?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'ta/pretty-parse-config' into jch
[git/jrn.git] / diff.c
blob7565f0719362ac2ff950dcb080cc961d64a6987b
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule-config.h"
17 #include "submodule.h"
18 #include "ll-merge.h"
19 #include "string-list.h"
20 #include "argv-array.h"
21
22 #ifdef NO_FAST_WORKING_DIRECTORY
23 #define FAST_WORKING_DIRECTORY 0
24 #else
25 #define FAST_WORKING_DIRECTORY 1
26 #endif
27
28 static int diff_detect_rename_default;
29 static int diff_rename_limit_default = 400;
30 static int diff_suppress_blank_empty;
31 static int diff_use_color_default = -1;
32 static int diff_context_default = 3;
33 static const char *diff_word_regex_cfg;
34 static const char *external_diff_cmd_cfg;
35 static const char *diff_order_file_cfg;
36 int diff_auto_refresh_index = 1;
37 static int diff_mnemonic_prefix;
38 static int diff_no_prefix;
39 static int diff_stat_graph_width;
40 static int diff_dirstat_permille_default = 30;
41 static struct diff_options default_diff_options;
42 static long diff_algorithm;
43
44 static char diff_colors[][COLOR_MAXLEN] = {
45 GIT_COLOR_RESET,
46 GIT_COLOR_NORMAL, /* PLAIN */
47 GIT_COLOR_BOLD, /* METAINFO */
48 GIT_COLOR_CYAN, /* FRAGINFO */
49 GIT_COLOR_RED, /* OLD */
50 GIT_COLOR_GREEN, /* NEW */
51 GIT_COLOR_YELLOW, /* COMMIT */
52 GIT_COLOR_BG_RED, /* WHITESPACE */
53 GIT_COLOR_NORMAL, /* FUNCINFO */
54 };
55
56 static int parse_diff_color_slot(const char *var)
57 {
58 if (!strcasecmp(var, "plain"))
59 return DIFF_PLAIN;
60 if (!strcasecmp(var, "meta"))
61 return DIFF_METAINFO;
62 if (!strcasecmp(var, "frag"))
63 return DIFF_FRAGINFO;
64 if (!strcasecmp(var, "old"))
65 return DIFF_FILE_OLD;
66 if (!strcasecmp(var, "new"))
67 return DIFF_FILE_NEW;
68 if (!strcasecmp(var, "commit"))
69 return DIFF_COMMIT;
70 if (!strcasecmp(var, "whitespace"))
71 return DIFF_WHITESPACE;
72 if (!strcasecmp(var, "func"))
73 return DIFF_FUNCINFO;
74 return -1;
75 }
76
77 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
78 struct strbuf *errmsg)
79 {
80 char *params_copy = xstrdup(params_string);
81 struct string_list params = STRING_LIST_INIT_NODUP;
82 int ret = 0;
83 int i;
84
85 if (*params_copy)
86 string_list_split_in_place(&params, params_copy, ',', -1);
87 for (i = 0; i < params.nr; i++) {
88 const char *p = params.items[i].string;
89 if (!strcmp(p, "changes")) {
90 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
91 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
92 } else if (!strcmp(p, "lines")) {
93 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
94 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
95 } else if (!strcmp(p, "files")) {
96 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
97 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
98 } else if (!strcmp(p, "noncumulative")) {
99 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
100 } else if (!strcmp(p, "cumulative")) {
101 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
102 } else if (isdigit(*p)) {
103 char *end;
104 int permille = strtoul(p, &end, 10) * 10;
105 if (*end == '.' && isdigit(*++end)) {
106 /* only use first digit */
107 permille += *end - '0';
108 /* .. and ignore any further digits */
109 while (isdigit(*++end))
110 ; /* nothing */
111 }
112 if (!*end)
113 options->dirstat_permille = permille;
114 else {
115 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
116 p);
117 ret++;
118 }
119 } else {
120 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
121 ret++;
122 }
123
124 }
125 string_list_clear(&params, 0);
126 free(params_copy);
127 return ret;
128 }
129
130 static int parse_submodule_params(struct diff_options *options, const char *value)
131 {
132 if (!strcmp(value, "log"))
133 DIFF_OPT_SET(options, SUBMODULE_LOG);
134 else if (!strcmp(value, "short"))
135 DIFF_OPT_CLR(options, SUBMODULE_LOG);
136 else
137 return -1;
138 return 0;
139 }
140
141 static int git_config_rename(const char *var, const char *value)
142 {
143 if (!value)
144 return DIFF_DETECT_RENAME;
145 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
146 return DIFF_DETECT_COPY;
147 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
148 }
149
150 long parse_algorithm_value(const char *value)
151 {
152 if (!value)
153 return -1;
154 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
155 return 0;
156 else if (!strcasecmp(value, "minimal"))
157 return XDF_NEED_MINIMAL;
158 else if (!strcasecmp(value, "patience"))
159 return XDF_PATIENCE_DIFF;
160 else if (!strcasecmp(value, "histogram"))
161 return XDF_HISTOGRAM_DIFF;
162 return -1;
163 }
164
165 /*
166 * These are to give UI layer defaults.
167 * The core-level commands such as git-diff-files should
168 * never be affected by the setting of diff.renames
169 * the user happens to have in the configuration file.
170 */
171 int git_diff_ui_config(const char *var, const char *value, void *cb)
172 {
173 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
174 diff_use_color_default = git_config_colorbool(var, value);
175 return 0;
176 }
177 if (!strcmp(var, "diff.context")) {
178 diff_context_default = git_config_int(var, value);
179 if (diff_context_default < 0)
180 return -1;
181 return 0;
182 }
183 if (!strcmp(var, "diff.renames")) {
184 diff_detect_rename_default = git_config_rename(var, value);
185 return 0;
186 }
187 if (!strcmp(var, "diff.autorefreshindex")) {
188 diff_auto_refresh_index = git_config_bool(var, value);
189 return 0;
190 }
191 if (!strcmp(var, "diff.mnemonicprefix")) {
192 diff_mnemonic_prefix = git_config_bool(var, value);
193 return 0;
194 }
195 if (!strcmp(var, "diff.noprefix")) {
196 diff_no_prefix = git_config_bool(var, value);
197 return 0;
198 }
199 if (!strcmp(var, "diff.statgraphwidth")) {
200 diff_stat_graph_width = git_config_int(var, value);
201 return 0;
202 }
203 if (!strcmp(var, "diff.external"))
204 return git_config_string(&external_diff_cmd_cfg, var, value);
205 if (!strcmp(var, "diff.wordregex"))
206 return git_config_string(&diff_word_regex_cfg, var, value);
207 if (!strcmp(var, "diff.orderfile"))
208 return git_config_pathname(&diff_order_file_cfg, var, value);
209
210 if (!strcmp(var, "diff.ignoresubmodules"))
211 handle_ignore_submodules_arg(&default_diff_options, value);
212
213 if (!strcmp(var, "diff.submodule")) {
214 if (parse_submodule_params(&default_diff_options, value))
215 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
216 value);
217 return 0;
218 }
219
220 if (!strcmp(var, "diff.algorithm")) {
221 diff_algorithm = parse_algorithm_value(value);
222 if (diff_algorithm < 0)
223 return -1;
224 return 0;
225 }
226
227 if (git_color_config(var, value, cb) < 0)
228 return -1;
229
230 return git_diff_basic_config(var, value, cb);
231 }
232
233 int git_diff_basic_config(const char *var, const char *value, void *cb)
234 {
235 const char *name;
236
237 if (!strcmp(var, "diff.renamelimit")) {
238 diff_rename_limit_default = git_config_int(var, value);
239 return 0;
240 }
241
242 if (userdiff_config(var, value) < 0)
243 return -1;
244
245 if (skip_prefix(var, "diff.color.", &name) ||
246 skip_prefix(var, "color.diff.", &name)) {
247 int slot = parse_diff_color_slot(name);
248 if (slot < 0)
249 return 0;
250 if (!value)
251 return config_error_nonbool(var);
252 color_parse(value, var, diff_colors[slot]);
253 return 0;
254 }
255
256 /* like GNU diff's --suppress-blank-empty option */
257 if (!strcmp(var, "diff.suppressblankempty") ||
258 /* for backwards compatibility */
259 !strcmp(var, "diff.suppress-blank-empty")) {
260 diff_suppress_blank_empty = git_config_bool(var, value);
261 return 0;
262 }
263
264 if (!strcmp(var, "diff.dirstat")) {
265 struct strbuf errmsg = STRBUF_INIT;
266 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
267 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
268 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
269 errmsg.buf);
270 strbuf_release(&errmsg);
271 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
272 return 0;
273 }
274
275 if (starts_with(var, "submodule."))
276 return parse_submodule_config_option(var, value);
277
278 return git_default_config(var, value, cb);
279 }
280
281 static char *quote_two(const char *one, const char *two)
282 {
283 int need_one = quote_c_style(one, NULL, NULL, 1);
284 int need_two = quote_c_style(two, NULL, NULL, 1);
285 struct strbuf res = STRBUF_INIT;
286
287 if (need_one + need_two) {
288 strbuf_addch(&res, '"');
289 quote_c_style(one, &res, NULL, 1);
290 quote_c_style(two, &res, NULL, 1);
291 strbuf_addch(&res, '"');
292 } else {
293 strbuf_addstr(&res, one);
294 strbuf_addstr(&res, two);
295 }
296 return strbuf_detach(&res, NULL);
297 }
298
299 static const char *external_diff(void)
300 {
301 static const char *external_diff_cmd = NULL;
302 static int done_preparing = 0;
303
304 if (done_preparing)
305 return external_diff_cmd;
306 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
307 if (!external_diff_cmd)
308 external_diff_cmd = external_diff_cmd_cfg;
309 done_preparing = 1;
310 return external_diff_cmd;
311 }
312
313 static struct diff_tempfile {
314 const char *name; /* filename external diff should read from */
315 char hex[41];
316 char mode[10];
317 char tmp_path[PATH_MAX];
318 } diff_temp[2];
319
320 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
321
322 struct emit_callback {
323 int color_diff;
324 unsigned ws_rule;
325 int blank_at_eof_in_preimage;
326 int blank_at_eof_in_postimage;
327 int lno_in_preimage;
328 int lno_in_postimage;
329 sane_truncate_fn truncate;
330 const char **label_path;
331 struct diff_words_data *diff_words;
332 struct diff_options *opt;
333 int *found_changesp;
334 struct strbuf *header;
335 };
336
337 static int count_lines(const char *data, int size)
338 {
339 int count, ch, completely_empty = 1, nl_just_seen = 0;
340 count = 0;
341 while (0 < size--) {
342 ch = *data++;
343 if (ch == '\n') {
344 count++;
345 nl_just_seen = 1;
346 completely_empty = 0;
347 }
348 else {
349 nl_just_seen = 0;
350 completely_empty = 0;
351 }
352 }
353 if (completely_empty)
354 return 0;
355 if (!nl_just_seen)
356 count++; /* no trailing newline */
357 return count;
358 }
359
360 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
361 {
362 if (!DIFF_FILE_VALID(one)) {
363 mf->ptr = (char *)""; /* does not matter */
364 mf->size = 0;
365 return 0;
366 }
367 else if (diff_populate_filespec(one, 0))
368 return -1;
369
370 mf->ptr = one->data;
371 mf->size = one->size;
372 return 0;
373 }
374
375 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
376 static unsigned long diff_filespec_size(struct diff_filespec *one)
377 {
378 if (!DIFF_FILE_VALID(one))
379 return 0;
380 diff_populate_filespec(one, 1);
381 return one->size;
382 }
383
384 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
385 {
386 char *ptr = mf->ptr;
387 long size = mf->size;
388 int cnt = 0;
389
390 if (!size)
391 return cnt;
392 ptr += size - 1; /* pointing at the very end */
393 if (*ptr != '\n')
394 ; /* incomplete line */
395 else
396 ptr--; /* skip the last LF */
397 while (mf->ptr < ptr) {
398 char *prev_eol;
399 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
400 if (*prev_eol == '\n')
401 break;
402 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
403 break;
404 cnt++;
405 ptr = prev_eol - 1;
406 }
407 return cnt;
408 }
409
410 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
411 struct emit_callback *ecbdata)
412 {
413 int l1, l2, at;
414 unsigned ws_rule = ecbdata->ws_rule;
415 l1 = count_trailing_blank(mf1, ws_rule);
416 l2 = count_trailing_blank(mf2, ws_rule);
417 if (l2 <= l1) {
418 ecbdata->blank_at_eof_in_preimage = 0;
419 ecbdata->blank_at_eof_in_postimage = 0;
420 return;
421 }
422 at = count_lines(mf1->ptr, mf1->size);
423 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
424
425 at = count_lines(mf2->ptr, mf2->size);
426 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
427 }
428
429 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
430 int first, const char *line, int len)
431 {
432 int has_trailing_newline, has_trailing_carriage_return;
433 int nofirst;
434 FILE *file = o->file;
435
436 fputs(diff_line_prefix(o), file);
437
438 if (len == 0) {
439 has_trailing_newline = (first == '\n');
440 has_trailing_carriage_return = (!has_trailing_newline &&
441 (first == '\r'));
442 nofirst = has_trailing_newline || has_trailing_carriage_return;
443 } else {
444 has_trailing_newline = (len > 0 && line[len-1] == '\n');
445 if (has_trailing_newline)
446 len--;
447 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
448 if (has_trailing_carriage_return)
449 len--;
450 nofirst = 0;
451 }
452
453 if (len || !nofirst) {
454 fputs(set, file);
455 if (!nofirst)
456 fputc(first, file);
457 fwrite(line, len, 1, file);
458 fputs(reset, file);
459 }
460 if (has_trailing_carriage_return)
461 fputc('\r', file);
462 if (has_trailing_newline)
463 fputc('\n', file);
464 }
465
466 static void emit_line(struct diff_options *o, const char *set, const char *reset,
467 const char *line, int len)
468 {
469 emit_line_0(o, set, reset, line[0], line+1, len-1);
470 }
471
472 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
473 {
474 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
475 ecbdata->blank_at_eof_in_preimage &&
476 ecbdata->blank_at_eof_in_postimage &&
477 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
478 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
479 return 0;
480 return ws_blank_line(line, len, ecbdata->ws_rule);
481 }
482
483 static void emit_add_line(const char *reset,
484 struct emit_callback *ecbdata,
485 const char *line, int len)
486 {
487 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
488 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
489
490 if (!*ws)
491 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
492 else if (new_blank_line_at_eof(ecbdata, line, len))
493 /* Blank line at EOF - paint '+' as well */
494 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
495 else {
496 /* Emit just the prefix, then the rest. */
497 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
498 ws_check_emit(line, len, ecbdata->ws_rule,
499 ecbdata->opt->file, set, reset, ws);
500 }
501 }
502
503 static void emit_hunk_header(struct emit_callback *ecbdata,
504 const char *line, int len)
505 {
506 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
507 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
508 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
509 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
510 static const char atat[2] = { '@', '@' };
511 const char *cp, *ep;
512 struct strbuf msgbuf = STRBUF_INIT;
513 int org_len = len;
514 int i = 1;
515
516 /*
517 * As a hunk header must begin with "@@ -<old>, +<new> @@",
518 * it always is at least 10 bytes long.
519 */
520 if (len < 10 ||
521 memcmp(line, atat, 2) ||
522 !(ep = memmem(line + 2, len - 2, atat, 2))) {
523 emit_line(ecbdata->opt, plain, reset, line, len);
524 return;
525 }
526 ep += 2; /* skip over @@ */
527
528 /* The hunk header in fraginfo color */
529 strbuf_addstr(&msgbuf, frag);
530 strbuf_add(&msgbuf, line, ep - line);
531 strbuf_addstr(&msgbuf, reset);
532
533 /*
534 * trailing "\r\n"
535 */
536 for ( ; i < 3; i++)
537 if (line[len - i] == '\r' || line[len - i] == '\n')
538 len--;
539
540 /* blank before the func header */
541 for (cp = ep; ep - line < len; ep++)
542 if (*ep != ' ' && *ep != '\t')
543 break;
544 if (ep != cp) {
545 strbuf_addstr(&msgbuf, plain);
546 strbuf_add(&msgbuf, cp, ep - cp);
547 strbuf_addstr(&msgbuf, reset);
548 }
549
550 if (ep < line + len) {
551 strbuf_addstr(&msgbuf, func);
552 strbuf_add(&msgbuf, ep, line + len - ep);
553 strbuf_addstr(&msgbuf, reset);
554 }
555
556 strbuf_add(&msgbuf, line + len, org_len - len);
557 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
558 strbuf_release(&msgbuf);
559 }
560
561 static struct diff_tempfile *claim_diff_tempfile(void) {
562 int i;
563 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
564 if (!diff_temp[i].name)
565 return diff_temp + i;
566 die("BUG: diff is failing to clean up its tempfiles");
567 }
568
569 static int remove_tempfile_installed;
570
571 static void remove_tempfile(void)
572 {
573 int i;
574 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
575 if (diff_temp[i].name == diff_temp[i].tmp_path)
576 unlink_or_warn(diff_temp[i].name);
577 diff_temp[i].name = NULL;
578 }
579 }
580
581 static void remove_tempfile_on_signal(int signo)
582 {
583 remove_tempfile();
584 sigchain_pop(signo);
585 raise(signo);
586 }
587
588 static void print_line_count(FILE *file, int count)
589 {
590 switch (count) {
591 case 0:
592 fprintf(file, "0,0");
593 break;
594 case 1:
595 fprintf(file, "1");
596 break;
597 default:
598 fprintf(file, "1,%d", count);
599 break;
600 }
601 }
602
603 static void emit_rewrite_lines(struct emit_callback *ecb,
604 int prefix, const char *data, int size)
605 {
606 const char *endp = NULL;
607 static const char *nneof = " No newline at end of file\n";
608 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
609 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
610
611 while (0 < size) {
612 int len;
613
614 endp = memchr(data, '\n', size);
615 len = endp ? (endp - data + 1) : size;
616 if (prefix != '+') {
617 ecb->lno_in_preimage++;
618 emit_line_0(ecb->opt, old, reset, '-',
619 data, len);
620 } else {
621 ecb->lno_in_postimage++;
622 emit_add_line(reset, ecb, data, len);
623 }
624 size -= len;
625 data += len;
626 }
627 if (!endp) {
628 const char *plain = diff_get_color(ecb->color_diff,
629 DIFF_PLAIN);
630 putc('\n', ecb->opt->file);
631 emit_line_0(ecb->opt, plain, reset, '\\',
632 nneof, strlen(nneof));
633 }
634 }
635
636 static void emit_rewrite_diff(const char *name_a,
637 const char *name_b,
638 struct diff_filespec *one,
639 struct diff_filespec *two,
640 struct userdiff_driver *textconv_one,
641 struct userdiff_driver *textconv_two,
642 struct diff_options *o)
643 {
644 int lc_a, lc_b;
645 const char *name_a_tab, *name_b_tab;
646 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
647 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
648 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
649 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
650 const char *a_prefix, *b_prefix;
651 char *data_one, *data_two;
652 size_t size_one, size_two;
653 struct emit_callback ecbdata;
654 const char *line_prefix = diff_line_prefix(o);
655
656 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
657 a_prefix = o->b_prefix;
658 b_prefix = o->a_prefix;
659 } else {
660 a_prefix = o->a_prefix;
661 b_prefix = o->b_prefix;
662 }
663
664 name_a += (*name_a == '/');
665 name_b += (*name_b == '/');
666 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
667 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
668
669 strbuf_reset(&a_name);
670 strbuf_reset(&b_name);
671 quote_two_c_style(&a_name, a_prefix, name_a, 0);
672 quote_two_c_style(&b_name, b_prefix, name_b, 0);
673
674 size_one = fill_textconv(textconv_one, one, &data_one);
675 size_two = fill_textconv(textconv_two, two, &data_two);
676
677 memset(&ecbdata, 0, sizeof(ecbdata));
678 ecbdata.color_diff = want_color(o->use_color);
679 ecbdata.found_changesp = &o->found_changes;
680 ecbdata.ws_rule = whitespace_rule(name_b);
681 ecbdata.opt = o;
682 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
683 mmfile_t mf1, mf2;
684 mf1.ptr = (char *)data_one;
685 mf2.ptr = (char *)data_two;
686 mf1.size = size_one;
687 mf2.size = size_two;
688 check_blank_at_eof(&mf1, &mf2, &ecbdata);
689 }
690 ecbdata.lno_in_preimage = 1;
691 ecbdata.lno_in_postimage = 1;
692
693 lc_a = count_lines(data_one, size_one);
694 lc_b = count_lines(data_two, size_two);
695 fprintf(o->file,
696 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
697 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
698 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
699 line_prefix, fraginfo);
700 if (!o->irreversible_delete)
701 print_line_count(o->file, lc_a);
702 else
703 fprintf(o->file, "?,?");
704 fprintf(o->file, " +");
705 print_line_count(o->file, lc_b);
706 fprintf(o->file, " @@%s\n", reset);
707 if (lc_a && !o->irreversible_delete)
708 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
709 if (lc_b)
710 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
711 if (textconv_one)
712 free((char *)data_one);
713 if (textconv_two)
714 free((char *)data_two);
715 }
716
717 struct diff_words_buffer {
718 mmfile_t text;
719 long alloc;
720 struct diff_words_orig {
721 const char *begin, *end;
722 } *orig;
723 int orig_nr, orig_alloc;
724 };
725
726 static void diff_words_append(char *line, unsigned long len,
727 struct diff_words_buffer *buffer)
728 {
729 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
730 line++;
731 len--;
732 memcpy(buffer->text.ptr + buffer->text.size, line, len);
733 buffer->text.size += len;
734 buffer->text.ptr[buffer->text.size] = '\0';
735 }
736
737 struct diff_words_style_elem {
738 const char *prefix;
739 const char *suffix;
740 const char *color; /* NULL; filled in by the setup code if
741 * color is enabled */
742 };
743
744 struct diff_words_style {
745 enum diff_words_type type;
746 struct diff_words_style_elem new, old, ctx;
747 const char *newline;
748 };
749
750 static struct diff_words_style diff_words_styles[] = {
751 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
752 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
753 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
754 };
755
756 struct diff_words_data {
757 struct diff_words_buffer minus, plus;
758 const char *current_plus;
759 int last_minus;
760 struct diff_options *opt;
761 regex_t *word_regex;
762 enum diff_words_type type;
763 struct diff_words_style *style;
764 };
765
766 static int fn_out_diff_words_write_helper(FILE *fp,
767 struct diff_words_style_elem *st_el,
768 const char *newline,
769 size_t count, const char *buf,
770 const char *line_prefix)
771 {
772 int print = 0;
773
774 while (count) {
775 char *p = memchr(buf, '\n', count);
776 if (print)
777 fputs(line_prefix, fp);
778 if (p != buf) {
779 if (st_el->color && fputs(st_el->color, fp) < 0)
780 return -1;
781 if (fputs(st_el->prefix, fp) < 0 ||
782 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
783 fputs(st_el->suffix, fp) < 0)
784 return -1;
785 if (st_el->color && *st_el->color
786 && fputs(GIT_COLOR_RESET, fp) < 0)
787 return -1;
788 }
789 if (!p)
790 return 0;
791 if (fputs(newline, fp) < 0)
792 return -1;
793 count -= p + 1 - buf;
794 buf = p + 1;
795 print = 1;
796 }
797 return 0;
798 }
799
800 /*
801 * '--color-words' algorithm can be described as:
802 *
803 * 1. collect a the minus/plus lines of a diff hunk, divided into
804 * minus-lines and plus-lines;
805 *
806 * 2. break both minus-lines and plus-lines into words and
807 * place them into two mmfile_t with one word for each line;
808 *
809 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
810 *
811 * And for the common parts of the both file, we output the plus side text.
812 * diff_words->current_plus is used to trace the current position of the plus file
813 * which printed. diff_words->last_minus is used to trace the last minus word
814 * printed.
815 *
816 * For '--graph' to work with '--color-words', we need to output the graph prefix
817 * on each line of color words output. Generally, there are two conditions on
818 * which we should output the prefix.
819 *
820 * 1. diff_words->last_minus == 0 &&
821 * diff_words->current_plus == diff_words->plus.text.ptr
822 *
823 * that is: the plus text must start as a new line, and if there is no minus
824 * word printed, a graph prefix must be printed.
825 *
826 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
827 * *(diff_words->current_plus - 1) == '\n'
828 *
829 * that is: a graph prefix must be printed following a '\n'
830 */
831 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
832 {
833 if ((diff_words->last_minus == 0 &&
834 diff_words->current_plus == diff_words->plus.text.ptr) ||
835 (diff_words->current_plus > diff_words->plus.text.ptr &&
836 *(diff_words->current_plus - 1) == '\n')) {
837 return 1;
838 } else {
839 return 0;
840 }
841 }
842
843 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
844 {
845 struct diff_words_data *diff_words = priv;
846 struct diff_words_style *style = diff_words->style;
847 int minus_first, minus_len, plus_first, plus_len;
848 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
849 struct diff_options *opt = diff_words->opt;
850 const char *line_prefix;
851
852 if (line[0] != '@' || parse_hunk_header(line, len,
853 &minus_first, &minus_len, &plus_first, &plus_len))
854 return;
855
856 assert(opt);
857 line_prefix = diff_line_prefix(opt);
858
859 /* POSIX requires that first be decremented by one if len == 0... */
860 if (minus_len) {
861 minus_begin = diff_words->minus.orig[minus_first].begin;
862 minus_end =
863 diff_words->minus.orig[minus_first + minus_len - 1].end;
864 } else
865 minus_begin = minus_end =
866 diff_words->minus.orig[minus_first].end;
867
868 if (plus_len) {
869 plus_begin = diff_words->plus.orig[plus_first].begin;
870 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
871 } else
872 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
873
874 if (color_words_output_graph_prefix(diff_words)) {
875 fputs(line_prefix, diff_words->opt->file);
876 }
877 if (diff_words->current_plus != plus_begin) {
878 fn_out_diff_words_write_helper(diff_words->opt->file,
879 &style->ctx, style->newline,
880 plus_begin - diff_words->current_plus,
881 diff_words->current_plus, line_prefix);
882 if (*(plus_begin - 1) == '\n')
883 fputs(line_prefix, diff_words->opt->file);
884 }
885 if (minus_begin != minus_end) {
886 fn_out_diff_words_write_helper(diff_words->opt->file,
887 &style->old, style->newline,
888 minus_end - minus_begin, minus_begin,
889 line_prefix);
890 }
891 if (plus_begin != plus_end) {
892 fn_out_diff_words_write_helper(diff_words->opt->file,
893 &style->new, style->newline,
894 plus_end - plus_begin, plus_begin,
895 line_prefix);
896 }
897
898 diff_words->current_plus = plus_end;
899 diff_words->last_minus = minus_first;
900 }
901
902 /* This function starts looking at *begin, and returns 0 iff a word was found. */
903 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
904 int *begin, int *end)
905 {
906 if (word_regex && *begin < buffer->size) {
907 regmatch_t match[1];
908 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
909 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
910 '\n', match[0].rm_eo - match[0].rm_so);
911 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
912 *begin += match[0].rm_so;
913 return *begin >= *end;
914 }
915 return -1;
916 }
917
918 /* find the next word */
919 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
920 (*begin)++;
921 if (*begin >= buffer->size)
922 return -1;
923
924 /* find the end of the word */
925 *end = *begin + 1;
926 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
927 (*end)++;
928
929 return 0;
930 }
931
932 /*
933 * This function splits the words in buffer->text, stores the list with
934 * newline separator into out, and saves the offsets of the original words
935 * in buffer->orig.
936 */
937 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
938 regex_t *word_regex)
939 {
940 int i, j;
941 long alloc = 0;
942
943 out->size = 0;
944 out->ptr = NULL;
945
946 /* fake an empty "0th" word */
947 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
948 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
949 buffer->orig_nr = 1;
950
951 for (i = 0; i < buffer->text.size; i++) {
952 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
953 return;
954
955 /* store original boundaries */
956 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
957 buffer->orig_alloc);
958 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
959 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
960 buffer->orig_nr++;
961
962 /* store one word */
963 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
964 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
965 out->ptr[out->size + j - i] = '\n';
966 out->size += j - i + 1;
967
968 i = j - 1;
969 }
970 }
971
972 /* this executes the word diff on the accumulated buffers */
973 static void diff_words_show(struct diff_words_data *diff_words)
974 {
975 xpparam_t xpp;
976 xdemitconf_t xecfg;
977 mmfile_t minus, plus;
978 struct diff_words_style *style = diff_words->style;
979
980 struct diff_options *opt = diff_words->opt;
981 const char *line_prefix;
982
983 assert(opt);
984 line_prefix = diff_line_prefix(opt);
985
986 /* special case: only removal */
987 if (!diff_words->plus.text.size) {
988 fputs(line_prefix, diff_words->opt->file);
989 fn_out_diff_words_write_helper(diff_words->opt->file,
990 &style->old, style->newline,
991 diff_words->minus.text.size,
992 diff_words->minus.text.ptr, line_prefix);
993 diff_words->minus.text.size = 0;
994 return;
995 }
996
997 diff_words->current_plus = diff_words->plus.text.ptr;
998 diff_words->last_minus = 0;
999
1000 memset(&xpp, 0, sizeof(xpp));
1001 memset(&xecfg, 0, sizeof(xecfg));
1002 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1003 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1004 xpp.flags = 0;
1005 /* as only the hunk header will be parsed, we need a 0-context */
1006 xecfg.ctxlen = 0;
1007 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1008 &xpp, &xecfg);
1009 free(minus.ptr);
1010 free(plus.ptr);
1011 if (diff_words->current_plus != diff_words->plus.text.ptr +
1012 diff_words->plus.text.size) {
1013 if (color_words_output_graph_prefix(diff_words))
1014 fputs(line_prefix, diff_words->opt->file);
1015 fn_out_diff_words_write_helper(diff_words->opt->file,
1016 &style->ctx, style->newline,
1017 diff_words->plus.text.ptr + diff_words->plus.text.size
1018 - diff_words->current_plus, diff_words->current_plus,
1019 line_prefix);
1020 }
1021 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1022 }
1023
1024 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1025 static void diff_words_flush(struct emit_callback *ecbdata)
1026 {
1027 if (ecbdata->diff_words->minus.text.size ||
1028 ecbdata->diff_words->plus.text.size)
1029 diff_words_show(ecbdata->diff_words);
1030 }
1031
1032 static void diff_filespec_load_driver(struct diff_filespec *one)
1033 {
1034 /* Use already-loaded driver */
1035 if (one->driver)
1036 return;
1037
1038 if (S_ISREG(one->mode))
1039 one->driver = userdiff_find_by_path(one->path);
1040
1041 /* Fallback to default settings */
1042 if (!one->driver)
1043 one->driver = userdiff_find_by_name("default");
1044 }
1045
1046 static const char *userdiff_word_regex(struct diff_filespec *one)
1047 {
1048 diff_filespec_load_driver(one);
1049 return one->driver->word_regex;
1050 }
1051
1052 static void init_diff_words_data(struct emit_callback *ecbdata,
1053 struct diff_options *orig_opts,
1054 struct diff_filespec *one,
1055 struct diff_filespec *two)
1056 {
1057 int i;
1058 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1059 memcpy(o, orig_opts, sizeof(struct diff_options));
1060
1061 ecbdata->diff_words =
1062 xcalloc(1, sizeof(struct diff_words_data));
1063 ecbdata->diff_words->type = o->word_diff;
1064 ecbdata->diff_words->opt = o;
1065 if (!o->word_regex)
1066 o->word_regex = userdiff_word_regex(one);
1067 if (!o->word_regex)
1068 o->word_regex = userdiff_word_regex(two);
1069 if (!o->word_regex)
1070 o->word_regex = diff_word_regex_cfg;
1071 if (o->word_regex) {
1072 ecbdata->diff_words->word_regex = (regex_t *)
1073 xmalloc(sizeof(regex_t));
1074 if (regcomp(ecbdata->diff_words->word_regex,
1075 o->word_regex,
1076 REG_EXTENDED | REG_NEWLINE))
1077 die ("Invalid regular expression: %s",
1078 o->word_regex);
1079 }
1080 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1081 if (o->word_diff == diff_words_styles[i].type) {
1082 ecbdata->diff_words->style =
1083 &diff_words_styles[i];
1084 break;
1085 }
1086 }
1087 if (want_color(o->use_color)) {
1088 struct diff_words_style *st = ecbdata->diff_words->style;
1089 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1090 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1091 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
1092 }
1093 }
1094
1095 static void free_diff_words_data(struct emit_callback *ecbdata)
1096 {
1097 if (ecbdata->diff_words) {
1098 diff_words_flush(ecbdata);
1099 free (ecbdata->diff_words->opt);
1100 free (ecbdata->diff_words->minus.text.ptr);
1101 free (ecbdata->diff_words->minus.orig);
1102 free (ecbdata->diff_words->plus.text.ptr);
1103 free (ecbdata->diff_words->plus.orig);
1104 if (ecbdata->diff_words->word_regex) {
1105 regfree(ecbdata->diff_words->word_regex);
1106 free(ecbdata->diff_words->word_regex);
1107 }
1108 free(ecbdata->diff_words);
1109 ecbdata->diff_words = NULL;
1110 }
1111 }
1112
1113 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1114 {
1115 if (want_color(diff_use_color))
1116 return diff_colors[ix];
1117 return "";
1118 }
1119
1120 const char *diff_line_prefix(struct diff_options *opt)
1121 {
1122 struct strbuf *msgbuf;
1123 if (!opt->output_prefix)
1124 return "";
1125
1126 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1127 return msgbuf->buf;
1128 }
1129
1130 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1131 {
1132 const char *cp;
1133 unsigned long allot;
1134 size_t l = len;
1135
1136 if (ecb->truncate)
1137 return ecb->truncate(line, len);
1138 cp = line;
1139 allot = l;
1140 while (0 < l) {
1141 (void) utf8_width(&cp, &l);
1142 if (!cp)
1143 break; /* truncated in the middle? */
1144 }
1145 return allot - l;
1146 }
1147
1148 static void find_lno(const char *line, struct emit_callback *ecbdata)
1149 {
1150 const char *p;
1151 ecbdata->lno_in_preimage = 0;
1152 ecbdata->lno_in_postimage = 0;
1153 p = strchr(line, '-');
1154 if (!p)
1155 return; /* cannot happen */
1156 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1157 p = strchr(p, '+');
1158 if (!p)
1159 return; /* cannot happen */
1160 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1161 }
1162
1163 static void fn_out_consume(void *priv, char *line, unsigned long len)
1164 {
1165 struct emit_callback *ecbdata = priv;
1166 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1167 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1168 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1169 struct diff_options *o = ecbdata->opt;
1170 const char *line_prefix = diff_line_prefix(o);
1171
1172 if (ecbdata->header) {
1173 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1174 strbuf_reset(ecbdata->header);
1175 ecbdata->header = NULL;
1176 }
1177 *(ecbdata->found_changesp) = 1;
1178
1179 if (ecbdata->label_path[0]) {
1180 const char *name_a_tab, *name_b_tab;
1181
1182 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1183 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1184
1185 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1186 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1187 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1188 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1189 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1190 }
1191
1192 if (diff_suppress_blank_empty
1193 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1194 line[0] = '\n';
1195 len = 1;
1196 }
1197
1198 if (line[0] == '@') {
1199 if (ecbdata->diff_words)
1200 diff_words_flush(ecbdata);
1201 len = sane_truncate_line(ecbdata, line, len);
1202 find_lno(line, ecbdata);
1203 emit_hunk_header(ecbdata, line, len);
1204 if (line[len-1] != '\n')
1205 putc('\n', ecbdata->opt->file);
1206 return;
1207 }
1208
1209 if (len < 1) {
1210 emit_line(ecbdata->opt, reset, reset, line, len);
1211 if (ecbdata->diff_words
1212 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1213 fputs("~\n", ecbdata->opt->file);
1214 return;
1215 }
1216
1217 if (ecbdata->diff_words) {
1218 if (line[0] == '-') {
1219 diff_words_append(line, len,
1220 &ecbdata->diff_words->minus);
1221 return;
1222 } else if (line[0] == '+') {
1223 diff_words_append(line, len,
1224 &ecbdata->diff_words->plus);
1225 return;
1226 } else if (starts_with(line, "\\ ")) {
1227 /*
1228 * Eat the "no newline at eof" marker as if we
1229 * saw a "+" or "-" line with nothing on it,
1230 * and return without diff_words_flush() to
1231 * defer processing. If this is the end of
1232 * preimage, more "+" lines may come after it.
1233 */
1234 return;
1235 }
1236 diff_words_flush(ecbdata);
1237 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1238 emit_line(ecbdata->opt, plain, reset, line, len);
1239 fputs("~\n", ecbdata->opt->file);
1240 } else {
1241 /*
1242 * Skip the prefix character, if any. With
1243 * diff_suppress_blank_empty, there may be
1244 * none.
1245 */
1246 if (line[0] != '\n') {
1247 line++;
1248 len--;
1249 }
1250 emit_line(ecbdata->opt, plain, reset, line, len);
1251 }
1252 return;
1253 }
1254
1255 if (line[0] != '+') {
1256 const char *color =
1257 diff_get_color(ecbdata->color_diff,
1258 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1259 ecbdata->lno_in_preimage++;
1260 if (line[0] == ' ')
1261 ecbdata->lno_in_postimage++;
1262 emit_line(ecbdata->opt, color, reset, line, len);
1263 } else {
1264 ecbdata->lno_in_postimage++;
1265 emit_add_line(reset, ecbdata, line + 1, len - 1);
1266 }
1267 }
1268
1269 static char *pprint_rename(const char *a, const char *b)
1270 {
1271 const char *old = a;
1272 const char *new = b;
1273 struct strbuf name = STRBUF_INIT;
1274 int pfx_length, sfx_length;
1275 int pfx_adjust_for_slash;
1276 int len_a = strlen(a);
1277 int len_b = strlen(b);
1278 int a_midlen, b_midlen;
1279 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1280 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1281
1282 if (qlen_a || qlen_b) {
1283 quote_c_style(a, &name, NULL, 0);
1284 strbuf_addstr(&name, " => ");
1285 quote_c_style(b, &name, NULL, 0);
1286 return strbuf_detach(&name, NULL);
1287 }
1288
1289 /* Find common prefix */
1290 pfx_length = 0;
1291 while (*old && *new && *old == *new) {
1292 if (*old == '/')
1293 pfx_length = old - a + 1;
1294 old++;
1295 new++;
1296 }
1297
1298 /* Find common suffix */
1299 old = a + len_a;
1300 new = b + len_b;
1301 sfx_length = 0;
1302 /*
1303 * If there is a common prefix, it must end in a slash. In
1304 * that case we let this loop run 1 into the prefix to see the
1305 * same slash.
1306 *
1307 * If there is no common prefix, we cannot do this as it would
1308 * underrun the input strings.
1309 */
1310 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1311 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1312 b + pfx_length - pfx_adjust_for_slash <= new &&
1313 *old == *new) {
1314 if (*old == '/')
1315 sfx_length = len_a - (old - a);
1316 old--;
1317 new--;
1318 }
1319
1320 /*
1321 * pfx{mid-a => mid-b}sfx
1322 * {pfx-a => pfx-b}sfx
1323 * pfx{sfx-a => sfx-b}
1324 * name-a => name-b
1325 */
1326 a_midlen = len_a - pfx_length - sfx_length;
1327 b_midlen = len_b - pfx_length - sfx_length;
1328 if (a_midlen < 0)
1329 a_midlen = 0;
1330 if (b_midlen < 0)
1331 b_midlen = 0;
1332
1333 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1334 if (pfx_length + sfx_length) {
1335 strbuf_add(&name, a, pfx_length);
1336 strbuf_addch(&name, '{');
1337 }
1338 strbuf_add(&name, a + pfx_length, a_midlen);
1339 strbuf_addstr(&name, " => ");
1340 strbuf_add(&name, b + pfx_length, b_midlen);
1341 if (pfx_length + sfx_length) {
1342 strbuf_addch(&name, '}');
1343 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1344 }
1345 return strbuf_detach(&name, NULL);
1346 }
1347
1348 struct diffstat_t {
1349 int nr;
1350 int alloc;
1351 struct diffstat_file {
1352 char *from_name;
1353 char *name;
1354 char *print_name;
1355 unsigned is_unmerged:1;
1356 unsigned is_binary:1;
1357 unsigned is_renamed:1;
1358 unsigned is_interesting:1;
1359 uintmax_t added, deleted;
1360 } **files;
1361 };
1362
1363 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1364 const char *name_a,
1365 const char *name_b)
1366 {
1367 struct diffstat_file *x;
1368 x = xcalloc(1, sizeof(*x));
1369 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1370 diffstat->files[diffstat->nr++] = x;
1371 if (name_b) {
1372 x->from_name = xstrdup(name_a);
1373 x->name = xstrdup(name_b);
1374 x->is_renamed = 1;
1375 }
1376 else {
1377 x->from_name = NULL;
1378 x->name = xstrdup(name_a);
1379 }
1380 return x;
1381 }
1382
1383 static void diffstat_consume(void *priv, char *line, unsigned long len)
1384 {
1385 struct diffstat_t *diffstat = priv;
1386 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1387
1388 if (line[0] == '+')
1389 x->added++;
1390 else if (line[0] == '-')
1391 x->deleted++;
1392 }
1393
1394 const char mime_boundary_leader[] = "------------";
1395
1396 static int scale_linear(int it, int width, int max_change)
1397 {
1398 if (!it)
1399 return 0;
1400 /*
1401 * make sure that at least one '-' or '+' is printed if
1402 * there is any change to this path. The easiest way is to
1403 * scale linearly as if the alloted width is one column shorter
1404 * than it is, and then add 1 to the result.
1405 */
1406 return 1 + (it * (width - 1) / max_change);
1407 }
1408
1409 static void show_name(FILE *file,
1410 const char *prefix, const char *name, int len)
1411 {
1412 fprintf(file, " %s%-*s |", prefix, len, name);
1413 }
1414
1415 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1416 {
1417 if (cnt <= 0)
1418 return;
1419 fprintf(file, "%s", set);
1420 while (cnt--)
1421 putc(ch, file);
1422 fprintf(file, "%s", reset);
1423 }
1424
1425 static void fill_print_name(struct diffstat_file *file)
1426 {
1427 char *pname;
1428
1429 if (file->print_name)
1430 return;
1431
1432 if (!file->is_renamed) {
1433 struct strbuf buf = STRBUF_INIT;
1434 if (quote_c_style(file->name, &buf, NULL, 0)) {
1435 pname = strbuf_detach(&buf, NULL);
1436 } else {
1437 pname = file->name;
1438 strbuf_release(&buf);
1439 }
1440 } else {
1441 pname = pprint_rename(file->from_name, file->name);
1442 }
1443 file->print_name = pname;
1444 }
1445
1446 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1447 {
1448 struct strbuf sb = STRBUF_INIT;
1449 int ret;
1450
1451 if (!files) {
1452 assert(insertions == 0 && deletions == 0);
1453 return fprintf(fp, "%s\n", " 0 files changed");
1454 }
1455
1456 strbuf_addf(&sb,
1457 (files == 1) ? " %d file changed" : " %d files changed",
1458 files);
1459
1460 /*
1461 * For binary diff, the caller may want to print "x files
1462 * changed" with insertions == 0 && deletions == 0.
1463 *
1464 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1465 * is probably less confusing (i.e skip over "2 files changed
1466 * but nothing about added/removed lines? Is this a bug in Git?").
1467 */
1468 if (insertions || deletions == 0) {
1469 strbuf_addf(&sb,
1470 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1471 insertions);
1472 }
1473
1474 if (deletions || insertions == 0) {
1475 strbuf_addf(&sb,
1476 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1477 deletions);
1478 }
1479 strbuf_addch(&sb, '\n');
1480 ret = fputs(sb.buf, fp);
1481 strbuf_release(&sb);
1482 return ret;
1483 }
1484
1485 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1486 {
1487 int i, len, add, del, adds = 0, dels = 0;
1488 uintmax_t max_change = 0, max_len = 0;
1489 int total_files = data->nr, count;
1490 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1491 const char *reset, *add_c, *del_c;
1492 const char *line_prefix = "";
1493 int extra_shown = 0;
1494
1495 if (data->nr == 0)
1496 return;
1497
1498 line_prefix = diff_line_prefix(options);
1499 count = options->stat_count ? options->stat_count : data->nr;
1500
1501 reset = diff_get_color_opt(options, DIFF_RESET);
1502 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1503 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1504
1505 /*
1506 * Find the longest filename and max number of changes
1507 */
1508 for (i = 0; (i < count) && (i < data->nr); i++) {
1509 struct diffstat_file *file = data->files[i];
1510 uintmax_t change = file->added + file->deleted;
1511
1512 if (!file->is_interesting && (change == 0)) {
1513 count++; /* not shown == room for one more */
1514 continue;
1515 }
1516 fill_print_name(file);
1517 len = strlen(file->print_name);
1518 if (max_len < len)
1519 max_len = len;
1520
1521 if (file->is_unmerged) {
1522 /* "Unmerged" is 8 characters */
1523 bin_width = bin_width < 8 ? 8 : bin_width;
1524 continue;
1525 }
1526 if (file->is_binary) {
1527 /* "Bin XXX -> YYY bytes" */
1528 int w = 14 + decimal_width(file->added)
1529 + decimal_width(file->deleted);
1530 bin_width = bin_width < w ? w : bin_width;
1531 /* Display change counts aligned with "Bin" */
1532 number_width = 3;
1533 continue;
1534 }
1535
1536 if (max_change < change)
1537 max_change = change;
1538 }
1539 count = i; /* where we can stop scanning in data->files[] */
1540
1541 /*
1542 * We have width = stat_width or term_columns() columns total.
1543 * We want a maximum of min(max_len, stat_name_width) for the name part.
1544 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1545 * We also need 1 for " " and 4 + decimal_width(max_change)
1546 * for " | NNNN " and one the empty column at the end, altogether
1547 * 6 + decimal_width(max_change).
1548 *
1549 * If there's not enough space, we will use the smaller of
1550 * stat_name_width (if set) and 5/8*width for the filename,
1551 * and the rest for constant elements + graph part, but no more
1552 * than stat_graph_width for the graph part.
1553 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1554 * for the standard terminal size).
1555 *
1556 * In other words: stat_width limits the maximum width, and
1557 * stat_name_width fixes the maximum width of the filename,
1558 * and is also used to divide available columns if there
1559 * aren't enough.
1560 *
1561 * Binary files are displayed with "Bin XXX -> YYY bytes"
1562 * instead of the change count and graph. This part is treated
1563 * similarly to the graph part, except that it is not
1564 * "scaled". If total width is too small to accommodate the
1565 * guaranteed minimum width of the filename part and the
1566 * separators and this message, this message will "overflow"
1567 * making the line longer than the maximum width.
1568 */
1569
1570 if (options->stat_width == -1)
1571 width = term_columns() - options->output_prefix_length;
1572 else
1573 width = options->stat_width ? options->stat_width : 80;
1574 number_width = decimal_width(max_change) > number_width ?
1575 decimal_width(max_change) : number_width;
1576
1577 if (options->stat_graph_width == -1)
1578 options->stat_graph_width = diff_stat_graph_width;
1579
1580 /*
1581 * Guarantee 3/8*16==6 for the graph part
1582 * and 5/8*16==10 for the filename part
1583 */
1584 if (width < 16 + 6 + number_width)
1585 width = 16 + 6 + number_width;
1586
1587 /*
1588 * First assign sizes that are wanted, ignoring available width.
1589 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1590 * starting from "XXX" should fit in graph_width.
1591 */
1592 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1593 if (options->stat_graph_width &&
1594 options->stat_graph_width < graph_width)
1595 graph_width = options->stat_graph_width;
1596
1597 name_width = (options->stat_name_width > 0 &&
1598 options->stat_name_width < max_len) ?
1599 options->stat_name_width : max_len;
1600
1601 /*
1602 * Adjust adjustable widths not to exceed maximum width
1603 */
1604 if (name_width + number_width + 6 + graph_width > width) {
1605 if (graph_width > width * 3/8 - number_width - 6) {
1606 graph_width = width * 3/8 - number_width - 6;
1607 if (graph_width < 6)
1608 graph_width = 6;
1609 }
1610
1611 if (options->stat_graph_width &&
1612 graph_width > options->stat_graph_width)
1613 graph_width = options->stat_graph_width;
1614 if (name_width > width - number_width - 6 - graph_width)
1615 name_width = width - number_width - 6 - graph_width;
1616 else
1617 graph_width = width - number_width - 6 - name_width;
1618 }
1619
1620 /*
1621 * From here name_width is the width of the name area,
1622 * and graph_width is the width of the graph area.
1623 * max_change is used to scale graph properly.
1624 */
1625 for (i = 0; i < count; i++) {
1626 const char *prefix = "";
1627 struct diffstat_file *file = data->files[i];
1628 char *name = file->print_name;
1629 uintmax_t added = file->added;
1630 uintmax_t deleted = file->deleted;
1631 int name_len;
1632
1633 if (!file->is_interesting && (added + deleted == 0))
1634 continue;
1635
1636 /*
1637 * "scale" the filename
1638 */
1639 len = name_width;
1640 name_len = strlen(name);
1641 if (name_width < name_len) {
1642 char *slash;
1643 prefix = "...";
1644 len -= 3;
1645 name += name_len - len;
1646 slash = strchr(name, '/');
1647 if (slash)
1648 name = slash;
1649 }
1650
1651 if (file->is_binary) {
1652 fprintf(options->file, "%s", line_prefix);
1653 show_name(options->file, prefix, name, len);
1654 fprintf(options->file, " %*s", number_width, "Bin");
1655 if (!added && !deleted) {
1656 putc('\n', options->file);
1657 continue;
1658 }
1659 fprintf(options->file, " %s%"PRIuMAX"%s",
1660 del_c, deleted, reset);
1661 fprintf(options->file, " -> ");
1662 fprintf(options->file, "%s%"PRIuMAX"%s",
1663 add_c, added, reset);
1664 fprintf(options->file, " bytes");
1665 fprintf(options->file, "\n");
1666 continue;
1667 }
1668 else if (file->is_unmerged) {
1669 fprintf(options->file, "%s", line_prefix);
1670 show_name(options->file, prefix, name, len);
1671 fprintf(options->file, " Unmerged\n");
1672 continue;
1673 }
1674
1675 /*
1676 * scale the add/delete
1677 */
1678 add = added;
1679 del = deleted;
1680
1681 if (graph_width <= max_change) {
1682 int total = scale_linear(add + del, graph_width, max_change);
1683 if (total < 2 && add && del)
1684 /* width >= 2 due to the sanity check */
1685 total = 2;
1686 if (add < del) {
1687 add = scale_linear(add, graph_width, max_change);
1688 del = total - add;
1689 } else {
1690 del = scale_linear(del, graph_width, max_change);
1691 add = total - del;
1692 }
1693 }
1694 fprintf(options->file, "%s", line_prefix);
1695 show_name(options->file, prefix, name, len);
1696 fprintf(options->file, " %*"PRIuMAX"%s",
1697 number_width, added + deleted,
1698 added + deleted ? " " : "");
1699 show_graph(options->file, '+', add, add_c, reset);
1700 show_graph(options->file, '-', del, del_c, reset);
1701 fprintf(options->file, "\n");
1702 }
1703
1704 for (i = 0; i < data->nr; i++) {
1705 struct diffstat_file *file = data->files[i];
1706 uintmax_t added = file->added;
1707 uintmax_t deleted = file->deleted;
1708
1709 if (file->is_unmerged ||
1710 (!file->is_interesting && (added + deleted == 0))) {
1711 total_files--;
1712 continue;
1713 }
1714
1715 if (!file->is_binary) {
1716 adds += added;
1717 dels += deleted;
1718 }
1719 if (i < count)
1720 continue;
1721 if (!extra_shown)
1722 fprintf(options->file, "%s ...\n", line_prefix);
1723 extra_shown = 1;
1724 }
1725 fprintf(options->file, "%s", line_prefix);
1726 print_stat_summary(options->file, total_files, adds, dels);
1727 }
1728
1729 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1730 {
1731 int i, adds = 0, dels = 0, total_files = data->nr;
1732
1733 if (data->nr == 0)
1734 return;
1735
1736 for (i = 0; i < data->nr; i++) {
1737 int added = data->files[i]->added;
1738 int deleted= data->files[i]->deleted;
1739
1740 if (data->files[i]->is_unmerged ||
1741 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1742 total_files--;
1743 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1744 adds += added;
1745 dels += deleted;
1746 }
1747 }
1748 fprintf(options->file, "%s", diff_line_prefix(options));
1749 print_stat_summary(options->file, total_files, adds, dels);
1750 }
1751
1752 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1753 {
1754 int i;
1755
1756 if (data->nr == 0)
1757 return;
1758
1759 for (i = 0; i < data->nr; i++) {
1760 struct diffstat_file *file = data->files[i];
1761
1762 fprintf(options->file, "%s", diff_line_prefix(options));
1763
1764 if (file->is_binary)
1765 fprintf(options->file, "-\t-\t");
1766 else
1767 fprintf(options->file,
1768 "%"PRIuMAX"\t%"PRIuMAX"\t",
1769 file->added, file->deleted);
1770 if (options->line_termination) {
1771 fill_print_name(file);
1772 if (!file->is_renamed)
1773 write_name_quoted(file->name, options->file,
1774 options->line_termination);
1775 else {
1776 fputs(file->print_name, options->file);
1777 putc(options->line_termination, options->file);
1778 }
1779 } else {
1780 if (file->is_renamed) {
1781 putc('\0', options->file);
1782 write_name_quoted(file->from_name, options->file, '\0');
1783 }
1784 write_name_quoted(file->name, options->file, '\0');
1785 }
1786 }
1787 }
1788
1789 struct dirstat_file {
1790 const char *name;
1791 unsigned long changed;
1792 };
1793
1794 struct dirstat_dir {
1795 struct dirstat_file *files;
1796 int alloc, nr, permille, cumulative;
1797 };
1798
1799 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1800 unsigned long changed, const char *base, int baselen)
1801 {
1802 unsigned long this_dir = 0;
1803 unsigned int sources = 0;
1804 const char *line_prefix = diff_line_prefix(opt);
1805
1806 while (dir->nr) {
1807 struct dirstat_file *f = dir->files;
1808 int namelen = strlen(f->name);
1809 unsigned long this;
1810 char *slash;
1811
1812 if (namelen < baselen)
1813 break;
1814 if (memcmp(f->name, base, baselen))
1815 break;
1816 slash = strchr(f->name + baselen, '/');
1817 if (slash) {
1818 int newbaselen = slash + 1 - f->name;
1819 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1820 sources++;
1821 } else {
1822 this = f->changed;
1823 dir->files++;
1824 dir->nr--;
1825 sources += 2;
1826 }
1827 this_dir += this;
1828 }
1829
1830 /*
1831 * We don't report dirstat's for
1832 * - the top level
1833 * - or cases where everything came from a single directory
1834 * under this directory (sources == 1).
1835 */
1836 if (baselen && sources != 1) {
1837 if (this_dir) {
1838 int permille = this_dir * 1000 / changed;
1839 if (permille >= dir->permille) {
1840 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1841 permille / 10, permille % 10, baselen, base);
1842 if (!dir->cumulative)
1843 return 0;
1844 }
1845 }
1846 }
1847 return this_dir;
1848 }
1849
1850 static int dirstat_compare(const void *_a, const void *_b)
1851 {
1852 const struct dirstat_file *a = _a;
1853 const struct dirstat_file *b = _b;
1854 return strcmp(a->name, b->name);
1855 }
1856
1857 static void show_dirstat(struct diff_options *options)
1858 {
1859 int i;
1860 unsigned long changed;
1861 struct dirstat_dir dir;
1862 struct diff_queue_struct *q = &diff_queued_diff;
1863
1864 dir.files = NULL;
1865 dir.alloc = 0;
1866 dir.nr = 0;
1867 dir.permille = options->dirstat_permille;
1868 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1869
1870 changed = 0;
1871 for (i = 0; i < q->nr; i++) {
1872 struct diff_filepair *p = q->queue[i];
1873 const char *name;
1874 unsigned long copied, added, damage;
1875 int content_changed;
1876
1877 name = p->two->path ? p->two->path : p->one->path;
1878
1879 if (p->one->sha1_valid && p->two->sha1_valid)
1880 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1881 else
1882 content_changed = 1;
1883
1884 if (!content_changed) {
1885 /*
1886 * The SHA1 has not changed, so pre-/post-content is
1887 * identical. We can therefore skip looking at the
1888 * file contents altogether.
1889 */
1890 damage = 0;
1891 goto found_damage;
1892 }
1893
1894 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1895 /*
1896 * In --dirstat-by-file mode, we don't really need to
1897 * look at the actual file contents at all.
1898 * The fact that the SHA1 changed is enough for us to
1899 * add this file to the list of results
1900 * (with each file contributing equal damage).
1901 */
1902 damage = 1;
1903 goto found_damage;
1904 }
1905
1906 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1907 diff_populate_filespec(p->one, 0);
1908 diff_populate_filespec(p->two, 0);
1909 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1910 &copied, &added);
1911 diff_free_filespec_data(p->one);
1912 diff_free_filespec_data(p->two);
1913 } else if (DIFF_FILE_VALID(p->one)) {
1914 diff_populate_filespec(p->one, 1);
1915 copied = added = 0;
1916 diff_free_filespec_data(p->one);
1917 } else if (DIFF_FILE_VALID(p->two)) {
1918 diff_populate_filespec(p->two, 1);
1919 copied = 0;
1920 added = p->two->size;
1921 diff_free_filespec_data(p->two);
1922 } else
1923 continue;
1924
1925 /*
1926 * Original minus copied is the removed material,
1927 * added is the new material. They are both damages
1928 * made to the preimage.
1929 * If the resulting damage is zero, we know that
1930 * diffcore_count_changes() considers the two entries to
1931 * be identical, but since content_changed is true, we
1932 * know that there must have been _some_ kind of change,
1933 * so we force all entries to have damage > 0.
1934 */
1935 damage = (p->one->size - copied) + added;
1936 if (!damage)
1937 damage = 1;
1938
1939 found_damage:
1940 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1941 dir.files[dir.nr].name = name;
1942 dir.files[dir.nr].changed = damage;
1943 changed += damage;
1944 dir.nr++;
1945 }
1946
1947 /* This can happen even with many files, if everything was renames */
1948 if (!changed)
1949 return;
1950
1951 /* Show all directories with more than x% of the changes */
1952 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1953 gather_dirstat(options, &dir, changed, "", 0);
1954 }
1955
1956 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1957 {
1958 int i;
1959 unsigned long changed;
1960 struct dirstat_dir dir;
1961
1962 if (data->nr == 0)
1963 return;
1964
1965 dir.files = NULL;
1966 dir.alloc = 0;
1967 dir.nr = 0;
1968 dir.permille = options->dirstat_permille;
1969 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1970
1971 changed = 0;
1972 for (i = 0; i < data->nr; i++) {
1973 struct diffstat_file *file = data->files[i];
1974 unsigned long damage = file->added + file->deleted;
1975 if (file->is_binary)
1976 /*
1977 * binary files counts bytes, not lines. Must find some
1978 * way to normalize binary bytes vs. textual lines.
1979 * The following heuristic assumes that there are 64
1980 * bytes per "line".
1981 * This is stupid and ugly, but very cheap...
1982 */
1983 damage = (damage + 63) / 64;
1984 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1985 dir.files[dir.nr].name = file->name;
1986 dir.files[dir.nr].changed = damage;
1987 changed += damage;
1988 dir.nr++;
1989 }
1990
1991 /* This can happen even with many files, if everything was renames */
1992 if (!changed)
1993 return;
1994
1995 /* Show all directories with more than x% of the changes */
1996 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1997 gather_dirstat(options, &dir, changed, "", 0);
1998 }
1999
2000 static void free_diffstat_info(struct diffstat_t *diffstat)
2001 {
2002 int i;
2003 for (i = 0; i < diffstat->nr; i++) {
2004 struct diffstat_file *f = diffstat->files[i];
2005 if (f->name != f->print_name)
2006 free(f->print_name);
2007 free(f->name);
2008 free(f->from_name);
2009 free(f);
2010 }
2011 free(diffstat->files);
2012 }
2013
2014 struct checkdiff_t {
2015 const char *filename;
2016 int lineno;
2017 int conflict_marker_size;
2018 struct diff_options *o;
2019 unsigned ws_rule;
2020 unsigned status;
2021 };
2022
2023 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2024 {
2025 char firstchar;
2026 int cnt;
2027
2028 if (len < marker_size + 1)
2029 return 0;
2030 firstchar = line[0];
2031 switch (firstchar) {
2032 case '=': case '>': case '<': case '|':
2033 break;
2034 default:
2035 return 0;
2036 }
2037 for (cnt = 1; cnt < marker_size; cnt++)
2038 if (line[cnt] != firstchar)
2039 return 0;
2040 /* line[1] thru line[marker_size-1] are same as firstchar */
2041 if (len < marker_size + 1 || !isspace(line[marker_size]))
2042 return 0;
2043 return 1;
2044 }
2045
2046 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2047 {
2048 struct checkdiff_t *data = priv;
2049 int marker_size = data->conflict_marker_size;
2050 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2051 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2052 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2053 char *err;
2054 const char *line_prefix;
2055
2056 assert(data->o);
2057 line_prefix = diff_line_prefix(data->o);
2058
2059 if (line[0] == '+') {
2060 unsigned bad;
2061 data->lineno++;
2062 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2063 data->status |= 1;
2064 fprintf(data->o->file,
2065 "%s%s:%d: leftover conflict marker\n",
2066 line_prefix, data->filename, data->lineno);
2067 }
2068 bad = ws_check(line + 1, len - 1, data->ws_rule);
2069 if (!bad)
2070 return;
2071 data->status |= bad;
2072 err = whitespace_error_string(bad);
2073 fprintf(data->o->file, "%s%s:%d: %s.\n",
2074 line_prefix, data->filename, data->lineno, err);
2075 free(err);
2076 emit_line(data->o, set, reset, line, 1);
2077 ws_check_emit(line + 1, len - 1, data->ws_rule,
2078 data->o->file, set, reset, ws);
2079 } else if (line[0] == ' ') {
2080 data->lineno++;
2081 } else if (line[0] == '@') {
2082 char *plus = strchr(line, '+');
2083 if (plus)
2084 data->lineno = strtol(plus, NULL, 10) - 1;
2085 else
2086 die("invalid diff");
2087 }
2088 }
2089
2090 static unsigned char *deflate_it(char *data,
2091 unsigned long size,
2092 unsigned long *result_size)
2093 {
2094 int bound;
2095 unsigned char *deflated;
2096 git_zstream stream;
2097
2098 memset(&stream, 0, sizeof(stream));
2099 git_deflate_init(&stream, zlib_compression_level);
2100 bound = git_deflate_bound(&stream, size);
2101 deflated = xmalloc(bound);
2102 stream.next_out = deflated;
2103 stream.avail_out = bound;
2104
2105 stream.next_in = (unsigned char *)data;
2106 stream.avail_in = size;
2107 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2108 ; /* nothing */
2109 git_deflate_end(&stream);
2110 *result_size = stream.total_out;
2111 return deflated;
2112 }
2113
2114 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2115 const char *prefix)
2116 {
2117 void *cp;
2118 void *delta;
2119 void *deflated;
2120 void *data;
2121 unsigned long orig_size;
2122 unsigned long delta_size;
2123 unsigned long deflate_size;
2124 unsigned long data_size;
2125
2126 /* We could do deflated delta, or we could do just deflated two,
2127 * whichever is smaller.
2128 */
2129 delta = NULL;
2130 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2131 if (one->size && two->size) {
2132 delta = diff_delta(one->ptr, one->size,
2133 two->ptr, two->size,
2134 &delta_size, deflate_size);
2135 if (delta) {
2136 void *to_free = delta;
2137 orig_size = delta_size;
2138 delta = deflate_it(delta, delta_size, &delta_size);
2139 free(to_free);
2140 }
2141 }
2142
2143 if (delta && delta_size < deflate_size) {
2144 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2145 free(deflated);
2146 data = delta;
2147 data_size = delta_size;
2148 }
2149 else {
2150 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2151 free(delta);
2152 data = deflated;
2153 data_size = deflate_size;
2154 }
2155
2156 /* emit data encoded in base85 */
2157 cp = data;
2158 while (data_size) {
2159 int bytes = (52 < data_size) ? 52 : data_size;
2160 char line[70];
2161 data_size -= bytes;
2162 if (bytes <= 26)
2163 line[0] = bytes + 'A' - 1;
2164 else
2165 line[0] = bytes - 26 + 'a' - 1;
2166 encode_85(line + 1, cp, bytes);
2167 cp = (char *) cp + bytes;
2168 fprintf(file, "%s", prefix);
2169 fputs(line, file);
2170 fputc('\n', file);
2171 }
2172 fprintf(file, "%s\n", prefix);
2173 free(data);
2174 }
2175
2176 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2177 const char *prefix)
2178 {
2179 fprintf(file, "%sGIT binary patch\n", prefix);
2180 emit_binary_diff_body(file, one, two, prefix);
2181 emit_binary_diff_body(file, two, one, prefix);
2182 }
2183
2184 int diff_filespec_is_binary(struct diff_filespec *one)
2185 {
2186 if (one->is_binary == -1) {
2187 diff_filespec_load_driver(one);
2188 if (one->driver->binary != -1)
2189 one->is_binary = one->driver->binary;
2190 else {
2191 if (!one->data && DIFF_FILE_VALID(one))
2192 diff_populate_filespec(one, 0);
2193 if (one->data)
2194 one->is_binary = buffer_is_binary(one->data,
2195 one->size);
2196 if (one->is_binary == -1)
2197 one->is_binary = 0;
2198 }
2199 }
2200 return one->is_binary;
2201 }
2202
2203 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2204 {
2205 diff_filespec_load_driver(one);
2206 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2207 }
2208
2209 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2210 {
2211 if (!options->a_prefix)
2212 options->a_prefix = a;
2213 if (!options->b_prefix)
2214 options->b_prefix = b;
2215 }
2216
2217 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2218 {
2219 if (!DIFF_FILE_VALID(one))
2220 return NULL;
2221
2222 diff_filespec_load_driver(one);
2223 return userdiff_get_textconv(one->driver);
2224 }
2225
2226 static void builtin_diff(const char *name_a,
2227 const char *name_b,
2228 struct diff_filespec *one,
2229 struct diff_filespec *two,
2230 const char *xfrm_msg,
2231 int must_show_header,
2232 struct diff_options *o,
2233 int complete_rewrite)
2234 {
2235 mmfile_t mf1, mf2;
2236 const char *lbl[2];
2237 char *a_one, *b_two;
2238 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2239 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2240 const char *a_prefix, *b_prefix;
2241 struct userdiff_driver *textconv_one = NULL;
2242 struct userdiff_driver *textconv_two = NULL;
2243 struct strbuf header = STRBUF_INIT;
2244 const char *line_prefix = diff_line_prefix(o);
2245
2246 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2247 (!one->mode || S_ISGITLINK(one->mode)) &&
2248 (!two->mode || S_ISGITLINK(two->mode))) {
2249 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2250 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2251 show_submodule_summary(o->file, one->path ? one->path : two->path,
2252 line_prefix,
2253 one->sha1, two->sha1, two->dirty_submodule,
2254 meta, del, add, reset);
2255 return;
2256 }
2257
2258 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2259 textconv_one = get_textconv(one);
2260 textconv_two = get_textconv(two);
2261 }
2262
2263 diff_set_mnemonic_prefix(o, "a/", "b/");
2264 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2265 a_prefix = o->b_prefix;
2266 b_prefix = o->a_prefix;
2267 } else {
2268 a_prefix = o->a_prefix;
2269 b_prefix = o->b_prefix;
2270 }
2271
2272 /* Never use a non-valid filename anywhere if at all possible */
2273 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2274 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2275
2276 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2277 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2278 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2279 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2280 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2281 if (lbl[0][0] == '/') {
2282 /* /dev/null */
2283 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2284 if (xfrm_msg)
2285 strbuf_addstr(&header, xfrm_msg);
2286 must_show_header = 1;
2287 }
2288 else if (lbl[1][0] == '/') {
2289 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2290 if (xfrm_msg)
2291 strbuf_addstr(&header, xfrm_msg);
2292 must_show_header = 1;
2293 }
2294 else {
2295 if (one->mode != two->mode) {
2296 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2297 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2298 must_show_header = 1;
2299 }
2300 if (xfrm_msg)
2301 strbuf_addstr(&header, xfrm_msg);
2302
2303 /*
2304 * we do not run diff between different kind
2305 * of objects.
2306 */
2307 if ((one->mode ^ two->mode) & S_IFMT)
2308 goto free_ab_and_return;
2309 if (complete_rewrite &&
2310 (textconv_one || !diff_filespec_is_binary(one)) &&
2311 (textconv_two || !diff_filespec_is_binary(two))) {
2312 fprintf(o->file, "%s", header.buf);
2313 strbuf_reset(&header);
2314 emit_rewrite_diff(name_a, name_b, one, two,
2315 textconv_one, textconv_two, o);
2316 o->found_changes = 1;
2317 goto free_ab_and_return;
2318 }
2319 }
2320
2321 if (o->irreversible_delete && lbl[1][0] == '/') {
2322 fprintf(o->file, "%s", header.buf);
2323 strbuf_reset(&header);
2324 goto free_ab_and_return;
2325 } else if (!DIFF_OPT_TST(o, TEXT) &&
2326 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2327 (!textconv_two && diff_filespec_is_binary(two)) )) {
2328 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2329 die("unable to read files to diff");
2330 /* Quite common confusing case */
2331 if (mf1.size == mf2.size &&
2332 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2333 if (must_show_header)
2334 fprintf(o->file, "%s", header.buf);
2335 goto free_ab_and_return;
2336 }
2337 fprintf(o->file, "%s", header.buf);
2338 strbuf_reset(&header);
2339 if (DIFF_OPT_TST(o, BINARY))
2340 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2341 else
2342 fprintf(o->file, "%sBinary files %s and %s differ\n",
2343 line_prefix, lbl[0], lbl[1]);
2344 o->found_changes = 1;
2345 } else {
2346 /* Crazy xdl interfaces.. */
2347 const char *diffopts = getenv("GIT_DIFF_OPTS");
2348 const char *v;
2349 xpparam_t xpp;
2350 xdemitconf_t xecfg;
2351 struct emit_callback ecbdata;
2352 const struct userdiff_funcname *pe;
2353
2354 if (must_show_header) {
2355 fprintf(o->file, "%s", header.buf);
2356 strbuf_reset(&header);
2357 }
2358
2359 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2360 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2361
2362 pe = diff_funcname_pattern(one);
2363 if (!pe)
2364 pe = diff_funcname_pattern(two);
2365
2366 memset(&xpp, 0, sizeof(xpp));
2367 memset(&xecfg, 0, sizeof(xecfg));
2368 memset(&ecbdata, 0, sizeof(ecbdata));
2369 ecbdata.label_path = lbl;
2370 ecbdata.color_diff = want_color(o->use_color);
2371 ecbdata.found_changesp = &o->found_changes;
2372 ecbdata.ws_rule = whitespace_rule(name_b);
2373 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2374 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2375 ecbdata.opt = o;
2376 ecbdata.header = header.len ? &header : NULL;
2377 xpp.flags = o->xdl_opts;
2378 xecfg.ctxlen = o->context;
2379 xecfg.interhunkctxlen = o->interhunkcontext;
2380 xecfg.flags = XDL_EMIT_FUNCNAMES;
2381 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2382 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2383 if (pe)
2384 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2385 if (!diffopts)
2386 ;
2387 else if (skip_prefix(diffopts, "--unified=", &v))
2388 xecfg.ctxlen = strtoul(v, NULL, 10);
2389 else if (skip_prefix(diffopts, "-u", &v))
2390 xecfg.ctxlen = strtoul(v, NULL, 10);
2391 if (o->word_diff)
2392 init_diff_words_data(&ecbdata, o, one, two);
2393 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2394 &xpp, &xecfg);
2395 if (o->word_diff)
2396 free_diff_words_data(&ecbdata);
2397 if (textconv_one)
2398 free(mf1.ptr);
2399 if (textconv_two)
2400 free(mf2.ptr);
2401 xdiff_clear_find_func(&xecfg);
2402 }
2403
2404 free_ab_and_return:
2405 strbuf_release(&header);
2406 diff_free_filespec_data(one);
2407 diff_free_filespec_data(two);
2408 free(a_one);
2409 free(b_two);
2410 return;
2411 }
2412
2413 static void builtin_diffstat(const char *name_a, const char *name_b,
2414 struct diff_filespec *one,
2415 struct diff_filespec *two,
2416 struct diffstat_t *diffstat,
2417 struct diff_options *o,
2418 struct diff_filepair *p)
2419 {
2420 mmfile_t mf1, mf2;
2421 struct diffstat_file *data;
2422 int same_contents;
2423 int complete_rewrite = 0;
2424
2425 if (!DIFF_PAIR_UNMERGED(p)) {
2426 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2427 complete_rewrite = 1;
2428 }
2429
2430 data = diffstat_add(diffstat, name_a, name_b);
2431 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2432
2433 if (!one || !two) {
2434 data->is_unmerged = 1;
2435 return;
2436 }
2437
2438 same_contents = !hashcmp(one->sha1, two->sha1);
2439
2440 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2441 data->is_binary = 1;
2442 if (same_contents) {
2443 data->added = 0;
2444 data->deleted = 0;
2445 } else {
2446 data->added = diff_filespec_size(two);
2447 data->deleted = diff_filespec_size(one);
2448 }
2449 }
2450
2451 else if (complete_rewrite) {
2452 diff_populate_filespec(one, 0);
2453 diff_populate_filespec(two, 0);
2454 data->deleted = count_lines(one->data, one->size);
2455 data->added = count_lines(two->data, two->size);
2456 }
2457
2458 else if (!same_contents) {
2459 /* Crazy xdl interfaces.. */
2460 xpparam_t xpp;
2461 xdemitconf_t xecfg;
2462
2463 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2464 die("unable to read files to diff");
2465
2466 memset(&xpp, 0, sizeof(xpp));
2467 memset(&xecfg, 0, sizeof(xecfg));
2468 xpp.flags = o->xdl_opts;
2469 xecfg.ctxlen = o->context;
2470 xecfg.interhunkctxlen = o->interhunkcontext;
2471 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2472 &xpp, &xecfg);
2473 }
2474
2475 diff_free_filespec_data(one);
2476 diff_free_filespec_data(two);
2477 }
2478
2479 static void builtin_checkdiff(const char *name_a, const char *name_b,
2480 const char *attr_path,
2481 struct diff_filespec *one,
2482 struct diff_filespec *two,
2483 struct diff_options *o)
2484 {
2485 mmfile_t mf1, mf2;
2486 struct checkdiff_t data;
2487
2488 if (!two)
2489 return;
2490
2491 memset(&data, 0, sizeof(data));
2492 data.filename = name_b ? name_b : name_a;
2493 data.lineno = 0;
2494 data.o = o;
2495 data.ws_rule = whitespace_rule(attr_path);
2496 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2497
2498 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2499 die("unable to read files to diff");
2500
2501 /*
2502 * All the other codepaths check both sides, but not checking
2503 * the "old" side here is deliberate. We are checking the newly
2504 * introduced changes, and as long as the "new" side is text, we
2505 * can and should check what it introduces.
2506 */
2507 if (diff_filespec_is_binary(two))
2508 goto free_and_return;
2509 else {
2510 /* Crazy xdl interfaces.. */
2511 xpparam_t xpp;
2512 xdemitconf_t xecfg;
2513
2514 memset(&xpp, 0, sizeof(xpp));
2515 memset(&xecfg, 0, sizeof(xecfg));
2516 xecfg.ctxlen = 1; /* at least one context line */
2517 xpp.flags = 0;
2518 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2519 &xpp, &xecfg);
2520
2521 if (data.ws_rule & WS_BLANK_AT_EOF) {
2522 struct emit_callback ecbdata;
2523 int blank_at_eof;
2524
2525 ecbdata.ws_rule = data.ws_rule;
2526 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2527 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2528
2529 if (blank_at_eof) {
2530 static char *err;
2531 if (!err)
2532 err = whitespace_error_string(WS_BLANK_AT_EOF);
2533 fprintf(o->file, "%s:%d: %s.\n",
2534 data.filename, blank_at_eof, err);
2535 data.status = 1; /* report errors */
2536 }
2537 }
2538 }
2539 free_and_return:
2540 diff_free_filespec_data(one);
2541 diff_free_filespec_data(two);
2542 if (data.status)
2543 DIFF_OPT_SET(o, CHECK_FAILED);
2544 }
2545
2546 struct diff_filespec *alloc_filespec(const char *path)
2547 {
2548 int namelen = strlen(path);
2549 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2550
2551 memset(spec, 0, sizeof(*spec));
2552 spec->path = (char *)(spec + 1);
2553 memcpy(spec->path, path, namelen+1);
2554 spec->count = 1;
2555 spec->is_binary = -1;
2556 return spec;
2557 }
2558
2559 void free_filespec(struct diff_filespec *spec)
2560 {
2561 if (!--spec->count) {
2562 diff_free_filespec_data(spec);
2563 free(spec);
2564 }
2565 }
2566
2567 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2568 int sha1_valid, unsigned short mode)
2569 {
2570 if (mode) {
2571 spec->mode = canon_mode(mode);
2572 hashcpy(spec->sha1, sha1);
2573 spec->sha1_valid = sha1_valid;
2574 }
2575 }
2576
2577 /*
2578 * Given a name and sha1 pair, if the index tells us the file in
2579 * the work tree has that object contents, return true, so that
2580 * prepare_temp_file() does not have to inflate and extract.
2581 */
2582 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2583 {
2584 const struct cache_entry *ce;
2585 struct stat st;
2586 int pos, len;
2587
2588 /*
2589 * We do not read the cache ourselves here, because the
2590 * benchmark with my previous version that always reads cache
2591 * shows that it makes things worse for diff-tree comparing
2592 * two linux-2.6 kernel trees in an already checked out work
2593 * tree. This is because most diff-tree comparisons deal with
2594 * only a small number of files, while reading the cache is
2595 * expensive for a large project, and its cost outweighs the
2596 * savings we get by not inflating the object to a temporary
2597 * file. Practically, this code only helps when we are used
2598 * by diff-cache --cached, which does read the cache before
2599 * calling us.
2600 */
2601 if (!active_cache)
2602 return 0;
2603
2604 /* We want to avoid the working directory if our caller
2605 * doesn't need the data in a normal file, this system
2606 * is rather slow with its stat/open/mmap/close syscalls,
2607 * and the object is contained in a pack file. The pack
2608 * is probably already open and will be faster to obtain
2609 * the data through than the working directory. Loose
2610 * objects however would tend to be slower as they need
2611 * to be individually opened and inflated.
2612 */
2613 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2614 return 0;
2615
2616 len = strlen(name);
2617 pos = cache_name_pos(name, len);
2618 if (pos < 0)
2619 return 0;
2620 ce = active_cache[pos];
2621
2622 /*
2623 * This is not the sha1 we are looking for, or
2624 * unreusable because it is not a regular file.
2625 */
2626 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2627 return 0;
2628
2629 /*
2630 * If ce is marked as "assume unchanged", there is no
2631 * guarantee that work tree matches what we are looking for.
2632 */
2633 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2634 return 0;
2635
2636 /*
2637 * If ce matches the file in the work tree, we can reuse it.
2638 */
2639 if (ce_uptodate(ce) ||
2640 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2641 return 1;
2642
2643 return 0;
2644 }
2645
2646 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2647 {
2648 int len;
2649 char *data = xmalloc(100), *dirty = "";
2650
2651 /* Are we looking at the work tree? */
2652 if (s->dirty_submodule)
2653 dirty = "-dirty";
2654
2655 len = snprintf(data, 100,
2656 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2657 s->data = data;
2658 s->size = len;
2659 s->should_free = 1;
2660 if (size_only) {
2661 s->data = NULL;
2662 free(data);
2663 }
2664 return 0;
2665 }
2666
2667 /*
2668 * While doing rename detection and pickaxe operation, we may need to
2669 * grab the data for the blob (or file) for our own in-core comparison.
2670 * diff_filespec has data and size fields for this purpose.
2671 */
2672 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2673 {
2674 int err = 0;
2675 /*
2676 * demote FAIL to WARN to allow inspecting the situation
2677 * instead of refusing.
2678 */
2679 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
2680 ? SAFE_CRLF_WARN
2681 : safe_crlf);
2682
2683 if (!DIFF_FILE_VALID(s))
2684 die("internal error: asking to populate invalid file.");
2685 if (S_ISDIR(s->mode))
2686 return -1;
2687
2688 if (s->data)
2689 return 0;
2690
2691 if (size_only && 0 < s->size)
2692 return 0;
2693
2694 if (S_ISGITLINK(s->mode))
2695 return diff_populate_gitlink(s, size_only);
2696
2697 if (!s->sha1_valid ||
2698 reuse_worktree_file(s->path, s->sha1, 0)) {
2699 struct strbuf buf = STRBUF_INIT;
2700 struct stat st;
2701 int fd;
2702
2703 if (lstat(s->path, &st) < 0) {
2704 if (errno == ENOENT) {
2705 err_empty:
2706 err = -1;
2707 empty:
2708 s->data = (char *)"";
2709 s->size = 0;
2710 return err;
2711 }
2712 }
2713 s->size = xsize_t(st.st_size);
2714 if (!s->size)
2715 goto empty;
2716 if (S_ISLNK(st.st_mode)) {
2717 struct strbuf sb = STRBUF_INIT;
2718
2719 if (strbuf_readlink(&sb, s->path, s->size))
2720 goto err_empty;
2721 s->size = sb.len;
2722 s->data = strbuf_detach(&sb, NULL);
2723 s->should_free = 1;
2724 return 0;
2725 }
2726 if (size_only)
2727 return 0;
2728 fd = open(s->path, O_RDONLY);
2729 if (fd < 0)
2730 goto err_empty;
2731 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2732 close(fd);
2733 s->should_munmap = 1;
2734
2735 /*
2736 * Convert from working tree format to canonical git format
2737 */
2738 if (convert_to_git(s->path, s->data, s->size, &buf, crlf_warn)) {
2739 size_t size = 0;
2740 munmap(s->data, s->size);
2741 s->should_munmap = 0;
2742 s->data = strbuf_detach(&buf, &size);
2743 s->size = size;
2744 s->should_free = 1;
2745 }
2746 }
2747 else {
2748 enum object_type type;
2749 if (size_only) {
2750 type = sha1_object_info(s->sha1, &s->size);
2751 if (type < 0)
2752 die("unable to read %s", sha1_to_hex(s->sha1));
2753 } else {
2754 s->data = read_sha1_file(s->sha1, &type, &s->size);
2755 if (!s->data)
2756 die("unable to read %s", sha1_to_hex(s->sha1));
2757 s->should_free = 1;
2758 }
2759 }
2760 return 0;
2761 }
2762
2763 void diff_free_filespec_blob(struct diff_filespec *s)
2764 {
2765 if (s->should_free)
2766 free(s->data);
2767 else if (s->should_munmap)
2768 munmap(s->data, s->size);
2769
2770 if (s->should_free || s->should_munmap) {
2771 s->should_free = s->should_munmap = 0;
2772 s->data = NULL;
2773 }
2774 }
2775
2776 void diff_free_filespec_data(struct diff_filespec *s)
2777 {
2778 diff_free_filespec_blob(s);
2779 free(s->cnt_data);
2780 s->cnt_data = NULL;
2781 }
2782
2783 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2784 void *blob,
2785 unsigned long size,
2786 const unsigned char *sha1,
2787 int mode)
2788 {
2789 int fd;
2790 struct strbuf buf = STRBUF_INIT;
2791 struct strbuf template = STRBUF_INIT;
2792 char *path_dup = xstrdup(path);
2793 const char *base = basename(path_dup);
2794
2795 /* Generate "XXXXXX_basename.ext" */
2796 strbuf_addstr(&template, "XXXXXX_");
2797 strbuf_addstr(&template, base);
2798
2799 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2800 strlen(base) + 1);
2801 if (fd < 0)
2802 die_errno("unable to create temp-file");
2803 if (convert_to_working_tree(path,
2804 (const char *)blob, (size_t)size, &buf)) {
2805 blob = buf.buf;
2806 size = buf.len;
2807 }
2808 if (write_in_full(fd, blob, size) != size)
2809 die_errno("unable to write temp-file");
2810 close(fd);
2811 temp->name = temp->tmp_path;
2812 strcpy(temp->hex, sha1_to_hex(sha1));
2813 temp->hex[40] = 0;
2814 sprintf(temp->mode, "%06o", mode);
2815 strbuf_release(&buf);
2816 strbuf_release(&template);
2817 free(path_dup);
2818 }
2819
2820 static struct diff_tempfile *prepare_temp_file(const char *name,
2821 struct diff_filespec *one)
2822 {
2823 struct diff_tempfile *temp = claim_diff_tempfile();
2824
2825 if (!DIFF_FILE_VALID(one)) {
2826 not_a_valid_file:
2827 /* A '-' entry produces this for file-2, and
2828 * a '+' entry produces this for file-1.
2829 */
2830 temp->name = "/dev/null";
2831 strcpy(temp->hex, ".");
2832 strcpy(temp->mode, ".");
2833 return temp;
2834 }
2835
2836 if (!remove_tempfile_installed) {
2837 atexit(remove_tempfile);
2838 sigchain_push_common(remove_tempfile_on_signal);
2839 remove_tempfile_installed = 1;
2840 }
2841
2842 if (!S_ISGITLINK(one->mode) &&
2843 (!one->sha1_valid ||
2844 reuse_worktree_file(name, one->sha1, 1))) {
2845 struct stat st;
2846 if (lstat(name, &st) < 0) {
2847 if (errno == ENOENT)
2848 goto not_a_valid_file;
2849 die_errno("stat(%s)", name);
2850 }
2851 if (S_ISLNK(st.st_mode)) {
2852 struct strbuf sb = STRBUF_INIT;
2853 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2854 die_errno("readlink(%s)", name);
2855 prep_temp_blob(name, temp, sb.buf, sb.len,
2856 (one->sha1_valid ?
2857 one->sha1 : null_sha1),
2858 (one->sha1_valid ?
2859 one->mode : S_IFLNK));
2860 strbuf_release(&sb);
2861 }
2862 else {
2863 /* we can borrow from the file in the work tree */
2864 temp->name = name;
2865 if (!one->sha1_valid)
2866 strcpy(temp->hex, sha1_to_hex(null_sha1));
2867 else
2868 strcpy(temp->hex, sha1_to_hex(one->sha1));
2869 /* Even though we may sometimes borrow the
2870 * contents from the work tree, we always want
2871 * one->mode. mode is trustworthy even when
2872 * !(one->sha1_valid), as long as
2873 * DIFF_FILE_VALID(one).
2874 */
2875 sprintf(temp->mode, "%06o", one->mode);
2876 }
2877 return temp;
2878 }
2879 else {
2880 if (diff_populate_filespec(one, 0))
2881 die("cannot read data blob for %s", one->path);
2882 prep_temp_blob(name, temp, one->data, one->size,
2883 one->sha1, one->mode);
2884 }
2885 return temp;
2886 }
2887
2888 static void add_external_diff_name(struct argv_array *argv,
2889 const char *name,
2890 struct diff_filespec *df)
2891 {
2892 struct diff_tempfile *temp = prepare_temp_file(name, df);
2893 argv_array_push(argv, temp->name);
2894 argv_array_push(argv, temp->hex);
2895 argv_array_push(argv, temp->mode);
2896 }
2897
2898 /* An external diff command takes:
2899 *
2900 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2901 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2902 *
2903 */
2904 static void run_external_diff(const char *pgm,
2905 const char *name,
2906 const char *other,
2907 struct diff_filespec *one,
2908 struct diff_filespec *two,
2909 const char *xfrm_msg,
2910 int complete_rewrite,
2911 struct diff_options *o)
2912 {
2913 struct argv_array argv = ARGV_ARRAY_INIT;
2914 struct argv_array env = ARGV_ARRAY_INIT;
2915 struct diff_queue_struct *q = &diff_queued_diff;
2916
2917 argv_array_push(&argv, pgm);
2918 argv_array_push(&argv, name);
2919
2920 if (one && two) {
2921 add_external_diff_name(&argv, name, one);
2922 if (!other)
2923 add_external_diff_name(&argv, name, two);
2924 else {
2925 add_external_diff_name(&argv, other, two);
2926 argv_array_push(&argv, other);
2927 argv_array_push(&argv, xfrm_msg);
2928 }
2929 }
2930
2931 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
2932 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
2933
2934 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
2935 die(_("external diff died, stopping at %s"), name);
2936
2937 remove_tempfile();
2938 argv_array_clear(&argv);
2939 argv_array_clear(&env);
2940 }
2941
2942 static int similarity_index(struct diff_filepair *p)
2943 {
2944 return p->score * 100 / MAX_SCORE;
2945 }
2946
2947 static void fill_metainfo(struct strbuf *msg,
2948 const char *name,
2949 const char *other,
2950 struct diff_filespec *one,
2951 struct diff_filespec *two,
2952 struct diff_options *o,
2953 struct diff_filepair *p,
2954 int *must_show_header,
2955 int use_color)
2956 {
2957 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2958 const char *reset = diff_get_color(use_color, DIFF_RESET);
2959 const char *line_prefix = diff_line_prefix(o);
2960
2961 *must_show_header = 1;
2962 strbuf_init(msg, PATH_MAX * 2 + 300);
2963 switch (p->status) {
2964 case DIFF_STATUS_COPIED:
2965 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2966 line_prefix, set, similarity_index(p));
2967 strbuf_addf(msg, "%s\n%s%scopy from ",
2968 reset, line_prefix, set);
2969 quote_c_style(name, msg, NULL, 0);
2970 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2971 quote_c_style(other, msg, NULL, 0);
2972 strbuf_addf(msg, "%s\n", reset);
2973 break;
2974 case DIFF_STATUS_RENAMED:
2975 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2976 line_prefix, set, similarity_index(p));
2977 strbuf_addf(msg, "%s\n%s%srename from ",
2978 reset, line_prefix, set);
2979 quote_c_style(name, msg, NULL, 0);
2980 strbuf_addf(msg, "%s\n%s%srename to ",
2981 reset, line_prefix, set);
2982 quote_c_style(other, msg, NULL, 0);
2983 strbuf_addf(msg, "%s\n", reset);
2984 break;
2985 case DIFF_STATUS_MODIFIED:
2986 if (p->score) {
2987 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2988 line_prefix,
2989 set, similarity_index(p), reset);
2990 break;
2991 }
2992 /* fallthru */
2993 default:
2994 *must_show_header = 0;
2995 }
2996 if (one && two && hashcmp(one->sha1, two->sha1)) {
2997 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2998
2999 if (DIFF_OPT_TST(o, BINARY)) {
3000 mmfile_t mf;
3001 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3002 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3003 abbrev = 40;
3004 }
3005 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
3006 find_unique_abbrev(one->sha1, abbrev));
3007 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
3008 if (one->mode == two->mode)
3009 strbuf_addf(msg, " %06o", one->mode);
3010 strbuf_addf(msg, "%s\n", reset);
3011 }
3012 }
3013
3014 static void run_diff_cmd(const char *pgm,
3015 const char *name,
3016 const char *other,
3017 const char *attr_path,
3018 struct diff_filespec *one,
3019 struct diff_filespec *two,
3020 struct strbuf *msg,
3021 struct diff_options *o,
3022 struct diff_filepair *p)
3023 {
3024 const char *xfrm_msg = NULL;
3025 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3026 int must_show_header = 0;
3027
3028
3029 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3030 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3031 if (drv && drv->external)
3032 pgm = drv->external;
3033 }
3034
3035 if (msg) {
3036 /*
3037 * don't use colors when the header is intended for an
3038 * external diff driver
3039 */
3040 fill_metainfo(msg, name, other, one, two, o, p,
3041 &must_show_header,
3042 want_color(o->use_color) && !pgm);
3043 xfrm_msg = msg->len ? msg->buf : NULL;
3044 }
3045
3046 if (pgm) {
3047 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3048 complete_rewrite, o);
3049 return;
3050 }
3051 if (one && two)
3052 builtin_diff(name, other ? other : name,
3053 one, two, xfrm_msg, must_show_header,
3054 o, complete_rewrite);
3055 else
3056 fprintf(o->file, "* Unmerged path %s\n", name);
3057 }
3058
3059 static void diff_fill_sha1_info(struct diff_filespec *one)
3060 {
3061 if (DIFF_FILE_VALID(one)) {
3062 if (!one->sha1_valid) {
3063 struct stat st;
3064 if (one->is_stdin) {
3065 hashcpy(one->sha1, null_sha1);
3066 return;
3067 }
3068 if (lstat(one->path, &st) < 0)
3069 die_errno("stat '%s'", one->path);
3070 if (index_path(one->sha1, one->path, &st, 0))
3071 die("cannot hash %s", one->path);
3072 }
3073 }
3074 else
3075 hashclr(one->sha1);
3076 }
3077
3078 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3079 {
3080 /* Strip the prefix but do not molest /dev/null and absolute paths */
3081 if (*namep && **namep != '/') {
3082 *namep += prefix_length;
3083 if (**namep == '/')
3084 ++*namep;
3085 }
3086 if (*otherp && **otherp != '/') {
3087 *otherp += prefix_length;
3088 if (**otherp == '/')
3089 ++*otherp;
3090 }
3091 }
3092
3093 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3094 {
3095 const char *pgm = external_diff();
3096 struct strbuf msg;
3097 struct diff_filespec *one = p->one;
3098 struct diff_filespec *two = p->two;
3099 const char *name;
3100 const char *other;
3101 const char *attr_path;
3102
3103 name = p->one->path;
3104 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3105 attr_path = name;
3106 if (o->prefix_length)
3107 strip_prefix(o->prefix_length, &name, &other);
3108
3109 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3110 pgm = NULL;
3111
3112 if (DIFF_PAIR_UNMERGED(p)) {
3113 run_diff_cmd(pgm, name, NULL, attr_path,
3114 NULL, NULL, NULL, o, p);
3115 return;
3116 }
3117
3118 diff_fill_sha1_info(one);
3119 diff_fill_sha1_info(two);
3120
3121 if (!pgm &&
3122 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3123 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3124 /*
3125 * a filepair that changes between file and symlink
3126 * needs to be split into deletion and creation.
3127 */
3128 struct diff_filespec *null = alloc_filespec(two->path);
3129 run_diff_cmd(NULL, name, other, attr_path,
3130 one, null, &msg, o, p);
3131 free(null);
3132 strbuf_release(&msg);
3133
3134 null = alloc_filespec(one->path);
3135 run_diff_cmd(NULL, name, other, attr_path,
3136 null, two, &msg, o, p);
3137 free(null);
3138 }
3139 else
3140 run_diff_cmd(pgm, name, other, attr_path,
3141 one, two, &msg, o, p);
3142
3143 strbuf_release(&msg);
3144 }
3145
3146 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3147 struct diffstat_t *diffstat)
3148 {
3149 const char *name;
3150 const char *other;
3151
3152 if (DIFF_PAIR_UNMERGED(p)) {
3153 /* unmerged */
3154 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3155 return;
3156 }
3157
3158 name = p->one->path;
3159 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3160
3161 if (o->prefix_length)
3162 strip_prefix(o->prefix_length, &name, &other);
3163
3164 diff_fill_sha1_info(p->one);
3165 diff_fill_sha1_info(p->two);
3166
3167 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3168 }
3169
3170 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3171 {
3172 const char *name;
3173 const char *other;
3174 const char *attr_path;
3175
3176 if (DIFF_PAIR_UNMERGED(p)) {
3177 /* unmerged */
3178 return;
3179 }
3180
3181 name = p->one->path;
3182 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3183 attr_path = other ? other : name;
3184
3185 if (o->prefix_length)
3186 strip_prefix(o->prefix_length, &name, &other);
3187
3188 diff_fill_sha1_info(p->one);
3189 diff_fill_sha1_info(p->two);
3190
3191 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3192 }
3193
3194 void diff_setup(struct diff_options *options)
3195 {
3196 memcpy(options, &default_diff_options, sizeof(*options));
3197
3198 options->file = stdout;
3199
3200 options->line_termination = '\n';
3201 options->break_opt = -1;
3202 options->rename_limit = -1;
3203 options->dirstat_permille = diff_dirstat_permille_default;
3204 options->context = diff_context_default;
3205 DIFF_OPT_SET(options, RENAME_EMPTY);
3206
3207 /* pathchange left =NULL by default */
3208 options->change = diff_change;
3209 options->add_remove = diff_addremove;
3210 options->use_color = diff_use_color_default;
3211 options->detect_rename = diff_detect_rename_default;
3212 options->xdl_opts |= diff_algorithm;
3213
3214 options->orderfile = diff_order_file_cfg;
3215
3216 if (diff_no_prefix) {
3217 options->a_prefix = options->b_prefix = "";
3218 } else if (!diff_mnemonic_prefix) {
3219 options->a_prefix = "a/";
3220 options->b_prefix = "b/";
3221 }
3222 }
3223
3224 void diff_setup_done(struct diff_options *options)
3225 {
3226 int count = 0;
3227
3228 if (options->set_default)
3229 options->set_default(options);
3230
3231 if (options->output_format & DIFF_FORMAT_NAME)
3232 count++;
3233 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3234 count++;
3235 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3236 count++;
3237 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3238 count++;
3239 if (count > 1)
3240 die("--name-only, --name-status, --check and -s are mutually exclusive");
3241
3242 /*
3243 * Most of the time we can say "there are changes"
3244 * only by checking if there are changed paths, but
3245 * --ignore-whitespace* options force us to look
3246 * inside contents.
3247 */
3248
3249 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3250 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3251 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3252 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3253 else
3254 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3255
3256 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3257 options->detect_rename = DIFF_DETECT_COPY;
3258
3259 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3260 options->prefix = NULL;
3261 if (options->prefix)
3262 options->prefix_length = strlen(options->prefix);
3263 else
3264 options->prefix_length = 0;
3265
3266 if (options->output_format & (DIFF_FORMAT_NAME |
3267 DIFF_FORMAT_NAME_STATUS |
3268 DIFF_FORMAT_CHECKDIFF |
3269 DIFF_FORMAT_NO_OUTPUT))
3270 options->output_format &= ~(DIFF_FORMAT_RAW |
3271 DIFF_FORMAT_NUMSTAT |
3272 DIFF_FORMAT_DIFFSTAT |
3273 DIFF_FORMAT_SHORTSTAT |
3274 DIFF_FORMAT_DIRSTAT |
3275 DIFF_FORMAT_SUMMARY |
3276 DIFF_FORMAT_PATCH);
3277
3278 /*
3279 * These cases always need recursive; we do not drop caller-supplied
3280 * recursive bits for other formats here.
3281 */
3282 if (options->output_format & (DIFF_FORMAT_PATCH |
3283 DIFF_FORMAT_NUMSTAT |
3284 DIFF_FORMAT_DIFFSTAT |
3285 DIFF_FORMAT_SHORTSTAT |
3286 DIFF_FORMAT_DIRSTAT |
3287 DIFF_FORMAT_SUMMARY |
3288 DIFF_FORMAT_CHECKDIFF))
3289 DIFF_OPT_SET(options, RECURSIVE);
3290 /*
3291 * Also pickaxe would not work very well if you do not say recursive
3292 */
3293 if (options->pickaxe)
3294 DIFF_OPT_SET(options, RECURSIVE);
3295 /*
3296 * When patches are generated, submodules diffed against the work tree
3297 * must be checked for dirtiness too so it can be shown in the output
3298 */
3299 if (options->output_format & DIFF_FORMAT_PATCH)
3300 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3301
3302 if (options->detect_rename && options->rename_limit < 0)
3303 options->rename_limit = diff_rename_limit_default;
3304 if (options->setup & DIFF_SETUP_USE_CACHE) {
3305 if (!active_cache)
3306 /* read-cache does not die even when it fails
3307 * so it is safe for us to do this here. Also
3308 * it does not smudge active_cache or active_nr
3309 * when it fails, so we do not have to worry about
3310 * cleaning it up ourselves either.
3311 */
3312 read_cache();
3313 }
3314 if (options->abbrev <= 0 || 40 < options->abbrev)
3315 options->abbrev = 40; /* full */
3316
3317 /*
3318 * It does not make sense to show the first hit we happened
3319 * to have found. It does not make sense not to return with
3320 * exit code in such a case either.
3321 */
3322 if (DIFF_OPT_TST(options, QUICK)) {
3323 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3324 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3325 }
3326
3327 options->diff_path_counter = 0;
3328
3329 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3330 die(_("--follow requires exactly one pathspec"));
3331 }
3332
3333 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3334 {
3335 char c, *eq;
3336 int len;
3337
3338 if (*arg != '-')
3339 return 0;
3340 c = *++arg;
3341 if (!c)
3342 return 0;
3343 if (c == arg_short) {
3344 c = *++arg;
3345 if (!c)
3346 return 1;
3347 if (val && isdigit(c)) {
3348 char *end;
3349 int n = strtoul(arg, &end, 10);
3350 if (*end)
3351 return 0;
3352 *val = n;
3353 return 1;
3354 }
3355 return 0;
3356 }
3357 if (c != '-')
3358 return 0;
3359 arg++;
3360 eq = strchrnul(arg, '=');
3361 len = eq - arg;
3362 if (!len || strncmp(arg, arg_long, len))
3363 return 0;
3364 if (*eq) {
3365 int n;
3366 char *end;
3367 if (!isdigit(*++eq))
3368 return 0;
3369 n = strtoul(eq, &end, 10);
3370 if (*end)
3371 return 0;
3372 *val = n;
3373 }
3374 return 1;
3375 }
3376
3377 static int diff_scoreopt_parse(const char *opt);
3378
3379 static inline int short_opt(char opt, const char **argv,
3380 const char **optarg)
3381 {
3382 const char *arg = argv[0];
3383 if (arg[0] != '-' || arg[1] != opt)
3384 return 0;
3385 if (arg[2] != '\0') {
3386 *optarg = arg + 2;
3387 return 1;
3388 }
3389 if (!argv[1])
3390 die("Option '%c' requires a value", opt);
3391 *optarg = argv[1];
3392 return 2;
3393 }
3394
3395 int parse_long_opt(const char *opt, const char **argv,
3396 const char **optarg)
3397 {
3398 const char *arg = argv[0];
3399 if (!skip_prefix(arg, "--", &arg))
3400 return 0;
3401 if (!skip_prefix(arg, opt, &arg))
3402 return 0;
3403 if (*arg == '=') { /* stuck form: --option=value */
3404 *optarg = arg + 1;
3405 return 1;
3406 }
3407 if (*arg != '\0')
3408 return 0;
3409 /* separate form: --option value */
3410 if (!argv[1])
3411 die("Option '--%s' requires a value", opt);
3412 *optarg = argv[1];
3413 return 2;
3414 }
3415
3416 static int stat_opt(struct diff_options *options, const char **av)
3417 {
3418 const char *arg = av[0];
3419 char *end;
3420 int width = options->stat_width;
3421 int name_width = options->stat_name_width;
3422 int graph_width = options->stat_graph_width;
3423 int count = options->stat_count;
3424 int argcount = 1;
3425
3426 if (!skip_prefix(arg, "--stat", &arg))
3427 die("BUG: stat option does not begin with --stat: %s", arg);
3428 end = (char *)arg;
3429
3430 switch (*arg) {
3431 case '-':
3432 if (skip_prefix(arg, "-width", &arg)) {
3433 if (*arg == '=')
3434 width = strtoul(arg + 1, &end, 10);
3435 else if (!*arg && !av[1])
3436 die("Option '--stat-width' requires a value");
3437 else if (!*arg) {
3438 width = strtoul(av[1], &end, 10);
3439 argcount = 2;
3440 }
3441 } else if (skip_prefix(arg, "-name-width", &arg)) {
3442 if (*arg == '=')
3443 name_width = strtoul(arg + 1, &end, 10);
3444 else if (!*arg && !av[1])
3445 die("Option '--stat-name-width' requires a value");
3446 else if (!*arg) {
3447 name_width = strtoul(av[1], &end, 10);
3448 argcount = 2;
3449 }
3450 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3451 if (*arg == '=')
3452 graph_width = strtoul(arg + 1, &end, 10);
3453 else if (!*arg && !av[1])
3454 die("Option '--stat-graph-width' requires a value");
3455 else if (!*arg) {
3456 graph_width = strtoul(av[1], &end, 10);
3457 argcount = 2;
3458 }
3459 } else if (skip_prefix(arg, "-count", &arg)) {
3460 if (*arg == '=')
3461 count = strtoul(arg + 1, &end, 10);
3462 else if (!*arg && !av[1])
3463 die("Option '--stat-count' requires a value");
3464 else if (!*arg) {
3465 count = strtoul(av[1], &end, 10);
3466 argcount = 2;
3467 }
3468 }
3469 break;
3470 case '=':
3471 width = strtoul(arg+1, &end, 10);
3472 if (*end == ',')
3473 name_width = strtoul(end+1, &end, 10);
3474 if (*end == ',')
3475 count = strtoul(end+1, &end, 10);
3476 }
3477
3478 /* Important! This checks all the error cases! */
3479 if (*end)
3480 return 0;
3481 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3482 options->stat_name_width = name_width;
3483 options->stat_graph_width = graph_width;
3484 options->stat_width = width;
3485 options->stat_count = count;
3486 return argcount;
3487 }
3488
3489 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3490 {
3491 struct strbuf errmsg = STRBUF_INIT;
3492 if (parse_dirstat_params(options, params, &errmsg))
3493 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3494 errmsg.buf);
3495 strbuf_release(&errmsg);
3496 /*
3497 * The caller knows a dirstat-related option is given from the command
3498 * line; allow it to say "return this_function();"
3499 */
3500 options->output_format |= DIFF_FORMAT_DIRSTAT;
3501 return 1;
3502 }
3503
3504 static int parse_submodule_opt(struct diff_options *options, const char *value)
3505 {
3506 if (parse_submodule_params(options, value))
3507 die(_("Failed to parse --submodule option parameter: '%s'"),
3508 value);
3509 return 1;
3510 }
3511
3512 static const char diff_status_letters[] = {
3513 DIFF_STATUS_ADDED,
3514 DIFF_STATUS_COPIED,
3515 DIFF_STATUS_DELETED,
3516 DIFF_STATUS_MODIFIED,
3517 DIFF_STATUS_RENAMED,
3518 DIFF_STATUS_TYPE_CHANGED,
3519 DIFF_STATUS_UNKNOWN,
3520 DIFF_STATUS_UNMERGED,
3521 DIFF_STATUS_FILTER_AON,
3522 DIFF_STATUS_FILTER_BROKEN,
3523 '\0',
3524 };
3525
3526 static unsigned int filter_bit['Z' + 1];
3527
3528 static void prepare_filter_bits(void)
3529 {
3530 int i;
3531
3532 if (!filter_bit[DIFF_STATUS_ADDED]) {
3533 for (i = 0; diff_status_letters[i]; i++)
3534 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3535 }
3536 }
3537
3538 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3539 {
3540 return opt->filter & filter_bit[(int) status];
3541 }
3542
3543 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3544 {
3545 int i, optch;
3546
3547 prepare_filter_bits();
3548
3549 /*
3550 * If there is a negation e.g. 'd' in the input, and we haven't
3551 * initialized the filter field with another --diff-filter, start
3552 * from full set of bits, except for AON.
3553 */
3554 if (!opt->filter) {
3555 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3556 if (optch < 'a' || 'z' < optch)
3557 continue;
3558 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3559 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3560 break;
3561 }
3562 }
3563
3564 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3565 unsigned int bit;
3566 int negate;
3567
3568 if ('a' <= optch && optch <= 'z') {
3569 negate = 1;
3570 optch = toupper(optch);
3571 } else {
3572 negate = 0;
3573 }
3574
3575 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
3576 if (!bit)
3577 return optarg[i];
3578 if (negate)
3579 opt->filter &= ~bit;
3580 else
3581 opt->filter |= bit;
3582 }
3583 return 0;
3584 }
3585
3586 static void enable_patch_output(int *fmt) {
3587 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
3588 *fmt |= DIFF_FORMAT_PATCH;
3589 }
3590
3591 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3592 {
3593 const char *arg = av[0];
3594 const char *optarg;
3595 int argcount;
3596
3597 /* Output format options */
3598 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
3599 || opt_arg(arg, 'U', "unified", &options->context))
3600 enable_patch_output(&options->output_format);
3601 else if (!strcmp(arg, "--raw"))
3602 options->output_format |= DIFF_FORMAT_RAW;
3603 else if (!strcmp(arg, "--patch-with-raw")) {
3604 enable_patch_output(&options->output_format);
3605 options->output_format |= DIFF_FORMAT_RAW;
3606 } else if (!strcmp(arg, "--numstat"))
3607 options->output_format |= DIFF_FORMAT_NUMSTAT;
3608 else if (!strcmp(arg, "--shortstat"))
3609 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3610 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3611 return parse_dirstat_opt(options, "");
3612 else if (skip_prefix(arg, "-X", &arg))
3613 return parse_dirstat_opt(options, arg);
3614 else if (skip_prefix(arg, "--dirstat=", &arg))
3615 return parse_dirstat_opt(options, arg);
3616 else if (!strcmp(arg, "--cumulative"))
3617 return parse_dirstat_opt(options, "cumulative");
3618 else if (!strcmp(arg, "--dirstat-by-file"))
3619 return parse_dirstat_opt(options, "files");
3620 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
3621 parse_dirstat_opt(options, "files");
3622 return parse_dirstat_opt(options, arg);
3623 }
3624 else if (!strcmp(arg, "--check"))
3625 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3626 else if (!strcmp(arg, "--summary"))
3627 options->output_format |= DIFF_FORMAT_SUMMARY;
3628 else if (!strcmp(arg, "--patch-with-stat")) {
3629 enable_patch_output(&options->output_format);
3630 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3631 } else if (!strcmp(arg, "--name-only"))
3632 options->output_format |= DIFF_FORMAT_NAME;
3633 else if (!strcmp(arg, "--name-status"))
3634 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3635 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
3636 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3637 else if (starts_with(arg, "--stat"))
3638 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3639 return stat_opt(options, av);
3640
3641 /* renames options */
3642 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
3643 !strcmp(arg, "--break-rewrites")) {
3644 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3645 return error("invalid argument to -B: %s", arg+2);
3646 }
3647 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
3648 !strcmp(arg, "--find-renames")) {
3649 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3650 return error("invalid argument to -M: %s", arg+2);
3651 options->detect_rename = DIFF_DETECT_RENAME;
3652 }
3653 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3654 options->irreversible_delete = 1;
3655 }
3656 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
3657 !strcmp(arg, "--find-copies")) {
3658 if (options->detect_rename == DIFF_DETECT_COPY)
3659 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3660 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3661 return error("invalid argument to -C: %s", arg+2);
3662 options->detect_rename = DIFF_DETECT_COPY;
3663 }
3664 else if (!strcmp(arg, "--no-renames"))
3665 options->detect_rename = 0;
3666 else if (!strcmp(arg, "--rename-empty"))
3667 DIFF_OPT_SET(options, RENAME_EMPTY);
3668 else if (!strcmp(arg, "--no-rename-empty"))
3669 DIFF_OPT_CLR(options, RENAME_EMPTY);
3670 else if (!strcmp(arg, "--relative"))
3671 DIFF_OPT_SET(options, RELATIVE_NAME);
3672 else if (skip_prefix(arg, "--relative=", &arg)) {
3673 DIFF_OPT_SET(options, RELATIVE_NAME);
3674 options->prefix = arg;
3675 }
3676
3677 /* xdiff options */
3678 else if (!strcmp(arg, "--minimal"))
3679 DIFF_XDL_SET(options, NEED_MINIMAL);
3680 else if (!strcmp(arg, "--no-minimal"))
3681 DIFF_XDL_CLR(options, NEED_MINIMAL);
3682 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3683 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3684 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3685 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3686 else if (!strcmp(arg, "--ignore-space-at-eol"))
3687 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3688 else if (!strcmp(arg, "--ignore-blank-lines"))
3689 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
3690 else if (!strcmp(arg, "--patience"))
3691 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3692 else if (!strcmp(arg, "--histogram"))
3693 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3694 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
3695 long value = parse_algorithm_value(optarg);
3696 if (value < 0)
3697 return error("option diff-algorithm accepts \"myers\", "
3698 "\"minimal\", \"patience\" and \"histogram\"");
3699 /* clear out previous settings */
3700 DIFF_XDL_CLR(options, NEED_MINIMAL);
3701 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3702 options->xdl_opts |= value;
3703 return argcount;
3704 }
3705
3706 /* flags options */
3707 else if (!strcmp(arg, "--binary")) {
3708 enable_patch_output(&options->output_format);
3709 DIFF_OPT_SET(options, BINARY);
3710 }
3711 else if (!strcmp(arg, "--full-index"))
3712 DIFF_OPT_SET(options, FULL_INDEX);
3713 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3714 DIFF_OPT_SET(options, TEXT);
3715 else if (!strcmp(arg, "-R"))
3716 DIFF_OPT_SET(options, REVERSE_DIFF);
3717 else if (!strcmp(arg, "--find-copies-harder"))
3718 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3719 else if (!strcmp(arg, "--follow"))
3720 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3721 else if (!strcmp(arg, "--no-follow"))
3722 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3723 else if (!strcmp(arg, "--color"))
3724 options->use_color = 1;
3725 else if (skip_prefix(arg, "--color=", &arg)) {
3726 int value = git_config_colorbool(NULL, arg);
3727 if (value < 0)
3728 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3729 options->use_color = value;
3730 }
3731 else if (!strcmp(arg, "--no-color"))
3732 options->use_color = 0;
3733 else if (!strcmp(arg, "--color-words")) {
3734 options->use_color = 1;
3735 options->word_diff = DIFF_WORDS_COLOR;
3736 }
3737 else if (skip_prefix(arg, "--color-words=", &arg)) {
3738 options->use_color = 1;
3739 options->word_diff = DIFF_WORDS_COLOR;
3740 options->word_regex = arg;
3741 }
3742 else if (!strcmp(arg, "--word-diff")) {
3743 if (options->word_diff == DIFF_WORDS_NONE)
3744 options->word_diff = DIFF_WORDS_PLAIN;
3745 }
3746 else if (skip_prefix(arg, "--word-diff=", &arg)) {
3747 if (!strcmp(arg, "plain"))
3748 options->word_diff = DIFF_WORDS_PLAIN;
3749 else if (!strcmp(arg, "color")) {
3750 options->use_color = 1;
3751 options->word_diff = DIFF_WORDS_COLOR;
3752 }
3753 else if (!strcmp(arg, "porcelain"))
3754 options->word_diff = DIFF_WORDS_PORCELAIN;
3755 else if (!strcmp(arg, "none"))
3756 options->word_diff = DIFF_WORDS_NONE;
3757 else
3758 die("bad --word-diff argument: %s", arg);
3759 }
3760 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3761 if (options->word_diff == DIFF_WORDS_NONE)
3762 options->word_diff = DIFF_WORDS_PLAIN;
3763 options->word_regex = optarg;
3764 return argcount;
3765 }
3766 else if (!strcmp(arg, "--exit-code"))
3767 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3768 else if (!strcmp(arg, "--quiet"))
3769 DIFF_OPT_SET(options, QUICK);
3770 else if (!strcmp(arg, "--ext-diff"))
3771 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3772 else if (!strcmp(arg, "--no-ext-diff"))
3773 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3774 else if (!strcmp(arg, "--textconv"))
3775 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3776 else if (!strcmp(arg, "--no-textconv"))
3777 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3778 else if (!strcmp(arg, "--ignore-submodules")) {
3779 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3780 handle_ignore_submodules_arg(options, "all");
3781 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
3782 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3783 handle_ignore_submodules_arg(options, arg);
3784 } else if (!strcmp(arg, "--submodule"))
3785 DIFF_OPT_SET(options, SUBMODULE_LOG);
3786 else if (skip_prefix(arg, "--submodule=", &arg))
3787 return parse_submodule_opt(options, arg);
3788
3789 /* misc options */
3790 else if (!strcmp(arg, "-z"))
3791 options->line_termination = 0;
3792 else if ((argcount = short_opt('l', av, &optarg))) {
3793 options->rename_limit = strtoul(optarg, NULL, 10);
3794 return argcount;
3795 }
3796 else if ((argcount = short_opt('S', av, &optarg))) {
3797 options->pickaxe = optarg;
3798 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3799 return argcount;
3800 } else if ((argcount = short_opt('G', av, &optarg))) {
3801 options->pickaxe = optarg;
3802 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3803 return argcount;
3804 }
3805 else if (!strcmp(arg, "--pickaxe-all"))
3806 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3807 else if (!strcmp(arg, "--pickaxe-regex"))
3808 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3809 else if ((argcount = short_opt('O', av, &optarg))) {
3810 options->orderfile = optarg;
3811 return argcount;
3812 }
3813 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3814 int offending = parse_diff_filter_opt(optarg, options);
3815 if (offending)
3816 die("unknown change class '%c' in --diff-filter=%s",
3817 offending, optarg);
3818 return argcount;
3819 }
3820 else if (!strcmp(arg, "--abbrev"))
3821 options->abbrev = DEFAULT_ABBREV;
3822 else if (skip_prefix(arg, "--abbrev=", &arg)) {
3823 options->abbrev = strtoul(arg, NULL, 10);
3824 if (options->abbrev < MINIMUM_ABBREV)
3825 options->abbrev = MINIMUM_ABBREV;
3826 else if (40 < options->abbrev)
3827 options->abbrev = 40;
3828 }
3829 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3830 options->a_prefix = optarg;
3831 return argcount;
3832 }
3833 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3834 options->b_prefix = optarg;
3835 return argcount;
3836 }
3837 else if (!strcmp(arg, "--no-prefix"))
3838 options->a_prefix = options->b_prefix = "";
3839 else if (opt_arg(arg, '\0', "inter-hunk-context",
3840 &options->interhunkcontext))
3841 ;
3842 else if (!strcmp(arg, "-W"))
3843 DIFF_OPT_SET(options, FUNCCONTEXT);
3844 else if (!strcmp(arg, "--function-context"))
3845 DIFF_OPT_SET(options, FUNCCONTEXT);
3846 else if (!strcmp(arg, "--no-function-context"))
3847 DIFF_OPT_CLR(options, FUNCCONTEXT);
3848 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3849 options->file = fopen(optarg, "w");
3850 if (!options->file)
3851 die_errno("Could not open '%s'", optarg);
3852 options->close_file = 1;
3853 return argcount;
3854 } else
3855 return 0;
3856 return 1;
3857 }
3858
3859 int parse_rename_score(const char **cp_p)
3860 {
3861 unsigned long num, scale;
3862 int ch, dot;
3863 const char *cp = *cp_p;
3864
3865 num = 0;
3866 scale = 1;
3867 dot = 0;
3868 for (;;) {
3869 ch = *cp;
3870 if ( !dot && ch == '.' ) {
3871 scale = 1;
3872 dot = 1;
3873 } else if ( ch == '%' ) {
3874 scale = dot ? scale*100 : 100;
3875 cp++; /* % is always at the end */
3876 break;
3877 } else if ( ch >= '0' && ch <= '9' ) {
3878 if ( scale < 100000 ) {
3879 scale *= 10;
3880 num = (num*10) + (ch-'0');
3881 }
3882 } else {
3883 break;
3884 }
3885 cp++;
3886 }
3887 *cp_p = cp;
3888
3889 /* user says num divided by scale and we say internally that
3890 * is MAX_SCORE * num / scale.
3891 */
3892 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3893 }
3894
3895 static int diff_scoreopt_parse(const char *opt)
3896 {
3897 int opt1, opt2, cmd;
3898
3899 if (*opt++ != '-')
3900 return -1;
3901 cmd = *opt++;
3902 if (cmd == '-') {
3903 /* convert the long-form arguments into short-form versions */
3904 if (skip_prefix(opt, "break-rewrites", &opt)) {
3905 if (*opt == 0 || *opt++ == '=')
3906 cmd = 'B';
3907 } else if (skip_prefix(opt, "find-copies", &opt)) {
3908 if (*opt == 0 || *opt++ == '=')
3909 cmd = 'C';
3910 } else if (skip_prefix(opt, "find-renames", &opt)) {
3911 if (*opt == 0 || *opt++ == '=')
3912 cmd = 'M';
3913 }
3914 }
3915 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3916 return -1; /* that is not a -M, -C, or -B option */
3917
3918 opt1 = parse_rename_score(&opt);
3919 if (cmd != 'B')
3920 opt2 = 0;
3921 else {
3922 if (*opt == 0)
3923 opt2 = 0;
3924 else if (*opt != '/')
3925 return -1; /* we expect -B80/99 or -B80 */
3926 else {
3927 opt++;
3928 opt2 = parse_rename_score(&opt);
3929 }
3930 }
3931 if (*opt != 0)
3932 return -1;
3933 return opt1 | (opt2 << 16);
3934 }
3935
3936 struct diff_queue_struct diff_queued_diff;
3937
3938 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3939 {
3940 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
3941 queue->queue[queue->nr++] = dp;
3942 }
3943
3944 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3945 struct diff_filespec *one,
3946 struct diff_filespec *two)
3947 {
3948 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3949 dp->one = one;
3950 dp->two = two;
3951 if (queue)
3952 diff_q(queue, dp);
3953 return dp;
3954 }
3955
3956 void diff_free_filepair(struct diff_filepair *p)
3957 {
3958 free_filespec(p->one);
3959 free_filespec(p->two);
3960 free(p);
3961 }
3962
3963 /* This is different from find_unique_abbrev() in that
3964 * it stuffs the result with dots for alignment.
3965 */
3966 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3967 {
3968 int abblen;
3969 const char *abbrev;
3970 if (len == 40)
3971 return sha1_to_hex(sha1);
3972
3973 abbrev = find_unique_abbrev(sha1, len);
3974 abblen = strlen(abbrev);
3975 if (abblen < 37) {
3976 static char hex[41];
3977 if (len < abblen && abblen <= len + 2)
3978 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3979 else
3980 sprintf(hex, "%s...", abbrev);
3981 return hex;
3982 }
3983 return sha1_to_hex(sha1);
3984 }
3985
3986 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3987 {
3988 int line_termination = opt->line_termination;
3989 int inter_name_termination = line_termination ? '\t' : '\0';
3990
3991 fprintf(opt->file, "%s", diff_line_prefix(opt));
3992 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3993 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3994 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3995 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3996 }
3997 if (p->score) {
3998 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3999 inter_name_termination);
4000 } else {
4001 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4002 }
4003
4004 if (p->status == DIFF_STATUS_COPIED ||
4005 p->status == DIFF_STATUS_RENAMED) {
4006 const char *name_a, *name_b;
4007 name_a = p->one->path;
4008 name_b = p->two->path;
4009 strip_prefix(opt->prefix_length, &name_a, &name_b);
4010 write_name_quoted(name_a, opt->file, inter_name_termination);
4011 write_name_quoted(name_b, opt->file, line_termination);
4012 } else {
4013 const char *name_a, *name_b;
4014 name_a = p->one->mode ? p->one->path : p->two->path;
4015 name_b = NULL;
4016 strip_prefix(opt->prefix_length, &name_a, &name_b);
4017 write_name_quoted(name_a, opt->file, line_termination);
4018 }
4019 }
4020
4021 int diff_unmodified_pair(struct diff_filepair *p)
4022 {
4023 /* This function is written stricter than necessary to support
4024 * the currently implemented transformers, but the idea is to
4025 * let transformers to produce diff_filepairs any way they want,
4026 * and filter and clean them up here before producing the output.
4027 */
4028 struct diff_filespec *one = p->one, *two = p->two;
4029
4030 if (DIFF_PAIR_UNMERGED(p))
4031 return 0; /* unmerged is interesting */
4032
4033 /* deletion, addition, mode or type change
4034 * and rename are all interesting.
4035 */
4036 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4037 DIFF_PAIR_MODE_CHANGED(p) ||
4038 strcmp(one->path, two->path))
4039 return 0;
4040
4041 /* both are valid and point at the same path. that is, we are
4042 * dealing with a change.
4043 */
4044 if (one->sha1_valid && two->sha1_valid &&
4045 !hashcmp(one->sha1, two->sha1) &&
4046 !one->dirty_submodule && !two->dirty_submodule)
4047 return 1; /* no change */
4048 if (!one->sha1_valid && !two->sha1_valid)
4049 return 1; /* both look at the same file on the filesystem. */
4050 return 0;
4051 }
4052
4053 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4054 {
4055 if (diff_unmodified_pair(p))
4056 return;
4057
4058 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4059 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4060 return; /* no tree diffs in patch format */
4061
4062 run_diff(p, o);
4063 }
4064
4065 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4066 struct diffstat_t *diffstat)
4067 {
4068 if (diff_unmodified_pair(p))
4069 return;
4070
4071 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4072 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4073 return; /* no useful stat for tree diffs */
4074
4075 run_diffstat(p, o, diffstat);
4076 }
4077
4078 static void diff_flush_checkdiff(struct diff_filepair *p,
4079 struct diff_options *o)
4080 {
4081 if (diff_unmodified_pair(p))
4082 return;
4083
4084 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4085 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4086 return; /* nothing to check in tree diffs */
4087
4088 run_checkdiff(p, o);
4089 }
4090
4091 int diff_queue_is_empty(void)
4092 {
4093 struct diff_queue_struct *q = &diff_queued_diff;
4094 int i;
4095 for (i = 0; i < q->nr; i++)
4096 if (!diff_unmodified_pair(q->queue[i]))
4097 return 0;
4098 return 1;
4099 }
4100
4101 #if DIFF_DEBUG
4102 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4103 {
4104 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4105 x, one ? one : "",
4106 s->path,
4107 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4108 s->mode,
4109 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
4110 fprintf(stderr, "queue[%d] %s size %lu\n",
4111 x, one ? one : "",
4112 s->size);
4113 }
4114
4115 void diff_debug_filepair(const struct diff_filepair *p, int i)
4116 {
4117 diff_debug_filespec(p->one, i, "one");
4118 diff_debug_filespec(p->two, i, "two");
4119 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4120 p->score, p->status ? p->status : '?',
4121 p->one->rename_used, p->broken_pair);
4122 }
4123
4124 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4125 {
4126 int i;
4127 if (msg)
4128 fprintf(stderr, "%s\n", msg);
4129 fprintf(stderr, "q->nr = %d\n", q->nr);
4130 for (i = 0; i < q->nr; i++) {
4131 struct diff_filepair *p = q->queue[i];
4132 diff_debug_filepair(p, i);
4133 }
4134 }
4135 #endif
4136
4137 static void diff_resolve_rename_copy(void)
4138 {
4139 int i;
4140 struct diff_filepair *p;
4141 struct diff_queue_struct *q = &diff_queued_diff;
4142
4143 diff_debug_queue("resolve-rename-copy", q);
4144
4145 for (i = 0; i < q->nr; i++) {
4146 p = q->queue[i];
4147 p->status = 0; /* undecided */
4148 if (DIFF_PAIR_UNMERGED(p))
4149 p->status = DIFF_STATUS_UNMERGED;
4150 else if (!DIFF_FILE_VALID(p->one))
4151 p->status = DIFF_STATUS_ADDED;
4152 else if (!DIFF_FILE_VALID(p->two))
4153 p->status = DIFF_STATUS_DELETED;
4154 else if (DIFF_PAIR_TYPE_CHANGED(p))
4155 p->status = DIFF_STATUS_TYPE_CHANGED;
4156
4157 /* from this point on, we are dealing with a pair
4158 * whose both sides are valid and of the same type, i.e.
4159 * either in-place edit or rename/copy edit.
4160 */
4161 else if (DIFF_PAIR_RENAME(p)) {
4162 /*
4163 * A rename might have re-connected a broken
4164 * pair up, causing the pathnames to be the
4165 * same again. If so, that's not a rename at
4166 * all, just a modification..
4167 *
4168 * Otherwise, see if this source was used for
4169 * multiple renames, in which case we decrement
4170 * the count, and call it a copy.
4171 */
4172 if (!strcmp(p->one->path, p->two->path))
4173 p->status = DIFF_STATUS_MODIFIED;
4174 else if (--p->one->rename_used > 0)
4175 p->status = DIFF_STATUS_COPIED;
4176 else
4177 p->status = DIFF_STATUS_RENAMED;
4178 }
4179 else if (hashcmp(p->one->sha1, p->two->sha1) ||
4180 p->one->mode != p->two->mode ||
4181 p->one->dirty_submodule ||
4182 p->two->dirty_submodule ||
4183 is_null_sha1(p->one->sha1))
4184 p->status = DIFF_STATUS_MODIFIED;
4185 else {
4186 /* This is a "no-change" entry and should not
4187 * happen anymore, but prepare for broken callers.
4188 */
4189 error("feeding unmodified %s to diffcore",
4190 p->one->path);
4191 p->status = DIFF_STATUS_UNKNOWN;
4192 }
4193 }
4194 diff_debug_queue("resolve-rename-copy done", q);
4195 }
4196
4197 static int check_pair_status(struct diff_filepair *p)
4198 {
4199 switch (p->status) {
4200 case DIFF_STATUS_UNKNOWN:
4201 return 0;
4202 case 0:
4203 die("internal error in diff-resolve-rename-copy");
4204 default:
4205 return 1;
4206 }
4207 }
4208
4209 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4210 {
4211 int fmt = opt->output_format;
4212
4213 if (fmt & DIFF_FORMAT_CHECKDIFF)
4214 diff_flush_checkdiff(p, opt);
4215 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4216 diff_flush_raw(p, opt);
4217 else if (fmt & DIFF_FORMAT_NAME) {
4218 const char *name_a, *name_b;
4219 name_a = p->two->path;
4220 name_b = NULL;
4221 strip_prefix(opt->prefix_length, &name_a, &name_b);
4222 write_name_quoted(name_a, opt->file, opt->line_termination);
4223 }
4224 }
4225
4226 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4227 {
4228 if (fs->mode)
4229 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4230 else
4231 fprintf(file, " %s ", newdelete);
4232 write_name_quoted(fs->path, file, '\n');
4233 }
4234
4235
4236 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4237 const char *line_prefix)
4238 {
4239 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4240 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4241 p->two->mode, show_name ? ' ' : '\n');
4242 if (show_name) {
4243 write_name_quoted(p->two->path, file, '\n');
4244 }
4245 }
4246 }
4247
4248 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4249 const char *line_prefix)
4250 {
4251 char *names = pprint_rename(p->one->path, p->two->path);
4252
4253 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4254 free(names);
4255 show_mode_change(file, p, 0, line_prefix);
4256 }
4257
4258 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4259 {
4260 FILE *file = opt->file;
4261 const char *line_prefix = diff_line_prefix(opt);
4262
4263 switch(p->status) {
4264 case DIFF_STATUS_DELETED:
4265 fputs(line_prefix, file);
4266 show_file_mode_name(file, "delete", p->one);
4267 break;
4268 case DIFF_STATUS_ADDED:
4269 fputs(line_prefix, file);
4270 show_file_mode_name(file, "create", p->two);
4271 break;
4272 case DIFF_STATUS_COPIED:
4273 fputs(line_prefix, file);
4274 show_rename_copy(file, "copy", p, line_prefix);
4275 break;
4276 case DIFF_STATUS_RENAMED:
4277 fputs(line_prefix, file);
4278 show_rename_copy(file, "rename", p, line_prefix);
4279 break;
4280 default:
4281 if (p->score) {
4282 fprintf(file, "%s rewrite ", line_prefix);
4283 write_name_quoted(p->two->path, file, ' ');
4284 fprintf(file, "(%d%%)\n", similarity_index(p));
4285 }
4286 show_mode_change(file, p, !p->score, line_prefix);
4287 break;
4288 }
4289 }
4290
4291 struct patch_id_t {
4292 git_SHA_CTX *ctx;
4293 int patchlen;
4294 };
4295
4296 static int remove_space(char *line, int len)
4297 {
4298 int i;
4299 char *dst = line;
4300 unsigned char c;
4301
4302 for (i = 0; i < len; i++)
4303 if (!isspace((c = line[i])))
4304 *dst++ = c;
4305
4306 return dst - line;
4307 }
4308
4309 static void patch_id_consume(void *priv, char *line, unsigned long len)
4310 {
4311 struct patch_id_t *data = priv;
4312 int new_len;
4313
4314 /* Ignore line numbers when computing the SHA1 of the patch */
4315 if (starts_with(line, "@@ -"))
4316 return;
4317
4318 new_len = remove_space(line, len);
4319
4320 git_SHA1_Update(data->ctx, line, new_len);
4321 data->patchlen += new_len;
4322 }
4323
4324 /* returns 0 upon success, and writes result into sha1 */
4325 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4326 {
4327 struct diff_queue_struct *q = &diff_queued_diff;
4328 int i;
4329 git_SHA_CTX ctx;
4330 struct patch_id_t data;
4331 char buffer[PATH_MAX * 4 + 20];
4332
4333 git_SHA1_Init(&ctx);
4334 memset(&data, 0, sizeof(struct patch_id_t));
4335 data.ctx = &ctx;
4336
4337 for (i = 0; i < q->nr; i++) {
4338 xpparam_t xpp;
4339 xdemitconf_t xecfg;
4340 mmfile_t mf1, mf2;
4341 struct diff_filepair *p = q->queue[i];
4342 int len1, len2;
4343
4344 memset(&xpp, 0, sizeof(xpp));
4345 memset(&xecfg, 0, sizeof(xecfg));
4346 if (p->status == 0)
4347 return error("internal diff status error");
4348 if (p->status == DIFF_STATUS_UNKNOWN)
4349 continue;
4350 if (diff_unmodified_pair(p))
4351 continue;
4352 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4353 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4354 continue;
4355 if (DIFF_PAIR_UNMERGED(p))
4356 continue;
4357
4358 diff_fill_sha1_info(p->one);
4359 diff_fill_sha1_info(p->two);
4360 if (fill_mmfile(&mf1, p->one) < 0 ||
4361 fill_mmfile(&mf2, p->two) < 0)
4362 return error("unable to read files to diff");
4363
4364 len1 = remove_space(p->one->path, strlen(p->one->path));
4365 len2 = remove_space(p->two->path, strlen(p->two->path));
4366 if (p->one->mode == 0)
4367 len1 = snprintf(buffer, sizeof(buffer),
4368 "diff--gita/%.*sb/%.*s"
4369 "newfilemode%06o"
4370 "---/dev/null"
4371 "+++b/%.*s",
4372 len1, p->one->path,
4373 len2, p->two->path,
4374 p->two->mode,
4375 len2, p->two->path);
4376 else if (p->two->mode == 0)
4377 len1 = snprintf(buffer, sizeof(buffer),
4378 "diff--gita/%.*sb/%.*s"
4379 "deletedfilemode%06o"
4380 "---a/%.*s"
4381 "+++/dev/null",
4382 len1, p->one->path,
4383 len2, p->two->path,
4384 p->one->mode,
4385 len1, p->one->path);
4386 else
4387 len1 = snprintf(buffer, sizeof(buffer),
4388 "diff--gita/%.*sb/%.*s"
4389 "---a/%.*s"
4390 "+++b/%.*s",
4391 len1, p->one->path,
4392 len2, p->two->path,
4393 len1, p->one->path,
4394 len2, p->two->path);
4395 git_SHA1_Update(&ctx, buffer, len1);
4396
4397 if (diff_filespec_is_binary(p->one) ||
4398 diff_filespec_is_binary(p->two)) {
4399 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4400 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4401 continue;
4402 }
4403
4404 xpp.flags = 0;
4405 xecfg.ctxlen = 3;
4406 xecfg.flags = 0;
4407 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4408 &xpp, &xecfg);
4409 }
4410
4411 git_SHA1_Final(sha1, &ctx);
4412 return 0;
4413 }
4414
4415 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4416 {
4417 struct diff_queue_struct *q = &diff_queued_diff;
4418 int i;
4419 int result = diff_get_patch_id(options, sha1);
4420
4421 for (i = 0; i < q->nr; i++)
4422 diff_free_filepair(q->queue[i]);
4423
4424 free(q->queue);
4425 DIFF_QUEUE_CLEAR(q);
4426
4427 return result;
4428 }
4429
4430 static int is_summary_empty(const struct diff_queue_struct *q)
4431 {
4432 int i;
4433
4434 for (i = 0; i < q->nr; i++) {
4435 const struct diff_filepair *p = q->queue[i];
4436
4437 switch (p->status) {
4438 case DIFF_STATUS_DELETED:
4439 case DIFF_STATUS_ADDED:
4440 case DIFF_STATUS_COPIED:
4441 case DIFF_STATUS_RENAMED:
4442 return 0;
4443 default:
4444 if (p->score)
4445 return 0;
4446 if (p->one->mode && p->two->mode &&
4447 p->one->mode != p->two->mode)
4448 return 0;
4449 break;
4450 }
4451 }
4452 return 1;
4453 }
4454
4455 static const char rename_limit_warning[] =
4456 "inexact rename detection was skipped due to too many files.";
4457
4458 static const char degrade_cc_to_c_warning[] =
4459 "only found copies from modified paths due to too many files.";
4460
4461 static const char rename_limit_advice[] =
4462 "you may want to set your %s variable to at least "
4463 "%d and retry the command.";
4464
4465 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4466 {
4467 if (degraded_cc)
4468 warning(degrade_cc_to_c_warning);
4469 else if (needed)
4470 warning(rename_limit_warning);
4471 else
4472 return;
4473 if (0 < needed && needed < 32767)
4474 warning(rename_limit_advice, varname, needed);
4475 }
4476
4477 void diff_flush(struct diff_options *options)
4478 {
4479 struct diff_queue_struct *q = &diff_queued_diff;
4480 int i, output_format = options->output_format;
4481 int separator = 0;
4482 int dirstat_by_line = 0;
4483
4484 /*
4485 * Order: raw, stat, summary, patch
4486 * or: name/name-status/checkdiff (other bits clear)
4487 */
4488 if (!q->nr)
4489 goto free_queue;
4490
4491 if (output_format & (DIFF_FORMAT_RAW |
4492 DIFF_FORMAT_NAME |
4493 DIFF_FORMAT_NAME_STATUS |
4494 DIFF_FORMAT_CHECKDIFF)) {
4495 for (i = 0; i < q->nr; i++) {
4496 struct diff_filepair *p = q->queue[i];
4497 if (check_pair_status(p))
4498 flush_one_pair(p, options);
4499 }
4500 separator++;
4501 }
4502
4503 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4504 dirstat_by_line = 1;
4505
4506 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4507 dirstat_by_line) {
4508 struct diffstat_t diffstat;
4509
4510 memset(&diffstat, 0, sizeof(struct diffstat_t));
4511 for (i = 0; i < q->nr; i++) {
4512 struct diff_filepair *p = q->queue[i];
4513 if (check_pair_status(p))
4514 diff_flush_stat(p, options, &diffstat);
4515 }
4516 if (output_format & DIFF_FORMAT_NUMSTAT)
4517 show_numstat(&diffstat, options);
4518 if (output_format & DIFF_FORMAT_DIFFSTAT)
4519 show_stats(&diffstat, options);
4520 if (output_format & DIFF_FORMAT_SHORTSTAT)
4521 show_shortstats(&diffstat, options);
4522 if (output_format & DIFF_FORMAT_DIRSTAT)
4523 show_dirstat_by_line(&diffstat, options);
4524 free_diffstat_info(&diffstat);
4525 separator++;
4526 }
4527 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4528 show_dirstat(options);
4529
4530 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4531 for (i = 0; i < q->nr; i++) {
4532 diff_summary(options, q->queue[i]);
4533 }
4534 separator++;
4535 }
4536
4537 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4538 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4539 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4540 /*
4541 * run diff_flush_patch for the exit status. setting
4542 * options->file to /dev/null should be safe, because we
4543 * aren't supposed to produce any output anyway.
4544 */
4545 if (options->close_file)
4546 fclose(options->file);
4547 options->file = fopen("/dev/null", "w");
4548 if (!options->file)
4549 die_errno("Could not open /dev/null");
4550 options->close_file = 1;
4551 for (i = 0; i < q->nr; i++) {
4552 struct diff_filepair *p = q->queue[i];
4553 if (check_pair_status(p))
4554 diff_flush_patch(p, options);
4555 if (options->found_changes)
4556 break;
4557 }
4558 }
4559
4560 if (output_format & DIFF_FORMAT_PATCH) {
4561 if (separator) {
4562 fprintf(options->file, "%s%c",
4563 diff_line_prefix(options),
4564 options->line_termination);
4565 if (options->stat_sep) {
4566 /* attach patch instead of inline */
4567 fputs(options->stat_sep, options->file);
4568 }
4569 }
4570
4571 for (i = 0; i < q->nr; i++) {
4572 struct diff_filepair *p = q->queue[i];
4573 if (check_pair_status(p))
4574 diff_flush_patch(p, options);
4575 }
4576 }
4577
4578 if (output_format & DIFF_FORMAT_CALLBACK)
4579 options->format_callback(q, options, options->format_callback_data);
4580
4581 for (i = 0; i < q->nr; i++)
4582 diff_free_filepair(q->queue[i]);
4583 free_queue:
4584 free(q->queue);
4585 DIFF_QUEUE_CLEAR(q);
4586 if (options->close_file)
4587 fclose(options->file);
4588
4589 /*
4590 * Report the content-level differences with HAS_CHANGES;
4591 * diff_addremove/diff_change does not set the bit when
4592 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4593 */
4594 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4595 if (options->found_changes)
4596 DIFF_OPT_SET(options, HAS_CHANGES);
4597 else
4598 DIFF_OPT_CLR(options, HAS_CHANGES);
4599 }
4600 }
4601
4602 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
4603 {
4604 return (((p->status == DIFF_STATUS_MODIFIED) &&
4605 ((p->score &&
4606 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
4607 (!p->score &&
4608 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
4609 ((p->status != DIFF_STATUS_MODIFIED) &&
4610 filter_bit_tst(p->status, options)));
4611 }
4612
4613 static void diffcore_apply_filter(struct diff_options *options)
4614 {
4615 int i;
4616 struct diff_queue_struct *q = &diff_queued_diff;
4617 struct diff_queue_struct outq;
4618
4619 DIFF_QUEUE_CLEAR(&outq);
4620
4621 if (!options->filter)
4622 return;
4623
4624 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
4625 int found;
4626 for (i = found = 0; !found && i < q->nr; i++) {
4627 if (match_filter(options, q->queue[i]))
4628 found++;
4629 }
4630 if (found)
4631 return;
4632
4633 /* otherwise we will clear the whole queue
4634 * by copying the empty outq at the end of this
4635 * function, but first clear the current entries
4636 * in the queue.
4637 */
4638 for (i = 0; i < q->nr; i++)
4639 diff_free_filepair(q->queue[i]);
4640 }
4641 else {
4642 /* Only the matching ones */
4643 for (i = 0; i < q->nr; i++) {
4644 struct diff_filepair *p = q->queue[i];
4645 if (match_filter(options, p))
4646 diff_q(&outq, p);
4647 else
4648 diff_free_filepair(p);
4649 }
4650 }
4651 free(q->queue);
4652 *q = outq;
4653 }
4654
4655 /* Check whether two filespecs with the same mode and size are identical */
4656 static int diff_filespec_is_identical(struct diff_filespec *one,
4657 struct diff_filespec *two)
4658 {
4659 if (S_ISGITLINK(one->mode))
4660 return 0;
4661 if (diff_populate_filespec(one, 0))
4662 return 0;
4663 if (diff_populate_filespec(two, 0))
4664 return 0;
4665 return !memcmp(one->data, two->data, one->size);
4666 }
4667
4668 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
4669 {
4670 if (p->done_skip_stat_unmatch)
4671 return p->skip_stat_unmatch_result;
4672
4673 p->done_skip_stat_unmatch = 1;
4674 p->skip_stat_unmatch_result = 0;
4675 /*
4676 * 1. Entries that come from stat info dirtiness
4677 * always have both sides (iow, not create/delete),
4678 * one side of the object name is unknown, with
4679 * the same mode and size. Keep the ones that
4680 * do not match these criteria. They have real
4681 * differences.
4682 *
4683 * 2. At this point, the file is known to be modified,
4684 * with the same mode and size, and the object
4685 * name of one side is unknown. Need to inspect
4686 * the identical contents.
4687 */
4688 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4689 !DIFF_FILE_VALID(p->two) ||
4690 (p->one->sha1_valid && p->two->sha1_valid) ||
4691 (p->one->mode != p->two->mode) ||
4692 diff_populate_filespec(p->one, 1) ||
4693 diff_populate_filespec(p->two, 1) ||
4694 (p->one->size != p->two->size) ||
4695 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4696 p->skip_stat_unmatch_result = 1;
4697 return p->skip_stat_unmatch_result;
4698 }
4699
4700 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4701 {
4702 int i;
4703 struct diff_queue_struct *q = &diff_queued_diff;
4704 struct diff_queue_struct outq;
4705 DIFF_QUEUE_CLEAR(&outq);
4706
4707 for (i = 0; i < q->nr; i++) {
4708 struct diff_filepair *p = q->queue[i];
4709
4710 if (diff_filespec_check_stat_unmatch(p))
4711 diff_q(&outq, p);
4712 else {
4713 /*
4714 * The caller can subtract 1 from skip_stat_unmatch
4715 * to determine how many paths were dirty only
4716 * due to stat info mismatch.
4717 */
4718 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4719 diffopt->skip_stat_unmatch++;
4720 diff_free_filepair(p);
4721 }
4722 }
4723 free(q->queue);
4724 *q = outq;
4725 }
4726
4727 static int diffnamecmp(const void *a_, const void *b_)
4728 {
4729 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4730 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4731 const char *name_a, *name_b;
4732
4733 name_a = a->one ? a->one->path : a->two->path;
4734 name_b = b->one ? b->one->path : b->two->path;
4735 return strcmp(name_a, name_b);
4736 }
4737
4738 void diffcore_fix_diff_index(struct diff_options *options)
4739 {
4740 struct diff_queue_struct *q = &diff_queued_diff;
4741 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4742 }
4743
4744 void diffcore_std(struct diff_options *options)
4745 {
4746 /* NOTE please keep the following in sync with diff_tree_combined() */
4747 if (options->skip_stat_unmatch)
4748 diffcore_skip_stat_unmatch(options);
4749 if (!options->found_follow) {
4750 /* See try_to_follow_renames() in tree-diff.c */
4751 if (options->break_opt != -1)
4752 diffcore_break(options->break_opt);
4753 if (options->detect_rename)
4754 diffcore_rename(options);
4755 if (options->break_opt != -1)
4756 diffcore_merge_broken();
4757 }
4758 if (options->pickaxe)
4759 diffcore_pickaxe(options);
4760 if (options->orderfile)
4761 diffcore_order(options->orderfile);
4762 if (!options->found_follow)
4763 /* See try_to_follow_renames() in tree-diff.c */
4764 diff_resolve_rename_copy();
4765 diffcore_apply_filter(options);
4766
4767 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4768 DIFF_OPT_SET(options, HAS_CHANGES);
4769 else
4770 DIFF_OPT_CLR(options, HAS_CHANGES);
4771
4772 options->found_follow = 0;
4773 }
4774
4775 int diff_result_code(struct diff_options *opt, int status)
4776 {
4777 int result = 0;
4778
4779 diff_warn_rename_limit("diff.renameLimit",
4780 opt->needed_rename_limit,
4781 opt->degraded_cc_to_c);
4782 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4783 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4784 return status;
4785 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4786 DIFF_OPT_TST(opt, HAS_CHANGES))
4787 result |= 01;
4788 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4789 DIFF_OPT_TST(opt, CHECK_FAILED))
4790 result |= 02;
4791 return result;
4792 }
4793
4794 int diff_can_quit_early(struct diff_options *opt)
4795 {
4796 return (DIFF_OPT_TST(opt, QUICK) &&
4797 !opt->filter &&
4798 DIFF_OPT_TST(opt, HAS_CHANGES));
4799 }
4800
4801 /*
4802 * Shall changes to this submodule be ignored?
4803 *
4804 * Submodule changes can be configured to be ignored separately for each path,
4805 * but that configuration can be overridden from the command line.
4806 */
4807 static int is_submodule_ignored(const char *path, struct diff_options *options)
4808 {
4809 int ignored = 0;
4810 unsigned orig_flags = options->flags;
4811 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4812 set_diffopt_flags_from_submodule_config(options, path);
4813 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4814 ignored = 1;
4815 options->flags = orig_flags;
4816 return ignored;
4817 }
4818
4819 void diff_addremove(struct diff_options *options,
4820 int addremove, unsigned mode,
4821 const unsigned char *sha1,
4822 int sha1_valid,
4823 const char *concatpath, unsigned dirty_submodule)
4824 {
4825 struct diff_filespec *one, *two;
4826
4827 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4828 return;
4829
4830 /* This may look odd, but it is a preparation for
4831 * feeding "there are unchanged files which should
4832 * not produce diffs, but when you are doing copy
4833 * detection you would need them, so here they are"
4834 * entries to the diff-core. They will be prefixed
4835 * with something like '=' or '*' (I haven't decided
4836 * which but should not make any difference).
4837 * Feeding the same new and old to diff_change()
4838 * also has the same effect.
4839 * Before the final output happens, they are pruned after
4840 * merged into rename/copy pairs as appropriate.
4841 */
4842 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4843 addremove = (addremove == '+' ? '-' :
4844 addremove == '-' ? '+' : addremove);
4845
4846 if (options->prefix &&
4847 strncmp(concatpath, options->prefix, options->prefix_length))
4848 return;
4849
4850 one = alloc_filespec(concatpath);
4851 two = alloc_filespec(concatpath);
4852
4853 if (addremove != '+')
4854 fill_filespec(one, sha1, sha1_valid, mode);
4855 if (addremove != '-') {
4856 fill_filespec(two, sha1, sha1_valid, mode);
4857 two->dirty_submodule = dirty_submodule;
4858 }
4859
4860 diff_queue(&diff_queued_diff, one, two);
4861 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4862 DIFF_OPT_SET(options, HAS_CHANGES);
4863 }
4864
4865 void diff_change(struct diff_options *options,
4866 unsigned old_mode, unsigned new_mode,
4867 const unsigned char *old_sha1,
4868 const unsigned char *new_sha1,
4869 int old_sha1_valid, int new_sha1_valid,
4870 const char *concatpath,
4871 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4872 {
4873 struct diff_filespec *one, *two;
4874 struct diff_filepair *p;
4875
4876 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4877 is_submodule_ignored(concatpath, options))
4878 return;
4879
4880 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4881 unsigned tmp;
4882 const unsigned char *tmp_c;
4883 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4884 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4885 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
4886 new_sha1_valid = tmp;
4887 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4888 new_dirty_submodule = tmp;
4889 }
4890
4891 if (options->prefix &&
4892 strncmp(concatpath, options->prefix, options->prefix_length))
4893 return;
4894
4895 one = alloc_filespec(concatpath);
4896 two = alloc_filespec(concatpath);
4897 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
4898 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
4899 one->dirty_submodule = old_dirty_submodule;
4900 two->dirty_submodule = new_dirty_submodule;
4901 p = diff_queue(&diff_queued_diff, one, two);
4902
4903 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4904 return;
4905
4906 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
4907 !diff_filespec_check_stat_unmatch(p))
4908 return;
4909
4910 DIFF_OPT_SET(options, HAS_CHANGES);
4911 }
4912
4913 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4914 {
4915 struct diff_filepair *pair;
4916 struct diff_filespec *one, *two;
4917
4918 if (options->prefix &&
4919 strncmp(path, options->prefix, options->prefix_length))
4920 return NULL;
4921
4922 one = alloc_filespec(path);
4923 two = alloc_filespec(path);
4924 pair = diff_queue(&diff_queued_diff, one, two);
4925 pair->is_unmerged = 1;
4926 return pair;
4927 }
4928
4929 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4930 size_t *outsize)
4931 {
4932 struct diff_tempfile *temp;
4933 const char *argv[3];
4934 const char **arg = argv;
4935 struct child_process child;
4936 struct strbuf buf = STRBUF_INIT;
4937 int err = 0;
4938
4939 temp = prepare_temp_file(spec->path, spec);
4940 *arg++ = pgm;
4941 *arg++ = temp->name;
4942 *arg = NULL;
4943
4944 memset(&child, 0, sizeof(child));
4945 child.use_shell = 1;
4946 child.argv = argv;
4947 child.out = -1;
4948 if (start_command(&child)) {
4949 remove_tempfile();
4950 return NULL;
4951 }
4952
4953 if (strbuf_read(&buf, child.out, 0) < 0)
4954 err = error("error reading from textconv command '%s'", pgm);
4955 close(child.out);
4956
4957 if (finish_command(&child) || err) {
4958 strbuf_release(&buf);
4959 remove_tempfile();
4960 return NULL;
4961 }
4962 remove_tempfile();
4963
4964 return strbuf_detach(&buf, outsize);
4965 }
4966
4967 size_t fill_textconv(struct userdiff_driver *driver,
4968 struct diff_filespec *df,
4969 char **outbuf)
4970 {
4971 size_t size;
4972
4973 if (!driver || !driver->textconv) {
4974 if (!DIFF_FILE_VALID(df)) {
4975 *outbuf = "";
4976 return 0;
4977 }
4978 if (diff_populate_filespec(df, 0))
4979 die("unable to read files to diff");
4980 *outbuf = df->data;
4981 return df->size;
4982 }
4983
4984 if (driver->textconv_cache && df->sha1_valid) {
4985 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4986 &size);
4987 if (*outbuf)
4988 return size;
4989 }
4990
4991 *outbuf = run_textconv(driver->textconv, df, &size);
4992 if (!*outbuf)
4993 die("unable to read files to diff");
4994
4995 if (driver->textconv_cache && df->sha1_valid) {
4996 /* ignore errors, as we might be in a readonly repository */
4997 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4998 size);
4999 /*
5000 * we could save up changes and flush them all at the end,
5001 * but we would need an extra call after all diffing is done.
5002 * Since generating a cache entry is the slow path anyway,
5003 * this extra overhead probably isn't a big deal.
5004 */
5005 notes_cache_write(driver->textconv_cache);
5006 }
5007
5008 return size;
5009 }
5010
5011 void setup_diff_pager(struct diff_options *opt)
5012 {
5013 /*
5014 * If the user asked for our exit code, then either they want --quiet
5015 * or --exit-code. We should definitely not bother with a pager in the
5016 * former case, as we will generate no output. Since we still properly
5017 * report our exit code even when a pager is run, we _could_ run a
5018 * pager with --exit-code. But since we have not done so historically,
5019 * and because it is easy to find people oneline advising "git diff
5020 * --exit-code" in hooks and other scripts, we do not do so.
5021 */
5022 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5023 check_pager_config("diff") != 0)
5024 setup_pager();
5025 }
Jonathan Nieder’s miscellaneous git branches