?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
add xsnprintf helper function
[git/git-svn.git] / diff.c
blob08508f6a2017eb35a350268ea78a44cfc1d867e4
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "tempfile.h"
6 #include "quote.h"
7 #include "diff.h"
8 #include "diffcore.h"
9 #include "delta.h"
10 #include "xdiff-interface.h"
11 #include "color.h"
12 #include "attr.h"
13 #include "run-command.h"
14 #include "utf8.h"
15 #include "userdiff.h"
16 #include "sigchain.h"
17 #include "submodule-config.h"
18 #include "submodule.h"
19 #include "ll-merge.h"
20 #include "string-list.h"
21 #include "argv-array.h"
22
23 #ifdef NO_FAST_WORKING_DIRECTORY
24 #define FAST_WORKING_DIRECTORY 0
25 #else
26 #define FAST_WORKING_DIRECTORY 1
27 #endif
28
29 static int diff_detect_rename_default;
30 static int diff_rename_limit_default = 400;
31 static int diff_suppress_blank_empty;
32 static int diff_use_color_default = -1;
33 static int diff_context_default = 3;
34 static const char *diff_word_regex_cfg;
35 static const char *external_diff_cmd_cfg;
36 static const char *diff_order_file_cfg;
37 int diff_auto_refresh_index = 1;
38 static int diff_mnemonic_prefix;
39 static int diff_no_prefix;
40 static int diff_stat_graph_width;
41 static int diff_dirstat_permille_default = 30;
42 static struct diff_options default_diff_options;
43 static long diff_algorithm;
44
45 static char diff_colors[][COLOR_MAXLEN] = {
46 GIT_COLOR_RESET,
47 GIT_COLOR_NORMAL, /* CONTEXT */
48 GIT_COLOR_BOLD, /* METAINFO */
49 GIT_COLOR_CYAN, /* FRAGINFO */
50 GIT_COLOR_RED, /* OLD */
51 GIT_COLOR_GREEN, /* NEW */
52 GIT_COLOR_YELLOW, /* COMMIT */
53 GIT_COLOR_BG_RED, /* WHITESPACE */
54 GIT_COLOR_NORMAL, /* FUNCINFO */
55 };
56
57 static int parse_diff_color_slot(const char *var)
58 {
59 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
60 return DIFF_CONTEXT;
61 if (!strcasecmp(var, "meta"))
62 return DIFF_METAINFO;
63 if (!strcasecmp(var, "frag"))
64 return DIFF_FRAGINFO;
65 if (!strcasecmp(var, "old"))
66 return DIFF_FILE_OLD;
67 if (!strcasecmp(var, "new"))
68 return DIFF_FILE_NEW;
69 if (!strcasecmp(var, "commit"))
70 return DIFF_COMMIT;
71 if (!strcasecmp(var, "whitespace"))
72 return DIFF_WHITESPACE;
73 if (!strcasecmp(var, "func"))
74 return DIFF_FUNCINFO;
75 return -1;
76 }
77
78 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
79 struct strbuf *errmsg)
80 {
81 char *params_copy = xstrdup(params_string);
82 struct string_list params = STRING_LIST_INIT_NODUP;
83 int ret = 0;
84 int i;
85
86 if (*params_copy)
87 string_list_split_in_place(&params, params_copy, ',', -1);
88 for (i = 0; i < params.nr; i++) {
89 const char *p = params.items[i].string;
90 if (!strcmp(p, "changes")) {
91 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
92 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
93 } else if (!strcmp(p, "lines")) {
94 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
95 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
96 } else if (!strcmp(p, "files")) {
97 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
98 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
99 } else if (!strcmp(p, "noncumulative")) {
100 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
101 } else if (!strcmp(p, "cumulative")) {
102 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
103 } else if (isdigit(*p)) {
104 char *end;
105 int permille = strtoul(p, &end, 10) * 10;
106 if (*end == '.' && isdigit(*++end)) {
107 /* only use first digit */
108 permille += *end - '0';
109 /* .. and ignore any further digits */
110 while (isdigit(*++end))
111 ; /* nothing */
112 }
113 if (!*end)
114 options->dirstat_permille = permille;
115 else {
116 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
117 p);
118 ret++;
119 }
120 } else {
121 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
122 ret++;
123 }
124
125 }
126 string_list_clear(&params, 0);
127 free(params_copy);
128 return ret;
129 }
130
131 static int parse_submodule_params(struct diff_options *options, const char *value)
132 {
133 if (!strcmp(value, "log"))
134 DIFF_OPT_SET(options, SUBMODULE_LOG);
135 else if (!strcmp(value, "short"))
136 DIFF_OPT_CLR(options, SUBMODULE_LOG);
137 else
138 return -1;
139 return 0;
140 }
141
142 static int git_config_rename(const char *var, const char *value)
143 {
144 if (!value)
145 return DIFF_DETECT_RENAME;
146 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
147 return DIFF_DETECT_COPY;
148 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
149 }
150
151 long parse_algorithm_value(const char *value)
152 {
153 if (!value)
154 return -1;
155 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
156 return 0;
157 else if (!strcasecmp(value, "minimal"))
158 return XDF_NEED_MINIMAL;
159 else if (!strcasecmp(value, "patience"))
160 return XDF_PATIENCE_DIFF;
161 else if (!strcasecmp(value, "histogram"))
162 return XDF_HISTOGRAM_DIFF;
163 return -1;
164 }
165
166 /*
167 * These are to give UI layer defaults.
168 * The core-level commands such as git-diff-files should
169 * never be affected by the setting of diff.renames
170 * the user happens to have in the configuration file.
171 */
172 int git_diff_ui_config(const char *var, const char *value, void *cb)
173 {
174 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
175 diff_use_color_default = git_config_colorbool(var, value);
176 return 0;
177 }
178 if (!strcmp(var, "diff.context")) {
179 diff_context_default = git_config_int(var, value);
180 if (diff_context_default < 0)
181 return -1;
182 return 0;
183 }
184 if (!strcmp(var, "diff.renames")) {
185 diff_detect_rename_default = git_config_rename(var, value);
186 return 0;
187 }
188 if (!strcmp(var, "diff.autorefreshindex")) {
189 diff_auto_refresh_index = git_config_bool(var, value);
190 return 0;
191 }
192 if (!strcmp(var, "diff.mnemonicprefix")) {
193 diff_mnemonic_prefix = git_config_bool(var, value);
194 return 0;
195 }
196 if (!strcmp(var, "diff.noprefix")) {
197 diff_no_prefix = git_config_bool(var, value);
198 return 0;
199 }
200 if (!strcmp(var, "diff.statgraphwidth")) {
201 diff_stat_graph_width = git_config_int(var, value);
202 return 0;
203 }
204 if (!strcmp(var, "diff.external"))
205 return git_config_string(&external_diff_cmd_cfg, var, value);
206 if (!strcmp(var, "diff.wordregex"))
207 return git_config_string(&diff_word_regex_cfg, var, value);
208 if (!strcmp(var, "diff.orderfile"))
209 return git_config_pathname(&diff_order_file_cfg, var, value);
210
211 if (!strcmp(var, "diff.ignoresubmodules"))
212 handle_ignore_submodules_arg(&default_diff_options, value);
213
214 if (!strcmp(var, "diff.submodule")) {
215 if (parse_submodule_params(&default_diff_options, value))
216 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
217 value);
218 return 0;
219 }
220
221 if (!strcmp(var, "diff.algorithm")) {
222 diff_algorithm = parse_algorithm_value(value);
223 if (diff_algorithm < 0)
224 return -1;
225 return 0;
226 }
227
228 if (git_color_config(var, value, cb) < 0)
229 return -1;
230
231 return git_diff_basic_config(var, value, cb);
232 }
233
234 int git_diff_basic_config(const char *var, const char *value, void *cb)
235 {
236 const char *name;
237
238 if (!strcmp(var, "diff.renamelimit")) {
239 diff_rename_limit_default = git_config_int(var, value);
240 return 0;
241 }
242
243 if (userdiff_config(var, value) < 0)
244 return -1;
245
246 if (skip_prefix(var, "diff.color.", &name) ||
247 skip_prefix(var, "color.diff.", &name)) {
248 int slot = parse_diff_color_slot(name);
249 if (slot < 0)
250 return 0;
251 if (!value)
252 return config_error_nonbool(var);
253 return color_parse(value, diff_colors[slot]);
254 }
255
256 /* like GNU diff's --suppress-blank-empty option */
257 if (!strcmp(var, "diff.suppressblankempty") ||
258 /* for backwards compatibility */
259 !strcmp(var, "diff.suppress-blank-empty")) {
260 diff_suppress_blank_empty = git_config_bool(var, value);
261 return 0;
262 }
263
264 if (!strcmp(var, "diff.dirstat")) {
265 struct strbuf errmsg = STRBUF_INIT;
266 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
267 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
268 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
269 errmsg.buf);
270 strbuf_release(&errmsg);
271 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
272 return 0;
273 }
274
275 if (starts_with(var, "submodule."))
276 return parse_submodule_config_option(var, value);
277
278 return git_default_config(var, value, cb);
279 }
280
281 static char *quote_two(const char *one, const char *two)
282 {
283 int need_one = quote_c_style(one, NULL, NULL, 1);
284 int need_two = quote_c_style(two, NULL, NULL, 1);
285 struct strbuf res = STRBUF_INIT;
286
287 if (need_one + need_two) {
288 strbuf_addch(&res, '"');
289 quote_c_style(one, &res, NULL, 1);
290 quote_c_style(two, &res, NULL, 1);
291 strbuf_addch(&res, '"');
292 } else {
293 strbuf_addstr(&res, one);
294 strbuf_addstr(&res, two);
295 }
296 return strbuf_detach(&res, NULL);
297 }
298
299 static const char *external_diff(void)
300 {
301 static const char *external_diff_cmd = NULL;
302 static int done_preparing = 0;
303
304 if (done_preparing)
305 return external_diff_cmd;
306 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
307 if (!external_diff_cmd)
308 external_diff_cmd = external_diff_cmd_cfg;
309 done_preparing = 1;
310 return external_diff_cmd;
311 }
312
313 /*
314 * Keep track of files used for diffing. Sometimes such an entry
315 * refers to a temporary file, sometimes to an existing file, and
316 * sometimes to "/dev/null".
317 */
318 static struct diff_tempfile {
319 /*
320 * filename external diff should read from, or NULL if this
321 * entry is currently not in use:
322 */
323 const char *name;
324
325 char hex[41];
326 char mode[10];
327
328 /*
329 * If this diff_tempfile instance refers to a temporary file,
330 * this tempfile object is used to manage its lifetime.
331 */
332 struct tempfile tempfile;
333 } diff_temp[2];
334
335 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
336
337 struct emit_callback {
338 int color_diff;
339 unsigned ws_rule;
340 int blank_at_eof_in_preimage;
341 int blank_at_eof_in_postimage;
342 int lno_in_preimage;
343 int lno_in_postimage;
344 sane_truncate_fn truncate;
345 const char **label_path;
346 struct diff_words_data *diff_words;
347 struct diff_options *opt;
348 int *found_changesp;
349 struct strbuf *header;
350 };
351
352 static int count_lines(const char *data, int size)
353 {
354 int count, ch, completely_empty = 1, nl_just_seen = 0;
355 count = 0;
356 while (0 < size--) {
357 ch = *data++;
358 if (ch == '\n') {
359 count++;
360 nl_just_seen = 1;
361 completely_empty = 0;
362 }
363 else {
364 nl_just_seen = 0;
365 completely_empty = 0;
366 }
367 }
368 if (completely_empty)
369 return 0;
370 if (!nl_just_seen)
371 count++; /* no trailing newline */
372 return count;
373 }
374
375 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
376 {
377 if (!DIFF_FILE_VALID(one)) {
378 mf->ptr = (char *)""; /* does not matter */
379 mf->size = 0;
380 return 0;
381 }
382 else if (diff_populate_filespec(one, 0))
383 return -1;
384
385 mf->ptr = one->data;
386 mf->size = one->size;
387 return 0;
388 }
389
390 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
391 static unsigned long diff_filespec_size(struct diff_filespec *one)
392 {
393 if (!DIFF_FILE_VALID(one))
394 return 0;
395 diff_populate_filespec(one, CHECK_SIZE_ONLY);
396 return one->size;
397 }
398
399 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
400 {
401 char *ptr = mf->ptr;
402 long size = mf->size;
403 int cnt = 0;
404
405 if (!size)
406 return cnt;
407 ptr += size - 1; /* pointing at the very end */
408 if (*ptr != '\n')
409 ; /* incomplete line */
410 else
411 ptr--; /* skip the last LF */
412 while (mf->ptr < ptr) {
413 char *prev_eol;
414 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
415 if (*prev_eol == '\n')
416 break;
417 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
418 break;
419 cnt++;
420 ptr = prev_eol - 1;
421 }
422 return cnt;
423 }
424
425 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
426 struct emit_callback *ecbdata)
427 {
428 int l1, l2, at;
429 unsigned ws_rule = ecbdata->ws_rule;
430 l1 = count_trailing_blank(mf1, ws_rule);
431 l2 = count_trailing_blank(mf2, ws_rule);
432 if (l2 <= l1) {
433 ecbdata->blank_at_eof_in_preimage = 0;
434 ecbdata->blank_at_eof_in_postimage = 0;
435 return;
436 }
437 at = count_lines(mf1->ptr, mf1->size);
438 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
439
440 at = count_lines(mf2->ptr, mf2->size);
441 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
442 }
443
444 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
445 int first, const char *line, int len)
446 {
447 int has_trailing_newline, has_trailing_carriage_return;
448 int nofirst;
449 FILE *file = o->file;
450
451 fputs(diff_line_prefix(o), file);
452
453 if (len == 0) {
454 has_trailing_newline = (first == '\n');
455 has_trailing_carriage_return = (!has_trailing_newline &&
456 (first == '\r'));
457 nofirst = has_trailing_newline || has_trailing_carriage_return;
458 } else {
459 has_trailing_newline = (len > 0 && line[len-1] == '\n');
460 if (has_trailing_newline)
461 len--;
462 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
463 if (has_trailing_carriage_return)
464 len--;
465 nofirst = 0;
466 }
467
468 if (len || !nofirst) {
469 fputs(set, file);
470 if (!nofirst)
471 fputc(first, file);
472 fwrite(line, len, 1, file);
473 fputs(reset, file);
474 }
475 if (has_trailing_carriage_return)
476 fputc('\r', file);
477 if (has_trailing_newline)
478 fputc('\n', file);
479 }
480
481 static void emit_line(struct diff_options *o, const char *set, const char *reset,
482 const char *line, int len)
483 {
484 emit_line_0(o, set, reset, line[0], line+1, len-1);
485 }
486
487 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
488 {
489 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
490 ecbdata->blank_at_eof_in_preimage &&
491 ecbdata->blank_at_eof_in_postimage &&
492 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
493 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
494 return 0;
495 return ws_blank_line(line, len, ecbdata->ws_rule);
496 }
497
498 static void emit_line_checked(const char *reset,
499 struct emit_callback *ecbdata,
500 const char *line, int len,
501 enum color_diff color,
502 unsigned ws_error_highlight,
503 char sign)
504 {
505 const char *set = diff_get_color(ecbdata->color_diff, color);
506 const char *ws = NULL;
507
508 if (ecbdata->opt->ws_error_highlight & ws_error_highlight) {
509 ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
510 if (!*ws)
511 ws = NULL;
512 }
513
514 if (!ws)
515 emit_line_0(ecbdata->opt, set, reset, sign, line, len);
516 else if (sign == '+' && new_blank_line_at_eof(ecbdata, line, len))
517 /* Blank line at EOF - paint '+' as well */
518 emit_line_0(ecbdata->opt, ws, reset, sign, line, len);
519 else {
520 /* Emit just the prefix, then the rest. */
521 emit_line_0(ecbdata->opt, set, reset, sign, "", 0);
522 ws_check_emit(line, len, ecbdata->ws_rule,
523 ecbdata->opt->file, set, reset, ws);
524 }
525 }
526
527 static void emit_add_line(const char *reset,
528 struct emit_callback *ecbdata,
529 const char *line, int len)
530 {
531 emit_line_checked(reset, ecbdata, line, len,
532 DIFF_FILE_NEW, WSEH_NEW, '+');
533 }
534
535 static void emit_del_line(const char *reset,
536 struct emit_callback *ecbdata,
537 const char *line, int len)
538 {
539 emit_line_checked(reset, ecbdata, line, len,
540 DIFF_FILE_OLD, WSEH_OLD, '-');
541 }
542
543 static void emit_context_line(const char *reset,
544 struct emit_callback *ecbdata,
545 const char *line, int len)
546 {
547 emit_line_checked(reset, ecbdata, line, len,
548 DIFF_CONTEXT, WSEH_CONTEXT, ' ');
549 }
550
551 static void emit_hunk_header(struct emit_callback *ecbdata,
552 const char *line, int len)
553 {
554 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
555 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
556 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
557 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
558 static const char atat[2] = { '@', '@' };
559 const char *cp, *ep;
560 struct strbuf msgbuf = STRBUF_INIT;
561 int org_len = len;
562 int i = 1;
563
564 /*
565 * As a hunk header must begin with "@@ -<old>, +<new> @@",
566 * it always is at least 10 bytes long.
567 */
568 if (len < 10 ||
569 memcmp(line, atat, 2) ||
570 !(ep = memmem(line + 2, len - 2, atat, 2))) {
571 emit_line(ecbdata->opt, context, reset, line, len);
572 return;
573 }
574 ep += 2; /* skip over @@ */
575
576 /* The hunk header in fraginfo color */
577 strbuf_addstr(&msgbuf, frag);
578 strbuf_add(&msgbuf, line, ep - line);
579 strbuf_addstr(&msgbuf, reset);
580
581 /*
582 * trailing "\r\n"
583 */
584 for ( ; i < 3; i++)
585 if (line[len - i] == '\r' || line[len - i] == '\n')
586 len--;
587
588 /* blank before the func header */
589 for (cp = ep; ep - line < len; ep++)
590 if (*ep != ' ' && *ep != '\t')
591 break;
592 if (ep != cp) {
593 strbuf_addstr(&msgbuf, context);
594 strbuf_add(&msgbuf, cp, ep - cp);
595 strbuf_addstr(&msgbuf, reset);
596 }
597
598 if (ep < line + len) {
599 strbuf_addstr(&msgbuf, func);
600 strbuf_add(&msgbuf, ep, line + len - ep);
601 strbuf_addstr(&msgbuf, reset);
602 }
603
604 strbuf_add(&msgbuf, line + len, org_len - len);
605 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
606 strbuf_release(&msgbuf);
607 }
608
609 static struct diff_tempfile *claim_diff_tempfile(void) {
610 int i;
611 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
612 if (!diff_temp[i].name)
613 return diff_temp + i;
614 die("BUG: diff is failing to clean up its tempfiles");
615 }
616
617 static void remove_tempfile(void)
618 {
619 int i;
620 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
621 if (is_tempfile_active(&diff_temp[i].tempfile))
622 delete_tempfile(&diff_temp[i].tempfile);
623 diff_temp[i].name = NULL;
624 }
625 }
626
627 static void print_line_count(FILE *file, int count)
628 {
629 switch (count) {
630 case 0:
631 fprintf(file, "0,0");
632 break;
633 case 1:
634 fprintf(file, "1");
635 break;
636 default:
637 fprintf(file, "1,%d", count);
638 break;
639 }
640 }
641
642 static void emit_rewrite_lines(struct emit_callback *ecb,
643 int prefix, const char *data, int size)
644 {
645 const char *endp = NULL;
646 static const char *nneof = " No newline at end of file\n";
647 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
648
649 while (0 < size) {
650 int len;
651
652 endp = memchr(data, '\n', size);
653 len = endp ? (endp - data + 1) : size;
654 if (prefix != '+') {
655 ecb->lno_in_preimage++;
656 emit_del_line(reset, ecb, data, len);
657 } else {
658 ecb->lno_in_postimage++;
659 emit_add_line(reset, ecb, data, len);
660 }
661 size -= len;
662 data += len;
663 }
664 if (!endp) {
665 const char *context = diff_get_color(ecb->color_diff,
666 DIFF_CONTEXT);
667 putc('\n', ecb->opt->file);
668 emit_line_0(ecb->opt, context, reset, '\\',
669 nneof, strlen(nneof));
670 }
671 }
672
673 static void emit_rewrite_diff(const char *name_a,
674 const char *name_b,
675 struct diff_filespec *one,
676 struct diff_filespec *two,
677 struct userdiff_driver *textconv_one,
678 struct userdiff_driver *textconv_two,
679 struct diff_options *o)
680 {
681 int lc_a, lc_b;
682 const char *name_a_tab, *name_b_tab;
683 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
684 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
685 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
686 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
687 const char *a_prefix, *b_prefix;
688 char *data_one, *data_two;
689 size_t size_one, size_two;
690 struct emit_callback ecbdata;
691 const char *line_prefix = diff_line_prefix(o);
692
693 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
694 a_prefix = o->b_prefix;
695 b_prefix = o->a_prefix;
696 } else {
697 a_prefix = o->a_prefix;
698 b_prefix = o->b_prefix;
699 }
700
701 name_a += (*name_a == '/');
702 name_b += (*name_b == '/');
703 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
704 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
705
706 strbuf_reset(&a_name);
707 strbuf_reset(&b_name);
708 quote_two_c_style(&a_name, a_prefix, name_a, 0);
709 quote_two_c_style(&b_name, b_prefix, name_b, 0);
710
711 size_one = fill_textconv(textconv_one, one, &data_one);
712 size_two = fill_textconv(textconv_two, two, &data_two);
713
714 memset(&ecbdata, 0, sizeof(ecbdata));
715 ecbdata.color_diff = want_color(o->use_color);
716 ecbdata.found_changesp = &o->found_changes;
717 ecbdata.ws_rule = whitespace_rule(name_b);
718 ecbdata.opt = o;
719 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
720 mmfile_t mf1, mf2;
721 mf1.ptr = (char *)data_one;
722 mf2.ptr = (char *)data_two;
723 mf1.size = size_one;
724 mf2.size = size_two;
725 check_blank_at_eof(&mf1, &mf2, &ecbdata);
726 }
727 ecbdata.lno_in_preimage = 1;
728 ecbdata.lno_in_postimage = 1;
729
730 lc_a = count_lines(data_one, size_one);
731 lc_b = count_lines(data_two, size_two);
732 fprintf(o->file,
733 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
734 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
735 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
736 line_prefix, fraginfo);
737 if (!o->irreversible_delete)
738 print_line_count(o->file, lc_a);
739 else
740 fprintf(o->file, "?,?");
741 fprintf(o->file, " +");
742 print_line_count(o->file, lc_b);
743 fprintf(o->file, " @@%s\n", reset);
744 if (lc_a && !o->irreversible_delete)
745 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
746 if (lc_b)
747 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
748 if (textconv_one)
749 free((char *)data_one);
750 if (textconv_two)
751 free((char *)data_two);
752 }
753
754 struct diff_words_buffer {
755 mmfile_t text;
756 long alloc;
757 struct diff_words_orig {
758 const char *begin, *end;
759 } *orig;
760 int orig_nr, orig_alloc;
761 };
762
763 static void diff_words_append(char *line, unsigned long len,
764 struct diff_words_buffer *buffer)
765 {
766 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
767 line++;
768 len--;
769 memcpy(buffer->text.ptr + buffer->text.size, line, len);
770 buffer->text.size += len;
771 buffer->text.ptr[buffer->text.size] = '\0';
772 }
773
774 struct diff_words_style_elem {
775 const char *prefix;
776 const char *suffix;
777 const char *color; /* NULL; filled in by the setup code if
778 * color is enabled */
779 };
780
781 struct diff_words_style {
782 enum diff_words_type type;
783 struct diff_words_style_elem new, old, ctx;
784 const char *newline;
785 };
786
787 static struct diff_words_style diff_words_styles[] = {
788 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
789 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
790 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
791 };
792
793 struct diff_words_data {
794 struct diff_words_buffer minus, plus;
795 const char *current_plus;
796 int last_minus;
797 struct diff_options *opt;
798 regex_t *word_regex;
799 enum diff_words_type type;
800 struct diff_words_style *style;
801 };
802
803 static int fn_out_diff_words_write_helper(FILE *fp,
804 struct diff_words_style_elem *st_el,
805 const char *newline,
806 size_t count, const char *buf,
807 const char *line_prefix)
808 {
809 int print = 0;
810
811 while (count) {
812 char *p = memchr(buf, '\n', count);
813 if (print)
814 fputs(line_prefix, fp);
815 if (p != buf) {
816 if (st_el->color && fputs(st_el->color, fp) < 0)
817 return -1;
818 if (fputs(st_el->prefix, fp) < 0 ||
819 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
820 fputs(st_el->suffix, fp) < 0)
821 return -1;
822 if (st_el->color && *st_el->color
823 && fputs(GIT_COLOR_RESET, fp) < 0)
824 return -1;
825 }
826 if (!p)
827 return 0;
828 if (fputs(newline, fp) < 0)
829 return -1;
830 count -= p + 1 - buf;
831 buf = p + 1;
832 print = 1;
833 }
834 return 0;
835 }
836
837 /*
838 * '--color-words' algorithm can be described as:
839 *
840 * 1. collect a the minus/plus lines of a diff hunk, divided into
841 * minus-lines and plus-lines;
842 *
843 * 2. break both minus-lines and plus-lines into words and
844 * place them into two mmfile_t with one word for each line;
845 *
846 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
847 *
848 * And for the common parts of the both file, we output the plus side text.
849 * diff_words->current_plus is used to trace the current position of the plus file
850 * which printed. diff_words->last_minus is used to trace the last minus word
851 * printed.
852 *
853 * For '--graph' to work with '--color-words', we need to output the graph prefix
854 * on each line of color words output. Generally, there are two conditions on
855 * which we should output the prefix.
856 *
857 * 1. diff_words->last_minus == 0 &&
858 * diff_words->current_plus == diff_words->plus.text.ptr
859 *
860 * that is: the plus text must start as a new line, and if there is no minus
861 * word printed, a graph prefix must be printed.
862 *
863 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
864 * *(diff_words->current_plus - 1) == '\n'
865 *
866 * that is: a graph prefix must be printed following a '\n'
867 */
868 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
869 {
870 if ((diff_words->last_minus == 0 &&
871 diff_words->current_plus == diff_words->plus.text.ptr) ||
872 (diff_words->current_plus > diff_words->plus.text.ptr &&
873 *(diff_words->current_plus - 1) == '\n')) {
874 return 1;
875 } else {
876 return 0;
877 }
878 }
879
880 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
881 {
882 struct diff_words_data *diff_words = priv;
883 struct diff_words_style *style = diff_words->style;
884 int minus_first, minus_len, plus_first, plus_len;
885 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
886 struct diff_options *opt = diff_words->opt;
887 const char *line_prefix;
888
889 if (line[0] != '@' || parse_hunk_header(line, len,
890 &minus_first, &minus_len, &plus_first, &plus_len))
891 return;
892
893 assert(opt);
894 line_prefix = diff_line_prefix(opt);
895
896 /* POSIX requires that first be decremented by one if len == 0... */
897 if (minus_len) {
898 minus_begin = diff_words->minus.orig[minus_first].begin;
899 minus_end =
900 diff_words->minus.orig[minus_first + minus_len - 1].end;
901 } else
902 minus_begin = minus_end =
903 diff_words->minus.orig[minus_first].end;
904
905 if (plus_len) {
906 plus_begin = diff_words->plus.orig[plus_first].begin;
907 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
908 } else
909 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
910
911 if (color_words_output_graph_prefix(diff_words)) {
912 fputs(line_prefix, diff_words->opt->file);
913 }
914 if (diff_words->current_plus != plus_begin) {
915 fn_out_diff_words_write_helper(diff_words->opt->file,
916 &style->ctx, style->newline,
917 plus_begin - diff_words->current_plus,
918 diff_words->current_plus, line_prefix);
919 if (*(plus_begin - 1) == '\n')
920 fputs(line_prefix, diff_words->opt->file);
921 }
922 if (minus_begin != minus_end) {
923 fn_out_diff_words_write_helper(diff_words->opt->file,
924 &style->old, style->newline,
925 minus_end - minus_begin, minus_begin,
926 line_prefix);
927 }
928 if (plus_begin != plus_end) {
929 fn_out_diff_words_write_helper(diff_words->opt->file,
930 &style->new, style->newline,
931 plus_end - plus_begin, plus_begin,
932 line_prefix);
933 }
934
935 diff_words->current_plus = plus_end;
936 diff_words->last_minus = minus_first;
937 }
938
939 /* This function starts looking at *begin, and returns 0 iff a word was found. */
940 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
941 int *begin, int *end)
942 {
943 if (word_regex && *begin < buffer->size) {
944 regmatch_t match[1];
945 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
946 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
947 '\n', match[0].rm_eo - match[0].rm_so);
948 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
949 *begin += match[0].rm_so;
950 return *begin >= *end;
951 }
952 return -1;
953 }
954
955 /* find the next word */
956 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
957 (*begin)++;
958 if (*begin >= buffer->size)
959 return -1;
960
961 /* find the end of the word */
962 *end = *begin + 1;
963 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
964 (*end)++;
965
966 return 0;
967 }
968
969 /*
970 * This function splits the words in buffer->text, stores the list with
971 * newline separator into out, and saves the offsets of the original words
972 * in buffer->orig.
973 */
974 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
975 regex_t *word_regex)
976 {
977 int i, j;
978 long alloc = 0;
979
980 out->size = 0;
981 out->ptr = NULL;
982
983 /* fake an empty "0th" word */
984 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
985 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
986 buffer->orig_nr = 1;
987
988 for (i = 0; i < buffer->text.size; i++) {
989 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
990 return;
991
992 /* store original boundaries */
993 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
994 buffer->orig_alloc);
995 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
996 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
997 buffer->orig_nr++;
998
999 /* store one word */
1000 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1001 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1002 out->ptr[out->size + j - i] = '\n';
1003 out->size += j - i + 1;
1004
1005 i = j - 1;
1006 }
1007 }
1008
1009 /* this executes the word diff on the accumulated buffers */
1010 static void diff_words_show(struct diff_words_data *diff_words)
1011 {
1012 xpparam_t xpp;
1013 xdemitconf_t xecfg;
1014 mmfile_t minus, plus;
1015 struct diff_words_style *style = diff_words->style;
1016
1017 struct diff_options *opt = diff_words->opt;
1018 const char *line_prefix;
1019
1020 assert(opt);
1021 line_prefix = diff_line_prefix(opt);
1022
1023 /* special case: only removal */
1024 if (!diff_words->plus.text.size) {
1025 fputs(line_prefix, diff_words->opt->file);
1026 fn_out_diff_words_write_helper(diff_words->opt->file,
1027 &style->old, style->newline,
1028 diff_words->minus.text.size,
1029 diff_words->minus.text.ptr, line_prefix);
1030 diff_words->minus.text.size = 0;
1031 return;
1032 }
1033
1034 diff_words->current_plus = diff_words->plus.text.ptr;
1035 diff_words->last_minus = 0;
1036
1037 memset(&xpp, 0, sizeof(xpp));
1038 memset(&xecfg, 0, sizeof(xecfg));
1039 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1040 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1041 xpp.flags = 0;
1042 /* as only the hunk header will be parsed, we need a 0-context */
1043 xecfg.ctxlen = 0;
1044 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1045 &xpp, &xecfg);
1046 free(minus.ptr);
1047 free(plus.ptr);
1048 if (diff_words->current_plus != diff_words->plus.text.ptr +
1049 diff_words->plus.text.size) {
1050 if (color_words_output_graph_prefix(diff_words))
1051 fputs(line_prefix, diff_words->opt->file);
1052 fn_out_diff_words_write_helper(diff_words->opt->file,
1053 &style->ctx, style->newline,
1054 diff_words->plus.text.ptr + diff_words->plus.text.size
1055 - diff_words->current_plus, diff_words->current_plus,
1056 line_prefix);
1057 }
1058 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1059 }
1060
1061 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1062 static void diff_words_flush(struct emit_callback *ecbdata)
1063 {
1064 if (ecbdata->diff_words->minus.text.size ||
1065 ecbdata->diff_words->plus.text.size)
1066 diff_words_show(ecbdata->diff_words);
1067 }
1068
1069 static void diff_filespec_load_driver(struct diff_filespec *one)
1070 {
1071 /* Use already-loaded driver */
1072 if (one->driver)
1073 return;
1074
1075 if (S_ISREG(one->mode))
1076 one->driver = userdiff_find_by_path(one->path);
1077
1078 /* Fallback to default settings */
1079 if (!one->driver)
1080 one->driver = userdiff_find_by_name("default");
1081 }
1082
1083 static const char *userdiff_word_regex(struct diff_filespec *one)
1084 {
1085 diff_filespec_load_driver(one);
1086 return one->driver->word_regex;
1087 }
1088
1089 static void init_diff_words_data(struct emit_callback *ecbdata,
1090 struct diff_options *orig_opts,
1091 struct diff_filespec *one,
1092 struct diff_filespec *two)
1093 {
1094 int i;
1095 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1096 memcpy(o, orig_opts, sizeof(struct diff_options));
1097
1098 ecbdata->diff_words =
1099 xcalloc(1, sizeof(struct diff_words_data));
1100 ecbdata->diff_words->type = o->word_diff;
1101 ecbdata->diff_words->opt = o;
1102 if (!o->word_regex)
1103 o->word_regex = userdiff_word_regex(one);
1104 if (!o->word_regex)
1105 o->word_regex = userdiff_word_regex(two);
1106 if (!o->word_regex)
1107 o->word_regex = diff_word_regex_cfg;
1108 if (o->word_regex) {
1109 ecbdata->diff_words->word_regex = (regex_t *)
1110 xmalloc(sizeof(regex_t));
1111 if (regcomp(ecbdata->diff_words->word_regex,
1112 o->word_regex,
1113 REG_EXTENDED | REG_NEWLINE))
1114 die ("Invalid regular expression: %s",
1115 o->word_regex);
1116 }
1117 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1118 if (o->word_diff == diff_words_styles[i].type) {
1119 ecbdata->diff_words->style =
1120 &diff_words_styles[i];
1121 break;
1122 }
1123 }
1124 if (want_color(o->use_color)) {
1125 struct diff_words_style *st = ecbdata->diff_words->style;
1126 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1127 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1128 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1129 }
1130 }
1131
1132 static void free_diff_words_data(struct emit_callback *ecbdata)
1133 {
1134 if (ecbdata->diff_words) {
1135 diff_words_flush(ecbdata);
1136 free (ecbdata->diff_words->opt);
1137 free (ecbdata->diff_words->minus.text.ptr);
1138 free (ecbdata->diff_words->minus.orig);
1139 free (ecbdata->diff_words->plus.text.ptr);
1140 free (ecbdata->diff_words->plus.orig);
1141 if (ecbdata->diff_words->word_regex) {
1142 regfree(ecbdata->diff_words->word_regex);
1143 free(ecbdata->diff_words->word_regex);
1144 }
1145 free(ecbdata->diff_words);
1146 ecbdata->diff_words = NULL;
1147 }
1148 }
1149
1150 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1151 {
1152 if (want_color(diff_use_color))
1153 return diff_colors[ix];
1154 return "";
1155 }
1156
1157 const char *diff_line_prefix(struct diff_options *opt)
1158 {
1159 struct strbuf *msgbuf;
1160 if (!opt->output_prefix)
1161 return "";
1162
1163 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1164 return msgbuf->buf;
1165 }
1166
1167 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1168 {
1169 const char *cp;
1170 unsigned long allot;
1171 size_t l = len;
1172
1173 if (ecb->truncate)
1174 return ecb->truncate(line, len);
1175 cp = line;
1176 allot = l;
1177 while (0 < l) {
1178 (void) utf8_width(&cp, &l);
1179 if (!cp)
1180 break; /* truncated in the middle? */
1181 }
1182 return allot - l;
1183 }
1184
1185 static void find_lno(const char *line, struct emit_callback *ecbdata)
1186 {
1187 const char *p;
1188 ecbdata->lno_in_preimage = 0;
1189 ecbdata->lno_in_postimage = 0;
1190 p = strchr(line, '-');
1191 if (!p)
1192 return; /* cannot happen */
1193 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1194 p = strchr(p, '+');
1195 if (!p)
1196 return; /* cannot happen */
1197 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1198 }
1199
1200 static void fn_out_consume(void *priv, char *line, unsigned long len)
1201 {
1202 struct emit_callback *ecbdata = priv;
1203 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1204 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1205 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1206 struct diff_options *o = ecbdata->opt;
1207 const char *line_prefix = diff_line_prefix(o);
1208
1209 if (ecbdata->header) {
1210 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1211 strbuf_reset(ecbdata->header);
1212 ecbdata->header = NULL;
1213 }
1214 *(ecbdata->found_changesp) = 1;
1215
1216 if (ecbdata->label_path[0]) {
1217 const char *name_a_tab, *name_b_tab;
1218
1219 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1220 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1221
1222 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1223 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1224 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1225 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1226 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1227 }
1228
1229 if (diff_suppress_blank_empty
1230 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1231 line[0] = '\n';
1232 len = 1;
1233 }
1234
1235 if (line[0] == '@') {
1236 if (ecbdata->diff_words)
1237 diff_words_flush(ecbdata);
1238 len = sane_truncate_line(ecbdata, line, len);
1239 find_lno(line, ecbdata);
1240 emit_hunk_header(ecbdata, line, len);
1241 if (line[len-1] != '\n')
1242 putc('\n', ecbdata->opt->file);
1243 return;
1244 }
1245
1246 if (len < 1) {
1247 emit_line(ecbdata->opt, reset, reset, line, len);
1248 if (ecbdata->diff_words
1249 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1250 fputs("~\n", ecbdata->opt->file);
1251 return;
1252 }
1253
1254 if (ecbdata->diff_words) {
1255 if (line[0] == '-') {
1256 diff_words_append(line, len,
1257 &ecbdata->diff_words->minus);
1258 return;
1259 } else if (line[0] == '+') {
1260 diff_words_append(line, len,
1261 &ecbdata->diff_words->plus);
1262 return;
1263 } else if (starts_with(line, "\\ ")) {
1264 /*
1265 * Eat the "no newline at eof" marker as if we
1266 * saw a "+" or "-" line with nothing on it,
1267 * and return without diff_words_flush() to
1268 * defer processing. If this is the end of
1269 * preimage, more "+" lines may come after it.
1270 */
1271 return;
1272 }
1273 diff_words_flush(ecbdata);
1274 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1275 emit_line(ecbdata->opt, context, reset, line, len);
1276 fputs("~\n", ecbdata->opt->file);
1277 } else {
1278 /*
1279 * Skip the prefix character, if any. With
1280 * diff_suppress_blank_empty, there may be
1281 * none.
1282 */
1283 if (line[0] != '\n') {
1284 line++;
1285 len--;
1286 }
1287 emit_line(ecbdata->opt, context, reset, line, len);
1288 }
1289 return;
1290 }
1291
1292 switch (line[0]) {
1293 case '+':
1294 ecbdata->lno_in_postimage++;
1295 emit_add_line(reset, ecbdata, line + 1, len - 1);
1296 break;
1297 case '-':
1298 ecbdata->lno_in_preimage++;
1299 emit_del_line(reset, ecbdata, line + 1, len - 1);
1300 break;
1301 case ' ':
1302 ecbdata->lno_in_postimage++;
1303 ecbdata->lno_in_preimage++;
1304 emit_context_line(reset, ecbdata, line + 1, len - 1);
1305 break;
1306 default:
1307 /* incomplete line at the end */
1308 ecbdata->lno_in_preimage++;
1309 emit_line(ecbdata->opt,
1310 diff_get_color(ecbdata->color_diff, DIFF_CONTEXT),
1311 reset, line, len);
1312 break;
1313 }
1314 }
1315
1316 static char *pprint_rename(const char *a, const char *b)
1317 {
1318 const char *old = a;
1319 const char *new = b;
1320 struct strbuf name = STRBUF_INIT;
1321 int pfx_length, sfx_length;
1322 int pfx_adjust_for_slash;
1323 int len_a = strlen(a);
1324 int len_b = strlen(b);
1325 int a_midlen, b_midlen;
1326 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1327 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1328
1329 if (qlen_a || qlen_b) {
1330 quote_c_style(a, &name, NULL, 0);
1331 strbuf_addstr(&name, " => ");
1332 quote_c_style(b, &name, NULL, 0);
1333 return strbuf_detach(&name, NULL);
1334 }
1335
1336 /* Find common prefix */
1337 pfx_length = 0;
1338 while (*old && *new && *old == *new) {
1339 if (*old == '/')
1340 pfx_length = old - a + 1;
1341 old++;
1342 new++;
1343 }
1344
1345 /* Find common suffix */
1346 old = a + len_a;
1347 new = b + len_b;
1348 sfx_length = 0;
1349 /*
1350 * If there is a common prefix, it must end in a slash. In
1351 * that case we let this loop run 1 into the prefix to see the
1352 * same slash.
1353 *
1354 * If there is no common prefix, we cannot do this as it would
1355 * underrun the input strings.
1356 */
1357 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1358 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1359 b + pfx_length - pfx_adjust_for_slash <= new &&
1360 *old == *new) {
1361 if (*old == '/')
1362 sfx_length = len_a - (old - a);
1363 old--;
1364 new--;
1365 }
1366
1367 /*
1368 * pfx{mid-a => mid-b}sfx
1369 * {pfx-a => pfx-b}sfx
1370 * pfx{sfx-a => sfx-b}
1371 * name-a => name-b
1372 */
1373 a_midlen = len_a - pfx_length - sfx_length;
1374 b_midlen = len_b - pfx_length - sfx_length;
1375 if (a_midlen < 0)
1376 a_midlen = 0;
1377 if (b_midlen < 0)
1378 b_midlen = 0;
1379
1380 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1381 if (pfx_length + sfx_length) {
1382 strbuf_add(&name, a, pfx_length);
1383 strbuf_addch(&name, '{');
1384 }
1385 strbuf_add(&name, a + pfx_length, a_midlen);
1386 strbuf_addstr(&name, " => ");
1387 strbuf_add(&name, b + pfx_length, b_midlen);
1388 if (pfx_length + sfx_length) {
1389 strbuf_addch(&name, '}');
1390 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1391 }
1392 return strbuf_detach(&name, NULL);
1393 }
1394
1395 struct diffstat_t {
1396 int nr;
1397 int alloc;
1398 struct diffstat_file {
1399 char *from_name;
1400 char *name;
1401 char *print_name;
1402 unsigned is_unmerged:1;
1403 unsigned is_binary:1;
1404 unsigned is_renamed:1;
1405 unsigned is_interesting:1;
1406 uintmax_t added, deleted;
1407 } **files;
1408 };
1409
1410 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1411 const char *name_a,
1412 const char *name_b)
1413 {
1414 struct diffstat_file *x;
1415 x = xcalloc(1, sizeof(*x));
1416 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1417 diffstat->files[diffstat->nr++] = x;
1418 if (name_b) {
1419 x->from_name = xstrdup(name_a);
1420 x->name = xstrdup(name_b);
1421 x->is_renamed = 1;
1422 }
1423 else {
1424 x->from_name = NULL;
1425 x->name = xstrdup(name_a);
1426 }
1427 return x;
1428 }
1429
1430 static void diffstat_consume(void *priv, char *line, unsigned long len)
1431 {
1432 struct diffstat_t *diffstat = priv;
1433 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1434
1435 if (line[0] == '+')
1436 x->added++;
1437 else if (line[0] == '-')
1438 x->deleted++;
1439 }
1440
1441 const char mime_boundary_leader[] = "------------";
1442
1443 static int scale_linear(int it, int width, int max_change)
1444 {
1445 if (!it)
1446 return 0;
1447 /*
1448 * make sure that at least one '-' or '+' is printed if
1449 * there is any change to this path. The easiest way is to
1450 * scale linearly as if the alloted width is one column shorter
1451 * than it is, and then add 1 to the result.
1452 */
1453 return 1 + (it * (width - 1) / max_change);
1454 }
1455
1456 static void show_name(FILE *file,
1457 const char *prefix, const char *name, int len)
1458 {
1459 fprintf(file, " %s%-*s |", prefix, len, name);
1460 }
1461
1462 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1463 {
1464 if (cnt <= 0)
1465 return;
1466 fprintf(file, "%s", set);
1467 while (cnt--)
1468 putc(ch, file);
1469 fprintf(file, "%s", reset);
1470 }
1471
1472 static void fill_print_name(struct diffstat_file *file)
1473 {
1474 char *pname;
1475
1476 if (file->print_name)
1477 return;
1478
1479 if (!file->is_renamed) {
1480 struct strbuf buf = STRBUF_INIT;
1481 if (quote_c_style(file->name, &buf, NULL, 0)) {
1482 pname = strbuf_detach(&buf, NULL);
1483 } else {
1484 pname = file->name;
1485 strbuf_release(&buf);
1486 }
1487 } else {
1488 pname = pprint_rename(file->from_name, file->name);
1489 }
1490 file->print_name = pname;
1491 }
1492
1493 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1494 {
1495 struct strbuf sb = STRBUF_INIT;
1496 int ret;
1497
1498 if (!files) {
1499 assert(insertions == 0 && deletions == 0);
1500 return fprintf(fp, "%s\n", " 0 files changed");
1501 }
1502
1503 strbuf_addf(&sb,
1504 (files == 1) ? " %d file changed" : " %d files changed",
1505 files);
1506
1507 /*
1508 * For binary diff, the caller may want to print "x files
1509 * changed" with insertions == 0 && deletions == 0.
1510 *
1511 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1512 * is probably less confusing (i.e skip over "2 files changed
1513 * but nothing about added/removed lines? Is this a bug in Git?").
1514 */
1515 if (insertions || deletions == 0) {
1516 strbuf_addf(&sb,
1517 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1518 insertions);
1519 }
1520
1521 if (deletions || insertions == 0) {
1522 strbuf_addf(&sb,
1523 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1524 deletions);
1525 }
1526 strbuf_addch(&sb, '\n');
1527 ret = fputs(sb.buf, fp);
1528 strbuf_release(&sb);
1529 return ret;
1530 }
1531
1532 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1533 {
1534 int i, len, add, del, adds = 0, dels = 0;
1535 uintmax_t max_change = 0, max_len = 0;
1536 int total_files = data->nr, count;
1537 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1538 const char *reset, *add_c, *del_c;
1539 const char *line_prefix = "";
1540 int extra_shown = 0;
1541
1542 if (data->nr == 0)
1543 return;
1544
1545 line_prefix = diff_line_prefix(options);
1546 count = options->stat_count ? options->stat_count : data->nr;
1547
1548 reset = diff_get_color_opt(options, DIFF_RESET);
1549 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1550 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1551
1552 /*
1553 * Find the longest filename and max number of changes
1554 */
1555 for (i = 0; (i < count) && (i < data->nr); i++) {
1556 struct diffstat_file *file = data->files[i];
1557 uintmax_t change = file->added + file->deleted;
1558
1559 if (!file->is_interesting && (change == 0)) {
1560 count++; /* not shown == room for one more */
1561 continue;
1562 }
1563 fill_print_name(file);
1564 len = strlen(file->print_name);
1565 if (max_len < len)
1566 max_len = len;
1567
1568 if (file->is_unmerged) {
1569 /* "Unmerged" is 8 characters */
1570 bin_width = bin_width < 8 ? 8 : bin_width;
1571 continue;
1572 }
1573 if (file->is_binary) {
1574 /* "Bin XXX -> YYY bytes" */
1575 int w = 14 + decimal_width(file->added)
1576 + decimal_width(file->deleted);
1577 bin_width = bin_width < w ? w : bin_width;
1578 /* Display change counts aligned with "Bin" */
1579 number_width = 3;
1580 continue;
1581 }
1582
1583 if (max_change < change)
1584 max_change = change;
1585 }
1586 count = i; /* where we can stop scanning in data->files[] */
1587
1588 /*
1589 * We have width = stat_width or term_columns() columns total.
1590 * We want a maximum of min(max_len, stat_name_width) for the name part.
1591 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1592 * We also need 1 for " " and 4 + decimal_width(max_change)
1593 * for " | NNNN " and one the empty column at the end, altogether
1594 * 6 + decimal_width(max_change).
1595 *
1596 * If there's not enough space, we will use the smaller of
1597 * stat_name_width (if set) and 5/8*width for the filename,
1598 * and the rest for constant elements + graph part, but no more
1599 * than stat_graph_width for the graph part.
1600 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1601 * for the standard terminal size).
1602 *
1603 * In other words: stat_width limits the maximum width, and
1604 * stat_name_width fixes the maximum width of the filename,
1605 * and is also used to divide available columns if there
1606 * aren't enough.
1607 *
1608 * Binary files are displayed with "Bin XXX -> YYY bytes"
1609 * instead of the change count and graph. This part is treated
1610 * similarly to the graph part, except that it is not
1611 * "scaled". If total width is too small to accommodate the
1612 * guaranteed minimum width of the filename part and the
1613 * separators and this message, this message will "overflow"
1614 * making the line longer than the maximum width.
1615 */
1616
1617 if (options->stat_width == -1)
1618 width = term_columns() - options->output_prefix_length;
1619 else
1620 width = options->stat_width ? options->stat_width : 80;
1621 number_width = decimal_width(max_change) > number_width ?
1622 decimal_width(max_change) : number_width;
1623
1624 if (options->stat_graph_width == -1)
1625 options->stat_graph_width = diff_stat_graph_width;
1626
1627 /*
1628 * Guarantee 3/8*16==6 for the graph part
1629 * and 5/8*16==10 for the filename part
1630 */
1631 if (width < 16 + 6 + number_width)
1632 width = 16 + 6 + number_width;
1633
1634 /*
1635 * First assign sizes that are wanted, ignoring available width.
1636 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1637 * starting from "XXX" should fit in graph_width.
1638 */
1639 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1640 if (options->stat_graph_width &&
1641 options->stat_graph_width < graph_width)
1642 graph_width = options->stat_graph_width;
1643
1644 name_width = (options->stat_name_width > 0 &&
1645 options->stat_name_width < max_len) ?
1646 options->stat_name_width : max_len;
1647
1648 /*
1649 * Adjust adjustable widths not to exceed maximum width
1650 */
1651 if (name_width + number_width + 6 + graph_width > width) {
1652 if (graph_width > width * 3/8 - number_width - 6) {
1653 graph_width = width * 3/8 - number_width - 6;
1654 if (graph_width < 6)
1655 graph_width = 6;
1656 }
1657
1658 if (options->stat_graph_width &&
1659 graph_width > options->stat_graph_width)
1660 graph_width = options->stat_graph_width;
1661 if (name_width > width - number_width - 6 - graph_width)
1662 name_width = width - number_width - 6 - graph_width;
1663 else
1664 graph_width = width - number_width - 6 - name_width;
1665 }
1666
1667 /*
1668 * From here name_width is the width of the name area,
1669 * and graph_width is the width of the graph area.
1670 * max_change is used to scale graph properly.
1671 */
1672 for (i = 0; i < count; i++) {
1673 const char *prefix = "";
1674 struct diffstat_file *file = data->files[i];
1675 char *name = file->print_name;
1676 uintmax_t added = file->added;
1677 uintmax_t deleted = file->deleted;
1678 int name_len;
1679
1680 if (!file->is_interesting && (added + deleted == 0))
1681 continue;
1682
1683 /*
1684 * "scale" the filename
1685 */
1686 len = name_width;
1687 name_len = strlen(name);
1688 if (name_width < name_len) {
1689 char *slash;
1690 prefix = "...";
1691 len -= 3;
1692 name += name_len - len;
1693 slash = strchr(name, '/');
1694 if (slash)
1695 name = slash;
1696 }
1697
1698 if (file->is_binary) {
1699 fprintf(options->file, "%s", line_prefix);
1700 show_name(options->file, prefix, name, len);
1701 fprintf(options->file, " %*s", number_width, "Bin");
1702 if (!added && !deleted) {
1703 putc('\n', options->file);
1704 continue;
1705 }
1706 fprintf(options->file, " %s%"PRIuMAX"%s",
1707 del_c, deleted, reset);
1708 fprintf(options->file, " -> ");
1709 fprintf(options->file, "%s%"PRIuMAX"%s",
1710 add_c, added, reset);
1711 fprintf(options->file, " bytes");
1712 fprintf(options->file, "\n");
1713 continue;
1714 }
1715 else if (file->is_unmerged) {
1716 fprintf(options->file, "%s", line_prefix);
1717 show_name(options->file, prefix, name, len);
1718 fprintf(options->file, " Unmerged\n");
1719 continue;
1720 }
1721
1722 /*
1723 * scale the add/delete
1724 */
1725 add = added;
1726 del = deleted;
1727
1728 if (graph_width <= max_change) {
1729 int total = scale_linear(add + del, graph_width, max_change);
1730 if (total < 2 && add && del)
1731 /* width >= 2 due to the sanity check */
1732 total = 2;
1733 if (add < del) {
1734 add = scale_linear(add, graph_width, max_change);
1735 del = total - add;
1736 } else {
1737 del = scale_linear(del, graph_width, max_change);
1738 add = total - del;
1739 }
1740 }
1741 fprintf(options->file, "%s", line_prefix);
1742 show_name(options->file, prefix, name, len);
1743 fprintf(options->file, " %*"PRIuMAX"%s",
1744 number_width, added + deleted,
1745 added + deleted ? " " : "");
1746 show_graph(options->file, '+', add, add_c, reset);
1747 show_graph(options->file, '-', del, del_c, reset);
1748 fprintf(options->file, "\n");
1749 }
1750
1751 for (i = 0; i < data->nr; i++) {
1752 struct diffstat_file *file = data->files[i];
1753 uintmax_t added = file->added;
1754 uintmax_t deleted = file->deleted;
1755
1756 if (file->is_unmerged ||
1757 (!file->is_interesting && (added + deleted == 0))) {
1758 total_files--;
1759 continue;
1760 }
1761
1762 if (!file->is_binary) {
1763 adds += added;
1764 dels += deleted;
1765 }
1766 if (i < count)
1767 continue;
1768 if (!extra_shown)
1769 fprintf(options->file, "%s ...\n", line_prefix);
1770 extra_shown = 1;
1771 }
1772 fprintf(options->file, "%s", line_prefix);
1773 print_stat_summary(options->file, total_files, adds, dels);
1774 }
1775
1776 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1777 {
1778 int i, adds = 0, dels = 0, total_files = data->nr;
1779
1780 if (data->nr == 0)
1781 return;
1782
1783 for (i = 0; i < data->nr; i++) {
1784 int added = data->files[i]->added;
1785 int deleted= data->files[i]->deleted;
1786
1787 if (data->files[i]->is_unmerged ||
1788 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1789 total_files--;
1790 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1791 adds += added;
1792 dels += deleted;
1793 }
1794 }
1795 fprintf(options->file, "%s", diff_line_prefix(options));
1796 print_stat_summary(options->file, total_files, adds, dels);
1797 }
1798
1799 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1800 {
1801 int i;
1802
1803 if (data->nr == 0)
1804 return;
1805
1806 for (i = 0; i < data->nr; i++) {
1807 struct diffstat_file *file = data->files[i];
1808
1809 fprintf(options->file, "%s", diff_line_prefix(options));
1810
1811 if (file->is_binary)
1812 fprintf(options->file, "-\t-\t");
1813 else
1814 fprintf(options->file,
1815 "%"PRIuMAX"\t%"PRIuMAX"\t",
1816 file->added, file->deleted);
1817 if (options->line_termination) {
1818 fill_print_name(file);
1819 if (!file->is_renamed)
1820 write_name_quoted(file->name, options->file,
1821 options->line_termination);
1822 else {
1823 fputs(file->print_name, options->file);
1824 putc(options->line_termination, options->file);
1825 }
1826 } else {
1827 if (file->is_renamed) {
1828 putc('\0', options->file);
1829 write_name_quoted(file->from_name, options->file, '\0');
1830 }
1831 write_name_quoted(file->name, options->file, '\0');
1832 }
1833 }
1834 }
1835
1836 struct dirstat_file {
1837 const char *name;
1838 unsigned long changed;
1839 };
1840
1841 struct dirstat_dir {
1842 struct dirstat_file *files;
1843 int alloc, nr, permille, cumulative;
1844 };
1845
1846 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1847 unsigned long changed, const char *base, int baselen)
1848 {
1849 unsigned long this_dir = 0;
1850 unsigned int sources = 0;
1851 const char *line_prefix = diff_line_prefix(opt);
1852
1853 while (dir->nr) {
1854 struct dirstat_file *f = dir->files;
1855 int namelen = strlen(f->name);
1856 unsigned long this;
1857 char *slash;
1858
1859 if (namelen < baselen)
1860 break;
1861 if (memcmp(f->name, base, baselen))
1862 break;
1863 slash = strchr(f->name + baselen, '/');
1864 if (slash) {
1865 int newbaselen = slash + 1 - f->name;
1866 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1867 sources++;
1868 } else {
1869 this = f->changed;
1870 dir->files++;
1871 dir->nr--;
1872 sources += 2;
1873 }
1874 this_dir += this;
1875 }
1876
1877 /*
1878 * We don't report dirstat's for
1879 * - the top level
1880 * - or cases where everything came from a single directory
1881 * under this directory (sources == 1).
1882 */
1883 if (baselen && sources != 1) {
1884 if (this_dir) {
1885 int permille = this_dir * 1000 / changed;
1886 if (permille >= dir->permille) {
1887 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1888 permille / 10, permille % 10, baselen, base);
1889 if (!dir->cumulative)
1890 return 0;
1891 }
1892 }
1893 }
1894 return this_dir;
1895 }
1896
1897 static int dirstat_compare(const void *_a, const void *_b)
1898 {
1899 const struct dirstat_file *a = _a;
1900 const struct dirstat_file *b = _b;
1901 return strcmp(a->name, b->name);
1902 }
1903
1904 static void show_dirstat(struct diff_options *options)
1905 {
1906 int i;
1907 unsigned long changed;
1908 struct dirstat_dir dir;
1909 struct diff_queue_struct *q = &diff_queued_diff;
1910
1911 dir.files = NULL;
1912 dir.alloc = 0;
1913 dir.nr = 0;
1914 dir.permille = options->dirstat_permille;
1915 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1916
1917 changed = 0;
1918 for (i = 0; i < q->nr; i++) {
1919 struct diff_filepair *p = q->queue[i];
1920 const char *name;
1921 unsigned long copied, added, damage;
1922 int content_changed;
1923
1924 name = p->two->path ? p->two->path : p->one->path;
1925
1926 if (p->one->sha1_valid && p->two->sha1_valid)
1927 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1928 else
1929 content_changed = 1;
1930
1931 if (!content_changed) {
1932 /*
1933 * The SHA1 has not changed, so pre-/post-content is
1934 * identical. We can therefore skip looking at the
1935 * file contents altogether.
1936 */
1937 damage = 0;
1938 goto found_damage;
1939 }
1940
1941 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1942 /*
1943 * In --dirstat-by-file mode, we don't really need to
1944 * look at the actual file contents at all.
1945 * The fact that the SHA1 changed is enough for us to
1946 * add this file to the list of results
1947 * (with each file contributing equal damage).
1948 */
1949 damage = 1;
1950 goto found_damage;
1951 }
1952
1953 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1954 diff_populate_filespec(p->one, 0);
1955 diff_populate_filespec(p->two, 0);
1956 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1957 &copied, &added);
1958 diff_free_filespec_data(p->one);
1959 diff_free_filespec_data(p->two);
1960 } else if (DIFF_FILE_VALID(p->one)) {
1961 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
1962 copied = added = 0;
1963 diff_free_filespec_data(p->one);
1964 } else if (DIFF_FILE_VALID(p->two)) {
1965 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
1966 copied = 0;
1967 added = p->two->size;
1968 diff_free_filespec_data(p->two);
1969 } else
1970 continue;
1971
1972 /*
1973 * Original minus copied is the removed material,
1974 * added is the new material. They are both damages
1975 * made to the preimage.
1976 * If the resulting damage is zero, we know that
1977 * diffcore_count_changes() considers the two entries to
1978 * be identical, but since content_changed is true, we
1979 * know that there must have been _some_ kind of change,
1980 * so we force all entries to have damage > 0.
1981 */
1982 damage = (p->one->size - copied) + added;
1983 if (!damage)
1984 damage = 1;
1985
1986 found_damage:
1987 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1988 dir.files[dir.nr].name = 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 show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2004 {
2005 int i;
2006 unsigned long changed;
2007 struct dirstat_dir dir;
2008
2009 if (data->nr == 0)
2010 return;
2011
2012 dir.files = NULL;
2013 dir.alloc = 0;
2014 dir.nr = 0;
2015 dir.permille = options->dirstat_permille;
2016 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2017
2018 changed = 0;
2019 for (i = 0; i < data->nr; i++) {
2020 struct diffstat_file *file = data->files[i];
2021 unsigned long damage = file->added + file->deleted;
2022 if (file->is_binary)
2023 /*
2024 * binary files counts bytes, not lines. Must find some
2025 * way to normalize binary bytes vs. textual lines.
2026 * The following heuristic assumes that there are 64
2027 * bytes per "line".
2028 * This is stupid and ugly, but very cheap...
2029 */
2030 damage = (damage + 63) / 64;
2031 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2032 dir.files[dir.nr].name = file->name;
2033 dir.files[dir.nr].changed = damage;
2034 changed += damage;
2035 dir.nr++;
2036 }
2037
2038 /* This can happen even with many files, if everything was renames */
2039 if (!changed)
2040 return;
2041
2042 /* Show all directories with more than x% of the changes */
2043 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
2044 gather_dirstat(options, &dir, changed, "", 0);
2045 }
2046
2047 static void free_diffstat_info(struct diffstat_t *diffstat)
2048 {
2049 int i;
2050 for (i = 0; i < diffstat->nr; i++) {
2051 struct diffstat_file *f = diffstat->files[i];
2052 if (f->name != f->print_name)
2053 free(f->print_name);
2054 free(f->name);
2055 free(f->from_name);
2056 free(f);
2057 }
2058 free(diffstat->files);
2059 }
2060
2061 struct checkdiff_t {
2062 const char *filename;
2063 int lineno;
2064 int conflict_marker_size;
2065 struct diff_options *o;
2066 unsigned ws_rule;
2067 unsigned status;
2068 };
2069
2070 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2071 {
2072 char firstchar;
2073 int cnt;
2074
2075 if (len < marker_size + 1)
2076 return 0;
2077 firstchar = line[0];
2078 switch (firstchar) {
2079 case '=': case '>': case '<': case '|':
2080 break;
2081 default:
2082 return 0;
2083 }
2084 for (cnt = 1; cnt < marker_size; cnt++)
2085 if (line[cnt] != firstchar)
2086 return 0;
2087 /* line[1] thru line[marker_size-1] are same as firstchar */
2088 if (len < marker_size + 1 || !isspace(line[marker_size]))
2089 return 0;
2090 return 1;
2091 }
2092
2093 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2094 {
2095 struct checkdiff_t *data = priv;
2096 int marker_size = data->conflict_marker_size;
2097 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2098 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2099 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2100 char *err;
2101 const char *line_prefix;
2102
2103 assert(data->o);
2104 line_prefix = diff_line_prefix(data->o);
2105
2106 if (line[0] == '+') {
2107 unsigned bad;
2108 data->lineno++;
2109 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2110 data->status |= 1;
2111 fprintf(data->o->file,
2112 "%s%s:%d: leftover conflict marker\n",
2113 line_prefix, data->filename, data->lineno);
2114 }
2115 bad = ws_check(line + 1, len - 1, data->ws_rule);
2116 if (!bad)
2117 return;
2118 data->status |= bad;
2119 err = whitespace_error_string(bad);
2120 fprintf(data->o->file, "%s%s:%d: %s.\n",
2121 line_prefix, data->filename, data->lineno, err);
2122 free(err);
2123 emit_line(data->o, set, reset, line, 1);
2124 ws_check_emit(line + 1, len - 1, data->ws_rule,
2125 data->o->file, set, reset, ws);
2126 } else if (line[0] == ' ') {
2127 data->lineno++;
2128 } else if (line[0] == '@') {
2129 char *plus = strchr(line, '+');
2130 if (plus)
2131 data->lineno = strtol(plus, NULL, 10) - 1;
2132 else
2133 die("invalid diff");
2134 }
2135 }
2136
2137 static unsigned char *deflate_it(char *data,
2138 unsigned long size,
2139 unsigned long *result_size)
2140 {
2141 int bound;
2142 unsigned char *deflated;
2143 git_zstream stream;
2144
2145 git_deflate_init(&stream, zlib_compression_level);
2146 bound = git_deflate_bound(&stream, size);
2147 deflated = xmalloc(bound);
2148 stream.next_out = deflated;
2149 stream.avail_out = bound;
2150
2151 stream.next_in = (unsigned char *)data;
2152 stream.avail_in = size;
2153 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2154 ; /* nothing */
2155 git_deflate_end(&stream);
2156 *result_size = stream.total_out;
2157 return deflated;
2158 }
2159
2160 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2161 const char *prefix)
2162 {
2163 void *cp;
2164 void *delta;
2165 void *deflated;
2166 void *data;
2167 unsigned long orig_size;
2168 unsigned long delta_size;
2169 unsigned long deflate_size;
2170 unsigned long data_size;
2171
2172 /* We could do deflated delta, or we could do just deflated two,
2173 * whichever is smaller.
2174 */
2175 delta = NULL;
2176 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2177 if (one->size && two->size) {
2178 delta = diff_delta(one->ptr, one->size,
2179 two->ptr, two->size,
2180 &delta_size, deflate_size);
2181 if (delta) {
2182 void *to_free = delta;
2183 orig_size = delta_size;
2184 delta = deflate_it(delta, delta_size, &delta_size);
2185 free(to_free);
2186 }
2187 }
2188
2189 if (delta && delta_size < deflate_size) {
2190 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2191 free(deflated);
2192 data = delta;
2193 data_size = delta_size;
2194 }
2195 else {
2196 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2197 free(delta);
2198 data = deflated;
2199 data_size = deflate_size;
2200 }
2201
2202 /* emit data encoded in base85 */
2203 cp = data;
2204 while (data_size) {
2205 int bytes = (52 < data_size) ? 52 : data_size;
2206 char line[70];
2207 data_size -= bytes;
2208 if (bytes <= 26)
2209 line[0] = bytes + 'A' - 1;
2210 else
2211 line[0] = bytes - 26 + 'a' - 1;
2212 encode_85(line + 1, cp, bytes);
2213 cp = (char *) cp + bytes;
2214 fprintf(file, "%s", prefix);
2215 fputs(line, file);
2216 fputc('\n', file);
2217 }
2218 fprintf(file, "%s\n", prefix);
2219 free(data);
2220 }
2221
2222 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2223 const char *prefix)
2224 {
2225 fprintf(file, "%sGIT binary patch\n", prefix);
2226 emit_binary_diff_body(file, one, two, prefix);
2227 emit_binary_diff_body(file, two, one, prefix);
2228 }
2229
2230 int diff_filespec_is_binary(struct diff_filespec *one)
2231 {
2232 if (one->is_binary == -1) {
2233 diff_filespec_load_driver(one);
2234 if (one->driver->binary != -1)
2235 one->is_binary = one->driver->binary;
2236 else {
2237 if (!one->data && DIFF_FILE_VALID(one))
2238 diff_populate_filespec(one, CHECK_BINARY);
2239 if (one->is_binary == -1 && one->data)
2240 one->is_binary = buffer_is_binary(one->data,
2241 one->size);
2242 if (one->is_binary == -1)
2243 one->is_binary = 0;
2244 }
2245 }
2246 return one->is_binary;
2247 }
2248
2249 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2250 {
2251 diff_filespec_load_driver(one);
2252 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2253 }
2254
2255 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2256 {
2257 if (!options->a_prefix)
2258 options->a_prefix = a;
2259 if (!options->b_prefix)
2260 options->b_prefix = b;
2261 }
2262
2263 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2264 {
2265 if (!DIFF_FILE_VALID(one))
2266 return NULL;
2267
2268 diff_filespec_load_driver(one);
2269 return userdiff_get_textconv(one->driver);
2270 }
2271
2272 static void builtin_diff(const char *name_a,
2273 const char *name_b,
2274 struct diff_filespec *one,
2275 struct diff_filespec *two,
2276 const char *xfrm_msg,
2277 int must_show_header,
2278 struct diff_options *o,
2279 int complete_rewrite)
2280 {
2281 mmfile_t mf1, mf2;
2282 const char *lbl[2];
2283 char *a_one, *b_two;
2284 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2285 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2286 const char *a_prefix, *b_prefix;
2287 struct userdiff_driver *textconv_one = NULL;
2288 struct userdiff_driver *textconv_two = NULL;
2289 struct strbuf header = STRBUF_INIT;
2290 const char *line_prefix = diff_line_prefix(o);
2291
2292 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2293 (!one->mode || S_ISGITLINK(one->mode)) &&
2294 (!two->mode || S_ISGITLINK(two->mode))) {
2295 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2296 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2297 show_submodule_summary(o->file, one->path ? one->path : two->path,
2298 line_prefix,
2299 one->sha1, two->sha1, two->dirty_submodule,
2300 meta, del, add, reset);
2301 return;
2302 }
2303
2304 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2305 textconv_one = get_textconv(one);
2306 textconv_two = get_textconv(two);
2307 }
2308
2309 diff_set_mnemonic_prefix(o, "a/", "b/");
2310 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2311 a_prefix = o->b_prefix;
2312 b_prefix = o->a_prefix;
2313 } else {
2314 a_prefix = o->a_prefix;
2315 b_prefix = o->b_prefix;
2316 }
2317
2318 /* Never use a non-valid filename anywhere if at all possible */
2319 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2320 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2321
2322 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2323 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2324 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2325 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2326 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2327 if (lbl[0][0] == '/') {
2328 /* /dev/null */
2329 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2330 if (xfrm_msg)
2331 strbuf_addstr(&header, xfrm_msg);
2332 must_show_header = 1;
2333 }
2334 else if (lbl[1][0] == '/') {
2335 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2336 if (xfrm_msg)
2337 strbuf_addstr(&header, xfrm_msg);
2338 must_show_header = 1;
2339 }
2340 else {
2341 if (one->mode != two->mode) {
2342 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2343 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2344 must_show_header = 1;
2345 }
2346 if (xfrm_msg)
2347 strbuf_addstr(&header, xfrm_msg);
2348
2349 /*
2350 * we do not run diff between different kind
2351 * of objects.
2352 */
2353 if ((one->mode ^ two->mode) & S_IFMT)
2354 goto free_ab_and_return;
2355 if (complete_rewrite &&
2356 (textconv_one || !diff_filespec_is_binary(one)) &&
2357 (textconv_two || !diff_filespec_is_binary(two))) {
2358 fprintf(o->file, "%s", header.buf);
2359 strbuf_reset(&header);
2360 emit_rewrite_diff(name_a, name_b, one, two,
2361 textconv_one, textconv_two, o);
2362 o->found_changes = 1;
2363 goto free_ab_and_return;
2364 }
2365 }
2366
2367 if (o->irreversible_delete && lbl[1][0] == '/') {
2368 fprintf(o->file, "%s", header.buf);
2369 strbuf_reset(&header);
2370 goto free_ab_and_return;
2371 } else if (!DIFF_OPT_TST(o, TEXT) &&
2372 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2373 (!textconv_two && diff_filespec_is_binary(two)) )) {
2374 if (!one->data && !two->data &&
2375 S_ISREG(one->mode) && S_ISREG(two->mode) &&
2376 !DIFF_OPT_TST(o, BINARY)) {
2377 if (!hashcmp(one->sha1, two->sha1)) {
2378 if (must_show_header)
2379 fprintf(o->file, "%s", header.buf);
2380 goto free_ab_and_return;
2381 }
2382 fprintf(o->file, "%s", header.buf);
2383 fprintf(o->file, "%sBinary files %s and %s differ\n",
2384 line_prefix, lbl[0], lbl[1]);
2385 goto free_ab_and_return;
2386 }
2387 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2388 die("unable to read files to diff");
2389 /* Quite common confusing case */
2390 if (mf1.size == mf2.size &&
2391 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2392 if (must_show_header)
2393 fprintf(o->file, "%s", header.buf);
2394 goto free_ab_and_return;
2395 }
2396 fprintf(o->file, "%s", header.buf);
2397 strbuf_reset(&header);
2398 if (DIFF_OPT_TST(o, BINARY))
2399 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2400 else
2401 fprintf(o->file, "%sBinary files %s and %s differ\n",
2402 line_prefix, lbl[0], lbl[1]);
2403 o->found_changes = 1;
2404 } else {
2405 /* Crazy xdl interfaces.. */
2406 const char *diffopts = getenv("GIT_DIFF_OPTS");
2407 const char *v;
2408 xpparam_t xpp;
2409 xdemitconf_t xecfg;
2410 struct emit_callback ecbdata;
2411 const struct userdiff_funcname *pe;
2412
2413 if (must_show_header) {
2414 fprintf(o->file, "%s", header.buf);
2415 strbuf_reset(&header);
2416 }
2417
2418 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2419 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2420
2421 pe = diff_funcname_pattern(one);
2422 if (!pe)
2423 pe = diff_funcname_pattern(two);
2424
2425 memset(&xpp, 0, sizeof(xpp));
2426 memset(&xecfg, 0, sizeof(xecfg));
2427 memset(&ecbdata, 0, sizeof(ecbdata));
2428 ecbdata.label_path = lbl;
2429 ecbdata.color_diff = want_color(o->use_color);
2430 ecbdata.found_changesp = &o->found_changes;
2431 ecbdata.ws_rule = whitespace_rule(name_b);
2432 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2433 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2434 ecbdata.opt = o;
2435 ecbdata.header = header.len ? &header : NULL;
2436 xpp.flags = o->xdl_opts;
2437 xecfg.ctxlen = o->context;
2438 xecfg.interhunkctxlen = o->interhunkcontext;
2439 xecfg.flags = XDL_EMIT_FUNCNAMES;
2440 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2441 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2442 if (pe)
2443 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2444 if (!diffopts)
2445 ;
2446 else if (skip_prefix(diffopts, "--unified=", &v))
2447 xecfg.ctxlen = strtoul(v, NULL, 10);
2448 else if (skip_prefix(diffopts, "-u", &v))
2449 xecfg.ctxlen = strtoul(v, NULL, 10);
2450 if (o->word_diff)
2451 init_diff_words_data(&ecbdata, o, one, two);
2452 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2453 &xpp, &xecfg);
2454 if (o->word_diff)
2455 free_diff_words_data(&ecbdata);
2456 if (textconv_one)
2457 free(mf1.ptr);
2458 if (textconv_two)
2459 free(mf2.ptr);
2460 xdiff_clear_find_func(&xecfg);
2461 }
2462
2463 free_ab_and_return:
2464 strbuf_release(&header);
2465 diff_free_filespec_data(one);
2466 diff_free_filespec_data(two);
2467 free(a_one);
2468 free(b_two);
2469 return;
2470 }
2471
2472 static void builtin_diffstat(const char *name_a, const char *name_b,
2473 struct diff_filespec *one,
2474 struct diff_filespec *two,
2475 struct diffstat_t *diffstat,
2476 struct diff_options *o,
2477 struct diff_filepair *p)
2478 {
2479 mmfile_t mf1, mf2;
2480 struct diffstat_file *data;
2481 int same_contents;
2482 int complete_rewrite = 0;
2483
2484 if (!DIFF_PAIR_UNMERGED(p)) {
2485 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2486 complete_rewrite = 1;
2487 }
2488
2489 data = diffstat_add(diffstat, name_a, name_b);
2490 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2491
2492 if (!one || !two) {
2493 data->is_unmerged = 1;
2494 return;
2495 }
2496
2497 same_contents = !hashcmp(one->sha1, two->sha1);
2498
2499 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2500 data->is_binary = 1;
2501 if (same_contents) {
2502 data->added = 0;
2503 data->deleted = 0;
2504 } else {
2505 data->added = diff_filespec_size(two);
2506 data->deleted = diff_filespec_size(one);
2507 }
2508 }
2509
2510 else if (complete_rewrite) {
2511 diff_populate_filespec(one, 0);
2512 diff_populate_filespec(two, 0);
2513 data->deleted = count_lines(one->data, one->size);
2514 data->added = count_lines(two->data, two->size);
2515 }
2516
2517 else if (!same_contents) {
2518 /* Crazy xdl interfaces.. */
2519 xpparam_t xpp;
2520 xdemitconf_t xecfg;
2521
2522 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2523 die("unable to read files to diff");
2524
2525 memset(&xpp, 0, sizeof(xpp));
2526 memset(&xecfg, 0, sizeof(xecfg));
2527 xpp.flags = o->xdl_opts;
2528 xecfg.ctxlen = o->context;
2529 xecfg.interhunkctxlen = o->interhunkcontext;
2530 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2531 &xpp, &xecfg);
2532 }
2533
2534 diff_free_filespec_data(one);
2535 diff_free_filespec_data(two);
2536 }
2537
2538 static void builtin_checkdiff(const char *name_a, const char *name_b,
2539 const char *attr_path,
2540 struct diff_filespec *one,
2541 struct diff_filespec *two,
2542 struct diff_options *o)
2543 {
2544 mmfile_t mf1, mf2;
2545 struct checkdiff_t data;
2546
2547 if (!two)
2548 return;
2549
2550 memset(&data, 0, sizeof(data));
2551 data.filename = name_b ? name_b : name_a;
2552 data.lineno = 0;
2553 data.o = o;
2554 data.ws_rule = whitespace_rule(attr_path);
2555 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2556
2557 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2558 die("unable to read files to diff");
2559
2560 /*
2561 * All the other codepaths check both sides, but not checking
2562 * the "old" side here is deliberate. We are checking the newly
2563 * introduced changes, and as long as the "new" side is text, we
2564 * can and should check what it introduces.
2565 */
2566 if (diff_filespec_is_binary(two))
2567 goto free_and_return;
2568 else {
2569 /* Crazy xdl interfaces.. */
2570 xpparam_t xpp;
2571 xdemitconf_t xecfg;
2572
2573 memset(&xpp, 0, sizeof(xpp));
2574 memset(&xecfg, 0, sizeof(xecfg));
2575 xecfg.ctxlen = 1; /* at least one context line */
2576 xpp.flags = 0;
2577 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2578 &xpp, &xecfg);
2579
2580 if (data.ws_rule & WS_BLANK_AT_EOF) {
2581 struct emit_callback ecbdata;
2582 int blank_at_eof;
2583
2584 ecbdata.ws_rule = data.ws_rule;
2585 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2586 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2587
2588 if (blank_at_eof) {
2589 static char *err;
2590 if (!err)
2591 err = whitespace_error_string(WS_BLANK_AT_EOF);
2592 fprintf(o->file, "%s:%d: %s.\n",
2593 data.filename, blank_at_eof, err);
2594 data.status = 1; /* report errors */
2595 }
2596 }
2597 }
2598 free_and_return:
2599 diff_free_filespec_data(one);
2600 diff_free_filespec_data(two);
2601 if (data.status)
2602 DIFF_OPT_SET(o, CHECK_FAILED);
2603 }
2604
2605 struct diff_filespec *alloc_filespec(const char *path)
2606 {
2607 int namelen = strlen(path);
2608 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2609
2610 memset(spec, 0, sizeof(*spec));
2611 spec->path = (char *)(spec + 1);
2612 memcpy(spec->path, path, namelen+1);
2613 spec->count = 1;
2614 spec->is_binary = -1;
2615 return spec;
2616 }
2617
2618 void free_filespec(struct diff_filespec *spec)
2619 {
2620 if (!--spec->count) {
2621 diff_free_filespec_data(spec);
2622 free(spec);
2623 }
2624 }
2625
2626 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2627 int sha1_valid, unsigned short mode)
2628 {
2629 if (mode) {
2630 spec->mode = canon_mode(mode);
2631 hashcpy(spec->sha1, sha1);
2632 spec->sha1_valid = sha1_valid;
2633 }
2634 }
2635
2636 /*
2637 * Given a name and sha1 pair, if the index tells us the file in
2638 * the work tree has that object contents, return true, so that
2639 * prepare_temp_file() does not have to inflate and extract.
2640 */
2641 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2642 {
2643 const struct cache_entry *ce;
2644 struct stat st;
2645 int pos, len;
2646
2647 /*
2648 * We do not read the cache ourselves here, because the
2649 * benchmark with my previous version that always reads cache
2650 * shows that it makes things worse for diff-tree comparing
2651 * two linux-2.6 kernel trees in an already checked out work
2652 * tree. This is because most diff-tree comparisons deal with
2653 * only a small number of files, while reading the cache is
2654 * expensive for a large project, and its cost outweighs the
2655 * savings we get by not inflating the object to a temporary
2656 * file. Practically, this code only helps when we are used
2657 * by diff-cache --cached, which does read the cache before
2658 * calling us.
2659 */
2660 if (!active_cache)
2661 return 0;
2662
2663 /* We want to avoid the working directory if our caller
2664 * doesn't need the data in a normal file, this system
2665 * is rather slow with its stat/open/mmap/close syscalls,
2666 * and the object is contained in a pack file. The pack
2667 * is probably already open and will be faster to obtain
2668 * the data through than the working directory. Loose
2669 * objects however would tend to be slower as they need
2670 * to be individually opened and inflated.
2671 */
2672 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2673 return 0;
2674
2675 len = strlen(name);
2676 pos = cache_name_pos(name, len);
2677 if (pos < 0)
2678 return 0;
2679 ce = active_cache[pos];
2680
2681 /*
2682 * This is not the sha1 we are looking for, or
2683 * unreusable because it is not a regular file.
2684 */
2685 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2686 return 0;
2687
2688 /*
2689 * If ce is marked as "assume unchanged", there is no
2690 * guarantee that work tree matches what we are looking for.
2691 */
2692 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2693 return 0;
2694
2695 /*
2696 * If ce matches the file in the work tree, we can reuse it.
2697 */
2698 if (ce_uptodate(ce) ||
2699 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2700 return 1;
2701
2702 return 0;
2703 }
2704
2705 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2706 {
2707 int len;
2708 char *data = xmalloc(100), *dirty = "";
2709
2710 /* Are we looking at the work tree? */
2711 if (s->dirty_submodule)
2712 dirty = "-dirty";
2713
2714 len = snprintf(data, 100,
2715 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2716 s->data = data;
2717 s->size = len;
2718 s->should_free = 1;
2719 if (size_only) {
2720 s->data = NULL;
2721 free(data);
2722 }
2723 return 0;
2724 }
2725
2726 /*
2727 * While doing rename detection and pickaxe operation, we may need to
2728 * grab the data for the blob (or file) for our own in-core comparison.
2729 * diff_filespec has data and size fields for this purpose.
2730 */
2731 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
2732 {
2733 int size_only = flags & CHECK_SIZE_ONLY;
2734 int err = 0;
2735 /*
2736 * demote FAIL to WARN to allow inspecting the situation
2737 * instead of refusing.
2738 */
2739 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
2740 ? SAFE_CRLF_WARN
2741 : safe_crlf);
2742
2743 if (!DIFF_FILE_VALID(s))
2744 die("internal error: asking to populate invalid file.");
2745 if (S_ISDIR(s->mode))
2746 return -1;
2747
2748 if (s->data)
2749 return 0;
2750
2751 if (size_only && 0 < s->size)
2752 return 0;
2753
2754 if (S_ISGITLINK(s->mode))
2755 return diff_populate_gitlink(s, size_only);
2756
2757 if (!s->sha1_valid ||
2758 reuse_worktree_file(s->path, s->sha1, 0)) {
2759 struct strbuf buf = STRBUF_INIT;
2760 struct stat st;
2761 int fd;
2762
2763 if (lstat(s->path, &st) < 0) {
2764 if (errno == ENOENT) {
2765 err_empty:
2766 err = -1;
2767 empty:
2768 s->data = (char *)"";
2769 s->size = 0;
2770 return err;
2771 }
2772 }
2773 s->size = xsize_t(st.st_size);
2774 if (!s->size)
2775 goto empty;
2776 if (S_ISLNK(st.st_mode)) {
2777 struct strbuf sb = STRBUF_INIT;
2778
2779 if (strbuf_readlink(&sb, s->path, s->size))
2780 goto err_empty;
2781 s->size = sb.len;
2782 s->data = strbuf_detach(&sb, NULL);
2783 s->should_free = 1;
2784 return 0;
2785 }
2786 if (size_only)
2787 return 0;
2788 if ((flags & CHECK_BINARY) &&
2789 s->size > big_file_threshold && s->is_binary == -1) {
2790 s->is_binary = 1;
2791 return 0;
2792 }
2793 fd = open(s->path, O_RDONLY);
2794 if (fd < 0)
2795 goto err_empty;
2796 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2797 close(fd);
2798 s->should_munmap = 1;
2799
2800 /*
2801 * Convert from working tree format to canonical git format
2802 */
2803 if (convert_to_git(s->path, s->data, s->size, &buf, crlf_warn)) {
2804 size_t size = 0;
2805 munmap(s->data, s->size);
2806 s->should_munmap = 0;
2807 s->data = strbuf_detach(&buf, &size);
2808 s->size = size;
2809 s->should_free = 1;
2810 }
2811 }
2812 else {
2813 enum object_type type;
2814 if (size_only || (flags & CHECK_BINARY)) {
2815 type = sha1_object_info(s->sha1, &s->size);
2816 if (type < 0)
2817 die("unable to read %s", sha1_to_hex(s->sha1));
2818 if (size_only)
2819 return 0;
2820 if (s->size > big_file_threshold && s->is_binary == -1) {
2821 s->is_binary = 1;
2822 return 0;
2823 }
2824 }
2825 s->data = read_sha1_file(s->sha1, &type, &s->size);
2826 if (!s->data)
2827 die("unable to read %s", sha1_to_hex(s->sha1));
2828 s->should_free = 1;
2829 }
2830 return 0;
2831 }
2832
2833 void diff_free_filespec_blob(struct diff_filespec *s)
2834 {
2835 if (s->should_free)
2836 free(s->data);
2837 else if (s->should_munmap)
2838 munmap(s->data, s->size);
2839
2840 if (s->should_free || s->should_munmap) {
2841 s->should_free = s->should_munmap = 0;
2842 s->data = NULL;
2843 }
2844 }
2845
2846 void diff_free_filespec_data(struct diff_filespec *s)
2847 {
2848 diff_free_filespec_blob(s);
2849 free(s->cnt_data);
2850 s->cnt_data = NULL;
2851 }
2852
2853 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2854 void *blob,
2855 unsigned long size,
2856 const unsigned char *sha1,
2857 int mode)
2858 {
2859 int fd;
2860 struct strbuf buf = STRBUF_INIT;
2861 struct strbuf template = STRBUF_INIT;
2862 char *path_dup = xstrdup(path);
2863 const char *base = basename(path_dup);
2864
2865 /* Generate "XXXXXX_basename.ext" */
2866 strbuf_addstr(&template, "XXXXXX_");
2867 strbuf_addstr(&template, base);
2868
2869 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
2870 if (fd < 0)
2871 die_errno("unable to create temp-file");
2872 if (convert_to_working_tree(path,
2873 (const char *)blob, (size_t)size, &buf)) {
2874 blob = buf.buf;
2875 size = buf.len;
2876 }
2877 if (write_in_full(fd, blob, size) != size)
2878 die_errno("unable to write temp-file");
2879 close_tempfile(&temp->tempfile);
2880 temp->name = get_tempfile_path(&temp->tempfile);
2881 strcpy(temp->hex, sha1_to_hex(sha1));
2882 temp->hex[40] = 0;
2883 sprintf(temp->mode, "%06o", mode);
2884 strbuf_release(&buf);
2885 strbuf_release(&template);
2886 free(path_dup);
2887 }
2888
2889 static struct diff_tempfile *prepare_temp_file(const char *name,
2890 struct diff_filespec *one)
2891 {
2892 struct diff_tempfile *temp = claim_diff_tempfile();
2893
2894 if (!DIFF_FILE_VALID(one)) {
2895 not_a_valid_file:
2896 /* A '-' entry produces this for file-2, and
2897 * a '+' entry produces this for file-1.
2898 */
2899 temp->name = "/dev/null";
2900 strcpy(temp->hex, ".");
2901 strcpy(temp->mode, ".");
2902 return temp;
2903 }
2904
2905 if (!S_ISGITLINK(one->mode) &&
2906 (!one->sha1_valid ||
2907 reuse_worktree_file(name, one->sha1, 1))) {
2908 struct stat st;
2909 if (lstat(name, &st) < 0) {
2910 if (errno == ENOENT)
2911 goto not_a_valid_file;
2912 die_errno("stat(%s)", name);
2913 }
2914 if (S_ISLNK(st.st_mode)) {
2915 struct strbuf sb = STRBUF_INIT;
2916 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2917 die_errno("readlink(%s)", name);
2918 prep_temp_blob(name, temp, sb.buf, sb.len,
2919 (one->sha1_valid ?
2920 one->sha1 : null_sha1),
2921 (one->sha1_valid ?
2922 one->mode : S_IFLNK));
2923 strbuf_release(&sb);
2924 }
2925 else {
2926 /* we can borrow from the file in the work tree */
2927 temp->name = name;
2928 if (!one->sha1_valid)
2929 strcpy(temp->hex, sha1_to_hex(null_sha1));
2930 else
2931 strcpy(temp->hex, sha1_to_hex(one->sha1));
2932 /* Even though we may sometimes borrow the
2933 * contents from the work tree, we always want
2934 * one->mode. mode is trustworthy even when
2935 * !(one->sha1_valid), as long as
2936 * DIFF_FILE_VALID(one).
2937 */
2938 sprintf(temp->mode, "%06o", one->mode);
2939 }
2940 return temp;
2941 }
2942 else {
2943 if (diff_populate_filespec(one, 0))
2944 die("cannot read data blob for %s", one->path);
2945 prep_temp_blob(name, temp, one->data, one->size,
2946 one->sha1, one->mode);
2947 }
2948 return temp;
2949 }
2950
2951 static void add_external_diff_name(struct argv_array *argv,
2952 const char *name,
2953 struct diff_filespec *df)
2954 {
2955 struct diff_tempfile *temp = prepare_temp_file(name, df);
2956 argv_array_push(argv, temp->name);
2957 argv_array_push(argv, temp->hex);
2958 argv_array_push(argv, temp->mode);
2959 }
2960
2961 /* An external diff command takes:
2962 *
2963 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2964 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2965 *
2966 */
2967 static void run_external_diff(const char *pgm,
2968 const char *name,
2969 const char *other,
2970 struct diff_filespec *one,
2971 struct diff_filespec *two,
2972 const char *xfrm_msg,
2973 int complete_rewrite,
2974 struct diff_options *o)
2975 {
2976 struct argv_array argv = ARGV_ARRAY_INIT;
2977 struct argv_array env = ARGV_ARRAY_INIT;
2978 struct diff_queue_struct *q = &diff_queued_diff;
2979
2980 argv_array_push(&argv, pgm);
2981 argv_array_push(&argv, name);
2982
2983 if (one && two) {
2984 add_external_diff_name(&argv, name, one);
2985 if (!other)
2986 add_external_diff_name(&argv, name, two);
2987 else {
2988 add_external_diff_name(&argv, other, two);
2989 argv_array_push(&argv, other);
2990 argv_array_push(&argv, xfrm_msg);
2991 }
2992 }
2993
2994 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
2995 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
2996
2997 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
2998 die(_("external diff died, stopping at %s"), name);
2999
3000 remove_tempfile();
3001 argv_array_clear(&argv);
3002 argv_array_clear(&env);
3003 }
3004
3005 static int similarity_index(struct diff_filepair *p)
3006 {
3007 return p->score * 100 / MAX_SCORE;
3008 }
3009
3010 static void fill_metainfo(struct strbuf *msg,
3011 const char *name,
3012 const char *other,
3013 struct diff_filespec *one,
3014 struct diff_filespec *two,
3015 struct diff_options *o,
3016 struct diff_filepair *p,
3017 int *must_show_header,
3018 int use_color)
3019 {
3020 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3021 const char *reset = diff_get_color(use_color, DIFF_RESET);
3022 const char *line_prefix = diff_line_prefix(o);
3023
3024 *must_show_header = 1;
3025 strbuf_init(msg, PATH_MAX * 2 + 300);
3026 switch (p->status) {
3027 case DIFF_STATUS_COPIED:
3028 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3029 line_prefix, set, similarity_index(p));
3030 strbuf_addf(msg, "%s\n%s%scopy from ",
3031 reset, line_prefix, set);
3032 quote_c_style(name, msg, NULL, 0);
3033 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3034 quote_c_style(other, msg, NULL, 0);
3035 strbuf_addf(msg, "%s\n", reset);
3036 break;
3037 case DIFF_STATUS_RENAMED:
3038 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3039 line_prefix, set, similarity_index(p));
3040 strbuf_addf(msg, "%s\n%s%srename from ",
3041 reset, line_prefix, set);
3042 quote_c_style(name, msg, NULL, 0);
3043 strbuf_addf(msg, "%s\n%s%srename to ",
3044 reset, line_prefix, set);
3045 quote_c_style(other, msg, NULL, 0);
3046 strbuf_addf(msg, "%s\n", reset);
3047 break;
3048 case DIFF_STATUS_MODIFIED:
3049 if (p->score) {
3050 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3051 line_prefix,
3052 set, similarity_index(p), reset);
3053 break;
3054 }
3055 /* fallthru */
3056 default:
3057 *must_show_header = 0;
3058 }
3059 if (one && two && hashcmp(one->sha1, two->sha1)) {
3060 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3061
3062 if (DIFF_OPT_TST(o, BINARY)) {
3063 mmfile_t mf;
3064 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3065 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3066 abbrev = 40;
3067 }
3068 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
3069 find_unique_abbrev(one->sha1, abbrev));
3070 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
3071 if (one->mode == two->mode)
3072 strbuf_addf(msg, " %06o", one->mode);
3073 strbuf_addf(msg, "%s\n", reset);
3074 }
3075 }
3076
3077 static void run_diff_cmd(const char *pgm,
3078 const char *name,
3079 const char *other,
3080 const char *attr_path,
3081 struct diff_filespec *one,
3082 struct diff_filespec *two,
3083 struct strbuf *msg,
3084 struct diff_options *o,
3085 struct diff_filepair *p)
3086 {
3087 const char *xfrm_msg = NULL;
3088 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3089 int must_show_header = 0;
3090
3091
3092 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3093 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3094 if (drv && drv->external)
3095 pgm = drv->external;
3096 }
3097
3098 if (msg) {
3099 /*
3100 * don't use colors when the header is intended for an
3101 * external diff driver
3102 */
3103 fill_metainfo(msg, name, other, one, two, o, p,
3104 &must_show_header,
3105 want_color(o->use_color) && !pgm);
3106 xfrm_msg = msg->len ? msg->buf : NULL;
3107 }
3108
3109 if (pgm) {
3110 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3111 complete_rewrite, o);
3112 return;
3113 }
3114 if (one && two)
3115 builtin_diff(name, other ? other : name,
3116 one, two, xfrm_msg, must_show_header,
3117 o, complete_rewrite);
3118 else
3119 fprintf(o->file, "* Unmerged path %s\n", name);
3120 }
3121
3122 static void diff_fill_sha1_info(struct diff_filespec *one)
3123 {
3124 if (DIFF_FILE_VALID(one)) {
3125 if (!one->sha1_valid) {
3126 struct stat st;
3127 if (one->is_stdin) {
3128 hashcpy(one->sha1, null_sha1);
3129 return;
3130 }
3131 if (lstat(one->path, &st) < 0)
3132 die_errno("stat '%s'", one->path);
3133 if (index_path(one->sha1, one->path, &st, 0))
3134 die("cannot hash %s", one->path);
3135 }
3136 }
3137 else
3138 hashclr(one->sha1);
3139 }
3140
3141 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3142 {
3143 /* Strip the prefix but do not molest /dev/null and absolute paths */
3144 if (*namep && **namep != '/') {
3145 *namep += prefix_length;
3146 if (**namep == '/')
3147 ++*namep;
3148 }
3149 if (*otherp && **otherp != '/') {
3150 *otherp += prefix_length;
3151 if (**otherp == '/')
3152 ++*otherp;
3153 }
3154 }
3155
3156 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3157 {
3158 const char *pgm = external_diff();
3159 struct strbuf msg;
3160 struct diff_filespec *one = p->one;
3161 struct diff_filespec *two = p->two;
3162 const char *name;
3163 const char *other;
3164 const char *attr_path;
3165
3166 name = p->one->path;
3167 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3168 attr_path = name;
3169 if (o->prefix_length)
3170 strip_prefix(o->prefix_length, &name, &other);
3171
3172 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3173 pgm = NULL;
3174
3175 if (DIFF_PAIR_UNMERGED(p)) {
3176 run_diff_cmd(pgm, name, NULL, attr_path,
3177 NULL, NULL, NULL, o, p);
3178 return;
3179 }
3180
3181 diff_fill_sha1_info(one);
3182 diff_fill_sha1_info(two);
3183
3184 if (!pgm &&
3185 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3186 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3187 /*
3188 * a filepair that changes between file and symlink
3189 * needs to be split into deletion and creation.
3190 */
3191 struct diff_filespec *null = alloc_filespec(two->path);
3192 run_diff_cmd(NULL, name, other, attr_path,
3193 one, null, &msg, o, p);
3194 free(null);
3195 strbuf_release(&msg);
3196
3197 null = alloc_filespec(one->path);
3198 run_diff_cmd(NULL, name, other, attr_path,
3199 null, two, &msg, o, p);
3200 free(null);
3201 }
3202 else
3203 run_diff_cmd(pgm, name, other, attr_path,
3204 one, two, &msg, o, p);
3205
3206 strbuf_release(&msg);
3207 }
3208
3209 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3210 struct diffstat_t *diffstat)
3211 {
3212 const char *name;
3213 const char *other;
3214
3215 if (DIFF_PAIR_UNMERGED(p)) {
3216 /* unmerged */
3217 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3218 return;
3219 }
3220
3221 name = p->one->path;
3222 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3223
3224 if (o->prefix_length)
3225 strip_prefix(o->prefix_length, &name, &other);
3226
3227 diff_fill_sha1_info(p->one);
3228 diff_fill_sha1_info(p->two);
3229
3230 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3231 }
3232
3233 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3234 {
3235 const char *name;
3236 const char *other;
3237 const char *attr_path;
3238
3239 if (DIFF_PAIR_UNMERGED(p)) {
3240 /* unmerged */
3241 return;
3242 }
3243
3244 name = p->one->path;
3245 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3246 attr_path = other ? other : name;
3247
3248 if (o->prefix_length)
3249 strip_prefix(o->prefix_length, &name, &other);
3250
3251 diff_fill_sha1_info(p->one);
3252 diff_fill_sha1_info(p->two);
3253
3254 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3255 }
3256
3257 void diff_setup(struct diff_options *options)
3258 {
3259 memcpy(options, &default_diff_options, sizeof(*options));
3260
3261 options->file = stdout;
3262
3263 options->line_termination = '\n';
3264 options->break_opt = -1;
3265 options->rename_limit = -1;
3266 options->dirstat_permille = diff_dirstat_permille_default;
3267 options->context = diff_context_default;
3268 options->ws_error_highlight = WSEH_NEW;
3269 DIFF_OPT_SET(options, RENAME_EMPTY);
3270
3271 /* pathchange left =NULL by default */
3272 options->change = diff_change;
3273 options->add_remove = diff_addremove;
3274 options->use_color = diff_use_color_default;
3275 options->detect_rename = diff_detect_rename_default;
3276 options->xdl_opts |= diff_algorithm;
3277
3278 options->orderfile = diff_order_file_cfg;
3279
3280 if (diff_no_prefix) {
3281 options->a_prefix = options->b_prefix = "";
3282 } else if (!diff_mnemonic_prefix) {
3283 options->a_prefix = "a/";
3284 options->b_prefix = "b/";
3285 }
3286 }
3287
3288 void diff_setup_done(struct diff_options *options)
3289 {
3290 int count = 0;
3291
3292 if (options->set_default)
3293 options->set_default(options);
3294
3295 if (options->output_format & DIFF_FORMAT_NAME)
3296 count++;
3297 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3298 count++;
3299 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3300 count++;
3301 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3302 count++;
3303 if (count > 1)
3304 die("--name-only, --name-status, --check and -s are mutually exclusive");
3305
3306 /*
3307 * Most of the time we can say "there are changes"
3308 * only by checking if there are changed paths, but
3309 * --ignore-whitespace* options force us to look
3310 * inside contents.
3311 */
3312
3313 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3314 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3315 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3316 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3317 else
3318 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3319
3320 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3321 options->detect_rename = DIFF_DETECT_COPY;
3322
3323 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3324 options->prefix = NULL;
3325 if (options->prefix)
3326 options->prefix_length = strlen(options->prefix);
3327 else
3328 options->prefix_length = 0;
3329
3330 if (options->output_format & (DIFF_FORMAT_NAME |
3331 DIFF_FORMAT_NAME_STATUS |
3332 DIFF_FORMAT_CHECKDIFF |
3333 DIFF_FORMAT_NO_OUTPUT))
3334 options->output_format &= ~(DIFF_FORMAT_RAW |
3335 DIFF_FORMAT_NUMSTAT |
3336 DIFF_FORMAT_DIFFSTAT |
3337 DIFF_FORMAT_SHORTSTAT |
3338 DIFF_FORMAT_DIRSTAT |
3339 DIFF_FORMAT_SUMMARY |
3340 DIFF_FORMAT_PATCH);
3341
3342 /*
3343 * These cases always need recursive; we do not drop caller-supplied
3344 * recursive bits for other formats here.
3345 */
3346 if (options->output_format & (DIFF_FORMAT_PATCH |
3347 DIFF_FORMAT_NUMSTAT |
3348 DIFF_FORMAT_DIFFSTAT |
3349 DIFF_FORMAT_SHORTSTAT |
3350 DIFF_FORMAT_DIRSTAT |
3351 DIFF_FORMAT_SUMMARY |
3352 DIFF_FORMAT_CHECKDIFF))
3353 DIFF_OPT_SET(options, RECURSIVE);
3354 /*
3355 * Also pickaxe would not work very well if you do not say recursive
3356 */
3357 if (options->pickaxe)
3358 DIFF_OPT_SET(options, RECURSIVE);
3359 /*
3360 * When patches are generated, submodules diffed against the work tree
3361 * must be checked for dirtiness too so it can be shown in the output
3362 */
3363 if (options->output_format & DIFF_FORMAT_PATCH)
3364 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3365
3366 if (options->detect_rename && options->rename_limit < 0)
3367 options->rename_limit = diff_rename_limit_default;
3368 if (options->setup & DIFF_SETUP_USE_CACHE) {
3369 if (!active_cache)
3370 /* read-cache does not die even when it fails
3371 * so it is safe for us to do this here. Also
3372 * it does not smudge active_cache or active_nr
3373 * when it fails, so we do not have to worry about
3374 * cleaning it up ourselves either.
3375 */
3376 read_cache();
3377 }
3378 if (options->abbrev <= 0 || 40 < options->abbrev)
3379 options->abbrev = 40; /* full */
3380
3381 /*
3382 * It does not make sense to show the first hit we happened
3383 * to have found. It does not make sense not to return with
3384 * exit code in such a case either.
3385 */
3386 if (DIFF_OPT_TST(options, QUICK)) {
3387 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3388 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3389 }
3390
3391 options->diff_path_counter = 0;
3392
3393 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3394 die(_("--follow requires exactly one pathspec"));
3395 }
3396
3397 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3398 {
3399 char c, *eq;
3400 int len;
3401
3402 if (*arg != '-')
3403 return 0;
3404 c = *++arg;
3405 if (!c)
3406 return 0;
3407 if (c == arg_short) {
3408 c = *++arg;
3409 if (!c)
3410 return 1;
3411 if (val && isdigit(c)) {
3412 char *end;
3413 int n = strtoul(arg, &end, 10);
3414 if (*end)
3415 return 0;
3416 *val = n;
3417 return 1;
3418 }
3419 return 0;
3420 }
3421 if (c != '-')
3422 return 0;
3423 arg++;
3424 eq = strchrnul(arg, '=');
3425 len = eq - arg;
3426 if (!len || strncmp(arg, arg_long, len))
3427 return 0;
3428 if (*eq) {
3429 int n;
3430 char *end;
3431 if (!isdigit(*++eq))
3432 return 0;
3433 n = strtoul(eq, &end, 10);
3434 if (*end)
3435 return 0;
3436 *val = n;
3437 }
3438 return 1;
3439 }
3440
3441 static int diff_scoreopt_parse(const char *opt);
3442
3443 static inline int short_opt(char opt, const char **argv,
3444 const char **optarg)
3445 {
3446 const char *arg = argv[0];
3447 if (arg[0] != '-' || arg[1] != opt)
3448 return 0;
3449 if (arg[2] != '\0') {
3450 *optarg = arg + 2;
3451 return 1;
3452 }
3453 if (!argv[1])
3454 die("Option '%c' requires a value", opt);
3455 *optarg = argv[1];
3456 return 2;
3457 }
3458
3459 int parse_long_opt(const char *opt, const char **argv,
3460 const char **optarg)
3461 {
3462 const char *arg = argv[0];
3463 if (!skip_prefix(arg, "--", &arg))
3464 return 0;
3465 if (!skip_prefix(arg, opt, &arg))
3466 return 0;
3467 if (*arg == '=') { /* stuck form: --option=value */
3468 *optarg = arg + 1;
3469 return 1;
3470 }
3471 if (*arg != '\0')
3472 return 0;
3473 /* separate form: --option value */
3474 if (!argv[1])
3475 die("Option '--%s' requires a value", opt);
3476 *optarg = argv[1];
3477 return 2;
3478 }
3479
3480 static int stat_opt(struct diff_options *options, const char **av)
3481 {
3482 const char *arg = av[0];
3483 char *end;
3484 int width = options->stat_width;
3485 int name_width = options->stat_name_width;
3486 int graph_width = options->stat_graph_width;
3487 int count = options->stat_count;
3488 int argcount = 1;
3489
3490 if (!skip_prefix(arg, "--stat", &arg))
3491 die("BUG: stat option does not begin with --stat: %s", arg);
3492 end = (char *)arg;
3493
3494 switch (*arg) {
3495 case '-':
3496 if (skip_prefix(arg, "-width", &arg)) {
3497 if (*arg == '=')
3498 width = strtoul(arg + 1, &end, 10);
3499 else if (!*arg && !av[1])
3500 die("Option '--stat-width' requires a value");
3501 else if (!*arg) {
3502 width = strtoul(av[1], &end, 10);
3503 argcount = 2;
3504 }
3505 } else if (skip_prefix(arg, "-name-width", &arg)) {
3506 if (*arg == '=')
3507 name_width = strtoul(arg + 1, &end, 10);
3508 else if (!*arg && !av[1])
3509 die("Option '--stat-name-width' requires a value");
3510 else if (!*arg) {
3511 name_width = strtoul(av[1], &end, 10);
3512 argcount = 2;
3513 }
3514 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3515 if (*arg == '=')
3516 graph_width = strtoul(arg + 1, &end, 10);
3517 else if (!*arg && !av[1])
3518 die("Option '--stat-graph-width' requires a value");
3519 else if (!*arg) {
3520 graph_width = strtoul(av[1], &end, 10);
3521 argcount = 2;
3522 }
3523 } else if (skip_prefix(arg, "-count", &arg)) {
3524 if (*arg == '=')
3525 count = strtoul(arg + 1, &end, 10);
3526 else if (!*arg && !av[1])
3527 die("Option '--stat-count' requires a value");
3528 else if (!*arg) {
3529 count = strtoul(av[1], &end, 10);
3530 argcount = 2;
3531 }
3532 }
3533 break;
3534 case '=':
3535 width = strtoul(arg+1, &end, 10);
3536 if (*end == ',')
3537 name_width = strtoul(end+1, &end, 10);
3538 if (*end == ',')
3539 count = strtoul(end+1, &end, 10);
3540 }
3541
3542 /* Important! This checks all the error cases! */
3543 if (*end)
3544 return 0;
3545 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3546 options->stat_name_width = name_width;
3547 options->stat_graph_width = graph_width;
3548 options->stat_width = width;
3549 options->stat_count = count;
3550 return argcount;
3551 }
3552
3553 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3554 {
3555 struct strbuf errmsg = STRBUF_INIT;
3556 if (parse_dirstat_params(options, params, &errmsg))
3557 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3558 errmsg.buf);
3559 strbuf_release(&errmsg);
3560 /*
3561 * The caller knows a dirstat-related option is given from the command
3562 * line; allow it to say "return this_function();"
3563 */
3564 options->output_format |= DIFF_FORMAT_DIRSTAT;
3565 return 1;
3566 }
3567
3568 static int parse_submodule_opt(struct diff_options *options, const char *value)
3569 {
3570 if (parse_submodule_params(options, value))
3571 die(_("Failed to parse --submodule option parameter: '%s'"),
3572 value);
3573 return 1;
3574 }
3575
3576 static const char diff_status_letters[] = {
3577 DIFF_STATUS_ADDED,
3578 DIFF_STATUS_COPIED,
3579 DIFF_STATUS_DELETED,
3580 DIFF_STATUS_MODIFIED,
3581 DIFF_STATUS_RENAMED,
3582 DIFF_STATUS_TYPE_CHANGED,
3583 DIFF_STATUS_UNKNOWN,
3584 DIFF_STATUS_UNMERGED,
3585 DIFF_STATUS_FILTER_AON,
3586 DIFF_STATUS_FILTER_BROKEN,
3587 '\0',
3588 };
3589
3590 static unsigned int filter_bit['Z' + 1];
3591
3592 static void prepare_filter_bits(void)
3593 {
3594 int i;
3595
3596 if (!filter_bit[DIFF_STATUS_ADDED]) {
3597 for (i = 0; diff_status_letters[i]; i++)
3598 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3599 }
3600 }
3601
3602 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3603 {
3604 return opt->filter & filter_bit[(int) status];
3605 }
3606
3607 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3608 {
3609 int i, optch;
3610
3611 prepare_filter_bits();
3612
3613 /*
3614 * If there is a negation e.g. 'd' in the input, and we haven't
3615 * initialized the filter field with another --diff-filter, start
3616 * from full set of bits, except for AON.
3617 */
3618 if (!opt->filter) {
3619 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3620 if (optch < 'a' || 'z' < optch)
3621 continue;
3622 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3623 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3624 break;
3625 }
3626 }
3627
3628 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3629 unsigned int bit;
3630 int negate;
3631
3632 if ('a' <= optch && optch <= 'z') {
3633 negate = 1;
3634 optch = toupper(optch);
3635 } else {
3636 negate = 0;
3637 }
3638
3639 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
3640 if (!bit)
3641 return optarg[i];
3642 if (negate)
3643 opt->filter &= ~bit;
3644 else
3645 opt->filter |= bit;
3646 }
3647 return 0;
3648 }
3649
3650 static void enable_patch_output(int *fmt) {
3651 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
3652 *fmt |= DIFF_FORMAT_PATCH;
3653 }
3654
3655 static int parse_one_token(const char **arg, const char *token)
3656 {
3657 const char *rest;
3658 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
3659 *arg = rest;
3660 return 1;
3661 }
3662 return 0;
3663 }
3664
3665 static int parse_ws_error_highlight(struct diff_options *opt, const char *arg)
3666 {
3667 const char *orig_arg = arg;
3668 unsigned val = 0;
3669 while (*arg) {
3670 if (parse_one_token(&arg, "none"))
3671 val = 0;
3672 else if (parse_one_token(&arg, "default"))
3673 val = WSEH_NEW;
3674 else if (parse_one_token(&arg, "all"))
3675 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
3676 else if (parse_one_token(&arg, "new"))
3677 val |= WSEH_NEW;
3678 else if (parse_one_token(&arg, "old"))
3679 val |= WSEH_OLD;
3680 else if (parse_one_token(&arg, "context"))
3681 val |= WSEH_CONTEXT;
3682 else {
3683 error("unknown value after ws-error-highlight=%.*s",
3684 (int)(arg - orig_arg), orig_arg);
3685 return 0;
3686 }
3687 if (*arg)
3688 arg++;
3689 }
3690 opt->ws_error_highlight = val;
3691 return 1;
3692 }
3693
3694 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3695 {
3696 const char *arg = av[0];
3697 const char *optarg;
3698 int argcount;
3699
3700 /* Output format options */
3701 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
3702 || opt_arg(arg, 'U', "unified", &options->context))
3703 enable_patch_output(&options->output_format);
3704 else if (!strcmp(arg, "--raw"))
3705 options->output_format |= DIFF_FORMAT_RAW;
3706 else if (!strcmp(arg, "--patch-with-raw")) {
3707 enable_patch_output(&options->output_format);
3708 options->output_format |= DIFF_FORMAT_RAW;
3709 } else if (!strcmp(arg, "--numstat"))
3710 options->output_format |= DIFF_FORMAT_NUMSTAT;
3711 else if (!strcmp(arg, "--shortstat"))
3712 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3713 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3714 return parse_dirstat_opt(options, "");
3715 else if (skip_prefix(arg, "-X", &arg))
3716 return parse_dirstat_opt(options, arg);
3717 else if (skip_prefix(arg, "--dirstat=", &arg))
3718 return parse_dirstat_opt(options, arg);
3719 else if (!strcmp(arg, "--cumulative"))
3720 return parse_dirstat_opt(options, "cumulative");
3721 else if (!strcmp(arg, "--dirstat-by-file"))
3722 return parse_dirstat_opt(options, "files");
3723 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
3724 parse_dirstat_opt(options, "files");
3725 return parse_dirstat_opt(options, arg);
3726 }
3727 else if (!strcmp(arg, "--check"))
3728 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3729 else if (!strcmp(arg, "--summary"))
3730 options->output_format |= DIFF_FORMAT_SUMMARY;
3731 else if (!strcmp(arg, "--patch-with-stat")) {
3732 enable_patch_output(&options->output_format);
3733 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3734 } else if (!strcmp(arg, "--name-only"))
3735 options->output_format |= DIFF_FORMAT_NAME;
3736 else if (!strcmp(arg, "--name-status"))
3737 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3738 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
3739 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3740 else if (starts_with(arg, "--stat"))
3741 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3742 return stat_opt(options, av);
3743
3744 /* renames options */
3745 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
3746 !strcmp(arg, "--break-rewrites")) {
3747 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3748 return error("invalid argument to -B: %s", arg+2);
3749 }
3750 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
3751 !strcmp(arg, "--find-renames")) {
3752 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3753 return error("invalid argument to -M: %s", arg+2);
3754 options->detect_rename = DIFF_DETECT_RENAME;
3755 }
3756 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3757 options->irreversible_delete = 1;
3758 }
3759 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
3760 !strcmp(arg, "--find-copies")) {
3761 if (options->detect_rename == DIFF_DETECT_COPY)
3762 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3763 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3764 return error("invalid argument to -C: %s", arg+2);
3765 options->detect_rename = DIFF_DETECT_COPY;
3766 }
3767 else if (!strcmp(arg, "--no-renames"))
3768 options->detect_rename = 0;
3769 else if (!strcmp(arg, "--rename-empty"))
3770 DIFF_OPT_SET(options, RENAME_EMPTY);
3771 else if (!strcmp(arg, "--no-rename-empty"))
3772 DIFF_OPT_CLR(options, RENAME_EMPTY);
3773 else if (!strcmp(arg, "--relative"))
3774 DIFF_OPT_SET(options, RELATIVE_NAME);
3775 else if (skip_prefix(arg, "--relative=", &arg)) {
3776 DIFF_OPT_SET(options, RELATIVE_NAME);
3777 options->prefix = arg;
3778 }
3779
3780 /* xdiff options */
3781 else if (!strcmp(arg, "--minimal"))
3782 DIFF_XDL_SET(options, NEED_MINIMAL);
3783 else if (!strcmp(arg, "--no-minimal"))
3784 DIFF_XDL_CLR(options, NEED_MINIMAL);
3785 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3786 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3787 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3788 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3789 else if (!strcmp(arg, "--ignore-space-at-eol"))
3790 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3791 else if (!strcmp(arg, "--ignore-blank-lines"))
3792 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
3793 else if (!strcmp(arg, "--patience"))
3794 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3795 else if (!strcmp(arg, "--histogram"))
3796 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3797 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
3798 long value = parse_algorithm_value(optarg);
3799 if (value < 0)
3800 return error("option diff-algorithm accepts \"myers\", "
3801 "\"minimal\", \"patience\" and \"histogram\"");
3802 /* clear out previous settings */
3803 DIFF_XDL_CLR(options, NEED_MINIMAL);
3804 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3805 options->xdl_opts |= value;
3806 return argcount;
3807 }
3808
3809 /* flags options */
3810 else if (!strcmp(arg, "--binary")) {
3811 enable_patch_output(&options->output_format);
3812 DIFF_OPT_SET(options, BINARY);
3813 }
3814 else if (!strcmp(arg, "--full-index"))
3815 DIFF_OPT_SET(options, FULL_INDEX);
3816 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3817 DIFF_OPT_SET(options, TEXT);
3818 else if (!strcmp(arg, "-R"))
3819 DIFF_OPT_SET(options, REVERSE_DIFF);
3820 else if (!strcmp(arg, "--find-copies-harder"))
3821 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3822 else if (!strcmp(arg, "--follow"))
3823 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3824 else if (!strcmp(arg, "--no-follow")) {
3825 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3826 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
3827 } else if (!strcmp(arg, "--color"))
3828 options->use_color = 1;
3829 else if (skip_prefix(arg, "--color=", &arg)) {
3830 int value = git_config_colorbool(NULL, arg);
3831 if (value < 0)
3832 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3833 options->use_color = value;
3834 }
3835 else if (!strcmp(arg, "--no-color"))
3836 options->use_color = 0;
3837 else if (!strcmp(arg, "--color-words")) {
3838 options->use_color = 1;
3839 options->word_diff = DIFF_WORDS_COLOR;
3840 }
3841 else if (skip_prefix(arg, "--color-words=", &arg)) {
3842 options->use_color = 1;
3843 options->word_diff = DIFF_WORDS_COLOR;
3844 options->word_regex = arg;
3845 }
3846 else if (!strcmp(arg, "--word-diff")) {
3847 if (options->word_diff == DIFF_WORDS_NONE)
3848 options->word_diff = DIFF_WORDS_PLAIN;
3849 }
3850 else if (skip_prefix(arg, "--word-diff=", &arg)) {
3851 if (!strcmp(arg, "plain"))
3852 options->word_diff = DIFF_WORDS_PLAIN;
3853 else if (!strcmp(arg, "color")) {
3854 options->use_color = 1;
3855 options->word_diff = DIFF_WORDS_COLOR;
3856 }
3857 else if (!strcmp(arg, "porcelain"))
3858 options->word_diff = DIFF_WORDS_PORCELAIN;
3859 else if (!strcmp(arg, "none"))
3860 options->word_diff = DIFF_WORDS_NONE;
3861 else
3862 die("bad --word-diff argument: %s", arg);
3863 }
3864 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3865 if (options->word_diff == DIFF_WORDS_NONE)
3866 options->word_diff = DIFF_WORDS_PLAIN;
3867 options->word_regex = optarg;
3868 return argcount;
3869 }
3870 else if (!strcmp(arg, "--exit-code"))
3871 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3872 else if (!strcmp(arg, "--quiet"))
3873 DIFF_OPT_SET(options, QUICK);
3874 else if (!strcmp(arg, "--ext-diff"))
3875 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3876 else if (!strcmp(arg, "--no-ext-diff"))
3877 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3878 else if (!strcmp(arg, "--textconv"))
3879 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3880 else if (!strcmp(arg, "--no-textconv"))
3881 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3882 else if (!strcmp(arg, "--ignore-submodules")) {
3883 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3884 handle_ignore_submodules_arg(options, "all");
3885 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
3886 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3887 handle_ignore_submodules_arg(options, arg);
3888 } else if (!strcmp(arg, "--submodule"))
3889 DIFF_OPT_SET(options, SUBMODULE_LOG);
3890 else if (skip_prefix(arg, "--submodule=", &arg))
3891 return parse_submodule_opt(options, arg);
3892 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
3893 return parse_ws_error_highlight(options, arg);
3894
3895 /* misc options */
3896 else if (!strcmp(arg, "-z"))
3897 options->line_termination = 0;
3898 else if ((argcount = short_opt('l', av, &optarg))) {
3899 options->rename_limit = strtoul(optarg, NULL, 10);
3900 return argcount;
3901 }
3902 else if ((argcount = short_opt('S', av, &optarg))) {
3903 options->pickaxe = optarg;
3904 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3905 return argcount;
3906 } else if ((argcount = short_opt('G', av, &optarg))) {
3907 options->pickaxe = optarg;
3908 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3909 return argcount;
3910 }
3911 else if (!strcmp(arg, "--pickaxe-all"))
3912 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3913 else if (!strcmp(arg, "--pickaxe-regex"))
3914 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3915 else if ((argcount = short_opt('O', av, &optarg))) {
3916 options->orderfile = optarg;
3917 return argcount;
3918 }
3919 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3920 int offending = parse_diff_filter_opt(optarg, options);
3921 if (offending)
3922 die("unknown change class '%c' in --diff-filter=%s",
3923 offending, optarg);
3924 return argcount;
3925 }
3926 else if (!strcmp(arg, "--abbrev"))
3927 options->abbrev = DEFAULT_ABBREV;
3928 else if (skip_prefix(arg, "--abbrev=", &arg)) {
3929 options->abbrev = strtoul(arg, NULL, 10);
3930 if (options->abbrev < MINIMUM_ABBREV)
3931 options->abbrev = MINIMUM_ABBREV;
3932 else if (40 < options->abbrev)
3933 options->abbrev = 40;
3934 }
3935 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3936 options->a_prefix = optarg;
3937 return argcount;
3938 }
3939 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3940 options->b_prefix = optarg;
3941 return argcount;
3942 }
3943 else if (!strcmp(arg, "--no-prefix"))
3944 options->a_prefix = options->b_prefix = "";
3945 else if (opt_arg(arg, '\0', "inter-hunk-context",
3946 &options->interhunkcontext))
3947 ;
3948 else if (!strcmp(arg, "-W"))
3949 DIFF_OPT_SET(options, FUNCCONTEXT);
3950 else if (!strcmp(arg, "--function-context"))
3951 DIFF_OPT_SET(options, FUNCCONTEXT);
3952 else if (!strcmp(arg, "--no-function-context"))
3953 DIFF_OPT_CLR(options, FUNCCONTEXT);
3954 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3955 options->file = fopen(optarg, "w");
3956 if (!options->file)
3957 die_errno("Could not open '%s'", optarg);
3958 options->close_file = 1;
3959 return argcount;
3960 } else
3961 return 0;
3962 return 1;
3963 }
3964
3965 int parse_rename_score(const char **cp_p)
3966 {
3967 unsigned long num, scale;
3968 int ch, dot;
3969 const char *cp = *cp_p;
3970
3971 num = 0;
3972 scale = 1;
3973 dot = 0;
3974 for (;;) {
3975 ch = *cp;
3976 if ( !dot && ch == '.' ) {
3977 scale = 1;
3978 dot = 1;
3979 } else if ( ch == '%' ) {
3980 scale = dot ? scale*100 : 100;
3981 cp++; /* % is always at the end */
3982 break;
3983 } else if ( ch >= '0' && ch <= '9' ) {
3984 if ( scale < 100000 ) {
3985 scale *= 10;
3986 num = (num*10) + (ch-'0');
3987 }
3988 } else {
3989 break;
3990 }
3991 cp++;
3992 }
3993 *cp_p = cp;
3994
3995 /* user says num divided by scale and we say internally that
3996 * is MAX_SCORE * num / scale.
3997 */
3998 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3999 }
4000
4001 static int diff_scoreopt_parse(const char *opt)
4002 {
4003 int opt1, opt2, cmd;
4004
4005 if (*opt++ != '-')
4006 return -1;
4007 cmd = *opt++;
4008 if (cmd == '-') {
4009 /* convert the long-form arguments into short-form versions */
4010 if (skip_prefix(opt, "break-rewrites", &opt)) {
4011 if (*opt == 0 || *opt++ == '=')
4012 cmd = 'B';
4013 } else if (skip_prefix(opt, "find-copies", &opt)) {
4014 if (*opt == 0 || *opt++ == '=')
4015 cmd = 'C';
4016 } else if (skip_prefix(opt, "find-renames", &opt)) {
4017 if (*opt == 0 || *opt++ == '=')
4018 cmd = 'M';
4019 }
4020 }
4021 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4022 return -1; /* that is not a -M, -C, or -B option */
4023
4024 opt1 = parse_rename_score(&opt);
4025 if (cmd != 'B')
4026 opt2 = 0;
4027 else {
4028 if (*opt == 0)
4029 opt2 = 0;
4030 else if (*opt != '/')
4031 return -1; /* we expect -B80/99 or -B80 */
4032 else {
4033 opt++;
4034 opt2 = parse_rename_score(&opt);
4035 }
4036 }
4037 if (*opt != 0)
4038 return -1;
4039 return opt1 | (opt2 << 16);
4040 }
4041
4042 struct diff_queue_struct diff_queued_diff;
4043
4044 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4045 {
4046 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4047 queue->queue[queue->nr++] = dp;
4048 }
4049
4050 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4051 struct diff_filespec *one,
4052 struct diff_filespec *two)
4053 {
4054 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4055 dp->one = one;
4056 dp->two = two;
4057 if (queue)
4058 diff_q(queue, dp);
4059 return dp;
4060 }
4061
4062 void diff_free_filepair(struct diff_filepair *p)
4063 {
4064 free_filespec(p->one);
4065 free_filespec(p->two);
4066 free(p);
4067 }
4068
4069 /* This is different from find_unique_abbrev() in that
4070 * it stuffs the result with dots for alignment.
4071 */
4072 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
4073 {
4074 int abblen;
4075 const char *abbrev;
4076 if (len == 40)
4077 return sha1_to_hex(sha1);
4078
4079 abbrev = find_unique_abbrev(sha1, len);
4080 abblen = strlen(abbrev);
4081 if (abblen < 37) {
4082 static char hex[41];
4083 if (len < abblen && abblen <= len + 2)
4084 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
4085 else
4086 sprintf(hex, "%s...", abbrev);
4087 return hex;
4088 }
4089 return sha1_to_hex(sha1);
4090 }
4091
4092 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4093 {
4094 int line_termination = opt->line_termination;
4095 int inter_name_termination = line_termination ? '\t' : '\0';
4096
4097 fprintf(opt->file, "%s", diff_line_prefix(opt));
4098 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4099 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4100 diff_unique_abbrev(p->one->sha1, opt->abbrev));
4101 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
4102 }
4103 if (p->score) {
4104 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4105 inter_name_termination);
4106 } else {
4107 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4108 }
4109
4110 if (p->status == DIFF_STATUS_COPIED ||
4111 p->status == DIFF_STATUS_RENAMED) {
4112 const char *name_a, *name_b;
4113 name_a = p->one->path;
4114 name_b = p->two->path;
4115 strip_prefix(opt->prefix_length, &name_a, &name_b);
4116 write_name_quoted(name_a, opt->file, inter_name_termination);
4117 write_name_quoted(name_b, opt->file, line_termination);
4118 } else {
4119 const char *name_a, *name_b;
4120 name_a = p->one->mode ? p->one->path : p->two->path;
4121 name_b = NULL;
4122 strip_prefix(opt->prefix_length, &name_a, &name_b);
4123 write_name_quoted(name_a, opt->file, line_termination);
4124 }
4125 }
4126
4127 int diff_unmodified_pair(struct diff_filepair *p)
4128 {
4129 /* This function is written stricter than necessary to support
4130 * the currently implemented transformers, but the idea is to
4131 * let transformers to produce diff_filepairs any way they want,
4132 * and filter and clean them up here before producing the output.
4133 */
4134 struct diff_filespec *one = p->one, *two = p->two;
4135
4136 if (DIFF_PAIR_UNMERGED(p))
4137 return 0; /* unmerged is interesting */
4138
4139 /* deletion, addition, mode or type change
4140 * and rename are all interesting.
4141 */
4142 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4143 DIFF_PAIR_MODE_CHANGED(p) ||
4144 strcmp(one->path, two->path))
4145 return 0;
4146
4147 /* both are valid and point at the same path. that is, we are
4148 * dealing with a change.
4149 */
4150 if (one->sha1_valid && two->sha1_valid &&
4151 !hashcmp(one->sha1, two->sha1) &&
4152 !one->dirty_submodule && !two->dirty_submodule)
4153 return 1; /* no change */
4154 if (!one->sha1_valid && !two->sha1_valid)
4155 return 1; /* both look at the same file on the filesystem. */
4156 return 0;
4157 }
4158
4159 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4160 {
4161 if (diff_unmodified_pair(p))
4162 return;
4163
4164 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4165 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4166 return; /* no tree diffs in patch format */
4167
4168 run_diff(p, o);
4169 }
4170
4171 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4172 struct diffstat_t *diffstat)
4173 {
4174 if (diff_unmodified_pair(p))
4175 return;
4176
4177 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4178 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4179 return; /* no useful stat for tree diffs */
4180
4181 run_diffstat(p, o, diffstat);
4182 }
4183
4184 static void diff_flush_checkdiff(struct diff_filepair *p,
4185 struct diff_options *o)
4186 {
4187 if (diff_unmodified_pair(p))
4188 return;
4189
4190 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4191 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4192 return; /* nothing to check in tree diffs */
4193
4194 run_checkdiff(p, o);
4195 }
4196
4197 int diff_queue_is_empty(void)
4198 {
4199 struct diff_queue_struct *q = &diff_queued_diff;
4200 int i;
4201 for (i = 0; i < q->nr; i++)
4202 if (!diff_unmodified_pair(q->queue[i]))
4203 return 0;
4204 return 1;
4205 }
4206
4207 #if DIFF_DEBUG
4208 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4209 {
4210 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4211 x, one ? one : "",
4212 s->path,
4213 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4214 s->mode,
4215 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
4216 fprintf(stderr, "queue[%d] %s size %lu\n",
4217 x, one ? one : "",
4218 s->size);
4219 }
4220
4221 void diff_debug_filepair(const struct diff_filepair *p, int i)
4222 {
4223 diff_debug_filespec(p->one, i, "one");
4224 diff_debug_filespec(p->two, i, "two");
4225 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4226 p->score, p->status ? p->status : '?',
4227 p->one->rename_used, p->broken_pair);
4228 }
4229
4230 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4231 {
4232 int i;
4233 if (msg)
4234 fprintf(stderr, "%s\n", msg);
4235 fprintf(stderr, "q->nr = %d\n", q->nr);
4236 for (i = 0; i < q->nr; i++) {
4237 struct diff_filepair *p = q->queue[i];
4238 diff_debug_filepair(p, i);
4239 }
4240 }
4241 #endif
4242
4243 static void diff_resolve_rename_copy(void)
4244 {
4245 int i;
4246 struct diff_filepair *p;
4247 struct diff_queue_struct *q = &diff_queued_diff;
4248
4249 diff_debug_queue("resolve-rename-copy", q);
4250
4251 for (i = 0; i < q->nr; i++) {
4252 p = q->queue[i];
4253 p->status = 0; /* undecided */
4254 if (DIFF_PAIR_UNMERGED(p))
4255 p->status = DIFF_STATUS_UNMERGED;
4256 else if (!DIFF_FILE_VALID(p->one))
4257 p->status = DIFF_STATUS_ADDED;
4258 else if (!DIFF_FILE_VALID(p->two))
4259 p->status = DIFF_STATUS_DELETED;
4260 else if (DIFF_PAIR_TYPE_CHANGED(p))
4261 p->status = DIFF_STATUS_TYPE_CHANGED;
4262
4263 /* from this point on, we are dealing with a pair
4264 * whose both sides are valid and of the same type, i.e.
4265 * either in-place edit or rename/copy edit.
4266 */
4267 else if (DIFF_PAIR_RENAME(p)) {
4268 /*
4269 * A rename might have re-connected a broken
4270 * pair up, causing the pathnames to be the
4271 * same again. If so, that's not a rename at
4272 * all, just a modification..
4273 *
4274 * Otherwise, see if this source was used for
4275 * multiple renames, in which case we decrement
4276 * the count, and call it a copy.
4277 */
4278 if (!strcmp(p->one->path, p->two->path))
4279 p->status = DIFF_STATUS_MODIFIED;
4280 else if (--p->one->rename_used > 0)
4281 p->status = DIFF_STATUS_COPIED;
4282 else
4283 p->status = DIFF_STATUS_RENAMED;
4284 }
4285 else if (hashcmp(p->one->sha1, p->two->sha1) ||
4286 p->one->mode != p->two->mode ||
4287 p->one->dirty_submodule ||
4288 p->two->dirty_submodule ||
4289 is_null_sha1(p->one->sha1))
4290 p->status = DIFF_STATUS_MODIFIED;
4291 else {
4292 /* This is a "no-change" entry and should not
4293 * happen anymore, but prepare for broken callers.
4294 */
4295 error("feeding unmodified %s to diffcore",
4296 p->one->path);
4297 p->status = DIFF_STATUS_UNKNOWN;
4298 }
4299 }
4300 diff_debug_queue("resolve-rename-copy done", q);
4301 }
4302
4303 static int check_pair_status(struct diff_filepair *p)
4304 {
4305 switch (p->status) {
4306 case DIFF_STATUS_UNKNOWN:
4307 return 0;
4308 case 0:
4309 die("internal error in diff-resolve-rename-copy");
4310 default:
4311 return 1;
4312 }
4313 }
4314
4315 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4316 {
4317 int fmt = opt->output_format;
4318
4319 if (fmt & DIFF_FORMAT_CHECKDIFF)
4320 diff_flush_checkdiff(p, opt);
4321 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4322 diff_flush_raw(p, opt);
4323 else if (fmt & DIFF_FORMAT_NAME) {
4324 const char *name_a, *name_b;
4325 name_a = p->two->path;
4326 name_b = NULL;
4327 strip_prefix(opt->prefix_length, &name_a, &name_b);
4328 write_name_quoted(name_a, opt->file, opt->line_termination);
4329 }
4330 }
4331
4332 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4333 {
4334 if (fs->mode)
4335 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4336 else
4337 fprintf(file, " %s ", newdelete);
4338 write_name_quoted(fs->path, file, '\n');
4339 }
4340
4341
4342 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4343 const char *line_prefix)
4344 {
4345 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4346 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4347 p->two->mode, show_name ? ' ' : '\n');
4348 if (show_name) {
4349 write_name_quoted(p->two->path, file, '\n');
4350 }
4351 }
4352 }
4353
4354 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4355 const char *line_prefix)
4356 {
4357 char *names = pprint_rename(p->one->path, p->two->path);
4358
4359 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4360 free(names);
4361 show_mode_change(file, p, 0, line_prefix);
4362 }
4363
4364 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4365 {
4366 FILE *file = opt->file;
4367 const char *line_prefix = diff_line_prefix(opt);
4368
4369 switch(p->status) {
4370 case DIFF_STATUS_DELETED:
4371 fputs(line_prefix, file);
4372 show_file_mode_name(file, "delete", p->one);
4373 break;
4374 case DIFF_STATUS_ADDED:
4375 fputs(line_prefix, file);
4376 show_file_mode_name(file, "create", p->two);
4377 break;
4378 case DIFF_STATUS_COPIED:
4379 fputs(line_prefix, file);
4380 show_rename_copy(file, "copy", p, line_prefix);
4381 break;
4382 case DIFF_STATUS_RENAMED:
4383 fputs(line_prefix, file);
4384 show_rename_copy(file, "rename", p, line_prefix);
4385 break;
4386 default:
4387 if (p->score) {
4388 fprintf(file, "%s rewrite ", line_prefix);
4389 write_name_quoted(p->two->path, file, ' ');
4390 fprintf(file, "(%d%%)\n", similarity_index(p));
4391 }
4392 show_mode_change(file, p, !p->score, line_prefix);
4393 break;
4394 }
4395 }
4396
4397 struct patch_id_t {
4398 git_SHA_CTX *ctx;
4399 int patchlen;
4400 };
4401
4402 static int remove_space(char *line, int len)
4403 {
4404 int i;
4405 char *dst = line;
4406 unsigned char c;
4407
4408 for (i = 0; i < len; i++)
4409 if (!isspace((c = line[i])))
4410 *dst++ = c;
4411
4412 return dst - line;
4413 }
4414
4415 static void patch_id_consume(void *priv, char *line, unsigned long len)
4416 {
4417 struct patch_id_t *data = priv;
4418 int new_len;
4419
4420 /* Ignore line numbers when computing the SHA1 of the patch */
4421 if (starts_with(line, "@@ -"))
4422 return;
4423
4424 new_len = remove_space(line, len);
4425
4426 git_SHA1_Update(data->ctx, line, new_len);
4427 data->patchlen += new_len;
4428 }
4429
4430 /* returns 0 upon success, and writes result into sha1 */
4431 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4432 {
4433 struct diff_queue_struct *q = &diff_queued_diff;
4434 int i;
4435 git_SHA_CTX ctx;
4436 struct patch_id_t data;
4437 char buffer[PATH_MAX * 4 + 20];
4438
4439 git_SHA1_Init(&ctx);
4440 memset(&data, 0, sizeof(struct patch_id_t));
4441 data.ctx = &ctx;
4442
4443 for (i = 0; i < q->nr; i++) {
4444 xpparam_t xpp;
4445 xdemitconf_t xecfg;
4446 mmfile_t mf1, mf2;
4447 struct diff_filepair *p = q->queue[i];
4448 int len1, len2;
4449
4450 memset(&xpp, 0, sizeof(xpp));
4451 memset(&xecfg, 0, sizeof(xecfg));
4452 if (p->status == 0)
4453 return error("internal diff status error");
4454 if (p->status == DIFF_STATUS_UNKNOWN)
4455 continue;
4456 if (diff_unmodified_pair(p))
4457 continue;
4458 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4459 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4460 continue;
4461 if (DIFF_PAIR_UNMERGED(p))
4462 continue;
4463
4464 diff_fill_sha1_info(p->one);
4465 diff_fill_sha1_info(p->two);
4466 if (fill_mmfile(&mf1, p->one) < 0 ||
4467 fill_mmfile(&mf2, p->two) < 0)
4468 return error("unable to read files to diff");
4469
4470 len1 = remove_space(p->one->path, strlen(p->one->path));
4471 len2 = remove_space(p->two->path, strlen(p->two->path));
4472 if (p->one->mode == 0)
4473 len1 = snprintf(buffer, sizeof(buffer),
4474 "diff--gita/%.*sb/%.*s"
4475 "newfilemode%06o"
4476 "---/dev/null"
4477 "+++b/%.*s",
4478 len1, p->one->path,
4479 len2, p->two->path,
4480 p->two->mode,
4481 len2, p->two->path);
4482 else if (p->two->mode == 0)
4483 len1 = snprintf(buffer, sizeof(buffer),
4484 "diff--gita/%.*sb/%.*s"
4485 "deletedfilemode%06o"
4486 "---a/%.*s"
4487 "+++/dev/null",
4488 len1, p->one->path,
4489 len2, p->two->path,
4490 p->one->mode,
4491 len1, p->one->path);
4492 else
4493 len1 = snprintf(buffer, sizeof(buffer),
4494 "diff--gita/%.*sb/%.*s"
4495 "---a/%.*s"
4496 "+++b/%.*s",
4497 len1, p->one->path,
4498 len2, p->two->path,
4499 len1, p->one->path,
4500 len2, p->two->path);
4501 git_SHA1_Update(&ctx, buffer, len1);
4502
4503 if (diff_filespec_is_binary(p->one) ||
4504 diff_filespec_is_binary(p->two)) {
4505 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4506 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4507 continue;
4508 }
4509
4510 xpp.flags = 0;
4511 xecfg.ctxlen = 3;
4512 xecfg.flags = 0;
4513 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4514 &xpp, &xecfg);
4515 }
4516
4517 git_SHA1_Final(sha1, &ctx);
4518 return 0;
4519 }
4520
4521 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4522 {
4523 struct diff_queue_struct *q = &diff_queued_diff;
4524 int i;
4525 int result = diff_get_patch_id(options, sha1);
4526
4527 for (i = 0; i < q->nr; i++)
4528 diff_free_filepair(q->queue[i]);
4529
4530 free(q->queue);
4531 DIFF_QUEUE_CLEAR(q);
4532
4533 return result;
4534 }
4535
4536 static int is_summary_empty(const struct diff_queue_struct *q)
4537 {
4538 int i;
4539
4540 for (i = 0; i < q->nr; i++) {
4541 const struct diff_filepair *p = q->queue[i];
4542
4543 switch (p->status) {
4544 case DIFF_STATUS_DELETED:
4545 case DIFF_STATUS_ADDED:
4546 case DIFF_STATUS_COPIED:
4547 case DIFF_STATUS_RENAMED:
4548 return 0;
4549 default:
4550 if (p->score)
4551 return 0;
4552 if (p->one->mode && p->two->mode &&
4553 p->one->mode != p->two->mode)
4554 return 0;
4555 break;
4556 }
4557 }
4558 return 1;
4559 }
4560
4561 static const char rename_limit_warning[] =
4562 "inexact rename detection was skipped due to too many files.";
4563
4564 static const char degrade_cc_to_c_warning[] =
4565 "only found copies from modified paths due to too many files.";
4566
4567 static const char rename_limit_advice[] =
4568 "you may want to set your %s variable to at least "
4569 "%d and retry the command.";
4570
4571 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4572 {
4573 if (degraded_cc)
4574 warning(degrade_cc_to_c_warning);
4575 else if (needed)
4576 warning(rename_limit_warning);
4577 else
4578 return;
4579 if (0 < needed && needed < 32767)
4580 warning(rename_limit_advice, varname, needed);
4581 }
4582
4583 void diff_flush(struct diff_options *options)
4584 {
4585 struct diff_queue_struct *q = &diff_queued_diff;
4586 int i, output_format = options->output_format;
4587 int separator = 0;
4588 int dirstat_by_line = 0;
4589
4590 /*
4591 * Order: raw, stat, summary, patch
4592 * or: name/name-status/checkdiff (other bits clear)
4593 */
4594 if (!q->nr)
4595 goto free_queue;
4596
4597 if (output_format & (DIFF_FORMAT_RAW |
4598 DIFF_FORMAT_NAME |
4599 DIFF_FORMAT_NAME_STATUS |
4600 DIFF_FORMAT_CHECKDIFF)) {
4601 for (i = 0; i < q->nr; i++) {
4602 struct diff_filepair *p = q->queue[i];
4603 if (check_pair_status(p))
4604 flush_one_pair(p, options);
4605 }
4606 separator++;
4607 }
4608
4609 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4610 dirstat_by_line = 1;
4611
4612 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4613 dirstat_by_line) {
4614 struct diffstat_t diffstat;
4615
4616 memset(&diffstat, 0, sizeof(struct diffstat_t));
4617 for (i = 0; i < q->nr; i++) {
4618 struct diff_filepair *p = q->queue[i];
4619 if (check_pair_status(p))
4620 diff_flush_stat(p, options, &diffstat);
4621 }
4622 if (output_format & DIFF_FORMAT_NUMSTAT)
4623 show_numstat(&diffstat, options);
4624 if (output_format & DIFF_FORMAT_DIFFSTAT)
4625 show_stats(&diffstat, options);
4626 if (output_format & DIFF_FORMAT_SHORTSTAT)
4627 show_shortstats(&diffstat, options);
4628 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
4629 show_dirstat_by_line(&diffstat, options);
4630 free_diffstat_info(&diffstat);
4631 separator++;
4632 }
4633 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4634 show_dirstat(options);
4635
4636 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4637 for (i = 0; i < q->nr; i++) {
4638 diff_summary(options, q->queue[i]);
4639 }
4640 separator++;
4641 }
4642
4643 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4644 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4645 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4646 /*
4647 * run diff_flush_patch for the exit status. setting
4648 * options->file to /dev/null should be safe, because we
4649 * aren't supposed to produce any output anyway.
4650 */
4651 if (options->close_file)
4652 fclose(options->file);
4653 options->file = fopen("/dev/null", "w");
4654 if (!options->file)
4655 die_errno("Could not open /dev/null");
4656 options->close_file = 1;
4657 for (i = 0; i < q->nr; i++) {
4658 struct diff_filepair *p = q->queue[i];
4659 if (check_pair_status(p))
4660 diff_flush_patch(p, options);
4661 if (options->found_changes)
4662 break;
4663 }
4664 }
4665
4666 if (output_format & DIFF_FORMAT_PATCH) {
4667 if (separator) {
4668 fprintf(options->file, "%s%c",
4669 diff_line_prefix(options),
4670 options->line_termination);
4671 if (options->stat_sep) {
4672 /* attach patch instead of inline */
4673 fputs(options->stat_sep, options->file);
4674 }
4675 }
4676
4677 for (i = 0; i < q->nr; i++) {
4678 struct diff_filepair *p = q->queue[i];
4679 if (check_pair_status(p))
4680 diff_flush_patch(p, options);
4681 }
4682 }
4683
4684 if (output_format & DIFF_FORMAT_CALLBACK)
4685 options->format_callback(q, options, options->format_callback_data);
4686
4687 for (i = 0; i < q->nr; i++)
4688 diff_free_filepair(q->queue[i]);
4689 free_queue:
4690 free(q->queue);
4691 DIFF_QUEUE_CLEAR(q);
4692 if (options->close_file)
4693 fclose(options->file);
4694
4695 /*
4696 * Report the content-level differences with HAS_CHANGES;
4697 * diff_addremove/diff_change does not set the bit when
4698 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4699 */
4700 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4701 if (options->found_changes)
4702 DIFF_OPT_SET(options, HAS_CHANGES);
4703 else
4704 DIFF_OPT_CLR(options, HAS_CHANGES);
4705 }
4706 }
4707
4708 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
4709 {
4710 return (((p->status == DIFF_STATUS_MODIFIED) &&
4711 ((p->score &&
4712 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
4713 (!p->score &&
4714 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
4715 ((p->status != DIFF_STATUS_MODIFIED) &&
4716 filter_bit_tst(p->status, options)));
4717 }
4718
4719 static void diffcore_apply_filter(struct diff_options *options)
4720 {
4721 int i;
4722 struct diff_queue_struct *q = &diff_queued_diff;
4723 struct diff_queue_struct outq;
4724
4725 DIFF_QUEUE_CLEAR(&outq);
4726
4727 if (!options->filter)
4728 return;
4729
4730 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
4731 int found;
4732 for (i = found = 0; !found && i < q->nr; i++) {
4733 if (match_filter(options, q->queue[i]))
4734 found++;
4735 }
4736 if (found)
4737 return;
4738
4739 /* otherwise we will clear the whole queue
4740 * by copying the empty outq at the end of this
4741 * function, but first clear the current entries
4742 * in the queue.
4743 */
4744 for (i = 0; i < q->nr; i++)
4745 diff_free_filepair(q->queue[i]);
4746 }
4747 else {
4748 /* Only the matching ones */
4749 for (i = 0; i < q->nr; i++) {
4750 struct diff_filepair *p = q->queue[i];
4751 if (match_filter(options, p))
4752 diff_q(&outq, p);
4753 else
4754 diff_free_filepair(p);
4755 }
4756 }
4757 free(q->queue);
4758 *q = outq;
4759 }
4760
4761 /* Check whether two filespecs with the same mode and size are identical */
4762 static int diff_filespec_is_identical(struct diff_filespec *one,
4763 struct diff_filespec *two)
4764 {
4765 if (S_ISGITLINK(one->mode))
4766 return 0;
4767 if (diff_populate_filespec(one, 0))
4768 return 0;
4769 if (diff_populate_filespec(two, 0))
4770 return 0;
4771 return !memcmp(one->data, two->data, one->size);
4772 }
4773
4774 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
4775 {
4776 if (p->done_skip_stat_unmatch)
4777 return p->skip_stat_unmatch_result;
4778
4779 p->done_skip_stat_unmatch = 1;
4780 p->skip_stat_unmatch_result = 0;
4781 /*
4782 * 1. Entries that come from stat info dirtiness
4783 * always have both sides (iow, not create/delete),
4784 * one side of the object name is unknown, with
4785 * the same mode and size. Keep the ones that
4786 * do not match these criteria. They have real
4787 * differences.
4788 *
4789 * 2. At this point, the file is known to be modified,
4790 * with the same mode and size, and the object
4791 * name of one side is unknown. Need to inspect
4792 * the identical contents.
4793 */
4794 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4795 !DIFF_FILE_VALID(p->two) ||
4796 (p->one->sha1_valid && p->two->sha1_valid) ||
4797 (p->one->mode != p->two->mode) ||
4798 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
4799 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
4800 (p->one->size != p->two->size) ||
4801 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4802 p->skip_stat_unmatch_result = 1;
4803 return p->skip_stat_unmatch_result;
4804 }
4805
4806 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4807 {
4808 int i;
4809 struct diff_queue_struct *q = &diff_queued_diff;
4810 struct diff_queue_struct outq;
4811 DIFF_QUEUE_CLEAR(&outq);
4812
4813 for (i = 0; i < q->nr; i++) {
4814 struct diff_filepair *p = q->queue[i];
4815
4816 if (diff_filespec_check_stat_unmatch(p))
4817 diff_q(&outq, p);
4818 else {
4819 /*
4820 * The caller can subtract 1 from skip_stat_unmatch
4821 * to determine how many paths were dirty only
4822 * due to stat info mismatch.
4823 */
4824 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4825 diffopt->skip_stat_unmatch++;
4826 diff_free_filepair(p);
4827 }
4828 }
4829 free(q->queue);
4830 *q = outq;
4831 }
4832
4833 static int diffnamecmp(const void *a_, const void *b_)
4834 {
4835 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4836 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4837 const char *name_a, *name_b;
4838
4839 name_a = a->one ? a->one->path : a->two->path;
4840 name_b = b->one ? b->one->path : b->two->path;
4841 return strcmp(name_a, name_b);
4842 }
4843
4844 void diffcore_fix_diff_index(struct diff_options *options)
4845 {
4846 struct diff_queue_struct *q = &diff_queued_diff;
4847 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4848 }
4849
4850 void diffcore_std(struct diff_options *options)
4851 {
4852 /* NOTE please keep the following in sync with diff_tree_combined() */
4853 if (options->skip_stat_unmatch)
4854 diffcore_skip_stat_unmatch(options);
4855 if (!options->found_follow) {
4856 /* See try_to_follow_renames() in tree-diff.c */
4857 if (options->break_opt != -1)
4858 diffcore_break(options->break_opt);
4859 if (options->detect_rename)
4860 diffcore_rename(options);
4861 if (options->break_opt != -1)
4862 diffcore_merge_broken();
4863 }
4864 if (options->pickaxe)
4865 diffcore_pickaxe(options);
4866 if (options->orderfile)
4867 diffcore_order(options->orderfile);
4868 if (!options->found_follow)
4869 /* See try_to_follow_renames() in tree-diff.c */
4870 diff_resolve_rename_copy();
4871 diffcore_apply_filter(options);
4872
4873 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4874 DIFF_OPT_SET(options, HAS_CHANGES);
4875 else
4876 DIFF_OPT_CLR(options, HAS_CHANGES);
4877
4878 options->found_follow = 0;
4879 }
4880
4881 int diff_result_code(struct diff_options *opt, int status)
4882 {
4883 int result = 0;
4884
4885 diff_warn_rename_limit("diff.renameLimit",
4886 opt->needed_rename_limit,
4887 opt->degraded_cc_to_c);
4888 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4889 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4890 return status;
4891 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4892 DIFF_OPT_TST(opt, HAS_CHANGES))
4893 result |= 01;
4894 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4895 DIFF_OPT_TST(opt, CHECK_FAILED))
4896 result |= 02;
4897 return result;
4898 }
4899
4900 int diff_can_quit_early(struct diff_options *opt)
4901 {
4902 return (DIFF_OPT_TST(opt, QUICK) &&
4903 !opt->filter &&
4904 DIFF_OPT_TST(opt, HAS_CHANGES));
4905 }
4906
4907 /*
4908 * Shall changes to this submodule be ignored?
4909 *
4910 * Submodule changes can be configured to be ignored separately for each path,
4911 * but that configuration can be overridden from the command line.
4912 */
4913 static int is_submodule_ignored(const char *path, struct diff_options *options)
4914 {
4915 int ignored = 0;
4916 unsigned orig_flags = options->flags;
4917 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4918 set_diffopt_flags_from_submodule_config(options, path);
4919 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4920 ignored = 1;
4921 options->flags = orig_flags;
4922 return ignored;
4923 }
4924
4925 void diff_addremove(struct diff_options *options,
4926 int addremove, unsigned mode,
4927 const unsigned char *sha1,
4928 int sha1_valid,
4929 const char *concatpath, unsigned dirty_submodule)
4930 {
4931 struct diff_filespec *one, *two;
4932
4933 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4934 return;
4935
4936 /* This may look odd, but it is a preparation for
4937 * feeding "there are unchanged files which should
4938 * not produce diffs, but when you are doing copy
4939 * detection you would need them, so here they are"
4940 * entries to the diff-core. They will be prefixed
4941 * with something like '=' or '*' (I haven't decided
4942 * which but should not make any difference).
4943 * Feeding the same new and old to diff_change()
4944 * also has the same effect.
4945 * Before the final output happens, they are pruned after
4946 * merged into rename/copy pairs as appropriate.
4947 */
4948 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4949 addremove = (addremove == '+' ? '-' :
4950 addremove == '-' ? '+' : addremove);
4951
4952 if (options->prefix &&
4953 strncmp(concatpath, options->prefix, options->prefix_length))
4954 return;
4955
4956 one = alloc_filespec(concatpath);
4957 two = alloc_filespec(concatpath);
4958
4959 if (addremove != '+')
4960 fill_filespec(one, sha1, sha1_valid, mode);
4961 if (addremove != '-') {
4962 fill_filespec(two, sha1, sha1_valid, mode);
4963 two->dirty_submodule = dirty_submodule;
4964 }
4965
4966 diff_queue(&diff_queued_diff, one, two);
4967 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4968 DIFF_OPT_SET(options, HAS_CHANGES);
4969 }
4970
4971 void diff_change(struct diff_options *options,
4972 unsigned old_mode, unsigned new_mode,
4973 const unsigned char *old_sha1,
4974 const unsigned char *new_sha1,
4975 int old_sha1_valid, int new_sha1_valid,
4976 const char *concatpath,
4977 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4978 {
4979 struct diff_filespec *one, *two;
4980 struct diff_filepair *p;
4981
4982 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4983 is_submodule_ignored(concatpath, options))
4984 return;
4985
4986 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4987 unsigned tmp;
4988 const unsigned char *tmp_c;
4989 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4990 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4991 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
4992 new_sha1_valid = tmp;
4993 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4994 new_dirty_submodule = tmp;
4995 }
4996
4997 if (options->prefix &&
4998 strncmp(concatpath, options->prefix, options->prefix_length))
4999 return;
5000
5001 one = alloc_filespec(concatpath);
5002 two = alloc_filespec(concatpath);
5003 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
5004 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
5005 one->dirty_submodule = old_dirty_submodule;
5006 two->dirty_submodule = new_dirty_submodule;
5007 p = diff_queue(&diff_queued_diff, one, two);
5008
5009 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5010 return;
5011
5012 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5013 !diff_filespec_check_stat_unmatch(p))
5014 return;
5015
5016 DIFF_OPT_SET(options, HAS_CHANGES);
5017 }
5018
5019 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5020 {
5021 struct diff_filepair *pair;
5022 struct diff_filespec *one, *two;
5023
5024 if (options->prefix &&
5025 strncmp(path, options->prefix, options->prefix_length))
5026 return NULL;
5027
5028 one = alloc_filespec(path);
5029 two = alloc_filespec(path);
5030 pair = diff_queue(&diff_queued_diff, one, two);
5031 pair->is_unmerged = 1;
5032 return pair;
5033 }
5034
5035 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5036 size_t *outsize)
5037 {
5038 struct diff_tempfile *temp;
5039 const char *argv[3];
5040 const char **arg = argv;
5041 struct child_process child = CHILD_PROCESS_INIT;
5042 struct strbuf buf = STRBUF_INIT;
5043 int err = 0;
5044
5045 temp = prepare_temp_file(spec->path, spec);
5046 *arg++ = pgm;
5047 *arg++ = temp->name;
5048 *arg = NULL;
5049
5050 child.use_shell = 1;
5051 child.argv = argv;
5052 child.out = -1;
5053 if (start_command(&child)) {
5054 remove_tempfile();
5055 return NULL;
5056 }
5057
5058 if (strbuf_read(&buf, child.out, 0) < 0)
5059 err = error("error reading from textconv command '%s'", pgm);
5060 close(child.out);
5061
5062 if (finish_command(&child) || err) {
5063 strbuf_release(&buf);
5064 remove_tempfile();
5065 return NULL;
5066 }
5067 remove_tempfile();
5068
5069 return strbuf_detach(&buf, outsize);
5070 }
5071
5072 size_t fill_textconv(struct userdiff_driver *driver,
5073 struct diff_filespec *df,
5074 char **outbuf)
5075 {
5076 size_t size;
5077
5078 if (!driver || !driver->textconv) {
5079 if (!DIFF_FILE_VALID(df)) {
5080 *outbuf = "";
5081 return 0;
5082 }
5083 if (diff_populate_filespec(df, 0))
5084 die("unable to read files to diff");
5085 *outbuf = df->data;
5086 return df->size;
5087 }
5088
5089 if (driver->textconv_cache && df->sha1_valid) {
5090 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
5091 &size);
5092 if (*outbuf)
5093 return size;
5094 }
5095
5096 *outbuf = run_textconv(driver->textconv, df, &size);
5097 if (!*outbuf)
5098 die("unable to read files to diff");
5099
5100 if (driver->textconv_cache && df->sha1_valid) {
5101 /* ignore errors, as we might be in a readonly repository */
5102 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
5103 size);
5104 /*
5105 * we could save up changes and flush them all at the end,
5106 * but we would need an extra call after all diffing is done.
5107 * Since generating a cache entry is the slow path anyway,
5108 * this extra overhead probably isn't a big deal.
5109 */
5110 notes_cache_write(driver->textconv_cache);
5111 }
5112
5113 return size;
5114 }
5115
5116 void setup_diff_pager(struct diff_options *opt)
5117 {
5118 /*
5119 * If the user asked for our exit code, then either they want --quiet
5120 * or --exit-code. We should definitely not bother with a pager in the
5121 * former case, as we will generate no output. Since we still properly
5122 * report our exit code even when a pager is run, we _could_ run a
5123 * pager with --exit-code. But since we have not done so historically,
5124 * and because it is easy to find people oneline advising "git diff
5125 * --exit-code" in hooks and other scripts, we do not do so.
5126 */
5127 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5128 check_pager_config("diff") != 0)
5129 setup_pager();
5130 }
Eric Wong's git-svn branch