![[tokkee]](http://tokkee.org/images/avatar.png){kind=avatar}
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 */
47 };
49 static int parse_diff_color_slot(const char *var, int ofs)
50 {
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;
68 }
70 static int parse_dirstat_params(struct diff_options *options, const char *params,
71 struct strbuf *errmsg)
72 {
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 */
100 }
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++;
107 }
108 } else {
109 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%.*s'\n"),
110 p_len, p);
111 ret++;
112 }
114 p += p_len;
116 if (*p)
117 p++; /* more parameters, swallow separator */
118 }
119 return ret;
120 }
122 static int git_config_rename(const char *var, const char *value)
123 {
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;
129 }
131 /*
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.
136 */
137 int git_diff_ui_config(const char *var, const char *value, void *cb)
138 {
139 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
140 diff_use_color_default = git_config_colorbool(var, value, -1);
141 return 0;
142 }
143 if (!strcmp(var, "diff.renames")) {
144 diff_detect_rename_default = git_config_rename(var, value);
145 return 0;
146 }
147 if (!strcmp(var, "diff.autorefreshindex")) {
148 diff_auto_refresh_index = git_config_bool(var, value);
149 return 0;
150 }
151 if (!strcmp(var, "diff.mnemonicprefix")) {
152 diff_mnemonic_prefix = git_config_bool(var, value);
153 return 0;
154 }
155 if (!strcmp(var, "diff.noprefix")) {
156 diff_no_prefix = git_config_bool(var, value);
157 return 0;
158 }
159 if (!strcmp(var, "diff.external"))
160 return git_config_string(&external_diff_cmd_cfg, var, value);
161 if (!strcmp(var, "diff.wordregex"))
162 return git_config_string(&diff_word_regex_cfg, var, value);
164 if (!strcmp(var, "diff.ignoresubmodules"))
165 handle_ignore_submodules_arg(&default_diff_options, value);
167 return git_diff_basic_config(var, value, cb);
168 }
170 int git_diff_basic_config(const char *var, const char *value, void *cb)
171 {
172 if (!strcmp(var, "diff.renamelimit")) {
173 diff_rename_limit_default = git_config_int(var, value);
174 return 0;
175 }
177 switch (userdiff_config(var, value)) {
178 case 0: break;
179 case -1: return -1;
180 default: return 0;
181 }
183 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
184 int slot = parse_diff_color_slot(var, 11);
185 if (slot < 0)
186 return 0;
187 if (!value)
188 return config_error_nonbool(var);
189 color_parse(value, var, diff_colors[slot]);
190 return 0;
191 }
193 /* like GNU diff's --suppress-blank-empty option */
194 if (!strcmp(var, "diff.suppressblankempty") ||
195 /* for backwards compatibility */
196 !strcmp(var, "diff.suppress-blank-empty")) {
197 diff_suppress_blank_empty = git_config_bool(var, value);
198 return 0;
199 }
201 if (!strcmp(var, "diff.dirstat")) {
202 struct strbuf errmsg = STRBUF_INIT;
203 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
204 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
205 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
206 errmsg.buf);
207 strbuf_release(&errmsg);
208 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
209 return 0;
210 }
212 if (!prefixcmp(var, "submodule."))
213 return parse_submodule_config_option(var, value);
215 return git_color_default_config(var, value, cb);
216 }
218 static char *quote_two(const char *one, const char *two)
219 {
220 int need_one = quote_c_style(one, NULL, NULL, 1);
221 int need_two = quote_c_style(two, NULL, NULL, 1);
222 struct strbuf res = STRBUF_INIT;
224 if (need_one + need_two) {
225 strbuf_addch(&res, '"');
226 quote_c_style(one, &res, NULL, 1);
227 quote_c_style(two, &res, NULL, 1);
228 strbuf_addch(&res, '"');
229 } else {
230 strbuf_addstr(&res, one);
231 strbuf_addstr(&res, two);
232 }
233 return strbuf_detach(&res, NULL);
234 }
236 static const char *external_diff(void)
237 {
238 static const char *external_diff_cmd = NULL;
239 static int done_preparing = 0;
241 if (done_preparing)
242 return external_diff_cmd;
243 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
244 if (!external_diff_cmd)
245 external_diff_cmd = external_diff_cmd_cfg;
246 done_preparing = 1;
247 return external_diff_cmd;
248 }
250 static struct diff_tempfile {
251 const char *name; /* filename external diff should read from */
252 char hex[41];
253 char mode[10];
254 char tmp_path[PATH_MAX];
255 } diff_temp[2];
257 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
259 struct emit_callback {
260 int color_diff;
261 unsigned ws_rule;
262 int blank_at_eof_in_preimage;
263 int blank_at_eof_in_postimage;
264 int lno_in_preimage;
265 int lno_in_postimage;
266 sane_truncate_fn truncate;
267 const char **label_path;
268 struct diff_words_data *diff_words;
269 struct diff_options *opt;
270 int *found_changesp;
271 struct strbuf *header;
272 };
274 static int count_lines(const char *data, int size)
275 {
276 int count, ch, completely_empty = 1, nl_just_seen = 0;
277 count = 0;
278 while (0 < size--) {
279 ch = *data++;
280 if (ch == '\n') {
281 count++;
282 nl_just_seen = 1;
283 completely_empty = 0;
284 }
285 else {
286 nl_just_seen = 0;
287 completely_empty = 0;
288 }
289 }
290 if (completely_empty)
291 return 0;
292 if (!nl_just_seen)
293 count++; /* no trailing newline */
294 return count;
295 }
297 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
298 {
299 if (!DIFF_FILE_VALID(one)) {
300 mf->ptr = (char *)""; /* does not matter */
301 mf->size = 0;
302 return 0;
303 }
304 else if (diff_populate_filespec(one, 0))
305 return -1;
307 mf->ptr = one->data;
308 mf->size = one->size;
309 return 0;
310 }
312 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
313 static unsigned long diff_filespec_size(struct diff_filespec *one)
314 {
315 if (!DIFF_FILE_VALID(one))
316 return 0;
317 diff_populate_filespec(one, 1);
318 return one->size;
319 }
321 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
322 {
323 char *ptr = mf->ptr;
324 long size = mf->size;
325 int cnt = 0;
327 if (!size)
328 return cnt;
329 ptr += size - 1; /* pointing at the very end */
330 if (*ptr != '\n')
331 ; /* incomplete line */
332 else
333 ptr--; /* skip the last LF */
334 while (mf->ptr < ptr) {
335 char *prev_eol;
336 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
337 if (*prev_eol == '\n')
338 break;
339 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
340 break;
341 cnt++;
342 ptr = prev_eol - 1;
343 }
344 return cnt;
345 }
347 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
348 struct emit_callback *ecbdata)
349 {
350 int l1, l2, at;
351 unsigned ws_rule = ecbdata->ws_rule;
352 l1 = count_trailing_blank(mf1, ws_rule);
353 l2 = count_trailing_blank(mf2, ws_rule);
354 if (l2 <= l1) {
355 ecbdata->blank_at_eof_in_preimage = 0;
356 ecbdata->blank_at_eof_in_postimage = 0;
357 return;
358 }
359 at = count_lines(mf1->ptr, mf1->size);
360 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
362 at = count_lines(mf2->ptr, mf2->size);
363 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
364 }
366 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
367 int first, const char *line, int len)
368 {
369 int has_trailing_newline, has_trailing_carriage_return;
370 int nofirst;
371 FILE *file = o->file;
373 if (o->output_prefix) {
374 struct strbuf *msg = NULL;
375 msg = o->output_prefix(o, o->output_prefix_data);
376 assert(msg);
377 fwrite(msg->buf, msg->len, 1, file);
378 }
380 if (len == 0) {
381 has_trailing_newline = (first == '\n');
382 has_trailing_carriage_return = (!has_trailing_newline &&
383 (first == '\r'));
384 nofirst = has_trailing_newline || has_trailing_carriage_return;
385 } else {
386 has_trailing_newline = (len > 0 && line[len-1] == '\n');
387 if (has_trailing_newline)
388 len--;
389 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
390 if (has_trailing_carriage_return)
391 len--;
392 nofirst = 0;
393 }
395 if (len || !nofirst) {
396 fputs(set, file);
397 if (!nofirst)
398 fputc(first, file);
399 fwrite(line, len, 1, file);
400 fputs(reset, file);
401 }
402 if (has_trailing_carriage_return)
403 fputc('\r', file);
404 if (has_trailing_newline)
405 fputc('\n', file);
406 }
408 static void emit_line(struct diff_options *o, const char *set, const char *reset,
409 const char *line, int len)
410 {
411 emit_line_0(o, set, reset, line[0], line+1, len-1);
412 }
414 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
415 {
416 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
417 ecbdata->blank_at_eof_in_preimage &&
418 ecbdata->blank_at_eof_in_postimage &&
419 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
420 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
421 return 0;
422 return ws_blank_line(line, len, ecbdata->ws_rule);
423 }
425 static void emit_add_line(const char *reset,
426 struct emit_callback *ecbdata,
427 const char *line, int len)
428 {
429 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
430 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
432 if (!*ws)
433 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
434 else if (new_blank_line_at_eof(ecbdata, line, len))
435 /* Blank line at EOF - paint '+' as well */
436 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
437 else {
438 /* Emit just the prefix, then the rest. */
439 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
440 ws_check_emit(line, len, ecbdata->ws_rule,
441 ecbdata->opt->file, set, reset, ws);
442 }
443 }
445 static void emit_hunk_header(struct emit_callback *ecbdata,
446 const char *line, int len)
447 {
448 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
449 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
450 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
451 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
452 static const char atat[2] = { '@', '@' };
453 const char *cp, *ep;
454 struct strbuf msgbuf = STRBUF_INIT;
455 int org_len = len;
456 int i = 1;
458 /*
459 * As a hunk header must begin with "@@ -<old>, +<new> @@",
460 * it always is at least 10 bytes long.
461 */
462 if (len < 10 ||
463 memcmp(line, atat, 2) ||
464 !(ep = memmem(line + 2, len - 2, atat, 2))) {
465 emit_line(ecbdata->opt, plain, reset, line, len);
466 return;
467 }
468 ep += 2; /* skip over @@ */
470 /* The hunk header in fraginfo color */
471 strbuf_add(&msgbuf, frag, strlen(frag));
472 strbuf_add(&msgbuf, line, ep - line);
473 strbuf_add(&msgbuf, reset, strlen(reset));
475 /*
476 * trailing "\r\n"
477 */
478 for ( ; i < 3; i++)
479 if (line[len - i] == '\r' || line[len - i] == '\n')
480 len--;
482 /* blank before the func header */
483 for (cp = ep; ep - line < len; ep++)
484 if (*ep != ' ' && *ep != '\t')
485 break;
486 if (ep != cp) {
487 strbuf_add(&msgbuf, plain, strlen(plain));
488 strbuf_add(&msgbuf, cp, ep - cp);
489 strbuf_add(&msgbuf, reset, strlen(reset));
490 }
492 if (ep < line + len) {
493 strbuf_add(&msgbuf, func, strlen(func));
494 strbuf_add(&msgbuf, ep, line + len - ep);
495 strbuf_add(&msgbuf, reset, strlen(reset));
496 }
498 strbuf_add(&msgbuf, line + len, org_len - len);
499 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
500 strbuf_release(&msgbuf);
501 }
503 static struct diff_tempfile *claim_diff_tempfile(void) {
504 int i;
505 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
506 if (!diff_temp[i].name)
507 return diff_temp + i;
508 die("BUG: diff is failing to clean up its tempfiles");
509 }
511 static int remove_tempfile_installed;
513 static void remove_tempfile(void)
514 {
515 int i;
516 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
517 if (diff_temp[i].name == diff_temp[i].tmp_path)
518 unlink_or_warn(diff_temp[i].name);
519 diff_temp[i].name = NULL;
520 }
521 }
523 static void remove_tempfile_on_signal(int signo)
524 {
525 remove_tempfile();
526 sigchain_pop(signo);
527 raise(signo);
528 }
530 static void print_line_count(FILE *file, int count)
531 {
532 switch (count) {
533 case 0:
534 fprintf(file, "0,0");
535 break;
536 case 1:
537 fprintf(file, "1");
538 break;
539 default:
540 fprintf(file, "1,%d", count);
541 break;
542 }
543 }
545 static void emit_rewrite_lines(struct emit_callback *ecb,
546 int prefix, const char *data, int size)
547 {
548 const char *endp = NULL;
549 static const char *nneof = " No newline at end of file\n";
550 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
551 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
553 while (0 < size) {
554 int len;
556 endp = memchr(data, '\n', size);
557 len = endp ? (endp - data + 1) : size;
558 if (prefix != '+') {
559 ecb->lno_in_preimage++;
560 emit_line_0(ecb->opt, old, reset, '-',
561 data, len);
562 } else {
563 ecb->lno_in_postimage++;
564 emit_add_line(reset, ecb, data, len);
565 }
566 size -= len;
567 data += len;
568 }
569 if (!endp) {
570 const char *plain = diff_get_color(ecb->color_diff,
571 DIFF_PLAIN);
572 emit_line_0(ecb->opt, plain, reset, '\\',
573 nneof, strlen(nneof));
574 }
575 }
577 static void emit_rewrite_diff(const char *name_a,
578 const char *name_b,
579 struct diff_filespec *one,
580 struct diff_filespec *two,
581 struct userdiff_driver *textconv_one,
582 struct userdiff_driver *textconv_two,
583 struct diff_options *o)
584 {
585 int lc_a, lc_b;
586 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
587 const char *name_a_tab, *name_b_tab;
588 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
589 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
590 const char *reset = diff_get_color(color_diff, DIFF_RESET);
591 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
592 const char *a_prefix, *b_prefix;
593 char *data_one, *data_two;
594 size_t size_one, size_two;
595 struct emit_callback ecbdata;
596 char *line_prefix = "";
597 struct strbuf *msgbuf;
599 if (o && o->output_prefix) {
600 msgbuf = o->output_prefix(o, o->output_prefix_data);
601 line_prefix = msgbuf->buf;
602 }
604 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
605 a_prefix = o->b_prefix;
606 b_prefix = o->a_prefix;
607 } else {
608 a_prefix = o->a_prefix;
609 b_prefix = o->b_prefix;
610 }
612 name_a += (*name_a == '/');
613 name_b += (*name_b == '/');
614 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
615 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
617 strbuf_reset(&a_name);
618 strbuf_reset(&b_name);
619 quote_two_c_style(&a_name, a_prefix, name_a, 0);
620 quote_two_c_style(&b_name, b_prefix, name_b, 0);
622 size_one = fill_textconv(textconv_one, one, &data_one);
623 size_two = fill_textconv(textconv_two, two, &data_two);
625 memset(&ecbdata, 0, sizeof(ecbdata));
626 ecbdata.color_diff = color_diff;
627 ecbdata.found_changesp = &o->found_changes;
628 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
629 ecbdata.opt = o;
630 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
631 mmfile_t mf1, mf2;
632 mf1.ptr = (char *)data_one;
633 mf2.ptr = (char *)data_two;
634 mf1.size = size_one;
635 mf2.size = size_two;
636 check_blank_at_eof(&mf1, &mf2, &ecbdata);
637 }
638 ecbdata.lno_in_preimage = 1;
639 ecbdata.lno_in_postimage = 1;
641 lc_a = count_lines(data_one, size_one);
642 lc_b = count_lines(data_two, size_two);
643 fprintf(o->file,
644 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
645 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
646 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
647 line_prefix, fraginfo);
648 if (!o->irreversible_delete)
649 print_line_count(o->file, lc_a);
650 else
651 fprintf(o->file, "?,?");
652 fprintf(o->file, " +");
653 print_line_count(o->file, lc_b);
654 fprintf(o->file, " @@%s\n", reset);
655 if (lc_a && !o->irreversible_delete)
656 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
657 if (lc_b)
658 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
659 if (textconv_one)
660 free((char *)data_one);
661 if (textconv_two)
662 free((char *)data_two);
663 }
665 struct diff_words_buffer {
666 mmfile_t text;
667 long alloc;
668 struct diff_words_orig {
669 const char *begin, *end;
670 } *orig;
671 int orig_nr, orig_alloc;
672 };
674 static void diff_words_append(char *line, unsigned long len,
675 struct diff_words_buffer *buffer)
676 {
677 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
678 line++;
679 len--;
680 memcpy(buffer->text.ptr + buffer->text.size, line, len);
681 buffer->text.size += len;
682 buffer->text.ptr[buffer->text.size] = '\0';
683 }
685 struct diff_words_style_elem {
686 const char *prefix;
687 const char *suffix;
688 const char *color; /* NULL; filled in by the setup code if
689 * color is enabled */
690 };
692 struct diff_words_style {
693 enum diff_words_type type;
694 struct diff_words_style_elem new, old, ctx;
695 const char *newline;
696 };
698 static struct diff_words_style diff_words_styles[] = {
699 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
700 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
701 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
702 };
704 struct diff_words_data {
705 struct diff_words_buffer minus, plus;
706 const char *current_plus;
707 int last_minus;
708 struct diff_options *opt;
709 regex_t *word_regex;
710 enum diff_words_type type;
711 struct diff_words_style *style;
712 };
714 static int fn_out_diff_words_write_helper(FILE *fp,
715 struct diff_words_style_elem *st_el,
716 const char *newline,
717 size_t count, const char *buf,
718 const char *line_prefix)
719 {
720 int print = 0;
722 while (count) {
723 char *p = memchr(buf, '\n', count);
724 if (print)
725 fputs(line_prefix, fp);
726 if (p != buf) {
727 if (st_el->color && fputs(st_el->color, fp) < 0)
728 return -1;
729 if (fputs(st_el->prefix, fp) < 0 ||
730 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
731 fputs(st_el->suffix, fp) < 0)
732 return -1;
733 if (st_el->color && *st_el->color
734 && fputs(GIT_COLOR_RESET, fp) < 0)
735 return -1;
736 }
737 if (!p)
738 return 0;
739 if (fputs(newline, fp) < 0)
740 return -1;
741 count -= p + 1 - buf;
742 buf = p + 1;
743 print = 1;
744 }
745 return 0;
746 }
748 /*
749 * '--color-words' algorithm can be described as:
750 *
751 * 1. collect a the minus/plus lines of a diff hunk, divided into
752 * minus-lines and plus-lines;
753 *
754 * 2. break both minus-lines and plus-lines into words and
755 * place them into two mmfile_t with one word for each line;
756 *
757 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
758 *
759 * And for the common parts of the both file, we output the plus side text.
760 * diff_words->current_plus is used to trace the current position of the plus file
761 * which printed. diff_words->last_minus is used to trace the last minus word
762 * printed.
763 *
764 * For '--graph' to work with '--color-words', we need to output the graph prefix
765 * on each line of color words output. Generally, there are two conditions on
766 * which we should output the prefix.
767 *
768 * 1. diff_words->last_minus == 0 &&
769 * diff_words->current_plus == diff_words->plus.text.ptr
770 *
771 * that is: the plus text must start as a new line, and if there is no minus
772 * word printed, a graph prefix must be printed.
773 *
774 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
775 * *(diff_words->current_plus - 1) == '\n'
776 *
777 * that is: a graph prefix must be printed following a '\n'
778 */
779 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
780 {
781 if ((diff_words->last_minus == 0 &&
782 diff_words->current_plus == diff_words->plus.text.ptr) ||
783 (diff_words->current_plus > diff_words->plus.text.ptr &&
784 *(diff_words->current_plus - 1) == '\n')) {
785 return 1;
786 } else {
787 return 0;
788 }
789 }
791 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
792 {
793 struct diff_words_data *diff_words = priv;
794 struct diff_words_style *style = diff_words->style;
795 int minus_first, minus_len, plus_first, plus_len;
796 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
797 struct diff_options *opt = diff_words->opt;
798 struct strbuf *msgbuf;
799 char *line_prefix = "";
801 if (line[0] != '@' || parse_hunk_header(line, len,
802 &minus_first, &minus_len, &plus_first, &plus_len))
803 return;
805 assert(opt);
806 if (opt->output_prefix) {
807 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
808 line_prefix = msgbuf->buf;
809 }
811 /* POSIX requires that first be decremented by one if len == 0... */
812 if (minus_len) {
813 minus_begin = diff_words->minus.orig[minus_first].begin;
814 minus_end =
815 diff_words->minus.orig[minus_first + minus_len - 1].end;
816 } else
817 minus_begin = minus_end =
818 diff_words->minus.orig[minus_first].end;
820 if (plus_len) {
821 plus_begin = diff_words->plus.orig[plus_first].begin;
822 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
823 } else
824 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
826 if (color_words_output_graph_prefix(diff_words)) {
827 fputs(line_prefix, diff_words->opt->file);
828 }
829 if (diff_words->current_plus != plus_begin) {
830 fn_out_diff_words_write_helper(diff_words->opt->file,
831 &style->ctx, style->newline,
832 plus_begin - diff_words->current_plus,
833 diff_words->current_plus, line_prefix);
834 if (*(plus_begin - 1) == '\n')
835 fputs(line_prefix, diff_words->opt->file);
836 }
837 if (minus_begin != minus_end) {
838 fn_out_diff_words_write_helper(diff_words->opt->file,
839 &style->old, style->newline,
840 minus_end - minus_begin, minus_begin,
841 line_prefix);
842 }
843 if (plus_begin != plus_end) {
844 fn_out_diff_words_write_helper(diff_words->opt->file,
845 &style->new, style->newline,
846 plus_end - plus_begin, plus_begin,
847 line_prefix);
848 }
850 diff_words->current_plus = plus_end;
851 diff_words->last_minus = minus_first;
852 }
854 /* This function starts looking at *begin, and returns 0 iff a word was found. */
855 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
856 int *begin, int *end)
857 {
858 if (word_regex && *begin < buffer->size) {
859 regmatch_t match[1];
860 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
861 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
862 '\n', match[0].rm_eo - match[0].rm_so);
863 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
864 *begin += match[0].rm_so;
865 return *begin >= *end;
866 }
867 return -1;
868 }
870 /* find the next word */
871 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
872 (*begin)++;
873 if (*begin >= buffer->size)
874 return -1;
876 /* find the end of the word */
877 *end = *begin + 1;
878 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
879 (*end)++;
881 return 0;
882 }
884 /*
885 * This function splits the words in buffer->text, stores the list with
886 * newline separator into out, and saves the offsets of the original words
887 * in buffer->orig.
888 */
889 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
890 regex_t *word_regex)
891 {
892 int i, j;
893 long alloc = 0;
895 out->size = 0;
896 out->ptr = NULL;
898 /* fake an empty "0th" word */
899 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
900 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
901 buffer->orig_nr = 1;
903 for (i = 0; i < buffer->text.size; i++) {
904 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
905 return;
907 /* store original boundaries */
908 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
909 buffer->orig_alloc);
910 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
911 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
912 buffer->orig_nr++;
914 /* store one word */
915 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
916 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
917 out->ptr[out->size + j - i] = '\n';
918 out->size += j - i + 1;
920 i = j - 1;
921 }
922 }
924 /* this executes the word diff on the accumulated buffers */
925 static void diff_words_show(struct diff_words_data *diff_words)
926 {
927 xpparam_t xpp;
928 xdemitconf_t xecfg;
929 mmfile_t minus, plus;
930 struct diff_words_style *style = diff_words->style;
932 struct diff_options *opt = diff_words->opt;
933 struct strbuf *msgbuf;
934 char *line_prefix = "";
936 assert(opt);
937 if (opt->output_prefix) {
938 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
939 line_prefix = msgbuf->buf;
940 }
942 /* special case: only removal */
943 if (!diff_words->plus.text.size) {
944 fputs(line_prefix, diff_words->opt->file);
945 fn_out_diff_words_write_helper(diff_words->opt->file,
946 &style->old, style->newline,
947 diff_words->minus.text.size,
948 diff_words->minus.text.ptr, line_prefix);
949 diff_words->minus.text.size = 0;
950 return;
951 }
953 diff_words->current_plus = diff_words->plus.text.ptr;
954 diff_words->last_minus = 0;
956 memset(&xpp, 0, sizeof(xpp));
957 memset(&xecfg, 0, sizeof(xecfg));
958 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
959 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
960 xpp.flags = 0;
961 /* as only the hunk header will be parsed, we need a 0-context */
962 xecfg.ctxlen = 0;
963 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
964 &xpp, &xecfg);
965 free(minus.ptr);
966 free(plus.ptr);
967 if (diff_words->current_plus != diff_words->plus.text.ptr +
968 diff_words->plus.text.size) {
969 if (color_words_output_graph_prefix(diff_words))
970 fputs(line_prefix, diff_words->opt->file);
971 fn_out_diff_words_write_helper(diff_words->opt->file,
972 &style->ctx, style->newline,
973 diff_words->plus.text.ptr + diff_words->plus.text.size
974 - diff_words->current_plus, diff_words->current_plus,
975 line_prefix);
976 }
977 diff_words->minus.text.size = diff_words->plus.text.size = 0;
978 }
980 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
981 static void diff_words_flush(struct emit_callback *ecbdata)
982 {
983 if (ecbdata->diff_words->minus.text.size ||
984 ecbdata->diff_words->plus.text.size)
985 diff_words_show(ecbdata->diff_words);
986 }
988 static void free_diff_words_data(struct emit_callback *ecbdata)
989 {
990 if (ecbdata->diff_words) {
991 diff_words_flush(ecbdata);
992 free (ecbdata->diff_words->minus.text.ptr);
993 free (ecbdata->diff_words->minus.orig);
994 free (ecbdata->diff_words->plus.text.ptr);
995 free (ecbdata->diff_words->plus.orig);
996 if (ecbdata->diff_words->word_regex) {
997 regfree(ecbdata->diff_words->word_regex);
998 free(ecbdata->diff_words->word_regex);
999 }
1000 free(ecbdata->diff_words);
1001 ecbdata->diff_words = NULL;
1002 }
1003 }
1005 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1006 {
1007 if (diff_use_color)
1008 return diff_colors[ix];
1009 return "";
1010 }
1012 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1013 {
1014 const char *cp;
1015 unsigned long allot;
1016 size_t l = len;
1018 if (ecb->truncate)
1019 return ecb->truncate(line, len);
1020 cp = line;
1021 allot = l;
1022 while (0 < l) {
1023 (void) utf8_width(&cp, &l);
1024 if (!cp)
1025 break; /* truncated in the middle? */
1026 }
1027 return allot - l;
1028 }
1030 static void find_lno(const char *line, struct emit_callback *ecbdata)
1031 {
1032 const char *p;
1033 ecbdata->lno_in_preimage = 0;
1034 ecbdata->lno_in_postimage = 0;
1035 p = strchr(line, '-');
1036 if (!p)
1037 return; /* cannot happen */
1038 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1039 p = strchr(p, '+');
1040 if (!p)
1041 return; /* cannot happen */
1042 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1043 }
1045 static void fn_out_consume(void *priv, char *line, unsigned long len)
1046 {
1047 struct emit_callback *ecbdata = priv;
1048 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1049 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1050 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1051 struct diff_options *o = ecbdata->opt;
1052 char *line_prefix = "";
1053 struct strbuf *msgbuf;
1055 if (o && o->output_prefix) {
1056 msgbuf = o->output_prefix(o, o->output_prefix_data);
1057 line_prefix = msgbuf->buf;
1058 }
1060 if (ecbdata->header) {
1061 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1062 strbuf_reset(ecbdata->header);
1063 ecbdata->header = NULL;
1064 }
1065 *(ecbdata->found_changesp) = 1;
1067 if (ecbdata->label_path[0]) {
1068 const char *name_a_tab, *name_b_tab;
1070 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1071 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1073 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1074 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1075 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1076 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1077 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1078 }
1080 if (diff_suppress_blank_empty
1081 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1082 line[0] = '\n';
1083 len = 1;
1084 }
1086 if (line[0] == '@') {
1087 if (ecbdata->diff_words)
1088 diff_words_flush(ecbdata);
1089 len = sane_truncate_line(ecbdata, line, len);
1090 find_lno(line, ecbdata);
1091 emit_hunk_header(ecbdata, line, len);
1092 if (line[len-1] != '\n')
1093 putc('\n', ecbdata->opt->file);
1094 return;
1095 }
1097 if (len < 1) {
1098 emit_line(ecbdata->opt, reset, reset, line, len);
1099 if (ecbdata->diff_words
1100 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1101 fputs("~\n", ecbdata->opt->file);
1102 return;
1103 }
1105 if (ecbdata->diff_words) {
1106 if (line[0] == '-') {
1107 diff_words_append(line, len,
1108 &ecbdata->diff_words->minus);
1109 return;
1110 } else if (line[0] == '+') {
1111 diff_words_append(line, len,
1112 &ecbdata->diff_words->plus);
1113 return;
1114 }
1115 diff_words_flush(ecbdata);
1116 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1117 emit_line(ecbdata->opt, plain, reset, line, len);
1118 fputs("~\n", ecbdata->opt->file);
1119 } else {
1120 /* don't print the prefix character */
1121 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1122 }
1123 return;
1124 }
1126 if (line[0] != '+') {
1127 const char *color =
1128 diff_get_color(ecbdata->color_diff,
1129 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1130 ecbdata->lno_in_preimage++;
1131 if (line[0] == ' ')
1132 ecbdata->lno_in_postimage++;
1133 emit_line(ecbdata->opt, color, reset, line, len);
1134 } else {
1135 ecbdata->lno_in_postimage++;
1136 emit_add_line(reset, ecbdata, line + 1, len - 1);
1137 }
1138 }
1140 static char *pprint_rename(const char *a, const char *b)
1141 {
1142 const char *old = a;
1143 const char *new = b;
1144 struct strbuf name = STRBUF_INIT;
1145 int pfx_length, sfx_length;
1146 int len_a = strlen(a);
1147 int len_b = strlen(b);
1148 int a_midlen, b_midlen;
1149 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1150 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1152 if (qlen_a || qlen_b) {
1153 quote_c_style(a, &name, NULL, 0);
1154 strbuf_addstr(&name, " => ");
1155 quote_c_style(b, &name, NULL, 0);
1156 return strbuf_detach(&name, NULL);
1157 }
1159 /* Find common prefix */
1160 pfx_length = 0;
1161 while (*old && *new && *old == *new) {
1162 if (*old == '/')
1163 pfx_length = old - a + 1;
1164 old++;
1165 new++;
1166 }
1168 /* Find common suffix */
1169 old = a + len_a;
1170 new = b + len_b;
1171 sfx_length = 0;
1172 while (a <= old && b <= new && *old == *new) {
1173 if (*old == '/')
1174 sfx_length = len_a - (old - a);
1175 old--;
1176 new--;
1177 }
1179 /*
1180 * pfx{mid-a => mid-b}sfx
1181 * {pfx-a => pfx-b}sfx
1182 * pfx{sfx-a => sfx-b}
1183 * name-a => name-b
1184 */
1185 a_midlen = len_a - pfx_length - sfx_length;
1186 b_midlen = len_b - pfx_length - sfx_length;
1187 if (a_midlen < 0)
1188 a_midlen = 0;
1189 if (b_midlen < 0)
1190 b_midlen = 0;
1192 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1193 if (pfx_length + sfx_length) {
1194 strbuf_add(&name, a, pfx_length);
1195 strbuf_addch(&name, '{');
1196 }
1197 strbuf_add(&name, a + pfx_length, a_midlen);
1198 strbuf_addstr(&name, " => ");
1199 strbuf_add(&name, b + pfx_length, b_midlen);
1200 if (pfx_length + sfx_length) {
1201 strbuf_addch(&name, '}');
1202 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1203 }
1204 return strbuf_detach(&name, NULL);
1205 }
1207 struct diffstat_t {
1208 int nr;
1209 int alloc;
1210 struct diffstat_file {
1211 char *from_name;
1212 char *name;
1213 char *print_name;
1214 unsigned is_unmerged:1;
1215 unsigned is_binary:1;
1216 unsigned is_renamed:1;
1217 uintmax_t added, deleted;
1218 } **files;
1219 };
1221 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1222 const char *name_a,
1223 const char *name_b)
1224 {
1225 struct diffstat_file *x;
1226 x = xcalloc(sizeof (*x), 1);
1227 if (diffstat->nr == diffstat->alloc) {
1228 diffstat->alloc = alloc_nr(diffstat->alloc);
1229 diffstat->files = xrealloc(diffstat->files,
1230 diffstat->alloc * sizeof(x));
1231 }
1232 diffstat->files[diffstat->nr++] = x;
1233 if (name_b) {
1234 x->from_name = xstrdup(name_a);
1235 x->name = xstrdup(name_b);
1236 x->is_renamed = 1;
1237 }
1238 else {
1239 x->from_name = NULL;
1240 x->name = xstrdup(name_a);
1241 }
1242 return x;
1243 }
1245 static void diffstat_consume(void *priv, char *line, unsigned long len)
1246 {
1247 struct diffstat_t *diffstat = priv;
1248 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1250 if (line[0] == '+')
1251 x->added++;
1252 else if (line[0] == '-')
1253 x->deleted++;
1254 }
1256 const char mime_boundary_leader[] = "------------";
1258 static int scale_linear(int it, int width, int max_change)
1259 {
1260 /*
1261 * make sure that at least one '-' is printed if there were deletions,
1262 * and likewise for '+'.
1263 */
1264 if (max_change < 2)
1265 return it;
1266 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1267 }
1269 static void show_name(FILE *file,
1270 const char *prefix, const char *name, int len)
1271 {
1272 fprintf(file, " %s%-*s |", prefix, len, name);
1273 }
1275 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1276 {
1277 if (cnt <= 0)
1278 return;
1279 fprintf(file, "%s", set);
1280 while (cnt--)
1281 putc(ch, file);
1282 fprintf(file, "%s", reset);
1283 }
1285 static void fill_print_name(struct diffstat_file *file)
1286 {
1287 char *pname;
1289 if (file->print_name)
1290 return;
1292 if (!file->is_renamed) {
1293 struct strbuf buf = STRBUF_INIT;
1294 if (quote_c_style(file->name, &buf, NULL, 0)) {
1295 pname = strbuf_detach(&buf, NULL);
1296 } else {
1297 pname = file->name;
1298 strbuf_release(&buf);
1299 }
1300 } else {
1301 pname = pprint_rename(file->from_name, file->name);
1302 }
1303 file->print_name = pname;
1304 }
1306 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1307 {
1308 int i, len, add, del, adds = 0, dels = 0;
1309 uintmax_t max_change = 0, max_len = 0;
1310 int total_files = data->nr;
1311 int width, name_width;
1312 const char *reset, *add_c, *del_c;
1313 const char *line_prefix = "";
1314 struct strbuf *msg = NULL;
1316 if (data->nr == 0)
1317 return;
1319 if (options->output_prefix) {
1320 msg = options->output_prefix(options, options->output_prefix_data);
1321 line_prefix = msg->buf;
1322 }
1324 width = options->stat_width ? options->stat_width : 80;
1325 name_width = options->stat_name_width ? options->stat_name_width : 50;
1327 /* Sanity: give at least 5 columns to the graph,
1328 * but leave at least 10 columns for the name.
1329 */
1330 if (width < 25)
1331 width = 25;
1332 if (name_width < 10)
1333 name_width = 10;
1334 else if (width < name_width + 15)
1335 name_width = width - 15;
1337 /* Find the longest filename and max number of changes */
1338 reset = diff_get_color_opt(options, DIFF_RESET);
1339 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1340 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1342 for (i = 0; i < data->nr; i++) {
1343 struct diffstat_file *file = data->files[i];
1344 uintmax_t change = file->added + file->deleted;
1345 fill_print_name(file);
1346 len = strlen(file->print_name);
1347 if (max_len < len)
1348 max_len = len;
1350 if (file->is_binary || file->is_unmerged)
1351 continue;
1352 if (max_change < change)
1353 max_change = change;
1354 }
1356 /* Compute the width of the graph part;
1357 * 10 is for one blank at the beginning of the line plus
1358 * " | count " between the name and the graph.
1359 *
1360 * From here on, name_width is the width of the name area,
1361 * and width is the width of the graph area.
1362 */
1363 name_width = (name_width < max_len) ? name_width : max_len;
1364 if (width < (name_width + 10) + max_change)
1365 width = width - (name_width + 10);
1366 else
1367 width = max_change;
1369 for (i = 0; i < data->nr; i++) {
1370 const char *prefix = "";
1371 char *name = data->files[i]->print_name;
1372 uintmax_t added = data->files[i]->added;
1373 uintmax_t deleted = data->files[i]->deleted;
1374 int name_len;
1376 /*
1377 * "scale" the filename
1378 */
1379 len = name_width;
1380 name_len = strlen(name);
1381 if (name_width < name_len) {
1382 char *slash;
1383 prefix = "...";
1384 len -= 3;
1385 name += name_len - len;
1386 slash = strchr(name, '/');
1387 if (slash)
1388 name = slash;
1389 }
1391 if (data->files[i]->is_binary) {
1392 fprintf(options->file, "%s", line_prefix);
1393 show_name(options->file, prefix, name, len);
1394 fprintf(options->file, " Bin ");
1395 fprintf(options->file, "%s%"PRIuMAX"%s",
1396 del_c, deleted, reset);
1397 fprintf(options->file, " -> ");
1398 fprintf(options->file, "%s%"PRIuMAX"%s",
1399 add_c, added, reset);
1400 fprintf(options->file, " bytes");
1401 fprintf(options->file, "\n");
1402 continue;
1403 }
1404 else if (data->files[i]->is_unmerged) {
1405 fprintf(options->file, "%s", line_prefix);
1406 show_name(options->file, prefix, name, len);
1407 fprintf(options->file, " Unmerged\n");
1408 continue;
1409 }
1410 else if (!data->files[i]->is_renamed &&
1411 (added + deleted == 0)) {
1412 total_files--;
1413 continue;
1414 }
1416 /*
1417 * scale the add/delete
1418 */
1419 add = added;
1420 del = deleted;
1421 adds += add;
1422 dels += del;
1424 if (width <= max_change) {
1425 add = scale_linear(add, width, max_change);
1426 del = scale_linear(del, width, max_change);
1427 }
1428 fprintf(options->file, "%s", line_prefix);
1429 show_name(options->file, prefix, name, len);
1430 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1431 added + deleted ? " " : "");
1432 show_graph(options->file, '+', add, add_c, reset);
1433 show_graph(options->file, '-', del, del_c, reset);
1434 fprintf(options->file, "\n");
1435 }
1436 fprintf(options->file, "%s", line_prefix);
1437 fprintf(options->file,
1438 " %d files changed, %d insertions(+), %d deletions(-)\n",
1439 total_files, adds, dels);
1440 }
1442 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1443 {
1444 int i, adds = 0, dels = 0, total_files = data->nr;
1446 if (data->nr == 0)
1447 return;
1449 for (i = 0; i < data->nr; i++) {
1450 if (!data->files[i]->is_binary &&
1451 !data->files[i]->is_unmerged) {
1452 int added = data->files[i]->added;
1453 int deleted= data->files[i]->deleted;
1454 if (!data->files[i]->is_renamed &&
1455 (added + deleted == 0)) {
1456 total_files--;
1457 } else {
1458 adds += added;
1459 dels += deleted;
1460 }
1461 }
1462 }
1463 if (options->output_prefix) {
1464 struct strbuf *msg = NULL;
1465 msg = options->output_prefix(options,
1466 options->output_prefix_data);
1467 fprintf(options->file, "%s", msg->buf);
1468 }
1469 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1470 total_files, adds, dels);
1471 }
1473 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1474 {
1475 int i;
1477 if (data->nr == 0)
1478 return;
1480 for (i = 0; i < data->nr; i++) {
1481 struct diffstat_file *file = data->files[i];
1483 if (options->output_prefix) {
1484 struct strbuf *msg = NULL;
1485 msg = options->output_prefix(options,
1486 options->output_prefix_data);
1487 fprintf(options->file, "%s", msg->buf);
1488 }
1490 if (file->is_binary)
1491 fprintf(options->file, "-\t-\t");
1492 else
1493 fprintf(options->file,
1494 "%"PRIuMAX"\t%"PRIuMAX"\t",
1495 file->added, file->deleted);
1496 if (options->line_termination) {
1497 fill_print_name(file);
1498 if (!file->is_renamed)
1499 write_name_quoted(file->name, options->file,
1500 options->line_termination);
1501 else {
1502 fputs(file->print_name, options->file);
1503 putc(options->line_termination, options->file);
1504 }
1505 } else {
1506 if (file->is_renamed) {
1507 putc('\0', options->file);
1508 write_name_quoted(file->from_name, options->file, '\0');
1509 }
1510 write_name_quoted(file->name, options->file, '\0');
1511 }
1512 }
1513 }
1515 struct dirstat_file {
1516 const char *name;
1517 unsigned long changed;
1518 };
1520 struct dirstat_dir {
1521 struct dirstat_file *files;
1522 int alloc, nr, permille, cumulative;
1523 };
1525 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1526 unsigned long changed, const char *base, int baselen)
1527 {
1528 unsigned long this_dir = 0;
1529 unsigned int sources = 0;
1530 const char *line_prefix = "";
1531 struct strbuf *msg = NULL;
1533 if (opt->output_prefix) {
1534 msg = opt->output_prefix(opt, opt->output_prefix_data);
1535 line_prefix = msg->buf;
1536 }
1538 while (dir->nr) {
1539 struct dirstat_file *f = dir->files;
1540 int namelen = strlen(f->name);
1541 unsigned long this;
1542 char *slash;
1544 if (namelen < baselen)
1545 break;
1546 if (memcmp(f->name, base, baselen))
1547 break;
1548 slash = strchr(f->name + baselen, '/');
1549 if (slash) {
1550 int newbaselen = slash + 1 - f->name;
1551 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1552 sources++;
1553 } else {
1554 this = f->changed;
1555 dir->files++;
1556 dir->nr--;
1557 sources += 2;
1558 }
1559 this_dir += this;
1560 }
1562 /*
1563 * We don't report dirstat's for
1564 * - the top level
1565 * - or cases where everything came from a single directory
1566 * under this directory (sources == 1).
1567 */
1568 if (baselen && sources != 1) {
1569 if (this_dir) {
1570 int permille = this_dir * 1000 / changed;
1571 if (permille >= dir->permille) {
1572 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1573 permille / 10, permille % 10, baselen, base);
1574 if (!dir->cumulative)
1575 return 0;
1576 }
1577 }
1578 }
1579 return this_dir;
1580 }
1582 static int dirstat_compare(const void *_a, const void *_b)
1583 {
1584 const struct dirstat_file *a = _a;
1585 const struct dirstat_file *b = _b;
1586 return strcmp(a->name, b->name);
1587 }
1589 static void show_dirstat(struct diff_options *options)
1590 {
1591 int i;
1592 unsigned long changed;
1593 struct dirstat_dir dir;
1594 struct diff_queue_struct *q = &diff_queued_diff;
1596 dir.files = NULL;
1597 dir.alloc = 0;
1598 dir.nr = 0;
1599 dir.permille = options->dirstat_permille;
1600 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1602 changed = 0;
1603 for (i = 0; i < q->nr; i++) {
1604 struct diff_filepair *p = q->queue[i];
1605 const char *name;
1606 unsigned long copied, added, damage;
1607 int content_changed;
1609 name = p->two->path ? p->two->path : p->one->path;
1611 if (p->one->sha1_valid && p->two->sha1_valid)
1612 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1613 else
1614 content_changed = 1;
1616 if (!content_changed) {
1617 /*
1618 * The SHA1 has not changed, so pre-/post-content is
1619 * identical. We can therefore skip looking at the
1620 * file contents altogether.
1621 */
1622 damage = 0;
1623 goto found_damage;
1624 }
1626 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1627 /*
1628 * In --dirstat-by-file mode, we don't really need to
1629 * look at the actual file contents at all.
1630 * The fact that the SHA1 changed is enough for us to
1631 * add this file to the list of results
1632 * (with each file contributing equal damage).
1633 */
1634 damage = 1;
1635 goto found_damage;
1636 }
1638 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1639 diff_populate_filespec(p->one, 0);
1640 diff_populate_filespec(p->two, 0);
1641 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1642 &copied, &added);
1643 diff_free_filespec_data(p->one);
1644 diff_free_filespec_data(p->two);
1645 } else if (DIFF_FILE_VALID(p->one)) {
1646 diff_populate_filespec(p->one, 1);
1647 copied = added = 0;
1648 diff_free_filespec_data(p->one);
1649 } else if (DIFF_FILE_VALID(p->two)) {
1650 diff_populate_filespec(p->two, 1);
1651 copied = 0;
1652 added = p->two->size;
1653 diff_free_filespec_data(p->two);
1654 } else
1655 continue;
1657 /*
1658 * Original minus copied is the removed material,
1659 * added is the new material. They are both damages
1660 * made to the preimage.
1661 * If the resulting damage is zero, we know that
1662 * diffcore_count_changes() considers the two entries to
1663 * be identical, but since content_changed is true, we
1664 * know that there must have been _some_ kind of change,
1665 * so we force all entries to have damage > 0.
1666 */
1667 damage = (p->one->size - copied) + added;
1668 if (!damage)
1669 damage = 1;
1671 found_damage:
1672 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1673 dir.files[dir.nr].name = name;
1674 dir.files[dir.nr].changed = damage;
1675 changed += damage;
1676 dir.nr++;
1677 }
1679 /* This can happen even with many files, if everything was renames */
1680 if (!changed)
1681 return;
1683 /* Show all directories with more than x% of the changes */
1684 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1685 gather_dirstat(options, &dir, changed, "", 0);
1686 }
1688 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1689 {
1690 int i;
1691 unsigned long changed;
1692 struct dirstat_dir dir;
1694 if (data->nr == 0)
1695 return;
1697 dir.files = NULL;
1698 dir.alloc = 0;
1699 dir.nr = 0;
1700 dir.permille = options->dirstat_permille;
1701 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1703 changed = 0;
1704 for (i = 0; i < data->nr; i++) {
1705 struct diffstat_file *file = data->files[i];
1706 unsigned long damage = file->added + file->deleted;
1707 if (file->is_binary)
1708 /*
1709 * binary files counts bytes, not lines. Must find some
1710 * way to normalize binary bytes vs. textual lines.
1711 * The following heuristic assumes that there are 64
1712 * bytes per "line".
1713 * This is stupid and ugly, but very cheap...
1714 */
1715 damage = (damage + 63) / 64;
1716 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1717 dir.files[dir.nr].name = file->name;
1718 dir.files[dir.nr].changed = damage;
1719 changed += damage;
1720 dir.nr++;
1721 }
1723 /* This can happen even with many files, if everything was renames */
1724 if (!changed)
1725 return;
1727 /* Show all directories with more than x% of the changes */
1728 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1729 gather_dirstat(options, &dir, changed, "", 0);
1730 }
1732 static void free_diffstat_info(struct diffstat_t *diffstat)
1733 {
1734 int i;
1735 for (i = 0; i < diffstat->nr; i++) {
1736 struct diffstat_file *f = diffstat->files[i];
1737 if (f->name != f->print_name)
1738 free(f->print_name);
1739 free(f->name);
1740 free(f->from_name);
1741 free(f);
1742 }
1743 free(diffstat->files);
1744 }
1746 struct checkdiff_t {
1747 const char *filename;
1748 int lineno;
1749 int conflict_marker_size;
1750 struct diff_options *o;
1751 unsigned ws_rule;
1752 unsigned status;
1753 };
1755 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1756 {
1757 char firstchar;
1758 int cnt;
1760 if (len < marker_size + 1)
1761 return 0;
1762 firstchar = line[0];
1763 switch (firstchar) {
1764 case '=': case '>': case '<': case '|':
1765 break;
1766 default:
1767 return 0;
1768 }
1769 for (cnt = 1; cnt < marker_size; cnt++)
1770 if (line[cnt] != firstchar)
1771 return 0;
1772 /* line[1] thru line[marker_size-1] are same as firstchar */
1773 if (len < marker_size + 1 || !isspace(line[marker_size]))
1774 return 0;
1775 return 1;
1776 }
1778 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1779 {
1780 struct checkdiff_t *data = priv;
1781 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1782 int marker_size = data->conflict_marker_size;
1783 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1784 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1785 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1786 char *err;
1787 char *line_prefix = "";
1788 struct strbuf *msgbuf;
1790 assert(data->o);
1791 if (data->o->output_prefix) {
1792 msgbuf = data->o->output_prefix(data->o,
1793 data->o->output_prefix_data);
1794 line_prefix = msgbuf->buf;
1795 }
1797 if (line[0] == '+') {
1798 unsigned bad;
1799 data->lineno++;
1800 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1801 data->status |= 1;
1802 fprintf(data->o->file,
1803 "%s%s:%d: leftover conflict marker\n",
1804 line_prefix, data->filename, data->lineno);
1805 }
1806 bad = ws_check(line + 1, len - 1, data->ws_rule);
1807 if (!bad)
1808 return;
1809 data->status |= bad;
1810 err = whitespace_error_string(bad);
1811 fprintf(data->o->file, "%s%s:%d: %s.\n",
1812 line_prefix, data->filename, data->lineno, err);
1813 free(err);
1814 emit_line(data->o, set, reset, line, 1);
1815 ws_check_emit(line + 1, len - 1, data->ws_rule,
1816 data->o->file, set, reset, ws);
1817 } else if (line[0] == ' ') {
1818 data->lineno++;
1819 } else if (line[0] == '@') {
1820 char *plus = strchr(line, '+');
1821 if (plus)
1822 data->lineno = strtol(plus, NULL, 10) - 1;
1823 else
1824 die("invalid diff");
1825 }
1826 }
1828 static unsigned char *deflate_it(char *data,
1829 unsigned long size,
1830 unsigned long *result_size)
1831 {
1832 int bound;
1833 unsigned char *deflated;
1834 z_stream stream;
1836 memset(&stream, 0, sizeof(stream));
1837 deflateInit(&stream, zlib_compression_level);
1838 bound = deflateBound(&stream, size);
1839 deflated = xmalloc(bound);
1840 stream.next_out = deflated;
1841 stream.avail_out = bound;
1843 stream.next_in = (unsigned char *)data;
1844 stream.avail_in = size;
1845 while (deflate(&stream, Z_FINISH) == Z_OK)
1846 ; /* nothing */
1847 deflateEnd(&stream);
1848 *result_size = stream.total_out;
1849 return deflated;
1850 }
1852 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1853 {
1854 void *cp;
1855 void *delta;
1856 void *deflated;
1857 void *data;
1858 unsigned long orig_size;
1859 unsigned long delta_size;
1860 unsigned long deflate_size;
1861 unsigned long data_size;
1863 /* We could do deflated delta, or we could do just deflated two,
1864 * whichever is smaller.
1865 */
1866 delta = NULL;
1867 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1868 if (one->size && two->size) {
1869 delta = diff_delta(one->ptr, one->size,
1870 two->ptr, two->size,
1871 &delta_size, deflate_size);
1872 if (delta) {
1873 void *to_free = delta;
1874 orig_size = delta_size;
1875 delta = deflate_it(delta, delta_size, &delta_size);
1876 free(to_free);
1877 }
1878 }
1880 if (delta && delta_size < deflate_size) {
1881 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1882 free(deflated);
1883 data = delta;
1884 data_size = delta_size;
1885 }
1886 else {
1887 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1888 free(delta);
1889 data = deflated;
1890 data_size = deflate_size;
1891 }
1893 /* emit data encoded in base85 */
1894 cp = data;
1895 while (data_size) {
1896 int bytes = (52 < data_size) ? 52 : data_size;
1897 char line[70];
1898 data_size -= bytes;
1899 if (bytes <= 26)
1900 line[0] = bytes + 'A' - 1;
1901 else
1902 line[0] = bytes - 26 + 'a' - 1;
1903 encode_85(line + 1, cp, bytes);
1904 cp = (char *) cp + bytes;
1905 fprintf(file, "%s", prefix);
1906 fputs(line, file);
1907 fputc('\n', file);
1908 }
1909 fprintf(file, "%s\n", prefix);
1910 free(data);
1911 }
1913 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1914 {
1915 fprintf(file, "%sGIT binary patch\n", prefix);
1916 emit_binary_diff_body(file, one, two, prefix);
1917 emit_binary_diff_body(file, two, one, prefix);
1918 }
1920 static void diff_filespec_load_driver(struct diff_filespec *one)
1921 {
1922 /* Use already-loaded driver */
1923 if (one->driver)
1924 return;
1926 if (S_ISREG(one->mode))
1927 one->driver = userdiff_find_by_path(one->path);
1929 /* Fallback to default settings */
1930 if (!one->driver)
1931 one->driver = userdiff_find_by_name("default");
1932 }
1934 int diff_filespec_is_binary(struct diff_filespec *one)
1935 {
1936 if (one->is_binary == -1) {
1937 diff_filespec_load_driver(one);
1938 if (one->driver->binary != -1)
1939 one->is_binary = one->driver->binary;
1940 else {
1941 if (!one->data && DIFF_FILE_VALID(one))
1942 diff_populate_filespec(one, 0);
1943 if (one->data)
1944 one->is_binary = buffer_is_binary(one->data,
1945 one->size);
1946 if (one->is_binary == -1)
1947 one->is_binary = 0;
1948 }
1949 }
1950 return one->is_binary;
1951 }
1953 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1954 {
1955 diff_filespec_load_driver(one);
1956 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1957 }
1959 static const char *userdiff_word_regex(struct diff_filespec *one)
1960 {
1961 diff_filespec_load_driver(one);
1962 return one->driver->word_regex;
1963 }
1965 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1966 {
1967 if (!options->a_prefix)
1968 options->a_prefix = a;
1969 if (!options->b_prefix)
1970 options->b_prefix = b;
1971 }
1973 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1974 {
1975 if (!DIFF_FILE_VALID(one))
1976 return NULL;
1978 diff_filespec_load_driver(one);
1979 return userdiff_get_textconv(one->driver);
1980 }
1982 static void builtin_diff(const char *name_a,
1983 const char *name_b,
1984 struct diff_filespec *one,
1985 struct diff_filespec *two,
1986 const char *xfrm_msg,
1987 int must_show_header,
1988 struct diff_options *o,
1989 int complete_rewrite)
1990 {
1991 mmfile_t mf1, mf2;
1992 const char *lbl[2];
1993 char *a_one, *b_two;
1994 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1995 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1996 const char *a_prefix, *b_prefix;
1997 struct userdiff_driver *textconv_one = NULL;
1998 struct userdiff_driver *textconv_two = NULL;
1999 struct strbuf header = STRBUF_INIT;
2000 struct strbuf *msgbuf;
2001 char *line_prefix = "";
2003 if (o->output_prefix) {
2004 msgbuf = o->output_prefix(o, o->output_prefix_data);
2005 line_prefix = msgbuf->buf;
2006 }
2008 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2009 (!one->mode || S_ISGITLINK(one->mode)) &&
2010 (!two->mode || S_ISGITLINK(two->mode))) {
2011 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2012 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2013 show_submodule_summary(o->file, one ? one->path : two->path,
2014 one->sha1, two->sha1, two->dirty_submodule,
2015 del, add, reset);
2016 return;
2017 }
2019 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2020 textconv_one = get_textconv(one);
2021 textconv_two = get_textconv(two);
2022 }
2024 diff_set_mnemonic_prefix(o, "a/", "b/");
2025 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2026 a_prefix = o->b_prefix;
2027 b_prefix = o->a_prefix;
2028 } else {
2029 a_prefix = o->a_prefix;
2030 b_prefix = o->b_prefix;
2031 }
2033 /* Never use a non-valid filename anywhere if at all possible */
2034 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2035 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2037 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2038 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2039 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2040 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2041 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2042 if (lbl[0][0] == '/') {
2043 /* /dev/null */
2044 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2045 if (xfrm_msg)
2046 strbuf_addstr(&header, xfrm_msg);
2047 must_show_header = 1;
2048 }
2049 else if (lbl[1][0] == '/') {
2050 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2051 if (xfrm_msg)
2052 strbuf_addstr(&header, xfrm_msg);
2053 must_show_header = 1;
2054 }
2055 else {
2056 if (one->mode != two->mode) {
2057 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2058 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2059 must_show_header = 1;
2060 }
2061 if (xfrm_msg)
2062 strbuf_addstr(&header, xfrm_msg);
2064 /*
2065 * we do not run diff between different kind
2066 * of objects.
2067 */
2068 if ((one->mode ^ two->mode) & S_IFMT)
2069 goto free_ab_and_return;
2070 if (complete_rewrite &&
2071 (textconv_one || !diff_filespec_is_binary(one)) &&
2072 (textconv_two || !diff_filespec_is_binary(two))) {
2073 fprintf(o->file, "%s", header.buf);
2074 strbuf_reset(&header);
2075 emit_rewrite_diff(name_a, name_b, one, two,
2076 textconv_one, textconv_two, o);
2077 o->found_changes = 1;
2078 goto free_ab_and_return;
2079 }
2080 }
2082 if (o->irreversible_delete && lbl[1][0] == '/') {
2083 fprintf(o->file, "%s", header.buf);
2084 strbuf_reset(&header);
2085 goto free_ab_and_return;
2086 } else if (!DIFF_OPT_TST(o, TEXT) &&
2087 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2088 (!textconv_two && diff_filespec_is_binary(two)) )) {
2089 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2090 die("unable to read files to diff");
2091 /* Quite common confusing case */
2092 if (mf1.size == mf2.size &&
2093 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2094 if (must_show_header)
2095 fprintf(o->file, "%s", header.buf);
2096 goto free_ab_and_return;
2097 }
2098 fprintf(o->file, "%s", header.buf);
2099 strbuf_reset(&header);
2100 if (DIFF_OPT_TST(o, BINARY))
2101 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2102 else
2103 fprintf(o->file, "%sBinary files %s and %s differ\n",
2104 line_prefix, lbl[0], lbl[1]);
2105 o->found_changes = 1;
2106 } else {
2107 /* Crazy xdl interfaces.. */
2108 const char *diffopts = getenv("GIT_DIFF_OPTS");
2109 xpparam_t xpp;
2110 xdemitconf_t xecfg;
2111 struct emit_callback ecbdata;
2112 const struct userdiff_funcname *pe;
2114 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2115 fprintf(o->file, "%s", header.buf);
2116 strbuf_reset(&header);
2117 }
2119 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2120 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2122 pe = diff_funcname_pattern(one);
2123 if (!pe)
2124 pe = diff_funcname_pattern(two);
2126 memset(&xpp, 0, sizeof(xpp));
2127 memset(&xecfg, 0, sizeof(xecfg));
2128 memset(&ecbdata, 0, sizeof(ecbdata));
2129 ecbdata.label_path = lbl;
2130 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2131 ecbdata.found_changesp = &o->found_changes;
2132 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2133 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2134 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2135 ecbdata.opt = o;
2136 ecbdata.header = header.len ? &header : NULL;
2137 xpp.flags = o->xdl_opts;
2138 xecfg.ctxlen = o->context;
2139 xecfg.interhunkctxlen = o->interhunkcontext;
2140 xecfg.flags = XDL_EMIT_FUNCNAMES;
2141 if (pe)
2142 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2143 if (!diffopts)
2144 ;
2145 else if (!prefixcmp(diffopts, "--unified="))
2146 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2147 else if (!prefixcmp(diffopts, "-u"))
2148 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2149 if (o->word_diff) {
2150 int i;
2152 ecbdata.diff_words =
2153 xcalloc(1, sizeof(struct diff_words_data));
2154 ecbdata.diff_words->type = o->word_diff;
2155 ecbdata.diff_words->opt = o;
2156 if (!o->word_regex)
2157 o->word_regex = userdiff_word_regex(one);
2158 if (!o->word_regex)
2159 o->word_regex = userdiff_word_regex(two);
2160 if (!o->word_regex)
2161 o->word_regex = diff_word_regex_cfg;
2162 if (o->word_regex) {
2163 ecbdata.diff_words->word_regex = (regex_t *)
2164 xmalloc(sizeof(regex_t));
2165 if (regcomp(ecbdata.diff_words->word_regex,
2166 o->word_regex,
2167 REG_EXTENDED | REG_NEWLINE))
2168 die ("Invalid regular expression: %s",
2169 o->word_regex);
2170 }
2171 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2172 if (o->word_diff == diff_words_styles[i].type) {
2173 ecbdata.diff_words->style =
2174 &diff_words_styles[i];
2175 break;
2176 }
2177 }
2178 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2179 struct diff_words_style *st = ecbdata.diff_words->style;
2180 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2181 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2182 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2183 }
2184 }
2185 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2186 &xpp, &xecfg);
2187 if (o->word_diff)
2188 free_diff_words_data(&ecbdata);
2189 if (textconv_one)
2190 free(mf1.ptr);
2191 if (textconv_two)
2192 free(mf2.ptr);
2193 xdiff_clear_find_func(&xecfg);
2194 }
2196 free_ab_and_return:
2197 strbuf_release(&header);
2198 diff_free_filespec_data(one);
2199 diff_free_filespec_data(two);
2200 free(a_one);
2201 free(b_two);
2202 return;
2203 }
2205 static void builtin_diffstat(const char *name_a, const char *name_b,
2206 struct diff_filespec *one,
2207 struct diff_filespec *two,
2208 struct diffstat_t *diffstat,
2209 struct diff_options *o,
2210 int complete_rewrite)
2211 {
2212 mmfile_t mf1, mf2;
2213 struct diffstat_file *data;
2215 data = diffstat_add(diffstat, name_a, name_b);
2217 if (!one || !two) {
2218 data->is_unmerged = 1;
2219 return;
2220 }
2222 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2223 data->is_binary = 1;
2224 data->added = diff_filespec_size(two);
2225 data->deleted = diff_filespec_size(one);
2226 }
2228 else if (complete_rewrite) {
2229 diff_populate_filespec(one, 0);
2230 diff_populate_filespec(two, 0);
2231 data->deleted = count_lines(one->data, one->size);
2232 data->added = count_lines(two->data, two->size);
2233 }
2235 else {
2236 /* Crazy xdl interfaces.. */
2237 xpparam_t xpp;
2238 xdemitconf_t xecfg;
2240 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2241 die("unable to read files to diff");
2243 memset(&xpp, 0, sizeof(xpp));
2244 memset(&xecfg, 0, sizeof(xecfg));
2245 xpp.flags = o->xdl_opts;
2246 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2247 &xpp, &xecfg);
2248 }
2250 diff_free_filespec_data(one);
2251 diff_free_filespec_data(two);
2252 }
2254 static void builtin_checkdiff(const char *name_a, const char *name_b,
2255 const char *attr_path,
2256 struct diff_filespec *one,
2257 struct diff_filespec *two,
2258 struct diff_options *o)
2259 {
2260 mmfile_t mf1, mf2;
2261 struct checkdiff_t data;
2263 if (!two)
2264 return;
2266 memset(&data, 0, sizeof(data));
2267 data.filename = name_b ? name_b : name_a;
2268 data.lineno = 0;
2269 data.o = o;
2270 data.ws_rule = whitespace_rule(attr_path);
2271 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2273 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2274 die("unable to read files to diff");
2276 /*
2277 * All the other codepaths check both sides, but not checking
2278 * the "old" side here is deliberate. We are checking the newly
2279 * introduced changes, and as long as the "new" side is text, we
2280 * can and should check what it introduces.
2281 */
2282 if (diff_filespec_is_binary(two))
2283 goto free_and_return;
2284 else {
2285 /* Crazy xdl interfaces.. */
2286 xpparam_t xpp;
2287 xdemitconf_t xecfg;
2289 memset(&xpp, 0, sizeof(xpp));
2290 memset(&xecfg, 0, sizeof(xecfg));
2291 xecfg.ctxlen = 1; /* at least one context line */
2292 xpp.flags = 0;
2293 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2294 &xpp, &xecfg);
2296 if (data.ws_rule & WS_BLANK_AT_EOF) {
2297 struct emit_callback ecbdata;
2298 int blank_at_eof;
2300 ecbdata.ws_rule = data.ws_rule;
2301 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2302 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2304 if (blank_at_eof) {
2305 static char *err;
2306 if (!err)
2307 err = whitespace_error_string(WS_BLANK_AT_EOF);
2308 fprintf(o->file, "%s:%d: %s.\n",
2309 data.filename, blank_at_eof, err);
2310 data.status = 1; /* report errors */
2311 }
2312 }
2313 }
2314 free_and_return:
2315 diff_free_filespec_data(one);
2316 diff_free_filespec_data(two);
2317 if (data.status)
2318 DIFF_OPT_SET(o, CHECK_FAILED);
2319 }
2321 struct diff_filespec *alloc_filespec(const char *path)
2322 {
2323 int namelen = strlen(path);
2324 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2326 memset(spec, 0, sizeof(*spec));
2327 spec->path = (char *)(spec + 1);
2328 memcpy(spec->path, path, namelen+1);
2329 spec->count = 1;
2330 spec->is_binary = -1;
2331 return spec;
2332 }
2334 void free_filespec(struct diff_filespec *spec)
2335 {
2336 if (!--spec->count) {
2337 diff_free_filespec_data(spec);
2338 free(spec);
2339 }
2340 }
2342 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2343 unsigned short mode)
2344 {
2345 if (mode) {
2346 spec->mode = canon_mode(mode);
2347 hashcpy(spec->sha1, sha1);
2348 spec->sha1_valid = !is_null_sha1(sha1);
2349 }
2350 }
2352 /*
2353 * Given a name and sha1 pair, if the index tells us the file in
2354 * the work tree has that object contents, return true, so that
2355 * prepare_temp_file() does not have to inflate and extract.
2356 */
2357 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2358 {
2359 struct cache_entry *ce;
2360 struct stat st;
2361 int pos, len;
2363 /*
2364 * We do not read the cache ourselves here, because the
2365 * benchmark with my previous version that always reads cache
2366 * shows that it makes things worse for diff-tree comparing
2367 * two linux-2.6 kernel trees in an already checked out work
2368 * tree. This is because most diff-tree comparisons deal with
2369 * only a small number of files, while reading the cache is
2370 * expensive for a large project, and its cost outweighs the
2371 * savings we get by not inflating the object to a temporary
2372 * file. Practically, this code only helps when we are used
2373 * by diff-cache --cached, which does read the cache before
2374 * calling us.
2375 */
2376 if (!active_cache)
2377 return 0;
2379 /* We want to avoid the working directory if our caller
2380 * doesn't need the data in a normal file, this system
2381 * is rather slow with its stat/open/mmap/close syscalls,
2382 * and the object is contained in a pack file. The pack
2383 * is probably already open and will be faster to obtain
2384 * the data through than the working directory. Loose
2385 * objects however would tend to be slower as they need
2386 * to be individually opened and inflated.
2387 */
2388 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2389 return 0;
2391 len = strlen(name);
2392 pos = cache_name_pos(name, len);
2393 if (pos < 0)
2394 return 0;
2395 ce = active_cache[pos];
2397 /*
2398 * This is not the sha1 we are looking for, or
2399 * unreusable because it is not a regular file.
2400 */
2401 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2402 return 0;
2404 /*
2405 * If ce is marked as "assume unchanged", there is no
2406 * guarantee that work tree matches what we are looking for.
2407 */
2408 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2409 return 0;
2411 /*
2412 * If ce matches the file in the work tree, we can reuse it.
2413 */
2414 if (ce_uptodate(ce) ||
2415 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2416 return 1;
2418 return 0;
2419 }
2421 static int populate_from_stdin(struct diff_filespec *s)
2422 {
2423 struct strbuf buf = STRBUF_INIT;
2424 size_t size = 0;
2426 if (strbuf_read(&buf, 0, 0) < 0)
2427 return error("error while reading from stdin %s",
2428 strerror(errno));
2430 s->should_munmap = 0;
2431 s->data = strbuf_detach(&buf, &size);
2432 s->size = size;
2433 s->should_free = 1;
2434 return 0;
2435 }
2437 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2438 {
2439 int len;
2440 char *data = xmalloc(100), *dirty = "";
2442 /* Are we looking at the work tree? */
2443 if (s->dirty_submodule)
2444 dirty = "-dirty";
2446 len = snprintf(data, 100,
2447 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2448 s->data = data;
2449 s->size = len;
2450 s->should_free = 1;
2451 if (size_only) {
2452 s->data = NULL;
2453 free(data);
2454 }
2455 return 0;
2456 }
2458 /*
2459 * While doing rename detection and pickaxe operation, we may need to
2460 * grab the data for the blob (or file) for our own in-core comparison.
2461 * diff_filespec has data and size fields for this purpose.
2462 */
2463 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2464 {
2465 int err = 0;
2466 if (!DIFF_FILE_VALID(s))
2467 die("internal error: asking to populate invalid file.");
2468 if (S_ISDIR(s->mode))
2469 return -1;
2471 if (s->data)
2472 return 0;
2474 if (size_only && 0 < s->size)
2475 return 0;
2477 if (S_ISGITLINK(s->mode))
2478 return diff_populate_gitlink(s, size_only);
2480 if (!s->sha1_valid ||
2481 reuse_worktree_file(s->path, s->sha1, 0)) {
2482 struct strbuf buf = STRBUF_INIT;
2483 struct stat st;
2484 int fd;
2486 if (!strcmp(s->path, "-"))
2487 return populate_from_stdin(s);
2489 if (lstat(s->path, &st) < 0) {
2490 if (errno == ENOENT) {
2491 err_empty:
2492 err = -1;
2493 empty:
2494 s->data = (char *)"";
2495 s->size = 0;
2496 return err;
2497 }
2498 }
2499 s->size = xsize_t(st.st_size);
2500 if (!s->size)
2501 goto empty;
2502 if (S_ISLNK(st.st_mode)) {
2503 struct strbuf sb = STRBUF_INIT;
2505 if (strbuf_readlink(&sb, s->path, s->size))
2506 goto err_empty;
2507 s->size = sb.len;
2508 s->data = strbuf_detach(&sb, NULL);
2509 s->should_free = 1;
2510 return 0;
2511 }
2512 if (size_only)
2513 return 0;
2514 fd = open(s->path, O_RDONLY);
2515 if (fd < 0)
2516 goto err_empty;
2517 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2518 close(fd);
2519 s->should_munmap = 1;
2521 /*
2522 * Convert from working tree format to canonical git format
2523 */
2524 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2525 size_t size = 0;
2526 munmap(s->data, s->size);
2527 s->should_munmap = 0;
2528 s->data = strbuf_detach(&buf, &size);
2529 s->size = size;
2530 s->should_free = 1;
2531 }
2532 }
2533 else {
2534 enum object_type type;
2535 if (size_only) {
2536 type = sha1_object_info(s->sha1, &s->size);
2537 if (type < 0)
2538 die("unable to read %s", sha1_to_hex(s->sha1));
2539 } else {
2540 s->data = read_sha1_file(s->sha1, &type, &s->size);
2541 if (!s->data)
2542 die("unable to read %s", sha1_to_hex(s->sha1));
2543 s->should_free = 1;
2544 }
2545 }
2546 return 0;
2547 }
2549 void diff_free_filespec_blob(struct diff_filespec *s)
2550 {
2551 if (s->should_free)
2552 free(s->data);
2553 else if (s->should_munmap)
2554 munmap(s->data, s->size);
2556 if (s->should_free || s->should_munmap) {
2557 s->should_free = s->should_munmap = 0;
2558 s->data = NULL;
2559 }
2560 }
2562 void diff_free_filespec_data(struct diff_filespec *s)
2563 {
2564 diff_free_filespec_blob(s);
2565 free(s->cnt_data);
2566 s->cnt_data = NULL;
2567 }
2569 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2570 void *blob,
2571 unsigned long size,
2572 const unsigned char *sha1,
2573 int mode)
2574 {
2575 int fd;
2576 struct strbuf buf = STRBUF_INIT;
2577 struct strbuf template = STRBUF_INIT;
2578 char *path_dup = xstrdup(path);
2579 const char *base = basename(path_dup);
2581 /* Generate "XXXXXX_basename.ext" */
2582 strbuf_addstr(&template, "XXXXXX_");
2583 strbuf_addstr(&template, base);
2585 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2586 strlen(base) + 1);
2587 if (fd < 0)
2588 die_errno("unable to create temp-file");
2589 if (convert_to_working_tree(path,
2590 (const char *)blob, (size_t)size, &buf)) {
2591 blob = buf.buf;
2592 size = buf.len;
2593 }
2594 if (write_in_full(fd, blob, size) != size)
2595 die_errno("unable to write temp-file");
2596 close(fd);
2597 temp->name = temp->tmp_path;
2598 strcpy(temp->hex, sha1_to_hex(sha1));
2599 temp->hex[40] = 0;
2600 sprintf(temp->mode, "%06o", mode);
2601 strbuf_release(&buf);
2602 strbuf_release(&template);
2603 free(path_dup);
2604 }
2606 static struct diff_tempfile *prepare_temp_file(const char *name,
2607 struct diff_filespec *one)
2608 {
2609 struct diff_tempfile *temp = claim_diff_tempfile();
2611 if (!DIFF_FILE_VALID(one)) {
2612 not_a_valid_file:
2613 /* A '-' entry produces this for file-2, and
2614 * a '+' entry produces this for file-1.
2615 */
2616 temp->name = "/dev/null";
2617 strcpy(temp->hex, ".");
2618 strcpy(temp->mode, ".");
2619 return temp;
2620 }
2622 if (!remove_tempfile_installed) {
2623 atexit(remove_tempfile);
2624 sigchain_push_common(remove_tempfile_on_signal);
2625 remove_tempfile_installed = 1;
2626 }
2628 if (!one->sha1_valid ||
2629 reuse_worktree_file(name, one->sha1, 1)) {
2630 struct stat st;
2631 if (lstat(name, &st) < 0) {
2632 if (errno == ENOENT)
2633 goto not_a_valid_file;
2634 die_errno("stat(%s)", name);
2635 }
2636 if (S_ISLNK(st.st_mode)) {
2637 struct strbuf sb = STRBUF_INIT;
2638 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2639 die_errno("readlink(%s)", name);
2640 prep_temp_blob(name, temp, sb.buf, sb.len,
2641 (one->sha1_valid ?
2642 one->sha1 : null_sha1),
2643 (one->sha1_valid ?
2644 one->mode : S_IFLNK));
2645 strbuf_release(&sb);
2646 }
2647 else {
2648 /* we can borrow from the file in the work tree */
2649 temp->name = name;
2650 if (!one->sha1_valid)
2651 strcpy(temp->hex, sha1_to_hex(null_sha1));
2652 else
2653 strcpy(temp->hex, sha1_to_hex(one->sha1));
2654 /* Even though we may sometimes borrow the
2655 * contents from the work tree, we always want
2656 * one->mode. mode is trustworthy even when
2657 * !(one->sha1_valid), as long as
2658 * DIFF_FILE_VALID(one).
2659 */
2660 sprintf(temp->mode, "%06o", one->mode);
2661 }
2662 return temp;
2663 }
2664 else {
2665 if (diff_populate_filespec(one, 0))
2666 die("cannot read data blob for %s", one->path);
2667 prep_temp_blob(name, temp, one->data, one->size,
2668 one->sha1, one->mode);
2669 }
2670 return temp;
2671 }
2673 /* An external diff command takes:
2674 *
2675 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2676 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2677 *
2678 */
2679 static void run_external_diff(const char *pgm,
2680 const char *name,
2681 const char *other,
2682 struct diff_filespec *one,
2683 struct diff_filespec *two,
2684 const char *xfrm_msg,
2685 int complete_rewrite)
2686 {
2687 const char *spawn_arg[10];
2688 int retval;
2689 const char **arg = &spawn_arg[0];
2691 if (one && two) {
2692 struct diff_tempfile *temp_one, *temp_two;
2693 const char *othername = (other ? other : name);
2694 temp_one = prepare_temp_file(name, one);
2695 temp_two = prepare_temp_file(othername, two);
2696 *arg++ = pgm;
2697 *arg++ = name;
2698 *arg++ = temp_one->name;
2699 *arg++ = temp_one->hex;
2700 *arg++ = temp_one->mode;
2701 *arg++ = temp_two->name;
2702 *arg++ = temp_two->hex;
2703 *arg++ = temp_two->mode;
2704 if (other) {
2705 *arg++ = other;
2706 *arg++ = xfrm_msg;
2707 }
2708 } else {
2709 *arg++ = pgm;
2710 *arg++ = name;
2711 }
2712 *arg = NULL;
2713 fflush(NULL);
2714 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2715 remove_tempfile();
2716 if (retval) {
2717 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2718 exit(1);
2719 }
2720 }
2722 static int similarity_index(struct diff_filepair *p)
2723 {
2724 return p->score * 100 / MAX_SCORE;
2725 }
2727 static void fill_metainfo(struct strbuf *msg,
2728 const char *name,
2729 const char *other,
2730 struct diff_filespec *one,
2731 struct diff_filespec *two,
2732 struct diff_options *o,
2733 struct diff_filepair *p,
2734 int *must_show_header,
2735 int use_color)
2736 {
2737 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2738 const char *reset = diff_get_color(use_color, DIFF_RESET);
2739 struct strbuf *msgbuf;
2740 char *line_prefix = "";
2742 *must_show_header = 1;
2743 if (o->output_prefix) {
2744 msgbuf = o->output_prefix(o, o->output_prefix_data);
2745 line_prefix = msgbuf->buf;
2746 }
2747 strbuf_init(msg, PATH_MAX * 2 + 300);
2748 switch (p->status) {
2749 case DIFF_STATUS_COPIED:
2750 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2751 line_prefix, set, similarity_index(p));
2752 strbuf_addf(msg, "%s\n%s%scopy from ",
2753 reset, line_prefix, set);
2754 quote_c_style(name, msg, NULL, 0);
2755 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2756 quote_c_style(other, msg, NULL, 0);
2757 strbuf_addf(msg, "%s\n", reset);
2758 break;
2759 case DIFF_STATUS_RENAMED:
2760 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2761 line_prefix, set, similarity_index(p));
2762 strbuf_addf(msg, "%s\n%s%srename from ",
2763 reset, line_prefix, set);
2764 quote_c_style(name, msg, NULL, 0);
2765 strbuf_addf(msg, "%s\n%s%srename to ",
2766 reset, line_prefix, set);
2767 quote_c_style(other, msg, NULL, 0);
2768 strbuf_addf(msg, "%s\n", reset);
2769 break;
2770 case DIFF_STATUS_MODIFIED:
2771 if (p->score) {
2772 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2773 line_prefix,
2774 set, similarity_index(p), reset);
2775 break;
2776 }
2777 /* fallthru */
2778 default:
2779 *must_show_header = 0;
2780 }
2781 if (one && two && hashcmp(one->sha1, two->sha1)) {
2782 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2784 if (DIFF_OPT_TST(o, BINARY)) {
2785 mmfile_t mf;
2786 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2787 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2788 abbrev = 40;
2789 }
2790 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2791 find_unique_abbrev(one->sha1, abbrev));
2792 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2793 if (one->mode == two->mode)
2794 strbuf_addf(msg, " %06o", one->mode);
2795 strbuf_addf(msg, "%s\n", reset);
2796 }
2797 }
2799 static void run_diff_cmd(const char *pgm,
2800 const char *name,
2801 const char *other,
2802 const char *attr_path,
2803 struct diff_filespec *one,
2804 struct diff_filespec *two,
2805 struct strbuf *msg,
2806 struct diff_options *o,
2807 struct diff_filepair *p)
2808 {
2809 const char *xfrm_msg = NULL;
2810 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2811 int must_show_header = 0;
2813 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2814 pgm = NULL;
2815 else {
2816 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2817 if (drv && drv->external)
2818 pgm = drv->external;
2819 }
2821 if (msg) {
2822 /*
2823 * don't use colors when the header is intended for an
2824 * external diff driver
2825 */
2826 fill_metainfo(msg, name, other, one, two, o, p,
2827 &must_show_header,
2828 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2829 xfrm_msg = msg->len ? msg->buf : NULL;
2830 }
2832 if (pgm) {
2833 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2834 complete_rewrite);
2835 return;
2836 }
2837 if (one && two)
2838 builtin_diff(name, other ? other : name,
2839 one, two, xfrm_msg, must_show_header,
2840 o, complete_rewrite);
2841 else
2842 fprintf(o->file, "* Unmerged path %s\n", name);
2843 }
2845 static void diff_fill_sha1_info(struct diff_filespec *one)
2846 {
2847 if (DIFF_FILE_VALID(one)) {
2848 if (!one->sha1_valid) {
2849 struct stat st;
2850 if (!strcmp(one->path, "-")) {
2851 hashcpy(one->sha1, null_sha1);
2852 return;
2853 }
2854 if (lstat(one->path, &st) < 0)
2855 die_errno("stat '%s'", one->path);
2856 if (index_path(one->sha1, one->path, &st, 0))
2857 die("cannot hash %s", one->path);
2858 }
2859 }
2860 else
2861 hashclr(one->sha1);
2862 }
2864 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2865 {
2866 /* Strip the prefix but do not molest /dev/null and absolute paths */
2867 if (*namep && **namep != '/') {
2868 *namep += prefix_length;
2869 if (**namep == '/')
2870 ++*namep;
2871 }
2872 if (*otherp && **otherp != '/') {
2873 *otherp += prefix_length;
2874 if (**otherp == '/')
2875 ++*otherp;
2876 }
2877 }
2879 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2880 {
2881 const char *pgm = external_diff();
2882 struct strbuf msg;
2883 struct diff_filespec *one = p->one;
2884 struct diff_filespec *two = p->two;
2885 const char *name;
2886 const char *other;
2887 const char *attr_path;
2889 name = p->one->path;
2890 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2891 attr_path = name;
2892 if (o->prefix_length)
2893 strip_prefix(o->prefix_length, &name, &other);
2895 if (DIFF_PAIR_UNMERGED(p)) {
2896 run_diff_cmd(pgm, name, NULL, attr_path,
2897 NULL, NULL, NULL, o, p);
2898 return;
2899 }
2901 diff_fill_sha1_info(one);
2902 diff_fill_sha1_info(two);
2904 if (!pgm &&
2905 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2906 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2907 /*
2908 * a filepair that changes between file and symlink
2909 * needs to be split into deletion and creation.
2910 */
2911 struct diff_filespec *null = alloc_filespec(two->path);
2912 run_diff_cmd(NULL, name, other, attr_path,
2913 one, null, &msg, o, p);
2914 free(null);
2915 strbuf_release(&msg);
2917 null = alloc_filespec(one->path);
2918 run_diff_cmd(NULL, name, other, attr_path,
2919 null, two, &msg, o, p);
2920 free(null);
2921 }
2922 else
2923 run_diff_cmd(pgm, name, other, attr_path,
2924 one, two, &msg, o, p);
2926 strbuf_release(&msg);
2927 }
2929 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2930 struct diffstat_t *diffstat)
2931 {
2932 const char *name;
2933 const char *other;
2934 int complete_rewrite = 0;
2936 if (DIFF_PAIR_UNMERGED(p)) {
2937 /* unmerged */
2938 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2939 return;
2940 }
2942 name = p->one->path;
2943 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2945 if (o->prefix_length)
2946 strip_prefix(o->prefix_length, &name, &other);
2948 diff_fill_sha1_info(p->one);
2949 diff_fill_sha1_info(p->two);
2951 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2952 complete_rewrite = 1;
2953 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2954 }
2956 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2957 {
2958 const char *name;
2959 const char *other;
2960 const char *attr_path;
2962 if (DIFF_PAIR_UNMERGED(p)) {
2963 /* unmerged */
2964 return;
2965 }
2967 name = p->one->path;
2968 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2969 attr_path = other ? other : name;
2971 if (o->prefix_length)
2972 strip_prefix(o->prefix_length, &name, &other);
2974 diff_fill_sha1_info(p->one);
2975 diff_fill_sha1_info(p->two);
2977 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2978 }
2980 void diff_setup(struct diff_options *options)
2981 {
2982 memcpy(options, &default_diff_options, sizeof(*options));
2984 options->file = stdout;
2986 options->line_termination = '\n';
2987 options->break_opt = -1;
2988 options->rename_limit = -1;
2989 options->dirstat_permille = diff_dirstat_permille_default;
2990 options->context = 3;
2992 options->change = diff_change;
2993 options->add_remove = diff_addremove;
2994 if (diff_use_color_default > 0)
2995 DIFF_OPT_SET(options, COLOR_DIFF);
2996 options->detect_rename = diff_detect_rename_default;
2998 if (diff_no_prefix) {
2999 options->a_prefix = options->b_prefix = "";
3000 } else if (!diff_mnemonic_prefix) {
3001 options->a_prefix = "a/";
3002 options->b_prefix = "b/";
3003 }
3004 }
3006 int diff_setup_done(struct diff_options *options)
3007 {
3008 int count = 0;
3010 if (options->output_format & DIFF_FORMAT_NAME)
3011 count++;
3012 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3013 count++;
3014 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3015 count++;
3016 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3017 count++;
3018 if (count > 1)
3019 die("--name-only, --name-status, --check and -s are mutually exclusive");
3021 /*
3022 * Most of the time we can say "there are changes"
3023 * only by checking if there are changed paths, but
3024 * --ignore-whitespace* options force us to look
3025 * inside contents.
3026 */
3028 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3029 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3030 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3031 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3032 else
3033 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3035 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3036 options->detect_rename = DIFF_DETECT_COPY;
3038 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3039 options->prefix = NULL;
3040 if (options->prefix)
3041 options->prefix_length = strlen(options->prefix);
3042 else
3043 options->prefix_length = 0;
3045 if (options->output_format & (DIFF_FORMAT_NAME |
3046 DIFF_FORMAT_NAME_STATUS |
3047 DIFF_FORMAT_CHECKDIFF |
3048 DIFF_FORMAT_NO_OUTPUT))
3049 options->output_format &= ~(DIFF_FORMAT_RAW |
3050 DIFF_FORMAT_NUMSTAT |
3051 DIFF_FORMAT_DIFFSTAT |
3052 DIFF_FORMAT_SHORTSTAT |
3053 DIFF_FORMAT_DIRSTAT |
3054 DIFF_FORMAT_SUMMARY |
3055 DIFF_FORMAT_PATCH);
3057 /*
3058 * These cases always need recursive; we do not drop caller-supplied
3059 * recursive bits for other formats here.
3060 */
3061 if (options->output_format & (DIFF_FORMAT_PATCH |
3062 DIFF_FORMAT_NUMSTAT |
3063 DIFF_FORMAT_DIFFSTAT |
3064 DIFF_FORMAT_SHORTSTAT |
3065 DIFF_FORMAT_DIRSTAT |
3066 DIFF_FORMAT_SUMMARY |
3067 DIFF_FORMAT_CHECKDIFF))
3068 DIFF_OPT_SET(options, RECURSIVE);
3069 /*
3070 * Also pickaxe would not work very well if you do not say recursive
3071 */
3072 if (options->pickaxe)
3073 DIFF_OPT_SET(options, RECURSIVE);
3074 /*
3075 * When patches are generated, submodules diffed against the work tree
3076 * must be checked for dirtiness too so it can be shown in the output
3077 */
3078 if (options->output_format & DIFF_FORMAT_PATCH)
3079 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3081 if (options->detect_rename && options->rename_limit < 0)
3082 options->rename_limit = diff_rename_limit_default;
3083 if (options->setup & DIFF_SETUP_USE_CACHE) {
3084 if (!active_cache)
3085 /* read-cache does not die even when it fails
3086 * so it is safe for us to do this here. Also
3087 * it does not smudge active_cache or active_nr
3088 * when it fails, so we do not have to worry about
3089 * cleaning it up ourselves either.
3090 */
3091 read_cache();
3092 }
3093 if (options->abbrev <= 0 || 40 < options->abbrev)
3094 options->abbrev = 40; /* full */
3096 /*
3097 * It does not make sense to show the first hit we happened
3098 * to have found. It does not make sense not to return with
3099 * exit code in such a case either.
3100 */
3101 if (DIFF_OPT_TST(options, QUICK)) {
3102 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3103 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3104 }
3106 return 0;
3107 }
3109 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3110 {
3111 char c, *eq;
3112 int len;
3114 if (*arg != '-')
3115 return 0;
3116 c = *++arg;
3117 if (!c)
3118 return 0;
3119 if (c == arg_short) {
3120 c = *++arg;
3121 if (!c)
3122 return 1;
3123 if (val && isdigit(c)) {
3124 char *end;
3125 int n = strtoul(arg, &end, 10);
3126 if (*end)
3127 return 0;
3128 *val = n;
3129 return 1;
3130 }
3131 return 0;
3132 }
3133 if (c != '-')
3134 return 0;
3135 arg++;
3136 eq = strchr(arg, '=');
3137 if (eq)
3138 len = eq - arg;
3139 else
3140 len = strlen(arg);
3141 if (!len || strncmp(arg, arg_long, len))
3142 return 0;
3143 if (eq) {
3144 int n;
3145 char *end;
3146 if (!isdigit(*++eq))
3147 return 0;
3148 n = strtoul(eq, &end, 10);
3149 if (*end)
3150 return 0;
3151 *val = n;
3152 }
3153 return 1;
3154 }
3156 static int diff_scoreopt_parse(const char *opt);
3158 static inline int short_opt(char opt, const char **argv,
3159 const char **optarg)
3160 {
3161 const char *arg = argv[0];
3162 if (arg[0] != '-' || arg[1] != opt)
3163 return 0;
3164 if (arg[2] != '\0') {
3165 *optarg = arg + 2;
3166 return 1;
3167 }
3168 if (!argv[1])
3169 die("Option '%c' requires a value", opt);
3170 *optarg = argv[1];
3171 return 2;
3172 }
3174 int parse_long_opt(const char *opt, const char **argv,
3175 const char **optarg)
3176 {
3177 const char *arg = argv[0];
3178 if (arg[0] != '-' || arg[1] != '-')
3179 return 0;
3180 arg += strlen("--");
3181 if (prefixcmp(arg, opt))
3182 return 0;
3183 arg += strlen(opt);
3184 if (*arg == '=') { /* sticked form: --option=value */
3185 *optarg = arg + 1;
3186 return 1;
3187 }
3188 if (*arg != '\0')
3189 return 0;
3190 /* separate form: --option value */
3191 if (!argv[1])
3192 die("Option '--%s' requires a value", opt);
3193 *optarg = argv[1];
3194 return 2;
3195 }
3197 static int stat_opt(struct diff_options *options, const char **av)
3198 {
3199 const char *arg = av[0];
3200 char *end;
3201 int width = options->stat_width;
3202 int name_width = options->stat_name_width;
3203 int argcount = 1;
3205 arg += strlen("--stat");
3206 end = (char *)arg;
3208 switch (*arg) {
3209 case '-':
3210 if (!prefixcmp(arg, "-width")) {
3211 arg += strlen("-width");
3212 if (*arg == '=')
3213 width = strtoul(arg + 1, &end, 10);
3214 else if (!*arg && !av[1])
3215 die("Option '--stat-width' requires a value");
3216 else if (!*arg) {
3217 width = strtoul(av[1], &end, 10);
3218 argcount = 2;
3219 }
3220 } else if (!prefixcmp(arg, "-name-width")) {
3221 arg += strlen("-name-width");
3222 if (*arg == '=')
3223 name_width = strtoul(arg + 1, &end, 10);
3224 else if (!*arg && !av[1])
3225 die("Option '--stat-name-width' requires a value");
3226 else if (!*arg) {
3227 name_width = strtoul(av[1], &end, 10);
3228 argcount = 2;
3229 }
3230 }
3231 break;
3232 case '=':
3233 width = strtoul(arg+1, &end, 10);
3234 if (*end == ',')
3235 name_width = strtoul(end+1, &end, 10);
3236 }
3238 /* Important! This checks all the error cases! */
3239 if (*end)
3240 return 0;
3241 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3242 options->stat_name_width = name_width;
3243 options->stat_width = width;
3244 return argcount;
3245 }
3247 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3248 {
3249 struct strbuf errmsg = STRBUF_INIT;
3250 if (parse_dirstat_params(options, params, &errmsg))
3251 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3252 errmsg.buf);
3253 strbuf_release(&errmsg);
3254 /*
3255 * The caller knows a dirstat-related option is given from the command
3256 * line; allow it to say "return this_function();"
3257 */
3258 options->output_format |= DIFF_FORMAT_DIRSTAT;
3259 return 1;
3260 }
3262 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3263 {
3264 const char *arg = av[0];
3265 const char *optarg;
3266 int argcount;
3268 /* Output format options */
3269 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3270 options->output_format |= DIFF_FORMAT_PATCH;
3271 else if (opt_arg(arg, 'U', "unified", &options->context))
3272 options->output_format |= DIFF_FORMAT_PATCH;
3273 else if (!strcmp(arg, "--raw"))
3274 options->output_format |= DIFF_FORMAT_RAW;
3275 else if (!strcmp(arg, "--patch-with-raw"))
3276 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3277 else if (!strcmp(arg, "--numstat"))
3278 options->output_format |= DIFF_FORMAT_NUMSTAT;
3279 else if (!strcmp(arg, "--shortstat"))
3280 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3281 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3282 return parse_dirstat_opt(options, "");
3283 else if (!prefixcmp(arg, "-X"))
3284 return parse_dirstat_opt(options, arg + 2);
3285 else if (!prefixcmp(arg, "--dirstat="))
3286 return parse_dirstat_opt(options, arg + 10);
3287 else if (!strcmp(arg, "--cumulative"))
3288 return parse_dirstat_opt(options, "cumulative");
3289 else if (!strcmp(arg, "--dirstat-by-file"))
3290 return parse_dirstat_opt(options, "files");
3291 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3292 parse_dirstat_opt(options, "files");
3293 return parse_dirstat_opt(options, arg + 18);
3294 }
3295 else if (!strcmp(arg, "--check"))
3296 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3297 else if (!strcmp(arg, "--summary"))
3298 options->output_format |= DIFF_FORMAT_SUMMARY;
3299 else if (!strcmp(arg, "--patch-with-stat"))
3300 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3301 else if (!strcmp(arg, "--name-only"))
3302 options->output_format |= DIFF_FORMAT_NAME;
3303 else if (!strcmp(arg, "--name-status"))
3304 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3305 else if (!strcmp(arg, "-s"))
3306 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3307 else if (!prefixcmp(arg, "--stat"))
3308 /* --stat, --stat-width, or --stat-name-width */
3309 return stat_opt(options, av);
3311 /* renames options */
3312 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3313 !strcmp(arg, "--break-rewrites")) {
3314 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3315 return error("invalid argument to -B: %s", arg+2);
3316 }
3317 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3318 !strcmp(arg, "--find-renames")) {
3319 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3320 return error("invalid argument to -M: %s", arg+2);
3321 options->detect_rename = DIFF_DETECT_RENAME;
3322 }
3323 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3324 options->irreversible_delete = 1;
3325 }
3326 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3327 !strcmp(arg, "--find-copies")) {
3328 if (options->detect_rename == DIFF_DETECT_COPY)
3329 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3330 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3331 return error("invalid argument to -C: %s", arg+2);
3332 options->detect_rename = DIFF_DETECT_COPY;
3333 }
3334 else if (!strcmp(arg, "--no-renames"))
3335 options->detect_rename = 0;
3336 else if (!strcmp(arg, "--relative"))
3337 DIFF_OPT_SET(options, RELATIVE_NAME);
3338 else if (!prefixcmp(arg, "--relative=")) {
3339 DIFF_OPT_SET(options, RELATIVE_NAME);
3340 options->prefix = arg + 11;
3341 }
3343 /* xdiff options */
3344 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3345 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3346 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3347 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3348 else if (!strcmp(arg, "--ignore-space-at-eol"))
3349 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3350 else if (!strcmp(arg, "--patience"))
3351 DIFF_XDL_SET(options, PATIENCE_DIFF);
3353 /* flags options */
3354 else if (!strcmp(arg, "--binary")) {
3355 options->output_format |= DIFF_FORMAT_PATCH;
3356 DIFF_OPT_SET(options, BINARY);
3357 }
3358 else if (!strcmp(arg, "--full-index"))
3359 DIFF_OPT_SET(options, FULL_INDEX);
3360 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3361 DIFF_OPT_SET(options, TEXT);
3362 else if (!strcmp(arg, "-R"))
3363 DIFF_OPT_SET(options, REVERSE_DIFF);
3364 else if (!strcmp(arg, "--find-copies-harder"))
3365 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3366 else if (!strcmp(arg, "--follow"))
3367 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3368 else if (!strcmp(arg, "--color"))
3369 DIFF_OPT_SET(options, COLOR_DIFF);
3370 else if (!prefixcmp(arg, "--color=")) {
3371 int value = git_config_colorbool(NULL, arg+8, -1);
3372 if (value == 0)
3373 DIFF_OPT_CLR(options, COLOR_DIFF);
3374 else if (value > 0)
3375 DIFF_OPT_SET(options, COLOR_DIFF);
3376 else
3377 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3378 }
3379 else if (!strcmp(arg, "--no-color"))
3380 DIFF_OPT_CLR(options, COLOR_DIFF);
3381 else if (!strcmp(arg, "--color-words")) {
3382 DIFF_OPT_SET(options, COLOR_DIFF);
3383 options->word_diff = DIFF_WORDS_COLOR;
3384 }
3385 else if (!prefixcmp(arg, "--color-words=")) {
3386 DIFF_OPT_SET(options, COLOR_DIFF);
3387 options->word_diff = DIFF_WORDS_COLOR;
3388 options->word_regex = arg + 14;
3389 }
3390 else if (!strcmp(arg, "--word-diff")) {
3391 if (options->word_diff == DIFF_WORDS_NONE)
3392 options->word_diff = DIFF_WORDS_PLAIN;
3393 }
3394 else if (!prefixcmp(arg, "--word-diff=")) {
3395 const char *type = arg + 12;
3396 if (!strcmp(type, "plain"))
3397 options->word_diff = DIFF_WORDS_PLAIN;
3398 else if (!strcmp(type, "color")) {
3399 DIFF_OPT_SET(options, COLOR_DIFF);
3400 options->word_diff = DIFF_WORDS_COLOR;
3401 }
3402 else if (!strcmp(type, "porcelain"))
3403 options->word_diff = DIFF_WORDS_PORCELAIN;
3404 else if (!strcmp(type, "none"))
3405 options->word_diff = DIFF_WORDS_NONE;
3406 else
3407 die("bad --word-diff argument: %s", type);
3408 }
3409 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3410 if (options->word_diff == DIFF_WORDS_NONE)
3411 options->word_diff = DIFF_WORDS_PLAIN;
3412 options->word_regex = optarg;
3413 return argcount;
3414 }
3415 else if (!strcmp(arg, "--exit-code"))
3416 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3417 else if (!strcmp(arg, "--quiet"))
3418 DIFF_OPT_SET(options, QUICK);
3419 else if (!strcmp(arg, "--ext-diff"))
3420 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3421 else if (!strcmp(arg, "--no-ext-diff"))
3422 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3423 else if (!strcmp(arg, "--textconv"))
3424 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3425 else if (!strcmp(arg, "--no-textconv"))
3426 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3427 else if (!strcmp(arg, "--ignore-submodules")) {
3428 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3429 handle_ignore_submodules_arg(options, "all");
3430 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3431 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3432 handle_ignore_submodules_arg(options, arg + 20);
3433 } else if (!strcmp(arg, "--submodule"))
3434 DIFF_OPT_SET(options, SUBMODULE_LOG);
3435 else if (!prefixcmp(arg, "--submodule=")) {
3436 if (!strcmp(arg + 12, "log"))
3437 DIFF_OPT_SET(options, SUBMODULE_LOG);
3438 }
3440 /* misc options */
3441 else if (!strcmp(arg, "-z"))
3442 options->line_termination = 0;
3443 else if ((argcount = short_opt('l', av, &optarg))) {
3444 options->rename_limit = strtoul(optarg, NULL, 10);
3445 return argcount;
3446 }
3447 else if ((argcount = short_opt('S', av, &optarg))) {
3448 options->pickaxe = optarg;
3449 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3450 return argcount;
3451 } else if ((argcount = short_opt('G', av, &optarg))) {
3452 options->pickaxe = optarg;
3453 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3454 return argcount;
3455 }
3456 else if (!strcmp(arg, "--pickaxe-all"))
3457 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3458 else if (!strcmp(arg, "--pickaxe-regex"))
3459 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3460 else if ((argcount = short_opt('O', av, &optarg))) {
3461 options->orderfile = optarg;
3462 return argcount;
3463 }
3464 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3465 options->filter = optarg;
3466 return argcount;
3467 }
3468 else if (!strcmp(arg, "--abbrev"))
3469 options->abbrev = DEFAULT_ABBREV;
3470 else if (!prefixcmp(arg, "--abbrev=")) {
3471 options->abbrev = strtoul(arg + 9, NULL, 10);
3472 if (options->abbrev < MINIMUM_ABBREV)
3473 options->abbrev = MINIMUM_ABBREV;
3474 else if (40 < options->abbrev)
3475 options->abbrev = 40;
3476 }
3477 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3478 options->a_prefix = optarg;
3479 return argcount;
3480 }
3481 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3482 options->b_prefix = optarg;
3483 return argcount;
3484 }
3485 else if (!strcmp(arg, "--no-prefix"))
3486 options->a_prefix = options->b_prefix = "";
3487 else if (opt_arg(arg, '\0', "inter-hunk-context",
3488 &options->interhunkcontext))
3489 ;
3490 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3491 options->file = fopen(optarg, "w");
3492 if (!options->file)
3493 die_errno("Could not open '%s'", optarg);
3494 options->close_file = 1;
3495 return argcount;
3496 } else
3497 return 0;
3498 return 1;
3499 }
3501 int parse_rename_score(const char **cp_p)
3502 {
3503 unsigned long num, scale;
3504 int ch, dot;
3505 const char *cp = *cp_p;
3507 num = 0;
3508 scale = 1;
3509 dot = 0;
3510 for (;;) {
3511 ch = *cp;
3512 if ( !dot && ch == '.' ) {
3513 scale = 1;
3514 dot = 1;
3515 } else if ( ch == '%' ) {
3516 scale = dot ? scale*100 : 100;
3517 cp++; /* % is always at the end */
3518 break;
3519 } else if ( ch >= '0' && ch <= '9' ) {
3520 if ( scale < 100000 ) {
3521 scale *= 10;
3522 num = (num*10) + (ch-'0');
3523 }
3524 } else {
3525 break;
3526 }
3527 cp++;
3528 }
3529 *cp_p = cp;
3531 /* user says num divided by scale and we say internally that
3532 * is MAX_SCORE * num / scale.
3533 */
3534 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3535 }
3537 static int diff_scoreopt_parse(const char *opt)
3538 {
3539 int opt1, opt2, cmd;
3541 if (*opt++ != '-')
3542 return -1;
3543 cmd = *opt++;
3544 if (cmd == '-') {
3545 /* convert the long-form arguments into short-form versions */
3546 if (!prefixcmp(opt, "break-rewrites")) {
3547 opt += strlen("break-rewrites");
3548 if (*opt == 0 || *opt++ == '=')
3549 cmd = 'B';
3550 } else if (!prefixcmp(opt, "find-copies")) {
3551 opt += strlen("find-copies");
3552 if (*opt == 0 || *opt++ == '=')
3553 cmd = 'C';
3554 } else if (!prefixcmp(opt, "find-renames")) {
3555 opt += strlen("find-renames");
3556 if (*opt == 0 || *opt++ == '=')
3557 cmd = 'M';
3558 }
3559 }
3560 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3561 return -1; /* that is not a -M, -C nor -B option */
3563 opt1 = parse_rename_score(&opt);
3564 if (cmd != 'B')
3565 opt2 = 0;
3566 else {
3567 if (*opt == 0)
3568 opt2 = 0;
3569 else if (*opt != '/')
3570 return -1; /* we expect -B80/99 or -B80 */
3571 else {
3572 opt++;
3573 opt2 = parse_rename_score(&opt);
3574 }
3575 }
3576 if (*opt != 0)
3577 return -1;
3578 return opt1 | (opt2 << 16);
3579 }
3581 struct diff_queue_struct diff_queued_diff;
3583 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3584 {
3585 if (queue->alloc <= queue->nr) {
3586 queue->alloc = alloc_nr(queue->alloc);
3587 queue->queue = xrealloc(queue->queue,
3588 sizeof(dp) * queue->alloc);
3589 }
3590 queue->queue[queue->nr++] = dp;
3591 }
3593 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3594 struct diff_filespec *one,
3595 struct diff_filespec *two)
3596 {
3597 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3598 dp->one = one;
3599 dp->two = two;
3600 if (queue)
3601 diff_q(queue, dp);
3602 return dp;
3603 }
3605 void diff_free_filepair(struct diff_filepair *p)
3606 {
3607 free_filespec(p->one);
3608 free_filespec(p->two);
3609 free(p);
3610 }
3612 /* This is different from find_unique_abbrev() in that
3613 * it stuffs the result with dots for alignment.
3614 */
3615 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3616 {
3617 int abblen;
3618 const char *abbrev;
3619 if (len == 40)
3620 return sha1_to_hex(sha1);
3622 abbrev = find_unique_abbrev(sha1, len);
3623 abblen = strlen(abbrev);
3624 if (abblen < 37) {
3625 static char hex[41];
3626 if (len < abblen && abblen <= len + 2)
3627 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3628 else
3629 sprintf(hex, "%s...", abbrev);
3630 return hex;
3631 }
3632 return sha1_to_hex(sha1);
3633 }
3635 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3636 {
3637 int line_termination = opt->line_termination;
3638 int inter_name_termination = line_termination ? '\t' : '\0';
3639 if (opt->output_prefix) {
3640 struct strbuf *msg = NULL;
3641 msg = opt->output_prefix(opt, opt->output_prefix_data);
3642 fprintf(opt->file, "%s", msg->buf);
3643 }
3645 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3646 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3647 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3648 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3649 }
3650 if (p->score) {
3651 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3652 inter_name_termination);
3653 } else {
3654 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3655 }
3657 if (p->status == DIFF_STATUS_COPIED ||
3658 p->status == DIFF_STATUS_RENAMED) {
3659 const char *name_a, *name_b;
3660 name_a = p->one->path;
3661 name_b = p->two->path;
3662 strip_prefix(opt->prefix_length, &name_a, &name_b);
3663 write_name_quoted(name_a, opt->file, inter_name_termination);
3664 write_name_quoted(name_b, opt->file, line_termination);
3665 } else {
3666 const char *name_a, *name_b;
3667 name_a = p->one->mode ? p->one->path : p->two->path;
3668 name_b = NULL;
3669 strip_prefix(opt->prefix_length, &name_a, &name_b);
3670 write_name_quoted(name_a, opt->file, line_termination);
3671 }
3672 }
3674 int diff_unmodified_pair(struct diff_filepair *p)
3675 {
3676 /* This function is written stricter than necessary to support
3677 * the currently implemented transformers, but the idea is to
3678 * let transformers to produce diff_filepairs any way they want,
3679 * and filter and clean them up here before producing the output.
3680 */
3681 struct diff_filespec *one = p->one, *two = p->two;
3683 if (DIFF_PAIR_UNMERGED(p))
3684 return 0; /* unmerged is interesting */
3686 /* deletion, addition, mode or type change
3687 * and rename are all interesting.
3688 */
3689 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3690 DIFF_PAIR_MODE_CHANGED(p) ||
3691 strcmp(one->path, two->path))
3692 return 0;
3694 /* both are valid and point at the same path. that is, we are
3695 * dealing with a change.
3696 */
3697 if (one->sha1_valid && two->sha1_valid &&
3698 !hashcmp(one->sha1, two->sha1) &&
3699 !one->dirty_submodule && !two->dirty_submodule)
3700 return 1; /* no change */
3701 if (!one->sha1_valid && !two->sha1_valid)
3702 return 1; /* both look at the same file on the filesystem. */
3703 return 0;
3704 }
3706 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3707 {
3708 if (diff_unmodified_pair(p))
3709 return;
3711 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3712 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3713 return; /* no tree diffs in patch format */
3715 run_diff(p, o);
3716 }
3718 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3719 struct diffstat_t *diffstat)
3720 {
3721 if (diff_unmodified_pair(p))
3722 return;
3724 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3725 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3726 return; /* no useful stat for tree diffs */
3728 run_diffstat(p, o, diffstat);
3729 }
3731 static void diff_flush_checkdiff(struct diff_filepair *p,
3732 struct diff_options *o)
3733 {
3734 if (diff_unmodified_pair(p))
3735 return;
3737 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3738 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3739 return; /* nothing to check in tree diffs */
3741 run_checkdiff(p, o);
3742 }
3744 int diff_queue_is_empty(void)
3745 {
3746 struct diff_queue_struct *q = &diff_queued_diff;
3747 int i;
3748 for (i = 0; i < q->nr; i++)
3749 if (!diff_unmodified_pair(q->queue[i]))
3750 return 0;
3751 return 1;
3752 }
3754 #if DIFF_DEBUG
3755 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3756 {
3757 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3758 x, one ? one : "",
3759 s->path,
3760 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3761 s->mode,
3762 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3763 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3764 x, one ? one : "",
3765 s->size, s->xfrm_flags);
3766 }
3768 void diff_debug_filepair(const struct diff_filepair *p, int i)
3769 {
3770 diff_debug_filespec(p->one, i, "one");
3771 diff_debug_filespec(p->two, i, "two");
3772 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3773 p->score, p->status ? p->status : '?',
3774 p->one->rename_used, p->broken_pair);
3775 }
3777 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3778 {
3779 int i;
3780 if (msg)
3781 fprintf(stderr, "%s\n", msg);
3782 fprintf(stderr, "q->nr = %d\n", q->nr);
3783 for (i = 0; i < q->nr; i++) {
3784 struct diff_filepair *p = q->queue[i];
3785 diff_debug_filepair(p, i);
3786 }
3787 }
3788 #endif
3790 static void diff_resolve_rename_copy(void)
3791 {
3792 int i;
3793 struct diff_filepair *p;
3794 struct diff_queue_struct *q = &diff_queued_diff;
3796 diff_debug_queue("resolve-rename-copy", q);
3798 for (i = 0; i < q->nr; i++) {
3799 p = q->queue[i];
3800 p->status = 0; /* undecided */
3801 if (DIFF_PAIR_UNMERGED(p))
3802 p->status = DIFF_STATUS_UNMERGED;
3803 else if (!DIFF_FILE_VALID(p->one))
3804 p->status = DIFF_STATUS_ADDED;
3805 else if (!DIFF_FILE_VALID(p->two))
3806 p->status = DIFF_STATUS_DELETED;
3807 else if (DIFF_PAIR_TYPE_CHANGED(p))
3808 p->status = DIFF_STATUS_TYPE_CHANGED;
3810 /* from this point on, we are dealing with a pair
3811 * whose both sides are valid and of the same type, i.e.
3812 * either in-place edit or rename/copy edit.
3813 */
3814 else if (DIFF_PAIR_RENAME(p)) {
3815 /*
3816 * A rename might have re-connected a broken
3817 * pair up, causing the pathnames to be the
3818 * same again. If so, that's not a rename at
3819 * all, just a modification..
3820 *
3821 * Otherwise, see if this source was used for
3822 * multiple renames, in which case we decrement
3823 * the count, and call it a copy.
3824 */
3825 if (!strcmp(p->one->path, p->two->path))
3826 p->status = DIFF_STATUS_MODIFIED;
3827 else if (--p->one->rename_used > 0)
3828 p->status = DIFF_STATUS_COPIED;
3829 else
3830 p->status = DIFF_STATUS_RENAMED;
3831 }
3832 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3833 p->one->mode != p->two->mode ||
3834 p->one->dirty_submodule ||
3835 p->two->dirty_submodule ||
3836 is_null_sha1(p->one->sha1))
3837 p->status = DIFF_STATUS_MODIFIED;
3838 else {
3839 /* This is a "no-change" entry and should not
3840 * happen anymore, but prepare for broken callers.
3841 */
3842 error("feeding unmodified %s to diffcore",
3843 p->one->path);
3844 p->status = DIFF_STATUS_UNKNOWN;
3845 }
3846 }
3847 diff_debug_queue("resolve-rename-copy done", q);
3848 }
3850 static int check_pair_status(struct diff_filepair *p)
3851 {
3852 switch (p->status) {
3853 case DIFF_STATUS_UNKNOWN:
3854 return 0;
3855 case 0:
3856 die("internal error in diff-resolve-rename-copy");
3857 default:
3858 return 1;
3859 }
3860 }
3862 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3863 {
3864 int fmt = opt->output_format;
3866 if (fmt & DIFF_FORMAT_CHECKDIFF)
3867 diff_flush_checkdiff(p, opt);
3868 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3869 diff_flush_raw(p, opt);
3870 else if (fmt & DIFF_FORMAT_NAME) {
3871 const char *name_a, *name_b;
3872 name_a = p->two->path;
3873 name_b = NULL;
3874 strip_prefix(opt->prefix_length, &name_a, &name_b);
3875 write_name_quoted(name_a, opt->file, opt->line_termination);
3876 }
3877 }
3879 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3880 {
3881 if (fs->mode)
3882 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3883 else
3884 fprintf(file, " %s ", newdelete);
3885 write_name_quoted(fs->path, file, '\n');
3886 }
3889 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3890 const char *line_prefix)
3891 {
3892 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3893 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3894 p->two->mode, show_name ? ' ' : '\n');
3895 if (show_name) {
3896 write_name_quoted(p->two->path, file, '\n');
3897 }
3898 }
3899 }
3901 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3902 const char *line_prefix)
3903 {
3904 char *names = pprint_rename(p->one->path, p->two->path);
3906 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3907 free(names);
3908 show_mode_change(file, p, 0, line_prefix);
3909 }
3911 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3912 {
3913 FILE *file = opt->file;
3914 char *line_prefix = "";
3916 if (opt->output_prefix) {
3917 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3918 line_prefix = buf->buf;
3919 }
3921 switch(p->status) {
3922 case DIFF_STATUS_DELETED:
3923 fputs(line_prefix, file);
3924 show_file_mode_name(file, "delete", p->one);
3925 break;
3926 case DIFF_STATUS_ADDED:
3927 fputs(line_prefix, file);
3928 show_file_mode_name(file, "create", p->two);
3929 break;
3930 case DIFF_STATUS_COPIED:
3931 fputs(line_prefix, file);
3932 show_rename_copy(file, "copy", p, line_prefix);
3933 break;
3934 case DIFF_STATUS_RENAMED:
3935 fputs(line_prefix, file);
3936 show_rename_copy(file, "rename", p, line_prefix);
3937 break;
3938 default:
3939 if (p->score) {
3940 fprintf(file, "%s rewrite ", line_prefix);
3941 write_name_quoted(p->two->path, file, ' ');
3942 fprintf(file, "(%d%%)\n", similarity_index(p));
3943 }
3944 show_mode_change(file, p, !p->score, line_prefix);
3945 break;
3946 }
3947 }
3949 struct patch_id_t {
3950 git_SHA_CTX *ctx;
3951 int patchlen;
3952 };
3954 static int remove_space(char *line, int len)
3955 {
3956 int i;
3957 char *dst = line;
3958 unsigned char c;
3960 for (i = 0; i < len; i++)
3961 if (!isspace((c = line[i])))
3962 *dst++ = c;
3964 return dst - line;
3965 }
3967 static void patch_id_consume(void *priv, char *line, unsigned long len)
3968 {
3969 struct patch_id_t *data = priv;
3970 int new_len;
3972 /* Ignore line numbers when computing the SHA1 of the patch */
3973 if (!prefixcmp(line, "@@ -"))
3974 return;
3976 new_len = remove_space(line, len);
3978 git_SHA1_Update(data->ctx, line, new_len);
3979 data->patchlen += new_len;
3980 }
3982 /* returns 0 upon success, and writes result into sha1 */
3983 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3984 {
3985 struct diff_queue_struct *q = &diff_queued_diff;
3986 int i;
3987 git_SHA_CTX ctx;
3988 struct patch_id_t data;
3989 char buffer[PATH_MAX * 4 + 20];
3991 git_SHA1_Init(&ctx);
3992 memset(&data, 0, sizeof(struct patch_id_t));
3993 data.ctx = &ctx;
3995 for (i = 0; i < q->nr; i++) {
3996 xpparam_t xpp;
3997 xdemitconf_t xecfg;
3998 mmfile_t mf1, mf2;
3999 struct diff_filepair *p = q->queue[i];
4000 int len1, len2;
4002 memset(&xpp, 0, sizeof(xpp));
4003 memset(&xecfg, 0, sizeof(xecfg));
4004 if (p->status == 0)
4005 return error("internal diff status error");
4006 if (p->status == DIFF_STATUS_UNKNOWN)
4007 continue;
4008 if (diff_unmodified_pair(p))
4009 continue;
4010 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4011 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4012 continue;
4013 if (DIFF_PAIR_UNMERGED(p))
4014 continue;
4016 diff_fill_sha1_info(p->one);
4017 diff_fill_sha1_info(p->two);
4018 if (fill_mmfile(&mf1, p->one) < 0 ||
4019 fill_mmfile(&mf2, p->two) < 0)
4020 return error("unable to read files to diff");
4022 len1 = remove_space(p->one->path, strlen(p->one->path));
4023 len2 = remove_space(p->two->path, strlen(p->two->path));
4024 if (p->one->mode == 0)
4025 len1 = snprintf(buffer, sizeof(buffer),
4026 "diff--gita/%.*sb/%.*s"
4027 "newfilemode%06o"
4028 "---/dev/null"
4029 "+++b/%.*s",
4030 len1, p->one->path,
4031 len2, p->two->path,
4032 p->two->mode,
4033 len2, p->two->path);
4034 else if (p->two->mode == 0)
4035 len1 = snprintf(buffer, sizeof(buffer),
4036 "diff--gita/%.*sb/%.*s"
4037 "deletedfilemode%06o"
4038 "---a/%.*s"
4039 "+++/dev/null",
4040 len1, p->one->path,
4041 len2, p->two->path,
4042 p->one->mode,
4043 len1, p->one->path);
4044 else
4045 len1 = snprintf(buffer, sizeof(buffer),
4046 "diff--gita/%.*sb/%.*s"
4047 "---a/%.*s"
4048 "+++b/%.*s",
4049 len1, p->one->path,
4050 len2, p->two->path,
4051 len1, p->one->path,
4052 len2, p->two->path);
4053 git_SHA1_Update(&ctx, buffer, len1);
4055 if (diff_filespec_is_binary(p->one) ||
4056 diff_filespec_is_binary(p->two)) {
4057 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4058 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4059 continue;
4060 }
4062 xpp.flags = 0;
4063 xecfg.ctxlen = 3;
4064 xecfg.flags = 0;
4065 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4066 &xpp, &xecfg);
4067 }
4069 git_SHA1_Final(sha1, &ctx);
4070 return 0;
4071 }
4073 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4074 {
4075 struct diff_queue_struct *q = &diff_queued_diff;
4076 int i;
4077 int result = diff_get_patch_id(options, sha1);
4079 for (i = 0; i < q->nr; i++)
4080 diff_free_filepair(q->queue[i]);
4082 free(q->queue);
4083 DIFF_QUEUE_CLEAR(q);
4085 return result;
4086 }
4088 static int is_summary_empty(const struct diff_queue_struct *q)
4089 {
4090 int i;
4092 for (i = 0; i < q->nr; i++) {
4093 const struct diff_filepair *p = q->queue[i];
4095 switch (p->status) {
4096 case DIFF_STATUS_DELETED:
4097 case DIFF_STATUS_ADDED:
4098 case DIFF_STATUS_COPIED:
4099 case DIFF_STATUS_RENAMED:
4100 return 0;
4101 default:
4102 if (p->score)
4103 return 0;
4104 if (p->one->mode && p->two->mode &&
4105 p->one->mode != p->two->mode)
4106 return 0;
4107 break;
4108 }
4109 }
4110 return 1;
4111 }
4113 static const char rename_limit_warning[] =
4114 "inexact rename detection was skipped due to too many files.";
4116 static const char degrade_cc_to_c_warning[] =
4117 "only found copies from modified paths due to too many files.";
4119 static const char rename_limit_advice[] =
4120 "you may want to set your %s variable to at least "
4121 "%d and retry the command.";
4123 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4124 {
4125 if (degraded_cc)
4126 warning(degrade_cc_to_c_warning);
4127 else if (needed)
4128 warning(rename_limit_warning);
4129 else
4130 return;
4131 if (0 < needed && needed < 32767)
4132 warning(rename_limit_advice, varname, needed);
4133 }
4135 void diff_flush(struct diff_options *options)
4136 {
4137 struct diff_queue_struct *q = &diff_queued_diff;
4138 int i, output_format = options->output_format;
4139 int separator = 0;
4140 int dirstat_by_line = 0;
4142 /*
4143 * Order: raw, stat, summary, patch
4144 * or: name/name-status/checkdiff (other bits clear)
4145 */
4146 if (!q->nr)
4147 goto free_queue;
4149 if (output_format & (DIFF_FORMAT_RAW |
4150 DIFF_FORMAT_NAME |
4151 DIFF_FORMAT_NAME_STATUS |
4152 DIFF_FORMAT_CHECKDIFF)) {
4153 for (i = 0; i < q->nr; i++) {
4154 struct diff_filepair *p = q->queue[i];
4155 if (check_pair_status(p))
4156 flush_one_pair(p, options);
4157 }
4158 separator++;
4159 }
4161 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4162 dirstat_by_line = 1;
4164 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4165 dirstat_by_line) {
4166 struct diffstat_t diffstat;
4168 memset(&diffstat, 0, sizeof(struct diffstat_t));
4169 for (i = 0; i < q->nr; i++) {
4170 struct diff_filepair *p = q->queue[i];
4171 if (check_pair_status(p))
4172 diff_flush_stat(p, options, &diffstat);
4173 }
4174 if (output_format & DIFF_FORMAT_NUMSTAT)
4175 show_numstat(&diffstat, options);
4176 if (output_format & DIFF_FORMAT_DIFFSTAT)
4177 show_stats(&diffstat, options);
4178 if (output_format & DIFF_FORMAT_SHORTSTAT)
4179 show_shortstats(&diffstat, options);
4180 if (output_format & DIFF_FORMAT_DIRSTAT)
4181 show_dirstat_by_line(&diffstat, options);
4182 free_diffstat_info(&diffstat);
4183 separator++;
4184 }
4185 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4186 show_dirstat(options);
4188 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4189 for (i = 0; i < q->nr; i++) {
4190 diff_summary(options, q->queue[i]);
4191 }
4192 separator++;
4193 }
4195 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4196 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4197 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4198 /*
4199 * run diff_flush_patch for the exit status. setting
4200 * options->file to /dev/null should be safe, becaue we
4201 * aren't supposed to produce any output anyway.
4202 */
4203 if (options->close_file)
4204 fclose(options->file);
4205 options->file = fopen("/dev/null", "w");
4206 if (!options->file)
4207 die_errno("Could not open /dev/null");
4208 options->close_file = 1;
4209 for (i = 0; i < q->nr; i++) {
4210 struct diff_filepair *p = q->queue[i];
4211 if (check_pair_status(p))
4212 diff_flush_patch(p, options);
4213 if (options->found_changes)
4214 break;
4215 }
4216 }
4218 if (output_format & DIFF_FORMAT_PATCH) {
4219 if (separator) {
4220 putc(options->line_termination, options->file);
4221 if (options->stat_sep) {
4222 /* attach patch instead of inline */
4223 fputs(options->stat_sep, options->file);
4224 }
4225 }
4227 for (i = 0; i < q->nr; i++) {
4228 struct diff_filepair *p = q->queue[i];
4229 if (check_pair_status(p))
4230 diff_flush_patch(p, options);
4231 }
4232 }
4234 if (output_format & DIFF_FORMAT_CALLBACK)
4235 options->format_callback(q, options, options->format_callback_data);
4237 for (i = 0; i < q->nr; i++)
4238 diff_free_filepair(q->queue[i]);
4239 free_queue:
4240 free(q->queue);
4241 DIFF_QUEUE_CLEAR(q);
4242 if (options->close_file)
4243 fclose(options->file);
4245 /*
4246 * Report the content-level differences with HAS_CHANGES;
4247 * diff_addremove/diff_change does not set the bit when
4248 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4249 */
4250 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4251 if (options->found_changes)
4252 DIFF_OPT_SET(options, HAS_CHANGES);
4253 else
4254 DIFF_OPT_CLR(options, HAS_CHANGES);
4255 }
4256 }
4258 static void diffcore_apply_filter(const char *filter)
4259 {
4260 int i;
4261 struct diff_queue_struct *q = &diff_queued_diff;
4262 struct diff_queue_struct outq;
4263 DIFF_QUEUE_CLEAR(&outq);
4265 if (!filter)
4266 return;
4268 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4269 int found;
4270 for (i = found = 0; !found && i < q->nr; i++) {
4271 struct diff_filepair *p = q->queue[i];
4272 if (((p->status == DIFF_STATUS_MODIFIED) &&
4273 ((p->score &&
4274 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4275 (!p->score &&
4276 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4277 ((p->status != DIFF_STATUS_MODIFIED) &&
4278 strchr(filter, p->status)))
4279 found++;
4280 }
4281 if (found)
4282 return;
4284 /* otherwise we will clear the whole queue
4285 * by copying the empty outq at the end of this
4286 * function, but first clear the current entries
4287 * in the queue.
4288 */
4289 for (i = 0; i < q->nr; i++)
4290 diff_free_filepair(q->queue[i]);
4291 }
4292 else {
4293 /* Only the matching ones */
4294 for (i = 0; i < q->nr; i++) {
4295 struct diff_filepair *p = q->queue[i];
4297 if (((p->status == DIFF_STATUS_MODIFIED) &&
4298 ((p->score &&
4299 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4300 (!p->score &&
4301 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4302 ((p->status != DIFF_STATUS_MODIFIED) &&
4303 strchr(filter, p->status)))
4304 diff_q(&outq, p);
4305 else
4306 diff_free_filepair(p);
4307 }
4308 }
4309 free(q->queue);
4310 *q = outq;
4311 }
4313 /* Check whether two filespecs with the same mode and size are identical */
4314 static int diff_filespec_is_identical(struct diff_filespec *one,
4315 struct diff_filespec *two)
4316 {
4317 if (S_ISGITLINK(one->mode))
4318 return 0;
4319 if (diff_populate_filespec(one, 0))
4320 return 0;
4321 if (diff_populate_filespec(two, 0))
4322 return 0;
4323 return !memcmp(one->data, two->data, one->size);
4324 }
4326 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4327 {
4328 int i;
4329 struct diff_queue_struct *q = &diff_queued_diff;
4330 struct diff_queue_struct outq;
4331 DIFF_QUEUE_CLEAR(&outq);
4333 for (i = 0; i < q->nr; i++) {
4334 struct diff_filepair *p = q->queue[i];
4336 /*
4337 * 1. Entries that come from stat info dirtiness
4338 * always have both sides (iow, not create/delete),
4339 * one side of the object name is unknown, with
4340 * the same mode and size. Keep the ones that
4341 * do not match these criteria. They have real
4342 * differences.
4343 *
4344 * 2. At this point, the file is known to be modified,
4345 * with the same mode and size, and the object
4346 * name of one side is unknown. Need to inspect
4347 * the identical contents.
4348 */
4349 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4350 !DIFF_FILE_VALID(p->two) ||
4351 (p->one->sha1_valid && p->two->sha1_valid) ||
4352 (p->one->mode != p->two->mode) ||
4353 diff_populate_filespec(p->one, 1) ||
4354 diff_populate_filespec(p->two, 1) ||
4355 (p->one->size != p->two->size) ||
4356 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4357 diff_q(&outq, p);
4358 else {
4359 /*
4360 * The caller can subtract 1 from skip_stat_unmatch
4361 * to determine how many paths were dirty only
4362 * due to stat info mismatch.
4363 */
4364 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4365 diffopt->skip_stat_unmatch++;
4366 diff_free_filepair(p);
4367 }
4368 }
4369 free(q->queue);
4370 *q = outq;
4371 }
4373 static int diffnamecmp(const void *a_, const void *b_)
4374 {
4375 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4376 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4377 const char *name_a, *name_b;
4379 name_a = a->one ? a->one->path : a->two->path;
4380 name_b = b->one ? b->one->path : b->two->path;
4381 return strcmp(name_a, name_b);
4382 }
4384 void diffcore_fix_diff_index(struct diff_options *options)
4385 {
4386 struct diff_queue_struct *q = &diff_queued_diff;
4387 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4388 }
4390 void diffcore_std(struct diff_options *options)
4391 {
4392 if (options->skip_stat_unmatch)
4393 diffcore_skip_stat_unmatch(options);
4394 if (!options->found_follow) {
4395 /* See try_to_follow_renames() in tree-diff.c */
4396 if (options->break_opt != -1)
4397 diffcore_break(options->break_opt);
4398 if (options->detect_rename)
4399 diffcore_rename(options);
4400 if (options->break_opt != -1)
4401 diffcore_merge_broken();
4402 }
4403 if (options->pickaxe)
4404 diffcore_pickaxe(options);
4405 if (options->orderfile)
4406 diffcore_order(options->orderfile);
4407 if (!options->found_follow)
4408 /* See try_to_follow_renames() in tree-diff.c */
4409 diff_resolve_rename_copy();
4410 diffcore_apply_filter(options->filter);
4412 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4413 DIFF_OPT_SET(options, HAS_CHANGES);
4414 else
4415 DIFF_OPT_CLR(options, HAS_CHANGES);
4417 options->found_follow = 0;
4418 }
4420 int diff_result_code(struct diff_options *opt, int status)
4421 {
4422 int result = 0;
4424 diff_warn_rename_limit("diff.renamelimit",
4425 opt->needed_rename_limit,
4426 opt->degraded_cc_to_c);
4427 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4428 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4429 return status;
4430 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4431 DIFF_OPT_TST(opt, HAS_CHANGES))
4432 result |= 01;
4433 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4434 DIFF_OPT_TST(opt, CHECK_FAILED))
4435 result |= 02;
4436 return result;
4437 }
4439 /*
4440 * Shall changes to this submodule be ignored?
4441 *
4442 * Submodule changes can be configured to be ignored separately for each path,
4443 * but that configuration can be overridden from the command line.
4444 */
4445 static int is_submodule_ignored(const char *path, struct diff_options *options)
4446 {
4447 int ignored = 0;
4448 unsigned orig_flags = options->flags;
4449 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4450 set_diffopt_flags_from_submodule_config(options, path);
4451 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4452 ignored = 1;
4453 options->flags = orig_flags;
4454 return ignored;
4455 }
4457 void diff_addremove(struct diff_options *options,
4458 int addremove, unsigned mode,
4459 const unsigned char *sha1,
4460 const char *concatpath, unsigned dirty_submodule)
4461 {
4462 struct diff_filespec *one, *two;
4464 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4465 return;
4467 /* This may look odd, but it is a preparation for
4468 * feeding "there are unchanged files which should
4469 * not produce diffs, but when you are doing copy
4470 * detection you would need them, so here they are"
4471 * entries to the diff-core. They will be prefixed
4472 * with something like '=' or '*' (I haven't decided
4473 * which but should not make any difference).
4474 * Feeding the same new and old to diff_change()
4475 * also has the same effect.
4476 * Before the final output happens, they are pruned after
4477 * merged into rename/copy pairs as appropriate.
4478 */
4479 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4480 addremove = (addremove == '+' ? '-' :
4481 addremove == '-' ? '+' : addremove);
4483 if (options->prefix &&
4484 strncmp(concatpath, options->prefix, options->prefix_length))
4485 return;
4487 one = alloc_filespec(concatpath);
4488 two = alloc_filespec(concatpath);
4490 if (addremove != '+')
4491 fill_filespec(one, sha1, mode);
4492 if (addremove != '-') {
4493 fill_filespec(two, sha1, mode);
4494 two->dirty_submodule = dirty_submodule;
4495 }
4497 diff_queue(&diff_queued_diff, one, two);
4498 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4499 DIFF_OPT_SET(options, HAS_CHANGES);
4500 }
4502 void diff_change(struct diff_options *options,
4503 unsigned old_mode, unsigned new_mode,
4504 const unsigned char *old_sha1,
4505 const unsigned char *new_sha1,
4506 const char *concatpath,
4507 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4508 {
4509 struct diff_filespec *one, *two;
4511 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4512 is_submodule_ignored(concatpath, options))
4513 return;
4515 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4516 unsigned tmp;
4517 const unsigned char *tmp_c;
4518 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4519 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4520 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4521 new_dirty_submodule = tmp;
4522 }
4524 if (options->prefix &&
4525 strncmp(concatpath, options->prefix, options->prefix_length))
4526 return;
4528 one = alloc_filespec(concatpath);
4529 two = alloc_filespec(concatpath);
4530 fill_filespec(one, old_sha1, old_mode);
4531 fill_filespec(two, new_sha1, new_mode);
4532 one->dirty_submodule = old_dirty_submodule;
4533 two->dirty_submodule = new_dirty_submodule;
4535 diff_queue(&diff_queued_diff, one, two);
4536 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4537 DIFF_OPT_SET(options, HAS_CHANGES);
4538 }
4540 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4541 {
4542 struct diff_filepair *pair;
4543 struct diff_filespec *one, *two;
4545 if (options->prefix &&
4546 strncmp(path, options->prefix, options->prefix_length))
4547 return NULL;
4549 one = alloc_filespec(path);
4550 two = alloc_filespec(path);
4551 pair = diff_queue(&diff_queued_diff, one, two);
4552 pair->is_unmerged = 1;
4553 return pair;
4554 }
4556 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4557 size_t *outsize)
4558 {
4559 struct diff_tempfile *temp;
4560 const char *argv[3];
4561 const char **arg = argv;
4562 struct child_process child;
4563 struct strbuf buf = STRBUF_INIT;
4564 int err = 0;
4566 temp = prepare_temp_file(spec->path, spec);
4567 *arg++ = pgm;
4568 *arg++ = temp->name;
4569 *arg = NULL;
4571 memset(&child, 0, sizeof(child));
4572 child.use_shell = 1;
4573 child.argv = argv;
4574 child.out = -1;
4575 if (start_command(&child)) {
4576 remove_tempfile();
4577 return NULL;
4578 }
4580 if (strbuf_read(&buf, child.out, 0) < 0)
4581 err = error("error reading from textconv command '%s'", pgm);
4582 close(child.out);
4584 if (finish_command(&child) || err) {
4585 strbuf_release(&buf);
4586 remove_tempfile();
4587 return NULL;
4588 }
4589 remove_tempfile();
4591 return strbuf_detach(&buf, outsize);
4592 }
4594 size_t fill_textconv(struct userdiff_driver *driver,
4595 struct diff_filespec *df,
4596 char **outbuf)
4597 {
4598 size_t size;
4600 if (!driver || !driver->textconv) {
4601 if (!DIFF_FILE_VALID(df)) {
4602 *outbuf = "";
4603 return 0;
4604 }
4605 if (diff_populate_filespec(df, 0))
4606 die("unable to read files to diff");
4607 *outbuf = df->data;
4608 return df->size;
4609 }
4611 if (driver->textconv_cache && df->sha1_valid) {
4612 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4613 &size);
4614 if (*outbuf)
4615 return size;
4616 }
4618 *outbuf = run_textconv(driver->textconv, df, &size);
4619 if (!*outbuf)
4620 die("unable to read files to diff");
4622 if (driver->textconv_cache && df->sha1_valid) {
4623 /* ignore errors, as we might be in a readonly repository */
4624 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4625 size);
4626 /*
4627 * we could save up changes and flush them all at the end,
4628 * but we would need an extra call after all diffing is done.
4629 * Since generating a cache entry is the slow path anyway,
4630 * this extra overhead probably isn't a big deal.
4631 */
4632 notes_cache_write(driver->textconv_cache);
4633 }
4635 return size;
4636 }