?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'mg/gpg-interface-using-status' into next
[git/mjg.git] / diff.c
blob80f47529eb3d665f08ae65f1d943649c2466a56a
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"
18 #include "string-list.h"
19
20 #ifdef NO_FAST_WORKING_DIRECTORY
21 #define FAST_WORKING_DIRECTORY 0
22 #else
23 #define FAST_WORKING_DIRECTORY 1
24 #endif
25
26 static int diff_detect_rename_default;
27 static int diff_rename_limit_default = 400;
28 static int diff_suppress_blank_empty;
29 static int diff_use_color_default = -1;
30 static int diff_context_default = 3;
31 static const char *diff_word_regex_cfg;
32 static const char *external_diff_cmd_cfg;
33 int diff_auto_refresh_index = 1;
34 static int diff_mnemonic_prefix;
35 static int diff_no_prefix;
36 static int diff_stat_graph_width;
37 static int diff_dirstat_permille_default = 30;
38 static struct diff_options default_diff_options;
39 static long diff_algorithm;
40
41 static char diff_colors[][COLOR_MAXLEN] = {
42 GIT_COLOR_RESET,
43 GIT_COLOR_NORMAL, /* PLAIN */
44 GIT_COLOR_BOLD, /* METAINFO */
45 GIT_COLOR_CYAN, /* FRAGINFO */
46 GIT_COLOR_RED, /* OLD */
47 GIT_COLOR_GREEN, /* NEW */
48 GIT_COLOR_YELLOW, /* COMMIT */
49 GIT_COLOR_BG_RED, /* WHITESPACE */
50 GIT_COLOR_NORMAL, /* FUNCINFO */
51 };
52
53 static int parse_diff_color_slot(const char *var, int ofs)
54 {
55 if (!strcasecmp(var+ofs, "plain"))
56 return DIFF_PLAIN;
57 if (!strcasecmp(var+ofs, "meta"))
58 return DIFF_METAINFO;
59 if (!strcasecmp(var+ofs, "frag"))
60 return DIFF_FRAGINFO;
61 if (!strcasecmp(var+ofs, "old"))
62 return DIFF_FILE_OLD;
63 if (!strcasecmp(var+ofs, "new"))
64 return DIFF_FILE_NEW;
65 if (!strcasecmp(var+ofs, "commit"))
66 return DIFF_COMMIT;
67 if (!strcasecmp(var+ofs, "whitespace"))
68 return DIFF_WHITESPACE;
69 if (!strcasecmp(var+ofs, "func"))
70 return DIFF_FUNCINFO;
71 return -1;
72 }
73
74 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
75 struct strbuf *errmsg)
76 {
77 char *params_copy = xstrdup(params_string);
78 struct string_list params = STRING_LIST_INIT_NODUP;
79 int ret = 0;
80 int i;
81
82 if (*params_copy)
83 string_list_split_in_place(&params, params_copy, ',', -1);
84 for (i = 0; i < params.nr; i++) {
85 const char *p = params.items[i].string;
86 if (!strcmp(p, "changes")) {
87 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
88 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
89 } else if (!strcmp(p, "lines")) {
90 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
91 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
92 } else if (!strcmp(p, "files")) {
93 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
94 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
95 } else if (!strcmp(p, "noncumulative")) {
96 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
97 } else if (!strcmp(p, "cumulative")) {
98 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
99 } else if (isdigit(*p)) {
100 char *end;
101 int permille = strtoul(p, &end, 10) * 10;
102 if (*end == '.' && isdigit(*++end)) {
103 /* only use first digit */
104 permille += *end - '0';
105 /* .. and ignore any further digits */
106 while (isdigit(*++end))
107 ; /* nothing */
108 }
109 if (!*end)
110 options->dirstat_permille = permille;
111 else {
112 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
113 p);
114 ret++;
115 }
116 } else {
117 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
118 ret++;
119 }
120
121 }
122 string_list_clear(&params, 0);
123 free(params_copy);
124 return ret;
125 }
126
127 static int parse_submodule_params(struct diff_options *options, const char *value)
128 {
129 if (!strcmp(value, "log"))
130 DIFF_OPT_SET(options, SUBMODULE_LOG);
131 else if (!strcmp(value, "short"))
132 DIFF_OPT_CLR(options, SUBMODULE_LOG);
133 else
134 return -1;
135 return 0;
136 }
137
138 static int git_config_rename(const char *var, const char *value)
139 {
140 if (!value)
141 return DIFF_DETECT_RENAME;
142 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
143 return DIFF_DETECT_COPY;
144 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
145 }
146
147 long parse_algorithm_value(const char *value)
148 {
149 if (!value)
150 return -1;
151 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
152 return 0;
153 else if (!strcasecmp(value, "minimal"))
154 return XDF_NEED_MINIMAL;
155 else if (!strcasecmp(value, "patience"))
156 return XDF_PATIENCE_DIFF;
157 else if (!strcasecmp(value, "histogram"))
158 return XDF_HISTOGRAM_DIFF;
159 return -1;
160 }
161
162 /*
163 * These are to give UI layer defaults.
164 * The core-level commands such as git-diff-files should
165 * never be affected by the setting of diff.renames
166 * the user happens to have in the configuration file.
167 */
168 int git_diff_ui_config(const char *var, const char *value, void *cb)
169 {
170 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
171 diff_use_color_default = git_config_colorbool(var, value);
172 return 0;
173 }
174 if (!strcmp(var, "diff.context")) {
175 diff_context_default = git_config_int(var, value);
176 if (diff_context_default < 0)
177 return -1;
178 return 0;
179 }
180 if (!strcmp(var, "diff.renames")) {
181 diff_detect_rename_default = git_config_rename(var, value);
182 return 0;
183 }
184 if (!strcmp(var, "diff.autorefreshindex")) {
185 diff_auto_refresh_index = git_config_bool(var, value);
186 return 0;
187 }
188 if (!strcmp(var, "diff.mnemonicprefix")) {
189 diff_mnemonic_prefix = git_config_bool(var, value);
190 return 0;
191 }
192 if (!strcmp(var, "diff.noprefix")) {
193 diff_no_prefix = git_config_bool(var, value);
194 return 0;
195 }
196 if (!strcmp(var, "diff.statgraphwidth")) {
197 diff_stat_graph_width = git_config_int(var, value);
198 return 0;
199 }
200 if (!strcmp(var, "diff.external"))
201 return git_config_string(&external_diff_cmd_cfg, var, value);
202 if (!strcmp(var, "diff.wordregex"))
203 return git_config_string(&diff_word_regex_cfg, var, value);
204
205 if (!strcmp(var, "diff.ignoresubmodules"))
206 handle_ignore_submodules_arg(&default_diff_options, value);
207
208 if (!strcmp(var, "diff.submodule")) {
209 if (parse_submodule_params(&default_diff_options, value))
210 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
211 value);
212 return 0;
213 }
214
215 if (!strcmp(var, "diff.algorithm")) {
216 diff_algorithm = parse_algorithm_value(value);
217 if (diff_algorithm < 0)
218 return -1;
219 return 0;
220 }
221
222 if (git_color_config(var, value, cb) < 0)
223 return -1;
224
225 return git_diff_basic_config(var, value, cb);
226 }
227
228 int git_diff_basic_config(const char *var, const char *value, void *cb)
229 {
230 if (!strcmp(var, "diff.renamelimit")) {
231 diff_rename_limit_default = git_config_int(var, value);
232 return 0;
233 }
234
235 if (userdiff_config(var, value) < 0)
236 return -1;
237
238 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
239 int slot = parse_diff_color_slot(var, 11);
240 if (slot < 0)
241 return 0;
242 if (!value)
243 return config_error_nonbool(var);
244 color_parse(value, var, diff_colors[slot]);
245 return 0;
246 }
247
248 /* like GNU diff's --suppress-blank-empty option */
249 if (!strcmp(var, "diff.suppressblankempty") ||
250 /* for backwards compatibility */
251 !strcmp(var, "diff.suppress-blank-empty")) {
252 diff_suppress_blank_empty = git_config_bool(var, value);
253 return 0;
254 }
255
256 if (!strcmp(var, "diff.dirstat")) {
257 struct strbuf errmsg = STRBUF_INIT;
258 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
259 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
260 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
261 errmsg.buf);
262 strbuf_release(&errmsg);
263 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
264 return 0;
265 }
266
267 if (!prefixcmp(var, "submodule."))
268 return parse_submodule_config_option(var, value);
269
270 return git_default_config(var, value, cb);
271 }
272
273 static char *quote_two(const char *one, const char *two)
274 {
275 int need_one = quote_c_style(one, NULL, NULL, 1);
276 int need_two = quote_c_style(two, NULL, NULL, 1);
277 struct strbuf res = STRBUF_INIT;
278
279 if (need_one + need_two) {
280 strbuf_addch(&res, '"');
281 quote_c_style(one, &res, NULL, 1);
282 quote_c_style(two, &res, NULL, 1);
283 strbuf_addch(&res, '"');
284 } else {
285 strbuf_addstr(&res, one);
286 strbuf_addstr(&res, two);
287 }
288 return strbuf_detach(&res, NULL);
289 }
290
291 static const char *external_diff(void)
292 {
293 static const char *external_diff_cmd = NULL;
294 static int done_preparing = 0;
295
296 if (done_preparing)
297 return external_diff_cmd;
298 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
299 if (!external_diff_cmd)
300 external_diff_cmd = external_diff_cmd_cfg;
301 done_preparing = 1;
302 return external_diff_cmd;
303 }
304
305 static struct diff_tempfile {
306 const char *name; /* filename external diff should read from */
307 char hex[41];
308 char mode[10];
309 char tmp_path[PATH_MAX];
310 } diff_temp[2];
311
312 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
313
314 struct emit_callback {
315 int color_diff;
316 unsigned ws_rule;
317 int blank_at_eof_in_preimage;
318 int blank_at_eof_in_postimage;
319 int lno_in_preimage;
320 int lno_in_postimage;
321 sane_truncate_fn truncate;
322 const char **label_path;
323 struct diff_words_data *diff_words;
324 struct diff_options *opt;
325 int *found_changesp;
326 struct strbuf *header;
327 };
328
329 static int count_lines(const char *data, int size)
330 {
331 int count, ch, completely_empty = 1, nl_just_seen = 0;
332 count = 0;
333 while (0 < size--) {
334 ch = *data++;
335 if (ch == '\n') {
336 count++;
337 nl_just_seen = 1;
338 completely_empty = 0;
339 }
340 else {
341 nl_just_seen = 0;
342 completely_empty = 0;
343 }
344 }
345 if (completely_empty)
346 return 0;
347 if (!nl_just_seen)
348 count++; /* no trailing newline */
349 return count;
350 }
351
352 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
353 {
354 if (!DIFF_FILE_VALID(one)) {
355 mf->ptr = (char *)""; /* does not matter */
356 mf->size = 0;
357 return 0;
358 }
359 else if (diff_populate_filespec(one, 0))
360 return -1;
361
362 mf->ptr = one->data;
363 mf->size = one->size;
364 return 0;
365 }
366
367 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
368 static unsigned long diff_filespec_size(struct diff_filespec *one)
369 {
370 if (!DIFF_FILE_VALID(one))
371 return 0;
372 diff_populate_filespec(one, 1);
373 return one->size;
374 }
375
376 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
377 {
378 char *ptr = mf->ptr;
379 long size = mf->size;
380 int cnt = 0;
381
382 if (!size)
383 return cnt;
384 ptr += size - 1; /* pointing at the very end */
385 if (*ptr != '\n')
386 ; /* incomplete line */
387 else
388 ptr--; /* skip the last LF */
389 while (mf->ptr < ptr) {
390 char *prev_eol;
391 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
392 if (*prev_eol == '\n')
393 break;
394 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
395 break;
396 cnt++;
397 ptr = prev_eol - 1;
398 }
399 return cnt;
400 }
401
402 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
403 struct emit_callback *ecbdata)
404 {
405 int l1, l2, at;
406 unsigned ws_rule = ecbdata->ws_rule;
407 l1 = count_trailing_blank(mf1, ws_rule);
408 l2 = count_trailing_blank(mf2, ws_rule);
409 if (l2 <= l1) {
410 ecbdata->blank_at_eof_in_preimage = 0;
411 ecbdata->blank_at_eof_in_postimage = 0;
412 return;
413 }
414 at = count_lines(mf1->ptr, mf1->size);
415 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
416
417 at = count_lines(mf2->ptr, mf2->size);
418 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
419 }
420
421 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
422 int first, const char *line, int len)
423 {
424 int has_trailing_newline, has_trailing_carriage_return;
425 int nofirst;
426 FILE *file = o->file;
427
428 fputs(diff_line_prefix(o), file);
429
430 if (len == 0) {
431 has_trailing_newline = (first == '\n');
432 has_trailing_carriage_return = (!has_trailing_newline &&
433 (first == '\r'));
434 nofirst = has_trailing_newline || has_trailing_carriage_return;
435 } else {
436 has_trailing_newline = (len > 0 && line[len-1] == '\n');
437 if (has_trailing_newline)
438 len--;
439 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
440 if (has_trailing_carriage_return)
441 len--;
442 nofirst = 0;
443 }
444
445 if (len || !nofirst) {
446 fputs(set, file);
447 if (!nofirst)
448 fputc(first, file);
449 fwrite(line, len, 1, file);
450 fputs(reset, file);
451 }
452 if (has_trailing_carriage_return)
453 fputc('\r', file);
454 if (has_trailing_newline)
455 fputc('\n', file);
456 }
457
458 static void emit_line(struct diff_options *o, const char *set, const char *reset,
459 const char *line, int len)
460 {
461 emit_line_0(o, set, reset, line[0], line+1, len-1);
462 }
463
464 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
465 {
466 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
467 ecbdata->blank_at_eof_in_preimage &&
468 ecbdata->blank_at_eof_in_postimage &&
469 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
470 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
471 return 0;
472 return ws_blank_line(line, len, ecbdata->ws_rule);
473 }
474
475 static void emit_add_line(const char *reset,
476 struct emit_callback *ecbdata,
477 const char *line, int len)
478 {
479 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
480 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
481
482 if (!*ws)
483 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
484 else if (new_blank_line_at_eof(ecbdata, line, len))
485 /* Blank line at EOF - paint '+' as well */
486 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
487 else {
488 /* Emit just the prefix, then the rest. */
489 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
490 ws_check_emit(line, len, ecbdata->ws_rule,
491 ecbdata->opt->file, set, reset, ws);
492 }
493 }
494
495 static void emit_hunk_header(struct emit_callback *ecbdata,
496 const char *line, int len)
497 {
498 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
499 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
500 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
501 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
502 static const char atat[2] = { '@', '@' };
503 const char *cp, *ep;
504 struct strbuf msgbuf = STRBUF_INIT;
505 int org_len = len;
506 int i = 1;
507
508 /*
509 * As a hunk header must begin with "@@ -<old>, +<new> @@",
510 * it always is at least 10 bytes long.
511 */
512 if (len < 10 ||
513 memcmp(line, atat, 2) ||
514 !(ep = memmem(line + 2, len - 2, atat, 2))) {
515 emit_line(ecbdata->opt, plain, reset, line, len);
516 return;
517 }
518 ep += 2; /* skip over @@ */
519
520 /* The hunk header in fraginfo color */
521 strbuf_add(&msgbuf, frag, strlen(frag));
522 strbuf_add(&msgbuf, line, ep - line);
523 strbuf_add(&msgbuf, reset, strlen(reset));
524
525 /*
526 * trailing "\r\n"
527 */
528 for ( ; i < 3; i++)
529 if (line[len - i] == '\r' || line[len - i] == '\n')
530 len--;
531
532 /* blank before the func header */
533 for (cp = ep; ep - line < len; ep++)
534 if (*ep != ' ' && *ep != '\t')
535 break;
536 if (ep != cp) {
537 strbuf_add(&msgbuf, plain, strlen(plain));
538 strbuf_add(&msgbuf, cp, ep - cp);
539 strbuf_add(&msgbuf, reset, strlen(reset));
540 }
541
542 if (ep < line + len) {
543 strbuf_add(&msgbuf, func, strlen(func));
544 strbuf_add(&msgbuf, ep, line + len - ep);
545 strbuf_add(&msgbuf, reset, strlen(reset));
546 }
547
548 strbuf_add(&msgbuf, line + len, org_len - len);
549 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
550 strbuf_release(&msgbuf);
551 }
552
553 static struct diff_tempfile *claim_diff_tempfile(void) {
554 int i;
555 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
556 if (!diff_temp[i].name)
557 return diff_temp + i;
558 die("BUG: diff is failing to clean up its tempfiles");
559 }
560
561 static int remove_tempfile_installed;
562
563 static void remove_tempfile(void)
564 {
565 int i;
566 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
567 if (diff_temp[i].name == diff_temp[i].tmp_path)
568 unlink_or_warn(diff_temp[i].name);
569 diff_temp[i].name = NULL;
570 }
571 }
572
573 static void remove_tempfile_on_signal(int signo)
574 {
575 remove_tempfile();
576 sigchain_pop(signo);
577 raise(signo);
578 }
579
580 static void print_line_count(FILE *file, int count)
581 {
582 switch (count) {
583 case 0:
584 fprintf(file, "0,0");
585 break;
586 case 1:
587 fprintf(file, "1");
588 break;
589 default:
590 fprintf(file, "1,%d", count);
591 break;
592 }
593 }
594
595 static void emit_rewrite_lines(struct emit_callback *ecb,
596 int prefix, const char *data, int size)
597 {
598 const char *endp = NULL;
599 static const char *nneof = " No newline at end of file\n";
600 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
601 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
602
603 while (0 < size) {
604 int len;
605
606 endp = memchr(data, '\n', size);
607 len = endp ? (endp - data + 1) : size;
608 if (prefix != '+') {
609 ecb->lno_in_preimage++;
610 emit_line_0(ecb->opt, old, reset, '-',
611 data, len);
612 } else {
613 ecb->lno_in_postimage++;
614 emit_add_line(reset, ecb, data, len);
615 }
616 size -= len;
617 data += len;
618 }
619 if (!endp) {
620 const char *plain = diff_get_color(ecb->color_diff,
621 DIFF_PLAIN);
622 putc('\n', ecb->opt->file);
623 emit_line_0(ecb->opt, plain, reset, '\\',
624 nneof, strlen(nneof));
625 }
626 }
627
628 static void emit_rewrite_diff(const char *name_a,
629 const char *name_b,
630 struct diff_filespec *one,
631 struct diff_filespec *two,
632 struct userdiff_driver *textconv_one,
633 struct userdiff_driver *textconv_two,
634 struct diff_options *o)
635 {
636 int lc_a, lc_b;
637 const char *name_a_tab, *name_b_tab;
638 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
639 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
640 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
641 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
642 const char *a_prefix, *b_prefix;
643 char *data_one, *data_two;
644 size_t size_one, size_two;
645 struct emit_callback ecbdata;
646 const char *line_prefix = diff_line_prefix(o);
647
648 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
649 a_prefix = o->b_prefix;
650 b_prefix = o->a_prefix;
651 } else {
652 a_prefix = o->a_prefix;
653 b_prefix = o->b_prefix;
654 }
655
656 name_a += (*name_a == '/');
657 name_b += (*name_b == '/');
658 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
659 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
660
661 strbuf_reset(&a_name);
662 strbuf_reset(&b_name);
663 quote_two_c_style(&a_name, a_prefix, name_a, 0);
664 quote_two_c_style(&b_name, b_prefix, name_b, 0);
665
666 size_one = fill_textconv(textconv_one, one, &data_one);
667 size_two = fill_textconv(textconv_two, two, &data_two);
668
669 memset(&ecbdata, 0, sizeof(ecbdata));
670 ecbdata.color_diff = want_color(o->use_color);
671 ecbdata.found_changesp = &o->found_changes;
672 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
673 ecbdata.opt = o;
674 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
675 mmfile_t mf1, mf2;
676 mf1.ptr = (char *)data_one;
677 mf2.ptr = (char *)data_two;
678 mf1.size = size_one;
679 mf2.size = size_two;
680 check_blank_at_eof(&mf1, &mf2, &ecbdata);
681 }
682 ecbdata.lno_in_preimage = 1;
683 ecbdata.lno_in_postimage = 1;
684
685 lc_a = count_lines(data_one, size_one);
686 lc_b = count_lines(data_two, size_two);
687 fprintf(o->file,
688 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
689 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
690 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
691 line_prefix, fraginfo);
692 if (!o->irreversible_delete)
693 print_line_count(o->file, lc_a);
694 else
695 fprintf(o->file, "?,?");
696 fprintf(o->file, " +");
697 print_line_count(o->file, lc_b);
698 fprintf(o->file, " @@%s\n", reset);
699 if (lc_a && !o->irreversible_delete)
700 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
701 if (lc_b)
702 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
703 if (textconv_one)
704 free((char *)data_one);
705 if (textconv_two)
706 free((char *)data_two);
707 }
708
709 struct diff_words_buffer {
710 mmfile_t text;
711 long alloc;
712 struct diff_words_orig {
713 const char *begin, *end;
714 } *orig;
715 int orig_nr, orig_alloc;
716 };
717
718 static void diff_words_append(char *line, unsigned long len,
719 struct diff_words_buffer *buffer)
720 {
721 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
722 line++;
723 len--;
724 memcpy(buffer->text.ptr + buffer->text.size, line, len);
725 buffer->text.size += len;
726 buffer->text.ptr[buffer->text.size] = '\0';
727 }
728
729 struct diff_words_style_elem {
730 const char *prefix;
731 const char *suffix;
732 const char *color; /* NULL; filled in by the setup code if
733 * color is enabled */
734 };
735
736 struct diff_words_style {
737 enum diff_words_type type;
738 struct diff_words_style_elem new, old, ctx;
739 const char *newline;
740 };
741
742 static struct diff_words_style diff_words_styles[] = {
743 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
744 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
745 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
746 };
747
748 struct diff_words_data {
749 struct diff_words_buffer minus, plus;
750 const char *current_plus;
751 int last_minus;
752 struct diff_options *opt;
753 regex_t *word_regex;
754 enum diff_words_type type;
755 struct diff_words_style *style;
756 };
757
758 static int fn_out_diff_words_write_helper(FILE *fp,
759 struct diff_words_style_elem *st_el,
760 const char *newline,
761 size_t count, const char *buf,
762 const char *line_prefix)
763 {
764 int print = 0;
765
766 while (count) {
767 char *p = memchr(buf, '\n', count);
768 if (print)
769 fputs(line_prefix, fp);
770 if (p != buf) {
771 if (st_el->color && fputs(st_el->color, fp) < 0)
772 return -1;
773 if (fputs(st_el->prefix, fp) < 0 ||
774 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
775 fputs(st_el->suffix, fp) < 0)
776 return -1;
777 if (st_el->color && *st_el->color
778 && fputs(GIT_COLOR_RESET, fp) < 0)
779 return -1;
780 }
781 if (!p)
782 return 0;
783 if (fputs(newline, fp) < 0)
784 return -1;
785 count -= p + 1 - buf;
786 buf = p + 1;
787 print = 1;
788 }
789 return 0;
790 }
791
792 /*
793 * '--color-words' algorithm can be described as:
794 *
795 * 1. collect a the minus/plus lines of a diff hunk, divided into
796 * minus-lines and plus-lines;
797 *
798 * 2. break both minus-lines and plus-lines into words and
799 * place them into two mmfile_t with one word for each line;
800 *
801 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
802 *
803 * And for the common parts of the both file, we output the plus side text.
804 * diff_words->current_plus is used to trace the current position of the plus file
805 * which printed. diff_words->last_minus is used to trace the last minus word
806 * printed.
807 *
808 * For '--graph' to work with '--color-words', we need to output the graph prefix
809 * on each line of color words output. Generally, there are two conditions on
810 * which we should output the prefix.
811 *
812 * 1. diff_words->last_minus == 0 &&
813 * diff_words->current_plus == diff_words->plus.text.ptr
814 *
815 * that is: the plus text must start as a new line, and if there is no minus
816 * word printed, a graph prefix must be printed.
817 *
818 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
819 * *(diff_words->current_plus - 1) == '\n'
820 *
821 * that is: a graph prefix must be printed following a '\n'
822 */
823 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
824 {
825 if ((diff_words->last_minus == 0 &&
826 diff_words->current_plus == diff_words->plus.text.ptr) ||
827 (diff_words->current_plus > diff_words->plus.text.ptr &&
828 *(diff_words->current_plus - 1) == '\n')) {
829 return 1;
830 } else {
831 return 0;
832 }
833 }
834
835 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
836 {
837 struct diff_words_data *diff_words = priv;
838 struct diff_words_style *style = diff_words->style;
839 int minus_first, minus_len, plus_first, plus_len;
840 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
841 struct diff_options *opt = diff_words->opt;
842 const char *line_prefix;
843
844 if (line[0] != '@' || parse_hunk_header(line, len,
845 &minus_first, &minus_len, &plus_first, &plus_len))
846 return;
847
848 assert(opt);
849 line_prefix = diff_line_prefix(opt);
850
851 /* POSIX requires that first be decremented by one if len == 0... */
852 if (minus_len) {
853 minus_begin = diff_words->minus.orig[minus_first].begin;
854 minus_end =
855 diff_words->minus.orig[minus_first + minus_len - 1].end;
856 } else
857 minus_begin = minus_end =
858 diff_words->minus.orig[minus_first].end;
859
860 if (plus_len) {
861 plus_begin = diff_words->plus.orig[plus_first].begin;
862 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
863 } else
864 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
865
866 if (color_words_output_graph_prefix(diff_words)) {
867 fputs(line_prefix, diff_words->opt->file);
868 }
869 if (diff_words->current_plus != plus_begin) {
870 fn_out_diff_words_write_helper(diff_words->opt->file,
871 &style->ctx, style->newline,
872 plus_begin - diff_words->current_plus,
873 diff_words->current_plus, line_prefix);
874 if (*(plus_begin - 1) == '\n')
875 fputs(line_prefix, diff_words->opt->file);
876 }
877 if (minus_begin != minus_end) {
878 fn_out_diff_words_write_helper(diff_words->opt->file,
879 &style->old, style->newline,
880 minus_end - minus_begin, minus_begin,
881 line_prefix);
882 }
883 if (plus_begin != plus_end) {
884 fn_out_diff_words_write_helper(diff_words->opt->file,
885 &style->new, style->newline,
886 plus_end - plus_begin, plus_begin,
887 line_prefix);
888 }
889
890 diff_words->current_plus = plus_end;
891 diff_words->last_minus = minus_first;
892 }
893
894 /* This function starts looking at *begin, and returns 0 iff a word was found. */
895 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
896 int *begin, int *end)
897 {
898 if (word_regex && *begin < buffer->size) {
899 regmatch_t match[1];
900 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
901 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
902 '\n', match[0].rm_eo - match[0].rm_so);
903 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
904 *begin += match[0].rm_so;
905 return *begin >= *end;
906 }
907 return -1;
908 }
909
910 /* find the next word */
911 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
912 (*begin)++;
913 if (*begin >= buffer->size)
914 return -1;
915
916 /* find the end of the word */
917 *end = *begin + 1;
918 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
919 (*end)++;
920
921 return 0;
922 }
923
924 /*
925 * This function splits the words in buffer->text, stores the list with
926 * newline separator into out, and saves the offsets of the original words
927 * in buffer->orig.
928 */
929 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
930 regex_t *word_regex)
931 {
932 int i, j;
933 long alloc = 0;
934
935 out->size = 0;
936 out->ptr = NULL;
937
938 /* fake an empty "0th" word */
939 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
940 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
941 buffer->orig_nr = 1;
942
943 for (i = 0; i < buffer->text.size; i++) {
944 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
945 return;
946
947 /* store original boundaries */
948 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
949 buffer->orig_alloc);
950 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
951 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
952 buffer->orig_nr++;
953
954 /* store one word */
955 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
956 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
957 out->ptr[out->size + j - i] = '\n';
958 out->size += j - i + 1;
959
960 i = j - 1;
961 }
962 }
963
964 /* this executes the word diff on the accumulated buffers */
965 static void diff_words_show(struct diff_words_data *diff_words)
966 {
967 xpparam_t xpp;
968 xdemitconf_t xecfg;
969 mmfile_t minus, plus;
970 struct diff_words_style *style = diff_words->style;
971
972 struct diff_options *opt = diff_words->opt;
973 const char *line_prefix;
974
975 assert(opt);
976 line_prefix = diff_line_prefix(opt);
977
978 /* special case: only removal */
979 if (!diff_words->plus.text.size) {
980 fputs(line_prefix, diff_words->opt->file);
981 fn_out_diff_words_write_helper(diff_words->opt->file,
982 &style->old, style->newline,
983 diff_words->minus.text.size,
984 diff_words->minus.text.ptr, line_prefix);
985 diff_words->minus.text.size = 0;
986 return;
987 }
988
989 diff_words->current_plus = diff_words->plus.text.ptr;
990 diff_words->last_minus = 0;
991
992 memset(&xpp, 0, sizeof(xpp));
993 memset(&xecfg, 0, sizeof(xecfg));
994 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
995 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
996 xpp.flags = 0;
997 /* as only the hunk header will be parsed, we need a 0-context */
998 xecfg.ctxlen = 0;
999 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1000 &xpp, &xecfg);
1001 free(minus.ptr);
1002 free(plus.ptr);
1003 if (diff_words->current_plus != diff_words->plus.text.ptr +
1004 diff_words->plus.text.size) {
1005 if (color_words_output_graph_prefix(diff_words))
1006 fputs(line_prefix, diff_words->opt->file);
1007 fn_out_diff_words_write_helper(diff_words->opt->file,
1008 &style->ctx, style->newline,
1009 diff_words->plus.text.ptr + diff_words->plus.text.size
1010 - diff_words->current_plus, diff_words->current_plus,
1011 line_prefix);
1012 }
1013 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1014 }
1015
1016 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1017 static void diff_words_flush(struct emit_callback *ecbdata)
1018 {
1019 if (ecbdata->diff_words->minus.text.size ||
1020 ecbdata->diff_words->plus.text.size)
1021 diff_words_show(ecbdata->diff_words);
1022 }
1023
1024 static void diff_filespec_load_driver(struct diff_filespec *one)
1025 {
1026 /* Use already-loaded driver */
1027 if (one->driver)
1028 return;
1029
1030 if (S_ISREG(one->mode))
1031 one->driver = userdiff_find_by_path(one->path);
1032
1033 /* Fallback to default settings */
1034 if (!one->driver)
1035 one->driver = userdiff_find_by_name("default");
1036 }
1037
1038 static const char *userdiff_word_regex(struct diff_filespec *one)
1039 {
1040 diff_filespec_load_driver(one);
1041 return one->driver->word_regex;
1042 }
1043
1044 static void init_diff_words_data(struct emit_callback *ecbdata,
1045 struct diff_options *orig_opts,
1046 struct diff_filespec *one,
1047 struct diff_filespec *two)
1048 {
1049 int i;
1050 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1051 memcpy(o, orig_opts, sizeof(struct diff_options));
1052
1053 ecbdata->diff_words =
1054 xcalloc(1, sizeof(struct diff_words_data));
1055 ecbdata->diff_words->type = o->word_diff;
1056 ecbdata->diff_words->opt = o;
1057 if (!o->word_regex)
1058 o->word_regex = userdiff_word_regex(one);
1059 if (!o->word_regex)
1060 o->word_regex = userdiff_word_regex(two);
1061 if (!o->word_regex)
1062 o->word_regex = diff_word_regex_cfg;
1063 if (o->word_regex) {
1064 ecbdata->diff_words->word_regex = (regex_t *)
1065 xmalloc(sizeof(regex_t));
1066 if (regcomp(ecbdata->diff_words->word_regex,
1067 o->word_regex,
1068 REG_EXTENDED | REG_NEWLINE))
1069 die ("Invalid regular expression: %s",
1070 o->word_regex);
1071 }
1072 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1073 if (o->word_diff == diff_words_styles[i].type) {
1074 ecbdata->diff_words->style =
1075 &diff_words_styles[i];
1076 break;
1077 }
1078 }
1079 if (want_color(o->use_color)) {
1080 struct diff_words_style *st = ecbdata->diff_words->style;
1081 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1082 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1083 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
1084 }
1085 }
1086
1087 static void free_diff_words_data(struct emit_callback *ecbdata)
1088 {
1089 if (ecbdata->diff_words) {
1090 diff_words_flush(ecbdata);
1091 free (ecbdata->diff_words->opt);
1092 free (ecbdata->diff_words->minus.text.ptr);
1093 free (ecbdata->diff_words->minus.orig);
1094 free (ecbdata->diff_words->plus.text.ptr);
1095 free (ecbdata->diff_words->plus.orig);
1096 if (ecbdata->diff_words->word_regex) {
1097 regfree(ecbdata->diff_words->word_regex);
1098 free(ecbdata->diff_words->word_regex);
1099 }
1100 free(ecbdata->diff_words);
1101 ecbdata->diff_words = NULL;
1102 }
1103 }
1104
1105 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1106 {
1107 if (want_color(diff_use_color))
1108 return diff_colors[ix];
1109 return "";
1110 }
1111
1112 const char *diff_line_prefix(struct diff_options *opt)
1113 {
1114 struct strbuf *msgbuf;
1115 if (!opt->output_prefix)
1116 return "";
1117
1118 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1119 return msgbuf->buf;
1120 }
1121
1122 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1123 {
1124 const char *cp;
1125 unsigned long allot;
1126 size_t l = len;
1127
1128 if (ecb->truncate)
1129 return ecb->truncate(line, len);
1130 cp = line;
1131 allot = l;
1132 while (0 < l) {
1133 (void) utf8_width(&cp, &l);
1134 if (!cp)
1135 break; /* truncated in the middle? */
1136 }
1137 return allot - l;
1138 }
1139
1140 static void find_lno(const char *line, struct emit_callback *ecbdata)
1141 {
1142 const char *p;
1143 ecbdata->lno_in_preimage = 0;
1144 ecbdata->lno_in_postimage = 0;
1145 p = strchr(line, '-');
1146 if (!p)
1147 return; /* cannot happen */
1148 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1149 p = strchr(p, '+');
1150 if (!p)
1151 return; /* cannot happen */
1152 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1153 }
1154
1155 static void fn_out_consume(void *priv, char *line, unsigned long len)
1156 {
1157 struct emit_callback *ecbdata = priv;
1158 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1159 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1160 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1161 struct diff_options *o = ecbdata->opt;
1162 const char *line_prefix = diff_line_prefix(o);
1163
1164 if (ecbdata->header) {
1165 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1166 strbuf_reset(ecbdata->header);
1167 ecbdata->header = NULL;
1168 }
1169 *(ecbdata->found_changesp) = 1;
1170
1171 if (ecbdata->label_path[0]) {
1172 const char *name_a_tab, *name_b_tab;
1173
1174 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1175 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1176
1177 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1178 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1179 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1180 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1181 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1182 }
1183
1184 if (diff_suppress_blank_empty
1185 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1186 line[0] = '\n';
1187 len = 1;
1188 }
1189
1190 if (line[0] == '@') {
1191 if (ecbdata->diff_words)
1192 diff_words_flush(ecbdata);
1193 len = sane_truncate_line(ecbdata, line, len);
1194 find_lno(line, ecbdata);
1195 emit_hunk_header(ecbdata, line, len);
1196 if (line[len-1] != '\n')
1197 putc('\n', ecbdata->opt->file);
1198 return;
1199 }
1200
1201 if (len < 1) {
1202 emit_line(ecbdata->opt, reset, reset, line, len);
1203 if (ecbdata->diff_words
1204 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1205 fputs("~\n", ecbdata->opt->file);
1206 return;
1207 }
1208
1209 if (ecbdata->diff_words) {
1210 if (line[0] == '-') {
1211 diff_words_append(line, len,
1212 &ecbdata->diff_words->minus);
1213 return;
1214 } else if (line[0] == '+') {
1215 diff_words_append(line, len,
1216 &ecbdata->diff_words->plus);
1217 return;
1218 } else if (!prefixcmp(line, "\\ ")) {
1219 /*
1220 * Eat the "no newline at eof" marker as if we
1221 * saw a "+" or "-" line with nothing on it,
1222 * and return without diff_words_flush() to
1223 * defer processing. If this is the end of
1224 * preimage, more "+" lines may come after it.
1225 */
1226 return;
1227 }
1228 diff_words_flush(ecbdata);
1229 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1230 emit_line(ecbdata->opt, plain, reset, line, len);
1231 fputs("~\n", ecbdata->opt->file);
1232 } else {
1233 /*
1234 * Skip the prefix character, if any. With
1235 * diff_suppress_blank_empty, there may be
1236 * none.
1237 */
1238 if (line[0] != '\n') {
1239 line++;
1240 len--;
1241 }
1242 emit_line(ecbdata->opt, plain, reset, line, len);
1243 }
1244 return;
1245 }
1246
1247 if (line[0] != '+') {
1248 const char *color =
1249 diff_get_color(ecbdata->color_diff,
1250 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1251 ecbdata->lno_in_preimage++;
1252 if (line[0] == ' ')
1253 ecbdata->lno_in_postimage++;
1254 emit_line(ecbdata->opt, color, reset, line, len);
1255 } else {
1256 ecbdata->lno_in_postimage++;
1257 emit_add_line(reset, ecbdata, line + 1, len - 1);
1258 }
1259 }
1260
1261 static char *pprint_rename(const char *a, const char *b)
1262 {
1263 const char *old = a;
1264 const char *new = b;
1265 struct strbuf name = STRBUF_INIT;
1266 int pfx_length, sfx_length;
1267 int len_a = strlen(a);
1268 int len_b = strlen(b);
1269 int a_midlen, b_midlen;
1270 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1271 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1272
1273 if (qlen_a || qlen_b) {
1274 quote_c_style(a, &name, NULL, 0);
1275 strbuf_addstr(&name, " => ");
1276 quote_c_style(b, &name, NULL, 0);
1277 return strbuf_detach(&name, NULL);
1278 }
1279
1280 /* Find common prefix */
1281 pfx_length = 0;
1282 while (*old && *new && *old == *new) {
1283 if (*old == '/')
1284 pfx_length = old - a + 1;
1285 old++;
1286 new++;
1287 }
1288
1289 /* Find common suffix */
1290 old = a + len_a;
1291 new = b + len_b;
1292 sfx_length = 0;
1293 /*
1294 * Note:
1295 * if pfx_length is 0, old/new will never reach a - 1 because it
1296 * would mean the whole string is common suffix. But then, the
1297 * whole string would also be a common prefix, and we would not
1298 * have pfx_length equals 0.
1299 */
1300 while (a + pfx_length - 1 <= old &&
1301 b + pfx_length - 1 <= new &&
1302 *old == *new) {
1303 if (*old == '/')
1304 sfx_length = len_a - (old - a);
1305 old--;
1306 new--;
1307 }
1308
1309 /*
1310 * pfx{mid-a => mid-b}sfx
1311 * {pfx-a => pfx-b}sfx
1312 * pfx{sfx-a => sfx-b}
1313 * name-a => name-b
1314 */
1315 a_midlen = len_a - pfx_length - sfx_length;
1316 b_midlen = len_b - pfx_length - sfx_length;
1317 if (a_midlen < 0)
1318 a_midlen = 0;
1319 if (b_midlen < 0)
1320 b_midlen = 0;
1321
1322 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1323 if (pfx_length + sfx_length) {
1324 strbuf_add(&name, a, pfx_length);
1325 strbuf_addch(&name, '{');
1326 }
1327 strbuf_add(&name, a + pfx_length, a_midlen);
1328 strbuf_addstr(&name, " => ");
1329 strbuf_add(&name, b + pfx_length, b_midlen);
1330 if (pfx_length + sfx_length) {
1331 strbuf_addch(&name, '}');
1332 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1333 }
1334 return strbuf_detach(&name, NULL);
1335 }
1336
1337 struct diffstat_t {
1338 int nr;
1339 int alloc;
1340 struct diffstat_file {
1341 char *from_name;
1342 char *name;
1343 char *print_name;
1344 unsigned is_unmerged:1;
1345 unsigned is_binary:1;
1346 unsigned is_renamed:1;
1347 unsigned is_interesting:1;
1348 uintmax_t added, deleted;
1349 } **files;
1350 };
1351
1352 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1353 const char *name_a,
1354 const char *name_b)
1355 {
1356 struct diffstat_file *x;
1357 x = xcalloc(sizeof (*x), 1);
1358 if (diffstat->nr == diffstat->alloc) {
1359 diffstat->alloc = alloc_nr(diffstat->alloc);
1360 diffstat->files = xrealloc(diffstat->files,
1361 diffstat->alloc * sizeof(x));
1362 }
1363 diffstat->files[diffstat->nr++] = x;
1364 if (name_b) {
1365 x->from_name = xstrdup(name_a);
1366 x->name = xstrdup(name_b);
1367 x->is_renamed = 1;
1368 }
1369 else {
1370 x->from_name = NULL;
1371 x->name = xstrdup(name_a);
1372 }
1373 return x;
1374 }
1375
1376 static void diffstat_consume(void *priv, char *line, unsigned long len)
1377 {
1378 struct diffstat_t *diffstat = priv;
1379 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1380
1381 if (line[0] == '+')
1382 x->added++;
1383 else if (line[0] == '-')
1384 x->deleted++;
1385 }
1386
1387 const char mime_boundary_leader[] = "------------";
1388
1389 static int scale_linear(int it, int width, int max_change)
1390 {
1391 if (!it)
1392 return 0;
1393 /*
1394 * make sure that at least one '-' or '+' is printed if
1395 * there is any change to this path. The easiest way is to
1396 * scale linearly as if the alloted width is one column shorter
1397 * than it is, and then add 1 to the result.
1398 */
1399 return 1 + (it * (width - 1) / max_change);
1400 }
1401
1402 static void show_name(FILE *file,
1403 const char *prefix, const char *name, int len)
1404 {
1405 fprintf(file, " %s%-*s |", prefix, len, name);
1406 }
1407
1408 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1409 {
1410 if (cnt <= 0)
1411 return;
1412 fprintf(file, "%s", set);
1413 while (cnt--)
1414 putc(ch, file);
1415 fprintf(file, "%s", reset);
1416 }
1417
1418 static void fill_print_name(struct diffstat_file *file)
1419 {
1420 char *pname;
1421
1422 if (file->print_name)
1423 return;
1424
1425 if (!file->is_renamed) {
1426 struct strbuf buf = STRBUF_INIT;
1427 if (quote_c_style(file->name, &buf, NULL, 0)) {
1428 pname = strbuf_detach(&buf, NULL);
1429 } else {
1430 pname = file->name;
1431 strbuf_release(&buf);
1432 }
1433 } else {
1434 pname = pprint_rename(file->from_name, file->name);
1435 }
1436 file->print_name = pname;
1437 }
1438
1439 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1440 {
1441 struct strbuf sb = STRBUF_INIT;
1442 int ret;
1443
1444 if (!files) {
1445 assert(insertions == 0 && deletions == 0);
1446 return fprintf(fp, "%s\n", " 0 files changed");
1447 }
1448
1449 strbuf_addf(&sb,
1450 (files == 1) ? " %d file changed" : " %d files changed",
1451 files);
1452
1453 /*
1454 * For binary diff, the caller may want to print "x files
1455 * changed" with insertions == 0 && deletions == 0.
1456 *
1457 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1458 * is probably less confusing (i.e skip over "2 files changed
1459 * but nothing about added/removed lines? Is this a bug in Git?").
1460 */
1461 if (insertions || deletions == 0) {
1462 /*
1463 * TRANSLATORS: "+" in (+) is a line addition marker;
1464 * do not translate it.
1465 */
1466 strbuf_addf(&sb,
1467 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1468 insertions);
1469 }
1470
1471 if (deletions || insertions == 0) {
1472 /*
1473 * TRANSLATORS: "-" in (-) is a line removal marker;
1474 * do not translate it.
1475 */
1476 strbuf_addf(&sb,
1477 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1478 deletions);
1479 }
1480 strbuf_addch(&sb, '\n');
1481 ret = fputs(sb.buf, fp);
1482 strbuf_release(&sb);
1483 return ret;
1484 }
1485
1486 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1487 {
1488 int i, len, add, del, adds = 0, dels = 0;
1489 uintmax_t max_change = 0, max_len = 0;
1490 int total_files = data->nr, count;
1491 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1492 const char *reset, *add_c, *del_c;
1493 const char *line_prefix = "";
1494 int extra_shown = 0;
1495
1496 if (data->nr == 0)
1497 return;
1498
1499 line_prefix = diff_line_prefix(options);
1500 count = options->stat_count ? options->stat_count : data->nr;
1501
1502 reset = diff_get_color_opt(options, DIFF_RESET);
1503 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1504 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1505
1506 /*
1507 * Find the longest filename and max number of changes
1508 */
1509 for (i = 0; (i < count) && (i < data->nr); i++) {
1510 struct diffstat_file *file = data->files[i];
1511 uintmax_t change = file->added + file->deleted;
1512
1513 if (!file->is_interesting && (change == 0)) {
1514 count++; /* not shown == room for one more */
1515 continue;
1516 }
1517 fill_print_name(file);
1518 len = strlen(file->print_name);
1519 if (max_len < len)
1520 max_len = len;
1521
1522 if (file->is_unmerged) {
1523 /* "Unmerged" is 8 characters */
1524 bin_width = bin_width < 8 ? 8 : bin_width;
1525 continue;
1526 }
1527 if (file->is_binary) {
1528 /* "Bin XXX -> YYY bytes" */
1529 int w = 14 + decimal_width(file->added)
1530 + decimal_width(file->deleted);
1531 bin_width = bin_width < w ? w : bin_width;
1532 /* Display change counts aligned with "Bin" */
1533 number_width = 3;
1534 continue;
1535 }
1536
1537 if (max_change < change)
1538 max_change = change;
1539 }
1540 count = i; /* where we can stop scanning in data->files[] */
1541
1542 /*
1543 * We have width = stat_width or term_columns() columns total.
1544 * We want a maximum of min(max_len, stat_name_width) for the name part.
1545 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1546 * We also need 1 for " " and 4 + decimal_width(max_change)
1547 * for " | NNNN " and one the empty column at the end, altogether
1548 * 6 + decimal_width(max_change).
1549 *
1550 * If there's not enough space, we will use the smaller of
1551 * stat_name_width (if set) and 5/8*width for the filename,
1552 * and the rest for constant elements + graph part, but no more
1553 * than stat_graph_width for the graph part.
1554 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1555 * for the standard terminal size).
1556 *
1557 * In other words: stat_width limits the maximum width, and
1558 * stat_name_width fixes the maximum width of the filename,
1559 * and is also used to divide available columns if there
1560 * aren't enough.
1561 *
1562 * Binary files are displayed with "Bin XXX -> YYY bytes"
1563 * instead of the change count and graph. This part is treated
1564 * similarly to the graph part, except that it is not
1565 * "scaled". If total width is too small to accomodate the
1566 * guaranteed minimum width of the filename part and the
1567 * separators and this message, this message will "overflow"
1568 * making the line longer than the maximum width.
1569 */
1570
1571 if (options->stat_width == -1)
1572 width = term_columns() - options->output_prefix_length;
1573 else
1574 width = options->stat_width ? options->stat_width : 80;
1575 number_width = decimal_width(max_change) > number_width ?
1576 decimal_width(max_change) : number_width;
1577
1578 if (options->stat_graph_width == -1)
1579 options->stat_graph_width = diff_stat_graph_width;
1580
1581 /*
1582 * Guarantee 3/8*16==6 for the graph part
1583 * and 5/8*16==10 for the filename part
1584 */
1585 if (width < 16 + 6 + number_width)
1586 width = 16 + 6 + number_width;
1587
1588 /*
1589 * First assign sizes that are wanted, ignoring available width.
1590 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1591 * starting from "XXX" should fit in graph_width.
1592 */
1593 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1594 if (options->stat_graph_width &&
1595 options->stat_graph_width < graph_width)
1596 graph_width = options->stat_graph_width;
1597
1598 name_width = (options->stat_name_width > 0 &&
1599 options->stat_name_width < max_len) ?
1600 options->stat_name_width : max_len;
1601
1602 /*
1603 * Adjust adjustable widths not to exceed maximum width
1604 */
1605 if (name_width + number_width + 6 + graph_width > width) {
1606 if (graph_width > width * 3/8 - number_width - 6) {
1607 graph_width = width * 3/8 - number_width - 6;
1608 if (graph_width < 6)
1609 graph_width = 6;
1610 }
1611
1612 if (options->stat_graph_width &&
1613 graph_width > options->stat_graph_width)
1614 graph_width = options->stat_graph_width;
1615 if (name_width > width - number_width - 6 - graph_width)
1616 name_width = width - number_width - 6 - graph_width;
1617 else
1618 graph_width = width - number_width - 6 - name_width;
1619 }
1620
1621 /*
1622 * From here name_width is the width of the name area,
1623 * and graph_width is the width of the graph area.
1624 * max_change is used to scale graph properly.
1625 */
1626 for (i = 0; i < count; i++) {
1627 const char *prefix = "";
1628 struct diffstat_file *file = data->files[i];
1629 char *name = file->print_name;
1630 uintmax_t added = file->added;
1631 uintmax_t deleted = file->deleted;
1632 int name_len;
1633
1634 if (!file->is_interesting && (added + deleted == 0))
1635 continue;
1636
1637 /*
1638 * "scale" the filename
1639 */
1640 len = name_width;
1641 name_len = strlen(name);
1642 if (name_width < name_len) {
1643 char *slash;
1644 prefix = "...";
1645 len -= 3;
1646 name += name_len - len;
1647 slash = strchr(name, '/');
1648 if (slash)
1649 name = slash;
1650 }
1651
1652 if (file->is_binary) {
1653 fprintf(options->file, "%s", line_prefix);
1654 show_name(options->file, prefix, name, len);
1655 fprintf(options->file, " %*s", number_width, "Bin");
1656 if (!added && !deleted) {
1657 putc('\n', options->file);
1658 continue;
1659 }
1660 fprintf(options->file, " %s%"PRIuMAX"%s",
1661 del_c, deleted, reset);
1662 fprintf(options->file, " -> ");
1663 fprintf(options->file, "%s%"PRIuMAX"%s",
1664 add_c, added, reset);
1665 fprintf(options->file, " bytes");
1666 fprintf(options->file, "\n");
1667 continue;
1668 }
1669 else if (file->is_unmerged) {
1670 fprintf(options->file, "%s", line_prefix);
1671 show_name(options->file, prefix, name, len);
1672 fprintf(options->file, " Unmerged\n");
1673 continue;
1674 }
1675
1676 /*
1677 * scale the add/delete
1678 */
1679 add = added;
1680 del = deleted;
1681
1682 if (graph_width <= max_change) {
1683 int total = add + del;
1684
1685 total = scale_linear(add + del, graph_width, max_change);
1686 if (total < 2 && add && del)
1687 /* width >= 2 due to the sanity check */
1688 total = 2;
1689 if (add < del) {
1690 add = scale_linear(add, graph_width, max_change);
1691 del = total - add;
1692 } else {
1693 del = scale_linear(del, graph_width, max_change);
1694 add = total - del;
1695 }
1696 }
1697 fprintf(options->file, "%s", line_prefix);
1698 show_name(options->file, prefix, name, len);
1699 fprintf(options->file, " %*"PRIuMAX"%s",
1700 number_width, added + deleted,
1701 added + deleted ? " " : "");
1702 show_graph(options->file, '+', add, add_c, reset);
1703 show_graph(options->file, '-', del, del_c, reset);
1704 fprintf(options->file, "\n");
1705 }
1706
1707 for (i = 0; i < data->nr; i++) {
1708 struct diffstat_file *file = data->files[i];
1709 uintmax_t added = file->added;
1710 uintmax_t deleted = file->deleted;
1711
1712 if (file->is_unmerged ||
1713 (!file->is_interesting && (added + deleted == 0))) {
1714 total_files--;
1715 continue;
1716 }
1717
1718 if (!file->is_binary) {
1719 adds += added;
1720 dels += deleted;
1721 }
1722 if (i < count)
1723 continue;
1724 if (!extra_shown)
1725 fprintf(options->file, "%s ...\n", line_prefix);
1726 extra_shown = 1;
1727 }
1728 fprintf(options->file, "%s", line_prefix);
1729 print_stat_summary(options->file, total_files, adds, dels);
1730 }
1731
1732 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1733 {
1734 int i, adds = 0, dels = 0, total_files = data->nr;
1735
1736 if (data->nr == 0)
1737 return;
1738
1739 for (i = 0; i < data->nr; i++) {
1740 int added = data->files[i]->added;
1741 int deleted= data->files[i]->deleted;
1742
1743 if (data->files[i]->is_unmerged ||
1744 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1745 total_files--;
1746 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1747 adds += added;
1748 dels += deleted;
1749 }
1750 }
1751 fprintf(options->file, "%s", diff_line_prefix(options));
1752 print_stat_summary(options->file, total_files, adds, dels);
1753 }
1754
1755 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1756 {
1757 int i;
1758
1759 if (data->nr == 0)
1760 return;
1761
1762 for (i = 0; i < data->nr; i++) {
1763 struct diffstat_file *file = data->files[i];
1764
1765 fprintf(options->file, "%s", diff_line_prefix(options));
1766
1767 if (file->is_binary)
1768 fprintf(options->file, "-\t-\t");
1769 else
1770 fprintf(options->file,
1771 "%"PRIuMAX"\t%"PRIuMAX"\t",
1772 file->added, file->deleted);
1773 if (options->line_termination) {
1774 fill_print_name(file);
1775 if (!file->is_renamed)
1776 write_name_quoted(file->name, options->file,
1777 options->line_termination);
1778 else {
1779 fputs(file->print_name, options->file);
1780 putc(options->line_termination, options->file);
1781 }
1782 } else {
1783 if (file->is_renamed) {
1784 putc('\0', options->file);
1785 write_name_quoted(file->from_name, options->file, '\0');
1786 }
1787 write_name_quoted(file->name, options->file, '\0');
1788 }
1789 }
1790 }
1791
1792 struct dirstat_file {
1793 const char *name;
1794 unsigned long changed;
1795 };
1796
1797 struct dirstat_dir {
1798 struct dirstat_file *files;
1799 int alloc, nr, permille, cumulative;
1800 };
1801
1802 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1803 unsigned long changed, const char *base, int baselen)
1804 {
1805 unsigned long this_dir = 0;
1806 unsigned int sources = 0;
1807 const char *line_prefix = diff_line_prefix(opt);
1808
1809 while (dir->nr) {
1810 struct dirstat_file *f = dir->files;
1811 int namelen = strlen(f->name);
1812 unsigned long this;
1813 char *slash;
1814
1815 if (namelen < baselen)
1816 break;
1817 if (memcmp(f->name, base, baselen))
1818 break;
1819 slash = strchr(f->name + baselen, '/');
1820 if (slash) {
1821 int newbaselen = slash + 1 - f->name;
1822 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1823 sources++;
1824 } else {
1825 this = f->changed;
1826 dir->files++;
1827 dir->nr--;
1828 sources += 2;
1829 }
1830 this_dir += this;
1831 }
1832
1833 /*
1834 * We don't report dirstat's for
1835 * - the top level
1836 * - or cases where everything came from a single directory
1837 * under this directory (sources == 1).
1838 */
1839 if (baselen && sources != 1) {
1840 if (this_dir) {
1841 int permille = this_dir * 1000 / changed;
1842 if (permille >= dir->permille) {
1843 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1844 permille / 10, permille % 10, baselen, base);
1845 if (!dir->cumulative)
1846 return 0;
1847 }
1848 }
1849 }
1850 return this_dir;
1851 }
1852
1853 static int dirstat_compare(const void *_a, const void *_b)
1854 {
1855 const struct dirstat_file *a = _a;
1856 const struct dirstat_file *b = _b;
1857 return strcmp(a->name, b->name);
1858 }
1859
1860 static void show_dirstat(struct diff_options *options)
1861 {
1862 int i;
1863 unsigned long changed;
1864 struct dirstat_dir dir;
1865 struct diff_queue_struct *q = &diff_queued_diff;
1866
1867 dir.files = NULL;
1868 dir.alloc = 0;
1869 dir.nr = 0;
1870 dir.permille = options->dirstat_permille;
1871 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1872
1873 changed = 0;
1874 for (i = 0; i < q->nr; i++) {
1875 struct diff_filepair *p = q->queue[i];
1876 const char *name;
1877 unsigned long copied, added, damage;
1878 int content_changed;
1879
1880 name = p->two->path ? p->two->path : p->one->path;
1881
1882 if (p->one->sha1_valid && p->two->sha1_valid)
1883 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1884 else
1885 content_changed = 1;
1886
1887 if (!content_changed) {
1888 /*
1889 * The SHA1 has not changed, so pre-/post-content is
1890 * identical. We can therefore skip looking at the
1891 * file contents altogether.
1892 */
1893 damage = 0;
1894 goto found_damage;
1895 }
1896
1897 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1898 /*
1899 * In --dirstat-by-file mode, we don't really need to
1900 * look at the actual file contents at all.
1901 * The fact that the SHA1 changed is enough for us to
1902 * add this file to the list of results
1903 * (with each file contributing equal damage).
1904 */
1905 damage = 1;
1906 goto found_damage;
1907 }
1908
1909 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1910 diff_populate_filespec(p->one, 0);
1911 diff_populate_filespec(p->two, 0);
1912 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1913 &copied, &added);
1914 diff_free_filespec_data(p->one);
1915 diff_free_filespec_data(p->two);
1916 } else if (DIFF_FILE_VALID(p->one)) {
1917 diff_populate_filespec(p->one, 1);
1918 copied = added = 0;
1919 diff_free_filespec_data(p->one);
1920 } else if (DIFF_FILE_VALID(p->two)) {
1921 diff_populate_filespec(p->two, 1);
1922 copied = 0;
1923 added = p->two->size;
1924 diff_free_filespec_data(p->two);
1925 } else
1926 continue;
1927
1928 /*
1929 * Original minus copied is the removed material,
1930 * added is the new material. They are both damages
1931 * made to the preimage.
1932 * If the resulting damage is zero, we know that
1933 * diffcore_count_changes() considers the two entries to
1934 * be identical, but since content_changed is true, we
1935 * know that there must have been _some_ kind of change,
1936 * so we force all entries to have damage > 0.
1937 */
1938 damage = (p->one->size - copied) + added;
1939 if (!damage)
1940 damage = 1;
1941
1942 found_damage:
1943 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1944 dir.files[dir.nr].name = name;
1945 dir.files[dir.nr].changed = damage;
1946 changed += damage;
1947 dir.nr++;
1948 }
1949
1950 /* This can happen even with many files, if everything was renames */
1951 if (!changed)
1952 return;
1953
1954 /* Show all directories with more than x% of the changes */
1955 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1956 gather_dirstat(options, &dir, changed, "", 0);
1957 }
1958
1959 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1960 {
1961 int i;
1962 unsigned long changed;
1963 struct dirstat_dir dir;
1964
1965 if (data->nr == 0)
1966 return;
1967
1968 dir.files = NULL;
1969 dir.alloc = 0;
1970 dir.nr = 0;
1971 dir.permille = options->dirstat_permille;
1972 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1973
1974 changed = 0;
1975 for (i = 0; i < data->nr; i++) {
1976 struct diffstat_file *file = data->files[i];
1977 unsigned long damage = file->added + file->deleted;
1978 if (file->is_binary)
1979 /*
1980 * binary files counts bytes, not lines. Must find some
1981 * way to normalize binary bytes vs. textual lines.
1982 * The following heuristic assumes that there are 64
1983 * bytes per "line".
1984 * This is stupid and ugly, but very cheap...
1985 */
1986 damage = (damage + 63) / 64;
1987 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1988 dir.files[dir.nr].name = file->name;
1989 dir.files[dir.nr].changed = damage;
1990 changed += damage;
1991 dir.nr++;
1992 }
1993
1994 /* This can happen even with many files, if everything was renames */
1995 if (!changed)
1996 return;
1997
1998 /* Show all directories with more than x% of the changes */
1999 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
2000 gather_dirstat(options, &dir, changed, "", 0);
2001 }
2002
2003 static void free_diffstat_info(struct diffstat_t *diffstat)
2004 {
2005 int i;
2006 for (i = 0; i < diffstat->nr; i++) {
2007 struct diffstat_file *f = diffstat->files[i];
2008 if (f->name != f->print_name)
2009 free(f->print_name);
2010 free(f->name);
2011 free(f->from_name);
2012 free(f);
2013 }
2014 free(diffstat->files);
2015 }
2016
2017 struct checkdiff_t {
2018 const char *filename;
2019 int lineno;
2020 int conflict_marker_size;
2021 struct diff_options *o;
2022 unsigned ws_rule;
2023 unsigned status;
2024 };
2025
2026 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2027 {
2028 char firstchar;
2029 int cnt;
2030
2031 if (len < marker_size + 1)
2032 return 0;
2033 firstchar = line[0];
2034 switch (firstchar) {
2035 case '=': case '>': case '<': case '|':
2036 break;
2037 default:
2038 return 0;
2039 }
2040 for (cnt = 1; cnt < marker_size; cnt++)
2041 if (line[cnt] != firstchar)
2042 return 0;
2043 /* line[1] thru line[marker_size-1] are same as firstchar */
2044 if (len < marker_size + 1 || !isspace(line[marker_size]))
2045 return 0;
2046 return 1;
2047 }
2048
2049 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2050 {
2051 struct checkdiff_t *data = priv;
2052 int marker_size = data->conflict_marker_size;
2053 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2054 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2055 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2056 char *err;
2057 const char *line_prefix;
2058
2059 assert(data->o);
2060 line_prefix = diff_line_prefix(data->o);
2061
2062 if (line[0] == '+') {
2063 unsigned bad;
2064 data->lineno++;
2065 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2066 data->status |= 1;
2067 fprintf(data->o->file,
2068 "%s%s:%d: leftover conflict marker\n",
2069 line_prefix, data->filename, data->lineno);
2070 }
2071 bad = ws_check(line + 1, len - 1, data->ws_rule);
2072 if (!bad)
2073 return;
2074 data->status |= bad;
2075 err = whitespace_error_string(bad);
2076 fprintf(data->o->file, "%s%s:%d: %s.\n",
2077 line_prefix, data->filename, data->lineno, err);
2078 free(err);
2079 emit_line(data->o, set, reset, line, 1);
2080 ws_check_emit(line + 1, len - 1, data->ws_rule,
2081 data->o->file, set, reset, ws);
2082 } else if (line[0] == ' ') {
2083 data->lineno++;
2084 } else if (line[0] == '@') {
2085 char *plus = strchr(line, '+');
2086 if (plus)
2087 data->lineno = strtol(plus, NULL, 10) - 1;
2088 else
2089 die("invalid diff");
2090 }
2091 }
2092
2093 static unsigned char *deflate_it(char *data,
2094 unsigned long size,
2095 unsigned long *result_size)
2096 {
2097 int bound;
2098 unsigned char *deflated;
2099 git_zstream stream;
2100
2101 memset(&stream, 0, sizeof(stream));
2102 git_deflate_init(&stream, zlib_compression_level);
2103 bound = git_deflate_bound(&stream, size);
2104 deflated = xmalloc(bound);
2105 stream.next_out = deflated;
2106 stream.avail_out = bound;
2107
2108 stream.next_in = (unsigned char *)data;
2109 stream.avail_in = size;
2110 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2111 ; /* nothing */
2112 git_deflate_end(&stream);
2113 *result_size = stream.total_out;
2114 return deflated;
2115 }
2116
2117 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2118 const char *prefix)
2119 {
2120 void *cp;
2121 void *delta;
2122 void *deflated;
2123 void *data;
2124 unsigned long orig_size;
2125 unsigned long delta_size;
2126 unsigned long deflate_size;
2127 unsigned long data_size;
2128
2129 /* We could do deflated delta, or we could do just deflated two,
2130 * whichever is smaller.
2131 */
2132 delta = NULL;
2133 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2134 if (one->size && two->size) {
2135 delta = diff_delta(one->ptr, one->size,
2136 two->ptr, two->size,
2137 &delta_size, deflate_size);
2138 if (delta) {
2139 void *to_free = delta;
2140 orig_size = delta_size;
2141 delta = deflate_it(delta, delta_size, &delta_size);
2142 free(to_free);
2143 }
2144 }
2145
2146 if (delta && delta_size < deflate_size) {
2147 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2148 free(deflated);
2149 data = delta;
2150 data_size = delta_size;
2151 }
2152 else {
2153 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2154 free(delta);
2155 data = deflated;
2156 data_size = deflate_size;
2157 }
2158
2159 /* emit data encoded in base85 */
2160 cp = data;
2161 while (data_size) {
2162 int bytes = (52 < data_size) ? 52 : data_size;
2163 char line[70];
2164 data_size -= bytes;
2165 if (bytes <= 26)
2166 line[0] = bytes + 'A' - 1;
2167 else
2168 line[0] = bytes - 26 + 'a' - 1;
2169 encode_85(line + 1, cp, bytes);
2170 cp = (char *) cp + bytes;
2171 fprintf(file, "%s", prefix);
2172 fputs(line, file);
2173 fputc('\n', file);
2174 }
2175 fprintf(file, "%s\n", prefix);
2176 free(data);
2177 }
2178
2179 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2180 const char *prefix)
2181 {
2182 fprintf(file, "%sGIT binary patch\n", prefix);
2183 emit_binary_diff_body(file, one, two, prefix);
2184 emit_binary_diff_body(file, two, one, prefix);
2185 }
2186
2187 int diff_filespec_is_binary(struct diff_filespec *one)
2188 {
2189 if (one->is_binary == -1) {
2190 diff_filespec_load_driver(one);
2191 if (one->driver->binary != -1)
2192 one->is_binary = one->driver->binary;
2193 else {
2194 if (!one->data && DIFF_FILE_VALID(one))
2195 diff_populate_filespec(one, 0);
2196 if (one->data)
2197 one->is_binary = buffer_is_binary(one->data,
2198 one->size);
2199 if (one->is_binary == -1)
2200 one->is_binary = 0;
2201 }
2202 }
2203 return one->is_binary;
2204 }
2205
2206 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2207 {
2208 diff_filespec_load_driver(one);
2209 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2210 }
2211
2212 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2213 {
2214 if (!options->a_prefix)
2215 options->a_prefix = a;
2216 if (!options->b_prefix)
2217 options->b_prefix = b;
2218 }
2219
2220 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2221 {
2222 if (!DIFF_FILE_VALID(one))
2223 return NULL;
2224
2225 diff_filespec_load_driver(one);
2226 return userdiff_get_textconv(one->driver);
2227 }
2228
2229 static void builtin_diff(const char *name_a,
2230 const char *name_b,
2231 struct diff_filespec *one,
2232 struct diff_filespec *two,
2233 const char *xfrm_msg,
2234 int must_show_header,
2235 struct diff_options *o,
2236 int complete_rewrite)
2237 {
2238 mmfile_t mf1, mf2;
2239 const char *lbl[2];
2240 char *a_one, *b_two;
2241 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2242 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2243 const char *a_prefix, *b_prefix;
2244 struct userdiff_driver *textconv_one = NULL;
2245 struct userdiff_driver *textconv_two = NULL;
2246 struct strbuf header = STRBUF_INIT;
2247 const char *line_prefix = diff_line_prefix(o);
2248
2249 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2250 (!one->mode || S_ISGITLINK(one->mode)) &&
2251 (!two->mode || S_ISGITLINK(two->mode))) {
2252 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2253 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2254 show_submodule_summary(o->file, one ? one->path : two->path,
2255 one->sha1, two->sha1, two->dirty_submodule,
2256 meta, del, add, reset);
2257 return;
2258 }
2259
2260 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2261 textconv_one = get_textconv(one);
2262 textconv_two = get_textconv(two);
2263 }
2264
2265 diff_set_mnemonic_prefix(o, "a/", "b/");
2266 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2267 a_prefix = o->b_prefix;
2268 b_prefix = o->a_prefix;
2269 } else {
2270 a_prefix = o->a_prefix;
2271 b_prefix = o->b_prefix;
2272 }
2273
2274 /* Never use a non-valid filename anywhere if at all possible */
2275 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2276 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2277
2278 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2279 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2280 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2281 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2282 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2283 if (lbl[0][0] == '/') {
2284 /* /dev/null */
2285 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2286 if (xfrm_msg)
2287 strbuf_addstr(&header, xfrm_msg);
2288 must_show_header = 1;
2289 }
2290 else if (lbl[1][0] == '/') {
2291 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2292 if (xfrm_msg)
2293 strbuf_addstr(&header, xfrm_msg);
2294 must_show_header = 1;
2295 }
2296 else {
2297 if (one->mode != two->mode) {
2298 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2299 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2300 must_show_header = 1;
2301 }
2302 if (xfrm_msg)
2303 strbuf_addstr(&header, xfrm_msg);
2304
2305 /*
2306 * we do not run diff between different kind
2307 * of objects.
2308 */
2309 if ((one->mode ^ two->mode) & S_IFMT)
2310 goto free_ab_and_return;
2311 if (complete_rewrite &&
2312 (textconv_one || !diff_filespec_is_binary(one)) &&
2313 (textconv_two || !diff_filespec_is_binary(two))) {
2314 fprintf(o->file, "%s", header.buf);
2315 strbuf_reset(&header);
2316 emit_rewrite_diff(name_a, name_b, one, two,
2317 textconv_one, textconv_two, o);
2318 o->found_changes = 1;
2319 goto free_ab_and_return;
2320 }
2321 }
2322
2323 if (o->irreversible_delete && lbl[1][0] == '/') {
2324 fprintf(o->file, "%s", header.buf);
2325 strbuf_reset(&header);
2326 goto free_ab_and_return;
2327 } else if (!DIFF_OPT_TST(o, TEXT) &&
2328 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2329 (!textconv_two && diff_filespec_is_binary(two)) )) {
2330 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2331 die("unable to read files to diff");
2332 /* Quite common confusing case */
2333 if (mf1.size == mf2.size &&
2334 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2335 if (must_show_header)
2336 fprintf(o->file, "%s", header.buf);
2337 goto free_ab_and_return;
2338 }
2339 fprintf(o->file, "%s", header.buf);
2340 strbuf_reset(&header);
2341 if (DIFF_OPT_TST(o, BINARY))
2342 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2343 else
2344 fprintf(o->file, "%sBinary files %s and %s differ\n",
2345 line_prefix, lbl[0], lbl[1]);
2346 o->found_changes = 1;
2347 } else {
2348 /* Crazy xdl interfaces.. */
2349 const char *diffopts = getenv("GIT_DIFF_OPTS");
2350 xpparam_t xpp;
2351 xdemitconf_t xecfg;
2352 struct emit_callback ecbdata;
2353 const struct userdiff_funcname *pe;
2354
2355 if (must_show_header) {
2356 fprintf(o->file, "%s", header.buf);
2357 strbuf_reset(&header);
2358 }
2359
2360 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2361 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2362
2363 pe = diff_funcname_pattern(one);
2364 if (!pe)
2365 pe = diff_funcname_pattern(two);
2366
2367 memset(&xpp, 0, sizeof(xpp));
2368 memset(&xecfg, 0, sizeof(xecfg));
2369 memset(&ecbdata, 0, sizeof(ecbdata));
2370 ecbdata.label_path = lbl;
2371 ecbdata.color_diff = want_color(o->use_color);
2372 ecbdata.found_changesp = &o->found_changes;
2373 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2374 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2375 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2376 ecbdata.opt = o;
2377 ecbdata.header = header.len ? &header : NULL;
2378 xpp.flags = o->xdl_opts;
2379 xecfg.ctxlen = o->context;
2380 xecfg.interhunkctxlen = o->interhunkcontext;
2381 xecfg.flags = XDL_EMIT_FUNCNAMES;
2382 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2383 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2384 if (pe)
2385 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2386 if (!diffopts)
2387 ;
2388 else if (!prefixcmp(diffopts, "--unified="))
2389 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2390 else if (!prefixcmp(diffopts, "-u"))
2391 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2392 if (o->word_diff)
2393 init_diff_words_data(&ecbdata, o, one, two);
2394 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2395 &xpp, &xecfg);
2396 if (o->word_diff)
2397 free_diff_words_data(&ecbdata);
2398 if (textconv_one)
2399 free(mf1.ptr);
2400 if (textconv_two)
2401 free(mf2.ptr);
2402 xdiff_clear_find_func(&xecfg);
2403 }
2404
2405 free_ab_and_return:
2406 strbuf_release(&header);
2407 diff_free_filespec_data(one);
2408 diff_free_filespec_data(two);
2409 free(a_one);
2410 free(b_two);
2411 return;
2412 }
2413
2414 static void builtin_diffstat(const char *name_a, const char *name_b,
2415 struct diff_filespec *one,
2416 struct diff_filespec *two,
2417 struct diffstat_t *diffstat,
2418 struct diff_options *o,
2419 struct diff_filepair *p)
2420 {
2421 mmfile_t mf1, mf2;
2422 struct diffstat_file *data;
2423 int same_contents;
2424 int complete_rewrite = 0;
2425
2426 if (!DIFF_PAIR_UNMERGED(p)) {
2427 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2428 complete_rewrite = 1;
2429 }
2430
2431 data = diffstat_add(diffstat, name_a, name_b);
2432 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2433
2434 if (!one || !two) {
2435 data->is_unmerged = 1;
2436 return;
2437 }
2438
2439 same_contents = !hashcmp(one->sha1, two->sha1);
2440
2441 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2442 data->is_binary = 1;
2443 if (same_contents) {
2444 data->added = 0;
2445 data->deleted = 0;
2446 } else {
2447 data->added = diff_filespec_size(two);
2448 data->deleted = diff_filespec_size(one);
2449 }
2450 }
2451
2452 else if (complete_rewrite) {
2453 diff_populate_filespec(one, 0);
2454 diff_populate_filespec(two, 0);
2455 data->deleted = count_lines(one->data, one->size);
2456 data->added = count_lines(two->data, two->size);
2457 }
2458
2459 else if (!same_contents) {
2460 /* Crazy xdl interfaces.. */
2461 xpparam_t xpp;
2462 xdemitconf_t xecfg;
2463
2464 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2465 die("unable to read files to diff");
2466
2467 memset(&xpp, 0, sizeof(xpp));
2468 memset(&xecfg, 0, sizeof(xecfg));
2469 xpp.flags = o->xdl_opts;
2470 xecfg.ctxlen = o->context;
2471 xecfg.interhunkctxlen = o->interhunkcontext;
2472 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2473 &xpp, &xecfg);
2474 }
2475
2476 diff_free_filespec_data(one);
2477 diff_free_filespec_data(two);
2478 }
2479
2480 static void builtin_checkdiff(const char *name_a, const char *name_b,
2481 const char *attr_path,
2482 struct diff_filespec *one,
2483 struct diff_filespec *two,
2484 struct diff_options *o)
2485 {
2486 mmfile_t mf1, mf2;
2487 struct checkdiff_t data;
2488
2489 if (!two)
2490 return;
2491
2492 memset(&data, 0, sizeof(data));
2493 data.filename = name_b ? name_b : name_a;
2494 data.lineno = 0;
2495 data.o = o;
2496 data.ws_rule = whitespace_rule(attr_path);
2497 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2498
2499 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2500 die("unable to read files to diff");
2501
2502 /*
2503 * All the other codepaths check both sides, but not checking
2504 * the "old" side here is deliberate. We are checking the newly
2505 * introduced changes, and as long as the "new" side is text, we
2506 * can and should check what it introduces.
2507 */
2508 if (diff_filespec_is_binary(two))
2509 goto free_and_return;
2510 else {
2511 /* Crazy xdl interfaces.. */
2512 xpparam_t xpp;
2513 xdemitconf_t xecfg;
2514
2515 memset(&xpp, 0, sizeof(xpp));
2516 memset(&xecfg, 0, sizeof(xecfg));
2517 xecfg.ctxlen = 1; /* at least one context line */
2518 xpp.flags = 0;
2519 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2520 &xpp, &xecfg);
2521
2522 if (data.ws_rule & WS_BLANK_AT_EOF) {
2523 struct emit_callback ecbdata;
2524 int blank_at_eof;
2525
2526 ecbdata.ws_rule = data.ws_rule;
2527 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2528 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2529
2530 if (blank_at_eof) {
2531 static char *err;
2532 if (!err)
2533 err = whitespace_error_string(WS_BLANK_AT_EOF);
2534 fprintf(o->file, "%s:%d: %s.\n",
2535 data.filename, blank_at_eof, err);
2536 data.status = 1; /* report errors */
2537 }
2538 }
2539 }
2540 free_and_return:
2541 diff_free_filespec_data(one);
2542 diff_free_filespec_data(two);
2543 if (data.status)
2544 DIFF_OPT_SET(o, CHECK_FAILED);
2545 }
2546
2547 struct diff_filespec *alloc_filespec(const char *path)
2548 {
2549 int namelen = strlen(path);
2550 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2551
2552 memset(spec, 0, sizeof(*spec));
2553 spec->path = (char *)(spec + 1);
2554 memcpy(spec->path, path, namelen+1);
2555 spec->count = 1;
2556 spec->is_binary = -1;
2557 return spec;
2558 }
2559
2560 void free_filespec(struct diff_filespec *spec)
2561 {
2562 if (!--spec->count) {
2563 diff_free_filespec_data(spec);
2564 free(spec);
2565 }
2566 }
2567
2568 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2569 int sha1_valid, unsigned short mode)
2570 {
2571 if (mode) {
2572 spec->mode = canon_mode(mode);
2573 hashcpy(spec->sha1, sha1);
2574 spec->sha1_valid = sha1_valid;
2575 }
2576 }
2577
2578 /*
2579 * Given a name and sha1 pair, if the index tells us the file in
2580 * the work tree has that object contents, return true, so that
2581 * prepare_temp_file() does not have to inflate and extract.
2582 */
2583 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2584 {
2585 struct cache_entry *ce;
2586 struct stat st;
2587 int pos, len;
2588
2589 /*
2590 * We do not read the cache ourselves here, because the
2591 * benchmark with my previous version that always reads cache
2592 * shows that it makes things worse for diff-tree comparing
2593 * two linux-2.6 kernel trees in an already checked out work
2594 * tree. This is because most diff-tree comparisons deal with
2595 * only a small number of files, while reading the cache is
2596 * expensive for a large project, and its cost outweighs the
2597 * savings we get by not inflating the object to a temporary
2598 * file. Practically, this code only helps when we are used
2599 * by diff-cache --cached, which does read the cache before
2600 * calling us.
2601 */
2602 if (!active_cache)
2603 return 0;
2604
2605 /* We want to avoid the working directory if our caller
2606 * doesn't need the data in a normal file, this system
2607 * is rather slow with its stat/open/mmap/close syscalls,
2608 * and the object is contained in a pack file. The pack
2609 * is probably already open and will be faster to obtain
2610 * the data through than the working directory. Loose
2611 * objects however would tend to be slower as they need
2612 * to be individually opened and inflated.
2613 */
2614 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2615 return 0;
2616
2617 len = strlen(name);
2618 pos = cache_name_pos(name, len);
2619 if (pos < 0)
2620 return 0;
2621 ce = active_cache[pos];
2622
2623 /*
2624 * This is not the sha1 we are looking for, or
2625 * unreusable because it is not a regular file.
2626 */
2627 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2628 return 0;
2629
2630 /*
2631 * If ce is marked as "assume unchanged", there is no
2632 * guarantee that work tree matches what we are looking for.
2633 */
2634 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2635 return 0;
2636
2637 /*
2638 * If ce matches the file in the work tree, we can reuse it.
2639 */
2640 if (ce_uptodate(ce) ||
2641 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2642 return 1;
2643
2644 return 0;
2645 }
2646
2647 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2648 {
2649 int len;
2650 char *data = xmalloc(100), *dirty = "";
2651
2652 /* Are we looking at the work tree? */
2653 if (s->dirty_submodule)
2654 dirty = "-dirty";
2655
2656 len = snprintf(data, 100,
2657 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2658 s->data = data;
2659 s->size = len;
2660 s->should_free = 1;
2661 if (size_only) {
2662 s->data = NULL;
2663 free(data);
2664 }
2665 return 0;
2666 }
2667
2668 /*
2669 * While doing rename detection and pickaxe operation, we may need to
2670 * grab the data for the blob (or file) for our own in-core comparison.
2671 * diff_filespec has data and size fields for this purpose.
2672 */
2673 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2674 {
2675 int err = 0;
2676 if (!DIFF_FILE_VALID(s))
2677 die("internal error: asking to populate invalid file.");
2678 if (S_ISDIR(s->mode))
2679 return -1;
2680
2681 if (s->data)
2682 return 0;
2683
2684 if (size_only && 0 < s->size)
2685 return 0;
2686
2687 if (S_ISGITLINK(s->mode))
2688 return diff_populate_gitlink(s, size_only);
2689
2690 if (!s->sha1_valid ||
2691 reuse_worktree_file(s->path, s->sha1, 0)) {
2692 struct strbuf buf = STRBUF_INIT;
2693 struct stat st;
2694 int fd;
2695
2696 if (lstat(s->path, &st) < 0) {
2697 if (errno == ENOENT) {
2698 err_empty:
2699 err = -1;
2700 empty:
2701 s->data = (char *)"";
2702 s->size = 0;
2703 return err;
2704 }
2705 }
2706 s->size = xsize_t(st.st_size);
2707 if (!s->size)
2708 goto empty;
2709 if (S_ISLNK(st.st_mode)) {
2710 struct strbuf sb = STRBUF_INIT;
2711
2712 if (strbuf_readlink(&sb, s->path, s->size))
2713 goto err_empty;
2714 s->size = sb.len;
2715 s->data = strbuf_detach(&sb, NULL);
2716 s->should_free = 1;
2717 return 0;
2718 }
2719 if (size_only)
2720 return 0;
2721 fd = open(s->path, O_RDONLY);
2722 if (fd < 0)
2723 goto err_empty;
2724 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2725 close(fd);
2726 s->should_munmap = 1;
2727
2728 /*
2729 * Convert from working tree format to canonical git format
2730 */
2731 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2732 size_t size = 0;
2733 munmap(s->data, s->size);
2734 s->should_munmap = 0;
2735 s->data = strbuf_detach(&buf, &size);
2736 s->size = size;
2737 s->should_free = 1;
2738 }
2739 }
2740 else {
2741 enum object_type type;
2742 if (size_only) {
2743 type = sha1_object_info(s->sha1, &s->size);
2744 if (type < 0)
2745 die("unable to read %s", sha1_to_hex(s->sha1));
2746 } else {
2747 s->data = read_sha1_file(s->sha1, &type, &s->size);
2748 if (!s->data)
2749 die("unable to read %s", sha1_to_hex(s->sha1));
2750 s->should_free = 1;
2751 }
2752 }
2753 return 0;
2754 }
2755
2756 void diff_free_filespec_blob(struct diff_filespec *s)
2757 {
2758 if (s->should_free)
2759 free(s->data);
2760 else if (s->should_munmap)
2761 munmap(s->data, s->size);
2762
2763 if (s->should_free || s->should_munmap) {
2764 s->should_free = s->should_munmap = 0;
2765 s->data = NULL;
2766 }
2767 }
2768
2769 void diff_free_filespec_data(struct diff_filespec *s)
2770 {
2771 diff_free_filespec_blob(s);
2772 free(s->cnt_data);
2773 s->cnt_data = NULL;
2774 }
2775
2776 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2777 void *blob,
2778 unsigned long size,
2779 const unsigned char *sha1,
2780 int mode)
2781 {
2782 int fd;
2783 struct strbuf buf = STRBUF_INIT;
2784 struct strbuf template = STRBUF_INIT;
2785 char *path_dup = xstrdup(path);
2786 const char *base = basename(path_dup);
2787
2788 /* Generate "XXXXXX_basename.ext" */
2789 strbuf_addstr(&template, "XXXXXX_");
2790 strbuf_addstr(&template, base);
2791
2792 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2793 strlen(base) + 1);
2794 if (fd < 0)
2795 die_errno("unable to create temp-file");
2796 if (convert_to_working_tree(path,
2797 (const char *)blob, (size_t)size, &buf)) {
2798 blob = buf.buf;
2799 size = buf.len;
2800 }
2801 if (write_in_full(fd, blob, size) != size)
2802 die_errno("unable to write temp-file");
2803 close(fd);
2804 temp->name = temp->tmp_path;
2805 strcpy(temp->hex, sha1_to_hex(sha1));
2806 temp->hex[40] = 0;
2807 sprintf(temp->mode, "%06o", mode);
2808 strbuf_release(&buf);
2809 strbuf_release(&template);
2810 free(path_dup);
2811 }
2812
2813 static struct diff_tempfile *prepare_temp_file(const char *name,
2814 struct diff_filespec *one)
2815 {
2816 struct diff_tempfile *temp = claim_diff_tempfile();
2817
2818 if (!DIFF_FILE_VALID(one)) {
2819 not_a_valid_file:
2820 /* A '-' entry produces this for file-2, and
2821 * a '+' entry produces this for file-1.
2822 */
2823 temp->name = "/dev/null";
2824 strcpy(temp->hex, ".");
2825 strcpy(temp->mode, ".");
2826 return temp;
2827 }
2828
2829 if (!remove_tempfile_installed) {
2830 atexit(remove_tempfile);
2831 sigchain_push_common(remove_tempfile_on_signal);
2832 remove_tempfile_installed = 1;
2833 }
2834
2835 if (!one->sha1_valid ||
2836 reuse_worktree_file(name, one->sha1, 1)) {
2837 struct stat st;
2838 if (lstat(name, &st) < 0) {
2839 if (errno == ENOENT)
2840 goto not_a_valid_file;
2841 die_errno("stat(%s)", name);
2842 }
2843 if (S_ISLNK(st.st_mode)) {
2844 struct strbuf sb = STRBUF_INIT;
2845 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2846 die_errno("readlink(%s)", name);
2847 prep_temp_blob(name, temp, sb.buf, sb.len,
2848 (one->sha1_valid ?
2849 one->sha1 : null_sha1),
2850 (one->sha1_valid ?
2851 one->mode : S_IFLNK));
2852 strbuf_release(&sb);
2853 }
2854 else {
2855 /* we can borrow from the file in the work tree */
2856 temp->name = name;
2857 if (!one->sha1_valid)
2858 strcpy(temp->hex, sha1_to_hex(null_sha1));
2859 else
2860 strcpy(temp->hex, sha1_to_hex(one->sha1));
2861 /* Even though we may sometimes borrow the
2862 * contents from the work tree, we always want
2863 * one->mode. mode is trustworthy even when
2864 * !(one->sha1_valid), as long as
2865 * DIFF_FILE_VALID(one).
2866 */
2867 sprintf(temp->mode, "%06o", one->mode);
2868 }
2869 return temp;
2870 }
2871 else {
2872 if (diff_populate_filespec(one, 0))
2873 die("cannot read data blob for %s", one->path);
2874 prep_temp_blob(name, temp, one->data, one->size,
2875 one->sha1, one->mode);
2876 }
2877 return temp;
2878 }
2879
2880 /* An external diff command takes:
2881 *
2882 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2883 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2884 *
2885 */
2886 static void run_external_diff(const char *pgm,
2887 const char *name,
2888 const char *other,
2889 struct diff_filespec *one,
2890 struct diff_filespec *two,
2891 const char *xfrm_msg,
2892 int complete_rewrite)
2893 {
2894 const char *spawn_arg[10];
2895 int retval;
2896 const char **arg = &spawn_arg[0];
2897
2898 if (one && two) {
2899 struct diff_tempfile *temp_one, *temp_two;
2900 const char *othername = (other ? other : name);
2901 temp_one = prepare_temp_file(name, one);
2902 temp_two = prepare_temp_file(othername, two);
2903 *arg++ = pgm;
2904 *arg++ = name;
2905 *arg++ = temp_one->name;
2906 *arg++ = temp_one->hex;
2907 *arg++ = temp_one->mode;
2908 *arg++ = temp_two->name;
2909 *arg++ = temp_two->hex;
2910 *arg++ = temp_two->mode;
2911 if (other) {
2912 *arg++ = other;
2913 *arg++ = xfrm_msg;
2914 }
2915 } else {
2916 *arg++ = pgm;
2917 *arg++ = name;
2918 }
2919 *arg = NULL;
2920 fflush(NULL);
2921 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2922 remove_tempfile();
2923 if (retval) {
2924 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2925 exit(1);
2926 }
2927 }
2928
2929 static int similarity_index(struct diff_filepair *p)
2930 {
2931 return p->score * 100 / MAX_SCORE;
2932 }
2933
2934 static void fill_metainfo(struct strbuf *msg,
2935 const char *name,
2936 const char *other,
2937 struct diff_filespec *one,
2938 struct diff_filespec *two,
2939 struct diff_options *o,
2940 struct diff_filepair *p,
2941 int *must_show_header,
2942 int use_color)
2943 {
2944 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2945 const char *reset = diff_get_color(use_color, DIFF_RESET);
2946 const char *line_prefix = diff_line_prefix(o);
2947
2948 *must_show_header = 1;
2949 strbuf_init(msg, PATH_MAX * 2 + 300);
2950 switch (p->status) {
2951 case DIFF_STATUS_COPIED:
2952 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2953 line_prefix, set, similarity_index(p));
2954 strbuf_addf(msg, "%s\n%s%scopy from ",
2955 reset, line_prefix, set);
2956 quote_c_style(name, msg, NULL, 0);
2957 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2958 quote_c_style(other, msg, NULL, 0);
2959 strbuf_addf(msg, "%s\n", reset);
2960 break;
2961 case DIFF_STATUS_RENAMED:
2962 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2963 line_prefix, set, similarity_index(p));
2964 strbuf_addf(msg, "%s\n%s%srename from ",
2965 reset, line_prefix, set);
2966 quote_c_style(name, msg, NULL, 0);
2967 strbuf_addf(msg, "%s\n%s%srename to ",
2968 reset, line_prefix, set);
2969 quote_c_style(other, msg, NULL, 0);
2970 strbuf_addf(msg, "%s\n", reset);
2971 break;
2972 case DIFF_STATUS_MODIFIED:
2973 if (p->score) {
2974 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2975 line_prefix,
2976 set, similarity_index(p), reset);
2977 break;
2978 }
2979 /* fallthru */
2980 default:
2981 *must_show_header = 0;
2982 }
2983 if (one && two && hashcmp(one->sha1, two->sha1)) {
2984 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2985
2986 if (DIFF_OPT_TST(o, BINARY)) {
2987 mmfile_t mf;
2988 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2989 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2990 abbrev = 40;
2991 }
2992 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2993 find_unique_abbrev(one->sha1, abbrev));
2994 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2995 if (one->mode == two->mode)
2996 strbuf_addf(msg, " %06o", one->mode);
2997 strbuf_addf(msg, "%s\n", reset);
2998 }
2999 }
3000
3001 static void run_diff_cmd(const char *pgm,
3002 const char *name,
3003 const char *other,
3004 const char *attr_path,
3005 struct diff_filespec *one,
3006 struct diff_filespec *two,
3007 struct strbuf *msg,
3008 struct diff_options *o,
3009 struct diff_filepair *p)
3010 {
3011 const char *xfrm_msg = NULL;
3012 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3013 int must_show_header = 0;
3014
3015
3016 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3017 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3018 if (drv && drv->external)
3019 pgm = drv->external;
3020 }
3021
3022 if (msg) {
3023 /*
3024 * don't use colors when the header is intended for an
3025 * external diff driver
3026 */
3027 fill_metainfo(msg, name, other, one, two, o, p,
3028 &must_show_header,
3029 want_color(o->use_color) && !pgm);
3030 xfrm_msg = msg->len ? msg->buf : NULL;
3031 }
3032
3033 if (pgm) {
3034 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3035 complete_rewrite);
3036 return;
3037 }
3038 if (one && two)
3039 builtin_diff(name, other ? other : name,
3040 one, two, xfrm_msg, must_show_header,
3041 o, complete_rewrite);
3042 else
3043 fprintf(o->file, "* Unmerged path %s\n", name);
3044 }
3045
3046 static void diff_fill_sha1_info(struct diff_filespec *one)
3047 {
3048 if (DIFF_FILE_VALID(one)) {
3049 if (!one->sha1_valid) {
3050 struct stat st;
3051 if (one->is_stdin) {
3052 hashcpy(one->sha1, null_sha1);
3053 return;
3054 }
3055 if (lstat(one->path, &st) < 0)
3056 die_errno("stat '%s'", one->path);
3057 if (index_path(one->sha1, one->path, &st, 0))
3058 die("cannot hash %s", one->path);
3059 }
3060 }
3061 else
3062 hashclr(one->sha1);
3063 }
3064
3065 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3066 {
3067 /* Strip the prefix but do not molest /dev/null and absolute paths */
3068 if (*namep && **namep != '/') {
3069 *namep += prefix_length;
3070 if (**namep == '/')
3071 ++*namep;
3072 }
3073 if (*otherp && **otherp != '/') {
3074 *otherp += prefix_length;
3075 if (**otherp == '/')
3076 ++*otherp;
3077 }
3078 }
3079
3080 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3081 {
3082 const char *pgm = external_diff();
3083 struct strbuf msg;
3084 struct diff_filespec *one = p->one;
3085 struct diff_filespec *two = p->two;
3086 const char *name;
3087 const char *other;
3088 const char *attr_path;
3089
3090 name = p->one->path;
3091 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3092 attr_path = name;
3093 if (o->prefix_length)
3094 strip_prefix(o->prefix_length, &name, &other);
3095
3096 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3097 pgm = NULL;
3098
3099 if (DIFF_PAIR_UNMERGED(p)) {
3100 run_diff_cmd(pgm, name, NULL, attr_path,
3101 NULL, NULL, NULL, o, p);
3102 return;
3103 }
3104
3105 diff_fill_sha1_info(one);
3106 diff_fill_sha1_info(two);
3107
3108 if (!pgm &&
3109 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3110 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3111 /*
3112 * a filepair that changes between file and symlink
3113 * needs to be split into deletion and creation.
3114 */
3115 struct diff_filespec *null = alloc_filespec(two->path);
3116 run_diff_cmd(NULL, name, other, attr_path,
3117 one, null, &msg, o, p);
3118 free(null);
3119 strbuf_release(&msg);
3120
3121 null = alloc_filespec(one->path);
3122 run_diff_cmd(NULL, name, other, attr_path,
3123 null, two, &msg, o, p);
3124 free(null);
3125 }
3126 else
3127 run_diff_cmd(pgm, name, other, attr_path,
3128 one, two, &msg, o, p);
3129
3130 strbuf_release(&msg);
3131 }
3132
3133 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3134 struct diffstat_t *diffstat)
3135 {
3136 const char *name;
3137 const char *other;
3138
3139 if (DIFF_PAIR_UNMERGED(p)) {
3140 /* unmerged */
3141 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3142 return;
3143 }
3144
3145 name = p->one->path;
3146 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3147
3148 if (o->prefix_length)
3149 strip_prefix(o->prefix_length, &name, &other);
3150
3151 diff_fill_sha1_info(p->one);
3152 diff_fill_sha1_info(p->two);
3153
3154 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3155 }
3156
3157 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3158 {
3159 const char *name;
3160 const char *other;
3161 const char *attr_path;
3162
3163 if (DIFF_PAIR_UNMERGED(p)) {
3164 /* unmerged */
3165 return;
3166 }
3167
3168 name = p->one->path;
3169 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3170 attr_path = other ? other : name;
3171
3172 if (o->prefix_length)
3173 strip_prefix(o->prefix_length, &name, &other);
3174
3175 diff_fill_sha1_info(p->one);
3176 diff_fill_sha1_info(p->two);
3177
3178 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3179 }
3180
3181 void diff_setup(struct diff_options *options)
3182 {
3183 memcpy(options, &default_diff_options, sizeof(*options));
3184
3185 options->file = stdout;
3186
3187 options->line_termination = '\n';
3188 options->break_opt = -1;
3189 options->rename_limit = -1;
3190 options->dirstat_permille = diff_dirstat_permille_default;
3191 options->context = diff_context_default;
3192 DIFF_OPT_SET(options, RENAME_EMPTY);
3193
3194 options->change = diff_change;
3195 options->add_remove = diff_addremove;
3196 options->use_color = diff_use_color_default;
3197 options->detect_rename = diff_detect_rename_default;
3198 options->xdl_opts |= diff_algorithm;
3199
3200 if (diff_no_prefix) {
3201 options->a_prefix = options->b_prefix = "";
3202 } else if (!diff_mnemonic_prefix) {
3203 options->a_prefix = "a/";
3204 options->b_prefix = "b/";
3205 }
3206 }
3207
3208 void diff_setup_done(struct diff_options *options)
3209 {
3210 int count = 0;
3211
3212 if (options->output_format & DIFF_FORMAT_NAME)
3213 count++;
3214 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3215 count++;
3216 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3217 count++;
3218 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3219 count++;
3220 if (count > 1)
3221 die("--name-only, --name-status, --check and -s are mutually exclusive");
3222
3223 /*
3224 * Most of the time we can say "there are changes"
3225 * only by checking if there are changed paths, but
3226 * --ignore-whitespace* options force us to look
3227 * inside contents.
3228 */
3229
3230 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3231 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3232 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3233 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3234 else
3235 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3236
3237 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3238 options->detect_rename = DIFF_DETECT_COPY;
3239
3240 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3241 options->prefix = NULL;
3242 if (options->prefix)
3243 options->prefix_length = strlen(options->prefix);
3244 else
3245 options->prefix_length = 0;
3246
3247 if (options->output_format & (DIFF_FORMAT_NAME |
3248 DIFF_FORMAT_NAME_STATUS |
3249 DIFF_FORMAT_CHECKDIFF |
3250 DIFF_FORMAT_NO_OUTPUT))
3251 options->output_format &= ~(DIFF_FORMAT_RAW |
3252 DIFF_FORMAT_NUMSTAT |
3253 DIFF_FORMAT_DIFFSTAT |
3254 DIFF_FORMAT_SHORTSTAT |
3255 DIFF_FORMAT_DIRSTAT |
3256 DIFF_FORMAT_SUMMARY |
3257 DIFF_FORMAT_PATCH);
3258
3259 /*
3260 * These cases always need recursive; we do not drop caller-supplied
3261 * recursive bits for other formats here.
3262 */
3263 if (options->output_format & (DIFF_FORMAT_PATCH |
3264 DIFF_FORMAT_NUMSTAT |
3265 DIFF_FORMAT_DIFFSTAT |
3266 DIFF_FORMAT_SHORTSTAT |
3267 DIFF_FORMAT_DIRSTAT |
3268 DIFF_FORMAT_SUMMARY |
3269 DIFF_FORMAT_CHECKDIFF))
3270 DIFF_OPT_SET(options, RECURSIVE);
3271 /*
3272 * Also pickaxe would not work very well if you do not say recursive
3273 */
3274 if (options->pickaxe)
3275 DIFF_OPT_SET(options, RECURSIVE);
3276 /*
3277 * When patches are generated, submodules diffed against the work tree
3278 * must be checked for dirtiness too so it can be shown in the output
3279 */
3280 if (options->output_format & DIFF_FORMAT_PATCH)
3281 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3282
3283 if (options->detect_rename && options->rename_limit < 0)
3284 options->rename_limit = diff_rename_limit_default;
3285 if (options->setup & DIFF_SETUP_USE_CACHE) {
3286 if (!active_cache)
3287 /* read-cache does not die even when it fails
3288 * so it is safe for us to do this here. Also
3289 * it does not smudge active_cache or active_nr
3290 * when it fails, so we do not have to worry about
3291 * cleaning it up ourselves either.
3292 */
3293 read_cache();
3294 }
3295 if (options->abbrev <= 0 || 40 < options->abbrev)
3296 options->abbrev = 40; /* full */
3297
3298 /*
3299 * It does not make sense to show the first hit we happened
3300 * to have found. It does not make sense not to return with
3301 * exit code in such a case either.
3302 */
3303 if (DIFF_OPT_TST(options, QUICK)) {
3304 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3305 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3306 }
3307 }
3308
3309 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3310 {
3311 char c, *eq;
3312 int len;
3313
3314 if (*arg != '-')
3315 return 0;
3316 c = *++arg;
3317 if (!c)
3318 return 0;
3319 if (c == arg_short) {
3320 c = *++arg;
3321 if (!c)
3322 return 1;
3323 if (val && isdigit(c)) {
3324 char *end;
3325 int n = strtoul(arg, &end, 10);
3326 if (*end)
3327 return 0;
3328 *val = n;
3329 return 1;
3330 }
3331 return 0;
3332 }
3333 if (c != '-')
3334 return 0;
3335 arg++;
3336 eq = strchr(arg, '=');
3337 if (eq)
3338 len = eq - arg;
3339 else
3340 len = strlen(arg);
3341 if (!len || strncmp(arg, arg_long, len))
3342 return 0;
3343 if (eq) {
3344 int n;
3345 char *end;
3346 if (!isdigit(*++eq))
3347 return 0;
3348 n = strtoul(eq, &end, 10);
3349 if (*end)
3350 return 0;
3351 *val = n;
3352 }
3353 return 1;
3354 }
3355
3356 static int diff_scoreopt_parse(const char *opt);
3357
3358 static inline int short_opt(char opt, const char **argv,
3359 const char **optarg)
3360 {
3361 const char *arg = argv[0];
3362 if (arg[0] != '-' || arg[1] != opt)
3363 return 0;
3364 if (arg[2] != '\0') {
3365 *optarg = arg + 2;
3366 return 1;
3367 }
3368 if (!argv[1])
3369 die("Option '%c' requires a value", opt);
3370 *optarg = argv[1];
3371 return 2;
3372 }
3373
3374 int parse_long_opt(const char *opt, const char **argv,
3375 const char **optarg)
3376 {
3377 const char *arg = argv[0];
3378 if (arg[0] != '-' || arg[1] != '-')
3379 return 0;
3380 arg += strlen("--");
3381 if (prefixcmp(arg, opt))
3382 return 0;
3383 arg += strlen(opt);
3384 if (*arg == '=') { /* sticked form: --option=value */
3385 *optarg = arg + 1;
3386 return 1;
3387 }
3388 if (*arg != '\0')
3389 return 0;
3390 /* separate form: --option value */
3391 if (!argv[1])
3392 die("Option '--%s' requires a value", opt);
3393 *optarg = argv[1];
3394 return 2;
3395 }
3396
3397 static int stat_opt(struct diff_options *options, const char **av)
3398 {
3399 const char *arg = av[0];
3400 char *end;
3401 int width = options->stat_width;
3402 int name_width = options->stat_name_width;
3403 int graph_width = options->stat_graph_width;
3404 int count = options->stat_count;
3405 int argcount = 1;
3406
3407 arg += strlen("--stat");
3408 end = (char *)arg;
3409
3410 switch (*arg) {
3411 case '-':
3412 if (!prefixcmp(arg, "-width")) {
3413 arg += strlen("-width");
3414 if (*arg == '=')
3415 width = strtoul(arg + 1, &end, 10);
3416 else if (!*arg && !av[1])
3417 die("Option '--stat-width' requires a value");
3418 else if (!*arg) {
3419 width = strtoul(av[1], &end, 10);
3420 argcount = 2;
3421 }
3422 } else if (!prefixcmp(arg, "-name-width")) {
3423 arg += strlen("-name-width");
3424 if (*arg == '=')
3425 name_width = strtoul(arg + 1, &end, 10);
3426 else if (!*arg && !av[1])
3427 die("Option '--stat-name-width' requires a value");
3428 else if (!*arg) {
3429 name_width = strtoul(av[1], &end, 10);
3430 argcount = 2;
3431 }
3432 } else if (!prefixcmp(arg, "-graph-width")) {
3433 arg += strlen("-graph-width");
3434 if (*arg == '=')
3435 graph_width = strtoul(arg + 1, &end, 10);
3436 else if (!*arg && !av[1])
3437 die("Option '--stat-graph-width' requires a value");
3438 else if (!*arg) {
3439 graph_width = strtoul(av[1], &end, 10);
3440 argcount = 2;
3441 }
3442 } else if (!prefixcmp(arg, "-count")) {
3443 arg += strlen("-count");
3444 if (*arg == '=')
3445 count = strtoul(arg + 1, &end, 10);
3446 else if (!*arg && !av[1])
3447 die("Option '--stat-count' requires a value");
3448 else if (!*arg) {
3449 count = strtoul(av[1], &end, 10);
3450 argcount = 2;
3451 }
3452 }
3453 break;
3454 case '=':
3455 width = strtoul(arg+1, &end, 10);
3456 if (*end == ',')
3457 name_width = strtoul(end+1, &end, 10);
3458 if (*end == ',')
3459 count = strtoul(end+1, &end, 10);
3460 }
3461
3462 /* Important! This checks all the error cases! */
3463 if (*end)
3464 return 0;
3465 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3466 options->stat_name_width = name_width;
3467 options->stat_graph_width = graph_width;
3468 options->stat_width = width;
3469 options->stat_count = count;
3470 return argcount;
3471 }
3472
3473 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3474 {
3475 struct strbuf errmsg = STRBUF_INIT;
3476 if (parse_dirstat_params(options, params, &errmsg))
3477 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3478 errmsg.buf);
3479 strbuf_release(&errmsg);
3480 /*
3481 * The caller knows a dirstat-related option is given from the command
3482 * line; allow it to say "return this_function();"
3483 */
3484 options->output_format |= DIFF_FORMAT_DIRSTAT;
3485 return 1;
3486 }
3487
3488 static int parse_submodule_opt(struct diff_options *options, const char *value)
3489 {
3490 if (parse_submodule_params(options, value))
3491 die(_("Failed to parse --submodule option parameter: '%s'"),
3492 value);
3493 return 1;
3494 }
3495
3496 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3497 {
3498 const char *arg = av[0];
3499 const char *optarg;
3500 int argcount;
3501
3502 /* Output format options */
3503 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3504 options->output_format |= DIFF_FORMAT_PATCH;
3505 else if (opt_arg(arg, 'U', "unified", &options->context))
3506 options->output_format |= DIFF_FORMAT_PATCH;
3507 else if (!strcmp(arg, "--raw"))
3508 options->output_format |= DIFF_FORMAT_RAW;
3509 else if (!strcmp(arg, "--patch-with-raw"))
3510 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3511 else if (!strcmp(arg, "--numstat"))
3512 options->output_format |= DIFF_FORMAT_NUMSTAT;
3513 else if (!strcmp(arg, "--shortstat"))
3514 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3515 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3516 return parse_dirstat_opt(options, "");
3517 else if (!prefixcmp(arg, "-X"))
3518 return parse_dirstat_opt(options, arg + 2);
3519 else if (!prefixcmp(arg, "--dirstat="))
3520 return parse_dirstat_opt(options, arg + 10);
3521 else if (!strcmp(arg, "--cumulative"))
3522 return parse_dirstat_opt(options, "cumulative");
3523 else if (!strcmp(arg, "--dirstat-by-file"))
3524 return parse_dirstat_opt(options, "files");
3525 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3526 parse_dirstat_opt(options, "files");
3527 return parse_dirstat_opt(options, arg + 18);
3528 }
3529 else if (!strcmp(arg, "--check"))
3530 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3531 else if (!strcmp(arg, "--summary"))
3532 options->output_format |= DIFF_FORMAT_SUMMARY;
3533 else if (!strcmp(arg, "--patch-with-stat"))
3534 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3535 else if (!strcmp(arg, "--name-only"))
3536 options->output_format |= DIFF_FORMAT_NAME;
3537 else if (!strcmp(arg, "--name-status"))
3538 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3539 else if (!strcmp(arg, "-s"))
3540 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3541 else if (!prefixcmp(arg, "--stat"))
3542 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3543 return stat_opt(options, av);
3544
3545 /* renames options */
3546 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3547 !strcmp(arg, "--break-rewrites")) {
3548 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3549 return error("invalid argument to -B: %s", arg+2);
3550 }
3551 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3552 !strcmp(arg, "--find-renames")) {
3553 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3554 return error("invalid argument to -M: %s", arg+2);
3555 options->detect_rename = DIFF_DETECT_RENAME;
3556 }
3557 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3558 options->irreversible_delete = 1;
3559 }
3560 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3561 !strcmp(arg, "--find-copies")) {
3562 if (options->detect_rename == DIFF_DETECT_COPY)
3563 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3564 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3565 return error("invalid argument to -C: %s", arg+2);
3566 options->detect_rename = DIFF_DETECT_COPY;
3567 }
3568 else if (!strcmp(arg, "--no-renames"))
3569 options->detect_rename = 0;
3570 else if (!strcmp(arg, "--rename-empty"))
3571 DIFF_OPT_SET(options, RENAME_EMPTY);
3572 else if (!strcmp(arg, "--no-rename-empty"))
3573 DIFF_OPT_CLR(options, RENAME_EMPTY);
3574 else if (!strcmp(arg, "--relative"))
3575 DIFF_OPT_SET(options, RELATIVE_NAME);
3576 else if (!prefixcmp(arg, "--relative=")) {
3577 DIFF_OPT_SET(options, RELATIVE_NAME);
3578 options->prefix = arg + 11;
3579 }
3580
3581 /* xdiff options */
3582 else if (!strcmp(arg, "--minimal"))
3583 DIFF_XDL_SET(options, NEED_MINIMAL);
3584 else if (!strcmp(arg, "--no-minimal"))
3585 DIFF_XDL_CLR(options, NEED_MINIMAL);
3586 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3587 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3588 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3589 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3590 else if (!strcmp(arg, "--ignore-space-at-eol"))
3591 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3592 else if (!strcmp(arg, "--patience"))
3593 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3594 else if (!strcmp(arg, "--histogram"))
3595 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3596 else if (!prefixcmp(arg, "--diff-algorithm=")) {
3597 long value = parse_algorithm_value(arg+17);
3598 if (value < 0)
3599 return error("option diff-algorithm accepts \"myers\", "
3600 "\"minimal\", \"patience\" and \"histogram\"");
3601 /* clear out previous settings */
3602 DIFF_XDL_CLR(options, NEED_MINIMAL);
3603 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3604 options->xdl_opts |= value;
3605 }
3606
3607 /* flags options */
3608 else if (!strcmp(arg, "--binary")) {
3609 options->output_format |= DIFF_FORMAT_PATCH;
3610 DIFF_OPT_SET(options, BINARY);
3611 }
3612 else if (!strcmp(arg, "--full-index"))
3613 DIFF_OPT_SET(options, FULL_INDEX);
3614 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3615 DIFF_OPT_SET(options, TEXT);
3616 else if (!strcmp(arg, "-R"))
3617 DIFF_OPT_SET(options, REVERSE_DIFF);
3618 else if (!strcmp(arg, "--find-copies-harder"))
3619 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3620 else if (!strcmp(arg, "--follow"))
3621 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3622 else if (!strcmp(arg, "--no-follow"))
3623 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3624 else if (!strcmp(arg, "--color"))
3625 options->use_color = 1;
3626 else if (!prefixcmp(arg, "--color=")) {
3627 int value = git_config_colorbool(NULL, arg+8);
3628 if (value < 0)
3629 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3630 options->use_color = value;
3631 }
3632 else if (!strcmp(arg, "--no-color"))
3633 options->use_color = 0;
3634 else if (!strcmp(arg, "--color-words")) {
3635 options->use_color = 1;
3636 options->word_diff = DIFF_WORDS_COLOR;
3637 }
3638 else if (!prefixcmp(arg, "--color-words=")) {
3639 options->use_color = 1;
3640 options->word_diff = DIFF_WORDS_COLOR;
3641 options->word_regex = arg + 14;
3642 }
3643 else if (!strcmp(arg, "--word-diff")) {
3644 if (options->word_diff == DIFF_WORDS_NONE)
3645 options->word_diff = DIFF_WORDS_PLAIN;
3646 }
3647 else if (!prefixcmp(arg, "--word-diff=")) {
3648 const char *type = arg + 12;
3649 if (!strcmp(type, "plain"))
3650 options->word_diff = DIFF_WORDS_PLAIN;
3651 else if (!strcmp(type, "color")) {
3652 options->use_color = 1;
3653 options->word_diff = DIFF_WORDS_COLOR;
3654 }
3655 else if (!strcmp(type, "porcelain"))
3656 options->word_diff = DIFF_WORDS_PORCELAIN;
3657 else if (!strcmp(type, "none"))
3658 options->word_diff = DIFF_WORDS_NONE;
3659 else
3660 die("bad --word-diff argument: %s", type);
3661 }
3662 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3663 if (options->word_diff == DIFF_WORDS_NONE)
3664 options->word_diff = DIFF_WORDS_PLAIN;
3665 options->word_regex = optarg;
3666 return argcount;
3667 }
3668 else if (!strcmp(arg, "--exit-code"))
3669 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3670 else if (!strcmp(arg, "--quiet"))
3671 DIFF_OPT_SET(options, QUICK);
3672 else if (!strcmp(arg, "--ext-diff"))
3673 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3674 else if (!strcmp(arg, "--no-ext-diff"))
3675 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3676 else if (!strcmp(arg, "--textconv"))
3677 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3678 else if (!strcmp(arg, "--no-textconv"))
3679 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3680 else if (!strcmp(arg, "--ignore-submodules")) {
3681 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3682 handle_ignore_submodules_arg(options, "all");
3683 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3684 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3685 handle_ignore_submodules_arg(options, arg + 20);
3686 } else if (!strcmp(arg, "--submodule"))
3687 DIFF_OPT_SET(options, SUBMODULE_LOG);
3688 else if (!prefixcmp(arg, "--submodule="))
3689 return parse_submodule_opt(options, arg + 12);
3690
3691 /* misc options */
3692 else if (!strcmp(arg, "-z"))
3693 options->line_termination = 0;
3694 else if ((argcount = short_opt('l', av, &optarg))) {
3695 options->rename_limit = strtoul(optarg, NULL, 10);
3696 return argcount;
3697 }
3698 else if ((argcount = short_opt('S', av, &optarg))) {
3699 options->pickaxe = optarg;
3700 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3701 return argcount;
3702 } else if ((argcount = short_opt('G', av, &optarg))) {
3703 options->pickaxe = optarg;
3704 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3705 return argcount;
3706 }
3707 else if (!strcmp(arg, "--pickaxe-all"))
3708 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3709 else if (!strcmp(arg, "--pickaxe-regex"))
3710 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3711 else if ((argcount = short_opt('O', av, &optarg))) {
3712 options->orderfile = optarg;
3713 return argcount;
3714 }
3715 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3716 options->filter = optarg;
3717 return argcount;
3718 }
3719 else if (!strcmp(arg, "--abbrev"))
3720 options->abbrev = DEFAULT_ABBREV;
3721 else if (!prefixcmp(arg, "--abbrev=")) {
3722 options->abbrev = strtoul(arg + 9, NULL, 10);
3723 if (options->abbrev < MINIMUM_ABBREV)
3724 options->abbrev = MINIMUM_ABBREV;
3725 else if (40 < options->abbrev)
3726 options->abbrev = 40;
3727 }
3728 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3729 options->a_prefix = optarg;
3730 return argcount;
3731 }
3732 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3733 options->b_prefix = optarg;
3734 return argcount;
3735 }
3736 else if (!strcmp(arg, "--no-prefix"))
3737 options->a_prefix = options->b_prefix = "";
3738 else if (opt_arg(arg, '\0', "inter-hunk-context",
3739 &options->interhunkcontext))
3740 ;
3741 else if (!strcmp(arg, "-W"))
3742 DIFF_OPT_SET(options, FUNCCONTEXT);
3743 else if (!strcmp(arg, "--function-context"))
3744 DIFF_OPT_SET(options, FUNCCONTEXT);
3745 else if (!strcmp(arg, "--no-function-context"))
3746 DIFF_OPT_CLR(options, FUNCCONTEXT);
3747 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3748 options->file = fopen(optarg, "w");
3749 if (!options->file)
3750 die_errno("Could not open '%s'", optarg);
3751 options->close_file = 1;
3752 return argcount;
3753 } else
3754 return 0;
3755 return 1;
3756 }
3757
3758 int parse_rename_score(const char **cp_p)
3759 {
3760 unsigned long num, scale;
3761 int ch, dot;
3762 const char *cp = *cp_p;
3763
3764 num = 0;
3765 scale = 1;
3766 dot = 0;
3767 for (;;) {
3768 ch = *cp;
3769 if ( !dot && ch == '.' ) {
3770 scale = 1;
3771 dot = 1;
3772 } else if ( ch == '%' ) {
3773 scale = dot ? scale*100 : 100;
3774 cp++; /* % is always at the end */
3775 break;
3776 } else if ( ch >= '0' && ch <= '9' ) {
3777 if ( scale < 100000 ) {
3778 scale *= 10;
3779 num = (num*10) + (ch-'0');
3780 }
3781 } else {
3782 break;
3783 }
3784 cp++;
3785 }
3786 *cp_p = cp;
3787
3788 /* user says num divided by scale and we say internally that
3789 * is MAX_SCORE * num / scale.
3790 */
3791 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3792 }
3793
3794 static int diff_scoreopt_parse(const char *opt)
3795 {
3796 int opt1, opt2, cmd;
3797
3798 if (*opt++ != '-')
3799 return -1;
3800 cmd = *opt++;
3801 if (cmd == '-') {
3802 /* convert the long-form arguments into short-form versions */
3803 if (!prefixcmp(opt, "break-rewrites")) {
3804 opt += strlen("break-rewrites");
3805 if (*opt == 0 || *opt++ == '=')
3806 cmd = 'B';
3807 } else if (!prefixcmp(opt, "find-copies")) {
3808 opt += strlen("find-copies");
3809 if (*opt == 0 || *opt++ == '=')
3810 cmd = 'C';
3811 } else if (!prefixcmp(opt, "find-renames")) {
3812 opt += strlen("find-renames");
3813 if (*opt == 0 || *opt++ == '=')
3814 cmd = 'M';
3815 }
3816 }
3817 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3818 return -1; /* that is not a -M, -C nor -B option */
3819
3820 opt1 = parse_rename_score(&opt);
3821 if (cmd != 'B')
3822 opt2 = 0;
3823 else {
3824 if (*opt == 0)
3825 opt2 = 0;
3826 else if (*opt != '/')
3827 return -1; /* we expect -B80/99 or -B80 */
3828 else {
3829 opt++;
3830 opt2 = parse_rename_score(&opt);
3831 }
3832 }
3833 if (*opt != 0)
3834 return -1;
3835 return opt1 | (opt2 << 16);
3836 }
3837
3838 struct diff_queue_struct diff_queued_diff;
3839
3840 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3841 {
3842 if (queue->alloc <= queue->nr) {
3843 queue->alloc = alloc_nr(queue->alloc);
3844 queue->queue = xrealloc(queue->queue,
3845 sizeof(dp) * queue->alloc);
3846 }
3847 queue->queue[queue->nr++] = dp;
3848 }
3849
3850 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3851 struct diff_filespec *one,
3852 struct diff_filespec *two)
3853 {
3854 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3855 dp->one = one;
3856 dp->two = two;
3857 if (queue)
3858 diff_q(queue, dp);
3859 return dp;
3860 }
3861
3862 void diff_free_filepair(struct diff_filepair *p)
3863 {
3864 free_filespec(p->one);
3865 free_filespec(p->two);
3866 free(p);
3867 }
3868
3869 /* This is different from find_unique_abbrev() in that
3870 * it stuffs the result with dots for alignment.
3871 */
3872 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3873 {
3874 int abblen;
3875 const char *abbrev;
3876 if (len == 40)
3877 return sha1_to_hex(sha1);
3878
3879 abbrev = find_unique_abbrev(sha1, len);
3880 abblen = strlen(abbrev);
3881 if (abblen < 37) {
3882 static char hex[41];
3883 if (len < abblen && abblen <= len + 2)
3884 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3885 else
3886 sprintf(hex, "%s...", abbrev);
3887 return hex;
3888 }
3889 return sha1_to_hex(sha1);
3890 }
3891
3892 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3893 {
3894 int line_termination = opt->line_termination;
3895 int inter_name_termination = line_termination ? '\t' : '\0';
3896
3897 fprintf(opt->file, "%s", diff_line_prefix(opt));
3898 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3899 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3900 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3901 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3902 }
3903 if (p->score) {
3904 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3905 inter_name_termination);
3906 } else {
3907 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3908 }
3909
3910 if (p->status == DIFF_STATUS_COPIED ||
3911 p->status == DIFF_STATUS_RENAMED) {
3912 const char *name_a, *name_b;
3913 name_a = p->one->path;
3914 name_b = p->two->path;
3915 strip_prefix(opt->prefix_length, &name_a, &name_b);
3916 write_name_quoted(name_a, opt->file, inter_name_termination);
3917 write_name_quoted(name_b, opt->file, line_termination);
3918 } else {
3919 const char *name_a, *name_b;
3920 name_a = p->one->mode ? p->one->path : p->two->path;
3921 name_b = NULL;
3922 strip_prefix(opt->prefix_length, &name_a, &name_b);
3923 write_name_quoted(name_a, opt->file, line_termination);
3924 }
3925 }
3926
3927 int diff_unmodified_pair(struct diff_filepair *p)
3928 {
3929 /* This function is written stricter than necessary to support
3930 * the currently implemented transformers, but the idea is to
3931 * let transformers to produce diff_filepairs any way they want,
3932 * and filter and clean them up here before producing the output.
3933 */
3934 struct diff_filespec *one = p->one, *two = p->two;
3935
3936 if (DIFF_PAIR_UNMERGED(p))
3937 return 0; /* unmerged is interesting */
3938
3939 /* deletion, addition, mode or type change
3940 * and rename are all interesting.
3941 */
3942 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3943 DIFF_PAIR_MODE_CHANGED(p) ||
3944 strcmp(one->path, two->path))
3945 return 0;
3946
3947 /* both are valid and point at the same path. that is, we are
3948 * dealing with a change.
3949 */
3950 if (one->sha1_valid && two->sha1_valid &&
3951 !hashcmp(one->sha1, two->sha1) &&
3952 !one->dirty_submodule && !two->dirty_submodule)
3953 return 1; /* no change */
3954 if (!one->sha1_valid && !two->sha1_valid)
3955 return 1; /* both look at the same file on the filesystem. */
3956 return 0;
3957 }
3958
3959 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3960 {
3961 if (diff_unmodified_pair(p))
3962 return;
3963
3964 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3965 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3966 return; /* no tree diffs in patch format */
3967
3968 run_diff(p, o);
3969 }
3970
3971 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3972 struct diffstat_t *diffstat)
3973 {
3974 if (diff_unmodified_pair(p))
3975 return;
3976
3977 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3978 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3979 return; /* no useful stat for tree diffs */
3980
3981 run_diffstat(p, o, diffstat);
3982 }
3983
3984 static void diff_flush_checkdiff(struct diff_filepair *p,
3985 struct diff_options *o)
3986 {
3987 if (diff_unmodified_pair(p))
3988 return;
3989
3990 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3991 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3992 return; /* nothing to check in tree diffs */
3993
3994 run_checkdiff(p, o);
3995 }
3996
3997 int diff_queue_is_empty(void)
3998 {
3999 struct diff_queue_struct *q = &diff_queued_diff;
4000 int i;
4001 for (i = 0; i < q->nr; i++)
4002 if (!diff_unmodified_pair(q->queue[i]))
4003 return 0;
4004 return 1;
4005 }
4006
4007 #if DIFF_DEBUG
4008 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4009 {
4010 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4011 x, one ? one : "",
4012 s->path,
4013 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4014 s->mode,
4015 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
4016 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
4017 x, one ? one : "",
4018 s->size, s->xfrm_flags);
4019 }
4020
4021 void diff_debug_filepair(const struct diff_filepair *p, int i)
4022 {
4023 diff_debug_filespec(p->one, i, "one");
4024 diff_debug_filespec(p->two, i, "two");
4025 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4026 p->score, p->status ? p->status : '?',
4027 p->one->rename_used, p->broken_pair);
4028 }
4029
4030 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4031 {
4032 int i;
4033 if (msg)
4034 fprintf(stderr, "%s\n", msg);
4035 fprintf(stderr, "q->nr = %d\n", q->nr);
4036 for (i = 0; i < q->nr; i++) {
4037 struct diff_filepair *p = q->queue[i];
4038 diff_debug_filepair(p, i);
4039 }
4040 }
4041 #endif
4042
4043 static void diff_resolve_rename_copy(void)
4044 {
4045 int i;
4046 struct diff_filepair *p;
4047 struct diff_queue_struct *q = &diff_queued_diff;
4048
4049 diff_debug_queue("resolve-rename-copy", q);
4050
4051 for (i = 0; i < q->nr; i++) {
4052 p = q->queue[i];
4053 p->status = 0; /* undecided */
4054 if (DIFF_PAIR_UNMERGED(p))
4055 p->status = DIFF_STATUS_UNMERGED;
4056 else if (!DIFF_FILE_VALID(p->one))
4057 p->status = DIFF_STATUS_ADDED;
4058 else if (!DIFF_FILE_VALID(p->two))
4059 p->status = DIFF_STATUS_DELETED;
4060 else if (DIFF_PAIR_TYPE_CHANGED(p))
4061 p->status = DIFF_STATUS_TYPE_CHANGED;
4062
4063 /* from this point on, we are dealing with a pair
4064 * whose both sides are valid and of the same type, i.e.
4065 * either in-place edit or rename/copy edit.
4066 */
4067 else if (DIFF_PAIR_RENAME(p)) {
4068 /*
4069 * A rename might have re-connected a broken
4070 * pair up, causing the pathnames to be the
4071 * same again. If so, that's not a rename at
4072 * all, just a modification..
4073 *
4074 * Otherwise, see if this source was used for
4075 * multiple renames, in which case we decrement
4076 * the count, and call it a copy.
4077 */
4078 if (!strcmp(p->one->path, p->two->path))
4079 p->status = DIFF_STATUS_MODIFIED;
4080 else if (--p->one->rename_used > 0)
4081 p->status = DIFF_STATUS_COPIED;
4082 else
4083 p->status = DIFF_STATUS_RENAMED;
4084 }
4085 else if (hashcmp(p->one->sha1, p->two->sha1) ||
4086 p->one->mode != p->two->mode ||
4087 p->one->dirty_submodule ||
4088 p->two->dirty_submodule ||
4089 is_null_sha1(p->one->sha1))
4090 p->status = DIFF_STATUS_MODIFIED;
4091 else {
4092 /* This is a "no-change" entry and should not
4093 * happen anymore, but prepare for broken callers.
4094 */
4095 error("feeding unmodified %s to diffcore",
4096 p->one->path);
4097 p->status = DIFF_STATUS_UNKNOWN;
4098 }
4099 }
4100 diff_debug_queue("resolve-rename-copy done", q);
4101 }
4102
4103 static int check_pair_status(struct diff_filepair *p)
4104 {
4105 switch (p->status) {
4106 case DIFF_STATUS_UNKNOWN:
4107 return 0;
4108 case 0:
4109 die("internal error in diff-resolve-rename-copy");
4110 default:
4111 return 1;
4112 }
4113 }
4114
4115 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4116 {
4117 int fmt = opt->output_format;
4118
4119 if (fmt & DIFF_FORMAT_CHECKDIFF)
4120 diff_flush_checkdiff(p, opt);
4121 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4122 diff_flush_raw(p, opt);
4123 else if (fmt & DIFF_FORMAT_NAME) {
4124 const char *name_a, *name_b;
4125 name_a = p->two->path;
4126 name_b = NULL;
4127 strip_prefix(opt->prefix_length, &name_a, &name_b);
4128 write_name_quoted(name_a, opt->file, opt->line_termination);
4129 }
4130 }
4131
4132 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4133 {
4134 if (fs->mode)
4135 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4136 else
4137 fprintf(file, " %s ", newdelete);
4138 write_name_quoted(fs->path, file, '\n');
4139 }
4140
4141
4142 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4143 const char *line_prefix)
4144 {
4145 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4146 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4147 p->two->mode, show_name ? ' ' : '\n');
4148 if (show_name) {
4149 write_name_quoted(p->two->path, file, '\n');
4150 }
4151 }
4152 }
4153
4154 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4155 const char *line_prefix)
4156 {
4157 char *names = pprint_rename(p->one->path, p->two->path);
4158
4159 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4160 free(names);
4161 show_mode_change(file, p, 0, line_prefix);
4162 }
4163
4164 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4165 {
4166 FILE *file = opt->file;
4167 const char *line_prefix = diff_line_prefix(opt);
4168
4169 switch(p->status) {
4170 case DIFF_STATUS_DELETED:
4171 fputs(line_prefix, file);
4172 show_file_mode_name(file, "delete", p->one);
4173 break;
4174 case DIFF_STATUS_ADDED:
4175 fputs(line_prefix, file);
4176 show_file_mode_name(file, "create", p->two);
4177 break;
4178 case DIFF_STATUS_COPIED:
4179 fputs(line_prefix, file);
4180 show_rename_copy(file, "copy", p, line_prefix);
4181 break;
4182 case DIFF_STATUS_RENAMED:
4183 fputs(line_prefix, file);
4184 show_rename_copy(file, "rename", p, line_prefix);
4185 break;
4186 default:
4187 if (p->score) {
4188 fprintf(file, "%s rewrite ", line_prefix);
4189 write_name_quoted(p->two->path, file, ' ');
4190 fprintf(file, "(%d%%)\n", similarity_index(p));
4191 }
4192 show_mode_change(file, p, !p->score, line_prefix);
4193 break;
4194 }
4195 }
4196
4197 struct patch_id_t {
4198 git_SHA_CTX *ctx;
4199 int patchlen;
4200 };
4201
4202 static int remove_space(char *line, int len)
4203 {
4204 int i;
4205 char *dst = line;
4206 unsigned char c;
4207
4208 for (i = 0; i < len; i++)
4209 if (!isspace((c = line[i])))
4210 *dst++ = c;
4211
4212 return dst - line;
4213 }
4214
4215 static void patch_id_consume(void *priv, char *line, unsigned long len)
4216 {
4217 struct patch_id_t *data = priv;
4218 int new_len;
4219
4220 /* Ignore line numbers when computing the SHA1 of the patch */
4221 if (!prefixcmp(line, "@@ -"))
4222 return;
4223
4224 new_len = remove_space(line, len);
4225
4226 git_SHA1_Update(data->ctx, line, new_len);
4227 data->patchlen += new_len;
4228 }
4229
4230 /* returns 0 upon success, and writes result into sha1 */
4231 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4232 {
4233 struct diff_queue_struct *q = &diff_queued_diff;
4234 int i;
4235 git_SHA_CTX ctx;
4236 struct patch_id_t data;
4237 char buffer[PATH_MAX * 4 + 20];
4238
4239 git_SHA1_Init(&ctx);
4240 memset(&data, 0, sizeof(struct patch_id_t));
4241 data.ctx = &ctx;
4242
4243 for (i = 0; i < q->nr; i++) {
4244 xpparam_t xpp;
4245 xdemitconf_t xecfg;
4246 mmfile_t mf1, mf2;
4247 struct diff_filepair *p = q->queue[i];
4248 int len1, len2;
4249
4250 memset(&xpp, 0, sizeof(xpp));
4251 memset(&xecfg, 0, sizeof(xecfg));
4252 if (p->status == 0)
4253 return error("internal diff status error");
4254 if (p->status == DIFF_STATUS_UNKNOWN)
4255 continue;
4256 if (diff_unmodified_pair(p))
4257 continue;
4258 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4259 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4260 continue;
4261 if (DIFF_PAIR_UNMERGED(p))
4262 continue;
4263
4264 diff_fill_sha1_info(p->one);
4265 diff_fill_sha1_info(p->two);
4266 if (fill_mmfile(&mf1, p->one) < 0 ||
4267 fill_mmfile(&mf2, p->two) < 0)
4268 return error("unable to read files to diff");
4269
4270 len1 = remove_space(p->one->path, strlen(p->one->path));
4271 len2 = remove_space(p->two->path, strlen(p->two->path));
4272 if (p->one->mode == 0)
4273 len1 = snprintf(buffer, sizeof(buffer),
4274 "diff--gita/%.*sb/%.*s"
4275 "newfilemode%06o"
4276 "---/dev/null"
4277 "+++b/%.*s",
4278 len1, p->one->path,
4279 len2, p->two->path,
4280 p->two->mode,
4281 len2, p->two->path);
4282 else if (p->two->mode == 0)
4283 len1 = snprintf(buffer, sizeof(buffer),
4284 "diff--gita/%.*sb/%.*s"
4285 "deletedfilemode%06o"
4286 "---a/%.*s"
4287 "+++/dev/null",
4288 len1, p->one->path,
4289 len2, p->two->path,
4290 p->one->mode,
4291 len1, p->one->path);
4292 else
4293 len1 = snprintf(buffer, sizeof(buffer),
4294 "diff--gita/%.*sb/%.*s"
4295 "---a/%.*s"
4296 "+++b/%.*s",
4297 len1, p->one->path,
4298 len2, p->two->path,
4299 len1, p->one->path,
4300 len2, p->two->path);
4301 git_SHA1_Update(&ctx, buffer, len1);
4302
4303 if (diff_filespec_is_binary(p->one) ||
4304 diff_filespec_is_binary(p->two)) {
4305 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4306 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4307 continue;
4308 }
4309
4310 xpp.flags = 0;
4311 xecfg.ctxlen = 3;
4312 xecfg.flags = 0;
4313 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4314 &xpp, &xecfg);
4315 }
4316
4317 git_SHA1_Final(sha1, &ctx);
4318 return 0;
4319 }
4320
4321 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4322 {
4323 struct diff_queue_struct *q = &diff_queued_diff;
4324 int i;
4325 int result = diff_get_patch_id(options, sha1);
4326
4327 for (i = 0; i < q->nr; i++)
4328 diff_free_filepair(q->queue[i]);
4329
4330 free(q->queue);
4331 DIFF_QUEUE_CLEAR(q);
4332
4333 return result;
4334 }
4335
4336 static int is_summary_empty(const struct diff_queue_struct *q)
4337 {
4338 int i;
4339
4340 for (i = 0; i < q->nr; i++) {
4341 const struct diff_filepair *p = q->queue[i];
4342
4343 switch (p->status) {
4344 case DIFF_STATUS_DELETED:
4345 case DIFF_STATUS_ADDED:
4346 case DIFF_STATUS_COPIED:
4347 case DIFF_STATUS_RENAMED:
4348 return 0;
4349 default:
4350 if (p->score)
4351 return 0;
4352 if (p->one->mode && p->two->mode &&
4353 p->one->mode != p->two->mode)
4354 return 0;
4355 break;
4356 }
4357 }
4358 return 1;
4359 }
4360
4361 static const char rename_limit_warning[] =
4362 "inexact rename detection was skipped due to too many files.";
4363
4364 static const char degrade_cc_to_c_warning[] =
4365 "only found copies from modified paths due to too many files.";
4366
4367 static const char rename_limit_advice[] =
4368 "you may want to set your %s variable to at least "
4369 "%d and retry the command.";
4370
4371 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4372 {
4373 if (degraded_cc)
4374 warning(degrade_cc_to_c_warning);
4375 else if (needed)
4376 warning(rename_limit_warning);
4377 else
4378 return;
4379 if (0 < needed && needed < 32767)
4380 warning(rename_limit_advice, varname, needed);
4381 }
4382
4383 void diff_flush(struct diff_options *options)
4384 {
4385 struct diff_queue_struct *q = &diff_queued_diff;
4386 int i, output_format = options->output_format;
4387 int separator = 0;
4388 int dirstat_by_line = 0;
4389
4390 /*
4391 * Order: raw, stat, summary, patch
4392 * or: name/name-status/checkdiff (other bits clear)
4393 */
4394 if (!q->nr)
4395 goto free_queue;
4396
4397 if (output_format & (DIFF_FORMAT_RAW |
4398 DIFF_FORMAT_NAME |
4399 DIFF_FORMAT_NAME_STATUS |
4400 DIFF_FORMAT_CHECKDIFF)) {
4401 for (i = 0; i < q->nr; i++) {
4402 struct diff_filepair *p = q->queue[i];
4403 if (check_pair_status(p))
4404 flush_one_pair(p, options);
4405 }
4406 separator++;
4407 }
4408
4409 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4410 dirstat_by_line = 1;
4411
4412 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4413 dirstat_by_line) {
4414 struct diffstat_t diffstat;
4415
4416 memset(&diffstat, 0, sizeof(struct diffstat_t));
4417 for (i = 0; i < q->nr; i++) {
4418 struct diff_filepair *p = q->queue[i];
4419 if (check_pair_status(p))
4420 diff_flush_stat(p, options, &diffstat);
4421 }
4422 if (output_format & DIFF_FORMAT_NUMSTAT)
4423 show_numstat(&diffstat, options);
4424 if (output_format & DIFF_FORMAT_DIFFSTAT)
4425 show_stats(&diffstat, options);
4426 if (output_format & DIFF_FORMAT_SHORTSTAT)
4427 show_shortstats(&diffstat, options);
4428 if (output_format & DIFF_FORMAT_DIRSTAT)
4429 show_dirstat_by_line(&diffstat, options);
4430 free_diffstat_info(&diffstat);
4431 separator++;
4432 }
4433 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4434 show_dirstat(options);
4435
4436 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4437 for (i = 0; i < q->nr; i++) {
4438 diff_summary(options, q->queue[i]);
4439 }
4440 separator++;
4441 }
4442
4443 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4444 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4445 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4446 /*
4447 * run diff_flush_patch for the exit status. setting
4448 * options->file to /dev/null should be safe, becaue we
4449 * aren't supposed to produce any output anyway.
4450 */
4451 if (options->close_file)
4452 fclose(options->file);
4453 options->file = fopen("/dev/null", "w");
4454 if (!options->file)
4455 die_errno("Could not open /dev/null");
4456 options->close_file = 1;
4457 for (i = 0; i < q->nr; i++) {
4458 struct diff_filepair *p = q->queue[i];
4459 if (check_pair_status(p))
4460 diff_flush_patch(p, options);
4461 if (options->found_changes)
4462 break;
4463 }
4464 }
4465
4466 if (output_format & DIFF_FORMAT_PATCH) {
4467 if (separator) {
4468 fprintf(options->file, "%s%c",
4469 diff_line_prefix(options),
4470 options->line_termination);
4471 if (options->stat_sep) {
4472 /* attach patch instead of inline */
4473 fputs(options->stat_sep, options->file);
4474 }
4475 }
4476
4477 for (i = 0; i < q->nr; i++) {
4478 struct diff_filepair *p = q->queue[i];
4479 if (check_pair_status(p))
4480 diff_flush_patch(p, options);
4481 }
4482 }
4483
4484 if (output_format & DIFF_FORMAT_CALLBACK)
4485 options->format_callback(q, options, options->format_callback_data);
4486
4487 for (i = 0; i < q->nr; i++)
4488 diff_free_filepair(q->queue[i]);
4489 free_queue:
4490 free(q->queue);
4491 DIFF_QUEUE_CLEAR(q);
4492 if (options->close_file)
4493 fclose(options->file);
4494
4495 /*
4496 * Report the content-level differences with HAS_CHANGES;
4497 * diff_addremove/diff_change does not set the bit when
4498 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4499 */
4500 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4501 if (options->found_changes)
4502 DIFF_OPT_SET(options, HAS_CHANGES);
4503 else
4504 DIFF_OPT_CLR(options, HAS_CHANGES);
4505 }
4506 }
4507
4508 static void diffcore_apply_filter(const char *filter)
4509 {
4510 int i;
4511 struct diff_queue_struct *q = &diff_queued_diff;
4512 struct diff_queue_struct outq;
4513 DIFF_QUEUE_CLEAR(&outq);
4514
4515 if (!filter)
4516 return;
4517
4518 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4519 int found;
4520 for (i = found = 0; !found && i < q->nr; i++) {
4521 struct diff_filepair *p = q->queue[i];
4522 if (((p->status == DIFF_STATUS_MODIFIED) &&
4523 ((p->score &&
4524 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4525 (!p->score &&
4526 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4527 ((p->status != DIFF_STATUS_MODIFIED) &&
4528 strchr(filter, p->status)))
4529 found++;
4530 }
4531 if (found)
4532 return;
4533
4534 /* otherwise we will clear the whole queue
4535 * by copying the empty outq at the end of this
4536 * function, but first clear the current entries
4537 * in the queue.
4538 */
4539 for (i = 0; i < q->nr; i++)
4540 diff_free_filepair(q->queue[i]);
4541 }
4542 else {
4543 /* Only the matching ones */
4544 for (i = 0; i < q->nr; i++) {
4545 struct diff_filepair *p = q->queue[i];
4546
4547 if (((p->status == DIFF_STATUS_MODIFIED) &&
4548 ((p->score &&
4549 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4550 (!p->score &&
4551 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4552 ((p->status != DIFF_STATUS_MODIFIED) &&
4553 strchr(filter, p->status)))
4554 diff_q(&outq, p);
4555 else
4556 diff_free_filepair(p);
4557 }
4558 }
4559 free(q->queue);
4560 *q = outq;
4561 }
4562
4563 /* Check whether two filespecs with the same mode and size are identical */
4564 static int diff_filespec_is_identical(struct diff_filespec *one,
4565 struct diff_filespec *two)
4566 {
4567 if (S_ISGITLINK(one->mode))
4568 return 0;
4569 if (diff_populate_filespec(one, 0))
4570 return 0;
4571 if (diff_populate_filespec(two, 0))
4572 return 0;
4573 return !memcmp(one->data, two->data, one->size);
4574 }
4575
4576 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4577 {
4578 int i;
4579 struct diff_queue_struct *q = &diff_queued_diff;
4580 struct diff_queue_struct outq;
4581 DIFF_QUEUE_CLEAR(&outq);
4582
4583 for (i = 0; i < q->nr; i++) {
4584 struct diff_filepair *p = q->queue[i];
4585
4586 /*
4587 * 1. Entries that come from stat info dirtiness
4588 * always have both sides (iow, not create/delete),
4589 * one side of the object name is unknown, with
4590 * the same mode and size. Keep the ones that
4591 * do not match these criteria. They have real
4592 * differences.
4593 *
4594 * 2. At this point, the file is known to be modified,
4595 * with the same mode and size, and the object
4596 * name of one side is unknown. Need to inspect
4597 * the identical contents.
4598 */
4599 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4600 !DIFF_FILE_VALID(p->two) ||
4601 (p->one->sha1_valid && p->two->sha1_valid) ||
4602 (p->one->mode != p->two->mode) ||
4603 diff_populate_filespec(p->one, 1) ||
4604 diff_populate_filespec(p->two, 1) ||
4605 (p->one->size != p->two->size) ||
4606 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4607 diff_q(&outq, p);
4608 else {
4609 /*
4610 * The caller can subtract 1 from skip_stat_unmatch
4611 * to determine how many paths were dirty only
4612 * due to stat info mismatch.
4613 */
4614 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4615 diffopt->skip_stat_unmatch++;
4616 diff_free_filepair(p);
4617 }
4618 }
4619 free(q->queue);
4620 *q = outq;
4621 }
4622
4623 static int diffnamecmp(const void *a_, const void *b_)
4624 {
4625 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4626 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4627 const char *name_a, *name_b;
4628
4629 name_a = a->one ? a->one->path : a->two->path;
4630 name_b = b->one ? b->one->path : b->two->path;
4631 return strcmp(name_a, name_b);
4632 }
4633
4634 void diffcore_fix_diff_index(struct diff_options *options)
4635 {
4636 struct diff_queue_struct *q = &diff_queued_diff;
4637 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4638 }
4639
4640 void diffcore_std(struct diff_options *options)
4641 {
4642 if (options->skip_stat_unmatch)
4643 diffcore_skip_stat_unmatch(options);
4644 if (!options->found_follow) {
4645 /* See try_to_follow_renames() in tree-diff.c */
4646 if (options->break_opt != -1)
4647 diffcore_break(options->break_opt);
4648 if (options->detect_rename)
4649 diffcore_rename(options);
4650 if (options->break_opt != -1)
4651 diffcore_merge_broken();
4652 }
4653 if (options->pickaxe)
4654 diffcore_pickaxe(options);
4655 if (options->orderfile)
4656 diffcore_order(options->orderfile);
4657 if (!options->found_follow)
4658 /* See try_to_follow_renames() in tree-diff.c */
4659 diff_resolve_rename_copy();
4660 diffcore_apply_filter(options->filter);
4661
4662 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4663 DIFF_OPT_SET(options, HAS_CHANGES);
4664 else
4665 DIFF_OPT_CLR(options, HAS_CHANGES);
4666
4667 options->found_follow = 0;
4668 }
4669
4670 int diff_result_code(struct diff_options *opt, int status)
4671 {
4672 int result = 0;
4673
4674 diff_warn_rename_limit("diff.renamelimit",
4675 opt->needed_rename_limit,
4676 opt->degraded_cc_to_c);
4677 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4678 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4679 return status;
4680 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4681 DIFF_OPT_TST(opt, HAS_CHANGES))
4682 result |= 01;
4683 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4684 DIFF_OPT_TST(opt, CHECK_FAILED))
4685 result |= 02;
4686 return result;
4687 }
4688
4689 int diff_can_quit_early(struct diff_options *opt)
4690 {
4691 return (DIFF_OPT_TST(opt, QUICK) &&
4692 !opt->filter &&
4693 DIFF_OPT_TST(opt, HAS_CHANGES));
4694 }
4695
4696 /*
4697 * Shall changes to this submodule be ignored?
4698 *
4699 * Submodule changes can be configured to be ignored separately for each path,
4700 * but that configuration can be overridden from the command line.
4701 */
4702 static int is_submodule_ignored(const char *path, struct diff_options *options)
4703 {
4704 int ignored = 0;
4705 unsigned orig_flags = options->flags;
4706 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4707 set_diffopt_flags_from_submodule_config(options, path);
4708 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4709 ignored = 1;
4710 options->flags = orig_flags;
4711 return ignored;
4712 }
4713
4714 void diff_addremove(struct diff_options *options,
4715 int addremove, unsigned mode,
4716 const unsigned char *sha1,
4717 int sha1_valid,
4718 const char *concatpath, unsigned dirty_submodule)
4719 {
4720 struct diff_filespec *one, *two;
4721
4722 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4723 return;
4724
4725 /* This may look odd, but it is a preparation for
4726 * feeding "there are unchanged files which should
4727 * not produce diffs, but when you are doing copy
4728 * detection you would need them, so here they are"
4729 * entries to the diff-core. They will be prefixed
4730 * with something like '=' or '*' (I haven't decided
4731 * which but should not make any difference).
4732 * Feeding the same new and old to diff_change()
4733 * also has the same effect.
4734 * Before the final output happens, they are pruned after
4735 * merged into rename/copy pairs as appropriate.
4736 */
4737 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4738 addremove = (addremove == '+' ? '-' :
4739 addremove == '-' ? '+' : addremove);
4740
4741 if (options->prefix &&
4742 strncmp(concatpath, options->prefix, options->prefix_length))
4743 return;
4744
4745 one = alloc_filespec(concatpath);
4746 two = alloc_filespec(concatpath);
4747
4748 if (addremove != '+')
4749 fill_filespec(one, sha1, sha1_valid, mode);
4750 if (addremove != '-') {
4751 fill_filespec(two, sha1, sha1_valid, mode);
4752 two->dirty_submodule = dirty_submodule;
4753 }
4754
4755 diff_queue(&diff_queued_diff, one, two);
4756 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4757 DIFF_OPT_SET(options, HAS_CHANGES);
4758 }
4759
4760 void diff_change(struct diff_options *options,
4761 unsigned old_mode, unsigned new_mode,
4762 const unsigned char *old_sha1,
4763 const unsigned char *new_sha1,
4764 int old_sha1_valid, int new_sha1_valid,
4765 const char *concatpath,
4766 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4767 {
4768 struct diff_filespec *one, *two;
4769
4770 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4771 is_submodule_ignored(concatpath, options))
4772 return;
4773
4774 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4775 unsigned tmp;
4776 const unsigned char *tmp_c;
4777 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4778 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4779 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
4780 new_sha1_valid = tmp;
4781 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4782 new_dirty_submodule = tmp;
4783 }
4784
4785 if (options->prefix &&
4786 strncmp(concatpath, options->prefix, options->prefix_length))
4787 return;
4788
4789 one = alloc_filespec(concatpath);
4790 two = alloc_filespec(concatpath);
4791 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
4792 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
4793 one->dirty_submodule = old_dirty_submodule;
4794 two->dirty_submodule = new_dirty_submodule;
4795
4796 diff_queue(&diff_queued_diff, one, two);
4797 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4798 DIFF_OPT_SET(options, HAS_CHANGES);
4799 }
4800
4801 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4802 {
4803 struct diff_filepair *pair;
4804 struct diff_filespec *one, *two;
4805
4806 if (options->prefix &&
4807 strncmp(path, options->prefix, options->prefix_length))
4808 return NULL;
4809
4810 one = alloc_filespec(path);
4811 two = alloc_filespec(path);
4812 pair = diff_queue(&diff_queued_diff, one, two);
4813 pair->is_unmerged = 1;
4814 return pair;
4815 }
4816
4817 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4818 size_t *outsize)
4819 {
4820 struct diff_tempfile *temp;
4821 const char *argv[3];
4822 const char **arg = argv;
4823 struct child_process child;
4824 struct strbuf buf = STRBUF_INIT;
4825 int err = 0;
4826
4827 temp = prepare_temp_file(spec->path, spec);
4828 *arg++ = pgm;
4829 *arg++ = temp->name;
4830 *arg = NULL;
4831
4832 memset(&child, 0, sizeof(child));
4833 child.use_shell = 1;
4834 child.argv = argv;
4835 child.out = -1;
4836 if (start_command(&child)) {
4837 remove_tempfile();
4838 return NULL;
4839 }
4840
4841 if (strbuf_read(&buf, child.out, 0) < 0)
4842 err = error("error reading from textconv command '%s'", pgm);
4843 close(child.out);
4844
4845 if (finish_command(&child) || err) {
4846 strbuf_release(&buf);
4847 remove_tempfile();
4848 return NULL;
4849 }
4850 remove_tempfile();
4851
4852 return strbuf_detach(&buf, outsize);
4853 }
4854
4855 size_t fill_textconv(struct userdiff_driver *driver,
4856 struct diff_filespec *df,
4857 char **outbuf)
4858 {
4859 size_t size;
4860
4861 if (!driver || !driver->textconv) {
4862 if (!DIFF_FILE_VALID(df)) {
4863 *outbuf = "";
4864 return 0;
4865 }
4866 if (diff_populate_filespec(df, 0))
4867 die("unable to read files to diff");
4868 *outbuf = df->data;
4869 return df->size;
4870 }
4871
4872 if (driver->textconv_cache && df->sha1_valid) {
4873 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4874 &size);
4875 if (*outbuf)
4876 return size;
4877 }
4878
4879 *outbuf = run_textconv(driver->textconv, df, &size);
4880 if (!*outbuf)
4881 die("unable to read files to diff");
4882
4883 if (driver->textconv_cache && df->sha1_valid) {
4884 /* ignore errors, as we might be in a readonly repository */
4885 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4886 size);
4887 /*
4888 * we could save up changes and flush them all at the end,
4889 * but we would need an extra call after all diffing is done.
4890 * Since generating a cache entry is the slow path anyway,
4891 * this extra overhead probably isn't a big deal.
4892 */
4893 notes_cache_write(driver->textconv_cache);
4894 }
4895
4896 return size;
4897 }
4898
4899 void setup_diff_pager(struct diff_options *opt)
4900 {
4901 /*
4902 * If the user asked for our exit code, then either they want --quiet
4903 * or --exit-code. We should definitely not bother with a pager in the
4904 * former case, as we will generate no output. Since we still properly
4905 * report our exit code even when a pager is run, we _could_ run a
4906 * pager with --exit-code. But since we have not done so historically,
4907 * and because it is easy to find people oneline advising "git diff
4908 * --exit-code" in hooks and other scripts, we do not do so.
4909 */
4910 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4911 check_pager_config("diff") != 0)
4912 setup_pager();
4913 }
Michael J Gruber's git tree