?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
path: add a function to check for path suffix
[git/debian.git] / diff.c
blob1ccd96bbfde23a0f6d42104e81f78e0953c1bc93
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "config.h"
6 #include "tempfile.h"
7 #include "quote.h"
8 #include "diff.h"
9 #include "diffcore.h"
10 #include "delta.h"
11 #include "xdiff-interface.h"
12 #include "color.h"
13 #include "attr.h"
14 #include "run-command.h"
15 #include "utf8.h"
16 #include "object-store.h"
17 #include "userdiff.h"
18 #include "submodule-config.h"
19 #include "submodule.h"
20 #include "hashmap.h"
21 #include "ll-merge.h"
22 #include "string-list.h"
23 #include "argv-array.h"
24 #include "graph.h"
25 #include "packfile.h"
26 #include "parse-options.h"
27 #include "help.h"
28 #include "fetch-object.h"
29
30 #ifdef NO_FAST_WORKING_DIRECTORY
31 #define FAST_WORKING_DIRECTORY 0
32 #else
33 #define FAST_WORKING_DIRECTORY 1
34 #endif
35
36 static int diff_detect_rename_default;
37 static int diff_indent_heuristic = 1;
38 static int diff_rename_limit_default = 400;
39 static int diff_suppress_blank_empty;
40 static int diff_use_color_default = -1;
41 static int diff_color_moved_default;
42 static int diff_color_moved_ws_default;
43 static int diff_context_default = 3;
44 static int diff_interhunk_context_default;
45 static const char *diff_word_regex_cfg;
46 static const char *external_diff_cmd_cfg;
47 static const char *diff_order_file_cfg;
48 int diff_auto_refresh_index = 1;
49 static int diff_mnemonic_prefix;
50 static int diff_no_prefix;
51 static int diff_stat_graph_width;
52 static int diff_dirstat_permille_default = 30;
53 static struct diff_options default_diff_options;
54 static long diff_algorithm;
55 static unsigned ws_error_highlight_default = WSEH_NEW;
56
57 static char diff_colors[][COLOR_MAXLEN] = {
58 GIT_COLOR_RESET,
59 GIT_COLOR_NORMAL, /* CONTEXT */
60 GIT_COLOR_BOLD, /* METAINFO */
61 GIT_COLOR_CYAN, /* FRAGINFO */
62 GIT_COLOR_RED, /* OLD */
63 GIT_COLOR_GREEN, /* NEW */
64 GIT_COLOR_YELLOW, /* COMMIT */
65 GIT_COLOR_BG_RED, /* WHITESPACE */
66 GIT_COLOR_NORMAL, /* FUNCINFO */
67 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
68 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
69 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
70 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
71 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
72 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
73 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
74 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
75 GIT_COLOR_FAINT, /* CONTEXT_DIM */
76 GIT_COLOR_FAINT_RED, /* OLD_DIM */
77 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
78 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
79 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
80 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
81 };
82
83 static const char *color_diff_slots[] = {
84 [DIFF_CONTEXT] = "context",
85 [DIFF_METAINFO] = "meta",
86 [DIFF_FRAGINFO] = "frag",
87 [DIFF_FILE_OLD] = "old",
88 [DIFF_FILE_NEW] = "new",
89 [DIFF_COMMIT] = "commit",
90 [DIFF_WHITESPACE] = "whitespace",
91 [DIFF_FUNCINFO] = "func",
92 [DIFF_FILE_OLD_MOVED] = "oldMoved",
93 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
94 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
95 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
96 [DIFF_FILE_NEW_MOVED] = "newMoved",
97 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
98 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
99 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
100 [DIFF_CONTEXT_DIM] = "contextDimmed",
101 [DIFF_FILE_OLD_DIM] = "oldDimmed",
102 [DIFF_FILE_NEW_DIM] = "newDimmed",
103 [DIFF_CONTEXT_BOLD] = "contextBold",
104 [DIFF_FILE_OLD_BOLD] = "oldBold",
105 [DIFF_FILE_NEW_BOLD] = "newBold",
106 };
107
108 define_list_config_array_extra(color_diff_slots, {"plain"});
109
110 static int parse_diff_color_slot(const char *var)
111 {
112 if (!strcasecmp(var, "plain"))
113 return DIFF_CONTEXT;
114 return LOOKUP_CONFIG(color_diff_slots, var);
115 }
116
117 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
118 struct strbuf *errmsg)
119 {
120 char *params_copy = xstrdup(params_string);
121 struct string_list params = STRING_LIST_INIT_NODUP;
122 int ret = 0;
123 int i;
124
125 if (*params_copy)
126 string_list_split_in_place(&params, params_copy, ',', -1);
127 for (i = 0; i < params.nr; i++) {
128 const char *p = params.items[i].string;
129 if (!strcmp(p, "changes")) {
130 options->flags.dirstat_by_line = 0;
131 options->flags.dirstat_by_file = 0;
132 } else if (!strcmp(p, "lines")) {
133 options->flags.dirstat_by_line = 1;
134 options->flags.dirstat_by_file = 0;
135 } else if (!strcmp(p, "files")) {
136 options->flags.dirstat_by_line = 0;
137 options->flags.dirstat_by_file = 1;
138 } else if (!strcmp(p, "noncumulative")) {
139 options->flags.dirstat_cumulative = 0;
140 } else if (!strcmp(p, "cumulative")) {
141 options->flags.dirstat_cumulative = 1;
142 } else if (isdigit(*p)) {
143 char *end;
144 int permille = strtoul(p, &end, 10) * 10;
145 if (*end == '.' && isdigit(*++end)) {
146 /* only use first digit */
147 permille += *end - '0';
148 /* .. and ignore any further digits */
149 while (isdigit(*++end))
150 ; /* nothing */
151 }
152 if (!*end)
153 options->dirstat_permille = permille;
154 else {
155 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
156 p);
157 ret++;
158 }
159 } else {
160 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
161 ret++;
162 }
163
164 }
165 string_list_clear(&params, 0);
166 free(params_copy);
167 return ret;
168 }
169
170 static int parse_submodule_params(struct diff_options *options, const char *value)
171 {
172 if (!strcmp(value, "log"))
173 options->submodule_format = DIFF_SUBMODULE_LOG;
174 else if (!strcmp(value, "short"))
175 options->submodule_format = DIFF_SUBMODULE_SHORT;
176 else if (!strcmp(value, "diff"))
177 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
178 /*
179 * Please update $__git_diff_submodule_formats in
180 * git-completion.bash when you add new formats.
181 */
182 else
183 return -1;
184 return 0;
185 }
186
187 int git_config_rename(const char *var, const char *value)
188 {
189 if (!value)
190 return DIFF_DETECT_RENAME;
191 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
192 return DIFF_DETECT_COPY;
193 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
194 }
195
196 long parse_algorithm_value(const char *value)
197 {
198 if (!value)
199 return -1;
200 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
201 return 0;
202 else if (!strcasecmp(value, "minimal"))
203 return XDF_NEED_MINIMAL;
204 else if (!strcasecmp(value, "patience"))
205 return XDF_PATIENCE_DIFF;
206 else if (!strcasecmp(value, "histogram"))
207 return XDF_HISTOGRAM_DIFF;
208 /*
209 * Please update $__git_diff_algorithms in git-completion.bash
210 * when you add new algorithms.
211 */
212 return -1;
213 }
214
215 static int parse_one_token(const char **arg, const char *token)
216 {
217 const char *rest;
218 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
219 *arg = rest;
220 return 1;
221 }
222 return 0;
223 }
224
225 static int parse_ws_error_highlight(const char *arg)
226 {
227 const char *orig_arg = arg;
228 unsigned val = 0;
229
230 while (*arg) {
231 if (parse_one_token(&arg, "none"))
232 val = 0;
233 else if (parse_one_token(&arg, "default"))
234 val = WSEH_NEW;
235 else if (parse_one_token(&arg, "all"))
236 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
237 else if (parse_one_token(&arg, "new"))
238 val |= WSEH_NEW;
239 else if (parse_one_token(&arg, "old"))
240 val |= WSEH_OLD;
241 else if (parse_one_token(&arg, "context"))
242 val |= WSEH_CONTEXT;
243 else {
244 return -1 - (int)(arg - orig_arg);
245 }
246 if (*arg)
247 arg++;
248 }
249 return val;
250 }
251
252 /*
253 * These are to give UI layer defaults.
254 * The core-level commands such as git-diff-files should
255 * never be affected by the setting of diff.renames
256 * the user happens to have in the configuration file.
257 */
258 void init_diff_ui_defaults(void)
259 {
260 diff_detect_rename_default = DIFF_DETECT_RENAME;
261 }
262
263 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
264 {
265 if (!strcmp(var, "diff.indentheuristic"))
266 diff_indent_heuristic = git_config_bool(var, value);
267 return 0;
268 }
269
270 static int parse_color_moved(const char *arg)
271 {
272 switch (git_parse_maybe_bool(arg)) {
273 case 0:
274 return COLOR_MOVED_NO;
275 case 1:
276 return COLOR_MOVED_DEFAULT;
277 default:
278 break;
279 }
280
281 if (!strcmp(arg, "no"))
282 return COLOR_MOVED_NO;
283 else if (!strcmp(arg, "plain"))
284 return COLOR_MOVED_PLAIN;
285 else if (!strcmp(arg, "blocks"))
286 return COLOR_MOVED_BLOCKS;
287 else if (!strcmp(arg, "zebra"))
288 return COLOR_MOVED_ZEBRA;
289 else if (!strcmp(arg, "default"))
290 return COLOR_MOVED_DEFAULT;
291 else if (!strcmp(arg, "dimmed-zebra"))
292 return COLOR_MOVED_ZEBRA_DIM;
293 else if (!strcmp(arg, "dimmed_zebra"))
294 return COLOR_MOVED_ZEBRA_DIM;
295 else
296 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
297 }
298
299 static unsigned parse_color_moved_ws(const char *arg)
300 {
301 int ret = 0;
302 struct string_list l = STRING_LIST_INIT_DUP;
303 struct string_list_item *i;
304
305 string_list_split(&l, arg, ',', -1);
306
307 for_each_string_list_item(i, &l) {
308 struct strbuf sb = STRBUF_INIT;
309 strbuf_addstr(&sb, i->string);
310 strbuf_trim(&sb);
311
312 if (!strcmp(sb.buf, "no"))
313 ret = 0;
314 else if (!strcmp(sb.buf, "ignore-space-change"))
315 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
316 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
317 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
318 else if (!strcmp(sb.buf, "ignore-all-space"))
319 ret |= XDF_IGNORE_WHITESPACE;
320 else if (!strcmp(sb.buf, "allow-indentation-change"))
321 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
322 else {
323 ret |= COLOR_MOVED_WS_ERROR;
324 error(_("unknown color-moved-ws mode '%s', possible values are 'ignore-space-change', 'ignore-space-at-eol', 'ignore-all-space', 'allow-indentation-change'"), sb.buf);
325 }
326
327 strbuf_release(&sb);
328 }
329
330 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
331 (ret & XDF_WHITESPACE_FLAGS)) {
332 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
333 ret |= COLOR_MOVED_WS_ERROR;
334 }
335
336 string_list_clear(&l, 0);
337
338 return ret;
339 }
340
341 int git_diff_ui_config(const char *var, const char *value, void *cb)
342 {
343 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
344 diff_use_color_default = git_config_colorbool(var, value);
345 return 0;
346 }
347 if (!strcmp(var, "diff.colormoved")) {
348 int cm = parse_color_moved(value);
349 if (cm < 0)
350 return -1;
351 diff_color_moved_default = cm;
352 return 0;
353 }
354 if (!strcmp(var, "diff.colormovedws")) {
355 unsigned cm = parse_color_moved_ws(value);
356 if (cm & COLOR_MOVED_WS_ERROR)
357 return -1;
358 diff_color_moved_ws_default = cm;
359 return 0;
360 }
361 if (!strcmp(var, "diff.context")) {
362 diff_context_default = git_config_int(var, value);
363 if (diff_context_default < 0)
364 return -1;
365 return 0;
366 }
367 if (!strcmp(var, "diff.interhunkcontext")) {
368 diff_interhunk_context_default = git_config_int(var, value);
369 if (diff_interhunk_context_default < 0)
370 return -1;
371 return 0;
372 }
373 if (!strcmp(var, "diff.renames")) {
374 diff_detect_rename_default = git_config_rename(var, value);
375 return 0;
376 }
377 if (!strcmp(var, "diff.autorefreshindex")) {
378 diff_auto_refresh_index = git_config_bool(var, value);
379 return 0;
380 }
381 if (!strcmp(var, "diff.mnemonicprefix")) {
382 diff_mnemonic_prefix = git_config_bool(var, value);
383 return 0;
384 }
385 if (!strcmp(var, "diff.noprefix")) {
386 diff_no_prefix = git_config_bool(var, value);
387 return 0;
388 }
389 if (!strcmp(var, "diff.statgraphwidth")) {
390 diff_stat_graph_width = git_config_int(var, value);
391 return 0;
392 }
393 if (!strcmp(var, "diff.external"))
394 return git_config_string(&external_diff_cmd_cfg, var, value);
395 if (!strcmp(var, "diff.wordregex"))
396 return git_config_string(&diff_word_regex_cfg, var, value);
397 if (!strcmp(var, "diff.orderfile"))
398 return git_config_pathname(&diff_order_file_cfg, var, value);
399
400 if (!strcmp(var, "diff.ignoresubmodules"))
401 handle_ignore_submodules_arg(&default_diff_options, value);
402
403 if (!strcmp(var, "diff.submodule")) {
404 if (parse_submodule_params(&default_diff_options, value))
405 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
406 value);
407 return 0;
408 }
409
410 if (!strcmp(var, "diff.algorithm")) {
411 diff_algorithm = parse_algorithm_value(value);
412 if (diff_algorithm < 0)
413 return -1;
414 return 0;
415 }
416
417 if (!strcmp(var, "diff.wserrorhighlight")) {
418 int val = parse_ws_error_highlight(value);
419 if (val < 0)
420 return -1;
421 ws_error_highlight_default = val;
422 return 0;
423 }
424
425 if (git_color_config(var, value, cb) < 0)
426 return -1;
427
428 return git_diff_basic_config(var, value, cb);
429 }
430
431 int git_diff_basic_config(const char *var, const char *value, void *cb)
432 {
433 const char *name;
434
435 if (!strcmp(var, "diff.renamelimit")) {
436 diff_rename_limit_default = git_config_int(var, value);
437 return 0;
438 }
439
440 if (userdiff_config(var, value) < 0)
441 return -1;
442
443 if (skip_prefix(var, "diff.color.", &name) ||
444 skip_prefix(var, "color.diff.", &name)) {
445 int slot = parse_diff_color_slot(name);
446 if (slot < 0)
447 return 0;
448 if (!value)
449 return config_error_nonbool(var);
450 return color_parse(value, diff_colors[slot]);
451 }
452
453 /* like GNU diff's --suppress-blank-empty option */
454 if (!strcmp(var, "diff.suppressblankempty") ||
455 /* for backwards compatibility */
456 !strcmp(var, "diff.suppress-blank-empty")) {
457 diff_suppress_blank_empty = git_config_bool(var, value);
458 return 0;
459 }
460
461 if (!strcmp(var, "diff.dirstat")) {
462 struct strbuf errmsg = STRBUF_INIT;
463 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
464 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
465 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
466 errmsg.buf);
467 strbuf_release(&errmsg);
468 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
469 return 0;
470 }
471
472 if (git_diff_heuristic_config(var, value, cb) < 0)
473 return -1;
474
475 return git_default_config(var, value, cb);
476 }
477
478 static char *quote_two(const char *one, const char *two)
479 {
480 int need_one = quote_c_style(one, NULL, NULL, 1);
481 int need_two = quote_c_style(two, NULL, NULL, 1);
482 struct strbuf res = STRBUF_INIT;
483
484 if (need_one + need_two) {
485 strbuf_addch(&res, '"');
486 quote_c_style(one, &res, NULL, 1);
487 quote_c_style(two, &res, NULL, 1);
488 strbuf_addch(&res, '"');
489 } else {
490 strbuf_addstr(&res, one);
491 strbuf_addstr(&res, two);
492 }
493 return strbuf_detach(&res, NULL);
494 }
495
496 static const char *external_diff(void)
497 {
498 static const char *external_diff_cmd = NULL;
499 static int done_preparing = 0;
500
501 if (done_preparing)
502 return external_diff_cmd;
503 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
504 if (!external_diff_cmd)
505 external_diff_cmd = external_diff_cmd_cfg;
506 done_preparing = 1;
507 return external_diff_cmd;
508 }
509
510 /*
511 * Keep track of files used for diffing. Sometimes such an entry
512 * refers to a temporary file, sometimes to an existing file, and
513 * sometimes to "/dev/null".
514 */
515 static struct diff_tempfile {
516 /*
517 * filename external diff should read from, or NULL if this
518 * entry is currently not in use:
519 */
520 const char *name;
521
522 char hex[GIT_MAX_HEXSZ + 1];
523 char mode[10];
524
525 /*
526 * If this diff_tempfile instance refers to a temporary file,
527 * this tempfile object is used to manage its lifetime.
528 */
529 struct tempfile *tempfile;
530 } diff_temp[2];
531
532 struct emit_callback {
533 int color_diff;
534 unsigned ws_rule;
535 int blank_at_eof_in_preimage;
536 int blank_at_eof_in_postimage;
537 int lno_in_preimage;
538 int lno_in_postimage;
539 const char **label_path;
540 struct diff_words_data *diff_words;
541 struct diff_options *opt;
542 struct strbuf *header;
543 };
544
545 static int count_lines(const char *data, int size)
546 {
547 int count, ch, completely_empty = 1, nl_just_seen = 0;
548 count = 0;
549 while (0 < size--) {
550 ch = *data++;
551 if (ch == '\n') {
552 count++;
553 nl_just_seen = 1;
554 completely_empty = 0;
555 }
556 else {
557 nl_just_seen = 0;
558 completely_empty = 0;
559 }
560 }
561 if (completely_empty)
562 return 0;
563 if (!nl_just_seen)
564 count++; /* no trailing newline */
565 return count;
566 }
567
568 static int fill_mmfile(struct repository *r, mmfile_t *mf,
569 struct diff_filespec *one)
570 {
571 if (!DIFF_FILE_VALID(one)) {
572 mf->ptr = (char *)""; /* does not matter */
573 mf->size = 0;
574 return 0;
575 }
576 else if (diff_populate_filespec(r, one, 0))
577 return -1;
578
579 mf->ptr = one->data;
580 mf->size = one->size;
581 return 0;
582 }
583
584 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
585 static unsigned long diff_filespec_size(struct repository *r,
586 struct diff_filespec *one)
587 {
588 if (!DIFF_FILE_VALID(one))
589 return 0;
590 diff_populate_filespec(r, one, CHECK_SIZE_ONLY);
591 return one->size;
592 }
593
594 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
595 {
596 char *ptr = mf->ptr;
597 long size = mf->size;
598 int cnt = 0;
599
600 if (!size)
601 return cnt;
602 ptr += size - 1; /* pointing at the very end */
603 if (*ptr != '\n')
604 ; /* incomplete line */
605 else
606 ptr--; /* skip the last LF */
607 while (mf->ptr < ptr) {
608 char *prev_eol;
609 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
610 if (*prev_eol == '\n')
611 break;
612 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
613 break;
614 cnt++;
615 ptr = prev_eol - 1;
616 }
617 return cnt;
618 }
619
620 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
621 struct emit_callback *ecbdata)
622 {
623 int l1, l2, at;
624 unsigned ws_rule = ecbdata->ws_rule;
625 l1 = count_trailing_blank(mf1, ws_rule);
626 l2 = count_trailing_blank(mf2, ws_rule);
627 if (l2 <= l1) {
628 ecbdata->blank_at_eof_in_preimage = 0;
629 ecbdata->blank_at_eof_in_postimage = 0;
630 return;
631 }
632 at = count_lines(mf1->ptr, mf1->size);
633 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
634
635 at = count_lines(mf2->ptr, mf2->size);
636 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
637 }
638
639 static void emit_line_0(struct diff_options *o,
640 const char *set_sign, const char *set, unsigned reverse, const char *reset,
641 int first, const char *line, int len)
642 {
643 int has_trailing_newline, has_trailing_carriage_return;
644 int needs_reset = 0; /* at the end of the line */
645 FILE *file = o->file;
646
647 fputs(diff_line_prefix(o), file);
648
649 has_trailing_newline = (len > 0 && line[len-1] == '\n');
650 if (has_trailing_newline)
651 len--;
652
653 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
654 if (has_trailing_carriage_return)
655 len--;
656
657 if (!len && !first)
658 goto end_of_line;
659
660 if (reverse && want_color(o->use_color)) {
661 fputs(GIT_COLOR_REVERSE, file);
662 needs_reset = 1;
663 }
664
665 if (set_sign) {
666 fputs(set_sign, file);
667 needs_reset = 1;
668 }
669
670 if (first)
671 fputc(first, file);
672
673 if (!len)
674 goto end_of_line;
675
676 if (set) {
677 if (set_sign && set != set_sign)
678 fputs(reset, file);
679 fputs(set, file);
680 needs_reset = 1;
681 }
682 fwrite(line, len, 1, file);
683 needs_reset = 1; /* 'line' may contain color codes. */
684
685 end_of_line:
686 if (needs_reset)
687 fputs(reset, file);
688 if (has_trailing_carriage_return)
689 fputc('\r', file);
690 if (has_trailing_newline)
691 fputc('\n', file);
692 }
693
694 static void emit_line(struct diff_options *o, const char *set, const char *reset,
695 const char *line, int len)
696 {
697 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
698 }
699
700 enum diff_symbol {
701 DIFF_SYMBOL_BINARY_DIFF_HEADER,
702 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
703 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
704 DIFF_SYMBOL_BINARY_DIFF_BODY,
705 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
706 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
707 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
708 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
709 DIFF_SYMBOL_STATS_LINE,
710 DIFF_SYMBOL_WORD_DIFF,
711 DIFF_SYMBOL_STAT_SEP,
712 DIFF_SYMBOL_SUMMARY,
713 DIFF_SYMBOL_SUBMODULE_ADD,
714 DIFF_SYMBOL_SUBMODULE_DEL,
715 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
716 DIFF_SYMBOL_SUBMODULE_MODIFIED,
717 DIFF_SYMBOL_SUBMODULE_HEADER,
718 DIFF_SYMBOL_SUBMODULE_ERROR,
719 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
720 DIFF_SYMBOL_REWRITE_DIFF,
721 DIFF_SYMBOL_BINARY_FILES,
722 DIFF_SYMBOL_HEADER,
723 DIFF_SYMBOL_FILEPAIR_PLUS,
724 DIFF_SYMBOL_FILEPAIR_MINUS,
725 DIFF_SYMBOL_WORDS_PORCELAIN,
726 DIFF_SYMBOL_WORDS,
727 DIFF_SYMBOL_CONTEXT,
728 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
729 DIFF_SYMBOL_PLUS,
730 DIFF_SYMBOL_MINUS,
731 DIFF_SYMBOL_NO_LF_EOF,
732 DIFF_SYMBOL_CONTEXT_FRAGINFO,
733 DIFF_SYMBOL_CONTEXT_MARKER,
734 DIFF_SYMBOL_SEPARATOR
735 };
736 /*
737 * Flags for content lines:
738 * 0..12 are whitespace rules
739 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
740 * 16 is marking if the line is blank at EOF
741 */
742 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
743 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
744 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
745 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
746 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
747
748 /*
749 * This struct is used when we need to buffer the output of the diff output.
750 *
751 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
752 * into the pre/post image file. This pointer could be a union with the
753 * line pointer. By storing an offset into the file instead of the literal line,
754 * we can decrease the memory footprint for the buffered output. At first we
755 * may want to only have indirection for the content lines, but we could also
756 * enhance the state for emitting prefabricated lines, e.g. the similarity
757 * score line or hunk/file headers would only need to store a number or path
758 * and then the output can be constructed later on depending on state.
759 */
760 struct emitted_diff_symbol {
761 const char *line;
762 int len;
763 int flags;
764 int indent_off; /* Offset to first non-whitespace character */
765 int indent_width; /* The visual width of the indentation */
766 enum diff_symbol s;
767 };
768 #define EMITTED_DIFF_SYMBOL_INIT {NULL}
769
770 struct emitted_diff_symbols {
771 struct emitted_diff_symbol *buf;
772 int nr, alloc;
773 };
774 #define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
775
776 static void append_emitted_diff_symbol(struct diff_options *o,
777 struct emitted_diff_symbol *e)
778 {
779 struct emitted_diff_symbol *f;
780
781 ALLOC_GROW(o->emitted_symbols->buf,
782 o->emitted_symbols->nr + 1,
783 o->emitted_symbols->alloc);
784 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
785
786 memcpy(f, e, sizeof(struct emitted_diff_symbol));
787 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
788 }
789
790 struct moved_entry {
791 struct hashmap_entry ent;
792 const struct emitted_diff_symbol *es;
793 struct moved_entry *next_line;
794 };
795
796 struct moved_block {
797 struct moved_entry *match;
798 int wsd; /* The whitespace delta of this block */
799 };
800
801 static void moved_block_clear(struct moved_block *b)
802 {
803 memset(b, 0, sizeof(*b));
804 }
805
806 #define INDENT_BLANKLINE INT_MIN
807
808 static void fill_es_indent_data(struct emitted_diff_symbol *es)
809 {
810 unsigned int off = 0, i;
811 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
812 const char *s = es->line;
813 const int len = es->len;
814
815 /* skip any \v \f \r at start of indentation */
816 while (s[off] == '\f' || s[off] == '\v' ||
817 (s[off] == '\r' && off < len - 1))
818 off++;
819
820 /* calculate the visual width of indentation */
821 while(1) {
822 if (s[off] == ' ') {
823 width++;
824 off++;
825 } else if (s[off] == '\t') {
826 width += tab_width - (width % tab_width);
827 while (s[++off] == '\t')
828 width += tab_width;
829 } else {
830 break;
831 }
832 }
833
834 /* check if this line is blank */
835 for (i = off; i < len; i++)
836 if (!isspace(s[i]))
837 break;
838
839 if (i == len) {
840 es->indent_width = INDENT_BLANKLINE;
841 es->indent_off = len;
842 } else {
843 es->indent_off = off;
844 es->indent_width = width;
845 }
846 }
847
848 static int compute_ws_delta(const struct emitted_diff_symbol *a,
849 const struct emitted_diff_symbol *b,
850 int *out)
851 {
852 int a_len = a->len,
853 b_len = b->len,
854 a_off = a->indent_off,
855 a_width = a->indent_width,
856 b_off = b->indent_off,
857 b_width = b->indent_width;
858 int delta;
859
860 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE) {
861 *out = INDENT_BLANKLINE;
862 return 1;
863 }
864
865 if (a->s == DIFF_SYMBOL_PLUS)
866 delta = a_width - b_width;
867 else
868 delta = b_width - a_width;
869
870 if (a_len - a_off != b_len - b_off ||
871 memcmp(a->line + a_off, b->line + b_off, a_len - a_off))
872 return 0;
873
874 *out = delta;
875
876 return 1;
877 }
878
879 static int cmp_in_block_with_wsd(const struct diff_options *o,
880 const struct moved_entry *cur,
881 const struct moved_entry *match,
882 struct moved_block *pmb,
883 int n)
884 {
885 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
886 int al = cur->es->len, bl = match->es->len, cl = l->len;
887 const char *a = cur->es->line,
888 *b = match->es->line,
889 *c = l->line;
890 int a_off = cur->es->indent_off,
891 a_width = cur->es->indent_width,
892 c_off = l->indent_off,
893 c_width = l->indent_width;
894 int delta;
895
896 /*
897 * We need to check if 'cur' is equal to 'match'. As those
898 * are from the same (+/-) side, we do not need to adjust for
899 * indent changes. However these were found using fuzzy
900 * matching so we do have to check if they are equal. Here we
901 * just check the lengths. We delay calling memcmp() to check
902 * the contents until later as if the length comparison for a
903 * and c fails we can avoid the call all together.
904 */
905 if (al != bl)
906 return 1;
907
908 /* If 'l' and 'cur' are both blank then they match. */
909 if (a_width == INDENT_BLANKLINE && c_width == INDENT_BLANKLINE)
910 return 0;
911
912 /*
913 * The indent changes of the block are known and stored in pmb->wsd;
914 * however we need to check if the indent changes of the current line
915 * match those of the current block and that the text of 'l' and 'cur'
916 * after the indentation match.
917 */
918 if (cur->es->s == DIFF_SYMBOL_PLUS)
919 delta = a_width - c_width;
920 else
921 delta = c_width - a_width;
922
923 /*
924 * If the previous lines of this block were all blank then set its
925 * whitespace delta.
926 */
927 if (pmb->wsd == INDENT_BLANKLINE)
928 pmb->wsd = delta;
929
930 return !(delta == pmb->wsd && al - a_off == cl - c_off &&
931 !memcmp(a, b, al) && !
932 memcmp(a + a_off, c + c_off, al - a_off));
933 }
934
935 static int moved_entry_cmp(const void *hashmap_cmp_fn_data,
936 const void *entry,
937 const void *entry_or_key,
938 const void *keydata)
939 {
940 const struct diff_options *diffopt = hashmap_cmp_fn_data;
941 const struct moved_entry *a = entry;
942 const struct moved_entry *b = entry_or_key;
943 unsigned flags = diffopt->color_moved_ws_handling
944 & XDF_WHITESPACE_FLAGS;
945
946 if (diffopt->color_moved_ws_handling &
947 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
948 /*
949 * As there is not specific white space config given,
950 * we'd need to check for a new block, so ignore all
951 * white space. The setup of the white space
952 * configuration for the next block is done else where
953 */
954 flags |= XDF_IGNORE_WHITESPACE;
955
956 return !xdiff_compare_lines(a->es->line, a->es->len,
957 b->es->line, b->es->len,
958 flags);
959 }
960
961 static struct moved_entry *prepare_entry(struct diff_options *o,
962 int line_no)
963 {
964 struct moved_entry *ret = xmalloc(sizeof(*ret));
965 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
966 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
967
968 ret->ent.hash = xdiff_hash_string(l->line, l->len, flags);
969 ret->es = l;
970 ret->next_line = NULL;
971
972 return ret;
973 }
974
975 static void add_lines_to_move_detection(struct diff_options *o,
976 struct hashmap *add_lines,
977 struct hashmap *del_lines)
978 {
979 struct moved_entry *prev_line = NULL;
980
981 int n;
982 for (n = 0; n < o->emitted_symbols->nr; n++) {
983 struct hashmap *hm;
984 struct moved_entry *key;
985
986 switch (o->emitted_symbols->buf[n].s) {
987 case DIFF_SYMBOL_PLUS:
988 hm = add_lines;
989 break;
990 case DIFF_SYMBOL_MINUS:
991 hm = del_lines;
992 break;
993 default:
994 prev_line = NULL;
995 continue;
996 }
997
998 if (o->color_moved_ws_handling &
999 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1000 fill_es_indent_data(&o->emitted_symbols->buf[n]);
1001 key = prepare_entry(o, n);
1002 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
1003 prev_line->next_line = key;
1004
1005 hashmap_add(hm, key);
1006 prev_line = key;
1007 }
1008 }
1009
1010 static void pmb_advance_or_null(struct diff_options *o,
1011 struct moved_entry *match,
1012 struct hashmap *hm,
1013 struct moved_block *pmb,
1014 int pmb_nr)
1015 {
1016 int i;
1017 for (i = 0; i < pmb_nr; i++) {
1018 struct moved_entry *prev = pmb[i].match;
1019 struct moved_entry *cur = (prev && prev->next_line) ?
1020 prev->next_line : NULL;
1021 if (cur && !hm->cmpfn(o, cur, match, NULL)) {
1022 pmb[i].match = cur;
1023 } else {
1024 pmb[i].match = NULL;
1025 }
1026 }
1027 }
1028
1029 static void pmb_advance_or_null_multi_match(struct diff_options *o,
1030 struct moved_entry *match,
1031 struct hashmap *hm,
1032 struct moved_block *pmb,
1033 int pmb_nr, int n)
1034 {
1035 int i;
1036 char *got_match = xcalloc(1, pmb_nr);
1037
1038 for (; match; match = hashmap_get_next(hm, match)) {
1039 for (i = 0; i < pmb_nr; i++) {
1040 struct moved_entry *prev = pmb[i].match;
1041 struct moved_entry *cur = (prev && prev->next_line) ?
1042 prev->next_line : NULL;
1043 if (!cur)
1044 continue;
1045 if (!cmp_in_block_with_wsd(o, cur, match, &pmb[i], n))
1046 got_match[i] |= 1;
1047 }
1048 }
1049
1050 for (i = 0; i < pmb_nr; i++) {
1051 if (got_match[i]) {
1052 /* Advance to the next line */
1053 pmb[i].match = pmb[i].match->next_line;
1054 } else {
1055 moved_block_clear(&pmb[i]);
1056 }
1057 }
1058
1059 free(got_match);
1060 }
1061
1062 static int shrink_potential_moved_blocks(struct moved_block *pmb,
1063 int pmb_nr)
1064 {
1065 int lp, rp;
1066
1067 /* Shrink the set of potential block to the remaining running */
1068 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
1069 while (lp < pmb_nr && pmb[lp].match)
1070 lp++;
1071 /* lp points at the first NULL now */
1072
1073 while (rp > -1 && !pmb[rp].match)
1074 rp--;
1075 /* rp points at the last non-NULL */
1076
1077 if (lp < pmb_nr && rp > -1 && lp < rp) {
1078 pmb[lp] = pmb[rp];
1079 memset(&pmb[rp], 0, sizeof(pmb[rp]));
1080 rp--;
1081 lp++;
1082 }
1083 }
1084
1085 /* Remember the number of running sets */
1086 return rp + 1;
1087 }
1088
1089 /*
1090 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1091 *
1092 * Otherwise, if the last block has fewer alphanumeric characters than
1093 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1094 * that block.
1095 *
1096 * The last block consists of the (n - block_length)'th line up to but not
1097 * including the nth line.
1098 *
1099 * Returns 0 if the last block is empty or is unset by this function, non zero
1100 * otherwise.
1101 *
1102 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1103 * Think of a way to unify them.
1104 */
1105 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1106 {
1107 int i, alnum_count = 0;
1108 if (o->color_moved == COLOR_MOVED_PLAIN)
1109 return block_length;
1110 for (i = 1; i < block_length + 1; i++) {
1111 const char *c = o->emitted_symbols->buf[n - i].line;
1112 for (; *c; c++) {
1113 if (!isalnum(*c))
1114 continue;
1115 alnum_count++;
1116 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1117 return 1;
1118 }
1119 }
1120 for (i = 1; i < block_length + 1; i++)
1121 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
1122 return 0;
1123 }
1124
1125 /* Find blocks of moved code, delegate actual coloring decision to helper */
1126 static void mark_color_as_moved(struct diff_options *o,
1127 struct hashmap *add_lines,
1128 struct hashmap *del_lines)
1129 {
1130 struct moved_block *pmb = NULL; /* potentially moved blocks */
1131 int pmb_nr = 0, pmb_alloc = 0;
1132 int n, flipped_block = 0, block_length = 0;
1133
1134
1135 for (n = 0; n < o->emitted_symbols->nr; n++) {
1136 struct hashmap *hm = NULL;
1137 struct moved_entry *key;
1138 struct moved_entry *match = NULL;
1139 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1140 enum diff_symbol last_symbol = 0;
1141
1142 switch (l->s) {
1143 case DIFF_SYMBOL_PLUS:
1144 hm = del_lines;
1145 key = prepare_entry(o, n);
1146 match = hashmap_get(hm, key, NULL);
1147 free(key);
1148 break;
1149 case DIFF_SYMBOL_MINUS:
1150 hm = add_lines;
1151 key = prepare_entry(o, n);
1152 match = hashmap_get(hm, key, NULL);
1153 free(key);
1154 break;
1155 default:
1156 flipped_block = 0;
1157 }
1158
1159 if (!match) {
1160 int i;
1161
1162 adjust_last_block(o, n, block_length);
1163 for(i = 0; i < pmb_nr; i++)
1164 moved_block_clear(&pmb[i]);
1165 pmb_nr = 0;
1166 block_length = 0;
1167 flipped_block = 0;
1168 last_symbol = l->s;
1169 continue;
1170 }
1171
1172 if (o->color_moved == COLOR_MOVED_PLAIN) {
1173 last_symbol = l->s;
1174 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1175 continue;
1176 }
1177
1178 if (o->color_moved_ws_handling &
1179 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1180 pmb_advance_or_null_multi_match(o, match, hm, pmb, pmb_nr, n);
1181 else
1182 pmb_advance_or_null(o, match, hm, pmb, pmb_nr);
1183
1184 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
1185
1186 if (pmb_nr == 0) {
1187 /*
1188 * The current line is the start of a new block.
1189 * Setup the set of potential blocks.
1190 */
1191 for (; match; match = hashmap_get_next(hm, match)) {
1192 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1193 if (o->color_moved_ws_handling &
1194 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) {
1195 if (compute_ws_delta(l, match->es,
1196 &pmb[pmb_nr].wsd))
1197 pmb[pmb_nr++].match = match;
1198 } else {
1199 pmb[pmb_nr].wsd = 0;
1200 pmb[pmb_nr++].match = match;
1201 }
1202 }
1203
1204 if (adjust_last_block(o, n, block_length) &&
1205 pmb_nr && last_symbol != l->s)
1206 flipped_block = (flipped_block + 1) % 2;
1207 else
1208 flipped_block = 0;
1209
1210 block_length = 0;
1211 }
1212
1213 if (pmb_nr) {
1214 block_length++;
1215 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1216 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1217 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1218 }
1219 last_symbol = l->s;
1220 }
1221 adjust_last_block(o, n, block_length);
1222
1223 for(n = 0; n < pmb_nr; n++)
1224 moved_block_clear(&pmb[n]);
1225 free(pmb);
1226 }
1227
1228 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1229 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1230 static void dim_moved_lines(struct diff_options *o)
1231 {
1232 int n;
1233 for (n = 0; n < o->emitted_symbols->nr; n++) {
1234 struct emitted_diff_symbol *prev = (n != 0) ?
1235 &o->emitted_symbols->buf[n - 1] : NULL;
1236 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1237 struct emitted_diff_symbol *next =
1238 (n < o->emitted_symbols->nr - 1) ?
1239 &o->emitted_symbols->buf[n + 1] : NULL;
1240
1241 /* Not a plus or minus line? */
1242 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1243 continue;
1244
1245 /* Not a moved line? */
1246 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1247 continue;
1248
1249 /*
1250 * If prev or next are not a plus or minus line,
1251 * pretend they don't exist
1252 */
1253 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1254 prev->s != DIFF_SYMBOL_MINUS)
1255 prev = NULL;
1256 if (next && next->s != DIFF_SYMBOL_PLUS &&
1257 next->s != DIFF_SYMBOL_MINUS)
1258 next = NULL;
1259
1260 /* Inside a block? */
1261 if ((prev &&
1262 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1263 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1264 (next &&
1265 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1266 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1267 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1268 continue;
1269 }
1270
1271 /* Check if we are at an interesting bound: */
1272 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1273 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1274 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1275 continue;
1276 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1277 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1278 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1279 continue;
1280
1281 /*
1282 * The boundary to prev and next are not interesting,
1283 * so this line is not interesting as a whole
1284 */
1285 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1286 }
1287 }
1288
1289 static void emit_line_ws_markup(struct diff_options *o,
1290 const char *set_sign, const char *set,
1291 const char *reset,
1292 int sign_index, const char *line, int len,
1293 unsigned ws_rule, int blank_at_eof)
1294 {
1295 const char *ws = NULL;
1296 int sign = o->output_indicators[sign_index];
1297
1298 if (o->ws_error_highlight & ws_rule) {
1299 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1300 if (!*ws)
1301 ws = NULL;
1302 }
1303
1304 if (!ws && !set_sign)
1305 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1306 else if (!ws) {
1307 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1308 } else if (blank_at_eof)
1309 /* Blank line at EOF - paint '+' as well */
1310 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1311 else {
1312 /* Emit just the prefix, then the rest. */
1313 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1314 sign, "", 0);
1315 ws_check_emit(line, len, ws_rule,
1316 o->file, set, reset, ws);
1317 }
1318 }
1319
1320 static void emit_diff_symbol_from_struct(struct diff_options *o,
1321 struct emitted_diff_symbol *eds)
1322 {
1323 static const char *nneof = " No newline at end of file\n";
1324 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1325 struct strbuf sb = STRBUF_INIT;
1326
1327 enum diff_symbol s = eds->s;
1328 const char *line = eds->line;
1329 int len = eds->len;
1330 unsigned flags = eds->flags;
1331
1332 switch (s) {
1333 case DIFF_SYMBOL_NO_LF_EOF:
1334 context = diff_get_color_opt(o, DIFF_CONTEXT);
1335 reset = diff_get_color_opt(o, DIFF_RESET);
1336 putc('\n', o->file);
1337 emit_line_0(o, context, NULL, 0, reset, '\\',
1338 nneof, strlen(nneof));
1339 break;
1340 case DIFF_SYMBOL_SUBMODULE_HEADER:
1341 case DIFF_SYMBOL_SUBMODULE_ERROR:
1342 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1343 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1344 case DIFF_SYMBOL_SUMMARY:
1345 case DIFF_SYMBOL_STATS_LINE:
1346 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1347 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1348 emit_line(o, "", "", line, len);
1349 break;
1350 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1351 case DIFF_SYMBOL_CONTEXT_MARKER:
1352 context = diff_get_color_opt(o, DIFF_CONTEXT);
1353 reset = diff_get_color_opt(o, DIFF_RESET);
1354 emit_line(o, context, reset, line, len);
1355 break;
1356 case DIFF_SYMBOL_SEPARATOR:
1357 fprintf(o->file, "%s%c",
1358 diff_line_prefix(o),
1359 o->line_termination);
1360 break;
1361 case DIFF_SYMBOL_CONTEXT:
1362 set = diff_get_color_opt(o, DIFF_CONTEXT);
1363 reset = diff_get_color_opt(o, DIFF_RESET);
1364 set_sign = NULL;
1365 if (o->flags.dual_color_diffed_diffs) {
1366 char c = !len ? 0 : line[0];
1367
1368 if (c == '+')
1369 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1370 else if (c == '@')
1371 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1372 else if (c == '-')
1373 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1374 }
1375 emit_line_ws_markup(o, set_sign, set, reset,
1376 OUTPUT_INDICATOR_CONTEXT, line, len,
1377 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1378 break;
1379 case DIFF_SYMBOL_PLUS:
1380 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1381 DIFF_SYMBOL_MOVED_LINE_ALT |
1382 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1383 case DIFF_SYMBOL_MOVED_LINE |
1384 DIFF_SYMBOL_MOVED_LINE_ALT |
1385 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1386 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1387 break;
1388 case DIFF_SYMBOL_MOVED_LINE |
1389 DIFF_SYMBOL_MOVED_LINE_ALT:
1390 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1391 break;
1392 case DIFF_SYMBOL_MOVED_LINE |
1393 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1394 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1395 break;
1396 case DIFF_SYMBOL_MOVED_LINE:
1397 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1398 break;
1399 default:
1400 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1401 }
1402 reset = diff_get_color_opt(o, DIFF_RESET);
1403 if (!o->flags.dual_color_diffed_diffs)
1404 set_sign = NULL;
1405 else {
1406 char c = !len ? 0 : line[0];
1407
1408 set_sign = set;
1409 if (c == '-')
1410 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1411 else if (c == '@')
1412 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1413 else if (c == '+')
1414 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1415 else
1416 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1417 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1418 }
1419 emit_line_ws_markup(o, set_sign, set, reset,
1420 OUTPUT_INDICATOR_NEW, line, len,
1421 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1422 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1423 break;
1424 case DIFF_SYMBOL_MINUS:
1425 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1426 DIFF_SYMBOL_MOVED_LINE_ALT |
1427 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1428 case DIFF_SYMBOL_MOVED_LINE |
1429 DIFF_SYMBOL_MOVED_LINE_ALT |
1430 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1431 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1432 break;
1433 case DIFF_SYMBOL_MOVED_LINE |
1434 DIFF_SYMBOL_MOVED_LINE_ALT:
1435 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1436 break;
1437 case DIFF_SYMBOL_MOVED_LINE |
1438 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1439 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1440 break;
1441 case DIFF_SYMBOL_MOVED_LINE:
1442 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1443 break;
1444 default:
1445 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1446 }
1447 reset = diff_get_color_opt(o, DIFF_RESET);
1448 if (!o->flags.dual_color_diffed_diffs)
1449 set_sign = NULL;
1450 else {
1451 char c = !len ? 0 : line[0];
1452
1453 set_sign = set;
1454 if (c == '+')
1455 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1456 else if (c == '@')
1457 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1458 else if (c == '-')
1459 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1460 else
1461 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1462 }
1463 emit_line_ws_markup(o, set_sign, set, reset,
1464 OUTPUT_INDICATOR_OLD, line, len,
1465 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1466 break;
1467 case DIFF_SYMBOL_WORDS_PORCELAIN:
1468 context = diff_get_color_opt(o, DIFF_CONTEXT);
1469 reset = diff_get_color_opt(o, DIFF_RESET);
1470 emit_line(o, context, reset, line, len);
1471 fputs("~\n", o->file);
1472 break;
1473 case DIFF_SYMBOL_WORDS:
1474 context = diff_get_color_opt(o, DIFF_CONTEXT);
1475 reset = diff_get_color_opt(o, DIFF_RESET);
1476 /*
1477 * Skip the prefix character, if any. With
1478 * diff_suppress_blank_empty, there may be
1479 * none.
1480 */
1481 if (line[0] != '\n') {
1482 line++;
1483 len--;
1484 }
1485 emit_line(o, context, reset, line, len);
1486 break;
1487 case DIFF_SYMBOL_FILEPAIR_PLUS:
1488 meta = diff_get_color_opt(o, DIFF_METAINFO);
1489 reset = diff_get_color_opt(o, DIFF_RESET);
1490 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1491 line, reset,
1492 strchr(line, ' ') ? "\t" : "");
1493 break;
1494 case DIFF_SYMBOL_FILEPAIR_MINUS:
1495 meta = diff_get_color_opt(o, DIFF_METAINFO);
1496 reset = diff_get_color_opt(o, DIFF_RESET);
1497 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1498 line, reset,
1499 strchr(line, ' ') ? "\t" : "");
1500 break;
1501 case DIFF_SYMBOL_BINARY_FILES:
1502 case DIFF_SYMBOL_HEADER:
1503 fprintf(o->file, "%s", line);
1504 break;
1505 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1506 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1507 break;
1508 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1509 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1510 break;
1511 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1512 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1513 break;
1514 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1515 fputs(diff_line_prefix(o), o->file);
1516 fputc('\n', o->file);
1517 break;
1518 case DIFF_SYMBOL_REWRITE_DIFF:
1519 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1520 reset = diff_get_color_opt(o, DIFF_RESET);
1521 emit_line(o, fraginfo, reset, line, len);
1522 break;
1523 case DIFF_SYMBOL_SUBMODULE_ADD:
1524 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1525 reset = diff_get_color_opt(o, DIFF_RESET);
1526 emit_line(o, set, reset, line, len);
1527 break;
1528 case DIFF_SYMBOL_SUBMODULE_DEL:
1529 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1530 reset = diff_get_color_opt(o, DIFF_RESET);
1531 emit_line(o, set, reset, line, len);
1532 break;
1533 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1534 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1535 diff_line_prefix(o), line);
1536 break;
1537 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1538 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1539 diff_line_prefix(o), line);
1540 break;
1541 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1542 emit_line(o, "", "", " 0 files changed\n",
1543 strlen(" 0 files changed\n"));
1544 break;
1545 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1546 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1547 break;
1548 case DIFF_SYMBOL_WORD_DIFF:
1549 fprintf(o->file, "%.*s", len, line);
1550 break;
1551 case DIFF_SYMBOL_STAT_SEP:
1552 fputs(o->stat_sep, o->file);
1553 break;
1554 default:
1555 BUG("unknown diff symbol");
1556 }
1557 strbuf_release(&sb);
1558 }
1559
1560 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1561 const char *line, int len, unsigned flags)
1562 {
1563 struct emitted_diff_symbol e = {line, len, flags, 0, 0, s};
1564
1565 if (o->emitted_symbols)
1566 append_emitted_diff_symbol(o, &e);
1567 else
1568 emit_diff_symbol_from_struct(o, &e);
1569 }
1570
1571 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1572 {
1573 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1574 }
1575
1576 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1577 {
1578 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1579 }
1580
1581 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1582 {
1583 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1584 path, strlen(path), 0);
1585 }
1586
1587 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1588 {
1589 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1590 path, strlen(path), 0);
1591 }
1592
1593 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1594 {
1595 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1596 header, strlen(header), 0);
1597 }
1598
1599 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1600 {
1601 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1602 }
1603
1604 void diff_emit_submodule_pipethrough(struct diff_options *o,
1605 const char *line, int len)
1606 {
1607 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1608 }
1609
1610 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1611 {
1612 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1613 ecbdata->blank_at_eof_in_preimage &&
1614 ecbdata->blank_at_eof_in_postimage &&
1615 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1616 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1617 return 0;
1618 return ws_blank_line(line, len, ecbdata->ws_rule);
1619 }
1620
1621 static void emit_add_line(struct emit_callback *ecbdata,
1622 const char *line, int len)
1623 {
1624 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1625 if (new_blank_line_at_eof(ecbdata, line, len))
1626 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1627
1628 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1629 }
1630
1631 static void emit_del_line(struct emit_callback *ecbdata,
1632 const char *line, int len)
1633 {
1634 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1635 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1636 }
1637
1638 static void emit_context_line(struct emit_callback *ecbdata,
1639 const char *line, int len)
1640 {
1641 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1642 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1643 }
1644
1645 static void emit_hunk_header(struct emit_callback *ecbdata,
1646 const char *line, int len)
1647 {
1648 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1649 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1650 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1651 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1652 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1653 static const char atat[2] = { '@', '@' };
1654 const char *cp, *ep;
1655 struct strbuf msgbuf = STRBUF_INIT;
1656 int org_len = len;
1657 int i = 1;
1658
1659 /*
1660 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1661 * it always is at least 10 bytes long.
1662 */
1663 if (len < 10 ||
1664 memcmp(line, atat, 2) ||
1665 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1666 emit_diff_symbol(ecbdata->opt,
1667 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1668 return;
1669 }
1670 ep += 2; /* skip over @@ */
1671
1672 /* The hunk header in fraginfo color */
1673 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1674 strbuf_addstr(&msgbuf, reverse);
1675 strbuf_addstr(&msgbuf, frag);
1676 strbuf_add(&msgbuf, line, ep - line);
1677 strbuf_addstr(&msgbuf, reset);
1678
1679 /*
1680 * trailing "\r\n"
1681 */
1682 for ( ; i < 3; i++)
1683 if (line[len - i] == '\r' || line[len - i] == '\n')
1684 len--;
1685
1686 /* blank before the func header */
1687 for (cp = ep; ep - line < len; ep++)
1688 if (*ep != ' ' && *ep != '\t')
1689 break;
1690 if (ep != cp) {
1691 strbuf_addstr(&msgbuf, context);
1692 strbuf_add(&msgbuf, cp, ep - cp);
1693 strbuf_addstr(&msgbuf, reset);
1694 }
1695
1696 if (ep < line + len) {
1697 strbuf_addstr(&msgbuf, func);
1698 strbuf_add(&msgbuf, ep, line + len - ep);
1699 strbuf_addstr(&msgbuf, reset);
1700 }
1701
1702 strbuf_add(&msgbuf, line + len, org_len - len);
1703 strbuf_complete_line(&msgbuf);
1704 emit_diff_symbol(ecbdata->opt,
1705 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1706 strbuf_release(&msgbuf);
1707 }
1708
1709 static struct diff_tempfile *claim_diff_tempfile(void)
1710 {
1711 int i;
1712 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1713 if (!diff_temp[i].name)
1714 return diff_temp + i;
1715 BUG("diff is failing to clean up its tempfiles");
1716 }
1717
1718 static void remove_tempfile(void)
1719 {
1720 int i;
1721 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1722 if (is_tempfile_active(diff_temp[i].tempfile))
1723 delete_tempfile(&diff_temp[i].tempfile);
1724 diff_temp[i].name = NULL;
1725 }
1726 }
1727
1728 static void add_line_count(struct strbuf *out, int count)
1729 {
1730 switch (count) {
1731 case 0:
1732 strbuf_addstr(out, "0,0");
1733 break;
1734 case 1:
1735 strbuf_addstr(out, "1");
1736 break;
1737 default:
1738 strbuf_addf(out, "1,%d", count);
1739 break;
1740 }
1741 }
1742
1743 static void emit_rewrite_lines(struct emit_callback *ecb,
1744 int prefix, const char *data, int size)
1745 {
1746 const char *endp = NULL;
1747
1748 while (0 < size) {
1749 int len;
1750
1751 endp = memchr(data, '\n', size);
1752 len = endp ? (endp - data + 1) : size;
1753 if (prefix != '+') {
1754 ecb->lno_in_preimage++;
1755 emit_del_line(ecb, data, len);
1756 } else {
1757 ecb->lno_in_postimage++;
1758 emit_add_line(ecb, data, len);
1759 }
1760 size -= len;
1761 data += len;
1762 }
1763 if (!endp)
1764 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1765 }
1766
1767 static void emit_rewrite_diff(const char *name_a,
1768 const char *name_b,
1769 struct diff_filespec *one,
1770 struct diff_filespec *two,
1771 struct userdiff_driver *textconv_one,
1772 struct userdiff_driver *textconv_two,
1773 struct diff_options *o)
1774 {
1775 int lc_a, lc_b;
1776 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1777 const char *a_prefix, *b_prefix;
1778 char *data_one, *data_two;
1779 size_t size_one, size_two;
1780 struct emit_callback ecbdata;
1781 struct strbuf out = STRBUF_INIT;
1782
1783 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1784 a_prefix = o->b_prefix;
1785 b_prefix = o->a_prefix;
1786 } else {
1787 a_prefix = o->a_prefix;
1788 b_prefix = o->b_prefix;
1789 }
1790
1791 name_a += (*name_a == '/');
1792 name_b += (*name_b == '/');
1793
1794 strbuf_reset(&a_name);
1795 strbuf_reset(&b_name);
1796 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1797 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1798
1799 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1800 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1801
1802 memset(&ecbdata, 0, sizeof(ecbdata));
1803 ecbdata.color_diff = want_color(o->use_color);
1804 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1805 ecbdata.opt = o;
1806 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1807 mmfile_t mf1, mf2;
1808 mf1.ptr = (char *)data_one;
1809 mf2.ptr = (char *)data_two;
1810 mf1.size = size_one;
1811 mf2.size = size_two;
1812 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1813 }
1814 ecbdata.lno_in_preimage = 1;
1815 ecbdata.lno_in_postimage = 1;
1816
1817 lc_a = count_lines(data_one, size_one);
1818 lc_b = count_lines(data_two, size_two);
1819
1820 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1821 a_name.buf, a_name.len, 0);
1822 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1823 b_name.buf, b_name.len, 0);
1824
1825 strbuf_addstr(&out, "@@ -");
1826 if (!o->irreversible_delete)
1827 add_line_count(&out, lc_a);
1828 else
1829 strbuf_addstr(&out, "?,?");
1830 strbuf_addstr(&out, " +");
1831 add_line_count(&out, lc_b);
1832 strbuf_addstr(&out, " @@\n");
1833 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1834 strbuf_release(&out);
1835
1836 if (lc_a && !o->irreversible_delete)
1837 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1838 if (lc_b)
1839 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1840 if (textconv_one)
1841 free((char *)data_one);
1842 if (textconv_two)
1843 free((char *)data_two);
1844 }
1845
1846 struct diff_words_buffer {
1847 mmfile_t text;
1848 unsigned long alloc;
1849 struct diff_words_orig {
1850 const char *begin, *end;
1851 } *orig;
1852 int orig_nr, orig_alloc;
1853 };
1854
1855 static void diff_words_append(char *line, unsigned long len,
1856 struct diff_words_buffer *buffer)
1857 {
1858 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1859 line++;
1860 len--;
1861 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1862 buffer->text.size += len;
1863 buffer->text.ptr[buffer->text.size] = '\0';
1864 }
1865
1866 struct diff_words_style_elem {
1867 const char *prefix;
1868 const char *suffix;
1869 const char *color; /* NULL; filled in by the setup code if
1870 * color is enabled */
1871 };
1872
1873 struct diff_words_style {
1874 enum diff_words_type type;
1875 struct diff_words_style_elem new_word, old_word, ctx;
1876 const char *newline;
1877 };
1878
1879 static struct diff_words_style diff_words_styles[] = {
1880 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1881 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1882 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1883 };
1884
1885 struct diff_words_data {
1886 struct diff_words_buffer minus, plus;
1887 const char *current_plus;
1888 int last_minus;
1889 struct diff_options *opt;
1890 regex_t *word_regex;
1891 enum diff_words_type type;
1892 struct diff_words_style *style;
1893 };
1894
1895 static int fn_out_diff_words_write_helper(struct diff_options *o,
1896 struct diff_words_style_elem *st_el,
1897 const char *newline,
1898 size_t count, const char *buf)
1899 {
1900 int print = 0;
1901 struct strbuf sb = STRBUF_INIT;
1902
1903 while (count) {
1904 char *p = memchr(buf, '\n', count);
1905 if (print)
1906 strbuf_addstr(&sb, diff_line_prefix(o));
1907
1908 if (p != buf) {
1909 const char *reset = st_el->color && *st_el->color ?
1910 GIT_COLOR_RESET : NULL;
1911 if (st_el->color && *st_el->color)
1912 strbuf_addstr(&sb, st_el->color);
1913 strbuf_addstr(&sb, st_el->prefix);
1914 strbuf_add(&sb, buf, p ? p - buf : count);
1915 strbuf_addstr(&sb, st_el->suffix);
1916 if (reset)
1917 strbuf_addstr(&sb, reset);
1918 }
1919 if (!p)
1920 goto out;
1921
1922 strbuf_addstr(&sb, newline);
1923 count -= p + 1 - buf;
1924 buf = p + 1;
1925 print = 1;
1926 if (count) {
1927 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1928 sb.buf, sb.len, 0);
1929 strbuf_reset(&sb);
1930 }
1931 }
1932
1933 out:
1934 if (sb.len)
1935 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1936 sb.buf, sb.len, 0);
1937 strbuf_release(&sb);
1938 return 0;
1939 }
1940
1941 /*
1942 * '--color-words' algorithm can be described as:
1943 *
1944 * 1. collect the minus/plus lines of a diff hunk, divided into
1945 * minus-lines and plus-lines;
1946 *
1947 * 2. break both minus-lines and plus-lines into words and
1948 * place them into two mmfile_t with one word for each line;
1949 *
1950 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1951 *
1952 * And for the common parts of the both file, we output the plus side text.
1953 * diff_words->current_plus is used to trace the current position of the plus file
1954 * which printed. diff_words->last_minus is used to trace the last minus word
1955 * printed.
1956 *
1957 * For '--graph' to work with '--color-words', we need to output the graph prefix
1958 * on each line of color words output. Generally, there are two conditions on
1959 * which we should output the prefix.
1960 *
1961 * 1. diff_words->last_minus == 0 &&
1962 * diff_words->current_plus == diff_words->plus.text.ptr
1963 *
1964 * that is: the plus text must start as a new line, and if there is no minus
1965 * word printed, a graph prefix must be printed.
1966 *
1967 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1968 * *(diff_words->current_plus - 1) == '\n'
1969 *
1970 * that is: a graph prefix must be printed following a '\n'
1971 */
1972 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1973 {
1974 if ((diff_words->last_minus == 0 &&
1975 diff_words->current_plus == diff_words->plus.text.ptr) ||
1976 (diff_words->current_plus > diff_words->plus.text.ptr &&
1977 *(diff_words->current_plus - 1) == '\n')) {
1978 return 1;
1979 } else {
1980 return 0;
1981 }
1982 }
1983
1984 static void fn_out_diff_words_aux(void *priv,
1985 long minus_first, long minus_len,
1986 long plus_first, long plus_len,
1987 const char *func, long funclen)
1988 {
1989 struct diff_words_data *diff_words = priv;
1990 struct diff_words_style *style = diff_words->style;
1991 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1992 struct diff_options *opt = diff_words->opt;
1993 const char *line_prefix;
1994
1995 assert(opt);
1996 line_prefix = diff_line_prefix(opt);
1997
1998 /* POSIX requires that first be decremented by one if len == 0... */
1999 if (minus_len) {
2000 minus_begin = diff_words->minus.orig[minus_first].begin;
2001 minus_end =
2002 diff_words->minus.orig[minus_first + minus_len - 1].end;
2003 } else
2004 minus_begin = minus_end =
2005 diff_words->minus.orig[minus_first].end;
2006
2007 if (plus_len) {
2008 plus_begin = diff_words->plus.orig[plus_first].begin;
2009 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2010 } else
2011 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2012
2013 if (color_words_output_graph_prefix(diff_words)) {
2014 fputs(line_prefix, diff_words->opt->file);
2015 }
2016 if (diff_words->current_plus != plus_begin) {
2017 fn_out_diff_words_write_helper(diff_words->opt,
2018 &style->ctx, style->newline,
2019 plus_begin - diff_words->current_plus,
2020 diff_words->current_plus);
2021 }
2022 if (minus_begin != minus_end) {
2023 fn_out_diff_words_write_helper(diff_words->opt,
2024 &style->old_word, style->newline,
2025 minus_end - minus_begin, minus_begin);
2026 }
2027 if (plus_begin != plus_end) {
2028 fn_out_diff_words_write_helper(diff_words->opt,
2029 &style->new_word, style->newline,
2030 plus_end - plus_begin, plus_begin);
2031 }
2032
2033 diff_words->current_plus = plus_end;
2034 diff_words->last_minus = minus_first;
2035 }
2036
2037 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2038 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2039 int *begin, int *end)
2040 {
2041 if (word_regex && *begin < buffer->size) {
2042 regmatch_t match[1];
2043 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2044 buffer->size - *begin, 1, match, 0)) {
2045 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2046 '\n', match[0].rm_eo - match[0].rm_so);
2047 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2048 *begin += match[0].rm_so;
2049 return *begin >= *end;
2050 }
2051 return -1;
2052 }
2053
2054 /* find the next word */
2055 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2056 (*begin)++;
2057 if (*begin >= buffer->size)
2058 return -1;
2059
2060 /* find the end of the word */
2061 *end = *begin + 1;
2062 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2063 (*end)++;
2064
2065 return 0;
2066 }
2067
2068 /*
2069 * This function splits the words in buffer->text, stores the list with
2070 * newline separator into out, and saves the offsets of the original words
2071 * in buffer->orig.
2072 */
2073 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2074 regex_t *word_regex)
2075 {
2076 int i, j;
2077 long alloc = 0;
2078
2079 out->size = 0;
2080 out->ptr = NULL;
2081
2082 /* fake an empty "0th" word */
2083 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2084 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2085 buffer->orig_nr = 1;
2086
2087 for (i = 0; i < buffer->text.size; i++) {
2088 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2089 return;
2090
2091 /* store original boundaries */
2092 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2093 buffer->orig_alloc);
2094 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2095 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2096 buffer->orig_nr++;
2097
2098 /* store one word */
2099 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2100 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2101 out->ptr[out->size + j - i] = '\n';
2102 out->size += j - i + 1;
2103
2104 i = j - 1;
2105 }
2106 }
2107
2108 /* this executes the word diff on the accumulated buffers */
2109 static void diff_words_show(struct diff_words_data *diff_words)
2110 {
2111 xpparam_t xpp;
2112 xdemitconf_t xecfg;
2113 mmfile_t minus, plus;
2114 struct diff_words_style *style = diff_words->style;
2115
2116 struct diff_options *opt = diff_words->opt;
2117 const char *line_prefix;
2118
2119 assert(opt);
2120 line_prefix = diff_line_prefix(opt);
2121
2122 /* special case: only removal */
2123 if (!diff_words->plus.text.size) {
2124 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2125 line_prefix, strlen(line_prefix), 0);
2126 fn_out_diff_words_write_helper(diff_words->opt,
2127 &style->old_word, style->newline,
2128 diff_words->minus.text.size,
2129 diff_words->minus.text.ptr);
2130 diff_words->minus.text.size = 0;
2131 return;
2132 }
2133
2134 diff_words->current_plus = diff_words->plus.text.ptr;
2135 diff_words->last_minus = 0;
2136
2137 memset(&xpp, 0, sizeof(xpp));
2138 memset(&xecfg, 0, sizeof(xecfg));
2139 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2140 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2141 xpp.flags = 0;
2142 /* as only the hunk header will be parsed, we need a 0-context */
2143 xecfg.ctxlen = 0;
2144 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2145 diff_words, &xpp, &xecfg))
2146 die("unable to generate word diff");
2147 free(minus.ptr);
2148 free(plus.ptr);
2149 if (diff_words->current_plus != diff_words->plus.text.ptr +
2150 diff_words->plus.text.size) {
2151 if (color_words_output_graph_prefix(diff_words))
2152 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2153 line_prefix, strlen(line_prefix), 0);
2154 fn_out_diff_words_write_helper(diff_words->opt,
2155 &style->ctx, style->newline,
2156 diff_words->plus.text.ptr + diff_words->plus.text.size
2157 - diff_words->current_plus, diff_words->current_plus);
2158 }
2159 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2160 }
2161
2162 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2163 static void diff_words_flush(struct emit_callback *ecbdata)
2164 {
2165 struct diff_options *wo = ecbdata->diff_words->opt;
2166
2167 if (ecbdata->diff_words->minus.text.size ||
2168 ecbdata->diff_words->plus.text.size)
2169 diff_words_show(ecbdata->diff_words);
2170
2171 if (wo->emitted_symbols) {
2172 struct diff_options *o = ecbdata->opt;
2173 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2174 int i;
2175
2176 /*
2177 * NEEDSWORK:
2178 * Instead of appending each, concat all words to a line?
2179 */
2180 for (i = 0; i < wol->nr; i++)
2181 append_emitted_diff_symbol(o, &wol->buf[i]);
2182
2183 for (i = 0; i < wol->nr; i++)
2184 free((void *)wol->buf[i].line);
2185
2186 wol->nr = 0;
2187 }
2188 }
2189
2190 static void diff_filespec_load_driver(struct diff_filespec *one,
2191 struct index_state *istate)
2192 {
2193 /* Use already-loaded driver */
2194 if (one->driver)
2195 return;
2196
2197 if (S_ISREG(one->mode))
2198 one->driver = userdiff_find_by_path(istate, one->path);
2199
2200 /* Fallback to default settings */
2201 if (!one->driver)
2202 one->driver = userdiff_find_by_name("default");
2203 }
2204
2205 static const char *userdiff_word_regex(struct diff_filespec *one,
2206 struct index_state *istate)
2207 {
2208 diff_filespec_load_driver(one, istate);
2209 return one->driver->word_regex;
2210 }
2211
2212 static void init_diff_words_data(struct emit_callback *ecbdata,
2213 struct diff_options *orig_opts,
2214 struct diff_filespec *one,
2215 struct diff_filespec *two)
2216 {
2217 int i;
2218 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2219 memcpy(o, orig_opts, sizeof(struct diff_options));
2220
2221 ecbdata->diff_words =
2222 xcalloc(1, sizeof(struct diff_words_data));
2223 ecbdata->diff_words->type = o->word_diff;
2224 ecbdata->diff_words->opt = o;
2225
2226 if (orig_opts->emitted_symbols)
2227 o->emitted_symbols =
2228 xcalloc(1, sizeof(struct emitted_diff_symbols));
2229
2230 if (!o->word_regex)
2231 o->word_regex = userdiff_word_regex(one, o->repo->index);
2232 if (!o->word_regex)
2233 o->word_regex = userdiff_word_regex(two, o->repo->index);
2234 if (!o->word_regex)
2235 o->word_regex = diff_word_regex_cfg;
2236 if (o->word_regex) {
2237 ecbdata->diff_words->word_regex = (regex_t *)
2238 xmalloc(sizeof(regex_t));
2239 if (regcomp(ecbdata->diff_words->word_regex,
2240 o->word_regex,
2241 REG_EXTENDED | REG_NEWLINE))
2242 die("invalid regular expression: %s",
2243 o->word_regex);
2244 }
2245 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2246 if (o->word_diff == diff_words_styles[i].type) {
2247 ecbdata->diff_words->style =
2248 &diff_words_styles[i];
2249 break;
2250 }
2251 }
2252 if (want_color(o->use_color)) {
2253 struct diff_words_style *st = ecbdata->diff_words->style;
2254 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2255 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2256 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2257 }
2258 }
2259
2260 static void free_diff_words_data(struct emit_callback *ecbdata)
2261 {
2262 if (ecbdata->diff_words) {
2263 diff_words_flush(ecbdata);
2264 free (ecbdata->diff_words->opt->emitted_symbols);
2265 free (ecbdata->diff_words->opt);
2266 free (ecbdata->diff_words->minus.text.ptr);
2267 free (ecbdata->diff_words->minus.orig);
2268 free (ecbdata->diff_words->plus.text.ptr);
2269 free (ecbdata->diff_words->plus.orig);
2270 if (ecbdata->diff_words->word_regex) {
2271 regfree(ecbdata->diff_words->word_regex);
2272 free(ecbdata->diff_words->word_regex);
2273 }
2274 FREE_AND_NULL(ecbdata->diff_words);
2275 }
2276 }
2277
2278 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2279 {
2280 if (want_color(diff_use_color))
2281 return diff_colors[ix];
2282 return "";
2283 }
2284
2285 const char *diff_line_prefix(struct diff_options *opt)
2286 {
2287 struct strbuf *msgbuf;
2288 if (!opt->output_prefix)
2289 return "";
2290
2291 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2292 return msgbuf->buf;
2293 }
2294
2295 static unsigned long sane_truncate_line(char *line, unsigned long len)
2296 {
2297 const char *cp;
2298 unsigned long allot;
2299 size_t l = len;
2300
2301 cp = line;
2302 allot = l;
2303 while (0 < l) {
2304 (void) utf8_width(&cp, &l);
2305 if (!cp)
2306 break; /* truncated in the middle? */
2307 }
2308 return allot - l;
2309 }
2310
2311 static void find_lno(const char *line, struct emit_callback *ecbdata)
2312 {
2313 const char *p;
2314 ecbdata->lno_in_preimage = 0;
2315 ecbdata->lno_in_postimage = 0;
2316 p = strchr(line, '-');
2317 if (!p)
2318 return; /* cannot happen */
2319 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2320 p = strchr(p, '+');
2321 if (!p)
2322 return; /* cannot happen */
2323 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2324 }
2325
2326 static void fn_out_consume(void *priv, char *line, unsigned long len)
2327 {
2328 struct emit_callback *ecbdata = priv;
2329 struct diff_options *o = ecbdata->opt;
2330
2331 o->found_changes = 1;
2332
2333 if (ecbdata->header) {
2334 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2335 ecbdata->header->buf, ecbdata->header->len, 0);
2336 strbuf_reset(ecbdata->header);
2337 ecbdata->header = NULL;
2338 }
2339
2340 if (ecbdata->label_path[0]) {
2341 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2342 ecbdata->label_path[0],
2343 strlen(ecbdata->label_path[0]), 0);
2344 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2345 ecbdata->label_path[1],
2346 strlen(ecbdata->label_path[1]), 0);
2347 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2348 }
2349
2350 if (diff_suppress_blank_empty
2351 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2352 line[0] = '\n';
2353 len = 1;
2354 }
2355
2356 if (line[0] == '@') {
2357 if (ecbdata->diff_words)
2358 diff_words_flush(ecbdata);
2359 len = sane_truncate_line(line, len);
2360 find_lno(line, ecbdata);
2361 emit_hunk_header(ecbdata, line, len);
2362 return;
2363 }
2364
2365 if (ecbdata->diff_words) {
2366 enum diff_symbol s =
2367 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2368 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2369 if (line[0] == '-') {
2370 diff_words_append(line, len,
2371 &ecbdata->diff_words->minus);
2372 return;
2373 } else if (line[0] == '+') {
2374 diff_words_append(line, len,
2375 &ecbdata->diff_words->plus);
2376 return;
2377 } else if (starts_with(line, "\\ ")) {
2378 /*
2379 * Eat the "no newline at eof" marker as if we
2380 * saw a "+" or "-" line with nothing on it,
2381 * and return without diff_words_flush() to
2382 * defer processing. If this is the end of
2383 * preimage, more "+" lines may come after it.
2384 */
2385 return;
2386 }
2387 diff_words_flush(ecbdata);
2388 emit_diff_symbol(o, s, line, len, 0);
2389 return;
2390 }
2391
2392 switch (line[0]) {
2393 case '+':
2394 ecbdata->lno_in_postimage++;
2395 emit_add_line(ecbdata, line + 1, len - 1);
2396 break;
2397 case '-':
2398 ecbdata->lno_in_preimage++;
2399 emit_del_line(ecbdata, line + 1, len - 1);
2400 break;
2401 case ' ':
2402 ecbdata->lno_in_postimage++;
2403 ecbdata->lno_in_preimage++;
2404 emit_context_line(ecbdata, line + 1, len - 1);
2405 break;
2406 default:
2407 /* incomplete line at the end */
2408 ecbdata->lno_in_preimage++;
2409 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2410 line, len, 0);
2411 break;
2412 }
2413 }
2414
2415 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2416 {
2417 const char *old_name = a;
2418 const char *new_name = b;
2419 int pfx_length, sfx_length;
2420 int pfx_adjust_for_slash;
2421 int len_a = strlen(a);
2422 int len_b = strlen(b);
2423 int a_midlen, b_midlen;
2424 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2425 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2426
2427 if (qlen_a || qlen_b) {
2428 quote_c_style(a, name, NULL, 0);
2429 strbuf_addstr(name, " => ");
2430 quote_c_style(b, name, NULL, 0);
2431 return;
2432 }
2433
2434 /* Find common prefix */
2435 pfx_length = 0;
2436 while (*old_name && *new_name && *old_name == *new_name) {
2437 if (*old_name == '/')
2438 pfx_length = old_name - a + 1;
2439 old_name++;
2440 new_name++;
2441 }
2442
2443 /* Find common suffix */
2444 old_name = a + len_a;
2445 new_name = b + len_b;
2446 sfx_length = 0;
2447 /*
2448 * If there is a common prefix, it must end in a slash. In
2449 * that case we let this loop run 1 into the prefix to see the
2450 * same slash.
2451 *
2452 * If there is no common prefix, we cannot do this as it would
2453 * underrun the input strings.
2454 */
2455 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2456 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2457 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2458 *old_name == *new_name) {
2459 if (*old_name == '/')
2460 sfx_length = len_a - (old_name - a);
2461 old_name--;
2462 new_name--;
2463 }
2464
2465 /*
2466 * pfx{mid-a => mid-b}sfx
2467 * {pfx-a => pfx-b}sfx
2468 * pfx{sfx-a => sfx-b}
2469 * name-a => name-b
2470 */
2471 a_midlen = len_a - pfx_length - sfx_length;
2472 b_midlen = len_b - pfx_length - sfx_length;
2473 if (a_midlen < 0)
2474 a_midlen = 0;
2475 if (b_midlen < 0)
2476 b_midlen = 0;
2477
2478 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2479 if (pfx_length + sfx_length) {
2480 strbuf_add(name, a, pfx_length);
2481 strbuf_addch(name, '{');
2482 }
2483 strbuf_add(name, a + pfx_length, a_midlen);
2484 strbuf_addstr(name, " => ");
2485 strbuf_add(name, b + pfx_length, b_midlen);
2486 if (pfx_length + sfx_length) {
2487 strbuf_addch(name, '}');
2488 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2489 }
2490 }
2491
2492 struct diffstat_t {
2493 int nr;
2494 int alloc;
2495 struct diffstat_file {
2496 char *from_name;
2497 char *name;
2498 char *print_name;
2499 const char *comments;
2500 unsigned is_unmerged:1;
2501 unsigned is_binary:1;
2502 unsigned is_renamed:1;
2503 unsigned is_interesting:1;
2504 uintmax_t added, deleted;
2505 } **files;
2506 };
2507
2508 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2509 const char *name_a,
2510 const char *name_b)
2511 {
2512 struct diffstat_file *x;
2513 x = xcalloc(1, sizeof(*x));
2514 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2515 diffstat->files[diffstat->nr++] = x;
2516 if (name_b) {
2517 x->from_name = xstrdup(name_a);
2518 x->name = xstrdup(name_b);
2519 x->is_renamed = 1;
2520 }
2521 else {
2522 x->from_name = NULL;
2523 x->name = xstrdup(name_a);
2524 }
2525 return x;
2526 }
2527
2528 static void diffstat_consume(void *priv, char *line, unsigned long len)
2529 {
2530 struct diffstat_t *diffstat = priv;
2531 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2532
2533 if (line[0] == '+')
2534 x->added++;
2535 else if (line[0] == '-')
2536 x->deleted++;
2537 }
2538
2539 const char mime_boundary_leader[] = "------------";
2540
2541 static int scale_linear(int it, int width, int max_change)
2542 {
2543 if (!it)
2544 return 0;
2545 /*
2546 * make sure that at least one '-' or '+' is printed if
2547 * there is any change to this path. The easiest way is to
2548 * scale linearly as if the alloted width is one column shorter
2549 * than it is, and then add 1 to the result.
2550 */
2551 return 1 + (it * (width - 1) / max_change);
2552 }
2553
2554 static void show_graph(struct strbuf *out, char ch, int cnt,
2555 const char *set, const char *reset)
2556 {
2557 if (cnt <= 0)
2558 return;
2559 strbuf_addstr(out, set);
2560 strbuf_addchars(out, ch, cnt);
2561 strbuf_addstr(out, reset);
2562 }
2563
2564 static void fill_print_name(struct diffstat_file *file)
2565 {
2566 struct strbuf pname = STRBUF_INIT;
2567
2568 if (file->print_name)
2569 return;
2570
2571 if (file->is_renamed)
2572 pprint_rename(&pname, file->from_name, file->name);
2573 else
2574 quote_c_style(file->name, &pname, NULL, 0);
2575
2576 if (file->comments)
2577 strbuf_addf(&pname, " (%s)", file->comments);
2578
2579 file->print_name = strbuf_detach(&pname, NULL);
2580 }
2581
2582 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2583 int files, int insertions, int deletions)
2584 {
2585 struct strbuf sb = STRBUF_INIT;
2586
2587 if (!files) {
2588 assert(insertions == 0 && deletions == 0);
2589 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2590 NULL, 0, 0);
2591 return;
2592 }
2593
2594 strbuf_addf(&sb,
2595 (files == 1) ? " %d file changed" : " %d files changed",
2596 files);
2597
2598 /*
2599 * For binary diff, the caller may want to print "x files
2600 * changed" with insertions == 0 && deletions == 0.
2601 *
2602 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2603 * is probably less confusing (i.e skip over "2 files changed
2604 * but nothing about added/removed lines? Is this a bug in Git?").
2605 */
2606 if (insertions || deletions == 0) {
2607 strbuf_addf(&sb,
2608 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2609 insertions);
2610 }
2611
2612 if (deletions || insertions == 0) {
2613 strbuf_addf(&sb,
2614 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2615 deletions);
2616 }
2617 strbuf_addch(&sb, '\n');
2618 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2619 sb.buf, sb.len, 0);
2620 strbuf_release(&sb);
2621 }
2622
2623 void print_stat_summary(FILE *fp, int files,
2624 int insertions, int deletions)
2625 {
2626 struct diff_options o;
2627 memset(&o, 0, sizeof(o));
2628 o.file = fp;
2629
2630 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2631 }
2632
2633 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2634 {
2635 int i, len, add, del, adds = 0, dels = 0;
2636 uintmax_t max_change = 0, max_len = 0;
2637 int total_files = data->nr, count;
2638 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2639 const char *reset, *add_c, *del_c;
2640 int extra_shown = 0;
2641 const char *line_prefix = diff_line_prefix(options);
2642 struct strbuf out = STRBUF_INIT;
2643
2644 if (data->nr == 0)
2645 return;
2646
2647 count = options->stat_count ? options->stat_count : data->nr;
2648
2649 reset = diff_get_color_opt(options, DIFF_RESET);
2650 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2651 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2652
2653 /*
2654 * Find the longest filename and max number of changes
2655 */
2656 for (i = 0; (i < count) && (i < data->nr); i++) {
2657 struct diffstat_file *file = data->files[i];
2658 uintmax_t change = file->added + file->deleted;
2659
2660 if (!file->is_interesting && (change == 0)) {
2661 count++; /* not shown == room for one more */
2662 continue;
2663 }
2664 fill_print_name(file);
2665 len = strlen(file->print_name);
2666 if (max_len < len)
2667 max_len = len;
2668
2669 if (file->is_unmerged) {
2670 /* "Unmerged" is 8 characters */
2671 bin_width = bin_width < 8 ? 8 : bin_width;
2672 continue;
2673 }
2674 if (file->is_binary) {
2675 /* "Bin XXX -> YYY bytes" */
2676 int w = 14 + decimal_width(file->added)
2677 + decimal_width(file->deleted);
2678 bin_width = bin_width < w ? w : bin_width;
2679 /* Display change counts aligned with "Bin" */
2680 number_width = 3;
2681 continue;
2682 }
2683
2684 if (max_change < change)
2685 max_change = change;
2686 }
2687 count = i; /* where we can stop scanning in data->files[] */
2688
2689 /*
2690 * We have width = stat_width or term_columns() columns total.
2691 * We want a maximum of min(max_len, stat_name_width) for the name part.
2692 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2693 * We also need 1 for " " and 4 + decimal_width(max_change)
2694 * for " | NNNN " and one the empty column at the end, altogether
2695 * 6 + decimal_width(max_change).
2696 *
2697 * If there's not enough space, we will use the smaller of
2698 * stat_name_width (if set) and 5/8*width for the filename,
2699 * and the rest for constant elements + graph part, but no more
2700 * than stat_graph_width for the graph part.
2701 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2702 * for the standard terminal size).
2703 *
2704 * In other words: stat_width limits the maximum width, and
2705 * stat_name_width fixes the maximum width of the filename,
2706 * and is also used to divide available columns if there
2707 * aren't enough.
2708 *
2709 * Binary files are displayed with "Bin XXX -> YYY bytes"
2710 * instead of the change count and graph. This part is treated
2711 * similarly to the graph part, except that it is not
2712 * "scaled". If total width is too small to accommodate the
2713 * guaranteed minimum width of the filename part and the
2714 * separators and this message, this message will "overflow"
2715 * making the line longer than the maximum width.
2716 */
2717
2718 if (options->stat_width == -1)
2719 width = term_columns() - strlen(line_prefix);
2720 else
2721 width = options->stat_width ? options->stat_width : 80;
2722 number_width = decimal_width(max_change) > number_width ?
2723 decimal_width(max_change) : number_width;
2724
2725 if (options->stat_graph_width == -1)
2726 options->stat_graph_width = diff_stat_graph_width;
2727
2728 /*
2729 * Guarantee 3/8*16==6 for the graph part
2730 * and 5/8*16==10 for the filename part
2731 */
2732 if (width < 16 + 6 + number_width)
2733 width = 16 + 6 + number_width;
2734
2735 /*
2736 * First assign sizes that are wanted, ignoring available width.
2737 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2738 * starting from "XXX" should fit in graph_width.
2739 */
2740 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2741 if (options->stat_graph_width &&
2742 options->stat_graph_width < graph_width)
2743 graph_width = options->stat_graph_width;
2744
2745 name_width = (options->stat_name_width > 0 &&
2746 options->stat_name_width < max_len) ?
2747 options->stat_name_width : max_len;
2748
2749 /*
2750 * Adjust adjustable widths not to exceed maximum width
2751 */
2752 if (name_width + number_width + 6 + graph_width > width) {
2753 if (graph_width > width * 3/8 - number_width - 6) {
2754 graph_width = width * 3/8 - number_width - 6;
2755 if (graph_width < 6)
2756 graph_width = 6;
2757 }
2758
2759 if (options->stat_graph_width &&
2760 graph_width > options->stat_graph_width)
2761 graph_width = options->stat_graph_width;
2762 if (name_width > width - number_width - 6 - graph_width)
2763 name_width = width - number_width - 6 - graph_width;
2764 else
2765 graph_width = width - number_width - 6 - name_width;
2766 }
2767
2768 /*
2769 * From here name_width is the width of the name area,
2770 * and graph_width is the width of the graph area.
2771 * max_change is used to scale graph properly.
2772 */
2773 for (i = 0; i < count; i++) {
2774 const char *prefix = "";
2775 struct diffstat_file *file = data->files[i];
2776 char *name = file->print_name;
2777 uintmax_t added = file->added;
2778 uintmax_t deleted = file->deleted;
2779 int name_len;
2780
2781 if (!file->is_interesting && (added + deleted == 0))
2782 continue;
2783
2784 /*
2785 * "scale" the filename
2786 */
2787 len = name_width;
2788 name_len = strlen(name);
2789 if (name_width < name_len) {
2790 char *slash;
2791 prefix = "...";
2792 len -= 3;
2793 name += name_len - len;
2794 slash = strchr(name, '/');
2795 if (slash)
2796 name = slash;
2797 }
2798
2799 if (file->is_binary) {
2800 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2801 strbuf_addf(&out, " %*s", number_width, "Bin");
2802 if (!added && !deleted) {
2803 strbuf_addch(&out, '\n');
2804 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2805 out.buf, out.len, 0);
2806 strbuf_reset(&out);
2807 continue;
2808 }
2809 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2810 del_c, deleted, reset);
2811 strbuf_addstr(&out, " -> ");
2812 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2813 add_c, added, reset);
2814 strbuf_addstr(&out, " bytes\n");
2815 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2816 out.buf, out.len, 0);
2817 strbuf_reset(&out);
2818 continue;
2819 }
2820 else if (file->is_unmerged) {
2821 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2822 strbuf_addstr(&out, " Unmerged\n");
2823 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2824 out.buf, out.len, 0);
2825 strbuf_reset(&out);
2826 continue;
2827 }
2828
2829 /*
2830 * scale the add/delete
2831 */
2832 add = added;
2833 del = deleted;
2834
2835 if (graph_width <= max_change) {
2836 int total = scale_linear(add + del, graph_width, max_change);
2837 if (total < 2 && add && del)
2838 /* width >= 2 due to the sanity check */
2839 total = 2;
2840 if (add < del) {
2841 add = scale_linear(add, graph_width, max_change);
2842 del = total - add;
2843 } else {
2844 del = scale_linear(del, graph_width, max_change);
2845 add = total - del;
2846 }
2847 }
2848 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2849 strbuf_addf(&out, " %*"PRIuMAX"%s",
2850 number_width, added + deleted,
2851 added + deleted ? " " : "");
2852 show_graph(&out, '+', add, add_c, reset);
2853 show_graph(&out, '-', del, del_c, reset);
2854 strbuf_addch(&out, '\n');
2855 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2856 out.buf, out.len, 0);
2857 strbuf_reset(&out);
2858 }
2859
2860 for (i = 0; i < data->nr; i++) {
2861 struct diffstat_file *file = data->files[i];
2862 uintmax_t added = file->added;
2863 uintmax_t deleted = file->deleted;
2864
2865 if (file->is_unmerged ||
2866 (!file->is_interesting && (added + deleted == 0))) {
2867 total_files--;
2868 continue;
2869 }
2870
2871 if (!file->is_binary) {
2872 adds += added;
2873 dels += deleted;
2874 }
2875 if (i < count)
2876 continue;
2877 if (!extra_shown)
2878 emit_diff_symbol(options,
2879 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2880 NULL, 0, 0);
2881 extra_shown = 1;
2882 }
2883
2884 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2885 strbuf_release(&out);
2886 }
2887
2888 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2889 {
2890 int i, adds = 0, dels = 0, total_files = data->nr;
2891
2892 if (data->nr == 0)
2893 return;
2894
2895 for (i = 0; i < data->nr; i++) {
2896 int added = data->files[i]->added;
2897 int deleted = data->files[i]->deleted;
2898
2899 if (data->files[i]->is_unmerged ||
2900 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2901 total_files--;
2902 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2903 adds += added;
2904 dels += deleted;
2905 }
2906 }
2907 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2908 }
2909
2910 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2911 {
2912 int i;
2913
2914 if (data->nr == 0)
2915 return;
2916
2917 for (i = 0; i < data->nr; i++) {
2918 struct diffstat_file *file = data->files[i];
2919
2920 fprintf(options->file, "%s", diff_line_prefix(options));
2921
2922 if (file->is_binary)
2923 fprintf(options->file, "-\t-\t");
2924 else
2925 fprintf(options->file,
2926 "%"PRIuMAX"\t%"PRIuMAX"\t",
2927 file->added, file->deleted);
2928 if (options->line_termination) {
2929 fill_print_name(file);
2930 if (!file->is_renamed)
2931 write_name_quoted(file->name, options->file,
2932 options->line_termination);
2933 else {
2934 fputs(file->print_name, options->file);
2935 putc(options->line_termination, options->file);
2936 }
2937 } else {
2938 if (file->is_renamed) {
2939 putc('\0', options->file);
2940 write_name_quoted(file->from_name, options->file, '\0');
2941 }
2942 write_name_quoted(file->name, options->file, '\0');
2943 }
2944 }
2945 }
2946
2947 struct dirstat_file {
2948 const char *name;
2949 unsigned long changed;
2950 };
2951
2952 struct dirstat_dir {
2953 struct dirstat_file *files;
2954 int alloc, nr, permille, cumulative;
2955 };
2956
2957 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2958 unsigned long changed, const char *base, int baselen)
2959 {
2960 unsigned long sum_changes = 0;
2961 unsigned int sources = 0;
2962 const char *line_prefix = diff_line_prefix(opt);
2963
2964 while (dir->nr) {
2965 struct dirstat_file *f = dir->files;
2966 int namelen = strlen(f->name);
2967 unsigned long changes;
2968 char *slash;
2969
2970 if (namelen < baselen)
2971 break;
2972 if (memcmp(f->name, base, baselen))
2973 break;
2974 slash = strchr(f->name + baselen, '/');
2975 if (slash) {
2976 int newbaselen = slash + 1 - f->name;
2977 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2978 sources++;
2979 } else {
2980 changes = f->changed;
2981 dir->files++;
2982 dir->nr--;
2983 sources += 2;
2984 }
2985 sum_changes += changes;
2986 }
2987
2988 /*
2989 * We don't report dirstat's for
2990 * - the top level
2991 * - or cases where everything came from a single directory
2992 * under this directory (sources == 1).
2993 */
2994 if (baselen && sources != 1) {
2995 if (sum_changes) {
2996 int permille = sum_changes * 1000 / changed;
2997 if (permille >= dir->permille) {
2998 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2999 permille / 10, permille % 10, baselen, base);
3000 if (!dir->cumulative)
3001 return 0;
3002 }
3003 }
3004 }
3005 return sum_changes;
3006 }
3007
3008 static int dirstat_compare(const void *_a, const void *_b)
3009 {
3010 const struct dirstat_file *a = _a;
3011 const struct dirstat_file *b = _b;
3012 return strcmp(a->name, b->name);
3013 }
3014
3015 static void show_dirstat(struct diff_options *options)
3016 {
3017 int i;
3018 unsigned long changed;
3019 struct dirstat_dir dir;
3020 struct diff_queue_struct *q = &diff_queued_diff;
3021
3022 dir.files = NULL;
3023 dir.alloc = 0;
3024 dir.nr = 0;
3025 dir.permille = options->dirstat_permille;
3026 dir.cumulative = options->flags.dirstat_cumulative;
3027
3028 changed = 0;
3029 for (i = 0; i < q->nr; i++) {
3030 struct diff_filepair *p = q->queue[i];
3031 const char *name;
3032 unsigned long copied, added, damage;
3033
3034 name = p->two->path ? p->two->path : p->one->path;
3035
3036 if (p->one->oid_valid && p->two->oid_valid &&
3037 oideq(&p->one->oid, &p->two->oid)) {
3038 /*
3039 * The SHA1 has not changed, so pre-/post-content is
3040 * identical. We can therefore skip looking at the
3041 * file contents altogether.
3042 */
3043 damage = 0;
3044 goto found_damage;
3045 }
3046
3047 if (options->flags.dirstat_by_file) {
3048 /*
3049 * In --dirstat-by-file mode, we don't really need to
3050 * look at the actual file contents at all.
3051 * The fact that the SHA1 changed is enough for us to
3052 * add this file to the list of results
3053 * (with each file contributing equal damage).
3054 */
3055 damage = 1;
3056 goto found_damage;
3057 }
3058
3059 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3060 diff_populate_filespec(options->repo, p->one, 0);
3061 diff_populate_filespec(options->repo, p->two, 0);
3062 diffcore_count_changes(options->repo,
3063 p->one, p->two, NULL, NULL,
3064 &copied, &added);
3065 diff_free_filespec_data(p->one);
3066 diff_free_filespec_data(p->two);
3067 } else if (DIFF_FILE_VALID(p->one)) {
3068 diff_populate_filespec(options->repo, p->one, CHECK_SIZE_ONLY);
3069 copied = added = 0;
3070 diff_free_filespec_data(p->one);
3071 } else if (DIFF_FILE_VALID(p->two)) {
3072 diff_populate_filespec(options->repo, p->two, CHECK_SIZE_ONLY);
3073 copied = 0;
3074 added = p->two->size;
3075 diff_free_filespec_data(p->two);
3076 } else
3077 continue;
3078
3079 /*
3080 * Original minus copied is the removed material,
3081 * added is the new material. They are both damages
3082 * made to the preimage.
3083 * If the resulting damage is zero, we know that
3084 * diffcore_count_changes() considers the two entries to
3085 * be identical, but since the oid changed, we
3086 * know that there must have been _some_ kind of change,
3087 * so we force all entries to have damage > 0.
3088 */
3089 damage = (p->one->size - copied) + added;
3090 if (!damage)
3091 damage = 1;
3092
3093 found_damage:
3094 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3095 dir.files[dir.nr].name = name;
3096 dir.files[dir.nr].changed = damage;
3097 changed += damage;
3098 dir.nr++;
3099 }
3100
3101 /* This can happen even with many files, if everything was renames */
3102 if (!changed)
3103 return;
3104
3105 /* Show all directories with more than x% of the changes */
3106 QSORT(dir.files, dir.nr, dirstat_compare);
3107 gather_dirstat(options, &dir, changed, "", 0);
3108 }
3109
3110 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3111 {
3112 int i;
3113 unsigned long changed;
3114 struct dirstat_dir dir;
3115
3116 if (data->nr == 0)
3117 return;
3118
3119 dir.files = NULL;
3120 dir.alloc = 0;
3121 dir.nr = 0;
3122 dir.permille = options->dirstat_permille;
3123 dir.cumulative = options->flags.dirstat_cumulative;
3124
3125 changed = 0;
3126 for (i = 0; i < data->nr; i++) {
3127 struct diffstat_file *file = data->files[i];
3128 unsigned long damage = file->added + file->deleted;
3129 if (file->is_binary)
3130 /*
3131 * binary files counts bytes, not lines. Must find some
3132 * way to normalize binary bytes vs. textual lines.
3133 * The following heuristic assumes that there are 64
3134 * bytes per "line".
3135 * This is stupid and ugly, but very cheap...
3136 */
3137 damage = DIV_ROUND_UP(damage, 64);
3138 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3139 dir.files[dir.nr].name = file->name;
3140 dir.files[dir.nr].changed = damage;
3141 changed += damage;
3142 dir.nr++;
3143 }
3144
3145 /* This can happen even with many files, if everything was renames */
3146 if (!changed)
3147 return;
3148
3149 /* Show all directories with more than x% of the changes */
3150 QSORT(dir.files, dir.nr, dirstat_compare);
3151 gather_dirstat(options, &dir, changed, "", 0);
3152 }
3153
3154 static void free_diffstat_info(struct diffstat_t *diffstat)
3155 {
3156 int i;
3157 for (i = 0; i < diffstat->nr; i++) {
3158 struct diffstat_file *f = diffstat->files[i];
3159 free(f->print_name);
3160 free(f->name);
3161 free(f->from_name);
3162 free(f);
3163 }
3164 free(diffstat->files);
3165 }
3166
3167 struct checkdiff_t {
3168 const char *filename;
3169 int lineno;
3170 int conflict_marker_size;
3171 struct diff_options *o;
3172 unsigned ws_rule;
3173 unsigned status;
3174 };
3175
3176 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3177 {
3178 char firstchar;
3179 int cnt;
3180
3181 if (len < marker_size + 1)
3182 return 0;
3183 firstchar = line[0];
3184 switch (firstchar) {
3185 case '=': case '>': case '<': case '|':
3186 break;
3187 default:
3188 return 0;
3189 }
3190 for (cnt = 1; cnt < marker_size; cnt++)
3191 if (line[cnt] != firstchar)
3192 return 0;
3193 /* line[1] thru line[marker_size-1] are same as firstchar */
3194 if (len < marker_size + 1 || !isspace(line[marker_size]))
3195 return 0;
3196 return 1;
3197 }
3198
3199 static void checkdiff_consume_hunk(void *priv,
3200 long ob, long on, long nb, long nn,
3201 const char *func, long funclen)
3202
3203 {
3204 struct checkdiff_t *data = priv;
3205 data->lineno = nb - 1;
3206 }
3207
3208 static void checkdiff_consume(void *priv, char *line, unsigned long len)
3209 {
3210 struct checkdiff_t *data = priv;
3211 int marker_size = data->conflict_marker_size;
3212 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3213 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3214 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3215 char *err;
3216 const char *line_prefix;
3217
3218 assert(data->o);
3219 line_prefix = diff_line_prefix(data->o);
3220
3221 if (line[0] == '+') {
3222 unsigned bad;
3223 data->lineno++;
3224 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3225 data->status |= 1;
3226 fprintf(data->o->file,
3227 "%s%s:%d: leftover conflict marker\n",
3228 line_prefix, data->filename, data->lineno);
3229 }
3230 bad = ws_check(line + 1, len - 1, data->ws_rule);
3231 if (!bad)
3232 return;
3233 data->status |= bad;
3234 err = whitespace_error_string(bad);
3235 fprintf(data->o->file, "%s%s:%d: %s.\n",
3236 line_prefix, data->filename, data->lineno, err);
3237 free(err);
3238 emit_line(data->o, set, reset, line, 1);
3239 ws_check_emit(line + 1, len - 1, data->ws_rule,
3240 data->o->file, set, reset, ws);
3241 } else if (line[0] == ' ') {
3242 data->lineno++;
3243 }
3244 }
3245
3246 static unsigned char *deflate_it(char *data,
3247 unsigned long size,
3248 unsigned long *result_size)
3249 {
3250 int bound;
3251 unsigned char *deflated;
3252 git_zstream stream;
3253
3254 git_deflate_init(&stream, zlib_compression_level);
3255 bound = git_deflate_bound(&stream, size);
3256 deflated = xmalloc(bound);
3257 stream.next_out = deflated;
3258 stream.avail_out = bound;
3259
3260 stream.next_in = (unsigned char *)data;
3261 stream.avail_in = size;
3262 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3263 ; /* nothing */
3264 git_deflate_end(&stream);
3265 *result_size = stream.total_out;
3266 return deflated;
3267 }
3268
3269 static void emit_binary_diff_body(struct diff_options *o,
3270 mmfile_t *one, mmfile_t *two)
3271 {
3272 void *cp;
3273 void *delta;
3274 void *deflated;
3275 void *data;
3276 unsigned long orig_size;
3277 unsigned long delta_size;
3278 unsigned long deflate_size;
3279 unsigned long data_size;
3280
3281 /* We could do deflated delta, or we could do just deflated two,
3282 * whichever is smaller.
3283 */
3284 delta = NULL;
3285 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3286 if (one->size && two->size) {
3287 delta = diff_delta(one->ptr, one->size,
3288 two->ptr, two->size,
3289 &delta_size, deflate_size);
3290 if (delta) {
3291 void *to_free = delta;
3292 orig_size = delta_size;
3293 delta = deflate_it(delta, delta_size, &delta_size);
3294 free(to_free);
3295 }
3296 }
3297
3298 if (delta && delta_size < deflate_size) {
3299 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3300 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3301 s, strlen(s), 0);
3302 free(s);
3303 free(deflated);
3304 data = delta;
3305 data_size = delta_size;
3306 } else {
3307 char *s = xstrfmt("%lu", two->size);
3308 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3309 s, strlen(s), 0);
3310 free(s);
3311 free(delta);
3312 data = deflated;
3313 data_size = deflate_size;
3314 }
3315
3316 /* emit data encoded in base85 */
3317 cp = data;
3318 while (data_size) {
3319 int len;
3320 int bytes = (52 < data_size) ? 52 : data_size;
3321 char line[71];
3322 data_size -= bytes;
3323 if (bytes <= 26)
3324 line[0] = bytes + 'A' - 1;
3325 else
3326 line[0] = bytes - 26 + 'a' - 1;
3327 encode_85(line + 1, cp, bytes);
3328 cp = (char *) cp + bytes;
3329
3330 len = strlen(line);
3331 line[len++] = '\n';
3332 line[len] = '\0';
3333
3334 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3335 line, len, 0);
3336 }
3337 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3338 free(data);
3339 }
3340
3341 static void emit_binary_diff(struct diff_options *o,
3342 mmfile_t *one, mmfile_t *two)
3343 {
3344 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3345 emit_binary_diff_body(o, one, two);
3346 emit_binary_diff_body(o, two, one);
3347 }
3348
3349 int diff_filespec_is_binary(struct repository *r,
3350 struct diff_filespec *one)
3351 {
3352 if (one->is_binary == -1) {
3353 diff_filespec_load_driver(one, r->index);
3354 if (one->driver->binary != -1)
3355 one->is_binary = one->driver->binary;
3356 else {
3357 if (!one->data && DIFF_FILE_VALID(one))
3358 diff_populate_filespec(r, one, CHECK_BINARY);
3359 if (one->is_binary == -1 && one->data)
3360 one->is_binary = buffer_is_binary(one->data,
3361 one->size);
3362 if (one->is_binary == -1)
3363 one->is_binary = 0;
3364 }
3365 }
3366 return one->is_binary;
3367 }
3368
3369 static const struct userdiff_funcname *
3370 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3371 {
3372 diff_filespec_load_driver(one, o->repo->index);
3373 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3374 }
3375
3376 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3377 {
3378 if (!options->a_prefix)
3379 options->a_prefix = a;
3380 if (!options->b_prefix)
3381 options->b_prefix = b;
3382 }
3383
3384 struct userdiff_driver *get_textconv(struct repository *r,
3385 struct diff_filespec *one)
3386 {
3387 if (!DIFF_FILE_VALID(one))
3388 return NULL;
3389
3390 diff_filespec_load_driver(one, r->index);
3391 return userdiff_get_textconv(r, one->driver);
3392 }
3393
3394 static void builtin_diff(const char *name_a,
3395 const char *name_b,
3396 struct diff_filespec *one,
3397 struct diff_filespec *two,
3398 const char *xfrm_msg,
3399 int must_show_header,
3400 struct diff_options *o,
3401 int complete_rewrite)
3402 {
3403 mmfile_t mf1, mf2;
3404 const char *lbl[2];
3405 char *a_one, *b_two;
3406 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3407 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3408 const char *a_prefix, *b_prefix;
3409 struct userdiff_driver *textconv_one = NULL;
3410 struct userdiff_driver *textconv_two = NULL;
3411 struct strbuf header = STRBUF_INIT;
3412 const char *line_prefix = diff_line_prefix(o);
3413
3414 diff_set_mnemonic_prefix(o, "a/", "b/");
3415 if (o->flags.reverse_diff) {
3416 a_prefix = o->b_prefix;
3417 b_prefix = o->a_prefix;
3418 } else {
3419 a_prefix = o->a_prefix;
3420 b_prefix = o->b_prefix;
3421 }
3422
3423 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3424 (!one->mode || S_ISGITLINK(one->mode)) &&
3425 (!two->mode || S_ISGITLINK(two->mode))) {
3426 show_submodule_summary(o, one->path ? one->path : two->path,
3427 &one->oid, &two->oid,
3428 two->dirty_submodule);
3429 return;
3430 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3431 (!one->mode || S_ISGITLINK(one->mode)) &&
3432 (!two->mode || S_ISGITLINK(two->mode))) {
3433 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3434 &one->oid, &two->oid,
3435 two->dirty_submodule);
3436 return;
3437 }
3438
3439 if (o->flags.allow_textconv) {
3440 textconv_one = get_textconv(o->repo, one);
3441 textconv_two = get_textconv(o->repo, two);
3442 }
3443
3444 /* Never use a non-valid filename anywhere if at all possible */
3445 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3446 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3447
3448 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3449 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3450 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3451 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3452 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3453 if (lbl[0][0] == '/') {
3454 /* /dev/null */
3455 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3456 if (xfrm_msg)
3457 strbuf_addstr(&header, xfrm_msg);
3458 must_show_header = 1;
3459 }
3460 else if (lbl[1][0] == '/') {
3461 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3462 if (xfrm_msg)
3463 strbuf_addstr(&header, xfrm_msg);
3464 must_show_header = 1;
3465 }
3466 else {
3467 if (one->mode != two->mode) {
3468 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3469 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3470 must_show_header = 1;
3471 }
3472 if (xfrm_msg)
3473 strbuf_addstr(&header, xfrm_msg);
3474
3475 /*
3476 * we do not run diff between different kind
3477 * of objects.
3478 */
3479 if ((one->mode ^ two->mode) & S_IFMT)
3480 goto free_ab_and_return;
3481 if (complete_rewrite &&
3482 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3483 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3484 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3485 header.buf, header.len, 0);
3486 strbuf_reset(&header);
3487 emit_rewrite_diff(name_a, name_b, one, two,
3488 textconv_one, textconv_two, o);
3489 o->found_changes = 1;
3490 goto free_ab_and_return;
3491 }
3492 }
3493
3494 if (o->irreversible_delete && lbl[1][0] == '/') {
3495 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3496 header.len, 0);
3497 strbuf_reset(&header);
3498 goto free_ab_and_return;
3499 } else if (!o->flags.text &&
3500 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3501 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3502 struct strbuf sb = STRBUF_INIT;
3503 if (!one->data && !two->data &&
3504 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3505 !o->flags.binary) {
3506 if (oideq(&one->oid, &two->oid)) {
3507 if (must_show_header)
3508 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3509 header.buf, header.len,
3510 0);
3511 goto free_ab_and_return;
3512 }
3513 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3514 header.buf, header.len, 0);
3515 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3516 diff_line_prefix(o), lbl[0], lbl[1]);
3517 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3518 sb.buf, sb.len, 0);
3519 strbuf_release(&sb);
3520 goto free_ab_and_return;
3521 }
3522 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3523 fill_mmfile(o->repo, &mf2, two) < 0)
3524 die("unable to read files to diff");
3525 /* Quite common confusing case */
3526 if (mf1.size == mf2.size &&
3527 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3528 if (must_show_header)
3529 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3530 header.buf, header.len, 0);
3531 goto free_ab_and_return;
3532 }
3533 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3534 strbuf_reset(&header);
3535 if (o->flags.binary)
3536 emit_binary_diff(o, &mf1, &mf2);
3537 else {
3538 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3539 diff_line_prefix(o), lbl[0], lbl[1]);
3540 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3541 sb.buf, sb.len, 0);
3542 strbuf_release(&sb);
3543 }
3544 o->found_changes = 1;
3545 } else {
3546 /* Crazy xdl interfaces.. */
3547 const char *diffopts;
3548 const char *v;
3549 xpparam_t xpp;
3550 xdemitconf_t xecfg;
3551 struct emit_callback ecbdata;
3552 const struct userdiff_funcname *pe;
3553
3554 if (must_show_header) {
3555 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3556 header.buf, header.len, 0);
3557 strbuf_reset(&header);
3558 }
3559
3560 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3561 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3562
3563 pe = diff_funcname_pattern(o, one);
3564 if (!pe)
3565 pe = diff_funcname_pattern(o, two);
3566
3567 memset(&xpp, 0, sizeof(xpp));
3568 memset(&xecfg, 0, sizeof(xecfg));
3569 memset(&ecbdata, 0, sizeof(ecbdata));
3570 if (o->flags.suppress_diff_headers)
3571 lbl[0] = NULL;
3572 ecbdata.label_path = lbl;
3573 ecbdata.color_diff = want_color(o->use_color);
3574 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3575 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3576 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3577 ecbdata.opt = o;
3578 if (header.len && !o->flags.suppress_diff_headers)
3579 ecbdata.header = &header;
3580 xpp.flags = o->xdl_opts;
3581 xpp.anchors = o->anchors;
3582 xpp.anchors_nr = o->anchors_nr;
3583 xecfg.ctxlen = o->context;
3584 xecfg.interhunkctxlen = o->interhunkcontext;
3585 xecfg.flags = XDL_EMIT_FUNCNAMES;
3586 if (o->flags.funccontext)
3587 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3588 if (pe)
3589 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3590
3591 diffopts = getenv("GIT_DIFF_OPTS");
3592 if (!diffopts)
3593 ;
3594 else if (skip_prefix(diffopts, "--unified=", &v))
3595 xecfg.ctxlen = strtoul(v, NULL, 10);
3596 else if (skip_prefix(diffopts, "-u", &v))
3597 xecfg.ctxlen = strtoul(v, NULL, 10);
3598
3599 if (o->word_diff)
3600 init_diff_words_data(&ecbdata, o, one, two);
3601 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3602 &ecbdata, &xpp, &xecfg))
3603 die("unable to generate diff for %s", one->path);
3604 if (o->word_diff)
3605 free_diff_words_data(&ecbdata);
3606 if (textconv_one)
3607 free(mf1.ptr);
3608 if (textconv_two)
3609 free(mf2.ptr);
3610 xdiff_clear_find_func(&xecfg);
3611 }
3612
3613 free_ab_and_return:
3614 strbuf_release(&header);
3615 diff_free_filespec_data(one);
3616 diff_free_filespec_data(two);
3617 free(a_one);
3618 free(b_two);
3619 return;
3620 }
3621
3622 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3623 {
3624 if (!is_renamed) {
3625 if (p->status == DIFF_STATUS_ADDED) {
3626 if (S_ISLNK(p->two->mode))
3627 return "new +l";
3628 else if ((p->two->mode & 0777) == 0755)
3629 return "new +x";
3630 else
3631 return "new";
3632 } else if (p->status == DIFF_STATUS_DELETED)
3633 return "gone";
3634 }
3635 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3636 return "mode -l";
3637 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3638 return "mode +l";
3639 else if ((p->one->mode & 0777) == 0644 &&
3640 (p->two->mode & 0777) == 0755)
3641 return "mode +x";
3642 else if ((p->one->mode & 0777) == 0755 &&
3643 (p->two->mode & 0777) == 0644)
3644 return "mode -x";
3645 return NULL;
3646 }
3647
3648 static void builtin_diffstat(const char *name_a, const char *name_b,
3649 struct diff_filespec *one,
3650 struct diff_filespec *two,
3651 struct diffstat_t *diffstat,
3652 struct diff_options *o,
3653 struct diff_filepair *p)
3654 {
3655 mmfile_t mf1, mf2;
3656 struct diffstat_file *data;
3657 int same_contents;
3658 int complete_rewrite = 0;
3659
3660 if (!DIFF_PAIR_UNMERGED(p)) {
3661 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3662 complete_rewrite = 1;
3663 }
3664
3665 data = diffstat_add(diffstat, name_a, name_b);
3666 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3667 if (o->flags.stat_with_summary)
3668 data->comments = get_compact_summary(p, data->is_renamed);
3669
3670 if (!one || !two) {
3671 data->is_unmerged = 1;
3672 return;
3673 }
3674
3675 same_contents = oideq(&one->oid, &two->oid);
3676
3677 if (diff_filespec_is_binary(o->repo, one) ||
3678 diff_filespec_is_binary(o->repo, two)) {
3679 data->is_binary = 1;
3680 if (same_contents) {
3681 data->added = 0;
3682 data->deleted = 0;
3683 } else {
3684 data->added = diff_filespec_size(o->repo, two);
3685 data->deleted = diff_filespec_size(o->repo, one);
3686 }
3687 }
3688
3689 else if (complete_rewrite) {
3690 diff_populate_filespec(o->repo, one, 0);
3691 diff_populate_filespec(o->repo, two, 0);
3692 data->deleted = count_lines(one->data, one->size);
3693 data->added = count_lines(two->data, two->size);
3694 }
3695
3696 else if (!same_contents) {
3697 /* Crazy xdl interfaces.. */
3698 xpparam_t xpp;
3699 xdemitconf_t xecfg;
3700
3701 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3702 fill_mmfile(o->repo, &mf2, two) < 0)
3703 die("unable to read files to diff");
3704
3705 memset(&xpp, 0, sizeof(xpp));
3706 memset(&xecfg, 0, sizeof(xecfg));
3707 xpp.flags = o->xdl_opts;
3708 xpp.anchors = o->anchors;
3709 xpp.anchors_nr = o->anchors_nr;
3710 xecfg.ctxlen = o->context;
3711 xecfg.interhunkctxlen = o->interhunkcontext;
3712 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
3713 diffstat_consume, diffstat, &xpp, &xecfg))
3714 die("unable to generate diffstat for %s", one->path);
3715 }
3716
3717 diff_free_filespec_data(one);
3718 diff_free_filespec_data(two);
3719 }
3720
3721 static void builtin_checkdiff(const char *name_a, const char *name_b,
3722 const char *attr_path,
3723 struct diff_filespec *one,
3724 struct diff_filespec *two,
3725 struct diff_options *o)
3726 {
3727 mmfile_t mf1, mf2;
3728 struct checkdiff_t data;
3729
3730 if (!two)
3731 return;
3732
3733 memset(&data, 0, sizeof(data));
3734 data.filename = name_b ? name_b : name_a;
3735 data.lineno = 0;
3736 data.o = o;
3737 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3738 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3739
3740 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3741 fill_mmfile(o->repo, &mf2, two) < 0)
3742 die("unable to read files to diff");
3743
3744 /*
3745 * All the other codepaths check both sides, but not checking
3746 * the "old" side here is deliberate. We are checking the newly
3747 * introduced changes, and as long as the "new" side is text, we
3748 * can and should check what it introduces.
3749 */
3750 if (diff_filespec_is_binary(o->repo, two))
3751 goto free_and_return;
3752 else {
3753 /* Crazy xdl interfaces.. */
3754 xpparam_t xpp;
3755 xdemitconf_t xecfg;
3756
3757 memset(&xpp, 0, sizeof(xpp));
3758 memset(&xecfg, 0, sizeof(xecfg));
3759 xecfg.ctxlen = 1; /* at least one context line */
3760 xpp.flags = 0;
3761 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3762 checkdiff_consume, &data,
3763 &xpp, &xecfg))
3764 die("unable to generate checkdiff for %s", one->path);
3765
3766 if (data.ws_rule & WS_BLANK_AT_EOF) {
3767 struct emit_callback ecbdata;
3768 int blank_at_eof;
3769
3770 ecbdata.ws_rule = data.ws_rule;
3771 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3772 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3773
3774 if (blank_at_eof) {
3775 static char *err;
3776 if (!err)
3777 err = whitespace_error_string(WS_BLANK_AT_EOF);
3778 fprintf(o->file, "%s:%d: %s.\n",
3779 data.filename, blank_at_eof, err);
3780 data.status = 1; /* report errors */
3781 }
3782 }
3783 }
3784 free_and_return:
3785 diff_free_filespec_data(one);
3786 diff_free_filespec_data(two);
3787 if (data.status)
3788 o->flags.check_failed = 1;
3789 }
3790
3791 struct diff_filespec *alloc_filespec(const char *path)
3792 {
3793 struct diff_filespec *spec;
3794
3795 FLEXPTR_ALLOC_STR(spec, path, path);
3796 spec->count = 1;
3797 spec->is_binary = -1;
3798 return spec;
3799 }
3800
3801 void free_filespec(struct diff_filespec *spec)
3802 {
3803 if (!--spec->count) {
3804 diff_free_filespec_data(spec);
3805 free(spec);
3806 }
3807 }
3808
3809 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3810 int oid_valid, unsigned short mode)
3811 {
3812 if (mode) {
3813 spec->mode = canon_mode(mode);
3814 oidcpy(&spec->oid, oid);
3815 spec->oid_valid = oid_valid;
3816 }
3817 }
3818
3819 /*
3820 * Given a name and sha1 pair, if the index tells us the file in
3821 * the work tree has that object contents, return true, so that
3822 * prepare_temp_file() does not have to inflate and extract.
3823 */
3824 static int reuse_worktree_file(struct index_state *istate,
3825 const char *name,
3826 const struct object_id *oid,
3827 int want_file)
3828 {
3829 const struct cache_entry *ce;
3830 struct stat st;
3831 int pos, len;
3832
3833 /*
3834 * We do not read the cache ourselves here, because the
3835 * benchmark with my previous version that always reads cache
3836 * shows that it makes things worse for diff-tree comparing
3837 * two linux-2.6 kernel trees in an already checked out work
3838 * tree. This is because most diff-tree comparisons deal with
3839 * only a small number of files, while reading the cache is
3840 * expensive for a large project, and its cost outweighs the
3841 * savings we get by not inflating the object to a temporary
3842 * file. Practically, this code only helps when we are used
3843 * by diff-cache --cached, which does read the cache before
3844 * calling us.
3845 */
3846 if (!istate->cache)
3847 return 0;
3848
3849 /* We want to avoid the working directory if our caller
3850 * doesn't need the data in a normal file, this system
3851 * is rather slow with its stat/open/mmap/close syscalls,
3852 * and the object is contained in a pack file. The pack
3853 * is probably already open and will be faster to obtain
3854 * the data through than the working directory. Loose
3855 * objects however would tend to be slower as they need
3856 * to be individually opened and inflated.
3857 */
3858 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3859 return 0;
3860
3861 /*
3862 * Similarly, if we'd have to convert the file contents anyway, that
3863 * makes the optimization not worthwhile.
3864 */
3865 if (!want_file && would_convert_to_git(istate, name))
3866 return 0;
3867
3868 len = strlen(name);
3869 pos = index_name_pos(istate, name, len);
3870 if (pos < 0)
3871 return 0;
3872 ce = istate->cache[pos];
3873
3874 /*
3875 * This is not the sha1 we are looking for, or
3876 * unreusable because it is not a regular file.
3877 */
3878 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3879 return 0;
3880
3881 /*
3882 * If ce is marked as "assume unchanged", there is no
3883 * guarantee that work tree matches what we are looking for.
3884 */
3885 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3886 return 0;
3887
3888 /*
3889 * If ce matches the file in the work tree, we can reuse it.
3890 */
3891 if (ce_uptodate(ce) ||
3892 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
3893 return 1;
3894
3895 return 0;
3896 }
3897
3898 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3899 {
3900 struct strbuf buf = STRBUF_INIT;
3901 char *dirty = "";
3902
3903 /* Are we looking at the work tree? */
3904 if (s->dirty_submodule)
3905 dirty = "-dirty";
3906
3907 strbuf_addf(&buf, "Subproject commit %s%s\n",
3908 oid_to_hex(&s->oid), dirty);
3909 s->size = buf.len;
3910 if (size_only) {
3911 s->data = NULL;
3912 strbuf_release(&buf);
3913 } else {
3914 s->data = strbuf_detach(&buf, NULL);
3915 s->should_free = 1;
3916 }
3917 return 0;
3918 }
3919
3920 /*
3921 * While doing rename detection and pickaxe operation, we may need to
3922 * grab the data for the blob (or file) for our own in-core comparison.
3923 * diff_filespec has data and size fields for this purpose.
3924 */
3925 int diff_populate_filespec(struct repository *r,
3926 struct diff_filespec *s,
3927 unsigned int flags)
3928 {
3929 int size_only = flags & CHECK_SIZE_ONLY;
3930 int err = 0;
3931 int conv_flags = global_conv_flags_eol;
3932 /*
3933 * demote FAIL to WARN to allow inspecting the situation
3934 * instead of refusing.
3935 */
3936 if (conv_flags & CONV_EOL_RNDTRP_DIE)
3937 conv_flags = CONV_EOL_RNDTRP_WARN;
3938
3939 if (!DIFF_FILE_VALID(s))
3940 die("internal error: asking to populate invalid file.");
3941 if (S_ISDIR(s->mode))
3942 return -1;
3943
3944 if (s->data)
3945 return 0;
3946
3947 if (size_only && 0 < s->size)
3948 return 0;
3949
3950 if (S_ISGITLINK(s->mode))
3951 return diff_populate_gitlink(s, size_only);
3952
3953 if (!s->oid_valid ||
3954 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
3955 struct strbuf buf = STRBUF_INIT;
3956 struct stat st;
3957 int fd;
3958
3959 if (lstat(s->path, &st) < 0) {
3960 err_empty:
3961 err = -1;
3962 empty:
3963 s->data = (char *)"";
3964 s->size = 0;
3965 return err;
3966 }
3967 s->size = xsize_t(st.st_size);
3968 if (!s->size)
3969 goto empty;
3970 if (S_ISLNK(st.st_mode)) {
3971 struct strbuf sb = STRBUF_INIT;
3972
3973 if (strbuf_readlink(&sb, s->path, s->size))
3974 goto err_empty;
3975 s->size = sb.len;
3976 s->data = strbuf_detach(&sb, NULL);
3977 s->should_free = 1;
3978 return 0;
3979 }
3980
3981 /*
3982 * Even if the caller would be happy with getting
3983 * only the size, we cannot return early at this
3984 * point if the path requires us to run the content
3985 * conversion.
3986 */
3987 if (size_only && !would_convert_to_git(r->index, s->path))
3988 return 0;
3989
3990 /*
3991 * Note: this check uses xsize_t(st.st_size) that may
3992 * not be the true size of the blob after it goes
3993 * through convert_to_git(). This may not strictly be
3994 * correct, but the whole point of big_file_threshold
3995 * and is_binary check being that we want to avoid
3996 * opening the file and inspecting the contents, this
3997 * is probably fine.
3998 */
3999 if ((flags & CHECK_BINARY) &&
4000 s->size > big_file_threshold && s->is_binary == -1) {
4001 s->is_binary = 1;
4002 return 0;
4003 }
4004 fd = open(s->path, O_RDONLY);
4005 if (fd < 0)
4006 goto err_empty;
4007 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4008 close(fd);
4009 s->should_munmap = 1;
4010
4011 /*
4012 * Convert from working tree format to canonical git format
4013 */
4014 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4015 size_t size = 0;
4016 munmap(s->data, s->size);
4017 s->should_munmap = 0;
4018 s->data = strbuf_detach(&buf, &size);
4019 s->size = size;
4020 s->should_free = 1;
4021 }
4022 }
4023 else {
4024 enum object_type type;
4025 if (size_only || (flags & CHECK_BINARY)) {
4026 type = oid_object_info(r, &s->oid, &s->size);
4027 if (type < 0)
4028 die("unable to read %s",
4029 oid_to_hex(&s->oid));
4030 if (size_only)
4031 return 0;
4032 if (s->size > big_file_threshold && s->is_binary == -1) {
4033 s->is_binary = 1;
4034 return 0;
4035 }
4036 }
4037 s->data = read_object_file(&s->oid, &type, &s->size);
4038 if (!s->data)
4039 die("unable to read %s", oid_to_hex(&s->oid));
4040 s->should_free = 1;
4041 }
4042 return 0;
4043 }
4044
4045 void diff_free_filespec_blob(struct diff_filespec *s)
4046 {
4047 if (s->should_free)
4048 free(s->data);
4049 else if (s->should_munmap)
4050 munmap(s->data, s->size);
4051
4052 if (s->should_free || s->should_munmap) {
4053 s->should_free = s->should_munmap = 0;
4054 s->data = NULL;
4055 }
4056 }
4057
4058 void diff_free_filespec_data(struct diff_filespec *s)
4059 {
4060 diff_free_filespec_blob(s);
4061 FREE_AND_NULL(s->cnt_data);
4062 }
4063
4064 static void prep_temp_blob(struct index_state *istate,
4065 const char *path, struct diff_tempfile *temp,
4066 void *blob,
4067 unsigned long size,
4068 const struct object_id *oid,
4069 int mode)
4070 {
4071 struct strbuf buf = STRBUF_INIT;
4072 struct strbuf tempfile = STRBUF_INIT;
4073 char *path_dup = xstrdup(path);
4074 const char *base = basename(path_dup);
4075
4076 /* Generate "XXXXXX_basename.ext" */
4077 strbuf_addstr(&tempfile, "XXXXXX_");
4078 strbuf_addstr(&tempfile, base);
4079
4080 temp->tempfile = mks_tempfile_ts(tempfile.buf, strlen(base) + 1);
4081 if (!temp->tempfile)
4082 die_errno("unable to create temp-file");
4083 if (convert_to_working_tree(istate, path,
4084 (const char *)blob, (size_t)size, &buf)) {
4085 blob = buf.buf;
4086 size = buf.len;
4087 }
4088 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4089 close_tempfile_gently(temp->tempfile))
4090 die_errno("unable to write temp-file");
4091 temp->name = get_tempfile_path(temp->tempfile);
4092 oid_to_hex_r(temp->hex, oid);
4093 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4094 strbuf_release(&buf);
4095 strbuf_release(&tempfile);
4096 free(path_dup);
4097 }
4098
4099 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4100 const char *name,
4101 struct diff_filespec *one)
4102 {
4103 struct diff_tempfile *temp = claim_diff_tempfile();
4104
4105 if (!DIFF_FILE_VALID(one)) {
4106 not_a_valid_file:
4107 /* A '-' entry produces this for file-2, and
4108 * a '+' entry produces this for file-1.
4109 */
4110 temp->name = "/dev/null";
4111 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4112 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4113 return temp;
4114 }
4115
4116 if (!S_ISGITLINK(one->mode) &&
4117 (!one->oid_valid ||
4118 reuse_worktree_file(r->index, name, &one->oid, 1))) {
4119 struct stat st;
4120 if (lstat(name, &st) < 0) {
4121 if (errno == ENOENT)
4122 goto not_a_valid_file;
4123 die_errno("stat(%s)", name);
4124 }
4125 if (S_ISLNK(st.st_mode)) {
4126 struct strbuf sb = STRBUF_INIT;
4127 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4128 die_errno("readlink(%s)", name);
4129 prep_temp_blob(r->index, name, temp, sb.buf, sb.len,
4130 (one->oid_valid ?
4131 &one->oid : &null_oid),
4132 (one->oid_valid ?
4133 one->mode : S_IFLNK));
4134 strbuf_release(&sb);
4135 }
4136 else {
4137 /* we can borrow from the file in the work tree */
4138 temp->name = name;
4139 if (!one->oid_valid)
4140 oid_to_hex_r(temp->hex, &null_oid);
4141 else
4142 oid_to_hex_r(temp->hex, &one->oid);
4143 /* Even though we may sometimes borrow the
4144 * contents from the work tree, we always want
4145 * one->mode. mode is trustworthy even when
4146 * !(one->oid_valid), as long as
4147 * DIFF_FILE_VALID(one).
4148 */
4149 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4150 }
4151 return temp;
4152 }
4153 else {
4154 if (diff_populate_filespec(r, one, 0))
4155 die("cannot read data blob for %s", one->path);
4156 prep_temp_blob(r->index, name, temp,
4157 one->data, one->size,
4158 &one->oid, one->mode);
4159 }
4160 return temp;
4161 }
4162
4163 static void add_external_diff_name(struct repository *r,
4164 struct argv_array *argv,
4165 const char *name,
4166 struct diff_filespec *df)
4167 {
4168 struct diff_tempfile *temp = prepare_temp_file(r, name, df);
4169 argv_array_push(argv, temp->name);
4170 argv_array_push(argv, temp->hex);
4171 argv_array_push(argv, temp->mode);
4172 }
4173
4174 /* An external diff command takes:
4175 *
4176 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4177 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4178 *
4179 */
4180 static void run_external_diff(const char *pgm,
4181 const char *name,
4182 const char *other,
4183 struct diff_filespec *one,
4184 struct diff_filespec *two,
4185 const char *xfrm_msg,
4186 struct diff_options *o)
4187 {
4188 struct argv_array argv = ARGV_ARRAY_INIT;
4189 struct argv_array env = ARGV_ARRAY_INIT;
4190 struct diff_queue_struct *q = &diff_queued_diff;
4191
4192 argv_array_push(&argv, pgm);
4193 argv_array_push(&argv, name);
4194
4195 if (one && two) {
4196 add_external_diff_name(o->repo, &argv, name, one);
4197 if (!other)
4198 add_external_diff_name(o->repo, &argv, name, two);
4199 else {
4200 add_external_diff_name(o->repo, &argv, other, two);
4201 argv_array_push(&argv, other);
4202 argv_array_push(&argv, xfrm_msg);
4203 }
4204 }
4205
4206 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
4207 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4208
4209 diff_free_filespec_data(one);
4210 diff_free_filespec_data(two);
4211 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
4212 die(_("external diff died, stopping at %s"), name);
4213
4214 remove_tempfile();
4215 argv_array_clear(&argv);
4216 argv_array_clear(&env);
4217 }
4218
4219 static int similarity_index(struct diff_filepair *p)
4220 {
4221 return p->score * 100 / MAX_SCORE;
4222 }
4223
4224 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4225 {
4226 if (startup_info->have_repository)
4227 return find_unique_abbrev(oid, abbrev);
4228 else {
4229 char *hex = oid_to_hex(oid);
4230 if (abbrev < 0)
4231 abbrev = FALLBACK_DEFAULT_ABBREV;
4232 if (abbrev > the_hash_algo->hexsz)
4233 BUG("oid abbreviation out of range: %d", abbrev);
4234 if (abbrev)
4235 hex[abbrev] = '\0';
4236 return hex;
4237 }
4238 }
4239
4240 static void fill_metainfo(struct strbuf *msg,
4241 const char *name,
4242 const char *other,
4243 struct diff_filespec *one,
4244 struct diff_filespec *two,
4245 struct diff_options *o,
4246 struct diff_filepair *p,
4247 int *must_show_header,
4248 int use_color)
4249 {
4250 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4251 const char *reset = diff_get_color(use_color, DIFF_RESET);
4252 const char *line_prefix = diff_line_prefix(o);
4253
4254 *must_show_header = 1;
4255 strbuf_init(msg, PATH_MAX * 2 + 300);
4256 switch (p->status) {
4257 case DIFF_STATUS_COPIED:
4258 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4259 line_prefix, set, similarity_index(p));
4260 strbuf_addf(msg, "%s\n%s%scopy from ",
4261 reset, line_prefix, set);
4262 quote_c_style(name, msg, NULL, 0);
4263 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4264 quote_c_style(other, msg, NULL, 0);
4265 strbuf_addf(msg, "%s\n", reset);
4266 break;
4267 case DIFF_STATUS_RENAMED:
4268 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4269 line_prefix, set, similarity_index(p));
4270 strbuf_addf(msg, "%s\n%s%srename from ",
4271 reset, line_prefix, set);
4272 quote_c_style(name, msg, NULL, 0);
4273 strbuf_addf(msg, "%s\n%s%srename to ",
4274 reset, line_prefix, set);
4275 quote_c_style(other, msg, NULL, 0);
4276 strbuf_addf(msg, "%s\n", reset);
4277 break;
4278 case DIFF_STATUS_MODIFIED:
4279 if (p->score) {
4280 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4281 line_prefix,
4282 set, similarity_index(p), reset);
4283 break;
4284 }
4285 /* fallthru */
4286 default:
4287 *must_show_header = 0;
4288 }
4289 if (one && two && !oideq(&one->oid, &two->oid)) {
4290 const unsigned hexsz = the_hash_algo->hexsz;
4291 int abbrev = o->flags.full_index ? hexsz : DEFAULT_ABBREV;
4292
4293 if (o->flags.binary) {
4294 mmfile_t mf;
4295 if ((!fill_mmfile(o->repo, &mf, one) &&
4296 diff_filespec_is_binary(o->repo, one)) ||
4297 (!fill_mmfile(o->repo, &mf, two) &&
4298 diff_filespec_is_binary(o->repo, two)))
4299 abbrev = hexsz;
4300 }
4301 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4302 diff_abbrev_oid(&one->oid, abbrev),
4303 diff_abbrev_oid(&two->oid, abbrev));
4304 if (one->mode == two->mode)
4305 strbuf_addf(msg, " %06o", one->mode);
4306 strbuf_addf(msg, "%s\n", reset);
4307 }
4308 }
4309
4310 static void run_diff_cmd(const char *pgm,
4311 const char *name,
4312 const char *other,
4313 const char *attr_path,
4314 struct diff_filespec *one,
4315 struct diff_filespec *two,
4316 struct strbuf *msg,
4317 struct diff_options *o,
4318 struct diff_filepair *p)
4319 {
4320 const char *xfrm_msg = NULL;
4321 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4322 int must_show_header = 0;
4323
4324
4325 if (o->flags.allow_external) {
4326 struct userdiff_driver *drv;
4327
4328 drv = userdiff_find_by_path(o->repo->index, attr_path);
4329 if (drv && drv->external)
4330 pgm = drv->external;
4331 }
4332
4333 if (msg) {
4334 /*
4335 * don't use colors when the header is intended for an
4336 * external diff driver
4337 */
4338 fill_metainfo(msg, name, other, one, two, o, p,
4339 &must_show_header,
4340 want_color(o->use_color) && !pgm);
4341 xfrm_msg = msg->len ? msg->buf : NULL;
4342 }
4343
4344 if (pgm) {
4345 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4346 return;
4347 }
4348 if (one && two)
4349 builtin_diff(name, other ? other : name,
4350 one, two, xfrm_msg, must_show_header,
4351 o, complete_rewrite);
4352 else
4353 fprintf(o->file, "* Unmerged path %s\n", name);
4354 }
4355
4356 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4357 {
4358 if (DIFF_FILE_VALID(one)) {
4359 if (!one->oid_valid) {
4360 struct stat st;
4361 if (one->is_stdin) {
4362 oidclr(&one->oid);
4363 return;
4364 }
4365 if (lstat(one->path, &st) < 0)
4366 die_errno("stat '%s'", one->path);
4367 if (index_path(istate, &one->oid, one->path, &st, 0))
4368 die("cannot hash %s", one->path);
4369 }
4370 }
4371 else
4372 oidclr(&one->oid);
4373 }
4374
4375 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4376 {
4377 /* Strip the prefix but do not molest /dev/null and absolute paths */
4378 if (*namep && !is_absolute_path(*namep)) {
4379 *namep += prefix_length;
4380 if (**namep == '/')
4381 ++*namep;
4382 }
4383 if (*otherp && !is_absolute_path(*otherp)) {
4384 *otherp += prefix_length;
4385 if (**otherp == '/')
4386 ++*otherp;
4387 }
4388 }
4389
4390 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4391 {
4392 const char *pgm = external_diff();
4393 struct strbuf msg;
4394 struct diff_filespec *one = p->one;
4395 struct diff_filespec *two = p->two;
4396 const char *name;
4397 const char *other;
4398 const char *attr_path;
4399
4400 name = one->path;
4401 other = (strcmp(name, two->path) ? two->path : NULL);
4402 attr_path = name;
4403 if (o->prefix_length)
4404 strip_prefix(o->prefix_length, &name, &other);
4405
4406 if (!o->flags.allow_external)
4407 pgm = NULL;
4408
4409 if (DIFF_PAIR_UNMERGED(p)) {
4410 run_diff_cmd(pgm, name, NULL, attr_path,
4411 NULL, NULL, NULL, o, p);
4412 return;
4413 }
4414
4415 diff_fill_oid_info(one, o->repo->index);
4416 diff_fill_oid_info(two, o->repo->index);
4417
4418 if (!pgm &&
4419 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4420 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4421 /*
4422 * a filepair that changes between file and symlink
4423 * needs to be split into deletion and creation.
4424 */
4425 struct diff_filespec *null = alloc_filespec(two->path);
4426 run_diff_cmd(NULL, name, other, attr_path,
4427 one, null, &msg,
4428 o, p);
4429 free(null);
4430 strbuf_release(&msg);
4431
4432 null = alloc_filespec(one->path);
4433 run_diff_cmd(NULL, name, other, attr_path,
4434 null, two, &msg, o, p);
4435 free(null);
4436 }
4437 else
4438 run_diff_cmd(pgm, name, other, attr_path,
4439 one, two, &msg, o, p);
4440
4441 strbuf_release(&msg);
4442 }
4443
4444 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4445 struct diffstat_t *diffstat)
4446 {
4447 const char *name;
4448 const char *other;
4449
4450 if (DIFF_PAIR_UNMERGED(p)) {
4451 /* unmerged */
4452 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4453 diffstat, o, p);
4454 return;
4455 }
4456
4457 name = p->one->path;
4458 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4459
4460 if (o->prefix_length)
4461 strip_prefix(o->prefix_length, &name, &other);
4462
4463 diff_fill_oid_info(p->one, o->repo->index);
4464 diff_fill_oid_info(p->two, o->repo->index);
4465
4466 builtin_diffstat(name, other, p->one, p->two,
4467 diffstat, o, p);
4468 }
4469
4470 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4471 {
4472 const char *name;
4473 const char *other;
4474 const char *attr_path;
4475
4476 if (DIFF_PAIR_UNMERGED(p)) {
4477 /* unmerged */
4478 return;
4479 }
4480
4481 name = p->one->path;
4482 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4483 attr_path = other ? other : name;
4484
4485 if (o->prefix_length)
4486 strip_prefix(o->prefix_length, &name, &other);
4487
4488 diff_fill_oid_info(p->one, o->repo->index);
4489 diff_fill_oid_info(p->two, o->repo->index);
4490
4491 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4492 }
4493
4494 static void prep_parse_options(struct diff_options *options);
4495
4496 void repo_diff_setup(struct repository *r, struct diff_options *options)
4497 {
4498 memcpy(options, &default_diff_options, sizeof(*options));
4499
4500 options->file = stdout;
4501 options->repo = r;
4502
4503 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4504 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4505 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4506 options->abbrev = DEFAULT_ABBREV;
4507 options->line_termination = '\n';
4508 options->break_opt = -1;
4509 options->rename_limit = -1;
4510 options->dirstat_permille = diff_dirstat_permille_default;
4511 options->context = diff_context_default;
4512 options->interhunkcontext = diff_interhunk_context_default;
4513 options->ws_error_highlight = ws_error_highlight_default;
4514 options->flags.rename_empty = 1;
4515 options->objfind = NULL;
4516
4517 /* pathchange left =NULL by default */
4518 options->change = diff_change;
4519 options->add_remove = diff_addremove;
4520 options->use_color = diff_use_color_default;
4521 options->detect_rename = diff_detect_rename_default;
4522 options->xdl_opts |= diff_algorithm;
4523 if (diff_indent_heuristic)
4524 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4525
4526 options->orderfile = diff_order_file_cfg;
4527
4528 if (diff_no_prefix) {
4529 options->a_prefix = options->b_prefix = "";
4530 } else if (!diff_mnemonic_prefix) {
4531 options->a_prefix = "a/";
4532 options->b_prefix = "b/";
4533 }
4534
4535 options->color_moved = diff_color_moved_default;
4536 options->color_moved_ws_handling = diff_color_moved_ws_default;
4537
4538 prep_parse_options(options);
4539 }
4540
4541 void diff_setup_done(struct diff_options *options)
4542 {
4543 unsigned check_mask = DIFF_FORMAT_NAME |
4544 DIFF_FORMAT_NAME_STATUS |
4545 DIFF_FORMAT_CHECKDIFF |
4546 DIFF_FORMAT_NO_OUTPUT;
4547 /*
4548 * This must be signed because we're comparing against a potentially
4549 * negative value.
4550 */
4551 const int hexsz = the_hash_algo->hexsz;
4552
4553 if (options->set_default)
4554 options->set_default(options);
4555
4556 if (HAS_MULTI_BITS(options->output_format & check_mask))
4557 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4558
4559 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4560 die(_("-G, -S and --find-object are mutually exclusive"));
4561
4562 /*
4563 * Most of the time we can say "there are changes"
4564 * only by checking if there are changed paths, but
4565 * --ignore-whitespace* options force us to look
4566 * inside contents.
4567 */
4568
4569 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS))
4570 options->flags.diff_from_contents = 1;
4571 else
4572 options->flags.diff_from_contents = 0;
4573
4574 if (options->flags.find_copies_harder)
4575 options->detect_rename = DIFF_DETECT_COPY;
4576
4577 if (!options->flags.relative_name)
4578 options->prefix = NULL;
4579 if (options->prefix)
4580 options->prefix_length = strlen(options->prefix);
4581 else
4582 options->prefix_length = 0;
4583
4584 if (options->output_format & (DIFF_FORMAT_NAME |
4585 DIFF_FORMAT_NAME_STATUS |
4586 DIFF_FORMAT_CHECKDIFF |
4587 DIFF_FORMAT_NO_OUTPUT))
4588 options->output_format &= ~(DIFF_FORMAT_RAW |
4589 DIFF_FORMAT_NUMSTAT |
4590 DIFF_FORMAT_DIFFSTAT |
4591 DIFF_FORMAT_SHORTSTAT |
4592 DIFF_FORMAT_DIRSTAT |
4593 DIFF_FORMAT_SUMMARY |
4594 DIFF_FORMAT_PATCH);
4595
4596 /*
4597 * These cases always need recursive; we do not drop caller-supplied
4598 * recursive bits for other formats here.
4599 */
4600 if (options->output_format & (DIFF_FORMAT_PATCH |
4601 DIFF_FORMAT_NUMSTAT |
4602 DIFF_FORMAT_DIFFSTAT |
4603 DIFF_FORMAT_SHORTSTAT |
4604 DIFF_FORMAT_DIRSTAT |
4605 DIFF_FORMAT_SUMMARY |
4606 DIFF_FORMAT_CHECKDIFF))
4607 options->flags.recursive = 1;
4608 /*
4609 * Also pickaxe would not work very well if you do not say recursive
4610 */
4611 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4612 options->flags.recursive = 1;
4613 /*
4614 * When patches are generated, submodules diffed against the work tree
4615 * must be checked for dirtiness too so it can be shown in the output
4616 */
4617 if (options->output_format & DIFF_FORMAT_PATCH)
4618 options->flags.dirty_submodules = 1;
4619
4620 if (options->detect_rename && options->rename_limit < 0)
4621 options->rename_limit = diff_rename_limit_default;
4622 if (hexsz < options->abbrev)
4623 options->abbrev = hexsz; /* full */
4624
4625 /*
4626 * It does not make sense to show the first hit we happened
4627 * to have found. It does not make sense not to return with
4628 * exit code in such a case either.
4629 */
4630 if (options->flags.quick) {
4631 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4632 options->flags.exit_with_status = 1;
4633 }
4634
4635 options->diff_path_counter = 0;
4636
4637 if (options->flags.follow_renames && options->pathspec.nr != 1)
4638 die(_("--follow requires exactly one pathspec"));
4639
4640 if (!options->use_color || external_diff())
4641 options->color_moved = 0;
4642
4643 FREE_AND_NULL(options->parseopts);
4644 }
4645
4646 int parse_long_opt(const char *opt, const char **argv,
4647 const char **optarg)
4648 {
4649 const char *arg = argv[0];
4650 if (!skip_prefix(arg, "--", &arg))
4651 return 0;
4652 if (!skip_prefix(arg, opt, &arg))
4653 return 0;
4654 if (*arg == '=') { /* stuck form: --option=value */
4655 *optarg = arg + 1;
4656 return 1;
4657 }
4658 if (*arg != '\0')
4659 return 0;
4660 /* separate form: --option value */
4661 if (!argv[1])
4662 die("Option '--%s' requires a value", opt);
4663 *optarg = argv[1];
4664 return 2;
4665 }
4666
4667 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4668 {
4669 struct diff_options *options = opt->value;
4670 int width = options->stat_width;
4671 int name_width = options->stat_name_width;
4672 int graph_width = options->stat_graph_width;
4673 int count = options->stat_count;
4674 char *end;
4675
4676 BUG_ON_OPT_NEG(unset);
4677
4678 if (!strcmp(opt->long_name, "stat")) {
4679 if (value) {
4680 width = strtoul(value, &end, 10);
4681 if (*end == ',')
4682 name_width = strtoul(end+1, &end, 10);
4683 if (*end == ',')
4684 count = strtoul(end+1, &end, 10);
4685 if (*end)
4686 return error(_("invalid --stat value: %s"), value);
4687 }
4688 } else if (!strcmp(opt->long_name, "stat-width")) {
4689 width = strtoul(value, &end, 10);
4690 if (*end)
4691 return error(_("%s expects a numerical value"),
4692 opt->long_name);
4693 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4694 name_width = strtoul(value, &end, 10);
4695 if (*end)
4696 return error(_("%s expects a numerical value"),
4697 opt->long_name);
4698 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4699 graph_width = strtoul(value, &end, 10);
4700 if (*end)
4701 return error(_("%s expects a numerical value"),
4702 opt->long_name);
4703 } else if (!strcmp(opt->long_name, "stat-count")) {
4704 count = strtoul(value, &end, 10);
4705 if (*end)
4706 return error(_("%s expects a numerical value"),
4707 opt->long_name);
4708 } else
4709 BUG("%s should not get here", opt->long_name);
4710
4711 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4712 options->stat_name_width = name_width;
4713 options->stat_graph_width = graph_width;
4714 options->stat_width = width;
4715 options->stat_count = count;
4716 return 0;
4717 }
4718
4719 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4720 {
4721 struct strbuf errmsg = STRBUF_INIT;
4722 if (parse_dirstat_params(options, params, &errmsg))
4723 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4724 errmsg.buf);
4725 strbuf_release(&errmsg);
4726 /*
4727 * The caller knows a dirstat-related option is given from the command
4728 * line; allow it to say "return this_function();"
4729 */
4730 options->output_format |= DIFF_FORMAT_DIRSTAT;
4731 return 1;
4732 }
4733
4734 static const char diff_status_letters[] = {
4735 DIFF_STATUS_ADDED,
4736 DIFF_STATUS_COPIED,
4737 DIFF_STATUS_DELETED,
4738 DIFF_STATUS_MODIFIED,
4739 DIFF_STATUS_RENAMED,
4740 DIFF_STATUS_TYPE_CHANGED,
4741 DIFF_STATUS_UNKNOWN,
4742 DIFF_STATUS_UNMERGED,
4743 DIFF_STATUS_FILTER_AON,
4744 DIFF_STATUS_FILTER_BROKEN,
4745 '\0',
4746 };
4747
4748 static unsigned int filter_bit['Z' + 1];
4749
4750 static void prepare_filter_bits(void)
4751 {
4752 int i;
4753
4754 if (!filter_bit[DIFF_STATUS_ADDED]) {
4755 for (i = 0; diff_status_letters[i]; i++)
4756 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4757 }
4758 }
4759
4760 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4761 {
4762 return opt->filter & filter_bit[(int) status];
4763 }
4764
4765 unsigned diff_filter_bit(char status)
4766 {
4767 prepare_filter_bits();
4768 return filter_bit[(int) status];
4769 }
4770
4771 static int diff_opt_diff_filter(const struct option *option,
4772 const char *optarg, int unset)
4773 {
4774 struct diff_options *opt = option->value;
4775 int i, optch;
4776
4777 BUG_ON_OPT_NEG(unset);
4778 prepare_filter_bits();
4779
4780 /*
4781 * If there is a negation e.g. 'd' in the input, and we haven't
4782 * initialized the filter field with another --diff-filter, start
4783 * from full set of bits, except for AON.
4784 */
4785 if (!opt->filter) {
4786 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4787 if (optch < 'a' || 'z' < optch)
4788 continue;
4789 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4790 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4791 break;
4792 }
4793 }
4794
4795 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4796 unsigned int bit;
4797 int negate;
4798
4799 if ('a' <= optch && optch <= 'z') {
4800 negate = 1;
4801 optch = toupper(optch);
4802 } else {
4803 negate = 0;
4804 }
4805
4806 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4807 if (!bit)
4808 return error(_("unknown change class '%c' in --diff-filter=%s"),
4809 optarg[i], optarg);
4810 if (negate)
4811 opt->filter &= ~bit;
4812 else
4813 opt->filter |= bit;
4814 }
4815 return 0;
4816 }
4817
4818 static void enable_patch_output(int *fmt)
4819 {
4820 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4821 *fmt |= DIFF_FORMAT_PATCH;
4822 }
4823
4824 static int diff_opt_ws_error_highlight(const struct option *option,
4825 const char *arg, int unset)
4826 {
4827 struct diff_options *opt = option->value;
4828 int val = parse_ws_error_highlight(arg);
4829
4830 BUG_ON_OPT_NEG(unset);
4831 if (val < 0)
4832 return error(_("unknown value after ws-error-highlight=%.*s"),
4833 -1 - val, arg);
4834 opt->ws_error_highlight = val;
4835 return 0;
4836 }
4837
4838 static int diff_opt_find_object(const struct option *option,
4839 const char *arg, int unset)
4840 {
4841 struct diff_options *opt = option->value;
4842 struct object_id oid;
4843
4844 BUG_ON_OPT_NEG(unset);
4845 if (get_oid(arg, &oid))
4846 return error(_("unable to resolve '%s'"), arg);
4847
4848 if (!opt->objfind)
4849 opt->objfind = xcalloc(1, sizeof(*opt->objfind));
4850
4851 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4852 opt->flags.recursive = 1;
4853 opt->flags.tree_in_recursive = 1;
4854 oidset_insert(opt->objfind, &oid);
4855 return 0;
4856 }
4857
4858 static int diff_opt_anchored(const struct option *opt,
4859 const char *arg, int unset)
4860 {
4861 struct diff_options *options = opt->value;
4862
4863 BUG_ON_OPT_NEG(unset);
4864 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4865 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
4866 options->anchors_alloc);
4867 options->anchors[options->anchors_nr++] = xstrdup(arg);
4868 return 0;
4869 }
4870
4871 static int diff_opt_binary(const struct option *opt,
4872 const char *arg, int unset)
4873 {
4874 struct diff_options *options = opt->value;
4875
4876 BUG_ON_OPT_NEG(unset);
4877 BUG_ON_OPT_ARG(arg);
4878 enable_patch_output(&options->output_format);
4879 options->flags.binary = 1;
4880 return 0;
4881 }
4882
4883 static int diff_opt_break_rewrites(const struct option *opt,
4884 const char *arg, int unset)
4885 {
4886 int *break_opt = opt->value;
4887 int opt1, opt2;
4888
4889 BUG_ON_OPT_NEG(unset);
4890 if (!arg)
4891 arg = "";
4892 opt1 = parse_rename_score(&arg);
4893 if (*arg == 0)
4894 opt2 = 0;
4895 else if (*arg != '/')
4896 return error(_("%s expects <n>/<m> form"), opt->long_name);
4897 else {
4898 arg++;
4899 opt2 = parse_rename_score(&arg);
4900 }
4901 if (*arg != 0)
4902 return error(_("%s expects <n>/<m> form"), opt->long_name);
4903 *break_opt = opt1 | (opt2 << 16);
4904 return 0;
4905 }
4906
4907 static int diff_opt_char(const struct option *opt,
4908 const char *arg, int unset)
4909 {
4910 char *value = opt->value;
4911
4912 BUG_ON_OPT_NEG(unset);
4913 if (arg[1])
4914 return error(_("%s expects a character, got '%s'"),
4915 opt->long_name, arg);
4916 *value = arg[0];
4917 return 0;
4918 }
4919
4920 static int diff_opt_color_moved(const struct option *opt,
4921 const char *arg, int unset)
4922 {
4923 struct diff_options *options = opt->value;
4924
4925 if (unset) {
4926 options->color_moved = COLOR_MOVED_NO;
4927 } else if (!arg) {
4928 if (diff_color_moved_default)
4929 options->color_moved = diff_color_moved_default;
4930 if (options->color_moved == COLOR_MOVED_NO)
4931 options->color_moved = COLOR_MOVED_DEFAULT;
4932 } else {
4933 int cm = parse_color_moved(arg);
4934 if (cm < 0)
4935 return error(_("bad --color-moved argument: %s"), arg);
4936 options->color_moved = cm;
4937 }
4938 return 0;
4939 }
4940
4941 static int diff_opt_color_moved_ws(const struct option *opt,
4942 const char *arg, int unset)
4943 {
4944 struct diff_options *options = opt->value;
4945 unsigned cm;
4946
4947 if (unset) {
4948 options->color_moved_ws_handling = 0;
4949 return 0;
4950 }
4951
4952 cm = parse_color_moved_ws(arg);
4953 if (cm & COLOR_MOVED_WS_ERROR)
4954 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
4955 options->color_moved_ws_handling = cm;
4956 return 0;
4957 }
4958
4959 static int diff_opt_color_words(const struct option *opt,
4960 const char *arg, int unset)
4961 {
4962 struct diff_options *options = opt->value;
4963
4964 BUG_ON_OPT_NEG(unset);
4965 options->use_color = 1;
4966 options->word_diff = DIFF_WORDS_COLOR;
4967 options->word_regex = arg;
4968 return 0;
4969 }
4970
4971 static int diff_opt_compact_summary(const struct option *opt,
4972 const char *arg, int unset)
4973 {
4974 struct diff_options *options = opt->value;
4975
4976 BUG_ON_OPT_ARG(arg);
4977 if (unset) {
4978 options->flags.stat_with_summary = 0;
4979 } else {
4980 options->flags.stat_with_summary = 1;
4981 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4982 }
4983 return 0;
4984 }
4985
4986 static int diff_opt_diff_algorithm(const struct option *opt,
4987 const char *arg, int unset)
4988 {
4989 struct diff_options *options = opt->value;
4990 long value = parse_algorithm_value(arg);
4991
4992 BUG_ON_OPT_NEG(unset);
4993 if (value < 0)
4994 return error(_("option diff-algorithm accepts \"myers\", "
4995 "\"minimal\", \"patience\" and \"histogram\""));
4996
4997 /* clear out previous settings */
4998 DIFF_XDL_CLR(options, NEED_MINIMAL);
4999 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
5000 options->xdl_opts |= value;
5001 return 0;
5002 }
5003
5004 static int diff_opt_dirstat(const struct option *opt,
5005 const char *arg, int unset)
5006 {
5007 struct diff_options *options = opt->value;
5008
5009 BUG_ON_OPT_NEG(unset);
5010 if (!strcmp(opt->long_name, "cumulative")) {
5011 if (arg)
5012 BUG("how come --cumulative take a value?");
5013 arg = "cumulative";
5014 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5015 parse_dirstat_opt(options, "files");
5016 parse_dirstat_opt(options, arg ? arg : "");
5017 return 0;
5018 }
5019
5020 static int diff_opt_find_copies(const struct option *opt,
5021 const char *arg, int unset)
5022 {
5023 struct diff_options *options = opt->value;
5024
5025 BUG_ON_OPT_NEG(unset);
5026 if (!arg)
5027 arg = "";
5028 options->rename_score = parse_rename_score(&arg);
5029 if (*arg != 0)
5030 return error(_("invalid argument to %s"), opt->long_name);
5031
5032 if (options->detect_rename == DIFF_DETECT_COPY)
5033 options->flags.find_copies_harder = 1;
5034 else
5035 options->detect_rename = DIFF_DETECT_COPY;
5036
5037 return 0;
5038 }
5039
5040 static int diff_opt_find_renames(const struct option *opt,
5041 const char *arg, int unset)
5042 {
5043 struct diff_options *options = opt->value;
5044
5045 BUG_ON_OPT_NEG(unset);
5046 if (!arg)
5047 arg = "";
5048 options->rename_score = parse_rename_score(&arg);
5049 if (*arg != 0)
5050 return error(_("invalid argument to %s"), opt->long_name);
5051
5052 options->detect_rename = DIFF_DETECT_RENAME;
5053 return 0;
5054 }
5055
5056 static int diff_opt_follow(const struct option *opt,
5057 const char *arg, int unset)
5058 {
5059 struct diff_options *options = opt->value;
5060
5061 BUG_ON_OPT_ARG(arg);
5062 if (unset) {
5063 options->flags.follow_renames = 0;
5064 options->flags.default_follow_renames = 0;
5065 } else {
5066 options->flags.follow_renames = 1;
5067 }
5068 return 0;
5069 }
5070
5071 static int diff_opt_ignore_submodules(const struct option *opt,
5072 const char *arg, int unset)
5073 {
5074 struct diff_options *options = opt->value;
5075
5076 BUG_ON_OPT_NEG(unset);
5077 if (!arg)
5078 arg = "all";
5079 options->flags.override_submodule_config = 1;
5080 handle_ignore_submodules_arg(options, arg);
5081 return 0;
5082 }
5083
5084 static int diff_opt_line_prefix(const struct option *opt,
5085 const char *optarg, int unset)
5086 {
5087 struct diff_options *options = opt->value;
5088
5089 BUG_ON_OPT_NEG(unset);
5090 options->line_prefix = optarg;
5091 options->line_prefix_length = strlen(options->line_prefix);
5092 graph_setup_line_prefix(options);
5093 return 0;
5094 }
5095
5096 static int diff_opt_no_prefix(const struct option *opt,
5097 const char *optarg, int unset)
5098 {
5099 struct diff_options *options = opt->value;
5100
5101 BUG_ON_OPT_NEG(unset);
5102 BUG_ON_OPT_ARG(optarg);
5103 options->a_prefix = "";
5104 options->b_prefix = "";
5105 return 0;
5106 }
5107
5108 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5109 const struct option *opt,
5110 const char *arg, int unset)
5111 {
5112 struct diff_options *options = opt->value;
5113 char *path;
5114
5115 BUG_ON_OPT_NEG(unset);
5116 path = prefix_filename(ctx->prefix, arg);
5117 options->file = xfopen(path, "w");
5118 options->close_file = 1;
5119 if (options->use_color != GIT_COLOR_ALWAYS)
5120 options->use_color = GIT_COLOR_NEVER;
5121 free(path);
5122 return 0;
5123 }
5124
5125 static int diff_opt_patience(const struct option *opt,
5126 const char *arg, int unset)
5127 {
5128 struct diff_options *options = opt->value;
5129 int i;
5130
5131 BUG_ON_OPT_NEG(unset);
5132 BUG_ON_OPT_ARG(arg);
5133 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5134 /*
5135 * Both --patience and --anchored use PATIENCE_DIFF
5136 * internally, so remove any anchors previously
5137 * specified.
5138 */
5139 for (i = 0; i < options->anchors_nr; i++)
5140 free(options->anchors[i]);
5141 options->anchors_nr = 0;
5142 return 0;
5143 }
5144
5145 static int diff_opt_pickaxe_regex(const struct option *opt,
5146 const char *arg, int unset)
5147 {
5148 struct diff_options *options = opt->value;
5149
5150 BUG_ON_OPT_NEG(unset);
5151 options->pickaxe = arg;
5152 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5153 return 0;
5154 }
5155
5156 static int diff_opt_pickaxe_string(const struct option *opt,
5157 const char *arg, int unset)
5158 {
5159 struct diff_options *options = opt->value;
5160
5161 BUG_ON_OPT_NEG(unset);
5162 options->pickaxe = arg;
5163 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5164 return 0;
5165 }
5166
5167 static int diff_opt_relative(const struct option *opt,
5168 const char *arg, int unset)
5169 {
5170 struct diff_options *options = opt->value;
5171
5172 BUG_ON_OPT_NEG(unset);
5173 options->flags.relative_name = 1;
5174 if (arg)
5175 options->prefix = arg;
5176 return 0;
5177 }
5178
5179 static int diff_opt_submodule(const struct option *opt,
5180 const char *arg, int unset)
5181 {
5182 struct diff_options *options = opt->value;
5183
5184 BUG_ON_OPT_NEG(unset);
5185 if (!arg)
5186 arg = "log";
5187 if (parse_submodule_params(options, arg))
5188 return error(_("failed to parse --submodule option parameter: '%s'"),
5189 arg);
5190 return 0;
5191 }
5192
5193 static int diff_opt_textconv(const struct option *opt,
5194 const char *arg, int unset)
5195 {
5196 struct diff_options *options = opt->value;
5197
5198 BUG_ON_OPT_ARG(arg);
5199 if (unset) {
5200 options->flags.allow_textconv = 0;
5201 } else {
5202 options->flags.allow_textconv = 1;
5203 options->flags.textconv_set_via_cmdline = 1;
5204 }
5205 return 0;
5206 }
5207
5208 static int diff_opt_unified(const struct option *opt,
5209 const char *arg, int unset)
5210 {
5211 struct diff_options *options = opt->value;
5212 char *s;
5213
5214 BUG_ON_OPT_NEG(unset);
5215
5216 if (arg) {
5217 options->context = strtol(arg, &s, 10);
5218 if (*s)
5219 return error(_("%s expects a numerical value"), "--unified");
5220 }
5221 enable_patch_output(&options->output_format);
5222
5223 return 0;
5224 }
5225
5226 static int diff_opt_word_diff(const struct option *opt,
5227 const char *arg, int unset)
5228 {
5229 struct diff_options *options = opt->value;
5230
5231 BUG_ON_OPT_NEG(unset);
5232 if (arg) {
5233 if (!strcmp(arg, "plain"))
5234 options->word_diff = DIFF_WORDS_PLAIN;
5235 else if (!strcmp(arg, "color")) {
5236 options->use_color = 1;
5237 options->word_diff = DIFF_WORDS_COLOR;
5238 }
5239 else if (!strcmp(arg, "porcelain"))
5240 options->word_diff = DIFF_WORDS_PORCELAIN;
5241 else if (!strcmp(arg, "none"))
5242 options->word_diff = DIFF_WORDS_NONE;
5243 else
5244 return error(_("bad --word-diff argument: %s"), arg);
5245 } else {
5246 if (options->word_diff == DIFF_WORDS_NONE)
5247 options->word_diff = DIFF_WORDS_PLAIN;
5248 }
5249 return 0;
5250 }
5251
5252 static int diff_opt_word_diff_regex(const struct option *opt,
5253 const char *arg, int unset)
5254 {
5255 struct diff_options *options = opt->value;
5256
5257 BUG_ON_OPT_NEG(unset);
5258 if (options->word_diff == DIFF_WORDS_NONE)
5259 options->word_diff = DIFF_WORDS_PLAIN;
5260 options->word_regex = arg;
5261 return 0;
5262 }
5263
5264 static void prep_parse_options(struct diff_options *options)
5265 {
5266 struct option parseopts[] = {
5267 OPT_GROUP(N_("Diff output format options")),
5268 OPT_BITOP('p', "patch", &options->output_format,
5269 N_("generate patch"),
5270 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5271 OPT_BIT_F('s', "no-patch", &options->output_format,
5272 N_("suppress diff output"),
5273 DIFF_FORMAT_NO_OUTPUT, PARSE_OPT_NONEG),
5274 OPT_BITOP('u', NULL, &options->output_format,
5275 N_("generate patch"),
5276 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5277 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5278 N_("generate diffs with <n> lines context"),
5279 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5280 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5281 N_("generate diffs with <n> lines context")),
5282 OPT_BIT_F(0, "raw", &options->output_format,
5283 N_("generate the diff in raw format"),
5284 DIFF_FORMAT_RAW, PARSE_OPT_NONEG),
5285 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5286 N_("synonym for '-p --raw'"),
5287 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5288 DIFF_FORMAT_NO_OUTPUT),
5289 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5290 N_("synonym for '-p --stat'"),
5291 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5292 DIFF_FORMAT_NO_OUTPUT),
5293 OPT_BIT_F(0, "numstat", &options->output_format,
5294 N_("machine friendly --stat"),
5295 DIFF_FORMAT_NUMSTAT, PARSE_OPT_NONEG),
5296 OPT_BIT_F(0, "shortstat", &options->output_format,
5297 N_("output only the last line of --stat"),
5298 DIFF_FORMAT_SHORTSTAT, PARSE_OPT_NONEG),
5299 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1,param2>..."),
5300 N_("output the distribution of relative amount of changes for each sub-directory"),
5301 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5302 diff_opt_dirstat),
5303 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5304 N_("synonym for --dirstat=cumulative"),
5305 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5306 diff_opt_dirstat),
5307 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1,param2>..."),
5308 N_("synonym for --dirstat=files,param1,param2..."),
5309 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5310 diff_opt_dirstat),
5311 OPT_BIT_F(0, "check", &options->output_format,
5312 N_("warn if changes introduce conflict markers or whitespace errors"),
5313 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5314 OPT_BIT_F(0, "summary", &options->output_format,
5315 N_("condensed summary such as creations, renames and mode changes"),
5316 DIFF_FORMAT_SUMMARY, PARSE_OPT_NONEG),
5317 OPT_BIT_F(0, "name-only", &options->output_format,
5318 N_("show only names of changed files"),
5319 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5320 OPT_BIT_F(0, "name-status", &options->output_format,
5321 N_("show only names and status of changed files"),
5322 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5323 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5324 N_("generate diffstat"),
5325 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5326 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5327 N_("generate diffstat with a given width"),
5328 PARSE_OPT_NONEG, diff_opt_stat),
5329 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5330 N_("generate diffstat with a given name width"),
5331 PARSE_OPT_NONEG, diff_opt_stat),
5332 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5333 N_("generate diffstat with a given graph width"),
5334 PARSE_OPT_NONEG, diff_opt_stat),
5335 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5336 N_("generate diffstat with limited lines"),
5337 PARSE_OPT_NONEG, diff_opt_stat),
5338 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5339 N_("generate compact summary in diffstat"),
5340 PARSE_OPT_NOARG, diff_opt_compact_summary),
5341 OPT_CALLBACK_F(0, "binary", options, NULL,
5342 N_("output a binary diff that can be applied"),
5343 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5344 OPT_BOOL(0, "full-index", &options->flags.full_index,
5345 N_("show full pre- and post-image object names on the \"index\" lines")),
5346 OPT_COLOR_FLAG(0, "color", &options->use_color,
5347 N_("show colored diff")),
5348 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5349 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5350 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5351 OPT_SET_INT('z', NULL, &options->line_termination,
5352 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5353 0),
5354 OPT__ABBREV(&options->abbrev),
5355 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5356 N_("show the given source prefix instead of \"a/\""),
5357 PARSE_OPT_NONEG),
5358 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5359 N_("show the given destination prefix instead of \"b/\""),
5360 PARSE_OPT_NONEG),
5361 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5362 N_("prepend an additional prefix to every line of output"),
5363 PARSE_OPT_NONEG, diff_opt_line_prefix),
5364 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5365 N_("do not show any source or destination prefix"),
5366 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5367 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5368 N_("show context between diff hunks up to the specified number of lines"),
5369 PARSE_OPT_NONEG),
5370 OPT_CALLBACK_F(0, "output-indicator-new",
5371 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5372 N_("<char>"),
5373 N_("specify the character to indicate a new line instead of '+'"),
5374 PARSE_OPT_NONEG, diff_opt_char),
5375 OPT_CALLBACK_F(0, "output-indicator-old",
5376 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5377 N_("<char>"),
5378 N_("specify the character to indicate an old line instead of '-'"),
5379 PARSE_OPT_NONEG, diff_opt_char),
5380 OPT_CALLBACK_F(0, "output-indicator-context",
5381 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5382 N_("<char>"),
5383 N_("specify the character to indicate a context instead of ' '"),
5384 PARSE_OPT_NONEG, diff_opt_char),
5385
5386 OPT_GROUP(N_("Diff rename options")),
5387 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5388 N_("break complete rewrite changes into pairs of delete and create"),
5389 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5390 diff_opt_break_rewrites),
5391 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5392 N_("detect renames"),
5393 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5394 diff_opt_find_renames),
5395 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5396 N_("omit the preimage for deletes"),
5397 1, PARSE_OPT_NONEG),
5398 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5399 N_("detect copies"),
5400 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5401 diff_opt_find_copies),
5402 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5403 N_("use unmodified files as source to find copies")),
5404 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5405 N_("disable rename detection"),
5406 0, PARSE_OPT_NONEG),
5407 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5408 N_("use empty blobs as rename source")),
5409 OPT_CALLBACK_F(0, "follow", options, NULL,
5410 N_("continue listing the history of a file beyond renames"),
5411 PARSE_OPT_NOARG, diff_opt_follow),
5412 OPT_INTEGER('l', NULL, &options->rename_limit,
5413 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5414
5415 OPT_GROUP(N_("Diff algorithm options")),
5416 OPT_BIT(0, "minimal", &options->xdl_opts,
5417 N_("produce the smallest possible diff"),
5418 XDF_NEED_MINIMAL),
5419 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5420 N_("ignore whitespace when comparing lines"),
5421 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5422 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5423 N_("ignore changes in amount of whitespace"),
5424 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5425 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5426 N_("ignore changes in whitespace at EOL"),
5427 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5428 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5429 N_("ignore carrier-return at the end of line"),
5430 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5431 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5432 N_("ignore changes whose lines are all blank"),
5433 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5434 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5435 N_("heuristic to shift diff hunk boundaries for easy reading"),
5436 XDF_INDENT_HEURISTIC),
5437 OPT_CALLBACK_F(0, "patience", options, NULL,
5438 N_("generate diff using the \"patience diff\" algorithm"),
5439 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5440 diff_opt_patience),
5441 OPT_BITOP(0, "histogram", &options->xdl_opts,
5442 N_("generate diff using the \"histogram diff\" algorithm"),
5443 XDF_HISTOGRAM_DIFF, XDF_DIFF_ALGORITHM_MASK),
5444 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5445 N_("choose a diff algorithm"),
5446 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5447 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5448 N_("generate diff using the \"anchored diff\" algorithm"),
5449 PARSE_OPT_NONEG, diff_opt_anchored),
5450 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5451 N_("show word diff, using <mode> to delimit changed words"),
5452 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5453 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5454 N_("use <regex> to decide what a word is"),
5455 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5456 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5457 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5458 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5459 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5460 N_("moved lines of code are colored differently"),
5461 PARSE_OPT_OPTARG, diff_opt_color_moved),
5462 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5463 N_("how white spaces are ignored in --color-moved"),
5464 0, diff_opt_color_moved_ws),
5465
5466 OPT_GROUP(N_("Other diff options")),
5467 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5468 N_("when run from subdir, exclude changes outside and show relative paths"),
5469 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5470 diff_opt_relative),
5471 OPT_BOOL('a', "text", &options->flags.text,
5472 N_("treat all files as text")),
5473 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5474 N_("swap two inputs, reverse the diff")),
5475 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5476 N_("exit with 1 if there were differences, 0 otherwise")),
5477 OPT_BOOL(0, "quiet", &options->flags.quick,
5478 N_("disable all output of the program")),
5479 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5480 N_("allow an external diff helper to be executed")),
5481 OPT_CALLBACK_F(0, "textconv", options, NULL,
5482 N_("run external text conversion filters when comparing binary files"),
5483 PARSE_OPT_NOARG, diff_opt_textconv),
5484 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5485 N_("ignore changes to submodules in the diff generation"),
5486 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5487 diff_opt_ignore_submodules),
5488 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5489 N_("specify how differences in submodules are shown"),
5490 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5491 diff_opt_submodule),
5492 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5493 N_("hide 'git add -N' entries from the index"),
5494 1, PARSE_OPT_NONEG),
5495 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5496 N_("treat 'git add -N' entries as real in the index"),
5497 0, PARSE_OPT_NONEG),
5498 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5499 N_("look for differences that change the number of occurrences of the specified string"),
5500 0, diff_opt_pickaxe_string),
5501 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5502 N_("look for differences that change the number of occurrences of the specified regex"),
5503 0, diff_opt_pickaxe_regex),
5504 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5505 N_("show all changes in the changeset with -S or -G"),
5506 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5507 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5508 N_("treat <string> in -S as extended POSIX regular expression"),
5509 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5510 OPT_FILENAME('O', NULL, &options->orderfile,
5511 N_("control the order in which files appear in the output")),
5512 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5513 N_("look for differences that change the number of occurrences of the specified object"),
5514 PARSE_OPT_NONEG, diff_opt_find_object),
5515 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5516 N_("select files by diff type"),
5517 PARSE_OPT_NONEG, diff_opt_diff_filter),
5518 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5519 N_("Output to a specific file"),
5520 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5521
5522 OPT_END()
5523 };
5524
5525 ALLOC_ARRAY(options->parseopts, ARRAY_SIZE(parseopts));
5526 memcpy(options->parseopts, parseopts, sizeof(parseopts));
5527 }
5528
5529 int diff_opt_parse(struct diff_options *options,
5530 const char **av, int ac, const char *prefix)
5531 {
5532 if (!prefix)
5533 prefix = "";
5534
5535 ac = parse_options(ac, av, prefix, options->parseopts, NULL,
5536 PARSE_OPT_KEEP_DASHDASH |
5537 PARSE_OPT_KEEP_UNKNOWN |
5538 PARSE_OPT_NO_INTERNAL_HELP |
5539 PARSE_OPT_ONE_SHOT |
5540 PARSE_OPT_STOP_AT_NON_OPTION);
5541
5542 return ac;
5543 }
5544
5545 int parse_rename_score(const char **cp_p)
5546 {
5547 unsigned long num, scale;
5548 int ch, dot;
5549 const char *cp = *cp_p;
5550
5551 num = 0;
5552 scale = 1;
5553 dot = 0;
5554 for (;;) {
5555 ch = *cp;
5556 if ( !dot && ch == '.' ) {
5557 scale = 1;
5558 dot = 1;
5559 } else if ( ch == '%' ) {
5560 scale = dot ? scale*100 : 100;
5561 cp++; /* % is always at the end */
5562 break;
5563 } else if ( ch >= '0' && ch <= '9' ) {
5564 if ( scale < 100000 ) {
5565 scale *= 10;
5566 num = (num*10) + (ch-'0');
5567 }
5568 } else {
5569 break;
5570 }
5571 cp++;
5572 }
5573 *cp_p = cp;
5574
5575 /* user says num divided by scale and we say internally that
5576 * is MAX_SCORE * num / scale.
5577 */
5578 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5579 }
5580
5581 struct diff_queue_struct diff_queued_diff;
5582
5583 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5584 {
5585 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5586 queue->queue[queue->nr++] = dp;
5587 }
5588
5589 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5590 struct diff_filespec *one,
5591 struct diff_filespec *two)
5592 {
5593 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5594 dp->one = one;
5595 dp->two = two;
5596 if (queue)
5597 diff_q(queue, dp);
5598 return dp;
5599 }
5600
5601 void diff_free_filepair(struct diff_filepair *p)
5602 {
5603 free_filespec(p->one);
5604 free_filespec(p->two);
5605 free(p);
5606 }
5607
5608 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5609 {
5610 int abblen;
5611 const char *abbrev;
5612
5613 /* Do we want all 40 hex characters? */
5614 if (len == the_hash_algo->hexsz)
5615 return oid_to_hex(oid);
5616
5617 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5618 abbrev = diff_abbrev_oid(oid, len);
5619
5620 if (!print_sha1_ellipsis())
5621 return abbrev;
5622
5623 abblen = strlen(abbrev);
5624
5625 /*
5626 * In well-behaved cases, where the abbreviated result is the
5627 * same as the requested length, append three dots after the
5628 * abbreviation (hence the whole logic is limited to the case
5629 * where abblen < 37); when the actual abbreviated result is a
5630 * bit longer than the requested length, we reduce the number
5631 * of dots so that they match the well-behaved ones. However,
5632 * if the actual abbreviation is longer than the requested
5633 * length by more than three, we give up on aligning, and add
5634 * three dots anyway, to indicate that the output is not the
5635 * full object name. Yes, this may be suboptimal, but this
5636 * appears only in "diff --raw --abbrev" output and it is not
5637 * worth the effort to change it now. Note that this would
5638 * likely to work fine when the automatic sizing of default
5639 * abbreviation length is used--we would be fed -1 in "len" in
5640 * that case, and will end up always appending three-dots, but
5641 * the automatic sizing is supposed to give abblen that ensures
5642 * uniqueness across all objects (statistically speaking).
5643 */
5644 if (abblen < the_hash_algo->hexsz - 3) {
5645 static char hex[GIT_MAX_HEXSZ + 1];
5646 if (len < abblen && abblen <= len + 2)
5647 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5648 else
5649 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5650 return hex;
5651 }
5652
5653 return oid_to_hex(oid);
5654 }
5655
5656 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5657 {
5658 int line_termination = opt->line_termination;
5659 int inter_name_termination = line_termination ? '\t' : '\0';
5660
5661 fprintf(opt->file, "%s", diff_line_prefix(opt));
5662 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5663 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5664 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5665 fprintf(opt->file, "%s ",
5666 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5667 }
5668 if (p->score) {
5669 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5670 inter_name_termination);
5671 } else {
5672 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5673 }
5674
5675 if (p->status == DIFF_STATUS_COPIED ||
5676 p->status == DIFF_STATUS_RENAMED) {
5677 const char *name_a, *name_b;
5678 name_a = p->one->path;
5679 name_b = p->two->path;
5680 strip_prefix(opt->prefix_length, &name_a, &name_b);
5681 write_name_quoted(name_a, opt->file, inter_name_termination);
5682 write_name_quoted(name_b, opt->file, line_termination);
5683 } else {
5684 const char *name_a, *name_b;
5685 name_a = p->one->mode ? p->one->path : p->two->path;
5686 name_b = NULL;
5687 strip_prefix(opt->prefix_length, &name_a, &name_b);
5688 write_name_quoted(name_a, opt->file, line_termination);
5689 }
5690 }
5691
5692 int diff_unmodified_pair(struct diff_filepair *p)
5693 {
5694 /* This function is written stricter than necessary to support
5695 * the currently implemented transformers, but the idea is to
5696 * let transformers to produce diff_filepairs any way they want,
5697 * and filter and clean them up here before producing the output.
5698 */
5699 struct diff_filespec *one = p->one, *two = p->two;
5700
5701 if (DIFF_PAIR_UNMERGED(p))
5702 return 0; /* unmerged is interesting */
5703
5704 /* deletion, addition, mode or type change
5705 * and rename are all interesting.
5706 */
5707 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5708 DIFF_PAIR_MODE_CHANGED(p) ||
5709 strcmp(one->path, two->path))
5710 return 0;
5711
5712 /* both are valid and point at the same path. that is, we are
5713 * dealing with a change.
5714 */
5715 if (one->oid_valid && two->oid_valid &&
5716 oideq(&one->oid, &two->oid) &&
5717 !one->dirty_submodule && !two->dirty_submodule)
5718 return 1; /* no change */
5719 if (!one->oid_valid && !two->oid_valid)
5720 return 1; /* both look at the same file on the filesystem. */
5721 return 0;
5722 }
5723
5724 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5725 {
5726 if (diff_unmodified_pair(p))
5727 return;
5728
5729 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5730 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5731 return; /* no tree diffs in patch format */
5732
5733 run_diff(p, o);
5734 }
5735
5736 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5737 struct diffstat_t *diffstat)
5738 {
5739 if (diff_unmodified_pair(p))
5740 return;
5741
5742 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5743 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5744 return; /* no useful stat for tree diffs */
5745
5746 run_diffstat(p, o, diffstat);
5747 }
5748
5749 static void diff_flush_checkdiff(struct diff_filepair *p,
5750 struct diff_options *o)
5751 {
5752 if (diff_unmodified_pair(p))
5753 return;
5754
5755 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5756 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5757 return; /* nothing to check in tree diffs */
5758
5759 run_checkdiff(p, o);
5760 }
5761
5762 int diff_queue_is_empty(void)
5763 {
5764 struct diff_queue_struct *q = &diff_queued_diff;
5765 int i;
5766 for (i = 0; i < q->nr; i++)
5767 if (!diff_unmodified_pair(q->queue[i]))
5768 return 0;
5769 return 1;
5770 }
5771
5772 #if DIFF_DEBUG
5773 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5774 {
5775 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5776 x, one ? one : "",
5777 s->path,
5778 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5779 s->mode,
5780 s->oid_valid ? oid_to_hex(&s->oid) : "");
5781 fprintf(stderr, "queue[%d] %s size %lu\n",
5782 x, one ? one : "",
5783 s->size);
5784 }
5785
5786 void diff_debug_filepair(const struct diff_filepair *p, int i)
5787 {
5788 diff_debug_filespec(p->one, i, "one");
5789 diff_debug_filespec(p->two, i, "two");
5790 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5791 p->score, p->status ? p->status : '?',
5792 p->one->rename_used, p->broken_pair);
5793 }
5794
5795 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5796 {
5797 int i;
5798 if (msg)
5799 fprintf(stderr, "%s\n", msg);
5800 fprintf(stderr, "q->nr = %d\n", q->nr);
5801 for (i = 0; i < q->nr; i++) {
5802 struct diff_filepair *p = q->queue[i];
5803 diff_debug_filepair(p, i);
5804 }
5805 }
5806 #endif
5807
5808 static void diff_resolve_rename_copy(void)
5809 {
5810 int i;
5811 struct diff_filepair *p;
5812 struct diff_queue_struct *q = &diff_queued_diff;
5813
5814 diff_debug_queue("resolve-rename-copy", q);
5815
5816 for (i = 0; i < q->nr; i++) {
5817 p = q->queue[i];
5818 p->status = 0; /* undecided */
5819 if (DIFF_PAIR_UNMERGED(p))
5820 p->status = DIFF_STATUS_UNMERGED;
5821 else if (!DIFF_FILE_VALID(p->one))
5822 p->status = DIFF_STATUS_ADDED;
5823 else if (!DIFF_FILE_VALID(p->two))
5824 p->status = DIFF_STATUS_DELETED;
5825 else if (DIFF_PAIR_TYPE_CHANGED(p))
5826 p->status = DIFF_STATUS_TYPE_CHANGED;
5827
5828 /* from this point on, we are dealing with a pair
5829 * whose both sides are valid and of the same type, i.e.
5830 * either in-place edit or rename/copy edit.
5831 */
5832 else if (DIFF_PAIR_RENAME(p)) {
5833 /*
5834 * A rename might have re-connected a broken
5835 * pair up, causing the pathnames to be the
5836 * same again. If so, that's not a rename at
5837 * all, just a modification..
5838 *
5839 * Otherwise, see if this source was used for
5840 * multiple renames, in which case we decrement
5841 * the count, and call it a copy.
5842 */
5843 if (!strcmp(p->one->path, p->two->path))
5844 p->status = DIFF_STATUS_MODIFIED;
5845 else if (--p->one->rename_used > 0)
5846 p->status = DIFF_STATUS_COPIED;
5847 else
5848 p->status = DIFF_STATUS_RENAMED;
5849 }
5850 else if (!oideq(&p->one->oid, &p->two->oid) ||
5851 p->one->mode != p->two->mode ||
5852 p->one->dirty_submodule ||
5853 p->two->dirty_submodule ||
5854 is_null_oid(&p->one->oid))
5855 p->status = DIFF_STATUS_MODIFIED;
5856 else {
5857 /* This is a "no-change" entry and should not
5858 * happen anymore, but prepare for broken callers.
5859 */
5860 error("feeding unmodified %s to diffcore",
5861 p->one->path);
5862 p->status = DIFF_STATUS_UNKNOWN;
5863 }
5864 }
5865 diff_debug_queue("resolve-rename-copy done", q);
5866 }
5867
5868 static int check_pair_status(struct diff_filepair *p)
5869 {
5870 switch (p->status) {
5871 case DIFF_STATUS_UNKNOWN:
5872 return 0;
5873 case 0:
5874 die("internal error in diff-resolve-rename-copy");
5875 default:
5876 return 1;
5877 }
5878 }
5879
5880 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5881 {
5882 int fmt = opt->output_format;
5883
5884 if (fmt & DIFF_FORMAT_CHECKDIFF)
5885 diff_flush_checkdiff(p, opt);
5886 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5887 diff_flush_raw(p, opt);
5888 else if (fmt & DIFF_FORMAT_NAME) {
5889 const char *name_a, *name_b;
5890 name_a = p->two->path;
5891 name_b = NULL;
5892 strip_prefix(opt->prefix_length, &name_a, &name_b);
5893 fprintf(opt->file, "%s", diff_line_prefix(opt));
5894 write_name_quoted(name_a, opt->file, opt->line_termination);
5895 }
5896 }
5897
5898 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5899 {
5900 struct strbuf sb = STRBUF_INIT;
5901 if (fs->mode)
5902 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5903 else
5904 strbuf_addf(&sb, " %s ", newdelete);
5905
5906 quote_c_style(fs->path, &sb, NULL, 0);
5907 strbuf_addch(&sb, '\n');
5908 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5909 sb.buf, sb.len, 0);
5910 strbuf_release(&sb);
5911 }
5912
5913 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5914 int show_name)
5915 {
5916 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5917 struct strbuf sb = STRBUF_INIT;
5918 strbuf_addf(&sb, " mode change %06o => %06o",
5919 p->one->mode, p->two->mode);
5920 if (show_name) {
5921 strbuf_addch(&sb, ' ');
5922 quote_c_style(p->two->path, &sb, NULL, 0);
5923 }
5924 strbuf_addch(&sb, '\n');
5925 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5926 sb.buf, sb.len, 0);
5927 strbuf_release(&sb);
5928 }
5929 }
5930
5931 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5932 struct diff_filepair *p)
5933 {
5934 struct strbuf sb = STRBUF_INIT;
5935 struct strbuf names = STRBUF_INIT;
5936
5937 pprint_rename(&names, p->one->path, p->two->path);
5938 strbuf_addf(&sb, " %s %s (%d%%)\n",
5939 renamecopy, names.buf, similarity_index(p));
5940 strbuf_release(&names);
5941 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5942 sb.buf, sb.len, 0);
5943 show_mode_change(opt, p, 0);
5944 strbuf_release(&sb);
5945 }
5946
5947 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5948 {
5949 switch(p->status) {
5950 case DIFF_STATUS_DELETED:
5951 show_file_mode_name(opt, "delete", p->one);
5952 break;
5953 case DIFF_STATUS_ADDED:
5954 show_file_mode_name(opt, "create", p->two);
5955 break;
5956 case DIFF_STATUS_COPIED:
5957 show_rename_copy(opt, "copy", p);
5958 break;
5959 case DIFF_STATUS_RENAMED:
5960 show_rename_copy(opt, "rename", p);
5961 break;
5962 default:
5963 if (p->score) {
5964 struct strbuf sb = STRBUF_INIT;
5965 strbuf_addstr(&sb, " rewrite ");
5966 quote_c_style(p->two->path, &sb, NULL, 0);
5967 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5968 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5969 sb.buf, sb.len, 0);
5970 strbuf_release(&sb);
5971 }
5972 show_mode_change(opt, p, !p->score);
5973 break;
5974 }
5975 }
5976
5977 struct patch_id_t {
5978 git_SHA_CTX *ctx;
5979 int patchlen;
5980 };
5981
5982 static int remove_space(char *line, int len)
5983 {
5984 int i;
5985 char *dst = line;
5986 unsigned char c;
5987
5988 for (i = 0; i < len; i++)
5989 if (!isspace((c = line[i])))
5990 *dst++ = c;
5991
5992 return dst - line;
5993 }
5994
5995 static void patch_id_consume(void *priv, char *line, unsigned long len)
5996 {
5997 struct patch_id_t *data = priv;
5998 int new_len;
5999
6000 new_len = remove_space(line, len);
6001
6002 git_SHA1_Update(data->ctx, line, new_len);
6003 data->patchlen += new_len;
6004 }
6005
6006 static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
6007 {
6008 git_SHA1_Update(ctx, str, strlen(str));
6009 }
6010
6011 static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
6012 {
6013 /* large enough for 2^32 in octal */
6014 char buf[12];
6015 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6016 git_SHA1_Update(ctx, buf, len);
6017 }
6018
6019 /* returns 0 upon success, and writes result into sha1 */
6020 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6021 {
6022 struct diff_queue_struct *q = &diff_queued_diff;
6023 int i;
6024 git_SHA_CTX ctx;
6025 struct patch_id_t data;
6026
6027 git_SHA1_Init(&ctx);
6028 memset(&data, 0, sizeof(struct patch_id_t));
6029 data.ctx = &ctx;
6030
6031 for (i = 0; i < q->nr; i++) {
6032 xpparam_t xpp;
6033 xdemitconf_t xecfg;
6034 mmfile_t mf1, mf2;
6035 struct diff_filepair *p = q->queue[i];
6036 int len1, len2;
6037
6038 memset(&xpp, 0, sizeof(xpp));
6039 memset(&xecfg, 0, sizeof(xecfg));
6040 if (p->status == 0)
6041 return error("internal diff status error");
6042 if (p->status == DIFF_STATUS_UNKNOWN)
6043 continue;
6044 if (diff_unmodified_pair(p))
6045 continue;
6046 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6047 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6048 continue;
6049 if (DIFF_PAIR_UNMERGED(p))
6050 continue;
6051
6052 diff_fill_oid_info(p->one, options->repo->index);
6053 diff_fill_oid_info(p->two, options->repo->index);
6054
6055 len1 = remove_space(p->one->path, strlen(p->one->path));
6056 len2 = remove_space(p->two->path, strlen(p->two->path));
6057 patch_id_add_string(&ctx, "diff--git");
6058 patch_id_add_string(&ctx, "a/");
6059 git_SHA1_Update(&ctx, p->one->path, len1);
6060 patch_id_add_string(&ctx, "b/");
6061 git_SHA1_Update(&ctx, p->two->path, len2);
6062
6063 if (p->one->mode == 0) {
6064 patch_id_add_string(&ctx, "newfilemode");
6065 patch_id_add_mode(&ctx, p->two->mode);
6066 patch_id_add_string(&ctx, "---/dev/null");
6067 patch_id_add_string(&ctx, "+++b/");
6068 git_SHA1_Update(&ctx, p->two->path, len2);
6069 } else if (p->two->mode == 0) {
6070 patch_id_add_string(&ctx, "deletedfilemode");
6071 patch_id_add_mode(&ctx, p->one->mode);
6072 patch_id_add_string(&ctx, "---a/");
6073 git_SHA1_Update(&ctx, p->one->path, len1);
6074 patch_id_add_string(&ctx, "+++/dev/null");
6075 } else {
6076 patch_id_add_string(&ctx, "---a/");
6077 git_SHA1_Update(&ctx, p->one->path, len1);
6078 patch_id_add_string(&ctx, "+++b/");
6079 git_SHA1_Update(&ctx, p->two->path, len2);
6080 }
6081
6082 if (diff_header_only)
6083 continue;
6084
6085 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6086 fill_mmfile(options->repo, &mf2, p->two) < 0)
6087 return error("unable to read files to diff");
6088
6089 if (diff_filespec_is_binary(options->repo, p->one) ||
6090 diff_filespec_is_binary(options->repo, p->two)) {
6091 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
6092 GIT_SHA1_HEXSZ);
6093 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
6094 GIT_SHA1_HEXSZ);
6095 continue;
6096 }
6097
6098 xpp.flags = 0;
6099 xecfg.ctxlen = 3;
6100 xecfg.flags = 0;
6101 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
6102 patch_id_consume, &data, &xpp, &xecfg))
6103 return error("unable to generate patch-id diff for %s",
6104 p->one->path);
6105 }
6106
6107 git_SHA1_Final(oid->hash, &ctx);
6108 return 0;
6109 }
6110
6111 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6112 {
6113 struct diff_queue_struct *q = &diff_queued_diff;
6114 int i;
6115 int result = diff_get_patch_id(options, oid, diff_header_only);
6116
6117 for (i = 0; i < q->nr; i++)
6118 diff_free_filepair(q->queue[i]);
6119
6120 free(q->queue);
6121 DIFF_QUEUE_CLEAR(q);
6122
6123 return result;
6124 }
6125
6126 static int is_summary_empty(const struct diff_queue_struct *q)
6127 {
6128 int i;
6129
6130 for (i = 0; i < q->nr; i++) {
6131 const struct diff_filepair *p = q->queue[i];
6132
6133 switch (p->status) {
6134 case DIFF_STATUS_DELETED:
6135 case DIFF_STATUS_ADDED:
6136 case DIFF_STATUS_COPIED:
6137 case DIFF_STATUS_RENAMED:
6138 return 0;
6139 default:
6140 if (p->score)
6141 return 0;
6142 if (p->one->mode && p->two->mode &&
6143 p->one->mode != p->two->mode)
6144 return 0;
6145 break;
6146 }
6147 }
6148 return 1;
6149 }
6150
6151 static const char rename_limit_warning[] =
6152 N_("inexact rename detection was skipped due to too many files.");
6153
6154 static const char degrade_cc_to_c_warning[] =
6155 N_("only found copies from modified paths due to too many files.");
6156
6157 static const char rename_limit_advice[] =
6158 N_("you may want to set your %s variable to at least "
6159 "%d and retry the command.");
6160
6161 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6162 {
6163 fflush(stdout);
6164 if (degraded_cc)
6165 warning(_(degrade_cc_to_c_warning));
6166 else if (needed)
6167 warning(_(rename_limit_warning));
6168 else
6169 return;
6170 if (0 < needed)
6171 warning(_(rename_limit_advice), varname, needed);
6172 }
6173
6174 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6175 {
6176 int i;
6177 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6178 struct diff_queue_struct *q = &diff_queued_diff;
6179
6180 if (WSEH_NEW & WS_RULE_MASK)
6181 BUG("WS rules bit mask overlaps with diff symbol flags");
6182
6183 if (o->color_moved)
6184 o->emitted_symbols = &esm;
6185
6186 for (i = 0; i < q->nr; i++) {
6187 struct diff_filepair *p = q->queue[i];
6188 if (check_pair_status(p))
6189 diff_flush_patch(p, o);
6190 }
6191
6192 if (o->emitted_symbols) {
6193 if (o->color_moved) {
6194 struct hashmap add_lines, del_lines;
6195
6196 if (o->color_moved_ws_handling &
6197 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
6198 o->color_moved_ws_handling |= XDF_IGNORE_WHITESPACE;
6199
6200 hashmap_init(&del_lines, moved_entry_cmp, o, 0);
6201 hashmap_init(&add_lines, moved_entry_cmp, o, 0);
6202
6203 add_lines_to_move_detection(o, &add_lines, &del_lines);
6204 mark_color_as_moved(o, &add_lines, &del_lines);
6205 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6206 dim_moved_lines(o);
6207
6208 hashmap_free(&add_lines, 1);
6209 hashmap_free(&del_lines, 1);
6210 }
6211
6212 for (i = 0; i < esm.nr; i++)
6213 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6214
6215 for (i = 0; i < esm.nr; i++)
6216 free((void *)esm.buf[i].line);
6217 esm.nr = 0;
6218
6219 o->emitted_symbols = NULL;
6220 }
6221 }
6222
6223 void diff_flush(struct diff_options *options)
6224 {
6225 struct diff_queue_struct *q = &diff_queued_diff;
6226 int i, output_format = options->output_format;
6227 int separator = 0;
6228 int dirstat_by_line = 0;
6229
6230 /*
6231 * Order: raw, stat, summary, patch
6232 * or: name/name-status/checkdiff (other bits clear)
6233 */
6234 if (!q->nr)
6235 goto free_queue;
6236
6237 if (output_format & (DIFF_FORMAT_RAW |
6238 DIFF_FORMAT_NAME |
6239 DIFF_FORMAT_NAME_STATUS |
6240 DIFF_FORMAT_CHECKDIFF)) {
6241 for (i = 0; i < q->nr; i++) {
6242 struct diff_filepair *p = q->queue[i];
6243 if (check_pair_status(p))
6244 flush_one_pair(p, options);
6245 }
6246 separator++;
6247 }
6248
6249 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6250 dirstat_by_line = 1;
6251
6252 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6253 dirstat_by_line) {
6254 struct diffstat_t diffstat;
6255
6256 memset(&diffstat, 0, sizeof(struct diffstat_t));
6257 for (i = 0; i < q->nr; i++) {
6258 struct diff_filepair *p = q->queue[i];
6259 if (check_pair_status(p))
6260 diff_flush_stat(p, options, &diffstat);
6261 }
6262 if (output_format & DIFF_FORMAT_NUMSTAT)
6263 show_numstat(&diffstat, options);
6264 if (output_format & DIFF_FORMAT_DIFFSTAT)
6265 show_stats(&diffstat, options);
6266 if (output_format & DIFF_FORMAT_SHORTSTAT)
6267 show_shortstats(&diffstat, options);
6268 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6269 show_dirstat_by_line(&diffstat, options);
6270 free_diffstat_info(&diffstat);
6271 separator++;
6272 }
6273 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6274 show_dirstat(options);
6275
6276 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6277 for (i = 0; i < q->nr; i++) {
6278 diff_summary(options, q->queue[i]);
6279 }
6280 separator++;
6281 }
6282
6283 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6284 options->flags.exit_with_status &&
6285 options->flags.diff_from_contents) {
6286 /*
6287 * run diff_flush_patch for the exit status. setting
6288 * options->file to /dev/null should be safe, because we
6289 * aren't supposed to produce any output anyway.
6290 */
6291 if (options->close_file)
6292 fclose(options->file);
6293 options->file = xfopen("/dev/null", "w");
6294 options->close_file = 1;
6295 options->color_moved = 0;
6296 for (i = 0; i < q->nr; i++) {
6297 struct diff_filepair *p = q->queue[i];
6298 if (check_pair_status(p))
6299 diff_flush_patch(p, options);
6300 if (options->found_changes)
6301 break;
6302 }
6303 }
6304
6305 if (output_format & DIFF_FORMAT_PATCH) {
6306 if (separator) {
6307 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6308 if (options->stat_sep)
6309 /* attach patch instead of inline */
6310 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6311 NULL, 0, 0);
6312 }
6313
6314 diff_flush_patch_all_file_pairs(options);
6315 }
6316
6317 if (output_format & DIFF_FORMAT_CALLBACK)
6318 options->format_callback(q, options, options->format_callback_data);
6319
6320 for (i = 0; i < q->nr; i++)
6321 diff_free_filepair(q->queue[i]);
6322 free_queue:
6323 free(q->queue);
6324 DIFF_QUEUE_CLEAR(q);
6325 if (options->close_file)
6326 fclose(options->file);
6327
6328 /*
6329 * Report the content-level differences with HAS_CHANGES;
6330 * diff_addremove/diff_change does not set the bit when
6331 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6332 */
6333 if (options->flags.diff_from_contents) {
6334 if (options->found_changes)
6335 options->flags.has_changes = 1;
6336 else
6337 options->flags.has_changes = 0;
6338 }
6339 }
6340
6341 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6342 {
6343 return (((p->status == DIFF_STATUS_MODIFIED) &&
6344 ((p->score &&
6345 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6346 (!p->score &&
6347 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6348 ((p->status != DIFF_STATUS_MODIFIED) &&
6349 filter_bit_tst(p->status, options)));
6350 }
6351
6352 static void diffcore_apply_filter(struct diff_options *options)
6353 {
6354 int i;
6355 struct diff_queue_struct *q = &diff_queued_diff;
6356 struct diff_queue_struct outq;
6357
6358 DIFF_QUEUE_CLEAR(&outq);
6359
6360 if (!options->filter)
6361 return;
6362
6363 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6364 int found;
6365 for (i = found = 0; !found && i < q->nr; i++) {
6366 if (match_filter(options, q->queue[i]))
6367 found++;
6368 }
6369 if (found)
6370 return;
6371
6372 /* otherwise we will clear the whole queue
6373 * by copying the empty outq at the end of this
6374 * function, but first clear the current entries
6375 * in the queue.
6376 */
6377 for (i = 0; i < q->nr; i++)
6378 diff_free_filepair(q->queue[i]);
6379 }
6380 else {
6381 /* Only the matching ones */
6382 for (i = 0; i < q->nr; i++) {
6383 struct diff_filepair *p = q->queue[i];
6384 if (match_filter(options, p))
6385 diff_q(&outq, p);
6386 else
6387 diff_free_filepair(p);
6388 }
6389 }
6390 free(q->queue);
6391 *q = outq;
6392 }
6393
6394 /* Check whether two filespecs with the same mode and size are identical */
6395 static int diff_filespec_is_identical(struct repository *r,
6396 struct diff_filespec *one,
6397 struct diff_filespec *two)
6398 {
6399 if (S_ISGITLINK(one->mode))
6400 return 0;
6401 if (diff_populate_filespec(r, one, 0))
6402 return 0;
6403 if (diff_populate_filespec(r, two, 0))
6404 return 0;
6405 return !memcmp(one->data, two->data, one->size);
6406 }
6407
6408 static int diff_filespec_check_stat_unmatch(struct repository *r,
6409 struct diff_filepair *p)
6410 {
6411 if (p->done_skip_stat_unmatch)
6412 return p->skip_stat_unmatch_result;
6413
6414 p->done_skip_stat_unmatch = 1;
6415 p->skip_stat_unmatch_result = 0;
6416 /*
6417 * 1. Entries that come from stat info dirtiness
6418 * always have both sides (iow, not create/delete),
6419 * one side of the object name is unknown, with
6420 * the same mode and size. Keep the ones that
6421 * do not match these criteria. They have real
6422 * differences.
6423 *
6424 * 2. At this point, the file is known to be modified,
6425 * with the same mode and size, and the object
6426 * name of one side is unknown. Need to inspect
6427 * the identical contents.
6428 */
6429 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6430 !DIFF_FILE_VALID(p->two) ||
6431 (p->one->oid_valid && p->two->oid_valid) ||
6432 (p->one->mode != p->two->mode) ||
6433 diff_populate_filespec(r, p->one, CHECK_SIZE_ONLY) ||
6434 diff_populate_filespec(r, p->two, CHECK_SIZE_ONLY) ||
6435 (p->one->size != p->two->size) ||
6436 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6437 p->skip_stat_unmatch_result = 1;
6438 return p->skip_stat_unmatch_result;
6439 }
6440
6441 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6442 {
6443 int i;
6444 struct diff_queue_struct *q = &diff_queued_diff;
6445 struct diff_queue_struct outq;
6446 DIFF_QUEUE_CLEAR(&outq);
6447
6448 for (i = 0; i < q->nr; i++) {
6449 struct diff_filepair *p = q->queue[i];
6450
6451 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6452 diff_q(&outq, p);
6453 else {
6454 /*
6455 * The caller can subtract 1 from skip_stat_unmatch
6456 * to determine how many paths were dirty only
6457 * due to stat info mismatch.
6458 */
6459 if (!diffopt->flags.no_index)
6460 diffopt->skip_stat_unmatch++;
6461 diff_free_filepair(p);
6462 }
6463 }
6464 free(q->queue);
6465 *q = outq;
6466 }
6467
6468 static int diffnamecmp(const void *a_, const void *b_)
6469 {
6470 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6471 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6472 const char *name_a, *name_b;
6473
6474 name_a = a->one ? a->one->path : a->two->path;
6475 name_b = b->one ? b->one->path : b->two->path;
6476 return strcmp(name_a, name_b);
6477 }
6478
6479 void diffcore_fix_diff_index(void)
6480 {
6481 struct diff_queue_struct *q = &diff_queued_diff;
6482 QSORT(q->queue, q->nr, diffnamecmp);
6483 }
6484
6485 static void add_if_missing(struct repository *r,
6486 struct oid_array *to_fetch,
6487 const struct diff_filespec *filespec)
6488 {
6489 if (filespec && filespec->oid_valid &&
6490 oid_object_info_extended(r, &filespec->oid, NULL,
6491 OBJECT_INFO_FOR_PREFETCH))
6492 oid_array_append(to_fetch, &filespec->oid);
6493 }
6494
6495 void diffcore_std(struct diff_options *options)
6496 {
6497 if (options->repo == the_repository &&
6498 repository_format_partial_clone) {
6499 /*
6500 * Prefetch the diff pairs that are about to be flushed.
6501 */
6502 int i;
6503 struct diff_queue_struct *q = &diff_queued_diff;
6504 struct oid_array to_fetch = OID_ARRAY_INIT;
6505
6506 for (i = 0; i < q->nr; i++) {
6507 struct diff_filepair *p = q->queue[i];
6508 add_if_missing(options->repo, &to_fetch, p->one);
6509 add_if_missing(options->repo, &to_fetch, p->two);
6510 }
6511 if (to_fetch.nr)
6512 /*
6513 * NEEDSWORK: Consider deduplicating the OIDs sent.
6514 */
6515 fetch_objects(repository_format_partial_clone,
6516 to_fetch.oid, to_fetch.nr);
6517 oid_array_clear(&to_fetch);
6518 }
6519
6520 /* NOTE please keep the following in sync with diff_tree_combined() */
6521 if (options->skip_stat_unmatch)
6522 diffcore_skip_stat_unmatch(options);
6523 if (!options->found_follow) {
6524 /* See try_to_follow_renames() in tree-diff.c */
6525 if (options->break_opt != -1)
6526 diffcore_break(options->repo,
6527 options->break_opt);
6528 if (options->detect_rename)
6529 diffcore_rename(options);
6530 if (options->break_opt != -1)
6531 diffcore_merge_broken();
6532 }
6533 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6534 diffcore_pickaxe(options);
6535 if (options->orderfile)
6536 diffcore_order(options->orderfile);
6537 if (!options->found_follow)
6538 /* See try_to_follow_renames() in tree-diff.c */
6539 diff_resolve_rename_copy();
6540 diffcore_apply_filter(options);
6541
6542 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6543 options->flags.has_changes = 1;
6544 else
6545 options->flags.has_changes = 0;
6546
6547 options->found_follow = 0;
6548 }
6549
6550 int diff_result_code(struct diff_options *opt, int status)
6551 {
6552 int result = 0;
6553
6554 diff_warn_rename_limit("diff.renameLimit",
6555 opt->needed_rename_limit,
6556 opt->degraded_cc_to_c);
6557 if (!opt->flags.exit_with_status &&
6558 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6559 return status;
6560 if (opt->flags.exit_with_status &&
6561 opt->flags.has_changes)
6562 result |= 01;
6563 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6564 opt->flags.check_failed)
6565 result |= 02;
6566 return result;
6567 }
6568
6569 int diff_can_quit_early(struct diff_options *opt)
6570 {
6571 return (opt->flags.quick &&
6572 !opt->filter &&
6573 opt->flags.has_changes);
6574 }
6575
6576 /*
6577 * Shall changes to this submodule be ignored?
6578 *
6579 * Submodule changes can be configured to be ignored separately for each path,
6580 * but that configuration can be overridden from the command line.
6581 */
6582 static int is_submodule_ignored(const char *path, struct diff_options *options)
6583 {
6584 int ignored = 0;
6585 struct diff_flags orig_flags = options->flags;
6586 if (!options->flags.override_submodule_config)
6587 set_diffopt_flags_from_submodule_config(options, path);
6588 if (options->flags.ignore_submodules)
6589 ignored = 1;
6590 options->flags = orig_flags;
6591 return ignored;
6592 }
6593
6594 void diff_addremove(struct diff_options *options,
6595 int addremove, unsigned mode,
6596 const struct object_id *oid,
6597 int oid_valid,
6598 const char *concatpath, unsigned dirty_submodule)
6599 {
6600 struct diff_filespec *one, *two;
6601
6602 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6603 return;
6604
6605 /* This may look odd, but it is a preparation for
6606 * feeding "there are unchanged files which should
6607 * not produce diffs, but when you are doing copy
6608 * detection you would need them, so here they are"
6609 * entries to the diff-core. They will be prefixed
6610 * with something like '=' or '*' (I haven't decided
6611 * which but should not make any difference).
6612 * Feeding the same new and old to diff_change()
6613 * also has the same effect.
6614 * Before the final output happens, they are pruned after
6615 * merged into rename/copy pairs as appropriate.
6616 */
6617 if (options->flags.reverse_diff)
6618 addremove = (addremove == '+' ? '-' :
6619 addremove == '-' ? '+' : addremove);
6620
6621 if (options->prefix &&
6622 strncmp(concatpath, options->prefix, options->prefix_length))
6623 return;
6624
6625 one = alloc_filespec(concatpath);
6626 two = alloc_filespec(concatpath);
6627
6628 if (addremove != '+')
6629 fill_filespec(one, oid, oid_valid, mode);
6630 if (addremove != '-') {
6631 fill_filespec(two, oid, oid_valid, mode);
6632 two->dirty_submodule = dirty_submodule;
6633 }
6634
6635 diff_queue(&diff_queued_diff, one, two);
6636 if (!options->flags.diff_from_contents)
6637 options->flags.has_changes = 1;
6638 }
6639
6640 void diff_change(struct diff_options *options,
6641 unsigned old_mode, unsigned new_mode,
6642 const struct object_id *old_oid,
6643 const struct object_id *new_oid,
6644 int old_oid_valid, int new_oid_valid,
6645 const char *concatpath,
6646 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6647 {
6648 struct diff_filespec *one, *two;
6649 struct diff_filepair *p;
6650
6651 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6652 is_submodule_ignored(concatpath, options))
6653 return;
6654
6655 if (options->flags.reverse_diff) {
6656 SWAP(old_mode, new_mode);
6657 SWAP(old_oid, new_oid);
6658 SWAP(old_oid_valid, new_oid_valid);
6659 SWAP(old_dirty_submodule, new_dirty_submodule);
6660 }
6661
6662 if (options->prefix &&
6663 strncmp(concatpath, options->prefix, options->prefix_length))
6664 return;
6665
6666 one = alloc_filespec(concatpath);
6667 two = alloc_filespec(concatpath);
6668 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6669 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6670 one->dirty_submodule = old_dirty_submodule;
6671 two->dirty_submodule = new_dirty_submodule;
6672 p = diff_queue(&diff_queued_diff, one, two);
6673
6674 if (options->flags.diff_from_contents)
6675 return;
6676
6677 if (options->flags.quick && options->skip_stat_unmatch &&
6678 !diff_filespec_check_stat_unmatch(options->repo, p))
6679 return;
6680
6681 options->flags.has_changes = 1;
6682 }
6683
6684 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6685 {
6686 struct diff_filepair *pair;
6687 struct diff_filespec *one, *two;
6688
6689 if (options->prefix &&
6690 strncmp(path, options->prefix, options->prefix_length))
6691 return NULL;
6692
6693 one = alloc_filespec(path);
6694 two = alloc_filespec(path);
6695 pair = diff_queue(&diff_queued_diff, one, two);
6696 pair->is_unmerged = 1;
6697 return pair;
6698 }
6699
6700 static char *run_textconv(struct repository *r,
6701 const char *pgm,
6702 struct diff_filespec *spec,
6703 size_t *outsize)
6704 {
6705 struct diff_tempfile *temp;
6706 const char *argv[3];
6707 const char **arg = argv;
6708 struct child_process child = CHILD_PROCESS_INIT;
6709 struct strbuf buf = STRBUF_INIT;
6710 int err = 0;
6711
6712 temp = prepare_temp_file(r, spec->path, spec);
6713 *arg++ = pgm;
6714 *arg++ = temp->name;
6715 *arg = NULL;
6716
6717 child.use_shell = 1;
6718 child.argv = argv;
6719 child.out = -1;
6720 if (start_command(&child)) {
6721 remove_tempfile();
6722 return NULL;
6723 }
6724
6725 if (strbuf_read(&buf, child.out, 0) < 0)
6726 err = error("error reading from textconv command '%s'", pgm);
6727 close(child.out);
6728
6729 if (finish_command(&child) || err) {
6730 strbuf_release(&buf);
6731 remove_tempfile();
6732 return NULL;
6733 }
6734 remove_tempfile();
6735
6736 return strbuf_detach(&buf, outsize);
6737 }
6738
6739 size_t fill_textconv(struct repository *r,
6740 struct userdiff_driver *driver,
6741 struct diff_filespec *df,
6742 char **outbuf)
6743 {
6744 size_t size;
6745
6746 if (!driver) {
6747 if (!DIFF_FILE_VALID(df)) {
6748 *outbuf = "";
6749 return 0;
6750 }
6751 if (diff_populate_filespec(r, df, 0))
6752 die("unable to read files to diff");
6753 *outbuf = df->data;
6754 return df->size;
6755 }
6756
6757 if (!driver->textconv)
6758 BUG("fill_textconv called with non-textconv driver");
6759
6760 if (driver->textconv_cache && df->oid_valid) {
6761 *outbuf = notes_cache_get(driver->textconv_cache,
6762 &df->oid,
6763 &size);
6764 if (*outbuf)
6765 return size;
6766 }
6767
6768 *outbuf = run_textconv(r, driver->textconv, df, &size);
6769 if (!*outbuf)
6770 die("unable to read files to diff");
6771
6772 if (driver->textconv_cache && df->oid_valid) {
6773 /* ignore errors, as we might be in a readonly repository */
6774 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6775 size);
6776 /*
6777 * we could save up changes and flush them all at the end,
6778 * but we would need an extra call after all diffing is done.
6779 * Since generating a cache entry is the slow path anyway,
6780 * this extra overhead probably isn't a big deal.
6781 */
6782 notes_cache_write(driver->textconv_cache);
6783 }
6784
6785 return size;
6786 }
6787
6788 int textconv_object(struct repository *r,
6789 const char *path,
6790 unsigned mode,
6791 const struct object_id *oid,
6792 int oid_valid,
6793 char **buf,
6794 unsigned long *buf_size)
6795 {
6796 struct diff_filespec *df;
6797 struct userdiff_driver *textconv;
6798
6799 df = alloc_filespec(path);
6800 fill_filespec(df, oid, oid_valid, mode);
6801 textconv = get_textconv(r, df);
6802 if (!textconv) {
6803 free_filespec(df);
6804 return 0;
6805 }
6806
6807 *buf_size = fill_textconv(r, textconv, df, buf);
6808 free_filespec(df);
6809 return 1;
6810 }
6811
6812 void setup_diff_pager(struct diff_options *opt)
6813 {
6814 /*
6815 * If the user asked for our exit code, then either they want --quiet
6816 * or --exit-code. We should definitely not bother with a pager in the
6817 * former case, as we will generate no output. Since we still properly
6818 * report our exit code even when a pager is run, we _could_ run a
6819 * pager with --exit-code. But since we have not done so historically,
6820 * and because it is easy to find people oneline advising "git diff
6821 * --exit-code" in hooks and other scripts, we do not do so.
6822 */
6823 if (!opt->flags.exit_with_status &&
6824 check_pager_config("diff") != 0)
6825 setup_pager();
6826 }
git for Debian