?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'jk/path-name-safety-2.4' into jk/path-name-safety-2.5
[git/debian.git] / diff.c
blob71f0274ce0df2e6af8cd144c78c64dbb694c1ab7
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18 #include "string-list.h"
19 #include "argv-array.h"
20
21 #ifdef NO_FAST_WORKING_DIRECTORY
22 #define FAST_WORKING_DIRECTORY 0
23 #else
24 #define FAST_WORKING_DIRECTORY 1
25 #endif
26
27 static int diff_detect_rename_default;
28 static int diff_rename_limit_default = 400;
29 static int diff_suppress_blank_empty;
30 static int diff_use_color_default = -1;
31 static int diff_context_default = 3;
32 static const char *diff_word_regex_cfg;
33 static const char *external_diff_cmd_cfg;
34 static const char *diff_order_file_cfg;
35 int diff_auto_refresh_index = 1;
36 static int diff_mnemonic_prefix;
37 static int diff_no_prefix;
38 static int diff_stat_graph_width;
39 static int diff_dirstat_permille_default = 30;
40 static struct diff_options default_diff_options;
41 static long diff_algorithm;
42
43 static char diff_colors[][COLOR_MAXLEN] = {
44 GIT_COLOR_RESET,
45 GIT_COLOR_NORMAL, /* CONTEXT */
46 GIT_COLOR_BOLD, /* METAINFO */
47 GIT_COLOR_CYAN, /* FRAGINFO */
48 GIT_COLOR_RED, /* OLD */
49 GIT_COLOR_GREEN, /* NEW */
50 GIT_COLOR_YELLOW, /* COMMIT */
51 GIT_COLOR_BG_RED, /* WHITESPACE */
52 GIT_COLOR_NORMAL, /* FUNCINFO */
53 };
54
55 static int parse_diff_color_slot(const char *var)
56 {
57 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
58 return DIFF_CONTEXT;
59 if (!strcasecmp(var, "meta"))
60 return DIFF_METAINFO;
61 if (!strcasecmp(var, "frag"))
62 return DIFF_FRAGINFO;
63 if (!strcasecmp(var, "old"))
64 return DIFF_FILE_OLD;
65 if (!strcasecmp(var, "new"))
66 return DIFF_FILE_NEW;
67 if (!strcasecmp(var, "commit"))
68 return DIFF_COMMIT;
69 if (!strcasecmp(var, "whitespace"))
70 return DIFF_WHITESPACE;
71 if (!strcasecmp(var, "func"))
72 return DIFF_FUNCINFO;
73 return -1;
74 }
75
76 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
77 struct strbuf *errmsg)
78 {
79 char *params_copy = xstrdup(params_string);
80 struct string_list params = STRING_LIST_INIT_NODUP;
81 int ret = 0;
82 int i;
83
84 if (*params_copy)
85 string_list_split_in_place(&params, params_copy, ',', -1);
86 for (i = 0; i < params.nr; i++) {
87 const char *p = params.items[i].string;
88 if (!strcmp(p, "changes")) {
89 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
90 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
91 } else if (!strcmp(p, "lines")) {
92 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
93 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
94 } else if (!strcmp(p, "files")) {
95 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
96 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
97 } else if (!strcmp(p, "noncumulative")) {
98 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
99 } else if (!strcmp(p, "cumulative")) {
100 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
101 } else if (isdigit(*p)) {
102 char *end;
103 int permille = strtoul(p, &end, 10) * 10;
104 if (*end == '.' && isdigit(*++end)) {
105 /* only use first digit */
106 permille += *end - '0';
107 /* .. and ignore any further digits */
108 while (isdigit(*++end))
109 ; /* nothing */
110 }
111 if (!*end)
112 options->dirstat_permille = permille;
113 else {
114 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
115 p);
116 ret++;
117 }
118 } else {
119 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
120 ret++;
121 }
122
123 }
124 string_list_clear(&params, 0);
125 free(params_copy);
126 return ret;
127 }
128
129 static int parse_submodule_params(struct diff_options *options, const char *value)
130 {
131 if (!strcmp(value, "log"))
132 DIFF_OPT_SET(options, SUBMODULE_LOG);
133 else if (!strcmp(value, "short"))
134 DIFF_OPT_CLR(options, SUBMODULE_LOG);
135 else
136 return -1;
137 return 0;
138 }
139
140 static int git_config_rename(const char *var, const char *value)
141 {
142 if (!value)
143 return DIFF_DETECT_RENAME;
144 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
145 return DIFF_DETECT_COPY;
146 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
147 }
148
149 long parse_algorithm_value(const char *value)
150 {
151 if (!value)
152 return -1;
153 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
154 return 0;
155 else if (!strcasecmp(value, "minimal"))
156 return XDF_NEED_MINIMAL;
157 else if (!strcasecmp(value, "patience"))
158 return XDF_PATIENCE_DIFF;
159 else if (!strcasecmp(value, "histogram"))
160 return XDF_HISTOGRAM_DIFF;
161 return -1;
162 }
163
164 /*
165 * These are to give UI layer defaults.
166 * The core-level commands such as git-diff-files should
167 * never be affected by the setting of diff.renames
168 * the user happens to have in the configuration file.
169 */
170 int git_diff_ui_config(const char *var, const char *value, void *cb)
171 {
172 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
173 diff_use_color_default = git_config_colorbool(var, value);
174 return 0;
175 }
176 if (!strcmp(var, "diff.context")) {
177 diff_context_default = git_config_int(var, value);
178 if (diff_context_default < 0)
179 return -1;
180 return 0;
181 }
182 if (!strcmp(var, "diff.renames")) {
183 diff_detect_rename_default = git_config_rename(var, value);
184 return 0;
185 }
186 if (!strcmp(var, "diff.autorefreshindex")) {
187 diff_auto_refresh_index = git_config_bool(var, value);
188 return 0;
189 }
190 if (!strcmp(var, "diff.mnemonicprefix")) {
191 diff_mnemonic_prefix = git_config_bool(var, value);
192 return 0;
193 }
194 if (!strcmp(var, "diff.noprefix")) {
195 diff_no_prefix = git_config_bool(var, value);
196 return 0;
197 }
198 if (!strcmp(var, "diff.statgraphwidth")) {
199 diff_stat_graph_width = git_config_int(var, value);
200 return 0;
201 }
202 if (!strcmp(var, "diff.external"))
203 return git_config_string(&external_diff_cmd_cfg, var, value);
204 if (!strcmp(var, "diff.wordregex"))
205 return git_config_string(&diff_word_regex_cfg, var, value);
206 if (!strcmp(var, "diff.orderfile"))
207 return git_config_pathname(&diff_order_file_cfg, var, value);
208
209 if (!strcmp(var, "diff.ignoresubmodules"))
210 handle_ignore_submodules_arg(&default_diff_options, value);
211
212 if (!strcmp(var, "diff.submodule")) {
213 if (parse_submodule_params(&default_diff_options, value))
214 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
215 value);
216 return 0;
217 }
218
219 if (!strcmp(var, "diff.algorithm")) {
220 diff_algorithm = parse_algorithm_value(value);
221 if (diff_algorithm < 0)
222 return -1;
223 return 0;
224 }
225
226 if (git_color_config(var, value, cb) < 0)
227 return -1;
228
229 return git_diff_basic_config(var, value, cb);
230 }
231
232 int git_diff_basic_config(const char *var, const char *value, void *cb)
233 {
234 const char *name;
235
236 if (!strcmp(var, "diff.renamelimit")) {
237 diff_rename_limit_default = git_config_int(var, value);
238 return 0;
239 }
240
241 if (userdiff_config(var, value) < 0)
242 return -1;
243
244 if (skip_prefix(var, "diff.color.", &name) ||
245 skip_prefix(var, "color.diff.", &name)) {
246 int slot = parse_diff_color_slot(name);
247 if (slot < 0)
248 return 0;
249 if (!value)
250 return config_error_nonbool(var);
251 return color_parse(value, diff_colors[slot]);
252 }
253
254 /* like GNU diff's --suppress-blank-empty option */
255 if (!strcmp(var, "diff.suppressblankempty") ||
256 /* for backwards compatibility */
257 !strcmp(var, "diff.suppress-blank-empty")) {
258 diff_suppress_blank_empty = git_config_bool(var, value);
259 return 0;
260 }
261
262 if (!strcmp(var, "diff.dirstat")) {
263 struct strbuf errmsg = STRBUF_INIT;
264 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
265 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
266 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
267 errmsg.buf);
268 strbuf_release(&errmsg);
269 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
270 return 0;
271 }
272
273 if (starts_with(var, "submodule."))
274 return parse_submodule_config_option(var, value);
275
276 return git_default_config(var, value, cb);
277 }
278
279 static char *quote_two(const char *one, const char *two)
280 {
281 int need_one = quote_c_style(one, NULL, NULL, 1);
282 int need_two = quote_c_style(two, NULL, NULL, 1);
283 struct strbuf res = STRBUF_INIT;
284
285 if (need_one + need_two) {
286 strbuf_addch(&res, '"');
287 quote_c_style(one, &res, NULL, 1);
288 quote_c_style(two, &res, NULL, 1);
289 strbuf_addch(&res, '"');
290 } else {
291 strbuf_addstr(&res, one);
292 strbuf_addstr(&res, two);
293 }
294 return strbuf_detach(&res, NULL);
295 }
296
297 static const char *external_diff(void)
298 {
299 static const char *external_diff_cmd = NULL;
300 static int done_preparing = 0;
301
302 if (done_preparing)
303 return external_diff_cmd;
304 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
305 if (!external_diff_cmd)
306 external_diff_cmd = external_diff_cmd_cfg;
307 done_preparing = 1;
308 return external_diff_cmd;
309 }
310
311 static struct diff_tempfile {
312 const char *name; /* filename external diff should read from */
313 char hex[41];
314 char mode[10];
315 char tmp_path[PATH_MAX];
316 } diff_temp[2];
317
318 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
319
320 struct emit_callback {
321 int color_diff;
322 unsigned ws_rule;
323 int blank_at_eof_in_preimage;
324 int blank_at_eof_in_postimage;
325 int lno_in_preimage;
326 int lno_in_postimage;
327 sane_truncate_fn truncate;
328 const char **label_path;
329 struct diff_words_data *diff_words;
330 struct diff_options *opt;
331 int *found_changesp;
332 struct strbuf *header;
333 };
334
335 static int count_lines(const char *data, int size)
336 {
337 int count, ch, completely_empty = 1, nl_just_seen = 0;
338 count = 0;
339 while (0 < size--) {
340 ch = *data++;
341 if (ch == '\n') {
342 count++;
343 nl_just_seen = 1;
344 completely_empty = 0;
345 }
346 else {
347 nl_just_seen = 0;
348 completely_empty = 0;
349 }
350 }
351 if (completely_empty)
352 return 0;
353 if (!nl_just_seen)
354 count++; /* no trailing newline */
355 return count;
356 }
357
358 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
359 {
360 if (!DIFF_FILE_VALID(one)) {
361 mf->ptr = (char *)""; /* does not matter */
362 mf->size = 0;
363 return 0;
364 }
365 else if (diff_populate_filespec(one, 0))
366 return -1;
367
368 mf->ptr = one->data;
369 mf->size = one->size;
370 return 0;
371 }
372
373 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
374 static unsigned long diff_filespec_size(struct diff_filespec *one)
375 {
376 if (!DIFF_FILE_VALID(one))
377 return 0;
378 diff_populate_filespec(one, CHECK_SIZE_ONLY);
379 return one->size;
380 }
381
382 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
383 {
384 char *ptr = mf->ptr;
385 long size = mf->size;
386 int cnt = 0;
387
388 if (!size)
389 return cnt;
390 ptr += size - 1; /* pointing at the very end */
391 if (*ptr != '\n')
392 ; /* incomplete line */
393 else
394 ptr--; /* skip the last LF */
395 while (mf->ptr < ptr) {
396 char *prev_eol;
397 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
398 if (*prev_eol == '\n')
399 break;
400 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
401 break;
402 cnt++;
403 ptr = prev_eol - 1;
404 }
405 return cnt;
406 }
407
408 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
409 struct emit_callback *ecbdata)
410 {
411 int l1, l2, at;
412 unsigned ws_rule = ecbdata->ws_rule;
413 l1 = count_trailing_blank(mf1, ws_rule);
414 l2 = count_trailing_blank(mf2, ws_rule);
415 if (l2 <= l1) {
416 ecbdata->blank_at_eof_in_preimage = 0;
417 ecbdata->blank_at_eof_in_postimage = 0;
418 return;
419 }
420 at = count_lines(mf1->ptr, mf1->size);
421 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
422
423 at = count_lines(mf2->ptr, mf2->size);
424 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
425 }
426
427 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
428 int first, const char *line, int len)
429 {
430 int has_trailing_newline, has_trailing_carriage_return;
431 int nofirst;
432 FILE *file = o->file;
433
434 fputs(diff_line_prefix(o), file);
435
436 if (len == 0) {
437 has_trailing_newline = (first == '\n');
438 has_trailing_carriage_return = (!has_trailing_newline &&
439 (first == '\r'));
440 nofirst = has_trailing_newline || has_trailing_carriage_return;
441 } else {
442 has_trailing_newline = (len > 0 && line[len-1] == '\n');
443 if (has_trailing_newline)
444 len--;
445 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
446 if (has_trailing_carriage_return)
447 len--;
448 nofirst = 0;
449 }
450
451 if (len || !nofirst) {
452 fputs(set, file);
453 if (!nofirst)
454 fputc(first, file);
455 fwrite(line, len, 1, file);
456 fputs(reset, file);
457 }
458 if (has_trailing_carriage_return)
459 fputc('\r', file);
460 if (has_trailing_newline)
461 fputc('\n', file);
462 }
463
464 static void emit_line(struct diff_options *o, const char *set, const char *reset,
465 const char *line, int len)
466 {
467 emit_line_0(o, set, reset, line[0], line+1, len-1);
468 }
469
470 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
471 {
472 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
473 ecbdata->blank_at_eof_in_preimage &&
474 ecbdata->blank_at_eof_in_postimage &&
475 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
476 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
477 return 0;
478 return ws_blank_line(line, len, ecbdata->ws_rule);
479 }
480
481 static void emit_line_checked(const char *reset,
482 struct emit_callback *ecbdata,
483 const char *line, int len,
484 enum color_diff color,
485 unsigned ws_error_highlight,
486 char sign)
487 {
488 const char *set = diff_get_color(ecbdata->color_diff, color);
489 const char *ws = NULL;
490
491 if (ecbdata->opt->ws_error_highlight & ws_error_highlight) {
492 ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
493 if (!*ws)
494 ws = NULL;
495 }
496
497 if (!ws)
498 emit_line_0(ecbdata->opt, set, reset, sign, line, len);
499 else if (sign == '+' && new_blank_line_at_eof(ecbdata, line, len))
500 /* Blank line at EOF - paint '+' as well */
501 emit_line_0(ecbdata->opt, ws, reset, sign, line, len);
502 else {
503 /* Emit just the prefix, then the rest. */
504 emit_line_0(ecbdata->opt, set, reset, sign, "", 0);
505 ws_check_emit(line, len, ecbdata->ws_rule,
506 ecbdata->opt->file, set, reset, ws);
507 }
508 }
509
510 static void emit_add_line(const char *reset,
511 struct emit_callback *ecbdata,
512 const char *line, int len)
513 {
514 emit_line_checked(reset, ecbdata, line, len,
515 DIFF_FILE_NEW, WSEH_NEW, '+');
516 }
517
518 static void emit_del_line(const char *reset,
519 struct emit_callback *ecbdata,
520 const char *line, int len)
521 {
522 emit_line_checked(reset, ecbdata, line, len,
523 DIFF_FILE_OLD, WSEH_OLD, '-');
524 }
525
526 static void emit_context_line(const char *reset,
527 struct emit_callback *ecbdata,
528 const char *line, int len)
529 {
530 emit_line_checked(reset, ecbdata, line, len,
531 DIFF_CONTEXT, WSEH_CONTEXT, ' ');
532 }
533
534 static void emit_hunk_header(struct emit_callback *ecbdata,
535 const char *line, int len)
536 {
537 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
538 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
539 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
540 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
541 static const char atat[2] = { '@', '@' };
542 const char *cp, *ep;
543 struct strbuf msgbuf = STRBUF_INIT;
544 int org_len = len;
545 int i = 1;
546
547 /*
548 * As a hunk header must begin with "@@ -<old>, +<new> @@",
549 * it always is at least 10 bytes long.
550 */
551 if (len < 10 ||
552 memcmp(line, atat, 2) ||
553 !(ep = memmem(line + 2, len - 2, atat, 2))) {
554 emit_line(ecbdata->opt, context, reset, line, len);
555 return;
556 }
557 ep += 2; /* skip over @@ */
558
559 /* The hunk header in fraginfo color */
560 strbuf_addstr(&msgbuf, frag);
561 strbuf_add(&msgbuf, line, ep - line);
562 strbuf_addstr(&msgbuf, reset);
563
564 /*
565 * trailing "\r\n"
566 */
567 for ( ; i < 3; i++)
568 if (line[len - i] == '\r' || line[len - i] == '\n')
569 len--;
570
571 /* blank before the func header */
572 for (cp = ep; ep - line < len; ep++)
573 if (*ep != ' ' && *ep != '\t')
574 break;
575 if (ep != cp) {
576 strbuf_addstr(&msgbuf, context);
577 strbuf_add(&msgbuf, cp, ep - cp);
578 strbuf_addstr(&msgbuf, reset);
579 }
580
581 if (ep < line + len) {
582 strbuf_addstr(&msgbuf, func);
583 strbuf_add(&msgbuf, ep, line + len - ep);
584 strbuf_addstr(&msgbuf, reset);
585 }
586
587 strbuf_add(&msgbuf, line + len, org_len - len);
588 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
589 strbuf_release(&msgbuf);
590 }
591
592 static struct diff_tempfile *claim_diff_tempfile(void) {
593 int i;
594 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
595 if (!diff_temp[i].name)
596 return diff_temp + i;
597 die("BUG: diff is failing to clean up its tempfiles");
598 }
599
600 static int remove_tempfile_installed;
601
602 static void remove_tempfile(void)
603 {
604 int i;
605 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
606 if (diff_temp[i].name == diff_temp[i].tmp_path)
607 unlink_or_warn(diff_temp[i].name);
608 diff_temp[i].name = NULL;
609 }
610 }
611
612 static void remove_tempfile_on_signal(int signo)
613 {
614 remove_tempfile();
615 sigchain_pop(signo);
616 raise(signo);
617 }
618
619 static void print_line_count(FILE *file, int count)
620 {
621 switch (count) {
622 case 0:
623 fprintf(file, "0,0");
624 break;
625 case 1:
626 fprintf(file, "1");
627 break;
628 default:
629 fprintf(file, "1,%d", count);
630 break;
631 }
632 }
633
634 static void emit_rewrite_lines(struct emit_callback *ecb,
635 int prefix, const char *data, int size)
636 {
637 const char *endp = NULL;
638 static const char *nneof = " No newline at end of file\n";
639 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
640
641 while (0 < size) {
642 int len;
643
644 endp = memchr(data, '\n', size);
645 len = endp ? (endp - data + 1) : size;
646 if (prefix != '+') {
647 ecb->lno_in_preimage++;
648 emit_del_line(reset, ecb, data, len);
649 } else {
650 ecb->lno_in_postimage++;
651 emit_add_line(reset, ecb, data, len);
652 }
653 size -= len;
654 data += len;
655 }
656 if (!endp) {
657 const char *context = diff_get_color(ecb->color_diff,
658 DIFF_CONTEXT);
659 putc('\n', ecb->opt->file);
660 emit_line_0(ecb->opt, context, reset, '\\',
661 nneof, strlen(nneof));
662 }
663 }
664
665 static void emit_rewrite_diff(const char *name_a,
666 const char *name_b,
667 struct diff_filespec *one,
668 struct diff_filespec *two,
669 struct userdiff_driver *textconv_one,
670 struct userdiff_driver *textconv_two,
671 struct diff_options *o)
672 {
673 int lc_a, lc_b;
674 const char *name_a_tab, *name_b_tab;
675 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
676 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
677 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
678 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
679 const char *a_prefix, *b_prefix;
680 char *data_one, *data_two;
681 size_t size_one, size_two;
682 struct emit_callback ecbdata;
683 const char *line_prefix = diff_line_prefix(o);
684
685 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
686 a_prefix = o->b_prefix;
687 b_prefix = o->a_prefix;
688 } else {
689 a_prefix = o->a_prefix;
690 b_prefix = o->b_prefix;
691 }
692
693 name_a += (*name_a == '/');
694 name_b += (*name_b == '/');
695 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
696 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
697
698 strbuf_reset(&a_name);
699 strbuf_reset(&b_name);
700 quote_two_c_style(&a_name, a_prefix, name_a, 0);
701 quote_two_c_style(&b_name, b_prefix, name_b, 0);
702
703 size_one = fill_textconv(textconv_one, one, &data_one);
704 size_two = fill_textconv(textconv_two, two, &data_two);
705
706 memset(&ecbdata, 0, sizeof(ecbdata));
707 ecbdata.color_diff = want_color(o->use_color);
708 ecbdata.found_changesp = &o->found_changes;
709 ecbdata.ws_rule = whitespace_rule(name_b);
710 ecbdata.opt = o;
711 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
712 mmfile_t mf1, mf2;
713 mf1.ptr = (char *)data_one;
714 mf2.ptr = (char *)data_two;
715 mf1.size = size_one;
716 mf2.size = size_two;
717 check_blank_at_eof(&mf1, &mf2, &ecbdata);
718 }
719 ecbdata.lno_in_preimage = 1;
720 ecbdata.lno_in_postimage = 1;
721
722 lc_a = count_lines(data_one, size_one);
723 lc_b = count_lines(data_two, size_two);
724 fprintf(o->file,
725 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
726 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
727 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
728 line_prefix, fraginfo);
729 if (!o->irreversible_delete)
730 print_line_count(o->file, lc_a);
731 else
732 fprintf(o->file, "?,?");
733 fprintf(o->file, " +");
734 print_line_count(o->file, lc_b);
735 fprintf(o->file, " @@%s\n", reset);
736 if (lc_a && !o->irreversible_delete)
737 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
738 if (lc_b)
739 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
740 if (textconv_one)
741 free((char *)data_one);
742 if (textconv_two)
743 free((char *)data_two);
744 }
745
746 struct diff_words_buffer {
747 mmfile_t text;
748 long alloc;
749 struct diff_words_orig {
750 const char *begin, *end;
751 } *orig;
752 int orig_nr, orig_alloc;
753 };
754
755 static void diff_words_append(char *line, unsigned long len,
756 struct diff_words_buffer *buffer)
757 {
758 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
759 line++;
760 len--;
761 memcpy(buffer->text.ptr + buffer->text.size, line, len);
762 buffer->text.size += len;
763 buffer->text.ptr[buffer->text.size] = '\0';
764 }
765
766 struct diff_words_style_elem {
767 const char *prefix;
768 const char *suffix;
769 const char *color; /* NULL; filled in by the setup code if
770 * color is enabled */
771 };
772
773 struct diff_words_style {
774 enum diff_words_type type;
775 struct diff_words_style_elem new, old, ctx;
776 const char *newline;
777 };
778
779 static struct diff_words_style diff_words_styles[] = {
780 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
781 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
782 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
783 };
784
785 struct diff_words_data {
786 struct diff_words_buffer minus, plus;
787 const char *current_plus;
788 int last_minus;
789 struct diff_options *opt;
790 regex_t *word_regex;
791 enum diff_words_type type;
792 struct diff_words_style *style;
793 };
794
795 static int fn_out_diff_words_write_helper(FILE *fp,
796 struct diff_words_style_elem *st_el,
797 const char *newline,
798 size_t count, const char *buf,
799 const char *line_prefix)
800 {
801 int print = 0;
802
803 while (count) {
804 char *p = memchr(buf, '\n', count);
805 if (print)
806 fputs(line_prefix, fp);
807 if (p != buf) {
808 if (st_el->color && fputs(st_el->color, fp) < 0)
809 return -1;
810 if (fputs(st_el->prefix, fp) < 0 ||
811 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
812 fputs(st_el->suffix, fp) < 0)
813 return -1;
814 if (st_el->color && *st_el->color
815 && fputs(GIT_COLOR_RESET, fp) < 0)
816 return -1;
817 }
818 if (!p)
819 return 0;
820 if (fputs(newline, fp) < 0)
821 return -1;
822 count -= p + 1 - buf;
823 buf = p + 1;
824 print = 1;
825 }
826 return 0;
827 }
828
829 /*
830 * '--color-words' algorithm can be described as:
831 *
832 * 1. collect a the minus/plus lines of a diff hunk, divided into
833 * minus-lines and plus-lines;
834 *
835 * 2. break both minus-lines and plus-lines into words and
836 * place them into two mmfile_t with one word for each line;
837 *
838 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
839 *
840 * And for the common parts of the both file, we output the plus side text.
841 * diff_words->current_plus is used to trace the current position of the plus file
842 * which printed. diff_words->last_minus is used to trace the last minus word
843 * printed.
844 *
845 * For '--graph' to work with '--color-words', we need to output the graph prefix
846 * on each line of color words output. Generally, there are two conditions on
847 * which we should output the prefix.
848 *
849 * 1. diff_words->last_minus == 0 &&
850 * diff_words->current_plus == diff_words->plus.text.ptr
851 *
852 * that is: the plus text must start as a new line, and if there is no minus
853 * word printed, a graph prefix must be printed.
854 *
855 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
856 * *(diff_words->current_plus - 1) == '\n'
857 *
858 * that is: a graph prefix must be printed following a '\n'
859 */
860 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
861 {
862 if ((diff_words->last_minus == 0 &&
863 diff_words->current_plus == diff_words->plus.text.ptr) ||
864 (diff_words->current_plus > diff_words->plus.text.ptr &&
865 *(diff_words->current_plus - 1) == '\n')) {
866 return 1;
867 } else {
868 return 0;
869 }
870 }
871
872 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
873 {
874 struct diff_words_data *diff_words = priv;
875 struct diff_words_style *style = diff_words->style;
876 int minus_first, minus_len, plus_first, plus_len;
877 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
878 struct diff_options *opt = diff_words->opt;
879 const char *line_prefix;
880
881 if (line[0] != '@' || parse_hunk_header(line, len,
882 &minus_first, &minus_len, &plus_first, &plus_len))
883 return;
884
885 assert(opt);
886 line_prefix = diff_line_prefix(opt);
887
888 /* POSIX requires that first be decremented by one if len == 0... */
889 if (minus_len) {
890 minus_begin = diff_words->minus.orig[minus_first].begin;
891 minus_end =
892 diff_words->minus.orig[minus_first + minus_len - 1].end;
893 } else
894 minus_begin = minus_end =
895 diff_words->minus.orig[minus_first].end;
896
897 if (plus_len) {
898 plus_begin = diff_words->plus.orig[plus_first].begin;
899 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
900 } else
901 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
902
903 if (color_words_output_graph_prefix(diff_words)) {
904 fputs(line_prefix, diff_words->opt->file);
905 }
906 if (diff_words->current_plus != plus_begin) {
907 fn_out_diff_words_write_helper(diff_words->opt->file,
908 &style->ctx, style->newline,
909 plus_begin - diff_words->current_plus,
910 diff_words->current_plus, line_prefix);
911 if (*(plus_begin - 1) == '\n')
912 fputs(line_prefix, diff_words->opt->file);
913 }
914 if (minus_begin != minus_end) {
915 fn_out_diff_words_write_helper(diff_words->opt->file,
916 &style->old, style->newline,
917 minus_end - minus_begin, minus_begin,
918 line_prefix);
919 }
920 if (plus_begin != plus_end) {
921 fn_out_diff_words_write_helper(diff_words->opt->file,
922 &style->new, style->newline,
923 plus_end - plus_begin, plus_begin,
924 line_prefix);
925 }
926
927 diff_words->current_plus = plus_end;
928 diff_words->last_minus = minus_first;
929 }
930
931 /* This function starts looking at *begin, and returns 0 iff a word was found. */
932 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
933 int *begin, int *end)
934 {
935 if (word_regex && *begin < buffer->size) {
936 regmatch_t match[1];
937 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
938 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
939 '\n', match[0].rm_eo - match[0].rm_so);
940 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
941 *begin += match[0].rm_so;
942 return *begin >= *end;
943 }
944 return -1;
945 }
946
947 /* find the next word */
948 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
949 (*begin)++;
950 if (*begin >= buffer->size)
951 return -1;
952
953 /* find the end of the word */
954 *end = *begin + 1;
955 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
956 (*end)++;
957
958 return 0;
959 }
960
961 /*
962 * This function splits the words in buffer->text, stores the list with
963 * newline separator into out, and saves the offsets of the original words
964 * in buffer->orig.
965 */
966 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
967 regex_t *word_regex)
968 {
969 int i, j;
970 long alloc = 0;
971
972 out->size = 0;
973 out->ptr = NULL;
974
975 /* fake an empty "0th" word */
976 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
977 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
978 buffer->orig_nr = 1;
979
980 for (i = 0; i < buffer->text.size; i++) {
981 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
982 return;
983
984 /* store original boundaries */
985 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
986 buffer->orig_alloc);
987 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
988 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
989 buffer->orig_nr++;
990
991 /* store one word */
992 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
993 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
994 out->ptr[out->size + j - i] = '\n';
995 out->size += j - i + 1;
996
997 i = j - 1;
998 }
999 }
1000
1001 /* this executes the word diff on the accumulated buffers */
1002 static void diff_words_show(struct diff_words_data *diff_words)
1003 {
1004 xpparam_t xpp;
1005 xdemitconf_t xecfg;
1006 mmfile_t minus, plus;
1007 struct diff_words_style *style = diff_words->style;
1008
1009 struct diff_options *opt = diff_words->opt;
1010 const char *line_prefix;
1011
1012 assert(opt);
1013 line_prefix = diff_line_prefix(opt);
1014
1015 /* special case: only removal */
1016 if (!diff_words->plus.text.size) {
1017 fputs(line_prefix, diff_words->opt->file);
1018 fn_out_diff_words_write_helper(diff_words->opt->file,
1019 &style->old, style->newline,
1020 diff_words->minus.text.size,
1021 diff_words->minus.text.ptr, line_prefix);
1022 diff_words->minus.text.size = 0;
1023 return;
1024 }
1025
1026 diff_words->current_plus = diff_words->plus.text.ptr;
1027 diff_words->last_minus = 0;
1028
1029 memset(&xpp, 0, sizeof(xpp));
1030 memset(&xecfg, 0, sizeof(xecfg));
1031 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1032 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1033 xpp.flags = 0;
1034 /* as only the hunk header will be parsed, we need a 0-context */
1035 xecfg.ctxlen = 0;
1036 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1037 &xpp, &xecfg))
1038 die("unable to generate word diff");
1039 free(minus.ptr);
1040 free(plus.ptr);
1041 if (diff_words->current_plus != diff_words->plus.text.ptr +
1042 diff_words->plus.text.size) {
1043 if (color_words_output_graph_prefix(diff_words))
1044 fputs(line_prefix, diff_words->opt->file);
1045 fn_out_diff_words_write_helper(diff_words->opt->file,
1046 &style->ctx, style->newline,
1047 diff_words->plus.text.ptr + diff_words->plus.text.size
1048 - diff_words->current_plus, diff_words->current_plus,
1049 line_prefix);
1050 }
1051 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1052 }
1053
1054 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1055 static void diff_words_flush(struct emit_callback *ecbdata)
1056 {
1057 if (ecbdata->diff_words->minus.text.size ||
1058 ecbdata->diff_words->plus.text.size)
1059 diff_words_show(ecbdata->diff_words);
1060 }
1061
1062 static void diff_filespec_load_driver(struct diff_filespec *one)
1063 {
1064 /* Use already-loaded driver */
1065 if (one->driver)
1066 return;
1067
1068 if (S_ISREG(one->mode))
1069 one->driver = userdiff_find_by_path(one->path);
1070
1071 /* Fallback to default settings */
1072 if (!one->driver)
1073 one->driver = userdiff_find_by_name("default");
1074 }
1075
1076 static const char *userdiff_word_regex(struct diff_filespec *one)
1077 {
1078 diff_filespec_load_driver(one);
1079 return one->driver->word_regex;
1080 }
1081
1082 static void init_diff_words_data(struct emit_callback *ecbdata,
1083 struct diff_options *orig_opts,
1084 struct diff_filespec *one,
1085 struct diff_filespec *two)
1086 {
1087 int i;
1088 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1089 memcpy(o, orig_opts, sizeof(struct diff_options));
1090
1091 ecbdata->diff_words =
1092 xcalloc(1, sizeof(struct diff_words_data));
1093 ecbdata->diff_words->type = o->word_diff;
1094 ecbdata->diff_words->opt = o;
1095 if (!o->word_regex)
1096 o->word_regex = userdiff_word_regex(one);
1097 if (!o->word_regex)
1098 o->word_regex = userdiff_word_regex(two);
1099 if (!o->word_regex)
1100 o->word_regex = diff_word_regex_cfg;
1101 if (o->word_regex) {
1102 ecbdata->diff_words->word_regex = (regex_t *)
1103 xmalloc(sizeof(regex_t));
1104 if (regcomp(ecbdata->diff_words->word_regex,
1105 o->word_regex,
1106 REG_EXTENDED | REG_NEWLINE))
1107 die ("Invalid regular expression: %s",
1108 o->word_regex);
1109 }
1110 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1111 if (o->word_diff == diff_words_styles[i].type) {
1112 ecbdata->diff_words->style =
1113 &diff_words_styles[i];
1114 break;
1115 }
1116 }
1117 if (want_color(o->use_color)) {
1118 struct diff_words_style *st = ecbdata->diff_words->style;
1119 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1120 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1121 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1122 }
1123 }
1124
1125 static void free_diff_words_data(struct emit_callback *ecbdata)
1126 {
1127 if (ecbdata->diff_words) {
1128 diff_words_flush(ecbdata);
1129 free (ecbdata->diff_words->opt);
1130 free (ecbdata->diff_words->minus.text.ptr);
1131 free (ecbdata->diff_words->minus.orig);
1132 free (ecbdata->diff_words->plus.text.ptr);
1133 free (ecbdata->diff_words->plus.orig);
1134 if (ecbdata->diff_words->word_regex) {
1135 regfree(ecbdata->diff_words->word_regex);
1136 free(ecbdata->diff_words->word_regex);
1137 }
1138 free(ecbdata->diff_words);
1139 ecbdata->diff_words = NULL;
1140 }
1141 }
1142
1143 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1144 {
1145 if (want_color(diff_use_color))
1146 return diff_colors[ix];
1147 return "";
1148 }
1149
1150 const char *diff_line_prefix(struct diff_options *opt)
1151 {
1152 struct strbuf *msgbuf;
1153 if (!opt->output_prefix)
1154 return "";
1155
1156 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1157 return msgbuf->buf;
1158 }
1159
1160 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1161 {
1162 const char *cp;
1163 unsigned long allot;
1164 size_t l = len;
1165
1166 if (ecb->truncate)
1167 return ecb->truncate(line, len);
1168 cp = line;
1169 allot = l;
1170 while (0 < l) {
1171 (void) utf8_width(&cp, &l);
1172 if (!cp)
1173 break; /* truncated in the middle? */
1174 }
1175 return allot - l;
1176 }
1177
1178 static void find_lno(const char *line, struct emit_callback *ecbdata)
1179 {
1180 const char *p;
1181 ecbdata->lno_in_preimage = 0;
1182 ecbdata->lno_in_postimage = 0;
1183 p = strchr(line, '-');
1184 if (!p)
1185 return; /* cannot happen */
1186 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1187 p = strchr(p, '+');
1188 if (!p)
1189 return; /* cannot happen */
1190 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1191 }
1192
1193 static void fn_out_consume(void *priv, char *line, unsigned long len)
1194 {
1195 struct emit_callback *ecbdata = priv;
1196 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1197 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1198 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1199 struct diff_options *o = ecbdata->opt;
1200 const char *line_prefix = diff_line_prefix(o);
1201
1202 if (ecbdata->header) {
1203 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1204 strbuf_reset(ecbdata->header);
1205 ecbdata->header = NULL;
1206 }
1207 *(ecbdata->found_changesp) = 1;
1208
1209 if (ecbdata->label_path[0]) {
1210 const char *name_a_tab, *name_b_tab;
1211
1212 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1213 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1214
1215 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1216 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1217 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1218 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1219 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1220 }
1221
1222 if (diff_suppress_blank_empty
1223 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1224 line[0] = '\n';
1225 len = 1;
1226 }
1227
1228 if (line[0] == '@') {
1229 if (ecbdata->diff_words)
1230 diff_words_flush(ecbdata);
1231 len = sane_truncate_line(ecbdata, line, len);
1232 find_lno(line, ecbdata);
1233 emit_hunk_header(ecbdata, line, len);
1234 if (line[len-1] != '\n')
1235 putc('\n', ecbdata->opt->file);
1236 return;
1237 }
1238
1239 if (len < 1) {
1240 emit_line(ecbdata->opt, reset, reset, line, len);
1241 if (ecbdata->diff_words
1242 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1243 fputs("~\n", ecbdata->opt->file);
1244 return;
1245 }
1246
1247 if (ecbdata->diff_words) {
1248 if (line[0] == '-') {
1249 diff_words_append(line, len,
1250 &ecbdata->diff_words->minus);
1251 return;
1252 } else if (line[0] == '+') {
1253 diff_words_append(line, len,
1254 &ecbdata->diff_words->plus);
1255 return;
1256 } else if (starts_with(line, "\\ ")) {
1257 /*
1258 * Eat the "no newline at eof" marker as if we
1259 * saw a "+" or "-" line with nothing on it,
1260 * and return without diff_words_flush() to
1261 * defer processing. If this is the end of
1262 * preimage, more "+" lines may come after it.
1263 */
1264 return;
1265 }
1266 diff_words_flush(ecbdata);
1267 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1268 emit_line(ecbdata->opt, context, reset, line, len);
1269 fputs("~\n", ecbdata->opt->file);
1270 } else {
1271 /*
1272 * Skip the prefix character, if any. With
1273 * diff_suppress_blank_empty, there may be
1274 * none.
1275 */
1276 if (line[0] != '\n') {
1277 line++;
1278 len--;
1279 }
1280 emit_line(ecbdata->opt, context, reset, line, len);
1281 }
1282 return;
1283 }
1284
1285 switch (line[0]) {
1286 case '+':
1287 ecbdata->lno_in_postimage++;
1288 emit_add_line(reset, ecbdata, line + 1, len - 1);
1289 break;
1290 case '-':
1291 ecbdata->lno_in_preimage++;
1292 emit_del_line(reset, ecbdata, line + 1, len - 1);
1293 break;
1294 case ' ':
1295 ecbdata->lno_in_postimage++;
1296 ecbdata->lno_in_preimage++;
1297 emit_context_line(reset, ecbdata, line + 1, len - 1);
1298 break;
1299 default:
1300 /* incomplete line at the end */
1301 ecbdata->lno_in_preimage++;
1302 emit_line(ecbdata->opt,
1303 diff_get_color(ecbdata->color_diff, DIFF_CONTEXT),
1304 reset, line, len);
1305 break;
1306 }
1307 }
1308
1309 static char *pprint_rename(const char *a, const char *b)
1310 {
1311 const char *old = a;
1312 const char *new = b;
1313 struct strbuf name = STRBUF_INIT;
1314 int pfx_length, sfx_length;
1315 int pfx_adjust_for_slash;
1316 int len_a = strlen(a);
1317 int len_b = strlen(b);
1318 int a_midlen, b_midlen;
1319 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1320 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1321
1322 if (qlen_a || qlen_b) {
1323 quote_c_style(a, &name, NULL, 0);
1324 strbuf_addstr(&name, " => ");
1325 quote_c_style(b, &name, NULL, 0);
1326 return strbuf_detach(&name, NULL);
1327 }
1328
1329 /* Find common prefix */
1330 pfx_length = 0;
1331 while (*old && *new && *old == *new) {
1332 if (*old == '/')
1333 pfx_length = old - a + 1;
1334 old++;
1335 new++;
1336 }
1337
1338 /* Find common suffix */
1339 old = a + len_a;
1340 new = b + len_b;
1341 sfx_length = 0;
1342 /*
1343 * If there is a common prefix, it must end in a slash. In
1344 * that case we let this loop run 1 into the prefix to see the
1345 * same slash.
1346 *
1347 * If there is no common prefix, we cannot do this as it would
1348 * underrun the input strings.
1349 */
1350 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1351 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1352 b + pfx_length - pfx_adjust_for_slash <= new &&
1353 *old == *new) {
1354 if (*old == '/')
1355 sfx_length = len_a - (old - a);
1356 old--;
1357 new--;
1358 }
1359
1360 /*
1361 * pfx{mid-a => mid-b}sfx
1362 * {pfx-a => pfx-b}sfx
1363 * pfx{sfx-a => sfx-b}
1364 * name-a => name-b
1365 */
1366 a_midlen = len_a - pfx_length - sfx_length;
1367 b_midlen = len_b - pfx_length - sfx_length;
1368 if (a_midlen < 0)
1369 a_midlen = 0;
1370 if (b_midlen < 0)
1371 b_midlen = 0;
1372
1373 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1374 if (pfx_length + sfx_length) {
1375 strbuf_add(&name, a, pfx_length);
1376 strbuf_addch(&name, '{');
1377 }
1378 strbuf_add(&name, a + pfx_length, a_midlen);
1379 strbuf_addstr(&name, " => ");
1380 strbuf_add(&name, b + pfx_length, b_midlen);
1381 if (pfx_length + sfx_length) {
1382 strbuf_addch(&name, '}');
1383 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1384 }
1385 return strbuf_detach(&name, NULL);
1386 }
1387
1388 struct diffstat_t {
1389 int nr;
1390 int alloc;
1391 struct diffstat_file {
1392 char *from_name;
1393 char *name;
1394 char *print_name;
1395 unsigned is_unmerged:1;
1396 unsigned is_binary:1;
1397 unsigned is_renamed:1;
1398 unsigned is_interesting:1;
1399 uintmax_t added, deleted;
1400 } **files;
1401 };
1402
1403 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1404 const char *name_a,
1405 const char *name_b)
1406 {
1407 struct diffstat_file *x;
1408 x = xcalloc(1, sizeof(*x));
1409 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1410 diffstat->files[diffstat->nr++] = x;
1411 if (name_b) {
1412 x->from_name = xstrdup(name_a);
1413 x->name = xstrdup(name_b);
1414 x->is_renamed = 1;
1415 }
1416 else {
1417 x->from_name = NULL;
1418 x->name = xstrdup(name_a);
1419 }
1420 return x;
1421 }
1422
1423 static void diffstat_consume(void *priv, char *line, unsigned long len)
1424 {
1425 struct diffstat_t *diffstat = priv;
1426 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1427
1428 if (line[0] == '+')
1429 x->added++;
1430 else if (line[0] == '-')
1431 x->deleted++;
1432 }
1433
1434 const char mime_boundary_leader[] = "------------";
1435
1436 static int scale_linear(int it, int width, int max_change)
1437 {
1438 if (!it)
1439 return 0;
1440 /*
1441 * make sure that at least one '-' or '+' is printed if
1442 * there is any change to this path. The easiest way is to
1443 * scale linearly as if the alloted width is one column shorter
1444 * than it is, and then add 1 to the result.
1445 */
1446 return 1 + (it * (width - 1) / max_change);
1447 }
1448
1449 static void show_name(FILE *file,
1450 const char *prefix, const char *name, int len)
1451 {
1452 fprintf(file, " %s%-*s |", prefix, len, name);
1453 }
1454
1455 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1456 {
1457 if (cnt <= 0)
1458 return;
1459 fprintf(file, "%s", set);
1460 while (cnt--)
1461 putc(ch, file);
1462 fprintf(file, "%s", reset);
1463 }
1464
1465 static void fill_print_name(struct diffstat_file *file)
1466 {
1467 char *pname;
1468
1469 if (file->print_name)
1470 return;
1471
1472 if (!file->is_renamed) {
1473 struct strbuf buf = STRBUF_INIT;
1474 if (quote_c_style(file->name, &buf, NULL, 0)) {
1475 pname = strbuf_detach(&buf, NULL);
1476 } else {
1477 pname = file->name;
1478 strbuf_release(&buf);
1479 }
1480 } else {
1481 pname = pprint_rename(file->from_name, file->name);
1482 }
1483 file->print_name = pname;
1484 }
1485
1486 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1487 {
1488 struct strbuf sb = STRBUF_INIT;
1489 int ret;
1490
1491 if (!files) {
1492 assert(insertions == 0 && deletions == 0);
1493 return fprintf(fp, "%s\n", " 0 files changed");
1494 }
1495
1496 strbuf_addf(&sb,
1497 (files == 1) ? " %d file changed" : " %d files changed",
1498 files);
1499
1500 /*
1501 * For binary diff, the caller may want to print "x files
1502 * changed" with insertions == 0 && deletions == 0.
1503 *
1504 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1505 * is probably less confusing (i.e skip over "2 files changed
1506 * but nothing about added/removed lines? Is this a bug in Git?").
1507 */
1508 if (insertions || deletions == 0) {
1509 strbuf_addf(&sb,
1510 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1511 insertions);
1512 }
1513
1514 if (deletions || insertions == 0) {
1515 strbuf_addf(&sb,
1516 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1517 deletions);
1518 }
1519 strbuf_addch(&sb, '\n');
1520 ret = fputs(sb.buf, fp);
1521 strbuf_release(&sb);
1522 return ret;
1523 }
1524
1525 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1526 {
1527 int i, len, add, del, adds = 0, dels = 0;
1528 uintmax_t max_change = 0, max_len = 0;
1529 int total_files = data->nr, count;
1530 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1531 const char *reset, *add_c, *del_c;
1532 const char *line_prefix = "";
1533 int extra_shown = 0;
1534
1535 if (data->nr == 0)
1536 return;
1537
1538 line_prefix = diff_line_prefix(options);
1539 count = options->stat_count ? options->stat_count : data->nr;
1540
1541 reset = diff_get_color_opt(options, DIFF_RESET);
1542 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1543 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1544
1545 /*
1546 * Find the longest filename and max number of changes
1547 */
1548 for (i = 0; (i < count) && (i < data->nr); i++) {
1549 struct diffstat_file *file = data->files[i];
1550 uintmax_t change = file->added + file->deleted;
1551
1552 if (!file->is_interesting && (change == 0)) {
1553 count++; /* not shown == room for one more */
1554 continue;
1555 }
1556 fill_print_name(file);
1557 len = strlen(file->print_name);
1558 if (max_len < len)
1559 max_len = len;
1560
1561 if (file->is_unmerged) {
1562 /* "Unmerged" is 8 characters */
1563 bin_width = bin_width < 8 ? 8 : bin_width;
1564 continue;
1565 }
1566 if (file->is_binary) {
1567 /* "Bin XXX -> YYY bytes" */
1568 int w = 14 + decimal_width(file->added)
1569 + decimal_width(file->deleted);
1570 bin_width = bin_width < w ? w : bin_width;
1571 /* Display change counts aligned with "Bin" */
1572 number_width = 3;
1573 continue;
1574 }
1575
1576 if (max_change < change)
1577 max_change = change;
1578 }
1579 count = i; /* where we can stop scanning in data->files[] */
1580
1581 /*
1582 * We have width = stat_width or term_columns() columns total.
1583 * We want a maximum of min(max_len, stat_name_width) for the name part.
1584 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1585 * We also need 1 for " " and 4 + decimal_width(max_change)
1586 * for " | NNNN " and one the empty column at the end, altogether
1587 * 6 + decimal_width(max_change).
1588 *
1589 * If there's not enough space, we will use the smaller of
1590 * stat_name_width (if set) and 5/8*width for the filename,
1591 * and the rest for constant elements + graph part, but no more
1592 * than stat_graph_width for the graph part.
1593 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1594 * for the standard terminal size).
1595 *
1596 * In other words: stat_width limits the maximum width, and
1597 * stat_name_width fixes the maximum width of the filename,
1598 * and is also used to divide available columns if there
1599 * aren't enough.
1600 *
1601 * Binary files are displayed with "Bin XXX -> YYY bytes"
1602 * instead of the change count and graph. This part is treated
1603 * similarly to the graph part, except that it is not
1604 * "scaled". If total width is too small to accommodate the
1605 * guaranteed minimum width of the filename part and the
1606 * separators and this message, this message will "overflow"
1607 * making the line longer than the maximum width.
1608 */
1609
1610 if (options->stat_width == -1)
1611 width = term_columns() - options->output_prefix_length;
1612 else
1613 width = options->stat_width ? options->stat_width : 80;
1614 number_width = decimal_width(max_change) > number_width ?
1615 decimal_width(max_change) : number_width;
1616
1617 if (options->stat_graph_width == -1)
1618 options->stat_graph_width = diff_stat_graph_width;
1619
1620 /*
1621 * Guarantee 3/8*16==6 for the graph part
1622 * and 5/8*16==10 for the filename part
1623 */
1624 if (width < 16 + 6 + number_width)
1625 width = 16 + 6 + number_width;
1626
1627 /*
1628 * First assign sizes that are wanted, ignoring available width.
1629 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1630 * starting from "XXX" should fit in graph_width.
1631 */
1632 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1633 if (options->stat_graph_width &&
1634 options->stat_graph_width < graph_width)
1635 graph_width = options->stat_graph_width;
1636
1637 name_width = (options->stat_name_width > 0 &&
1638 options->stat_name_width < max_len) ?
1639 options->stat_name_width : max_len;
1640
1641 /*
1642 * Adjust adjustable widths not to exceed maximum width
1643 */
1644 if (name_width + number_width + 6 + graph_width > width) {
1645 if (graph_width > width * 3/8 - number_width - 6) {
1646 graph_width = width * 3/8 - number_width - 6;
1647 if (graph_width < 6)
1648 graph_width = 6;
1649 }
1650
1651 if (options->stat_graph_width &&
1652 graph_width > options->stat_graph_width)
1653 graph_width = options->stat_graph_width;
1654 if (name_width > width - number_width - 6 - graph_width)
1655 name_width = width - number_width - 6 - graph_width;
1656 else
1657 graph_width = width - number_width - 6 - name_width;
1658 }
1659
1660 /*
1661 * From here name_width is the width of the name area,
1662 * and graph_width is the width of the graph area.
1663 * max_change is used to scale graph properly.
1664 */
1665 for (i = 0; i < count; i++) {
1666 const char *prefix = "";
1667 struct diffstat_file *file = data->files[i];
1668 char *name = file->print_name;
1669 uintmax_t added = file->added;
1670 uintmax_t deleted = file->deleted;
1671 int name_len;
1672
1673 if (!file->is_interesting && (added + deleted == 0))
1674 continue;
1675
1676 /*
1677 * "scale" the filename
1678 */
1679 len = name_width;
1680 name_len = strlen(name);
1681 if (name_width < name_len) {
1682 char *slash;
1683 prefix = "...";
1684 len -= 3;
1685 name += name_len - len;
1686 slash = strchr(name, '/');
1687 if (slash)
1688 name = slash;
1689 }
1690
1691 if (file->is_binary) {
1692 fprintf(options->file, "%s", line_prefix);
1693 show_name(options->file, prefix, name, len);
1694 fprintf(options->file, " %*s", number_width, "Bin");
1695 if (!added && !deleted) {
1696 putc('\n', options->file);
1697 continue;
1698 }
1699 fprintf(options->file, " %s%"PRIuMAX"%s",
1700 del_c, deleted, reset);
1701 fprintf(options->file, " -> ");
1702 fprintf(options->file, "%s%"PRIuMAX"%s",
1703 add_c, added, reset);
1704 fprintf(options->file, " bytes");
1705 fprintf(options->file, "\n");
1706 continue;
1707 }
1708 else if (file->is_unmerged) {
1709 fprintf(options->file, "%s", line_prefix);
1710 show_name(options->file, prefix, name, len);
1711 fprintf(options->file, " Unmerged\n");
1712 continue;
1713 }
1714
1715 /*
1716 * scale the add/delete
1717 */
1718 add = added;
1719 del = deleted;
1720
1721 if (graph_width <= max_change) {
1722 int total = scale_linear(add + del, graph_width, max_change);
1723 if (total < 2 && add && del)
1724 /* width >= 2 due to the sanity check */
1725 total = 2;
1726 if (add < del) {
1727 add = scale_linear(add, graph_width, max_change);
1728 del = total - add;
1729 } else {
1730 del = scale_linear(del, graph_width, max_change);
1731 add = total - del;
1732 }
1733 }
1734 fprintf(options->file, "%s", line_prefix);
1735 show_name(options->file, prefix, name, len);
1736 fprintf(options->file, " %*"PRIuMAX"%s",
1737 number_width, added + deleted,
1738 added + deleted ? " " : "");
1739 show_graph(options->file, '+', add, add_c, reset);
1740 show_graph(options->file, '-', del, del_c, reset);
1741 fprintf(options->file, "\n");
1742 }
1743
1744 for (i = 0; i < data->nr; i++) {
1745 struct diffstat_file *file = data->files[i];
1746 uintmax_t added = file->added;
1747 uintmax_t deleted = file->deleted;
1748
1749 if (file->is_unmerged ||
1750 (!file->is_interesting && (added + deleted == 0))) {
1751 total_files--;
1752 continue;
1753 }
1754
1755 if (!file->is_binary) {
1756 adds += added;
1757 dels += deleted;
1758 }
1759 if (i < count)
1760 continue;
1761 if (!extra_shown)
1762 fprintf(options->file, "%s ...\n", line_prefix);
1763 extra_shown = 1;
1764 }
1765 fprintf(options->file, "%s", line_prefix);
1766 print_stat_summary(options->file, total_files, adds, dels);
1767 }
1768
1769 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1770 {
1771 int i, adds = 0, dels = 0, total_files = data->nr;
1772
1773 if (data->nr == 0)
1774 return;
1775
1776 for (i = 0; i < data->nr; i++) {
1777 int added = data->files[i]->added;
1778 int deleted= data->files[i]->deleted;
1779
1780 if (data->files[i]->is_unmerged ||
1781 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1782 total_files--;
1783 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1784 adds += added;
1785 dels += deleted;
1786 }
1787 }
1788 fprintf(options->file, "%s", diff_line_prefix(options));
1789 print_stat_summary(options->file, total_files, adds, dels);
1790 }
1791
1792 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1793 {
1794 int i;
1795
1796 if (data->nr == 0)
1797 return;
1798
1799 for (i = 0; i < data->nr; i++) {
1800 struct diffstat_file *file = data->files[i];
1801
1802 fprintf(options->file, "%s", diff_line_prefix(options));
1803
1804 if (file->is_binary)
1805 fprintf(options->file, "-\t-\t");
1806 else
1807 fprintf(options->file,
1808 "%"PRIuMAX"\t%"PRIuMAX"\t",
1809 file->added, file->deleted);
1810 if (options->line_termination) {
1811 fill_print_name(file);
1812 if (!file->is_renamed)
1813 write_name_quoted(file->name, options->file,
1814 options->line_termination);
1815 else {
1816 fputs(file->print_name, options->file);
1817 putc(options->line_termination, options->file);
1818 }
1819 } else {
1820 if (file->is_renamed) {
1821 putc('\0', options->file);
1822 write_name_quoted(file->from_name, options->file, '\0');
1823 }
1824 write_name_quoted(file->name, options->file, '\0');
1825 }
1826 }
1827 }
1828
1829 struct dirstat_file {
1830 const char *name;
1831 unsigned long changed;
1832 };
1833
1834 struct dirstat_dir {
1835 struct dirstat_file *files;
1836 int alloc, nr, permille, cumulative;
1837 };
1838
1839 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1840 unsigned long changed, const char *base, int baselen)
1841 {
1842 unsigned long this_dir = 0;
1843 unsigned int sources = 0;
1844 const char *line_prefix = diff_line_prefix(opt);
1845
1846 while (dir->nr) {
1847 struct dirstat_file *f = dir->files;
1848 int namelen = strlen(f->name);
1849 unsigned long this;
1850 char *slash;
1851
1852 if (namelen < baselen)
1853 break;
1854 if (memcmp(f->name, base, baselen))
1855 break;
1856 slash = strchr(f->name + baselen, '/');
1857 if (slash) {
1858 int newbaselen = slash + 1 - f->name;
1859 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1860 sources++;
1861 } else {
1862 this = f->changed;
1863 dir->files++;
1864 dir->nr--;
1865 sources += 2;
1866 }
1867 this_dir += this;
1868 }
1869
1870 /*
1871 * We don't report dirstat's for
1872 * - the top level
1873 * - or cases where everything came from a single directory
1874 * under this directory (sources == 1).
1875 */
1876 if (baselen && sources != 1) {
1877 if (this_dir) {
1878 int permille = this_dir * 1000 / changed;
1879 if (permille >= dir->permille) {
1880 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1881 permille / 10, permille % 10, baselen, base);
1882 if (!dir->cumulative)
1883 return 0;
1884 }
1885 }
1886 }
1887 return this_dir;
1888 }
1889
1890 static int dirstat_compare(const void *_a, const void *_b)
1891 {
1892 const struct dirstat_file *a = _a;
1893 const struct dirstat_file *b = _b;
1894 return strcmp(a->name, b->name);
1895 }
1896
1897 static void show_dirstat(struct diff_options *options)
1898 {
1899 int i;
1900 unsigned long changed;
1901 struct dirstat_dir dir;
1902 struct diff_queue_struct *q = &diff_queued_diff;
1903
1904 dir.files = NULL;
1905 dir.alloc = 0;
1906 dir.nr = 0;
1907 dir.permille = options->dirstat_permille;
1908 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1909
1910 changed = 0;
1911 for (i = 0; i < q->nr; i++) {
1912 struct diff_filepair *p = q->queue[i];
1913 const char *name;
1914 unsigned long copied, added, damage;
1915 int content_changed;
1916
1917 name = p->two->path ? p->two->path : p->one->path;
1918
1919 if (p->one->sha1_valid && p->two->sha1_valid)
1920 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1921 else
1922 content_changed = 1;
1923
1924 if (!content_changed) {
1925 /*
1926 * The SHA1 has not changed, so pre-/post-content is
1927 * identical. We can therefore skip looking at the
1928 * file contents altogether.
1929 */
1930 damage = 0;
1931 goto found_damage;
1932 }
1933
1934 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1935 /*
1936 * In --dirstat-by-file mode, we don't really need to
1937 * look at the actual file contents at all.
1938 * The fact that the SHA1 changed is enough for us to
1939 * add this file to the list of results
1940 * (with each file contributing equal damage).
1941 */
1942 damage = 1;
1943 goto found_damage;
1944 }
1945
1946 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1947 diff_populate_filespec(p->one, 0);
1948 diff_populate_filespec(p->two, 0);
1949 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1950 &copied, &added);
1951 diff_free_filespec_data(p->one);
1952 diff_free_filespec_data(p->two);
1953 } else if (DIFF_FILE_VALID(p->one)) {
1954 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
1955 copied = added = 0;
1956 diff_free_filespec_data(p->one);
1957 } else if (DIFF_FILE_VALID(p->two)) {
1958 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
1959 copied = 0;
1960 added = p->two->size;
1961 diff_free_filespec_data(p->two);
1962 } else
1963 continue;
1964
1965 /*
1966 * Original minus copied is the removed material,
1967 * added is the new material. They are both damages
1968 * made to the preimage.
1969 * If the resulting damage is zero, we know that
1970 * diffcore_count_changes() considers the two entries to
1971 * be identical, but since content_changed is true, we
1972 * know that there must have been _some_ kind of change,
1973 * so we force all entries to have damage > 0.
1974 */
1975 damage = (p->one->size - copied) + added;
1976 if (!damage)
1977 damage = 1;
1978
1979 found_damage:
1980 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1981 dir.files[dir.nr].name = name;
1982 dir.files[dir.nr].changed = damage;
1983 changed += damage;
1984 dir.nr++;
1985 }
1986
1987 /* This can happen even with many files, if everything was renames */
1988 if (!changed)
1989 return;
1990
1991 /* Show all directories with more than x% of the changes */
1992 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1993 gather_dirstat(options, &dir, changed, "", 0);
1994 }
1995
1996 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1997 {
1998 int i;
1999 unsigned long changed;
2000 struct dirstat_dir dir;
2001
2002 if (data->nr == 0)
2003 return;
2004
2005 dir.files = NULL;
2006 dir.alloc = 0;
2007 dir.nr = 0;
2008 dir.permille = options->dirstat_permille;
2009 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2010
2011 changed = 0;
2012 for (i = 0; i < data->nr; i++) {
2013 struct diffstat_file *file = data->files[i];
2014 unsigned long damage = file->added + file->deleted;
2015 if (file->is_binary)
2016 /*
2017 * binary files counts bytes, not lines. Must find some
2018 * way to normalize binary bytes vs. textual lines.
2019 * The following heuristic assumes that there are 64
2020 * bytes per "line".
2021 * This is stupid and ugly, but very cheap...
2022 */
2023 damage = (damage + 63) / 64;
2024 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2025 dir.files[dir.nr].name = file->name;
2026 dir.files[dir.nr].changed = damage;
2027 changed += damage;
2028 dir.nr++;
2029 }
2030
2031 /* This can happen even with many files, if everything was renames */
2032 if (!changed)
2033 return;
2034
2035 /* Show all directories with more than x% of the changes */
2036 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
2037 gather_dirstat(options, &dir, changed, "", 0);
2038 }
2039
2040 static void free_diffstat_info(struct diffstat_t *diffstat)
2041 {
2042 int i;
2043 for (i = 0; i < diffstat->nr; i++) {
2044 struct diffstat_file *f = diffstat->files[i];
2045 if (f->name != f->print_name)
2046 free(f->print_name);
2047 free(f->name);
2048 free(f->from_name);
2049 free(f);
2050 }
2051 free(diffstat->files);
2052 }
2053
2054 struct checkdiff_t {
2055 const char *filename;
2056 int lineno;
2057 int conflict_marker_size;
2058 struct diff_options *o;
2059 unsigned ws_rule;
2060 unsigned status;
2061 };
2062
2063 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2064 {
2065 char firstchar;
2066 int cnt;
2067
2068 if (len < marker_size + 1)
2069 return 0;
2070 firstchar = line[0];
2071 switch (firstchar) {
2072 case '=': case '>': case '<': case '|':
2073 break;
2074 default:
2075 return 0;
2076 }
2077 for (cnt = 1; cnt < marker_size; cnt++)
2078 if (line[cnt] != firstchar)
2079 return 0;
2080 /* line[1] thru line[marker_size-1] are same as firstchar */
2081 if (len < marker_size + 1 || !isspace(line[marker_size]))
2082 return 0;
2083 return 1;
2084 }
2085
2086 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2087 {
2088 struct checkdiff_t *data = priv;
2089 int marker_size = data->conflict_marker_size;
2090 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2091 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2092 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2093 char *err;
2094 const char *line_prefix;
2095
2096 assert(data->o);
2097 line_prefix = diff_line_prefix(data->o);
2098
2099 if (line[0] == '+') {
2100 unsigned bad;
2101 data->lineno++;
2102 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2103 data->status |= 1;
2104 fprintf(data->o->file,
2105 "%s%s:%d: leftover conflict marker\n",
2106 line_prefix, data->filename, data->lineno);
2107 }
2108 bad = ws_check(line + 1, len - 1, data->ws_rule);
2109 if (!bad)
2110 return;
2111 data->status |= bad;
2112 err = whitespace_error_string(bad);
2113 fprintf(data->o->file, "%s%s:%d: %s.\n",
2114 line_prefix, data->filename, data->lineno, err);
2115 free(err);
2116 emit_line(data->o, set, reset, line, 1);
2117 ws_check_emit(line + 1, len - 1, data->ws_rule,
2118 data->o->file, set, reset, ws);
2119 } else if (line[0] == ' ') {
2120 data->lineno++;
2121 } else if (line[0] == '@') {
2122 char *plus = strchr(line, '+');
2123 if (plus)
2124 data->lineno = strtol(plus, NULL, 10) - 1;
2125 else
2126 die("invalid diff");
2127 }
2128 }
2129
2130 static unsigned char *deflate_it(char *data,
2131 unsigned long size,
2132 unsigned long *result_size)
2133 {
2134 int bound;
2135 unsigned char *deflated;
2136 git_zstream stream;
2137
2138 git_deflate_init(&stream, zlib_compression_level);
2139 bound = git_deflate_bound(&stream, size);
2140 deflated = xmalloc(bound);
2141 stream.next_out = deflated;
2142 stream.avail_out = bound;
2143
2144 stream.next_in = (unsigned char *)data;
2145 stream.avail_in = size;
2146 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2147 ; /* nothing */
2148 git_deflate_end(&stream);
2149 *result_size = stream.total_out;
2150 return deflated;
2151 }
2152
2153 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2154 const char *prefix)
2155 {
2156 void *cp;
2157 void *delta;
2158 void *deflated;
2159 void *data;
2160 unsigned long orig_size;
2161 unsigned long delta_size;
2162 unsigned long deflate_size;
2163 unsigned long data_size;
2164
2165 /* We could do deflated delta, or we could do just deflated two,
2166 * whichever is smaller.
2167 */
2168 delta = NULL;
2169 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2170 if (one->size && two->size) {
2171 delta = diff_delta(one->ptr, one->size,
2172 two->ptr, two->size,
2173 &delta_size, deflate_size);
2174 if (delta) {
2175 void *to_free = delta;
2176 orig_size = delta_size;
2177 delta = deflate_it(delta, delta_size, &delta_size);
2178 free(to_free);
2179 }
2180 }
2181
2182 if (delta && delta_size < deflate_size) {
2183 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2184 free(deflated);
2185 data = delta;
2186 data_size = delta_size;
2187 }
2188 else {
2189 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2190 free(delta);
2191 data = deflated;
2192 data_size = deflate_size;
2193 }
2194
2195 /* emit data encoded in base85 */
2196 cp = data;
2197 while (data_size) {
2198 int bytes = (52 < data_size) ? 52 : data_size;
2199 char line[70];
2200 data_size -= bytes;
2201 if (bytes <= 26)
2202 line[0] = bytes + 'A' - 1;
2203 else
2204 line[0] = bytes - 26 + 'a' - 1;
2205 encode_85(line + 1, cp, bytes);
2206 cp = (char *) cp + bytes;
2207 fprintf(file, "%s", prefix);
2208 fputs(line, file);
2209 fputc('\n', file);
2210 }
2211 fprintf(file, "%s\n", prefix);
2212 free(data);
2213 }
2214
2215 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2216 const char *prefix)
2217 {
2218 fprintf(file, "%sGIT binary patch\n", prefix);
2219 emit_binary_diff_body(file, one, two, prefix);
2220 emit_binary_diff_body(file, two, one, prefix);
2221 }
2222
2223 int diff_filespec_is_binary(struct diff_filespec *one)
2224 {
2225 if (one->is_binary == -1) {
2226 diff_filespec_load_driver(one);
2227 if (one->driver->binary != -1)
2228 one->is_binary = one->driver->binary;
2229 else {
2230 if (!one->data && DIFF_FILE_VALID(one))
2231 diff_populate_filespec(one, CHECK_BINARY);
2232 if (one->is_binary == -1 && one->data)
2233 one->is_binary = buffer_is_binary(one->data,
2234 one->size);
2235 if (one->is_binary == -1)
2236 one->is_binary = 0;
2237 }
2238 }
2239 return one->is_binary;
2240 }
2241
2242 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2243 {
2244 diff_filespec_load_driver(one);
2245 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2246 }
2247
2248 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2249 {
2250 if (!options->a_prefix)
2251 options->a_prefix = a;
2252 if (!options->b_prefix)
2253 options->b_prefix = b;
2254 }
2255
2256 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2257 {
2258 if (!DIFF_FILE_VALID(one))
2259 return NULL;
2260
2261 diff_filespec_load_driver(one);
2262 return userdiff_get_textconv(one->driver);
2263 }
2264
2265 static void builtin_diff(const char *name_a,
2266 const char *name_b,
2267 struct diff_filespec *one,
2268 struct diff_filespec *two,
2269 const char *xfrm_msg,
2270 int must_show_header,
2271 struct diff_options *o,
2272 int complete_rewrite)
2273 {
2274 mmfile_t mf1, mf2;
2275 const char *lbl[2];
2276 char *a_one, *b_two;
2277 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2278 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2279 const char *a_prefix, *b_prefix;
2280 struct userdiff_driver *textconv_one = NULL;
2281 struct userdiff_driver *textconv_two = NULL;
2282 struct strbuf header = STRBUF_INIT;
2283 const char *line_prefix = diff_line_prefix(o);
2284
2285 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2286 (!one->mode || S_ISGITLINK(one->mode)) &&
2287 (!two->mode || S_ISGITLINK(two->mode))) {
2288 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2289 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2290 show_submodule_summary(o->file, one->path ? one->path : two->path,
2291 line_prefix,
2292 one->sha1, two->sha1, two->dirty_submodule,
2293 meta, del, add, reset);
2294 return;
2295 }
2296
2297 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2298 textconv_one = get_textconv(one);
2299 textconv_two = get_textconv(two);
2300 }
2301
2302 diff_set_mnemonic_prefix(o, "a/", "b/");
2303 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2304 a_prefix = o->b_prefix;
2305 b_prefix = o->a_prefix;
2306 } else {
2307 a_prefix = o->a_prefix;
2308 b_prefix = o->b_prefix;
2309 }
2310
2311 /* Never use a non-valid filename anywhere if at all possible */
2312 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2313 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2314
2315 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2316 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2317 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2318 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2319 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2320 if (lbl[0][0] == '/') {
2321 /* /dev/null */
2322 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2323 if (xfrm_msg)
2324 strbuf_addstr(&header, xfrm_msg);
2325 must_show_header = 1;
2326 }
2327 else if (lbl[1][0] == '/') {
2328 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2329 if (xfrm_msg)
2330 strbuf_addstr(&header, xfrm_msg);
2331 must_show_header = 1;
2332 }
2333 else {
2334 if (one->mode != two->mode) {
2335 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2336 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2337 must_show_header = 1;
2338 }
2339 if (xfrm_msg)
2340 strbuf_addstr(&header, xfrm_msg);
2341
2342 /*
2343 * we do not run diff between different kind
2344 * of objects.
2345 */
2346 if ((one->mode ^ two->mode) & S_IFMT)
2347 goto free_ab_and_return;
2348 if (complete_rewrite &&
2349 (textconv_one || !diff_filespec_is_binary(one)) &&
2350 (textconv_two || !diff_filespec_is_binary(two))) {
2351 fprintf(o->file, "%s", header.buf);
2352 strbuf_reset(&header);
2353 emit_rewrite_diff(name_a, name_b, one, two,
2354 textconv_one, textconv_two, o);
2355 o->found_changes = 1;
2356 goto free_ab_and_return;
2357 }
2358 }
2359
2360 if (o->irreversible_delete && lbl[1][0] == '/') {
2361 fprintf(o->file, "%s", header.buf);
2362 strbuf_reset(&header);
2363 goto free_ab_and_return;
2364 } else if (!DIFF_OPT_TST(o, TEXT) &&
2365 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2366 (!textconv_two && diff_filespec_is_binary(two)) )) {
2367 if (!one->data && !two->data &&
2368 S_ISREG(one->mode) && S_ISREG(two->mode) &&
2369 !DIFF_OPT_TST(o, BINARY)) {
2370 if (!hashcmp(one->sha1, two->sha1)) {
2371 if (must_show_header)
2372 fprintf(o->file, "%s", header.buf);
2373 goto free_ab_and_return;
2374 }
2375 fprintf(o->file, "%s", header.buf);
2376 fprintf(o->file, "%sBinary files %s and %s differ\n",
2377 line_prefix, lbl[0], lbl[1]);
2378 goto free_ab_and_return;
2379 }
2380 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2381 die("unable to read files to diff");
2382 /* Quite common confusing case */
2383 if (mf1.size == mf2.size &&
2384 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2385 if (must_show_header)
2386 fprintf(o->file, "%s", header.buf);
2387 goto free_ab_and_return;
2388 }
2389 fprintf(o->file, "%s", header.buf);
2390 strbuf_reset(&header);
2391 if (DIFF_OPT_TST(o, BINARY))
2392 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2393 else
2394 fprintf(o->file, "%sBinary files %s and %s differ\n",
2395 line_prefix, lbl[0], lbl[1]);
2396 o->found_changes = 1;
2397 } else {
2398 /* Crazy xdl interfaces.. */
2399 const char *diffopts = getenv("GIT_DIFF_OPTS");
2400 const char *v;
2401 xpparam_t xpp;
2402 xdemitconf_t xecfg;
2403 struct emit_callback ecbdata;
2404 const struct userdiff_funcname *pe;
2405
2406 if (must_show_header) {
2407 fprintf(o->file, "%s", header.buf);
2408 strbuf_reset(&header);
2409 }
2410
2411 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2412 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2413
2414 pe = diff_funcname_pattern(one);
2415 if (!pe)
2416 pe = diff_funcname_pattern(two);
2417
2418 memset(&xpp, 0, sizeof(xpp));
2419 memset(&xecfg, 0, sizeof(xecfg));
2420 memset(&ecbdata, 0, sizeof(ecbdata));
2421 ecbdata.label_path = lbl;
2422 ecbdata.color_diff = want_color(o->use_color);
2423 ecbdata.found_changesp = &o->found_changes;
2424 ecbdata.ws_rule = whitespace_rule(name_b);
2425 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2426 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2427 ecbdata.opt = o;
2428 ecbdata.header = header.len ? &header : NULL;
2429 xpp.flags = o->xdl_opts;
2430 xecfg.ctxlen = o->context;
2431 xecfg.interhunkctxlen = o->interhunkcontext;
2432 xecfg.flags = XDL_EMIT_FUNCNAMES;
2433 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2434 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2435 if (pe)
2436 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2437 if (!diffopts)
2438 ;
2439 else if (skip_prefix(diffopts, "--unified=", &v))
2440 xecfg.ctxlen = strtoul(v, NULL, 10);
2441 else if (skip_prefix(diffopts, "-u", &v))
2442 xecfg.ctxlen = strtoul(v, NULL, 10);
2443 if (o->word_diff)
2444 init_diff_words_data(&ecbdata, o, one, two);
2445 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2446 &xpp, &xecfg))
2447 die("unable to generate diff for %s", one->path);
2448 if (o->word_diff)
2449 free_diff_words_data(&ecbdata);
2450 if (textconv_one)
2451 free(mf1.ptr);
2452 if (textconv_two)
2453 free(mf2.ptr);
2454 xdiff_clear_find_func(&xecfg);
2455 }
2456
2457 free_ab_and_return:
2458 strbuf_release(&header);
2459 diff_free_filespec_data(one);
2460 diff_free_filespec_data(two);
2461 free(a_one);
2462 free(b_two);
2463 return;
2464 }
2465
2466 static void builtin_diffstat(const char *name_a, const char *name_b,
2467 struct diff_filespec *one,
2468 struct diff_filespec *two,
2469 struct diffstat_t *diffstat,
2470 struct diff_options *o,
2471 struct diff_filepair *p)
2472 {
2473 mmfile_t mf1, mf2;
2474 struct diffstat_file *data;
2475 int same_contents;
2476 int complete_rewrite = 0;
2477
2478 if (!DIFF_PAIR_UNMERGED(p)) {
2479 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2480 complete_rewrite = 1;
2481 }
2482
2483 data = diffstat_add(diffstat, name_a, name_b);
2484 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2485
2486 if (!one || !two) {
2487 data->is_unmerged = 1;
2488 return;
2489 }
2490
2491 same_contents = !hashcmp(one->sha1, two->sha1);
2492
2493 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2494 data->is_binary = 1;
2495 if (same_contents) {
2496 data->added = 0;
2497 data->deleted = 0;
2498 } else {
2499 data->added = diff_filespec_size(two);
2500 data->deleted = diff_filespec_size(one);
2501 }
2502 }
2503
2504 else if (complete_rewrite) {
2505 diff_populate_filespec(one, 0);
2506 diff_populate_filespec(two, 0);
2507 data->deleted = count_lines(one->data, one->size);
2508 data->added = count_lines(two->data, two->size);
2509 }
2510
2511 else if (!same_contents) {
2512 /* Crazy xdl interfaces.. */
2513 xpparam_t xpp;
2514 xdemitconf_t xecfg;
2515
2516 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2517 die("unable to read files to diff");
2518
2519 memset(&xpp, 0, sizeof(xpp));
2520 memset(&xecfg, 0, sizeof(xecfg));
2521 xpp.flags = o->xdl_opts;
2522 xecfg.ctxlen = o->context;
2523 xecfg.interhunkctxlen = o->interhunkcontext;
2524 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2525 &xpp, &xecfg))
2526 die("unable to generate diffstat for %s", one->path);
2527 }
2528
2529 diff_free_filespec_data(one);
2530 diff_free_filespec_data(two);
2531 }
2532
2533 static void builtin_checkdiff(const char *name_a, const char *name_b,
2534 const char *attr_path,
2535 struct diff_filespec *one,
2536 struct diff_filespec *two,
2537 struct diff_options *o)
2538 {
2539 mmfile_t mf1, mf2;
2540 struct checkdiff_t data;
2541
2542 if (!two)
2543 return;
2544
2545 memset(&data, 0, sizeof(data));
2546 data.filename = name_b ? name_b : name_a;
2547 data.lineno = 0;
2548 data.o = o;
2549 data.ws_rule = whitespace_rule(attr_path);
2550 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2551
2552 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2553 die("unable to read files to diff");
2554
2555 /*
2556 * All the other codepaths check both sides, but not checking
2557 * the "old" side here is deliberate. We are checking the newly
2558 * introduced changes, and as long as the "new" side is text, we
2559 * can and should check what it introduces.
2560 */
2561 if (diff_filespec_is_binary(two))
2562 goto free_and_return;
2563 else {
2564 /* Crazy xdl interfaces.. */
2565 xpparam_t xpp;
2566 xdemitconf_t xecfg;
2567
2568 memset(&xpp, 0, sizeof(xpp));
2569 memset(&xecfg, 0, sizeof(xecfg));
2570 xecfg.ctxlen = 1; /* at least one context line */
2571 xpp.flags = 0;
2572 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2573 &xpp, &xecfg))
2574 die("unable to generate checkdiff for %s", one->path);
2575
2576 if (data.ws_rule & WS_BLANK_AT_EOF) {
2577 struct emit_callback ecbdata;
2578 int blank_at_eof;
2579
2580 ecbdata.ws_rule = data.ws_rule;
2581 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2582 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2583
2584 if (blank_at_eof) {
2585 static char *err;
2586 if (!err)
2587 err = whitespace_error_string(WS_BLANK_AT_EOF);
2588 fprintf(o->file, "%s:%d: %s.\n",
2589 data.filename, blank_at_eof, err);
2590 data.status = 1; /* report errors */
2591 }
2592 }
2593 }
2594 free_and_return:
2595 diff_free_filespec_data(one);
2596 diff_free_filespec_data(two);
2597 if (data.status)
2598 DIFF_OPT_SET(o, CHECK_FAILED);
2599 }
2600
2601 struct diff_filespec *alloc_filespec(const char *path)
2602 {
2603 int namelen = strlen(path);
2604 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2605
2606 memset(spec, 0, sizeof(*spec));
2607 spec->path = (char *)(spec + 1);
2608 memcpy(spec->path, path, namelen+1);
2609 spec->count = 1;
2610 spec->is_binary = -1;
2611 return spec;
2612 }
2613
2614 void free_filespec(struct diff_filespec *spec)
2615 {
2616 if (!--spec->count) {
2617 diff_free_filespec_data(spec);
2618 free(spec);
2619 }
2620 }
2621
2622 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2623 int sha1_valid, unsigned short mode)
2624 {
2625 if (mode) {
2626 spec->mode = canon_mode(mode);
2627 hashcpy(spec->sha1, sha1);
2628 spec->sha1_valid = sha1_valid;
2629 }
2630 }
2631
2632 /*
2633 * Given a name and sha1 pair, if the index tells us the file in
2634 * the work tree has that object contents, return true, so that
2635 * prepare_temp_file() does not have to inflate and extract.
2636 */
2637 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2638 {
2639 const struct cache_entry *ce;
2640 struct stat st;
2641 int pos, len;
2642
2643 /*
2644 * We do not read the cache ourselves here, because the
2645 * benchmark with my previous version that always reads cache
2646 * shows that it makes things worse for diff-tree comparing
2647 * two linux-2.6 kernel trees in an already checked out work
2648 * tree. This is because most diff-tree comparisons deal with
2649 * only a small number of files, while reading the cache is
2650 * expensive for a large project, and its cost outweighs the
2651 * savings we get by not inflating the object to a temporary
2652 * file. Practically, this code only helps when we are used
2653 * by diff-cache --cached, which does read the cache before
2654 * calling us.
2655 */
2656 if (!active_cache)
2657 return 0;
2658
2659 /* We want to avoid the working directory if our caller
2660 * doesn't need the data in a normal file, this system
2661 * is rather slow with its stat/open/mmap/close syscalls,
2662 * and the object is contained in a pack file. The pack
2663 * is probably already open and will be faster to obtain
2664 * the data through than the working directory. Loose
2665 * objects however would tend to be slower as they need
2666 * to be individually opened and inflated.
2667 */
2668 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2669 return 0;
2670
2671 len = strlen(name);
2672 pos = cache_name_pos(name, len);
2673 if (pos < 0)
2674 return 0;
2675 ce = active_cache[pos];
2676
2677 /*
2678 * This is not the sha1 we are looking for, or
2679 * unreusable because it is not a regular file.
2680 */
2681 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2682 return 0;
2683
2684 /*
2685 * If ce is marked as "assume unchanged", there is no
2686 * guarantee that work tree matches what we are looking for.
2687 */
2688 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2689 return 0;
2690
2691 /*
2692 * If ce matches the file in the work tree, we can reuse it.
2693 */
2694 if (ce_uptodate(ce) ||
2695 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2696 return 1;
2697
2698 return 0;
2699 }
2700
2701 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2702 {
2703 int len;
2704 char *data = xmalloc(100), *dirty = "";
2705
2706 /* Are we looking at the work tree? */
2707 if (s->dirty_submodule)
2708 dirty = "-dirty";
2709
2710 len = snprintf(data, 100,
2711 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2712 s->data = data;
2713 s->size = len;
2714 s->should_free = 1;
2715 if (size_only) {
2716 s->data = NULL;
2717 free(data);
2718 }
2719 return 0;
2720 }
2721
2722 /*
2723 * While doing rename detection and pickaxe operation, we may need to
2724 * grab the data for the blob (or file) for our own in-core comparison.
2725 * diff_filespec has data and size fields for this purpose.
2726 */
2727 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
2728 {
2729 int size_only = flags & CHECK_SIZE_ONLY;
2730 int err = 0;
2731 /*
2732 * demote FAIL to WARN to allow inspecting the situation
2733 * instead of refusing.
2734 */
2735 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
2736 ? SAFE_CRLF_WARN
2737 : safe_crlf);
2738
2739 if (!DIFF_FILE_VALID(s))
2740 die("internal error: asking to populate invalid file.");
2741 if (S_ISDIR(s->mode))
2742 return -1;
2743
2744 if (s->data)
2745 return 0;
2746
2747 if (size_only && 0 < s->size)
2748 return 0;
2749
2750 if (S_ISGITLINK(s->mode))
2751 return diff_populate_gitlink(s, size_only);
2752
2753 if (!s->sha1_valid ||
2754 reuse_worktree_file(s->path, s->sha1, 0)) {
2755 struct strbuf buf = STRBUF_INIT;
2756 struct stat st;
2757 int fd;
2758
2759 if (lstat(s->path, &st) < 0) {
2760 if (errno == ENOENT) {
2761 err_empty:
2762 err = -1;
2763 empty:
2764 s->data = (char *)"";
2765 s->size = 0;
2766 return err;
2767 }
2768 }
2769 s->size = xsize_t(st.st_size);
2770 if (!s->size)
2771 goto empty;
2772 if (S_ISLNK(st.st_mode)) {
2773 struct strbuf sb = STRBUF_INIT;
2774
2775 if (strbuf_readlink(&sb, s->path, s->size))
2776 goto err_empty;
2777 s->size = sb.len;
2778 s->data = strbuf_detach(&sb, NULL);
2779 s->should_free = 1;
2780 return 0;
2781 }
2782 if (size_only)
2783 return 0;
2784 if ((flags & CHECK_BINARY) &&
2785 s->size > big_file_threshold && s->is_binary == -1) {
2786 s->is_binary = 1;
2787 return 0;
2788 }
2789 fd = open(s->path, O_RDONLY);
2790 if (fd < 0)
2791 goto err_empty;
2792 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2793 close(fd);
2794 s->should_munmap = 1;
2795
2796 /*
2797 * Convert from working tree format to canonical git format
2798 */
2799 if (convert_to_git(s->path, s->data, s->size, &buf, crlf_warn)) {
2800 size_t size = 0;
2801 munmap(s->data, s->size);
2802 s->should_munmap = 0;
2803 s->data = strbuf_detach(&buf, &size);
2804 s->size = size;
2805 s->should_free = 1;
2806 }
2807 }
2808 else {
2809 enum object_type type;
2810 if (size_only || (flags & CHECK_BINARY)) {
2811 type = sha1_object_info(s->sha1, &s->size);
2812 if (type < 0)
2813 die("unable to read %s", sha1_to_hex(s->sha1));
2814 if (size_only)
2815 return 0;
2816 if (s->size > big_file_threshold && s->is_binary == -1) {
2817 s->is_binary = 1;
2818 return 0;
2819 }
2820 }
2821 s->data = read_sha1_file(s->sha1, &type, &s->size);
2822 if (!s->data)
2823 die("unable to read %s", sha1_to_hex(s->sha1));
2824 s->should_free = 1;
2825 }
2826 return 0;
2827 }
2828
2829 void diff_free_filespec_blob(struct diff_filespec *s)
2830 {
2831 if (s->should_free)
2832 free(s->data);
2833 else if (s->should_munmap)
2834 munmap(s->data, s->size);
2835
2836 if (s->should_free || s->should_munmap) {
2837 s->should_free = s->should_munmap = 0;
2838 s->data = NULL;
2839 }
2840 }
2841
2842 void diff_free_filespec_data(struct diff_filespec *s)
2843 {
2844 diff_free_filespec_blob(s);
2845 free(s->cnt_data);
2846 s->cnt_data = NULL;
2847 }
2848
2849 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2850 void *blob,
2851 unsigned long size,
2852 const unsigned char *sha1,
2853 int mode)
2854 {
2855 int fd;
2856 struct strbuf buf = STRBUF_INIT;
2857 struct strbuf template = STRBUF_INIT;
2858 char *path_dup = xstrdup(path);
2859 const char *base = basename(path_dup);
2860
2861 /* Generate "XXXXXX_basename.ext" */
2862 strbuf_addstr(&template, "XXXXXX_");
2863 strbuf_addstr(&template, base);
2864
2865 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2866 strlen(base) + 1);
2867 if (fd < 0)
2868 die_errno("unable to create temp-file");
2869 if (convert_to_working_tree(path,
2870 (const char *)blob, (size_t)size, &buf)) {
2871 blob = buf.buf;
2872 size = buf.len;
2873 }
2874 if (write_in_full(fd, blob, size) != size)
2875 die_errno("unable to write temp-file");
2876 close(fd);
2877 temp->name = temp->tmp_path;
2878 strcpy(temp->hex, sha1_to_hex(sha1));
2879 temp->hex[40] = 0;
2880 sprintf(temp->mode, "%06o", mode);
2881 strbuf_release(&buf);
2882 strbuf_release(&template);
2883 free(path_dup);
2884 }
2885
2886 static struct diff_tempfile *prepare_temp_file(const char *name,
2887 struct diff_filespec *one)
2888 {
2889 struct diff_tempfile *temp = claim_diff_tempfile();
2890
2891 if (!DIFF_FILE_VALID(one)) {
2892 not_a_valid_file:
2893 /* A '-' entry produces this for file-2, and
2894 * a '+' entry produces this for file-1.
2895 */
2896 temp->name = "/dev/null";
2897 strcpy(temp->hex, ".");
2898 strcpy(temp->mode, ".");
2899 return temp;
2900 }
2901
2902 if (!remove_tempfile_installed) {
2903 atexit(remove_tempfile);
2904 sigchain_push_common(remove_tempfile_on_signal);
2905 remove_tempfile_installed = 1;
2906 }
2907
2908 if (!S_ISGITLINK(one->mode) &&
2909 (!one->sha1_valid ||
2910 reuse_worktree_file(name, one->sha1, 1))) {
2911 struct stat st;
2912 if (lstat(name, &st) < 0) {
2913 if (errno == ENOENT)
2914 goto not_a_valid_file;
2915 die_errno("stat(%s)", name);
2916 }
2917 if (S_ISLNK(st.st_mode)) {
2918 struct strbuf sb = STRBUF_INIT;
2919 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2920 die_errno("readlink(%s)", name);
2921 prep_temp_blob(name, temp, sb.buf, sb.len,
2922 (one->sha1_valid ?
2923 one->sha1 : null_sha1),
2924 (one->sha1_valid ?
2925 one->mode : S_IFLNK));
2926 strbuf_release(&sb);
2927 }
2928 else {
2929 /* we can borrow from the file in the work tree */
2930 temp->name = name;
2931 if (!one->sha1_valid)
2932 strcpy(temp->hex, sha1_to_hex(null_sha1));
2933 else
2934 strcpy(temp->hex, sha1_to_hex(one->sha1));
2935 /* Even though we may sometimes borrow the
2936 * contents from the work tree, we always want
2937 * one->mode. mode is trustworthy even when
2938 * !(one->sha1_valid), as long as
2939 * DIFF_FILE_VALID(one).
2940 */
2941 sprintf(temp->mode, "%06o", one->mode);
2942 }
2943 return temp;
2944 }
2945 else {
2946 if (diff_populate_filespec(one, 0))
2947 die("cannot read data blob for %s", one->path);
2948 prep_temp_blob(name, temp, one->data, one->size,
2949 one->sha1, one->mode);
2950 }
2951 return temp;
2952 }
2953
2954 static void add_external_diff_name(struct argv_array *argv,
2955 const char *name,
2956 struct diff_filespec *df)
2957 {
2958 struct diff_tempfile *temp = prepare_temp_file(name, df);
2959 argv_array_push(argv, temp->name);
2960 argv_array_push(argv, temp->hex);
2961 argv_array_push(argv, temp->mode);
2962 }
2963
2964 /* An external diff command takes:
2965 *
2966 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2967 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2968 *
2969 */
2970 static void run_external_diff(const char *pgm,
2971 const char *name,
2972 const char *other,
2973 struct diff_filespec *one,
2974 struct diff_filespec *two,
2975 const char *xfrm_msg,
2976 int complete_rewrite,
2977 struct diff_options *o)
2978 {
2979 struct argv_array argv = ARGV_ARRAY_INIT;
2980 struct argv_array env = ARGV_ARRAY_INIT;
2981 struct diff_queue_struct *q = &diff_queued_diff;
2982
2983 argv_array_push(&argv, pgm);
2984 argv_array_push(&argv, name);
2985
2986 if (one && two) {
2987 add_external_diff_name(&argv, name, one);
2988 if (!other)
2989 add_external_diff_name(&argv, name, two);
2990 else {
2991 add_external_diff_name(&argv, other, two);
2992 argv_array_push(&argv, other);
2993 argv_array_push(&argv, xfrm_msg);
2994 }
2995 }
2996
2997 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
2998 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
2999
3000 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3001 die(_("external diff died, stopping at %s"), name);
3002
3003 remove_tempfile();
3004 argv_array_clear(&argv);
3005 argv_array_clear(&env);
3006 }
3007
3008 static int similarity_index(struct diff_filepair *p)
3009 {
3010 return p->score * 100 / MAX_SCORE;
3011 }
3012
3013 static void fill_metainfo(struct strbuf *msg,
3014 const char *name,
3015 const char *other,
3016 struct diff_filespec *one,
3017 struct diff_filespec *two,
3018 struct diff_options *o,
3019 struct diff_filepair *p,
3020 int *must_show_header,
3021 int use_color)
3022 {
3023 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3024 const char *reset = diff_get_color(use_color, DIFF_RESET);
3025 const char *line_prefix = diff_line_prefix(o);
3026
3027 *must_show_header = 1;
3028 strbuf_init(msg, PATH_MAX * 2 + 300);
3029 switch (p->status) {
3030 case DIFF_STATUS_COPIED:
3031 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3032 line_prefix, set, similarity_index(p));
3033 strbuf_addf(msg, "%s\n%s%scopy from ",
3034 reset, line_prefix, set);
3035 quote_c_style(name, msg, NULL, 0);
3036 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3037 quote_c_style(other, msg, NULL, 0);
3038 strbuf_addf(msg, "%s\n", reset);
3039 break;
3040 case DIFF_STATUS_RENAMED:
3041 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3042 line_prefix, set, similarity_index(p));
3043 strbuf_addf(msg, "%s\n%s%srename from ",
3044 reset, line_prefix, set);
3045 quote_c_style(name, msg, NULL, 0);
3046 strbuf_addf(msg, "%s\n%s%srename to ",
3047 reset, line_prefix, set);
3048 quote_c_style(other, msg, NULL, 0);
3049 strbuf_addf(msg, "%s\n", reset);
3050 break;
3051 case DIFF_STATUS_MODIFIED:
3052 if (p->score) {
3053 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3054 line_prefix,
3055 set, similarity_index(p), reset);
3056 break;
3057 }
3058 /* fallthru */
3059 default:
3060 *must_show_header = 0;
3061 }
3062 if (one && two && hashcmp(one->sha1, two->sha1)) {
3063 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3064
3065 if (DIFF_OPT_TST(o, BINARY)) {
3066 mmfile_t mf;
3067 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3068 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3069 abbrev = 40;
3070 }
3071 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
3072 find_unique_abbrev(one->sha1, abbrev));
3073 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
3074 if (one->mode == two->mode)
3075 strbuf_addf(msg, " %06o", one->mode);
3076 strbuf_addf(msg, "%s\n", reset);
3077 }
3078 }
3079
3080 static void run_diff_cmd(const char *pgm,
3081 const char *name,
3082 const char *other,
3083 const char *attr_path,
3084 struct diff_filespec *one,
3085 struct diff_filespec *two,
3086 struct strbuf *msg,
3087 struct diff_options *o,
3088 struct diff_filepair *p)
3089 {
3090 const char *xfrm_msg = NULL;
3091 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3092 int must_show_header = 0;
3093
3094
3095 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3096 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3097 if (drv && drv->external)
3098 pgm = drv->external;
3099 }
3100
3101 if (msg) {
3102 /*
3103 * don't use colors when the header is intended for an
3104 * external diff driver
3105 */
3106 fill_metainfo(msg, name, other, one, two, o, p,
3107 &must_show_header,
3108 want_color(o->use_color) && !pgm);
3109 xfrm_msg = msg->len ? msg->buf : NULL;
3110 }
3111
3112 if (pgm) {
3113 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3114 complete_rewrite, o);
3115 return;
3116 }
3117 if (one && two)
3118 builtin_diff(name, other ? other : name,
3119 one, two, xfrm_msg, must_show_header,
3120 o, complete_rewrite);
3121 else
3122 fprintf(o->file, "* Unmerged path %s\n", name);
3123 }
3124
3125 static void diff_fill_sha1_info(struct diff_filespec *one)
3126 {
3127 if (DIFF_FILE_VALID(one)) {
3128 if (!one->sha1_valid) {
3129 struct stat st;
3130 if (one->is_stdin) {
3131 hashcpy(one->sha1, null_sha1);
3132 return;
3133 }
3134 if (lstat(one->path, &st) < 0)
3135 die_errno("stat '%s'", one->path);
3136 if (index_path(one->sha1, one->path, &st, 0))
3137 die("cannot hash %s", one->path);
3138 }
3139 }
3140 else
3141 hashclr(one->sha1);
3142 }
3143
3144 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3145 {
3146 /* Strip the prefix but do not molest /dev/null and absolute paths */
3147 if (*namep && **namep != '/') {
3148 *namep += prefix_length;
3149 if (**namep == '/')
3150 ++*namep;
3151 }
3152 if (*otherp && **otherp != '/') {
3153 *otherp += prefix_length;
3154 if (**otherp == '/')
3155 ++*otherp;
3156 }
3157 }
3158
3159 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3160 {
3161 const char *pgm = external_diff();
3162 struct strbuf msg;
3163 struct diff_filespec *one = p->one;
3164 struct diff_filespec *two = p->two;
3165 const char *name;
3166 const char *other;
3167 const char *attr_path;
3168
3169 name = p->one->path;
3170 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3171 attr_path = name;
3172 if (o->prefix_length)
3173 strip_prefix(o->prefix_length, &name, &other);
3174
3175 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3176 pgm = NULL;
3177
3178 if (DIFF_PAIR_UNMERGED(p)) {
3179 run_diff_cmd(pgm, name, NULL, attr_path,
3180 NULL, NULL, NULL, o, p);
3181 return;
3182 }
3183
3184 diff_fill_sha1_info(one);
3185 diff_fill_sha1_info(two);
3186
3187 if (!pgm &&
3188 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3189 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3190 /*
3191 * a filepair that changes between file and symlink
3192 * needs to be split into deletion and creation.
3193 */
3194 struct diff_filespec *null = alloc_filespec(two->path);
3195 run_diff_cmd(NULL, name, other, attr_path,
3196 one, null, &msg, o, p);
3197 free(null);
3198 strbuf_release(&msg);
3199
3200 null = alloc_filespec(one->path);
3201 run_diff_cmd(NULL, name, other, attr_path,
3202 null, two, &msg, o, p);
3203 free(null);
3204 }
3205 else
3206 run_diff_cmd(pgm, name, other, attr_path,
3207 one, two, &msg, o, p);
3208
3209 strbuf_release(&msg);
3210 }
3211
3212 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3213 struct diffstat_t *diffstat)
3214 {
3215 const char *name;
3216 const char *other;
3217
3218 if (DIFF_PAIR_UNMERGED(p)) {
3219 /* unmerged */
3220 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3221 return;
3222 }
3223
3224 name = p->one->path;
3225 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3226
3227 if (o->prefix_length)
3228 strip_prefix(o->prefix_length, &name, &other);
3229
3230 diff_fill_sha1_info(p->one);
3231 diff_fill_sha1_info(p->two);
3232
3233 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3234 }
3235
3236 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3237 {
3238 const char *name;
3239 const char *other;
3240 const char *attr_path;
3241
3242 if (DIFF_PAIR_UNMERGED(p)) {
3243 /* unmerged */
3244 return;
3245 }
3246
3247 name = p->one->path;
3248 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3249 attr_path = other ? other : name;
3250
3251 if (o->prefix_length)
3252 strip_prefix(o->prefix_length, &name, &other);
3253
3254 diff_fill_sha1_info(p->one);
3255 diff_fill_sha1_info(p->two);
3256
3257 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3258 }
3259
3260 void diff_setup(struct diff_options *options)
3261 {
3262 memcpy(options, &default_diff_options, sizeof(*options));
3263
3264 options->file = stdout;
3265
3266 options->line_termination = '\n';
3267 options->break_opt = -1;
3268 options->rename_limit = -1;
3269 options->dirstat_permille = diff_dirstat_permille_default;
3270 options->context = diff_context_default;
3271 options->ws_error_highlight = WSEH_NEW;
3272 DIFF_OPT_SET(options, RENAME_EMPTY);
3273
3274 /* pathchange left =NULL by default */
3275 options->change = diff_change;
3276 options->add_remove = diff_addremove;
3277 options->use_color = diff_use_color_default;
3278 options->detect_rename = diff_detect_rename_default;
3279 options->xdl_opts |= diff_algorithm;
3280
3281 options->orderfile = diff_order_file_cfg;
3282
3283 if (diff_no_prefix) {
3284 options->a_prefix = options->b_prefix = "";
3285 } else if (!diff_mnemonic_prefix) {
3286 options->a_prefix = "a/";
3287 options->b_prefix = "b/";
3288 }
3289 }
3290
3291 void diff_setup_done(struct diff_options *options)
3292 {
3293 int count = 0;
3294
3295 if (options->set_default)
3296 options->set_default(options);
3297
3298 if (options->output_format & DIFF_FORMAT_NAME)
3299 count++;
3300 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3301 count++;
3302 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3303 count++;
3304 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3305 count++;
3306 if (count > 1)
3307 die("--name-only, --name-status, --check and -s are mutually exclusive");
3308
3309 /*
3310 * Most of the time we can say "there are changes"
3311 * only by checking if there are changed paths, but
3312 * --ignore-whitespace* options force us to look
3313 * inside contents.
3314 */
3315
3316 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3317 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3318 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3319 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3320 else
3321 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3322
3323 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3324 options->detect_rename = DIFF_DETECT_COPY;
3325
3326 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3327 options->prefix = NULL;
3328 if (options->prefix)
3329 options->prefix_length = strlen(options->prefix);
3330 else
3331 options->prefix_length = 0;
3332
3333 if (options->output_format & (DIFF_FORMAT_NAME |
3334 DIFF_FORMAT_NAME_STATUS |
3335 DIFF_FORMAT_CHECKDIFF |
3336 DIFF_FORMAT_NO_OUTPUT))
3337 options->output_format &= ~(DIFF_FORMAT_RAW |
3338 DIFF_FORMAT_NUMSTAT |
3339 DIFF_FORMAT_DIFFSTAT |
3340 DIFF_FORMAT_SHORTSTAT |
3341 DIFF_FORMAT_DIRSTAT |
3342 DIFF_FORMAT_SUMMARY |
3343 DIFF_FORMAT_PATCH);
3344
3345 /*
3346 * These cases always need recursive; we do not drop caller-supplied
3347 * recursive bits for other formats here.
3348 */
3349 if (options->output_format & (DIFF_FORMAT_PATCH |
3350 DIFF_FORMAT_NUMSTAT |
3351 DIFF_FORMAT_DIFFSTAT |
3352 DIFF_FORMAT_SHORTSTAT |
3353 DIFF_FORMAT_DIRSTAT |
3354 DIFF_FORMAT_SUMMARY |
3355 DIFF_FORMAT_CHECKDIFF))
3356 DIFF_OPT_SET(options, RECURSIVE);
3357 /*
3358 * Also pickaxe would not work very well if you do not say recursive
3359 */
3360 if (options->pickaxe)
3361 DIFF_OPT_SET(options, RECURSIVE);
3362 /*
3363 * When patches are generated, submodules diffed against the work tree
3364 * must be checked for dirtiness too so it can be shown in the output
3365 */
3366 if (options->output_format & DIFF_FORMAT_PATCH)
3367 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3368
3369 if (options->detect_rename && options->rename_limit < 0)
3370 options->rename_limit = diff_rename_limit_default;
3371 if (options->setup & DIFF_SETUP_USE_CACHE) {
3372 if (!active_cache)
3373 /* read-cache does not die even when it fails
3374 * so it is safe for us to do this here. Also
3375 * it does not smudge active_cache or active_nr
3376 * when it fails, so we do not have to worry about
3377 * cleaning it up ourselves either.
3378 */
3379 read_cache();
3380 }
3381 if (options->abbrev <= 0 || 40 < options->abbrev)
3382 options->abbrev = 40; /* full */
3383
3384 /*
3385 * It does not make sense to show the first hit we happened
3386 * to have found. It does not make sense not to return with
3387 * exit code in such a case either.
3388 */
3389 if (DIFF_OPT_TST(options, QUICK)) {
3390 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3391 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3392 }
3393
3394 options->diff_path_counter = 0;
3395
3396 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3397 die(_("--follow requires exactly one pathspec"));
3398 }
3399
3400 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3401 {
3402 char c, *eq;
3403 int len;
3404
3405 if (*arg != '-')
3406 return 0;
3407 c = *++arg;
3408 if (!c)
3409 return 0;
3410 if (c == arg_short) {
3411 c = *++arg;
3412 if (!c)
3413 return 1;
3414 if (val && isdigit(c)) {
3415 char *end;
3416 int n = strtoul(arg, &end, 10);
3417 if (*end)
3418 return 0;
3419 *val = n;
3420 return 1;
3421 }
3422 return 0;
3423 }
3424 if (c != '-')
3425 return 0;
3426 arg++;
3427 eq = strchrnul(arg, '=');
3428 len = eq - arg;
3429 if (!len || strncmp(arg, arg_long, len))
3430 return 0;
3431 if (*eq) {
3432 int n;
3433 char *end;
3434 if (!isdigit(*++eq))
3435 return 0;
3436 n = strtoul(eq, &end, 10);
3437 if (*end)
3438 return 0;
3439 *val = n;
3440 }
3441 return 1;
3442 }
3443
3444 static int diff_scoreopt_parse(const char *opt);
3445
3446 static inline int short_opt(char opt, const char **argv,
3447 const char **optarg)
3448 {
3449 const char *arg = argv[0];
3450 if (arg[0] != '-' || arg[1] != opt)
3451 return 0;
3452 if (arg[2] != '\0') {
3453 *optarg = arg + 2;
3454 return 1;
3455 }
3456 if (!argv[1])
3457 die("Option '%c' requires a value", opt);
3458 *optarg = argv[1];
3459 return 2;
3460 }
3461
3462 int parse_long_opt(const char *opt, const char **argv,
3463 const char **optarg)
3464 {
3465 const char *arg = argv[0];
3466 if (!skip_prefix(arg, "--", &arg))
3467 return 0;
3468 if (!skip_prefix(arg, opt, &arg))
3469 return 0;
3470 if (*arg == '=') { /* stuck form: --option=value */
3471 *optarg = arg + 1;
3472 return 1;
3473 }
3474 if (*arg != '\0')
3475 return 0;
3476 /* separate form: --option value */
3477 if (!argv[1])
3478 die("Option '--%s' requires a value", opt);
3479 *optarg = argv[1];
3480 return 2;
3481 }
3482
3483 static int stat_opt(struct diff_options *options, const char **av)
3484 {
3485 const char *arg = av[0];
3486 char *end;
3487 int width = options->stat_width;
3488 int name_width = options->stat_name_width;
3489 int graph_width = options->stat_graph_width;
3490 int count = options->stat_count;
3491 int argcount = 1;
3492
3493 if (!skip_prefix(arg, "--stat", &arg))
3494 die("BUG: stat option does not begin with --stat: %s", arg);
3495 end = (char *)arg;
3496
3497 switch (*arg) {
3498 case '-':
3499 if (skip_prefix(arg, "-width", &arg)) {
3500 if (*arg == '=')
3501 width = strtoul(arg + 1, &end, 10);
3502 else if (!*arg && !av[1])
3503 die("Option '--stat-width' requires a value");
3504 else if (!*arg) {
3505 width = strtoul(av[1], &end, 10);
3506 argcount = 2;
3507 }
3508 } else if (skip_prefix(arg, "-name-width", &arg)) {
3509 if (*arg == '=')
3510 name_width = strtoul(arg + 1, &end, 10);
3511 else if (!*arg && !av[1])
3512 die("Option '--stat-name-width' requires a value");
3513 else if (!*arg) {
3514 name_width = strtoul(av[1], &end, 10);
3515 argcount = 2;
3516 }
3517 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3518 if (*arg == '=')
3519 graph_width = strtoul(arg + 1, &end, 10);
3520 else if (!*arg && !av[1])
3521 die("Option '--stat-graph-width' requires a value");
3522 else if (!*arg) {
3523 graph_width = strtoul(av[1], &end, 10);
3524 argcount = 2;
3525 }
3526 } else if (skip_prefix(arg, "-count", &arg)) {
3527 if (*arg == '=')
3528 count = strtoul(arg + 1, &end, 10);
3529 else if (!*arg && !av[1])
3530 die("Option '--stat-count' requires a value");
3531 else if (!*arg) {
3532 count = strtoul(av[1], &end, 10);
3533 argcount = 2;
3534 }
3535 }
3536 break;
3537 case '=':
3538 width = strtoul(arg+1, &end, 10);
3539 if (*end == ',')
3540 name_width = strtoul(end+1, &end, 10);
3541 if (*end == ',')
3542 count = strtoul(end+1, &end, 10);
3543 }
3544
3545 /* Important! This checks all the error cases! */
3546 if (*end)
3547 return 0;
3548 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3549 options->stat_name_width = name_width;
3550 options->stat_graph_width = graph_width;
3551 options->stat_width = width;
3552 options->stat_count = count;
3553 return argcount;
3554 }
3555
3556 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3557 {
3558 struct strbuf errmsg = STRBUF_INIT;
3559 if (parse_dirstat_params(options, params, &errmsg))
3560 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3561 errmsg.buf);
3562 strbuf_release(&errmsg);
3563 /*
3564 * The caller knows a dirstat-related option is given from the command
3565 * line; allow it to say "return this_function();"
3566 */
3567 options->output_format |= DIFF_FORMAT_DIRSTAT;
3568 return 1;
3569 }
3570
3571 static int parse_submodule_opt(struct diff_options *options, const char *value)
3572 {
3573 if (parse_submodule_params(options, value))
3574 die(_("Failed to parse --submodule option parameter: '%s'"),
3575 value);
3576 return 1;
3577 }
3578
3579 static const char diff_status_letters[] = {
3580 DIFF_STATUS_ADDED,
3581 DIFF_STATUS_COPIED,
3582 DIFF_STATUS_DELETED,
3583 DIFF_STATUS_MODIFIED,
3584 DIFF_STATUS_RENAMED,
3585 DIFF_STATUS_TYPE_CHANGED,
3586 DIFF_STATUS_UNKNOWN,
3587 DIFF_STATUS_UNMERGED,
3588 DIFF_STATUS_FILTER_AON,
3589 DIFF_STATUS_FILTER_BROKEN,
3590 '\0',
3591 };
3592
3593 static unsigned int filter_bit['Z' + 1];
3594
3595 static void prepare_filter_bits(void)
3596 {
3597 int i;
3598
3599 if (!filter_bit[DIFF_STATUS_ADDED]) {
3600 for (i = 0; diff_status_letters[i]; i++)
3601 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3602 }
3603 }
3604
3605 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3606 {
3607 return opt->filter & filter_bit[(int) status];
3608 }
3609
3610 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3611 {
3612 int i, optch;
3613
3614 prepare_filter_bits();
3615
3616 /*
3617 * If there is a negation e.g. 'd' in the input, and we haven't
3618 * initialized the filter field with another --diff-filter, start
3619 * from full set of bits, except for AON.
3620 */
3621 if (!opt->filter) {
3622 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3623 if (optch < 'a' || 'z' < optch)
3624 continue;
3625 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3626 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3627 break;
3628 }
3629 }
3630
3631 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3632 unsigned int bit;
3633 int negate;
3634
3635 if ('a' <= optch && optch <= 'z') {
3636 negate = 1;
3637 optch = toupper(optch);
3638 } else {
3639 negate = 0;
3640 }
3641
3642 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
3643 if (!bit)
3644 return optarg[i];
3645 if (negate)
3646 opt->filter &= ~bit;
3647 else
3648 opt->filter |= bit;
3649 }
3650 return 0;
3651 }
3652
3653 static void enable_patch_output(int *fmt) {
3654 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
3655 *fmt |= DIFF_FORMAT_PATCH;
3656 }
3657
3658 static int parse_one_token(const char **arg, const char *token)
3659 {
3660 const char *rest;
3661 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
3662 *arg = rest;
3663 return 1;
3664 }
3665 return 0;
3666 }
3667
3668 static int parse_ws_error_highlight(struct diff_options *opt, const char *arg)
3669 {
3670 const char *orig_arg = arg;
3671 unsigned val = 0;
3672 while (*arg) {
3673 if (parse_one_token(&arg, "none"))
3674 val = 0;
3675 else if (parse_one_token(&arg, "default"))
3676 val = WSEH_NEW;
3677 else if (parse_one_token(&arg, "all"))
3678 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
3679 else if (parse_one_token(&arg, "new"))
3680 val |= WSEH_NEW;
3681 else if (parse_one_token(&arg, "old"))
3682 val |= WSEH_OLD;
3683 else if (parse_one_token(&arg, "context"))
3684 val |= WSEH_CONTEXT;
3685 else {
3686 error("unknown value after ws-error-highlight=%.*s",
3687 (int)(arg - orig_arg), orig_arg);
3688 return 0;
3689 }
3690 if (*arg)
3691 arg++;
3692 }
3693 opt->ws_error_highlight = val;
3694 return 1;
3695 }
3696
3697 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3698 {
3699 const char *arg = av[0];
3700 const char *optarg;
3701 int argcount;
3702
3703 /* Output format options */
3704 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
3705 || opt_arg(arg, 'U', "unified", &options->context))
3706 enable_patch_output(&options->output_format);
3707 else if (!strcmp(arg, "--raw"))
3708 options->output_format |= DIFF_FORMAT_RAW;
3709 else if (!strcmp(arg, "--patch-with-raw")) {
3710 enable_patch_output(&options->output_format);
3711 options->output_format |= DIFF_FORMAT_RAW;
3712 } else if (!strcmp(arg, "--numstat"))
3713 options->output_format |= DIFF_FORMAT_NUMSTAT;
3714 else if (!strcmp(arg, "--shortstat"))
3715 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3716 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3717 return parse_dirstat_opt(options, "");
3718 else if (skip_prefix(arg, "-X", &arg))
3719 return parse_dirstat_opt(options, arg);
3720 else if (skip_prefix(arg, "--dirstat=", &arg))
3721 return parse_dirstat_opt(options, arg);
3722 else if (!strcmp(arg, "--cumulative"))
3723 return parse_dirstat_opt(options, "cumulative");
3724 else if (!strcmp(arg, "--dirstat-by-file"))
3725 return parse_dirstat_opt(options, "files");
3726 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
3727 parse_dirstat_opt(options, "files");
3728 return parse_dirstat_opt(options, arg);
3729 }
3730 else if (!strcmp(arg, "--check"))
3731 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3732 else if (!strcmp(arg, "--summary"))
3733 options->output_format |= DIFF_FORMAT_SUMMARY;
3734 else if (!strcmp(arg, "--patch-with-stat")) {
3735 enable_patch_output(&options->output_format);
3736 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3737 } else if (!strcmp(arg, "--name-only"))
3738 options->output_format |= DIFF_FORMAT_NAME;
3739 else if (!strcmp(arg, "--name-status"))
3740 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3741 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
3742 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3743 else if (starts_with(arg, "--stat"))
3744 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3745 return stat_opt(options, av);
3746
3747 /* renames options */
3748 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
3749 !strcmp(arg, "--break-rewrites")) {
3750 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3751 return error("invalid argument to -B: %s", arg+2);
3752 }
3753 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
3754 !strcmp(arg, "--find-renames")) {
3755 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3756 return error("invalid argument to -M: %s", arg+2);
3757 options->detect_rename = DIFF_DETECT_RENAME;
3758 }
3759 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3760 options->irreversible_delete = 1;
3761 }
3762 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
3763 !strcmp(arg, "--find-copies")) {
3764 if (options->detect_rename == DIFF_DETECT_COPY)
3765 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3766 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3767 return error("invalid argument to -C: %s", arg+2);
3768 options->detect_rename = DIFF_DETECT_COPY;
3769 }
3770 else if (!strcmp(arg, "--no-renames"))
3771 options->detect_rename = 0;
3772 else if (!strcmp(arg, "--rename-empty"))
3773 DIFF_OPT_SET(options, RENAME_EMPTY);
3774 else if (!strcmp(arg, "--no-rename-empty"))
3775 DIFF_OPT_CLR(options, RENAME_EMPTY);
3776 else if (!strcmp(arg, "--relative"))
3777 DIFF_OPT_SET(options, RELATIVE_NAME);
3778 else if (skip_prefix(arg, "--relative=", &arg)) {
3779 DIFF_OPT_SET(options, RELATIVE_NAME);
3780 options->prefix = arg;
3781 }
3782
3783 /* xdiff options */
3784 else if (!strcmp(arg, "--minimal"))
3785 DIFF_XDL_SET(options, NEED_MINIMAL);
3786 else if (!strcmp(arg, "--no-minimal"))
3787 DIFF_XDL_CLR(options, NEED_MINIMAL);
3788 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3789 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3790 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3791 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3792 else if (!strcmp(arg, "--ignore-space-at-eol"))
3793 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3794 else if (!strcmp(arg, "--ignore-blank-lines"))
3795 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
3796 else if (!strcmp(arg, "--patience"))
3797 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3798 else if (!strcmp(arg, "--histogram"))
3799 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3800 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
3801 long value = parse_algorithm_value(optarg);
3802 if (value < 0)
3803 return error("option diff-algorithm accepts \"myers\", "
3804 "\"minimal\", \"patience\" and \"histogram\"");
3805 /* clear out previous settings */
3806 DIFF_XDL_CLR(options, NEED_MINIMAL);
3807 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3808 options->xdl_opts |= value;
3809 return argcount;
3810 }
3811
3812 /* flags options */
3813 else if (!strcmp(arg, "--binary")) {
3814 enable_patch_output(&options->output_format);
3815 DIFF_OPT_SET(options, BINARY);
3816 }
3817 else if (!strcmp(arg, "--full-index"))
3818 DIFF_OPT_SET(options, FULL_INDEX);
3819 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3820 DIFF_OPT_SET(options, TEXT);
3821 else if (!strcmp(arg, "-R"))
3822 DIFF_OPT_SET(options, REVERSE_DIFF);
3823 else if (!strcmp(arg, "--find-copies-harder"))
3824 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3825 else if (!strcmp(arg, "--follow"))
3826 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3827 else if (!strcmp(arg, "--no-follow"))
3828 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3829 else if (!strcmp(arg, "--color"))
3830 options->use_color = 1;
3831 else if (skip_prefix(arg, "--color=", &arg)) {
3832 int value = git_config_colorbool(NULL, arg);
3833 if (value < 0)
3834 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3835 options->use_color = value;
3836 }
3837 else if (!strcmp(arg, "--no-color"))
3838 options->use_color = 0;
3839 else if (!strcmp(arg, "--color-words")) {
3840 options->use_color = 1;
3841 options->word_diff = DIFF_WORDS_COLOR;
3842 }
3843 else if (skip_prefix(arg, "--color-words=", &arg)) {
3844 options->use_color = 1;
3845 options->word_diff = DIFF_WORDS_COLOR;
3846 options->word_regex = arg;
3847 }
3848 else if (!strcmp(arg, "--word-diff")) {
3849 if (options->word_diff == DIFF_WORDS_NONE)
3850 options->word_diff = DIFF_WORDS_PLAIN;
3851 }
3852 else if (skip_prefix(arg, "--word-diff=", &arg)) {
3853 if (!strcmp(arg, "plain"))
3854 options->word_diff = DIFF_WORDS_PLAIN;
3855 else if (!strcmp(arg, "color")) {
3856 options->use_color = 1;
3857 options->word_diff = DIFF_WORDS_COLOR;
3858 }
3859 else if (!strcmp(arg, "porcelain"))
3860 options->word_diff = DIFF_WORDS_PORCELAIN;
3861 else if (!strcmp(arg, "none"))
3862 options->word_diff = DIFF_WORDS_NONE;
3863 else
3864 die("bad --word-diff argument: %s", arg);
3865 }
3866 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3867 if (options->word_diff == DIFF_WORDS_NONE)
3868 options->word_diff = DIFF_WORDS_PLAIN;
3869 options->word_regex = optarg;
3870 return argcount;
3871 }
3872 else if (!strcmp(arg, "--exit-code"))
3873 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3874 else if (!strcmp(arg, "--quiet"))
3875 DIFF_OPT_SET(options, QUICK);
3876 else if (!strcmp(arg, "--ext-diff"))
3877 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3878 else if (!strcmp(arg, "--no-ext-diff"))
3879 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3880 else if (!strcmp(arg, "--textconv"))
3881 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3882 else if (!strcmp(arg, "--no-textconv"))
3883 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3884 else if (!strcmp(arg, "--ignore-submodules")) {
3885 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3886 handle_ignore_submodules_arg(options, "all");
3887 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
3888 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3889 handle_ignore_submodules_arg(options, arg);
3890 } else if (!strcmp(arg, "--submodule"))
3891 DIFF_OPT_SET(options, SUBMODULE_LOG);
3892 else if (skip_prefix(arg, "--submodule=", &arg))
3893 return parse_submodule_opt(options, arg);
3894 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
3895 return parse_ws_error_highlight(options, arg);
3896
3897 /* misc options */
3898 else if (!strcmp(arg, "-z"))
3899 options->line_termination = 0;
3900 else if ((argcount = short_opt('l', av, &optarg))) {
3901 options->rename_limit = strtoul(optarg, NULL, 10);
3902 return argcount;
3903 }
3904 else if ((argcount = short_opt('S', av, &optarg))) {
3905 options->pickaxe = optarg;
3906 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3907 return argcount;
3908 } else if ((argcount = short_opt('G', av, &optarg))) {
3909 options->pickaxe = optarg;
3910 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3911 return argcount;
3912 }
3913 else if (!strcmp(arg, "--pickaxe-all"))
3914 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3915 else if (!strcmp(arg, "--pickaxe-regex"))
3916 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3917 else if ((argcount = short_opt('O', av, &optarg))) {
3918 options->orderfile = optarg;
3919 return argcount;
3920 }
3921 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3922 int offending = parse_diff_filter_opt(optarg, options);
3923 if (offending)
3924 die("unknown change class '%c' in --diff-filter=%s",
3925 offending, optarg);
3926 return argcount;
3927 }
3928 else if (!strcmp(arg, "--abbrev"))
3929 options->abbrev = DEFAULT_ABBREV;
3930 else if (skip_prefix(arg, "--abbrev=", &arg)) {
3931 options->abbrev = strtoul(arg, NULL, 10);
3932 if (options->abbrev < MINIMUM_ABBREV)
3933 options->abbrev = MINIMUM_ABBREV;
3934 else if (40 < options->abbrev)
3935 options->abbrev = 40;
3936 }
3937 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3938 options->a_prefix = optarg;
3939 return argcount;
3940 }
3941 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3942 options->b_prefix = optarg;
3943 return argcount;
3944 }
3945 else if (!strcmp(arg, "--no-prefix"))
3946 options->a_prefix = options->b_prefix = "";
3947 else if (opt_arg(arg, '\0', "inter-hunk-context",
3948 &options->interhunkcontext))
3949 ;
3950 else if (!strcmp(arg, "-W"))
3951 DIFF_OPT_SET(options, FUNCCONTEXT);
3952 else if (!strcmp(arg, "--function-context"))
3953 DIFF_OPT_SET(options, FUNCCONTEXT);
3954 else if (!strcmp(arg, "--no-function-context"))
3955 DIFF_OPT_CLR(options, FUNCCONTEXT);
3956 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3957 options->file = fopen(optarg, "w");
3958 if (!options->file)
3959 die_errno("Could not open '%s'", optarg);
3960 options->close_file = 1;
3961 return argcount;
3962 } else
3963 return 0;
3964 return 1;
3965 }
3966
3967 int parse_rename_score(const char **cp_p)
3968 {
3969 unsigned long num, scale;
3970 int ch, dot;
3971 const char *cp = *cp_p;
3972
3973 num = 0;
3974 scale = 1;
3975 dot = 0;
3976 for (;;) {
3977 ch = *cp;
3978 if ( !dot && ch == '.' ) {
3979 scale = 1;
3980 dot = 1;
3981 } else if ( ch == '%' ) {
3982 scale = dot ? scale*100 : 100;
3983 cp++; /* % is always at the end */
3984 break;
3985 } else if ( ch >= '0' && ch <= '9' ) {
3986 if ( scale < 100000 ) {
3987 scale *= 10;
3988 num = (num*10) + (ch-'0');
3989 }
3990 } else {
3991 break;
3992 }
3993 cp++;
3994 }
3995 *cp_p = cp;
3996
3997 /* user says num divided by scale and we say internally that
3998 * is MAX_SCORE * num / scale.
3999 */
4000 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4001 }
4002
4003 static int diff_scoreopt_parse(const char *opt)
4004 {
4005 int opt1, opt2, cmd;
4006
4007 if (*opt++ != '-')
4008 return -1;
4009 cmd = *opt++;
4010 if (cmd == '-') {
4011 /* convert the long-form arguments into short-form versions */
4012 if (skip_prefix(opt, "break-rewrites", &opt)) {
4013 if (*opt == 0 || *opt++ == '=')
4014 cmd = 'B';
4015 } else if (skip_prefix(opt, "find-copies", &opt)) {
4016 if (*opt == 0 || *opt++ == '=')
4017 cmd = 'C';
4018 } else if (skip_prefix(opt, "find-renames", &opt)) {
4019 if (*opt == 0 || *opt++ == '=')
4020 cmd = 'M';
4021 }
4022 }
4023 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4024 return -1; /* that is not a -M, -C, or -B option */
4025
4026 opt1 = parse_rename_score(&opt);
4027 if (cmd != 'B')
4028 opt2 = 0;
4029 else {
4030 if (*opt == 0)
4031 opt2 = 0;
4032 else if (*opt != '/')
4033 return -1; /* we expect -B80/99 or -B80 */
4034 else {
4035 opt++;
4036 opt2 = parse_rename_score(&opt);
4037 }
4038 }
4039 if (*opt != 0)
4040 return -1;
4041 return opt1 | (opt2 << 16);
4042 }
4043
4044 struct diff_queue_struct diff_queued_diff;
4045
4046 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4047 {
4048 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4049 queue->queue[queue->nr++] = dp;
4050 }
4051
4052 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4053 struct diff_filespec *one,
4054 struct diff_filespec *two)
4055 {
4056 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4057 dp->one = one;
4058 dp->two = two;
4059 if (queue)
4060 diff_q(queue, dp);
4061 return dp;
4062 }
4063
4064 void diff_free_filepair(struct diff_filepair *p)
4065 {
4066 free_filespec(p->one);
4067 free_filespec(p->two);
4068 free(p);
4069 }
4070
4071 /* This is different from find_unique_abbrev() in that
4072 * it stuffs the result with dots for alignment.
4073 */
4074 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
4075 {
4076 int abblen;
4077 const char *abbrev;
4078 if (len == 40)
4079 return sha1_to_hex(sha1);
4080
4081 abbrev = find_unique_abbrev(sha1, len);
4082 abblen = strlen(abbrev);
4083 if (abblen < 37) {
4084 static char hex[41];
4085 if (len < abblen && abblen <= len + 2)
4086 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
4087 else
4088 sprintf(hex, "%s...", abbrev);
4089 return hex;
4090 }
4091 return sha1_to_hex(sha1);
4092 }
4093
4094 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4095 {
4096 int line_termination = opt->line_termination;
4097 int inter_name_termination = line_termination ? '\t' : '\0';
4098
4099 fprintf(opt->file, "%s", diff_line_prefix(opt));
4100 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4101 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4102 diff_unique_abbrev(p->one->sha1, opt->abbrev));
4103 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
4104 }
4105 if (p->score) {
4106 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4107 inter_name_termination);
4108 } else {
4109 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4110 }
4111
4112 if (p->status == DIFF_STATUS_COPIED ||
4113 p->status == DIFF_STATUS_RENAMED) {
4114 const char *name_a, *name_b;
4115 name_a = p->one->path;
4116 name_b = p->two->path;
4117 strip_prefix(opt->prefix_length, &name_a, &name_b);
4118 write_name_quoted(name_a, opt->file, inter_name_termination);
4119 write_name_quoted(name_b, opt->file, line_termination);
4120 } else {
4121 const char *name_a, *name_b;
4122 name_a = p->one->mode ? p->one->path : p->two->path;
4123 name_b = NULL;
4124 strip_prefix(opt->prefix_length, &name_a, &name_b);
4125 write_name_quoted(name_a, opt->file, line_termination);
4126 }
4127 }
4128
4129 int diff_unmodified_pair(struct diff_filepair *p)
4130 {
4131 /* This function is written stricter than necessary to support
4132 * the currently implemented transformers, but the idea is to
4133 * let transformers to produce diff_filepairs any way they want,
4134 * and filter and clean them up here before producing the output.
4135 */
4136 struct diff_filespec *one = p->one, *two = p->two;
4137
4138 if (DIFF_PAIR_UNMERGED(p))
4139 return 0; /* unmerged is interesting */
4140
4141 /* deletion, addition, mode or type change
4142 * and rename are all interesting.
4143 */
4144 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4145 DIFF_PAIR_MODE_CHANGED(p) ||
4146 strcmp(one->path, two->path))
4147 return 0;
4148
4149 /* both are valid and point at the same path. that is, we are
4150 * dealing with a change.
4151 */
4152 if (one->sha1_valid && two->sha1_valid &&
4153 !hashcmp(one->sha1, two->sha1) &&
4154 !one->dirty_submodule && !two->dirty_submodule)
4155 return 1; /* no change */
4156 if (!one->sha1_valid && !two->sha1_valid)
4157 return 1; /* both look at the same file on the filesystem. */
4158 return 0;
4159 }
4160
4161 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4162 {
4163 if (diff_unmodified_pair(p))
4164 return;
4165
4166 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4167 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4168 return; /* no tree diffs in patch format */
4169
4170 run_diff(p, o);
4171 }
4172
4173 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4174 struct diffstat_t *diffstat)
4175 {
4176 if (diff_unmodified_pair(p))
4177 return;
4178
4179 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4180 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4181 return; /* no useful stat for tree diffs */
4182
4183 run_diffstat(p, o, diffstat);
4184 }
4185
4186 static void diff_flush_checkdiff(struct diff_filepair *p,
4187 struct diff_options *o)
4188 {
4189 if (diff_unmodified_pair(p))
4190 return;
4191
4192 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4193 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4194 return; /* nothing to check in tree diffs */
4195
4196 run_checkdiff(p, o);
4197 }
4198
4199 int diff_queue_is_empty(void)
4200 {
4201 struct diff_queue_struct *q = &diff_queued_diff;
4202 int i;
4203 for (i = 0; i < q->nr; i++)
4204 if (!diff_unmodified_pair(q->queue[i]))
4205 return 0;
4206 return 1;
4207 }
4208
4209 #if DIFF_DEBUG
4210 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4211 {
4212 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4213 x, one ? one : "",
4214 s->path,
4215 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4216 s->mode,
4217 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
4218 fprintf(stderr, "queue[%d] %s size %lu\n",
4219 x, one ? one : "",
4220 s->size);
4221 }
4222
4223 void diff_debug_filepair(const struct diff_filepair *p, int i)
4224 {
4225 diff_debug_filespec(p->one, i, "one");
4226 diff_debug_filespec(p->two, i, "two");
4227 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4228 p->score, p->status ? p->status : '?',
4229 p->one->rename_used, p->broken_pair);
4230 }
4231
4232 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4233 {
4234 int i;
4235 if (msg)
4236 fprintf(stderr, "%s\n", msg);
4237 fprintf(stderr, "q->nr = %d\n", q->nr);
4238 for (i = 0; i < q->nr; i++) {
4239 struct diff_filepair *p = q->queue[i];
4240 diff_debug_filepair(p, i);
4241 }
4242 }
4243 #endif
4244
4245 static void diff_resolve_rename_copy(void)
4246 {
4247 int i;
4248 struct diff_filepair *p;
4249 struct diff_queue_struct *q = &diff_queued_diff;
4250
4251 diff_debug_queue("resolve-rename-copy", q);
4252
4253 for (i = 0; i < q->nr; i++) {
4254 p = q->queue[i];
4255 p->status = 0; /* undecided */
4256 if (DIFF_PAIR_UNMERGED(p))
4257 p->status = DIFF_STATUS_UNMERGED;
4258 else if (!DIFF_FILE_VALID(p->one))
4259 p->status = DIFF_STATUS_ADDED;
4260 else if (!DIFF_FILE_VALID(p->two))
4261 p->status = DIFF_STATUS_DELETED;
4262 else if (DIFF_PAIR_TYPE_CHANGED(p))
4263 p->status = DIFF_STATUS_TYPE_CHANGED;
4264
4265 /* from this point on, we are dealing with a pair
4266 * whose both sides are valid and of the same type, i.e.
4267 * either in-place edit or rename/copy edit.
4268 */
4269 else if (DIFF_PAIR_RENAME(p)) {
4270 /*
4271 * A rename might have re-connected a broken
4272 * pair up, causing the pathnames to be the
4273 * same again. If so, that's not a rename at
4274 * all, just a modification..
4275 *
4276 * Otherwise, see if this source was used for
4277 * multiple renames, in which case we decrement
4278 * the count, and call it a copy.
4279 */
4280 if (!strcmp(p->one->path, p->two->path))
4281 p->status = DIFF_STATUS_MODIFIED;
4282 else if (--p->one->rename_used > 0)
4283 p->status = DIFF_STATUS_COPIED;
4284 else
4285 p->status = DIFF_STATUS_RENAMED;
4286 }
4287 else if (hashcmp(p->one->sha1, p->two->sha1) ||
4288 p->one->mode != p->two->mode ||
4289 p->one->dirty_submodule ||
4290 p->two->dirty_submodule ||
4291 is_null_sha1(p->one->sha1))
4292 p->status = DIFF_STATUS_MODIFIED;
4293 else {
4294 /* This is a "no-change" entry and should not
4295 * happen anymore, but prepare for broken callers.
4296 */
4297 error("feeding unmodified %s to diffcore",
4298 p->one->path);
4299 p->status = DIFF_STATUS_UNKNOWN;
4300 }
4301 }
4302 diff_debug_queue("resolve-rename-copy done", q);
4303 }
4304
4305 static int check_pair_status(struct diff_filepair *p)
4306 {
4307 switch (p->status) {
4308 case DIFF_STATUS_UNKNOWN:
4309 return 0;
4310 case 0:
4311 die("internal error in diff-resolve-rename-copy");
4312 default:
4313 return 1;
4314 }
4315 }
4316
4317 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4318 {
4319 int fmt = opt->output_format;
4320
4321 if (fmt & DIFF_FORMAT_CHECKDIFF)
4322 diff_flush_checkdiff(p, opt);
4323 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4324 diff_flush_raw(p, opt);
4325 else if (fmt & DIFF_FORMAT_NAME) {
4326 const char *name_a, *name_b;
4327 name_a = p->two->path;
4328 name_b = NULL;
4329 strip_prefix(opt->prefix_length, &name_a, &name_b);
4330 write_name_quoted(name_a, opt->file, opt->line_termination);
4331 }
4332 }
4333
4334 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4335 {
4336 if (fs->mode)
4337 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4338 else
4339 fprintf(file, " %s ", newdelete);
4340 write_name_quoted(fs->path, file, '\n');
4341 }
4342
4343
4344 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4345 const char *line_prefix)
4346 {
4347 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4348 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4349 p->two->mode, show_name ? ' ' : '\n');
4350 if (show_name) {
4351 write_name_quoted(p->two->path, file, '\n');
4352 }
4353 }
4354 }
4355
4356 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4357 const char *line_prefix)
4358 {
4359 char *names = pprint_rename(p->one->path, p->two->path);
4360
4361 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4362 free(names);
4363 show_mode_change(file, p, 0, line_prefix);
4364 }
4365
4366 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4367 {
4368 FILE *file = opt->file;
4369 const char *line_prefix = diff_line_prefix(opt);
4370
4371 switch(p->status) {
4372 case DIFF_STATUS_DELETED:
4373 fputs(line_prefix, file);
4374 show_file_mode_name(file, "delete", p->one);
4375 break;
4376 case DIFF_STATUS_ADDED:
4377 fputs(line_prefix, file);
4378 show_file_mode_name(file, "create", p->two);
4379 break;
4380 case DIFF_STATUS_COPIED:
4381 fputs(line_prefix, file);
4382 show_rename_copy(file, "copy", p, line_prefix);
4383 break;
4384 case DIFF_STATUS_RENAMED:
4385 fputs(line_prefix, file);
4386 show_rename_copy(file, "rename", p, line_prefix);
4387 break;
4388 default:
4389 if (p->score) {
4390 fprintf(file, "%s rewrite ", line_prefix);
4391 write_name_quoted(p->two->path, file, ' ');
4392 fprintf(file, "(%d%%)\n", similarity_index(p));
4393 }
4394 show_mode_change(file, p, !p->score, line_prefix);
4395 break;
4396 }
4397 }
4398
4399 struct patch_id_t {
4400 git_SHA_CTX *ctx;
4401 int patchlen;
4402 };
4403
4404 static int remove_space(char *line, int len)
4405 {
4406 int i;
4407 char *dst = line;
4408 unsigned char c;
4409
4410 for (i = 0; i < len; i++)
4411 if (!isspace((c = line[i])))
4412 *dst++ = c;
4413
4414 return dst - line;
4415 }
4416
4417 static void patch_id_consume(void *priv, char *line, unsigned long len)
4418 {
4419 struct patch_id_t *data = priv;
4420 int new_len;
4421
4422 /* Ignore line numbers when computing the SHA1 of the patch */
4423 if (starts_with(line, "@@ -"))
4424 return;
4425
4426 new_len = remove_space(line, len);
4427
4428 git_SHA1_Update(data->ctx, line, new_len);
4429 data->patchlen += new_len;
4430 }
4431
4432 /* returns 0 upon success, and writes result into sha1 */
4433 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4434 {
4435 struct diff_queue_struct *q = &diff_queued_diff;
4436 int i;
4437 git_SHA_CTX ctx;
4438 struct patch_id_t data;
4439 char buffer[PATH_MAX * 4 + 20];
4440
4441 git_SHA1_Init(&ctx);
4442 memset(&data, 0, sizeof(struct patch_id_t));
4443 data.ctx = &ctx;
4444
4445 for (i = 0; i < q->nr; i++) {
4446 xpparam_t xpp;
4447 xdemitconf_t xecfg;
4448 mmfile_t mf1, mf2;
4449 struct diff_filepair *p = q->queue[i];
4450 int len1, len2;
4451
4452 memset(&xpp, 0, sizeof(xpp));
4453 memset(&xecfg, 0, sizeof(xecfg));
4454 if (p->status == 0)
4455 return error("internal diff status error");
4456 if (p->status == DIFF_STATUS_UNKNOWN)
4457 continue;
4458 if (diff_unmodified_pair(p))
4459 continue;
4460 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4461 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4462 continue;
4463 if (DIFF_PAIR_UNMERGED(p))
4464 continue;
4465
4466 diff_fill_sha1_info(p->one);
4467 diff_fill_sha1_info(p->two);
4468 if (fill_mmfile(&mf1, p->one) < 0 ||
4469 fill_mmfile(&mf2, p->two) < 0)
4470 return error("unable to read files to diff");
4471
4472 len1 = remove_space(p->one->path, strlen(p->one->path));
4473 len2 = remove_space(p->two->path, strlen(p->two->path));
4474 if (p->one->mode == 0)
4475 len1 = snprintf(buffer, sizeof(buffer),
4476 "diff--gita/%.*sb/%.*s"
4477 "newfilemode%06o"
4478 "---/dev/null"
4479 "+++b/%.*s",
4480 len1, p->one->path,
4481 len2, p->two->path,
4482 p->two->mode,
4483 len2, p->two->path);
4484 else if (p->two->mode == 0)
4485 len1 = snprintf(buffer, sizeof(buffer),
4486 "diff--gita/%.*sb/%.*s"
4487 "deletedfilemode%06o"
4488 "---a/%.*s"
4489 "+++/dev/null",
4490 len1, p->one->path,
4491 len2, p->two->path,
4492 p->one->mode,
4493 len1, p->one->path);
4494 else
4495 len1 = snprintf(buffer, sizeof(buffer),
4496 "diff--gita/%.*sb/%.*s"
4497 "---a/%.*s"
4498 "+++b/%.*s",
4499 len1, p->one->path,
4500 len2, p->two->path,
4501 len1, p->one->path,
4502 len2, p->two->path);
4503 git_SHA1_Update(&ctx, buffer, len1);
4504
4505 if (diff_filespec_is_binary(p->one) ||
4506 diff_filespec_is_binary(p->two)) {
4507 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4508 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4509 continue;
4510 }
4511
4512 xpp.flags = 0;
4513 xecfg.ctxlen = 3;
4514 xecfg.flags = 0;
4515 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4516 &xpp, &xecfg))
4517 return error("unable to generate patch-id diff for %s",
4518 p->one->path);
4519 }
4520
4521 git_SHA1_Final(sha1, &ctx);
4522 return 0;
4523 }
4524
4525 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4526 {
4527 struct diff_queue_struct *q = &diff_queued_diff;
4528 int i;
4529 int result = diff_get_patch_id(options, sha1);
4530
4531 for (i = 0; i < q->nr; i++)
4532 diff_free_filepair(q->queue[i]);
4533
4534 free(q->queue);
4535 DIFF_QUEUE_CLEAR(q);
4536
4537 return result;
4538 }
4539
4540 static int is_summary_empty(const struct diff_queue_struct *q)
4541 {
4542 int i;
4543
4544 for (i = 0; i < q->nr; i++) {
4545 const struct diff_filepair *p = q->queue[i];
4546
4547 switch (p->status) {
4548 case DIFF_STATUS_DELETED:
4549 case DIFF_STATUS_ADDED:
4550 case DIFF_STATUS_COPIED:
4551 case DIFF_STATUS_RENAMED:
4552 return 0;
4553 default:
4554 if (p->score)
4555 return 0;
4556 if (p->one->mode && p->two->mode &&
4557 p->one->mode != p->two->mode)
4558 return 0;
4559 break;
4560 }
4561 }
4562 return 1;
4563 }
4564
4565 static const char rename_limit_warning[] =
4566 "inexact rename detection was skipped due to too many files.";
4567
4568 static const char degrade_cc_to_c_warning[] =
4569 "only found copies from modified paths due to too many files.";
4570
4571 static const char rename_limit_advice[] =
4572 "you may want to set your %s variable to at least "
4573 "%d and retry the command.";
4574
4575 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4576 {
4577 if (degraded_cc)
4578 warning(degrade_cc_to_c_warning);
4579 else if (needed)
4580 warning(rename_limit_warning);
4581 else
4582 return;
4583 if (0 < needed && needed < 32767)
4584 warning(rename_limit_advice, varname, needed);
4585 }
4586
4587 void diff_flush(struct diff_options *options)
4588 {
4589 struct diff_queue_struct *q = &diff_queued_diff;
4590 int i, output_format = options->output_format;
4591 int separator = 0;
4592 int dirstat_by_line = 0;
4593
4594 /*
4595 * Order: raw, stat, summary, patch
4596 * or: name/name-status/checkdiff (other bits clear)
4597 */
4598 if (!q->nr)
4599 goto free_queue;
4600
4601 if (output_format & (DIFF_FORMAT_RAW |
4602 DIFF_FORMAT_NAME |
4603 DIFF_FORMAT_NAME_STATUS |
4604 DIFF_FORMAT_CHECKDIFF)) {
4605 for (i = 0; i < q->nr; i++) {
4606 struct diff_filepair *p = q->queue[i];
4607 if (check_pair_status(p))
4608 flush_one_pair(p, options);
4609 }
4610 separator++;
4611 }
4612
4613 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4614 dirstat_by_line = 1;
4615
4616 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4617 dirstat_by_line) {
4618 struct diffstat_t diffstat;
4619
4620 memset(&diffstat, 0, sizeof(struct diffstat_t));
4621 for (i = 0; i < q->nr; i++) {
4622 struct diff_filepair *p = q->queue[i];
4623 if (check_pair_status(p))
4624 diff_flush_stat(p, options, &diffstat);
4625 }
4626 if (output_format & DIFF_FORMAT_NUMSTAT)
4627 show_numstat(&diffstat, options);
4628 if (output_format & DIFF_FORMAT_DIFFSTAT)
4629 show_stats(&diffstat, options);
4630 if (output_format & DIFF_FORMAT_SHORTSTAT)
4631 show_shortstats(&diffstat, options);
4632 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
4633 show_dirstat_by_line(&diffstat, options);
4634 free_diffstat_info(&diffstat);
4635 separator++;
4636 }
4637 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4638 show_dirstat(options);
4639
4640 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4641 for (i = 0; i < q->nr; i++) {
4642 diff_summary(options, q->queue[i]);
4643 }
4644 separator++;
4645 }
4646
4647 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4648 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4649 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4650 /*
4651 * run diff_flush_patch for the exit status. setting
4652 * options->file to /dev/null should be safe, because we
4653 * aren't supposed to produce any output anyway.
4654 */
4655 if (options->close_file)
4656 fclose(options->file);
4657 options->file = fopen("/dev/null", "w");
4658 if (!options->file)
4659 die_errno("Could not open /dev/null");
4660 options->close_file = 1;
4661 for (i = 0; i < q->nr; i++) {
4662 struct diff_filepair *p = q->queue[i];
4663 if (check_pair_status(p))
4664 diff_flush_patch(p, options);
4665 if (options->found_changes)
4666 break;
4667 }
4668 }
4669
4670 if (output_format & DIFF_FORMAT_PATCH) {
4671 if (separator) {
4672 fprintf(options->file, "%s%c",
4673 diff_line_prefix(options),
4674 options->line_termination);
4675 if (options->stat_sep) {
4676 /* attach patch instead of inline */
4677 fputs(options->stat_sep, options->file);
4678 }
4679 }
4680
4681 for (i = 0; i < q->nr; i++) {
4682 struct diff_filepair *p = q->queue[i];
4683 if (check_pair_status(p))
4684 diff_flush_patch(p, options);
4685 }
4686 }
4687
4688 if (output_format & DIFF_FORMAT_CALLBACK)
4689 options->format_callback(q, options, options->format_callback_data);
4690
4691 for (i = 0; i < q->nr; i++)
4692 diff_free_filepair(q->queue[i]);
4693 free_queue:
4694 free(q->queue);
4695 DIFF_QUEUE_CLEAR(q);
4696 if (options->close_file)
4697 fclose(options->file);
4698
4699 /*
4700 * Report the content-level differences with HAS_CHANGES;
4701 * diff_addremove/diff_change does not set the bit when
4702 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4703 */
4704 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4705 if (options->found_changes)
4706 DIFF_OPT_SET(options, HAS_CHANGES);
4707 else
4708 DIFF_OPT_CLR(options, HAS_CHANGES);
4709 }
4710 }
4711
4712 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
4713 {
4714 return (((p->status == DIFF_STATUS_MODIFIED) &&
4715 ((p->score &&
4716 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
4717 (!p->score &&
4718 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
4719 ((p->status != DIFF_STATUS_MODIFIED) &&
4720 filter_bit_tst(p->status, options)));
4721 }
4722
4723 static void diffcore_apply_filter(struct diff_options *options)
4724 {
4725 int i;
4726 struct diff_queue_struct *q = &diff_queued_diff;
4727 struct diff_queue_struct outq;
4728
4729 DIFF_QUEUE_CLEAR(&outq);
4730
4731 if (!options->filter)
4732 return;
4733
4734 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
4735 int found;
4736 for (i = found = 0; !found && i < q->nr; i++) {
4737 if (match_filter(options, q->queue[i]))
4738 found++;
4739 }
4740 if (found)
4741 return;
4742
4743 /* otherwise we will clear the whole queue
4744 * by copying the empty outq at the end of this
4745 * function, but first clear the current entries
4746 * in the queue.
4747 */
4748 for (i = 0; i < q->nr; i++)
4749 diff_free_filepair(q->queue[i]);
4750 }
4751 else {
4752 /* Only the matching ones */
4753 for (i = 0; i < q->nr; i++) {
4754 struct diff_filepair *p = q->queue[i];
4755 if (match_filter(options, p))
4756 diff_q(&outq, p);
4757 else
4758 diff_free_filepair(p);
4759 }
4760 }
4761 free(q->queue);
4762 *q = outq;
4763 }
4764
4765 /* Check whether two filespecs with the same mode and size are identical */
4766 static int diff_filespec_is_identical(struct diff_filespec *one,
4767 struct diff_filespec *two)
4768 {
4769 if (S_ISGITLINK(one->mode))
4770 return 0;
4771 if (diff_populate_filespec(one, 0))
4772 return 0;
4773 if (diff_populate_filespec(two, 0))
4774 return 0;
4775 return !memcmp(one->data, two->data, one->size);
4776 }
4777
4778 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
4779 {
4780 if (p->done_skip_stat_unmatch)
4781 return p->skip_stat_unmatch_result;
4782
4783 p->done_skip_stat_unmatch = 1;
4784 p->skip_stat_unmatch_result = 0;
4785 /*
4786 * 1. Entries that come from stat info dirtiness
4787 * always have both sides (iow, not create/delete),
4788 * one side of the object name is unknown, with
4789 * the same mode and size. Keep the ones that
4790 * do not match these criteria. They have real
4791 * differences.
4792 *
4793 * 2. At this point, the file is known to be modified,
4794 * with the same mode and size, and the object
4795 * name of one side is unknown. Need to inspect
4796 * the identical contents.
4797 */
4798 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4799 !DIFF_FILE_VALID(p->two) ||
4800 (p->one->sha1_valid && p->two->sha1_valid) ||
4801 (p->one->mode != p->two->mode) ||
4802 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
4803 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
4804 (p->one->size != p->two->size) ||
4805 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4806 p->skip_stat_unmatch_result = 1;
4807 return p->skip_stat_unmatch_result;
4808 }
4809
4810 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4811 {
4812 int i;
4813 struct diff_queue_struct *q = &diff_queued_diff;
4814 struct diff_queue_struct outq;
4815 DIFF_QUEUE_CLEAR(&outq);
4816
4817 for (i = 0; i < q->nr; i++) {
4818 struct diff_filepair *p = q->queue[i];
4819
4820 if (diff_filespec_check_stat_unmatch(p))
4821 diff_q(&outq, p);
4822 else {
4823 /*
4824 * The caller can subtract 1 from skip_stat_unmatch
4825 * to determine how many paths were dirty only
4826 * due to stat info mismatch.
4827 */
4828 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4829 diffopt->skip_stat_unmatch++;
4830 diff_free_filepair(p);
4831 }
4832 }
4833 free(q->queue);
4834 *q = outq;
4835 }
4836
4837 static int diffnamecmp(const void *a_, const void *b_)
4838 {
4839 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4840 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4841 const char *name_a, *name_b;
4842
4843 name_a = a->one ? a->one->path : a->two->path;
4844 name_b = b->one ? b->one->path : b->two->path;
4845 return strcmp(name_a, name_b);
4846 }
4847
4848 void diffcore_fix_diff_index(struct diff_options *options)
4849 {
4850 struct diff_queue_struct *q = &diff_queued_diff;
4851 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4852 }
4853
4854 void diffcore_std(struct diff_options *options)
4855 {
4856 /* NOTE please keep the following in sync with diff_tree_combined() */
4857 if (options->skip_stat_unmatch)
4858 diffcore_skip_stat_unmatch(options);
4859 if (!options->found_follow) {
4860 /* See try_to_follow_renames() in tree-diff.c */
4861 if (options->break_opt != -1)
4862 diffcore_break(options->break_opt);
4863 if (options->detect_rename)
4864 diffcore_rename(options);
4865 if (options->break_opt != -1)
4866 diffcore_merge_broken();
4867 }
4868 if (options->pickaxe)
4869 diffcore_pickaxe(options);
4870 if (options->orderfile)
4871 diffcore_order(options->orderfile);
4872 if (!options->found_follow)
4873 /* See try_to_follow_renames() in tree-diff.c */
4874 diff_resolve_rename_copy();
4875 diffcore_apply_filter(options);
4876
4877 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4878 DIFF_OPT_SET(options, HAS_CHANGES);
4879 else
4880 DIFF_OPT_CLR(options, HAS_CHANGES);
4881
4882 options->found_follow = 0;
4883 }
4884
4885 int diff_result_code(struct diff_options *opt, int status)
4886 {
4887 int result = 0;
4888
4889 diff_warn_rename_limit("diff.renameLimit",
4890 opt->needed_rename_limit,
4891 opt->degraded_cc_to_c);
4892 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4893 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4894 return status;
4895 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4896 DIFF_OPT_TST(opt, HAS_CHANGES))
4897 result |= 01;
4898 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4899 DIFF_OPT_TST(opt, CHECK_FAILED))
4900 result |= 02;
4901 return result;
4902 }
4903
4904 int diff_can_quit_early(struct diff_options *opt)
4905 {
4906 return (DIFF_OPT_TST(opt, QUICK) &&
4907 !opt->filter &&
4908 DIFF_OPT_TST(opt, HAS_CHANGES));
4909 }
4910
4911 /*
4912 * Shall changes to this submodule be ignored?
4913 *
4914 * Submodule changes can be configured to be ignored separately for each path,
4915 * but that configuration can be overridden from the command line.
4916 */
4917 static int is_submodule_ignored(const char *path, struct diff_options *options)
4918 {
4919 int ignored = 0;
4920 unsigned orig_flags = options->flags;
4921 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4922 set_diffopt_flags_from_submodule_config(options, path);
4923 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4924 ignored = 1;
4925 options->flags = orig_flags;
4926 return ignored;
4927 }
4928
4929 void diff_addremove(struct diff_options *options,
4930 int addremove, unsigned mode,
4931 const unsigned char *sha1,
4932 int sha1_valid,
4933 const char *concatpath, unsigned dirty_submodule)
4934 {
4935 struct diff_filespec *one, *two;
4936
4937 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4938 return;
4939
4940 /* This may look odd, but it is a preparation for
4941 * feeding "there are unchanged files which should
4942 * not produce diffs, but when you are doing copy
4943 * detection you would need them, so here they are"
4944 * entries to the diff-core. They will be prefixed
4945 * with something like '=' or '*' (I haven't decided
4946 * which but should not make any difference).
4947 * Feeding the same new and old to diff_change()
4948 * also has the same effect.
4949 * Before the final output happens, they are pruned after
4950 * merged into rename/copy pairs as appropriate.
4951 */
4952 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4953 addremove = (addremove == '+' ? '-' :
4954 addremove == '-' ? '+' : addremove);
4955
4956 if (options->prefix &&
4957 strncmp(concatpath, options->prefix, options->prefix_length))
4958 return;
4959
4960 one = alloc_filespec(concatpath);
4961 two = alloc_filespec(concatpath);
4962
4963 if (addremove != '+')
4964 fill_filespec(one, sha1, sha1_valid, mode);
4965 if (addremove != '-') {
4966 fill_filespec(two, sha1, sha1_valid, mode);
4967 two->dirty_submodule = dirty_submodule;
4968 }
4969
4970 diff_queue(&diff_queued_diff, one, two);
4971 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4972 DIFF_OPT_SET(options, HAS_CHANGES);
4973 }
4974
4975 void diff_change(struct diff_options *options,
4976 unsigned old_mode, unsigned new_mode,
4977 const unsigned char *old_sha1,
4978 const unsigned char *new_sha1,
4979 int old_sha1_valid, int new_sha1_valid,
4980 const char *concatpath,
4981 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4982 {
4983 struct diff_filespec *one, *two;
4984 struct diff_filepair *p;
4985
4986 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4987 is_submodule_ignored(concatpath, options))
4988 return;
4989
4990 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4991 unsigned tmp;
4992 const unsigned char *tmp_c;
4993 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4994 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4995 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
4996 new_sha1_valid = tmp;
4997 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4998 new_dirty_submodule = tmp;
4999 }
5000
5001 if (options->prefix &&
5002 strncmp(concatpath, options->prefix, options->prefix_length))
5003 return;
5004
5005 one = alloc_filespec(concatpath);
5006 two = alloc_filespec(concatpath);
5007 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
5008 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
5009 one->dirty_submodule = old_dirty_submodule;
5010 two->dirty_submodule = new_dirty_submodule;
5011 p = diff_queue(&diff_queued_diff, one, two);
5012
5013 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5014 return;
5015
5016 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5017 !diff_filespec_check_stat_unmatch(p))
5018 return;
5019
5020 DIFF_OPT_SET(options, HAS_CHANGES);
5021 }
5022
5023 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5024 {
5025 struct diff_filepair *pair;
5026 struct diff_filespec *one, *two;
5027
5028 if (options->prefix &&
5029 strncmp(path, options->prefix, options->prefix_length))
5030 return NULL;
5031
5032 one = alloc_filespec(path);
5033 two = alloc_filespec(path);
5034 pair = diff_queue(&diff_queued_diff, one, two);
5035 pair->is_unmerged = 1;
5036 return pair;
5037 }
5038
5039 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5040 size_t *outsize)
5041 {
5042 struct diff_tempfile *temp;
5043 const char *argv[3];
5044 const char **arg = argv;
5045 struct child_process child = CHILD_PROCESS_INIT;
5046 struct strbuf buf = STRBUF_INIT;
5047 int err = 0;
5048
5049 temp = prepare_temp_file(spec->path, spec);
5050 *arg++ = pgm;
5051 *arg++ = temp->name;
5052 *arg = NULL;
5053
5054 child.use_shell = 1;
5055 child.argv = argv;
5056 child.out = -1;
5057 if (start_command(&child)) {
5058 remove_tempfile();
5059 return NULL;
5060 }
5061
5062 if (strbuf_read(&buf, child.out, 0) < 0)
5063 err = error("error reading from textconv command '%s'", pgm);
5064 close(child.out);
5065
5066 if (finish_command(&child) || err) {
5067 strbuf_release(&buf);
5068 remove_tempfile();
5069 return NULL;
5070 }
5071 remove_tempfile();
5072
5073 return strbuf_detach(&buf, outsize);
5074 }
5075
5076 size_t fill_textconv(struct userdiff_driver *driver,
5077 struct diff_filespec *df,
5078 char **outbuf)
5079 {
5080 size_t size;
5081
5082 if (!driver || !driver->textconv) {
5083 if (!DIFF_FILE_VALID(df)) {
5084 *outbuf = "";
5085 return 0;
5086 }
5087 if (diff_populate_filespec(df, 0))
5088 die("unable to read files to diff");
5089 *outbuf = df->data;
5090 return df->size;
5091 }
5092
5093 if (driver->textconv_cache && df->sha1_valid) {
5094 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
5095 &size);
5096 if (*outbuf)
5097 return size;
5098 }
5099
5100 *outbuf = run_textconv(driver->textconv, df, &size);
5101 if (!*outbuf)
5102 die("unable to read files to diff");
5103
5104 if (driver->textconv_cache && df->sha1_valid) {
5105 /* ignore errors, as we might be in a readonly repository */
5106 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
5107 size);
5108 /*
5109 * we could save up changes and flush them all at the end,
5110 * but we would need an extra call after all diffing is done.
5111 * Since generating a cache entry is the slow path anyway,
5112 * this extra overhead probably isn't a big deal.
5113 */
5114 notes_cache_write(driver->textconv_cache);
5115 }
5116
5117 return size;
5118 }
5119
5120 void setup_diff_pager(struct diff_options *opt)
5121 {
5122 /*
5123 * If the user asked for our exit code, then either they want --quiet
5124 * or --exit-code. We should definitely not bother with a pager in the
5125 * former case, as we will generate no output. Since we still properly
5126 * report our exit code even when a pager is run, we _could_ run a
5127 * pager with --exit-code. But since we have not done so historically,
5128 * and because it is easy to find people oneline advising "git diff
5129 * --exit-code" in hooks and other scripts, we do not do so.
5130 */
5131 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5132 check_pager_config("diff") != 0)
5133 setup_pager();
5134 }
git for Debian