words_flush(ecbdata);
994 free (ecbdata->diff_words->minus.text.ptr);
995 free (ecbdata->diff_words->minus.orig);
996 free (ecbdata->diff_words->plus.text.ptr);
997 free (ecbdata->diff_words->plus.orig);
998 if (ecbdata->diff_words->word_regex) {
999 regfree(ecbdata->diff_words->word_regex);
1000 free(ecbdata->diff_words->word_regex);
1002 free(ecbdata->diff_words);
1003 ecbdata->diff_words = NULL;
1007 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1009 if (want_color(diff_use_color))
1010 return diff_colors[ix];
1011 return "";
1014 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1016 const char *cp;
1017 unsigned long allot;
1018 size_t l = len;
1020 if (ecb->truncate)
1021 return ecb->truncate(line, len);
1022 cp = line;
1023 allot = l;
1024 while (0 < l) {
1025 (void) utf8_width(&cp, &l);
1026 if (!cp)
1027 break; /* truncated in the middle? */
1029 return allot - l;
1032 static void find_lno(const char *line, struct emit_callback *ecbdata)
1034 const char *p;
1035 ecbdata->lno_in_preimage = 0;
1036 ecbdata->lno_in_postimage = 0;
1037 p = strchr(line, '-');
1038 if (!p)
1039 return; /* cannot happen */
1040 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1041 p = strchr(p, '+');
1042 if (!p)
1043 return; /* cannot happen */
1044 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1047 static void fn_out_consume(void *priv, char *line, unsigned long len)
1049 struct emit_callback *ecbdata = priv;
1050 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1051 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1052 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1053 struct diff_options *o = ecbdata->opt;
1054 char *line_prefix = "";
1055 struct strbuf *msgbuf;
1057 if (o && o->output_prefix) {
1058 msgbuf = o->output_prefix(o, o->output_prefix_data);
1059 line_prefix = msgbuf->buf;
1062 if (ecbdata->header) {
1063 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1064 strbuf_reset(ecbdata->header);
1065 ecbdata->header = NULL;
1067 *(ecbdata->found_changesp) = 1;
1069 if (ecbdata->label_path[0]) {
1070 const char *name_a_tab, *name_b_tab;
1072 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1073 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1075 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1076 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1077 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1078 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1079 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1082 if (diff_suppress_blank_empty
1083 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1084 line[0] = '\n';
1085 len = 1;
1088 if (line[0] == '@') {
1089 if (ecbdata->diff_words)
1090 diff_words_flush(ecbdata);
1091 len = sane_truncate_line(ecbdata, line, len);
1092 find_lno(line, ecbdata);
1093 emit_hunk_header(ecbdata, line, len);
1094 if (line[len-1] != '\n')
1095 putc('\n', ecbdata->opt->file);
1096 return;
1099 if (len < 1) {
1100 emit_line(ecbdata->opt, reset, reset, line, len);
1101 if (ecbdata->diff_words
1102 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1103 fputs("~\n", ecbdata->opt->file);
1104 return;
1107 if (ecbdata->diff_words) {
1108 if (line[0] == '-') {
1109 diff_words_append(line, len,
1110 &ecbdata->diff_words->minus);
1111 return;
1112 } else if (line[0] == '+') {
1113 diff_words_append(line, len,
1114 &ecbdata->diff_words->plus);
1115 return;
1116 } else if (!prefixcmp(line, "\\ ")) {
1118 * Eat the "no newline at eof" marker as if we
1119 * saw a "+" or "-" line with nothing on it,
1120 * and return without diff_words_flush() to
1121 * defer processing. If this is the end of
1122 * preimage, more "+" lines may come after it.
1124 return;
1126 diff_words_flush(ecbdata);
1127 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1128 emit_line(ecbdata->opt, plain, reset, line, len);
1129 fputs("~\n", ecbdata->opt->file);
1130 } else {
1132 * Skip the prefix character, if any. With
1133 * diff_suppress_blank_empty, there may be
1134 * none.
1136 if (line[0] != '\n') {
1137 line++;
1138 len--;
1140 emit_line(ecbdata->opt, plain, reset, line, len);
1142 return;
1145 if (line[0] != '+') {
1146 const char *color =
1147 diff_get_color(ecbdata->color_diff,
1148 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1149 ecbdata->lno_in_preimage++;
1150 if (line[0] == ' ')
1151 ecbdata->lno_in_postimage++;
1152 emit_line(ecbdata->opt, color, reset, line, len);
1153 } else {
1154 ecbdata->lno_in_postimage++;
1155 emit_add_line(reset, ecbdata, line + 1, len - 1);
1159 static char *pprint_rename(const char *a, const char *b)
1161 const char *old = a;
1162 const char *new = b;
1163 struct strbuf name = STRBUF_INIT;
1164 int pfx_length, sfx_length;
1165 int len_a = strlen(a);
1166 int len_b = strlen(b);
1167 int a_midlen, b_midlen;
1168 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1169 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1171 if (qlen_a || qlen_b) {
1172 quote_c_style(a, &name, NULL, 0);
1173 strbuf_addstr(&name, " => ");
1174 quote_c_style(b, &name, NULL, 0);
1175 return strbuf_detach(&name, NULL);
1178 /* Find common prefix */
1179 pfx_length = 0;
1180 while (*old && *new && *old == *new) {
1181 if (*old == '/')
1182 pfx_length = old - a + 1;
1183 old++;
1184 new++;
1187 /* Find common suffix */
1188 old = a + len_a;
1189 new = b + len_b;
1190 sfx_length = 0;
1191 while (a <= old && b <= new && *old == *new) {
1192 if (*old == '/')
1193 sfx_length = len_a - (old - a);
1194 old--;
1195 new--;
1199 * pfx{mid-a => mid-b}sfx
1200 * {pfx-a => pfx-b}sfx
1201 * pfx{sfx-a => sfx-b}
1202 * name-a => name-b
1204 a_midlen = len_a - pfx_length - sfx_length;
1205 b_midlen = len_b - pfx_length - sfx_length;
1206 if (a_midlen < 0)
1207 a_midlen = 0;
1208 if (b_midlen < 0)
1209 b_midlen = 0;
1211 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1212 if (pfx_length + sfx_length) {
1213 strbuf_add(&name, a, pfx_length);
1214 strbuf_addch(&name, '{');
1216 strbuf_add(&name, a + pfx_length, a_midlen);
1217 strbuf_addstr(&name, " => ");
1218 strbuf_add(&name, b + pfx_length, b_midlen);
1219 if (pfx_length + sfx_length) {
1220 strbuf_addch(&name, '}');
1221 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1223 return strbuf_detach(&name, NULL);
1226 struct diffstat_t {
1227 int nr;
1228 int alloc;
1229 struct diffstat_file {
1230 char *from_name;
1231 char *name;
1232 char *print_name;
1233 unsigned is_unmerged:1;
1234 unsigned is_binary:1;
1235 unsigned is_renamed:1;
1236 uintmax_t added, deleted;
1237 } **files;
1240 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1241 const char *name_a,
1242 const char *name_b)
1244 struct diffstat_file *x;
1245 x = xcalloc(sizeof (*x), 1);
1246 if (diffstat->nr == diffstat->alloc) {
1247 diffstat->alloc = alloc_nr(diffstat->alloc);
1248 diffstat->files = xrealloc(diffstat->files,
1249 diffstat->alloc * sizeof(x));
1251 diffstat->files[diffstat->nr++] = x;
1252 if (name_b) {
1253 x->from_name = xstrdup(name_a);
1254 x->name = xstrdup(name_b);
1255 x->is_renamed = 1;
1257 else {
1258 x->from_name = NULL;
1259 x->name = xstrdup(name_a);
1261 return x;
1264 static void diffstat_consume(void *priv, char *line, unsigned long len)
1266 struct diffstat_t *diffstat = priv;
1267 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1269 if (line[0] == '+')
1270 x->added++;
1271 else if (line[0] == '-')
1272 x->deleted++;
1275 const char mime_boundary_leader[] = "------------";
1277 static int scale_linear(int it, int width, int max_change)
1280 * make sure that at least one '-' is printed if there were deletions,
1281 * and likewise for '+'.
1283 if (max_change < 2)
1284 return it;
1285 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1288 static void show_name(FILE *file,
1289 const char *prefix, const char *name, int len)
1291 fprintf(file, " %s%-*s |", prefix, len, name);
1294 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1296 if (cnt <= 0)
1297 return;
1298 fprintf(file, "%s", set);
1299 while (cnt--)
1300 putc(ch, file);
1301 fprintf(file, "%s", reset);
1304 static void fill_print_name(struct diffstat_file *file)
1306 char *pname;
1308 if (file->print_name)
1309 return;
1311 if (!file->is_renamed) {
1312 struct strbuf buf = STRBUF_INIT;
1313 if (quote_c_style(file->name, &buf, NULL, 0)) {
1314 pname = strbuf_detach(&buf, NULL);
1315 } else {
1316 pname = file->name;
1317 strbuf_release(&buf);
1319 } else {
1320 pname = pprint_rename(file->from_name, file->name);
1322 file->print_name = pname;
1325 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1327 int i, len, add, del, adds = 0, dels = 0;
1328 uintmax_t max_change = 0, max_len = 0;
1329 int total_files = data->nr;
1330 int width, name_width, count;
1331 const char *reset, *add_c, *del_c;
1332 const char *line_prefix = "";
1333 int extra_shown = 0;
1334 struct strbuf *msg = NULL;
1336 if (data->nr == 0)
1337 return;
1339 if (options->output_prefix) {
1340 msg = options->output_prefix(options, options->output_prefix_data);
1341 line_prefix = msg->buf;
1344 width = options->stat_width ? options->stat_width : 80;
1345 name_width = options->stat_name_width ? options->stat_name_width : 50;
1346 count = options->stat_count ? options->stat_count : data->nr;
1348 /* Sanity: give at least 5 columns to the graph,
1349 * but leave at least 10 columns for the name.
1351 if (width < 25)
1352 width = 25;
1353 if (name_width < 10)
1354 name_width = 10;
1355 else if (width < name_width + 15)
1356 name_width = width - 15;
1358 /* Find the longest filename and max number of changes */
1359 reset = diff_get_color_opt(options, DIFF_RESET);
1360 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1361 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1363 for (i = 0; (i < count) && (i < data->nr); i++) {
1364 struct diffstat_file *file = data->files[i];
1365 uintmax_t change = file->added + file->deleted;
1366 if (!data->files[i]->is_renamed &&
1367 (change == 0)) {
1368 count++; /* not shown == room for one more */
1369 continue;
1371 fill_print_name(file);
1372 len = strlen(file->print_name);
1373 if (max_len < len)
1374 max_len = len;
1376 if (file->is_binary || file->is_unmerged)
1377 continue;
1378 if (max_change < change)
1379 max_change = change;
1381 count = i; /* min(count, data->nr) */
1383 /* Compute the width of the graph part;
1384 * 10 is for one blank at the beginning of the line plus
1385 * " | count " between the name and the graph.
1387 * From here on, name_width is the width of the name area,
1388 * and width is the width of the graph area.
1390 name_width = (name_width < max_len) ? name_width : max_len;
1391 if (width < (name_width + 10) + max_change)
1392 width = width - (name_width + 10);
1393 else
1394 width = max_change;
1396 for (i = 0; i < count; i++) {
1397 const char *prefix = "";
1398 char *name = data->files[i]->print_name;
1399 uintmax_t added = data->files[i]->added;
1400 uintmax_t deleted = data->files[i]->deleted;
1401 int name_len;
1403 if (!data->files[i]->is_renamed &&
1404 (added + deleted == 0)) {
1405 total_files--;
1406 continue;
1409 * "scale" the filename
1411 len = name_width;
1412 name_len = strlen(name);
1413 if (name_width < name_len) {
1414 char *slash;
1415 prefix = "...";
1416 len -= 3;
1417 name += name_len - len;
1418 slash = strchr(name, '/');
1419 if (slash)
1420 name = slash;
1423 if (data->files[i]->is_binary) {
1424 fprintf(options->file, "%s", line_prefix);
1425 show_name(options->file, prefix, name, len);
1426 fprintf(options->file, " Bin ");
1427 fprintf(options->file, "%s%"PRIuMAX"%s",
1428 del_c, deleted, reset);
1429 fprintf(options->file, " -> ");
1430 fprintf(options->file, "%s%"PRIuMAX"%s",
1431 add_c, added, reset);
1432 fprintf(options->file, " bytes");
1433 fprintf(options->file, "\n");
1434 continue;
1436 else if (data->files[i]->is_unmerged) {
1437 fprintf(options->file, "%s", line_prefix);
1438 show_name(options->file, prefix, name, len);
1439 fprintf(options->file, " Unmerged\n");
1440 continue;
1444 * scale the add/delete
1446 add = added;
1447 del = deleted;
1448 adds += add;
1449 dels += del;
1451 if (width <= max_change) {
1452 add = scale_linear(add, width, max_change);
1453 del = scale_linear(del, width, max_change);
1455 fprintf(options->file, "%s", line_prefix);
1456 show_name(options->file, prefix, name, len);
1457 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1458 added + deleted ? " " : "");
1459 show_graph(options->file, '+', add, add_c, reset);
1460 show_graph(options->file, '-', del, del_c, reset);
1461 fprintf(options->file, "\n");
1463 for (i = count; i < data->nr; i++) {
1464 uintmax_t added = data->files[i]->added;
1465 uintmax_t deleted = data->files[i]->deleted;
1466 if (!data->files[i]->is_renamed &&
1467 (added + deleted == 0)) {
1468 total_files--;
1469 continue;
1471 adds += added;
1472 dels += deleted;
1473 if (!extra_shown)
1474 fprintf(options->file, "%s ...\n", line_prefix);
1475 extra_shown = 1;
1477 fprintf(options->file, "%s", line_prefix);
1478 fprintf(options->file,
1479 " %d files changed, %d insertions(+), %d deletions(-)\n",
1480 total_files, adds, dels);
1483 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1485 int i, adds = 0, dels = 0, total_files = data->nr;
1487 if (data->nr == 0)
1488 return;
1490 for (i = 0; i < data->nr; i++) {
1491 if (!data->files[i]->is_binary &&
1492 !data->files[i]->is_unmerged) {
1493 int added = data->files[i]->added;
1494 int deleted= data->files[i]->deleted;
1495 if (!data->files[i]->is_renamed &&
1496 (added + deleted == 0)) {
1497 total_files--;
1498 } else {
1499 adds += added;
1500 dels += deleted;
1504 if (options->output_prefix) {
1505 struct strbuf *msg = NULL;
1506 msg = options->output_prefix(options,
1507 options->output_prefix_data);
1508 fprintf(options->file, "%s", msg->buf);
1510 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1511 total_files, adds, dels);
1514 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1516 int i;
1518 if (data->nr == 0)
1519 return;
1521 for (i = 0; i < data->nr; i++) {
1522 struct diffstat_file *file = data->files[i];
1524 if (options->output_prefix) {
1525 struct strbuf *msg = NULL;
1526 msg = options->output_prefix(options,
1527 options->output_prefix_data);
1528 fprintf(options->file, "%s", msg->buf);
1531 if (file->is_binary)
1532 fprintf(options->file, "-\t-\t");
1533 else
1534 fprintf(options->file,
1535 "%"PRIuMAX"\t%"PRIuMAX"\t",
1536 file->added, file->deleted);
1537 if (options->line_termination) {
1538 fill_print_name(file);
1539 if (!file->is_renamed)
1540 write_name_quoted(file->name, options->file,
1541 options->line_termination);
1542 else {
1543 fputs(file->print_name, options->file);
1544 putc(options->line_termination, options->file);
1546 } else {
1547 if (file->is_renamed) {
1548 putc('\0', options->file);
1549 write_name_quoted(file->from_name, options->file, '\0');
1551 write_name_quoted(file->name, options->file, '\0');
1556 struct dirstat_file {
1557 const char *name;
1558 unsigned long changed;
1561 struct dirstat_dir {
1562 struct dirstat_file *files;
1563 int alloc, nr, permille, cumulative;
1566 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1567 unsigned long changed, const char *base, int baselen)
1569 unsigned long this_dir = 0;
1570 unsigned int sources = 0;
1571 const char *line_prefix = "";
1572 struct strbuf *msg = NULL;
1574 if (opt->output_prefix) {
1575 msg = opt->output_prefix(opt, opt->output_prefix_data);
1576 line_prefix = msg->buf;
1579 while (dir->nr) {
1580 struct dirstat_file *f = dir->files;
1581 int namelen = strlen(f->name);
1582 unsigned long this;
1583 char *slash;
1585 if (namelen < baselen)
1586 break;
1587 if (memcmp(f->name, base, baselen))
1588 break;
1589 slash = strchr(f->name + baselen, '/');
1590 if (slash) {
1591 int newbaselen = slash + 1 - f->name;
1592 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1593 sources++;
1594 } else {
1595 this = f->changed;
1596 dir->files++;
1597 dir->nr--;
1598 sources += 2;
1600 this_dir += this;
1604 * We don't report dirstat's for
1605 * - the top level
1606 * - or cases where everything came from a single directory
1607 * under this directory (sources == 1).
1609 if (baselen && sources != 1) {
1610 if (this_dir) {
1611 int permille = this_dir * 1000 / changed;
1612 if (permille >= dir->permille) {
1613 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1614 permille / 10, permille % 10, baselen, base);
1615 if (!dir->cumulative)
1616 return 0;
1620 return this_dir;
1623 static int dirstat_compare(const void *_a, const void *_b)
1625 const struct dirstat_file *a = _a;
1626 const struct dirstat_file *b = _b;
1627 return strcmp(a->name, b->name);
1630 static void show_dirstat(struct diff_options *options)
1632 int i;
1633 unsigned long changed;
1634 struct dirstat_dir dir;
1635 struct diff_queue_struct *q = &diff_queued_diff;
1637 dir.files = NULL;
1638 dir.alloc = 0;
1639 dir.nr = 0;
1640 dir.permille = options->dirstat_permille;
1641 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1643 changed = 0;
1644 for (i = 0; i < q->nr; i++) {
1645 struct diff_filepair *p = q->queue[i];
1646 const char *name;
1647 unsigned long copied, added, damage;
1648 int content_changed;
1650 name = p->two->path ? p->two->path : p->one->path;
1652 if (p->one->sha1_valid && p->two->sha1_valid)
1653 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1654 else
1655 content_changed = 1;
1657 if (!content_changed) {
1659 * The SHA1 has not changed, so pre-/post-content is
1660 * identical. We can therefore skip looking at the
1661 * file contents altogether.
1663 damage = 0;
1664 goto found_damage;
1667 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1669 * In --dirstat-by-file mode, we don't really need to
1670 * look at the actual file contents at all.
1671 * The fact that the SHA1 changed is enough for us to
1672 * add this file to the list of results
1673 * (with each file contributing equal damage).
1675 damage = 1;
1676 goto found_damage;
1679 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1680 diff_populate_filespec(p->one, 0);
1681 diff_populate_filespec(p->two, 0);
1682 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1683 &copied, &added);
1684 diff_free_filespec_data(p->one);
1685 diff_free_filespec_data(p->two);
1686 } else if (DIFF_FILE_VALID(p->one)) {
1687 diff_populate_filespec(p->one, 1);
1688 copied = added = 0;
1689 diff_free_filespec_data(p->one);
1690 } else if (DIFF_FILE_VALID(p->two)) {
1691 diff_populate_filespec(p->two, 1);
1692 copied = 0;
1693 added = p->two->size;
1694 diff_free_filespec_data(p->two);
1695 } else
1696 continue;
1699 * Original minus copied is the removed material,
1700 * added is the new material. They are both damages
1701 * made to the preimage.
1702 * If the resulting damage is zero, we know that
1703 * diffcore_count_changes() considers the two entries to
1704 * be identical, but since content_changed is true, we
1705 * know that there must have been _some_ kind of change,
1706 * so we force all entries to have damage > 0.
1708 damage = (p->one->size - copied) + added;
1709 if (!damage)
1710 damage = 1;
1712 found_damage:
1713 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1714 dir.files[dir.nr].name = name;
1715 dir.files[dir.nr].changed = damage;
1716 changed += damage;
1717 dir.nr++;
1720 /* This can happen even with many files, if everything was renames */
1721 if (!changed)
1722 return;
1724 /* Show all directories with more than x% of the changes */
1725 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1726 gather_dirstat(options, &dir, changed, "", 0);
1729 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1731 int i;
1732 unsigned long changed;
1733 struct dirstat_dir dir;
1735 if (data->nr == 0)
1736 return;
1738 dir.files = NULL;
1739 dir.alloc = 0;
1740 dir.nr = 0;
1741 dir.permille = options->dirstat_permille;
1742 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1744 changed = 0;
1745 for (i = 0; i < data->nr; i++) {
1746 struct diffstat_file *file = data->files[i];
1747 unsigned long damage = file->added + file->deleted;
1748 if (file->is_binary)
1750 * binary files counts bytes, not lines. Must find some
1751 * way to normalize binary bytes vs. textual lines.
1752 * The following heuristic assumes that there are 64
1753 * bytes per "line".
1754 * This is stupid and ugly, but very cheap...
1756 damage = (damage + 63) / 64;
1757 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1758 dir.files[dir.nr].name = file->name;
1759 dir.files[dir.nr].changed = damage;
1760 changed += damage;
1761 dir.nr++;
1764 /* This can happen even with many files, if everything was renames */
1765 if (!changed)
1766 return;
1768 /* Show all directories with more than x% of the changes */
1769 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1770 gather_dirstat(options, &dir, changed, "", 0);
1773 static void free_diffstat_info(struct diffstat_t *diffstat)
1775 int i;
1776 for (i = 0; i < diffstat->nr; i++) {
1777 struct diffstat_file *f = diffstat->files[i];
1778 if (f->name != f->print_name)
1779 free(f->print_name);
1780 free(f->name);
1781 free(f->from_name);
1782 free(f);
1784 free(diffstat->files);
1787 struct checkdiff_t {
1788 const char *filename;
1789 int lineno;
1790 int conflict_marker_size;
1791 struct diff_options *o;
1792 unsigned ws_rule;
1793 unsigned status;
1796 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1798 char firstchar;
1799 int cnt;
1801 if (len < marker_size + 1)
1802 return 0;
1803 firstchar = line[0];
1804 switch (firstchar) {
1805 case '=': case '>': case '<': case '|':
1806 break;
1807 default:
1808 return 0;
1810 for (cnt = 1; cnt < marker_size; cnt++)
1811 if (line[cnt] != firstchar)
1812 return 0;
1813 /* line[1] thru line[marker_size-1] are same as firstchar */
1814 if (len < marker_size + 1 || !isspace(line[marker_size]))
1815 return 0;
1816 return 1;
1819 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1821 struct checkdiff_t *data = priv;
1822 int marker_size = data->conflict_marker_size;
1823 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
1824 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
1825 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
1826 char *err;
1827 char *line_prefix = "";
1828 struct strbuf *msgbuf;
1830 assert(data->o);
1831 if (data->o->output_prefix) {
1832 msgbuf = data->o->output_prefix(data->o,
1833 data->o->output_prefix_data);
1834 line_prefix = msgbuf->buf;
1837 if (line[0] == '+') {
1838 unsigned bad;
1839 data->lineno++;
1840 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1841 data->status |= 1;
1842 fprintf(data->o->file,
1843 "%s%s:%d: leftover conflict marker\n",
1844 line_prefix, data->filename, data->lineno);
1846 bad = ws_check(line + 1, len - 1, data->ws_rule);
1847 if (!bad)
1848 return;
1849 data->status |= bad;
1850 err = whitespace_error_string(bad);
1851 fprintf(data->o->file, "%s%s:%d: %s.\n",
1852 line_prefix, data->filename, data->lineno, err);
1853 free(err);
1854 emit_line(data->o, set, reset, line, 1);
1855 ws_check_emit(line + 1, len - 1, data->ws_rule,
1856 data->o->file, set, reset, ws);
1857 } else if (line[0] == ' ') {
1858 data->lineno++;
1859 } else if (line[0] == '@') {
1860 char *plus = strchr(line, '+');
1861 if (plus)
1862 data->lineno = strtol(plus, NULL, 10) - 1;
1863 else
1864 die("invalid diff");
1868 static unsigned char *deflate_it(char *data,
1869 unsigned long size,
1870 unsigned long *result_size)
1872 int bound;
1873 unsigned char *deflated;
1874 git_zstream stream;
1876 memset(&stream, 0, sizeof(stream));
1877 git_deflate_init(&stream, zlib_compression_level);
1878 bound = git_deflate_bound(&stream, size);
1879 deflated = xmalloc(bound);
1880 stream.next_out = deflated;
1881 stream.avail_out = bound;
1883 stream.next_in = (unsigned char *)data;
1884 stream.avail_in = size;
1885 while (git_deflate(&stream, Z_FINISH) == Z_OK)
1886 ; /* nothing */
1887 git_deflate_end(&stream);
1888 *result_size = stream.total_out;
1889 return deflated;
1892 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1894 void *cp;
1895 void *delta;
1896 void *deflated;
1897 void *data;
1898 unsigned long orig_size;
1899 unsigned long delta_size;
1900 unsigned long deflate_size;
1901 unsigned long data_size;
1903 /* We could do deflated delta, or we could do just deflated two,
1904 * whichever is smaller.
1906 delta = NULL;
1907 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1908 if (one->size && two->size) {
1909 delta = diff_delta(one->ptr, one->size,
1910 two->ptr, two->size,
1911 &delta_size, deflate_size);
1912 if (delta) {
1913 void *to_free = delta;
1914 orig_size = delta_size;
1915 delta = deflate_it(delta, delta_size, &delta_size);
1916 free(to_free);
1920 if (delta && delta_size < deflate_size) {
1921 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1922 free(deflated);
1923 data = delta;
1924 data_size = delta_size;
1926 else {
1927 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1928 free(delta);
1929 data = deflated;
1930 data_size = deflate_size;
1933 /* emit data encoded in base85 */
1934 cp = data;
1935 while (data_size) {
1936 int bytes = (52 < data_size) ? 52 : data_size;
1937 char line[70];
1938 data_size -= bytes;
1939 if (bytes <= 26)
1940 line[0] = bytes + 'A' - 1;
1941 else
1942 line[0] = bytes - 26 + 'a' - 1;
1943 encode_85(line + 1, cp, bytes);
1944 cp = (char *) cp + bytes;
1945 fprintf(file, "%s", prefix);
1946 fputs(line, file);
1947 fputc('\n', file);
1949 fprintf(file, "%s\n", prefix);
1950 free(data);
1953 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1955 fprintf(file, "%sGIT binary patch\n", prefix);
1956 emit_binary_diff_body(file, one, two, prefix);
1957 emit_binary_diff_body(file, two, one, prefix);
1960 static void diff_filespec_load_driver(struct diff_filespec *one)
1962 /* Use already-loaded driver */
1963 if (one->driver)
1964 return;
1966 if (S_ISREG(one->mode))
1967 one->driver = userdiff_find_by_path(one->path);
1969 /* Fallback to default settings */
1970 if (!one->driver)
1971 one->driver = userdiff_find_by_name("default");
1974 int diff_filespec_is_binary(struct diff_filespec *one)
1976 if (one->is_binary == -1) {
1977 diff_filespec_load_driver(one);
1978 if (one->driver->binary != -1)
1979 one->is_binary = one->driver->binary;
1980 else {
1981 if (!one->data && DIFF_FILE_VALID(one))
1982 diff_populate_filespec(one, 0);
1983 if (one->data)
1984 one->is_binary = buffer_is_binary(one->data,
1985 one->size);
1986 if (one->is_binary == -1)
1987 one->is_binary = 0;
1990 return one->is_binary;
1993 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1995 diff_filespec_load_driver(one);
1996 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1999 static const char *userdiff_word_regex(struct diff_filespec *one)
2001 diff_filespec_load_driver(one);
2002 return one->driver->word_regex;
2005 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2007 if (!options->a_prefix)
2008 options->a_prefix = a;
2009 if (!options->b_prefix)
2010 options->b_prefix = b;
2013 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2015 if (!DIFF_FILE_VALID(one))
2016 return NULL;
2018 diff_filespec_load_driver(one);
2019 return userdiff_get_textconv(one->driver);
2022 static void builtin_diff(const char *name_a,
2023 const char *name_b,
2024 struct diff_filespec *one,
2025 struct diff_filespec *two,
2026 const char *xfrm_msg,
2027 int must_show_header,
2028 struct diff_options *o,
2029 int complete_rewrite)
2031 mmfile_t mf1, mf2;
2032 const char *lbl[2];
2033 char *a_one, *b_two;
2034 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
2035 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2036 const char *a_prefix, *b_prefix;
2037 struct userdiff_driver *textconv_one = NULL;
2038 struct userdiff_driver *textconv_two = NULL;
2039 struct strbuf header = STRBUF_INIT;
2040 struct strbuf *msgbuf;
2041 char *line_prefix = "";
2043 if (o->output_prefix) {
2044 msgbuf = o->output_prefix(o, o->output_prefix_data);
2045 line_prefix = msgbuf->buf;
2048 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2049 (!one->mode || S_ISGITLINK(one->mode)) &&
2050 (!two->mode || S_ISGITLINK(two->mode))) {
2051 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2052 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2053 show_submodule_summary(o->file, one ? one->path : two->path,
2054 one->sha1, two->sha1, two->dirty_submodule,
2055 del, add, reset);
2056 return;
2059 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2060 textconv_one = get_textconv(one);
2061 textconv_two = get_textconv(two);
2064 diff_set_mnemonic_prefix(o, "a/", "b/");
2065 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2066 a_prefix = o->b_prefix;
2067 b_prefix = o->a_prefix;
2068 } else {
2069 a_prefix = o->a_prefix;
2070 b_prefix = o->b_prefix;
2073 /* Never use a non-valid filename anywhere if at all possible */
2074 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2075 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2077 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2078 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2079 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2080 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2081 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2082 if (lbl[0][0] == '/') {
2083 /* /dev/null */
2084 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2085 if (xfrm_msg)
2086 strbuf_addstr(&header, xfrm_msg);
2087 must_show_header = 1;
2089 else if (lbl[1][0] == '/') {
2090 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2091 if (xfrm_msg)
2092 strbuf_addstr(&header, xfrm_msg);
2093 must_show_header = 1;
2095 else {
2096 if (one->mode != two->mode) {
2097 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2098 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2099 must_show_header = 1;
2101 if (xfrm_msg)
2102 strbuf_addstr(&header, xfrm_msg);
2105 * we do not run diff between different kind
2106 * of objects.
2108 if ((one->mode ^ two->mode) & S_IFMT)
2109 goto free_ab_and_return;
2110 if (complete_rewrite &&
2111 (textconv_one || !diff_filespec_is_binary(one)) &&
2112 (textconv_two || !diff_filespec_is_binary(two))) {
2113 fprintf(o->file, "%s", header.buf);
2114 strbuf_reset(&header);
2115 emit_rewrite_diff(name_a, name_b, one, two,
2116 textconv_one, textconv_two, o);
2117 o->found_changes = 1;
2118 goto free_ab_and_return;
2122 if (o->irreversible_delete && lbl[1][0] == '/') {
2123 fprintf(o->file, "%s", header.buf);
2124 strbuf_reset(&header);
2125 goto free_ab_and_return;
2126 } else if (!DIFF_OPT_TST(o, TEXT) &&
2127 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2128 (!textconv_two && diff_filespec_is_binary(two)) )) {
2129 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2130 die("unable to read files to diff");
2131 /* Quite common confusing case */
2132 if (mf1.size == mf2.size &&
2133 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2134 if (must_show_header)
2135 fprintf(o->file, "%s", header.buf);
2136 goto free_ab_and_return;
2138 fprintf(o->file, "%s", header.buf);
2139 strbuf_reset(&header);
2140 if (DIFF_OPT_TST(o, BINARY))
2141 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2142 else
2143 fprintf(o->file, "%sBinary files %s and %s differ\n",
2144 line_prefix, lbl[0], lbl[1]);
2145 o->found_changes = 1;
2146 } else {
2147 /* Crazy xdl interfaces.. */
2148 const char *diffopts = getenv("GIT_DIFF_OPTS");
2149 xpparam_t xpp;
2150 xdemitconf_t xecfg;
2151 struct emit_callback ecbdata;
2152 const struct userdiff_funcname *pe;
2154 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2155 fprintf(o->file, "%s", header.buf);
2156 strbuf_reset(&header);
2159 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2160 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2162 pe = diff_funcname_pattern(one);
2163 if (!pe)
2164 pe = diff_funcname_pattern(two);
2166 memset(&xpp, 0, sizeof(xpp));
2167 memset(&xecfg, 0, sizeof(xecfg));
2168 memset(&ecbdata, 0, sizeof(ecbdata));
2169 ecbdata.label_path = lbl;
2170 ecbdata.color_diff = want_color(o->use_color);
2171 ecbdata.found_changesp = &o->found_changes;
2172 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2173 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2174 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2175 ecbdata.opt = o;
2176 ecbdata.header = header.len ? &header : NULL;
2177 xpp.flags = o->xdl_opts;
2178 xecfg.ctxlen = o->context;
2179 xecfg.interhunkctxlen = o->interhunkcontext;
2180 xecfg.flags = XDL_EMIT_FUNCNAMES;
2181 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2182 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2183 if (pe)
2184 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2185 if (!diffopts)
2187 else if (!prefixcmp(diffopts, "--unified="))
2188 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2189 else if (!prefixcmp(diffopts, "-u"))
2190 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2191 if (o->word_diff) {
2192 int i;
2194 ecbdata.diff_words =
2195 xcalloc(1, sizeof(struct diff_words_data));
2196 ecbdata.diff_words->type = o->word_diff;
2197 ecbdata.diff_words->opt = o;
2198 if (!o->word_regex)
2199 o->word_regex = userdiff_word_regex(one);
2200 if (!o->word_regex)
2201 o->word_regex = userdiff_word_regex(two);
2202 if (!o->word_regex)
2203 o->word_regex = diff_word_regex_cfg;
2204 if (o->word_regex) {
2205 ecbdata.diff_words->word_regex = (regex_t *)
2206 xmalloc(sizeof(regex_t));
2207 if (regcomp(ecbdata.diff_words->word_regex,
2208 o->word_regex,
2209 REG_EXTENDED | REG_NEWLINE))
2210 die ("Invalid regular expression: %s",
2211 o->word_regex);
2213 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2214 if (o->word_diff == diff_words_styles[i].type) {
2215 ecbdata.diff_words->style =
2216 &diff_words_styles[i];
2217 break;
2220 if (want_color(o->use_color)) {
2221 struct diff_words_style *st = ecbdata.diff_words->style;
2222 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2223 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2224 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2227 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2228 &xpp, &xecfg);
2229 if (o->word_diff)
2230 free_diff_words_data(&ecbdata);
2231 if (textconv_one)
2232 free(mf1.ptr);
2233 if (textconv_two)
2234 free(mf2.ptr);
2235 xdiff_clear_find_func(&xecfg);
2238 free_ab_and_return:
2239 strbuf_release(&header);
2240 diff_free_filespec_data(one);
2241 diff_free_filespec_data(two);
2242 free(a_one);
2243 free(b_two);
2244 return;
2247 static void builtin_diffstat(const char *name_a, const char *name_b,
2248 struct diff_filespec *one,
2249 struct diff_filespec *two,
2250 struct diffstat_t *diffstat,
2251 struct diff_options *o,
2252 int complete_rewrite)
2254 mmfile_t mf1, mf2;
2255 struct diffstat_file *data;
2257 data = diffstat_add(diffstat, name_a, name_b);
2259 if (!one || !two) {
2260 data->is_unmerged = 1;
2261 return;
2264 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2265 data->is_binary = 1;
2266 data->added = diff_filespec_size(two);
2267 data->deleted = diff_filespec_size(one);
2270 else if (complete_rewrite) {
2271 diff_populate_filespec(one, 0);
2272 diff_populate_filespec(two, 0);
2273 data->deleted = count_lines(one->data, one->size);
2274 data->added = count_lines(two->data, two->size);
2277 else {
2278 /* Crazy xdl interfaces.. */
2279 xpparam_t xpp;
2280 xdemitconf_t xecfg;
2282 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2283 die("unable to read files to diff");
2285 memset(&xpp, 0, sizeof(xpp));
2286 memset(&xecfg, 0, sizeof(xecfg));
2287 xpp.flags = o->xdl_opts;
2288 xecfg.ctxlen = o->context;
2289 xecfg.interhunkctxlen = o->interhunkcontext;
2290 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2291 &xpp, &xecfg);
2294 diff_free_filespec_data(one);
2295 diff_free_filespec_data(two);
2298 static void builtin_checkdiff(const char *name_a, const char *name_b,
2299 const char *attr_path,
2300 struct diff_filespec *one,
2301 struct diff_filespec *two,
2302 struct diff_options *o)
2304 mmfile_t mf1, mf2;
2305 struct checkdiff_t data;
2307 if (!two)
2308 return;
2310 memset(&data, 0, sizeof(data));
2311 data.filename = name_b ? name_b : name_a;
2312 data.lineno = 0;
2313 data.o = o;
2314 data.ws_rule = whitespace_rule(attr_path);
2315 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2317 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2318 die("unable to read files to diff");
2321 * All the other codepaths check both sides, but not checking
2322 * the "old" side here is deliberate. We are checking the newly
2323 * introduced changes, and as long as the "new" side is text, we
2324 * can and should check what it introduces.
2326 if (diff_filespec_is_binary(two))
2327 goto free_and_return;
2328 else {
2329 /* Crazy xdl interfaces.. */
2330 xpparam_t xpp;
2331 xdemitconf_t xecfg;
2333 memset(&xpp, 0, sizeof(xpp));
2334 memset(&xecfg, 0, sizeof(xecfg));
2335 xecfg.ctxlen = 1; /* at least one context line */
2336 xpp.flags = 0;
2337 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2338 &xpp, &xecfg);
2340 if (data.ws_rule & WS_BLANK_AT_EOF) {
2341 struct emit_callback ecbdata;
2342 int blank_at_eof;
2344 ecbdata.ws_rule = data.ws_rule;
2345 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2346 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2348 if (blank_at_eof) {
2349 static char *err;
2350 if (!err)
2351 err = whitespace_error_string(WS_BLANK_AT_EOF);
2352 fprintf(o->file, "%s:%d: %s.\n",
2353 data.filename, blank_at_eof, err);
2354 data.status = 1; /* report errors */
2358 free_and_return:
2359 diff_free_filespec_data(one);
2360 diff_free_filespec_data(two);
2361 if (data.status)
2362 DIFF_OPT_SET(o, CHECK_FAILED);
2365 struct diff_filespec *alloc_filespec(const char *path)
2367 int namelen = strlen(path);
2368 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2370 memset(spec, 0, sizeof(*spec));
2371 spec->path = (char *)(spec + 1);
2372 memcpy(spec->path, path, namelen+1);
2373 spec->count = 1;
2374 spec->is_binary = -1;
2375 return spec;
2378 void free_filespec(struct diff_filespec *spec)
2380 if (!--spec->count) {
2381 diff_free_filespec_data(spec);
2382 free(spec);
2386 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2387 unsigned short mode)
2389 if (mode) {
2390 spec->mode = canon_mode(mode);
2391 hashcpy(spec->sha1, sha1);
2392 spec->sha1_valid = !is_null_sha1(sha1);
2397 * Given a name and sha1 pair, if the index tells us the file in
2398 * the work tree has that object contents, return true, so that
2399 * prepare_temp_file() does not have to inflate and extract.
2401 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2403 struct cache_entry *ce;
2404 struct stat st;
2405 int pos, len;
2408 * We do not read the cache ourselves here, because the
2409 * benchmark with my previous version that always reads cache
2410 * shows that it makes things worse for diff-tree comparing
2411 * two linux-2.6 kernel trees in an already checked out work
2412 * tree. This is because most diff-tree comparisons deal with
2413 * only a small number of files, while reading the cache is
2414 * expensive for a large project, and its cost outweighs the
2415 * savings we get by not inflating the object to a temporary
2416 * file. Practically, this code only helps when we are used
2417 * by diff-cache --cached, which does read the cache before
2418 * calling us.
2420 if (!active_cache)
2421 return 0;
2423 /* We want to avoid the working directory if our caller
2424 * doesn't need the data in a normal file, this system
2425 * is rather slow with its stat/open/mmap/close syscalls,
2426 * and the object is contained in a pack file. The pack
2427 * is probably already open and will be faster to obtain
2428 * the data through than the working directory. Loose
2429 * objects however would tend to be slower as they need
2430 * to be individually opened and inflated.
2432 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2433 return 0;
2435 len = strlen(name);
2436 pos = cache_name_pos(name, len);
2437 if (pos < 0)
2438 return 0;
2439 ce = active_cache[pos];
2442 * This is not the sha1 we are looking for, or
2443 * unreusable because it is not a regular file.
2445 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2446 return 0;
2449 * If ce is marked as "assume unchanged", there is no
2450 * guarantee that work tree matches what we are looking for.
2452 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2453 return 0;
2456 * If ce matches the file in the work tree, we can reuse it.
2458 if (ce_uptodate(ce) ||
2459 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2460 return 1;
2462 return 0;
2465 static int populate_from_stdin(struct diff_filespec *s)
2467 struct strbuf buf = STRBUF_INIT;
2468 size_t size = 0;
2470 if (strbuf_read(&buf, 0, 0) < 0)
2471 return error("error while reading from stdin %s",
2472 strerror(errno));
2474 s->should_munmap = 0;
2475 s->data = strbuf_detach(&buf, &size);
2476 s->size = size;
2477 s->should_free = 1;
2478 return 0;
2481 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2483 int len;
2484 char *data = xmalloc(100), *dirty = "";
2486 /* Are we looking at the work tree? */
2487 if (s->dirty_submodule)
2488 dirty = "-dirty";
2490 len = snprintf(data, 100,
2491 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2492 s->data = data;
2493 s->size = len;
2494 s->should_free = 1;
2495 if (size_only) {
2496 s->data = NULL;
2497 free(data);
2499 return 0;
2503 * While doing rename detection and pickaxe operation, we may need to
2504 * grab the data for the blob (or file) for our own in-core comparison.
2505 * diff_filespec has data and size fields for this purpose.
2507 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2509 int err = 0;
2510 if (!DIFF_FILE_VALID(s))
2511 die("internal error: asking to populate invalid file.");
2512 if (S_ISDIR(s->mode))
2513 return -1;
2515 if (s->data)
2516 return 0;
2518 if (size_only && 0 < s->size)
2519 return 0;
2521 if (S_ISGITLINK(s->mode))
2522 return diff_populate_gitlink(s, size_only);
2524 if (!s->sha1_valid ||
2525 reuse_worktree_file(s->path, s->sha1, 0)) {
2526 struct strbuf buf = STRBUF_INIT;
2527 struct stat st;
2528 int fd;
2530 if (!strcmp(s->path, "-"))
2531 return populate_from_stdin(s);
2533 if (lstat(s->path, &st) < 0) {
2534 if (errno == ENOENT) {
2535 err_empty:
2536 err = -1;
2537 empty:
2538 s->data = (char *)"";
2539 s->size = 0;
2540 return err;
2543 s->size = xsize_t(st.st_size);
2544 if (!s->size)
2545 goto empty;
2546 if (S_ISLNK(st.st_mode)) {
2547 struct strbuf sb = STRBUF_INIT;
2549 if (strbuf_readlink(&sb, s->path, s->size))
2550 goto err_empty;
2551 s->size = sb.len;
2552 s->data = strbuf_detach(&sb, NULL);
2553 s->should_free = 1;
2554 return 0;
2556 if (size_only)
2557 return 0;
2558 fd = open(s->path, O_RDONLY);
2559 if (fd < 0)
2560 goto err_empty;
2561 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2562 close(fd);
2563 s->should_munmap = 1;
2566 * Convert from working tree format to canonical git format
2568 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2569 size_t size = 0;
2570 munmap(s->data, s->size);
2571 s->should_munmap = 0;
2572 s->data = strbuf_detach(&buf, &size);
2573 s->size = size;
2574 s->should_free = 1;
2577 else {
2578 enum object_type type;
2579 if (size_only) {
2580 type = sha1_object_info(s->sha1, &s->size);
2581 if (type < 0)
2582 die("unable to read %s", sha1_to_hex(s->sha1));
2583 } else {
2584 s->data = read_sha1_file(s->sha1, &type, &s->size);
2585 if (!s->data)
2586 die("unable to read %s", sha1_to_hex(s->sha1));
2587 s->should_free = 1;
2590 return 0;
2593 void diff_free_filespec_blob(struct diff_filespec *s)
2595 if (s->should_free)
2596 free(s->data);
2597 else if (s->should_munmap)
2598 munmap(s->data, s->size);
2600 if (s->should_free || s->should_munmap) {
2601 s->should_free = s->should_munmap = 0;
2602 s->data = NULL;
2606 void diff_free_filespec_data(struct diff_filespec *s)
2608 diff_free_filespec_blob(s);
2609 free(s->cnt_data);
2610 s->cnt_data = NULL;
2613 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2614 void *blob,
2615 unsigned long size,
2616 const unsigned char *sha1,
2617 int mode)
2619 int fd;
2620 struct strbuf buf = STRBUF_INIT;
2621 struct strbuf template = STRBUF_INIT;
2622 char *path_dup = xstrdup(path);
2623 const char *base = basename(path_dup);
2625 /* Generate "XXXXXX_basename.ext" */
2626 strbuf_addstr(&template, "XXXXXX_");
2627 strbuf_addstr(&template, base);
2629 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2630 strlen(base) + 1);
2631 if (fd < 0)
2632 die_errno("unable to create temp-file");
2633 if (convert_to_working_tree(path,
2634 (const char *)blob, (size_t)size, &buf)) {
2635 blob = buf.buf;
2636 size = buf.len;
2638 if (write_in_full(fd, blob, size) != size)
2639 die_errno("unable to write temp-file");
2640 close(fd);
2641 temp->name = temp->tmp_path;
2642 strcpy(temp->hex, sha1_to_hex(sha1));
2643 temp->hex[40] = 0;
2644 sprintf(temp->mode, "%06o", mode);
2645 strbuf_release(&buf);
2646 strbuf_release(&template);
2647 free(path_dup);
2650 static struct diff_tempfile *prepare_temp_file(const char *name,
2651 struct diff_filespec *one)
2653 struct diff_tempfile *temp = claim_diff_tempfile();
2655 if (!DIFF_FILE_VALID(one)) {
2656 not_a_valid_file:
2657 /* A '-' entry produces this for file-2, and
2658 * a '+' entry produces this for file-1.
2660 temp->name = "/dev/null";
2661 strcpy(temp->hex, ".");
2662 strcpy(temp->mode, ".");
2663 return temp;
2666 if (!remove_tempfile_installed) {
2667 atexit(remove_tempfile);
2668 sigchain_push_common(remove_tempfile_on_signal);
2669 remove_tempfile_installed = 1;
2672 if (!one->sha1_valid ||
2673 reuse_worktree_file(name, one->sha1, 1)) {
2674 struct stat st;
2675 if (lstat(name, &st) < 0) {
2676 if (errno == ENOENT)
2677 goto not_a_valid_file;
2678 die_errno("stat(%s)", name);
2680 if (S_ISLNK(st.st_mode)) {
2681 struct strbuf sb = STRBUF_INIT;
2682 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2683 die_errno("readlink(%s)", name);
2684 prep_temp_blob(name, temp, sb.buf, sb.len,
2685 (one->sha1_valid ?
2686 one->sha1 : null_sha1),
2687 (one->sha1_valid ?
2688 one->mode : S_IFLNK));
2689 strbuf_release(&sb);
2691 else {
2692 /* we can borrow from the file in the work tree */
2693 temp->name = name;
2694 if (!one->sha1_valid)
2695 strcpy(temp->hex, sha1_to_hex(null_sha1));
2696 else
2697 strcpy(temp->hex, sha1_to_hex(one->sha1));
2698 /* Even though we may sometimes borrow the
2699 * contents from the work tree, we always want
2700 * one->mode. mode is trustworthy even when
2701 * !(one->sha1_valid), as long as
2702 * DIFF_FILE_VALID(one).
2704 sprintf(temp->mode, "%06o", one->mode);
2706 return temp;
2708 else {
2709 if (diff_populate_filespec(one, 0))
2710 die("cannot read data blob for %s", one->path);
2711 prep_temp_blob(name, temp, one->data, one->size,
2712 one->sha1, one->mode);
2714 return temp;
2717 /* An external diff command takes:
2719 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2720 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2723 static void run_external_diff(const char *pgm,
2724 const char *name,
2725 const char *other,
2726 struct diff_filespec *one,
2727 struct diff_filespec *two,
2728 const char *xfrm_msg,
2729 int complete_rewrite)
2731 const char *spawn_arg[10];
2732 int retval;
2733 const char **arg = &spawn_arg[0];
2735 if (one && two) {
2736 struct diff_tempfile *temp_one, *temp_two;
2737 const char *othername = (other ? other : name);
2738 temp_one = prepare_temp_file(name, one);
2739 temp_two = prepare_temp_file(othername, two);
2740 *arg++ = pgm;
2741 *arg++ = name;
2742 *arg++ = temp_one->name;
2743 *arg++ = temp_one->hex;
2744 *arg++ = temp_one->mode;
2745 *arg++ = temp_two->name;
2746 *arg++ = temp_two->hex;
2747 *arg++ = temp_two->mode;
2748 if (other) {
2749 *arg++ = other;
2750 *arg++ = xfrm_msg;
2752 } else {
2753 *arg++ = pgm;
2754 *arg++ = name;
2756 *arg = NULL;
2757 fflush(NULL);
2758 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2759 remove_tempfile();
2760 if (retval) {
2761 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2762 exit(1);
2766 static int similarity_index(struct diff_filepair *p)
2768 return p->score * 100 / MAX_SCORE;
2771 static void fill_metainfo(struct strbuf *msg,
2772 const char *name,
2773 const char *other,
2774 struct diff_filespec *one,
2775 struct diff_filespec *two,
2776 struct diff_options *o,
2777 struct diff_filepair *p,
2778 int *must_show_header,
2779 int use_color)
2781 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2782 const char *reset = diff_get_color(use_color, DIFF_RESET);
2783 struct strbuf *msgbuf;
2784 char *line_prefix = "";
2786 *must_show_header = 1;
2787 if (o->output_prefix) {
2788 msgbuf = o->output_prefix(o, o->output_prefix_data);
2789 line_prefix = msgbuf->buf;
2791 strbuf_init(msg, PATH_MAX * 2 + 300);
2792 switch (p->status) {
2793 case DIFF_STATUS_COPIED:
2794 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2795 line_prefix, set, similarity_index(p));
2796 strbuf_addf(msg, "%s\n%s%scopy from ",
2797 reset, line_prefix, set);
2798 quote_c_style(name, msg, NULL, 0);
2799 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2800 quote_c_style(other, msg, NULL, 0);
2801 strbuf_addf(msg, "%s\n", reset);
2802 break;
2803 case DIFF_STATUS_RENAMED:
2804 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2805 line_prefix, set, similarity_index(p));
2806 strbuf_addf(msg, "%s\n%s%srename from ",
2807 reset, line_prefix, set);
2808 quote_c_style(name, msg, NULL, 0);
2809 strbuf_addf(msg, "%s\n%s%srename to ",
2810 reset, line_prefix, set);
2811 quote_c_style(other, msg, NULL, 0);
2812 strbuf_addf(msg, "%s\n", reset);
2813 break;
2814 case DIFF_STATUS_MODIFIED:
2815 if (p->score) {
2816 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2817 line_prefix,
2818 set, similarity_index(p), reset);
2819 break;
2821 /* fallthru */
2822 default:
2823 *must_show_header = 0;
2825 if (one && two && hashcmp(one->sha1, two->sha1)) {
2826 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2828 if (DIFF_OPT_TST(o, BINARY)) {
2829 mmfile_t mf;
2830 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2831 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2832 abbrev = 40;
2834 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2835 find_unique_abbrev(one->sha1, abbrev));
2836 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2837 if (one->mode == two->mode)
2838 strbuf_addf(msg, " %06o", one->mode);
2839 strbuf_addf(msg, "%s\n", reset);
2843 static void run_diff_cmd(const char *pgm,
2844 const char *name,
2845 const char *other,
2846 const char *attr_path,
2847 struct diff_filespec *one,
2848 struct diff_filespec *two,
2849 struct strbuf *msg,
2850 struct diff_options *o,
2851 struct diff_filepair *p)
2853 const char *xfrm_msg = NULL;
2854 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2855 int must_show_header = 0;
2857 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2858 pgm = NULL;
2859 else {
2860 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2861 if (drv && drv->external)
2862 pgm = drv->external;
2865 if (msg) {
2867 * don't use colors when the header is intended for an
2868 * external diff driver
2870 fill_metainfo(msg, name, other, one, two, o, p,
2871 &must_show_header,
2872 want_color(o->use_color) && !pgm);
2873 xfrm_msg = msg->len ? msg->buf : NULL;
2876 if (pgm) {
2877 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2878 complete_rewrite);
2879 return;
2881 if (one && two)
2882 builtin_diff(name, other ? other : name,
2883 one, two, xfrm_msg, must_show_header,
2884 o, complete_rewrite);
2885 else
2886 fprintf(o->file, "* Unmerged path %s\n", name);
2889 static void diff_fill_sha1_info(struct diff_filespec *one)
2891 if (DIFF_FILE_VALID(one)) {
2892 if (!one->sha1_valid) {
2893 struct stat st;
2894 if (!strcmp(one->path, "-")) {
2895 hashcpy(one->sha1, null_sha1);
2896 return;
2898 if (lstat(one->path, &st) < 0)
2899 die_errno("stat '%s'", one->path);
2900 if (index_path(one->sha1, one->path, &st, 0))
2901 die("cannot hash %s", one->path);
2904 else
2905 hashclr(one->sha1);
2908 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2910 /* Strip the prefix but do not molest /dev/null and absolute paths */
2911 if (*namep && **namep != '/') {
2912 *namep += prefix_length;
2913 if (**namep == '/')
2914 ++*namep;
2916 if (*otherp && **otherp != '/') {
2917 *otherp += prefix_length;
2918 if (**otherp == '/')
2919 ++*otherp;
2923 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2925 const char *pgm = external_diff();
2926 struct strbuf msg;
2927 struct diff_filespec *one = p->one;
2928 struct diff_filespec *two = p->two;
2929 const char *name;
2930 const char *other;
2931 const char *attr_path;
2933 name = p->one->path;
2934 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2935 attr_path = name;
2936 if (o->prefix_length)
2937 strip_prefix(o->prefix_length, &name, &other);
2939 if (DIFF_PAIR_UNMERGED(p)) {
2940 run_diff_cmd(pgm, name, NULL, attr_path,
2941 NULL, NULL, NULL, o, p);
2942 return;
2945 diff_fill_sha1_info(one);
2946 diff_fill_sha1_info(two);
2948 if (!pgm &&
2949 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2950 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2952 * a filepair that changes between file and symlink
2953 * needs to be split into deletion and creation.
2955 struct diff_filespec *null = alloc_filespec(two->path);
2956 run_diff_cmd(NULL, name, other, attr_path,
2957 one, null, &msg, o, p);
2958 free(null);
2959 strbuf_release(&msg);
2961 null = alloc_filespec(one->path);
2962 run_diff_cmd(NULL, name, other, attr_path,
2963 null, two, &msg, o, p);
2964 free(null);
2966 else
2967 run_diff_cmd(pgm, name, other, attr_path,
2968 one, two, &msg, o, p);
2970 strbuf_release(&msg);
2973 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2974 struct diffstat_t *diffstat)
2976 const char *name;
2977 const char *other;
2978 int complete_rewrite = 0;
2980 if (DIFF_PAIR_UNMERGED(p)) {
2981 /* unmerged */
2982 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2983 return;
2986 name = p->one->path;
2987 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2989 if (o->prefix_length)
2990 strip_prefix(o->prefix_length, &name, &other);
2992 diff_fill_sha1_info(p->one);
2993 diff_fill_sha1_info(p->two);
2995 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2996 complete_rewrite = 1;
2997 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
3000 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3002 const char *name;
3003 const char *other;
3004 const char *attr_path;
3006 if (DIFF_PAIR_UNMERGED(p)) {
3007 /* unmerged */
3008 return;
3011 name = p->one->path;
3012 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3013 attr_path = other ? other : name;
3015 if (o->prefix_length)
3016 strip_prefix(o->prefix_length, &name, &other);
3018 diff_fill_sha1_info(p->one);
3019 diff_fill_sha1_info(p->two);
3021 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3024 void diff_setup(struct diff_options *options)
3026 memcpy(options, &default_diff_options, sizeof(*options));
3028 options->file = stdout;
3030 options->line_termination = '\n';
3031 options->break_opt = -1;
3032 options->rename_limit = -1;
3033 options->dirstat_permille = diff_dirstat_permille_default;
3034 options->context = 3;
3036 options->change = diff_change;
3037 options->add_remove = diff_addremove;
3038 options->use_color = diff_use_color_default;
3039 options->detect_rename = diff_detect_rename_default;
3041 if (diff_no_prefix) {
3042 options->a_prefix = options->b_prefix = "";
3043 } else if (!diff_mnemonic_prefix) {
3044 options->a_prefix = "a/";
3045 options->b_prefix = "b/";
3049 int diff_setup_done(struct diff_options *options)
3051 int count = 0;
3053 if (options->output_format & DIFF_FORMAT_NAME)
3054 count++;
3055 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3056 count++;
3057 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3058 count++;
3059 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3060 count++;
3061 if (count > 1)
3062 die("--name-only, --name-status, --check and -s are mutually exclusive");
3065 * Most of the time we can say "there are changes"
3066 * only by checking if there are changed paths, but
3067 * --ignore-whitespace* options force us to look
3068 * inside contents.
3071 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3072 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3073 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3074 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3075 else
3076 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3078 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3079 options->detect_rename = DIFF_DETECT_COPY;
3081 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3082 options->prefix = NULL;
3083 if (options->prefix)
3084 options->prefix_length = strlen(options->prefix);
3085 else
3086 options->prefix_length = 0;
3088 if (options->output_format & (DIFF_FORMAT_NAME |
3089 DIFF_FORMAT_NAME_STATUS |
3090 DIFF_FORMAT_CHECKDIFF |
3091 DIFF_FORMAT_NO_OUTPUT))
3092 options->output_format &= ~(DIFF_FORMAT_RAW |
3093 DIFF_FORMAT_NUMSTAT |
3094 DIFF_FORMAT_DIFFSTAT |
3095 DIFF_FORMAT_SHORTSTAT |
3096 DIFF_FORMAT_DIRSTAT |
3097 DIFF_FORMAT_SUMMARY |
3098 DIFF_FORMAT_PATCH);
3101 * These cases always need recursive; we do not drop caller-supplied
3102 * recursive bits for other formats here.
3104 if (options->output_format & (DIFF_FORMAT_PATCH |
3105 DIFF_FORMAT_NUMSTAT |
3106 DIFF_FORMAT_DIFFSTAT |
3107 DIFF_FORMAT_SHORTSTAT |
3108 DIFF_FORMAT_DIRSTAT |
3109 DIFF_FORMAT_SUMMARY |
3110 DIFF_FORMAT_CHECKDIFF))
3111 DIFF_OPT_SET(options, RECURSIVE);
3113 * Also pickaxe would not work very well if you do not say recursive
3115 if (options->pickaxe)
3116 DIFF_OPT_SET(options, RECURSIVE);
3118 * When patches are generated, submodules diffed against the work tree
3119 * must be checked for dirtiness too so it can be shown in the output
3121 if (options->output_format & DIFF_FORMAT_PATCH)
3122 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3124 if (options->detect_rename && options->rename_limit < 0)
3125 options->rename_limit = diff_rename_limit_default;
3126 if (options->setup & DIFF_SETUP_USE_CACHE) {
3127 if (!active_cache)
3128 /* read-cache does not die even when it fails
3129 * so it is safe for us to do this here. Also
3130 * it does not smudge active_cache or active_nr
3131 * when it fails, so we do not have to worry about
3132 * cleaning it up ourselves either.
3134 read_cache();
3136 if (options->abbrev <= 0 || 40 < options->abbrev)
3137 options->abbrev = 40; /* full */
3140 * It does not make sense to show the first hit we happened
3141 * to have found. It does not make sense not to return with
3142 * exit code in such a case either.
3144 if (DIFF_OPT_TST(options, QUICK)) {
3145 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3146 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3149 return 0;
3152 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3154 char c, *eq;
3155 int len;
3157 if (*arg != '-')
3158 return 0;
3159 c = *++arg;
3160 if (!c)
3161 return 0;
3162 if (c == arg_short) {
3163 c = *++arg;
3164 if (!c)
3165 return 1;
3166 if (val && isdigit(c)) {
3167 char *end;
3168 int n = strtoul(arg, &end, 10);
3169 if (*end)
3170 return 0;
3171 *val = n;
3172 return 1;
3174 return 0;
3176 if (c != '-')
3177 return 0;
3178 arg++;
3179 eq = strchr(arg, '=');
3180 if (eq)
3181 len = eq - arg;
3182 else
3183 len = strlen(arg);
3184 if (!len || strncmp(arg, arg_long, len))
3185 return 0;
3186 if (eq) {
3187 int n;
3188 char *end;
3189 if (!isdigit(*++eq))
3190 return 0;
3191 n = strtoul(eq, &end, 10);
3192 if (*end)
3193 return 0;
3194 *val = n;
3196 return 1;
3199 static int diff_scoreopt_parse(const char *opt);
3201 static inline int short_opt(char opt, const char **argv,
3202 const char **optarg)
3204 const char *arg = argv[0];
3205 if (arg[0] != '-' || arg[1] != opt)
3206 return 0;
3207 if (arg[2] != '\0') {
3208 *optarg = arg + 2;
3209 return 1;
3211 if (!argv[1])
3212 die("Option '%c' requires a value", opt);
3213 *optarg = argv[1];
3214 return 2;
3217 int parse_long_opt(const char *opt, const char **argv,
3218 const char **optarg)
3220 const char *arg = argv[0];
3221 if (arg[0] != '-' || arg[1] != '-')
3222 return 0;
3223 arg += strlen("--");
3224 if (prefixcmp(arg, opt))
3225 return 0;
3226 arg += strlen(opt);
3227 if (*arg == '=') { /* sticked form: --option=value */
3228 *optarg = arg + 1;
3229 return 1;
3231 if (*arg != '\0')
3232 return 0;
3233 /* separate form: --option value */
3234 if (!argv[1])
3235 die("Option '--%s' requires a value", opt);
3236 *optarg = argv[1];
3237 return 2;
3240 static int stat_opt(struct diff_options *options, const char **av)
3242 const char *arg = av[0];
3243 char *end;
3244 int width = options->stat_width;
3245 int name_width = options->stat_name_width;
3246 int count = options->stat_count;
3247 int argcount = 1;
3249 arg += strlen("--stat");
3250 end = (char *)arg;
3252 switch (*arg) {
3253 case '-':
3254 if (!prefixcmp(arg, "-width")) {
3255 arg += strlen("-width");
3256 if (*arg == '=')
3257 width = strtoul(arg + 1, &end, 10);
3258 else if (!*arg && !av[1])
3259 die("Option '--stat-width' requires a value");
3260 else if (!*arg) {
3261 width = strtoul(av[1], &end, 10);
3262 argcount = 2;
3264 } else if (!prefixcmp(arg, "-name-width")) {
3265 arg += strlen("-name-width");
3266 if (*arg == '=')
3267 name_width = strtoul(arg + 1, &end, 10);
3268 else if (!*arg && !av[1])
3269 die("Option '--stat-name-width' requires a value");
3270 else if (!*arg) {
3271 name_width = strtoul(av[1], &end, 10);
3272 argcount = 2;
3274 } else if (!prefixcmp(arg, "-count")) {
3275 arg += strlen("-count");
3276 if (*arg == '=')
3277 count = strtoul(arg + 1, &end, 10);
3278 else if (!*arg && !av[1])
3279 die("Option '--stat-count' requires a value");
3280 else if (!*arg) {
3281 count = strtoul(av[1], &end, 10);
3282 argcount = 2;
3285 break;
3286 case '=':
3287 width = strtoul(arg+1, &end, 10);
3288 if (*end == ',')
3289 name_width = strtoul(end+1, &end, 10);
3290 if (*end == ',')
3291 count = strtoul(end+1, &end, 10);
3294 /* Important! This checks all the error cases! */
3295 if (*end)
3296 return 0;
3297 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3298 options->stat_name_width = name_width;
3299 options->stat_width = width;
3300 options->stat_count = count;
3301 return argcount;
3304 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3306 struct strbuf errmsg = STRBUF_INIT;
3307 if (parse_dirstat_params(options, params, &errmsg))
3308 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3309 errmsg.buf);
3310 strbuf_release(&errmsg);
3312 * The caller knows a dirstat-related option is given from the command
3313 * line; allow it to say "return this_function();"
3315 options->output_format |= DIFF_FORMAT_DIRSTAT;
3316 return 1;
3319 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3321 const char *arg = av[0];
3322 const char *optarg;
3323 int argcount;
3325 /* Output format options */
3326 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3327 options->output_format |= DIFF_FORMAT_PATCH;
3328 else if (opt_arg(arg, 'U', "unified", &options->context))
3329 options->output_format |= DIFF_FORMAT_PATCH;
3330 else if (!strcmp(arg, "--raw"))
3331 options->output_format |= DIFF_FORMAT_RAW;
3332 else if (!strcmp(arg, "--patch-with-raw"))
3333 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3334 else if (!strcmp(arg, "--numstat"))
3335 options->output_format |= DIFF_FORMAT_NUMSTAT;
3336 else if (!strcmp(arg, "--shortstat"))
3337 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3338 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3339 return parse_dirstat_opt(options, "");
3340 else if (!prefixcmp(arg, "-X"))
3341 return parse_dirstat_opt(options, arg + 2);
3342 else if (!prefixcmp(arg, "--dirstat="))
3343 return parse_dirstat_opt(options, arg + 10);
3344 else if (!strcmp(arg, "--cumulative"))
3345 return parse_dirstat_opt(options, "cumulative");
3346 else if (!strcmp(arg, "--dirstat-by-file"))
3347 return parse_dirstat_opt(options, "files");
3348 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3349 parse_dirstat_opt(options, "files");
3350 return parse_dirstat_opt(options, arg + 18);
3352 else if (!strcmp(arg, "--check"))
3353 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3354 else if (!strcmp(arg, "--summary"))
3355 options->output_format |= DIFF_FORMAT_SUMMARY;
3356 else if (!strcmp(arg, "--patch-with-stat"))
3357 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3358 else if (!strcmp(arg, "--name-only"))
3359 options->output_format |= DIFF_FORMAT_NAME;
3360 else if (!strcmp(arg, "--name-status"))
3361 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3362 else if (!strcmp(arg, "-s"))
3363 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3364 else if (!prefixcmp(arg, "--stat"))
3365 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3366 return stat_opt(options, av);
3368 /* renames options */
3369 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3370 !strcmp(arg, "--break-rewrites")) {
3371 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3372 return error("invalid argument to -B: %s", arg+2);
3374 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3375 !strcmp(arg, "--find-renames")) {
3376 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3377 return error("invalid argument to -M: %s", arg+2);
3378 options->detect_rename = DIFF_DETECT_RENAME;
3380 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3381 options->irreversible_delete = 1;
3383 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3384 !strcmp(arg, "--find-copies")) {
3385 if (options->detect_rename == DIFF_DETECT_COPY)
3386 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3387 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3388 return error("invalid argument to -C: %s", arg+2);
3389 options->detect_rename = DIFF_DETECT_COPY;
3391 else if (!strcmp(arg, "--no-renames"))
3392 options->detect_rename = 0;
3393 else if (!strcmp(arg, "--relative"))
3394 DIFF_OPT_SET(options, RELATIVE_NAME);
3395 else if (!prefixcmp(arg, "--relative=")) {
3396 DIFF_OPT_SET(options, RELATIVE_NAME);
3397 options->prefix = arg + 11;
3400 /* xdiff options */
3401 else if (!strcmp(arg, "--minimal"))
3402 DIFF_XDL_SET(options, NEED_MINIMAL);
3403 else if (!strcmp(arg, "--no-minimal"))
3404 DIFF_XDL_CLR(options, NEED_MINIMAL);
3405 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3406 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3407 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3408 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3409 else if (!strcmp(arg, "--ignore-space-at-eol"))
3410 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3411 else if (!strcmp(arg, "--patience"))
3412 DIFF_XDL_SET(options, PATIENCE_DIFF);
3413 else if (!strcmp(arg, "--histogram"))
3414 DIFF_XDL_SET(options, HISTOGRAM_DIFF);
3416 /* flags options */
3417 else if (!strcmp(arg, "--binary")) {
3418 options->output_format |= DIFF_FORMAT_PATCH;
3419 DIFF_OPT_SET(options, BINARY);
3421 else if (!strcmp(arg, "--full-index"))
3422 DIFF_OPT_SET(options, FULL_INDEX);
3423 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3424 DIFF_OPT_SET(options, TEXT);
3425 else if (!strcmp(arg, "-R"))
3426 DIFF_OPT_SET(options, REVERSE_DIFF);
3427 else if (!strcmp(arg, "--find-copies-harder"))
3428 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3429 else if (!strcmp(arg, "--follow"))
3430 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3431 else if (!strcmp(arg, "--color"))
3432 options->use_color = 1;
3433 else if (!prefixcmp(arg, "--color=")) {
3434 int value = git_config_colorbool(NULL, arg+8);
3435 if (value < 0)
3436 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3437 options->use_color = value;
3439 else if (!strcmp(arg, "--no-color"))
3440 options->use_color = 0;
3441 else if (!strcmp(arg, "--color-words")) {
3442 options->use_color = 1;
3443 options->word_diff = DIFF_WORDS_COLOR;
3445 else if (!prefixcmp(arg, "--color-words=")) {
3446 options->use_color = 1;
3447 options->word_diff = DIFF_WORDS_COLOR;
3448 options->word_regex = arg + 14;
3450 else if (!strcmp(arg, "--word-diff")) {
3451 if (options->word_diff == DIFF_WORDS_NONE)
3452 options->word_diff = DIFF_WORDS_PLAIN;
3454 else if (!prefixcmp(arg, "--word-diff=")) {
3455 const char *type = arg + 12;
3456 if (!strcmp(type, "plain"))
3457 options->word_diff = DIFF_WORDS_PLAIN;
3458 else if (!strcmp(type, "color")) {
3459 options->use_color = 1;
3460 options->word_diff = DIFF_WORDS_COLOR;
3462 else if (!strcmp(type, "porcelain"))
3463 options->word_diff = DIFF_WORDS_PORCELAIN;
3464 else if (!strcmp(type, "none"))
3465 options->word_diff = DIFF_WORDS_NONE;
3466 else
3467 die("bad --word-diff argument: %s", type);
3469 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3470 if (options->word_diff == DIFF_WORDS_NONE)
3471 options->word_diff = DIFF_WORDS_PLAIN;
3472 options->word_regex = optarg;
3473 return argcount;
3475 else if (!strcmp(arg, "--exit-code"))
3476 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3477 else if (!strcmp(arg, "--quiet"))
3478 DIFF_OPT_SET(options, QUICK);
3479 else if (!strcmp(arg, "--ext-diff"))
3480 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3481 else if (!strcmp(arg, "--no-ext-diff"))
3482 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3483 else if (!strcmp(arg, "--textconv"))
3484 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3485 else if (!strcmp(arg, "--no-textconv"))
3486 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3487 else if (!strcmp(arg, "--ignore-submodules")) {
3488 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3489 handle_ignore_submodules_arg(options, "all");
3490 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3491 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3492 handle_ignore_submodules_arg(options, arg + 20);
3493 } else if (!strcmp(arg, "--submodule"))
3494 DIFF_OPT_SET(options, SUBMODULE_LOG);
3495 else if (!prefixcmp(arg, "--submodule=")) {
3496 if (!strcmp(arg + 12, "log"))
3497 DIFF_OPT_SET(options, SUBMODULE_LOG);
3500 /* misc options */
3501 else if (!strcmp(arg, "-z"))
3502 options->line_termination = 0;
3503 else if ((argcount = short_opt('l', av, &optarg))) {
3504 options->rename_limit = strtoul(optarg, NULL, 10);
3505 return argcount;
3507 else if ((argcount = short_opt('S', av, &optarg))) {
3508 options->pickaxe = optarg;
3509 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3510 return argcount;
3511 } else if ((argcount = short_opt('G', av, &optarg))) {
3512 options->pickaxe = optarg;
3513 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3514 return argcount;
3516 else if (!strcmp(arg, "--pickaxe-all"))
3517 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3518 else if (!strcmp(arg, "--pickaxe-regex"))
3519 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3520 else if ((argcount = short_opt('O', av, &optarg))) {
3521 options->orderfile = optarg;
3522 return argcount;
3524 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3525 options->filter = optarg;
3526 return argcount;
3528 else if (!strcmp(arg, "--abbrev"))
3529 options->abbrev = DEFAULT_ABBREV;
3530 else if (!prefixcmp(arg, "--abbrev=")) {
3531 options->abbrev = strtoul(arg + 9, NULL, 10);
3532 if (options->abbrev < MINIMUM_ABBREV)
3533 options->abbrev = MINIMUM_ABBREV;
3534 else if (40 < options->abbrev)
3535 options->abbrev = 40;
3537 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3538 options->a_prefix = optarg;
3539 return argcount;
3541 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3542 options->b_prefix = optarg;
3543 return argcount;
3545 else if (!strcmp(arg, "--no-prefix"))
3546 options->a_prefix = options->b_prefix = "";
3547 else if (opt_arg(arg, '\0', "inter-hunk-context",
3548 &options->interhunkcontext))
3550 else if (!strcmp(arg, "-W"))
3551 DIFF_OPT_SET(options, FUNCCONTEXT);
3552 else if (!strcmp(arg, "--function-context"))
3553 DIFF_OPT_SET(options, FUNCCONTEXT);
3554 else if (!strcmp(arg, "--no-function-context"))
3555 DIFF_OPT_CLR(options, FUNCCONTEXT);
3556 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3557 options->file = fopen(optarg, "w");
3558 if (!options->file)
3559 die_errno("Could not open '%s'", optarg);
3560 options->close_file = 1;
3561 return argcount;
3562 } else
3563 return 0;
3564 return 1;
3567 int parse_rename_score(const char **cp_p)
3569 unsigned long num, scale;
3570 int ch, dot;
3571 const char *cp = *cp_p;
3573 num = 0;
3574 scale = 1;
3575 dot = 0;
3576 for (;;) {
3577 ch = *cp;
3578 if ( !dot && ch == '.' ) {
3579 scale = 1;
3580 dot = 1;
3581 } else if ( ch == '%' ) {
3582 scale = dot ? scale*100 : 100;
3583 cp++; /* % is always at the end */
3584 break;
3585 } else if ( ch >= '0' && ch <= '9' ) {
3586 if ( scale < 100000 ) {
3587 scale *= 10;
3588 num = (num*10) + (ch-'0');
3590 } else {
3591 break;
3593 cp++;
3595 *cp_p = cp;
3597 /* user says num divided by scale and we say internally that
3598 * is MAX_SCORE * num / scale.
3600 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3603 static int diff_scoreopt_parse(const char *opt)
3605 int opt1, opt2, cmd;
3607 if (*opt++ != '-')
3608 return -1;
3609 cmd = *opt++;
3610 if (cmd == '-') {
3611 /* convert the long-form arguments into short-form versions */
3612 if (!prefixcmp(opt, "break-rewrites")) {
3613 opt += strlen("break-rewrites");
3614 if (*opt == 0 || *opt++ == '=')
3615 cmd = 'B';
3616 } else if (!prefixcmp(opt, "find-copies")) {
3617 opt += strlen("find-copies");
3618 if (*opt == 0 || *opt++ == '=')
3619 cmd = 'C';
3620 } else if (!prefixcmp(opt, "find-renames")) {
3621 opt += strlen("find-renames");
3622 if (*opt == 0 || *opt++ == '=')
3623 cmd = 'M';
3626 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3627 return -1; /* that is not a -M, -C nor -B option */
3629 opt1 = parse_rename_score(&opt);
3630 if (cmd != 'B')
3631 opt2 = 0;
3632 else {
3633 if (*opt == 0)
3634 opt2 = 0;
3635 else if (*opt != '/')
3636 return -1; /* we expect -B80/99 or -B80 */
3637 else {
3638 opt++;
3639 opt2 = parse_rename_score(&opt);
3642 if (*opt != 0)
3643 return -1;
3644 return opt1 | (opt2 << 16);
3647 struct diff_queue_struct diff_queued_diff;
3649 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3651 if (queue->alloc <= queue->nr) {
3652 queue->alloc = alloc_nr(queue->alloc);
3653 queue->queue = xrealloc(queue->queue,
3654 sizeof(dp) * queue->alloc);
3656 queue->queue[queue->nr++] = dp;
3659 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3660 struct diff_filespec *one,
3661 struct diff_filespec *two)
3663 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3664 dp->one = one;
3665 dp->two = two;
3666 if (queue)
3667 diff_q(queue, dp);
3668 return dp;
3671 void diff_free_filepair(struct diff_filepair *p)
3673 free_filespec(p->one);
3674 free_filespec(p->two);
3675 free(p);
3678 /* This is different from find_unique_abbrev() in that
3679 * it stuffs the result with dots for alignment.
3681 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3683 int abblen;
3684 const char *abbrev;
3685 if (len == 40)
3686 return sha1_to_hex(sha1);
3688 abbrev = find_unique_abbrev(sha1, len);
3689 abblen = strlen(abbrev);
3690 if (abblen < 37) {
3691 static char hex[41];
3692 if (len < abblen && abblen <= len + 2)
3693 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3694 else
3695 sprintf(hex, "%s...", abbrev);
3696 return hex;
3698 return sha1_to_hex(sha1);
3701 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3703 int line_termination = opt->line_termination;
3704 int inter_name_termination = line_termination ? '\t' : '\0';
3705 if (opt->output_prefix) {
3706 struct strbuf *msg = NULL;
3707 msg = opt->output_prefix(opt, opt->output_prefix_data);
3708 fprintf(opt->file, "%s", msg->buf);
3711 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3712 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3713 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3714 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3716 if (p->score) {
3717 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3718 inter_name_termination);
3719 } else {
3720 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3723 if (p->status == DIFF_STATUS_COPIED ||
3724 p->status == DIFF_STATUS_RENAMED) {
3725 const char *name_a, *name_b;
3726 name_a = p->one->path;
3727 name_b = p->two->path;
3728 strip_prefix(opt->prefix_length, &name_a, &name_b);
3729 write_name_quoted(name_a, opt->file, inter_name_termination);
3730 write_name_quoted(name_b, opt->file, line_termination);
3731 } else {
3732 const char *name_a, *name_b;
3733 name_a = p->one->mode ? p->one->path : p->two->path;
3734 name_b = NULL;
3735 strip_prefix(opt->prefix_length, &name_a, &name_b);
3736 write_name_quoted(name_a, opt->file, line_termination);
3740 int diff_unmodified_pair(struct diff_filepair *p)
3742 /* This function is written stricter than necessary to support
3743 * the currently implemented transformers, but the idea is to
3744 * let transformers to produce diff_filepairs any way they want,
3745 * and filter and clean them up here before producing the output.
3747 struct diff_filespec *one = p->one, *two = p->two;
3749 if (DIFF_PAIR_UNMERGED(p))
3750 return 0; /* unmerged is interesting */
3752 /* deletion, addition, mode or type change
3753 * and rename are all interesting.
3755 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3756 DIFF_PAIR_MODE_CHANGED(p) ||
3757 strcmp(one->path, two->path))
3758 return 0;
3760 /* both are valid and point at the same path. that is, we are
3761 * dealing with a change.
3763 if (one->sha1_valid && two->sha1_valid &&
3764 !hashcmp(one->sha1, two->sha1) &&
3765 !one->dirty_submodule && !two->dirty_submodule)
3766 return 1; /* no change */
3767 if (!one->sha1_valid && !two->sha1_valid)
3768 return 1; /* both look at the same file on the filesystem. */
3769 return 0;
3772 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3774 if (diff_unmodified_pair(p))
3775 return;
3777 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3778 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3779 return; /* no tree diffs in patch format */
3781 run_diff(p, o);
3784 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3785 struct diffstat_t *diffstat)
3787 if (diff_unmodified_pair(p))
3788 return;
3790 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3791 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3792 return; /* no useful stat for tree diffs */
3794 run_diffstat(p, o, diffstat);
3797 static void diff_flush_checkdiff(struct diff_filepair *p,
3798 struct diff_options *o)
3800 if (diff_unmodified_pair(p))
3801 return;
3803 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3804 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3805 return; /* nothing to check in tree diffs */
3807 run_checkdiff(p, o);
3810 int diff_queue_is_empty(void)
3812 struct diff_queue_struct *q = &diff_queued_diff;
3813 int i;
3814 for (i = 0; i < q->nr; i++)
3815 if (!diff_unmodified_pair(q->queue[i]))
3816 return 0;
3817 return 1;
3820 #if DIFF_DEBUG
3821 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3823 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3824 x, one ? one : "",
3825 s->path,
3826 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3827 s->mode,
3828 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3829 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3830 x, one ? one : "",
3831 s->size, s->xfrm_flags);
3834 void diff_debug_filepair(const struct diff_filepair *p, int i)
3836 diff_debug_filespec(p->one, i, "one");
3837 diff_debug_filespec(p->two, i, "two");
3838 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3839 p->score, p->status ? p->status : '?',
3840 p->one->rename_used, p->broken_pair);
3843 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3845 int i;
3846 if (msg)
3847 fprintf(stderr, "%s\n", msg);
3848 fprintf(stderr, "q->nr = %d\n", q->nr);
3849 for (i = 0; i < q->nr; i++) {
3850 struct diff_filepair *p = q->queue[i];
3851 diff_debug_filepair(p, i);
3854 #endif
3856 static void diff_resolve_rename_copy(void)
3858 int i;
3859 struct diff_filepair *p;
3860 struct diff_queue_struct *q = &diff_queued_diff;
3862 diff_debug_queue("resolve-rename-copy", q);
3864 for (i = 0; i < q->nr; i++) {
3865 p = q->queue[i];
3866 p->status = 0; /* undecided */
3867 if (DIFF_PAIR_UNMERGED(p))
3868 p->status = DIFF_STATUS_UNMERGED;
3869 else if (!DIFF_FILE_VALID(p->one))
3870 p->status = DIFF_STATUS_ADDED;
3871 else if (!DIFF_FILE_VALID(p->two))
3872 p->status = DIFF_STATUS_DELETED;
3873 else if (DIFF_PAIR_TYPE_CHANGED(p))
3874 p->status = DIFF_STATUS_TYPE_CHANGED;
3876 /* from this point on, we are dealing with a pair
3877 * whose both sides are valid and of the same type, i.e.
3878 * either in-place edit or rename/copy edit.
3880 else if (DIFF_PAIR_RENAME(p)) {
3882 * A rename might have re-connected a broken
3883 * pair up, causing the pathnames to be the
3884 * same again. If so, that's not a rename at
3885 * all, just a modification..
3887 * Otherwise, see if this source was used for
3888 * multiple renames, in which case we decrement
3889 * the count, and call it a copy.
3891 if (!strcmp(p->one->path, p->two->path))
3892 p->status = DIFF_STATUS_MODIFIED;
3893 else if (--p->one->rename_used > 0)
3894 p->status = DIFF_STATUS_COPIED;
3895 else
3896 p->status = DIFF_STATUS_RENAMED;
3898 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3899 p->one->mode != p->two->mode ||
3900 p->one->dirty_submodule ||
3901 p->two->dirty_submodule ||
3902 is_null_sha1(p->one->sha1))
3903 p->status = DIFF_STATUS_MODIFIED;
3904 else {
3905 /* This is a "no-change" entry and should not
3906 * happen anymore, but prepare for broken callers.
3908 error("feeding unmodified %s to diffcore",
3909 p->one->path);
3910 p->status = DIFF_STATUS_UNKNOWN;
3913 diff_debug_queue("resolve-rename-copy done", q);
3916 static int check_pair_status(struct diff_filepair *p)
3918 switch (p->status) {
3919 case DIFF_STATUS_UNKNOWN:
3920 return 0;
3921 case 0:
3922 die("internal error in diff-resolve-rename-copy");
3923 default:
3924 return 1;
3928 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3930 int fmt = opt->output_format;
3932 if (fmt & DIFF_FORMAT_CHECKDIFF)
3933 diff_flush_checkdiff(p, opt);
3934 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3935 diff_flush_raw(p, opt);
3936 else if (fmt & DIFF_FORMAT_NAME) {
3937 const char *name_a, *name_b;
3938 name_a = p->two->path;
3939 name_b = NULL;
3940 strip_prefix(opt->prefix_length, &name_a, &name_b);
3941 write_name_quoted(name_a, opt->file, opt->line_termination);
3945 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3947 if (fs->mode)
3948 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3949 else
3950 fprintf(file, " %s ", newdelete);
3951 write_name_quoted(fs->path, file, '\n');
3955 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3956 const char *line_prefix)
3958 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3959 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3960 p->two->mode, show_name ? ' ' : '\n');
3961 if (show_name) {
3962 write_name_quoted(p->two->path, file, '\n');
3967 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3968 const char *line_prefix)
3970 char *names = pprint_rename(p->one->path, p->two->path);
3972 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3973 free(names);
3974 show_mode_change(file, p, 0, line_prefix);
3977 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3979 FILE *file = opt->file;
3980 char *line_prefix = "";
3982 if (opt->output_prefix) {
3983 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3984 line_prefix = buf->buf;
3987 switch(p->status) {
3988 case DIFF_STATUS_DELETED:
3989 fputs(line_prefix, file);
3990 show_file_mode_name(file, "delete", p->one);
3991 break;
3992 case DIFF_STATUS_ADDED:
3993 fputs(line_prefix, file);
3994 show_file_mode_name(file, "create", p->two);
3995 break;
3996 case DIFF_STATUS_COPIED:
3997 fputs(line_prefix, file);
3998 show_rename_copy(file, "copy", p, line_prefix);
3999 break;
4000 case DIFF_STATUS_RENAMED:
4001 fputs(line_prefix, file);
4002 show_rename_copy(file, "rename", p, line_prefix);
4003 break;
4004 default:
4005 if (p->score) {
4006 fprintf(file, "%s rewrite ", line_prefix);
4007 write_name_quoted(p->two->path, file, ' ');
4008 fprintf(file, "(%d%%)\n", similarity_index(p));
4010 show_mode_change(file, p, !p->score, line_prefix);
4011 break;
4015 struct patch_id_t {
4016 git_SHA_CTX *ctx;
4017 int patchlen;
4020 static int remove_space(char *line, int len)
4022 int i;
4023 char *dst = line;
4024 unsigned char c;
4026 for (i = 0; i < len; i++)
4027 if (!isspace((c = line[i])))
4028 *dst++ = c;
4030 return dst - line;
4033 static void patch_id_consume(void *priv, char *line, unsigned long len)
4035 struct patch_id_t *data = priv;
4036 int new_len;
4038 /* Ignore line numbers when computing the SHA1 of the patch */
4039 if (!prefixcmp(line, "@@ -"))
4040 return;
4042 new_len = remove_space(line, len);
4044 git_SHA1_Update(data->ctx, line, new_len);
4045 data->patchlen += new_len;
4048 /* returns 0 upon success, and writes result into sha1 */
4049 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4051 struct diff_queue_struct *q = &diff_queued_diff;
4052 int i;
4053 git_SHA_CTX ctx;
4054 struct patch_id_t data;
4055 char buffer[PATH_MAX * 4 + 20];
4057 git_SHA1_Init(&ctx);
4058 memset(&data, 0, sizeof(struct patch_id_t));
4059 data.ctx = &ctx;
4061 for (i = 0; i < q->nr; i++) {
4062 xpparam_t xpp;
4063 xdemitconf_t xecfg;
4064 mmfile_t mf1, mf2;
4065 struct diff_filepair *p = q->queue[i];
4066 int len1, len2;
4068 memset(&xpp, 0, sizeof(xpp));
4069 memset(&xecfg, 0, sizeof(xecfg));
4070 if (p->status == 0)
4071 return error("internal diff status error");
4072 if (p->status == DIFF_STATUS_UNKNOWN)
4073 continue;
4074 if (diff_unmodified_pair(p))
4075 continue;
4076 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4077 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4078 continue;
4079 if (DIFF_PAIR_UNMERGED(p))
4080 continue;
4082 diff_fill_sha1_info(p->one);
4083 diff_fill_sha1_info(p->two);
4084 if (fill_mmfile(&mf1, p->one) < 0 ||
4085 fill_mmfile(&mf2, p->two) < 0)
4086 return error("unable to read files to diff");
4088 len1 = remove_space(p->one->path, strlen(p->one->path));
4089 len2 = remove_space(p->two->path, strlen(p->two->path));
4090 if (p->one->mode == 0)
4091 len1 = snprintf(buffer, sizeof(buffer),
4092 "diff--gita/%.*sb/%.*s"
4093 "newfilemode%06o"
4094 "---/dev/null"
4095 "+++b/%.*s",
4096 len1, p->one->path,
4097 len2, p->two->path,
4098 p->two->mode,
4099 len2, p->two->path);
4100 else if (p->two->mode == 0)
4101 len1 = snprintf(buffer, sizeof(buffer),
4102 "diff--gita/%.*sb/%.*s"
4103 "deletedfilemode%06o"
4104 "---a/%.*s"
4105 "+++/dev/null",
4106 len1, p->one->path,
4107 len2, p->two->path,
4108 p->one->mode,
4109 len1, p->one->path);
4110 else
4111 len1 = snprintf(buffer, sizeof(buffer),
4112 "diff--gita/%.*sb/%.*s"
4113 "---a/%.*s"
4114 "+++b/%.*s",
4115 len1, p->one->path,
4116 len2, p->two->path,
4117 len1, p->one->path,
4118 len2, p->two->path);
4119 git_SHA1_Update(&ctx, buffer, len1);
4121 if (diff_filespec_is_binary(p->one) ||
4122 diff_filespec_is_binary(p->two)) {
4123 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4124 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4125 continue;
4128 xpp.flags = 0;
4129 xecfg.ctxlen = 3;
4130 xecfg.flags = 0;
4131 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4132 &xpp, &xecfg);
4135 git_SHA1_Final(sha1, &ctx);
4136 return 0;
4139 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4141 struct diff_queue_struct *q = &diff_queued_diff;
4142 int i;
4143 int result = diff_get_patch_id(options, sha1);
4145 for (i = 0; i < q->nr; i++)
4146 diff_free_filepair(q->queue[i]);
4148 free(q->queue);
4149 DIFF_QUEUE_CLEAR(q);
4151 return result;
4154 static int is_summary_empty(const struct diff_queue_struct *q)
4156 int i;
4158 for (i = 0; i < q->nr; i++) {
4159 const struct diff_filepair *p = q->queue[i];
4161 switch (p->status) {
4162 case DIFF_STATUS_DELETED:
4163 case DIFF_STATUS_ADDED:
4164 case DIFF_STATUS_COPIED:
4165 case DIFF_STATUS_RENAMED:
4166 return 0;
4167 default:
4168 if (p->score)
4169 return 0;
4170 if (p->one->mode && p->two->mode &&
4171 p->one->mode != p->two->mode)
4172 return 0;
4173 break;
4176 return 1;
4179 static const char rename_limit_warning[] =
4180 "inexact rename detection was skipped due to too many files.";
4182 static const char degrade_cc_to_c_warning[] =
4183 "only found copies from modified paths due to too many files.";
4185 static const char rename_limit_advice[] =
4186 "you may want to set your %s variable to at least "
4187 "%d and retry the command.";
4189 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4191 if (degraded_cc)
4192 warning(degrade_cc_to_c_warning);
4193 else if (needed)
4194 warning(rename_limit_warning);
4195 else
4196 return;
4197 if (0 < needed && needed < 32767)
4198 warning(rename_limit_advice, varname, needed);
4201 void diff_flush(struct diff_options *options)
4203 struct diff_queue_struct *q = &diff_queued_diff;
4204 int i, output_format = options->output_format;
4205 int separator = 0;
4206 int dirstat_by_line = 0;
4209 * Order: raw, stat, summary, patch
4210 * or: name/name-status/checkdiff (other bits clear)
4212 if (!q->nr)
4213 goto free_queue;
4215 if (output_format & (DIFF_FORMAT_RAW |
4216 DIFF_FORMAT_NAME |
4217 DIFF_FORMAT_NAME_STATUS |
4218 DIFF_FORMAT_CHECKDIFF)) {
4219 for (i = 0; i < q->nr; i++) {
4220 struct diff_filepair *p = q->queue[i];
4221 if (check_pair_status(p))
4222 flush_one_pair(p, options);
4224 separator++;
4227 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4228 dirstat_by_line = 1;
4230 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4231 dirstat_by_line) {
4232 struct diffstat_t diffstat;
4234 memset(&diffstat, 0, sizeof(struct diffstat_t));
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_stat(p, options, &diffstat);
4240 if (output_format & DIFF_FORMAT_NUMSTAT)
4241 show_numstat(&diffstat, options);
4242 if (output_format & DIFF_FORMAT_DIFFSTAT)
4243 show_stats(&diffstat, options);
4244 if (output_format & DIFF_FORMAT_SHORTSTAT)
4245 show_shortstats(&diffstat, options);
4246 if (output_format & DIFF_FORMAT_DIRSTAT)
4247 show_dirstat_by_line(&diffstat, options);
4248 free_diffstat_info(&diffstat);
4249 separator++;
4251 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4252 show_dirstat(options);
4254 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4255 for (i = 0; i < q->nr; i++) {
4256 diff_summary(options, q->queue[i]);
4258 separator++;
4261 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4262 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4263 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4265 * run diff_flush_patch for the exit status. setting
4266 * options->file to /dev/null should be safe, becaue we
4267 * aren't supposed to produce any output anyway.
4269 if (options->close_file)
4270 fclose(options->file);
4271 options->file = fopen("/dev/null", "w");
4272 if (!options->file)
4273 die_errno("Could not open /dev/null");
4274 options->close_file = 1;
4275 for (i = 0; i < q->nr; i++) {
4276 struct diff_filepair *p = q->queue[i];
4277 if (check_pair_status(p))
4278 diff_flush_patch(p, options);
4279 if (options->found_changes)
4280 break;
4284 if (output_format & DIFF_FORMAT_PATCH) {
4285 if (separator) {
4286 putc(options->line_termination, options->file);
4287 if (options->stat_sep) {
4288 /* attach patch instead of inline */
4289 fputs(options->stat_sep, options->file);
4293 for (i = 0; i < q->nr; i++) {
4294 struct diff_filepair *p = q->queue[i];
4295 if (check_pair_status(p))
4296 diff_flush_patch(p, options);
4300 if (output_format & DIFF_FORMAT_CALLBACK)
4301 options->format_callback(q, options, options->format_callback_data);
4303 for (i = 0; i < q->nr; i++)
4304 diff_free_filepair(q->queue[i]);
4305 free_queue:
4306 free(q->queue);
4307 DIFF_QUEUE_CLEAR(q);
4308 if (options->close_file)
4309 fclose(options->file);
4312 * Report the content-level differences with HAS_CHANGES;
4313 * diff_addremove/diff_change does not set the bit when
4314 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4316 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4317 if (options->found_changes)
4318 DIFF_OPT_SET(options, HAS_CHANGES);
4319 else
4320 DIFF_OPT_CLR(options, HAS_CHANGES);
4324 static void diffcore_apply_filter(const char *filter)
4326 int i;
4327 struct diff_queue_struct *q = &diff_queued_diff;
4328 struct diff_queue_struct outq;
4329 DIFF_QUEUE_CLEAR(&outq);
4331 if (!filter)
4332 return;
4334 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4335 int found;
4336 for (i = found = 0; !found && i < q->nr; i++) {
4337 struct diff_filepair *p = q->queue[i];
4338 if (((p->status == DIFF_STATUS_MODIFIED) &&
4339 ((p->score &&
4340 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4341 (!p->score &&
4342 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4343 ((p->status != DIFF_STATUS_MODIFIED) &&
4344 strchr(filter, p->status)))
4345 found++;
4347 if (found)
4348 return;
4350 /* otherwise we will clear the whole queue
4351 * by copying the empty outq at the end of this
4352 * function, but first clear the current entries
4353 * in the queue.
4355 for (i = 0; i < q->nr; i++)
4356 diff_free_filepair(q->queue[i]);
4358 else {
4359 /* Only the matching ones */
4360 for (i = 0; i < q->nr; i++) {
4361 struct diff_filepair *p = q->queue[i];
4363 if (((p->status == DIFF_STATUS_MODIFIED) &&
4364 ((p->score &&
4365 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4366 (!p->score &&
4367 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4368 ((p->status != DIFF_STATUS_MODIFIED) &&
4369 strchr(filter, p->status)))
4370 diff_q(&outq, p);
4371 else
4372 diff_free_filepair(p);
4375 free(q->queue);
4376 *q = outq;
4379 /* Check whether two filespecs with the same mode and size are identical */
4380 static int diff_filespec_is_identical(struct diff_filespec *one,
4381 struct diff_filespec *two)
4383 if (S_ISGITLINK(one->mode))
4384 return 0;
4385 if (diff_populate_filespec(one, 0))
4386 return 0;
4387 if (diff_populate_filespec(two, 0))
4388 return 0;
4389 return !memcmp(one->data, two->data, one->size);
4392 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4394 int i;
4395 struct diff_queue_struct *q = &diff_queued_diff;
4396 struct diff_queue_struct outq;
4397 DIFF_QUEUE_CLEAR(&outq);
4399 for (i = 0; i < q->nr; i++) {
4400 struct diff_filepair *p = q->queue[i];
4403 * 1. Entries that come from stat info dirtiness
4404 * always have both sides (iow, not create/delete),
4405 * one side of the object name is unknown, with
4406 * the same mode and size. Keep the ones that
4407 * do not match these criteria. They have real
4408 * differences.
4410 * 2. At this point, the file is known to be modified,
4411 * with the same mode and size, and the object
4412 * name of one side is unknown. Need to inspect
4413 * the identical contents.
4415 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4416 !DIFF_FILE_VALID(p->two) ||
4417 (p->one->sha1_valid && p->two->sha1_valid) ||
4418 (p->one->mode != p->two->mode) ||
4419 diff_populate_filespec(p->one, 1) ||
4420 diff_populate_filespec(p->two, 1) ||
4421 (p->one->size != p->two->size) ||
4422 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4423 diff_q(&outq, p);
4424 else {
4426 * The caller can subtract 1 from skip_stat_unmatch
4427 * to determine how many paths were dirty only
4428 * due to stat info mismatch.
4430 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4431 diffopt->skip_stat_unmatch++;
4432 diff_free_filepair(p);
4435 free(q->queue);
4436 *q = outq;
4439 static int diffnamecmp(const void *a_, const void *b_)
4441 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4442 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4443 const char *name_a, *name_b;
4445 name_a = a->one ? a->one->path : a->two->path;
4446 name_b = b->one ? b->one->path : b->two->path;
4447 return strcmp(name_a, name_b);
4450 void diffcore_fix_diff_index(struct diff_options *options)
4452 struct diff_queue_struct *q = &diff_queued_diff;
4453 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4456 void diffcore_std(struct diff_options *options)
4458 if (options->skip_stat_unmatch)
4459 diffcore_skip_stat_unmatch(options);
4460 if (!options->found_follow) {
4461 /* See try_to_follow_renames() in tree-diff.c */
4462 if (options->break_opt != -1)
4463 diffcore_break(options->break_opt);
4464 if (options->detect_rename)
4465 diffcore_rename(options);
4466 if (options->break_opt != -1)
4467 diffcore_merge_broken();
4469 if (options->pickaxe)
4470 diffcore_pickaxe(options);
4471 if (options->orderfile)
4472 diffcore_order(options->orderfile);
4473 if (!options->found_follow)
4474 /* See try_to_follow_renames() in tree-diff.c */
4475 diff_resolve_rename_copy();
4476 diffcore_apply_filter(options->filter);
4478 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4479 DIFF_OPT_SET(options, HAS_CHANGES);
4480 else
4481 DIFF_OPT_CLR(options, HAS_CHANGES);
4483 options->found_follow = 0;
4486 int diff_result_code(struct diff_options *opt, int status)
4488 int result = 0;
4490 diff_warn_rename_limit("diff.renamelimit",
4491 opt->needed_rename_limit,
4492 opt->degraded_cc_to_c);
4493 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4494 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4495 return status;
4496 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4497 DIFF_OPT_TST(opt, HAS_CHANGES))
4498 result |= 01;
4499 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4500 DIFF_OPT_TST(opt, CHECK_FAILED))
4501 result |= 02;
4502 return result;
4505 int diff_can_quit_early(struct diff_options *opt)
4507 return (DIFF_OPT_TST(opt, QUICK) &&
4508 !opt->filter &&
4509 DIFF_OPT_TST(opt, HAS_CHANGES));
4513 * Shall changes to this submodule be ignored?
4515 * Submodule changes can be configured to be ignored separately for each path,
4516 * but that configuration can be overridden from the command line.
4518 static int is_submodule_ignored(const char *path, struct diff_options *options)
4520 int ignored = 0;
4521 unsigned orig_flags = options->flags;
4522 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4523 set_diffopt_flags_from_submodule_config(options, path);
4524 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4525 ignored = 1;
4526 options->flags = orig_flags;
4527 return ignored;
4530 void diff_addremove(struct diff_options *options,
4531 int addremove, unsigned mode,
4532 const unsigned char *sha1,
4533 const char *concatpath, unsigned dirty_submodule)
4535 struct diff_filespec *one, *two;
4537 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4538 return;
4540 /* This may look odd, but it is a preparation for
4541 * feeding "there are unchanged files which should
4542 * not produce diffs, but when you are doing copy
4543 * detection you would need them, so here they are"
4544 * entries to the diff-core. They will be prefixed
4545 * with something like '=' or '*' (I haven't decided
4546 * which but should not make any difference).
4547 * Feeding the same new and old to diff_change()
4548 * also has the same effect.
4549 * Before the final output happens, they are pruned after
4550 * merged into rename/copy pairs as appropriate.
4552 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4553 addremove = (addremove == '+' ? '-' :
4554 addremove == '-' ? '+' : addremove);
4556 if (options->prefix &&
4557 strncmp(concatpath, options->prefix, options->prefix_length))
4558 return;
4560 one = alloc_filespec(concatpath);
4561 two = alloc_filespec(concatpath);
4563 if (addremove != '+')
4564 fill_filespec(one, sha1, mode);
4565 if (addremove != '-') {
4566 fill_filespec(two, sha1, mode);
4567 two->dirty_submodule = dirty_submodule;
4570 diff_queue(&diff_queued_diff, one, two);
4571 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4572 DIFF_OPT_SET(options, HAS_CHANGES);
4575 void diff_change(struct diff_options *options,
4576 unsigned old_mode, unsigned new_mode,
4577 const unsigned char *old_sha1,
4578 const unsigned char *new_sha1,
4579 const char *concatpath,
4580 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4582 struct diff_filespec *one, *two;
4584 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4585 is_submodule_ignored(concatpath, options))
4586 return;
4588 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4589 unsigned tmp;
4590 const unsigned char *tmp_c;
4591 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4592 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4593 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4594 new_dirty_submodule = tmp;
4597 if (options->prefix &&
4598 strncmp(concatpath, options->prefix, options->prefix_length))
4599 return;
4601 one = alloc_filespec(concatpath);
4602 two = alloc_filespec(concatpath);
4603 fill_filespec(one, old_sha1, old_mode);
4604 fill_filespec(two, new_sha1, new_mode);
4605 one->dirty_submodule = old_dirty_submodule;
4606 two->dirty_submodule = new_dirty_submodule;
4608 diff_queue(&diff_queued_diff, one, two);
4609 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4610 DIFF_OPT_SET(options, HAS_CHANGES);
4613 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4615 struct diff_filepair *pair;
4616 struct diff_filespec *one, *two;
4618 if (options->prefix &&
4619 strncmp(path, options->prefix, options->prefix_length))
4620 return NULL;
4622 one = alloc_filespec(path);
4623 two = alloc_filespec(path);
4624 pair = diff_queue(&diff_queued_diff, one, two);
4625 pair->is_unmerged = 1;
4626 return pair;
4629 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4630 size_t *outsize)
4632 struct diff_tempfile *temp;
4633 const char *argv[3];
4634 const char **arg = argv;
4635 struct child_process child;
4636 struct strbuf buf = STRBUF_INIT;
4637 int err = 0;
4639 temp = prepare_temp_file(spec->path, spec);
4640 *arg++ = pgm;
4641 *arg++ = temp->name;
4642 *arg = NULL;
4644 memset(&child, 0, sizeof(child));
4645 child.use_shell = 1;
4646 child.argv = argv;
4647 child.out = -1;
4648 if (start_command(&child)) {
4649 remove_tempfile();
4650 return NULL;
4653 if (strbuf_read(&buf, child.out, 0) < 0)
4654 err = error("error reading from textconv command '%s'", pgm);
4655 close(child.out);
4657 if (finish_command(&child) || err) {
4658 strbuf_release(&buf);
4659 remove_tempfile();
4660 return NULL;
4662 remove_tempfile();
4664 return strbuf_detach(&buf, outsize);
4667 size_t fill_textconv(struct userdiff_driver *driver,
4668 struct diff_filespec *df,
4669 char **outbuf)
4671 size_t size;
4673 if (!driver || !driver->textconv) {
4674 if (!DIFF_FILE_VALID(df)) {
4675 *outbuf = "";
4676 return 0;
4678 if (diff_populate_filespec(df, 0))
4679 die("unable to read files to diff");
4680 *outbuf = df->data;
4681 return df->size;
4684 if (driver->textconv_cache && df->sha1_valid) {
4685 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4686 &size);
4687 if (*outbuf)
4688 return size;
4691 *outbuf = run_textconv(driver->textconv, df, &size);
4692 if (!*outbuf)
4693 die("unable to read files to diff");
4695 if (driver->textconv_cache && df->sha1_valid) {
4696 /* ignore errors, as we might be in a readonly repository */
4697 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4698 size);
4700 * we could save up changes and flush them all at the end,
4701 * but we would need an extra call after all diffing is done.
4702 * Since generating a cache entry is the slow path anyway,
4703 * this extra overhead probably isn't a big deal.
4705 notes_cache_write(driver->textconv_cache);
4708 return size;