?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
allow do_submodule_path to work even if submodule isn't checked out
[git.git] / diff.c
blob1380bbe250ad333a68f267a2fa3e345233fd0e1b
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 "submodule-config.h"
17 #include "submodule.h"
18 #include "ll-merge.h"
19 #include "string-list.h"
20 #include "argv-array.h"
21 #include "graph.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_compaction_heuristic; /* experimental */
31 static int diff_rename_limit_default = 400;
32 static int diff_suppress_blank_empty;
33 static int diff_use_color_default = -1;
34 static int diff_context_default = 3;
35 static const char *diff_word_regex_cfg;
36 static const char *external_diff_cmd_cfg;
37 static const char *diff_order_file_cfg;
38 int diff_auto_refresh_index = 1;
39 static int diff_mnemonic_prefix;
40 static int diff_no_prefix;
41 static int diff_stat_graph_width;
42 static int diff_dirstat_permille_default = 30;
43 static struct diff_options default_diff_options;
44 static long diff_algorithm;
45
46 static char diff_colors[][COLOR_MAXLEN] = {
47 GIT_COLOR_RESET,
48 GIT_COLOR_NORMAL, /* CONTEXT */
49 GIT_COLOR_BOLD, /* METAINFO */
50 GIT_COLOR_CYAN, /* FRAGINFO */
51 GIT_COLOR_RED, /* OLD */
52 GIT_COLOR_GREEN, /* NEW */
53 GIT_COLOR_YELLOW, /* COMMIT */
54 GIT_COLOR_BG_RED, /* WHITESPACE */
55 GIT_COLOR_NORMAL, /* FUNCINFO */
56 };
57
58 static int parse_diff_color_slot(const char *var)
59 {
60 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
61 return DIFF_CONTEXT;
62 if (!strcasecmp(var, "meta"))
63 return DIFF_METAINFO;
64 if (!strcasecmp(var, "frag"))
65 return DIFF_FRAGINFO;
66 if (!strcasecmp(var, "old"))
67 return DIFF_FILE_OLD;
68 if (!strcasecmp(var, "new"))
69 return DIFF_FILE_NEW;
70 if (!strcasecmp(var, "commit"))
71 return DIFF_COMMIT;
72 if (!strcasecmp(var, "whitespace"))
73 return DIFF_WHITESPACE;
74 if (!strcasecmp(var, "func"))
75 return DIFF_FUNCINFO;
76 return -1;
77 }
78
79 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
80 struct strbuf *errmsg)
81 {
82 char *params_copy = xstrdup(params_string);
83 struct string_list params = STRING_LIST_INIT_NODUP;
84 int ret = 0;
85 int i;
86
87 if (*params_copy)
88 string_list_split_in_place(&params, params_copy, ',', -1);
89 for (i = 0; i < params.nr; i++) {
90 const char *p = params.items[i].string;
91 if (!strcmp(p, "changes")) {
92 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
93 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
94 } else if (!strcmp(p, "lines")) {
95 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
96 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
97 } else if (!strcmp(p, "files")) {
98 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
99 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
100 } else if (!strcmp(p, "noncumulative")) {
101 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
102 } else if (!strcmp(p, "cumulative")) {
103 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
104 } else if (isdigit(*p)) {
105 char *end;
106 int permille = strtoul(p, &end, 10) * 10;
107 if (*end == '.' && isdigit(*++end)) {
108 /* only use first digit */
109 permille += *end - '0';
110 /* .. and ignore any further digits */
111 while (isdigit(*++end))
112 ; /* nothing */
113 }
114 if (!*end)
115 options->dirstat_permille = permille;
116 else {
117 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
118 p);
119 ret++;
120 }
121 } else {
122 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
123 ret++;
124 }
125
126 }
127 string_list_clear(&params, 0);
128 free(params_copy);
129 return ret;
130 }
131
132 static int parse_submodule_params(struct diff_options *options, const char *value)
133 {
134 if (!strcmp(value, "log"))
135 options->submodule_format = DIFF_SUBMODULE_LOG;
136 else if (!strcmp(value, "short"))
137 options->submodule_format = DIFF_SUBMODULE_SHORT;
138 else
139 return -1;
140 return 0;
141 }
142
143 static int git_config_rename(const char *var, const char *value)
144 {
145 if (!value)
146 return DIFF_DETECT_RENAME;
147 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
148 return DIFF_DETECT_COPY;
149 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
150 }
151
152 long parse_algorithm_value(const char *value)
153 {
154 if (!value)
155 return -1;
156 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
157 return 0;
158 else if (!strcasecmp(value, "minimal"))
159 return XDF_NEED_MINIMAL;
160 else if (!strcasecmp(value, "patience"))
161 return XDF_PATIENCE_DIFF;
162 else if (!strcasecmp(value, "histogram"))
163 return XDF_HISTOGRAM_DIFF;
164 return -1;
165 }
166
167 /*
168 * These are to give UI layer defaults.
169 * The core-level commands such as git-diff-files should
170 * never be affected by the setting of diff.renames
171 * the user happens to have in the configuration file.
172 */
173 void init_diff_ui_defaults(void)
174 {
175 diff_detect_rename_default = 1;
176 }
177
178 int git_diff_ui_config(const char *var, const char *value, void *cb)
179 {
180 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
181 diff_use_color_default = git_config_colorbool(var, value);
182 return 0;
183 }
184 if (!strcmp(var, "diff.context")) {
185 diff_context_default = git_config_int(var, value);
186 if (diff_context_default < 0)
187 return -1;
188 return 0;
189 }
190 if (!strcmp(var, "diff.renames")) {
191 diff_detect_rename_default = git_config_rename(var, value);
192 return 0;
193 }
194 if (!strcmp(var, "diff.compactionheuristic")) {
195 diff_compaction_heuristic = git_config_bool(var, value);
196 return 0;
197 }
198 if (!strcmp(var, "diff.autorefreshindex")) {
199 diff_auto_refresh_index = git_config_bool(var, value);
200 return 0;
201 }
202 if (!strcmp(var, "diff.mnemonicprefix")) {
203 diff_mnemonic_prefix = git_config_bool(var, value);
204 return 0;
205 }
206 if (!strcmp(var, "diff.noprefix")) {
207 diff_no_prefix = git_config_bool(var, value);
208 return 0;
209 }
210 if (!strcmp(var, "diff.statgraphwidth")) {
211 diff_stat_graph_width = git_config_int(var, value);
212 return 0;
213 }
214 if (!strcmp(var, "diff.external"))
215 return git_config_string(&external_diff_cmd_cfg, var, value);
216 if (!strcmp(var, "diff.wordregex"))
217 return git_config_string(&diff_word_regex_cfg, var, value);
218 if (!strcmp(var, "diff.orderfile"))
219 return git_config_pathname(&diff_order_file_cfg, var, value);
220
221 if (!strcmp(var, "diff.ignoresubmodules"))
222 handle_ignore_submodules_arg(&default_diff_options, value);
223
224 if (!strcmp(var, "diff.submodule")) {
225 if (parse_submodule_params(&default_diff_options, value))
226 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
227 value);
228 return 0;
229 }
230
231 if (!strcmp(var, "diff.algorithm")) {
232 diff_algorithm = parse_algorithm_value(value);
233 if (diff_algorithm < 0)
234 return -1;
235 return 0;
236 }
237
238 if (git_color_config(var, value, cb) < 0)
239 return -1;
240
241 return git_diff_basic_config(var, value, cb);
242 }
243
244 int git_diff_basic_config(const char *var, const char *value, void *cb)
245 {
246 const char *name;
247
248 if (!strcmp(var, "diff.renamelimit")) {
249 diff_rename_limit_default = git_config_int(var, value);
250 return 0;
251 }
252
253 if (userdiff_config(var, value) < 0)
254 return -1;
255
256 if (skip_prefix(var, "diff.color.", &name) ||
257 skip_prefix(var, "color.diff.", &name)) {
258 int slot = parse_diff_color_slot(name);
259 if (slot < 0)
260 return 0;
261 if (!value)
262 return config_error_nonbool(var);
263 return color_parse(value, diff_colors[slot]);
264 }
265
266 /* like GNU diff's --suppress-blank-empty option */
267 if (!strcmp(var, "diff.suppressblankempty") ||
268 /* for backwards compatibility */
269 !strcmp(var, "diff.suppress-blank-empty")) {
270 diff_suppress_blank_empty = git_config_bool(var, value);
271 return 0;
272 }
273
274 if (!strcmp(var, "diff.dirstat")) {
275 struct strbuf errmsg = STRBUF_INIT;
276 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
277 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
278 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
279 errmsg.buf);
280 strbuf_release(&errmsg);
281 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
282 return 0;
283 }
284
285 if (starts_with(var, "submodule."))
286 return parse_submodule_config_option(var, value);
287
288 return git_default_config(var, value, cb);
289 }
290
291 static char *quote_two(const char *one, const char *two)
292 {
293 int need_one = quote_c_style(one, NULL, NULL, 1);
294 int need_two = quote_c_style(two, NULL, NULL, 1);
295 struct strbuf res = STRBUF_INIT;
296
297 if (need_one + need_two) {
298 strbuf_addch(&res, '"');
299 quote_c_style(one, &res, NULL, 1);
300 quote_c_style(two, &res, NULL, 1);
301 strbuf_addch(&res, '"');
302 } else {
303 strbuf_addstr(&res, one);
304 strbuf_addstr(&res, two);
305 }
306 return strbuf_detach(&res, NULL);
307 }
308
309 static const char *external_diff(void)
310 {
311 static const char *external_diff_cmd = NULL;
312 static int done_preparing = 0;
313
314 if (done_preparing)
315 return external_diff_cmd;
316 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
317 if (!external_diff_cmd)
318 external_diff_cmd = external_diff_cmd_cfg;
319 done_preparing = 1;
320 return external_diff_cmd;
321 }
322
323 /*
324 * Keep track of files used for diffing. Sometimes such an entry
325 * refers to a temporary file, sometimes to an existing file, and
326 * sometimes to "/dev/null".
327 */
328 static struct diff_tempfile {
329 /*
330 * filename external diff should read from, or NULL if this
331 * entry is currently not in use:
332 */
333 const char *name;
334
335 char hex[GIT_SHA1_HEXSZ + 1];
336 char mode[10];
337
338 /*
339 * If this diff_tempfile instance refers to a temporary file,
340 * this tempfile object is used to manage its lifetime.
341 */
342 struct tempfile tempfile;
343 } diff_temp[2];
344
345 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
346
347 struct emit_callback {
348 int color_diff;
349 unsigned ws_rule;
350 int blank_at_eof_in_preimage;
351 int blank_at_eof_in_postimage;
352 int lno_in_preimage;
353 int lno_in_postimage;
354 sane_truncate_fn truncate;
355 const char **label_path;
356 struct diff_words_data *diff_words;
357 struct diff_options *opt;
358 int *found_changesp;
359 struct strbuf *header;
360 };
361
362 static int count_lines(const char *data, int size)
363 {
364 int count, ch, completely_empty = 1, nl_just_seen = 0;
365 count = 0;
366 while (0 < size--) {
367 ch = *data++;
368 if (ch == '\n') {
369 count++;
370 nl_just_seen = 1;
371 completely_empty = 0;
372 }
373 else {
374 nl_just_seen = 0;
375 completely_empty = 0;
376 }
377 }
378 if (completely_empty)
379 return 0;
380 if (!nl_just_seen)
381 count++; /* no trailing newline */
382 return count;
383 }
384
385 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
386 {
387 if (!DIFF_FILE_VALID(one)) {
388 mf->ptr = (char *)""; /* does not matter */
389 mf->size = 0;
390 return 0;
391 }
392 else if (diff_populate_filespec(one, 0))
393 return -1;
394
395 mf->ptr = one->data;
396 mf->size = one->size;
397 return 0;
398 }
399
400 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
401 static unsigned long diff_filespec_size(struct diff_filespec *one)
402 {
403 if (!DIFF_FILE_VALID(one))
404 return 0;
405 diff_populate_filespec(one, CHECK_SIZE_ONLY);
406 return one->size;
407 }
408
409 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
410 {
411 char *ptr = mf->ptr;
412 long size = mf->size;
413 int cnt = 0;
414
415 if (!size)
416 return cnt;
417 ptr += size - 1; /* pointing at the very end */
418 if (*ptr != '\n')
419 ; /* incomplete line */
420 else
421 ptr--; /* skip the last LF */
422 while (mf->ptr < ptr) {
423 char *prev_eol;
424 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
425 if (*prev_eol == '\n')
426 break;
427 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
428 break;
429 cnt++;
430 ptr = prev_eol - 1;
431 }
432 return cnt;
433 }
434
435 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
436 struct emit_callback *ecbdata)
437 {
438 int l1, l2, at;
439 unsigned ws_rule = ecbdata->ws_rule;
440 l1 = count_trailing_blank(mf1, ws_rule);
441 l2 = count_trailing_blank(mf2, ws_rule);
442 if (l2 <= l1) {
443 ecbdata->blank_at_eof_in_preimage = 0;
444 ecbdata->blank_at_eof_in_postimage = 0;
445 return;
446 }
447 at = count_lines(mf1->ptr, mf1->size);
448 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
449
450 at = count_lines(mf2->ptr, mf2->size);
451 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
452 }
453
454 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
455 int first, const char *line, int len)
456 {
457 int has_trailing_newline, has_trailing_carriage_return;
458 int nofirst;
459 FILE *file = o->file;
460
461 fputs(diff_line_prefix(o), file);
462
463 if (len == 0) {
464 has_trailing_newline = (first == '\n');
465 has_trailing_carriage_return = (!has_trailing_newline &&
466 (first == '\r'));
467 nofirst = has_trailing_newline || has_trailing_carriage_return;
468 } else {
469 has_trailing_newline = (len > 0 && line[len-1] == '\n');
470 if (has_trailing_newline)
471 len--;
472 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
473 if (has_trailing_carriage_return)
474 len--;
475 nofirst = 0;
476 }
477
478 if (len || !nofirst) {
479 fputs(set, file);
480 if (!nofirst)
481 fputc(first, file);
482 fwrite(line, len, 1, file);
483 fputs(reset, file);
484 }
485 if (has_trailing_carriage_return)
486 fputc('\r', file);
487 if (has_trailing_newline)
488 fputc('\n', file);
489 }
490
491 static void emit_line(struct diff_options *o, const char *set, const char *reset,
492 const char *line, int len)
493 {
494 emit_line_0(o, set, reset, line[0], line+1, len-1);
495 }
496
497 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
498 {
499 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
500 ecbdata->blank_at_eof_in_preimage &&
501 ecbdata->blank_at_eof_in_postimage &&
502 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
503 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
504 return 0;
505 return ws_blank_line(line, len, ecbdata->ws_rule);
506 }
507
508 static void emit_line_checked(const char *reset,
509 struct emit_callback *ecbdata,
510 const char *line, int len,
511 enum color_diff color,
512 unsigned ws_error_highlight,
513 char sign)
514 {
515 const char *set = diff_get_color(ecbdata->color_diff, color);
516 const char *ws = NULL;
517
518 if (ecbdata->opt->ws_error_highlight & ws_error_highlight) {
519 ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
520 if (!*ws)
521 ws = NULL;
522 }
523
524 if (!ws)
525 emit_line_0(ecbdata->opt, set, reset, sign, line, len);
526 else if (sign == '+' && new_blank_line_at_eof(ecbdata, line, len))
527 /* Blank line at EOF - paint '+' as well */
528 emit_line_0(ecbdata->opt, ws, reset, sign, line, len);
529 else {
530 /* Emit just the prefix, then the rest. */
531 emit_line_0(ecbdata->opt, set, reset, sign, "", 0);
532 ws_check_emit(line, len, ecbdata->ws_rule,
533 ecbdata->opt->file, set, reset, ws);
534 }
535 }
536
537 static void emit_add_line(const char *reset,
538 struct emit_callback *ecbdata,
539 const char *line, int len)
540 {
541 emit_line_checked(reset, ecbdata, line, len,
542 DIFF_FILE_NEW, WSEH_NEW, '+');
543 }
544
545 static void emit_del_line(const char *reset,
546 struct emit_callback *ecbdata,
547 const char *line, int len)
548 {
549 emit_line_checked(reset, ecbdata, line, len,
550 DIFF_FILE_OLD, WSEH_OLD, '-');
551 }
552
553 static void emit_context_line(const char *reset,
554 struct emit_callback *ecbdata,
555 const char *line, int len)
556 {
557 emit_line_checked(reset, ecbdata, line, len,
558 DIFF_CONTEXT, WSEH_CONTEXT, ' ');
559 }
560
561 static void emit_hunk_header(struct emit_callback *ecbdata,
562 const char *line, int len)
563 {
564 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
565 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
566 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
567 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
568 static const char atat[2] = { '@', '@' };
569 const char *cp, *ep;
570 struct strbuf msgbuf = STRBUF_INIT;
571 int org_len = len;
572 int i = 1;
573
574 /*
575 * As a hunk header must begin with "@@ -<old>, +<new> @@",
576 * it always is at least 10 bytes long.
577 */
578 if (len < 10 ||
579 memcmp(line, atat, 2) ||
580 !(ep = memmem(line + 2, len - 2, atat, 2))) {
581 emit_line(ecbdata->opt, context, reset, line, len);
582 return;
583 }
584 ep += 2; /* skip over @@ */
585
586 /* The hunk header in fraginfo color */
587 strbuf_addstr(&msgbuf, frag);
588 strbuf_add(&msgbuf, line, ep - line);
589 strbuf_addstr(&msgbuf, reset);
590
591 /*
592 * trailing "\r\n"
593 */
594 for ( ; i < 3; i++)
595 if (line[len - i] == '\r' || line[len - i] == '\n')
596 len--;
597
598 /* blank before the func header */
599 for (cp = ep; ep - line < len; ep++)
600 if (*ep != ' ' && *ep != '\t')
601 break;
602 if (ep != cp) {
603 strbuf_addstr(&msgbuf, context);
604 strbuf_add(&msgbuf, cp, ep - cp);
605 strbuf_addstr(&msgbuf, reset);
606 }
607
608 if (ep < line + len) {
609 strbuf_addstr(&msgbuf, func);
610 strbuf_add(&msgbuf, ep, line + len - ep);
611 strbuf_addstr(&msgbuf, reset);
612 }
613
614 strbuf_add(&msgbuf, line + len, org_len - len);
615 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
616 strbuf_release(&msgbuf);
617 }
618
619 static struct diff_tempfile *claim_diff_tempfile(void) {
620 int i;
621 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
622 if (!diff_temp[i].name)
623 return diff_temp + i;
624 die("BUG: diff is failing to clean up its tempfiles");
625 }
626
627 static void remove_tempfile(void)
628 {
629 int i;
630 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
631 if (is_tempfile_active(&diff_temp[i].tempfile))
632 delete_tempfile(&diff_temp[i].tempfile);
633 diff_temp[i].name = NULL;
634 }
635 }
636
637 static void print_line_count(FILE *file, int count)
638 {
639 switch (count) {
640 case 0:
641 fprintf(file, "0,0");
642 break;
643 case 1:
644 fprintf(file, "1");
645 break;
646 default:
647 fprintf(file, "1,%d", count);
648 break;
649 }
650 }
651
652 static void emit_rewrite_lines(struct emit_callback *ecb,
653 int prefix, const char *data, int size)
654 {
655 const char *endp = NULL;
656 static const char *nneof = " No newline at end of file\n";
657 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
658
659 while (0 < size) {
660 int len;
661
662 endp = memchr(data, '\n', size);
663 len = endp ? (endp - data + 1) : size;
664 if (prefix != '+') {
665 ecb->lno_in_preimage++;
666 emit_del_line(reset, ecb, data, len);
667 } else {
668 ecb->lno_in_postimage++;
669 emit_add_line(reset, ecb, data, len);
670 }
671 size -= len;
672 data += len;
673 }
674 if (!endp) {
675 const char *context = diff_get_color(ecb->color_diff,
676 DIFF_CONTEXT);
677 putc('\n', ecb->opt->file);
678 emit_line_0(ecb->opt, context, reset, '\\',
679 nneof, strlen(nneof));
680 }
681 }
682
683 static void emit_rewrite_diff(const char *name_a,
684 const char *name_b,
685 struct diff_filespec *one,
686 struct diff_filespec *two,
687 struct userdiff_driver *textconv_one,
688 struct userdiff_driver *textconv_two,
689 struct diff_options *o)
690 {
691 int lc_a, lc_b;
692 const char *name_a_tab, *name_b_tab;
693 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
694 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
695 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
696 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
697 const char *a_prefix, *b_prefix;
698 char *data_one, *data_two;
699 size_t size_one, size_two;
700 struct emit_callback ecbdata;
701 const char *line_prefix = diff_line_prefix(o);
702
703 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
704 a_prefix = o->b_prefix;
705 b_prefix = o->a_prefix;
706 } else {
707 a_prefix = o->a_prefix;
708 b_prefix = o->b_prefix;
709 }
710
711 name_a += (*name_a == '/');
712 name_b += (*name_b == '/');
713 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
714 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
715
716 strbuf_reset(&a_name);
717 strbuf_reset(&b_name);
718 quote_two_c_style(&a_name, a_prefix, name_a, 0);
719 quote_two_c_style(&b_name, b_prefix, name_b, 0);
720
721 size_one = fill_textconv(textconv_one, one, &data_one);
722 size_two = fill_textconv(textconv_two, two, &data_two);
723
724 memset(&ecbdata, 0, sizeof(ecbdata));
725 ecbdata.color_diff = want_color(o->use_color);
726 ecbdata.found_changesp = &o->found_changes;
727 ecbdata.ws_rule = whitespace_rule(name_b);
728 ecbdata.opt = o;
729 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
730 mmfile_t mf1, mf2;
731 mf1.ptr = (char *)data_one;
732 mf2.ptr = (char *)data_two;
733 mf1.size = size_one;
734 mf2.size = size_two;
735 check_blank_at_eof(&mf1, &mf2, &ecbdata);
736 }
737 ecbdata.lno_in_preimage = 1;
738 ecbdata.lno_in_postimage = 1;
739
740 lc_a = count_lines(data_one, size_one);
741 lc_b = count_lines(data_two, size_two);
742 fprintf(o->file,
743 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
744 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
745 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
746 line_prefix, fraginfo);
747 if (!o->irreversible_delete)
748 print_line_count(o->file, lc_a);
749 else
750 fprintf(o->file, "?,?");
751 fprintf(o->file, " +");
752 print_line_count(o->file, lc_b);
753 fprintf(o->file, " @@%s\n", reset);
754 if (lc_a && !o->irreversible_delete)
755 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
756 if (lc_b)
757 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
758 if (textconv_one)
759 free((char *)data_one);
760 if (textconv_two)
761 free((char *)data_two);
762 }
763
764 struct diff_words_buffer {
765 mmfile_t text;
766 long alloc;
767 struct diff_words_orig {
768 const char *begin, *end;
769 } *orig;
770 int orig_nr, orig_alloc;
771 };
772
773 static void diff_words_append(char *line, unsigned long len,
774 struct diff_words_buffer *buffer)
775 {
776 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
777 line++;
778 len--;
779 memcpy(buffer->text.ptr + buffer->text.size, line, len);
780 buffer->text.size += len;
781 buffer->text.ptr[buffer->text.size] = '\0';
782 }
783
784 struct diff_words_style_elem {
785 const char *prefix;
786 const char *suffix;
787 const char *color; /* NULL; filled in by the setup code if
788 * color is enabled */
789 };
790
791 struct diff_words_style {
792 enum diff_words_type type;
793 struct diff_words_style_elem new, old, ctx;
794 const char *newline;
795 };
796
797 static struct diff_words_style diff_words_styles[] = {
798 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
799 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
800 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
801 };
802
803 struct diff_words_data {
804 struct diff_words_buffer minus, plus;
805 const char *current_plus;
806 int last_minus;
807 struct diff_options *opt;
808 regex_t *word_regex;
809 enum diff_words_type type;
810 struct diff_words_style *style;
811 };
812
813 static int fn_out_diff_words_write_helper(FILE *fp,
814 struct diff_words_style_elem *st_el,
815 const char *newline,
816 size_t count, const char *buf,
817 const char *line_prefix)
818 {
819 int print = 0;
820
821 while (count) {
822 char *p = memchr(buf, '\n', count);
823 if (print)
824 fputs(line_prefix, fp);
825 if (p != buf) {
826 if (st_el->color && fputs(st_el->color, fp) < 0)
827 return -1;
828 if (fputs(st_el->prefix, fp) < 0 ||
829 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
830 fputs(st_el->suffix, fp) < 0)
831 return -1;
832 if (st_el->color && *st_el->color
833 && fputs(GIT_COLOR_RESET, fp) < 0)
834 return -1;
835 }
836 if (!p)
837 return 0;
838 if (fputs(newline, fp) < 0)
839 return -1;
840 count -= p + 1 - buf;
841 buf = p + 1;
842 print = 1;
843 }
844 return 0;
845 }
846
847 /*
848 * '--color-words' algorithm can be described as:
849 *
850 * 1. collect a the minus/plus lines of a diff hunk, divided into
851 * minus-lines and plus-lines;
852 *
853 * 2. break both minus-lines and plus-lines into words and
854 * place them into two mmfile_t with one word for each line;
855 *
856 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
857 *
858 * And for the common parts of the both file, we output the plus side text.
859 * diff_words->current_plus is used to trace the current position of the plus file
860 * which printed. diff_words->last_minus is used to trace the last minus word
861 * printed.
862 *
863 * For '--graph' to work with '--color-words', we need to output the graph prefix
864 * on each line of color words output. Generally, there are two conditions on
865 * which we should output the prefix.
866 *
867 * 1. diff_words->last_minus == 0 &&
868 * diff_words->current_plus == diff_words->plus.text.ptr
869 *
870 * that is: the plus text must start as a new line, and if there is no minus
871 * word printed, a graph prefix must be printed.
872 *
873 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
874 * *(diff_words->current_plus - 1) == '\n'
875 *
876 * that is: a graph prefix must be printed following a '\n'
877 */
878 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
879 {
880 if ((diff_words->last_minus == 0 &&
881 diff_words->current_plus == diff_words->plus.text.ptr) ||
882 (diff_words->current_plus > diff_words->plus.text.ptr &&
883 *(diff_words->current_plus - 1) == '\n')) {
884 return 1;
885 } else {
886 return 0;
887 }
888 }
889
890 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
891 {
892 struct diff_words_data *diff_words = priv;
893 struct diff_words_style *style = diff_words->style;
894 int minus_first, minus_len, plus_first, plus_len;
895 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
896 struct diff_options *opt = diff_words->opt;
897 const char *line_prefix;
898
899 if (line[0] != '@' || parse_hunk_header(line, len,
900 &minus_first, &minus_len, &plus_first, &plus_len))
901 return;
902
903 assert(opt);
904 line_prefix = diff_line_prefix(opt);
905
906 /* POSIX requires that first be decremented by one if len == 0... */
907 if (minus_len) {
908 minus_begin = diff_words->minus.orig[minus_first].begin;
909 minus_end =
910 diff_words->minus.orig[minus_first + minus_len - 1].end;
911 } else
912 minus_begin = minus_end =
913 diff_words->minus.orig[minus_first].end;
914
915 if (plus_len) {
916 plus_begin = diff_words->plus.orig[plus_first].begin;
917 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
918 } else
919 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
920
921 if (color_words_output_graph_prefix(diff_words)) {
922 fputs(line_prefix, diff_words->opt->file);
923 }
924 if (diff_words->current_plus != plus_begin) {
925 fn_out_diff_words_write_helper(diff_words->opt->file,
926 &style->ctx, style->newline,
927 plus_begin - diff_words->current_plus,
928 diff_words->current_plus, line_prefix);
929 if (*(plus_begin - 1) == '\n')
930 fputs(line_prefix, diff_words->opt->file);
931 }
932 if (minus_begin != minus_end) {
933 fn_out_diff_words_write_helper(diff_words->opt->file,
934 &style->old, style->newline,
935 minus_end - minus_begin, minus_begin,
936 line_prefix);
937 }
938 if (plus_begin != plus_end) {
939 fn_out_diff_words_write_helper(diff_words->opt->file,
940 &style->new, style->newline,
941 plus_end - plus_begin, plus_begin,
942 line_prefix);
943 }
944
945 diff_words->current_plus = plus_end;
946 diff_words->last_minus = minus_first;
947 }
948
949 /* This function starts looking at *begin, and returns 0 iff a word was found. */
950 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
951 int *begin, int *end)
952 {
953 if (word_regex && *begin < buffer->size) {
954 regmatch_t match[1];
955 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
956 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
957 '\n', match[0].rm_eo - match[0].rm_so);
958 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
959 *begin += match[0].rm_so;
960 return *begin >= *end;
961 }
962 return -1;
963 }
964
965 /* find the next word */
966 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
967 (*begin)++;
968 if (*begin >= buffer->size)
969 return -1;
970
971 /* find the end of the word */
972 *end = *begin + 1;
973 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
974 (*end)++;
975
976 return 0;
977 }
978
979 /*
980 * This function splits the words in buffer->text, stores the list with
981 * newline separator into out, and saves the offsets of the original words
982 * in buffer->orig.
983 */
984 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
985 regex_t *word_regex)
986 {
987 int i, j;
988 long alloc = 0;
989
990 out->size = 0;
991 out->ptr = NULL;
992
993 /* fake an empty "0th" word */
994 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
995 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
996 buffer->orig_nr = 1;
997
998 for (i = 0; i < buffer->text.size; i++) {
999 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1000 return;
1001
1002 /* store original boundaries */
1003 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1004 buffer->orig_alloc);
1005 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1006 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1007 buffer->orig_nr++;
1008
1009 /* store one word */
1010 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1011 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1012 out->ptr[out->size + j - i] = '\n';
1013 out->size += j - i + 1;
1014
1015 i = j - 1;
1016 }
1017 }
1018
1019 /* this executes the word diff on the accumulated buffers */
1020 static void diff_words_show(struct diff_words_data *diff_words)
1021 {
1022 xpparam_t xpp;
1023 xdemitconf_t xecfg;
1024 mmfile_t minus, plus;
1025 struct diff_words_style *style = diff_words->style;
1026
1027 struct diff_options *opt = diff_words->opt;
1028 const char *line_prefix;
1029
1030 assert(opt);
1031 line_prefix = diff_line_prefix(opt);
1032
1033 /* special case: only removal */
1034 if (!diff_words->plus.text.size) {
1035 fputs(line_prefix, diff_words->opt->file);
1036 fn_out_diff_words_write_helper(diff_words->opt->file,
1037 &style->old, style->newline,
1038 diff_words->minus.text.size,
1039 diff_words->minus.text.ptr, line_prefix);
1040 diff_words->minus.text.size = 0;
1041 return;
1042 }
1043
1044 diff_words->current_plus = diff_words->plus.text.ptr;
1045 diff_words->last_minus = 0;
1046
1047 memset(&xpp, 0, sizeof(xpp));
1048 memset(&xecfg, 0, sizeof(xecfg));
1049 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1050 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1051 xpp.flags = 0;
1052 /* as only the hunk header will be parsed, we need a 0-context */
1053 xecfg.ctxlen = 0;
1054 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1055 &xpp, &xecfg))
1056 die("unable to generate word diff");
1057 free(minus.ptr);
1058 free(plus.ptr);
1059 if (diff_words->current_plus != diff_words->plus.text.ptr +
1060 diff_words->plus.text.size) {
1061 if (color_words_output_graph_prefix(diff_words))
1062 fputs(line_prefix, diff_words->opt->file);
1063 fn_out_diff_words_write_helper(diff_words->opt->file,
1064 &style->ctx, style->newline,
1065 diff_words->plus.text.ptr + diff_words->plus.text.size
1066 - diff_words->current_plus, diff_words->current_plus,
1067 line_prefix);
1068 }
1069 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1070 }
1071
1072 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1073 static void diff_words_flush(struct emit_callback *ecbdata)
1074 {
1075 if (ecbdata->diff_words->minus.text.size ||
1076 ecbdata->diff_words->plus.text.size)
1077 diff_words_show(ecbdata->diff_words);
1078 }
1079
1080 static void diff_filespec_load_driver(struct diff_filespec *one)
1081 {
1082 /* Use already-loaded driver */
1083 if (one->driver)
1084 return;
1085
1086 if (S_ISREG(one->mode))
1087 one->driver = userdiff_find_by_path(one->path);
1088
1089 /* Fallback to default settings */
1090 if (!one->driver)
1091 one->driver = userdiff_find_by_name("default");
1092 }
1093
1094 static const char *userdiff_word_regex(struct diff_filespec *one)
1095 {
1096 diff_filespec_load_driver(one);
1097 return one->driver->word_regex;
1098 }
1099
1100 static void init_diff_words_data(struct emit_callback *ecbdata,
1101 struct diff_options *orig_opts,
1102 struct diff_filespec *one,
1103 struct diff_filespec *two)
1104 {
1105 int i;
1106 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1107 memcpy(o, orig_opts, sizeof(struct diff_options));
1108
1109 ecbdata->diff_words =
1110 xcalloc(1, sizeof(struct diff_words_data));
1111 ecbdata->diff_words->type = o->word_diff;
1112 ecbdata->diff_words->opt = o;
1113 if (!o->word_regex)
1114 o->word_regex = userdiff_word_regex(one);
1115 if (!o->word_regex)
1116 o->word_regex = userdiff_word_regex(two);
1117 if (!o->word_regex)
1118 o->word_regex = diff_word_regex_cfg;
1119 if (o->word_regex) {
1120 ecbdata->diff_words->word_regex = (regex_t *)
1121 xmalloc(sizeof(regex_t));
1122 if (regcomp(ecbdata->diff_words->word_regex,
1123 o->word_regex,
1124 REG_EXTENDED | REG_NEWLINE))
1125 die ("Invalid regular expression: %s",
1126 o->word_regex);
1127 }
1128 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1129 if (o->word_diff == diff_words_styles[i].type) {
1130 ecbdata->diff_words->style =
1131 &diff_words_styles[i];
1132 break;
1133 }
1134 }
1135 if (want_color(o->use_color)) {
1136 struct diff_words_style *st = ecbdata->diff_words->style;
1137 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1138 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1139 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1140 }
1141 }
1142
1143 static void free_diff_words_data(struct emit_callback *ecbdata)
1144 {
1145 if (ecbdata->diff_words) {
1146 diff_words_flush(ecbdata);
1147 free (ecbdata->diff_words->opt);
1148 free (ecbdata->diff_words->minus.text.ptr);
1149 free (ecbdata->diff_words->minus.orig);
1150 free (ecbdata->diff_words->plus.text.ptr);
1151 free (ecbdata->diff_words->plus.orig);
1152 if (ecbdata->diff_words->word_regex) {
1153 regfree(ecbdata->diff_words->word_regex);
1154 free(ecbdata->diff_words->word_regex);
1155 }
1156 free(ecbdata->diff_words);
1157 ecbdata->diff_words = NULL;
1158 }
1159 }
1160
1161 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1162 {
1163 if (want_color(diff_use_color))
1164 return diff_colors[ix];
1165 return "";
1166 }
1167
1168 const char *diff_line_prefix(struct diff_options *opt)
1169 {
1170 struct strbuf *msgbuf;
1171 if (!opt->output_prefix)
1172 return "";
1173
1174 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1175 return msgbuf->buf;
1176 }
1177
1178 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1179 {
1180 const char *cp;
1181 unsigned long allot;
1182 size_t l = len;
1183
1184 if (ecb->truncate)
1185 return ecb->truncate(line, len);
1186 cp = line;
1187 allot = l;
1188 while (0 < l) {
1189 (void) utf8_width(&cp, &l);
1190 if (!cp)
1191 break; /* truncated in the middle? */
1192 }
1193 return allot - l;
1194 }
1195
1196 static void find_lno(const char *line, struct emit_callback *ecbdata)
1197 {
1198 const char *p;
1199 ecbdata->lno_in_preimage = 0;
1200 ecbdata->lno_in_postimage = 0;
1201 p = strchr(line, '-');
1202 if (!p)
1203 return; /* cannot happen */
1204 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1205 p = strchr(p, '+');
1206 if (!p)
1207 return; /* cannot happen */
1208 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1209 }
1210
1211 static void fn_out_consume(void *priv, char *line, unsigned long len)
1212 {
1213 struct emit_callback *ecbdata = priv;
1214 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1215 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1216 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1217 struct diff_options *o = ecbdata->opt;
1218 const char *line_prefix = diff_line_prefix(o);
1219
1220 if (ecbdata->header) {
1221 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1222 strbuf_reset(ecbdata->header);
1223 ecbdata->header = NULL;
1224 }
1225 *(ecbdata->found_changesp) = 1;
1226
1227 if (ecbdata->label_path[0]) {
1228 const char *name_a_tab, *name_b_tab;
1229
1230 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1231 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1232
1233 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1234 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1235 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1236 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1237 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1238 }
1239
1240 if (diff_suppress_blank_empty
1241 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1242 line[0] = '\n';
1243 len = 1;
1244 }
1245
1246 if (line[0] == '@') {
1247 if (ecbdata->diff_words)
1248 diff_words_flush(ecbdata);
1249 len = sane_truncate_line(ecbdata, line, len);
1250 find_lno(line, ecbdata);
1251 emit_hunk_header(ecbdata, line, len);
1252 if (line[len-1] != '\n')
1253 putc('\n', ecbdata->opt->file);
1254 return;
1255 }
1256
1257 if (len < 1) {
1258 emit_line(ecbdata->opt, reset, reset, line, len);
1259 if (ecbdata->diff_words
1260 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1261 fputs("~\n", ecbdata->opt->file);
1262 return;
1263 }
1264
1265 if (ecbdata->diff_words) {
1266 if (line[0] == '-') {
1267 diff_words_append(line, len,
1268 &ecbdata->diff_words->minus);
1269 return;
1270 } else if (line[0] == '+') {
1271 diff_words_append(line, len,
1272 &ecbdata->diff_words->plus);
1273 return;
1274 } else if (starts_with(line, "\\ ")) {
1275 /*
1276 * Eat the "no newline at eof" marker as if we
1277 * saw a "+" or "-" line with nothing on it,
1278 * and return without diff_words_flush() to
1279 * defer processing. If this is the end of
1280 * preimage, more "+" lines may come after it.
1281 */
1282 return;
1283 }
1284 diff_words_flush(ecbdata);
1285 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1286 emit_line(ecbdata->opt, context, reset, line, len);
1287 fputs("~\n", ecbdata->opt->file);
1288 } else {
1289 /*
1290 * Skip the prefix character, if any. With
1291 * diff_suppress_blank_empty, there may be
1292 * none.
1293 */
1294 if (line[0] != '\n') {
1295 line++;
1296 len--;
1297 }
1298 emit_line(ecbdata->opt, context, reset, line, len);
1299 }
1300 return;
1301 }
1302
1303 switch (line[0]) {
1304 case '+':
1305 ecbdata->lno_in_postimage++;
1306 emit_add_line(reset, ecbdata, line + 1, len - 1);
1307 break;
1308 case '-':
1309 ecbdata->lno_in_preimage++;
1310 emit_del_line(reset, ecbdata, line + 1, len - 1);
1311 break;
1312 case ' ':
1313 ecbdata->lno_in_postimage++;
1314 ecbdata->lno_in_preimage++;
1315 emit_context_line(reset, ecbdata, line + 1, len - 1);
1316 break;
1317 default:
1318 /* incomplete line at the end */
1319 ecbdata->lno_in_preimage++;
1320 emit_line(ecbdata->opt,
1321 diff_get_color(ecbdata->color_diff, DIFF_CONTEXT),
1322 reset, line, len);
1323 break;
1324 }
1325 }
1326
1327 static char *pprint_rename(const char *a, const char *b)
1328 {
1329 const char *old = a;
1330 const char *new = b;
1331 struct strbuf name = STRBUF_INIT;
1332 int pfx_length, sfx_length;
1333 int pfx_adjust_for_slash;
1334 int len_a = strlen(a);
1335 int len_b = strlen(b);
1336 int a_midlen, b_midlen;
1337 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1338 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1339
1340 if (qlen_a || qlen_b) {
1341 quote_c_style(a, &name, NULL, 0);
1342 strbuf_addstr(&name, " => ");
1343 quote_c_style(b, &name, NULL, 0);
1344 return strbuf_detach(&name, NULL);
1345 }
1346
1347 /* Find common prefix */
1348 pfx_length = 0;
1349 while (*old && *new && *old == *new) {
1350 if (*old == '/')
1351 pfx_length = old - a + 1;
1352 old++;
1353 new++;
1354 }
1355
1356 /* Find common suffix */
1357 old = a + len_a;
1358 new = b + len_b;
1359 sfx_length = 0;
1360 /*
1361 * If there is a common prefix, it must end in a slash. In
1362 * that case we let this loop run 1 into the prefix to see the
1363 * same slash.
1364 *
1365 * If there is no common prefix, we cannot do this as it would
1366 * underrun the input strings.
1367 */
1368 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1369 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1370 b + pfx_length - pfx_adjust_for_slash <= new &&
1371 *old == *new) {
1372 if (*old == '/')
1373 sfx_length = len_a - (old - a);
1374 old--;
1375 new--;
1376 }
1377
1378 /*
1379 * pfx{mid-a => mid-b}sfx
1380 * {pfx-a => pfx-b}sfx
1381 * pfx{sfx-a => sfx-b}
1382 * name-a => name-b
1383 */
1384 a_midlen = len_a - pfx_length - sfx_length;
1385 b_midlen = len_b - pfx_length - sfx_length;
1386 if (a_midlen < 0)
1387 a_midlen = 0;
1388 if (b_midlen < 0)
1389 b_midlen = 0;
1390
1391 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1392 if (pfx_length + sfx_length) {
1393 strbuf_add(&name, a, pfx_length);
1394 strbuf_addch(&name, '{');
1395 }
1396 strbuf_add(&name, a + pfx_length, a_midlen);
1397 strbuf_addstr(&name, " => ");
1398 strbuf_add(&name, b + pfx_length, b_midlen);
1399 if (pfx_length + sfx_length) {
1400 strbuf_addch(&name, '}');
1401 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1402 }
1403 return strbuf_detach(&name, NULL);
1404 }
1405
1406 struct diffstat_t {
1407 int nr;
1408 int alloc;
1409 struct diffstat_file {
1410 char *from_name;
1411 char *name;
1412 char *print_name;
1413 unsigned is_unmerged:1;
1414 unsigned is_binary:1;
1415 unsigned is_renamed:1;
1416 unsigned is_interesting:1;
1417 uintmax_t added, deleted;
1418 } **files;
1419 };
1420
1421 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1422 const char *name_a,
1423 const char *name_b)
1424 {
1425 struct diffstat_file *x;
1426 x = xcalloc(1, sizeof(*x));
1427 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1428 diffstat->files[diffstat->nr++] = x;
1429 if (name_b) {
1430 x->from_name = xstrdup(name_a);
1431 x->name = xstrdup(name_b);
1432 x->is_renamed = 1;
1433 }
1434 else {
1435 x->from_name = NULL;
1436 x->name = xstrdup(name_a);
1437 }
1438 return x;
1439 }
1440
1441 static void diffstat_consume(void *priv, char *line, unsigned long len)
1442 {
1443 struct diffstat_t *diffstat = priv;
1444 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1445
1446 if (line[0] == '+')
1447 x->added++;
1448 else if (line[0] == '-')
1449 x->deleted++;
1450 }
1451
1452 const char mime_boundary_leader[] = "------------";
1453
1454 static int scale_linear(int it, int width, int max_change)
1455 {
1456 if (!it)
1457 return 0;
1458 /*
1459 * make sure that at least one '-' or '+' is printed if
1460 * there is any change to this path. The easiest way is to
1461 * scale linearly as if the alloted width is one column shorter
1462 * than it is, and then add 1 to the result.
1463 */
1464 return 1 + (it * (width - 1) / max_change);
1465 }
1466
1467 static void show_name(FILE *file,
1468 const char *prefix, const char *name, int len)
1469 {
1470 fprintf(file, " %s%-*s |", prefix, len, name);
1471 }
1472
1473 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1474 {
1475 if (cnt <= 0)
1476 return;
1477 fprintf(file, "%s", set);
1478 while (cnt--)
1479 putc(ch, file);
1480 fprintf(file, "%s", reset);
1481 }
1482
1483 static void fill_print_name(struct diffstat_file *file)
1484 {
1485 char *pname;
1486
1487 if (file->print_name)
1488 return;
1489
1490 if (!file->is_renamed) {
1491 struct strbuf buf = STRBUF_INIT;
1492 if (quote_c_style(file->name, &buf, NULL, 0)) {
1493 pname = strbuf_detach(&buf, NULL);
1494 } else {
1495 pname = file->name;
1496 strbuf_release(&buf);
1497 }
1498 } else {
1499 pname = pprint_rename(file->from_name, file->name);
1500 }
1501 file->print_name = pname;
1502 }
1503
1504 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1505 {
1506 struct strbuf sb = STRBUF_INIT;
1507 int ret;
1508
1509 if (!files) {
1510 assert(insertions == 0 && deletions == 0);
1511 return fprintf(fp, "%s\n", " 0 files changed");
1512 }
1513
1514 strbuf_addf(&sb,
1515 (files == 1) ? " %d file changed" : " %d files changed",
1516 files);
1517
1518 /*
1519 * For binary diff, the caller may want to print "x files
1520 * changed" with insertions == 0 && deletions == 0.
1521 *
1522 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1523 * is probably less confusing (i.e skip over "2 files changed
1524 * but nothing about added/removed lines? Is this a bug in Git?").
1525 */
1526 if (insertions || deletions == 0) {
1527 strbuf_addf(&sb,
1528 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1529 insertions);
1530 }
1531
1532 if (deletions || insertions == 0) {
1533 strbuf_addf(&sb,
1534 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1535 deletions);
1536 }
1537 strbuf_addch(&sb, '\n');
1538 ret = fputs(sb.buf, fp);
1539 strbuf_release(&sb);
1540 return ret;
1541 }
1542
1543 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1544 {
1545 int i, len, add, del, adds = 0, dels = 0;
1546 uintmax_t max_change = 0, max_len = 0;
1547 int total_files = data->nr, count;
1548 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1549 const char *reset, *add_c, *del_c;
1550 const char *line_prefix = "";
1551 int extra_shown = 0;
1552
1553 if (data->nr == 0)
1554 return;
1555
1556 line_prefix = diff_line_prefix(options);
1557 count = options->stat_count ? options->stat_count : data->nr;
1558
1559 reset = diff_get_color_opt(options, DIFF_RESET);
1560 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1561 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1562
1563 /*
1564 * Find the longest filename and max number of changes
1565 */
1566 for (i = 0; (i < count) && (i < data->nr); i++) {
1567 struct diffstat_file *file = data->files[i];
1568 uintmax_t change = file->added + file->deleted;
1569
1570 if (!file->is_interesting && (change == 0)) {
1571 count++; /* not shown == room for one more */
1572 continue;
1573 }
1574 fill_print_name(file);
1575 len = strlen(file->print_name);
1576 if (max_len < len)
1577 max_len = len;
1578
1579 if (file->is_unmerged) {
1580 /* "Unmerged" is 8 characters */
1581 bin_width = bin_width < 8 ? 8 : bin_width;
1582 continue;
1583 }
1584 if (file->is_binary) {
1585 /* "Bin XXX -> YYY bytes" */
1586 int w = 14 + decimal_width(file->added)
1587 + decimal_width(file->deleted);
1588 bin_width = bin_width < w ? w : bin_width;
1589 /* Display change counts aligned with "Bin" */
1590 number_width = 3;
1591 continue;
1592 }
1593
1594 if (max_change < change)
1595 max_change = change;
1596 }
1597 count = i; /* where we can stop scanning in data->files[] */
1598
1599 /*
1600 * We have width = stat_width or term_columns() columns total.
1601 * We want a maximum of min(max_len, stat_name_width) for the name part.
1602 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1603 * We also need 1 for " " and 4 + decimal_width(max_change)
1604 * for " | NNNN " and one the empty column at the end, altogether
1605 * 6 + decimal_width(max_change).
1606 *
1607 * If there's not enough space, we will use the smaller of
1608 * stat_name_width (if set) and 5/8*width for the filename,
1609 * and the rest for constant elements + graph part, but no more
1610 * than stat_graph_width for the graph part.
1611 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1612 * for the standard terminal size).
1613 *
1614 * In other words: stat_width limits the maximum width, and
1615 * stat_name_width fixes the maximum width of the filename,
1616 * and is also used to divide available columns if there
1617 * aren't enough.
1618 *
1619 * Binary files are displayed with "Bin XXX -> YYY bytes"
1620 * instead of the change count and graph. This part is treated
1621 * similarly to the graph part, except that it is not
1622 * "scaled". If total width is too small to accommodate the
1623 * guaranteed minimum width of the filename part and the
1624 * separators and this message, this message will "overflow"
1625 * making the line longer than the maximum width.
1626 */
1627
1628 if (options->stat_width == -1)
1629 width = term_columns() - strlen(line_prefix);
1630 else
1631 width = options->stat_width ? options->stat_width : 80;
1632 number_width = decimal_width(max_change) > number_width ?
1633 decimal_width(max_change) : number_width;
1634
1635 if (options->stat_graph_width == -1)
1636 options->stat_graph_width = diff_stat_graph_width;
1637
1638 /*
1639 * Guarantee 3/8*16==6 for the graph part
1640 * and 5/8*16==10 for the filename part
1641 */
1642 if (width < 16 + 6 + number_width)
1643 width = 16 + 6 + number_width;
1644
1645 /*
1646 * First assign sizes that are wanted, ignoring available width.
1647 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1648 * starting from "XXX" should fit in graph_width.
1649 */
1650 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1651 if (options->stat_graph_width &&
1652 options->stat_graph_width < graph_width)
1653 graph_width = options->stat_graph_width;
1654
1655 name_width = (options->stat_name_width > 0 &&
1656 options->stat_name_width < max_len) ?
1657 options->stat_name_width : max_len;
1658
1659 /*
1660 * Adjust adjustable widths not to exceed maximum width
1661 */
1662 if (name_width + number_width + 6 + graph_width > width) {
1663 if (graph_width > width * 3/8 - number_width - 6) {
1664 graph_width = width * 3/8 - number_width - 6;
1665 if (graph_width < 6)
1666 graph_width = 6;
1667 }
1668
1669 if (options->stat_graph_width &&
1670 graph_width > options->stat_graph_width)
1671 graph_width = options->stat_graph_width;
1672 if (name_width > width - number_width - 6 - graph_width)
1673 name_width = width - number_width - 6 - graph_width;
1674 else
1675 graph_width = width - number_width - 6 - name_width;
1676 }
1677
1678 /*
1679 * From here name_width is the width of the name area,
1680 * and graph_width is the width of the graph area.
1681 * max_change is used to scale graph properly.
1682 */
1683 for (i = 0; i < count; i++) {
1684 const char *prefix = "";
1685 struct diffstat_file *file = data->files[i];
1686 char *name = file->print_name;
1687 uintmax_t added = file->added;
1688 uintmax_t deleted = file->deleted;
1689 int name_len;
1690
1691 if (!file->is_interesting && (added + deleted == 0))
1692 continue;
1693
1694 /*
1695 * "scale" the filename
1696 */
1697 len = name_width;
1698 name_len = strlen(name);
1699 if (name_width < name_len) {
1700 char *slash;
1701 prefix = "...";
1702 len -= 3;
1703 name += name_len - len;
1704 slash = strchr(name, '/');
1705 if (slash)
1706 name = slash;
1707 }
1708
1709 if (file->is_binary) {
1710 fprintf(options->file, "%s", line_prefix);
1711 show_name(options->file, prefix, name, len);
1712 fprintf(options->file, " %*s", number_width, "Bin");
1713 if (!added && !deleted) {
1714 putc('\n', options->file);
1715 continue;
1716 }
1717 fprintf(options->file, " %s%"PRIuMAX"%s",
1718 del_c, deleted, reset);
1719 fprintf(options->file, " -> ");
1720 fprintf(options->file, "%s%"PRIuMAX"%s",
1721 add_c, added, reset);
1722 fprintf(options->file, " bytes");
1723 fprintf(options->file, "\n");
1724 continue;
1725 }
1726 else if (file->is_unmerged) {
1727 fprintf(options->file, "%s", line_prefix);
1728 show_name(options->file, prefix, name, len);
1729 fprintf(options->file, " Unmerged\n");
1730 continue;
1731 }
1732
1733 /*
1734 * scale the add/delete
1735 */
1736 add = added;
1737 del = deleted;
1738
1739 if (graph_width <= max_change) {
1740 int total = scale_linear(add + del, graph_width, max_change);
1741 if (total < 2 && add && del)
1742 /* width >= 2 due to the sanity check */
1743 total = 2;
1744 if (add < del) {
1745 add = scale_linear(add, graph_width, max_change);
1746 del = total - add;
1747 } else {
1748 del = scale_linear(del, graph_width, max_change);
1749 add = total - del;
1750 }
1751 }
1752 fprintf(options->file, "%s", line_prefix);
1753 show_name(options->file, prefix, name, len);
1754 fprintf(options->file, " %*"PRIuMAX"%s",
1755 number_width, added + deleted,
1756 added + deleted ? " " : "");
1757 show_graph(options->file, '+', add, add_c, reset);
1758 show_graph(options->file, '-', del, del_c, reset);
1759 fprintf(options->file, "\n");
1760 }
1761
1762 for (i = 0; i < data->nr; i++) {
1763 struct diffstat_file *file = data->files[i];
1764 uintmax_t added = file->added;
1765 uintmax_t deleted = file->deleted;
1766
1767 if (file->is_unmerged ||
1768 (!file->is_interesting && (added + deleted == 0))) {
1769 total_files--;
1770 continue;
1771 }
1772
1773 if (!file->is_binary) {
1774 adds += added;
1775 dels += deleted;
1776 }
1777 if (i < count)
1778 continue;
1779 if (!extra_shown)
1780 fprintf(options->file, "%s ...\n", line_prefix);
1781 extra_shown = 1;
1782 }
1783 fprintf(options->file, "%s", line_prefix);
1784 print_stat_summary(options->file, total_files, adds, dels);
1785 }
1786
1787 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1788 {
1789 int i, adds = 0, dels = 0, total_files = data->nr;
1790
1791 if (data->nr == 0)
1792 return;
1793
1794 for (i = 0; i < data->nr; i++) {
1795 int added = data->files[i]->added;
1796 int deleted= data->files[i]->deleted;
1797
1798 if (data->files[i]->is_unmerged ||
1799 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1800 total_files--;
1801 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1802 adds += added;
1803 dels += deleted;
1804 }
1805 }
1806 fprintf(options->file, "%s", diff_line_prefix(options));
1807 print_stat_summary(options->file, total_files, adds, dels);
1808 }
1809
1810 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1811 {
1812 int i;
1813
1814 if (data->nr == 0)
1815 return;
1816
1817 for (i = 0; i < data->nr; i++) {
1818 struct diffstat_file *file = data->files[i];
1819
1820 fprintf(options->file, "%s", diff_line_prefix(options));
1821
1822 if (file->is_binary)
1823 fprintf(options->file, "-\t-\t");
1824 else
1825 fprintf(options->file,
1826 "%"PRIuMAX"\t%"PRIuMAX"\t",
1827 file->added, file->deleted);
1828 if (options->line_termination) {
1829 fill_print_name(file);
1830 if (!file->is_renamed)
1831 write_name_quoted(file->name, options->file,
1832 options->line_termination);
1833 else {
1834 fputs(file->print_name, options->file);
1835 putc(options->line_termination, options->file);
1836 }
1837 } else {
1838 if (file->is_renamed) {
1839 putc('\0', options->file);
1840 write_name_quoted(file->from_name, options->file, '\0');
1841 }
1842 write_name_quoted(file->name, options->file, '\0');
1843 }
1844 }
1845 }
1846
1847 struct dirstat_file {
1848 const char *name;
1849 unsigned long changed;
1850 };
1851
1852 struct dirstat_dir {
1853 struct dirstat_file *files;
1854 int alloc, nr, permille, cumulative;
1855 };
1856
1857 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1858 unsigned long changed, const char *base, int baselen)
1859 {
1860 unsigned long this_dir = 0;
1861 unsigned int sources = 0;
1862 const char *line_prefix = diff_line_prefix(opt);
1863
1864 while (dir->nr) {
1865 struct dirstat_file *f = dir->files;
1866 int namelen = strlen(f->name);
1867 unsigned long this;
1868 char *slash;
1869
1870 if (namelen < baselen)
1871 break;
1872 if (memcmp(f->name, base, baselen))
1873 break;
1874 slash = strchr(f->name + baselen, '/');
1875 if (slash) {
1876 int newbaselen = slash + 1 - f->name;
1877 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1878 sources++;
1879 } else {
1880 this = f->changed;
1881 dir->files++;
1882 dir->nr--;
1883 sources += 2;
1884 }
1885 this_dir += this;
1886 }
1887
1888 /*
1889 * We don't report dirstat's for
1890 * - the top level
1891 * - or cases where everything came from a single directory
1892 * under this directory (sources == 1).
1893 */
1894 if (baselen && sources != 1) {
1895 if (this_dir) {
1896 int permille = this_dir * 1000 / changed;
1897 if (permille >= dir->permille) {
1898 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1899 permille / 10, permille % 10, baselen, base);
1900 if (!dir->cumulative)
1901 return 0;
1902 }
1903 }
1904 }
1905 return this_dir;
1906 }
1907
1908 static int dirstat_compare(const void *_a, const void *_b)
1909 {
1910 const struct dirstat_file *a = _a;
1911 const struct dirstat_file *b = _b;
1912 return strcmp(a->name, b->name);
1913 }
1914
1915 static void show_dirstat(struct diff_options *options)
1916 {
1917 int i;
1918 unsigned long changed;
1919 struct dirstat_dir dir;
1920 struct diff_queue_struct *q = &diff_queued_diff;
1921
1922 dir.files = NULL;
1923 dir.alloc = 0;
1924 dir.nr = 0;
1925 dir.permille = options->dirstat_permille;
1926 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1927
1928 changed = 0;
1929 for (i = 0; i < q->nr; i++) {
1930 struct diff_filepair *p = q->queue[i];
1931 const char *name;
1932 unsigned long copied, added, damage;
1933 int content_changed;
1934
1935 name = p->two->path ? p->two->path : p->one->path;
1936
1937 if (p->one->oid_valid && p->two->oid_valid)
1938 content_changed = oidcmp(&p->one->oid, &p->two->oid);
1939 else
1940 content_changed = 1;
1941
1942 if (!content_changed) {
1943 /*
1944 * The SHA1 has not changed, so pre-/post-content is
1945 * identical. We can therefore skip looking at the
1946 * file contents altogether.
1947 */
1948 damage = 0;
1949 goto found_damage;
1950 }
1951
1952 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1953 /*
1954 * In --dirstat-by-file mode, we don't really need to
1955 * look at the actual file contents at all.
1956 * The fact that the SHA1 changed is enough for us to
1957 * add this file to the list of results
1958 * (with each file contributing equal damage).
1959 */
1960 damage = 1;
1961 goto found_damage;
1962 }
1963
1964 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1965 diff_populate_filespec(p->one, 0);
1966 diff_populate_filespec(p->two, 0);
1967 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1968 &copied, &added);
1969 diff_free_filespec_data(p->one);
1970 diff_free_filespec_data(p->two);
1971 } else if (DIFF_FILE_VALID(p->one)) {
1972 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
1973 copied = added = 0;
1974 diff_free_filespec_data(p->one);
1975 } else if (DIFF_FILE_VALID(p->two)) {
1976 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
1977 copied = 0;
1978 added = p->two->size;
1979 diff_free_filespec_data(p->two);
1980 } else
1981 continue;
1982
1983 /*
1984 * Original minus copied is the removed material,
1985 * added is the new material. They are both damages
1986 * made to the preimage.
1987 * If the resulting damage is zero, we know that
1988 * diffcore_count_changes() considers the two entries to
1989 * be identical, but since content_changed is true, we
1990 * know that there must have been _some_ kind of change,
1991 * so we force all entries to have damage > 0.
1992 */
1993 damage = (p->one->size - copied) + added;
1994 if (!damage)
1995 damage = 1;
1996
1997 found_damage:
1998 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1999 dir.files[dir.nr].name = name;
2000 dir.files[dir.nr].changed = damage;
2001 changed += damage;
2002 dir.nr++;
2003 }
2004
2005 /* This can happen even with many files, if everything was renames */
2006 if (!changed)
2007 return;
2008
2009 /* Show all directories with more than x% of the changes */
2010 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
2011 gather_dirstat(options, &dir, changed, "", 0);
2012 }
2013
2014 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2015 {
2016 int i;
2017 unsigned long changed;
2018 struct dirstat_dir dir;
2019
2020 if (data->nr == 0)
2021 return;
2022
2023 dir.files = NULL;
2024 dir.alloc = 0;
2025 dir.nr = 0;
2026 dir.permille = options->dirstat_permille;
2027 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2028
2029 changed = 0;
2030 for (i = 0; i < data->nr; i++) {
2031 struct diffstat_file *file = data->files[i];
2032 unsigned long damage = file->added + file->deleted;
2033 if (file->is_binary)
2034 /*
2035 * binary files counts bytes, not lines. Must find some
2036 * way to normalize binary bytes vs. textual lines.
2037 * The following heuristic assumes that there are 64
2038 * bytes per "line".
2039 * This is stupid and ugly, but very cheap...
2040 */
2041 damage = (damage + 63) / 64;
2042 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2043 dir.files[dir.nr].name = file->name;
2044 dir.files[dir.nr].changed = damage;
2045 changed += damage;
2046 dir.nr++;
2047 }
2048
2049 /* This can happen even with many files, if everything was renames */
2050 if (!changed)
2051 return;
2052
2053 /* Show all directories with more than x% of the changes */
2054 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
2055 gather_dirstat(options, &dir, changed, "", 0);
2056 }
2057
2058 static void free_diffstat_info(struct diffstat_t *diffstat)
2059 {
2060 int i;
2061 for (i = 0; i < diffstat->nr; i++) {
2062 struct diffstat_file *f = diffstat->files[i];
2063 if (f->name != f->print_name)
2064 free(f->print_name);
2065 free(f->name);
2066 free(f->from_name);
2067 free(f);
2068 }
2069 free(diffstat->files);
2070 }
2071
2072 struct checkdiff_t {
2073 const char *filename;
2074 int lineno;
2075 int conflict_marker_size;
2076 struct diff_options *o;
2077 unsigned ws_rule;
2078 unsigned status;
2079 };
2080
2081 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2082 {
2083 char firstchar;
2084 int cnt;
2085
2086 if (len < marker_size + 1)
2087 return 0;
2088 firstchar = line[0];
2089 switch (firstchar) {
2090 case '=': case '>': case '<': case '|':
2091 break;
2092 default:
2093 return 0;
2094 }
2095 for (cnt = 1; cnt < marker_size; cnt++)
2096 if (line[cnt] != firstchar)
2097 return 0;
2098 /* line[1] thru line[marker_size-1] are same as firstchar */
2099 if (len < marker_size + 1 || !isspace(line[marker_size]))
2100 return 0;
2101 return 1;
2102 }
2103
2104 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2105 {
2106 struct checkdiff_t *data = priv;
2107 int marker_size = data->conflict_marker_size;
2108 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2109 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2110 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2111 char *err;
2112 const char *line_prefix;
2113
2114 assert(data->o);
2115 line_prefix = diff_line_prefix(data->o);
2116
2117 if (line[0] == '+') {
2118 unsigned bad;
2119 data->lineno++;
2120 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2121 data->status |= 1;
2122 fprintf(data->o->file,
2123 "%s%s:%d: leftover conflict marker\n",
2124 line_prefix, data->filename, data->lineno);
2125 }
2126 bad = ws_check(line + 1, len - 1, data->ws_rule);
2127 if (!bad)
2128 return;
2129 data->status |= bad;
2130 err = whitespace_error_string(bad);
2131 fprintf(data->o->file, "%s%s:%d: %s.\n",
2132 line_prefix, data->filename, data->lineno, err);
2133 free(err);
2134 emit_line(data->o, set, reset, line, 1);
2135 ws_check_emit(line + 1, len - 1, data->ws_rule,
2136 data->o->file, set, reset, ws);
2137 } else if (line[0] == ' ') {
2138 data->lineno++;
2139 } else if (line[0] == '@') {
2140 char *plus = strchr(line, '+');
2141 if (plus)
2142 data->lineno = strtol(plus, NULL, 10) - 1;
2143 else
2144 die("invalid diff");
2145 }
2146 }
2147
2148 static unsigned char *deflate_it(char *data,
2149 unsigned long size,
2150 unsigned long *result_size)
2151 {
2152 int bound;
2153 unsigned char *deflated;
2154 git_zstream stream;
2155
2156 git_deflate_init(&stream, zlib_compression_level);
2157 bound = git_deflate_bound(&stream, size);
2158 deflated = xmalloc(bound);
2159 stream.next_out = deflated;
2160 stream.avail_out = bound;
2161
2162 stream.next_in = (unsigned char *)data;
2163 stream.avail_in = size;
2164 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2165 ; /* nothing */
2166 git_deflate_end(&stream);
2167 *result_size = stream.total_out;
2168 return deflated;
2169 }
2170
2171 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2172 const char *prefix)
2173 {
2174 void *cp;
2175 void *delta;
2176 void *deflated;
2177 void *data;
2178 unsigned long orig_size;
2179 unsigned long delta_size;
2180 unsigned long deflate_size;
2181 unsigned long data_size;
2182
2183 /* We could do deflated delta, or we could do just deflated two,
2184 * whichever is smaller.
2185 */
2186 delta = NULL;
2187 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2188 if (one->size && two->size) {
2189 delta = diff_delta(one->ptr, one->size,
2190 two->ptr, two->size,
2191 &delta_size, deflate_size);
2192 if (delta) {
2193 void *to_free = delta;
2194 orig_size = delta_size;
2195 delta = deflate_it(delta, delta_size, &delta_size);
2196 free(to_free);
2197 }
2198 }
2199
2200 if (delta && delta_size < deflate_size) {
2201 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2202 free(deflated);
2203 data = delta;
2204 data_size = delta_size;
2205 }
2206 else {
2207 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2208 free(delta);
2209 data = deflated;
2210 data_size = deflate_size;
2211 }
2212
2213 /* emit data encoded in base85 */
2214 cp = data;
2215 while (data_size) {
2216 int bytes = (52 < data_size) ? 52 : data_size;
2217 char line[70];
2218 data_size -= bytes;
2219 if (bytes <= 26)
2220 line[0] = bytes + 'A' - 1;
2221 else
2222 line[0] = bytes - 26 + 'a' - 1;
2223 encode_85(line + 1, cp, bytes);
2224 cp = (char *) cp + bytes;
2225 fprintf(file, "%s", prefix);
2226 fputs(line, file);
2227 fputc('\n', file);
2228 }
2229 fprintf(file, "%s\n", prefix);
2230 free(data);
2231 }
2232
2233 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2234 const char *prefix)
2235 {
2236 fprintf(file, "%sGIT binary patch\n", prefix);
2237 emit_binary_diff_body(file, one, two, prefix);
2238 emit_binary_diff_body(file, two, one, prefix);
2239 }
2240
2241 int diff_filespec_is_binary(struct diff_filespec *one)
2242 {
2243 if (one->is_binary == -1) {
2244 diff_filespec_load_driver(one);
2245 if (one->driver->binary != -1)
2246 one->is_binary = one->driver->binary;
2247 else {
2248 if (!one->data && DIFF_FILE_VALID(one))
2249 diff_populate_filespec(one, CHECK_BINARY);
2250 if (one->is_binary == -1 && one->data)
2251 one->is_binary = buffer_is_binary(one->data,
2252 one->size);
2253 if (one->is_binary == -1)
2254 one->is_binary = 0;
2255 }
2256 }
2257 return one->is_binary;
2258 }
2259
2260 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2261 {
2262 diff_filespec_load_driver(one);
2263 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2264 }
2265
2266 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2267 {
2268 if (!options->a_prefix)
2269 options->a_prefix = a;
2270 if (!options->b_prefix)
2271 options->b_prefix = b;
2272 }
2273
2274 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2275 {
2276 if (!DIFF_FILE_VALID(one))
2277 return NULL;
2278
2279 diff_filespec_load_driver(one);
2280 return userdiff_get_textconv(one->driver);
2281 }
2282
2283 static void builtin_diff(const char *name_a,
2284 const char *name_b,
2285 struct diff_filespec *one,
2286 struct diff_filespec *two,
2287 const char *xfrm_msg,
2288 int must_show_header,
2289 struct diff_options *o,
2290 int complete_rewrite)
2291 {
2292 mmfile_t mf1, mf2;
2293 const char *lbl[2];
2294 char *a_one, *b_two;
2295 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2296 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2297 const char *a_prefix, *b_prefix;
2298 struct userdiff_driver *textconv_one = NULL;
2299 struct userdiff_driver *textconv_two = NULL;
2300 struct strbuf header = STRBUF_INIT;
2301 const char *line_prefix = diff_line_prefix(o);
2302
2303 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
2304 (!one->mode || S_ISGITLINK(one->mode)) &&
2305 (!two->mode || S_ISGITLINK(two->mode))) {
2306 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2307 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2308 show_submodule_summary(o->file, one->path ? one->path : two->path,
2309 line_prefix,
2310 one->oid.hash, two->oid.hash,
2311 two->dirty_submodule,
2312 meta, del, add, reset);
2313 return;
2314 }
2315
2316 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2317 textconv_one = get_textconv(one);
2318 textconv_two = get_textconv(two);
2319 }
2320
2321 diff_set_mnemonic_prefix(o, "a/", "b/");
2322 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2323 a_prefix = o->b_prefix;
2324 b_prefix = o->a_prefix;
2325 } else {
2326 a_prefix = o->a_prefix;
2327 b_prefix = o->b_prefix;
2328 }
2329
2330 /* Never use a non-valid filename anywhere if at all possible */
2331 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2332 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2333
2334 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2335 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2336 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2337 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2338 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2339 if (lbl[0][0] == '/') {
2340 /* /dev/null */
2341 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2342 if (xfrm_msg)
2343 strbuf_addstr(&header, xfrm_msg);
2344 must_show_header = 1;
2345 }
2346 else if (lbl[1][0] == '/') {
2347 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2348 if (xfrm_msg)
2349 strbuf_addstr(&header, xfrm_msg);
2350 must_show_header = 1;
2351 }
2352 else {
2353 if (one->mode != two->mode) {
2354 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2355 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2356 must_show_header = 1;
2357 }
2358 if (xfrm_msg)
2359 strbuf_addstr(&header, xfrm_msg);
2360
2361 /*
2362 * we do not run diff between different kind
2363 * of objects.
2364 */
2365 if ((one->mode ^ two->mode) & S_IFMT)
2366 goto free_ab_and_return;
2367 if (complete_rewrite &&
2368 (textconv_one || !diff_filespec_is_binary(one)) &&
2369 (textconv_two || !diff_filespec_is_binary(two))) {
2370 fprintf(o->file, "%s", header.buf);
2371 strbuf_reset(&header);
2372 emit_rewrite_diff(name_a, name_b, one, two,
2373 textconv_one, textconv_two, o);
2374 o->found_changes = 1;
2375 goto free_ab_and_return;
2376 }
2377 }
2378
2379 if (o->irreversible_delete && lbl[1][0] == '/') {
2380 fprintf(o->file, "%s", header.buf);
2381 strbuf_reset(&header);
2382 goto free_ab_and_return;
2383 } else if (!DIFF_OPT_TST(o, TEXT) &&
2384 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2385 (!textconv_two && diff_filespec_is_binary(two)) )) {
2386 if (!one->data && !two->data &&
2387 S_ISREG(one->mode) && S_ISREG(two->mode) &&
2388 !DIFF_OPT_TST(o, BINARY)) {
2389 if (!oidcmp(&one->oid, &two->oid)) {
2390 if (must_show_header)
2391 fprintf(o->file, "%s", header.buf);
2392 goto free_ab_and_return;
2393 }
2394 fprintf(o->file, "%s", header.buf);
2395 fprintf(o->file, "%sBinary files %s and %s differ\n",
2396 line_prefix, lbl[0], lbl[1]);
2397 goto free_ab_and_return;
2398 }
2399 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2400 die("unable to read files to diff");
2401 /* Quite common confusing case */
2402 if (mf1.size == mf2.size &&
2403 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2404 if (must_show_header)
2405 fprintf(o->file, "%s", header.buf);
2406 goto free_ab_and_return;
2407 }
2408 fprintf(o->file, "%s", header.buf);
2409 strbuf_reset(&header);
2410 if (DIFF_OPT_TST(o, BINARY))
2411 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2412 else
2413 fprintf(o->file, "%sBinary files %s and %s differ\n",
2414 line_prefix, lbl[0], lbl[1]);
2415 o->found_changes = 1;
2416 } else {
2417 /* Crazy xdl interfaces.. */
2418 const char *diffopts = getenv("GIT_DIFF_OPTS");
2419 const char *v;
2420 xpparam_t xpp;
2421 xdemitconf_t xecfg;
2422 struct emit_callback ecbdata;
2423 const struct userdiff_funcname *pe;
2424
2425 if (must_show_header) {
2426 fprintf(o->file, "%s", header.buf);
2427 strbuf_reset(&header);
2428 }
2429
2430 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2431 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2432
2433 pe = diff_funcname_pattern(one);
2434 if (!pe)
2435 pe = diff_funcname_pattern(two);
2436
2437 memset(&xpp, 0, sizeof(xpp));
2438 memset(&xecfg, 0, sizeof(xecfg));
2439 memset(&ecbdata, 0, sizeof(ecbdata));
2440 ecbdata.label_path = lbl;
2441 ecbdata.color_diff = want_color(o->use_color);
2442 ecbdata.found_changesp = &o->found_changes;
2443 ecbdata.ws_rule = whitespace_rule(name_b);
2444 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2445 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2446 ecbdata.opt = o;
2447 ecbdata.header = header.len ? &header : NULL;
2448 xpp.flags = o->xdl_opts;
2449 xecfg.ctxlen = o->context;
2450 xecfg.interhunkctxlen = o->interhunkcontext;
2451 xecfg.flags = XDL_EMIT_FUNCNAMES;
2452 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2453 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2454 if (pe)
2455 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2456 if (!diffopts)
2457 ;
2458 else if (skip_prefix(diffopts, "--unified=", &v))
2459 xecfg.ctxlen = strtoul(v, NULL, 10);
2460 else if (skip_prefix(diffopts, "-u", &v))
2461 xecfg.ctxlen = strtoul(v, NULL, 10);
2462 if (o->word_diff)
2463 init_diff_words_data(&ecbdata, o, one, two);
2464 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2465 &xpp, &xecfg))
2466 die("unable to generate diff for %s", one->path);
2467 if (o->word_diff)
2468 free_diff_words_data(&ecbdata);
2469 if (textconv_one)
2470 free(mf1.ptr);
2471 if (textconv_two)
2472 free(mf2.ptr);
2473 xdiff_clear_find_func(&xecfg);
2474 }
2475
2476 free_ab_and_return:
2477 strbuf_release(&header);
2478 diff_free_filespec_data(one);
2479 diff_free_filespec_data(two);
2480 free(a_one);
2481 free(b_two);
2482 return;
2483 }
2484
2485 static void builtin_diffstat(const char *name_a, const char *name_b,
2486 struct diff_filespec *one,
2487 struct diff_filespec *two,
2488 struct diffstat_t *diffstat,
2489 struct diff_options *o,
2490 struct diff_filepair *p)
2491 {
2492 mmfile_t mf1, mf2;
2493 struct diffstat_file *data;
2494 int same_contents;
2495 int complete_rewrite = 0;
2496
2497 if (!DIFF_PAIR_UNMERGED(p)) {
2498 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2499 complete_rewrite = 1;
2500 }
2501
2502 data = diffstat_add(diffstat, name_a, name_b);
2503 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2504
2505 if (!one || !two) {
2506 data->is_unmerged = 1;
2507 return;
2508 }
2509
2510 same_contents = !oidcmp(&one->oid, &two->oid);
2511
2512 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2513 data->is_binary = 1;
2514 if (same_contents) {
2515 data->added = 0;
2516 data->deleted = 0;
2517 } else {
2518 data->added = diff_filespec_size(two);
2519 data->deleted = diff_filespec_size(one);
2520 }
2521 }
2522
2523 else if (complete_rewrite) {
2524 diff_populate_filespec(one, 0);
2525 diff_populate_filespec(two, 0);
2526 data->deleted = count_lines(one->data, one->size);
2527 data->added = count_lines(two->data, two->size);
2528 }
2529
2530 else if (!same_contents) {
2531 /* Crazy xdl interfaces.. */
2532 xpparam_t xpp;
2533 xdemitconf_t xecfg;
2534
2535 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2536 die("unable to read files to diff");
2537
2538 memset(&xpp, 0, sizeof(xpp));
2539 memset(&xecfg, 0, sizeof(xecfg));
2540 xpp.flags = o->xdl_opts;
2541 xecfg.ctxlen = o->context;
2542 xecfg.interhunkctxlen = o->interhunkcontext;
2543 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2544 &xpp, &xecfg))
2545 die("unable to generate diffstat for %s", one->path);
2546 }
2547
2548 diff_free_filespec_data(one);
2549 diff_free_filespec_data(two);
2550 }
2551
2552 static void builtin_checkdiff(const char *name_a, const char *name_b,
2553 const char *attr_path,
2554 struct diff_filespec *one,
2555 struct diff_filespec *two,
2556 struct diff_options *o)
2557 {
2558 mmfile_t mf1, mf2;
2559 struct checkdiff_t data;
2560
2561 if (!two)
2562 return;
2563
2564 memset(&data, 0, sizeof(data));
2565 data.filename = name_b ? name_b : name_a;
2566 data.lineno = 0;
2567 data.o = o;
2568 data.ws_rule = whitespace_rule(attr_path);
2569 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2570
2571 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2572 die("unable to read files to diff");
2573
2574 /*
2575 * All the other codepaths check both sides, but not checking
2576 * the "old" side here is deliberate. We are checking the newly
2577 * introduced changes, and as long as the "new" side is text, we
2578 * can and should check what it introduces.
2579 */
2580 if (diff_filespec_is_binary(two))
2581 goto free_and_return;
2582 else {
2583 /* Crazy xdl interfaces.. */
2584 xpparam_t xpp;
2585 xdemitconf_t xecfg;
2586
2587 memset(&xpp, 0, sizeof(xpp));
2588 memset(&xecfg, 0, sizeof(xecfg));
2589 xecfg.ctxlen = 1; /* at least one context line */
2590 xpp.flags = 0;
2591 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2592 &xpp, &xecfg))
2593 die("unable to generate checkdiff for %s", one->path);
2594
2595 if (data.ws_rule & WS_BLANK_AT_EOF) {
2596 struct emit_callback ecbdata;
2597 int blank_at_eof;
2598
2599 ecbdata.ws_rule = data.ws_rule;
2600 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2601 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2602
2603 if (blank_at_eof) {
2604 static char *err;
2605 if (!err)
2606 err = whitespace_error_string(WS_BLANK_AT_EOF);
2607 fprintf(o->file, "%s:%d: %s.\n",
2608 data.filename, blank_at_eof, err);
2609 data.status = 1; /* report errors */
2610 }
2611 }
2612 }
2613 free_and_return:
2614 diff_free_filespec_data(one);
2615 diff_free_filespec_data(two);
2616 if (data.status)
2617 DIFF_OPT_SET(o, CHECK_FAILED);
2618 }
2619
2620 struct diff_filespec *alloc_filespec(const char *path)
2621 {
2622 struct diff_filespec *spec;
2623
2624 FLEXPTR_ALLOC_STR(spec, path, path);
2625 spec->count = 1;
2626 spec->is_binary = -1;
2627 return spec;
2628 }
2629
2630 void free_filespec(struct diff_filespec *spec)
2631 {
2632 if (!--spec->count) {
2633 diff_free_filespec_data(spec);
2634 free(spec);
2635 }
2636 }
2637
2638 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2639 int sha1_valid, unsigned short mode)
2640 {
2641 if (mode) {
2642 spec->mode = canon_mode(mode);
2643 hashcpy(spec->oid.hash, sha1);
2644 spec->oid_valid = sha1_valid;
2645 }
2646 }
2647
2648 /*
2649 * Given a name and sha1 pair, if the index tells us the file in
2650 * the work tree has that object contents, return true, so that
2651 * prepare_temp_file() does not have to inflate and extract.
2652 */
2653 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2654 {
2655 const struct cache_entry *ce;
2656 struct stat st;
2657 int pos, len;
2658
2659 /*
2660 * We do not read the cache ourselves here, because the
2661 * benchmark with my previous version that always reads cache
2662 * shows that it makes things worse for diff-tree comparing
2663 * two linux-2.6 kernel trees in an already checked out work
2664 * tree. This is because most diff-tree comparisons deal with
2665 * only a small number of files, while reading the cache is
2666 * expensive for a large project, and its cost outweighs the
2667 * savings we get by not inflating the object to a temporary
2668 * file. Practically, this code only helps when we are used
2669 * by diff-cache --cached, which does read the cache before
2670 * calling us.
2671 */
2672 if (!active_cache)
2673 return 0;
2674
2675 /* We want to avoid the working directory if our caller
2676 * doesn't need the data in a normal file, this system
2677 * is rather slow with its stat/open/mmap/close syscalls,
2678 * and the object is contained in a pack file. The pack
2679 * is probably already open and will be faster to obtain
2680 * the data through than the working directory. Loose
2681 * objects however would tend to be slower as they need
2682 * to be individually opened and inflated.
2683 */
2684 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2685 return 0;
2686
2687 /*
2688 * Similarly, if we'd have to convert the file contents anyway, that
2689 * makes the optimization not worthwhile.
2690 */
2691 if (!want_file && would_convert_to_git(name))
2692 return 0;
2693
2694 len = strlen(name);
2695 pos = cache_name_pos(name, len);
2696 if (pos < 0)
2697 return 0;
2698 ce = active_cache[pos];
2699
2700 /*
2701 * This is not the sha1 we are looking for, or
2702 * unreusable because it is not a regular file.
2703 */
2704 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2705 return 0;
2706
2707 /*
2708 * If ce is marked as "assume unchanged", there is no
2709 * guarantee that work tree matches what we are looking for.
2710 */
2711 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2712 return 0;
2713
2714 /*
2715 * If ce matches the file in the work tree, we can reuse it.
2716 */
2717 if (ce_uptodate(ce) ||
2718 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2719 return 1;
2720
2721 return 0;
2722 }
2723
2724 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2725 {
2726 struct strbuf buf = STRBUF_INIT;
2727 char *dirty = "";
2728
2729 /* Are we looking at the work tree? */
2730 if (s->dirty_submodule)
2731 dirty = "-dirty";
2732
2733 strbuf_addf(&buf, "Subproject commit %s%s\n",
2734 oid_to_hex(&s->oid), dirty);
2735 s->size = buf.len;
2736 if (size_only) {
2737 s->data = NULL;
2738 strbuf_release(&buf);
2739 } else {
2740 s->data = strbuf_detach(&buf, NULL);
2741 s->should_free = 1;
2742 }
2743 return 0;
2744 }
2745
2746 /*
2747 * While doing rename detection and pickaxe operation, we may need to
2748 * grab the data for the blob (or file) for our own in-core comparison.
2749 * diff_filespec has data and size fields for this purpose.
2750 */
2751 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
2752 {
2753 int size_only = flags & CHECK_SIZE_ONLY;
2754 int err = 0;
2755 /*
2756 * demote FAIL to WARN to allow inspecting the situation
2757 * instead of refusing.
2758 */
2759 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
2760 ? SAFE_CRLF_WARN
2761 : safe_crlf);
2762
2763 if (!DIFF_FILE_VALID(s))
2764 die("internal error: asking to populate invalid file.");
2765 if (S_ISDIR(s->mode))
2766 return -1;
2767
2768 if (s->data)
2769 return 0;
2770
2771 if (size_only && 0 < s->size)
2772 return 0;
2773
2774 if (S_ISGITLINK(s->mode))
2775 return diff_populate_gitlink(s, size_only);
2776
2777 if (!s->oid_valid ||
2778 reuse_worktree_file(s->path, s->oid.hash, 0)) {
2779 struct strbuf buf = STRBUF_INIT;
2780 struct stat st;
2781 int fd;
2782
2783 if (lstat(s->path, &st) < 0) {
2784 if (errno == ENOENT) {
2785 err_empty:
2786 err = -1;
2787 empty:
2788 s->data = (char *)"";
2789 s->size = 0;
2790 return err;
2791 }
2792 }
2793 s->size = xsize_t(st.st_size);
2794 if (!s->size)
2795 goto empty;
2796 if (S_ISLNK(st.st_mode)) {
2797 struct strbuf sb = STRBUF_INIT;
2798
2799 if (strbuf_readlink(&sb, s->path, s->size))
2800 goto err_empty;
2801 s->size = sb.len;
2802 s->data = strbuf_detach(&sb, NULL);
2803 s->should_free = 1;
2804 return 0;
2805 }
2806 if (size_only)
2807 return 0;
2808 if ((flags & CHECK_BINARY) &&
2809 s->size > big_file_threshold && s->is_binary == -1) {
2810 s->is_binary = 1;
2811 return 0;
2812 }
2813 fd = open(s->path, O_RDONLY);
2814 if (fd < 0)
2815 goto err_empty;
2816 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2817 close(fd);
2818 s->should_munmap = 1;
2819
2820 /*
2821 * Convert from working tree format to canonical git format
2822 */
2823 if (convert_to_git(s->path, s->data, s->size, &buf, crlf_warn)) {
2824 size_t size = 0;
2825 munmap(s->data, s->size);
2826 s->should_munmap = 0;
2827 s->data = strbuf_detach(&buf, &size);
2828 s->size = size;
2829 s->should_free = 1;
2830 }
2831 }
2832 else {
2833 enum object_type type;
2834 if (size_only || (flags & CHECK_BINARY)) {
2835 type = sha1_object_info(s->oid.hash, &s->size);
2836 if (type < 0)
2837 die("unable to read %s",
2838 oid_to_hex(&s->oid));
2839 if (size_only)
2840 return 0;
2841 if (s->size > big_file_threshold && s->is_binary == -1) {
2842 s->is_binary = 1;
2843 return 0;
2844 }
2845 }
2846 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
2847 if (!s->data)
2848 die("unable to read %s", oid_to_hex(&s->oid));
2849 s->should_free = 1;
2850 }
2851 return 0;
2852 }
2853
2854 void diff_free_filespec_blob(struct diff_filespec *s)
2855 {
2856 if (s->should_free)
2857 free(s->data);
2858 else if (s->should_munmap)
2859 munmap(s->data, s->size);
2860
2861 if (s->should_free || s->should_munmap) {
2862 s->should_free = s->should_munmap = 0;
2863 s->data = NULL;
2864 }
2865 }
2866
2867 void diff_free_filespec_data(struct diff_filespec *s)
2868 {
2869 diff_free_filespec_blob(s);
2870 free(s->cnt_data);
2871 s->cnt_data = NULL;
2872 }
2873
2874 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2875 void *blob,
2876 unsigned long size,
2877 const struct object_id *oid,
2878 int mode)
2879 {
2880 int fd;
2881 struct strbuf buf = STRBUF_INIT;
2882 struct strbuf template = STRBUF_INIT;
2883 char *path_dup = xstrdup(path);
2884 const char *base = basename(path_dup);
2885
2886 /* Generate "XXXXXX_basename.ext" */
2887 strbuf_addstr(&template, "XXXXXX_");
2888 strbuf_addstr(&template, base);
2889
2890 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
2891 if (fd < 0)
2892 die_errno("unable to create temp-file");
2893 if (convert_to_working_tree(path,
2894 (const char *)blob, (size_t)size, &buf)) {
2895 blob = buf.buf;
2896 size = buf.len;
2897 }
2898 if (write_in_full(fd, blob, size) != size)
2899 die_errno("unable to write temp-file");
2900 close_tempfile(&temp->tempfile);
2901 temp->name = get_tempfile_path(&temp->tempfile);
2902 oid_to_hex_r(temp->hex, oid);
2903 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
2904 strbuf_release(&buf);
2905 strbuf_release(&template);
2906 free(path_dup);
2907 }
2908
2909 static struct diff_tempfile *prepare_temp_file(const char *name,
2910 struct diff_filespec *one)
2911 {
2912 struct diff_tempfile *temp = claim_diff_tempfile();
2913
2914 if (!DIFF_FILE_VALID(one)) {
2915 not_a_valid_file:
2916 /* A '-' entry produces this for file-2, and
2917 * a '+' entry produces this for file-1.
2918 */
2919 temp->name = "/dev/null";
2920 xsnprintf(temp->hex, sizeof(temp->hex), ".");
2921 xsnprintf(temp->mode, sizeof(temp->mode), ".");
2922 return temp;
2923 }
2924
2925 if (!S_ISGITLINK(one->mode) &&
2926 (!one->oid_valid ||
2927 reuse_worktree_file(name, one->oid.hash, 1))) {
2928 struct stat st;
2929 if (lstat(name, &st) < 0) {
2930 if (errno == ENOENT)
2931 goto not_a_valid_file;
2932 die_errno("stat(%s)", name);
2933 }
2934 if (S_ISLNK(st.st_mode)) {
2935 struct strbuf sb = STRBUF_INIT;
2936 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2937 die_errno("readlink(%s)", name);
2938 prep_temp_blob(name, temp, sb.buf, sb.len,
2939 (one->oid_valid ?
2940 &one->oid : &null_oid),
2941 (one->oid_valid ?
2942 one->mode : S_IFLNK));
2943 strbuf_release(&sb);
2944 }
2945 else {
2946 /* we can borrow from the file in the work tree */
2947 temp->name = name;
2948 if (!one->oid_valid)
2949 sha1_to_hex_r(temp->hex, null_sha1);
2950 else
2951 sha1_to_hex_r(temp->hex, one->oid.hash);
2952 /* Even though we may sometimes borrow the
2953 * contents from the work tree, we always want
2954 * one->mode. mode is trustworthy even when
2955 * !(one->sha1_valid), as long as
2956 * DIFF_FILE_VALID(one).
2957 */
2958 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
2959 }
2960 return temp;
2961 }
2962 else {
2963 if (diff_populate_filespec(one, 0))
2964 die("cannot read data blob for %s", one->path);
2965 prep_temp_blob(name, temp, one->data, one->size,
2966 &one->oid, one->mode);
2967 }
2968 return temp;
2969 }
2970
2971 static void add_external_diff_name(struct argv_array *argv,
2972 const char *name,
2973 struct diff_filespec *df)
2974 {
2975 struct diff_tempfile *temp = prepare_temp_file(name, df);
2976 argv_array_push(argv, temp->name);
2977 argv_array_push(argv, temp->hex);
2978 argv_array_push(argv, temp->mode);
2979 }
2980
2981 /* An external diff command takes:
2982 *
2983 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2984 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2985 *
2986 */
2987 static void run_external_diff(const char *pgm,
2988 const char *name,
2989 const char *other,
2990 struct diff_filespec *one,
2991 struct diff_filespec *two,
2992 const char *xfrm_msg,
2993 int complete_rewrite,
2994 struct diff_options *o)
2995 {
2996 struct argv_array argv = ARGV_ARRAY_INIT;
2997 struct argv_array env = ARGV_ARRAY_INIT;
2998 struct diff_queue_struct *q = &diff_queued_diff;
2999
3000 argv_array_push(&argv, pgm);
3001 argv_array_push(&argv, name);
3002
3003 if (one && two) {
3004 add_external_diff_name(&argv, name, one);
3005 if (!other)
3006 add_external_diff_name(&argv, name, two);
3007 else {
3008 add_external_diff_name(&argv, other, two);
3009 argv_array_push(&argv, other);
3010 argv_array_push(&argv, xfrm_msg);
3011 }
3012 }
3013
3014 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3015 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3016
3017 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3018 die(_("external diff died, stopping at %s"), name);
3019
3020 remove_tempfile();
3021 argv_array_clear(&argv);
3022 argv_array_clear(&env);
3023 }
3024
3025 static int similarity_index(struct diff_filepair *p)
3026 {
3027 return p->score * 100 / MAX_SCORE;
3028 }
3029
3030 static void fill_metainfo(struct strbuf *msg,
3031 const char *name,
3032 const char *other,
3033 struct diff_filespec *one,
3034 struct diff_filespec *two,
3035 struct diff_options *o,
3036 struct diff_filepair *p,
3037 int *must_show_header,
3038 int use_color)
3039 {
3040 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3041 const char *reset = diff_get_color(use_color, DIFF_RESET);
3042 const char *line_prefix = diff_line_prefix(o);
3043
3044 *must_show_header = 1;
3045 strbuf_init(msg, PATH_MAX * 2 + 300);
3046 switch (p->status) {
3047 case DIFF_STATUS_COPIED:
3048 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3049 line_prefix, set, similarity_index(p));
3050 strbuf_addf(msg, "%s\n%s%scopy from ",
3051 reset, line_prefix, set);
3052 quote_c_style(name, msg, NULL, 0);
3053 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3054 quote_c_style(other, msg, NULL, 0);
3055 strbuf_addf(msg, "%s\n", reset);
3056 break;
3057 case DIFF_STATUS_RENAMED:
3058 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3059 line_prefix, set, similarity_index(p));
3060 strbuf_addf(msg, "%s\n%s%srename from ",
3061 reset, line_prefix, set);
3062 quote_c_style(name, msg, NULL, 0);
3063 strbuf_addf(msg, "%s\n%s%srename to ",
3064 reset, line_prefix, set);
3065 quote_c_style(other, msg, NULL, 0);
3066 strbuf_addf(msg, "%s\n", reset);
3067 break;
3068 case DIFF_STATUS_MODIFIED:
3069 if (p->score) {
3070 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3071 line_prefix,
3072 set, similarity_index(p), reset);
3073 break;
3074 }
3075 /* fallthru */
3076 default:
3077 *must_show_header = 0;
3078 }
3079 if (one && two && oidcmp(&one->oid, &two->oid)) {
3080 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3081
3082 if (DIFF_OPT_TST(o, BINARY)) {
3083 mmfile_t mf;
3084 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3085 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3086 abbrev = 40;
3087 }
3088 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
3089 find_unique_abbrev(one->oid.hash, abbrev));
3090 strbuf_addstr(msg, find_unique_abbrev(two->oid.hash, abbrev));
3091 if (one->mode == two->mode)
3092 strbuf_addf(msg, " %06o", one->mode);
3093 strbuf_addf(msg, "%s\n", reset);
3094 }
3095 }
3096
3097 static void run_diff_cmd(const char *pgm,
3098 const char *name,
3099 const char *other,
3100 const char *attr_path,
3101 struct diff_filespec *one,
3102 struct diff_filespec *two,
3103 struct strbuf *msg,
3104 struct diff_options *o,
3105 struct diff_filepair *p)
3106 {
3107 const char *xfrm_msg = NULL;
3108 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3109 int must_show_header = 0;
3110
3111
3112 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3113 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3114 if (drv && drv->external)
3115 pgm = drv->external;
3116 }
3117
3118 if (msg) {
3119 /*
3120 * don't use colors when the header is intended for an
3121 * external diff driver
3122 */
3123 fill_metainfo(msg, name, other, one, two, o, p,
3124 &must_show_header,
3125 want_color(o->use_color) && !pgm);
3126 xfrm_msg = msg->len ? msg->buf : NULL;
3127 }
3128
3129 if (pgm) {
3130 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3131 complete_rewrite, o);
3132 return;
3133 }
3134 if (one && two)
3135 builtin_diff(name, other ? other : name,
3136 one, two, xfrm_msg, must_show_header,
3137 o, complete_rewrite);
3138 else
3139 fprintf(o->file, "* Unmerged path %s\n", name);
3140 }
3141
3142 static void diff_fill_sha1_info(struct diff_filespec *one)
3143 {
3144 if (DIFF_FILE_VALID(one)) {
3145 if (!one->oid_valid) {
3146 struct stat st;
3147 if (one->is_stdin) {
3148 oidclr(&one->oid);
3149 return;
3150 }
3151 if (lstat(one->path, &st) < 0)
3152 die_errno("stat '%s'", one->path);
3153 if (index_path(one->oid.hash, one->path, &st, 0))
3154 die("cannot hash %s", one->path);
3155 }
3156 }
3157 else
3158 oidclr(&one->oid);
3159 }
3160
3161 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3162 {
3163 /* Strip the prefix but do not molest /dev/null and absolute paths */
3164 if (*namep && **namep != '/') {
3165 *namep += prefix_length;
3166 if (**namep == '/')
3167 ++*namep;
3168 }
3169 if (*otherp && **otherp != '/') {
3170 *otherp += prefix_length;
3171 if (**otherp == '/')
3172 ++*otherp;
3173 }
3174 }
3175
3176 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3177 {
3178 const char *pgm = external_diff();
3179 struct strbuf msg;
3180 struct diff_filespec *one = p->one;
3181 struct diff_filespec *two = p->two;
3182 const char *name;
3183 const char *other;
3184 const char *attr_path;
3185
3186 name = p->one->path;
3187 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3188 attr_path = name;
3189 if (o->prefix_length)
3190 strip_prefix(o->prefix_length, &name, &other);
3191
3192 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3193 pgm = NULL;
3194
3195 if (DIFF_PAIR_UNMERGED(p)) {
3196 run_diff_cmd(pgm, name, NULL, attr_path,
3197 NULL, NULL, NULL, o, p);
3198 return;
3199 }
3200
3201 diff_fill_sha1_info(one);
3202 diff_fill_sha1_info(two);
3203
3204 if (!pgm &&
3205 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3206 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3207 /*
3208 * a filepair that changes between file and symlink
3209 * needs to be split into deletion and creation.
3210 */
3211 struct diff_filespec *null = alloc_filespec(two->path);
3212 run_diff_cmd(NULL, name, other, attr_path,
3213 one, null, &msg, o, p);
3214 free(null);
3215 strbuf_release(&msg);
3216
3217 null = alloc_filespec(one->path);
3218 run_diff_cmd(NULL, name, other, attr_path,
3219 null, two, &msg, o, p);
3220 free(null);
3221 }
3222 else
3223 run_diff_cmd(pgm, name, other, attr_path,
3224 one, two, &msg, o, p);
3225
3226 strbuf_release(&msg);
3227 }
3228
3229 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3230 struct diffstat_t *diffstat)
3231 {
3232 const char *name;
3233 const char *other;
3234
3235 if (DIFF_PAIR_UNMERGED(p)) {
3236 /* unmerged */
3237 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3238 return;
3239 }
3240
3241 name = p->one->path;
3242 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3243
3244 if (o->prefix_length)
3245 strip_prefix(o->prefix_length, &name, &other);
3246
3247 diff_fill_sha1_info(p->one);
3248 diff_fill_sha1_info(p->two);
3249
3250 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3251 }
3252
3253 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3254 {
3255 const char *name;
3256 const char *other;
3257 const char *attr_path;
3258
3259 if (DIFF_PAIR_UNMERGED(p)) {
3260 /* unmerged */
3261 return;
3262 }
3263
3264 name = p->one->path;
3265 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3266 attr_path = other ? other : name;
3267
3268 if (o->prefix_length)
3269 strip_prefix(o->prefix_length, &name, &other);
3270
3271 diff_fill_sha1_info(p->one);
3272 diff_fill_sha1_info(p->two);
3273
3274 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3275 }
3276
3277 void diff_setup(struct diff_options *options)
3278 {
3279 memcpy(options, &default_diff_options, sizeof(*options));
3280
3281 options->file = stdout;
3282
3283 options->line_termination = '\n';
3284 options->break_opt = -1;
3285 options->rename_limit = -1;
3286 options->dirstat_permille = diff_dirstat_permille_default;
3287 options->context = diff_context_default;
3288 options->ws_error_highlight = WSEH_NEW;
3289 DIFF_OPT_SET(options, RENAME_EMPTY);
3290
3291 /* pathchange left =NULL by default */
3292 options->change = diff_change;
3293 options->add_remove = diff_addremove;
3294 options->use_color = diff_use_color_default;
3295 options->detect_rename = diff_detect_rename_default;
3296 options->xdl_opts |= diff_algorithm;
3297 if (diff_compaction_heuristic)
3298 DIFF_XDL_SET(options, COMPACTION_HEURISTIC);
3299
3300 options->orderfile = diff_order_file_cfg;
3301
3302 if (diff_no_prefix) {
3303 options->a_prefix = options->b_prefix = "";
3304 } else if (!diff_mnemonic_prefix) {
3305 options->a_prefix = "a/";
3306 options->b_prefix = "b/";
3307 }
3308 }
3309
3310 void diff_setup_done(struct diff_options *options)
3311 {
3312 int count = 0;
3313
3314 if (options->set_default)
3315 options->set_default(options);
3316
3317 if (options->output_format & DIFF_FORMAT_NAME)
3318 count++;
3319 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3320 count++;
3321 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3322 count++;
3323 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3324 count++;
3325 if (count > 1)
3326 die("--name-only, --name-status, --check and -s are mutually exclusive");
3327
3328 /*
3329 * Most of the time we can say "there are changes"
3330 * only by checking if there are changed paths, but
3331 * --ignore-whitespace* options force us to look
3332 * inside contents.
3333 */
3334
3335 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3336 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3337 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3338 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3339 else
3340 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3341
3342 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3343 options->detect_rename = DIFF_DETECT_COPY;
3344
3345 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3346 options->prefix = NULL;
3347 if (options->prefix)
3348 options->prefix_length = strlen(options->prefix);
3349 else
3350 options->prefix_length = 0;
3351
3352 if (options->output_format & (DIFF_FORMAT_NAME |
3353 DIFF_FORMAT_NAME_STATUS |
3354 DIFF_FORMAT_CHECKDIFF |
3355 DIFF_FORMAT_NO_OUTPUT))
3356 options->output_format &= ~(DIFF_FORMAT_RAW |
3357 DIFF_FORMAT_NUMSTAT |
3358 DIFF_FORMAT_DIFFSTAT |
3359 DIFF_FORMAT_SHORTSTAT |
3360 DIFF_FORMAT_DIRSTAT |
3361 DIFF_FORMAT_SUMMARY |
3362 DIFF_FORMAT_PATCH);
3363
3364 /*
3365 * These cases always need recursive; we do not drop caller-supplied
3366 * recursive bits for other formats here.
3367 */
3368 if (options->output_format & (DIFF_FORMAT_PATCH |
3369 DIFF_FORMAT_NUMSTAT |
3370 DIFF_FORMAT_DIFFSTAT |
3371 DIFF_FORMAT_SHORTSTAT |
3372 DIFF_FORMAT_DIRSTAT |
3373 DIFF_FORMAT_SUMMARY |
3374 DIFF_FORMAT_CHECKDIFF))
3375 DIFF_OPT_SET(options, RECURSIVE);
3376 /*
3377 * Also pickaxe would not work very well if you do not say recursive
3378 */
3379 if (options->pickaxe)
3380 DIFF_OPT_SET(options, RECURSIVE);
3381 /*
3382 * When patches are generated, submodules diffed against the work tree
3383 * must be checked for dirtiness too so it can be shown in the output
3384 */
3385 if (options->output_format & DIFF_FORMAT_PATCH)
3386 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3387
3388 if (options->detect_rename && options->rename_limit < 0)
3389 options->rename_limit = diff_rename_limit_default;
3390 if (options->setup & DIFF_SETUP_USE_CACHE) {
3391 if (!active_cache)
3392 /* read-cache does not die even when it fails
3393 * so it is safe for us to do this here. Also
3394 * it does not smudge active_cache or active_nr
3395 * when it fails, so we do not have to worry about
3396 * cleaning it up ourselves either.
3397 */
3398 read_cache();
3399 }
3400 if (options->abbrev <= 0 || 40 < options->abbrev)
3401 options->abbrev = 40; /* full */
3402
3403 /*
3404 * It does not make sense to show the first hit we happened
3405 * to have found. It does not make sense not to return with
3406 * exit code in such a case either.
3407 */
3408 if (DIFF_OPT_TST(options, QUICK)) {
3409 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3410 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3411 }
3412
3413 options->diff_path_counter = 0;
3414
3415 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3416 die(_("--follow requires exactly one pathspec"));
3417 }
3418
3419 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3420 {
3421 char c, *eq;
3422 int len;
3423
3424 if (*arg != '-')
3425 return 0;
3426 c = *++arg;
3427 if (!c)
3428 return 0;
3429 if (c == arg_short) {
3430 c = *++arg;
3431 if (!c)
3432 return 1;
3433 if (val && isdigit(c)) {
3434 char *end;
3435 int n = strtoul(arg, &end, 10);
3436 if (*end)
3437 return 0;
3438 *val = n;
3439 return 1;
3440 }
3441 return 0;
3442 }
3443 if (c != '-')
3444 return 0;
3445 arg++;
3446 eq = strchrnul(arg, '=');
3447 len = eq - arg;
3448 if (!len || strncmp(arg, arg_long, len))
3449 return 0;
3450 if (*eq) {
3451 int n;
3452 char *end;
3453 if (!isdigit(*++eq))
3454 return 0;
3455 n = strtoul(eq, &end, 10);
3456 if (*end)
3457 return 0;
3458 *val = n;
3459 }
3460 return 1;
3461 }
3462
3463 static int diff_scoreopt_parse(const char *opt);
3464
3465 static inline int short_opt(char opt, const char **argv,
3466 const char **optarg)
3467 {
3468 const char *arg = argv[0];
3469 if (arg[0] != '-' || arg[1] != opt)
3470 return 0;
3471 if (arg[2] != '\0') {
3472 *optarg = arg + 2;
3473 return 1;
3474 }
3475 if (!argv[1])
3476 die("Option '%c' requires a value", opt);
3477 *optarg = argv[1];
3478 return 2;
3479 }
3480
3481 int parse_long_opt(const char *opt, const char **argv,
3482 const char **optarg)
3483 {
3484 const char *arg = argv[0];
3485 if (!skip_prefix(arg, "--", &arg))
3486 return 0;
3487 if (!skip_prefix(arg, opt, &arg))
3488 return 0;
3489 if (*arg == '=') { /* stuck form: --option=value */
3490 *optarg = arg + 1;
3491 return 1;
3492 }
3493 if (*arg != '\0')
3494 return 0;
3495 /* separate form: --option value */
3496 if (!argv[1])
3497 die("Option '--%s' requires a value", opt);
3498 *optarg = argv[1];
3499 return 2;
3500 }
3501
3502 static int stat_opt(struct diff_options *options, const char **av)
3503 {
3504 const char *arg = av[0];
3505 char *end;
3506 int width = options->stat_width;
3507 int name_width = options->stat_name_width;
3508 int graph_width = options->stat_graph_width;
3509 int count = options->stat_count;
3510 int argcount = 1;
3511
3512 if (!skip_prefix(arg, "--stat", &arg))
3513 die("BUG: stat option does not begin with --stat: %s", arg);
3514 end = (char *)arg;
3515
3516 switch (*arg) {
3517 case '-':
3518 if (skip_prefix(arg, "-width", &arg)) {
3519 if (*arg == '=')
3520 width = strtoul(arg + 1, &end, 10);
3521 else if (!*arg && !av[1])
3522 die("Option '--stat-width' requires a value");
3523 else if (!*arg) {
3524 width = strtoul(av[1], &end, 10);
3525 argcount = 2;
3526 }
3527 } else if (skip_prefix(arg, "-name-width", &arg)) {
3528 if (*arg == '=')
3529 name_width = strtoul(arg + 1, &end, 10);
3530 else if (!*arg && !av[1])
3531 die("Option '--stat-name-width' requires a value");
3532 else if (!*arg) {
3533 name_width = strtoul(av[1], &end, 10);
3534 argcount = 2;
3535 }
3536 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3537 if (*arg == '=')
3538 graph_width = strtoul(arg + 1, &end, 10);
3539 else if (!*arg && !av[1])
3540 die("Option '--stat-graph-width' requires a value");
3541 else if (!*arg) {
3542 graph_width = strtoul(av[1], &end, 10);
3543 argcount = 2;
3544 }
3545 } else if (skip_prefix(arg, "-count", &arg)) {
3546 if (*arg == '=')
3547 count = strtoul(arg + 1, &end, 10);
3548 else if (!*arg && !av[1])
3549 die("Option '--stat-count' requires a value");
3550 else if (!*arg) {
3551 count = strtoul(av[1], &end, 10);
3552 argcount = 2;
3553 }
3554 }
3555 break;
3556 case '=':
3557 width = strtoul(arg+1, &end, 10);
3558 if (*end == ',')
3559 name_width = strtoul(end+1, &end, 10);
3560 if (*end == ',')
3561 count = strtoul(end+1, &end, 10);
3562 }
3563
3564 /* Important! This checks all the error cases! */
3565 if (*end)
3566 return 0;
3567 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3568 options->stat_name_width = name_width;
3569 options->stat_graph_width = graph_width;
3570 options->stat_width = width;
3571 options->stat_count = count;
3572 return argcount;
3573 }
3574
3575 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3576 {
3577 struct strbuf errmsg = STRBUF_INIT;
3578 if (parse_dirstat_params(options, params, &errmsg))
3579 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3580 errmsg.buf);
3581 strbuf_release(&errmsg);
3582 /*
3583 * The caller knows a dirstat-related option is given from the command
3584 * line; allow it to say "return this_function();"
3585 */
3586 options->output_format |= DIFF_FORMAT_DIRSTAT;
3587 return 1;
3588 }
3589
3590 static int parse_submodule_opt(struct diff_options *options, const char *value)
3591 {
3592 if (parse_submodule_params(options, value))
3593 die(_("Failed to parse --submodule option parameter: '%s'"),
3594 value);
3595 return 1;
3596 }
3597
3598 static const char diff_status_letters[] = {
3599 DIFF_STATUS_ADDED,
3600 DIFF_STATUS_COPIED,
3601 DIFF_STATUS_DELETED,
3602 DIFF_STATUS_MODIFIED,
3603 DIFF_STATUS_RENAMED,
3604 DIFF_STATUS_TYPE_CHANGED,
3605 DIFF_STATUS_UNKNOWN,
3606 DIFF_STATUS_UNMERGED,
3607 DIFF_STATUS_FILTER_AON,
3608 DIFF_STATUS_FILTER_BROKEN,
3609 '\0',
3610 };
3611
3612 static unsigned int filter_bit['Z' + 1];
3613
3614 static void prepare_filter_bits(void)
3615 {
3616 int i;
3617
3618 if (!filter_bit[DIFF_STATUS_ADDED]) {
3619 for (i = 0; diff_status_letters[i]; i++)
3620 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3621 }
3622 }
3623
3624 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3625 {
3626 return opt->filter & filter_bit[(int) status];
3627 }
3628
3629 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3630 {
3631 int i, optch;
3632
3633 prepare_filter_bits();
3634
3635 /*
3636 * If there is a negation e.g. 'd' in the input, and we haven't
3637 * initialized the filter field with another --diff-filter, start
3638 * from full set of bits, except for AON.
3639 */
3640 if (!opt->filter) {
3641 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3642 if (optch < 'a' || 'z' < optch)
3643 continue;
3644 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3645 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3646 break;
3647 }
3648 }
3649
3650 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3651 unsigned int bit;
3652 int negate;
3653
3654 if ('a' <= optch && optch <= 'z') {
3655 negate = 1;
3656 optch = toupper(optch);
3657 } else {
3658 negate = 0;
3659 }
3660
3661 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
3662 if (!bit)
3663 return optarg[i];
3664 if (negate)
3665 opt->filter &= ~bit;
3666 else
3667 opt->filter |= bit;
3668 }
3669 return 0;
3670 }
3671
3672 static void enable_patch_output(int *fmt) {
3673 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
3674 *fmt |= DIFF_FORMAT_PATCH;
3675 }
3676
3677 static int parse_one_token(const char **arg, const char *token)
3678 {
3679 const char *rest;
3680 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
3681 *arg = rest;
3682 return 1;
3683 }
3684 return 0;
3685 }
3686
3687 static int parse_ws_error_highlight(struct diff_options *opt, const char *arg)
3688 {
3689 const char *orig_arg = arg;
3690 unsigned val = 0;
3691 while (*arg) {
3692 if (parse_one_token(&arg, "none"))
3693 val = 0;
3694 else if (parse_one_token(&arg, "default"))
3695 val = WSEH_NEW;
3696 else if (parse_one_token(&arg, "all"))
3697 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
3698 else if (parse_one_token(&arg, "new"))
3699 val |= WSEH_NEW;
3700 else if (parse_one_token(&arg, "old"))
3701 val |= WSEH_OLD;
3702 else if (parse_one_token(&arg, "context"))
3703 val |= WSEH_CONTEXT;
3704 else {
3705 error("unknown value after ws-error-highlight=%.*s",
3706 (int)(arg - orig_arg), orig_arg);
3707 return 0;
3708 }
3709 if (*arg)
3710 arg++;
3711 }
3712 opt->ws_error_highlight = val;
3713 return 1;
3714 }
3715
3716 int diff_opt_parse(struct diff_options *options,
3717 const char **av, int ac, const char *prefix)
3718 {
3719 const char *arg = av[0];
3720 const char *optarg;
3721 int argcount;
3722
3723 if (!prefix)
3724 prefix = "";
3725
3726 /* Output format options */
3727 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
3728 || opt_arg(arg, 'U', "unified", &options->context))
3729 enable_patch_output(&options->output_format);
3730 else if (!strcmp(arg, "--raw"))
3731 options->output_format |= DIFF_FORMAT_RAW;
3732 else if (!strcmp(arg, "--patch-with-raw")) {
3733 enable_patch_output(&options->output_format);
3734 options->output_format |= DIFF_FORMAT_RAW;
3735 } else if (!strcmp(arg, "--numstat"))
3736 options->output_format |= DIFF_FORMAT_NUMSTAT;
3737 else if (!strcmp(arg, "--shortstat"))
3738 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3739 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3740 return parse_dirstat_opt(options, "");
3741 else if (skip_prefix(arg, "-X", &arg))
3742 return parse_dirstat_opt(options, arg);
3743 else if (skip_prefix(arg, "--dirstat=", &arg))
3744 return parse_dirstat_opt(options, arg);
3745 else if (!strcmp(arg, "--cumulative"))
3746 return parse_dirstat_opt(options, "cumulative");
3747 else if (!strcmp(arg, "--dirstat-by-file"))
3748 return parse_dirstat_opt(options, "files");
3749 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
3750 parse_dirstat_opt(options, "files");
3751 return parse_dirstat_opt(options, arg);
3752 }
3753 else if (!strcmp(arg, "--check"))
3754 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3755 else if (!strcmp(arg, "--summary"))
3756 options->output_format |= DIFF_FORMAT_SUMMARY;
3757 else if (!strcmp(arg, "--patch-with-stat")) {
3758 enable_patch_output(&options->output_format);
3759 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3760 } else if (!strcmp(arg, "--name-only"))
3761 options->output_format |= DIFF_FORMAT_NAME;
3762 else if (!strcmp(arg, "--name-status"))
3763 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3764 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
3765 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3766 else if (starts_with(arg, "--stat"))
3767 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3768 return stat_opt(options, av);
3769
3770 /* renames options */
3771 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
3772 !strcmp(arg, "--break-rewrites")) {
3773 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3774 return error("invalid argument to -B: %s", arg+2);
3775 }
3776 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
3777 !strcmp(arg, "--find-renames")) {
3778 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3779 return error("invalid argument to -M: %s", arg+2);
3780 options->detect_rename = DIFF_DETECT_RENAME;
3781 }
3782 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3783 options->irreversible_delete = 1;
3784 }
3785 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
3786 !strcmp(arg, "--find-copies")) {
3787 if (options->detect_rename == DIFF_DETECT_COPY)
3788 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3789 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3790 return error("invalid argument to -C: %s", arg+2);
3791 options->detect_rename = DIFF_DETECT_COPY;
3792 }
3793 else if (!strcmp(arg, "--no-renames"))
3794 options->detect_rename = 0;
3795 else if (!strcmp(arg, "--rename-empty"))
3796 DIFF_OPT_SET(options, RENAME_EMPTY);
3797 else if (!strcmp(arg, "--no-rename-empty"))
3798 DIFF_OPT_CLR(options, RENAME_EMPTY);
3799 else if (!strcmp(arg, "--relative"))
3800 DIFF_OPT_SET(options, RELATIVE_NAME);
3801 else if (skip_prefix(arg, "--relative=", &arg)) {
3802 DIFF_OPT_SET(options, RELATIVE_NAME);
3803 options->prefix = arg;
3804 }
3805
3806 /* xdiff options */
3807 else if (!strcmp(arg, "--minimal"))
3808 DIFF_XDL_SET(options, NEED_MINIMAL);
3809 else if (!strcmp(arg, "--no-minimal"))
3810 DIFF_XDL_CLR(options, NEED_MINIMAL);
3811 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3812 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3813 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3814 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3815 else if (!strcmp(arg, "--ignore-space-at-eol"))
3816 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3817 else if (!strcmp(arg, "--ignore-blank-lines"))
3818 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
3819 else if (!strcmp(arg, "--compaction-heuristic"))
3820 DIFF_XDL_SET(options, COMPACTION_HEURISTIC);
3821 else if (!strcmp(arg, "--no-compaction-heuristic"))
3822 DIFF_XDL_CLR(options, COMPACTION_HEURISTIC);
3823 else if (!strcmp(arg, "--patience"))
3824 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3825 else if (!strcmp(arg, "--histogram"))
3826 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3827 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
3828 long value = parse_algorithm_value(optarg);
3829 if (value < 0)
3830 return error("option diff-algorithm accepts \"myers\", "
3831 "\"minimal\", \"patience\" and \"histogram\"");
3832 /* clear out previous settings */
3833 DIFF_XDL_CLR(options, NEED_MINIMAL);
3834 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3835 options->xdl_opts |= value;
3836 return argcount;
3837 }
3838
3839 /* flags options */
3840 else if (!strcmp(arg, "--binary")) {
3841 enable_patch_output(&options->output_format);
3842 DIFF_OPT_SET(options, BINARY);
3843 }
3844 else if (!strcmp(arg, "--full-index"))
3845 DIFF_OPT_SET(options, FULL_INDEX);
3846 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3847 DIFF_OPT_SET(options, TEXT);
3848 else if (!strcmp(arg, "-R"))
3849 DIFF_OPT_SET(options, REVERSE_DIFF);
3850 else if (!strcmp(arg, "--find-copies-harder"))
3851 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3852 else if (!strcmp(arg, "--follow"))
3853 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3854 else if (!strcmp(arg, "--no-follow")) {
3855 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3856 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
3857 } else if (!strcmp(arg, "--color"))
3858 options->use_color = 1;
3859 else if (skip_prefix(arg, "--color=", &arg)) {
3860 int value = git_config_colorbool(NULL, arg);
3861 if (value < 0)
3862 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3863 options->use_color = value;
3864 }
3865 else if (!strcmp(arg, "--no-color"))
3866 options->use_color = 0;
3867 else if (!strcmp(arg, "--color-words")) {
3868 options->use_color = 1;
3869 options->word_diff = DIFF_WORDS_COLOR;
3870 }
3871 else if (skip_prefix(arg, "--color-words=", &arg)) {
3872 options->use_color = 1;
3873 options->word_diff = DIFF_WORDS_COLOR;
3874 options->word_regex = arg;
3875 }
3876 else if (!strcmp(arg, "--word-diff")) {
3877 if (options->word_diff == DIFF_WORDS_NONE)
3878 options->word_diff = DIFF_WORDS_PLAIN;
3879 }
3880 else if (skip_prefix(arg, "--word-diff=", &arg)) {
3881 if (!strcmp(arg, "plain"))
3882 options->word_diff = DIFF_WORDS_PLAIN;
3883 else if (!strcmp(arg, "color")) {
3884 options->use_color = 1;
3885 options->word_diff = DIFF_WORDS_COLOR;
3886 }
3887 else if (!strcmp(arg, "porcelain"))
3888 options->word_diff = DIFF_WORDS_PORCELAIN;
3889 else if (!strcmp(arg, "none"))
3890 options->word_diff = DIFF_WORDS_NONE;
3891 else
3892 die("bad --word-diff argument: %s", arg);
3893 }
3894 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3895 if (options->word_diff == DIFF_WORDS_NONE)
3896 options->word_diff = DIFF_WORDS_PLAIN;
3897 options->word_regex = optarg;
3898 return argcount;
3899 }
3900 else if (!strcmp(arg, "--exit-code"))
3901 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3902 else if (!strcmp(arg, "--quiet"))
3903 DIFF_OPT_SET(options, QUICK);
3904 else if (!strcmp(arg, "--ext-diff"))
3905 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3906 else if (!strcmp(arg, "--no-ext-diff"))
3907 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3908 else if (!strcmp(arg, "--textconv"))
3909 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3910 else if (!strcmp(arg, "--no-textconv"))
3911 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3912 else if (!strcmp(arg, "--ignore-submodules")) {
3913 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3914 handle_ignore_submodules_arg(options, "all");
3915 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
3916 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3917 handle_ignore_submodules_arg(options, arg);
3918 } else if (!strcmp(arg, "--submodule"))
3919 options->submodule_format = DIFF_SUBMODULE_LOG;
3920 else if (skip_prefix(arg, "--submodule=", &arg))
3921 return parse_submodule_opt(options, arg);
3922 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
3923 return parse_ws_error_highlight(options, arg);
3924
3925 /* misc options */
3926 else if (!strcmp(arg, "-z"))
3927 options->line_termination = 0;
3928 else if ((argcount = short_opt('l', av, &optarg))) {
3929 options->rename_limit = strtoul(optarg, NULL, 10);
3930 return argcount;
3931 }
3932 else if ((argcount = short_opt('S', av, &optarg))) {
3933 options->pickaxe = optarg;
3934 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3935 return argcount;
3936 } else if ((argcount = short_opt('G', av, &optarg))) {
3937 options->pickaxe = optarg;
3938 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3939 return argcount;
3940 }
3941 else if (!strcmp(arg, "--pickaxe-all"))
3942 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3943 else if (!strcmp(arg, "--pickaxe-regex"))
3944 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3945 else if ((argcount = short_opt('O', av, &optarg))) {
3946 const char *path = prefix_filename(prefix, strlen(prefix), optarg);
3947 options->orderfile = xstrdup(path);
3948 return argcount;
3949 }
3950 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3951 int offending = parse_diff_filter_opt(optarg, options);
3952 if (offending)
3953 die("unknown change class '%c' in --diff-filter=%s",
3954 offending, optarg);
3955 return argcount;
3956 }
3957 else if (!strcmp(arg, "--abbrev"))
3958 options->abbrev = DEFAULT_ABBREV;
3959 else if (skip_prefix(arg, "--abbrev=", &arg)) {
3960 options->abbrev = strtoul(arg, NULL, 10);
3961 if (options->abbrev < MINIMUM_ABBREV)
3962 options->abbrev = MINIMUM_ABBREV;
3963 else if (40 < options->abbrev)
3964 options->abbrev = 40;
3965 }
3966 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3967 options->a_prefix = optarg;
3968 return argcount;
3969 }
3970 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
3971 options->line_prefix = optarg;
3972 options->line_prefix_length = strlen(options->line_prefix);
3973 graph_setup_line_prefix(options);
3974 return argcount;
3975 }
3976 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3977 options->b_prefix = optarg;
3978 return argcount;
3979 }
3980 else if (!strcmp(arg, "--no-prefix"))
3981 options->a_prefix = options->b_prefix = "";
3982 else if (opt_arg(arg, '\0', "inter-hunk-context",
3983 &options->interhunkcontext))
3984 ;
3985 else if (!strcmp(arg, "-W"))
3986 DIFF_OPT_SET(options, FUNCCONTEXT);
3987 else if (!strcmp(arg, "--function-context"))
3988 DIFF_OPT_SET(options, FUNCCONTEXT);
3989 else if (!strcmp(arg, "--no-function-context"))
3990 DIFF_OPT_CLR(options, FUNCCONTEXT);
3991 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3992 const char *path = prefix_filename(prefix, strlen(prefix), optarg);
3993 options->file = fopen(path, "w");
3994 if (!options->file)
3995 die_errno("Could not open '%s'", path);
3996 options->close_file = 1;
3997 if (options->use_color != GIT_COLOR_ALWAYS)
3998 options->use_color = GIT_COLOR_NEVER;
3999 return argcount;
4000 } else
4001 return 0;
4002 return 1;
4003 }
4004
4005 int parse_rename_score(const char **cp_p)
4006 {
4007 unsigned long num, scale;
4008 int ch, dot;
4009 const char *cp = *cp_p;
4010
4011 num = 0;
4012 scale = 1;
4013 dot = 0;
4014 for (;;) {
4015 ch = *cp;
4016 if ( !dot && ch == '.' ) {
4017 scale = 1;
4018 dot = 1;
4019 } else if ( ch == '%' ) {
4020 scale = dot ? scale*100 : 100;
4021 cp++; /* % is always at the end */
4022 break;
4023 } else if ( ch >= '0' && ch <= '9' ) {
4024 if ( scale < 100000 ) {
4025 scale *= 10;
4026 num = (num*10) + (ch-'0');
4027 }
4028 } else {
4029 break;
4030 }
4031 cp++;
4032 }
4033 *cp_p = cp;
4034
4035 /* user says num divided by scale and we say internally that
4036 * is MAX_SCORE * num / scale.
4037 */
4038 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4039 }
4040
4041 static int diff_scoreopt_parse(const char *opt)
4042 {
4043 int opt1, opt2, cmd;
4044
4045 if (*opt++ != '-')
4046 return -1;
4047 cmd = *opt++;
4048 if (cmd == '-') {
4049 /* convert the long-form arguments into short-form versions */
4050 if (skip_prefix(opt, "break-rewrites", &opt)) {
4051 if (*opt == 0 || *opt++ == '=')
4052 cmd = 'B';
4053 } else if (skip_prefix(opt, "find-copies", &opt)) {
4054 if (*opt == 0 || *opt++ == '=')
4055 cmd = 'C';
4056 } else if (skip_prefix(opt, "find-renames", &opt)) {
4057 if (*opt == 0 || *opt++ == '=')
4058 cmd = 'M';
4059 }
4060 }
4061 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4062 return -1; /* that is not a -M, -C, or -B option */
4063
4064 opt1 = parse_rename_score(&opt);
4065 if (cmd != 'B')
4066 opt2 = 0;
4067 else {
4068 if (*opt == 0)
4069 opt2 = 0;
4070 else if (*opt != '/')
4071 return -1; /* we expect -B80/99 or -B80 */
4072 else {
4073 opt++;
4074 opt2 = parse_rename_score(&opt);
4075 }
4076 }
4077 if (*opt != 0)
4078 return -1;
4079 return opt1 | (opt2 << 16);
4080 }
4081
4082 struct diff_queue_struct diff_queued_diff;
4083
4084 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4085 {
4086 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4087 queue->queue[queue->nr++] = dp;
4088 }
4089
4090 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4091 struct diff_filespec *one,
4092 struct diff_filespec *two)
4093 {
4094 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4095 dp->one = one;
4096 dp->two = two;
4097 if (queue)
4098 diff_q(queue, dp);
4099 return dp;
4100 }
4101
4102 void diff_free_filepair(struct diff_filepair *p)
4103 {
4104 free_filespec(p->one);
4105 free_filespec(p->two);
4106 free(p);
4107 }
4108
4109 /* This is different from find_unique_abbrev() in that
4110 * it stuffs the result with dots for alignment.
4111 */
4112 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
4113 {
4114 int abblen;
4115 const char *abbrev;
4116 if (len == 40)
4117 return sha1_to_hex(sha1);
4118
4119 abbrev = find_unique_abbrev(sha1, len);
4120 abblen = strlen(abbrev);
4121 if (abblen < 37) {
4122 static char hex[41];
4123 if (len < abblen && abblen <= len + 2)
4124 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
4125 else
4126 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
4127 return hex;
4128 }
4129 return sha1_to_hex(sha1);
4130 }
4131
4132 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4133 {
4134 int line_termination = opt->line_termination;
4135 int inter_name_termination = line_termination ? '\t' : '\0';
4136
4137 fprintf(opt->file, "%s", diff_line_prefix(opt));
4138 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4139 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4140 diff_unique_abbrev(p->one->oid.hash, opt->abbrev));
4141 fprintf(opt->file, "%s ",
4142 diff_unique_abbrev(p->two->oid.hash, opt->abbrev));
4143 }
4144 if (p->score) {
4145 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4146 inter_name_termination);
4147 } else {
4148 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4149 }
4150
4151 if (p->status == DIFF_STATUS_COPIED ||
4152 p->status == DIFF_STATUS_RENAMED) {
4153 const char *name_a, *name_b;
4154 name_a = p->one->path;
4155 name_b = p->two->path;
4156 strip_prefix(opt->prefix_length, &name_a, &name_b);
4157 write_name_quoted(name_a, opt->file, inter_name_termination);
4158 write_name_quoted(name_b, opt->file, line_termination);
4159 } else {
4160 const char *name_a, *name_b;
4161 name_a = p->one->mode ? p->one->path : p->two->path;
4162 name_b = NULL;
4163 strip_prefix(opt->prefix_length, &name_a, &name_b);
4164 write_name_quoted(name_a, opt->file, line_termination);
4165 }
4166 }
4167
4168 int diff_unmodified_pair(struct diff_filepair *p)
4169 {
4170 /* This function is written stricter than necessary to support
4171 * the currently implemented transformers, but the idea is to
4172 * let transformers to produce diff_filepairs any way they want,
4173 * and filter and clean them up here before producing the output.
4174 */
4175 struct diff_filespec *one = p->one, *two = p->two;
4176
4177 if (DIFF_PAIR_UNMERGED(p))
4178 return 0; /* unmerged is interesting */
4179
4180 /* deletion, addition, mode or type change
4181 * and rename are all interesting.
4182 */
4183 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4184 DIFF_PAIR_MODE_CHANGED(p) ||
4185 strcmp(one->path, two->path))
4186 return 0;
4187
4188 /* both are valid and point at the same path. that is, we are
4189 * dealing with a change.
4190 */
4191 if (one->oid_valid && two->oid_valid &&
4192 !oidcmp(&one->oid, &two->oid) &&
4193 !one->dirty_submodule && !two->dirty_submodule)
4194 return 1; /* no change */
4195 if (!one->oid_valid && !two->oid_valid)
4196 return 1; /* both look at the same file on the filesystem. */
4197 return 0;
4198 }
4199
4200 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4201 {
4202 if (diff_unmodified_pair(p))
4203 return;
4204
4205 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4206 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4207 return; /* no tree diffs in patch format */
4208
4209 run_diff(p, o);
4210 }
4211
4212 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4213 struct diffstat_t *diffstat)
4214 {
4215 if (diff_unmodified_pair(p))
4216 return;
4217
4218 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4219 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4220 return; /* no useful stat for tree diffs */
4221
4222 run_diffstat(p, o, diffstat);
4223 }
4224
4225 static void diff_flush_checkdiff(struct diff_filepair *p,
4226 struct diff_options *o)
4227 {
4228 if (diff_unmodified_pair(p))
4229 return;
4230
4231 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4232 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4233 return; /* nothing to check in tree diffs */
4234
4235 run_checkdiff(p, o);
4236 }
4237
4238 int diff_queue_is_empty(void)
4239 {
4240 struct diff_queue_struct *q = &diff_queued_diff;
4241 int i;
4242 for (i = 0; i < q->nr; i++)
4243 if (!diff_unmodified_pair(q->queue[i]))
4244 return 0;
4245 return 1;
4246 }
4247
4248 #if DIFF_DEBUG
4249 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4250 {
4251 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4252 x, one ? one : "",
4253 s->path,
4254 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4255 s->mode,
4256 s->oid_valid ? oid_to_hex(&s->oid) : "");
4257 fprintf(stderr, "queue[%d] %s size %lu\n",
4258 x, one ? one : "",
4259 s->size);
4260 }
4261
4262 void diff_debug_filepair(const struct diff_filepair *p, int i)
4263 {
4264 diff_debug_filespec(p->one, i, "one");
4265 diff_debug_filespec(p->two, i, "two");
4266 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4267 p->score, p->status ? p->status : '?',
4268 p->one->rename_used, p->broken_pair);
4269 }
4270
4271 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4272 {
4273 int i;
4274 if (msg)
4275 fprintf(stderr, "%s\n", msg);
4276 fprintf(stderr, "q->nr = %d\n", q->nr);
4277 for (i = 0; i < q->nr; i++) {
4278 struct diff_filepair *p = q->queue[i];
4279 diff_debug_filepair(p, i);
4280 }
4281 }
4282 #endif
4283
4284 static void diff_resolve_rename_copy(void)
4285 {
4286 int i;
4287 struct diff_filepair *p;
4288 struct diff_queue_struct *q = &diff_queued_diff;
4289
4290 diff_debug_queue("resolve-rename-copy", q);
4291
4292 for (i = 0; i < q->nr; i++) {
4293 p = q->queue[i];
4294 p->status = 0; /* undecided */
4295 if (DIFF_PAIR_UNMERGED(p))
4296 p->status = DIFF_STATUS_UNMERGED;
4297 else if (!DIFF_FILE_VALID(p->one))
4298 p->status = DIFF_STATUS_ADDED;
4299 else if (!DIFF_FILE_VALID(p->two))
4300 p->status = DIFF_STATUS_DELETED;
4301 else if (DIFF_PAIR_TYPE_CHANGED(p))
4302 p->status = DIFF_STATUS_TYPE_CHANGED;
4303
4304 /* from this point on, we are dealing with a pair
4305 * whose both sides are valid and of the same type, i.e.
4306 * either in-place edit or rename/copy edit.
4307 */
4308 else if (DIFF_PAIR_RENAME(p)) {
4309 /*
4310 * A rename might have re-connected a broken
4311 * pair up, causing the pathnames to be the
4312 * same again. If so, that's not a rename at
4313 * all, just a modification..
4314 *
4315 * Otherwise, see if this source was used for
4316 * multiple renames, in which case we decrement
4317 * the count, and call it a copy.
4318 */
4319 if (!strcmp(p->one->path, p->two->path))
4320 p->status = DIFF_STATUS_MODIFIED;
4321 else if (--p->one->rename_used > 0)
4322 p->status = DIFF_STATUS_COPIED;
4323 else
4324 p->status = DIFF_STATUS_RENAMED;
4325 }
4326 else if (oidcmp(&p->one->oid, &p->two->oid) ||
4327 p->one->mode != p->two->mode ||
4328 p->one->dirty_submodule ||
4329 p->two->dirty_submodule ||
4330 is_null_oid(&p->one->oid))
4331 p->status = DIFF_STATUS_MODIFIED;
4332 else {
4333 /* This is a "no-change" entry and should not
4334 * happen anymore, but prepare for broken callers.
4335 */
4336 error("feeding unmodified %s to diffcore",
4337 p->one->path);
4338 p->status = DIFF_STATUS_UNKNOWN;
4339 }
4340 }
4341 diff_debug_queue("resolve-rename-copy done", q);
4342 }
4343
4344 static int check_pair_status(struct diff_filepair *p)
4345 {
4346 switch (p->status) {
4347 case DIFF_STATUS_UNKNOWN:
4348 return 0;
4349 case 0:
4350 die("internal error in diff-resolve-rename-copy");
4351 default:
4352 return 1;
4353 }
4354 }
4355
4356 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4357 {
4358 int fmt = opt->output_format;
4359
4360 if (fmt & DIFF_FORMAT_CHECKDIFF)
4361 diff_flush_checkdiff(p, opt);
4362 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4363 diff_flush_raw(p, opt);
4364 else if (fmt & DIFF_FORMAT_NAME) {
4365 const char *name_a, *name_b;
4366 name_a = p->two->path;
4367 name_b = NULL;
4368 strip_prefix(opt->prefix_length, &name_a, &name_b);
4369 write_name_quoted(name_a, opt->file, opt->line_termination);
4370 }
4371 }
4372
4373 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4374 {
4375 if (fs->mode)
4376 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4377 else
4378 fprintf(file, " %s ", newdelete);
4379 write_name_quoted(fs->path, file, '\n');
4380 }
4381
4382
4383 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4384 const char *line_prefix)
4385 {
4386 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4387 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4388 p->two->mode, show_name ? ' ' : '\n');
4389 if (show_name) {
4390 write_name_quoted(p->two->path, file, '\n');
4391 }
4392 }
4393 }
4394
4395 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4396 const char *line_prefix)
4397 {
4398 char *names = pprint_rename(p->one->path, p->two->path);
4399
4400 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4401 free(names);
4402 show_mode_change(file, p, 0, line_prefix);
4403 }
4404
4405 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4406 {
4407 FILE *file = opt->file;
4408 const char *line_prefix = diff_line_prefix(opt);
4409
4410 switch(p->status) {
4411 case DIFF_STATUS_DELETED:
4412 fputs(line_prefix, file);
4413 show_file_mode_name(file, "delete", p->one);
4414 break;
4415 case DIFF_STATUS_ADDED:
4416 fputs(line_prefix, file);
4417 show_file_mode_name(file, "create", p->two);
4418 break;
4419 case DIFF_STATUS_COPIED:
4420 fputs(line_prefix, file);
4421 show_rename_copy(file, "copy", p, line_prefix);
4422 break;
4423 case DIFF_STATUS_RENAMED:
4424 fputs(line_prefix, file);
4425 show_rename_copy(file, "rename", p, line_prefix);
4426 break;
4427 default:
4428 if (p->score) {
4429 fprintf(file, "%s rewrite ", line_prefix);
4430 write_name_quoted(p->two->path, file, ' ');
4431 fprintf(file, "(%d%%)\n", similarity_index(p));
4432 }
4433 show_mode_change(file, p, !p->score, line_prefix);
4434 break;
4435 }
4436 }
4437
4438 struct patch_id_t {
4439 git_SHA_CTX *ctx;
4440 int patchlen;
4441 };
4442
4443 static int remove_space(char *line, int len)
4444 {
4445 int i;
4446 char *dst = line;
4447 unsigned char c;
4448
4449 for (i = 0; i < len; i++)
4450 if (!isspace((c = line[i])))
4451 *dst++ = c;
4452
4453 return dst - line;
4454 }
4455
4456 static void patch_id_consume(void *priv, char *line, unsigned long len)
4457 {
4458 struct patch_id_t *data = priv;
4459 int new_len;
4460
4461 /* Ignore line numbers when computing the SHA1 of the patch */
4462 if (starts_with(line, "@@ -"))
4463 return;
4464
4465 new_len = remove_space(line, len);
4466
4467 git_SHA1_Update(data->ctx, line, new_len);
4468 data->patchlen += new_len;
4469 }
4470
4471 /* returns 0 upon success, and writes result into sha1 */
4472 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4473 {
4474 struct diff_queue_struct *q = &diff_queued_diff;
4475 int i;
4476 git_SHA_CTX ctx;
4477 struct patch_id_t data;
4478 char buffer[PATH_MAX * 4 + 20];
4479
4480 git_SHA1_Init(&ctx);
4481 memset(&data, 0, sizeof(struct patch_id_t));
4482 data.ctx = &ctx;
4483
4484 for (i = 0; i < q->nr; i++) {
4485 xpparam_t xpp;
4486 xdemitconf_t xecfg;
4487 mmfile_t mf1, mf2;
4488 struct diff_filepair *p = q->queue[i];
4489 int len1, len2;
4490
4491 memset(&xpp, 0, sizeof(xpp));
4492 memset(&xecfg, 0, sizeof(xecfg));
4493 if (p->status == 0)
4494 return error("internal diff status error");
4495 if (p->status == DIFF_STATUS_UNKNOWN)
4496 continue;
4497 if (diff_unmodified_pair(p))
4498 continue;
4499 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4500 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4501 continue;
4502 if (DIFF_PAIR_UNMERGED(p))
4503 continue;
4504
4505 diff_fill_sha1_info(p->one);
4506 diff_fill_sha1_info(p->two);
4507 if (fill_mmfile(&mf1, p->one) < 0 ||
4508 fill_mmfile(&mf2, p->two) < 0)
4509 return error("unable to read files to diff");
4510
4511 len1 = remove_space(p->one->path, strlen(p->one->path));
4512 len2 = remove_space(p->two->path, strlen(p->two->path));
4513 if (p->one->mode == 0)
4514 len1 = snprintf(buffer, sizeof(buffer),
4515 "diff--gita/%.*sb/%.*s"
4516 "newfilemode%06o"
4517 "---/dev/null"
4518 "+++b/%.*s",
4519 len1, p->one->path,
4520 len2, p->two->path,
4521 p->two->mode,
4522 len2, p->two->path);
4523 else if (p->two->mode == 0)
4524 len1 = snprintf(buffer, sizeof(buffer),
4525 "diff--gita/%.*sb/%.*s"
4526 "deletedfilemode%06o"
4527 "---a/%.*s"
4528 "+++/dev/null",
4529 len1, p->one->path,
4530 len2, p->two->path,
4531 p->one->mode,
4532 len1, p->one->path);
4533 else
4534 len1 = snprintf(buffer, sizeof(buffer),
4535 "diff--gita/%.*sb/%.*s"
4536 "---a/%.*s"
4537 "+++b/%.*s",
4538 len1, p->one->path,
4539 len2, p->two->path,
4540 len1, p->one->path,
4541 len2, p->two->path);
4542 git_SHA1_Update(&ctx, buffer, len1);
4543
4544 if (diff_filespec_is_binary(p->one) ||
4545 diff_filespec_is_binary(p->two)) {
4546 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
4547 40);
4548 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
4549 40);
4550 continue;
4551 }
4552
4553 xpp.flags = 0;
4554 xecfg.ctxlen = 3;
4555 xecfg.flags = 0;
4556 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4557 &xpp, &xecfg))
4558 return error("unable to generate patch-id diff for %s",
4559 p->one->path);
4560 }
4561
4562 git_SHA1_Final(sha1, &ctx);
4563 return 0;
4564 }
4565
4566 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4567 {
4568 struct diff_queue_struct *q = &diff_queued_diff;
4569 int i;
4570 int result = diff_get_patch_id(options, sha1);
4571
4572 for (i = 0; i < q->nr; i++)
4573 diff_free_filepair(q->queue[i]);
4574
4575 free(q->queue);
4576 DIFF_QUEUE_CLEAR(q);
4577
4578 return result;
4579 }
4580
4581 static int is_summary_empty(const struct diff_queue_struct *q)
4582 {
4583 int i;
4584
4585 for (i = 0; i < q->nr; i++) {
4586 const struct diff_filepair *p = q->queue[i];
4587
4588 switch (p->status) {
4589 case DIFF_STATUS_DELETED:
4590 case DIFF_STATUS_ADDED:
4591 case DIFF_STATUS_COPIED:
4592 case DIFF_STATUS_RENAMED:
4593 return 0;
4594 default:
4595 if (p->score)
4596 return 0;
4597 if (p->one->mode && p->two->mode &&
4598 p->one->mode != p->two->mode)
4599 return 0;
4600 break;
4601 }
4602 }
4603 return 1;
4604 }
4605
4606 static const char rename_limit_warning[] =
4607 "inexact rename detection was skipped due to too many files.";
4608
4609 static const char degrade_cc_to_c_warning[] =
4610 "only found copies from modified paths due to too many files.";
4611
4612 static const char rename_limit_advice[] =
4613 "you may want to set your %s variable to at least "
4614 "%d and retry the command.";
4615
4616 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4617 {
4618 if (degraded_cc)
4619 warning(degrade_cc_to_c_warning);
4620 else if (needed)
4621 warning(rename_limit_warning);
4622 else
4623 return;
4624 if (0 < needed && needed < 32767)
4625 warning(rename_limit_advice, varname, needed);
4626 }
4627
4628 void diff_flush(struct diff_options *options)
4629 {
4630 struct diff_queue_struct *q = &diff_queued_diff;
4631 int i, output_format = options->output_format;
4632 int separator = 0;
4633 int dirstat_by_line = 0;
4634
4635 /*
4636 * Order: raw, stat, summary, patch
4637 * or: name/name-status/checkdiff (other bits clear)
4638 */
4639 if (!q->nr)
4640 goto free_queue;
4641
4642 if (output_format & (DIFF_FORMAT_RAW |
4643 DIFF_FORMAT_NAME |
4644 DIFF_FORMAT_NAME_STATUS |
4645 DIFF_FORMAT_CHECKDIFF)) {
4646 for (i = 0; i < q->nr; i++) {
4647 struct diff_filepair *p = q->queue[i];
4648 if (check_pair_status(p))
4649 flush_one_pair(p, options);
4650 }
4651 separator++;
4652 }
4653
4654 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4655 dirstat_by_line = 1;
4656
4657 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4658 dirstat_by_line) {
4659 struct diffstat_t diffstat;
4660
4661 memset(&diffstat, 0, sizeof(struct diffstat_t));
4662 for (i = 0; i < q->nr; i++) {
4663 struct diff_filepair *p = q->queue[i];
4664 if (check_pair_status(p))
4665 diff_flush_stat(p, options, &diffstat);
4666 }
4667 if (output_format & DIFF_FORMAT_NUMSTAT)
4668 show_numstat(&diffstat, options);
4669 if (output_format & DIFF_FORMAT_DIFFSTAT)
4670 show_stats(&diffstat, options);
4671 if (output_format & DIFF_FORMAT_SHORTSTAT)
4672 show_shortstats(&diffstat, options);
4673 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
4674 show_dirstat_by_line(&diffstat, options);
4675 free_diffstat_info(&diffstat);
4676 separator++;
4677 }
4678 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4679 show_dirstat(options);
4680
4681 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4682 for (i = 0; i < q->nr; i++) {
4683 diff_summary(options, q->queue[i]);
4684 }
4685 separator++;
4686 }
4687
4688 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4689 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4690 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4691 /*
4692 * run diff_flush_patch for the exit status. setting
4693 * options->file to /dev/null should be safe, because we
4694 * aren't supposed to produce any output anyway.
4695 */
4696 if (options->close_file)
4697 fclose(options->file);
4698 options->file = fopen("/dev/null", "w");
4699 if (!options->file)
4700 die_errno("Could not open /dev/null");
4701 options->close_file = 1;
4702 for (i = 0; i < q->nr; i++) {
4703 struct diff_filepair *p = q->queue[i];
4704 if (check_pair_status(p))
4705 diff_flush_patch(p, options);
4706 if (options->found_changes)
4707 break;
4708 }
4709 }
4710
4711 if (output_format & DIFF_FORMAT_PATCH) {
4712 if (separator) {
4713 fprintf(options->file, "%s%c",
4714 diff_line_prefix(options),
4715 options->line_termination);
4716 if (options->stat_sep) {
4717 /* attach patch instead of inline */
4718 fputs(options->stat_sep, options->file);
4719 }
4720 }
4721
4722 for (i = 0; i < q->nr; i++) {
4723 struct diff_filepair *p = q->queue[i];
4724 if (check_pair_status(p))
4725 diff_flush_patch(p, options);
4726 }
4727 }
4728
4729 if (output_format & DIFF_FORMAT_CALLBACK)
4730 options->format_callback(q, options, options->format_callback_data);
4731
4732 for (i = 0; i < q->nr; i++)
4733 diff_free_filepair(q->queue[i]);
4734 free_queue:
4735 free(q->queue);
4736 DIFF_QUEUE_CLEAR(q);
4737 if (options->close_file)
4738 fclose(options->file);
4739
4740 /*
4741 * Report the content-level differences with HAS_CHANGES;
4742 * diff_addremove/diff_change does not set the bit when
4743 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4744 */
4745 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4746 if (options->found_changes)
4747 DIFF_OPT_SET(options, HAS_CHANGES);
4748 else
4749 DIFF_OPT_CLR(options, HAS_CHANGES);
4750 }
4751 }
4752
4753 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
4754 {
4755 return (((p->status == DIFF_STATUS_MODIFIED) &&
4756 ((p->score &&
4757 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
4758 (!p->score &&
4759 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
4760 ((p->status != DIFF_STATUS_MODIFIED) &&
4761 filter_bit_tst(p->status, options)));
4762 }
4763
4764 static void diffcore_apply_filter(struct diff_options *options)
4765 {
4766 int i;
4767 struct diff_queue_struct *q = &diff_queued_diff;
4768 struct diff_queue_struct outq;
4769
4770 DIFF_QUEUE_CLEAR(&outq);
4771
4772 if (!options->filter)
4773 return;
4774
4775 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
4776 int found;
4777 for (i = found = 0; !found && i < q->nr; i++) {
4778 if (match_filter(options, q->queue[i]))
4779 found++;
4780 }
4781 if (found)
4782 return;
4783
4784 /* otherwise we will clear the whole queue
4785 * by copying the empty outq at the end of this
4786 * function, but first clear the current entries
4787 * in the queue.
4788 */
4789 for (i = 0; i < q->nr; i++)
4790 diff_free_filepair(q->queue[i]);
4791 }
4792 else {
4793 /* Only the matching ones */
4794 for (i = 0; i < q->nr; i++) {
4795 struct diff_filepair *p = q->queue[i];
4796 if (match_filter(options, p))
4797 diff_q(&outq, p);
4798 else
4799 diff_free_filepair(p);
4800 }
4801 }
4802 free(q->queue);
4803 *q = outq;
4804 }
4805
4806 /* Check whether two filespecs with the same mode and size are identical */
4807 static int diff_filespec_is_identical(struct diff_filespec *one,
4808 struct diff_filespec *two)
4809 {
4810 if (S_ISGITLINK(one->mode))
4811 return 0;
4812 if (diff_populate_filespec(one, 0))
4813 return 0;
4814 if (diff_populate_filespec(two, 0))
4815 return 0;
4816 return !memcmp(one->data, two->data, one->size);
4817 }
4818
4819 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
4820 {
4821 if (p->done_skip_stat_unmatch)
4822 return p->skip_stat_unmatch_result;
4823
4824 p->done_skip_stat_unmatch = 1;
4825 p->skip_stat_unmatch_result = 0;
4826 /*
4827 * 1. Entries that come from stat info dirtiness
4828 * always have both sides (iow, not create/delete),
4829 * one side of the object name is unknown, with
4830 * the same mode and size. Keep the ones that
4831 * do not match these criteria. They have real
4832 * differences.
4833 *
4834 * 2. At this point, the file is known to be modified,
4835 * with the same mode and size, and the object
4836 * name of one side is unknown. Need to inspect
4837 * the identical contents.
4838 */
4839 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4840 !DIFF_FILE_VALID(p->two) ||
4841 (p->one->oid_valid && p->two->oid_valid) ||
4842 (p->one->mode != p->two->mode) ||
4843 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
4844 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
4845 (p->one->size != p->two->size) ||
4846 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4847 p->skip_stat_unmatch_result = 1;
4848 return p->skip_stat_unmatch_result;
4849 }
4850
4851 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4852 {
4853 int i;
4854 struct diff_queue_struct *q = &diff_queued_diff;
4855 struct diff_queue_struct outq;
4856 DIFF_QUEUE_CLEAR(&outq);
4857
4858 for (i = 0; i < q->nr; i++) {
4859 struct diff_filepair *p = q->queue[i];
4860
4861 if (diff_filespec_check_stat_unmatch(p))
4862 diff_q(&outq, p);
4863 else {
4864 /*
4865 * The caller can subtract 1 from skip_stat_unmatch
4866 * to determine how many paths were dirty only
4867 * due to stat info mismatch.
4868 */
4869 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4870 diffopt->skip_stat_unmatch++;
4871 diff_free_filepair(p);
4872 }
4873 }
4874 free(q->queue);
4875 *q = outq;
4876 }
4877
4878 static int diffnamecmp(const void *a_, const void *b_)
4879 {
4880 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4881 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4882 const char *name_a, *name_b;
4883
4884 name_a = a->one ? a->one->path : a->two->path;
4885 name_b = b->one ? b->one->path : b->two->path;
4886 return strcmp(name_a, name_b);
4887 }
4888
4889 void diffcore_fix_diff_index(struct diff_options *options)
4890 {
4891 struct diff_queue_struct *q = &diff_queued_diff;
4892 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4893 }
4894
4895 void diffcore_std(struct diff_options *options)
4896 {
4897 /* NOTE please keep the following in sync with diff_tree_combined() */
4898 if (options->skip_stat_unmatch)
4899 diffcore_skip_stat_unmatch(options);
4900 if (!options->found_follow) {
4901 /* See try_to_follow_renames() in tree-diff.c */
4902 if (options->break_opt != -1)
4903 diffcore_break(options->break_opt);
4904 if (options->detect_rename)
4905 diffcore_rename(options);
4906 if (options->break_opt != -1)
4907 diffcore_merge_broken();
4908 }
4909 if (options->pickaxe)
4910 diffcore_pickaxe(options);
4911 if (options->orderfile)
4912 diffcore_order(options->orderfile);
4913 if (!options->found_follow)
4914 /* See try_to_follow_renames() in tree-diff.c */
4915 diff_resolve_rename_copy();
4916 diffcore_apply_filter(options);
4917
4918 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4919 DIFF_OPT_SET(options, HAS_CHANGES);
4920 else
4921 DIFF_OPT_CLR(options, HAS_CHANGES);
4922
4923 options->found_follow = 0;
4924 }
4925
4926 int diff_result_code(struct diff_options *opt, int status)
4927 {
4928 int result = 0;
4929
4930 diff_warn_rename_limit("diff.renameLimit",
4931 opt->needed_rename_limit,
4932 opt->degraded_cc_to_c);
4933 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4934 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4935 return status;
4936 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4937 DIFF_OPT_TST(opt, HAS_CHANGES))
4938 result |= 01;
4939 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4940 DIFF_OPT_TST(opt, CHECK_FAILED))
4941 result |= 02;
4942 return result;
4943 }
4944
4945 int diff_can_quit_early(struct diff_options *opt)
4946 {
4947 return (DIFF_OPT_TST(opt, QUICK) &&
4948 !opt->filter &&
4949 DIFF_OPT_TST(opt, HAS_CHANGES));
4950 }
4951
4952 /*
4953 * Shall changes to this submodule be ignored?
4954 *
4955 * Submodule changes can be configured to be ignored separately for each path,
4956 * but that configuration can be overridden from the command line.
4957 */
4958 static int is_submodule_ignored(const char *path, struct diff_options *options)
4959 {
4960 int ignored = 0;
4961 unsigned orig_flags = options->flags;
4962 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4963 set_diffopt_flags_from_submodule_config(options, path);
4964 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4965 ignored = 1;
4966 options->flags = orig_flags;
4967 return ignored;
4968 }
4969
4970 void diff_addremove(struct diff_options *options,
4971 int addremove, unsigned mode,
4972 const unsigned char *sha1,
4973 int sha1_valid,
4974 const char *concatpath, unsigned dirty_submodule)
4975 {
4976 struct diff_filespec *one, *two;
4977
4978 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4979 return;
4980
4981 /* This may look odd, but it is a preparation for
4982 * feeding "there are unchanged files which should
4983 * not produce diffs, but when you are doing copy
4984 * detection you would need them, so here they are"
4985 * entries to the diff-core. They will be prefixed
4986 * with something like '=' or '*' (I haven't decided
4987 * which but should not make any difference).
4988 * Feeding the same new and old to diff_change()
4989 * also has the same effect.
4990 * Before the final output happens, they are pruned after
4991 * merged into rename/copy pairs as appropriate.
4992 */
4993 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4994 addremove = (addremove == '+' ? '-' :
4995 addremove == '-' ? '+' : addremove);
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
5004 if (addremove != '+')
5005 fill_filespec(one, sha1, sha1_valid, mode);
5006 if (addremove != '-') {
5007 fill_filespec(two, sha1, sha1_valid, mode);
5008 two->dirty_submodule = dirty_submodule;
5009 }
5010
5011 diff_queue(&diff_queued_diff, one, two);
5012 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5013 DIFF_OPT_SET(options, HAS_CHANGES);
5014 }
5015
5016 void diff_change(struct diff_options *options,
5017 unsigned old_mode, unsigned new_mode,
5018 const unsigned char *old_sha1,
5019 const unsigned char *new_sha1,
5020 int old_sha1_valid, int new_sha1_valid,
5021 const char *concatpath,
5022 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5023 {
5024 struct diff_filespec *one, *two;
5025 struct diff_filepair *p;
5026
5027 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5028 is_submodule_ignored(concatpath, options))
5029 return;
5030
5031 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5032 unsigned tmp;
5033 const unsigned char *tmp_c;
5034 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
5035 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
5036 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
5037 new_sha1_valid = tmp;
5038 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
5039 new_dirty_submodule = tmp;
5040 }
5041
5042 if (options->prefix &&
5043 strncmp(concatpath, options->prefix, options->prefix_length))
5044 return;
5045
5046 one = alloc_filespec(concatpath);
5047 two = alloc_filespec(concatpath);
5048 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
5049 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
5050 one->dirty_submodule = old_dirty_submodule;
5051 two->dirty_submodule = new_dirty_submodule;
5052 p = diff_queue(&diff_queued_diff, one, two);
5053
5054 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5055 return;
5056
5057 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5058 !diff_filespec_check_stat_unmatch(p))
5059 return;
5060
5061 DIFF_OPT_SET(options, HAS_CHANGES);
5062 }
5063
5064 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5065 {
5066 struct diff_filepair *pair;
5067 struct diff_filespec *one, *two;
5068
5069 if (options->prefix &&
5070 strncmp(path, options->prefix, options->prefix_length))
5071 return NULL;
5072
5073 one = alloc_filespec(path);
5074 two = alloc_filespec(path);
5075 pair = diff_queue(&diff_queued_diff, one, two);
5076 pair->is_unmerged = 1;
5077 return pair;
5078 }
5079
5080 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5081 size_t *outsize)
5082 {
5083 struct diff_tempfile *temp;
5084 const char *argv[3];
5085 const char **arg = argv;
5086 struct child_process child = CHILD_PROCESS_INIT;
5087 struct strbuf buf = STRBUF_INIT;
5088 int err = 0;
5089
5090 temp = prepare_temp_file(spec->path, spec);
5091 *arg++ = pgm;
5092 *arg++ = temp->name;
5093 *arg = NULL;
5094
5095 child.use_shell = 1;
5096 child.argv = argv;
5097 child.out = -1;
5098 if (start_command(&child)) {
5099 remove_tempfile();
5100 return NULL;
5101 }
5102
5103 if (strbuf_read(&buf, child.out, 0) < 0)
5104 err = error("error reading from textconv command '%s'", pgm);
5105 close(child.out);
5106
5107 if (finish_command(&child) || err) {
5108 strbuf_release(&buf);
5109 remove_tempfile();
5110 return NULL;
5111 }
5112 remove_tempfile();
5113
5114 return strbuf_detach(&buf, outsize);
5115 }
5116
5117 size_t fill_textconv(struct userdiff_driver *driver,
5118 struct diff_filespec *df,
5119 char **outbuf)
5120 {
5121 size_t size;
5122
5123 if (!driver) {
5124 if (!DIFF_FILE_VALID(df)) {
5125 *outbuf = "";
5126 return 0;
5127 }
5128 if (diff_populate_filespec(df, 0))
5129 die("unable to read files to diff");
5130 *outbuf = df->data;
5131 return df->size;
5132 }
5133
5134 if (!driver->textconv)
5135 die("BUG: fill_textconv called with non-textconv driver");
5136
5137 if (driver->textconv_cache && df->oid_valid) {
5138 *outbuf = notes_cache_get(driver->textconv_cache,
5139 df->oid.hash,
5140 &size);
5141 if (*outbuf)
5142 return size;
5143 }
5144
5145 *outbuf = run_textconv(driver->textconv, df, &size);
5146 if (!*outbuf)
5147 die("unable to read files to diff");
5148
5149 if (driver->textconv_cache && df->oid_valid) {
5150 /* ignore errors, as we might be in a readonly repository */
5151 notes_cache_put(driver->textconv_cache, df->oid.hash, *outbuf,
5152 size);
5153 /*
5154 * we could save up changes and flush them all at the end,
5155 * but we would need an extra call after all diffing is done.
5156 * Since generating a cache entry is the slow path anyway,
5157 * this extra overhead probably isn't a big deal.
5158 */
5159 notes_cache_write(driver->textconv_cache);
5160 }
5161
5162 return size;
5163 }
5164
5165 void setup_diff_pager(struct diff_options *opt)
5166 {
5167 /*
5168 * If the user asked for our exit code, then either they want --quiet
5169 * or --exit-code. We should definitely not bother with a pager in the
5170 * former case, as we will generate no output. Since we still properly
5171 * report our exit code even when a pager is run, we _could_ run a
5172 * pager with --exit-code. But since we have not done so historically,
5173 * and because it is easy to find people oneline advising "git diff
5174 * --exit-code" in hooks and other scripts, we do not do so.
5175 */
5176 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5177 check_pager_config("diff") != 0)
5178 setup_pager();
5179 }
The Best VCS Around