![[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);
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 const char *name_a_tab, *name_b_tab;
587 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
588 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
589 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
590 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
591 const char *a_prefix, *b_prefix;
592 char *data_one, *data_two;
593 size_t size_one, size_two;
594 struct emit_callback ecbdata;
595 char *line_prefix = "";
596 struct strbuf *msgbuf;
598 if (o && o->output_prefix) {
599 msgbuf = o->output_prefix(o, o->output_prefix_data);
600 line_prefix = msgbuf->buf;
601 }
603 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
604 a_prefix = o->b_prefix;
605 b_prefix = o->a_prefix;
606 } else {
607 a_prefix = o->a_prefix;
608 b_prefix = o->b_prefix;
609 }
611 name_a += (*name_a == '/');
612 name_b += (*name_b == '/');
613 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
614 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
616 strbuf_reset(&a_name);
617 strbuf_reset(&b_name);
618 quote_two_c_style(&a_name, a_prefix, name_a, 0);
619 quote_two_c_style(&b_name, b_prefix, name_b, 0);
621 size_one = fill_textconv(textconv_one, one, &data_one);
622 size_two = fill_textconv(textconv_two, two, &data_two);
624 memset(&ecbdata, 0, sizeof(ecbdata));
625 ecbdata.color_diff = o->use_color > 0;
626 ecbdata.found_changesp = &o->found_changes;
627 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
628 ecbdata.opt = o;
629 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
630 mmfile_t mf1, mf2;
631 mf1.ptr = (char *)data_one;
632 mf2.ptr = (char *)data_two;
633 mf1.size = size_one;
634 mf2.size = size_two;
635 check_blank_at_eof(&mf1, &mf2, &ecbdata);
636 }
637 ecbdata.lno_in_preimage = 1;
638 ecbdata.lno_in_postimage = 1;
640 lc_a = count_lines(data_one, size_one);
641 lc_b = count_lines(data_two, size_two);
642 fprintf(o->file,
643 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
644 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
645 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
646 line_prefix, fraginfo);
647 if (!o->irreversible_delete)
648 print_line_count(o->file, lc_a);
649 else
650 fprintf(o->file, "?,?");
651 fprintf(o->file, " +");
652 print_line_count(o->file, lc_b);
653 fprintf(o->file, " @@%s\n", reset);
654 if (lc_a && !o->irreversible_delete)
655 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
656 if (lc_b)
657 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
658 if (textconv_one)
659 free((char *)data_one);
660 if (textconv_two)
661 free((char *)data_two);
662 }
664 struct diff_words_buffer {
665 mmfile_t text;
666 long alloc;
667 struct diff_words_orig {
668 const char *begin, *end;
669 } *orig;
670 int orig_nr, orig_alloc;
671 };
673 static void diff_words_append(char *line, unsigned long len,
674 struct diff_words_buffer *buffer)
675 {
676 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
677 line++;
678 len--;
679 memcpy(buffer->text.ptr + buffer->text.size, line, len);
680 buffer->text.size += len;
681 buffer->text.ptr[buffer->text.size] = '\0';
682 }
684 struct diff_words_style_elem {
685 const char *prefix;
686 const char *suffix;
687 const char *color; /* NULL; filled in by the setup code if
688 * color is enabled */
689 };
691 struct diff_words_style {
692 enum diff_words_type type;
693 struct diff_words_style_elem new, old, ctx;
694 const char *newline;
695 };
697 static struct diff_words_style diff_words_styles[] = {
698 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
699 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
700 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
701 };
703 struct diff_words_data {
704 struct diff_words_buffer minus, plus;
705 const char *current_plus;
706 int last_minus;
707 struct diff_options *opt;
708 regex_t *word_regex;
709 enum diff_words_type type;
710 struct diff_words_style *style;
711 };
713 static int fn_out_diff_words_write_helper(FILE *fp,
714 struct diff_words_style_elem *st_el,
715 const char *newline,
716 size_t count, const char *buf,
717 const char *line_prefix)
718 {
719 int print = 0;
721 while (count) {
722 char *p = memchr(buf, '\n', count);
723 if (print)
724 fputs(line_prefix, fp);
725 if (p != buf) {
726 if (st_el->color && fputs(st_el->color, fp) < 0)
727 return -1;
728 if (fputs(st_el->prefix, fp) < 0 ||
729 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
730 fputs(st_el->suffix, fp) < 0)
731 return -1;
732 if (st_el->color && *st_el->color
733 && fputs(GIT_COLOR_RESET, fp) < 0)
734 return -1;
735 }
736 if (!p)
737 return 0;
738 if (fputs(newline, fp) < 0)
739 return -1;
740 count -= p + 1 - buf;
741 buf = p + 1;
742 print = 1;
743 }
744 return 0;
745 }
747 /*
748 * '--color-words' algorithm can be described as:
749 *
750 * 1. collect a the minus/plus lines of a diff hunk, divided into
751 * minus-lines and plus-lines;
752 *
753 * 2. break both minus-lines and plus-lines into words and
754 * place them into two mmfile_t with one word for each line;
755 *
756 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
757 *
758 * And for the common parts of the both file, we output the plus side text.
759 * diff_words->current_plus is used to trace the current position of the plus file
760 * which printed. diff_words->last_minus is used to trace the last minus word
761 * printed.
762 *
763 * For '--graph' to work with '--color-words', we need to output the graph prefix
764 * on each line of color words output. Generally, there are two conditions on
765 * which we should output the prefix.
766 *
767 * 1. diff_words->last_minus == 0 &&
768 * diff_words->current_plus == diff_words->plus.text.ptr
769 *
770 * that is: the plus text must start as a new line, and if there is no minus
771 * word printed, a graph prefix must be printed.
772 *
773 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
774 * *(diff_words->current_plus - 1) == '\n'
775 *
776 * that is: a graph prefix must be printed following a '\n'
777 */
778 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
779 {
780 if ((diff_words->last_minus == 0 &&
781 diff_words->current_plus == diff_words->plus.text.ptr) ||
782 (diff_words->current_plus > diff_words->plus.text.ptr &&
783 *(diff_words->current_plus - 1) == '\n')) {
784 return 1;
785 } else {
786 return 0;
787 }
788 }
790 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
791 {
792 struct diff_words_data *diff_words = priv;
793 struct diff_words_style *style = diff_words->style;
794 int minus_first, minus_len, plus_first, plus_len;
795 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
796 struct diff_options *opt = diff_words->opt;
797 struct strbuf *msgbuf;
798 char *line_prefix = "";
800 if (line[0] != '@' || parse_hunk_header(line, len,
801 &minus_first, &minus_len, &plus_first, &plus_len))
802 return;
804 assert(opt);
805 if (opt->output_prefix) {
806 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
807 line_prefix = msgbuf->buf;
808 }
810 /* POSIX requires that first be decremented by one if len == 0... */
811 if (minus_len) {
812 minus_begin = diff_words->minus.orig[minus_first].begin;
813 minus_end =
814 diff_words->minus.orig[minus_first + minus_len - 1].end;
815 } else
816 minus_begin = minus_end =
817 diff_words->minus.orig[minus_first].end;
819 if (plus_len) {
820 plus_begin = diff_words->plus.orig[plus_first].begin;
821 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
822 } else
823 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
825 if (color_words_output_graph_prefix(diff_words)) {
826 fputs(line_prefix, diff_words->opt->file);
827 }
828 if (diff_words->current_plus != plus_begin) {
829 fn_out_diff_words_write_helper(diff_words->opt->file,
830 &style->ctx, style->newline,
831 plus_begin - diff_words->current_plus,
832 diff_words->current_plus, line_prefix);
833 if (*(plus_begin - 1) == '\n')
834 fputs(line_prefix, diff_words->opt->file);
835 }
836 if (minus_begin != minus_end) {
837 fn_out_diff_words_write_helper(diff_words->opt->file,
838 &style->old, style->newline,
839 minus_end - minus_begin, minus_begin,
840 line_prefix);
841 }
842 if (plus_begin != plus_end) {
843 fn_out_diff_words_write_helper(diff_words->opt->file,
844 &style->new, style->newline,
845 plus_end - plus_begin, plus_begin,
846 line_prefix);
847 }
849 diff_words->current_plus = plus_end;
850 diff_words->last_minus = minus_first;
851 }
853 /* This function starts looking at *begin, and returns 0 iff a word was found. */
854 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
855 int *begin, int *end)
856 {
857 if (word_regex && *begin < buffer->size) {
858 regmatch_t match[1];
859 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
860 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
861 '\n', match[0].rm_eo - match[0].rm_so);
862 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
863 *begin += match[0].rm_so;
864 return *begin >= *end;
865 }
866 return -1;
867 }
869 /* find the next word */
870 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
871 (*begin)++;
872 if (*begin >= buffer->size)
873 return -1;
875 /* find the end of the word */
876 *end = *begin + 1;
877 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
878 (*end)++;
880 return 0;
881 }
883 /*
884 * This function splits the words in buffer->text, stores the list with
885 * newline separator into out, and saves the offsets of the original words
886 * in buffer->orig.
887 */
888 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
889 regex_t *word_regex)
890 {
891 int i, j;
892 long alloc = 0;
894 out->size = 0;
895 out->ptr = NULL;
897 /* fake an empty "0th" word */
898 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
899 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
900 buffer->orig_nr = 1;
902 for (i = 0; i < buffer->text.size; i++) {
903 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
904 return;
906 /* store original boundaries */
907 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
908 buffer->orig_alloc);
909 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
910 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
911 buffer->orig_nr++;
913 /* store one word */
914 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
915 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
916 out->ptr[out->size + j - i] = '\n';
917 out->size += j - i + 1;
919 i = j - 1;
920 }
921 }
923 /* this executes the word diff on the accumulated buffers */
924 static void diff_words_show(struct diff_words_data *diff_words)
925 {
926 xpparam_t xpp;
927 xdemitconf_t xecfg;
928 mmfile_t minus, plus;
929 struct diff_words_style *style = diff_words->style;
931 struct diff_options *opt = diff_words->opt;
932 struct strbuf *msgbuf;
933 char *line_prefix = "";
935 assert(opt);
936 if (opt->output_prefix) {
937 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
938 line_prefix = msgbuf->buf;
939 }
941 /* special case: only removal */
942 if (!diff_words->plus.text.size) {
943 fputs(line_prefix, diff_words->opt->file);
944 fn_out_diff_words_write_helper(diff_words->opt->file,
945 &style->old, style->newline,
946 diff_words->minus.text.size,
947 diff_words->minus.text.ptr, line_prefix);
948 diff_words->minus.text.size = 0;
949 return;
950 }
952 diff_words->current_plus = diff_words->plus.text.ptr;
953 diff_words->last_minus = 0;
955 memset(&xpp, 0, sizeof(xpp));
956 memset(&xecfg, 0, sizeof(xecfg));
957 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
958 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
959 xpp.flags = 0;
960 /* as only the hunk header will be parsed, we need a 0-context */
961 xecfg.ctxlen = 0;
962 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
963 &xpp, &xecfg);
964 free(minus.ptr);
965 free(plus.ptr);
966 if (diff_words->current_plus != diff_words->plus.text.ptr +
967 diff_words->plus.text.size) {
968 if (color_words_output_graph_prefix(diff_words))
969 fputs(line_prefix, diff_words->opt->file);
970 fn_out_diff_words_write_helper(diff_words->opt->file,
971 &style->ctx, style->newline,
972 diff_words->plus.text.ptr + diff_words->plus.text.size
973 - diff_words->current_plus, diff_words->current_plus,
974 line_prefix);
975 }
976 diff_words->minus.text.size = diff_words->plus.text.size = 0;
977 }
979 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
980 static void diff_words_flush(struct emit_callback *ecbdata)
981 {
982 if (ecbdata->diff_words->minus.text.size ||
983 ecbdata->diff_words->plus.text.size)
984 diff_words_show(ecbdata->diff_words);
985 }
987 static void free_diff_words_data(struct emit_callback *ecbdata)
988 {
989 if (ecbdata->diff_words) {
990 diff_words_flush(ecbdata);
991 free (ecbdata->diff_words->minus.text.ptr);
992 free (ecbdata->diff_words->minus.orig);
993 free (ecbdata->diff_words->plus.text.ptr);
994 free (ecbdata->diff_words->plus.orig);
995 if (ecbdata->diff_words->word_regex) {
996 regfree(ecbdata->diff_words->word_regex);
997 free(ecbdata->diff_words->word_regex);
998 }
999 free(ecbdata->diff_words);
1000 ecbdata->diff_words = NULL;
1001 }
1002 }
1004 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1005 {
1006 if (diff_use_color > 0)
1007 return diff_colors[ix];
1008 return "";
1009 }
1011 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1012 {
1013 const char *cp;
1014 unsigned long allot;
1015 size_t l = len;
1017 if (ecb->truncate)
1018 return ecb->truncate(line, len);
1019 cp = line;
1020 allot = l;
1021 while (0 < l) {
1022 (void) utf8_width(&cp, &l);
1023 if (!cp)
1024 break; /* truncated in the middle? */
1025 }
1026 return allot - l;
1027 }
1029 static void find_lno(const char *line, struct emit_callback *ecbdata)
1030 {
1031 const char *p;
1032 ecbdata->lno_in_preimage = 0;
1033 ecbdata->lno_in_postimage = 0;
1034 p = strchr(line, '-');
1035 if (!p)
1036 return; /* cannot happen */
1037 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1038 p = strchr(p, '+');
1039 if (!p)
1040 return; /* cannot happen */
1041 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1042 }
1044 static void fn_out_consume(void *priv, char *line, unsigned long len)
1045 {
1046 struct emit_callback *ecbdata = priv;
1047 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1048 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1049 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1050 struct diff_options *o = ecbdata->opt;
1051 char *line_prefix = "";
1052 struct strbuf *msgbuf;
1054 if (o && o->output_prefix) {
1055 msgbuf = o->output_prefix(o, o->output_prefix_data);
1056 line_prefix = msgbuf->buf;
1057 }
1059 if (ecbdata->header) {
1060 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1061 strbuf_reset(ecbdata->header);
1062 ecbdata->header = NULL;
1063 }
1064 *(ecbdata->found_changesp) = 1;
1066 if (ecbdata->label_path[0]) {
1067 const char *name_a_tab, *name_b_tab;
1069 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1070 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1072 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1073 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1074 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1075 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1076 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1077 }
1079 if (diff_suppress_blank_empty
1080 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1081 line[0] = '\n';
1082 len = 1;
1083 }
1085 if (line[0] == '@') {
1086 if (ecbdata->diff_words)
1087 diff_words_flush(ecbdata);
1088 len = sane_truncate_line(ecbdata, line, len);
1089 find_lno(line, ecbdata);
1090 emit_hunk_header(ecbdata, line, len);
1091 if (line[len-1] != '\n')
1092 putc('\n', ecbdata->opt->file);
1093 return;
1094 }
1096 if (len < 1) {
1097 emit_line(ecbdata->opt, reset, reset, line, len);
1098 if (ecbdata->diff_words
1099 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1100 fputs("~\n", ecbdata->opt->file);
1101 return;
1102 }
1104 if (ecbdata->diff_words) {
1105 if (line[0] == '-') {
1106 diff_words_append(line, len,
1107 &ecbdata->diff_words->minus);
1108 return;
1109 } else if (line[0] == '+') {
1110 diff_words_append(line, len,
1111 &ecbdata->diff_words->plus);
1112 return;
1113 }
1114 diff_words_flush(ecbdata);
1115 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1116 emit_line(ecbdata->opt, plain, reset, line, len);
1117 fputs("~\n", ecbdata->opt->file);
1118 } else {
1119 /*
1120 * Skip the prefix character, if any. With
1121 * diff_suppress_blank_empty, there may be
1122 * none.
1123 */
1124 if (line[0] != '\n') {
1125 line++;
1126 len--;
1127 }
1128 emit_line(ecbdata->opt, plain, reset, line, len);
1129 }
1130 return;
1131 }
1133 if (line[0] != '+') {
1134 const char *color =
1135 diff_get_color(ecbdata->color_diff,
1136 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1137 ecbdata->lno_in_preimage++;
1138 if (line[0] == ' ')
1139 ecbdata->lno_in_postimage++;
1140 emit_line(ecbdata->opt, color, reset, line, len);
1141 } else {
1142 ecbdata->lno_in_postimage++;
1143 emit_add_line(reset, ecbdata, line + 1, len - 1);
1144 }
1145 }
1147 static char *pprint_rename(const char *a, const char *b)
1148 {
1149 const char *old = a;
1150 const char *new = b;
1151 struct strbuf name = STRBUF_INIT;
1152 int pfx_length, sfx_length;
1153 int len_a = strlen(a);
1154 int len_b = strlen(b);
1155 int a_midlen, b_midlen;
1156 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1157 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1159 if (qlen_a || qlen_b) {
1160 quote_c_style(a, &name, NULL, 0);
1161 strbuf_addstr(&name, " => ");
1162 quote_c_style(b, &name, NULL, 0);
1163 return strbuf_detach(&name, NULL);
1164 }
1166 /* Find common prefix */
1167 pfx_length = 0;
1168 while (*old && *new && *old == *new) {
1169 if (*old == '/')
1170 pfx_length = old - a + 1;
1171 old++;
1172 new++;
1173 }
1175 /* Find common suffix */
1176 old = a + len_a;
1177 new = b + len_b;
1178 sfx_length = 0;
1179 while (a <= old && b <= new && *old == *new) {
1180 if (*old == '/')
1181 sfx_length = len_a - (old - a);
1182 old--;
1183 new--;
1184 }
1186 /*
1187 * pfx{mid-a => mid-b}sfx
1188 * {pfx-a => pfx-b}sfx
1189 * pfx{sfx-a => sfx-b}
1190 * name-a => name-b
1191 */
1192 a_midlen = len_a - pfx_length - sfx_length;
1193 b_midlen = len_b - pfx_length - sfx_length;
1194 if (a_midlen < 0)
1195 a_midlen = 0;
1196 if (b_midlen < 0)
1197 b_midlen = 0;
1199 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1200 if (pfx_length + sfx_length) {
1201 strbuf_add(&name, a, pfx_length);
1202 strbuf_addch(&name, '{');
1203 }
1204 strbuf_add(&name, a + pfx_length, a_midlen);
1205 strbuf_addstr(&name, " => ");
1206 strbuf_add(&name, b + pfx_length, b_midlen);
1207 if (pfx_length + sfx_length) {
1208 strbuf_addch(&name, '}');
1209 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1210 }
1211 return strbuf_detach(&name, NULL);
1212 }
1214 struct diffstat_t {
1215 int nr;
1216 int alloc;
1217 struct diffstat_file {
1218 char *from_name;
1219 char *name;
1220 char *print_name;
1221 unsigned is_unmerged:1;
1222 unsigned is_binary:1;
1223 unsigned is_renamed:1;
1224 uintmax_t added, deleted;
1225 } **files;
1226 };
1228 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1229 const char *name_a,
1230 const char *name_b)
1231 {
1232 struct diffstat_file *x;
1233 x = xcalloc(sizeof (*x), 1);
1234 if (diffstat->nr == diffstat->alloc) {
1235 diffstat->alloc = alloc_nr(diffstat->alloc);
1236 diffstat->files = xrealloc(diffstat->files,
1237 diffstat->alloc * sizeof(x));
1238 }
1239 diffstat->files[diffstat->nr++] = x;
1240 if (name_b) {
1241 x->from_name = xstrdup(name_a);
1242 x->name = xstrdup(name_b);
1243 x->is_renamed = 1;
1244 }
1245 else {
1246 x->from_name = NULL;
1247 x->name = xstrdup(name_a);
1248 }
1249 return x;
1250 }
1252 static void diffstat_consume(void *priv, char *line, unsigned long len)
1253 {
1254 struct diffstat_t *diffstat = priv;
1255 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1257 if (line[0] == '+')
1258 x->added++;
1259 else if (line[0] == '-')
1260 x->deleted++;
1261 }
1263 const char mime_boundary_leader[] = "------------";
1265 static int scale_linear(int it, int width, int max_change)
1266 {
1267 /*
1268 * make sure that at least one '-' is printed if there were deletions,
1269 * and likewise for '+'.
1270 */
1271 if (max_change < 2)
1272 return it;
1273 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1274 }
1276 static void show_name(FILE *file,
1277 const char *prefix, const char *name, int len)
1278 {
1279 fprintf(file, " %s%-*s |", prefix, len, name);
1280 }
1282 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1283 {
1284 if (cnt <= 0)
1285 return;
1286 fprintf(file, "%s", set);
1287 while (cnt--)
1288 putc(ch, file);
1289 fprintf(file, "%s", reset);
1290 }
1292 static void fill_print_name(struct diffstat_file *file)
1293 {
1294 char *pname;
1296 if (file->print_name)
1297 return;
1299 if (!file->is_renamed) {
1300 struct strbuf buf = STRBUF_INIT;
1301 if (quote_c_style(file->name, &buf, NULL, 0)) {
1302 pname = strbuf_detach(&buf, NULL);
1303 } else {
1304 pname = file->name;
1305 strbuf_release(&buf);
1306 }
1307 } else {
1308 pname = pprint_rename(file->from_name, file->name);
1309 }
1310 file->print_name = pname;
1311 }
1313 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1314 {
1315 int i, len, add, del, adds = 0, dels = 0;
1316 uintmax_t max_change = 0, max_len = 0;
1317 int total_files = data->nr;
1318 int width, name_width;
1319 const char *reset, *add_c, *del_c;
1320 const char *line_prefix = "";
1321 struct strbuf *msg = NULL;
1323 if (data->nr == 0)
1324 return;
1326 if (options->output_prefix) {
1327 msg = options->output_prefix(options, options->output_prefix_data);
1328 line_prefix = msg->buf;
1329 }
1331 width = options->stat_width ? options->stat_width : 80;
1332 name_width = options->stat_name_width ? options->stat_name_width : 50;
1334 /* Sanity: give at least 5 columns to the graph,
1335 * but leave at least 10 columns for the name.
1336 */
1337 if (width < 25)
1338 width = 25;
1339 if (name_width < 10)
1340 name_width = 10;
1341 else if (width < name_width + 15)
1342 name_width = width - 15;
1344 /* Find the longest filename and max number of changes */
1345 reset = diff_get_color_opt(options, DIFF_RESET);
1346 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1347 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1349 for (i = 0; i < data->nr; i++) {
1350 struct diffstat_file *file = data->files[i];
1351 uintmax_t change = file->added + file->deleted;
1352 fill_print_name(file);
1353 len = strlen(file->print_name);
1354 if (max_len < len)
1355 max_len = len;
1357 if (file->is_binary || file->is_unmerged)
1358 continue;
1359 if (max_change < change)
1360 max_change = change;
1361 }
1363 /* Compute the width of the graph part;
1364 * 10 is for one blank at the beginning of the line plus
1365 * " | count " between the name and the graph.
1366 *
1367 * From here on, name_width is the width of the name area,
1368 * and width is the width of the graph area.
1369 */
1370 name_width = (name_width < max_len) ? name_width : max_len;
1371 if (width < (name_width + 10) + max_change)
1372 width = width - (name_width + 10);
1373 else
1374 width = max_change;
1376 for (i = 0; i < data->nr; i++) {
1377 const char *prefix = "";
1378 char *name = data->files[i]->print_name;
1379 uintmax_t added = data->files[i]->added;
1380 uintmax_t deleted = data->files[i]->deleted;
1381 int name_len;
1383 /*
1384 * "scale" the filename
1385 */
1386 len = name_width;
1387 name_len = strlen(name);
1388 if (name_width < name_len) {
1389 char *slash;
1390 prefix = "...";
1391 len -= 3;
1392 name += name_len - len;
1393 slash = strchr(name, '/');
1394 if (slash)
1395 name = slash;
1396 }
1398 if (data->files[i]->is_binary) {
1399 fprintf(options->file, "%s", line_prefix);
1400 show_name(options->file, prefix, name, len);
1401 fprintf(options->file, " Bin ");
1402 fprintf(options->file, "%s%"PRIuMAX"%s",
1403 del_c, deleted, reset);
1404 fprintf(options->file, " -> ");
1405 fprintf(options->file, "%s%"PRIuMAX"%s",
1406 add_c, added, reset);
1407 fprintf(options->file, " bytes");
1408 fprintf(options->file, "\n");
1409 continue;
1410 }
1411 else if (data->files[i]->is_unmerged) {
1412 fprintf(options->file, "%s", line_prefix);
1413 show_name(options->file, prefix, name, len);
1414 fprintf(options->file, " Unmerged\n");
1415 continue;
1416 }
1417 else if (!data->files[i]->is_renamed &&
1418 (added + deleted == 0)) {
1419 total_files--;
1420 continue;
1421 }
1423 /*
1424 * scale the add/delete
1425 */
1426 add = added;
1427 del = deleted;
1428 adds += add;
1429 dels += del;
1431 if (width <= max_change) {
1432 add = scale_linear(add, width, max_change);
1433 del = scale_linear(del, width, max_change);
1434 }
1435 fprintf(options->file, "%s", line_prefix);
1436 show_name(options->file, prefix, name, len);
1437 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1438 added + deleted ? " " : "");
1439 show_graph(options->file, '+', add, add_c, reset);
1440 show_graph(options->file, '-', del, del_c, reset);
1441 fprintf(options->file, "\n");
1442 }
1443 fprintf(options->file, "%s", line_prefix);
1444 fprintf(options->file,
1445 " %d files changed, %d insertions(+), %d deletions(-)\n",
1446 total_files, adds, dels);
1447 }
1449 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1450 {
1451 int i, adds = 0, dels = 0, total_files = data->nr;
1453 if (data->nr == 0)
1454 return;
1456 for (i = 0; i < data->nr; i++) {
1457 if (!data->files[i]->is_binary &&
1458 !data->files[i]->is_unmerged) {
1459 int added = data->files[i]->added;
1460 int deleted= data->files[i]->deleted;
1461 if (!data->files[i]->is_renamed &&
1462 (added + deleted == 0)) {
1463 total_files--;
1464 } else {
1465 adds += added;
1466 dels += deleted;
1467 }
1468 }
1469 }
1470 if (options->output_prefix) {
1471 struct strbuf *msg = NULL;
1472 msg = options->output_prefix(options,
1473 options->output_prefix_data);
1474 fprintf(options->file, "%s", msg->buf);
1475 }
1476 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1477 total_files, adds, dels);
1478 }
1480 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1481 {
1482 int i;
1484 if (data->nr == 0)
1485 return;
1487 for (i = 0; i < data->nr; i++) {
1488 struct diffstat_file *file = data->files[i];
1490 if (options->output_prefix) {
1491 struct strbuf *msg = NULL;
1492 msg = options->output_prefix(options,
1493 options->output_prefix_data);
1494 fprintf(options->file, "%s", msg->buf);
1495 }
1497 if (file->is_binary)
1498 fprintf(options->file, "-\t-\t");
1499 else
1500 fprintf(options->file,
1501 "%"PRIuMAX"\t%"PRIuMAX"\t",
1502 file->added, file->deleted);
1503 if (options->line_termination) {
1504 fill_print_name(file);
1505 if (!file->is_renamed)
1506 write_name_quoted(file->name, options->file,
1507 options->line_termination);
1508 else {
1509 fputs(file->print_name, options->file);
1510 putc(options->line_termination, options->file);
1511 }
1512 } else {
1513 if (file->is_renamed) {
1514 putc('\0', options->file);
1515 write_name_quoted(file->from_name, options->file, '\0');
1516 }
1517 write_name_quoted(file->name, options->file, '\0');
1518 }
1519 }
1520 }
1522 struct dirstat_file {
1523 const char *name;
1524 unsigned long changed;
1525 };
1527 struct dirstat_dir {
1528 struct dirstat_file *files;
1529 int alloc, nr, permille, cumulative;
1530 };
1532 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1533 unsigned long changed, const char *base, int baselen)
1534 {
1535 unsigned long this_dir = 0;
1536 unsigned int sources = 0;
1537 const char *line_prefix = "";
1538 struct strbuf *msg = NULL;
1540 if (opt->output_prefix) {
1541 msg = opt->output_prefix(opt, opt->output_prefix_data);
1542 line_prefix = msg->buf;
1543 }
1545 while (dir->nr) {
1546 struct dirstat_file *f = dir->files;
1547 int namelen = strlen(f->name);
1548 unsigned long this;
1549 char *slash;
1551 if (namelen < baselen)
1552 break;
1553 if (memcmp(f->name, base, baselen))
1554 break;
1555 slash = strchr(f->name + baselen, '/');
1556 if (slash) {
1557 int newbaselen = slash + 1 - f->name;
1558 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1559 sources++;
1560 } else {
1561 this = f->changed;
1562 dir->files++;
1563 dir->nr--;
1564 sources += 2;
1565 }
1566 this_dir += this;
1567 }
1569 /*
1570 * We don't report dirstat's for
1571 * - the top level
1572 * - or cases where everything came from a single directory
1573 * under this directory (sources == 1).
1574 */
1575 if (baselen && sources != 1) {
1576 if (this_dir) {
1577 int permille = this_dir * 1000 / changed;
1578 if (permille >= dir->permille) {
1579 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1580 permille / 10, permille % 10, baselen, base);
1581 if (!dir->cumulative)
1582 return 0;
1583 }
1584 }
1585 }
1586 return this_dir;
1587 }
1589 static int dirstat_compare(const void *_a, const void *_b)
1590 {
1591 const struct dirstat_file *a = _a;
1592 const struct dirstat_file *b = _b;
1593 return strcmp(a->name, b->name);
1594 }
1596 static void show_dirstat(struct diff_options *options)
1597 {
1598 int i;
1599 unsigned long changed;
1600 struct dirstat_dir dir;
1601 struct diff_queue_struct *q = &diff_queued_diff;
1603 dir.files = NULL;
1604 dir.alloc = 0;
1605 dir.nr = 0;
1606 dir.permille = options->dirstat_permille;
1607 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1609 changed = 0;
1610 for (i = 0; i < q->nr; i++) {
1611 struct diff_filepair *p = q->queue[i];
1612 const char *name;
1613 unsigned long copied, added, damage;
1614 int content_changed;
1616 name = p->two->path ? p->two->path : p->one->path;
1618 if (p->one->sha1_valid && p->two->sha1_valid)
1619 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1620 else
1621 content_changed = 1;
1623 if (!content_changed) {
1624 /*
1625 * The SHA1 has not changed, so pre-/post-content is
1626 * identical. We can therefore skip looking at the
1627 * file contents altogether.
1628 */
1629 damage = 0;
1630 goto found_damage;
1631 }
1633 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1634 /*
1635 * In --dirstat-by-file mode, we don't really need to
1636 * look at the actual file contents at all.
1637 * The fact that the SHA1 changed is enough for us to
1638 * add this file to the list of results
1639 * (with each file contributing equal damage).
1640 */
1641 damage = 1;
1642 goto found_damage;
1643 }
1645 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1646 diff_populate_filespec(p->one, 0);
1647 diff_populate_filespec(p->two, 0);
1648 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1649 &copied, &added);
1650 diff_free_filespec_data(p->one);
1651 diff_free_filespec_data(p->two);
1652 } else if (DIFF_FILE_VALID(p->one)) {
1653 diff_populate_filespec(p->one, 1);
1654 copied = added = 0;
1655 diff_free_filespec_data(p->one);
1656 } else if (DIFF_FILE_VALID(p->two)) {
1657 diff_populate_filespec(p->two, 1);
1658 copied = 0;
1659 added = p->two->size;
1660 diff_free_filespec_data(p->two);
1661 } else
1662 continue;
1664 /*
1665 * Original minus copied is the removed material,
1666 * added is the new material. They are both damages
1667 * made to the preimage.
1668 * If the resulting damage is zero, we know that
1669 * diffcore_count_changes() considers the two entries to
1670 * be identical, but since content_changed is true, we
1671 * know that there must have been _some_ kind of change,
1672 * so we force all entries to have damage > 0.
1673 */
1674 damage = (p->one->size - copied) + added;
1675 if (!damage)
1676 damage = 1;
1678 found_damage:
1679 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1680 dir.files[dir.nr].name = name;
1681 dir.files[dir.nr].changed = damage;
1682 changed += damage;
1683 dir.nr++;
1684 }
1686 /* This can happen even with many files, if everything was renames */
1687 if (!changed)
1688 return;
1690 /* Show all directories with more than x% of the changes */
1691 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1692 gather_dirstat(options, &dir, changed, "", 0);
1693 }
1695 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1696 {
1697 int i;
1698 unsigned long changed;
1699 struct dirstat_dir dir;
1701 if (data->nr == 0)
1702 return;
1704 dir.files = NULL;
1705 dir.alloc = 0;
1706 dir.nr = 0;
1707 dir.permille = options->dirstat_permille;
1708 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1710 changed = 0;
1711 for (i = 0; i < data->nr; i++) {
1712 struct diffstat_file *file = data->files[i];
1713 unsigned long damage = file->added + file->deleted;
1714 if (file->is_binary)
1715 /*
1716 * binary files counts bytes, not lines. Must find some
1717 * way to normalize binary bytes vs. textual lines.
1718 * The following heuristic assumes that there are 64
1719 * bytes per "line".
1720 * This is stupid and ugly, but very cheap...
1721 */
1722 damage = (damage + 63) / 64;
1723 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1724 dir.files[dir.nr].name = file->name;
1725 dir.files[dir.nr].changed = damage;
1726 changed += damage;
1727 dir.nr++;
1728 }
1730 /* This can happen even with many files, if everything was renames */
1731 if (!changed)
1732 return;
1734 /* Show all directories with more than x% of the changes */
1735 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1736 gather_dirstat(options, &dir, changed, "", 0);
1737 }
1739 static void free_diffstat_info(struct diffstat_t *diffstat)
1740 {
1741 int i;
1742 for (i = 0; i < diffstat->nr; i++) {
1743 struct diffstat_file *f = diffstat->files[i];
1744 if (f->name != f->print_name)
1745 free(f->print_name);
1746 free(f->name);
1747 free(f->from_name);
1748 free(f);
1749 }
1750 free(diffstat->files);
1751 }
1753 struct checkdiff_t {
1754 const char *filename;
1755 int lineno;
1756 int conflict_marker_size;
1757 struct diff_options *o;
1758 unsigned ws_rule;
1759 unsigned status;
1760 };
1762 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1763 {
1764 char firstchar;
1765 int cnt;
1767 if (len < marker_size + 1)
1768 return 0;
1769 firstchar = line[0];
1770 switch (firstchar) {
1771 case '=': case '>': case '<': case '|':
1772 break;
1773 default:
1774 return 0;
1775 }
1776 for (cnt = 1; cnt < marker_size; cnt++)
1777 if (line[cnt] != firstchar)
1778 return 0;
1779 /* line[1] thru line[marker_size-1] are same as firstchar */
1780 if (len < marker_size + 1 || !isspace(line[marker_size]))
1781 return 0;
1782 return 1;
1783 }
1785 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1786 {
1787 struct checkdiff_t *data = priv;
1788 int marker_size = data->conflict_marker_size;
1789 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
1790 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
1791 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
1792 char *err;
1793 char *line_prefix = "";
1794 struct strbuf *msgbuf;
1796 assert(data->o);
1797 if (data->o->output_prefix) {
1798 msgbuf = data->o->output_prefix(data->o,
1799 data->o->output_prefix_data);
1800 line_prefix = msgbuf->buf;
1801 }
1803 if (line[0] == '+') {
1804 unsigned bad;
1805 data->lineno++;
1806 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1807 data->status |= 1;
1808 fprintf(data->o->file,
1809 "%s%s:%d: leftover conflict marker\n",
1810 line_prefix, data->filename, data->lineno);
1811 }
1812 bad = ws_check(line + 1, len - 1, data->ws_rule);
1813 if (!bad)
1814 return;
1815 data->status |= bad;
1816 err = whitespace_error_string(bad);
1817 fprintf(data->o->file, "%s%s:%d: %s.\n",
1818 line_prefix, data->filename, data->lineno, err);
1819 free(err);
1820 emit_line(data->o, set, reset, line, 1);
1821 ws_check_emit(line + 1, len - 1, data->ws_rule,
1822 data->o->file, set, reset, ws);
1823 } else if (line[0] == ' ') {
1824 data->lineno++;
1825 } else if (line[0] == '@') {
1826 char *plus = strchr(line, '+');
1827 if (plus)
1828 data->lineno = strtol(plus, NULL, 10) - 1;
1829 else
1830 die("invalid diff");
1831 }
1832 }
1834 static unsigned char *deflate_it(char *data,
1835 unsigned long size,
1836 unsigned long *result_size)
1837 {
1838 int bound;
1839 unsigned char *deflated;
1840 git_zstream stream;
1842 memset(&stream, 0, sizeof(stream));
1843 git_deflate_init(&stream, zlib_compression_level);
1844 bound = git_deflate_bound(&stream, size);
1845 deflated = xmalloc(bound);
1846 stream.next_out = deflated;
1847 stream.avail_out = bound;
1849 stream.next_in = (unsigned char *)data;
1850 stream.avail_in = size;
1851 while (git_deflate(&stream, Z_FINISH) == Z_OK)
1852 ; /* nothing */
1853 git_deflate_end(&stream);
1854 *result_size = stream.total_out;
1855 return deflated;
1856 }
1858 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1859 {
1860 void *cp;
1861 void *delta;
1862 void *deflated;
1863 void *data;
1864 unsigned long orig_size;
1865 unsigned long delta_size;
1866 unsigned long deflate_size;
1867 unsigned long data_size;
1869 /* We could do deflated delta, or we could do just deflated two,
1870 * whichever is smaller.
1871 */
1872 delta = NULL;
1873 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1874 if (one->size && two->size) {
1875 delta = diff_delta(one->ptr, one->size,
1876 two->ptr, two->size,
1877 &delta_size, deflate_size);
1878 if (delta) {
1879 void *to_free = delta;
1880 orig_size = delta_size;
1881 delta = deflate_it(delta, delta_size, &delta_size);
1882 free(to_free);
1883 }
1884 }
1886 if (delta && delta_size < deflate_size) {
1887 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1888 free(deflated);
1889 data = delta;
1890 data_size = delta_size;
1891 }
1892 else {
1893 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1894 free(delta);
1895 data = deflated;
1896 data_size = deflate_size;
1897 }
1899 /* emit data encoded in base85 */
1900 cp = data;
1901 while (data_size) {
1902 int bytes = (52 < data_size) ? 52 : data_size;
1903 char line[70];
1904 data_size -= bytes;
1905 if (bytes <= 26)
1906 line[0] = bytes + 'A' - 1;
1907 else
1908 line[0] = bytes - 26 + 'a' - 1;
1909 encode_85(line + 1, cp, bytes);
1910 cp = (char *) cp + bytes;
1911 fprintf(file, "%s", prefix);
1912 fputs(line, file);
1913 fputc('\n', file);
1914 }
1915 fprintf(file, "%s\n", prefix);
1916 free(data);
1917 }
1919 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1920 {
1921 fprintf(file, "%sGIT binary patch\n", prefix);
1922 emit_binary_diff_body(file, one, two, prefix);
1923 emit_binary_diff_body(file, two, one, prefix);
1924 }
1926 static void diff_filespec_load_driver(struct diff_filespec *one)
1927 {
1928 /* Use already-loaded driver */
1929 if (one->driver)
1930 return;
1932 if (S_ISREG(one->mode))
1933 one->driver = userdiff_find_by_path(one->path);
1935 /* Fallback to default settings */
1936 if (!one->driver)
1937 one->driver = userdiff_find_by_name("default");
1938 }
1940 int diff_filespec_is_binary(struct diff_filespec *one)
1941 {
1942 if (one->is_binary == -1) {
1943 diff_filespec_load_driver(one);
1944 if (one->driver->binary != -1)
1945 one->is_binary = one->driver->binary;
1946 else {
1947 if (!one->data && DIFF_FILE_VALID(one))
1948 diff_populate_filespec(one, 0);
1949 if (one->data)
1950 one->is_binary = buffer_is_binary(one->data,
1951 one->size);
1952 if (one->is_binary == -1)
1953 one->is_binary = 0;
1954 }
1955 }
1956 return one->is_binary;
1957 }
1959 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1960 {
1961 diff_filespec_load_driver(one);
1962 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1963 }
1965 static const char *userdiff_word_regex(struct diff_filespec *one)
1966 {
1967 diff_filespec_load_driver(one);
1968 return one->driver->word_regex;
1969 }
1971 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1972 {
1973 if (!options->a_prefix)
1974 options->a_prefix = a;
1975 if (!options->b_prefix)
1976 options->b_prefix = b;
1977 }
1979 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1980 {
1981 if (!DIFF_FILE_VALID(one))
1982 return NULL;
1984 diff_filespec_load_driver(one);
1985 return userdiff_get_textconv(one->driver);
1986 }
1988 static void builtin_diff(const char *name_a,
1989 const char *name_b,
1990 struct diff_filespec *one,
1991 struct diff_filespec *two,
1992 const char *xfrm_msg,
1993 int must_show_header,
1994 struct diff_options *o,
1995 int complete_rewrite)
1996 {
1997 mmfile_t mf1, mf2;
1998 const char *lbl[2];
1999 char *a_one, *b_two;
2000 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
2001 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2002 const char *a_prefix, *b_prefix;
2003 struct userdiff_driver *textconv_one = NULL;
2004 struct userdiff_driver *textconv_two = NULL;
2005 struct strbuf header = STRBUF_INIT;
2006 struct strbuf *msgbuf;
2007 char *line_prefix = "";
2009 if (o->output_prefix) {
2010 msgbuf = o->output_prefix(o, o->output_prefix_data);
2011 line_prefix = msgbuf->buf;
2012 }
2014 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2015 (!one->mode || S_ISGITLINK(one->mode)) &&
2016 (!two->mode || S_ISGITLINK(two->mode))) {
2017 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2018 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2019 show_submodule_summary(o->file, one ? one->path : two->path,
2020 one->sha1, two->sha1, two->dirty_submodule,
2021 del, add, reset);
2022 return;
2023 }
2025 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2026 textconv_one = get_textconv(one);
2027 textconv_two = get_textconv(two);
2028 }
2030 diff_set_mnemonic_prefix(o, "a/", "b/");
2031 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2032 a_prefix = o->b_prefix;
2033 b_prefix = o->a_prefix;
2034 } else {
2035 a_prefix = o->a_prefix;
2036 b_prefix = o->b_prefix;
2037 }
2039 /* Never use a non-valid filename anywhere if at all possible */
2040 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2041 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2043 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2044 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2045 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2046 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2047 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2048 if (lbl[0][0] == '/') {
2049 /* /dev/null */
2050 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2051 if (xfrm_msg)
2052 strbuf_addstr(&header, xfrm_msg);
2053 must_show_header = 1;
2054 }
2055 else if (lbl[1][0] == '/') {
2056 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2057 if (xfrm_msg)
2058 strbuf_addstr(&header, xfrm_msg);
2059 must_show_header = 1;
2060 }
2061 else {
2062 if (one->mode != two->mode) {
2063 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2064 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2065 must_show_header = 1;
2066 }
2067 if (xfrm_msg)
2068 strbuf_addstr(&header, xfrm_msg);
2070 /*
2071 * we do not run diff between different kind
2072 * of objects.
2073 */
2074 if ((one->mode ^ two->mode) & S_IFMT)
2075 goto free_ab_and_return;
2076 if (complete_rewrite &&
2077 (textconv_one || !diff_filespec_is_binary(one)) &&
2078 (textconv_two || !diff_filespec_is_binary(two))) {
2079 fprintf(o->file, "%s", header.buf);
2080 strbuf_reset(&header);
2081 emit_rewrite_diff(name_a, name_b, one, two,
2082 textconv_one, textconv_two, o);
2083 o->found_changes = 1;
2084 goto free_ab_and_return;
2085 }
2086 }
2088 if (o->irreversible_delete && lbl[1][0] == '/') {
2089 fprintf(o->file, "%s", header.buf);
2090 strbuf_reset(&header);
2091 goto free_ab_and_return;
2092 } else if (!DIFF_OPT_TST(o, TEXT) &&
2093 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2094 (!textconv_two && diff_filespec_is_binary(two)) )) {
2095 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2096 die("unable to read files to diff");
2097 /* Quite common confusing case */
2098 if (mf1.size == mf2.size &&
2099 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2100 if (must_show_header)
2101 fprintf(o->file, "%s", header.buf);
2102 goto free_ab_and_return;
2103 }
2104 fprintf(o->file, "%s", header.buf);
2105 strbuf_reset(&header);
2106 if (DIFF_OPT_TST(o, BINARY))
2107 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2108 else
2109 fprintf(o->file, "%sBinary files %s and %s differ\n",
2110 line_prefix, lbl[0], lbl[1]);
2111 o->found_changes = 1;
2112 } else {
2113 /* Crazy xdl interfaces.. */
2114 const char *diffopts = getenv("GIT_DIFF_OPTS");
2115 xpparam_t xpp;
2116 xdemitconf_t xecfg;
2117 struct emit_callback ecbdata;
2118 const struct userdiff_funcname *pe;
2120 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2121 fprintf(o->file, "%s", header.buf);
2122 strbuf_reset(&header);
2123 }
2125 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2126 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2128 pe = diff_funcname_pattern(one);
2129 if (!pe)
2130 pe = diff_funcname_pattern(two);
2132 memset(&xpp, 0, sizeof(xpp));
2133 memset(&xecfg, 0, sizeof(xecfg));
2134 memset(&ecbdata, 0, sizeof(ecbdata));
2135 ecbdata.label_path = lbl;
2136 ecbdata.color_diff = o->use_color > 0;
2137 ecbdata.found_changesp = &o->found_changes;
2138 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2139 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2140 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2141 ecbdata.opt = o;
2142 ecbdata.header = header.len ? &header : NULL;
2143 xpp.flags = o->xdl_opts;
2144 xecfg.ctxlen = o->context;
2145 xecfg.interhunkctxlen = o->interhunkcontext;
2146 xecfg.flags = XDL_EMIT_FUNCNAMES;
2147 if (pe)
2148 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2149 if (!diffopts)
2150 ;
2151 else if (!prefixcmp(diffopts, "--unified="))
2152 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2153 else if (!prefixcmp(diffopts, "-u"))
2154 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2155 if (o->word_diff) {
2156 int i;
2158 ecbdata.diff_words =
2159 xcalloc(1, sizeof(struct diff_words_data));
2160 ecbdata.diff_words->type = o->word_diff;
2161 ecbdata.diff_words->opt = o;
2162 if (!o->word_regex)
2163 o->word_regex = userdiff_word_regex(one);
2164 if (!o->word_regex)
2165 o->word_regex = userdiff_word_regex(two);
2166 if (!o->word_regex)
2167 o->word_regex = diff_word_regex_cfg;
2168 if (o->word_regex) {
2169 ecbdata.diff_words->word_regex = (regex_t *)
2170 xmalloc(sizeof(regex_t));
2171 if (regcomp(ecbdata.diff_words->word_regex,
2172 o->word_regex,
2173 REG_EXTENDED | REG_NEWLINE))
2174 die ("Invalid regular expression: %s",
2175 o->word_regex);
2176 }
2177 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2178 if (o->word_diff == diff_words_styles[i].type) {
2179 ecbdata.diff_words->style =
2180 &diff_words_styles[i];
2181 break;
2182 }
2183 }
2184 if (o->use_color > 0) {
2185 struct diff_words_style *st = ecbdata.diff_words->style;
2186 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2187 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2188 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2189 }
2190 }
2191 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2192 &xpp, &xecfg);
2193 if (o->word_diff)
2194 free_diff_words_data(&ecbdata);
2195 if (textconv_one)
2196 free(mf1.ptr);
2197 if (textconv_two)
2198 free(mf2.ptr);
2199 xdiff_clear_find_func(&xecfg);
2200 }
2202 free_ab_and_return:
2203 strbuf_release(&header);
2204 diff_free_filespec_data(one);
2205 diff_free_filespec_data(two);
2206 free(a_one);
2207 free(b_two);
2208 return;
2209 }
2211 static void builtin_diffstat(const char *name_a, const char *name_b,
2212 struct diff_filespec *one,
2213 struct diff_filespec *two,
2214 struct diffstat_t *diffstat,
2215 struct diff_options *o,
2216 int complete_rewrite)
2217 {
2218 mmfile_t mf1, mf2;
2219 struct diffstat_file *data;
2221 data = diffstat_add(diffstat, name_a, name_b);
2223 if (!one || !two) {
2224 data->is_unmerged = 1;
2225 return;
2226 }
2228 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2229 data->is_binary = 1;
2230 data->added = diff_filespec_size(two);
2231 data->deleted = diff_filespec_size(one);
2232 }
2234 else if (complete_rewrite) {
2235 diff_populate_filespec(one, 0);
2236 diff_populate_filespec(two, 0);
2237 data->deleted = count_lines(one->data, one->size);
2238 data->added = count_lines(two->data, two->size);
2239 }
2241 else {
2242 /* Crazy xdl interfaces.. */
2243 xpparam_t xpp;
2244 xdemitconf_t xecfg;
2246 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2247 die("unable to read files to diff");
2249 memset(&xpp, 0, sizeof(xpp));
2250 memset(&xecfg, 0, sizeof(xecfg));
2251 xpp.flags = o->xdl_opts;
2252 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2253 &xpp, &xecfg);
2254 }
2256 diff_free_filespec_data(one);
2257 diff_free_filespec_data(two);
2258 }
2260 static void builtin_checkdiff(const char *name_a, const char *name_b,
2261 const char *attr_path,
2262 struct diff_filespec *one,
2263 struct diff_filespec *two,
2264 struct diff_options *o)
2265 {
2266 mmfile_t mf1, mf2;
2267 struct checkdiff_t data;
2269 if (!two)
2270 return;
2272 memset(&data, 0, sizeof(data));
2273 data.filename = name_b ? name_b : name_a;
2274 data.lineno = 0;
2275 data.o = o;
2276 data.ws_rule = whitespace_rule(attr_path);
2277 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2279 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2280 die("unable to read files to diff");
2282 /*
2283 * All the other codepaths check both sides, but not checking
2284 * the "old" side here is deliberate. We are checking the newly
2285 * introduced changes, and as long as the "new" side is text, we
2286 * can and should check what it introduces.
2287 */
2288 if (diff_filespec_is_binary(two))
2289 goto free_and_return;
2290 else {
2291 /* Crazy xdl interfaces.. */
2292 xpparam_t xpp;
2293 xdemitconf_t xecfg;
2295 memset(&xpp, 0, sizeof(xpp));
2296 memset(&xecfg, 0, sizeof(xecfg));
2297 xecfg.ctxlen = 1; /* at least one context line */
2298 xpp.flags = 0;
2299 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2300 &xpp, &xecfg);
2302 if (data.ws_rule & WS_BLANK_AT_EOF) {
2303 struct emit_callback ecbdata;
2304 int blank_at_eof;
2306 ecbdata.ws_rule = data.ws_rule;
2307 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2308 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2310 if (blank_at_eof) {
2311 static char *err;
2312 if (!err)
2313 err = whitespace_error_string(WS_BLANK_AT_EOF);
2314 fprintf(o->file, "%s:%d: %s.\n",
2315 data.filename, blank_at_eof, err);
2316 data.status = 1; /* report errors */
2317 }
2318 }
2319 }
2320 free_and_return:
2321 diff_free_filespec_data(one);
2322 diff_free_filespec_data(two);
2323 if (data.status)
2324 DIFF_OPT_SET(o, CHECK_FAILED);
2325 }
2327 struct diff_filespec *alloc_filespec(const char *path)
2328 {
2329 int namelen = strlen(path);
2330 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2332 memset(spec, 0, sizeof(*spec));
2333 spec->path = (char *)(spec + 1);
2334 memcpy(spec->path, path, namelen+1);
2335 spec->count = 1;
2336 spec->is_binary = -1;
2337 return spec;
2338 }
2340 void free_filespec(struct diff_filespec *spec)
2341 {
2342 if (!--spec->count) {
2343 diff_free_filespec_data(spec);
2344 free(spec);
2345 }
2346 }
2348 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2349 unsigned short mode)
2350 {
2351 if (mode) {
2352 spec->mode = canon_mode(mode);
2353 hashcpy(spec->sha1, sha1);
2354 spec->sha1_valid = !is_null_sha1(sha1);
2355 }
2356 }
2358 /*
2359 * Given a name and sha1 pair, if the index tells us the file in
2360 * the work tree has that object contents, return true, so that
2361 * prepare_temp_file() does not have to inflate and extract.
2362 */
2363 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2364 {
2365 struct cache_entry *ce;
2366 struct stat st;
2367 int pos, len;
2369 /*
2370 * We do not read the cache ourselves here, because the
2371 * benchmark with my previous version that always reads cache
2372 * shows that it makes things worse for diff-tree comparing
2373 * two linux-2.6 kernel trees in an already checked out work
2374 * tree. This is because most diff-tree comparisons deal with
2375 * only a small number of files, while reading the cache is
2376 * expensive for a large project, and its cost outweighs the
2377 * savings we get by not inflating the object to a temporary
2378 * file. Practically, this code only helps when we are used
2379 * by diff-cache --cached, which does read the cache before
2380 * calling us.
2381 */
2382 if (!active_cache)
2383 return 0;
2385 /* We want to avoid the working directory if our caller
2386 * doesn't need the data in a normal file, this system
2387 * is rather slow with its stat/open/mmap/close syscalls,
2388 * and the object is contained in a pack file. The pack
2389 * is probably already open and will be faster to obtain
2390 * the data through than the working directory. Loose
2391 * objects however would tend to be slower as they need
2392 * to be individually opened and inflated.
2393 */
2394 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2395 return 0;
2397 len = strlen(name);
2398 pos = cache_name_pos(name, len);
2399 if (pos < 0)
2400 return 0;
2401 ce = active_cache[pos];
2403 /*
2404 * This is not the sha1 we are looking for, or
2405 * unreusable because it is not a regular file.
2406 */
2407 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2408 return 0;
2410 /*
2411 * If ce is marked as "assume unchanged", there is no
2412 * guarantee that work tree matches what we are looking for.
2413 */
2414 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2415 return 0;
2417 /*
2418 * If ce matches the file in the work tree, we can reuse it.
2419 */
2420 if (ce_uptodate(ce) ||
2421 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2422 return 1;
2424 return 0;
2425 }
2427 static int populate_from_stdin(struct diff_filespec *s)
2428 {
2429 struct strbuf buf = STRBUF_INIT;
2430 size_t size = 0;
2432 if (strbuf_read(&buf, 0, 0) < 0)
2433 return error("error while reading from stdin %s",
2434 strerror(errno));
2436 s->should_munmap = 0;
2437 s->data = strbuf_detach(&buf, &size);
2438 s->size = size;
2439 s->should_free = 1;
2440 return 0;
2441 }
2443 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2444 {
2445 int len;
2446 char *data = xmalloc(100), *dirty = "";
2448 /* Are we looking at the work tree? */
2449 if (s->dirty_submodule)
2450 dirty = "-dirty";
2452 len = snprintf(data, 100,
2453 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2454 s->data = data;
2455 s->size = len;
2456 s->should_free = 1;
2457 if (size_only) {
2458 s->data = NULL;
2459 free(data);
2460 }
2461 return 0;
2462 }
2464 /*
2465 * While doing rename detection and pickaxe operation, we may need to
2466 * grab the data for the blob (or file) for our own in-core comparison.
2467 * diff_filespec has data and size fields for this purpose.
2468 */
2469 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2470 {
2471 int err = 0;
2472 if (!DIFF_FILE_VALID(s))
2473 die("internal error: asking to populate invalid file.");
2474 if (S_ISDIR(s->mode))
2475 return -1;
2477 if (s->data)
2478 return 0;
2480 if (size_only && 0 < s->size)
2481 return 0;
2483 if (S_ISGITLINK(s->mode))
2484 return diff_populate_gitlink(s, size_only);
2486 if (!s->sha1_valid ||
2487 reuse_worktree_file(s->path, s->sha1, 0)) {
2488 struct strbuf buf = STRBUF_INIT;
2489 struct stat st;
2490 int fd;
2492 if (!strcmp(s->path, "-"))
2493 return populate_from_stdin(s);
2495 if (lstat(s->path, &st) < 0) {
2496 if (errno == ENOENT) {
2497 err_empty:
2498 err = -1;
2499 empty:
2500 s->data = (char *)"";
2501 s->size = 0;
2502 return err;
2503 }
2504 }
2505 s->size = xsize_t(st.st_size);
2506 if (!s->size)
2507 goto empty;
2508 if (S_ISLNK(st.st_mode)) {
2509 struct strbuf sb = STRBUF_INIT;
2511 if (strbuf_readlink(&sb, s->path, s->size))
2512 goto err_empty;
2513 s->size = sb.len;
2514 s->data = strbuf_detach(&sb, NULL);
2515 s->should_free = 1;
2516 return 0;
2517 }
2518 if (size_only)
2519 return 0;
2520 fd = open(s->path, O_RDONLY);
2521 if (fd < 0)
2522 goto err_empty;
2523 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2524 close(fd);
2525 s->should_munmap = 1;
2527 /*
2528 * Convert from working tree format to canonical git format
2529 */
2530 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2531 size_t size = 0;
2532 munmap(s->data, s->size);
2533 s->should_munmap = 0;
2534 s->data = strbuf_detach(&buf, &size);
2535 s->size = size;
2536 s->should_free = 1;
2537 }
2538 }
2539 else {
2540 enum object_type type;
2541 if (size_only) {
2542 type = sha1_object_info(s->sha1, &s->size);
2543 if (type < 0)
2544 die("unable to read %s", sha1_to_hex(s->sha1));
2545 } else {
2546 s->data = read_sha1_file(s->sha1, &type, &s->size);
2547 if (!s->data)
2548 die("unable to read %s", sha1_to_hex(s->sha1));
2549 s->should_free = 1;
2550 }
2551 }
2552 return 0;
2553 }
2555 void diff_free_filespec_blob(struct diff_filespec *s)
2556 {
2557 if (s->should_free)
2558 free(s->data);
2559 else if (s->should_munmap)
2560 munmap(s->data, s->size);
2562 if (s->should_free || s->should_munmap) {
2563 s->should_free = s->should_munmap = 0;
2564 s->data = NULL;
2565 }
2566 }
2568 void diff_free_filespec_data(struct diff_filespec *s)
2569 {
2570 diff_free_filespec_blob(s);
2571 free(s->cnt_data);
2572 s->cnt_data = NULL;
2573 }
2575 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2576 void *blob,
2577 unsigned long size,
2578 const unsigned char *sha1,
2579 int mode)
2580 {
2581 int fd;
2582 struct strbuf buf = STRBUF_INIT;
2583 struct strbuf template = STRBUF_INIT;
2584 char *path_dup = xstrdup(path);
2585 const char *base = basename(path_dup);
2587 /* Generate "XXXXXX_basename.ext" */
2588 strbuf_addstr(&template, "XXXXXX_");
2589 strbuf_addstr(&template, base);
2591 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2592 strlen(base) + 1);
2593 if (fd < 0)
2594 die_errno("unable to create temp-file");
2595 if (convert_to_working_tree(path,
2596 (const char *)blob, (size_t)size, &buf)) {
2597 blob = buf.buf;
2598 size = buf.len;
2599 }
2600 if (write_in_full(fd, blob, size) != size)
2601 die_errno("unable to write temp-file");
2602 close(fd);
2603 temp->name = temp->tmp_path;
2604 strcpy(temp->hex, sha1_to_hex(sha1));
2605 temp->hex[40] = 0;
2606 sprintf(temp->mode, "%06o", mode);
2607 strbuf_release(&buf);
2608 strbuf_release(&template);
2609 free(path_dup);
2610 }
2612 static struct diff_tempfile *prepare_temp_file(const char *name,
2613 struct diff_filespec *one)
2614 {
2615 struct diff_tempfile *temp = claim_diff_tempfile();
2617 if (!DIFF_FILE_VALID(one)) {
2618 not_a_valid_file:
2619 /* A '-' entry produces this for file-2, and
2620 * a '+' entry produces this for file-1.
2621 */
2622 temp->name = "/dev/null";
2623 strcpy(temp->hex, ".");
2624 strcpy(temp->mode, ".");
2625 return temp;
2626 }
2628 if (!remove_tempfile_installed) {
2629 atexit(remove_tempfile);
2630 sigchain_push_common(remove_tempfile_on_signal);
2631 remove_tempfile_installed = 1;
2632 }
2634 if (!one->sha1_valid ||
2635 reuse_worktree_file(name, one->sha1, 1)) {
2636 struct stat st;
2637 if (lstat(name, &st) < 0) {
2638 if (errno == ENOENT)
2639 goto not_a_valid_file;
2640 die_errno("stat(%s)", name);
2641 }
2642 if (S_ISLNK(st.st_mode)) {
2643 struct strbuf sb = STRBUF_INIT;
2644 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2645 die_errno("readlink(%s)", name);
2646 prep_temp_blob(name, temp, sb.buf, sb.len,
2647 (one->sha1_valid ?
2648 one->sha1 : null_sha1),
2649 (one->sha1_valid ?
2650 one->mode : S_IFLNK));
2651 strbuf_release(&sb);
2652 }
2653 else {
2654 /* we can borrow from the file in the work tree */
2655 temp->name = name;
2656 if (!one->sha1_valid)
2657 strcpy(temp->hex, sha1_to_hex(null_sha1));
2658 else
2659 strcpy(temp->hex, sha1_to_hex(one->sha1));
2660 /* Even though we may sometimes borrow the
2661 * contents from the work tree, we always want
2662 * one->mode. mode is trustworthy even when
2663 * !(one->sha1_valid), as long as
2664 * DIFF_FILE_VALID(one).
2665 */
2666 sprintf(temp->mode, "%06o", one->mode);
2667 }
2668 return temp;
2669 }
2670 else {
2671 if (diff_populate_filespec(one, 0))
2672 die("cannot read data blob for %s", one->path);
2673 prep_temp_blob(name, temp, one->data, one->size,
2674 one->sha1, one->mode);
2675 }
2676 return temp;
2677 }
2679 /* An external diff command takes:
2680 *
2681 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2682 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2683 *
2684 */
2685 static void run_external_diff(const char *pgm,
2686 const char *name,
2687 const char *other,
2688 struct diff_filespec *one,
2689 struct diff_filespec *two,
2690 const char *xfrm_msg,
2691 int complete_rewrite)
2692 {
2693 const char *spawn_arg[10];
2694 int retval;
2695 const char **arg = &spawn_arg[0];
2697 if (one && two) {
2698 struct diff_tempfile *temp_one, *temp_two;
2699 const char *othername = (other ? other : name);
2700 temp_one = prepare_temp_file(name, one);
2701 temp_two = prepare_temp_file(othername, two);
2702 *arg++ = pgm;
2703 *arg++ = name;
2704 *arg++ = temp_one->name;
2705 *arg++ = temp_one->hex;
2706 *arg++ = temp_one->mode;
2707 *arg++ = temp_two->name;
2708 *arg++ = temp_two->hex;
2709 *arg++ = temp_two->mode;
2710 if (other) {
2711 *arg++ = other;
2712 *arg++ = xfrm_msg;
2713 }
2714 } else {
2715 *arg++ = pgm;
2716 *arg++ = name;
2717 }
2718 *arg = NULL;
2719 fflush(NULL);
2720 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2721 remove_tempfile();
2722 if (retval) {
2723 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2724 exit(1);
2725 }
2726 }
2728 static int similarity_index(struct diff_filepair *p)
2729 {
2730 return p->score * 100 / MAX_SCORE;
2731 }
2733 static void fill_metainfo(struct strbuf *msg,
2734 const char *name,
2735 const char *other,
2736 struct diff_filespec *one,
2737 struct diff_filespec *two,
2738 struct diff_options *o,
2739 struct diff_filepair *p,
2740 int *must_show_header,
2741 int use_color)
2742 {
2743 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2744 const char *reset = diff_get_color(use_color, DIFF_RESET);
2745 struct strbuf *msgbuf;
2746 char *line_prefix = "";
2748 *must_show_header = 1;
2749 if (o->output_prefix) {
2750 msgbuf = o->output_prefix(o, o->output_prefix_data);
2751 line_prefix = msgbuf->buf;
2752 }
2753 strbuf_init(msg, PATH_MAX * 2 + 300);
2754 switch (p->status) {
2755 case DIFF_STATUS_COPIED:
2756 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2757 line_prefix, set, similarity_index(p));
2758 strbuf_addf(msg, "%s\n%s%scopy from ",
2759 reset, line_prefix, set);
2760 quote_c_style(name, msg, NULL, 0);
2761 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2762 quote_c_style(other, msg, NULL, 0);
2763 strbuf_addf(msg, "%s\n", reset);
2764 break;
2765 case DIFF_STATUS_RENAMED:
2766 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2767 line_prefix, set, similarity_index(p));
2768 strbuf_addf(msg, "%s\n%s%srename from ",
2769 reset, line_prefix, set);
2770 quote_c_style(name, msg, NULL, 0);
2771 strbuf_addf(msg, "%s\n%s%srename to ",
2772 reset, line_prefix, set);
2773 quote_c_style(other, msg, NULL, 0);
2774 strbuf_addf(msg, "%s\n", reset);
2775 break;
2776 case DIFF_STATUS_MODIFIED:
2777 if (p->score) {
2778 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2779 line_prefix,
2780 set, similarity_index(p), reset);
2781 break;
2782 }
2783 /* fallthru */
2784 default:
2785 *must_show_header = 0;
2786 }
2787 if (one && two && hashcmp(one->sha1, two->sha1)) {
2788 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2790 if (DIFF_OPT_TST(o, BINARY)) {
2791 mmfile_t mf;
2792 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2793 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2794 abbrev = 40;
2795 }
2796 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2797 find_unique_abbrev(one->sha1, abbrev));
2798 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2799 if (one->mode == two->mode)
2800 strbuf_addf(msg, " %06o", one->mode);
2801 strbuf_addf(msg, "%s\n", reset);
2802 }
2803 }
2805 static void run_diff_cmd(const char *pgm,
2806 const char *name,
2807 const char *other,
2808 const char *attr_path,
2809 struct diff_filespec *one,
2810 struct diff_filespec *two,
2811 struct strbuf *msg,
2812 struct diff_options *o,
2813 struct diff_filepair *p)
2814 {
2815 const char *xfrm_msg = NULL;
2816 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2817 int must_show_header = 0;
2819 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2820 pgm = NULL;
2821 else {
2822 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2823 if (drv && drv->external)
2824 pgm = drv->external;
2825 }
2827 if (msg) {
2828 /*
2829 * don't use colors when the header is intended for an
2830 * external diff driver
2831 */
2832 fill_metainfo(msg, name, other, one, two, o, p,
2833 &must_show_header,
2834 o->use_color > 0 && !pgm);
2835 xfrm_msg = msg->len ? msg->buf : NULL;
2836 }
2838 if (pgm) {
2839 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2840 complete_rewrite);
2841 return;
2842 }
2843 if (one && two)
2844 builtin_diff(name, other ? other : name,
2845 one, two, xfrm_msg, must_show_header,
2846 o, complete_rewrite);
2847 else
2848 fprintf(o->file, "* Unmerged path %s\n", name);
2849 }
2851 static void diff_fill_sha1_info(struct diff_filespec *one)
2852 {
2853 if (DIFF_FILE_VALID(one)) {
2854 if (!one->sha1_valid) {
2855 struct stat st;
2856 if (!strcmp(one->path, "-")) {
2857 hashcpy(one->sha1, null_sha1);
2858 return;
2859 }
2860 if (lstat(one->path, &st) < 0)
2861 die_errno("stat '%s'", one->path);
2862 if (index_path(one->sha1, one->path, &st, 0))
2863 die("cannot hash %s", one->path);
2864 }
2865 }
2866 else
2867 hashclr(one->sha1);
2868 }
2870 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2871 {
2872 /* Strip the prefix but do not molest /dev/null and absolute paths */
2873 if (*namep && **namep != '/') {
2874 *namep += prefix_length;
2875 if (**namep == '/')
2876 ++*namep;
2877 }
2878 if (*otherp && **otherp != '/') {
2879 *otherp += prefix_length;
2880 if (**otherp == '/')
2881 ++*otherp;
2882 }
2883 }
2885 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2886 {
2887 const char *pgm = external_diff();
2888 struct strbuf msg;
2889 struct diff_filespec *one = p->one;
2890 struct diff_filespec *two = p->two;
2891 const char *name;
2892 const char *other;
2893 const char *attr_path;
2895 name = p->one->path;
2896 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2897 attr_path = name;
2898 if (o->prefix_length)
2899 strip_prefix(o->prefix_length, &name, &other);
2901 if (DIFF_PAIR_UNMERGED(p)) {
2902 run_diff_cmd(pgm, name, NULL, attr_path,
2903 NULL, NULL, NULL, o, p);
2904 return;
2905 }
2907 diff_fill_sha1_info(one);
2908 diff_fill_sha1_info(two);
2910 if (!pgm &&
2911 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2912 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2913 /*
2914 * a filepair that changes between file and symlink
2915 * needs to be split into deletion and creation.
2916 */
2917 struct diff_filespec *null = alloc_filespec(two->path);
2918 run_diff_cmd(NULL, name, other, attr_path,
2919 one, null, &msg, o, p);
2920 free(null);
2921 strbuf_release(&msg);
2923 null = alloc_filespec(one->path);
2924 run_diff_cmd(NULL, name, other, attr_path,
2925 null, two, &msg, o, p);
2926 free(null);
2927 }
2928 else
2929 run_diff_cmd(pgm, name, other, attr_path,
2930 one, two, &msg, o, p);
2932 strbuf_release(&msg);
2933 }
2935 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2936 struct diffstat_t *diffstat)
2937 {
2938 const char *name;
2939 const char *other;
2940 int complete_rewrite = 0;
2942 if (DIFF_PAIR_UNMERGED(p)) {
2943 /* unmerged */
2944 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2945 return;
2946 }
2948 name = p->one->path;
2949 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2951 if (o->prefix_length)
2952 strip_prefix(o->prefix_length, &name, &other);
2954 diff_fill_sha1_info(p->one);
2955 diff_fill_sha1_info(p->two);
2957 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2958 complete_rewrite = 1;
2959 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2960 }
2962 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2963 {
2964 const char *name;
2965 const char *other;
2966 const char *attr_path;
2968 if (DIFF_PAIR_UNMERGED(p)) {
2969 /* unmerged */
2970 return;
2971 }
2973 name = p->one->path;
2974 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2975 attr_path = other ? other : name;
2977 if (o->prefix_length)
2978 strip_prefix(o->prefix_length, &name, &other);
2980 diff_fill_sha1_info(p->one);
2981 diff_fill_sha1_info(p->two);
2983 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2984 }
2986 void diff_setup(struct diff_options *options)
2987 {
2988 memcpy(options, &default_diff_options, sizeof(*options));
2990 options->file = stdout;
2992 options->line_termination = '\n';
2993 options->break_opt = -1;
2994 options->rename_limit = -1;
2995 options->dirstat_permille = diff_dirstat_permille_default;
2996 options->context = 3;
2998 options->change = diff_change;
2999 options->add_remove = diff_addremove;
3000 options->use_color = diff_use_color_default;
3001 options->detect_rename = diff_detect_rename_default;
3003 if (diff_no_prefix) {
3004 options->a_prefix = options->b_prefix = "";
3005 } else if (!diff_mnemonic_prefix) {
3006 options->a_prefix = "a/";
3007 options->b_prefix = "b/";
3008 }
3009 }
3011 int diff_setup_done(struct diff_options *options)
3012 {
3013 int count = 0;
3015 if (options->output_format & DIFF_FORMAT_NAME)
3016 count++;
3017 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3018 count++;
3019 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3020 count++;
3021 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3022 count++;
3023 if (count > 1)
3024 die("--name-only, --name-status, --check and -s are mutually exclusive");
3026 /*
3027 * Most of the time we can say "there are changes"
3028 * only by checking if there are changed paths, but
3029 * --ignore-whitespace* options force us to look
3030 * inside contents.
3031 */
3033 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3034 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3035 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3036 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3037 else
3038 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3040 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3041 options->detect_rename = DIFF_DETECT_COPY;
3043 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3044 options->prefix = NULL;
3045 if (options->prefix)
3046 options->prefix_length = strlen(options->prefix);
3047 else
3048 options->prefix_length = 0;
3050 if (options->output_format & (DIFF_FORMAT_NAME |
3051 DIFF_FORMAT_NAME_STATUS |
3052 DIFF_FORMAT_CHECKDIFF |
3053 DIFF_FORMAT_NO_OUTPUT))
3054 options->output_format &= ~(DIFF_FORMAT_RAW |
3055 DIFF_FORMAT_NUMSTAT |
3056 DIFF_FORMAT_DIFFSTAT |
3057 DIFF_FORMAT_SHORTSTAT |
3058 DIFF_FORMAT_DIRSTAT |
3059 DIFF_FORMAT_SUMMARY |
3060 DIFF_FORMAT_PATCH);
3062 /*
3063 * These cases always need recursive; we do not drop caller-supplied
3064 * recursive bits for other formats here.
3065 */
3066 if (options->output_format & (DIFF_FORMAT_PATCH |
3067 DIFF_FORMAT_NUMSTAT |
3068 DIFF_FORMAT_DIFFSTAT |
3069 DIFF_FORMAT_SHORTSTAT |
3070 DIFF_FORMAT_DIRSTAT |
3071 DIFF_FORMAT_SUMMARY |
3072 DIFF_FORMAT_CHECKDIFF))
3073 DIFF_OPT_SET(options, RECURSIVE);
3074 /*
3075 * Also pickaxe would not work very well if you do not say recursive
3076 */
3077 if (options->pickaxe)
3078 DIFF_OPT_SET(options, RECURSIVE);
3079 /*
3080 * When patches are generated, submodules diffed against the work tree
3081 * must be checked for dirtiness too so it can be shown in the output
3082 */
3083 if (options->output_format & DIFF_FORMAT_PATCH)
3084 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3086 if (options->detect_rename && options->rename_limit < 0)
3087 options->rename_limit = diff_rename_limit_default;
3088 if (options->setup & DIFF_SETUP_USE_CACHE) {
3089 if (!active_cache)
3090 /* read-cache does not die even when it fails
3091 * so it is safe for us to do this here. Also
3092 * it does not smudge active_cache or active_nr
3093 * when it fails, so we do not have to worry about
3094 * cleaning it up ourselves either.
3095 */
3096 read_cache();
3097 }
3098 if (options->abbrev <= 0 || 40 < options->abbrev)
3099 options->abbrev = 40; /* full */
3101 /*
3102 * It does not make sense to show the first hit we happened
3103 * to have found. It does not make sense not to return with
3104 * exit code in such a case either.
3105 */
3106 if (DIFF_OPT_TST(options, QUICK)) {
3107 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3108 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3109 }
3111 return 0;
3112 }
3114 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3115 {
3116 char c, *eq;
3117 int len;
3119 if (*arg != '-')
3120 return 0;
3121 c = *++arg;
3122 if (!c)
3123 return 0;
3124 if (c == arg_short) {
3125 c = *++arg;
3126 if (!c)
3127 return 1;
3128 if (val && isdigit(c)) {
3129 char *end;
3130 int n = strtoul(arg, &end, 10);
3131 if (*end)
3132 return 0;
3133 *val = n;
3134 return 1;
3135 }
3136 return 0;
3137 }
3138 if (c != '-')
3139 return 0;
3140 arg++;
3141 eq = strchr(arg, '=');
3142 if (eq)
3143 len = eq - arg;
3144 else
3145 len = strlen(arg);
3146 if (!len || strncmp(arg, arg_long, len))
3147 return 0;
3148 if (eq) {
3149 int n;
3150 char *end;
3151 if (!isdigit(*++eq))
3152 return 0;
3153 n = strtoul(eq, &end, 10);
3154 if (*end)
3155 return 0;
3156 *val = n;
3157 }
3158 return 1;
3159 }
3161 static int diff_scoreopt_parse(const char *opt);
3163 static inline int short_opt(char opt, const char **argv,
3164 const char **optarg)
3165 {
3166 const char *arg = argv[0];
3167 if (arg[0] != '-' || arg[1] != opt)
3168 return 0;
3169 if (arg[2] != '\0') {
3170 *optarg = arg + 2;
3171 return 1;
3172 }
3173 if (!argv[1])
3174 die("Option '%c' requires a value", opt);
3175 *optarg = argv[1];
3176 return 2;
3177 }
3179 int parse_long_opt(const char *opt, const char **argv,
3180 const char **optarg)
3181 {
3182 const char *arg = argv[0];
3183 if (arg[0] != '-' || arg[1] != '-')
3184 return 0;
3185 arg += strlen("--");
3186 if (prefixcmp(arg, opt))
3187 return 0;
3188 arg += strlen(opt);
3189 if (*arg == '=') { /* sticked form: --option=value */
3190 *optarg = arg + 1;
3191 return 1;
3192 }
3193 if (*arg != '\0')
3194 return 0;
3195 /* separate form: --option value */
3196 if (!argv[1])
3197 die("Option '--%s' requires a value", opt);
3198 *optarg = argv[1];
3199 return 2;
3200 }
3202 static int stat_opt(struct diff_options *options, const char **av)
3203 {
3204 const char *arg = av[0];
3205 char *end;
3206 int width = options->stat_width;
3207 int name_width = options->stat_name_width;
3208 int argcount = 1;
3210 arg += strlen("--stat");
3211 end = (char *)arg;
3213 switch (*arg) {
3214 case '-':
3215 if (!prefixcmp(arg, "-width")) {
3216 arg += strlen("-width");
3217 if (*arg == '=')
3218 width = strtoul(arg + 1, &end, 10);
3219 else if (!*arg && !av[1])
3220 die("Option '--stat-width' requires a value");
3221 else if (!*arg) {
3222 width = strtoul(av[1], &end, 10);
3223 argcount = 2;
3224 }
3225 } else if (!prefixcmp(arg, "-name-width")) {
3226 arg += strlen("-name-width");
3227 if (*arg == '=')
3228 name_width = strtoul(arg + 1, &end, 10);
3229 else if (!*arg && !av[1])
3230 die("Option '--stat-name-width' requires a value");
3231 else if (!*arg) {
3232 name_width = strtoul(av[1], &end, 10);
3233 argcount = 2;
3234 }
3235 }
3236 break;
3237 case '=':
3238 width = strtoul(arg+1, &end, 10);
3239 if (*end == ',')
3240 name_width = strtoul(end+1, &end, 10);
3241 }
3243 /* Important! This checks all the error cases! */
3244 if (*end)
3245 return 0;
3246 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3247 options->stat_name_width = name_width;
3248 options->stat_width = width;
3249 return argcount;
3250 }
3252 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3253 {
3254 struct strbuf errmsg = STRBUF_INIT;
3255 if (parse_dirstat_params(options, params, &errmsg))
3256 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3257 errmsg.buf);
3258 strbuf_release(&errmsg);
3259 /*
3260 * The caller knows a dirstat-related option is given from the command
3261 * line; allow it to say "return this_function();"
3262 */
3263 options->output_format |= DIFF_FORMAT_DIRSTAT;
3264 return 1;
3265 }
3267 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3268 {
3269 const char *arg = av[0];
3270 const char *optarg;
3271 int argcount;
3273 /* Output format options */
3274 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3275 options->output_format |= DIFF_FORMAT_PATCH;
3276 else if (opt_arg(arg, 'U', "unified", &options->context))
3277 options->output_format |= DIFF_FORMAT_PATCH;
3278 else if (!strcmp(arg, "--raw"))
3279 options->output_format |= DIFF_FORMAT_RAW;
3280 else if (!strcmp(arg, "--patch-with-raw"))
3281 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3282 else if (!strcmp(arg, "--numstat"))
3283 options->output_format |= DIFF_FORMAT_NUMSTAT;
3284 else if (!strcmp(arg, "--shortstat"))
3285 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3286 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3287 return parse_dirstat_opt(options, "");
3288 else if (!prefixcmp(arg, "-X"))
3289 return parse_dirstat_opt(options, arg + 2);
3290 else if (!prefixcmp(arg, "--dirstat="))
3291 return parse_dirstat_opt(options, arg + 10);
3292 else if (!strcmp(arg, "--cumulative"))
3293 return parse_dirstat_opt(options, "cumulative");
3294 else if (!strcmp(arg, "--dirstat-by-file"))
3295 return parse_dirstat_opt(options, "files");
3296 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3297 parse_dirstat_opt(options, "files");
3298 return parse_dirstat_opt(options, arg + 18);
3299 }
3300 else if (!strcmp(arg, "--check"))
3301 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3302 else if (!strcmp(arg, "--summary"))
3303 options->output_format |= DIFF_FORMAT_SUMMARY;
3304 else if (!strcmp(arg, "--patch-with-stat"))
3305 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3306 else if (!strcmp(arg, "--name-only"))
3307 options->output_format |= DIFF_FORMAT_NAME;
3308 else if (!strcmp(arg, "--name-status"))
3309 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3310 else if (!strcmp(arg, "-s"))
3311 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3312 else if (!prefixcmp(arg, "--stat"))
3313 /* --stat, --stat-width, or --stat-name-width */
3314 return stat_opt(options, av);
3316 /* renames options */
3317 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3318 !strcmp(arg, "--break-rewrites")) {
3319 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3320 return error("invalid argument to -B: %s", arg+2);
3321 }
3322 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3323 !strcmp(arg, "--find-renames")) {
3324 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3325 return error("invalid argument to -M: %s", arg+2);
3326 options->detect_rename = DIFF_DETECT_RENAME;
3327 }
3328 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3329 options->irreversible_delete = 1;
3330 }
3331 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3332 !strcmp(arg, "--find-copies")) {
3333 if (options->detect_rename == DIFF_DETECT_COPY)
3334 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3335 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3336 return error("invalid argument to -C: %s", arg+2);
3337 options->detect_rename = DIFF_DETECT_COPY;
3338 }
3339 else if (!strcmp(arg, "--no-renames"))
3340 options->detect_rename = 0;
3341 else if (!strcmp(arg, "--relative"))
3342 DIFF_OPT_SET(options, RELATIVE_NAME);
3343 else if (!prefixcmp(arg, "--relative=")) {
3344 DIFF_OPT_SET(options, RELATIVE_NAME);
3345 options->prefix = arg + 11;
3346 }
3348 /* xdiff options */
3349 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3350 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3351 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3352 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3353 else if (!strcmp(arg, "--ignore-space-at-eol"))
3354 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3355 else if (!strcmp(arg, "--patience"))
3356 DIFF_XDL_SET(options, PATIENCE_DIFF);
3358 /* flags options */
3359 else if (!strcmp(arg, "--binary")) {
3360 options->output_format |= DIFF_FORMAT_PATCH;
3361 DIFF_OPT_SET(options, BINARY);
3362 }
3363 else if (!strcmp(arg, "--full-index"))
3364 DIFF_OPT_SET(options, FULL_INDEX);
3365 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3366 DIFF_OPT_SET(options, TEXT);
3367 else if (!strcmp(arg, "-R"))
3368 DIFF_OPT_SET(options, REVERSE_DIFF);
3369 else if (!strcmp(arg, "--find-copies-harder"))
3370 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3371 else if (!strcmp(arg, "--follow"))
3372 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3373 else if (!strcmp(arg, "--color"))
3374 options->use_color = 1;
3375 else if (!prefixcmp(arg, "--color=")) {
3376 int value = git_config_colorbool(NULL, arg+8);
3377 if (value == 0)
3378 options->use_color = 0;
3379 else if (value > 0)
3380 options->use_color = 1;
3381 else
3382 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3383 }
3384 else if (!strcmp(arg, "--no-color"))
3385 options->use_color = 0;
3386 else if (!strcmp(arg, "--color-words")) {
3387 options->use_color = 1;
3388 options->word_diff = DIFF_WORDS_COLOR;
3389 }
3390 else if (!prefixcmp(arg, "--color-words=")) {
3391 options->use_color = 1;
3392 options->word_diff = DIFF_WORDS_COLOR;
3393 options->word_regex = arg + 14;
3394 }
3395 else if (!strcmp(arg, "--word-diff")) {
3396 if (options->word_diff == DIFF_WORDS_NONE)
3397 options->word_diff = DIFF_WORDS_PLAIN;
3398 }
3399 else if (!prefixcmp(arg, "--word-diff=")) {
3400 const char *type = arg + 12;
3401 if (!strcmp(type, "plain"))
3402 options->word_diff = DIFF_WORDS_PLAIN;
3403 else if (!strcmp(type, "color")) {
3404 options->use_color = 1;
3405 options->word_diff = DIFF_WORDS_COLOR;
3406 }
3407 else if (!strcmp(type, "porcelain"))
3408 options->word_diff = DIFF_WORDS_PORCELAIN;
3409 else if (!strcmp(type, "none"))
3410 options->word_diff = DIFF_WORDS_NONE;
3411 else
3412 die("bad --word-diff argument: %s", type);
3413 }
3414 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3415 if (options->word_diff == DIFF_WORDS_NONE)
3416 options->word_diff = DIFF_WORDS_PLAIN;
3417 options->word_regex = optarg;
3418 return argcount;
3419 }
3420 else if (!strcmp(arg, "--exit-code"))
3421 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3422 else if (!strcmp(arg, "--quiet"))
3423 DIFF_OPT_SET(options, QUICK);
3424 else if (!strcmp(arg, "--ext-diff"))
3425 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3426 else if (!strcmp(arg, "--no-ext-diff"))
3427 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3428 else if (!strcmp(arg, "--textconv"))
3429 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3430 else if (!strcmp(arg, "--no-textconv"))
3431 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3432 else if (!strcmp(arg, "--ignore-submodules")) {
3433 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3434 handle_ignore_submodules_arg(options, "all");
3435 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3436 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3437 handle_ignore_submodules_arg(options, arg + 20);
3438 } else if (!strcmp(arg, "--submodule"))
3439 DIFF_OPT_SET(options, SUBMODULE_LOG);
3440 else if (!prefixcmp(arg, "--submodule=")) {
3441 if (!strcmp(arg + 12, "log"))
3442 DIFF_OPT_SET(options, SUBMODULE_LOG);
3443 }
3445 /* misc options */
3446 else if (!strcmp(arg, "-z"))
3447 options->line_termination = 0;
3448 else if ((argcount = short_opt('l', av, &optarg))) {
3449 options->rename_limit = strtoul(optarg, NULL, 10);
3450 return argcount;
3451 }
3452 else if ((argcount = short_opt('S', av, &optarg))) {
3453 options->pickaxe = optarg;
3454 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3455 return argcount;
3456 } else if ((argcount = short_opt('G', av, &optarg))) {
3457 options->pickaxe = optarg;
3458 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3459 return argcount;
3460 }
3461 else if (!strcmp(arg, "--pickaxe-all"))
3462 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3463 else if (!strcmp(arg, "--pickaxe-regex"))
3464 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3465 else if ((argcount = short_opt('O', av, &optarg))) {
3466 options->orderfile = optarg;
3467 return argcount;
3468 }
3469 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3470 options->filter = optarg;
3471 return argcount;
3472 }
3473 else if (!strcmp(arg, "--abbrev"))
3474 options->abbrev = DEFAULT_ABBREV;
3475 else if (!prefixcmp(arg, "--abbrev=")) {
3476 options->abbrev = strtoul(arg + 9, NULL, 10);
3477 if (options->abbrev < MINIMUM_ABBREV)
3478 options->abbrev = MINIMUM_ABBREV;
3479 else if (40 < options->abbrev)
3480 options->abbrev = 40;
3481 }
3482 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3483 options->a_prefix = optarg;
3484 return argcount;
3485 }
3486 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3487 options->b_prefix = optarg;
3488 return argcount;
3489 }
3490 else if (!strcmp(arg, "--no-prefix"))
3491 options->a_prefix = options->b_prefix = "";
3492 else if (opt_arg(arg, '\0', "inter-hunk-context",
3493 &options->interhunkcontext))
3494 ;
3495 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3496 options->file = fopen(optarg, "w");
3497 if (!options->file)
3498 die_errno("Could not open '%s'", optarg);
3499 options->close_file = 1;
3500 return argcount;
3501 } else
3502 return 0;
3503 return 1;
3504 }
3506 int parse_rename_score(const char **cp_p)
3507 {
3508 unsigned long num, scale;
3509 int ch, dot;
3510 const char *cp = *cp_p;
3512 num = 0;
3513 scale = 1;
3514 dot = 0;
3515 for (;;) {
3516 ch = *cp;
3517 if ( !dot && ch == '.' ) {
3518 scale = 1;
3519 dot = 1;
3520 } else if ( ch == '%' ) {
3521 scale = dot ? scale*100 : 100;
3522 cp++; /* % is always at the end */
3523 break;
3524 } else if ( ch >= '0' && ch <= '9' ) {
3525 if ( scale < 100000 ) {
3526 scale *= 10;
3527 num = (num*10) + (ch-'0');
3528 }
3529 } else {
3530 break;
3531 }
3532 cp++;
3533 }
3534 *cp_p = cp;
3536 /* user says num divided by scale and we say internally that
3537 * is MAX_SCORE * num / scale.
3538 */
3539 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3540 }
3542 static int diff_scoreopt_parse(const char *opt)
3543 {
3544 int opt1, opt2, cmd;
3546 if (*opt++ != '-')
3547 return -1;
3548 cmd = *opt++;
3549 if (cmd == '-') {
3550 /* convert the long-form arguments into short-form versions */
3551 if (!prefixcmp(opt, "break-rewrites")) {
3552 opt += strlen("break-rewrites");
3553 if (*opt == 0 || *opt++ == '=')
3554 cmd = 'B';
3555 } else if (!prefixcmp(opt, "find-copies")) {
3556 opt += strlen("find-copies");
3557 if (*opt == 0 || *opt++ == '=')
3558 cmd = 'C';
3559 } else if (!prefixcmp(opt, "find-renames")) {
3560 opt += strlen("find-renames");
3561 if (*opt == 0 || *opt++ == '=')
3562 cmd = 'M';
3563 }
3564 }
3565 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3566 return -1; /* that is not a -M, -C nor -B option */
3568 opt1 = parse_rename_score(&opt);
3569 if (cmd != 'B')
3570 opt2 = 0;
3571 else {
3572 if (*opt == 0)
3573 opt2 = 0;
3574 else if (*opt != '/')
3575 return -1; /* we expect -B80/99 or -B80 */
3576 else {
3577 opt++;
3578 opt2 = parse_rename_score(&opt);
3579 }
3580 }
3581 if (*opt != 0)
3582 return -1;
3583 return opt1 | (opt2 << 16);
3584 }
3586 struct diff_queue_struct diff_queued_diff;
3588 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3589 {
3590 if (queue->alloc <= queue->nr) {
3591 queue->alloc = alloc_nr(queue->alloc);
3592 queue->queue = xrealloc(queue->queue,
3593 sizeof(dp) * queue->alloc);
3594 }
3595 queue->queue[queue->nr++] = dp;
3596 }
3598 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3599 struct diff_filespec *one,
3600 struct diff_filespec *two)
3601 {
3602 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3603 dp->one = one;
3604 dp->two = two;
3605 if (queue)
3606 diff_q(queue, dp);
3607 return dp;
3608 }
3610 void diff_free_filepair(struct diff_filepair *p)
3611 {
3612 free_filespec(p->one);
3613 free_filespec(p->two);
3614 free(p);
3615 }
3617 /* This is different from find_unique_abbrev() in that
3618 * it stuffs the result with dots for alignment.
3619 */
3620 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3621 {
3622 int abblen;
3623 const char *abbrev;
3624 if (len == 40)
3625 return sha1_to_hex(sha1);
3627 abbrev = find_unique_abbrev(sha1, len);
3628 abblen = strlen(abbrev);
3629 if (abblen < 37) {
3630 static char hex[41];
3631 if (len < abblen && abblen <= len + 2)
3632 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3633 else
3634 sprintf(hex, "%s...", abbrev);
3635 return hex;
3636 }
3637 return sha1_to_hex(sha1);
3638 }
3640 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3641 {
3642 int line_termination = opt->line_termination;
3643 int inter_name_termination = line_termination ? '\t' : '\0';
3644 if (opt->output_prefix) {
3645 struct strbuf *msg = NULL;
3646 msg = opt->output_prefix(opt, opt->output_prefix_data);
3647 fprintf(opt->file, "%s", msg->buf);
3648 }
3650 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3651 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3652 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3653 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3654 }
3655 if (p->score) {
3656 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3657 inter_name_termination);
3658 } else {
3659 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3660 }
3662 if (p->status == DIFF_STATUS_COPIED ||
3663 p->status == DIFF_STATUS_RENAMED) {
3664 const char *name_a, *name_b;
3665 name_a = p->one->path;
3666 name_b = p->two->path;
3667 strip_prefix(opt->prefix_length, &name_a, &name_b);
3668 write_name_quoted(name_a, opt->file, inter_name_termination);
3669 write_name_quoted(name_b, opt->file, line_termination);
3670 } else {
3671 const char *name_a, *name_b;
3672 name_a = p->one->mode ? p->one->path : p->two->path;
3673 name_b = NULL;
3674 strip_prefix(opt->prefix_length, &name_a, &name_b);
3675 write_name_quoted(name_a, opt->file, line_termination);
3676 }
3677 }
3679 int diff_unmodified_pair(struct diff_filepair *p)
3680 {
3681 /* This function is written stricter than necessary to support
3682 * the currently implemented transformers, but the idea is to
3683 * let transformers to produce diff_filepairs any way they want,
3684 * and filter and clean them up here before producing the output.
3685 */
3686 struct diff_filespec *one = p->one, *two = p->two;
3688 if (DIFF_PAIR_UNMERGED(p))
3689 return 0; /* unmerged is interesting */
3691 /* deletion, addition, mode or type change
3692 * and rename are all interesting.
3693 */
3694 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3695 DIFF_PAIR_MODE_CHANGED(p) ||
3696 strcmp(one->path, two->path))
3697 return 0;
3699 /* both are valid and point at the same path. that is, we are
3700 * dealing with a change.
3701 */
3702 if (one->sha1_valid && two->sha1_valid &&
3703 !hashcmp(one->sha1, two->sha1) &&
3704 !one->dirty_submodule && !two->dirty_submodule)
3705 return 1; /* no change */
3706 if (!one->sha1_valid && !two->sha1_valid)
3707 return 1; /* both look at the same file on the filesystem. */
3708 return 0;
3709 }
3711 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3712 {
3713 if (diff_unmodified_pair(p))
3714 return;
3716 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3717 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3718 return; /* no tree diffs in patch format */
3720 run_diff(p, o);
3721 }
3723 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3724 struct diffstat_t *diffstat)
3725 {
3726 if (diff_unmodified_pair(p))
3727 return;
3729 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3730 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3731 return; /* no useful stat for tree diffs */
3733 run_diffstat(p, o, diffstat);
3734 }
3736 static void diff_flush_checkdiff(struct diff_filepair *p,
3737 struct diff_options *o)
3738 {
3739 if (diff_unmodified_pair(p))
3740 return;
3742 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3743 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3744 return; /* nothing to check in tree diffs */
3746 run_checkdiff(p, o);
3747 }
3749 int diff_queue_is_empty(void)
3750 {
3751 struct diff_queue_struct *q = &diff_queued_diff;
3752 int i;
3753 for (i = 0; i < q->nr; i++)
3754 if (!diff_unmodified_pair(q->queue[i]))
3755 return 0;
3756 return 1;
3757 }
3759 #if DIFF_DEBUG
3760 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3761 {
3762 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3763 x, one ? one : "",
3764 s->path,
3765 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3766 s->mode,
3767 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3768 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3769 x, one ? one : "",
3770 s->size, s->xfrm_flags);
3771 }
3773 void diff_debug_filepair(const struct diff_filepair *p, int i)
3774 {
3775 diff_debug_filespec(p->one, i, "one");
3776 diff_debug_filespec(p->two, i, "two");
3777 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3778 p->score, p->status ? p->status : '?',
3779 p->one->rename_used, p->broken_pair);
3780 }
3782 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3783 {
3784 int i;
3785 if (msg)
3786 fprintf(stderr, "%s\n", msg);
3787 fprintf(stderr, "q->nr = %d\n", q->nr);
3788 for (i = 0; i < q->nr; i++) {
3789 struct diff_filepair *p = q->queue[i];
3790 diff_debug_filepair(p, i);
3791 }
3792 }
3793 #endif
3795 static void diff_resolve_rename_copy(void)
3796 {
3797 int i;
3798 struct diff_filepair *p;
3799 struct diff_queue_struct *q = &diff_queued_diff;
3801 diff_debug_queue("resolve-rename-copy", q);
3803 for (i = 0; i < q->nr; i++) {
3804 p = q->queue[i];
3805 p->status = 0; /* undecided */
3806 if (DIFF_PAIR_UNMERGED(p))
3807 p->status = DIFF_STATUS_UNMERGED;
3808 else if (!DIFF_FILE_VALID(p->one))
3809 p->status = DIFF_STATUS_ADDED;
3810 else if (!DIFF_FILE_VALID(p->two))
3811 p->status = DIFF_STATUS_DELETED;
3812 else if (DIFF_PAIR_TYPE_CHANGED(p))
3813 p->status = DIFF_STATUS_TYPE_CHANGED;
3815 /* from this point on, we are dealing with a pair
3816 * whose both sides are valid and of the same type, i.e.
3817 * either in-place edit or rename/copy edit.
3818 */
3819 else if (DIFF_PAIR_RENAME(p)) {
3820 /*
3821 * A rename might have re-connected a broken
3822 * pair up, causing the pathnames to be the
3823 * same again. If so, that's not a rename at
3824 * all, just a modification..
3825 *
3826 * Otherwise, see if this source was used for
3827 * multiple renames, in which case we decrement
3828 * the count, and call it a copy.
3829 */
3830 if (!strcmp(p->one->path, p->two->path))
3831 p->status = DIFF_STATUS_MODIFIED;
3832 else if (--p->one->rename_used > 0)
3833 p->status = DIFF_STATUS_COPIED;
3834 else
3835 p->status = DIFF_STATUS_RENAMED;
3836 }
3837 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3838 p->one->mode != p->two->mode ||
3839 p->one->dirty_submodule ||
3840 p->two->dirty_submodule ||
3841 is_null_sha1(p->one->sha1))
3842 p->status = DIFF_STATUS_MODIFIED;
3843 else {
3844 /* This is a "no-change" entry and should not
3845 * happen anymore, but prepare for broken callers.
3846 */
3847 error("feeding unmodified %s to diffcore",
3848 p->one->path);
3849 p->status = DIFF_STATUS_UNKNOWN;
3850 }
3851 }
3852 diff_debug_queue("resolve-rename-copy done", q);
3853 }
3855 static int check_pair_status(struct diff_filepair *p)
3856 {
3857 switch (p->status) {
3858 case DIFF_STATUS_UNKNOWN:
3859 return 0;
3860 case 0:
3861 die("internal error in diff-resolve-rename-copy");
3862 default:
3863 return 1;
3864 }
3865 }
3867 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3868 {
3869 int fmt = opt->output_format;
3871 if (fmt & DIFF_FORMAT_CHECKDIFF)
3872 diff_flush_checkdiff(p, opt);
3873 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3874 diff_flush_raw(p, opt);
3875 else if (fmt & DIFF_FORMAT_NAME) {
3876 const char *name_a, *name_b;
3877 name_a = p->two->path;
3878 name_b = NULL;
3879 strip_prefix(opt->prefix_length, &name_a, &name_b);
3880 write_name_quoted(name_a, opt->file, opt->line_termination);
3881 }
3882 }
3884 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3885 {
3886 if (fs->mode)
3887 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3888 else
3889 fprintf(file, " %s ", newdelete);
3890 write_name_quoted(fs->path, file, '\n');
3891 }
3894 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3895 const char *line_prefix)
3896 {
3897 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3898 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3899 p->two->mode, show_name ? ' ' : '\n');
3900 if (show_name) {
3901 write_name_quoted(p->two->path, file, '\n');
3902 }
3903 }
3904 }
3906 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3907 const char *line_prefix)
3908 {
3909 char *names = pprint_rename(p->one->path, p->two->path);
3911 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3912 free(names);
3913 show_mode_change(file, p, 0, line_prefix);
3914 }
3916 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3917 {
3918 FILE *file = opt->file;
3919 char *line_prefix = "";
3921 if (opt->output_prefix) {
3922 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3923 line_prefix = buf->buf;
3924 }
3926 switch(p->status) {
3927 case DIFF_STATUS_DELETED:
3928 fputs(line_prefix, file);
3929 show_file_mode_name(file, "delete", p->one);
3930 break;
3931 case DIFF_STATUS_ADDED:
3932 fputs(line_prefix, file);
3933 show_file_mode_name(file, "create", p->two);
3934 break;
3935 case DIFF_STATUS_COPIED:
3936 fputs(line_prefix, file);
3937 show_rename_copy(file, "copy", p, line_prefix);
3938 break;
3939 case DIFF_STATUS_RENAMED:
3940 fputs(line_prefix, file);
3941 show_rename_copy(file, "rename", p, line_prefix);
3942 break;
3943 default:
3944 if (p->score) {
3945 fprintf(file, "%s rewrite ", line_prefix);
3946 write_name_quoted(p->two->path, file, ' ');
3947 fprintf(file, "(%d%%)\n", similarity_index(p));
3948 }
3949 show_mode_change(file, p, !p->score, line_prefix);
3950 break;
3951 }
3952 }
3954 struct patch_id_t {
3955 git_SHA_CTX *ctx;
3956 int patchlen;
3957 };
3959 static int remove_space(char *line, int len)
3960 {
3961 int i;
3962 char *dst = line;
3963 unsigned char c;
3965 for (i = 0; i < len; i++)
3966 if (!isspace((c = line[i])))
3967 *dst++ = c;
3969 return dst - line;
3970 }
3972 static void patch_id_consume(void *priv, char *line, unsigned long len)
3973 {
3974 struct patch_id_t *data = priv;
3975 int new_len;
3977 /* Ignore line numbers when computing the SHA1 of the patch */
3978 if (!prefixcmp(line, "@@ -"))
3979 return;
3981 new_len = remove_space(line, len);
3983 git_SHA1_Update(data->ctx, line, new_len);
3984 data->patchlen += new_len;
3985 }
3987 /* returns 0 upon success, and writes result into sha1 */
3988 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3989 {
3990 struct diff_queue_struct *q = &diff_queued_diff;
3991 int i;
3992 git_SHA_CTX ctx;
3993 struct patch_id_t data;
3994 char buffer[PATH_MAX * 4 + 20];
3996 git_SHA1_Init(&ctx);
3997 memset(&data, 0, sizeof(struct patch_id_t));
3998 data.ctx = &ctx;
4000 for (i = 0; i < q->nr; i++) {
4001 xpparam_t xpp;
4002 xdemitconf_t xecfg;
4003 mmfile_t mf1, mf2;
4004 struct diff_filepair *p = q->queue[i];
4005 int len1, len2;
4007 memset(&xpp, 0, sizeof(xpp));
4008 memset(&xecfg, 0, sizeof(xecfg));
4009 if (p->status == 0)
4010 return error("internal diff status error");
4011 if (p->status == DIFF_STATUS_UNKNOWN)
4012 continue;
4013 if (diff_unmodified_pair(p))
4014 continue;
4015 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4016 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4017 continue;
4018 if (DIFF_PAIR_UNMERGED(p))
4019 continue;
4021 diff_fill_sha1_info(p->one);
4022 diff_fill_sha1_info(p->two);
4023 if (fill_mmfile(&mf1, p->one) < 0 ||
4024 fill_mmfile(&mf2, p->two) < 0)
4025 return error("unable to read files to diff");
4027 len1 = remove_space(p->one->path, strlen(p->one->path));
4028 len2 = remove_space(p->two->path, strlen(p->two->path));
4029 if (p->one->mode == 0)
4030 len1 = snprintf(buffer, sizeof(buffer),
4031 "diff--gita/%.*sb/%.*s"
4032 "newfilemode%06o"
4033 "---/dev/null"
4034 "+++b/%.*s",
4035 len1, p->one->path,
4036 len2, p->two->path,
4037 p->two->mode,
4038 len2, p->two->path);
4039 else if (p->two->mode == 0)
4040 len1 = snprintf(buffer, sizeof(buffer),
4041 "diff--gita/%.*sb/%.*s"
4042 "deletedfilemode%06o"
4043 "---a/%.*s"
4044 "+++/dev/null",
4045 len1, p->one->path,
4046 len2, p->two->path,
4047 p->one->mode,
4048 len1, p->one->path);
4049 else
4050 len1 = snprintf(buffer, sizeof(buffer),
4051 "diff--gita/%.*sb/%.*s"
4052 "---a/%.*s"
4053 "+++b/%.*s",
4054 len1, p->one->path,
4055 len2, p->two->path,
4056 len1, p->one->path,
4057 len2, p->two->path);
4058 git_SHA1_Update(&ctx, buffer, len1);
4060 if (diff_filespec_is_binary(p->one) ||
4061 diff_filespec_is_binary(p->two)) {
4062 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4063 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4064 continue;
4065 }
4067 xpp.flags = 0;
4068 xecfg.ctxlen = 3;
4069 xecfg.flags = 0;
4070 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4071 &xpp, &xecfg);
4072 }
4074 git_SHA1_Final(sha1, &ctx);
4075 return 0;
4076 }
4078 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4079 {
4080 struct diff_queue_struct *q = &diff_queued_diff;
4081 int i;
4082 int result = diff_get_patch_id(options, sha1);
4084 for (i = 0; i < q->nr; i++)
4085 diff_free_filepair(q->queue[i]);
4087 free(q->queue);
4088 DIFF_QUEUE_CLEAR(q);
4090 return result;
4091 }
4093 static int is_summary_empty(const struct diff_queue_struct *q)
4094 {
4095 int i;
4097 for (i = 0; i < q->nr; i++) {
4098 const struct diff_filepair *p = q->queue[i];
4100 switch (p->status) {
4101 case DIFF_STATUS_DELETED:
4102 case DIFF_STATUS_ADDED:
4103 case DIFF_STATUS_COPIED:
4104 case DIFF_STATUS_RENAMED:
4105 return 0;
4106 default:
4107 if (p->score)
4108 return 0;
4109 if (p->one->mode && p->two->mode &&
4110 p->one->mode != p->two->mode)
4111 return 0;
4112 break;
4113 }
4114 }
4115 return 1;
4116 }
4118 static const char rename_limit_warning[] =
4119 "inexact rename detection was skipped due to too many files.";
4121 static const char degrade_cc_to_c_warning[] =
4122 "only found copies from modified paths due to too many files.";
4124 static const char rename_limit_advice[] =
4125 "you may want to set your %s variable to at least "
4126 "%d and retry the command.";
4128 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4129 {
4130 if (degraded_cc)
4131 warning(degrade_cc_to_c_warning);
4132 else if (needed)
4133 warning(rename_limit_warning);
4134 else
4135 return;
4136 if (0 < needed && needed < 32767)
4137 warning(rename_limit_advice, varname, needed);
4138 }
4140 void diff_flush(struct diff_options *options)
4141 {
4142 struct diff_queue_struct *q = &diff_queued_diff;
4143 int i, output_format = options->output_format;
4144 int separator = 0;
4145 int dirstat_by_line = 0;
4147 /*
4148 * Order: raw, stat, summary, patch
4149 * or: name/name-status/checkdiff (other bits clear)
4150 */
4151 if (!q->nr)
4152 goto free_queue;
4154 if (output_format & (DIFF_FORMAT_RAW |
4155 DIFF_FORMAT_NAME |
4156 DIFF_FORMAT_NAME_STATUS |
4157 DIFF_FORMAT_CHECKDIFF)) {
4158 for (i = 0; i < q->nr; i++) {
4159 struct diff_filepair *p = q->queue[i];
4160 if (check_pair_status(p))
4161 flush_one_pair(p, options);
4162 }
4163 separator++;
4164 }
4166 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4167 dirstat_by_line = 1;
4169 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4170 dirstat_by_line) {
4171 struct diffstat_t diffstat;
4173 memset(&diffstat, 0, sizeof(struct diffstat_t));
4174 for (i = 0; i < q->nr; i++) {
4175 struct diff_filepair *p = q->queue[i];
4176 if (check_pair_status(p))
4177 diff_flush_stat(p, options, &diffstat);
4178 }
4179 if (output_format & DIFF_FORMAT_NUMSTAT)
4180 show_numstat(&diffstat, options);
4181 if (output_format & DIFF_FORMAT_DIFFSTAT)
4182 show_stats(&diffstat, options);
4183 if (output_format & DIFF_FORMAT_SHORTSTAT)
4184 show_shortstats(&diffstat, options);
4185 if (output_format & DIFF_FORMAT_DIRSTAT)
4186 show_dirstat_by_line(&diffstat, options);
4187 free_diffstat_info(&diffstat);
4188 separator++;
4189 }
4190 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4191 show_dirstat(options);
4193 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4194 for (i = 0; i < q->nr; i++) {
4195 diff_summary(options, q->queue[i]);
4196 }
4197 separator++;
4198 }
4200 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4201 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4202 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4203 /*
4204 * run diff_flush_patch for the exit status. setting
4205 * options->file to /dev/null should be safe, becaue we
4206 * aren't supposed to produce any output anyway.
4207 */
4208 if (options->close_file)
4209 fclose(options->file);
4210 options->file = fopen("/dev/null", "w");
4211 if (!options->file)
4212 die_errno("Could not open /dev/null");
4213 options->close_file = 1;
4214 for (i = 0; i < q->nr; i++) {
4215 struct diff_filepair *p = q->queue[i];
4216 if (check_pair_status(p))
4217 diff_flush_patch(p, options);
4218 if (options->found_changes)
4219 break;
4220 }
4221 }
4223 if (output_format & DIFF_FORMAT_PATCH) {
4224 if (separator) {
4225 putc(options->line_termination, options->file);
4226 if (options->stat_sep) {
4227 /* attach patch instead of inline */
4228 fputs(options->stat_sep, options->file);
4229 }
4230 }
4232 for (i = 0; i < q->nr; i++) {
4233 struct diff_filepair *p = q->queue[i];
4234 if (check_pair_status(p))
4235 diff_flush_patch(p, options);
4236 }
4237 }
4239 if (output_format & DIFF_FORMAT_CALLBACK)
4240 options->format_callback(q, options, options->format_callback_data);
4242 for (i = 0; i < q->nr; i++)
4243 diff_free_filepair(q->queue[i]);
4244 free_queue:
4245 free(q->queue);
4246 DIFF_QUEUE_CLEAR(q);
4247 if (options->close_file)
4248 fclose(options->file);
4250 /*
4251 * Report the content-level differences with HAS_CHANGES;
4252 * diff_addremove/diff_change does not set the bit when
4253 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4254 */
4255 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4256 if (options->found_changes)
4257 DIFF_OPT_SET(options, HAS_CHANGES);
4258 else
4259 DIFF_OPT_CLR(options, HAS_CHANGES);
4260 }
4261 }
4263 static void diffcore_apply_filter(const char *filter)
4264 {
4265 int i;
4266 struct diff_queue_struct *q = &diff_queued_diff;
4267 struct diff_queue_struct outq;
4268 DIFF_QUEUE_CLEAR(&outq);
4270 if (!filter)
4271 return;
4273 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4274 int found;
4275 for (i = found = 0; !found && i < q->nr; i++) {
4276 struct diff_filepair *p = q->queue[i];
4277 if (((p->status == DIFF_STATUS_MODIFIED) &&
4278 ((p->score &&
4279 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4280 (!p->score &&
4281 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4282 ((p->status != DIFF_STATUS_MODIFIED) &&
4283 strchr(filter, p->status)))
4284 found++;
4285 }
4286 if (found)
4287 return;
4289 /* otherwise we will clear the whole queue
4290 * by copying the empty outq at the end of this
4291 * function, but first clear the current entries
4292 * in the queue.
4293 */
4294 for (i = 0; i < q->nr; i++)
4295 diff_free_filepair(q->queue[i]);
4296 }
4297 else {
4298 /* Only the matching ones */
4299 for (i = 0; i < q->nr; i++) {
4300 struct diff_filepair *p = q->queue[i];
4302 if (((p->status == DIFF_STATUS_MODIFIED) &&
4303 ((p->score &&
4304 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4305 (!p->score &&
4306 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4307 ((p->status != DIFF_STATUS_MODIFIED) &&
4308 strchr(filter, p->status)))
4309 diff_q(&outq, p);
4310 else
4311 diff_free_filepair(p);
4312 }
4313 }
4314 free(q->queue);
4315 *q = outq;
4316 }
4318 /* Check whether two filespecs with the same mode and size are identical */
4319 static int diff_filespec_is_identical(struct diff_filespec *one,
4320 struct diff_filespec *two)
4321 {
4322 if (S_ISGITLINK(one->mode))
4323 return 0;
4324 if (diff_populate_filespec(one, 0))
4325 return 0;
4326 if (diff_populate_filespec(two, 0))
4327 return 0;
4328 return !memcmp(one->data, two->data, one->size);
4329 }
4331 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4332 {
4333 int i;
4334 struct diff_queue_struct *q = &diff_queued_diff;
4335 struct diff_queue_struct outq;
4336 DIFF_QUEUE_CLEAR(&outq);
4338 for (i = 0; i < q->nr; i++) {
4339 struct diff_filepair *p = q->queue[i];
4341 /*
4342 * 1. Entries that come from stat info dirtiness
4343 * always have both sides (iow, not create/delete),
4344 * one side of the object name is unknown, with
4345 * the same mode and size. Keep the ones that
4346 * do not match these criteria. They have real
4347 * differences.
4348 *
4349 * 2. At this point, the file is known to be modified,
4350 * with the same mode and size, and the object
4351 * name of one side is unknown. Need to inspect
4352 * the identical contents.
4353 */
4354 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4355 !DIFF_FILE_VALID(p->two) ||
4356 (p->one->sha1_valid && p->two->sha1_valid) ||
4357 (p->one->mode != p->two->mode) ||
4358 diff_populate_filespec(p->one, 1) ||
4359 diff_populate_filespec(p->two, 1) ||
4360 (p->one->size != p->two->size) ||
4361 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4362 diff_q(&outq, p);
4363 else {
4364 /*
4365 * The caller can subtract 1 from skip_stat_unmatch
4366 * to determine how many paths were dirty only
4367 * due to stat info mismatch.
4368 */
4369 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4370 diffopt->skip_stat_unmatch++;
4371 diff_free_filepair(p);
4372 }
4373 }
4374 free(q->queue);
4375 *q = outq;
4376 }
4378 static int diffnamecmp(const void *a_, const void *b_)
4379 {
4380 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4381 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4382 const char *name_a, *name_b;
4384 name_a = a->one ? a->one->path : a->two->path;
4385 name_b = b->one ? b->one->path : b->two->path;
4386 return strcmp(name_a, name_b);
4387 }
4389 void diffcore_fix_diff_index(struct diff_options *options)
4390 {
4391 struct diff_queue_struct *q = &diff_queued_diff;
4392 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4393 }
4395 void diffcore_std(struct diff_options *options)
4396 {
4397 if (options->skip_stat_unmatch)
4398 diffcore_skip_stat_unmatch(options);
4399 if (!options->found_follow) {
4400 /* See try_to_follow_renames() in tree-diff.c */
4401 if (options->break_opt != -1)
4402 diffcore_break(options->break_opt);
4403 if (options->detect_rename)
4404 diffcore_rename(options);
4405 if (options->break_opt != -1)
4406 diffcore_merge_broken();
4407 }
4408 if (options->pickaxe)
4409 diffcore_pickaxe(options);
4410 if (options->orderfile)
4411 diffcore_order(options->orderfile);
4412 if (!options->found_follow)
4413 /* See try_to_follow_renames() in tree-diff.c */
4414 diff_resolve_rename_copy();
4415 diffcore_apply_filter(options->filter);
4417 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4418 DIFF_OPT_SET(options, HAS_CHANGES);
4419 else
4420 DIFF_OPT_CLR(options, HAS_CHANGES);
4422 options->found_follow = 0;
4423 }
4425 int diff_result_code(struct diff_options *opt, int status)
4426 {
4427 int result = 0;
4429 diff_warn_rename_limit("diff.renamelimit",
4430 opt->needed_rename_limit,
4431 opt->degraded_cc_to_c);
4432 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4433 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4434 return status;
4435 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4436 DIFF_OPT_TST(opt, HAS_CHANGES))
4437 result |= 01;
4438 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4439 DIFF_OPT_TST(opt, CHECK_FAILED))
4440 result |= 02;
4441 return result;
4442 }
4444 int diff_can_quit_early(struct diff_options *opt)
4445 {
4446 return (DIFF_OPT_TST(opt, QUICK) &&
4447 !opt->filter &&
4448 DIFF_OPT_TST(opt, HAS_CHANGES));
4449 }
4451 /*
4452 * Shall changes to this submodule be ignored?
4453 *
4454 * Submodule changes can be configured to be ignored separately for each path,
4455 * but that configuration can be overridden from the command line.
4456 */
4457 static int is_submodule_ignored(const char *path, struct diff_options *options)
4458 {
4459 int ignored = 0;
4460 unsigned orig_flags = options->flags;
4461 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4462 set_diffopt_flags_from_submodule_config(options, path);
4463 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4464 ignored = 1;
4465 options->flags = orig_flags;
4466 return ignored;
4467 }
4469 void diff_addremove(struct diff_options *options,
4470 int addremove, unsigned mode,
4471 const unsigned char *sha1,
4472 const char *concatpath, unsigned dirty_submodule)
4473 {
4474 struct diff_filespec *one, *two;
4476 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4477 return;
4479 /* This may look odd, but it is a preparation for
4480 * feeding "there are unchanged files which should
4481 * not produce diffs, but when you are doing copy
4482 * detection you would need them, so here they are"
4483 * entries to the diff-core. They will be prefixed
4484 * with something like '=' or '*' (I haven't decided
4485 * which but should not make any difference).
4486 * Feeding the same new and old to diff_change()
4487 * also has the same effect.
4488 * Before the final output happens, they are pruned after
4489 * merged into rename/copy pairs as appropriate.
4490 */
4491 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4492 addremove = (addremove == '+' ? '-' :
4493 addremove == '-' ? '+' : addremove);
4495 if (options->prefix &&
4496 strncmp(concatpath, options->prefix, options->prefix_length))
4497 return;
4499 one = alloc_filespec(concatpath);
4500 two = alloc_filespec(concatpath);
4502 if (addremove != '+')
4503 fill_filespec(one, sha1, mode);
4504 if (addremove != '-') {
4505 fill_filespec(two, sha1, mode);
4506 two->dirty_submodule = dirty_submodule;
4507 }
4509 diff_queue(&diff_queued_diff, one, two);
4510 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4511 DIFF_OPT_SET(options, HAS_CHANGES);
4512 }
4514 void diff_change(struct diff_options *options,
4515 unsigned old_mode, unsigned new_mode,
4516 const unsigned char *old_sha1,
4517 const unsigned char *new_sha1,
4518 const char *concatpath,
4519 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4520 {
4521 struct diff_filespec *one, *two;
4523 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4524 is_submodule_ignored(concatpath, options))
4525 return;
4527 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4528 unsigned tmp;
4529 const unsigned char *tmp_c;
4530 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4531 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4532 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4533 new_dirty_submodule = tmp;
4534 }
4536 if (options->prefix &&
4537 strncmp(concatpath, options->prefix, options->prefix_length))
4538 return;
4540 one = alloc_filespec(concatpath);
4541 two = alloc_filespec(concatpath);
4542 fill_filespec(one, old_sha1, old_mode);
4543 fill_filespec(two, new_sha1, new_mode);
4544 one->dirty_submodule = old_dirty_submodule;
4545 two->dirty_submodule = new_dirty_submodule;
4547 diff_queue(&diff_queued_diff, one, two);
4548 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4549 DIFF_OPT_SET(options, HAS_CHANGES);
4550 }
4552 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4553 {
4554 struct diff_filepair *pair;
4555 struct diff_filespec *one, *two;
4557 if (options->prefix &&
4558 strncmp(path, options->prefix, options->prefix_length))
4559 return NULL;
4561 one = alloc_filespec(path);
4562 two = alloc_filespec(path);
4563 pair = diff_queue(&diff_queued_diff, one, two);
4564 pair->is_unmerged = 1;
4565 return pair;
4566 }
4568 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4569 size_t *outsize)
4570 {
4571 struct diff_tempfile *temp;
4572 const char *argv[3];
4573 const char **arg = argv;
4574 struct child_process child;
4575 struct strbuf buf = STRBUF_INIT;
4576 int err = 0;
4578 temp = prepare_temp_file(spec->path, spec);
4579 *arg++ = pgm;
4580 *arg++ = temp->name;
4581 *arg = NULL;
4583 memset(&child, 0, sizeof(child));
4584 child.use_shell = 1;
4585 child.argv = argv;
4586 child.out = -1;
4587 if (start_command(&child)) {
4588 remove_tempfile();
4589 return NULL;
4590 }
4592 if (strbuf_read(&buf, child.out, 0) < 0)
4593 err = error("error reading from textconv command '%s'", pgm);
4594 close(child.out);
4596 if (finish_command(&child) || err) {
4597 strbuf_release(&buf);
4598 remove_tempfile();
4599 return NULL;
4600 }
4601 remove_tempfile();
4603 return strbuf_detach(&buf, outsize);
4604 }
4606 size_t fill_textconv(struct userdiff_driver *driver,
4607 struct diff_filespec *df,
4608 char **outbuf)
4609 {
4610 size_t size;
4612 if (!driver || !driver->textconv) {
4613 if (!DIFF_FILE_VALID(df)) {
4614 *outbuf = "";
4615 return 0;
4616 }
4617 if (diff_populate_filespec(df, 0))
4618 die("unable to read files to diff");
4619 *outbuf = df->data;
4620 return df->size;
4621 }
4623 if (driver->textconv_cache && df->sha1_valid) {
4624 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4625 &size);
4626 if (*outbuf)
4627 return size;
4628 }
4630 *outbuf = run_textconv(driver->textconv, df, &size);
4631 if (!*outbuf)
4632 die("unable to read files to diff");
4634 if (driver->textconv_cache && df->sha1_valid) {
4635 /* ignore errors, as we might be in a readonly repository */
4636 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4637 size);
4638 /*
4639 * we could save up changes and flush them all at the end,
4640 * but we would need an extra call after all diffing is done.
4641 * Since generating a cache entry is the slow path anyway,
4642 * this extra overhead probably isn't a big deal.
4643 */
4644 notes_cache_write(driver->textconv_cache);
4645 }
4647 return size;
4648 }