![[tokkee]](http://tokkee.org/images/avatar.png){kind=avatar}
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 400;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static int diff_dirstat_permille_default = 30;
35 static struct diff_options default_diff_options;
37 static char diff_colors[][COLOR_MAXLEN] = {
38 GIT_COLOR_RESET,
39 GIT_COLOR_NORMAL, /* PLAIN */
40 GIT_COLOR_BOLD, /* METAINFO */
41 GIT_COLOR_CYAN, /* FRAGINFO */
42 GIT_COLOR_RED, /* OLD */
43 GIT_COLOR_GREEN, /* NEW */
44 GIT_COLOR_YELLOW, /* COMMIT */
45 GIT_COLOR_BG_RED, /* WHITESPACE */
46 GIT_COLOR_NORMAL, /* FUNCINFO */
47 };
49 static int parse_diff_color_slot(const char *var, int ofs)
50 {
51 if (!strcasecmp(var+ofs, "plain"))
52 return DIFF_PLAIN;
53 if (!strcasecmp(var+ofs, "meta"))
54 return DIFF_METAINFO;
55 if (!strcasecmp(var+ofs, "frag"))
56 return DIFF_FRAGINFO;
57 if (!strcasecmp(var+ofs, "old"))
58 return DIFF_FILE_OLD;
59 if (!strcasecmp(var+ofs, "new"))
60 return DIFF_FILE_NEW;
61 if (!strcasecmp(var+ofs, "commit"))
62 return DIFF_COMMIT;
63 if (!strcasecmp(var+ofs, "whitespace"))
64 return DIFF_WHITESPACE;
65 if (!strcasecmp(var+ofs, "func"))
66 return DIFF_FUNCINFO;
67 return -1;
68 }
70 static int parse_dirstat_params(struct diff_options *options, const char *params,
71 struct strbuf *errmsg)
72 {
73 const char *p = params;
74 int p_len, ret = 0;
76 while (*p) {
77 p_len = strchrnul(p, ',') - p;
78 if (!memcmp(p, "changes", p_len)) {
79 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
80 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
81 } else if (!memcmp(p, "lines", p_len)) {
82 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
83 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
84 } else if (!memcmp(p, "files", p_len)) {
85 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
86 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
87 } else if (!memcmp(p, "noncumulative", p_len)) {
88 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
89 } else if (!memcmp(p, "cumulative", p_len)) {
90 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
91 } else if (isdigit(*p)) {
92 char *end;
93 int permille = strtoul(p, &end, 10) * 10;
94 if (*end == '.' && isdigit(*++end)) {
95 /* only use first digit */
96 permille += *end - '0';
97 /* .. and ignore any further digits */
98 while (isdigit(*++end))
99 ; /* nothing */
100 }
101 if (end - p == p_len)
102 options->dirstat_permille = permille;
103 else {
104 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%.*s'\n"),
105 p_len, p);
106 ret++;
107 }
108 } else {
109 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%.*s'\n"),
110 p_len, p);
111 ret++;
112 }
114 p += p_len;
116 if (*p)
117 p++; /* more parameters, swallow separator */
118 }
119 return ret;
120 }
122 static int git_config_rename(const char *var, const char *value)
123 {
124 if (!value)
125 return DIFF_DETECT_RENAME;
126 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
127 return DIFF_DETECT_COPY;
128 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
129 }
131 /*
132 * These are to give UI layer defaults.
133 * The core-level commands such as git-diff-files should
134 * never be affected by the setting of diff.renames
135 * the user happens to have in the configuration file.
136 */
137 int git_diff_ui_config(const char *var, const char *value, void *cb)
138 {
139 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
140 diff_use_color_default = git_config_colorbool(var, value, -1);
141 return 0;
142 }
143 if (!strcmp(var, "diff.renames")) {
144 diff_detect_rename_default = git_config_rename(var, value);
145 return 0;
146 }
147 if (!strcmp(var, "diff.autorefreshindex")) {
148 diff_auto_refresh_index = git_config_bool(var, value);
149 return 0;
150 }
151 if (!strcmp(var, "diff.mnemonicprefix")) {
152 diff_mnemonic_prefix = git_config_bool(var, value);
153 return 0;
154 }
155 if (!strcmp(var, "diff.noprefix")) {
156 diff_no_prefix = git_config_bool(var, value);
157 return 0;
158 }
159 if (!strcmp(var, "diff.external"))
160 return git_config_string(&external_diff_cmd_cfg, var, value);
161 if (!strcmp(var, "diff.wordregex"))
162 return git_config_string(&diff_word_regex_cfg, var, value);
164 if (!strcmp(var, "diff.ignoresubmodules"))
165 handle_ignore_submodules_arg(&default_diff_options, value);
167 return git_diff_basic_config(var, value, cb);
168 }
170 int git_diff_basic_config(const char *var, const char *value, void *cb)
171 {
172 if (!strcmp(var, "diff.renamelimit")) {
173 diff_rename_limit_default = git_config_int(var, value);
174 return 0;
175 }
177 switch (userdiff_config(var, value)) {
178 case 0: break;
179 case -1: return -1;
180 default: return 0;
181 }
183 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
184 int slot = parse_diff_color_slot(var, 11);
185 if (slot < 0)
186 return 0;
187 if (!value)
188 return config_error_nonbool(var);
189 color_parse(value, var, diff_colors[slot]);
190 return 0;
191 }
193 /* like GNU diff's --suppress-blank-empty option */
194 if (!strcmp(var, "diff.suppressblankempty") ||
195 /* for backwards compatibility */
196 !strcmp(var, "diff.suppress-blank-empty")) {
197 diff_suppress_blank_empty = git_config_bool(var, value);
198 return 0;
199 }
201 if (!strcmp(var, "diff.dirstat")) {
202 struct strbuf errmsg = STRBUF_INIT;
203 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
204 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
205 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
206 errmsg.buf);
207 strbuf_release(&errmsg);
208 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
209 return 0;
210 }
212 if (!prefixcmp(var, "submodule."))
213 return parse_submodule_config_option(var, value);
215 return git_color_default_config(var, value, cb);
216 }
218 static char *quote_two(const char *one, const char *two)
219 {
220 int need_one = quote_c_style(one, NULL, NULL, 1);
221 int need_two = quote_c_style(two, NULL, NULL, 1);
222 struct strbuf res = STRBUF_INIT;
224 if (need_one + need_two) {
225 strbuf_addch(&res, '"');
226 quote_c_style(one, &res, NULL, 1);
227 quote_c_style(two, &res, NULL, 1);
228 strbuf_addch(&res, '"');
229 } else {
230 strbuf_addstr(&res, one);
231 strbuf_addstr(&res, two);
232 }
233 return strbuf_detach(&res, NULL);
234 }
236 static const char *external_diff(void)
237 {
238 static const char *external_diff_cmd = NULL;
239 static int done_preparing = 0;
241 if (done_preparing)
242 return external_diff_cmd;
243 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
244 if (!external_diff_cmd)
245 external_diff_cmd = external_diff_cmd_cfg;
246 done_preparing = 1;
247 return external_diff_cmd;
248 }
250 static struct diff_tempfile {
251 const char *name; /* filename external diff should read from */
252 char hex[41];
253 char mode[10];
254 char tmp_path[PATH_MAX];
255 } diff_temp[2];
257 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
259 struct emit_callback {
260 int color_diff;
261 unsigned ws_rule;
262 int blank_at_eof_in_preimage;
263 int blank_at_eof_in_postimage;
264 int lno_in_preimage;
265 int lno_in_postimage;
266 sane_truncate_fn truncate;
267 const char **label_path;
268 struct diff_words_data *diff_words;
269 struct diff_options *opt;
270 int *found_changesp;
271 struct strbuf *header;
272 };
274 static int count_lines(const char *data, int size)
275 {
276 int count, ch, completely_empty = 1, nl_just_seen = 0;
277 count = 0;
278 while (0 < size--) {
279 ch = *data++;
280 if (ch == '\n') {
281 count++;
282 nl_just_seen = 1;
283 completely_empty = 0;
284 }
285 else {
286 nl_just_seen = 0;
287 completely_empty = 0;
288 }
289 }
290 if (completely_empty)
291 return 0;
292 if (!nl_just_seen)
293 count++; /* no trailing newline */
294 return count;
295 }
297 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
298 {
299 if (!DIFF_FILE_VALID(one)) {
300 mf->ptr = (char *)""; /* does not matter */
301 mf->size = 0;
302 return 0;
303 }
304 else if (diff_populate_filespec(one, 0))
305 return -1;
307 mf->ptr = one->data;
308 mf->size = one->size;
309 return 0;
310 }
312 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
313 static unsigned long diff_filespec_size(struct diff_filespec *one)
314 {
315 if (!DIFF_FILE_VALID(one))
316 return 0;
317 diff_populate_filespec(one, 1);
318 return one->size;
319 }
321 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
322 {
323 char *ptr = mf->ptr;
324 long size = mf->size;
325 int cnt = 0;
327 if (!size)
328 return cnt;
329 ptr += size - 1; /* pointing at the very end */
330 if (*ptr != '\n')
331 ; /* incomplete line */
332 else
333 ptr--; /* skip the last LF */
334 while (mf->ptr < ptr) {
335 char *prev_eol;
336 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
337 if (*prev_eol == '\n')
338 break;
339 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
340 break;
341 cnt++;
342 ptr = prev_eol - 1;
343 }
344 return cnt;
345 }
347 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
348 struct emit_callback *ecbdata)
349 {
350 int l1, l2, at;
351 unsigned ws_rule = ecbdata->ws_rule;
352 l1 = count_trailing_blank(mf1, ws_rule);
353 l2 = count_trailing_blank(mf2, ws_rule);
354 if (l2 <= l1) {
355 ecbdata->blank_at_eof_in_preimage = 0;
356 ecbdata->blank_at_eof_in_postimage = 0;
357 return;
358 }
359 at = count_lines(mf1->ptr, mf1->size);
360 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
362 at = count_lines(mf2->ptr, mf2->size);
363 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
364 }
366 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
367 int first, const char *line, int len)
368 {
369 int has_trailing_newline, has_trailing_carriage_return;
370 int nofirst;
371 FILE *file = o->file;
373 if (o->output_prefix) {
374 struct strbuf *msg = NULL;
375 msg = o->output_prefix(o, o->output_prefix_data);
376 assert(msg);
377 fwrite(msg->buf, msg->len, 1, file);
378 }
380 if (len == 0) {
381 has_trailing_newline = (first == '\n');
382 has_trailing_carriage_return = (!has_trailing_newline &&
383 (first == '\r'));
384 nofirst = has_trailing_newline || has_trailing_carriage_return;
385 } else {
386 has_trailing_newline = (len > 0 && line[len-1] == '\n');
387 if (has_trailing_newline)
388 len--;
389 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
390 if (has_trailing_carriage_return)
391 len--;
392 nofirst = 0;
393 }
395 if (len || !nofirst) {
396 fputs(set, file);
397 if (!nofirst)
398 fputc(first, file);
399 fwrite(line, len, 1, file);
400 fputs(reset, file);
401 }
402 if (has_trailing_carriage_return)
403 fputc('\r', file);
404 if (has_trailing_newline)
405 fputc('\n', file);
406 }
408 static void emit_line(struct diff_options *o, const char *set, const char *reset,
409 const char *line, int len)
410 {
411 emit_line_0(o, set, reset, line[0], line+1, len-1);
412 }
414 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
415 {
416 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
417 ecbdata->blank_at_eof_in_preimage &&
418 ecbdata->blank_at_eof_in_postimage &&
419 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
420 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
421 return 0;
422 return ws_blank_line(line, len, ecbdata->ws_rule);
423 }
425 static void emit_add_line(const char *reset,
426 struct emit_callback *ecbdata,
427 const char *line, int len)
428 {
429 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
430 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
432 if (!*ws)
433 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
434 else if (new_blank_line_at_eof(ecbdata, line, len))
435 /* Blank line at EOF - paint '+' as well */
436 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
437 else {
438 /* Emit just the prefix, then the rest. */
439 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
440 ws_check_emit(line, len, ecbdata->ws_rule,
441 ecbdata->opt->file, set, reset, ws);
442 }
443 }
445 static void emit_hunk_header(struct emit_callback *ecbdata,
446 const char *line, int len)
447 {
448 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
449 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
450 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
451 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
452 static const char atat[2] = { '@', '@' };
453 const char *cp, *ep;
454 struct strbuf msgbuf = STRBUF_INIT;
455 int org_len = len;
456 int i = 1;
458 /*
459 * As a hunk header must begin with "@@ -<old>, +<new> @@",
460 * it always is at least 10 bytes long.
461 */
462 if (len < 10 ||
463 memcmp(line, atat, 2) ||
464 !(ep = memmem(line + 2, len - 2, atat, 2))) {
465 emit_line(ecbdata->opt, plain, reset, line, len);
466 return;
467 }
468 ep += 2; /* skip over @@ */
470 /* The hunk header in fraginfo color */
471 strbuf_add(&msgbuf, frag, strlen(frag));
472 strbuf_add(&msgbuf, line, ep - line);
473 strbuf_add(&msgbuf, reset, strlen(reset));
475 /*
476 * trailing "\r\n"
477 */
478 for ( ; i < 3; i++)
479 if (line[len - i] == '\r' || line[len - i] == '\n')
480 len--;
482 /* blank before the func header */
483 for (cp = ep; ep - line < len; ep++)
484 if (*ep != ' ' && *ep != '\t')
485 break;
486 if (ep != cp) {
487 strbuf_add(&msgbuf, plain, strlen(plain));
488 strbuf_add(&msgbuf, cp, ep - cp);
489 strbuf_add(&msgbuf, reset, strlen(reset));
490 }
492 if (ep < line + len) {
493 strbuf_add(&msgbuf, func, strlen(func));
494 strbuf_add(&msgbuf, ep, line + len - ep);
495 strbuf_add(&msgbuf, reset, strlen(reset));
496 }
498 strbuf_add(&msgbuf, line + len, org_len - len);
499 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
500 strbuf_release(&msgbuf);
501 }
503 static struct diff_tempfile *claim_diff_tempfile(void) {
504 int i;
505 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
506 if (!diff_temp[i].name)
507 return diff_temp + i;
508 die("BUG: diff is failing to clean up its tempfiles");
509 }
511 static int remove_tempfile_installed;
513 static void remove_tempfile(void)
514 {
515 int i;
516 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
517 if (diff_temp[i].name == diff_temp[i].tmp_path)
518 unlink_or_warn(diff_temp[i].name);
519 diff_temp[i].name = NULL;
520 }
521 }
523 static void remove_tempfile_on_signal(int signo)
524 {
525 remove_tempfile();
526 sigchain_pop(signo);
527 raise(signo);
528 }
530 static void print_line_count(FILE *file, int count)
531 {
532 switch (count) {
533 case 0:
534 fprintf(file, "0,0");
535 break;
536 case 1:
537 fprintf(file, "1");
538 break;
539 default:
540 fprintf(file, "1,%d", count);
541 break;
542 }
543 }
545 static void emit_rewrite_lines(struct emit_callback *ecb,
546 int prefix, const char *data, int size)
547 {
548 const char *endp = NULL;
549 static const char *nneof = " No newline at end of file\n";
550 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
551 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
553 while (0 < size) {
554 int len;
556 endp = memchr(data, '\n', size);
557 len = endp ? (endp - data + 1) : size;
558 if (prefix != '+') {
559 ecb->lno_in_preimage++;
560 emit_line_0(ecb->opt, old, reset, '-',
561 data, len);
562 } else {
563 ecb->lno_in_postimage++;
564 emit_add_line(reset, ecb, data, len);
565 }
566 size -= len;
567 data += len;
568 }
569 if (!endp) {
570 const char *plain = diff_get_color(ecb->color_diff,
571 DIFF_PLAIN);
572 emit_line_0(ecb->opt, plain, reset, '\\',
573 nneof, strlen(nneof));
574 }
575 }
577 static void emit_rewrite_diff(const char *name_a,
578 const char *name_b,
579 struct diff_filespec *one,
580 struct diff_filespec *two,
581 struct userdiff_driver *textconv_one,
582 struct userdiff_driver *textconv_two,
583 struct diff_options *o)
584 {
585 int lc_a, lc_b;
586 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
587 const char *name_a_tab, *name_b_tab;
588 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
589 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
590 const char *reset = diff_get_color(color_diff, DIFF_RESET);
591 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
592 const char *a_prefix, *b_prefix;
593 char *data_one, *data_two;
594 size_t size_one, size_two;
595 struct emit_callback ecbdata;
596 char *line_prefix = "";
597 struct strbuf *msgbuf;
599 if (o && o->output_prefix) {
600 msgbuf = o->output_prefix(o, o->output_prefix_data);
601 line_prefix = msgbuf->buf;
602 }
604 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
605 a_prefix = o->b_prefix;
606 b_prefix = o->a_prefix;
607 } else {
608 a_prefix = o->a_prefix;
609 b_prefix = o->b_prefix;
610 }
612 name_a += (*name_a == '/');
613 name_b += (*name_b == '/');
614 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
615 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
617 strbuf_reset(&a_name);
618 strbuf_reset(&b_name);
619 quote_two_c_style(&a_name, a_prefix, name_a, 0);
620 quote_two_c_style(&b_name, b_prefix, name_b, 0);
622 size_one = fill_textconv(textconv_one, one, &data_one);
623 size_two = fill_textconv(textconv_two, two, &data_two);
625 memset(&ecbdata, 0, sizeof(ecbdata));
626 ecbdata.color_diff = color_diff;
627 ecbdata.found_changesp = &o->found_changes;
628 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
629 ecbdata.opt = o;
630 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
631 mmfile_t mf1, mf2;
632 mf1.ptr = (char *)data_one;
633 mf2.ptr = (char *)data_two;
634 mf1.size = size_one;
635 mf2.size = size_two;
636 check_blank_at_eof(&mf1, &mf2, &ecbdata);
637 }
638 ecbdata.lno_in_preimage = 1;
639 ecbdata.lno_in_postimage = 1;
641 lc_a = count_lines(data_one, size_one);
642 lc_b = count_lines(data_two, size_two);
643 fprintf(o->file,
644 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
645 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
646 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
647 line_prefix, fraginfo);
648 if (!o->irreversible_delete)
649 print_line_count(o->file, lc_a);
650 else
651 fprintf(o->file, "?,?");
652 fprintf(o->file, " +");
653 print_line_count(o->file, lc_b);
654 fprintf(o->file, " @@%s\n", reset);
655 if (lc_a && !o->irreversible_delete)
656 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
657 if (lc_b)
658 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
659 if (textconv_one)
660 free((char *)data_one);
661 if (textconv_two)
662 free((char *)data_two);
663 }
665 struct diff_words_buffer {
666 mmfile_t text;
667 long alloc;
668 struct diff_words_orig {
669 const char *begin, *end;
670 } *orig;
671 int orig_nr, orig_alloc;
672 };
674 static void diff_words_append(char *line, unsigned long len,
675 struct diff_words_buffer *buffer)
676 {
677 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
678 line++;
679 len--;
680 memcpy(buffer->text.ptr + buffer->text.size, line, len);
681 buffer->text.size += len;
682 buffer->text.ptr[buffer->text.size] = '\0';
683 }
685 struct diff_words_style_elem {
686 const char *prefix;
687 const char *suffix;
688 const char *color; /* NULL; filled in by the setup code if
689 * color is enabled */
690 };
692 struct diff_words_style {
693 enum diff_words_type type;
694 struct diff_words_style_elem new, old, ctx;
695 const char *newline;
696 };
698 static struct diff_words_style diff_words_styles[] = {
699 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
700 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
701 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
702 };
704 struct diff_words_data {
705 struct diff_words_buffer minus, plus;
706 const char *current_plus;
707 int last_minus;
708 struct diff_options *opt;
709 regex_t *word_regex;
710 enum diff_words_type type;
711 struct diff_words_style *style;
712 };
714 static int fn_out_diff_words_write_helper(FILE *fp,
715 struct diff_words_style_elem *st_el,
716 const char *newline,
717 size_t count, const char *buf,
718 const char *line_prefix)
719 {
720 int print = 0;
722 while (count) {
723 char *p = memchr(buf, '\n', count);
724 if (print)
725 fputs(line_prefix, fp);
726 if (p != buf) {
727 if (st_el->color && fputs(st_el->color, fp) < 0)
728 return -1;
729 if (fputs(st_el->prefix, fp) < 0 ||
730 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
731 fputs(st_el->suffix, fp) < 0)
732 return -1;
733 if (st_el->color && *st_el->color
734 && fputs(GIT_COLOR_RESET, fp) < 0)
735 return -1;
736 }
737 if (!p)
738 return 0;
739 if (fputs(newline, fp) < 0)
740 return -1;
741 count -= p + 1 - buf;
742 buf = p + 1;
743 print = 1;
744 }
745 return 0;
746 }
748 /*
749 * '--color-words' algorithm can be described as:
750 *
751 * 1. collect a the minus/plus lines of a diff hunk, divided into
752 * minus-lines and plus-lines;
753 *
754 * 2. break both minus-lines and plus-lines into words and
755 * place them into two mmfile_t with one word for each line;
756 *
757 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
758 *
759 * And for the common parts of the both file, we output the plus side text.
760 * diff_words->current_plus is used to trace the current position of the plus file
761 * which printed. diff_words->last_minus is used to trace the last minus word
762 * printed.
763 *
764 * For '--graph' to work with '--color-words', we need to output the graph prefix
765 * on each line of color words output. Generally, there are two conditions on
766 * which we should output the prefix.
767 *
768 * 1. diff_words->last_minus == 0 &&
769 * diff_words->current_plus == diff_words->plus.text.ptr
770 *
771 * that is: the plus text must start as a new line, and if there is no minus
772 * word printed, a graph prefix must be printed.
773 *
774 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
775 * *(diff_words->current_plus - 1) == '\n'
776 *
777 * that is: a graph prefix must be printed following a '\n'
778 */
779 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
780 {
781 if ((diff_words->last_minus == 0 &&
782 diff_words->current_plus == diff_words->plus.text.ptr) ||
783 (diff_words->current_plus > diff_words->plus.text.ptr &&
784 *(diff_words->current_plus - 1) == '\n')) {
785 return 1;
786 } else {
787 return 0;
788 }
789 }
791 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
792 {
793 struct diff_words_data *diff_words = priv;
794 struct diff_words_style *style = diff_words->style;
795 int minus_first, minus_len, plus_first, plus_len;
796 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
797 struct diff_options *opt = diff_words->opt;
798 struct strbuf *msgbuf;
799 char *line_prefix = "";
801 if (line[0] != '@' || parse_hunk_header(line, len,
802 &minus_first, &minus_len, &plus_first, &plus_len))
803 return;
805 assert(opt);
806 if (opt->output_prefix) {
807 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
808 line_prefix = msgbuf->buf;
809 }
811 /* POSIX requires that first be decremented by one if len == 0... */
812 if (minus_len) {
813 minus_begin = diff_words->minus.orig[minus_first].begin;
814 minus_end =
815 diff_words->minus.orig[minus_first + minus_len - 1].end;
816 } else
817 minus_begin = minus_end =
818 diff_words->minus.orig[minus_first].end;
820 if (plus_len) {
821 plus_begin = diff_words->plus.orig[plus_first].begin;
822 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
823 } else
824 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
826 if (color_words_output_graph_prefix(diff_words)) {
827 fputs(line_prefix, diff_words->opt->file);
828 }
829 if (diff_words->current_plus != plus_begin) {
830 fn_out_diff_words_write_helper(diff_words->opt->file,
831 &style->ctx, style->newline,
832 plus_begin - diff_words->current_plus,
833 diff_words->current_plus, line_prefix);
834 if (*(plus_begin - 1) == '\n')
835 fputs(line_prefix, diff_words->opt->file);
836 }
837 if (minus_begin != minus_end) {
838 fn_out_diff_words_write_helper(diff_words->opt->file,
839 &style->old, style->newline,
840 minus_end - minus_begin, minus_begin,
841 line_prefix);
842 }
843 if (plus_begin != plus_end) {
844 fn_out_diff_words_write_helper(diff_words->opt->file,
845 &style->new, style->newline,
846 plus_end - plus_begin, plus_begin,
847 line_prefix);
848 }
850 diff_words->current_plus = plus_end;
851 diff_words->last_minus = minus_first;
852 }
854 /* This function starts looking at *begin, and returns 0 iff a word was found. */
855 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
856 int *begin, int *end)
857 {
858 if (word_regex && *begin < buffer->size) {
859 regmatch_t match[1];
860 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
861 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
862 '\n', match[0].rm_eo - match[0].rm_so);
863 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
864 *begin += match[0].rm_so;
865 return *begin >= *end;
866 }
867 return -1;
868 }
870 /* find the next word */
871 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
872 (*begin)++;
873 if (*begin >= buffer->size)
874 return -1;
876 /* find the end of the word */
877 *end = *begin + 1;
878 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
879 (*end)++;
881 return 0;
882 }
884 /*
885 * This function splits the words in buffer->text, stores the list with
886 * newline separator into out, and saves the offsets of the original words
887 * in buffer->orig.
888 */
889 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
890 regex_t *word_regex)
891 {
892 int i, j;
893 long alloc = 0;
895 out->size = 0;
896 out->ptr = NULL;
898 /* fake an empty "0th" word */
899 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
900 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
901 buffer->orig_nr = 1;
903 for (i = 0; i < buffer->text.size; i++) {
904 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
905 return;
907 /* store original boundaries */
908 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
909 buffer->orig_alloc);
910 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
911 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
912 buffer->orig_nr++;
914 /* store one word */
915 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
916 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
917 out->ptr[out->size + j - i] = '\n';
918 out->size += j - i + 1;
920 i = j - 1;
921 }
922 }
924 /* this executes the word diff on the accumulated buffers */
925 static void diff_words_show(struct diff_words_data *diff_words)
926 {
927 xpparam_t xpp;
928 xdemitconf_t xecfg;
929 mmfile_t minus, plus;
930 struct diff_words_style *style = diff_words->style;
932 struct diff_options *opt = diff_words->opt;
933 struct strbuf *msgbuf;
934 char *line_prefix = "";
936 assert(opt);
937 if (opt->output_prefix) {
938 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
939 line_prefix = msgbuf->buf;
940 }
942 /* special case: only removal */
943 if (!diff_words->plus.text.size) {
944 fputs(line_prefix, diff_words->opt->file);
945 fn_out_diff_words_write_helper(diff_words->opt->file,
946 &style->old, style->newline,
947 diff_words->minus.text.size,
948 diff_words->minus.text.ptr, line_prefix);
949 diff_words->minus.text.size = 0;
950 return;
951 }
953 diff_words->current_plus = diff_words->plus.text.ptr;
954 diff_words->last_minus = 0;
956 memset(&xpp, 0, sizeof(xpp));
957 memset(&xecfg, 0, sizeof(xecfg));
958 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
959 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
960 xpp.flags = 0;
961 /* as only the hunk header will be parsed, we need a 0-context */
962 xecfg.ctxlen = 0;
963 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
964 &xpp, &xecfg);
965 free(minus.ptr);
966 free(plus.ptr);
967 if (diff_words->current_plus != diff_words->plus.text.ptr +
968 diff_words->plus.text.size) {
969 if (color_words_output_graph_prefix(diff_words))
970 fputs(line_prefix, diff_words->opt->file);
971 fn_out_diff_words_write_helper(diff_words->opt->file,
972 &style->ctx, style->newline,
973 diff_words->plus.text.ptr + diff_words->plus.text.size
974 - diff_words->current_plus, diff_words->current_plus,
975 line_prefix);
976 }
977 diff_words->minus.text.size = diff_words->plus.text.size = 0;
978 }
980 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
981 static void diff_words_flush(struct emit_callback *ecbdata)
982 {
983 if (ecbdata->diff_words->minus.text.size ||
984 ecbdata->diff_words->plus.text.size)
985 diff_words_show(ecbdata->diff_words);
986 }
988 static void free_diff_words_data(struct emit_callback *ecbdata)
989 {
990 if (ecbdata->diff_words) {
991 diff_words_flush(ecbdata);
992 free (ecbdata->diff_words->minus.text.ptr);
993 free (ecbdata->diff_words->minus.orig);
994 free (ecbdata->diff_words->plus.text.ptr);
995 free (ecbdata->diff_words->plus.orig);
996 if (ecbdata->diff_words->word_regex) {
997 regfree(ecbdata->diff_words->word_regex);
998 free(ecbdata->diff_words->word_regex);
999 }
1000 free(ecbdata->diff_words);
1001 ecbdata->diff_words = NULL;
1002 }
1003 }
1005 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1006 {
1007 if (diff_use_color)
1008 return diff_colors[ix];
1009 return "";
1010 }
1012 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1013 {
1014 const char *cp;
1015 unsigned long allot;
1016 size_t l = len;
1018 if (ecb->truncate)
1019 return ecb->truncate(line, len);
1020 cp = line;
1021 allot = l;
1022 while (0 < l) {
1023 (void) utf8_width(&cp, &l);
1024 if (!cp)
1025 break; /* truncated in the middle? */
1026 }
1027 return allot - l;
1028 }
1030 static void find_lno(const char *line, struct emit_callback *ecbdata)
1031 {
1032 const char *p;
1033 ecbdata->lno_in_preimage = 0;
1034 ecbdata->lno_in_postimage = 0;
1035 p = strchr(line, '-');
1036 if (!p)
1037 return; /* cannot happen */
1038 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1039 p = strchr(p, '+');
1040 if (!p)
1041 return; /* cannot happen */
1042 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1043 }
1045 static void fn_out_consume(void *priv, char *line, unsigned long len)
1046 {
1047 struct emit_callback *ecbdata = priv;
1048 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1049 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1050 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1051 struct diff_options *o = ecbdata->opt;
1052 char *line_prefix = "";
1053 struct strbuf *msgbuf;
1055 if (o && o->output_prefix) {
1056 msgbuf = o->output_prefix(o, o->output_prefix_data);
1057 line_prefix = msgbuf->buf;
1058 }
1060 if (ecbdata->header) {
1061 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1062 strbuf_reset(ecbdata->header);
1063 ecbdata->header = NULL;
1064 }
1065 *(ecbdata->found_changesp) = 1;
1067 if (ecbdata->label_path[0]) {
1068 const char *name_a_tab, *name_b_tab;
1070 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1071 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1073 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1074 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1075 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1076 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1077 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1078 }
1080 if (diff_suppress_blank_empty
1081 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1082 line[0] = '\n';
1083 len = 1;
1084 }
1086 if (line[0] == '@') {
1087 if (ecbdata->diff_words)
1088 diff_words_flush(ecbdata);
1089 len = sane_truncate_line(ecbdata, line, len);
1090 find_lno(line, ecbdata);
1091 emit_hunk_header(ecbdata, line, len);
1092 if (line[len-1] != '\n')
1093 putc('\n', ecbdata->opt->file);
1094 return;
1095 }
1097 if (len < 1) {
1098 emit_line(ecbdata->opt, reset, reset, line, len);
1099 if (ecbdata->diff_words
1100 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1101 fputs("~\n", ecbdata->opt->file);
1102 return;
1103 }
1105 if (ecbdata->diff_words) {
1106 if (line[0] == '-') {
1107 diff_words_append(line, len,
1108 &ecbdata->diff_words->minus);
1109 return;
1110 } else if (line[0] == '+') {
1111 diff_words_append(line, len,
1112 &ecbdata->diff_words->plus);
1113 return;
1114 } else if (!prefixcmp(line, "\\ ")) {
1115 /*
1116 * Eat the "no newline at eof" marker as if we
1117 * saw a "+" or "-" line with nothing on it,
1118 * and return without diff_words_flush() to
1119 * defer processing. If this is the end of
1120 * preimage, more "+" lines may come after it.
1121 */
1122 return;
1123 }
1124 diff_words_flush(ecbdata);
1125 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1126 emit_line(ecbdata->opt, plain, reset, line, len);
1127 fputs("~\n", ecbdata->opt->file);
1128 } else {
1129 /*
1130 * Skip the prefix character, if any. With
1131 * diff_suppress_blank_empty, there may be
1132 * none.
1133 */
1134 if (line[0] != '\n') {
1135 line++;
1136 len--;
1137 }
1138 emit_line(ecbdata->opt, plain, reset, line, len);
1139 }
1140 return;
1141 }
1143 if (line[0] != '+') {
1144 const char *color =
1145 diff_get_color(ecbdata->color_diff,
1146 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1147 ecbdata->lno_in_preimage++;
1148 if (line[0] == ' ')
1149 ecbdata->lno_in_postimage++;
1150 emit_line(ecbdata->opt, color, reset, line, len);
1151 } else {
1152 ecbdata->lno_in_postimage++;
1153 emit_add_line(reset, ecbdata, line + 1, len - 1);
1154 }
1155 }
1157 static char *pprint_rename(const char *a, const char *b)
1158 {
1159 const char *old = a;
1160 const char *new = b;
1161 struct strbuf name = STRBUF_INIT;
1162 int pfx_length, sfx_length;
1163 int len_a = strlen(a);
1164 int len_b = strlen(b);
1165 int a_midlen, b_midlen;
1166 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1167 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1169 if (qlen_a || qlen_b) {
1170 quote_c_style(a, &name, NULL, 0);
1171 strbuf_addstr(&name, " => ");
1172 quote_c_style(b, &name, NULL, 0);
1173 return strbuf_detach(&name, NULL);
1174 }
1176 /* Find common prefix */
1177 pfx_length = 0;
1178 while (*old && *new && *old == *new) {
1179 if (*old == '/')
1180 pfx_length = old - a + 1;
1181 old++;
1182 new++;
1183 }
1185 /* Find common suffix */
1186 old = a + len_a;
1187 new = b + len_b;
1188 sfx_length = 0;
1189 while (a <= old && b <= new && *old == *new) {
1190 if (*old == '/')
1191 sfx_length = len_a - (old - a);
1192 old--;
1193 new--;
1194 }
1196 /*
1197 * pfx{mid-a => mid-b}sfx
1198 * {pfx-a => pfx-b}sfx
1199 * pfx{sfx-a => sfx-b}
1200 * name-a => name-b
1201 */
1202 a_midlen = len_a - pfx_length - sfx_length;
1203 b_midlen = len_b - pfx_length - sfx_length;
1204 if (a_midlen < 0)
1205 a_midlen = 0;
1206 if (b_midlen < 0)
1207 b_midlen = 0;
1209 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1210 if (pfx_length + sfx_length) {
1211 strbuf_add(&name, a, pfx_length);
1212 strbuf_addch(&name, '{');
1213 }
1214 strbuf_add(&name, a + pfx_length, a_midlen);
1215 strbuf_addstr(&name, " => ");
1216 strbuf_add(&name, b + pfx_length, b_midlen);
1217 if (pfx_length + sfx_length) {
1218 strbuf_addch(&name, '}');
1219 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1220 }
1221 return strbuf_detach(&name, NULL);
1222 }
1224 struct diffstat_t {
1225 int nr;
1226 int alloc;
1227 struct diffstat_file {
1228 char *from_name;
1229 char *name;
1230 char *print_name;
1231 unsigned is_unmerged:1;
1232 unsigned is_binary:1;
1233 unsigned is_renamed:1;
1234 uintmax_t added, deleted;
1235 } **files;
1236 };
1238 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1239 const char *name_a,
1240 const char *name_b)
1241 {
1242 struct diffstat_file *x;
1243 x = xcalloc(sizeof (*x), 1);
1244 if (diffstat->nr == diffstat->alloc) {
1245 diffstat->alloc = alloc_nr(diffstat->alloc);
1246 diffstat->files = xrealloc(diffstat->files,
1247 diffstat->alloc * sizeof(x));
1248 }
1249 diffstat->files[diffstat->nr++] = x;
1250 if (name_b) {
1251 x->from_name = xstrdup(name_a);
1252 x->name = xstrdup(name_b);
1253 x->is_renamed = 1;
1254 }
1255 else {
1256 x->from_name = NULL;
1257 x->name = xstrdup(name_a);
1258 }
1259 return x;
1260 }
1262 static void diffstat_consume(void *priv, char *line, unsigned long len)
1263 {
1264 struct diffstat_t *diffstat = priv;
1265 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1267 if (line[0] == '+')
1268 x->added++;
1269 else if (line[0] == '-')
1270 x->deleted++;
1271 }
1273 const char mime_boundary_leader[] = "------------";
1275 static int scale_linear(int it, int width, int max_change)
1276 {
1277 /*
1278 * make sure that at least one '-' is printed if there were deletions,
1279 * and likewise for '+'.
1280 */
1281 if (max_change < 2)
1282 return it;
1283 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1284 }
1286 static void show_name(FILE *file,
1287 const char *prefix, const char *name, int len)
1288 {
1289 fprintf(file, " %s%-*s |", prefix, len, name);
1290 }
1292 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1293 {
1294 if (cnt <= 0)
1295 return;
1296 fprintf(file, "%s", set);
1297 while (cnt--)
1298 putc(ch, file);
1299 fprintf(file, "%s", reset);
1300 }
1302 static void fill_print_name(struct diffstat_file *file)
1303 {
1304 char *pname;
1306 if (file->print_name)
1307 return;
1309 if (!file->is_renamed) {
1310 struct strbuf buf = STRBUF_INIT;
1311 if (quote_c_style(file->name, &buf, NULL, 0)) {
1312 pname = strbuf_detach(&buf, NULL);
1313 } else {
1314 pname = file->name;
1315 strbuf_release(&buf);
1316 }
1317 } else {
1318 pname = pprint_rename(file->from_name, file->name);
1319 }
1320 file->print_name = pname;
1321 }
1323 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1324 {
1325 int i, len, add, del, adds = 0, dels = 0;
1326 uintmax_t max_change = 0, max_len = 0;
1327 int total_files = data->nr;
1328 int width, name_width;
1329 const char *reset, *add_c, *del_c;
1330 const char *line_prefix = "";
1331 struct strbuf *msg = NULL;
1333 if (data->nr == 0)
1334 return;
1336 if (options->output_prefix) {
1337 msg = options->output_prefix(options, options->output_prefix_data);
1338 line_prefix = msg->buf;
1339 }
1341 width = options->stat_width ? options->stat_width : 80;
1342 name_width = options->stat_name_width ? options->stat_name_width : 50;
1344 /* Sanity: give at least 5 columns to the graph,
1345 * but leave at least 10 columns for the name.
1346 */
1347 if (width < 25)
1348 width = 25;
1349 if (name_width < 10)
1350 name_width = 10;
1351 else if (width < name_width + 15)
1352 name_width = width - 15;
1354 /* Find the longest filename and max number of changes */
1355 reset = diff_get_color_opt(options, DIFF_RESET);
1356 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1357 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1359 for (i = 0; i < data->nr; i++) {
1360 struct diffstat_file *file = data->files[i];
1361 uintmax_t change = file->added + file->deleted;
1362 fill_print_name(file);
1363 len = strlen(file->print_name);
1364 if (max_len < len)
1365 max_len = len;
1367 if (file->is_binary || file->is_unmerged)
1368 continue;
1369 if (max_change < change)
1370 max_change = change;
1371 }
1373 /* Compute the width of the graph part;
1374 * 10 is for one blank at the beginning of the line plus
1375 * " | count " between the name and the graph.
1376 *
1377 * From here on, name_width is the width of the name area,
1378 * and width is the width of the graph area.
1379 */
1380 name_width = (name_width < max_len) ? name_width : max_len;
1381 if (width < (name_width + 10) + max_change)
1382 width = width - (name_width + 10);
1383 else
1384 width = max_change;
1386 for (i = 0; i < data->nr; i++) {
1387 const char *prefix = "";
1388 char *name = data->files[i]->print_name;
1389 uintmax_t added = data->files[i]->added;
1390 uintmax_t deleted = data->files[i]->deleted;
1391 int name_len;
1393 /*
1394 * "scale" the filename
1395 */
1396 len = name_width;
1397 name_len = strlen(name);
1398 if (name_width < name_len) {
1399 char *slash;
1400 prefix = "...";
1401 len -= 3;
1402 name += name_len - len;
1403 slash = strchr(name, '/');
1404 if (slash)
1405 name = slash;
1406 }
1408 if (data->files[i]->is_binary) {
1409 fprintf(options->file, "%s", line_prefix);
1410 show_name(options->file, prefix, name, len);
1411 fprintf(options->file, " Bin ");
1412 fprintf(options->file, "%s%"PRIuMAX"%s",
1413 del_c, deleted, reset);
1414 fprintf(options->file, " -> ");
1415 fprintf(options->file, "%s%"PRIuMAX"%s",
1416 add_c, added, reset);
1417 fprintf(options->file, " bytes");
1418 fprintf(options->file, "\n");
1419 continue;
1420 }
1421 else if (data->files[i]->is_unmerged) {
1422 fprintf(options->file, "%s", line_prefix);
1423 show_name(options->file, prefix, name, len);
1424 fprintf(options->file, " Unmerged\n");
1425 continue;
1426 }
1427 else if (!data->files[i]->is_renamed &&
1428 (added + deleted == 0)) {
1429 total_files--;
1430 continue;
1431 }
1433 /*
1434 * scale the add/delete
1435 */
1436 add = added;
1437 del = deleted;
1438 adds += add;
1439 dels += del;
1441 if (width <= max_change) {
1442 add = scale_linear(add, width, max_change);
1443 del = scale_linear(del, width, max_change);
1444 }
1445 fprintf(options->file, "%s", line_prefix);
1446 show_name(options->file, prefix, name, len);
1447 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1448 added + deleted ? " " : "");
1449 show_graph(options->file, '+', add, add_c, reset);
1450 show_graph(options->file, '-', del, del_c, reset);
1451 fprintf(options->file, "\n");
1452 }
1453 fprintf(options->file, "%s", line_prefix);
1454 fprintf(options->file,
1455 " %d files changed, %d insertions(+), %d deletions(-)\n",
1456 total_files, adds, dels);
1457 }
1459 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1460 {
1461 int i, adds = 0, dels = 0, total_files = data->nr;
1463 if (data->nr == 0)
1464 return;
1466 for (i = 0; i < data->nr; i++) {
1467 if (!data->files[i]->is_binary &&
1468 !data->files[i]->is_unmerged) {
1469 int added = data->files[i]->added;
1470 int deleted= data->files[i]->deleted;
1471 if (!data->files[i]->is_renamed &&
1472 (added + deleted == 0)) {
1473 total_files--;
1474 } else {
1475 adds += added;
1476 dels += deleted;
1477 }
1478 }
1479 }
1480 if (options->output_prefix) {
1481 struct strbuf *msg = NULL;
1482 msg = options->output_prefix(options,
1483 options->output_prefix_data);
1484 fprintf(options->file, "%s", msg->buf);
1485 }
1486 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1487 total_files, adds, dels);
1488 }
1490 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1491 {
1492 int i;
1494 if (data->nr == 0)
1495 return;
1497 for (i = 0; i < data->nr; i++) {
1498 struct diffstat_file *file = data->files[i];
1500 if (options->output_prefix) {
1501 struct strbuf *msg = NULL;
1502 msg = options->output_prefix(options,
1503 options->output_prefix_data);
1504 fprintf(options->file, "%s", msg->buf);
1505 }
1507 if (file->is_binary)
1508 fprintf(options->file, "-\t-\t");
1509 else
1510 fprintf(options->file,
1511 "%"PRIuMAX"\t%"PRIuMAX"\t",
1512 file->added, file->deleted);
1513 if (options->line_termination) {
1514 fill_print_name(file);
1515 if (!file->is_renamed)
1516 write_name_quoted(file->name, options->file,
1517 options->line_termination);
1518 else {
1519 fputs(file->print_name, options->file);
1520 putc(options->line_termination, options->file);
1521 }
1522 } else {
1523 if (file->is_renamed) {
1524 putc('\0', options->file);
1525 write_name_quoted(file->from_name, options->file, '\0');
1526 }
1527 write_name_quoted(file->name, options->file, '\0');
1528 }
1529 }
1530 }
1532 struct dirstat_file {
1533 const char *name;
1534 unsigned long changed;
1535 };
1537 struct dirstat_dir {
1538 struct dirstat_file *files;
1539 int alloc, nr, permille, cumulative;
1540 };
1542 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1543 unsigned long changed, const char *base, int baselen)
1544 {
1545 unsigned long this_dir = 0;
1546 unsigned int sources = 0;
1547 const char *line_prefix = "";
1548 struct strbuf *msg = NULL;
1550 if (opt->output_prefix) {
1551 msg = opt->output_prefix(opt, opt->output_prefix_data);
1552 line_prefix = msg->buf;
1553 }
1555 while (dir->nr) {
1556 struct dirstat_file *f = dir->files;
1557 int namelen = strlen(f->name);
1558 unsigned long this;
1559 char *slash;
1561 if (namelen < baselen)
1562 break;
1563 if (memcmp(f->name, base, baselen))
1564 break;
1565 slash = strchr(f->name + baselen, '/');
1566 if (slash) {
1567 int newbaselen = slash + 1 - f->name;
1568 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1569 sources++;
1570 } else {
1571 this = f->changed;
1572 dir->files++;
1573 dir->nr--;
1574 sources += 2;
1575 }
1576 this_dir += this;
1577 }
1579 /*
1580 * We don't report dirstat's for
1581 * - the top level
1582 * - or cases where everything came from a single directory
1583 * under this directory (sources == 1).
1584 */
1585 if (baselen && sources != 1) {
1586 if (this_dir) {
1587 int permille = this_dir * 1000 / changed;
1588 if (permille >= dir->permille) {
1589 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1590 permille / 10, permille % 10, baselen, base);
1591 if (!dir->cumulative)
1592 return 0;
1593 }
1594 }
1595 }
1596 return this_dir;
1597 }
1599 static int dirstat_compare(const void *_a, const void *_b)
1600 {
1601 const struct dirstat_file *a = _a;
1602 const struct dirstat_file *b = _b;
1603 return strcmp(a->name, b->name);
1604 }
1606 static void show_dirstat(struct diff_options *options)
1607 {
1608 int i;
1609 unsigned long changed;
1610 struct dirstat_dir dir;
1611 struct diff_queue_struct *q = &diff_queued_diff;
1613 dir.files = NULL;
1614 dir.alloc = 0;
1615 dir.nr = 0;
1616 dir.permille = options->dirstat_permille;
1617 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1619 changed = 0;
1620 for (i = 0; i < q->nr; i++) {
1621 struct diff_filepair *p = q->queue[i];
1622 const char *name;
1623 unsigned long copied, added, damage;
1624 int content_changed;
1626 name = p->two->path ? p->two->path : p->one->path;
1628 if (p->one->sha1_valid && p->two->sha1_valid)
1629 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1630 else
1631 content_changed = 1;
1633 if (!content_changed) {
1634 /*
1635 * The SHA1 has not changed, so pre-/post-content is
1636 * identical. We can therefore skip looking at the
1637 * file contents altogether.
1638 */
1639 damage = 0;
1640 goto found_damage;
1641 }
1643 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1644 /*
1645 * In --dirstat-by-file mode, we don't really need to
1646 * look at the actual file contents at all.
1647 * The fact that the SHA1 changed is enough for us to
1648 * add this file to the list of results
1649 * (with each file contributing equal damage).
1650 */
1651 damage = 1;
1652 goto found_damage;
1653 }
1655 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1656 diff_populate_filespec(p->one, 0);
1657 diff_populate_filespec(p->two, 0);
1658 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1659 &copied, &added);
1660 diff_free_filespec_data(p->one);
1661 diff_free_filespec_data(p->two);
1662 } else if (DIFF_FILE_VALID(p->one)) {
1663 diff_populate_filespec(p->one, 1);
1664 copied = added = 0;
1665 diff_free_filespec_data(p->one);
1666 } else if (DIFF_FILE_VALID(p->two)) {
1667 diff_populate_filespec(p->two, 1);
1668 copied = 0;
1669 added = p->two->size;
1670 diff_free_filespec_data(p->two);
1671 } else
1672 continue;
1674 /*
1675 * Original minus copied is the removed material,
1676 * added is the new material. They are both damages
1677 * made to the preimage.
1678 * If the resulting damage is zero, we know that
1679 * diffcore_count_changes() considers the two entries to
1680 * be identical, but since content_changed is true, we
1681 * know that there must have been _some_ kind of change,
1682 * so we force all entries to have damage > 0.
1683 */
1684 damage = (p->one->size - copied) + added;
1685 if (!damage)
1686 damage = 1;
1688 found_damage:
1689 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1690 dir.files[dir.nr].name = name;
1691 dir.files[dir.nr].changed = damage;
1692 changed += damage;
1693 dir.nr++;
1694 }
1696 /* This can happen even with many files, if everything was renames */
1697 if (!changed)
1698 return;
1700 /* Show all directories with more than x% of the changes */
1701 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1702 gather_dirstat(options, &dir, changed, "", 0);
1703 }
1705 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1706 {
1707 int i;
1708 unsigned long changed;
1709 struct dirstat_dir dir;
1711 if (data->nr == 0)
1712 return;
1714 dir.files = NULL;
1715 dir.alloc = 0;
1716 dir.nr = 0;
1717 dir.permille = options->dirstat_permille;
1718 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1720 changed = 0;
1721 for (i = 0; i < data->nr; i++) {
1722 struct diffstat_file *file = data->files[i];
1723 unsigned long damage = file->added + file->deleted;
1724 if (file->is_binary)
1725 /*
1726 * binary files counts bytes, not lines. Must find some
1727 * way to normalize binary bytes vs. textual lines.
1728 * The following heuristic assumes that there are 64
1729 * bytes per "line".
1730 * This is stupid and ugly, but very cheap...
1731 */
1732 damage = (damage + 63) / 64;
1733 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1734 dir.files[dir.nr].name = file->name;
1735 dir.files[dir.nr].changed = damage;
1736 changed += damage;
1737 dir.nr++;
1738 }
1740 /* This can happen even with many files, if everything was renames */
1741 if (!changed)
1742 return;
1744 /* Show all directories with more than x% of the changes */
1745 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1746 gather_dirstat(options, &dir, changed, "", 0);
1747 }
1749 static void free_diffstat_info(struct diffstat_t *diffstat)
1750 {
1751 int i;
1752 for (i = 0; i < diffstat->nr; i++) {
1753 struct diffstat_file *f = diffstat->files[i];
1754 if (f->name != f->print_name)
1755 free(f->print_name);
1756 free(f->name);
1757 free(f->from_name);
1758 free(f);
1759 }
1760 free(diffstat->files);
1761 }
1763 struct checkdiff_t {
1764 const char *filename;
1765 int lineno;
1766 int conflict_marker_size;
1767 struct diff_options *o;
1768 unsigned ws_rule;
1769 unsigned status;
1770 };
1772 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1773 {
1774 char firstchar;
1775 int cnt;
1777 if (len < marker_size + 1)
1778 return 0;
1779 firstchar = line[0];
1780 switch (firstchar) {
1781 case '=': case '>': case '<': case '|':
1782 break;
1783 default:
1784 return 0;
1785 }
1786 for (cnt = 1; cnt < marker_size; cnt++)
1787 if (line[cnt] != firstchar)
1788 return 0;
1789 /* line[1] thru line[marker_size-1] are same as firstchar */
1790 if (len < marker_size + 1 || !isspace(line[marker_size]))
1791 return 0;
1792 return 1;
1793 }
1795 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1796 {
1797 struct checkdiff_t *data = priv;
1798 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1799 int marker_size = data->conflict_marker_size;
1800 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1801 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1802 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1803 char *err;
1804 char *line_prefix = "";
1805 struct strbuf *msgbuf;
1807 assert(data->o);
1808 if (data->o->output_prefix) {
1809 msgbuf = data->o->output_prefix(data->o,
1810 data->o->output_prefix_data);
1811 line_prefix = msgbuf->buf;
1812 }
1814 if (line[0] == '+') {
1815 unsigned bad;
1816 data->lineno++;
1817 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1818 data->status |= 1;
1819 fprintf(data->o->file,
1820 "%s%s:%d: leftover conflict marker\n",
1821 line_prefix, data->filename, data->lineno);
1822 }
1823 bad = ws_check(line + 1, len - 1, data->ws_rule);
1824 if (!bad)
1825 return;
1826 data->status |= bad;
1827 err = whitespace_error_string(bad);
1828 fprintf(data->o->file, "%s%s:%d: %s.\n",
1829 line_prefix, data->filename, data->lineno, err);
1830 free(err);
1831 emit_line(data->o, set, reset, line, 1);
1832 ws_check_emit(line + 1, len - 1, data->ws_rule,
1833 data->o->file, set, reset, ws);
1834 } else if (line[0] == ' ') {
1835 data->lineno++;
1836 } else if (line[0] == '@') {
1837 char *plus = strchr(line, '+');
1838 if (plus)
1839 data->lineno = strtol(plus, NULL, 10) - 1;
1840 else
1841 die("invalid diff");
1842 }
1843 }
1845 static unsigned char *deflate_it(char *data,
1846 unsigned long size,
1847 unsigned long *result_size)
1848 {
1849 int bound;
1850 unsigned char *deflated;
1851 git_zstream stream;
1853 memset(&stream, 0, sizeof(stream));
1854 git_deflate_init(&stream, zlib_compression_level);
1855 bound = git_deflate_bound(&stream, size);
1856 deflated = xmalloc(bound);
1857 stream.next_out = deflated;
1858 stream.avail_out = bound;
1860 stream.next_in = (unsigned char *)data;
1861 stream.avail_in = size;
1862 while (git_deflate(&stream, Z_FINISH) == Z_OK)
1863 ; /* nothing */
1864 git_deflate_end(&stream);
1865 *result_size = stream.total_out;
1866 return deflated;
1867 }
1869 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1870 {
1871 void *cp;
1872 void *delta;
1873 void *deflated;
1874 void *data;
1875 unsigned long orig_size;
1876 unsigned long delta_size;
1877 unsigned long deflate_size;
1878 unsigned long data_size;
1880 /* We could do deflated delta, or we could do just deflated two,
1881 * whichever is smaller.
1882 */
1883 delta = NULL;
1884 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1885 if (one->size && two->size) {
1886 delta = diff_delta(one->ptr, one->size,
1887 two->ptr, two->size,
1888 &delta_size, deflate_size);
1889 if (delta) {
1890 void *to_free = delta;
1891 orig_size = delta_size;
1892 delta = deflate_it(delta, delta_size, &delta_size);
1893 free(to_free);
1894 }
1895 }
1897 if (delta && delta_size < deflate_size) {
1898 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1899 free(deflated);
1900 data = delta;
1901 data_size = delta_size;
1902 }
1903 else {
1904 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1905 free(delta);
1906 data = deflated;
1907 data_size = deflate_size;
1908 }
1910 /* emit data encoded in base85 */
1911 cp = data;
1912 while (data_size) {
1913 int bytes = (52 < data_size) ? 52 : data_size;
1914 char line[70];
1915 data_size -= bytes;
1916 if (bytes <= 26)
1917 line[0] = bytes + 'A' - 1;
1918 else
1919 line[0] = bytes - 26 + 'a' - 1;
1920 encode_85(line + 1, cp, bytes);
1921 cp = (char *) cp + bytes;
1922 fprintf(file, "%s", prefix);
1923 fputs(line, file);
1924 fputc('\n', file);
1925 }
1926 fprintf(file, "%s\n", prefix);
1927 free(data);
1928 }
1930 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1931 {
1932 fprintf(file, "%sGIT binary patch\n", prefix);
1933 emit_binary_diff_body(file, one, two, prefix);
1934 emit_binary_diff_body(file, two, one, prefix);
1935 }
1937 static void diff_filespec_load_driver(struct diff_filespec *one)
1938 {
1939 /* Use already-loaded driver */
1940 if (one->driver)
1941 return;
1943 if (S_ISREG(one->mode))
1944 one->driver = userdiff_find_by_path(one->path);
1946 /* Fallback to default settings */
1947 if (!one->driver)
1948 one->driver = userdiff_find_by_name("default");
1949 }
1951 int diff_filespec_is_binary(struct diff_filespec *one)
1952 {
1953 if (one->is_binary == -1) {
1954 diff_filespec_load_driver(one);
1955 if (one->driver->binary != -1)
1956 one->is_binary = one->driver->binary;
1957 else {
1958 if (!one->data && DIFF_FILE_VALID(one))
1959 diff_populate_filespec(one, 0);
1960 if (one->data)
1961 one->is_binary = buffer_is_binary(one->data,
1962 one->size);
1963 if (one->is_binary == -1)
1964 one->is_binary = 0;
1965 }
1966 }
1967 return one->is_binary;
1968 }
1970 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1971 {
1972 diff_filespec_load_driver(one);
1973 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1974 }
1976 static const char *userdiff_word_regex(struct diff_filespec *one)
1977 {
1978 diff_filespec_load_driver(one);
1979 return one->driver->word_regex;
1980 }
1982 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1983 {
1984 if (!options->a_prefix)
1985 options->a_prefix = a;
1986 if (!options->b_prefix)
1987 options->b_prefix = b;
1988 }
1990 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1991 {
1992 if (!DIFF_FILE_VALID(one))
1993 return NULL;
1995 diff_filespec_load_driver(one);
1996 return userdiff_get_textconv(one->driver);
1997 }
1999 static void builtin_diff(const char *name_a,
2000 const char *name_b,
2001 struct diff_filespec *one,
2002 struct diff_filespec *two,
2003 const char *xfrm_msg,
2004 int must_show_header,
2005 struct diff_options *o,
2006 int complete_rewrite)
2007 {
2008 mmfile_t mf1, mf2;
2009 const char *lbl[2];
2010 char *a_one, *b_two;
2011 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
2012 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2013 const char *a_prefix, *b_prefix;
2014 struct userdiff_driver *textconv_one = NULL;
2015 struct userdiff_driver *textconv_two = NULL;
2016 struct strbuf header = STRBUF_INIT;
2017 struct strbuf *msgbuf;
2018 char *line_prefix = "";
2020 if (o->output_prefix) {
2021 msgbuf = o->output_prefix(o, o->output_prefix_data);
2022 line_prefix = msgbuf->buf;
2023 }
2025 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2026 (!one->mode || S_ISGITLINK(one->mode)) &&
2027 (!two->mode || S_ISGITLINK(two->mode))) {
2028 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2029 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2030 show_submodule_summary(o->file, one ? one->path : two->path,
2031 one->sha1, two->sha1, two->dirty_submodule,
2032 del, add, reset);
2033 return;
2034 }
2036 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2037 textconv_one = get_textconv(one);
2038 textconv_two = get_textconv(two);
2039 }
2041 diff_set_mnemonic_prefix(o, "a/", "b/");
2042 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2043 a_prefix = o->b_prefix;
2044 b_prefix = o->a_prefix;
2045 } else {
2046 a_prefix = o->a_prefix;
2047 b_prefix = o->b_prefix;
2048 }
2050 /* Never use a non-valid filename anywhere if at all possible */
2051 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2052 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2054 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2055 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2056 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2057 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2058 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2059 if (lbl[0][0] == '/') {
2060 /* /dev/null */
2061 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2062 if (xfrm_msg)
2063 strbuf_addstr(&header, xfrm_msg);
2064 must_show_header = 1;
2065 }
2066 else if (lbl[1][0] == '/') {
2067 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2068 if (xfrm_msg)
2069 strbuf_addstr(&header, xfrm_msg);
2070 must_show_header = 1;
2071 }
2072 else {
2073 if (one->mode != two->mode) {
2074 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2075 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2076 must_show_header = 1;
2077 }
2078 if (xfrm_msg)
2079 strbuf_addstr(&header, xfrm_msg);
2081 /*
2082 * we do not run diff between different kind
2083 * of objects.
2084 */
2085 if ((one->mode ^ two->mode) & S_IFMT)
2086 goto free_ab_and_return;
2087 if (complete_rewrite &&
2088 (textconv_one || !diff_filespec_is_binary(one)) &&
2089 (textconv_two || !diff_filespec_is_binary(two))) {
2090 fprintf(o->file, "%s", header.buf);
2091 strbuf_reset(&header);
2092 emit_rewrite_diff(name_a, name_b, one, two,
2093 textconv_one, textconv_two, o);
2094 o->found_changes = 1;
2095 goto free_ab_and_return;
2096 }
2097 }
2099 if (o->irreversible_delete && lbl[1][0] == '/') {
2100 fprintf(o->file, "%s", header.buf);
2101 strbuf_reset(&header);
2102 goto free_ab_and_return;
2103 } else if (!DIFF_OPT_TST(o, TEXT) &&
2104 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2105 (!textconv_two && diff_filespec_is_binary(two)) )) {
2106 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2107 die("unable to read files to diff");
2108 /* Quite common confusing case */
2109 if (mf1.size == mf2.size &&
2110 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2111 if (must_show_header)
2112 fprintf(o->file, "%s", header.buf);
2113 goto free_ab_and_return;
2114 }
2115 fprintf(o->file, "%s", header.buf);
2116 strbuf_reset(&header);
2117 if (DIFF_OPT_TST(o, BINARY))
2118 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2119 else
2120 fprintf(o->file, "%sBinary files %s and %s differ\n",
2121 line_prefix, lbl[0], lbl[1]);
2122 o->found_changes = 1;
2123 } else {
2124 /* Crazy xdl interfaces.. */
2125 const char *diffopts = getenv("GIT_DIFF_OPTS");
2126 xpparam_t xpp;
2127 xdemitconf_t xecfg;
2128 struct emit_callback ecbdata;
2129 const struct userdiff_funcname *pe;
2131 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2132 fprintf(o->file, "%s", header.buf);
2133 strbuf_reset(&header);
2134 }
2136 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2137 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2139 pe = diff_funcname_pattern(one);
2140 if (!pe)
2141 pe = diff_funcname_pattern(two);
2143 memset(&xpp, 0, sizeof(xpp));
2144 memset(&xecfg, 0, sizeof(xecfg));
2145 memset(&ecbdata, 0, sizeof(ecbdata));
2146 ecbdata.label_path = lbl;
2147 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2148 ecbdata.found_changesp = &o->found_changes;
2149 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2150 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2151 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2152 ecbdata.opt = o;
2153 ecbdata.header = header.len ? &header : NULL;
2154 xpp.flags = o->xdl_opts;
2155 xecfg.ctxlen = o->context;
2156 xecfg.interhunkctxlen = o->interhunkcontext;
2157 xecfg.flags = XDL_EMIT_FUNCNAMES;
2158 if (pe)
2159 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2160 if (!diffopts)
2161 ;
2162 else if (!prefixcmp(diffopts, "--unified="))
2163 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2164 else if (!prefixcmp(diffopts, "-u"))
2165 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2166 if (o->word_diff) {
2167 int i;
2169 ecbdata.diff_words =
2170 xcalloc(1, sizeof(struct diff_words_data));
2171 ecbdata.diff_words->type = o->word_diff;
2172 ecbdata.diff_words->opt = o;
2173 if (!o->word_regex)
2174 o->word_regex = userdiff_word_regex(one);
2175 if (!o->word_regex)
2176 o->word_regex = userdiff_word_regex(two);
2177 if (!o->word_regex)
2178 o->word_regex = diff_word_regex_cfg;
2179 if (o->word_regex) {
2180 ecbdata.diff_words->word_regex = (regex_t *)
2181 xmalloc(sizeof(regex_t));
2182 if (regcomp(ecbdata.diff_words->word_regex,
2183 o->word_regex,
2184 REG_EXTENDED | REG_NEWLINE))
2185 die ("Invalid regular expression: %s",
2186 o->word_regex);
2187 }
2188 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2189 if (o->word_diff == diff_words_styles[i].type) {
2190 ecbdata.diff_words->style =
2191 &diff_words_styles[i];
2192 break;
2193 }
2194 }
2195 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2196 struct diff_words_style *st = ecbdata.diff_words->style;
2197 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2198 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2199 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2200 }
2201 }
2202 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2203 &xpp, &xecfg);
2204 if (o->word_diff)
2205 free_diff_words_data(&ecbdata);
2206 if (textconv_one)
2207 free(mf1.ptr);
2208 if (textconv_two)
2209 free(mf2.ptr);
2210 xdiff_clear_find_func(&xecfg);
2211 }
2213 free_ab_and_return:
2214 strbuf_release(&header);
2215 diff_free_filespec_data(one);
2216 diff_free_filespec_data(two);
2217 free(a_one);
2218 free(b_two);
2219 return;
2220 }
2222 static void builtin_diffstat(const char *name_a, const char *name_b,
2223 struct diff_filespec *one,
2224 struct diff_filespec *two,
2225 struct diffstat_t *diffstat,
2226 struct diff_options *o,
2227 int complete_rewrite)
2228 {
2229 mmfile_t mf1, mf2;
2230 struct diffstat_file *data;
2232 data = diffstat_add(diffstat, name_a, name_b);
2234 if (!one || !two) {
2235 data->is_unmerged = 1;
2236 return;
2237 }
2239 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2240 data->is_binary = 1;
2241 data->added = diff_filespec_size(two);
2242 data->deleted = diff_filespec_size(one);
2243 }
2245 else if (complete_rewrite) {
2246 diff_populate_filespec(one, 0);
2247 diff_populate_filespec(two, 0);
2248 data->deleted = count_lines(one->data, one->size);
2249 data->added = count_lines(two->data, two->size);
2250 }
2252 else {
2253 /* Crazy xdl interfaces.. */
2254 xpparam_t xpp;
2255 xdemitconf_t xecfg;
2257 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2258 die("unable to read files to diff");
2260 memset(&xpp, 0, sizeof(xpp));
2261 memset(&xecfg, 0, sizeof(xecfg));
2262 xpp.flags = o->xdl_opts;
2263 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2264 &xpp, &xecfg);
2265 }
2267 diff_free_filespec_data(one);
2268 diff_free_filespec_data(two);
2269 }
2271 static void builtin_checkdiff(const char *name_a, const char *name_b,
2272 const char *attr_path,
2273 struct diff_filespec *one,
2274 struct diff_filespec *two,
2275 struct diff_options *o)
2276 {
2277 mmfile_t mf1, mf2;
2278 struct checkdiff_t data;
2280 if (!two)
2281 return;
2283 memset(&data, 0, sizeof(data));
2284 data.filename = name_b ? name_b : name_a;
2285 data.lineno = 0;
2286 data.o = o;
2287 data.ws_rule = whitespace_rule(attr_path);
2288 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2290 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2291 die("unable to read files to diff");
2293 /*
2294 * All the other codepaths check both sides, but not checking
2295 * the "old" side here is deliberate. We are checking the newly
2296 * introduced changes, and as long as the "new" side is text, we
2297 * can and should check what it introduces.
2298 */
2299 if (diff_filespec_is_binary(two))
2300 goto free_and_return;
2301 else {
2302 /* Crazy xdl interfaces.. */
2303 xpparam_t xpp;
2304 xdemitconf_t xecfg;
2306 memset(&xpp, 0, sizeof(xpp));
2307 memset(&xecfg, 0, sizeof(xecfg));
2308 xecfg.ctxlen = 1; /* at least one context line */
2309 xpp.flags = 0;
2310 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2311 &xpp, &xecfg);
2313 if (data.ws_rule & WS_BLANK_AT_EOF) {
2314 struct emit_callback ecbdata;
2315 int blank_at_eof;
2317 ecbdata.ws_rule = data.ws_rule;
2318 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2319 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2321 if (blank_at_eof) {
2322 static char *err;
2323 if (!err)
2324 err = whitespace_error_string(WS_BLANK_AT_EOF);
2325 fprintf(o->file, "%s:%d: %s.\n",
2326 data.filename, blank_at_eof, err);
2327 data.status = 1; /* report errors */
2328 }
2329 }
2330 }
2331 free_and_return:
2332 diff_free_filespec_data(one);
2333 diff_free_filespec_data(two);
2334 if (data.status)
2335 DIFF_OPT_SET(o, CHECK_FAILED);
2336 }
2338 struct diff_filespec *alloc_filespec(const char *path)
2339 {
2340 int namelen = strlen(path);
2341 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2343 memset(spec, 0, sizeof(*spec));
2344 spec->path = (char *)(spec + 1);
2345 memcpy(spec->path, path, namelen+1);
2346 spec->count = 1;
2347 spec->is_binary = -1;
2348 return spec;
2349 }
2351 void free_filespec(struct diff_filespec *spec)
2352 {
2353 if (!--spec->count) {
2354 diff_free_filespec_data(spec);
2355 free(spec);
2356 }
2357 }
2359 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2360 unsigned short mode)
2361 {
2362 if (mode) {
2363 spec->mode = canon_mode(mode);
2364 hashcpy(spec->sha1, sha1);
2365 spec->sha1_valid = !is_null_sha1(sha1);
2366 }
2367 }
2369 /*
2370 * Given a name and sha1 pair, if the index tells us the file in
2371 * the work tree has that object contents, return true, so that
2372 * prepare_temp_file() does not have to inflate and extract.
2373 */
2374 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2375 {
2376 struct cache_entry *ce;
2377 struct stat st;
2378 int pos, len;
2380 /*
2381 * We do not read the cache ourselves here, because the
2382 * benchmark with my previous version that always reads cache
2383 * shows that it makes things worse for diff-tree comparing
2384 * two linux-2.6 kernel trees in an already checked out work
2385 * tree. This is because most diff-tree comparisons deal with
2386 * only a small number of files, while reading the cache is
2387 * expensive for a large project, and its cost outweighs the
2388 * savings we get by not inflating the object to a temporary
2389 * file. Practically, this code only helps when we are used
2390 * by diff-cache --cached, which does read the cache before
2391 * calling us.
2392 */
2393 if (!active_cache)
2394 return 0;
2396 /* We want to avoid the working directory if our caller
2397 * doesn't need the data in a normal file, this system
2398 * is rather slow with its stat/open/mmap/close syscalls,
2399 * and the object is contained in a pack file. The pack
2400 * is probably already open and will be faster to obtain
2401 * the data through than the working directory. Loose
2402 * objects however would tend to be slower as they need
2403 * to be individually opened and inflated.
2404 */
2405 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2406 return 0;
2408 len = strlen(name);
2409 pos = cache_name_pos(name, len);
2410 if (pos < 0)
2411 return 0;
2412 ce = active_cache[pos];
2414 /*
2415 * This is not the sha1 we are looking for, or
2416 * unreusable because it is not a regular file.
2417 */
2418 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2419 return 0;
2421 /*
2422 * If ce is marked as "assume unchanged", there is no
2423 * guarantee that work tree matches what we are looking for.
2424 */
2425 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2426 return 0;
2428 /*
2429 * If ce matches the file in the work tree, we can reuse it.
2430 */
2431 if (ce_uptodate(ce) ||
2432 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2433 return 1;
2435 return 0;
2436 }
2438 static int populate_from_stdin(struct diff_filespec *s)
2439 {
2440 struct strbuf buf = STRBUF_INIT;
2441 size_t size = 0;
2443 if (strbuf_read(&buf, 0, 0) < 0)
2444 return error("error while reading from stdin %s",
2445 strerror(errno));
2447 s->should_munmap = 0;
2448 s->data = strbuf_detach(&buf, &size);
2449 s->size = size;
2450 s->should_free = 1;
2451 return 0;
2452 }
2454 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2455 {
2456 int len;
2457 char *data = xmalloc(100), *dirty = "";
2459 /* Are we looking at the work tree? */
2460 if (s->dirty_submodule)
2461 dirty = "-dirty";
2463 len = snprintf(data, 100,
2464 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2465 s->data = data;
2466 s->size = len;
2467 s->should_free = 1;
2468 if (size_only) {
2469 s->data = NULL;
2470 free(data);
2471 }
2472 return 0;
2473 }
2475 /*
2476 * While doing rename detection and pickaxe operation, we may need to
2477 * grab the data for the blob (or file) for our own in-core comparison.
2478 * diff_filespec has data and size fields for this purpose.
2479 */
2480 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2481 {
2482 int err = 0;
2483 if (!DIFF_FILE_VALID(s))
2484 die("internal error: asking to populate invalid file.");
2485 if (S_ISDIR(s->mode))
2486 return -1;
2488 if (s->data)
2489 return 0;
2491 if (size_only && 0 < s->size)
2492 return 0;
2494 if (S_ISGITLINK(s->mode))
2495 return diff_populate_gitlink(s, size_only);
2497 if (!s->sha1_valid ||
2498 reuse_worktree_file(s->path, s->sha1, 0)) {
2499 struct strbuf buf = STRBUF_INIT;
2500 struct stat st;
2501 int fd;
2503 if (!strcmp(s->path, "-"))
2504 return populate_from_stdin(s);
2506 if (lstat(s->path, &st) < 0) {
2507 if (errno == ENOENT) {
2508 err_empty:
2509 err = -1;
2510 empty:
2511 s->data = (char *)"";
2512 s->size = 0;
2513 return err;
2514 }
2515 }
2516 s->size = xsize_t(st.st_size);
2517 if (!s->size)
2518 goto empty;
2519 if (S_ISLNK(st.st_mode)) {
2520 struct strbuf sb = STRBUF_INIT;
2522 if (strbuf_readlink(&sb, s->path, s->size))
2523 goto err_empty;
2524 s->size = sb.len;
2525 s->data = strbuf_detach(&sb, NULL);
2526 s->should_free = 1;
2527 return 0;
2528 }
2529 if (size_only)
2530 return 0;
2531 fd = open(s->path, O_RDONLY);
2532 if (fd < 0)
2533 goto err_empty;
2534 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2535 close(fd);
2536 s->should_munmap = 1;
2538 /*
2539 * Convert from working tree format to canonical git format
2540 */
2541 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2542 size_t size = 0;
2543 munmap(s->data, s->size);
2544 s->should_munmap = 0;
2545 s->data = strbuf_detach(&buf, &size);
2546 s->size = size;
2547 s->should_free = 1;
2548 }
2549 }
2550 else {
2551 enum object_type type;
2552 if (size_only) {
2553 type = sha1_object_info(s->sha1, &s->size);
2554 if (type < 0)
2555 die("unable to read %s", sha1_to_hex(s->sha1));
2556 } else {
2557 s->data = read_sha1_file(s->sha1, &type, &s->size);
2558 if (!s->data)
2559 die("unable to read %s", sha1_to_hex(s->sha1));
2560 s->should_free = 1;
2561 }
2562 }
2563 return 0;
2564 }
2566 void diff_free_filespec_blob(struct diff_filespec *s)
2567 {
2568 if (s->should_free)
2569 free(s->data);
2570 else if (s->should_munmap)
2571 munmap(s->data, s->size);
2573 if (s->should_free || s->should_munmap) {
2574 s->should_free = s->should_munmap = 0;
2575 s->data = NULL;
2576 }
2577 }
2579 void diff_free_filespec_data(struct diff_filespec *s)
2580 {
2581 diff_free_filespec_blob(s);
2582 free(s->cnt_data);
2583 s->cnt_data = NULL;
2584 }
2586 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2587 void *blob,
2588 unsigned long size,
2589 const unsigned char *sha1,
2590 int mode)
2591 {
2592 int fd;
2593 struct strbuf buf = STRBUF_INIT;
2594 struct strbuf template = STRBUF_INIT;
2595 char *path_dup = xstrdup(path);
2596 const char *base = basename(path_dup);
2598 /* Generate "XXXXXX_basename.ext" */
2599 strbuf_addstr(&template, "XXXXXX_");
2600 strbuf_addstr(&template, base);
2602 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2603 strlen(base) + 1);
2604 if (fd < 0)
2605 die_errno("unable to create temp-file");
2606 if (convert_to_working_tree(path,
2607 (const char *)blob, (size_t)size, &buf)) {
2608 blob = buf.buf;
2609 size = buf.len;
2610 }
2611 if (write_in_full(fd, blob, size) != size)
2612 die_errno("unable to write temp-file");
2613 close(fd);
2614 temp->name = temp->tmp_path;
2615 strcpy(temp->hex, sha1_to_hex(sha1));
2616 temp->hex[40] = 0;
2617 sprintf(temp->mode, "%06o", mode);
2618 strbuf_release(&buf);
2619 strbuf_release(&template);
2620 free(path_dup);
2621 }
2623 static struct diff_tempfile *prepare_temp_file(const char *name,
2624 struct diff_filespec *one)
2625 {
2626 struct diff_tempfile *temp = claim_diff_tempfile();
2628 if (!DIFF_FILE_VALID(one)) {
2629 not_a_valid_file:
2630 /* A '-' entry produces this for file-2, and
2631 * a '+' entry produces this for file-1.
2632 */
2633 temp->name = "/dev/null";
2634 strcpy(temp->hex, ".");
2635 strcpy(temp->mode, ".");
2636 return temp;
2637 }
2639 if (!remove_tempfile_installed) {
2640 atexit(remove_tempfile);
2641 sigchain_push_common(remove_tempfile_on_signal);
2642 remove_tempfile_installed = 1;
2643 }
2645 if (!one->sha1_valid ||
2646 reuse_worktree_file(name, one->sha1, 1)) {
2647 struct stat st;
2648 if (lstat(name, &st) < 0) {
2649 if (errno == ENOENT)
2650 goto not_a_valid_file;
2651 die_errno("stat(%s)", name);
2652 }
2653 if (S_ISLNK(st.st_mode)) {
2654 struct strbuf sb = STRBUF_INIT;
2655 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2656 die_errno("readlink(%s)", name);
2657 prep_temp_blob(name, temp, sb.buf, sb.len,
2658 (one->sha1_valid ?
2659 one->sha1 : null_sha1),
2660 (one->sha1_valid ?
2661 one->mode : S_IFLNK));
2662 strbuf_release(&sb);
2663 }
2664 else {
2665 /* we can borrow from the file in the work tree */
2666 temp->name = name;
2667 if (!one->sha1_valid)
2668 strcpy(temp->hex, sha1_to_hex(null_sha1));
2669 else
2670 strcpy(temp->hex, sha1_to_hex(one->sha1));
2671 /* Even though we may sometimes borrow the
2672 * contents from the work tree, we always want
2673 * one->mode. mode is trustworthy even when
2674 * !(one->sha1_valid), as long as
2675 * DIFF_FILE_VALID(one).
2676 */
2677 sprintf(temp->mode, "%06o", one->mode);
2678 }
2679 return temp;
2680 }
2681 else {
2682 if (diff_populate_filespec(one, 0))
2683 die("cannot read data blob for %s", one->path);
2684 prep_temp_blob(name, temp, one->data, one->size,
2685 one->sha1, one->mode);
2686 }
2687 return temp;
2688 }
2690 /* An external diff command takes:
2691 *
2692 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2693 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2694 *
2695 */
2696 static void run_external_diff(const char *pgm,
2697 const char *name,
2698 const char *other,
2699 struct diff_filespec *one,
2700 struct diff_filespec *two,
2701 const char *xfrm_msg,
2702 int complete_rewrite)
2703 {
2704 const char *spawn_arg[10];
2705 int retval;
2706 const char **arg = &spawn_arg[0];
2708 if (one && two) {
2709 struct diff_tempfile *temp_one, *temp_two;
2710 const char *othername = (other ? other : name);
2711 temp_one = prepare_temp_file(name, one);
2712 temp_two = prepare_temp_file(othername, two);
2713 *arg++ = pgm;
2714 *arg++ = name;
2715 *arg++ = temp_one->name;
2716 *arg++ = temp_one->hex;
2717 *arg++ = temp_one->mode;
2718 *arg++ = temp_two->name;
2719 *arg++ = temp_two->hex;
2720 *arg++ = temp_two->mode;
2721 if (other) {
2722 *arg++ = other;
2723 *arg++ = xfrm_msg;
2724 }
2725 } else {
2726 *arg++ = pgm;
2727 *arg++ = name;
2728 }
2729 *arg = NULL;
2730 fflush(NULL);
2731 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2732 remove_tempfile();
2733 if (retval) {
2734 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2735 exit(1);
2736 }
2737 }
2739 static int similarity_index(struct diff_filepair *p)
2740 {
2741 return p->score * 100 / MAX_SCORE;
2742 }
2744 static void fill_metainfo(struct strbuf *msg,
2745 const char *name,
2746 const char *other,
2747 struct diff_filespec *one,
2748 struct diff_filespec *two,
2749 struct diff_options *o,
2750 struct diff_filepair *p,
2751 int *must_show_header,
2752 int use_color)
2753 {
2754 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2755 const char *reset = diff_get_color(use_color, DIFF_RESET);
2756 struct strbuf *msgbuf;
2757 char *line_prefix = "";
2759 *must_show_header = 1;
2760 if (o->output_prefix) {
2761 msgbuf = o->output_prefix(o, o->output_prefix_data);
2762 line_prefix = msgbuf->buf;
2763 }
2764 strbuf_init(msg, PATH_MAX * 2 + 300);
2765 switch (p->status) {
2766 case DIFF_STATUS_COPIED:
2767 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2768 line_prefix, set, similarity_index(p));
2769 strbuf_addf(msg, "%s\n%s%scopy from ",
2770 reset, line_prefix, set);
2771 quote_c_style(name, msg, NULL, 0);
2772 strbuf_addf(msg, "%s\n%s%scopy to ", 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_RENAMED:
2777 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2778 line_prefix, set, similarity_index(p));
2779 strbuf_addf(msg, "%s\n%s%srename from ",
2780 reset, line_prefix, set);
2781 quote_c_style(name, msg, NULL, 0);
2782 strbuf_addf(msg, "%s\n%s%srename to ",
2783 reset, line_prefix, set);
2784 quote_c_style(other, msg, NULL, 0);
2785 strbuf_addf(msg, "%s\n", reset);
2786 break;
2787 case DIFF_STATUS_MODIFIED:
2788 if (p->score) {
2789 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2790 line_prefix,
2791 set, similarity_index(p), reset);
2792 break;
2793 }
2794 /* fallthru */
2795 default:
2796 *must_show_header = 0;
2797 }
2798 if (one && two && hashcmp(one->sha1, two->sha1)) {
2799 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2801 if (DIFF_OPT_TST(o, BINARY)) {
2802 mmfile_t mf;
2803 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2804 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2805 abbrev = 40;
2806 }
2807 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2808 find_unique_abbrev(one->sha1, abbrev));
2809 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2810 if (one->mode == two->mode)
2811 strbuf_addf(msg, " %06o", one->mode);
2812 strbuf_addf(msg, "%s\n", reset);
2813 }
2814 }
2816 static void run_diff_cmd(const char *pgm,
2817 const char *name,
2818 const char *other,
2819 const char *attr_path,
2820 struct diff_filespec *one,
2821 struct diff_filespec *two,
2822 struct strbuf *msg,
2823 struct diff_options *o,
2824 struct diff_filepair *p)
2825 {
2826 const char *xfrm_msg = NULL;
2827 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2828 int must_show_header = 0;
2830 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2831 pgm = NULL;
2832 else {
2833 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2834 if (drv && drv->external)
2835 pgm = drv->external;
2836 }
2838 if (msg) {
2839 /*
2840 * don't use colors when the header is intended for an
2841 * external diff driver
2842 */
2843 fill_metainfo(msg, name, other, one, two, o, p,
2844 &must_show_header,
2845 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2846 xfrm_msg = msg->len ? msg->buf : NULL;
2847 }
2849 if (pgm) {
2850 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2851 complete_rewrite);
2852 return;
2853 }
2854 if (one && two)
2855 builtin_diff(name, other ? other : name,
2856 one, two, xfrm_msg, must_show_header,
2857 o, complete_rewrite);
2858 else
2859 fprintf(o->file, "* Unmerged path %s\n", name);
2860 }
2862 static void diff_fill_sha1_info(struct diff_filespec *one)
2863 {
2864 if (DIFF_FILE_VALID(one)) {
2865 if (!one->sha1_valid) {
2866 struct stat st;
2867 if (!strcmp(one->path, "-")) {
2868 hashcpy(one->sha1, null_sha1);
2869 return;
2870 }
2871 if (lstat(one->path, &st) < 0)
2872 die_errno("stat '%s'", one->path);
2873 if (index_path(one->sha1, one->path, &st, 0))
2874 die("cannot hash %s", one->path);
2875 }
2876 }
2877 else
2878 hashclr(one->sha1);
2879 }
2881 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2882 {
2883 /* Strip the prefix but do not molest /dev/null and absolute paths */
2884 if (*namep && **namep != '/') {
2885 *namep += prefix_length;
2886 if (**namep == '/')
2887 ++*namep;
2888 }
2889 if (*otherp && **otherp != '/') {
2890 *otherp += prefix_length;
2891 if (**otherp == '/')
2892 ++*otherp;
2893 }
2894 }
2896 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2897 {
2898 const char *pgm = external_diff();
2899 struct strbuf msg;
2900 struct diff_filespec *one = p->one;
2901 struct diff_filespec *two = p->two;
2902 const char *name;
2903 const char *other;
2904 const char *attr_path;
2906 name = p->one->path;
2907 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2908 attr_path = name;
2909 if (o->prefix_length)
2910 strip_prefix(o->prefix_length, &name, &other);
2912 if (DIFF_PAIR_UNMERGED(p)) {
2913 run_diff_cmd(pgm, name, NULL, attr_path,
2914 NULL, NULL, NULL, o, p);
2915 return;
2916 }
2918 diff_fill_sha1_info(one);
2919 diff_fill_sha1_info(two);
2921 if (!pgm &&
2922 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2923 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2924 /*
2925 * a filepair that changes between file and symlink
2926 * needs to be split into deletion and creation.
2927 */
2928 struct diff_filespec *null = alloc_filespec(two->path);
2929 run_diff_cmd(NULL, name, other, attr_path,
2930 one, null, &msg, o, p);
2931 free(null);
2932 strbuf_release(&msg);
2934 null = alloc_filespec(one->path);
2935 run_diff_cmd(NULL, name, other, attr_path,
2936 null, two, &msg, o, p);
2937 free(null);
2938 }
2939 else
2940 run_diff_cmd(pgm, name, other, attr_path,
2941 one, two, &msg, o, p);
2943 strbuf_release(&msg);
2944 }
2946 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2947 struct diffstat_t *diffstat)
2948 {
2949 const char *name;
2950 const char *other;
2951 int complete_rewrite = 0;
2953 if (DIFF_PAIR_UNMERGED(p)) {
2954 /* unmerged */
2955 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2956 return;
2957 }
2959 name = p->one->path;
2960 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2962 if (o->prefix_length)
2963 strip_prefix(o->prefix_length, &name, &other);
2965 diff_fill_sha1_info(p->one);
2966 diff_fill_sha1_info(p->two);
2968 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2969 complete_rewrite = 1;
2970 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2971 }
2973 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2974 {
2975 const char *name;
2976 const char *other;
2977 const char *attr_path;
2979 if (DIFF_PAIR_UNMERGED(p)) {
2980 /* unmerged */
2981 return;
2982 }
2984 name = p->one->path;
2985 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2986 attr_path = other ? other : name;
2988 if (o->prefix_length)
2989 strip_prefix(o->prefix_length, &name, &other);
2991 diff_fill_sha1_info(p->one);
2992 diff_fill_sha1_info(p->two);
2994 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2995 }
2997 void diff_setup(struct diff_options *options)
2998 {
2999 memcpy(options, &default_diff_options, sizeof(*options));
3001 options->file = stdout;
3003 options->line_termination = '\n';
3004 options->break_opt = -1;
3005 options->rename_limit = -1;
3006 options->dirstat_permille = diff_dirstat_permille_default;
3007 options->context = 3;
3009 options->change = diff_change;
3010 options->add_remove = diff_addremove;
3011 if (diff_use_color_default > 0)
3012 DIFF_OPT_SET(options, COLOR_DIFF);
3013 options->detect_rename = diff_detect_rename_default;
3015 if (diff_no_prefix) {
3016 options->a_prefix = options->b_prefix = "";
3017 } else if (!diff_mnemonic_prefix) {
3018 options->a_prefix = "a/";
3019 options->b_prefix = "b/";
3020 }
3021 }
3023 int diff_setup_done(struct diff_options *options)
3024 {
3025 int count = 0;
3027 if (options->output_format & DIFF_FORMAT_NAME)
3028 count++;
3029 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3030 count++;
3031 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3032 count++;
3033 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3034 count++;
3035 if (count > 1)
3036 die("--name-only, --name-status, --check and -s are mutually exclusive");
3038 /*
3039 * Most of the time we can say "there are changes"
3040 * only by checking if there are changed paths, but
3041 * --ignore-whitespace* options force us to look
3042 * inside contents.
3043 */
3045 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3046 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3047 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3048 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3049 else
3050 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3052 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3053 options->detect_rename = DIFF_DETECT_COPY;
3055 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3056 options->prefix = NULL;
3057 if (options->prefix)
3058 options->prefix_length = strlen(options->prefix);
3059 else
3060 options->prefix_length = 0;
3062 if (options->output_format & (DIFF_FORMAT_NAME |
3063 DIFF_FORMAT_NAME_STATUS |
3064 DIFF_FORMAT_CHECKDIFF |
3065 DIFF_FORMAT_NO_OUTPUT))
3066 options->output_format &= ~(DIFF_FORMAT_RAW |
3067 DIFF_FORMAT_NUMSTAT |
3068 DIFF_FORMAT_DIFFSTAT |
3069 DIFF_FORMAT_SHORTSTAT |
3070 DIFF_FORMAT_DIRSTAT |
3071 DIFF_FORMAT_SUMMARY |
3072 DIFF_FORMAT_PATCH);
3074 /*
3075 * These cases always need recursive; we do not drop caller-supplied
3076 * recursive bits for other formats here.
3077 */
3078 if (options->output_format & (DIFF_FORMAT_PATCH |
3079 DIFF_FORMAT_NUMSTAT |
3080 DIFF_FORMAT_DIFFSTAT |
3081 DIFF_FORMAT_SHORTSTAT |
3082 DIFF_FORMAT_DIRSTAT |
3083 DIFF_FORMAT_SUMMARY |
3084 DIFF_FORMAT_CHECKDIFF))
3085 DIFF_OPT_SET(options, RECURSIVE);
3086 /*
3087 * Also pickaxe would not work very well if you do not say recursive
3088 */
3089 if (options->pickaxe)
3090 DIFF_OPT_SET(options, RECURSIVE);
3091 /*
3092 * When patches are generated, submodules diffed against the work tree
3093 * must be checked for dirtiness too so it can be shown in the output
3094 */
3095 if (options->output_format & DIFF_FORMAT_PATCH)
3096 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3098 if (options->detect_rename && options->rename_limit < 0)
3099 options->rename_limit = diff_rename_limit_default;
3100 if (options->setup & DIFF_SETUP_USE_CACHE) {
3101 if (!active_cache)
3102 /* read-cache does not die even when it fails
3103 * so it is safe for us to do this here. Also
3104 * it does not smudge active_cache or active_nr
3105 * when it fails, so we do not have to worry about
3106 * cleaning it up ourselves either.
3107 */
3108 read_cache();
3109 }
3110 if (options->abbrev <= 0 || 40 < options->abbrev)
3111 options->abbrev = 40; /* full */
3113 /*
3114 * It does not make sense to show the first hit we happened
3115 * to have found. It does not make sense not to return with
3116 * exit code in such a case either.
3117 */
3118 if (DIFF_OPT_TST(options, QUICK)) {
3119 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3120 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3121 }
3123 return 0;
3124 }
3126 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3127 {
3128 char c, *eq;
3129 int len;
3131 if (*arg != '-')
3132 return 0;
3133 c = *++arg;
3134 if (!c)
3135 return 0;
3136 if (c == arg_short) {
3137 c = *++arg;
3138 if (!c)
3139 return 1;
3140 if (val && isdigit(c)) {
3141 char *end;
3142 int n = strtoul(arg, &end, 10);
3143 if (*end)
3144 return 0;
3145 *val = n;
3146 return 1;
3147 }
3148 return 0;
3149 }
3150 if (c != '-')
3151 return 0;
3152 arg++;
3153 eq = strchr(arg, '=');
3154 if (eq)
3155 len = eq - arg;
3156 else
3157 len = strlen(arg);
3158 if (!len || strncmp(arg, arg_long, len))
3159 return 0;
3160 if (eq) {
3161 int n;
3162 char *end;
3163 if (!isdigit(*++eq))
3164 return 0;
3165 n = strtoul(eq, &end, 10);
3166 if (*end)
3167 return 0;
3168 *val = n;
3169 }
3170 return 1;
3171 }
3173 static int diff_scoreopt_parse(const char *opt);
3175 static inline int short_opt(char opt, const char **argv,
3176 const char **optarg)
3177 {
3178 const char *arg = argv[0];
3179 if (arg[0] != '-' || arg[1] != opt)
3180 return 0;
3181 if (arg[2] != '\0') {
3182 *optarg = arg + 2;
3183 return 1;
3184 }
3185 if (!argv[1])
3186 die("Option '%c' requires a value", opt);
3187 *optarg = argv[1];
3188 return 2;
3189 }
3191 int parse_long_opt(const char *opt, const char **argv,
3192 const char **optarg)
3193 {
3194 const char *arg = argv[0];
3195 if (arg[0] != '-' || arg[1] != '-')
3196 return 0;
3197 arg += strlen("--");
3198 if (prefixcmp(arg, opt))
3199 return 0;
3200 arg += strlen(opt);
3201 if (*arg == '=') { /* sticked form: --option=value */
3202 *optarg = arg + 1;
3203 return 1;
3204 }
3205 if (*arg != '\0')
3206 return 0;
3207 /* separate form: --option value */
3208 if (!argv[1])
3209 die("Option '--%s' requires a value", opt);
3210 *optarg = argv[1];
3211 return 2;
3212 }
3214 static int stat_opt(struct diff_options *options, const char **av)
3215 {
3216 const char *arg = av[0];
3217 char *end;
3218 int width = options->stat_width;
3219 int name_width = options->stat_name_width;
3220 int argcount = 1;
3222 arg += strlen("--stat");
3223 end = (char *)arg;
3225 switch (*arg) {
3226 case '-':
3227 if (!prefixcmp(arg, "-width")) {
3228 arg += strlen("-width");
3229 if (*arg == '=')
3230 width = strtoul(arg + 1, &end, 10);
3231 else if (!*arg && !av[1])
3232 die("Option '--stat-width' requires a value");
3233 else if (!*arg) {
3234 width = strtoul(av[1], &end, 10);
3235 argcount = 2;
3236 }
3237 } else if (!prefixcmp(arg, "-name-width")) {
3238 arg += strlen("-name-width");
3239 if (*arg == '=')
3240 name_width = strtoul(arg + 1, &end, 10);
3241 else if (!*arg && !av[1])
3242 die("Option '--stat-name-width' requires a value");
3243 else if (!*arg) {
3244 name_width = strtoul(av[1], &end, 10);
3245 argcount = 2;
3246 }
3247 }
3248 break;
3249 case '=':
3250 width = strtoul(arg+1, &end, 10);
3251 if (*end == ',')
3252 name_width = strtoul(end+1, &end, 10);
3253 }
3255 /* Important! This checks all the error cases! */
3256 if (*end)
3257 return 0;
3258 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3259 options->stat_name_width = name_width;
3260 options->stat_width = width;
3261 return argcount;
3262 }
3264 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3265 {
3266 struct strbuf errmsg = STRBUF_INIT;
3267 if (parse_dirstat_params(options, params, &errmsg))
3268 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3269 errmsg.buf);
3270 strbuf_release(&errmsg);
3271 /*
3272 * The caller knows a dirstat-related option is given from the command
3273 * line; allow it to say "return this_function();"
3274 */
3275 options->output_format |= DIFF_FORMAT_DIRSTAT;
3276 return 1;
3277 }
3279 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3280 {
3281 const char *arg = av[0];
3282 const char *optarg;
3283 int argcount;
3285 /* Output format options */
3286 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3287 options->output_format |= DIFF_FORMAT_PATCH;
3288 else if (opt_arg(arg, 'U', "unified", &options->context))
3289 options->output_format |= DIFF_FORMAT_PATCH;
3290 else if (!strcmp(arg, "--raw"))
3291 options->output_format |= DIFF_FORMAT_RAW;
3292 else if (!strcmp(arg, "--patch-with-raw"))
3293 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3294 else if (!strcmp(arg, "--numstat"))
3295 options->output_format |= DIFF_FORMAT_NUMSTAT;
3296 else if (!strcmp(arg, "--shortstat"))
3297 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3298 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3299 return parse_dirstat_opt(options, "");
3300 else if (!prefixcmp(arg, "-X"))
3301 return parse_dirstat_opt(options, arg + 2);
3302 else if (!prefixcmp(arg, "--dirstat="))
3303 return parse_dirstat_opt(options, arg + 10);
3304 else if (!strcmp(arg, "--cumulative"))
3305 return parse_dirstat_opt(options, "cumulative");
3306 else if (!strcmp(arg, "--dirstat-by-file"))
3307 return parse_dirstat_opt(options, "files");
3308 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3309 parse_dirstat_opt(options, "files");
3310 return parse_dirstat_opt(options, arg + 18);
3311 }
3312 else if (!strcmp(arg, "--check"))
3313 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3314 else if (!strcmp(arg, "--summary"))
3315 options->output_format |= DIFF_FORMAT_SUMMARY;
3316 else if (!strcmp(arg, "--patch-with-stat"))
3317 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3318 else if (!strcmp(arg, "--name-only"))
3319 options->output_format |= DIFF_FORMAT_NAME;
3320 else if (!strcmp(arg, "--name-status"))
3321 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3322 else if (!strcmp(arg, "-s"))
3323 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3324 else if (!prefixcmp(arg, "--stat"))
3325 /* --stat, --stat-width, or --stat-name-width */
3326 return stat_opt(options, av);
3328 /* renames options */
3329 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3330 !strcmp(arg, "--break-rewrites")) {
3331 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3332 return error("invalid argument to -B: %s", arg+2);
3333 }
3334 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3335 !strcmp(arg, "--find-renames")) {
3336 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3337 return error("invalid argument to -M: %s", arg+2);
3338 options->detect_rename = DIFF_DETECT_RENAME;
3339 }
3340 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3341 options->irreversible_delete = 1;
3342 }
3343 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3344 !strcmp(arg, "--find-copies")) {
3345 if (options->detect_rename == DIFF_DETECT_COPY)
3346 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3347 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3348 return error("invalid argument to -C: %s", arg+2);
3349 options->detect_rename = DIFF_DETECT_COPY;
3350 }
3351 else if (!strcmp(arg, "--no-renames"))
3352 options->detect_rename = 0;
3353 else if (!strcmp(arg, "--relative"))
3354 DIFF_OPT_SET(options, RELATIVE_NAME);
3355 else if (!prefixcmp(arg, "--relative=")) {
3356 DIFF_OPT_SET(options, RELATIVE_NAME);
3357 options->prefix = arg + 11;
3358 }
3360 /* xdiff options */
3361 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3362 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3363 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3364 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3365 else if (!strcmp(arg, "--ignore-space-at-eol"))
3366 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3367 else if (!strcmp(arg, "--patience"))
3368 DIFF_XDL_SET(options, PATIENCE_DIFF);
3370 /* flags options */
3371 else if (!strcmp(arg, "--binary")) {
3372 options->output_format |= DIFF_FORMAT_PATCH;
3373 DIFF_OPT_SET(options, BINARY);
3374 }
3375 else if (!strcmp(arg, "--full-index"))
3376 DIFF_OPT_SET(options, FULL_INDEX);
3377 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3378 DIFF_OPT_SET(options, TEXT);
3379 else if (!strcmp(arg, "-R"))
3380 DIFF_OPT_SET(options, REVERSE_DIFF);
3381 else if (!strcmp(arg, "--find-copies-harder"))
3382 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3383 else if (!strcmp(arg, "--follow"))
3384 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3385 else if (!strcmp(arg, "--color"))
3386 DIFF_OPT_SET(options, COLOR_DIFF);
3387 else if (!prefixcmp(arg, "--color=")) {
3388 int value = git_config_colorbool(NULL, arg+8, -1);
3389 if (value == 0)
3390 DIFF_OPT_CLR(options, COLOR_DIFF);
3391 else if (value > 0)
3392 DIFF_OPT_SET(options, COLOR_DIFF);
3393 else
3394 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3395 }
3396 else if (!strcmp(arg, "--no-color"))
3397 DIFF_OPT_CLR(options, COLOR_DIFF);
3398 else if (!strcmp(arg, "--color-words")) {
3399 DIFF_OPT_SET(options, COLOR_DIFF);
3400 options->word_diff = DIFF_WORDS_COLOR;
3401 }
3402 else if (!prefixcmp(arg, "--color-words=")) {
3403 DIFF_OPT_SET(options, COLOR_DIFF);
3404 options->word_diff = DIFF_WORDS_COLOR;
3405 options->word_regex = arg + 14;
3406 }
3407 else if (!strcmp(arg, "--word-diff")) {
3408 if (options->word_diff == DIFF_WORDS_NONE)
3409 options->word_diff = DIFF_WORDS_PLAIN;
3410 }
3411 else if (!prefixcmp(arg, "--word-diff=")) {
3412 const char *type = arg + 12;
3413 if (!strcmp(type, "plain"))
3414 options->word_diff = DIFF_WORDS_PLAIN;
3415 else if (!strcmp(type, "color")) {
3416 DIFF_OPT_SET(options, COLOR_DIFF);
3417 options->word_diff = DIFF_WORDS_COLOR;
3418 }
3419 else if (!strcmp(type, "porcelain"))
3420 options->word_diff = DIFF_WORDS_PORCELAIN;
3421 else if (!strcmp(type, "none"))
3422 options->word_diff = DIFF_WORDS_NONE;
3423 else
3424 die("bad --word-diff argument: %s", type);
3425 }
3426 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3427 if (options->word_diff == DIFF_WORDS_NONE)
3428 options->word_diff = DIFF_WORDS_PLAIN;
3429 options->word_regex = optarg;
3430 return argcount;
3431 }
3432 else if (!strcmp(arg, "--exit-code"))
3433 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3434 else if (!strcmp(arg, "--quiet"))
3435 DIFF_OPT_SET(options, QUICK);
3436 else if (!strcmp(arg, "--ext-diff"))
3437 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3438 else if (!strcmp(arg, "--no-ext-diff"))
3439 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3440 else if (!strcmp(arg, "--textconv"))
3441 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3442 else if (!strcmp(arg, "--no-textconv"))
3443 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3444 else if (!strcmp(arg, "--ignore-submodules")) {
3445 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3446 handle_ignore_submodules_arg(options, "all");
3447 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3448 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3449 handle_ignore_submodules_arg(options, arg + 20);
3450 } else if (!strcmp(arg, "--submodule"))
3451 DIFF_OPT_SET(options, SUBMODULE_LOG);
3452 else if (!prefixcmp(arg, "--submodule=")) {
3453 if (!strcmp(arg + 12, "log"))
3454 DIFF_OPT_SET(options, SUBMODULE_LOG);
3455 }
3457 /* misc options */
3458 else if (!strcmp(arg, "-z"))
3459 options->line_termination = 0;
3460 else if ((argcount = short_opt('l', av, &optarg))) {
3461 options->rename_limit = strtoul(optarg, NULL, 10);
3462 return argcount;
3463 }
3464 else if ((argcount = short_opt('S', av, &optarg))) {
3465 options->pickaxe = optarg;
3466 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3467 return argcount;
3468 } else if ((argcount = short_opt('G', av, &optarg))) {
3469 options->pickaxe = optarg;
3470 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3471 return argcount;
3472 }
3473 else if (!strcmp(arg, "--pickaxe-all"))
3474 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3475 else if (!strcmp(arg, "--pickaxe-regex"))
3476 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3477 else if ((argcount = short_opt('O', av, &optarg))) {
3478 options->orderfile = optarg;
3479 return argcount;
3480 }
3481 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3482 options->filter = optarg;
3483 return argcount;
3484 }
3485 else if (!strcmp(arg, "--abbrev"))
3486 options->abbrev = DEFAULT_ABBREV;
3487 else if (!prefixcmp(arg, "--abbrev=")) {
3488 options->abbrev = strtoul(arg + 9, NULL, 10);
3489 if (options->abbrev < MINIMUM_ABBREV)
3490 options->abbrev = MINIMUM_ABBREV;
3491 else if (40 < options->abbrev)
3492 options->abbrev = 40;
3493 }
3494 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3495 options->a_prefix = optarg;
3496 return argcount;
3497 }
3498 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3499 options->b_prefix = optarg;
3500 return argcount;
3501 }
3502 else if (!strcmp(arg, "--no-prefix"))
3503 options->a_prefix = options->b_prefix = "";
3504 else if (opt_arg(arg, '\0', "inter-hunk-context",
3505 &options->interhunkcontext))
3506 ;
3507 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3508 options->file = fopen(optarg, "w");
3509 if (!options->file)
3510 die_errno("Could not open '%s'", optarg);
3511 options->close_file = 1;
3512 return argcount;
3513 } else
3514 return 0;
3515 return 1;
3516 }
3518 int parse_rename_score(const char **cp_p)
3519 {
3520 unsigned long num, scale;
3521 int ch, dot;
3522 const char *cp = *cp_p;
3524 num = 0;
3525 scale = 1;
3526 dot = 0;
3527 for (;;) {
3528 ch = *cp;
3529 if ( !dot && ch == '.' ) {
3530 scale = 1;
3531 dot = 1;
3532 } else if ( ch == '%' ) {
3533 scale = dot ? scale*100 : 100;
3534 cp++; /* % is always at the end */
3535 break;
3536 } else if ( ch >= '0' && ch <= '9' ) {
3537 if ( scale < 100000 ) {
3538 scale *= 10;
3539 num = (num*10) + (ch-'0');
3540 }
3541 } else {
3542 break;
3543 }
3544 cp++;
3545 }
3546 *cp_p = cp;
3548 /* user says num divided by scale and we say internally that
3549 * is MAX_SCORE * num / scale.
3550 */
3551 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3552 }
3554 static int diff_scoreopt_parse(const char *opt)
3555 {
3556 int opt1, opt2, cmd;
3558 if (*opt++ != '-')
3559 return -1;
3560 cmd = *opt++;
3561 if (cmd == '-') {
3562 /* convert the long-form arguments into short-form versions */
3563 if (!prefixcmp(opt, "break-rewrites")) {
3564 opt += strlen("break-rewrites");
3565 if (*opt == 0 || *opt++ == '=')
3566 cmd = 'B';
3567 } else if (!prefixcmp(opt, "find-copies")) {
3568 opt += strlen("find-copies");
3569 if (*opt == 0 || *opt++ == '=')
3570 cmd = 'C';
3571 } else if (!prefixcmp(opt, "find-renames")) {
3572 opt += strlen("find-renames");
3573 if (*opt == 0 || *opt++ == '=')
3574 cmd = 'M';
3575 }
3576 }
3577 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3578 return -1; /* that is not a -M, -C nor -B option */
3580 opt1 = parse_rename_score(&opt);
3581 if (cmd != 'B')
3582 opt2 = 0;
3583 else {
3584 if (*opt == 0)
3585 opt2 = 0;
3586 else if (*opt != '/')
3587 return -1; /* we expect -B80/99 or -B80 */
3588 else {
3589 opt++;
3590 opt2 = parse_rename_score(&opt);
3591 }
3592 }
3593 if (*opt != 0)
3594 return -1;
3595 return opt1 | (opt2 << 16);
3596 }
3598 struct diff_queue_struct diff_queued_diff;
3600 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3601 {
3602 if (queue->alloc <= queue->nr) {
3603 queue->alloc = alloc_nr(queue->alloc);
3604 queue->queue = xrealloc(queue->queue,
3605 sizeof(dp) * queue->alloc);
3606 }
3607 queue->queue[queue->nr++] = dp;
3608 }
3610 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3611 struct diff_filespec *one,
3612 struct diff_filespec *two)
3613 {
3614 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3615 dp->one = one;
3616 dp->two = two;
3617 if (queue)
3618 diff_q(queue, dp);
3619 return dp;
3620 }
3622 void diff_free_filepair(struct diff_filepair *p)
3623 {
3624 free_filespec(p->one);
3625 free_filespec(p->two);
3626 free(p);
3627 }
3629 /* This is different from find_unique_abbrev() in that
3630 * it stuffs the result with dots for alignment.
3631 */
3632 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3633 {
3634 int abblen;
3635 const char *abbrev;
3636 if (len == 40)
3637 return sha1_to_hex(sha1);
3639 abbrev = find_unique_abbrev(sha1, len);
3640 abblen = strlen(abbrev);
3641 if (abblen < 37) {
3642 static char hex[41];
3643 if (len < abblen && abblen <= len + 2)
3644 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3645 else
3646 sprintf(hex, "%s...", abbrev);
3647 return hex;
3648 }
3649 return sha1_to_hex(sha1);
3650 }
3652 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3653 {
3654 int line_termination = opt->line_termination;
3655 int inter_name_termination = line_termination ? '\t' : '\0';
3656 if (opt->output_prefix) {
3657 struct strbuf *msg = NULL;
3658 msg = opt->output_prefix(opt, opt->output_prefix_data);
3659 fprintf(opt->file, "%s", msg->buf);
3660 }
3662 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3663 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3664 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3665 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3666 }
3667 if (p->score) {
3668 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3669 inter_name_termination);
3670 } else {
3671 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3672 }
3674 if (p->status == DIFF_STATUS_COPIED ||
3675 p->status == DIFF_STATUS_RENAMED) {
3676 const char *name_a, *name_b;
3677 name_a = p->one->path;
3678 name_b = p->two->path;
3679 strip_prefix(opt->prefix_length, &name_a, &name_b);
3680 write_name_quoted(name_a, opt->file, inter_name_termination);
3681 write_name_quoted(name_b, opt->file, line_termination);
3682 } else {
3683 const char *name_a, *name_b;
3684 name_a = p->one->mode ? p->one->path : p->two->path;
3685 name_b = NULL;
3686 strip_prefix(opt->prefix_length, &name_a, &name_b);
3687 write_name_quoted(name_a, opt->file, line_termination);
3688 }
3689 }
3691 int diff_unmodified_pair(struct diff_filepair *p)
3692 {
3693 /* This function is written stricter than necessary to support
3694 * the currently implemented transformers, but the idea is to
3695 * let transformers to produce diff_filepairs any way they want,
3696 * and filter and clean them up here before producing the output.
3697 */
3698 struct diff_filespec *one = p->one, *two = p->two;
3700 if (DIFF_PAIR_UNMERGED(p))
3701 return 0; /* unmerged is interesting */
3703 /* deletion, addition, mode or type change
3704 * and rename are all interesting.
3705 */
3706 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3707 DIFF_PAIR_MODE_CHANGED(p) ||
3708 strcmp(one->path, two->path))
3709 return 0;
3711 /* both are valid and point at the same path. that is, we are
3712 * dealing with a change.
3713 */
3714 if (one->sha1_valid && two->sha1_valid &&
3715 !hashcmp(one->sha1, two->sha1) &&
3716 !one->dirty_submodule && !two->dirty_submodule)
3717 return 1; /* no change */
3718 if (!one->sha1_valid && !two->sha1_valid)
3719 return 1; /* both look at the same file on the filesystem. */
3720 return 0;
3721 }
3723 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3724 {
3725 if (diff_unmodified_pair(p))
3726 return;
3728 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3729 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3730 return; /* no tree diffs in patch format */
3732 run_diff(p, o);
3733 }
3735 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3736 struct diffstat_t *diffstat)
3737 {
3738 if (diff_unmodified_pair(p))
3739 return;
3741 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3742 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3743 return; /* no useful stat for tree diffs */
3745 run_diffstat(p, o, diffstat);
3746 }
3748 static void diff_flush_checkdiff(struct diff_filepair *p,
3749 struct diff_options *o)
3750 {
3751 if (diff_unmodified_pair(p))
3752 return;
3754 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3755 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3756 return; /* nothing to check in tree diffs */
3758 run_checkdiff(p, o);
3759 }
3761 int diff_queue_is_empty(void)
3762 {
3763 struct diff_queue_struct *q = &diff_queued_diff;
3764 int i;
3765 for (i = 0; i < q->nr; i++)
3766 if (!diff_unmodified_pair(q->queue[i]))
3767 return 0;
3768 return 1;
3769 }
3771 #if DIFF_DEBUG
3772 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3773 {
3774 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3775 x, one ? one : "",
3776 s->path,
3777 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3778 s->mode,
3779 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3780 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3781 x, one ? one : "",
3782 s->size, s->xfrm_flags);
3783 }
3785 void diff_debug_filepair(const struct diff_filepair *p, int i)
3786 {
3787 diff_debug_filespec(p->one, i, "one");
3788 diff_debug_filespec(p->two, i, "two");
3789 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3790 p->score, p->status ? p->status : '?',
3791 p->one->rename_used, p->broken_pair);
3792 }
3794 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3795 {
3796 int i;
3797 if (msg)
3798 fprintf(stderr, "%s\n", msg);
3799 fprintf(stderr, "q->nr = %d\n", q->nr);
3800 for (i = 0; i < q->nr; i++) {
3801 struct diff_filepair *p = q->queue[i];
3802 diff_debug_filepair(p, i);
3803 }
3804 }
3805 #endif
3807 static void diff_resolve_rename_copy(void)
3808 {
3809 int i;
3810 struct diff_filepair *p;
3811 struct diff_queue_struct *q = &diff_queued_diff;
3813 diff_debug_queue("resolve-rename-copy", q);
3815 for (i = 0; i < q->nr; i++) {
3816 p = q->queue[i];
3817 p->status = 0; /* undecided */
3818 if (DIFF_PAIR_UNMERGED(p))
3819 p->status = DIFF_STATUS_UNMERGED;
3820 else if (!DIFF_FILE_VALID(p->one))
3821 p->status = DIFF_STATUS_ADDED;
3822 else if (!DIFF_FILE_VALID(p->two))
3823 p->status = DIFF_STATUS_DELETED;
3824 else if (DIFF_PAIR_TYPE_CHANGED(p))
3825 p->status = DIFF_STATUS_TYPE_CHANGED;
3827 /* from this point on, we are dealing with a pair
3828 * whose both sides are valid and of the same type, i.e.
3829 * either in-place edit or rename/copy edit.
3830 */
3831 else if (DIFF_PAIR_RENAME(p)) {
3832 /*
3833 * A rename might have re-connected a broken
3834 * pair up, causing the pathnames to be the
3835 * same again. If so, that's not a rename at
3836 * all, just a modification..
3837 *
3838 * Otherwise, see if this source was used for
3839 * multiple renames, in which case we decrement
3840 * the count, and call it a copy.
3841 */
3842 if (!strcmp(p->one->path, p->two->path))
3843 p->status = DIFF_STATUS_MODIFIED;
3844 else if (--p->one->rename_used > 0)
3845 p->status = DIFF_STATUS_COPIED;
3846 else
3847 p->status = DIFF_STATUS_RENAMED;
3848 }
3849 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3850 p->one->mode != p->two->mode ||
3851 p->one->dirty_submodule ||
3852 p->two->dirty_submodule ||
3853 is_null_sha1(p->one->sha1))
3854 p->status = DIFF_STATUS_MODIFIED;
3855 else {
3856 /* This is a "no-change" entry and should not
3857 * happen anymore, but prepare for broken callers.
3858 */
3859 error("feeding unmodified %s to diffcore",
3860 p->one->path);
3861 p->status = DIFF_STATUS_UNKNOWN;
3862 }
3863 }
3864 diff_debug_queue("resolve-rename-copy done", q);
3865 }
3867 static int check_pair_status(struct diff_filepair *p)
3868 {
3869 switch (p->status) {
3870 case DIFF_STATUS_UNKNOWN:
3871 return 0;
3872 case 0:
3873 die("internal error in diff-resolve-rename-copy");
3874 default:
3875 return 1;
3876 }
3877 }
3879 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3880 {
3881 int fmt = opt->output_format;
3883 if (fmt & DIFF_FORMAT_CHECKDIFF)
3884 diff_flush_checkdiff(p, opt);
3885 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3886 diff_flush_raw(p, opt);
3887 else if (fmt & DIFF_FORMAT_NAME) {
3888 const char *name_a, *name_b;
3889 name_a = p->two->path;
3890 name_b = NULL;
3891 strip_prefix(opt->prefix_length, &name_a, &name_b);
3892 write_name_quoted(name_a, opt->file, opt->line_termination);
3893 }
3894 }
3896 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3897 {
3898 if (fs->mode)
3899 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3900 else
3901 fprintf(file, " %s ", newdelete);
3902 write_name_quoted(fs->path, file, '\n');
3903 }
3906 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3907 const char *line_prefix)
3908 {
3909 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3910 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3911 p->two->mode, show_name ? ' ' : '\n');
3912 if (show_name) {
3913 write_name_quoted(p->two->path, file, '\n');
3914 }
3915 }
3916 }
3918 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3919 const char *line_prefix)
3920 {
3921 char *names = pprint_rename(p->one->path, p->two->path);
3923 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3924 free(names);
3925 show_mode_change(file, p, 0, line_prefix);
3926 }
3928 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3929 {
3930 FILE *file = opt->file;
3931 char *line_prefix = "";
3933 if (opt->output_prefix) {
3934 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3935 line_prefix = buf->buf;
3936 }
3938 switch(p->status) {
3939 case DIFF_STATUS_DELETED:
3940 fputs(line_prefix, file);
3941 show_file_mode_name(file, "delete", p->one);
3942 break;
3943 case DIFF_STATUS_ADDED:
3944 fputs(line_prefix, file);
3945 show_file_mode_name(file, "create", p->two);
3946 break;
3947 case DIFF_STATUS_COPIED:
3948 fputs(line_prefix, file);
3949 show_rename_copy(file, "copy", p, line_prefix);
3950 break;
3951 case DIFF_STATUS_RENAMED:
3952 fputs(line_prefix, file);
3953 show_rename_copy(file, "rename", p, line_prefix);
3954 break;
3955 default:
3956 if (p->score) {
3957 fprintf(file, "%s rewrite ", line_prefix);
3958 write_name_quoted(p->two->path, file, ' ');
3959 fprintf(file, "(%d%%)\n", similarity_index(p));
3960 }
3961 show_mode_change(file, p, !p->score, line_prefix);
3962 break;
3963 }
3964 }
3966 struct patch_id_t {
3967 git_SHA_CTX *ctx;
3968 int patchlen;
3969 };
3971 static int remove_space(char *line, int len)
3972 {
3973 int i;
3974 char *dst = line;
3975 unsigned char c;
3977 for (i = 0; i < len; i++)
3978 if (!isspace((c = line[i])))
3979 *dst++ = c;
3981 return dst - line;
3982 }
3984 static void patch_id_consume(void *priv, char *line, unsigned long len)
3985 {
3986 struct patch_id_t *data = priv;
3987 int new_len;
3989 /* Ignore line numbers when computing the SHA1 of the patch */
3990 if (!prefixcmp(line, "@@ -"))
3991 return;
3993 new_len = remove_space(line, len);
3995 git_SHA1_Update(data->ctx, line, new_len);
3996 data->patchlen += new_len;
3997 }
3999 /* returns 0 upon success, and writes result into sha1 */
4000 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4001 {
4002 struct diff_queue_struct *q = &diff_queued_diff;
4003 int i;
4004 git_SHA_CTX ctx;
4005 struct patch_id_t data;
4006 char buffer[PATH_MAX * 4 + 20];
4008 git_SHA1_Init(&ctx);
4009 memset(&data, 0, sizeof(struct patch_id_t));
4010 data.ctx = &ctx;
4012 for (i = 0; i < q->nr; i++) {
4013 xpparam_t xpp;
4014 xdemitconf_t xecfg;
4015 mmfile_t mf1, mf2;
4016 struct diff_filepair *p = q->queue[i];
4017 int len1, len2;
4019 memset(&xpp, 0, sizeof(xpp));
4020 memset(&xecfg, 0, sizeof(xecfg));
4021 if (p->status == 0)
4022 return error("internal diff status error");
4023 if (p->status == DIFF_STATUS_UNKNOWN)
4024 continue;
4025 if (diff_unmodified_pair(p))
4026 continue;
4027 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4028 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4029 continue;
4030 if (DIFF_PAIR_UNMERGED(p))
4031 continue;
4033 diff_fill_sha1_info(p->one);
4034 diff_fill_sha1_info(p->two);
4035 if (fill_mmfile(&mf1, p->one) < 0 ||
4036 fill_mmfile(&mf2, p->two) < 0)
4037 return error("unable to read files to diff");
4039 len1 = remove_space(p->one->path, strlen(p->one->path));
4040 len2 = remove_space(p->two->path, strlen(p->two->path));
4041 if (p->one->mode == 0)
4042 len1 = snprintf(buffer, sizeof(buffer),
4043 "diff--gita/%.*sb/%.*s"
4044 "newfilemode%06o"
4045 "---/dev/null"
4046 "+++b/%.*s",
4047 len1, p->one->path,
4048 len2, p->two->path,
4049 p->two->mode,
4050 len2, p->two->path);
4051 else if (p->two->mode == 0)
4052 len1 = snprintf(buffer, sizeof(buffer),
4053 "diff--gita/%.*sb/%.*s"
4054 "deletedfilemode%06o"
4055 "---a/%.*s"
4056 "+++/dev/null",
4057 len1, p->one->path,
4058 len2, p->two->path,
4059 p->one->mode,
4060 len1, p->one->path);
4061 else
4062 len1 = snprintf(buffer, sizeof(buffer),
4063 "diff--gita/%.*sb/%.*s"
4064 "---a/%.*s"
4065 "+++b/%.*s",
4066 len1, p->one->path,
4067 len2, p->two->path,
4068 len1, p->one->path,
4069 len2, p->two->path);
4070 git_SHA1_Update(&ctx, buffer, len1);
4072 if (diff_filespec_is_binary(p->one) ||
4073 diff_filespec_is_binary(p->two)) {
4074 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4075 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4076 continue;
4077 }
4079 xpp.flags = 0;
4080 xecfg.ctxlen = 3;
4081 xecfg.flags = 0;
4082 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4083 &xpp, &xecfg);
4084 }
4086 git_SHA1_Final(sha1, &ctx);
4087 return 0;
4088 }
4090 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4091 {
4092 struct diff_queue_struct *q = &diff_queued_diff;
4093 int i;
4094 int result = diff_get_patch_id(options, sha1);
4096 for (i = 0; i < q->nr; i++)
4097 diff_free_filepair(q->queue[i]);
4099 free(q->queue);
4100 DIFF_QUEUE_CLEAR(q);
4102 return result;
4103 }
4105 static int is_summary_empty(const struct diff_queue_struct *q)
4106 {
4107 int i;
4109 for (i = 0; i < q->nr; i++) {
4110 const struct diff_filepair *p = q->queue[i];
4112 switch (p->status) {
4113 case DIFF_STATUS_DELETED:
4114 case DIFF_STATUS_ADDED:
4115 case DIFF_STATUS_COPIED:
4116 case DIFF_STATUS_RENAMED:
4117 return 0;
4118 default:
4119 if (p->score)
4120 return 0;
4121 if (p->one->mode && p->two->mode &&
4122 p->one->mode != p->two->mode)
4123 return 0;
4124 break;
4125 }
4126 }
4127 return 1;
4128 }
4130 static const char rename_limit_warning[] =
4131 "inexact rename detection was skipped due to too many files.";
4133 static const char degrade_cc_to_c_warning[] =
4134 "only found copies from modified paths due to too many files.";
4136 static const char rename_limit_advice[] =
4137 "you may want to set your %s variable to at least "
4138 "%d and retry the command.";
4140 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4141 {
4142 if (degraded_cc)
4143 warning(degrade_cc_to_c_warning);
4144 else if (needed)
4145 warning(rename_limit_warning);
4146 else
4147 return;
4148 if (0 < needed && needed < 32767)
4149 warning(rename_limit_advice, varname, needed);
4150 }
4152 void diff_flush(struct diff_options *options)
4153 {
4154 struct diff_queue_struct *q = &diff_queued_diff;
4155 int i, output_format = options->output_format;
4156 int separator = 0;
4157 int dirstat_by_line = 0;
4159 /*
4160 * Order: raw, stat, summary, patch
4161 * or: name/name-status/checkdiff (other bits clear)
4162 */
4163 if (!q->nr)
4164 goto free_queue;
4166 if (output_format & (DIFF_FORMAT_RAW |
4167 DIFF_FORMAT_NAME |
4168 DIFF_FORMAT_NAME_STATUS |
4169 DIFF_FORMAT_CHECKDIFF)) {
4170 for (i = 0; i < q->nr; i++) {
4171 struct diff_filepair *p = q->queue[i];
4172 if (check_pair_status(p))
4173 flush_one_pair(p, options);
4174 }
4175 separator++;
4176 }
4178 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4179 dirstat_by_line = 1;
4181 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4182 dirstat_by_line) {
4183 struct diffstat_t diffstat;
4185 memset(&diffstat, 0, sizeof(struct diffstat_t));
4186 for (i = 0; i < q->nr; i++) {
4187 struct diff_filepair *p = q->queue[i];
4188 if (check_pair_status(p))
4189 diff_flush_stat(p, options, &diffstat);
4190 }
4191 if (output_format & DIFF_FORMAT_NUMSTAT)
4192 show_numstat(&diffstat, options);
4193 if (output_format & DIFF_FORMAT_DIFFSTAT)
4194 show_stats(&diffstat, options);
4195 if (output_format & DIFF_FORMAT_SHORTSTAT)
4196 show_shortstats(&diffstat, options);
4197 if (output_format & DIFF_FORMAT_DIRSTAT)
4198 show_dirstat_by_line(&diffstat, options);
4199 free_diffstat_info(&diffstat);
4200 separator++;
4201 }
4202 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4203 show_dirstat(options);
4205 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4206 for (i = 0; i < q->nr; i++) {
4207 diff_summary(options, q->queue[i]);
4208 }
4209 separator++;
4210 }
4212 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4213 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4214 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4215 /*
4216 * run diff_flush_patch for the exit status. setting
4217 * options->file to /dev/null should be safe, becaue we
4218 * aren't supposed to produce any output anyway.
4219 */
4220 if (options->close_file)
4221 fclose(options->file);
4222 options->file = fopen("/dev/null", "w");
4223 if (!options->file)
4224 die_errno("Could not open /dev/null");
4225 options->close_file = 1;
4226 for (i = 0; i < q->nr; i++) {
4227 struct diff_filepair *p = q->queue[i];
4228 if (check_pair_status(p))
4229 diff_flush_patch(p, options);
4230 if (options->found_changes)
4231 break;
4232 }
4233 }
4235 if (output_format & DIFF_FORMAT_PATCH) {
4236 if (separator) {
4237 putc(options->line_termination, options->file);
4238 if (options->stat_sep) {
4239 /* attach patch instead of inline */
4240 fputs(options->stat_sep, options->file);
4241 }
4242 }
4244 for (i = 0; i < q->nr; i++) {
4245 struct diff_filepair *p = q->queue[i];
4246 if (check_pair_status(p))
4247 diff_flush_patch(p, options);
4248 }
4249 }
4251 if (output_format & DIFF_FORMAT_CALLBACK)
4252 options->format_callback(q, options, options->format_callback_data);
4254 for (i = 0; i < q->nr; i++)
4255 diff_free_filepair(q->queue[i]);
4256 free_queue:
4257 free(q->queue);
4258 DIFF_QUEUE_CLEAR(q);
4259 if (options->close_file)
4260 fclose(options->file);
4262 /*
4263 * Report the content-level differences with HAS_CHANGES;
4264 * diff_addremove/diff_change does not set the bit when
4265 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4266 */
4267 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4268 if (options->found_changes)
4269 DIFF_OPT_SET(options, HAS_CHANGES);
4270 else
4271 DIFF_OPT_CLR(options, HAS_CHANGES);
4272 }
4273 }
4275 static void diffcore_apply_filter(const char *filter)
4276 {
4277 int i;
4278 struct diff_queue_struct *q = &diff_queued_diff;
4279 struct diff_queue_struct outq;
4280 DIFF_QUEUE_CLEAR(&outq);
4282 if (!filter)
4283 return;
4285 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4286 int found;
4287 for (i = found = 0; !found && i < q->nr; i++) {
4288 struct diff_filepair *p = q->queue[i];
4289 if (((p->status == DIFF_STATUS_MODIFIED) &&
4290 ((p->score &&
4291 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4292 (!p->score &&
4293 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4294 ((p->status != DIFF_STATUS_MODIFIED) &&
4295 strchr(filter, p->status)))
4296 found++;
4297 }
4298 if (found)
4299 return;
4301 /* otherwise we will clear the whole queue
4302 * by copying the empty outq at the end of this
4303 * function, but first clear the current entries
4304 * in the queue.
4305 */
4306 for (i = 0; i < q->nr; i++)
4307 diff_free_filepair(q->queue[i]);
4308 }
4309 else {
4310 /* Only the matching ones */
4311 for (i = 0; i < q->nr; i++) {
4312 struct diff_filepair *p = q->queue[i];
4314 if (((p->status == DIFF_STATUS_MODIFIED) &&
4315 ((p->score &&
4316 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4317 (!p->score &&
4318 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4319 ((p->status != DIFF_STATUS_MODIFIED) &&
4320 strchr(filter, p->status)))
4321 diff_q(&outq, p);
4322 else
4323 diff_free_filepair(p);
4324 }
4325 }
4326 free(q->queue);
4327 *q = outq;
4328 }
4330 /* Check whether two filespecs with the same mode and size are identical */
4331 static int diff_filespec_is_identical(struct diff_filespec *one,
4332 struct diff_filespec *two)
4333 {
4334 if (S_ISGITLINK(one->mode))
4335 return 0;
4336 if (diff_populate_filespec(one, 0))
4337 return 0;
4338 if (diff_populate_filespec(two, 0))
4339 return 0;
4340 return !memcmp(one->data, two->data, one->size);
4341 }
4343 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4344 {
4345 int i;
4346 struct diff_queue_struct *q = &diff_queued_diff;
4347 struct diff_queue_struct outq;
4348 DIFF_QUEUE_CLEAR(&outq);
4350 for (i = 0; i < q->nr; i++) {
4351 struct diff_filepair *p = q->queue[i];
4353 /*
4354 * 1. Entries that come from stat info dirtiness
4355 * always have both sides (iow, not create/delete),
4356 * one side of the object name is unknown, with
4357 * the same mode and size. Keep the ones that
4358 * do not match these criteria. They have real
4359 * differences.
4360 *
4361 * 2. At this point, the file is known to be modified,
4362 * with the same mode and size, and the object
4363 * name of one side is unknown. Need to inspect
4364 * the identical contents.
4365 */
4366 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4367 !DIFF_FILE_VALID(p->two) ||
4368 (p->one->sha1_valid && p->two->sha1_valid) ||
4369 (p->one->mode != p->two->mode) ||
4370 diff_populate_filespec(p->one, 1) ||
4371 diff_populate_filespec(p->two, 1) ||
4372 (p->one->size != p->two->size) ||
4373 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4374 diff_q(&outq, p);
4375 else {
4376 /*
4377 * The caller can subtract 1 from skip_stat_unmatch
4378 * to determine how many paths were dirty only
4379 * due to stat info mismatch.
4380 */
4381 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4382 diffopt->skip_stat_unmatch++;
4383 diff_free_filepair(p);
4384 }
4385 }
4386 free(q->queue);
4387 *q = outq;
4388 }
4390 static int diffnamecmp(const void *a_, const void *b_)
4391 {
4392 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4393 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4394 const char *name_a, *name_b;
4396 name_a = a->one ? a->one->path : a->two->path;
4397 name_b = b->one ? b->one->path : b->two->path;
4398 return strcmp(name_a, name_b);
4399 }
4401 void diffcore_fix_diff_index(struct diff_options *options)
4402 {
4403 struct diff_queue_struct *q = &diff_queued_diff;
4404 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4405 }
4407 void diffcore_std(struct diff_options *options)
4408 {
4409 if (options->skip_stat_unmatch)
4410 diffcore_skip_stat_unmatch(options);
4411 if (!options->found_follow) {
4412 /* See try_to_follow_renames() in tree-diff.c */
4413 if (options->break_opt != -1)
4414 diffcore_break(options->break_opt);
4415 if (options->detect_rename)
4416 diffcore_rename(options);
4417 if (options->break_opt != -1)
4418 diffcore_merge_broken();
4419 }
4420 if (options->pickaxe)
4421 diffcore_pickaxe(options);
4422 if (options->orderfile)
4423 diffcore_order(options->orderfile);
4424 if (!options->found_follow)
4425 /* See try_to_follow_renames() in tree-diff.c */
4426 diff_resolve_rename_copy();
4427 diffcore_apply_filter(options->filter);
4429 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4430 DIFF_OPT_SET(options, HAS_CHANGES);
4431 else
4432 DIFF_OPT_CLR(options, HAS_CHANGES);
4434 options->found_follow = 0;
4435 }
4437 int diff_result_code(struct diff_options *opt, int status)
4438 {
4439 int result = 0;
4441 diff_warn_rename_limit("diff.renamelimit",
4442 opt->needed_rename_limit,
4443 opt->degraded_cc_to_c);
4444 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4445 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4446 return status;
4447 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4448 DIFF_OPT_TST(opt, HAS_CHANGES))
4449 result |= 01;
4450 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4451 DIFF_OPT_TST(opt, CHECK_FAILED))
4452 result |= 02;
4453 return result;
4454 }
4456 int diff_can_quit_early(struct diff_options *opt)
4457 {
4458 return (DIFF_OPT_TST(opt, QUICK) &&
4459 !opt->filter &&
4460 DIFF_OPT_TST(opt, HAS_CHANGES));
4461 }
4463 /*
4464 * Shall changes to this submodule be ignored?
4465 *
4466 * Submodule changes can be configured to be ignored separately for each path,
4467 * but that configuration can be overridden from the command line.
4468 */
4469 static int is_submodule_ignored(const char *path, struct diff_options *options)
4470 {
4471 int ignored = 0;
4472 unsigned orig_flags = options->flags;
4473 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4474 set_diffopt_flags_from_submodule_config(options, path);
4475 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4476 ignored = 1;
4477 options->flags = orig_flags;
4478 return ignored;
4479 }
4481 void diff_addremove(struct diff_options *options,
4482 int addremove, unsigned mode,
4483 const unsigned char *sha1,
4484 const char *concatpath, unsigned dirty_submodule)
4485 {
4486 struct diff_filespec *one, *two;
4488 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4489 return;
4491 /* This may look odd, but it is a preparation for
4492 * feeding "there are unchanged files which should
4493 * not produce diffs, but when you are doing copy
4494 * detection you would need them, so here they are"
4495 * entries to the diff-core. They will be prefixed
4496 * with something like '=' or '*' (I haven't decided
4497 * which but should not make any difference).
4498 * Feeding the same new and old to diff_change()
4499 * also has the same effect.
4500 * Before the final output happens, they are pruned after
4501 * merged into rename/copy pairs as appropriate.
4502 */
4503 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4504 addremove = (addremove == '+' ? '-' :
4505 addremove == '-' ? '+' : addremove);
4507 if (options->prefix &&
4508 strncmp(concatpath, options->prefix, options->prefix_length))
4509 return;
4511 one = alloc_filespec(concatpath);
4512 two = alloc_filespec(concatpath);
4514 if (addremove != '+')
4515 fill_filespec(one, sha1, mode);
4516 if (addremove != '-') {
4517 fill_filespec(two, sha1, mode);
4518 two->dirty_submodule = dirty_submodule;
4519 }
4521 diff_queue(&diff_queued_diff, one, two);
4522 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4523 DIFF_OPT_SET(options, HAS_CHANGES);
4524 }
4526 void diff_change(struct diff_options *options,
4527 unsigned old_mode, unsigned new_mode,
4528 const unsigned char *old_sha1,
4529 const unsigned char *new_sha1,
4530 const char *concatpath,
4531 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4532 {
4533 struct diff_filespec *one, *two;
4535 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4536 is_submodule_ignored(concatpath, options))
4537 return;
4539 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4540 unsigned tmp;
4541 const unsigned char *tmp_c;
4542 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4543 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4544 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4545 new_dirty_submodule = tmp;
4546 }
4548 if (options->prefix &&
4549 strncmp(concatpath, options->prefix, options->prefix_length))
4550 return;
4552 one = alloc_filespec(concatpath);
4553 two = alloc_filespec(concatpath);
4554 fill_filespec(one, old_sha1, old_mode);
4555 fill_filespec(two, new_sha1, new_mode);
4556 one->dirty_submodule = old_dirty_submodule;
4557 two->dirty_submodule = new_dirty_submodule;
4559 diff_queue(&diff_queued_diff, one, two);
4560 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4561 DIFF_OPT_SET(options, HAS_CHANGES);
4562 }
4564 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4565 {
4566 struct diff_filepair *pair;
4567 struct diff_filespec *one, *two;
4569 if (options->prefix &&
4570 strncmp(path, options->prefix, options->prefix_length))
4571 return NULL;
4573 one = alloc_filespec(path);
4574 two = alloc_filespec(path);
4575 pair = diff_queue(&diff_queued_diff, one, two);
4576 pair->is_unmerged = 1;
4577 return pair;
4578 }
4580 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4581 size_t *outsize)
4582 {
4583 struct diff_tempfile *temp;
4584 const char *argv[3];
4585 const char **arg = argv;
4586 struct child_process child;
4587 struct strbuf buf = STRBUF_INIT;
4588 int err = 0;
4590 temp = prepare_temp_file(spec->path, spec);
4591 *arg++ = pgm;
4592 *arg++ = temp->name;
4593 *arg = NULL;
4595 memset(&child, 0, sizeof(child));
4596 child.use_shell = 1;
4597 child.argv = argv;
4598 child.out = -1;
4599 if (start_command(&child)) {
4600 remove_tempfile();
4601 return NULL;
4602 }
4604 if (strbuf_read(&buf, child.out, 0) < 0)
4605 err = error("error reading from textconv command '%s'", pgm);
4606 close(child.out);
4608 if (finish_command(&child) || err) {
4609 strbuf_release(&buf);
4610 remove_tempfile();
4611 return NULL;
4612 }
4613 remove_tempfile();
4615 return strbuf_detach(&buf, outsize);
4616 }
4618 size_t fill_textconv(struct userdiff_driver *driver,
4619 struct diff_filespec *df,
4620 char **outbuf)
4621 {
4622 size_t size;
4624 if (!driver || !driver->textconv) {
4625 if (!DIFF_FILE_VALID(df)) {
4626 *outbuf = "";
4627 return 0;
4628 }
4629 if (diff_populate_filespec(df, 0))
4630 die("unable to read files to diff");
4631 *outbuf = df->data;
4632 return df->size;
4633 }
4635 if (driver->textconv_cache && df->sha1_valid) {
4636 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4637 &size);
4638 if (*outbuf)
4639 return size;
4640 }
4642 *outbuf = run_textconv(driver->textconv, df, &size);
4643 if (!*outbuf)
4644 die("unable to read files to diff");
4646 if (driver->textconv_cache && df->sha1_valid) {
4647 /* ignore errors, as we might be in a readonly repository */
4648 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4649 size);
4650 /*
4651 * we could save up changes and flush them all at the end,
4652 * but we would need an extra call after all diffing is done.
4653 * Since generating a cache entry is the slow path anyway,
4654 * this extra overhead probably isn't a big deal.
4655 */
4656 notes_cache_write(driver->textconv_cache);
4657 }
4659 return size;
4660 }