?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
treewide: reduce includes of cache.h in other headers
[git.git] / diff.c
blobfa86d023196b8a95f9f413f96b7bd2586b47add9
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "abspath.h"
6 #include "alloc.h"
7 #include "config.h"
8 #include "convert.h"
9 #include "environment.h"
10 #include "gettext.h"
11 #include "tempfile.h"
12 #include "quote.h"
13 #include "diff.h"
14 #include "diffcore.h"
15 #include "delta.h"
16 #include "hex.h"
17 #include "xdiff-interface.h"
18 #include "color.h"
19 #include "attr.h"
20 #include "run-command.h"
21 #include "utf8.h"
22 #include "object-store.h"
23 #include "userdiff.h"
24 #include "submodule-config.h"
25 #include "submodule.h"
26 #include "hashmap.h"
27 #include "mem-pool.h"
28 #include "ll-merge.h"
29 #include "string-list.h"
30 #include "strvec.h"
31 #include "graph.h"
32 #include "oid-array.h"
33 #include "packfile.h"
34 #include "pager.h"
35 #include "parse-options.h"
36 #include "help.h"
37 #include "promisor-remote.h"
38 #include "dir.h"
39 #include "object-file.h"
40 #include "object-name.h"
41 #include "setup.h"
42 #include "strmap.h"
43 #include "wrapper.h"
44
45 #ifdef NO_FAST_WORKING_DIRECTORY
46 #define FAST_WORKING_DIRECTORY 0
47 #else
48 #define FAST_WORKING_DIRECTORY 1
49 #endif
50
51 static int diff_detect_rename_default;
52 static int diff_indent_heuristic = 1;
53 static int diff_rename_limit_default = 1000;
54 static int diff_suppress_blank_empty;
55 static int diff_use_color_default = -1;
56 static int diff_color_moved_default;
57 static int diff_color_moved_ws_default;
58 static int diff_context_default = 3;
59 static int diff_interhunk_context_default;
60 static const char *diff_word_regex_cfg;
61 static const char *external_diff_cmd_cfg;
62 static const char *diff_order_file_cfg;
63 int diff_auto_refresh_index = 1;
64 static int diff_mnemonic_prefix;
65 static int diff_no_prefix;
66 static int diff_relative;
67 static int diff_stat_graph_width;
68 static int diff_dirstat_permille_default = 30;
69 static struct diff_options default_diff_options;
70 static long diff_algorithm;
71 static unsigned ws_error_highlight_default = WSEH_NEW;
72
73 static char diff_colors[][COLOR_MAXLEN] = {
74 GIT_COLOR_RESET,
75 GIT_COLOR_NORMAL, /* CONTEXT */
76 GIT_COLOR_BOLD, /* METAINFO */
77 GIT_COLOR_CYAN, /* FRAGINFO */
78 GIT_COLOR_RED, /* OLD */
79 GIT_COLOR_GREEN, /* NEW */
80 GIT_COLOR_YELLOW, /* COMMIT */
81 GIT_COLOR_BG_RED, /* WHITESPACE */
82 GIT_COLOR_NORMAL, /* FUNCINFO */
83 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
84 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
85 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
86 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
87 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
88 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
89 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
90 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
91 GIT_COLOR_FAINT, /* CONTEXT_DIM */
92 GIT_COLOR_FAINT_RED, /* OLD_DIM */
93 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
94 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
95 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
96 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
97 };
98
99 static const char *color_diff_slots[] = {
100 [DIFF_CONTEXT] = "context",
101 [DIFF_METAINFO] = "meta",
102 [DIFF_FRAGINFO] = "frag",
103 [DIFF_FILE_OLD] = "old",
104 [DIFF_FILE_NEW] = "new",
105 [DIFF_COMMIT] = "commit",
106 [DIFF_WHITESPACE] = "whitespace",
107 [DIFF_FUNCINFO] = "func",
108 [DIFF_FILE_OLD_MOVED] = "oldMoved",
109 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
110 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
111 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
112 [DIFF_FILE_NEW_MOVED] = "newMoved",
113 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
114 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
115 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
116 [DIFF_CONTEXT_DIM] = "contextDimmed",
117 [DIFF_FILE_OLD_DIM] = "oldDimmed",
118 [DIFF_FILE_NEW_DIM] = "newDimmed",
119 [DIFF_CONTEXT_BOLD] = "contextBold",
120 [DIFF_FILE_OLD_BOLD] = "oldBold",
121 [DIFF_FILE_NEW_BOLD] = "newBold",
122 };
123
124 define_list_config_array_extra(color_diff_slots, {"plain"});
125
126 static int parse_diff_color_slot(const char *var)
127 {
128 if (!strcasecmp(var, "plain"))
129 return DIFF_CONTEXT;
130 return LOOKUP_CONFIG(color_diff_slots, var);
131 }
132
133 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
134 struct strbuf *errmsg)
135 {
136 char *params_copy = xstrdup(params_string);
137 struct string_list params = STRING_LIST_INIT_NODUP;
138 int ret = 0;
139 int i;
140
141 if (*params_copy)
142 string_list_split_in_place(&params, params_copy, ',', -1);
143 for (i = 0; i < params.nr; i++) {
144 const char *p = params.items[i].string;
145 if (!strcmp(p, "changes")) {
146 options->flags.dirstat_by_line = 0;
147 options->flags.dirstat_by_file = 0;
148 } else if (!strcmp(p, "lines")) {
149 options->flags.dirstat_by_line = 1;
150 options->flags.dirstat_by_file = 0;
151 } else if (!strcmp(p, "files")) {
152 options->flags.dirstat_by_line = 0;
153 options->flags.dirstat_by_file = 1;
154 } else if (!strcmp(p, "noncumulative")) {
155 options->flags.dirstat_cumulative = 0;
156 } else if (!strcmp(p, "cumulative")) {
157 options->flags.dirstat_cumulative = 1;
158 } else if (isdigit(*p)) {
159 char *end;
160 int permille = strtoul(p, &end, 10) * 10;
161 if (*end == '.' && isdigit(*++end)) {
162 /* only use first digit */
163 permille += *end - '0';
164 /* .. and ignore any further digits */
165 while (isdigit(*++end))
166 ; /* nothing */
167 }
168 if (!*end)
169 options->dirstat_permille = permille;
170 else {
171 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
172 p);
173 ret++;
174 }
175 } else {
176 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
177 ret++;
178 }
179
180 }
181 string_list_clear(&params, 0);
182 free(params_copy);
183 return ret;
184 }
185
186 static int parse_submodule_params(struct diff_options *options, const char *value)
187 {
188 if (!strcmp(value, "log"))
189 options->submodule_format = DIFF_SUBMODULE_LOG;
190 else if (!strcmp(value, "short"))
191 options->submodule_format = DIFF_SUBMODULE_SHORT;
192 else if (!strcmp(value, "diff"))
193 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
194 /*
195 * Please update $__git_diff_submodule_formats in
196 * git-completion.bash when you add new formats.
197 */
198 else
199 return -1;
200 return 0;
201 }
202
203 int git_config_rename(const char *var, const char *value)
204 {
205 if (!value)
206 return DIFF_DETECT_RENAME;
207 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
208 return DIFF_DETECT_COPY;
209 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
210 }
211
212 long parse_algorithm_value(const char *value)
213 {
214 if (!value)
215 return -1;
216 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
217 return 0;
218 else if (!strcasecmp(value, "minimal"))
219 return XDF_NEED_MINIMAL;
220 else if (!strcasecmp(value, "patience"))
221 return XDF_PATIENCE_DIFF;
222 else if (!strcasecmp(value, "histogram"))
223 return XDF_HISTOGRAM_DIFF;
224 /*
225 * Please update $__git_diff_algorithms in git-completion.bash
226 * when you add new algorithms.
227 */
228 return -1;
229 }
230
231 static int parse_one_token(const char **arg, const char *token)
232 {
233 const char *rest;
234 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
235 *arg = rest;
236 return 1;
237 }
238 return 0;
239 }
240
241 static int parse_ws_error_highlight(const char *arg)
242 {
243 const char *orig_arg = arg;
244 unsigned val = 0;
245
246 while (*arg) {
247 if (parse_one_token(&arg, "none"))
248 val = 0;
249 else if (parse_one_token(&arg, "default"))
250 val = WSEH_NEW;
251 else if (parse_one_token(&arg, "all"))
252 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
253 else if (parse_one_token(&arg, "new"))
254 val |= WSEH_NEW;
255 else if (parse_one_token(&arg, "old"))
256 val |= WSEH_OLD;
257 else if (parse_one_token(&arg, "context"))
258 val |= WSEH_CONTEXT;
259 else {
260 return -1 - (int)(arg - orig_arg);
261 }
262 if (*arg)
263 arg++;
264 }
265 return val;
266 }
267
268 /*
269 * These are to give UI layer defaults.
270 * The core-level commands such as git-diff-files should
271 * never be affected by the setting of diff.renames
272 * the user happens to have in the configuration file.
273 */
274 void init_diff_ui_defaults(void)
275 {
276 diff_detect_rename_default = DIFF_DETECT_RENAME;
277 }
278
279 int git_diff_heuristic_config(const char *var, const char *value,
280 void *cb UNUSED)
281 {
282 if (!strcmp(var, "diff.indentheuristic"))
283 diff_indent_heuristic = git_config_bool(var, value);
284 return 0;
285 }
286
287 static int parse_color_moved(const char *arg)
288 {
289 switch (git_parse_maybe_bool(arg)) {
290 case 0:
291 return COLOR_MOVED_NO;
292 case 1:
293 return COLOR_MOVED_DEFAULT;
294 default:
295 break;
296 }
297
298 if (!strcmp(arg, "no"))
299 return COLOR_MOVED_NO;
300 else if (!strcmp(arg, "plain"))
301 return COLOR_MOVED_PLAIN;
302 else if (!strcmp(arg, "blocks"))
303 return COLOR_MOVED_BLOCKS;
304 else if (!strcmp(arg, "zebra"))
305 return COLOR_MOVED_ZEBRA;
306 else if (!strcmp(arg, "default"))
307 return COLOR_MOVED_DEFAULT;
308 else if (!strcmp(arg, "dimmed-zebra"))
309 return COLOR_MOVED_ZEBRA_DIM;
310 else if (!strcmp(arg, "dimmed_zebra"))
311 return COLOR_MOVED_ZEBRA_DIM;
312 else
313 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
314 }
315
316 static unsigned parse_color_moved_ws(const char *arg)
317 {
318 int ret = 0;
319 struct string_list l = STRING_LIST_INIT_DUP;
320 struct string_list_item *i;
321
322 string_list_split(&l, arg, ',', -1);
323
324 for_each_string_list_item(i, &l) {
325 struct strbuf sb = STRBUF_INIT;
326 strbuf_addstr(&sb, i->string);
327 strbuf_trim(&sb);
328
329 if (!strcmp(sb.buf, "no"))
330 ret = 0;
331 else if (!strcmp(sb.buf, "ignore-space-change"))
332 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
333 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
334 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
335 else if (!strcmp(sb.buf, "ignore-all-space"))
336 ret |= XDF_IGNORE_WHITESPACE;
337 else if (!strcmp(sb.buf, "allow-indentation-change"))
338 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
339 else {
340 ret |= COLOR_MOVED_WS_ERROR;
341 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);
342 }
343
344 strbuf_release(&sb);
345 }
346
347 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
348 (ret & XDF_WHITESPACE_FLAGS)) {
349 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
350 ret |= COLOR_MOVED_WS_ERROR;
351 }
352
353 string_list_clear(&l, 0);
354
355 return ret;
356 }
357
358 int git_diff_ui_config(const char *var, const char *value, void *cb)
359 {
360 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
361 diff_use_color_default = git_config_colorbool(var, value);
362 return 0;
363 }
364 if (!strcmp(var, "diff.colormoved")) {
365 int cm = parse_color_moved(value);
366 if (cm < 0)
367 return -1;
368 diff_color_moved_default = cm;
369 return 0;
370 }
371 if (!strcmp(var, "diff.colormovedws")) {
372 unsigned cm = parse_color_moved_ws(value);
373 if (cm & COLOR_MOVED_WS_ERROR)
374 return -1;
375 diff_color_moved_ws_default = cm;
376 return 0;
377 }
378 if (!strcmp(var, "diff.context")) {
379 diff_context_default = git_config_int(var, value);
380 if (diff_context_default < 0)
381 return -1;
382 return 0;
383 }
384 if (!strcmp(var, "diff.interhunkcontext")) {
385 diff_interhunk_context_default = git_config_int(var, value);
386 if (diff_interhunk_context_default < 0)
387 return -1;
388 return 0;
389 }
390 if (!strcmp(var, "diff.renames")) {
391 diff_detect_rename_default = git_config_rename(var, value);
392 return 0;
393 }
394 if (!strcmp(var, "diff.autorefreshindex")) {
395 diff_auto_refresh_index = git_config_bool(var, value);
396 return 0;
397 }
398 if (!strcmp(var, "diff.mnemonicprefix")) {
399 diff_mnemonic_prefix = git_config_bool(var, value);
400 return 0;
401 }
402 if (!strcmp(var, "diff.noprefix")) {
403 diff_no_prefix = git_config_bool(var, value);
404 return 0;
405 }
406 if (!strcmp(var, "diff.relative")) {
407 diff_relative = git_config_bool(var, value);
408 return 0;
409 }
410 if (!strcmp(var, "diff.statgraphwidth")) {
411 diff_stat_graph_width = git_config_int(var, value);
412 return 0;
413 }
414 if (!strcmp(var, "diff.external"))
415 return git_config_string(&external_diff_cmd_cfg, var, value);
416 if (!strcmp(var, "diff.wordregex"))
417 return git_config_string(&diff_word_regex_cfg, var, value);
418 if (!strcmp(var, "diff.orderfile"))
419 return git_config_pathname(&diff_order_file_cfg, var, value);
420
421 if (!strcmp(var, "diff.ignoresubmodules"))
422 handle_ignore_submodules_arg(&default_diff_options, value);
423
424 if (!strcmp(var, "diff.submodule")) {
425 if (parse_submodule_params(&default_diff_options, value))
426 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
427 value);
428 return 0;
429 }
430
431 if (!strcmp(var, "diff.algorithm")) {
432 diff_algorithm = parse_algorithm_value(value);
433 if (diff_algorithm < 0)
434 return -1;
435 return 0;
436 }
437
438 if (git_color_config(var, value, cb) < 0)
439 return -1;
440
441 return git_diff_basic_config(var, value, cb);
442 }
443
444 int git_diff_basic_config(const char *var, const char *value, void *cb)
445 {
446 const char *name;
447
448 if (!strcmp(var, "diff.renamelimit")) {
449 diff_rename_limit_default = git_config_int(var, value);
450 return 0;
451 }
452
453 if (userdiff_config(var, value) < 0)
454 return -1;
455
456 if (skip_prefix(var, "diff.color.", &name) ||
457 skip_prefix(var, "color.diff.", &name)) {
458 int slot = parse_diff_color_slot(name);
459 if (slot < 0)
460 return 0;
461 if (!value)
462 return config_error_nonbool(var);
463 return color_parse(value, diff_colors[slot]);
464 }
465
466 if (!strcmp(var, "diff.wserrorhighlight")) {
467 int val = parse_ws_error_highlight(value);
468 if (val < 0)
469 return -1;
470 ws_error_highlight_default = val;
471 return 0;
472 }
473
474 /* like GNU diff's --suppress-blank-empty option */
475 if (!strcmp(var, "diff.suppressblankempty") ||
476 /* for backwards compatibility */
477 !strcmp(var, "diff.suppress-blank-empty")) {
478 diff_suppress_blank_empty = git_config_bool(var, value);
479 return 0;
480 }
481
482 if (!strcmp(var, "diff.dirstat")) {
483 struct strbuf errmsg = STRBUF_INIT;
484 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
485 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
486 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
487 errmsg.buf);
488 strbuf_release(&errmsg);
489 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
490 return 0;
491 }
492
493 if (git_diff_heuristic_config(var, value, cb) < 0)
494 return -1;
495
496 return git_default_config(var, value, cb);
497 }
498
499 static char *quote_two(const char *one, const char *two)
500 {
501 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
502 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
503 struct strbuf res = STRBUF_INIT;
504
505 if (need_one + need_two) {
506 strbuf_addch(&res, '"');
507 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
508 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
509 strbuf_addch(&res, '"');
510 } else {
511 strbuf_addstr(&res, one);
512 strbuf_addstr(&res, two);
513 }
514 return strbuf_detach(&res, NULL);
515 }
516
517 static const char *external_diff(void)
518 {
519 static const char *external_diff_cmd = NULL;
520 static int done_preparing = 0;
521
522 if (done_preparing)
523 return external_diff_cmd;
524 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
525 if (!external_diff_cmd)
526 external_diff_cmd = external_diff_cmd_cfg;
527 done_preparing = 1;
528 return external_diff_cmd;
529 }
530
531 /*
532 * Keep track of files used for diffing. Sometimes such an entry
533 * refers to a temporary file, sometimes to an existing file, and
534 * sometimes to "/dev/null".
535 */
536 static struct diff_tempfile {
537 /*
538 * filename external diff should read from, or NULL if this
539 * entry is currently not in use:
540 */
541 const char *name;
542
543 char hex[GIT_MAX_HEXSZ + 1];
544 char mode[10];
545
546 /*
547 * If this diff_tempfile instance refers to a temporary file,
548 * this tempfile object is used to manage its lifetime.
549 */
550 struct tempfile *tempfile;
551 } diff_temp[2];
552
553 struct emit_callback {
554 int color_diff;
555 unsigned ws_rule;
556 int blank_at_eof_in_preimage;
557 int blank_at_eof_in_postimage;
558 int lno_in_preimage;
559 int lno_in_postimage;
560 const char **label_path;
561 struct diff_words_data *diff_words;
562 struct diff_options *opt;
563 struct strbuf *header;
564 };
565
566 static int count_lines(const char *data, int size)
567 {
568 int count, ch, completely_empty = 1, nl_just_seen = 0;
569 count = 0;
570 while (0 < size--) {
571 ch = *data++;
572 if (ch == '\n') {
573 count++;
574 nl_just_seen = 1;
575 completely_empty = 0;
576 }
577 else {
578 nl_just_seen = 0;
579 completely_empty = 0;
580 }
581 }
582 if (completely_empty)
583 return 0;
584 if (!nl_just_seen)
585 count++; /* no trailing newline */
586 return count;
587 }
588
589 static int fill_mmfile(struct repository *r, mmfile_t *mf,
590 struct diff_filespec *one)
591 {
592 if (!DIFF_FILE_VALID(one)) {
593 mf->ptr = (char *)""; /* does not matter */
594 mf->size = 0;
595 return 0;
596 }
597 else if (diff_populate_filespec(r, one, NULL))
598 return -1;
599
600 mf->ptr = one->data;
601 mf->size = one->size;
602 return 0;
603 }
604
605 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
606 static unsigned long diff_filespec_size(struct repository *r,
607 struct diff_filespec *one)
608 {
609 struct diff_populate_filespec_options dpf_options = {
610 .check_size_only = 1,
611 };
612
613 if (!DIFF_FILE_VALID(one))
614 return 0;
615 diff_populate_filespec(r, one, &dpf_options);
616 return one->size;
617 }
618
619 static int count_trailing_blank(mmfile_t *mf)
620 {
621 char *ptr = mf->ptr;
622 long size = mf->size;
623 int cnt = 0;
624
625 if (!size)
626 return cnt;
627 ptr += size - 1; /* pointing at the very end */
628 if (*ptr != '\n')
629 ; /* incomplete line */
630 else
631 ptr--; /* skip the last LF */
632 while (mf->ptr < ptr) {
633 char *prev_eol;
634 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
635 if (*prev_eol == '\n')
636 break;
637 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
638 break;
639 cnt++;
640 ptr = prev_eol - 1;
641 }
642 return cnt;
643 }
644
645 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
646 struct emit_callback *ecbdata)
647 {
648 int l1, l2, at;
649 l1 = count_trailing_blank(mf1);
650 l2 = count_trailing_blank(mf2);
651 if (l2 <= l1) {
652 ecbdata->blank_at_eof_in_preimage = 0;
653 ecbdata->blank_at_eof_in_postimage = 0;
654 return;
655 }
656 at = count_lines(mf1->ptr, mf1->size);
657 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
658
659 at = count_lines(mf2->ptr, mf2->size);
660 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
661 }
662
663 static void emit_line_0(struct diff_options *o,
664 const char *set_sign, const char *set, unsigned reverse, const char *reset,
665 int first, const char *line, int len)
666 {
667 int has_trailing_newline, has_trailing_carriage_return;
668 int needs_reset = 0; /* at the end of the line */
669 FILE *file = o->file;
670
671 fputs(diff_line_prefix(o), file);
672
673 has_trailing_newline = (len > 0 && line[len-1] == '\n');
674 if (has_trailing_newline)
675 len--;
676
677 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
678 if (has_trailing_carriage_return)
679 len--;
680
681 if (!len && !first)
682 goto end_of_line;
683
684 if (reverse && want_color(o->use_color)) {
685 fputs(GIT_COLOR_REVERSE, file);
686 needs_reset = 1;
687 }
688
689 if (set_sign) {
690 fputs(set_sign, file);
691 needs_reset = 1;
692 }
693
694 if (first)
695 fputc(first, file);
696
697 if (!len)
698 goto end_of_line;
699
700 if (set) {
701 if (set_sign && set != set_sign)
702 fputs(reset, file);
703 fputs(set, file);
704 needs_reset = 1;
705 }
706 fwrite(line, len, 1, file);
707 needs_reset = 1; /* 'line' may contain color codes. */
708
709 end_of_line:
710 if (needs_reset)
711 fputs(reset, file);
712 if (has_trailing_carriage_return)
713 fputc('\r', file);
714 if (has_trailing_newline)
715 fputc('\n', file);
716 }
717
718 static void emit_line(struct diff_options *o, const char *set, const char *reset,
719 const char *line, int len)
720 {
721 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
722 }
723
724 enum diff_symbol {
725 DIFF_SYMBOL_BINARY_DIFF_HEADER,
726 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
727 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
728 DIFF_SYMBOL_BINARY_DIFF_BODY,
729 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
730 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
731 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
732 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
733 DIFF_SYMBOL_STATS_LINE,
734 DIFF_SYMBOL_WORD_DIFF,
735 DIFF_SYMBOL_STAT_SEP,
736 DIFF_SYMBOL_SUMMARY,
737 DIFF_SYMBOL_SUBMODULE_ADD,
738 DIFF_SYMBOL_SUBMODULE_DEL,
739 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
740 DIFF_SYMBOL_SUBMODULE_MODIFIED,
741 DIFF_SYMBOL_SUBMODULE_HEADER,
742 DIFF_SYMBOL_SUBMODULE_ERROR,
743 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
744 DIFF_SYMBOL_REWRITE_DIFF,
745 DIFF_SYMBOL_BINARY_FILES,
746 DIFF_SYMBOL_HEADER,
747 DIFF_SYMBOL_FILEPAIR_PLUS,
748 DIFF_SYMBOL_FILEPAIR_MINUS,
749 DIFF_SYMBOL_WORDS_PORCELAIN,
750 DIFF_SYMBOL_WORDS,
751 DIFF_SYMBOL_CONTEXT,
752 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
753 DIFF_SYMBOL_PLUS,
754 DIFF_SYMBOL_MINUS,
755 DIFF_SYMBOL_NO_LF_EOF,
756 DIFF_SYMBOL_CONTEXT_FRAGINFO,
757 DIFF_SYMBOL_CONTEXT_MARKER,
758 DIFF_SYMBOL_SEPARATOR
759 };
760 /*
761 * Flags for content lines:
762 * 0..12 are whitespace rules
763 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
764 * 16 is marking if the line is blank at EOF
765 */
766 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
767 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
768 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
769 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
770 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
771
772 /*
773 * This struct is used when we need to buffer the output of the diff output.
774 *
775 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
776 * into the pre/post image file. This pointer could be a union with the
777 * line pointer. By storing an offset into the file instead of the literal line,
778 * we can decrease the memory footprint for the buffered output. At first we
779 * may want to only have indirection for the content lines, but we could also
780 * enhance the state for emitting prefabricated lines, e.g. the similarity
781 * score line or hunk/file headers would only need to store a number or path
782 * and then the output can be constructed later on depending on state.
783 */
784 struct emitted_diff_symbol {
785 const char *line;
786 int len;
787 int flags;
788 int indent_off; /* Offset to first non-whitespace character */
789 int indent_width; /* The visual width of the indentation */
790 unsigned id;
791 enum diff_symbol s;
792 };
793 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
794
795 struct emitted_diff_symbols {
796 struct emitted_diff_symbol *buf;
797 int nr, alloc;
798 };
799 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
800
801 static void append_emitted_diff_symbol(struct diff_options *o,
802 struct emitted_diff_symbol *e)
803 {
804 struct emitted_diff_symbol *f;
805
806 ALLOC_GROW(o->emitted_symbols->buf,
807 o->emitted_symbols->nr + 1,
808 o->emitted_symbols->alloc);
809 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
810
811 memcpy(f, e, sizeof(struct emitted_diff_symbol));
812 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
813 }
814
815 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
816 {
817 if (!e)
818 return;
819 free(e->buf);
820 free(e);
821 }
822
823 struct moved_entry {
824 const struct emitted_diff_symbol *es;
825 struct moved_entry *next_line;
826 struct moved_entry *next_match;
827 };
828
829 struct moved_block {
830 struct moved_entry *match;
831 int wsd; /* The whitespace delta of this block */
832 };
833
834 #define INDENT_BLANKLINE INT_MIN
835
836 static void fill_es_indent_data(struct emitted_diff_symbol *es)
837 {
838 unsigned int off = 0, i;
839 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
840 const char *s = es->line;
841 const int len = es->len;
842
843 /* skip any \v \f \r at start of indentation */
844 while (s[off] == '\f' || s[off] == '\v' ||
845 (s[off] == '\r' && off < len - 1))
846 off++;
847
848 /* calculate the visual width of indentation */
849 while(1) {
850 if (s[off] == ' ') {
851 width++;
852 off++;
853 } else if (s[off] == '\t') {
854 width += tab_width - (width % tab_width);
855 while (s[++off] == '\t')
856 width += tab_width;
857 } else {
858 break;
859 }
860 }
861
862 /* check if this line is blank */
863 for (i = off; i < len; i++)
864 if (!isspace(s[i]))
865 break;
866
867 if (i == len) {
868 es->indent_width = INDENT_BLANKLINE;
869 es->indent_off = len;
870 } else {
871 es->indent_off = off;
872 es->indent_width = width;
873 }
874 }
875
876 static int compute_ws_delta(const struct emitted_diff_symbol *a,
877 const struct emitted_diff_symbol *b)
878 {
879 int a_width = a->indent_width,
880 b_width = b->indent_width;
881
882 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
883 return INDENT_BLANKLINE;
884
885 return a_width - b_width;
886 }
887
888 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
889 const struct emitted_diff_symbol *l,
890 struct moved_block *pmb)
891 {
892 int a_width = cur->es->indent_width, b_width = l->indent_width;
893 int delta;
894
895 /* The text of each line must match */
896 if (cur->es->id != l->id)
897 return 1;
898
899 /*
900 * If 'l' and 'cur' are both blank then we don't need to check the
901 * indent. We only need to check cur as we know the strings match.
902 * */
903 if (a_width == INDENT_BLANKLINE)
904 return 0;
905
906 /*
907 * The indent changes of the block are known and stored in pmb->wsd;
908 * however we need to check if the indent changes of the current line
909 * match those of the current block.
910 */
911 delta = b_width - a_width;
912
913 /*
914 * If the previous lines of this block were all blank then set its
915 * whitespace delta.
916 */
917 if (pmb->wsd == INDENT_BLANKLINE)
918 pmb->wsd = delta;
919
920 return delta != pmb->wsd;
921 }
922
923 struct interned_diff_symbol {
924 struct hashmap_entry ent;
925 struct emitted_diff_symbol *es;
926 };
927
928 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
929 const struct hashmap_entry *eptr,
930 const struct hashmap_entry *entry_or_key,
931 const void *keydata UNUSED)
932 {
933 const struct diff_options *diffopt = hashmap_cmp_fn_data;
934 const struct emitted_diff_symbol *a, *b;
935 unsigned flags = diffopt->color_moved_ws_handling
936 & XDF_WHITESPACE_FLAGS;
937
938 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
939 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
940
941 return !xdiff_compare_lines(a->line + a->indent_off,
942 a->len - a->indent_off,
943 b->line + b->indent_off,
944 b->len - b->indent_off, flags);
945 }
946
947 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
948 struct interned_diff_symbol *s)
949 {
950 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
951 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
952 l->len - l->indent_off, flags);
953
954 hashmap_entry_init(&s->ent, hash);
955 s->es = l;
956 }
957
958 struct moved_entry_list {
959 struct moved_entry *add, *del;
960 };
961
962 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
963 struct mem_pool *entry_mem_pool)
964 {
965 struct moved_entry *prev_line = NULL;
966 struct mem_pool interned_pool;
967 struct hashmap interned_map;
968 struct moved_entry_list *entry_list = NULL;
969 size_t entry_list_alloc = 0;
970 unsigned id = 0;
971 int n;
972
973 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
974 mem_pool_init(&interned_pool, 1024 * 1024);
975
976 for (n = 0; n < o->emitted_symbols->nr; n++) {
977 struct interned_diff_symbol key;
978 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
979 struct interned_diff_symbol *s;
980 struct moved_entry *entry;
981
982 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
983 prev_line = NULL;
984 continue;
985 }
986
987 if (o->color_moved_ws_handling &
988 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
989 fill_es_indent_data(l);
990
991 prepare_entry(o, l, &key);
992 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
993 if (s) {
994 l->id = s->es->id;
995 } else {
996 l->id = id;
997 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
998 hashmap_add(&interned_map,
999 memcpy(mem_pool_alloc(&interned_pool,
1000 sizeof(key)),
1001 &key, sizeof(key)));
1002 }
1003 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1004 entry->es = l;
1005 entry->next_line = NULL;
1006 if (prev_line && prev_line->es->s == l->s)
1007 prev_line->next_line = entry;
1008 prev_line = entry;
1009 if (l->s == DIFF_SYMBOL_PLUS) {
1010 entry->next_match = entry_list[l->id].add;
1011 entry_list[l->id].add = entry;
1012 } else {
1013 entry->next_match = entry_list[l->id].del;
1014 entry_list[l->id].del = entry;
1015 }
1016 }
1017
1018 hashmap_clear(&interned_map);
1019 mem_pool_discard(&interned_pool, 0);
1020
1021 return entry_list;
1022 }
1023
1024 static void pmb_advance_or_null(struct diff_options *o,
1025 struct emitted_diff_symbol *l,
1026 struct moved_block *pmb,
1027 int *pmb_nr)
1028 {
1029 int i, j;
1030
1031 for (i = 0, j = 0; i < *pmb_nr; i++) {
1032 int match;
1033 struct moved_entry *prev = pmb[i].match;
1034 struct moved_entry *cur = (prev && prev->next_line) ?
1035 prev->next_line : NULL;
1036
1037 if (o->color_moved_ws_handling &
1038 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1039 match = cur &&
1040 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1041 else
1042 match = cur && cur->es->id == l->id;
1043
1044 if (match) {
1045 pmb[j] = pmb[i];
1046 pmb[j++].match = cur;
1047 }
1048 }
1049 *pmb_nr = j;
1050 }
1051
1052 static void fill_potential_moved_blocks(struct diff_options *o,
1053 struct moved_entry *match,
1054 struct emitted_diff_symbol *l,
1055 struct moved_block **pmb_p,
1056 int *pmb_alloc_p, int *pmb_nr_p)
1057
1058 {
1059 struct moved_block *pmb = *pmb_p;
1060 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1061
1062 /*
1063 * The current line is the start of a new block.
1064 * Setup the set of potential blocks.
1065 */
1066 for (; match; match = match->next_match) {
1067 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1068 if (o->color_moved_ws_handling &
1069 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1070 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1071 else
1072 pmb[pmb_nr].wsd = 0;
1073 pmb[pmb_nr++].match = match;
1074 }
1075
1076 *pmb_p = pmb;
1077 *pmb_alloc_p = pmb_alloc;
1078 *pmb_nr_p = pmb_nr;
1079 }
1080
1081 /*
1082 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1083 *
1084 * Otherwise, if the last block has fewer alphanumeric characters than
1085 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1086 * that block.
1087 *
1088 * The last block consists of the (n - block_length)'th line up to but not
1089 * including the nth line.
1090 *
1091 * Returns 0 if the last block is empty or is unset by this function, non zero
1092 * otherwise.
1093 *
1094 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1095 * Think of a way to unify them.
1096 */
1097 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1098 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1099 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1100 {
1101 int i, alnum_count = 0;
1102 if (o->color_moved == COLOR_MOVED_PLAIN)
1103 return block_length;
1104 for (i = 1; i < block_length + 1; i++) {
1105 const char *c = o->emitted_symbols->buf[n - i].line;
1106 for (; *c; c++) {
1107 if (!isalnum(*c))
1108 continue;
1109 alnum_count++;
1110 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1111 return 1;
1112 }
1113 }
1114 for (i = 1; i < block_length + 1; i++)
1115 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1116 return 0;
1117 }
1118
1119 /* Find blocks of moved code, delegate actual coloring decision to helper */
1120 static void mark_color_as_moved(struct diff_options *o,
1121 struct moved_entry_list *entry_list)
1122 {
1123 struct moved_block *pmb = NULL; /* potentially moved blocks */
1124 int pmb_nr = 0, pmb_alloc = 0;
1125 int n, flipped_block = 0, block_length = 0;
1126 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1127
1128
1129 for (n = 0; n < o->emitted_symbols->nr; n++) {
1130 struct moved_entry *match = NULL;
1131 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1132
1133 switch (l->s) {
1134 case DIFF_SYMBOL_PLUS:
1135 match = entry_list[l->id].del;
1136 break;
1137 case DIFF_SYMBOL_MINUS:
1138 match = entry_list[l->id].add;
1139 break;
1140 default:
1141 flipped_block = 0;
1142 }
1143
1144 if (pmb_nr && (!match || l->s != moved_symbol)) {
1145 if (!adjust_last_block(o, n, block_length) &&
1146 block_length > 1) {
1147 /*
1148 * Rewind in case there is another match
1149 * starting at the second line of the block
1150 */
1151 match = NULL;
1152 n -= block_length;
1153 }
1154 pmb_nr = 0;
1155 block_length = 0;
1156 flipped_block = 0;
1157 }
1158 if (!match) {
1159 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1160 continue;
1161 }
1162
1163 if (o->color_moved == COLOR_MOVED_PLAIN) {
1164 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1165 continue;
1166 }
1167
1168 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1169
1170 if (pmb_nr == 0) {
1171 int contiguous = adjust_last_block(o, n, block_length);
1172
1173 if (!contiguous && block_length > 1)
1174 /*
1175 * Rewind in case there is another match
1176 * starting at the second line of the block
1177 */
1178 n -= block_length;
1179 else
1180 fill_potential_moved_blocks(o, match, l,
1181 &pmb, &pmb_alloc,
1182 &pmb_nr);
1183
1184 if (contiguous && pmb_nr && moved_symbol == l->s)
1185 flipped_block = (flipped_block + 1) % 2;
1186 else
1187 flipped_block = 0;
1188
1189 if (pmb_nr)
1190 moved_symbol = l->s;
1191 else
1192 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1193
1194 block_length = 0;
1195 }
1196
1197 if (pmb_nr) {
1198 block_length++;
1199 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1200 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1201 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1202 }
1203 }
1204 adjust_last_block(o, n, block_length);
1205
1206 free(pmb);
1207 }
1208
1209 static void dim_moved_lines(struct diff_options *o)
1210 {
1211 int n;
1212 for (n = 0; n < o->emitted_symbols->nr; n++) {
1213 struct emitted_diff_symbol *prev = (n != 0) ?
1214 &o->emitted_symbols->buf[n - 1] : NULL;
1215 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1216 struct emitted_diff_symbol *next =
1217 (n < o->emitted_symbols->nr - 1) ?
1218 &o->emitted_symbols->buf[n + 1] : NULL;
1219
1220 /* Not a plus or minus line? */
1221 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1222 continue;
1223
1224 /* Not a moved line? */
1225 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1226 continue;
1227
1228 /*
1229 * If prev or next are not a plus or minus line,
1230 * pretend they don't exist
1231 */
1232 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1233 prev->s != DIFF_SYMBOL_MINUS)
1234 prev = NULL;
1235 if (next && next->s != DIFF_SYMBOL_PLUS &&
1236 next->s != DIFF_SYMBOL_MINUS)
1237 next = NULL;
1238
1239 /* Inside a block? */
1240 if ((prev &&
1241 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1242 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1243 (next &&
1244 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1245 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1246 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1247 continue;
1248 }
1249
1250 /* Check if we are at an interesting bound: */
1251 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1252 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1253 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1254 continue;
1255 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1256 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1257 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1258 continue;
1259
1260 /*
1261 * The boundary to prev and next are not interesting,
1262 * so this line is not interesting as a whole
1263 */
1264 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1265 }
1266 }
1267
1268 static void emit_line_ws_markup(struct diff_options *o,
1269 const char *set_sign, const char *set,
1270 const char *reset,
1271 int sign_index, const char *line, int len,
1272 unsigned ws_rule, int blank_at_eof)
1273 {
1274 const char *ws = NULL;
1275 int sign = o->output_indicators[sign_index];
1276
1277 if (o->ws_error_highlight & ws_rule) {
1278 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1279 if (!*ws)
1280 ws = NULL;
1281 }
1282
1283 if (!ws && !set_sign)
1284 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1285 else if (!ws) {
1286 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1287 } else if (blank_at_eof)
1288 /* Blank line at EOF - paint '+' as well */
1289 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1290 else {
1291 /* Emit just the prefix, then the rest. */
1292 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1293 sign, "", 0);
1294 ws_check_emit(line, len, ws_rule,
1295 o->file, set, reset, ws);
1296 }
1297 }
1298
1299 static void emit_diff_symbol_from_struct(struct diff_options *o,
1300 struct emitted_diff_symbol *eds)
1301 {
1302 static const char *nneof = " No newline at end of file\n";
1303 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1304
1305 enum diff_symbol s = eds->s;
1306 const char *line = eds->line;
1307 int len = eds->len;
1308 unsigned flags = eds->flags;
1309
1310 switch (s) {
1311 case DIFF_SYMBOL_NO_LF_EOF:
1312 context = diff_get_color_opt(o, DIFF_CONTEXT);
1313 reset = diff_get_color_opt(o, DIFF_RESET);
1314 putc('\n', o->file);
1315 emit_line_0(o, context, NULL, 0, reset, '\\',
1316 nneof, strlen(nneof));
1317 break;
1318 case DIFF_SYMBOL_SUBMODULE_HEADER:
1319 case DIFF_SYMBOL_SUBMODULE_ERROR:
1320 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1321 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1322 case DIFF_SYMBOL_SUMMARY:
1323 case DIFF_SYMBOL_STATS_LINE:
1324 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1325 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1326 emit_line(o, "", "", line, len);
1327 break;
1328 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1329 case DIFF_SYMBOL_CONTEXT_MARKER:
1330 context = diff_get_color_opt(o, DIFF_CONTEXT);
1331 reset = diff_get_color_opt(o, DIFF_RESET);
1332 emit_line(o, context, reset, line, len);
1333 break;
1334 case DIFF_SYMBOL_SEPARATOR:
1335 fprintf(o->file, "%s%c",
1336 diff_line_prefix(o),
1337 o->line_termination);
1338 break;
1339 case DIFF_SYMBOL_CONTEXT:
1340 set = diff_get_color_opt(o, DIFF_CONTEXT);
1341 reset = diff_get_color_opt(o, DIFF_RESET);
1342 set_sign = NULL;
1343 if (o->flags.dual_color_diffed_diffs) {
1344 char c = !len ? 0 : line[0];
1345
1346 if (c == '+')
1347 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1348 else if (c == '@')
1349 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1350 else if (c == '-')
1351 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1352 }
1353 emit_line_ws_markup(o, set_sign, set, reset,
1354 OUTPUT_INDICATOR_CONTEXT, line, len,
1355 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1356 break;
1357 case DIFF_SYMBOL_PLUS:
1358 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1359 DIFF_SYMBOL_MOVED_LINE_ALT |
1360 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1361 case DIFF_SYMBOL_MOVED_LINE |
1362 DIFF_SYMBOL_MOVED_LINE_ALT |
1363 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1364 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1365 break;
1366 case DIFF_SYMBOL_MOVED_LINE |
1367 DIFF_SYMBOL_MOVED_LINE_ALT:
1368 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1369 break;
1370 case DIFF_SYMBOL_MOVED_LINE |
1371 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1372 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1373 break;
1374 case DIFF_SYMBOL_MOVED_LINE:
1375 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1376 break;
1377 default:
1378 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1379 }
1380 reset = diff_get_color_opt(o, DIFF_RESET);
1381 if (!o->flags.dual_color_diffed_diffs)
1382 set_sign = NULL;
1383 else {
1384 char c = !len ? 0 : line[0];
1385
1386 set_sign = set;
1387 if (c == '-')
1388 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1389 else if (c == '@')
1390 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1391 else if (c == '+')
1392 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1393 else
1394 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1395 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1396 }
1397 emit_line_ws_markup(o, set_sign, set, reset,
1398 OUTPUT_INDICATOR_NEW, line, len,
1399 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1400 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1401 break;
1402 case DIFF_SYMBOL_MINUS:
1403 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1404 DIFF_SYMBOL_MOVED_LINE_ALT |
1405 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1406 case DIFF_SYMBOL_MOVED_LINE |
1407 DIFF_SYMBOL_MOVED_LINE_ALT |
1408 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1409 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1410 break;
1411 case DIFF_SYMBOL_MOVED_LINE |
1412 DIFF_SYMBOL_MOVED_LINE_ALT:
1413 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1414 break;
1415 case DIFF_SYMBOL_MOVED_LINE |
1416 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1417 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1418 break;
1419 case DIFF_SYMBOL_MOVED_LINE:
1420 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1421 break;
1422 default:
1423 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1424 }
1425 reset = diff_get_color_opt(o, DIFF_RESET);
1426 if (!o->flags.dual_color_diffed_diffs)
1427 set_sign = NULL;
1428 else {
1429 char c = !len ? 0 : line[0];
1430
1431 set_sign = set;
1432 if (c == '+')
1433 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1434 else if (c == '@')
1435 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1436 else if (c == '-')
1437 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1438 else
1439 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1440 }
1441 emit_line_ws_markup(o, set_sign, set, reset,
1442 OUTPUT_INDICATOR_OLD, line, len,
1443 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1444 break;
1445 case DIFF_SYMBOL_WORDS_PORCELAIN:
1446 context = diff_get_color_opt(o, DIFF_CONTEXT);
1447 reset = diff_get_color_opt(o, DIFF_RESET);
1448 emit_line(o, context, reset, line, len);
1449 fputs("~\n", o->file);
1450 break;
1451 case DIFF_SYMBOL_WORDS:
1452 context = diff_get_color_opt(o, DIFF_CONTEXT);
1453 reset = diff_get_color_opt(o, DIFF_RESET);
1454 /*
1455 * Skip the prefix character, if any. With
1456 * diff_suppress_blank_empty, there may be
1457 * none.
1458 */
1459 if (line[0] != '\n') {
1460 line++;
1461 len--;
1462 }
1463 emit_line(o, context, reset, line, len);
1464 break;
1465 case DIFF_SYMBOL_FILEPAIR_PLUS:
1466 meta = diff_get_color_opt(o, DIFF_METAINFO);
1467 reset = diff_get_color_opt(o, DIFF_RESET);
1468 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1469 line, reset,
1470 strchr(line, ' ') ? "\t" : "");
1471 break;
1472 case DIFF_SYMBOL_FILEPAIR_MINUS:
1473 meta = diff_get_color_opt(o, DIFF_METAINFO);
1474 reset = diff_get_color_opt(o, DIFF_RESET);
1475 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1476 line, reset,
1477 strchr(line, ' ') ? "\t" : "");
1478 break;
1479 case DIFF_SYMBOL_BINARY_FILES:
1480 case DIFF_SYMBOL_HEADER:
1481 fprintf(o->file, "%s", line);
1482 break;
1483 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1484 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1485 break;
1486 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1487 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1488 break;
1489 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1490 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1491 break;
1492 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1493 fputs(diff_line_prefix(o), o->file);
1494 fputc('\n', o->file);
1495 break;
1496 case DIFF_SYMBOL_REWRITE_DIFF:
1497 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1498 reset = diff_get_color_opt(o, DIFF_RESET);
1499 emit_line(o, fraginfo, reset, line, len);
1500 break;
1501 case DIFF_SYMBOL_SUBMODULE_ADD:
1502 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1503 reset = diff_get_color_opt(o, DIFF_RESET);
1504 emit_line(o, set, reset, line, len);
1505 break;
1506 case DIFF_SYMBOL_SUBMODULE_DEL:
1507 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1508 reset = diff_get_color_opt(o, DIFF_RESET);
1509 emit_line(o, set, reset, line, len);
1510 break;
1511 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1512 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1513 diff_line_prefix(o), line);
1514 break;
1515 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1516 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1517 diff_line_prefix(o), line);
1518 break;
1519 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1520 emit_line(o, "", "", " 0 files changed\n",
1521 strlen(" 0 files changed\n"));
1522 break;
1523 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1524 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1525 break;
1526 case DIFF_SYMBOL_WORD_DIFF:
1527 fprintf(o->file, "%.*s", len, line);
1528 break;
1529 case DIFF_SYMBOL_STAT_SEP:
1530 fputs(o->stat_sep, o->file);
1531 break;
1532 default:
1533 BUG("unknown diff symbol");
1534 }
1535 }
1536
1537 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1538 const char *line, int len, unsigned flags)
1539 {
1540 struct emitted_diff_symbol e = {
1541 .line = line, .len = len, .flags = flags, .s = s
1542 };
1543
1544 if (o->emitted_symbols)
1545 append_emitted_diff_symbol(o, &e);
1546 else
1547 emit_diff_symbol_from_struct(o, &e);
1548 }
1549
1550 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1551 {
1552 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1553 }
1554
1555 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1556 {
1557 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1558 }
1559
1560 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1561 {
1562 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1563 path, strlen(path), 0);
1564 }
1565
1566 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1567 {
1568 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1569 path, strlen(path), 0);
1570 }
1571
1572 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1573 {
1574 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1575 header, strlen(header), 0);
1576 }
1577
1578 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1579 {
1580 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1581 }
1582
1583 void diff_emit_submodule_pipethrough(struct diff_options *o,
1584 const char *line, int len)
1585 {
1586 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1587 }
1588
1589 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1590 {
1591 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1592 ecbdata->blank_at_eof_in_preimage &&
1593 ecbdata->blank_at_eof_in_postimage &&
1594 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1595 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1596 return 0;
1597 return ws_blank_line(line, len);
1598 }
1599
1600 static void emit_add_line(struct emit_callback *ecbdata,
1601 const char *line, int len)
1602 {
1603 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1604 if (new_blank_line_at_eof(ecbdata, line, len))
1605 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1606
1607 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1608 }
1609
1610 static void emit_del_line(struct emit_callback *ecbdata,
1611 const char *line, int len)
1612 {
1613 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1614 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1615 }
1616
1617 static void emit_context_line(struct emit_callback *ecbdata,
1618 const char *line, int len)
1619 {
1620 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1621 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1622 }
1623
1624 static void emit_hunk_header(struct emit_callback *ecbdata,
1625 const char *line, int len)
1626 {
1627 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1628 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1629 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1630 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1631 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1632 static const char atat[2] = { '@', '@' };
1633 const char *cp, *ep;
1634 struct strbuf msgbuf = STRBUF_INIT;
1635 int org_len = len;
1636 int i = 1;
1637
1638 /*
1639 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1640 * it always is at least 10 bytes long.
1641 */
1642 if (len < 10 ||
1643 memcmp(line, atat, 2) ||
1644 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1645 emit_diff_symbol(ecbdata->opt,
1646 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1647 return;
1648 }
1649 ep += 2; /* skip over @@ */
1650
1651 /* The hunk header in fraginfo color */
1652 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1653 strbuf_addstr(&msgbuf, reverse);
1654 strbuf_addstr(&msgbuf, frag);
1655 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1656 strbuf_add(&msgbuf, atat, sizeof(atat));
1657 else
1658 strbuf_add(&msgbuf, line, ep - line);
1659 strbuf_addstr(&msgbuf, reset);
1660
1661 /*
1662 * trailing "\r\n"
1663 */
1664 for ( ; i < 3; i++)
1665 if (line[len - i] == '\r' || line[len - i] == '\n')
1666 len--;
1667
1668 /* blank before the func header */
1669 for (cp = ep; ep - line < len; ep++)
1670 if (*ep != ' ' && *ep != '\t')
1671 break;
1672 if (ep != cp) {
1673 strbuf_addstr(&msgbuf, context);
1674 strbuf_add(&msgbuf, cp, ep - cp);
1675 strbuf_addstr(&msgbuf, reset);
1676 }
1677
1678 if (ep < line + len) {
1679 strbuf_addstr(&msgbuf, func);
1680 strbuf_add(&msgbuf, ep, line + len - ep);
1681 strbuf_addstr(&msgbuf, reset);
1682 }
1683
1684 strbuf_add(&msgbuf, line + len, org_len - len);
1685 strbuf_complete_line(&msgbuf);
1686 emit_diff_symbol(ecbdata->opt,
1687 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1688 strbuf_release(&msgbuf);
1689 }
1690
1691 static struct diff_tempfile *claim_diff_tempfile(void)
1692 {
1693 int i;
1694 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1695 if (!diff_temp[i].name)
1696 return diff_temp + i;
1697 BUG("diff is failing to clean up its tempfiles");
1698 }
1699
1700 static void remove_tempfile(void)
1701 {
1702 int i;
1703 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1704 if (is_tempfile_active(diff_temp[i].tempfile))
1705 delete_tempfile(&diff_temp[i].tempfile);
1706 diff_temp[i].name = NULL;
1707 }
1708 }
1709
1710 static void add_line_count(struct strbuf *out, int count)
1711 {
1712 switch (count) {
1713 case 0:
1714 strbuf_addstr(out, "0,0");
1715 break;
1716 case 1:
1717 strbuf_addstr(out, "1");
1718 break;
1719 default:
1720 strbuf_addf(out, "1,%d", count);
1721 break;
1722 }
1723 }
1724
1725 static void emit_rewrite_lines(struct emit_callback *ecb,
1726 int prefix, const char *data, int size)
1727 {
1728 const char *endp = NULL;
1729
1730 while (0 < size) {
1731 int len;
1732
1733 endp = memchr(data, '\n', size);
1734 len = endp ? (endp - data + 1) : size;
1735 if (prefix != '+') {
1736 ecb->lno_in_preimage++;
1737 emit_del_line(ecb, data, len);
1738 } else {
1739 ecb->lno_in_postimage++;
1740 emit_add_line(ecb, data, len);
1741 }
1742 size -= len;
1743 data += len;
1744 }
1745 if (!endp)
1746 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1747 }
1748
1749 static void emit_rewrite_diff(const char *name_a,
1750 const char *name_b,
1751 struct diff_filespec *one,
1752 struct diff_filespec *two,
1753 struct userdiff_driver *textconv_one,
1754 struct userdiff_driver *textconv_two,
1755 struct diff_options *o)
1756 {
1757 int lc_a, lc_b;
1758 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1759 const char *a_prefix, *b_prefix;
1760 char *data_one, *data_two;
1761 size_t size_one, size_two;
1762 struct emit_callback ecbdata;
1763 struct strbuf out = STRBUF_INIT;
1764
1765 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1766 a_prefix = o->b_prefix;
1767 b_prefix = o->a_prefix;
1768 } else {
1769 a_prefix = o->a_prefix;
1770 b_prefix = o->b_prefix;
1771 }
1772
1773 name_a += (*name_a == '/');
1774 name_b += (*name_b == '/');
1775
1776 strbuf_reset(&a_name);
1777 strbuf_reset(&b_name);
1778 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1779 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1780
1781 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1782 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1783
1784 memset(&ecbdata, 0, sizeof(ecbdata));
1785 ecbdata.color_diff = want_color(o->use_color);
1786 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1787 ecbdata.opt = o;
1788 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1789 mmfile_t mf1, mf2;
1790 mf1.ptr = (char *)data_one;
1791 mf2.ptr = (char *)data_two;
1792 mf1.size = size_one;
1793 mf2.size = size_two;
1794 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1795 }
1796 ecbdata.lno_in_preimage = 1;
1797 ecbdata.lno_in_postimage = 1;
1798
1799 lc_a = count_lines(data_one, size_one);
1800 lc_b = count_lines(data_two, size_two);
1801
1802 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1803 a_name.buf, a_name.len, 0);
1804 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1805 b_name.buf, b_name.len, 0);
1806
1807 strbuf_addstr(&out, "@@ -");
1808 if (!o->irreversible_delete)
1809 add_line_count(&out, lc_a);
1810 else
1811 strbuf_addstr(&out, "?,?");
1812 strbuf_addstr(&out, " +");
1813 add_line_count(&out, lc_b);
1814 strbuf_addstr(&out, " @@\n");
1815 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1816 strbuf_release(&out);
1817
1818 if (lc_a && !o->irreversible_delete)
1819 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1820 if (lc_b)
1821 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1822 if (textconv_one)
1823 free((char *)data_one);
1824 if (textconv_two)
1825 free((char *)data_two);
1826 }
1827
1828 struct diff_words_buffer {
1829 mmfile_t text;
1830 unsigned long alloc;
1831 struct diff_words_orig {
1832 const char *begin, *end;
1833 } *orig;
1834 int orig_nr, orig_alloc;
1835 };
1836
1837 static void diff_words_append(char *line, unsigned long len,
1838 struct diff_words_buffer *buffer)
1839 {
1840 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1841 line++;
1842 len--;
1843 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1844 buffer->text.size += len;
1845 buffer->text.ptr[buffer->text.size] = '\0';
1846 }
1847
1848 struct diff_words_style_elem {
1849 const char *prefix;
1850 const char *suffix;
1851 const char *color; /* NULL; filled in by the setup code if
1852 * color is enabled */
1853 };
1854
1855 struct diff_words_style {
1856 enum diff_words_type type;
1857 struct diff_words_style_elem new_word, old_word, ctx;
1858 const char *newline;
1859 };
1860
1861 static struct diff_words_style diff_words_styles[] = {
1862 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1863 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1864 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1865 };
1866
1867 struct diff_words_data {
1868 struct diff_words_buffer minus, plus;
1869 const char *current_plus;
1870 int last_minus;
1871 struct diff_options *opt;
1872 regex_t *word_regex;
1873 enum diff_words_type type;
1874 struct diff_words_style *style;
1875 };
1876
1877 static int fn_out_diff_words_write_helper(struct diff_options *o,
1878 struct diff_words_style_elem *st_el,
1879 const char *newline,
1880 size_t count, const char *buf)
1881 {
1882 int print = 0;
1883 struct strbuf sb = STRBUF_INIT;
1884
1885 while (count) {
1886 char *p = memchr(buf, '\n', count);
1887 if (print)
1888 strbuf_addstr(&sb, diff_line_prefix(o));
1889
1890 if (p != buf) {
1891 const char *reset = st_el->color && *st_el->color ?
1892 GIT_COLOR_RESET : NULL;
1893 if (st_el->color && *st_el->color)
1894 strbuf_addstr(&sb, st_el->color);
1895 strbuf_addstr(&sb, st_el->prefix);
1896 strbuf_add(&sb, buf, p ? p - buf : count);
1897 strbuf_addstr(&sb, st_el->suffix);
1898 if (reset)
1899 strbuf_addstr(&sb, reset);
1900 }
1901 if (!p)
1902 goto out;
1903
1904 strbuf_addstr(&sb, newline);
1905 count -= p + 1 - buf;
1906 buf = p + 1;
1907 print = 1;
1908 if (count) {
1909 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1910 sb.buf, sb.len, 0);
1911 strbuf_reset(&sb);
1912 }
1913 }
1914
1915 out:
1916 if (sb.len)
1917 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1918 sb.buf, sb.len, 0);
1919 strbuf_release(&sb);
1920 return 0;
1921 }
1922
1923 /*
1924 * '--color-words' algorithm can be described as:
1925 *
1926 * 1. collect the minus/plus lines of a diff hunk, divided into
1927 * minus-lines and plus-lines;
1928 *
1929 * 2. break both minus-lines and plus-lines into words and
1930 * place them into two mmfile_t with one word for each line;
1931 *
1932 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1933 *
1934 * And for the common parts of the both file, we output the plus side text.
1935 * diff_words->current_plus is used to trace the current position of the plus file
1936 * which printed. diff_words->last_minus is used to trace the last minus word
1937 * printed.
1938 *
1939 * For '--graph' to work with '--color-words', we need to output the graph prefix
1940 * on each line of color words output. Generally, there are two conditions on
1941 * which we should output the prefix.
1942 *
1943 * 1. diff_words->last_minus == 0 &&
1944 * diff_words->current_plus == diff_words->plus.text.ptr
1945 *
1946 * that is: the plus text must start as a new line, and if there is no minus
1947 * word printed, a graph prefix must be printed.
1948 *
1949 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1950 * *(diff_words->current_plus - 1) == '\n'
1951 *
1952 * that is: a graph prefix must be printed following a '\n'
1953 */
1954 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1955 {
1956 if ((diff_words->last_minus == 0 &&
1957 diff_words->current_plus == diff_words->plus.text.ptr) ||
1958 (diff_words->current_plus > diff_words->plus.text.ptr &&
1959 *(diff_words->current_plus - 1) == '\n')) {
1960 return 1;
1961 } else {
1962 return 0;
1963 }
1964 }
1965
1966 static void fn_out_diff_words_aux(void *priv,
1967 long minus_first, long minus_len,
1968 long plus_first, long plus_len,
1969 const char *func UNUSED, long funclen UNUSED)
1970 {
1971 struct diff_words_data *diff_words = priv;
1972 struct diff_words_style *style = diff_words->style;
1973 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1974 struct diff_options *opt = diff_words->opt;
1975 const char *line_prefix;
1976
1977 assert(opt);
1978 line_prefix = diff_line_prefix(opt);
1979
1980 /* POSIX requires that first be decremented by one if len == 0... */
1981 if (minus_len) {
1982 minus_begin = diff_words->minus.orig[minus_first].begin;
1983 minus_end =
1984 diff_words->minus.orig[minus_first + minus_len - 1].end;
1985 } else
1986 minus_begin = minus_end =
1987 diff_words->minus.orig[minus_first].end;
1988
1989 if (plus_len) {
1990 plus_begin = diff_words->plus.orig[plus_first].begin;
1991 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1992 } else
1993 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1994
1995 if (color_words_output_graph_prefix(diff_words)) {
1996 fputs(line_prefix, diff_words->opt->file);
1997 }
1998 if (diff_words->current_plus != plus_begin) {
1999 fn_out_diff_words_write_helper(diff_words->opt,
2000 &style->ctx, style->newline,
2001 plus_begin - diff_words->current_plus,
2002 diff_words->current_plus);
2003 }
2004 if (minus_begin != minus_end) {
2005 fn_out_diff_words_write_helper(diff_words->opt,
2006 &style->old_word, style->newline,
2007 minus_end - minus_begin, minus_begin);
2008 }
2009 if (plus_begin != plus_end) {
2010 fn_out_diff_words_write_helper(diff_words->opt,
2011 &style->new_word, style->newline,
2012 plus_end - plus_begin, plus_begin);
2013 }
2014
2015 diff_words->current_plus = plus_end;
2016 diff_words->last_minus = minus_first;
2017 }
2018
2019 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2020 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2021 int *begin, int *end)
2022 {
2023 while (word_regex && *begin < buffer->size) {
2024 regmatch_t match[1];
2025 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2026 buffer->size - *begin, 1, match, 0)) {
2027 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2028 '\n', match[0].rm_eo - match[0].rm_so);
2029 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2030 *begin += match[0].rm_so;
2031 if (*begin == *end)
2032 (*begin)++;
2033 else
2034 return *begin > *end;
2035 } else {
2036 return -1;
2037 }
2038 }
2039
2040 /* find the next word */
2041 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2042 (*begin)++;
2043 if (*begin >= buffer->size)
2044 return -1;
2045
2046 /* find the end of the word */
2047 *end = *begin + 1;
2048 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2049 (*end)++;
2050
2051 return 0;
2052 }
2053
2054 /*
2055 * This function splits the words in buffer->text, stores the list with
2056 * newline separator into out, and saves the offsets of the original words
2057 * in buffer->orig.
2058 */
2059 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2060 regex_t *word_regex)
2061 {
2062 int i, j;
2063 long alloc = 0;
2064
2065 out->size = 0;
2066 out->ptr = NULL;
2067
2068 /* fake an empty "0th" word */
2069 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2070 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2071 buffer->orig_nr = 1;
2072
2073 for (i = 0; i < buffer->text.size; i++) {
2074 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2075 return;
2076
2077 /* store original boundaries */
2078 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2079 buffer->orig_alloc);
2080 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2081 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2082 buffer->orig_nr++;
2083
2084 /* store one word */
2085 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2086 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2087 out->ptr[out->size + j - i] = '\n';
2088 out->size += j - i + 1;
2089
2090 i = j - 1;
2091 }
2092 }
2093
2094 /* this executes the word diff on the accumulated buffers */
2095 static void diff_words_show(struct diff_words_data *diff_words)
2096 {
2097 xpparam_t xpp;
2098 xdemitconf_t xecfg;
2099 mmfile_t minus, plus;
2100 struct diff_words_style *style = diff_words->style;
2101
2102 struct diff_options *opt = diff_words->opt;
2103 const char *line_prefix;
2104
2105 assert(opt);
2106 line_prefix = diff_line_prefix(opt);
2107
2108 /* special case: only removal */
2109 if (!diff_words->plus.text.size) {
2110 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2111 line_prefix, strlen(line_prefix), 0);
2112 fn_out_diff_words_write_helper(diff_words->opt,
2113 &style->old_word, style->newline,
2114 diff_words->minus.text.size,
2115 diff_words->minus.text.ptr);
2116 diff_words->minus.text.size = 0;
2117 return;
2118 }
2119
2120 diff_words->current_plus = diff_words->plus.text.ptr;
2121 diff_words->last_minus = 0;
2122
2123 memset(&xpp, 0, sizeof(xpp));
2124 memset(&xecfg, 0, sizeof(xecfg));
2125 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2126 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2127 xpp.flags = 0;
2128 /* as only the hunk header will be parsed, we need a 0-context */
2129 xecfg.ctxlen = 0;
2130 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2131 diff_words, &xpp, &xecfg))
2132 die("unable to generate word diff");
2133 free(minus.ptr);
2134 free(plus.ptr);
2135 if (diff_words->current_plus != diff_words->plus.text.ptr +
2136 diff_words->plus.text.size) {
2137 if (color_words_output_graph_prefix(diff_words))
2138 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2139 line_prefix, strlen(line_prefix), 0);
2140 fn_out_diff_words_write_helper(diff_words->opt,
2141 &style->ctx, style->newline,
2142 diff_words->plus.text.ptr + diff_words->plus.text.size
2143 - diff_words->current_plus, diff_words->current_plus);
2144 }
2145 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2146 }
2147
2148 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2149 static void diff_words_flush(struct emit_callback *ecbdata)
2150 {
2151 struct diff_options *wo = ecbdata->diff_words->opt;
2152
2153 if (ecbdata->diff_words->minus.text.size ||
2154 ecbdata->diff_words->plus.text.size)
2155 diff_words_show(ecbdata->diff_words);
2156
2157 if (wo->emitted_symbols) {
2158 struct diff_options *o = ecbdata->opt;
2159 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2160 int i;
2161
2162 /*
2163 * NEEDSWORK:
2164 * Instead of appending each, concat all words to a line?
2165 */
2166 for (i = 0; i < wol->nr; i++)
2167 append_emitted_diff_symbol(o, &wol->buf[i]);
2168
2169 for (i = 0; i < wol->nr; i++)
2170 free((void *)wol->buf[i].line);
2171
2172 wol->nr = 0;
2173 }
2174 }
2175
2176 static void diff_filespec_load_driver(struct diff_filespec *one,
2177 struct index_state *istate)
2178 {
2179 /* Use already-loaded driver */
2180 if (one->driver)
2181 return;
2182
2183 if (S_ISREG(one->mode))
2184 one->driver = userdiff_find_by_path(istate, one->path);
2185
2186 /* Fallback to default settings */
2187 if (!one->driver)
2188 one->driver = userdiff_find_by_name("default");
2189 }
2190
2191 static const char *userdiff_word_regex(struct diff_filespec *one,
2192 struct index_state *istate)
2193 {
2194 diff_filespec_load_driver(one, istate);
2195 return one->driver->word_regex;
2196 }
2197
2198 static void init_diff_words_data(struct emit_callback *ecbdata,
2199 struct diff_options *orig_opts,
2200 struct diff_filespec *one,
2201 struct diff_filespec *two)
2202 {
2203 int i;
2204 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2205 memcpy(o, orig_opts, sizeof(struct diff_options));
2206
2207 CALLOC_ARRAY(ecbdata->diff_words, 1);
2208 ecbdata->diff_words->type = o->word_diff;
2209 ecbdata->diff_words->opt = o;
2210
2211 if (orig_opts->emitted_symbols)
2212 CALLOC_ARRAY(o->emitted_symbols, 1);
2213
2214 if (!o->word_regex)
2215 o->word_regex = userdiff_word_regex(one, o->repo->index);
2216 if (!o->word_regex)
2217 o->word_regex = userdiff_word_regex(two, o->repo->index);
2218 if (!o->word_regex)
2219 o->word_regex = diff_word_regex_cfg;
2220 if (o->word_regex) {
2221 ecbdata->diff_words->word_regex = (regex_t *)
2222 xmalloc(sizeof(regex_t));
2223 if (regcomp(ecbdata->diff_words->word_regex,
2224 o->word_regex,
2225 REG_EXTENDED | REG_NEWLINE))
2226 die("invalid regular expression: %s",
2227 o->word_regex);
2228 }
2229 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2230 if (o->word_diff == diff_words_styles[i].type) {
2231 ecbdata->diff_words->style =
2232 &diff_words_styles[i];
2233 break;
2234 }
2235 }
2236 if (want_color(o->use_color)) {
2237 struct diff_words_style *st = ecbdata->diff_words->style;
2238 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2239 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2240 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2241 }
2242 }
2243
2244 static void free_diff_words_data(struct emit_callback *ecbdata)
2245 {
2246 if (ecbdata->diff_words) {
2247 diff_words_flush(ecbdata);
2248 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2249 free (ecbdata->diff_words->opt);
2250 free (ecbdata->diff_words->minus.text.ptr);
2251 free (ecbdata->diff_words->minus.orig);
2252 free (ecbdata->diff_words->plus.text.ptr);
2253 free (ecbdata->diff_words->plus.orig);
2254 if (ecbdata->diff_words->word_regex) {
2255 regfree(ecbdata->diff_words->word_regex);
2256 free(ecbdata->diff_words->word_regex);
2257 }
2258 FREE_AND_NULL(ecbdata->diff_words);
2259 }
2260 }
2261
2262 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2263 {
2264 if (want_color(diff_use_color))
2265 return diff_colors[ix];
2266 return "";
2267 }
2268
2269 const char *diff_line_prefix(struct diff_options *opt)
2270 {
2271 struct strbuf *msgbuf;
2272 if (!opt->output_prefix)
2273 return "";
2274
2275 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2276 return msgbuf->buf;
2277 }
2278
2279 static unsigned long sane_truncate_line(char *line, unsigned long len)
2280 {
2281 const char *cp;
2282 unsigned long allot;
2283 size_t l = len;
2284
2285 cp = line;
2286 allot = l;
2287 while (0 < l) {
2288 (void) utf8_width(&cp, &l);
2289 if (!cp)
2290 break; /* truncated in the middle? */
2291 }
2292 return allot - l;
2293 }
2294
2295 static void find_lno(const char *line, struct emit_callback *ecbdata)
2296 {
2297 const char *p;
2298 ecbdata->lno_in_preimage = 0;
2299 ecbdata->lno_in_postimage = 0;
2300 p = strchr(line, '-');
2301 if (!p)
2302 return; /* cannot happen */
2303 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2304 p = strchr(p, '+');
2305 if (!p)
2306 return; /* cannot happen */
2307 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2308 }
2309
2310 static int fn_out_consume(void *priv, char *line, unsigned long len)
2311 {
2312 struct emit_callback *ecbdata = priv;
2313 struct diff_options *o = ecbdata->opt;
2314
2315 o->found_changes = 1;
2316
2317 if (ecbdata->header) {
2318 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2319 ecbdata->header->buf, ecbdata->header->len, 0);
2320 strbuf_reset(ecbdata->header);
2321 ecbdata->header = NULL;
2322 }
2323
2324 if (ecbdata->label_path[0]) {
2325 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2326 ecbdata->label_path[0],
2327 strlen(ecbdata->label_path[0]), 0);
2328 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2329 ecbdata->label_path[1],
2330 strlen(ecbdata->label_path[1]), 0);
2331 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2332 }
2333
2334 if (diff_suppress_blank_empty
2335 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2336 line[0] = '\n';
2337 len = 1;
2338 }
2339
2340 if (line[0] == '@') {
2341 if (ecbdata->diff_words)
2342 diff_words_flush(ecbdata);
2343 len = sane_truncate_line(line, len);
2344 find_lno(line, ecbdata);
2345 emit_hunk_header(ecbdata, line, len);
2346 return 0;
2347 }
2348
2349 if (ecbdata->diff_words) {
2350 enum diff_symbol s =
2351 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2352 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2353 if (line[0] == '-') {
2354 diff_words_append(line, len,
2355 &ecbdata->diff_words->minus);
2356 return 0;
2357 } else if (line[0] == '+') {
2358 diff_words_append(line, len,
2359 &ecbdata->diff_words->plus);
2360 return 0;
2361 } else if (starts_with(line, "\\ ")) {
2362 /*
2363 * Eat the "no newline at eof" marker as if we
2364 * saw a "+" or "-" line with nothing on it,
2365 * and return without diff_words_flush() to
2366 * defer processing. If this is the end of
2367 * preimage, more "+" lines may come after it.
2368 */
2369 return 0;
2370 }
2371 diff_words_flush(ecbdata);
2372 emit_diff_symbol(o, s, line, len, 0);
2373 return 0;
2374 }
2375
2376 switch (line[0]) {
2377 case '+':
2378 ecbdata->lno_in_postimage++;
2379 emit_add_line(ecbdata, line + 1, len - 1);
2380 break;
2381 case '-':
2382 ecbdata->lno_in_preimage++;
2383 emit_del_line(ecbdata, line + 1, len - 1);
2384 break;
2385 case ' ':
2386 ecbdata->lno_in_postimage++;
2387 ecbdata->lno_in_preimage++;
2388 emit_context_line(ecbdata, line + 1, len - 1);
2389 break;
2390 default:
2391 /* incomplete line at the end */
2392 ecbdata->lno_in_preimage++;
2393 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2394 line, len, 0);
2395 break;
2396 }
2397 return 0;
2398 }
2399
2400 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2401 {
2402 const char *old_name = a;
2403 const char *new_name = b;
2404 int pfx_length, sfx_length;
2405 int pfx_adjust_for_slash;
2406 int len_a = strlen(a);
2407 int len_b = strlen(b);
2408 int a_midlen, b_midlen;
2409 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2410 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2411
2412 if (qlen_a || qlen_b) {
2413 quote_c_style(a, name, NULL, 0);
2414 strbuf_addstr(name, " => ");
2415 quote_c_style(b, name, NULL, 0);
2416 return;
2417 }
2418
2419 /* Find common prefix */
2420 pfx_length = 0;
2421 while (*old_name && *new_name && *old_name == *new_name) {
2422 if (*old_name == '/')
2423 pfx_length = old_name - a + 1;
2424 old_name++;
2425 new_name++;
2426 }
2427
2428 /* Find common suffix */
2429 old_name = a + len_a;
2430 new_name = b + len_b;
2431 sfx_length = 0;
2432 /*
2433 * If there is a common prefix, it must end in a slash. In
2434 * that case we let this loop run 1 into the prefix to see the
2435 * same slash.
2436 *
2437 * If there is no common prefix, we cannot do this as it would
2438 * underrun the input strings.
2439 */
2440 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2441 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2442 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2443 *old_name == *new_name) {
2444 if (*old_name == '/')
2445 sfx_length = len_a - (old_name - a);
2446 old_name--;
2447 new_name--;
2448 }
2449
2450 /*
2451 * pfx{mid-a => mid-b}sfx
2452 * {pfx-a => pfx-b}sfx
2453 * pfx{sfx-a => sfx-b}
2454 * name-a => name-b
2455 */
2456 a_midlen = len_a - pfx_length - sfx_length;
2457 b_midlen = len_b - pfx_length - sfx_length;
2458 if (a_midlen < 0)
2459 a_midlen = 0;
2460 if (b_midlen < 0)
2461 b_midlen = 0;
2462
2463 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2464 if (pfx_length + sfx_length) {
2465 strbuf_add(name, a, pfx_length);
2466 strbuf_addch(name, '{');
2467 }
2468 strbuf_add(name, a + pfx_length, a_midlen);
2469 strbuf_addstr(name, " => ");
2470 strbuf_add(name, b + pfx_length, b_midlen);
2471 if (pfx_length + sfx_length) {
2472 strbuf_addch(name, '}');
2473 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2474 }
2475 }
2476
2477 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2478 const char *name_a,
2479 const char *name_b)
2480 {
2481 struct diffstat_file *x;
2482 CALLOC_ARRAY(x, 1);
2483 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2484 diffstat->files[diffstat->nr++] = x;
2485 if (name_b) {
2486 x->from_name = xstrdup(name_a);
2487 x->name = xstrdup(name_b);
2488 x->is_renamed = 1;
2489 }
2490 else {
2491 x->from_name = NULL;
2492 x->name = xstrdup(name_a);
2493 }
2494 return x;
2495 }
2496
2497 static int diffstat_consume(void *priv, char *line, unsigned long len)
2498 {
2499 struct diffstat_t *diffstat = priv;
2500 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2501
2502 if (!len)
2503 BUG("xdiff fed us an empty line");
2504
2505 if (line[0] == '+')
2506 x->added++;
2507 else if (line[0] == '-')
2508 x->deleted++;
2509 return 0;
2510 }
2511
2512 const char mime_boundary_leader[] = "------------";
2513
2514 static int scale_linear(int it, int width, int max_change)
2515 {
2516 if (!it)
2517 return 0;
2518 /*
2519 * make sure that at least one '-' or '+' is printed if
2520 * there is any change to this path. The easiest way is to
2521 * scale linearly as if the allotted width is one column shorter
2522 * than it is, and then add 1 to the result.
2523 */
2524 return 1 + (it * (width - 1) / max_change);
2525 }
2526
2527 static void show_graph(struct strbuf *out, char ch, int cnt,
2528 const char *set, const char *reset)
2529 {
2530 if (cnt <= 0)
2531 return;
2532 strbuf_addstr(out, set);
2533 strbuf_addchars(out, ch, cnt);
2534 strbuf_addstr(out, reset);
2535 }
2536
2537 static void fill_print_name(struct diffstat_file *file)
2538 {
2539 struct strbuf pname = STRBUF_INIT;
2540
2541 if (file->print_name)
2542 return;
2543
2544 if (file->is_renamed)
2545 pprint_rename(&pname, file->from_name, file->name);
2546 else
2547 quote_c_style(file->name, &pname, NULL, 0);
2548
2549 if (file->comments)
2550 strbuf_addf(&pname, " (%s)", file->comments);
2551
2552 file->print_name = strbuf_detach(&pname, NULL);
2553 }
2554
2555 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2556 int files, int insertions, int deletions)
2557 {
2558 struct strbuf sb = STRBUF_INIT;
2559
2560 if (!files) {
2561 assert(insertions == 0 && deletions == 0);
2562 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2563 NULL, 0, 0);
2564 return;
2565 }
2566
2567 strbuf_addf(&sb,
2568 (files == 1) ? " %d file changed" : " %d files changed",
2569 files);
2570
2571 /*
2572 * For binary diff, the caller may want to print "x files
2573 * changed" with insertions == 0 && deletions == 0.
2574 *
2575 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2576 * is probably less confusing (i.e skip over "2 files changed
2577 * but nothing about added/removed lines? Is this a bug in Git?").
2578 */
2579 if (insertions || deletions == 0) {
2580 strbuf_addf(&sb,
2581 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2582 insertions);
2583 }
2584
2585 if (deletions || insertions == 0) {
2586 strbuf_addf(&sb,
2587 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2588 deletions);
2589 }
2590 strbuf_addch(&sb, '\n');
2591 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2592 sb.buf, sb.len, 0);
2593 strbuf_release(&sb);
2594 }
2595
2596 void print_stat_summary(FILE *fp, int files,
2597 int insertions, int deletions)
2598 {
2599 struct diff_options o;
2600 memset(&o, 0, sizeof(o));
2601 o.file = fp;
2602
2603 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2604 }
2605
2606 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2607 {
2608 int i, len, add, del, adds = 0, dels = 0;
2609 uintmax_t max_change = 0, max_len = 0;
2610 int total_files = data->nr, count;
2611 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2612 const char *reset, *add_c, *del_c;
2613 int extra_shown = 0;
2614 const char *line_prefix = diff_line_prefix(options);
2615 struct strbuf out = STRBUF_INIT;
2616
2617 if (data->nr == 0)
2618 return;
2619
2620 count = options->stat_count ? options->stat_count : data->nr;
2621
2622 reset = diff_get_color_opt(options, DIFF_RESET);
2623 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2624 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2625
2626 /*
2627 * Find the longest filename and max number of changes
2628 */
2629 for (i = 0; (i < count) && (i < data->nr); i++) {
2630 struct diffstat_file *file = data->files[i];
2631 uintmax_t change = file->added + file->deleted;
2632
2633 if (!file->is_interesting && (change == 0)) {
2634 count++; /* not shown == room for one more */
2635 continue;
2636 }
2637 fill_print_name(file);
2638 len = utf8_strwidth(file->print_name);
2639 if (max_len < len)
2640 max_len = len;
2641
2642 if (file->is_unmerged) {
2643 /* "Unmerged" is 8 characters */
2644 bin_width = bin_width < 8 ? 8 : bin_width;
2645 continue;
2646 }
2647 if (file->is_binary) {
2648 /* "Bin XXX -> YYY bytes" */
2649 int w = 14 + decimal_width(file->added)
2650 + decimal_width(file->deleted);
2651 bin_width = bin_width < w ? w : bin_width;
2652 /* Display change counts aligned with "Bin" */
2653 number_width = 3;
2654 continue;
2655 }
2656
2657 if (max_change < change)
2658 max_change = change;
2659 }
2660 count = i; /* where we can stop scanning in data->files[] */
2661
2662 /*
2663 * We have width = stat_width or term_columns() columns total.
2664 * We want a maximum of min(max_len, stat_name_width) for the name part.
2665 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2666 * We also need 1 for " " and 4 + decimal_width(max_change)
2667 * for " | NNNN " and one the empty column at the end, altogether
2668 * 6 + decimal_width(max_change).
2669 *
2670 * If there's not enough space, we will use the smaller of
2671 * stat_name_width (if set) and 5/8*width for the filename,
2672 * and the rest for constant elements + graph part, but no more
2673 * than stat_graph_width for the graph part.
2674 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2675 * for the standard terminal size).
2676 *
2677 * In other words: stat_width limits the maximum width, and
2678 * stat_name_width fixes the maximum width of the filename,
2679 * and is also used to divide available columns if there
2680 * aren't enough.
2681 *
2682 * Binary files are displayed with "Bin XXX -> YYY bytes"
2683 * instead of the change count and graph. This part is treated
2684 * similarly to the graph part, except that it is not
2685 * "scaled". If total width is too small to accommodate the
2686 * guaranteed minimum width of the filename part and the
2687 * separators and this message, this message will "overflow"
2688 * making the line longer than the maximum width.
2689 */
2690
2691 /*
2692 * NEEDSWORK: line_prefix is often used for "log --graph" output
2693 * and contains ANSI-colored string. utf8_strnwidth() should be
2694 * used to correctly count the display width instead of strlen().
2695 */
2696 if (options->stat_width == -1)
2697 width = term_columns() - strlen(line_prefix);
2698 else
2699 width = options->stat_width ? options->stat_width : 80;
2700 number_width = decimal_width(max_change) > number_width ?
2701 decimal_width(max_change) : number_width;
2702
2703 if (options->stat_graph_width == -1)
2704 options->stat_graph_width = diff_stat_graph_width;
2705
2706 /*
2707 * Guarantee 3/8*16==6 for the graph part
2708 * and 5/8*16==10 for the filename part
2709 */
2710 if (width < 16 + 6 + number_width)
2711 width = 16 + 6 + number_width;
2712
2713 /*
2714 * First assign sizes that are wanted, ignoring available width.
2715 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2716 * starting from "XXX" should fit in graph_width.
2717 */
2718 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2719 if (options->stat_graph_width &&
2720 options->stat_graph_width < graph_width)
2721 graph_width = options->stat_graph_width;
2722
2723 name_width = (options->stat_name_width > 0 &&
2724 options->stat_name_width < max_len) ?
2725 options->stat_name_width : max_len;
2726
2727 /*
2728 * Adjust adjustable widths not to exceed maximum width
2729 */
2730 if (name_width + number_width + 6 + graph_width > width) {
2731 if (graph_width > width * 3/8 - number_width - 6) {
2732 graph_width = width * 3/8 - number_width - 6;
2733 if (graph_width < 6)
2734 graph_width = 6;
2735 }
2736
2737 if (options->stat_graph_width &&
2738 graph_width > options->stat_graph_width)
2739 graph_width = options->stat_graph_width;
2740 if (name_width > width - number_width - 6 - graph_width)
2741 name_width = width - number_width - 6 - graph_width;
2742 else
2743 graph_width = width - number_width - 6 - name_width;
2744 }
2745
2746 /*
2747 * From here name_width is the width of the name area,
2748 * and graph_width is the width of the graph area.
2749 * max_change is used to scale graph properly.
2750 */
2751 for (i = 0; i < count; i++) {
2752 const char *prefix = "";
2753 struct diffstat_file *file = data->files[i];
2754 char *name = file->print_name;
2755 uintmax_t added = file->added;
2756 uintmax_t deleted = file->deleted;
2757 int name_len, padding;
2758
2759 if (!file->is_interesting && (added + deleted == 0))
2760 continue;
2761
2762 /*
2763 * "scale" the filename
2764 */
2765 len = name_width;
2766 name_len = utf8_strwidth(name);
2767 if (name_width < name_len) {
2768 char *slash;
2769 prefix = "...";
2770 len -= 3;
2771 /*
2772 * NEEDSWORK: (name_len - len) counts the display
2773 * width, which would be shorter than the byte
2774 * length of the corresponding substring.
2775 * Advancing "name" by that number of bytes does
2776 * *NOT* skip over that many columns, so it is
2777 * very likely that chomping the pathname at the
2778 * slash we will find starting from "name" will
2779 * leave the resulting string still too long.
2780 */
2781 name += name_len - len;
2782 slash = strchr(name, '/');
2783 if (slash)
2784 name = slash;
2785 }
2786 padding = len - utf8_strwidth(name);
2787 if (padding < 0)
2788 padding = 0;
2789
2790 if (file->is_binary) {
2791 strbuf_addf(&out, " %s%s%*s | %*s",
2792 prefix, name, padding, "",
2793 number_width, "Bin");
2794 if (!added && !deleted) {
2795 strbuf_addch(&out, '\n');
2796 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2797 out.buf, out.len, 0);
2798 strbuf_reset(&out);
2799 continue;
2800 }
2801 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2802 del_c, deleted, reset);
2803 strbuf_addstr(&out, " -> ");
2804 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2805 add_c, added, reset);
2806 strbuf_addstr(&out, " bytes\n");
2807 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2808 out.buf, out.len, 0);
2809 strbuf_reset(&out);
2810 continue;
2811 }
2812 else if (file->is_unmerged) {
2813 strbuf_addf(&out, " %s%s%*s | %*s",
2814 prefix, name, padding, "",
2815 number_width, "Unmerged\n");
2816 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2817 out.buf, out.len, 0);
2818 strbuf_reset(&out);
2819 continue;
2820 }
2821
2822 /*
2823 * scale the add/delete
2824 */
2825 add = added;
2826 del = deleted;
2827
2828 if (graph_width <= max_change) {
2829 int total = scale_linear(add + del, graph_width, max_change);
2830 if (total < 2 && add && del)
2831 /* width >= 2 due to the sanity check */
2832 total = 2;
2833 if (add < del) {
2834 add = scale_linear(add, graph_width, max_change);
2835 del = total - add;
2836 } else {
2837 del = scale_linear(del, graph_width, max_change);
2838 add = total - del;
2839 }
2840 }
2841 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2842 prefix, name, padding, "",
2843 number_width, added + deleted,
2844 added + deleted ? " " : "");
2845 show_graph(&out, '+', add, add_c, reset);
2846 show_graph(&out, '-', del, del_c, reset);
2847 strbuf_addch(&out, '\n');
2848 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2849 out.buf, out.len, 0);
2850 strbuf_reset(&out);
2851 }
2852
2853 for (i = 0; i < data->nr; i++) {
2854 struct diffstat_file *file = data->files[i];
2855 uintmax_t added = file->added;
2856 uintmax_t deleted = file->deleted;
2857
2858 if (file->is_unmerged ||
2859 (!file->is_interesting && (added + deleted == 0))) {
2860 total_files--;
2861 continue;
2862 }
2863
2864 if (!file->is_binary) {
2865 adds += added;
2866 dels += deleted;
2867 }
2868 if (i < count)
2869 continue;
2870 if (!extra_shown)
2871 emit_diff_symbol(options,
2872 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2873 NULL, 0, 0);
2874 extra_shown = 1;
2875 }
2876
2877 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2878 strbuf_release(&out);
2879 }
2880
2881 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2882 {
2883 int i, adds = 0, dels = 0, total_files = data->nr;
2884
2885 if (data->nr == 0)
2886 return;
2887
2888 for (i = 0; i < data->nr; i++) {
2889 int added = data->files[i]->added;
2890 int deleted = data->files[i]->deleted;
2891
2892 if (data->files[i]->is_unmerged ||
2893 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2894 total_files--;
2895 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2896 adds += added;
2897 dels += deleted;
2898 }
2899 }
2900 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2901 }
2902
2903 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2904 {
2905 int i;
2906
2907 if (data->nr == 0)
2908 return;
2909
2910 for (i = 0; i < data->nr; i++) {
2911 struct diffstat_file *file = data->files[i];
2912
2913 fprintf(options->file, "%s", diff_line_prefix(options));
2914
2915 if (file->is_binary)
2916 fprintf(options->file, "-\t-\t");
2917 else
2918 fprintf(options->file,
2919 "%"PRIuMAX"\t%"PRIuMAX"\t",
2920 file->added, file->deleted);
2921 if (options->line_termination) {
2922 fill_print_name(file);
2923 if (!file->is_renamed)
2924 write_name_quoted(file->name, options->file,
2925 options->line_termination);
2926 else {
2927 fputs(file->print_name, options->file);
2928 putc(options->line_termination, options->file);
2929 }
2930 } else {
2931 if (file->is_renamed) {
2932 putc('\0', options->file);
2933 write_name_quoted(file->from_name, options->file, '\0');
2934 }
2935 write_name_quoted(file->name, options->file, '\0');
2936 }
2937 }
2938 }
2939
2940 struct dirstat_file {
2941 const char *name;
2942 unsigned long changed;
2943 };
2944
2945 struct dirstat_dir {
2946 struct dirstat_file *files;
2947 int alloc, nr, permille, cumulative;
2948 };
2949
2950 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2951 unsigned long changed, const char *base, int baselen)
2952 {
2953 unsigned long sum_changes = 0;
2954 unsigned int sources = 0;
2955 const char *line_prefix = diff_line_prefix(opt);
2956
2957 while (dir->nr) {
2958 struct dirstat_file *f = dir->files;
2959 int namelen = strlen(f->name);
2960 unsigned long changes;
2961 char *slash;
2962
2963 if (namelen < baselen)
2964 break;
2965 if (memcmp(f->name, base, baselen))
2966 break;
2967 slash = strchr(f->name + baselen, '/');
2968 if (slash) {
2969 int newbaselen = slash + 1 - f->name;
2970 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2971 sources++;
2972 } else {
2973 changes = f->changed;
2974 dir->files++;
2975 dir->nr--;
2976 sources += 2;
2977 }
2978 sum_changes += changes;
2979 }
2980
2981 /*
2982 * We don't report dirstat's for
2983 * - the top level
2984 * - or cases where everything came from a single directory
2985 * under this directory (sources == 1).
2986 */
2987 if (baselen && sources != 1) {
2988 if (sum_changes) {
2989 int permille = sum_changes * 1000 / changed;
2990 if (permille >= dir->permille) {
2991 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2992 permille / 10, permille % 10, baselen, base);
2993 if (!dir->cumulative)
2994 return 0;
2995 }
2996 }
2997 }
2998 return sum_changes;
2999 }
3000
3001 static int dirstat_compare(const void *_a, const void *_b)
3002 {
3003 const struct dirstat_file *a = _a;
3004 const struct dirstat_file *b = _b;
3005 return strcmp(a->name, b->name);
3006 }
3007
3008 static void show_dirstat(struct diff_options *options)
3009 {
3010 int i;
3011 unsigned long changed;
3012 struct dirstat_dir dir;
3013 struct diff_queue_struct *q = &diff_queued_diff;
3014
3015 dir.files = NULL;
3016 dir.alloc = 0;
3017 dir.nr = 0;
3018 dir.permille = options->dirstat_permille;
3019 dir.cumulative = options->flags.dirstat_cumulative;
3020
3021 changed = 0;
3022 for (i = 0; i < q->nr; i++) {
3023 struct diff_filepair *p = q->queue[i];
3024 const char *name;
3025 unsigned long copied, added, damage;
3026 struct diff_populate_filespec_options dpf_options = {
3027 .check_size_only = 1,
3028 };
3029
3030 name = p->two->path ? p->two->path : p->one->path;
3031
3032 if (p->one->oid_valid && p->two->oid_valid &&
3033 oideq(&p->one->oid, &p->two->oid)) {
3034 /*
3035 * The SHA1 has not changed, so pre-/post-content is
3036 * identical. We can therefore skip looking at the
3037 * file contents altogether.
3038 */
3039 damage = 0;
3040 goto found_damage;
3041 }
3042
3043 if (options->flags.dirstat_by_file) {
3044 /*
3045 * In --dirstat-by-file mode, we don't really need to
3046 * look at the actual file contents at all.
3047 * The fact that the SHA1 changed is enough for us to
3048 * add this file to the list of results
3049 * (with each file contributing equal damage).
3050 */
3051 damage = 1;
3052 goto found_damage;
3053 }
3054
3055 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3056 diff_populate_filespec(options->repo, p->one, NULL);
3057 diff_populate_filespec(options->repo, p->two, NULL);
3058 diffcore_count_changes(options->repo,
3059 p->one, p->two, NULL, NULL,
3060 &copied, &added);
3061 diff_free_filespec_data(p->one);
3062 diff_free_filespec_data(p->two);
3063 } else if (DIFF_FILE_VALID(p->one)) {
3064 diff_populate_filespec(options->repo, p->one, &dpf_options);
3065 copied = added = 0;
3066 diff_free_filespec_data(p->one);
3067 } else if (DIFF_FILE_VALID(p->two)) {
3068 diff_populate_filespec(options->repo, p->two, &dpf_options);
3069 copied = 0;
3070 added = p->two->size;
3071 diff_free_filespec_data(p->two);
3072 } else
3073 continue;
3074
3075 /*
3076 * Original minus copied is the removed material,
3077 * added is the new material. They are both damages
3078 * made to the preimage.
3079 * If the resulting damage is zero, we know that
3080 * diffcore_count_changes() considers the two entries to
3081 * be identical, but since the oid changed, we
3082 * know that there must have been _some_ kind of change,
3083 * so we force all entries to have damage > 0.
3084 */
3085 damage = (p->one->size - copied) + added;
3086 if (!damage)
3087 damage = 1;
3088
3089 found_damage:
3090 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3091 dir.files[dir.nr].name = name;
3092 dir.files[dir.nr].changed = damage;
3093 changed += damage;
3094 dir.nr++;
3095 }
3096
3097 /* This can happen even with many files, if everything was renames */
3098 if (!changed)
3099 return;
3100
3101 /* Show all directories with more than x% of the changes */
3102 QSORT(dir.files, dir.nr, dirstat_compare);
3103 gather_dirstat(options, &dir, changed, "", 0);
3104 }
3105
3106 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3107 {
3108 int i;
3109 unsigned long changed;
3110 struct dirstat_dir dir;
3111
3112 if (data->nr == 0)
3113 return;
3114
3115 dir.files = NULL;
3116 dir.alloc = 0;
3117 dir.nr = 0;
3118 dir.permille = options->dirstat_permille;
3119 dir.cumulative = options->flags.dirstat_cumulative;
3120
3121 changed = 0;
3122 for (i = 0; i < data->nr; i++) {
3123 struct diffstat_file *file = data->files[i];
3124 unsigned long damage = file->added + file->deleted;
3125 if (file->is_binary)
3126 /*
3127 * binary files counts bytes, not lines. Must find some
3128 * way to normalize binary bytes vs. textual lines.
3129 * The following heuristic assumes that there are 64
3130 * bytes per "line".
3131 * This is stupid and ugly, but very cheap...
3132 */
3133 damage = DIV_ROUND_UP(damage, 64);
3134 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3135 dir.files[dir.nr].name = file->name;
3136 dir.files[dir.nr].changed = damage;
3137 changed += damage;
3138 dir.nr++;
3139 }
3140
3141 /* This can happen even with many files, if everything was renames */
3142 if (!changed)
3143 return;
3144
3145 /* Show all directories with more than x% of the changes */
3146 QSORT(dir.files, dir.nr, dirstat_compare);
3147 gather_dirstat(options, &dir, changed, "", 0);
3148 }
3149
3150 static void free_diffstat_file(struct diffstat_file *f)
3151 {
3152 free(f->print_name);
3153 free(f->name);
3154 free(f->from_name);
3155 free(f);
3156 }
3157
3158 void free_diffstat_info(struct diffstat_t *diffstat)
3159 {
3160 int i;
3161 for (i = 0; i < diffstat->nr; i++)
3162 free_diffstat_file(diffstat->files[i]);
3163 free(diffstat->files);
3164 }
3165
3166 struct checkdiff_t {
3167 const char *filename;
3168 int lineno;
3169 int conflict_marker_size;
3170 struct diff_options *o;
3171 unsigned ws_rule;
3172 unsigned status;
3173 };
3174
3175 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3176 {
3177 char firstchar;
3178 int cnt;
3179
3180 if (len < marker_size + 1)
3181 return 0;
3182 firstchar = line[0];
3183 switch (firstchar) {
3184 case '=': case '>': case '<': case '|':
3185 break;
3186 default:
3187 return 0;
3188 }
3189 for (cnt = 1; cnt < marker_size; cnt++)
3190 if (line[cnt] != firstchar)
3191 return 0;
3192 /* line[1] through line[marker_size-1] are same as firstchar */
3193 if (len < marker_size + 1 || !isspace(line[marker_size]))
3194 return 0;
3195 return 1;
3196 }
3197
3198 static void checkdiff_consume_hunk(void *priv,
3199 long ob UNUSED, long on UNUSED,
3200 long nb, long nn UNUSED,
3201 const char *func UNUSED, long funclen UNUSED)
3202
3203 {
3204 struct checkdiff_t *data = priv;
3205 data->lineno = nb - 1;
3206 }
3207
3208 static int 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 0;
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 return 0;
3245 }
3246
3247 static unsigned char *deflate_it(char *data,
3248 unsigned long size,
3249 unsigned long *result_size)
3250 {
3251 int bound;
3252 unsigned char *deflated;
3253 git_zstream stream;
3254
3255 git_deflate_init(&stream, zlib_compression_level);
3256 bound = git_deflate_bound(&stream, size);
3257 deflated = xmalloc(bound);
3258 stream.next_out = deflated;
3259 stream.avail_out = bound;
3260
3261 stream.next_in = (unsigned char *)data;
3262 stream.avail_in = size;
3263 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3264 ; /* nothing */
3265 git_deflate_end(&stream);
3266 *result_size = stream.total_out;
3267 return deflated;
3268 }
3269
3270 static void emit_binary_diff_body(struct diff_options *o,
3271 mmfile_t *one, mmfile_t *two)
3272 {
3273 void *cp;
3274 void *delta;
3275 void *deflated;
3276 void *data;
3277 unsigned long orig_size;
3278 unsigned long delta_size;
3279 unsigned long deflate_size;
3280 unsigned long data_size;
3281
3282 /* We could do deflated delta, or we could do just deflated two,
3283 * whichever is smaller.
3284 */
3285 delta = NULL;
3286 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3287 if (one->size && two->size) {
3288 delta = diff_delta(one->ptr, one->size,
3289 two->ptr, two->size,
3290 &delta_size, deflate_size);
3291 if (delta) {
3292 void *to_free = delta;
3293 orig_size = delta_size;
3294 delta = deflate_it(delta, delta_size, &delta_size);
3295 free(to_free);
3296 }
3297 }
3298
3299 if (delta && delta_size < deflate_size) {
3300 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3301 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3302 s, strlen(s), 0);
3303 free(s);
3304 free(deflated);
3305 data = delta;
3306 data_size = delta_size;
3307 } else {
3308 char *s = xstrfmt("%lu", two->size);
3309 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3310 s, strlen(s), 0);
3311 free(s);
3312 free(delta);
3313 data = deflated;
3314 data_size = deflate_size;
3315 }
3316
3317 /* emit data encoded in base85 */
3318 cp = data;
3319 while (data_size) {
3320 int len;
3321 int bytes = (52 < data_size) ? 52 : data_size;
3322 char line[71];
3323 data_size -= bytes;
3324 if (bytes <= 26)
3325 line[0] = bytes + 'A' - 1;
3326 else
3327 line[0] = bytes - 26 + 'a' - 1;
3328 encode_85(line + 1, cp, bytes);
3329 cp = (char *) cp + bytes;
3330
3331 len = strlen(line);
3332 line[len++] = '\n';
3333 line[len] = '\0';
3334
3335 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3336 line, len, 0);
3337 }
3338 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3339 free(data);
3340 }
3341
3342 static void emit_binary_diff(struct diff_options *o,
3343 mmfile_t *one, mmfile_t *two)
3344 {
3345 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3346 emit_binary_diff_body(o, one, two);
3347 emit_binary_diff_body(o, two, one);
3348 }
3349
3350 int diff_filespec_is_binary(struct repository *r,
3351 struct diff_filespec *one)
3352 {
3353 struct diff_populate_filespec_options dpf_options = {
3354 .check_binary = 1,
3355 };
3356
3357 if (one->is_binary == -1) {
3358 diff_filespec_load_driver(one, r->index);
3359 if (one->driver->binary != -1)
3360 one->is_binary = one->driver->binary;
3361 else {
3362 if (!one->data && DIFF_FILE_VALID(one))
3363 diff_populate_filespec(r, one, &dpf_options);
3364 if (one->is_binary == -1 && one->data)
3365 one->is_binary = buffer_is_binary(one->data,
3366 one->size);
3367 if (one->is_binary == -1)
3368 one->is_binary = 0;
3369 }
3370 }
3371 return one->is_binary;
3372 }
3373
3374 static const struct userdiff_funcname *
3375 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3376 {
3377 diff_filespec_load_driver(one, o->repo->index);
3378 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3379 }
3380
3381 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3382 {
3383 if (!options->a_prefix)
3384 options->a_prefix = a;
3385 if (!options->b_prefix)
3386 options->b_prefix = b;
3387 }
3388
3389 struct userdiff_driver *get_textconv(struct repository *r,
3390 struct diff_filespec *one)
3391 {
3392 if (!DIFF_FILE_VALID(one))
3393 return NULL;
3394
3395 diff_filespec_load_driver(one, r->index);
3396 return userdiff_get_textconv(r, one->driver);
3397 }
3398
3399 static struct string_list *additional_headers(struct diff_options *o,
3400 const char *path)
3401 {
3402 if (!o->additional_path_headers)
3403 return NULL;
3404 return strmap_get(o->additional_path_headers, path);
3405 }
3406
3407 static void add_formatted_header(struct strbuf *msg,
3408 const char *header,
3409 const char *line_prefix,
3410 const char *meta,
3411 const char *reset)
3412 {
3413 const char *next, *newline;
3414
3415 for (next = header; *next; next = newline) {
3416 newline = strchrnul(next, '\n');
3417 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3418 (int)(newline - next), next, reset);
3419 if (*newline)
3420 newline++;
3421 }
3422 }
3423
3424 static void add_formatted_headers(struct strbuf *msg,
3425 struct string_list *more_headers,
3426 const char *line_prefix,
3427 const char *meta,
3428 const char *reset)
3429 {
3430 int i;
3431
3432 for (i = 0; i < more_headers->nr; i++)
3433 add_formatted_header(msg, more_headers->items[i].string,
3434 line_prefix, meta, reset);
3435 }
3436
3437 static int diff_filepair_is_phoney(struct diff_filespec *one,
3438 struct diff_filespec *two)
3439 {
3440 /*
3441 * This function specifically looks for pairs injected by
3442 * create_filepairs_for_header_only_notifications(). Such
3443 * pairs are "phoney" in that they do not represent any
3444 * content or even mode difference, but were inserted because
3445 * diff_queued_diff previously had no pair associated with
3446 * that path but we needed some pair to avoid losing the
3447 * "remerge CONFLICT" header associated with the path.
3448 */
3449 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3450 }
3451
3452 static int set_diff_algorithm(struct diff_options *opts,
3453 const char *alg)
3454 {
3455 long value = parse_algorithm_value(alg);
3456
3457 if (value < 0)
3458 return -1;
3459
3460 /* clear out previous settings */
3461 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3462 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3463 opts->xdl_opts |= value;
3464
3465 return 0;
3466 }
3467
3468 static void builtin_diff(const char *name_a,
3469 const char *name_b,
3470 struct diff_filespec *one,
3471 struct diff_filespec *two,
3472 const char *xfrm_msg,
3473 int must_show_header,
3474 struct diff_options *o,
3475 int complete_rewrite)
3476 {
3477 mmfile_t mf1, mf2;
3478 const char *lbl[2];
3479 char *a_one, *b_two;
3480 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3481 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3482 const char *a_prefix, *b_prefix;
3483 struct userdiff_driver *textconv_one = NULL;
3484 struct userdiff_driver *textconv_two = NULL;
3485 struct strbuf header = STRBUF_INIT;
3486 const char *line_prefix = diff_line_prefix(o);
3487
3488 diff_set_mnemonic_prefix(o, "a/", "b/");
3489 if (o->flags.reverse_diff) {
3490 a_prefix = o->b_prefix;
3491 b_prefix = o->a_prefix;
3492 } else {
3493 a_prefix = o->a_prefix;
3494 b_prefix = o->b_prefix;
3495 }
3496
3497 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3498 (!one->mode || S_ISGITLINK(one->mode)) &&
3499 (!two->mode || S_ISGITLINK(two->mode)) &&
3500 (!diff_filepair_is_phoney(one, two))) {
3501 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3502 &one->oid, &two->oid,
3503 two->dirty_submodule);
3504 return;
3505 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3506 (!one->mode || S_ISGITLINK(one->mode)) &&
3507 (!two->mode || S_ISGITLINK(two->mode)) &&
3508 (!diff_filepair_is_phoney(one, two))) {
3509 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3510 &one->oid, &two->oid,
3511 two->dirty_submodule);
3512 return;
3513 }
3514
3515 if (o->flags.allow_textconv) {
3516 textconv_one = get_textconv(o->repo, one);
3517 textconv_two = get_textconv(o->repo, two);
3518 }
3519
3520 /* Never use a non-valid filename anywhere if at all possible */
3521 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3522 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3523
3524 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3525 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3526 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3527 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3528 if (diff_filepair_is_phoney(one, two)) {
3529 /*
3530 * We should only reach this point for pairs generated from
3531 * create_filepairs_for_header_only_notifications(). For
3532 * these, we want to avoid the "/dev/null" special casing
3533 * above, because we do not want such pairs shown as either
3534 * "new file" or "deleted file" below.
3535 */
3536 lbl[0] = a_one;
3537 lbl[1] = b_two;
3538 }
3539 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3540 if (lbl[0][0] == '/') {
3541 /* /dev/null */
3542 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3543 if (xfrm_msg)
3544 strbuf_addstr(&header, xfrm_msg);
3545 must_show_header = 1;
3546 }
3547 else if (lbl[1][0] == '/') {
3548 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3549 if (xfrm_msg)
3550 strbuf_addstr(&header, xfrm_msg);
3551 must_show_header = 1;
3552 }
3553 else {
3554 if (one->mode != two->mode) {
3555 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3556 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3557 must_show_header = 1;
3558 }
3559 if (xfrm_msg)
3560 strbuf_addstr(&header, xfrm_msg);
3561
3562 /*
3563 * we do not run diff between different kind
3564 * of objects.
3565 */
3566 if ((one->mode ^ two->mode) & S_IFMT)
3567 goto free_ab_and_return;
3568 if (complete_rewrite &&
3569 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3570 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3571 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3572 header.buf, header.len, 0);
3573 strbuf_reset(&header);
3574 emit_rewrite_diff(name_a, name_b, one, two,
3575 textconv_one, textconv_two, o);
3576 o->found_changes = 1;
3577 goto free_ab_and_return;
3578 }
3579 }
3580
3581 if (o->irreversible_delete && lbl[1][0] == '/') {
3582 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3583 header.len, 0);
3584 strbuf_reset(&header);
3585 goto free_ab_and_return;
3586 } else if (!o->flags.text &&
3587 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3588 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3589 struct strbuf sb = STRBUF_INIT;
3590 if (!one->data && !two->data &&
3591 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3592 !o->flags.binary) {
3593 if (oideq(&one->oid, &two->oid)) {
3594 if (must_show_header)
3595 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3596 header.buf, header.len,
3597 0);
3598 goto free_ab_and_return;
3599 }
3600 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3601 header.buf, header.len, 0);
3602 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3603 diff_line_prefix(o), lbl[0], lbl[1]);
3604 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3605 sb.buf, sb.len, 0);
3606 strbuf_release(&sb);
3607 goto free_ab_and_return;
3608 }
3609 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3610 fill_mmfile(o->repo, &mf2, two) < 0)
3611 die("unable to read files to diff");
3612 /* Quite common confusing case */
3613 if (mf1.size == mf2.size &&
3614 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3615 if (must_show_header)
3616 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3617 header.buf, header.len, 0);
3618 goto free_ab_and_return;
3619 }
3620 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3621 strbuf_reset(&header);
3622 if (o->flags.binary)
3623 emit_binary_diff(o, &mf1, &mf2);
3624 else {
3625 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3626 diff_line_prefix(o), lbl[0], lbl[1]);
3627 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3628 sb.buf, sb.len, 0);
3629 strbuf_release(&sb);
3630 }
3631 o->found_changes = 1;
3632 } else {
3633 /* Crazy xdl interfaces.. */
3634 const char *diffopts;
3635 const char *v;
3636 xpparam_t xpp;
3637 xdemitconf_t xecfg;
3638 struct emit_callback ecbdata;
3639 const struct userdiff_funcname *pe;
3640
3641 if (must_show_header) {
3642 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3643 header.buf, header.len, 0);
3644 strbuf_reset(&header);
3645 }
3646
3647 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3648 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3649
3650 pe = diff_funcname_pattern(o, one);
3651 if (!pe)
3652 pe = diff_funcname_pattern(o, two);
3653
3654 memset(&xpp, 0, sizeof(xpp));
3655 memset(&xecfg, 0, sizeof(xecfg));
3656 memset(&ecbdata, 0, sizeof(ecbdata));
3657 if (o->flags.suppress_diff_headers)
3658 lbl[0] = NULL;
3659 ecbdata.label_path = lbl;
3660 ecbdata.color_diff = want_color(o->use_color);
3661 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3662 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3663 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3664 ecbdata.opt = o;
3665 if (header.len && !o->flags.suppress_diff_headers)
3666 ecbdata.header = &header;
3667 xpp.flags = o->xdl_opts;
3668 xpp.ignore_regex = o->ignore_regex;
3669 xpp.ignore_regex_nr = o->ignore_regex_nr;
3670 xpp.anchors = o->anchors;
3671 xpp.anchors_nr = o->anchors_nr;
3672 xecfg.ctxlen = o->context;
3673 xecfg.interhunkctxlen = o->interhunkcontext;
3674 xecfg.flags = XDL_EMIT_FUNCNAMES;
3675 if (o->flags.funccontext)
3676 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3677 if (pe)
3678 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3679
3680 diffopts = getenv("GIT_DIFF_OPTS");
3681 if (!diffopts)
3682 ;
3683 else if (skip_prefix(diffopts, "--unified=", &v))
3684 xecfg.ctxlen = strtoul(v, NULL, 10);
3685 else if (skip_prefix(diffopts, "-u", &v))
3686 xecfg.ctxlen = strtoul(v, NULL, 10);
3687
3688 if (o->word_diff)
3689 init_diff_words_data(&ecbdata, o, one, two);
3690 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3691 &ecbdata, &xpp, &xecfg))
3692 die("unable to generate diff for %s", one->path);
3693 if (o->word_diff)
3694 free_diff_words_data(&ecbdata);
3695 if (textconv_one)
3696 free(mf1.ptr);
3697 if (textconv_two)
3698 free(mf2.ptr);
3699 xdiff_clear_find_func(&xecfg);
3700 }
3701
3702 free_ab_and_return:
3703 strbuf_release(&header);
3704 diff_free_filespec_data(one);
3705 diff_free_filespec_data(two);
3706 free(a_one);
3707 free(b_two);
3708 return;
3709 }
3710
3711 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3712 {
3713 if (!is_renamed) {
3714 if (p->status == DIFF_STATUS_ADDED) {
3715 if (S_ISLNK(p->two->mode))
3716 return "new +l";
3717 else if ((p->two->mode & 0777) == 0755)
3718 return "new +x";
3719 else
3720 return "new";
3721 } else if (p->status == DIFF_STATUS_DELETED)
3722 return "gone";
3723 }
3724 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3725 return "mode -l";
3726 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3727 return "mode +l";
3728 else if ((p->one->mode & 0777) == 0644 &&
3729 (p->two->mode & 0777) == 0755)
3730 return "mode +x";
3731 else if ((p->one->mode & 0777) == 0755 &&
3732 (p->two->mode & 0777) == 0644)
3733 return "mode -x";
3734 return NULL;
3735 }
3736
3737 static void builtin_diffstat(const char *name_a, const char *name_b,
3738 struct diff_filespec *one,
3739 struct diff_filespec *two,
3740 struct diffstat_t *diffstat,
3741 struct diff_options *o,
3742 struct diff_filepair *p)
3743 {
3744 mmfile_t mf1, mf2;
3745 struct diffstat_file *data;
3746 int may_differ;
3747 int complete_rewrite = 0;
3748
3749 if (!DIFF_PAIR_UNMERGED(p)) {
3750 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3751 complete_rewrite = 1;
3752 }
3753
3754 data = diffstat_add(diffstat, name_a, name_b);
3755 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3756 if (o->flags.stat_with_summary)
3757 data->comments = get_compact_summary(p, data->is_renamed);
3758
3759 if (!one || !two) {
3760 data->is_unmerged = 1;
3761 return;
3762 }
3763
3764 /* saves some reads if true, not a guarantee of diff outcome */
3765 may_differ = !(one->oid_valid && two->oid_valid &&
3766 oideq(&one->oid, &two->oid));
3767
3768 if (diff_filespec_is_binary(o->repo, one) ||
3769 diff_filespec_is_binary(o->repo, two)) {
3770 data->is_binary = 1;
3771 if (!may_differ) {
3772 data->added = 0;
3773 data->deleted = 0;
3774 } else {
3775 data->added = diff_filespec_size(o->repo, two);
3776 data->deleted = diff_filespec_size(o->repo, one);
3777 }
3778 }
3779
3780 else if (complete_rewrite) {
3781 diff_populate_filespec(o->repo, one, NULL);
3782 diff_populate_filespec(o->repo, two, NULL);
3783 data->deleted = count_lines(one->data, one->size);
3784 data->added = count_lines(two->data, two->size);
3785 }
3786
3787 else if (may_differ) {
3788 /* Crazy xdl interfaces.. */
3789 xpparam_t xpp;
3790 xdemitconf_t xecfg;
3791
3792 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3793 fill_mmfile(o->repo, &mf2, two) < 0)
3794 die("unable to read files to diff");
3795
3796 memset(&xpp, 0, sizeof(xpp));
3797 memset(&xecfg, 0, sizeof(xecfg));
3798 xpp.flags = o->xdl_opts;
3799 xpp.ignore_regex = o->ignore_regex;
3800 xpp.ignore_regex_nr = o->ignore_regex_nr;
3801 xpp.anchors = o->anchors;
3802 xpp.anchors_nr = o->anchors_nr;
3803 xecfg.ctxlen = o->context;
3804 xecfg.interhunkctxlen = o->interhunkcontext;
3805 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3806 if (xdi_diff_outf(&mf1, &mf2, NULL,
3807 diffstat_consume, diffstat, &xpp, &xecfg))
3808 die("unable to generate diffstat for %s", one->path);
3809
3810 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3811 struct diffstat_file *file =
3812 diffstat->files[diffstat->nr - 1];
3813 /*
3814 * Omit diffstats of modified files where nothing changed.
3815 * Even if may_differ, this might be the case due to
3816 * ignoring whitespace changes, etc.
3817 *
3818 * But note that we special-case additions, deletions,
3819 * renames, and mode changes as adding an empty file,
3820 * for example is still of interest.
3821 */
3822 if ((p->status == DIFF_STATUS_MODIFIED)
3823 && !file->added
3824 && !file->deleted
3825 && one->mode == two->mode) {
3826 free_diffstat_file(file);
3827 diffstat->nr--;
3828 }
3829 }
3830 }
3831
3832 diff_free_filespec_data(one);
3833 diff_free_filespec_data(two);
3834 }
3835
3836 static void builtin_checkdiff(const char *name_a, const char *name_b,
3837 const char *attr_path,
3838 struct diff_filespec *one,
3839 struct diff_filespec *two,
3840 struct diff_options *o)
3841 {
3842 mmfile_t mf1, mf2;
3843 struct checkdiff_t data;
3844
3845 if (!two)
3846 return;
3847
3848 memset(&data, 0, sizeof(data));
3849 data.filename = name_b ? name_b : name_a;
3850 data.lineno = 0;
3851 data.o = o;
3852 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3853 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3854
3855 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3856 fill_mmfile(o->repo, &mf2, two) < 0)
3857 die("unable to read files to diff");
3858
3859 /*
3860 * All the other codepaths check both sides, but not checking
3861 * the "old" side here is deliberate. We are checking the newly
3862 * introduced changes, and as long as the "new" side is text, we
3863 * can and should check what it introduces.
3864 */
3865 if (diff_filespec_is_binary(o->repo, two))
3866 goto free_and_return;
3867 else {
3868 /* Crazy xdl interfaces.. */
3869 xpparam_t xpp;
3870 xdemitconf_t xecfg;
3871
3872 memset(&xpp, 0, sizeof(xpp));
3873 memset(&xecfg, 0, sizeof(xecfg));
3874 xecfg.ctxlen = 1; /* at least one context line */
3875 xpp.flags = 0;
3876 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3877 checkdiff_consume, &data,
3878 &xpp, &xecfg))
3879 die("unable to generate checkdiff for %s", one->path);
3880
3881 if (data.ws_rule & WS_BLANK_AT_EOF) {
3882 struct emit_callback ecbdata;
3883 int blank_at_eof;
3884
3885 ecbdata.ws_rule = data.ws_rule;
3886 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3887 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3888
3889 if (blank_at_eof) {
3890 static char *err;
3891 if (!err)
3892 err = whitespace_error_string(WS_BLANK_AT_EOF);
3893 fprintf(o->file, "%s:%d: %s.\n",
3894 data.filename, blank_at_eof, err);
3895 data.status = 1; /* report errors */
3896 }
3897 }
3898 }
3899 free_and_return:
3900 diff_free_filespec_data(one);
3901 diff_free_filespec_data(two);
3902 if (data.status)
3903 o->flags.check_failed = 1;
3904 }
3905
3906 struct diff_filespec *alloc_filespec(const char *path)
3907 {
3908 struct diff_filespec *spec;
3909
3910 FLEXPTR_ALLOC_STR(spec, path, path);
3911 spec->count = 1;
3912 spec->is_binary = -1;
3913 return spec;
3914 }
3915
3916 void free_filespec(struct diff_filespec *spec)
3917 {
3918 if (!--spec->count) {
3919 diff_free_filespec_data(spec);
3920 free(spec);
3921 }
3922 }
3923
3924 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3925 int oid_valid, unsigned short mode)
3926 {
3927 if (mode) {
3928 spec->mode = canon_mode(mode);
3929 oidcpy(&spec->oid, oid);
3930 spec->oid_valid = oid_valid;
3931 }
3932 }
3933
3934 /*
3935 * Given a name and sha1 pair, if the index tells us the file in
3936 * the work tree has that object contents, return true, so that
3937 * prepare_temp_file() does not have to inflate and extract.
3938 */
3939 static int reuse_worktree_file(struct index_state *istate,
3940 const char *name,
3941 const struct object_id *oid,
3942 int want_file)
3943 {
3944 const struct cache_entry *ce;
3945 struct stat st;
3946 int pos, len;
3947
3948 /*
3949 * We do not read the cache ourselves here, because the
3950 * benchmark with my previous version that always reads cache
3951 * shows that it makes things worse for diff-tree comparing
3952 * two linux-2.6 kernel trees in an already checked out work
3953 * tree. This is because most diff-tree comparisons deal with
3954 * only a small number of files, while reading the cache is
3955 * expensive for a large project, and its cost outweighs the
3956 * savings we get by not inflating the object to a temporary
3957 * file. Practically, this code only helps when we are used
3958 * by diff-cache --cached, which does read the cache before
3959 * calling us.
3960 */
3961 if (!istate->cache)
3962 return 0;
3963
3964 /* We want to avoid the working directory if our caller
3965 * doesn't need the data in a normal file, this system
3966 * is rather slow with its stat/open/mmap/close syscalls,
3967 * and the object is contained in a pack file. The pack
3968 * is probably already open and will be faster to obtain
3969 * the data through than the working directory. Loose
3970 * objects however would tend to be slower as they need
3971 * to be individually opened and inflated.
3972 */
3973 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3974 return 0;
3975
3976 /*
3977 * Similarly, if we'd have to convert the file contents anyway, that
3978 * makes the optimization not worthwhile.
3979 */
3980 if (!want_file && would_convert_to_git(istate, name))
3981 return 0;
3982
3983 /*
3984 * If this path does not match our sparse-checkout definition,
3985 * then the file will not be in the working directory.
3986 */
3987 if (!path_in_sparse_checkout(name, istate))
3988 return 0;
3989
3990 len = strlen(name);
3991 pos = index_name_pos(istate, name, len);
3992 if (pos < 0)
3993 return 0;
3994 ce = istate->cache[pos];
3995
3996 /*
3997 * This is not the sha1 we are looking for, or
3998 * unreusable because it is not a regular file.
3999 */
4000 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4001 return 0;
4002
4003 /*
4004 * If ce is marked as "assume unchanged", there is no
4005 * guarantee that work tree matches what we are looking for.
4006 */
4007 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4008 return 0;
4009
4010 /*
4011 * If ce matches the file in the work tree, we can reuse it.
4012 */
4013 if (ce_uptodate(ce) ||
4014 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4015 return 1;
4016
4017 return 0;
4018 }
4019
4020 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4021 {
4022 struct strbuf buf = STRBUF_INIT;
4023 char *dirty = "";
4024
4025 /* Are we looking at the work tree? */
4026 if (s->dirty_submodule)
4027 dirty = "-dirty";
4028
4029 strbuf_addf(&buf, "Subproject commit %s%s\n",
4030 oid_to_hex(&s->oid), dirty);
4031 s->size = buf.len;
4032 if (size_only) {
4033 s->data = NULL;
4034 strbuf_release(&buf);
4035 } else {
4036 s->data = strbuf_detach(&buf, NULL);
4037 s->should_free = 1;
4038 }
4039 return 0;
4040 }
4041
4042 /*
4043 * While doing rename detection and pickaxe operation, we may need to
4044 * grab the data for the blob (or file) for our own in-core comparison.
4045 * diff_filespec has data and size fields for this purpose.
4046 */
4047 int diff_populate_filespec(struct repository *r,
4048 struct diff_filespec *s,
4049 const struct diff_populate_filespec_options *options)
4050 {
4051 int size_only = options ? options->check_size_only : 0;
4052 int check_binary = options ? options->check_binary : 0;
4053 int err = 0;
4054 int conv_flags = global_conv_flags_eol;
4055 /*
4056 * demote FAIL to WARN to allow inspecting the situation
4057 * instead of refusing.
4058 */
4059 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4060 conv_flags = CONV_EOL_RNDTRP_WARN;
4061
4062 if (!DIFF_FILE_VALID(s))
4063 die("internal error: asking to populate invalid file.");
4064 if (S_ISDIR(s->mode))
4065 return -1;
4066
4067 if (s->data)
4068 return 0;
4069
4070 if (size_only && 0 < s->size)
4071 return 0;
4072
4073 if (S_ISGITLINK(s->mode))
4074 return diff_populate_gitlink(s, size_only);
4075
4076 if (!s->oid_valid ||
4077 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4078 struct strbuf buf = STRBUF_INIT;
4079 struct stat st;
4080 int fd;
4081
4082 if (lstat(s->path, &st) < 0) {
4083 err_empty:
4084 err = -1;
4085 empty:
4086 s->data = (char *)"";
4087 s->size = 0;
4088 return err;
4089 }
4090 s->size = xsize_t(st.st_size);
4091 if (!s->size)
4092 goto empty;
4093 if (S_ISLNK(st.st_mode)) {
4094 struct strbuf sb = STRBUF_INIT;
4095
4096 if (strbuf_readlink(&sb, s->path, s->size))
4097 goto err_empty;
4098 s->size = sb.len;
4099 s->data = strbuf_detach(&sb, NULL);
4100 s->should_free = 1;
4101 return 0;
4102 }
4103
4104 /*
4105 * Even if the caller would be happy with getting
4106 * only the size, we cannot return early at this
4107 * point if the path requires us to run the content
4108 * conversion.
4109 */
4110 if (size_only && !would_convert_to_git(r->index, s->path))
4111 return 0;
4112
4113 /*
4114 * Note: this check uses xsize_t(st.st_size) that may
4115 * not be the true size of the blob after it goes
4116 * through convert_to_git(). This may not strictly be
4117 * correct, but the whole point of big_file_threshold
4118 * and is_binary check being that we want to avoid
4119 * opening the file and inspecting the contents, this
4120 * is probably fine.
4121 */
4122 if (check_binary &&
4123 s->size > big_file_threshold && s->is_binary == -1) {
4124 s->is_binary = 1;
4125 return 0;
4126 }
4127 fd = open(s->path, O_RDONLY);
4128 if (fd < 0)
4129 goto err_empty;
4130 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4131 close(fd);
4132 s->should_munmap = 1;
4133
4134 /*
4135 * Convert from working tree format to canonical git format
4136 */
4137 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4138 size_t size = 0;
4139 munmap(s->data, s->size);
4140 s->should_munmap = 0;
4141 s->data = strbuf_detach(&buf, &size);
4142 s->size = size;
4143 s->should_free = 1;
4144 }
4145 }
4146 else {
4147 struct object_info info = {
4148 .sizep = &s->size
4149 };
4150
4151 if (!(size_only || check_binary))
4152 /*
4153 * Set contentp, since there is no chance that merely
4154 * the size is sufficient.
4155 */
4156 info.contentp = &s->data;
4157
4158 if (options && options->missing_object_cb) {
4159 if (!oid_object_info_extended(r, &s->oid, &info,
4160 OBJECT_INFO_LOOKUP_REPLACE |
4161 OBJECT_INFO_SKIP_FETCH_OBJECT))
4162 goto object_read;
4163 options->missing_object_cb(options->missing_object_data);
4164 }
4165 if (oid_object_info_extended(r, &s->oid, &info,
4166 OBJECT_INFO_LOOKUP_REPLACE))
4167 die("unable to read %s", oid_to_hex(&s->oid));
4168
4169 object_read:
4170 if (size_only || check_binary) {
4171 if (size_only)
4172 return 0;
4173 if (s->size > big_file_threshold && s->is_binary == -1) {
4174 s->is_binary = 1;
4175 return 0;
4176 }
4177 }
4178 if (!info.contentp) {
4179 info.contentp = &s->data;
4180 if (oid_object_info_extended(r, &s->oid, &info,
4181 OBJECT_INFO_LOOKUP_REPLACE))
4182 die("unable to read %s", oid_to_hex(&s->oid));
4183 }
4184 s->should_free = 1;
4185 }
4186 return 0;
4187 }
4188
4189 void diff_free_filespec_blob(struct diff_filespec *s)
4190 {
4191 if (s->should_free)
4192 free(s->data);
4193 else if (s->should_munmap)
4194 munmap(s->data, s->size);
4195
4196 if (s->should_free || s->should_munmap) {
4197 s->should_free = s->should_munmap = 0;
4198 s->data = NULL;
4199 }
4200 }
4201
4202 void diff_free_filespec_data(struct diff_filespec *s)
4203 {
4204 if (!s)
4205 return;
4206
4207 diff_free_filespec_blob(s);
4208 FREE_AND_NULL(s->cnt_data);
4209 }
4210
4211 static void prep_temp_blob(struct index_state *istate,
4212 const char *path, struct diff_tempfile *temp,
4213 void *blob,
4214 unsigned long size,
4215 const struct object_id *oid,
4216 int mode)
4217 {
4218 struct strbuf buf = STRBUF_INIT;
4219 char *path_dup = xstrdup(path);
4220 const char *base = basename(path_dup);
4221 struct checkout_metadata meta;
4222
4223 init_checkout_metadata(&meta, NULL, NULL, oid);
4224
4225 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4226 if (!temp->tempfile)
4227 die_errno("unable to create temp-file");
4228 if (convert_to_working_tree(istate, path,
4229 (const char *)blob, (size_t)size, &buf, &meta)) {
4230 blob = buf.buf;
4231 size = buf.len;
4232 }
4233 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4234 close_tempfile_gently(temp->tempfile))
4235 die_errno("unable to write temp-file");
4236 temp->name = get_tempfile_path(temp->tempfile);
4237 oid_to_hex_r(temp->hex, oid);
4238 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4239 strbuf_release(&buf);
4240 free(path_dup);
4241 }
4242
4243 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4244 struct diff_filespec *one)
4245 {
4246 struct diff_tempfile *temp = claim_diff_tempfile();
4247
4248 if (!DIFF_FILE_VALID(one)) {
4249 not_a_valid_file:
4250 /* A '-' entry produces this for file-2, and
4251 * a '+' entry produces this for file-1.
4252 */
4253 temp->name = "/dev/null";
4254 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4255 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4256 return temp;
4257 }
4258
4259 if (!S_ISGITLINK(one->mode) &&
4260 (!one->oid_valid ||
4261 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4262 struct stat st;
4263 if (lstat(one->path, &st) < 0) {
4264 if (errno == ENOENT)
4265 goto not_a_valid_file;
4266 die_errno("stat(%s)", one->path);
4267 }
4268 if (S_ISLNK(st.st_mode)) {
4269 struct strbuf sb = STRBUF_INIT;
4270 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4271 die_errno("readlink(%s)", one->path);
4272 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4273 (one->oid_valid ?
4274 &one->oid : null_oid()),
4275 (one->oid_valid ?
4276 one->mode : S_IFLNK));
4277 strbuf_release(&sb);
4278 }
4279 else {
4280 /* we can borrow from the file in the work tree */
4281 temp->name = one->path;
4282 if (!one->oid_valid)
4283 oid_to_hex_r(temp->hex, null_oid());
4284 else
4285 oid_to_hex_r(temp->hex, &one->oid);
4286 /* Even though we may sometimes borrow the
4287 * contents from the work tree, we always want
4288 * one->mode. mode is trustworthy even when
4289 * !(one->oid_valid), as long as
4290 * DIFF_FILE_VALID(one).
4291 */
4292 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4293 }
4294 return temp;
4295 }
4296 else {
4297 if (diff_populate_filespec(r, one, NULL))
4298 die("cannot read data blob for %s", one->path);
4299 prep_temp_blob(r->index, one->path, temp,
4300 one->data, one->size,
4301 &one->oid, one->mode);
4302 }
4303 return temp;
4304 }
4305
4306 static void add_external_diff_name(struct repository *r,
4307 struct strvec *argv,
4308 struct diff_filespec *df)
4309 {
4310 struct diff_tempfile *temp = prepare_temp_file(r, df);
4311 strvec_push(argv, temp->name);
4312 strvec_push(argv, temp->hex);
4313 strvec_push(argv, temp->mode);
4314 }
4315
4316 /* An external diff command takes:
4317 *
4318 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4319 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4320 *
4321 */
4322 static void run_external_diff(const char *pgm,
4323 const char *name,
4324 const char *other,
4325 struct diff_filespec *one,
4326 struct diff_filespec *two,
4327 const char *xfrm_msg,
4328 struct diff_options *o)
4329 {
4330 struct child_process cmd = CHILD_PROCESS_INIT;
4331 struct diff_queue_struct *q = &diff_queued_diff;
4332
4333 strvec_push(&cmd.args, pgm);
4334 strvec_push(&cmd.args, name);
4335
4336 if (one && two) {
4337 add_external_diff_name(o->repo, &cmd.args, one);
4338 add_external_diff_name(o->repo, &cmd.args, two);
4339 if (other) {
4340 strvec_push(&cmd.args, other);
4341 strvec_push(&cmd.args, xfrm_msg);
4342 }
4343 }
4344
4345 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4346 ++o->diff_path_counter);
4347 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4348
4349 diff_free_filespec_data(one);
4350 diff_free_filespec_data(two);
4351 cmd.use_shell = 1;
4352 if (run_command(&cmd))
4353 die(_("external diff died, stopping at %s"), name);
4354
4355 remove_tempfile();
4356 }
4357
4358 static int similarity_index(struct diff_filepair *p)
4359 {
4360 return p->score * 100 / MAX_SCORE;
4361 }
4362
4363 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4364 {
4365 if (startup_info->have_repository)
4366 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4367 else {
4368 char *hex = oid_to_hex(oid);
4369 if (abbrev < 0)
4370 abbrev = FALLBACK_DEFAULT_ABBREV;
4371 if (abbrev > the_hash_algo->hexsz)
4372 BUG("oid abbreviation out of range: %d", abbrev);
4373 if (abbrev)
4374 hex[abbrev] = '\0';
4375 return hex;
4376 }
4377 }
4378
4379 static void fill_metainfo(struct strbuf *msg,
4380 const char *name,
4381 const char *other,
4382 struct diff_filespec *one,
4383 struct diff_filespec *two,
4384 struct diff_options *o,
4385 struct diff_filepair *p,
4386 int *must_show_header,
4387 int use_color)
4388 {
4389 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4390 const char *reset = diff_get_color(use_color, DIFF_RESET);
4391 const char *line_prefix = diff_line_prefix(o);
4392 struct string_list *more_headers = NULL;
4393
4394 *must_show_header = 1;
4395 strbuf_init(msg, PATH_MAX * 2 + 300);
4396 switch (p->status) {
4397 case DIFF_STATUS_COPIED:
4398 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4399 line_prefix, set, similarity_index(p));
4400 strbuf_addf(msg, "%s\n%s%scopy from ",
4401 reset, line_prefix, set);
4402 quote_c_style(name, msg, NULL, 0);
4403 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4404 quote_c_style(other, msg, NULL, 0);
4405 strbuf_addf(msg, "%s\n", reset);
4406 break;
4407 case DIFF_STATUS_RENAMED:
4408 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4409 line_prefix, set, similarity_index(p));
4410 strbuf_addf(msg, "%s\n%s%srename from ",
4411 reset, line_prefix, set);
4412 quote_c_style(name, msg, NULL, 0);
4413 strbuf_addf(msg, "%s\n%s%srename to ",
4414 reset, line_prefix, set);
4415 quote_c_style(other, msg, NULL, 0);
4416 strbuf_addf(msg, "%s\n", reset);
4417 break;
4418 case DIFF_STATUS_MODIFIED:
4419 if (p->score) {
4420 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4421 line_prefix,
4422 set, similarity_index(p), reset);
4423 break;
4424 }
4425 /* fallthru */
4426 default:
4427 *must_show_header = 0;
4428 }
4429 if ((more_headers = additional_headers(o, name))) {
4430 add_formatted_headers(msg, more_headers,
4431 line_prefix, set, reset);
4432 *must_show_header = 1;
4433 }
4434 if (one && two && !oideq(&one->oid, &two->oid)) {
4435 const unsigned hexsz = the_hash_algo->hexsz;
4436 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4437
4438 if (o->flags.full_index)
4439 abbrev = hexsz;
4440
4441 if (o->flags.binary) {
4442 mmfile_t mf;
4443 if ((!fill_mmfile(o->repo, &mf, one) &&
4444 diff_filespec_is_binary(o->repo, one)) ||
4445 (!fill_mmfile(o->repo, &mf, two) &&
4446 diff_filespec_is_binary(o->repo, two)))
4447 abbrev = hexsz;
4448 }
4449 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4450 diff_abbrev_oid(&one->oid, abbrev),
4451 diff_abbrev_oid(&two->oid, abbrev));
4452 if (one->mode == two->mode)
4453 strbuf_addf(msg, " %06o", one->mode);
4454 strbuf_addf(msg, "%s\n", reset);
4455 }
4456 }
4457
4458 static void run_diff_cmd(const char *pgm,
4459 const char *name,
4460 const char *other,
4461 const char *attr_path,
4462 struct diff_filespec *one,
4463 struct diff_filespec *two,
4464 struct strbuf *msg,
4465 struct diff_options *o,
4466 struct diff_filepair *p)
4467 {
4468 const char *xfrm_msg = NULL;
4469 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4470 int must_show_header = 0;
4471 struct userdiff_driver *drv = NULL;
4472
4473 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4474 drv = userdiff_find_by_path(o->repo->index, attr_path);
4475
4476 if (o->flags.allow_external && drv && drv->external)
4477 pgm = drv->external;
4478
4479 if (msg) {
4480 /*
4481 * don't use colors when the header is intended for an
4482 * external diff driver
4483 */
4484 fill_metainfo(msg, name, other, one, two, o, p,
4485 &must_show_header,
4486 want_color(o->use_color) && !pgm);
4487 xfrm_msg = msg->len ? msg->buf : NULL;
4488 }
4489
4490 if (pgm) {
4491 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4492 return;
4493 }
4494 if (one && two) {
4495 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4496 set_diff_algorithm(o, drv->algorithm);
4497
4498 builtin_diff(name, other ? other : name,
4499 one, two, xfrm_msg, must_show_header,
4500 o, complete_rewrite);
4501 } else {
4502 fprintf(o->file, "* Unmerged path %s\n", name);
4503 }
4504 }
4505
4506 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4507 {
4508 if (DIFF_FILE_VALID(one)) {
4509 if (!one->oid_valid) {
4510 struct stat st;
4511 if (one->is_stdin) {
4512 oidclr(&one->oid);
4513 return;
4514 }
4515 if (lstat(one->path, &st) < 0)
4516 die_errno("stat '%s'", one->path);
4517 if (index_path(istate, &one->oid, one->path, &st, 0))
4518 die("cannot hash %s", one->path);
4519 }
4520 }
4521 else
4522 oidclr(&one->oid);
4523 }
4524
4525 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4526 {
4527 /* Strip the prefix but do not molest /dev/null and absolute paths */
4528 if (*namep && !is_absolute_path(*namep)) {
4529 *namep += prefix_length;
4530 if (**namep == '/')
4531 ++*namep;
4532 }
4533 if (*otherp && !is_absolute_path(*otherp)) {
4534 *otherp += prefix_length;
4535 if (**otherp == '/')
4536 ++*otherp;
4537 }
4538 }
4539
4540 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4541 {
4542 const char *pgm = external_diff();
4543 struct strbuf msg;
4544 struct diff_filespec *one = p->one;
4545 struct diff_filespec *two = p->two;
4546 const char *name;
4547 const char *other;
4548 const char *attr_path;
4549
4550 name = one->path;
4551 other = (strcmp(name, two->path) ? two->path : NULL);
4552 attr_path = name;
4553 if (o->prefix_length)
4554 strip_prefix(o->prefix_length, &name, &other);
4555
4556 if (!o->flags.allow_external)
4557 pgm = NULL;
4558
4559 if (DIFF_PAIR_UNMERGED(p)) {
4560 run_diff_cmd(pgm, name, NULL, attr_path,
4561 NULL, NULL, NULL, o, p);
4562 return;
4563 }
4564
4565 diff_fill_oid_info(one, o->repo->index);
4566 diff_fill_oid_info(two, o->repo->index);
4567
4568 if (!pgm &&
4569 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4570 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4571 /*
4572 * a filepair that changes between file and symlink
4573 * needs to be split into deletion and creation.
4574 */
4575 struct diff_filespec *null = alloc_filespec(two->path);
4576 run_diff_cmd(NULL, name, other, attr_path,
4577 one, null, &msg,
4578 o, p);
4579 free(null);
4580 strbuf_release(&msg);
4581
4582 null = alloc_filespec(one->path);
4583 run_diff_cmd(NULL, name, other, attr_path,
4584 null, two, &msg, o, p);
4585 free(null);
4586 }
4587 else
4588 run_diff_cmd(pgm, name, other, attr_path,
4589 one, two, &msg, o, p);
4590
4591 strbuf_release(&msg);
4592 }
4593
4594 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4595 struct diffstat_t *diffstat)
4596 {
4597 const char *name;
4598 const char *other;
4599
4600 if (!o->ignore_driver_algorithm) {
4601 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4602 p->one->path);
4603
4604 if (drv && drv->algorithm)
4605 set_diff_algorithm(o, drv->algorithm);
4606 }
4607
4608 if (DIFF_PAIR_UNMERGED(p)) {
4609 /* unmerged */
4610 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4611 diffstat, o, p);
4612 return;
4613 }
4614
4615 name = p->one->path;
4616 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4617
4618 if (o->prefix_length)
4619 strip_prefix(o->prefix_length, &name, &other);
4620
4621 diff_fill_oid_info(p->one, o->repo->index);
4622 diff_fill_oid_info(p->two, o->repo->index);
4623
4624 builtin_diffstat(name, other, p->one, p->two,
4625 diffstat, o, p);
4626 }
4627
4628 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4629 {
4630 const char *name;
4631 const char *other;
4632 const char *attr_path;
4633
4634 if (DIFF_PAIR_UNMERGED(p)) {
4635 /* unmerged */
4636 return;
4637 }
4638
4639 name = p->one->path;
4640 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4641 attr_path = other ? other : name;
4642
4643 if (o->prefix_length)
4644 strip_prefix(o->prefix_length, &name, &other);
4645
4646 diff_fill_oid_info(p->one, o->repo->index);
4647 diff_fill_oid_info(p->two, o->repo->index);
4648
4649 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4650 }
4651
4652 void repo_diff_setup(struct repository *r, struct diff_options *options)
4653 {
4654 memcpy(options, &default_diff_options, sizeof(*options));
4655
4656 options->file = stdout;
4657 options->repo = r;
4658
4659 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4660 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4661 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4662 options->abbrev = DEFAULT_ABBREV;
4663 options->line_termination = '\n';
4664 options->break_opt = -1;
4665 options->rename_limit = -1;
4666 options->dirstat_permille = diff_dirstat_permille_default;
4667 options->context = diff_context_default;
4668 options->interhunkcontext = diff_interhunk_context_default;
4669 options->ws_error_highlight = ws_error_highlight_default;
4670 options->flags.rename_empty = 1;
4671 options->flags.relative_name = diff_relative;
4672 options->objfind = NULL;
4673
4674 /* pathchange left =NULL by default */
4675 options->change = diff_change;
4676 options->add_remove = diff_addremove;
4677 options->use_color = diff_use_color_default;
4678 options->detect_rename = diff_detect_rename_default;
4679 options->xdl_opts |= diff_algorithm;
4680 if (diff_indent_heuristic)
4681 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4682
4683 options->orderfile = diff_order_file_cfg;
4684
4685 if (!options->flags.ignore_submodule_set)
4686 options->flags.ignore_untracked_in_submodules = 1;
4687
4688 if (diff_no_prefix) {
4689 options->a_prefix = options->b_prefix = "";
4690 } else if (!diff_mnemonic_prefix) {
4691 options->a_prefix = "a/";
4692 options->b_prefix = "b/";
4693 }
4694
4695 options->color_moved = diff_color_moved_default;
4696 options->color_moved_ws_handling = diff_color_moved_ws_default;
4697 }
4698
4699 static const char diff_status_letters[] = {
4700 DIFF_STATUS_ADDED,
4701 DIFF_STATUS_COPIED,
4702 DIFF_STATUS_DELETED,
4703 DIFF_STATUS_MODIFIED,
4704 DIFF_STATUS_RENAMED,
4705 DIFF_STATUS_TYPE_CHANGED,
4706 DIFF_STATUS_UNKNOWN,
4707 DIFF_STATUS_UNMERGED,
4708 DIFF_STATUS_FILTER_AON,
4709 DIFF_STATUS_FILTER_BROKEN,
4710 '\0',
4711 };
4712
4713 static unsigned int filter_bit['Z' + 1];
4714
4715 static void prepare_filter_bits(void)
4716 {
4717 int i;
4718
4719 if (!filter_bit[DIFF_STATUS_ADDED]) {
4720 for (i = 0; diff_status_letters[i]; i++)
4721 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4722 }
4723 }
4724
4725 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4726 {
4727 return opt->filter & filter_bit[(int) status];
4728 }
4729
4730 unsigned diff_filter_bit(char status)
4731 {
4732 prepare_filter_bits();
4733 return filter_bit[(int) status];
4734 }
4735
4736 void diff_setup_done(struct diff_options *options)
4737 {
4738 unsigned check_mask = DIFF_FORMAT_NAME |
4739 DIFF_FORMAT_NAME_STATUS |
4740 DIFF_FORMAT_CHECKDIFF |
4741 DIFF_FORMAT_NO_OUTPUT;
4742 /*
4743 * This must be signed because we're comparing against a potentially
4744 * negative value.
4745 */
4746 const int hexsz = the_hash_algo->hexsz;
4747
4748 if (options->set_default)
4749 options->set_default(options);
4750
4751 if (HAS_MULTI_BITS(options->output_format & check_mask))
4752 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4753 "--name-only", "--name-status", "--check", "-s");
4754
4755 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4756 die(_("options '%s', '%s', and '%s' cannot be used together"),
4757 "-G", "-S", "--find-object");
4758
4759 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4760 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4761 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4762
4763 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4764 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4765 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4766
4767 /*
4768 * Most of the time we can say "there are changes"
4769 * only by checking if there are changed paths, but
4770 * --ignore-whitespace* options force us to look
4771 * inside contents.
4772 */
4773
4774 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4775 options->ignore_regex_nr)
4776 options->flags.diff_from_contents = 1;
4777 else
4778 options->flags.diff_from_contents = 0;
4779
4780 if (options->flags.find_copies_harder)
4781 options->detect_rename = DIFF_DETECT_COPY;
4782
4783 if (!options->flags.relative_name)
4784 options->prefix = NULL;
4785 if (options->prefix)
4786 options->prefix_length = strlen(options->prefix);
4787 else
4788 options->prefix_length = 0;
4789
4790 if (options->output_format & (DIFF_FORMAT_NAME |
4791 DIFF_FORMAT_NAME_STATUS |
4792 DIFF_FORMAT_CHECKDIFF |
4793 DIFF_FORMAT_NO_OUTPUT))
4794 options->output_format &= ~(DIFF_FORMAT_RAW |
4795 DIFF_FORMAT_NUMSTAT |
4796 DIFF_FORMAT_DIFFSTAT |
4797 DIFF_FORMAT_SHORTSTAT |
4798 DIFF_FORMAT_DIRSTAT |
4799 DIFF_FORMAT_SUMMARY |
4800 DIFF_FORMAT_PATCH);
4801
4802 /*
4803 * These cases always need recursive; we do not drop caller-supplied
4804 * recursive bits for other formats here.
4805 */
4806 if (options->output_format & (DIFF_FORMAT_PATCH |
4807 DIFF_FORMAT_NUMSTAT |
4808 DIFF_FORMAT_DIFFSTAT |
4809 DIFF_FORMAT_SHORTSTAT |
4810 DIFF_FORMAT_DIRSTAT |
4811 DIFF_FORMAT_SUMMARY |
4812 DIFF_FORMAT_CHECKDIFF))
4813 options->flags.recursive = 1;
4814 /*
4815 * Also pickaxe would not work very well if you do not say recursive
4816 */
4817 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4818 options->flags.recursive = 1;
4819 /*
4820 * When patches are generated, submodules diffed against the work tree
4821 * must be checked for dirtiness too so it can be shown in the output
4822 */
4823 if (options->output_format & DIFF_FORMAT_PATCH)
4824 options->flags.dirty_submodules = 1;
4825
4826 if (options->detect_rename && options->rename_limit < 0)
4827 options->rename_limit = diff_rename_limit_default;
4828 if (hexsz < options->abbrev)
4829 options->abbrev = hexsz; /* full */
4830
4831 /*
4832 * It does not make sense to show the first hit we happened
4833 * to have found. It does not make sense not to return with
4834 * exit code in such a case either.
4835 */
4836 if (options->flags.quick) {
4837 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4838 options->flags.exit_with_status = 1;
4839 }
4840
4841 options->diff_path_counter = 0;
4842
4843 if (options->flags.follow_renames && options->pathspec.nr != 1)
4844 die(_("--follow requires exactly one pathspec"));
4845
4846 if (!options->use_color || external_diff())
4847 options->color_moved = 0;
4848
4849 if (options->filter_not) {
4850 if (!options->filter)
4851 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4852 options->filter &= ~options->filter_not;
4853 }
4854 }
4855
4856 int parse_long_opt(const char *opt, const char **argv,
4857 const char **optarg)
4858 {
4859 const char *arg = argv[0];
4860 if (!skip_prefix(arg, "--", &arg))
4861 return 0;
4862 if (!skip_prefix(arg, opt, &arg))
4863 return 0;
4864 if (*arg == '=') { /* stuck form: --option=value */
4865 *optarg = arg + 1;
4866 return 1;
4867 }
4868 if (*arg != '\0')
4869 return 0;
4870 /* separate form: --option value */
4871 if (!argv[1])
4872 die("Option '--%s' requires a value", opt);
4873 *optarg = argv[1];
4874 return 2;
4875 }
4876
4877 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4878 {
4879 struct diff_options *options = opt->value;
4880 int width = options->stat_width;
4881 int name_width = options->stat_name_width;
4882 int graph_width = options->stat_graph_width;
4883 int count = options->stat_count;
4884 char *end;
4885
4886 BUG_ON_OPT_NEG(unset);
4887
4888 if (!strcmp(opt->long_name, "stat")) {
4889 if (value) {
4890 width = strtoul(value, &end, 10);
4891 if (*end == ',')
4892 name_width = strtoul(end+1, &end, 10);
4893 if (*end == ',')
4894 count = strtoul(end+1, &end, 10);
4895 if (*end)
4896 return error(_("invalid --stat value: %s"), value);
4897 }
4898 } else if (!strcmp(opt->long_name, "stat-width")) {
4899 width = strtoul(value, &end, 10);
4900 if (*end)
4901 return error(_("%s expects a numerical value"),
4902 opt->long_name);
4903 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4904 name_width = strtoul(value, &end, 10);
4905 if (*end)
4906 return error(_("%s expects a numerical value"),
4907 opt->long_name);
4908 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4909 graph_width = strtoul(value, &end, 10);
4910 if (*end)
4911 return error(_("%s expects a numerical value"),
4912 opt->long_name);
4913 } else if (!strcmp(opt->long_name, "stat-count")) {
4914 count = strtoul(value, &end, 10);
4915 if (*end)
4916 return error(_("%s expects a numerical value"),
4917 opt->long_name);
4918 } else
4919 BUG("%s should not get here", opt->long_name);
4920
4921 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4922 options->stat_name_width = name_width;
4923 options->stat_graph_width = graph_width;
4924 options->stat_width = width;
4925 options->stat_count = count;
4926 return 0;
4927 }
4928
4929 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4930 {
4931 struct strbuf errmsg = STRBUF_INIT;
4932 if (parse_dirstat_params(options, params, &errmsg))
4933 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4934 errmsg.buf);
4935 strbuf_release(&errmsg);
4936 /*
4937 * The caller knows a dirstat-related option is given from the command
4938 * line; allow it to say "return this_function();"
4939 */
4940 options->output_format |= DIFF_FORMAT_DIRSTAT;
4941 return 1;
4942 }
4943
4944 static int diff_opt_diff_filter(const struct option *option,
4945 const char *optarg, int unset)
4946 {
4947 struct diff_options *opt = option->value;
4948 int i, optch;
4949
4950 BUG_ON_OPT_NEG(unset);
4951 prepare_filter_bits();
4952
4953 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4954 unsigned int bit;
4955 int negate;
4956
4957 if ('a' <= optch && optch <= 'z') {
4958 negate = 1;
4959 optch = toupper(optch);
4960 } else {
4961 negate = 0;
4962 }
4963
4964 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4965 if (!bit)
4966 return error(_("unknown change class '%c' in --diff-filter=%s"),
4967 optarg[i], optarg);
4968 if (negate)
4969 opt->filter_not |= bit;
4970 else
4971 opt->filter |= bit;
4972 }
4973 return 0;
4974 }
4975
4976 static void enable_patch_output(int *fmt)
4977 {
4978 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4979 *fmt |= DIFF_FORMAT_PATCH;
4980 }
4981
4982 static int diff_opt_ws_error_highlight(const struct option *option,
4983 const char *arg, int unset)
4984 {
4985 struct diff_options *opt = option->value;
4986 int val = parse_ws_error_highlight(arg);
4987
4988 BUG_ON_OPT_NEG(unset);
4989 if (val < 0)
4990 return error(_("unknown value after ws-error-highlight=%.*s"),
4991 -1 - val, arg);
4992 opt->ws_error_highlight = val;
4993 return 0;
4994 }
4995
4996 static int diff_opt_find_object(const struct option *option,
4997 const char *arg, int unset)
4998 {
4999 struct diff_options *opt = option->value;
5000 struct object_id oid;
5001
5002 BUG_ON_OPT_NEG(unset);
5003 if (repo_get_oid(the_repository, arg, &oid))
5004 return error(_("unable to resolve '%s'"), arg);
5005
5006 if (!opt->objfind)
5007 CALLOC_ARRAY(opt->objfind, 1);
5008
5009 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5010 opt->flags.recursive = 1;
5011 opt->flags.tree_in_recursive = 1;
5012 oidset_insert(opt->objfind, &oid);
5013 return 0;
5014 }
5015
5016 static int diff_opt_anchored(const struct option *opt,
5017 const char *arg, int unset)
5018 {
5019 struct diff_options *options = opt->value;
5020
5021 BUG_ON_OPT_NEG(unset);
5022 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5023 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5024 options->anchors_alloc);
5025 options->anchors[options->anchors_nr++] = xstrdup(arg);
5026 return 0;
5027 }
5028
5029 static int diff_opt_binary(const struct option *opt,
5030 const char *arg, int unset)
5031 {
5032 struct diff_options *options = opt->value;
5033
5034 BUG_ON_OPT_NEG(unset);
5035 BUG_ON_OPT_ARG(arg);
5036 enable_patch_output(&options->output_format);
5037 options->flags.binary = 1;
5038 return 0;
5039 }
5040
5041 static int diff_opt_break_rewrites(const struct option *opt,
5042 const char *arg, int unset)
5043 {
5044 int *break_opt = opt->value;
5045 int opt1, opt2;
5046
5047 BUG_ON_OPT_NEG(unset);
5048 if (!arg)
5049 arg = "";
5050 opt1 = parse_rename_score(&arg);
5051 if (*arg == 0)
5052 opt2 = 0;
5053 else if (*arg != '/')
5054 return error(_("%s expects <n>/<m> form"), opt->long_name);
5055 else {
5056 arg++;
5057 opt2 = parse_rename_score(&arg);
5058 }
5059 if (*arg != 0)
5060 return error(_("%s expects <n>/<m> form"), opt->long_name);
5061 *break_opt = opt1 | (opt2 << 16);
5062 return 0;
5063 }
5064
5065 static int diff_opt_char(const struct option *opt,
5066 const char *arg, int unset)
5067 {
5068 char *value = opt->value;
5069
5070 BUG_ON_OPT_NEG(unset);
5071 if (arg[1])
5072 return error(_("%s expects a character, got '%s'"),
5073 opt->long_name, arg);
5074 *value = arg[0];
5075 return 0;
5076 }
5077
5078 static int diff_opt_color_moved(const struct option *opt,
5079 const char *arg, int unset)
5080 {
5081 struct diff_options *options = opt->value;
5082
5083 if (unset) {
5084 options->color_moved = COLOR_MOVED_NO;
5085 } else if (!arg) {
5086 if (diff_color_moved_default)
5087 options->color_moved = diff_color_moved_default;
5088 if (options->color_moved == COLOR_MOVED_NO)
5089 options->color_moved = COLOR_MOVED_DEFAULT;
5090 } else {
5091 int cm = parse_color_moved(arg);
5092 if (cm < 0)
5093 return error(_("bad --color-moved argument: %s"), arg);
5094 options->color_moved = cm;
5095 }
5096 return 0;
5097 }
5098
5099 static int diff_opt_color_moved_ws(const struct option *opt,
5100 const char *arg, int unset)
5101 {
5102 struct diff_options *options = opt->value;
5103 unsigned cm;
5104
5105 if (unset) {
5106 options->color_moved_ws_handling = 0;
5107 return 0;
5108 }
5109
5110 cm = parse_color_moved_ws(arg);
5111 if (cm & COLOR_MOVED_WS_ERROR)
5112 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5113 options->color_moved_ws_handling = cm;
5114 return 0;
5115 }
5116
5117 static int diff_opt_color_words(const struct option *opt,
5118 const char *arg, int unset)
5119 {
5120 struct diff_options *options = opt->value;
5121
5122 BUG_ON_OPT_NEG(unset);
5123 options->use_color = 1;
5124 options->word_diff = DIFF_WORDS_COLOR;
5125 options->word_regex = arg;
5126 return 0;
5127 }
5128
5129 static int diff_opt_compact_summary(const struct option *opt,
5130 const char *arg, int unset)
5131 {
5132 struct diff_options *options = opt->value;
5133
5134 BUG_ON_OPT_ARG(arg);
5135 if (unset) {
5136 options->flags.stat_with_summary = 0;
5137 } else {
5138 options->flags.stat_with_summary = 1;
5139 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5140 }
5141 return 0;
5142 }
5143
5144 static int diff_opt_diff_algorithm(const struct option *opt,
5145 const char *arg, int unset)
5146 {
5147 struct diff_options *options = opt->value;
5148
5149 BUG_ON_OPT_NEG(unset);
5150
5151 if (set_diff_algorithm(options, arg))
5152 return error(_("option diff-algorithm accepts \"myers\", "
5153 "\"minimal\", \"patience\" and \"histogram\""));
5154
5155 options->ignore_driver_algorithm = 1;
5156
5157 return 0;
5158 }
5159
5160 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5161 const char *arg, int unset)
5162 {
5163 struct diff_options *options = opt->value;
5164
5165 BUG_ON_OPT_NEG(unset);
5166 BUG_ON_OPT_ARG(arg);
5167
5168 if (set_diff_algorithm(options, opt->long_name))
5169 BUG("available diff algorithms include \"myers\", "
5170 "\"minimal\", \"patience\" and \"histogram\"");
5171
5172 options->ignore_driver_algorithm = 1;
5173
5174 return 0;
5175 }
5176
5177 static int diff_opt_dirstat(const struct option *opt,
5178 const char *arg, int unset)
5179 {
5180 struct diff_options *options = opt->value;
5181
5182 BUG_ON_OPT_NEG(unset);
5183 if (!strcmp(opt->long_name, "cumulative")) {
5184 if (arg)
5185 BUG("how come --cumulative take a value?");
5186 arg = "cumulative";
5187 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5188 parse_dirstat_opt(options, "files");
5189 parse_dirstat_opt(options, arg ? arg : "");
5190 return 0;
5191 }
5192
5193 static int diff_opt_find_copies(const struct option *opt,
5194 const char *arg, int unset)
5195 {
5196 struct diff_options *options = opt->value;
5197
5198 BUG_ON_OPT_NEG(unset);
5199 if (!arg)
5200 arg = "";
5201 options->rename_score = parse_rename_score(&arg);
5202 if (*arg != 0)
5203 return error(_("invalid argument to %s"), opt->long_name);
5204
5205 if (options->detect_rename == DIFF_DETECT_COPY)
5206 options->flags.find_copies_harder = 1;
5207 else
5208 options->detect_rename = DIFF_DETECT_COPY;
5209
5210 return 0;
5211 }
5212
5213 static int diff_opt_find_renames(const struct option *opt,
5214 const char *arg, int unset)
5215 {
5216 struct diff_options *options = opt->value;
5217
5218 BUG_ON_OPT_NEG(unset);
5219 if (!arg)
5220 arg = "";
5221 options->rename_score = parse_rename_score(&arg);
5222 if (*arg != 0)
5223 return error(_("invalid argument to %s"), opt->long_name);
5224
5225 options->detect_rename = DIFF_DETECT_RENAME;
5226 return 0;
5227 }
5228
5229 static int diff_opt_follow(const struct option *opt,
5230 const char *arg, int unset)
5231 {
5232 struct diff_options *options = opt->value;
5233
5234 BUG_ON_OPT_ARG(arg);
5235 if (unset) {
5236 options->flags.follow_renames = 0;
5237 options->flags.default_follow_renames = 0;
5238 } else {
5239 options->flags.follow_renames = 1;
5240 }
5241 return 0;
5242 }
5243
5244 static int diff_opt_ignore_submodules(const struct option *opt,
5245 const char *arg, int unset)
5246 {
5247 struct diff_options *options = opt->value;
5248
5249 BUG_ON_OPT_NEG(unset);
5250 if (!arg)
5251 arg = "all";
5252 options->flags.override_submodule_config = 1;
5253 handle_ignore_submodules_arg(options, arg);
5254 return 0;
5255 }
5256
5257 static int diff_opt_line_prefix(const struct option *opt,
5258 const char *optarg, int unset)
5259 {
5260 struct diff_options *options = opt->value;
5261
5262 BUG_ON_OPT_NEG(unset);
5263 options->line_prefix = optarg;
5264 options->line_prefix_length = strlen(options->line_prefix);
5265 graph_setup_line_prefix(options);
5266 return 0;
5267 }
5268
5269 static int diff_opt_no_prefix(const struct option *opt,
5270 const char *optarg, int unset)
5271 {
5272 struct diff_options *options = opt->value;
5273
5274 BUG_ON_OPT_NEG(unset);
5275 BUG_ON_OPT_ARG(optarg);
5276 options->a_prefix = "";
5277 options->b_prefix = "";
5278 return 0;
5279 }
5280
5281 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5282 const struct option *opt,
5283 const char *arg, int unset)
5284 {
5285 struct diff_options *options = opt->value;
5286 char *path;
5287
5288 BUG_ON_OPT_NEG(unset);
5289 path = prefix_filename(ctx->prefix, arg);
5290 options->file = xfopen(path, "w");
5291 options->close_file = 1;
5292 if (options->use_color != GIT_COLOR_ALWAYS)
5293 options->use_color = GIT_COLOR_NEVER;
5294 free(path);
5295 return 0;
5296 }
5297
5298 static int diff_opt_patience(const struct option *opt,
5299 const char *arg, int unset)
5300 {
5301 struct diff_options *options = opt->value;
5302 int i;
5303
5304 BUG_ON_OPT_NEG(unset);
5305 BUG_ON_OPT_ARG(arg);
5306 /*
5307 * Both --patience and --anchored use PATIENCE_DIFF
5308 * internally, so remove any anchors previously
5309 * specified.
5310 */
5311 for (i = 0; i < options->anchors_nr; i++)
5312 free(options->anchors[i]);
5313 options->anchors_nr = 0;
5314 options->ignore_driver_algorithm = 1;
5315
5316 return set_diff_algorithm(options, "patience");
5317 }
5318
5319 static int diff_opt_ignore_regex(const struct option *opt,
5320 const char *arg, int unset)
5321 {
5322 struct diff_options *options = opt->value;
5323 regex_t *regex;
5324
5325 BUG_ON_OPT_NEG(unset);
5326 regex = xmalloc(sizeof(*regex));
5327 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5328 return error(_("invalid regex given to -I: '%s'"), arg);
5329 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5330 options->ignore_regex_alloc);
5331 options->ignore_regex[options->ignore_regex_nr++] = regex;
5332 return 0;
5333 }
5334
5335 static int diff_opt_pickaxe_regex(const struct option *opt,
5336 const char *arg, int unset)
5337 {
5338 struct diff_options *options = opt->value;
5339
5340 BUG_ON_OPT_NEG(unset);
5341 options->pickaxe = arg;
5342 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5343 return 0;
5344 }
5345
5346 static int diff_opt_pickaxe_string(const struct option *opt,
5347 const char *arg, int unset)
5348 {
5349 struct diff_options *options = opt->value;
5350
5351 BUG_ON_OPT_NEG(unset);
5352 options->pickaxe = arg;
5353 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5354 return 0;
5355 }
5356
5357 static int diff_opt_relative(const struct option *opt,
5358 const char *arg, int unset)
5359 {
5360 struct diff_options *options = opt->value;
5361
5362 options->flags.relative_name = !unset;
5363 if (arg)
5364 options->prefix = arg;
5365 return 0;
5366 }
5367
5368 static int diff_opt_submodule(const struct option *opt,
5369 const char *arg, int unset)
5370 {
5371 struct diff_options *options = opt->value;
5372
5373 BUG_ON_OPT_NEG(unset);
5374 if (!arg)
5375 arg = "log";
5376 if (parse_submodule_params(options, arg))
5377 return error(_("failed to parse --submodule option parameter: '%s'"),
5378 arg);
5379 return 0;
5380 }
5381
5382 static int diff_opt_textconv(const struct option *opt,
5383 const char *arg, int unset)
5384 {
5385 struct diff_options *options = opt->value;
5386
5387 BUG_ON_OPT_ARG(arg);
5388 if (unset) {
5389 options->flags.allow_textconv = 0;
5390 } else {
5391 options->flags.allow_textconv = 1;
5392 options->flags.textconv_set_via_cmdline = 1;
5393 }
5394 return 0;
5395 }
5396
5397 static int diff_opt_unified(const struct option *opt,
5398 const char *arg, int unset)
5399 {
5400 struct diff_options *options = opt->value;
5401 char *s;
5402
5403 BUG_ON_OPT_NEG(unset);
5404
5405 if (arg) {
5406 options->context = strtol(arg, &s, 10);
5407 if (*s)
5408 return error(_("%s expects a numerical value"), "--unified");
5409 }
5410 enable_patch_output(&options->output_format);
5411
5412 return 0;
5413 }
5414
5415 static int diff_opt_word_diff(const struct option *opt,
5416 const char *arg, int unset)
5417 {
5418 struct diff_options *options = opt->value;
5419
5420 BUG_ON_OPT_NEG(unset);
5421 if (arg) {
5422 if (!strcmp(arg, "plain"))
5423 options->word_diff = DIFF_WORDS_PLAIN;
5424 else if (!strcmp(arg, "color")) {
5425 options->use_color = 1;
5426 options->word_diff = DIFF_WORDS_COLOR;
5427 }
5428 else if (!strcmp(arg, "porcelain"))
5429 options->word_diff = DIFF_WORDS_PORCELAIN;
5430 else if (!strcmp(arg, "none"))
5431 options->word_diff = DIFF_WORDS_NONE;
5432 else
5433 return error(_("bad --word-diff argument: %s"), arg);
5434 } else {
5435 if (options->word_diff == DIFF_WORDS_NONE)
5436 options->word_diff = DIFF_WORDS_PLAIN;
5437 }
5438 return 0;
5439 }
5440
5441 static int diff_opt_word_diff_regex(const struct option *opt,
5442 const char *arg, int unset)
5443 {
5444 struct diff_options *options = opt->value;
5445
5446 BUG_ON_OPT_NEG(unset);
5447 if (options->word_diff == DIFF_WORDS_NONE)
5448 options->word_diff = DIFF_WORDS_PLAIN;
5449 options->word_regex = arg;
5450 return 0;
5451 }
5452
5453 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5454 {
5455 struct diff_options *options = opt->value;
5456
5457 BUG_ON_OPT_NEG(unset);
5458 if (!strcmp(opt->long_name, "skip-to"))
5459 options->skip_instead_of_rotate = 1;
5460 else
5461 options->skip_instead_of_rotate = 0;
5462 options->rotate_to = arg;
5463 return 0;
5464 }
5465
5466 struct option *add_diff_options(const struct option *opts,
5467 struct diff_options *options)
5468 {
5469 struct option parseopts[] = {
5470 OPT_GROUP(N_("Diff output format options")),
5471 OPT_BITOP('p', "patch", &options->output_format,
5472 N_("generate patch"),
5473 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5474 OPT_BIT_F('s', "no-patch", &options->output_format,
5475 N_("suppress diff output"),
5476 DIFF_FORMAT_NO_OUTPUT, PARSE_OPT_NONEG),
5477 OPT_BITOP('u', NULL, &options->output_format,
5478 N_("generate patch"),
5479 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5480 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5481 N_("generate diffs with <n> lines context"),
5482 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5483 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5484 N_("generate diffs with <n> lines context")),
5485 OPT_BIT_F(0, "raw", &options->output_format,
5486 N_("generate the diff in raw format"),
5487 DIFF_FORMAT_RAW, PARSE_OPT_NONEG),
5488 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5489 N_("synonym for '-p --raw'"),
5490 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5491 DIFF_FORMAT_NO_OUTPUT),
5492 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5493 N_("synonym for '-p --stat'"),
5494 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5495 DIFF_FORMAT_NO_OUTPUT),
5496 OPT_BIT_F(0, "numstat", &options->output_format,
5497 N_("machine friendly --stat"),
5498 DIFF_FORMAT_NUMSTAT, PARSE_OPT_NONEG),
5499 OPT_BIT_F(0, "shortstat", &options->output_format,
5500 N_("output only the last line of --stat"),
5501 DIFF_FORMAT_SHORTSTAT, PARSE_OPT_NONEG),
5502 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1,param2>..."),
5503 N_("output the distribution of relative amount of changes for each sub-directory"),
5504 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5505 diff_opt_dirstat),
5506 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5507 N_("synonym for --dirstat=cumulative"),
5508 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5509 diff_opt_dirstat),
5510 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1,param2>..."),
5511 N_("synonym for --dirstat=files,param1,param2..."),
5512 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5513 diff_opt_dirstat),
5514 OPT_BIT_F(0, "check", &options->output_format,
5515 N_("warn if changes introduce conflict markers or whitespace errors"),
5516 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5517 OPT_BIT_F(0, "summary", &options->output_format,
5518 N_("condensed summary such as creations, renames and mode changes"),
5519 DIFF_FORMAT_SUMMARY, PARSE_OPT_NONEG),
5520 OPT_BIT_F(0, "name-only", &options->output_format,
5521 N_("show only names of changed files"),
5522 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5523 OPT_BIT_F(0, "name-status", &options->output_format,
5524 N_("show only names and status of changed files"),
5525 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5526 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5527 N_("generate diffstat"),
5528 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5529 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5530 N_("generate diffstat with a given width"),
5531 PARSE_OPT_NONEG, diff_opt_stat),
5532 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5533 N_("generate diffstat with a given name width"),
5534 PARSE_OPT_NONEG, diff_opt_stat),
5535 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5536 N_("generate diffstat with a given graph width"),
5537 PARSE_OPT_NONEG, diff_opt_stat),
5538 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5539 N_("generate diffstat with limited lines"),
5540 PARSE_OPT_NONEG, diff_opt_stat),
5541 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5542 N_("generate compact summary in diffstat"),
5543 PARSE_OPT_NOARG, diff_opt_compact_summary),
5544 OPT_CALLBACK_F(0, "binary", options, NULL,
5545 N_("output a binary diff that can be applied"),
5546 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5547 OPT_BOOL(0, "full-index", &options->flags.full_index,
5548 N_("show full pre- and post-image object names on the \"index\" lines")),
5549 OPT_COLOR_FLAG(0, "color", &options->use_color,
5550 N_("show colored diff")),
5551 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5552 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5553 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5554 OPT_SET_INT('z', NULL, &options->line_termination,
5555 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5556 0),
5557 OPT__ABBREV(&options->abbrev),
5558 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5559 N_("show the given source prefix instead of \"a/\""),
5560 PARSE_OPT_NONEG),
5561 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5562 N_("show the given destination prefix instead of \"b/\""),
5563 PARSE_OPT_NONEG),
5564 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5565 N_("prepend an additional prefix to every line of output"),
5566 PARSE_OPT_NONEG, diff_opt_line_prefix),
5567 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5568 N_("do not show any source or destination prefix"),
5569 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5570 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5571 N_("show context between diff hunks up to the specified number of lines"),
5572 PARSE_OPT_NONEG),
5573 OPT_CALLBACK_F(0, "output-indicator-new",
5574 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5575 N_("<char>"),
5576 N_("specify the character to indicate a new line instead of '+'"),
5577 PARSE_OPT_NONEG, diff_opt_char),
5578 OPT_CALLBACK_F(0, "output-indicator-old",
5579 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5580 N_("<char>"),
5581 N_("specify the character to indicate an old line instead of '-'"),
5582 PARSE_OPT_NONEG, diff_opt_char),
5583 OPT_CALLBACK_F(0, "output-indicator-context",
5584 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5585 N_("<char>"),
5586 N_("specify the character to indicate a context instead of ' '"),
5587 PARSE_OPT_NONEG, diff_opt_char),
5588
5589 OPT_GROUP(N_("Diff rename options")),
5590 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5591 N_("break complete rewrite changes into pairs of delete and create"),
5592 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5593 diff_opt_break_rewrites),
5594 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5595 N_("detect renames"),
5596 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5597 diff_opt_find_renames),
5598 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5599 N_("omit the preimage for deletes"),
5600 1, PARSE_OPT_NONEG),
5601 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5602 N_("detect copies"),
5603 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5604 diff_opt_find_copies),
5605 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5606 N_("use unmodified files as source to find copies")),
5607 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5608 N_("disable rename detection"),
5609 0, PARSE_OPT_NONEG),
5610 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5611 N_("use empty blobs as rename source")),
5612 OPT_CALLBACK_F(0, "follow", options, NULL,
5613 N_("continue listing the history of a file beyond renames"),
5614 PARSE_OPT_NOARG, diff_opt_follow),
5615 OPT_INTEGER('l', NULL, &options->rename_limit,
5616 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5617
5618 OPT_GROUP(N_("Diff algorithm options")),
5619 OPT_CALLBACK_F(0, "minimal", options, NULL,
5620 N_("produce the smallest possible diff"),
5621 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5622 diff_opt_diff_algorithm_no_arg),
5623 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5624 N_("ignore whitespace when comparing lines"),
5625 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5626 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5627 N_("ignore changes in amount of whitespace"),
5628 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5629 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5630 N_("ignore changes in whitespace at EOL"),
5631 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5632 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5633 N_("ignore carrier-return at the end of line"),
5634 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5635 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5636 N_("ignore changes whose lines are all blank"),
5637 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5638 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5639 N_("ignore changes whose all lines match <regex>"),
5640 0, diff_opt_ignore_regex),
5641 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5642 N_("heuristic to shift diff hunk boundaries for easy reading"),
5643 XDF_INDENT_HEURISTIC),
5644 OPT_CALLBACK_F(0, "patience", options, NULL,
5645 N_("generate diff using the \"patience diff\" algorithm"),
5646 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5647 diff_opt_patience),
5648 OPT_CALLBACK_F(0, "histogram", options, NULL,
5649 N_("generate diff using the \"histogram diff\" algorithm"),
5650 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5651 diff_opt_diff_algorithm_no_arg),
5652 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5653 N_("choose a diff algorithm"),
5654 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5655 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5656 N_("generate diff using the \"anchored diff\" algorithm"),
5657 PARSE_OPT_NONEG, diff_opt_anchored),
5658 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5659 N_("show word diff, using <mode> to delimit changed words"),
5660 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5661 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5662 N_("use <regex> to decide what a word is"),
5663 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5664 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5665 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5666 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5667 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5668 N_("moved lines of code are colored differently"),
5669 PARSE_OPT_OPTARG, diff_opt_color_moved),
5670 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5671 N_("how white spaces are ignored in --color-moved"),
5672 0, diff_opt_color_moved_ws),
5673
5674 OPT_GROUP(N_("Other diff options")),
5675 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5676 N_("when run from subdir, exclude changes outside and show relative paths"),
5677 PARSE_OPT_OPTARG,
5678 diff_opt_relative),
5679 OPT_BOOL('a', "text", &options->flags.text,
5680 N_("treat all files as text")),
5681 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5682 N_("swap two inputs, reverse the diff")),
5683 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5684 N_("exit with 1 if there were differences, 0 otherwise")),
5685 OPT_BOOL(0, "quiet", &options->flags.quick,
5686 N_("disable all output of the program")),
5687 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5688 N_("allow an external diff helper to be executed")),
5689 OPT_CALLBACK_F(0, "textconv", options, NULL,
5690 N_("run external text conversion filters when comparing binary files"),
5691 PARSE_OPT_NOARG, diff_opt_textconv),
5692 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5693 N_("ignore changes to submodules in the diff generation"),
5694 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5695 diff_opt_ignore_submodules),
5696 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5697 N_("specify how differences in submodules are shown"),
5698 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5699 diff_opt_submodule),
5700 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5701 N_("hide 'git add -N' entries from the index"),
5702 1, PARSE_OPT_NONEG),
5703 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5704 N_("treat 'git add -N' entries as real in the index"),
5705 0, PARSE_OPT_NONEG),
5706 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5707 N_("look for differences that change the number of occurrences of the specified string"),
5708 0, diff_opt_pickaxe_string),
5709 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5710 N_("look for differences that change the number of occurrences of the specified regex"),
5711 0, diff_opt_pickaxe_regex),
5712 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5713 N_("show all changes in the changeset with -S or -G"),
5714 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5715 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5716 N_("treat <string> in -S as extended POSIX regular expression"),
5717 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5718 OPT_FILENAME('O', NULL, &options->orderfile,
5719 N_("control the order in which files appear in the output")),
5720 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5721 N_("show the change in the specified path first"),
5722 PARSE_OPT_NONEG, diff_opt_rotate_to),
5723 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5724 N_("skip the output to the specified path"),
5725 PARSE_OPT_NONEG, diff_opt_rotate_to),
5726 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5727 N_("look for differences that change the number of occurrences of the specified object"),
5728 PARSE_OPT_NONEG, diff_opt_find_object),
5729 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5730 N_("select files by diff type"),
5731 PARSE_OPT_NONEG, diff_opt_diff_filter),
5732 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5733 N_("output to a specific file"),
5734 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5735
5736 OPT_END()
5737 };
5738
5739 return parse_options_concat(opts, parseopts);
5740 }
5741
5742 int diff_opt_parse(struct diff_options *options,
5743 const char **av, int ac, const char *prefix)
5744 {
5745 struct option no_options[] = { OPT_END() };
5746 struct option *parseopts = add_diff_options(no_options, options);
5747
5748 if (!prefix)
5749 prefix = "";
5750
5751 ac = parse_options(ac, av, prefix, parseopts, NULL,
5752 PARSE_OPT_KEEP_DASHDASH |
5753 PARSE_OPT_KEEP_UNKNOWN_OPT |
5754 PARSE_OPT_NO_INTERNAL_HELP |
5755 PARSE_OPT_ONE_SHOT |
5756 PARSE_OPT_STOP_AT_NON_OPTION);
5757 free(parseopts);
5758
5759 return ac;
5760 }
5761
5762 int parse_rename_score(const char **cp_p)
5763 {
5764 unsigned long num, scale;
5765 int ch, dot;
5766 const char *cp = *cp_p;
5767
5768 num = 0;
5769 scale = 1;
5770 dot = 0;
5771 for (;;) {
5772 ch = *cp;
5773 if ( !dot && ch == '.' ) {
5774 scale = 1;
5775 dot = 1;
5776 } else if ( ch == '%' ) {
5777 scale = dot ? scale*100 : 100;
5778 cp++; /* % is always at the end */
5779 break;
5780 } else if ( ch >= '0' && ch <= '9' ) {
5781 if ( scale < 100000 ) {
5782 scale *= 10;
5783 num = (num*10) + (ch-'0');
5784 }
5785 } else {
5786 break;
5787 }
5788 cp++;
5789 }
5790 *cp_p = cp;
5791
5792 /* user says num divided by scale and we say internally that
5793 * is MAX_SCORE * num / scale.
5794 */
5795 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5796 }
5797
5798 struct diff_queue_struct diff_queued_diff;
5799
5800 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5801 {
5802 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5803 queue->queue[queue->nr++] = dp;
5804 }
5805
5806 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5807 struct diff_filespec *one,
5808 struct diff_filespec *two)
5809 {
5810 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5811 dp->one = one;
5812 dp->two = two;
5813 if (queue)
5814 diff_q(queue, dp);
5815 return dp;
5816 }
5817
5818 void diff_free_filepair(struct diff_filepair *p)
5819 {
5820 free_filespec(p->one);
5821 free_filespec(p->two);
5822 free(p);
5823 }
5824
5825 void diff_free_queue(struct diff_queue_struct *q)
5826 {
5827 for (int i = 0; i < q->nr; i++)
5828 diff_free_filepair(q->queue[i]);
5829 free(q->queue);
5830 }
5831
5832 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5833 {
5834 int abblen;
5835 const char *abbrev;
5836
5837 /* Do we want all 40 hex characters? */
5838 if (len == the_hash_algo->hexsz)
5839 return oid_to_hex(oid);
5840
5841 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5842 abbrev = diff_abbrev_oid(oid, len);
5843
5844 if (!print_sha1_ellipsis())
5845 return abbrev;
5846
5847 abblen = strlen(abbrev);
5848
5849 /*
5850 * In well-behaved cases, where the abbreviated result is the
5851 * same as the requested length, append three dots after the
5852 * abbreviation (hence the whole logic is limited to the case
5853 * where abblen < 37); when the actual abbreviated result is a
5854 * bit longer than the requested length, we reduce the number
5855 * of dots so that they match the well-behaved ones. However,
5856 * if the actual abbreviation is longer than the requested
5857 * length by more than three, we give up on aligning, and add
5858 * three dots anyway, to indicate that the output is not the
5859 * full object name. Yes, this may be suboptimal, but this
5860 * appears only in "diff --raw --abbrev" output and it is not
5861 * worth the effort to change it now. Note that this would
5862 * likely to work fine when the automatic sizing of default
5863 * abbreviation length is used--we would be fed -1 in "len" in
5864 * that case, and will end up always appending three-dots, but
5865 * the automatic sizing is supposed to give abblen that ensures
5866 * uniqueness across all objects (statistically speaking).
5867 */
5868 if (abblen < the_hash_algo->hexsz - 3) {
5869 static char hex[GIT_MAX_HEXSZ + 1];
5870 if (len < abblen && abblen <= len + 2)
5871 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5872 else
5873 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5874 return hex;
5875 }
5876
5877 return oid_to_hex(oid);
5878 }
5879
5880 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5881 {
5882 int line_termination = opt->line_termination;
5883 int inter_name_termination = line_termination ? '\t' : '\0';
5884
5885 fprintf(opt->file, "%s", diff_line_prefix(opt));
5886 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5887 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5888 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5889 fprintf(opt->file, "%s ",
5890 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5891 }
5892 if (p->score) {
5893 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5894 inter_name_termination);
5895 } else {
5896 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5897 }
5898
5899 if (p->status == DIFF_STATUS_COPIED ||
5900 p->status == DIFF_STATUS_RENAMED) {
5901 const char *name_a, *name_b;
5902 name_a = p->one->path;
5903 name_b = p->two->path;
5904 strip_prefix(opt->prefix_length, &name_a, &name_b);
5905 write_name_quoted(name_a, opt->file, inter_name_termination);
5906 write_name_quoted(name_b, opt->file, line_termination);
5907 } else {
5908 const char *name_a, *name_b;
5909 name_a = p->one->mode ? p->one->path : p->two->path;
5910 name_b = NULL;
5911 strip_prefix(opt->prefix_length, &name_a, &name_b);
5912 write_name_quoted(name_a, opt->file, line_termination);
5913 }
5914 }
5915
5916 int diff_unmodified_pair(struct diff_filepair *p)
5917 {
5918 /* This function is written stricter than necessary to support
5919 * the currently implemented transformers, but the idea is to
5920 * let transformers to produce diff_filepairs any way they want,
5921 * and filter and clean them up here before producing the output.
5922 */
5923 struct diff_filespec *one = p->one, *two = p->two;
5924
5925 if (DIFF_PAIR_UNMERGED(p))
5926 return 0; /* unmerged is interesting */
5927
5928 /* deletion, addition, mode or type change
5929 * and rename are all interesting.
5930 */
5931 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5932 DIFF_PAIR_MODE_CHANGED(p) ||
5933 strcmp(one->path, two->path))
5934 return 0;
5935
5936 /* both are valid and point at the same path. that is, we are
5937 * dealing with a change.
5938 */
5939 if (one->oid_valid && two->oid_valid &&
5940 oideq(&one->oid, &two->oid) &&
5941 !one->dirty_submodule && !two->dirty_submodule)
5942 return 1; /* no change */
5943 if (!one->oid_valid && !two->oid_valid)
5944 return 1; /* both look at the same file on the filesystem. */
5945 return 0;
5946 }
5947
5948 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5949 {
5950 int include_conflict_headers =
5951 (additional_headers(o, p->one->path) &&
5952 !o->pickaxe_opts &&
5953 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
5954
5955 /*
5956 * Check if we can return early without showing a diff. Note that
5957 * diff_filepair only stores {oid, path, mode, is_valid}
5958 * information for each path, and thus diff_unmodified_pair() only
5959 * considers those bits of info. However, we do not want pairs
5960 * created by create_filepairs_for_header_only_notifications()
5961 * (which always look like unmodified pairs) to be ignored, so
5962 * return early if both p is unmodified AND we don't want to
5963 * include_conflict_headers.
5964 */
5965 if (diff_unmodified_pair(p) && !include_conflict_headers)
5966 return;
5967
5968 /* Actually, we can also return early to avoid showing tree diffs */
5969 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5970 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5971 return;
5972
5973 run_diff(p, o);
5974 }
5975
5976 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5977 struct diffstat_t *diffstat)
5978 {
5979 if (diff_unmodified_pair(p))
5980 return;
5981
5982 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5983 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5984 return; /* no useful stat for tree diffs */
5985
5986 run_diffstat(p, o, diffstat);
5987 }
5988
5989 static void diff_flush_checkdiff(struct diff_filepair *p,
5990 struct diff_options *o)
5991 {
5992 if (diff_unmodified_pair(p))
5993 return;
5994
5995 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5996 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5997 return; /* nothing to check in tree diffs */
5998
5999 run_checkdiff(p, o);
6000 }
6001
6002 int diff_queue_is_empty(struct diff_options *o)
6003 {
6004 struct diff_queue_struct *q = &diff_queued_diff;
6005 int i;
6006 int include_conflict_headers =
6007 (o->additional_path_headers &&
6008 strmap_get_size(o->additional_path_headers) &&
6009 !o->pickaxe_opts &&
6010 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6011
6012 if (include_conflict_headers)
6013 return 0;
6014
6015 for (i = 0; i < q->nr; i++)
6016 if (!diff_unmodified_pair(q->queue[i]))
6017 return 0;
6018 return 1;
6019 }
6020
6021 #if DIFF_DEBUG
6022 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6023 {
6024 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6025 x, one ? one : "",
6026 s->path,
6027 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6028 s->mode,
6029 s->oid_valid ? oid_to_hex(&s->oid) : "");
6030 fprintf(stderr, "queue[%d] %s size %lu\n",
6031 x, one ? one : "",
6032 s->size);
6033 }
6034
6035 void diff_debug_filepair(const struct diff_filepair *p, int i)
6036 {
6037 diff_debug_filespec(p->one, i, "one");
6038 diff_debug_filespec(p->two, i, "two");
6039 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6040 p->score, p->status ? p->status : '?',
6041 p->one->rename_used, p->broken_pair);
6042 }
6043
6044 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6045 {
6046 int i;
6047 if (msg)
6048 fprintf(stderr, "%s\n", msg);
6049 fprintf(stderr, "q->nr = %d\n", q->nr);
6050 for (i = 0; i < q->nr; i++) {
6051 struct diff_filepair *p = q->queue[i];
6052 diff_debug_filepair(p, i);
6053 }
6054 }
6055 #endif
6056
6057 static void diff_resolve_rename_copy(void)
6058 {
6059 int i;
6060 struct diff_filepair *p;
6061 struct diff_queue_struct *q = &diff_queued_diff;
6062
6063 diff_debug_queue("resolve-rename-copy", q);
6064
6065 for (i = 0; i < q->nr; i++) {
6066 p = q->queue[i];
6067 p->status = 0; /* undecided */
6068 if (DIFF_PAIR_UNMERGED(p))
6069 p->status = DIFF_STATUS_UNMERGED;
6070 else if (!DIFF_FILE_VALID(p->one))
6071 p->status = DIFF_STATUS_ADDED;
6072 else if (!DIFF_FILE_VALID(p->two))
6073 p->status = DIFF_STATUS_DELETED;
6074 else if (DIFF_PAIR_TYPE_CHANGED(p))
6075 p->status = DIFF_STATUS_TYPE_CHANGED;
6076
6077 /* from this point on, we are dealing with a pair
6078 * whose both sides are valid and of the same type, i.e.
6079 * either in-place edit or rename/copy edit.
6080 */
6081 else if (DIFF_PAIR_RENAME(p)) {
6082 /*
6083 * A rename might have re-connected a broken
6084 * pair up, causing the pathnames to be the
6085 * same again. If so, that's not a rename at
6086 * all, just a modification..
6087 *
6088 * Otherwise, see if this source was used for
6089 * multiple renames, in which case we decrement
6090 * the count, and call it a copy.
6091 */
6092 if (!strcmp(p->one->path, p->two->path))
6093 p->status = DIFF_STATUS_MODIFIED;
6094 else if (--p->one->rename_used > 0)
6095 p->status = DIFF_STATUS_COPIED;
6096 else
6097 p->status = DIFF_STATUS_RENAMED;
6098 }
6099 else if (!oideq(&p->one->oid, &p->two->oid) ||
6100 p->one->mode != p->two->mode ||
6101 p->one->dirty_submodule ||
6102 p->two->dirty_submodule ||
6103 is_null_oid(&p->one->oid))
6104 p->status = DIFF_STATUS_MODIFIED;
6105 else {
6106 /* This is a "no-change" entry and should not
6107 * happen anymore, but prepare for broken callers.
6108 */
6109 error("feeding unmodified %s to diffcore",
6110 p->one->path);
6111 p->status = DIFF_STATUS_UNKNOWN;
6112 }
6113 }
6114 diff_debug_queue("resolve-rename-copy done", q);
6115 }
6116
6117 static int check_pair_status(struct diff_filepair *p)
6118 {
6119 switch (p->status) {
6120 case DIFF_STATUS_UNKNOWN:
6121 return 0;
6122 case 0:
6123 die("internal error in diff-resolve-rename-copy");
6124 default:
6125 return 1;
6126 }
6127 }
6128
6129 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6130 {
6131 int fmt = opt->output_format;
6132
6133 if (fmt & DIFF_FORMAT_CHECKDIFF)
6134 diff_flush_checkdiff(p, opt);
6135 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6136 diff_flush_raw(p, opt);
6137 else if (fmt & DIFF_FORMAT_NAME) {
6138 const char *name_a, *name_b;
6139 name_a = p->two->path;
6140 name_b = NULL;
6141 strip_prefix(opt->prefix_length, &name_a, &name_b);
6142 fprintf(opt->file, "%s", diff_line_prefix(opt));
6143 write_name_quoted(name_a, opt->file, opt->line_termination);
6144 }
6145 }
6146
6147 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6148 {
6149 struct strbuf sb = STRBUF_INIT;
6150 if (fs->mode)
6151 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6152 else
6153 strbuf_addf(&sb, " %s ", newdelete);
6154
6155 quote_c_style(fs->path, &sb, NULL, 0);
6156 strbuf_addch(&sb, '\n');
6157 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6158 sb.buf, sb.len, 0);
6159 strbuf_release(&sb);
6160 }
6161
6162 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6163 int show_name)
6164 {
6165 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6166 struct strbuf sb = STRBUF_INIT;
6167 strbuf_addf(&sb, " mode change %06o => %06o",
6168 p->one->mode, p->two->mode);
6169 if (show_name) {
6170 strbuf_addch(&sb, ' ');
6171 quote_c_style(p->two->path, &sb, NULL, 0);
6172 }
6173 strbuf_addch(&sb, '\n');
6174 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6175 sb.buf, sb.len, 0);
6176 strbuf_release(&sb);
6177 }
6178 }
6179
6180 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6181 struct diff_filepair *p)
6182 {
6183 struct strbuf sb = STRBUF_INIT;
6184 struct strbuf names = STRBUF_INIT;
6185
6186 pprint_rename(&names, p->one->path, p->two->path);
6187 strbuf_addf(&sb, " %s %s (%d%%)\n",
6188 renamecopy, names.buf, similarity_index(p));
6189 strbuf_release(&names);
6190 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6191 sb.buf, sb.len, 0);
6192 show_mode_change(opt, p, 0);
6193 strbuf_release(&sb);
6194 }
6195
6196 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6197 {
6198 switch(p->status) {
6199 case DIFF_STATUS_DELETED:
6200 show_file_mode_name(opt, "delete", p->one);
6201 break;
6202 case DIFF_STATUS_ADDED:
6203 show_file_mode_name(opt, "create", p->two);
6204 break;
6205 case DIFF_STATUS_COPIED:
6206 show_rename_copy(opt, "copy", p);
6207 break;
6208 case DIFF_STATUS_RENAMED:
6209 show_rename_copy(opt, "rename", p);
6210 break;
6211 default:
6212 if (p->score) {
6213 struct strbuf sb = STRBUF_INIT;
6214 strbuf_addstr(&sb, " rewrite ");
6215 quote_c_style(p->two->path, &sb, NULL, 0);
6216 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6217 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6218 sb.buf, sb.len, 0);
6219 strbuf_release(&sb);
6220 }
6221 show_mode_change(opt, p, !p->score);
6222 break;
6223 }
6224 }
6225
6226 struct patch_id_t {
6227 git_hash_ctx *ctx;
6228 int patchlen;
6229 };
6230
6231 static int remove_space(char *line, int len)
6232 {
6233 int i;
6234 char *dst = line;
6235 unsigned char c;
6236
6237 for (i = 0; i < len; i++)
6238 if (!isspace((c = line[i])))
6239 *dst++ = c;
6240
6241 return dst - line;
6242 }
6243
6244 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6245 {
6246 unsigned char hash[GIT_MAX_RAWSZ];
6247 unsigned short carry = 0;
6248 int i;
6249
6250 the_hash_algo->final_fn(hash, ctx);
6251 the_hash_algo->init_fn(ctx);
6252 /* 20-byte sum, with carry */
6253 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6254 carry += result->hash[i] + hash[i];
6255 result->hash[i] = carry;
6256 carry >>= 8;
6257 }
6258 }
6259
6260 static int patch_id_consume(void *priv, char *line, unsigned long len)
6261 {
6262 struct patch_id_t *data = priv;
6263 int new_len;
6264
6265 if (len > 12 && starts_with(line, "\\ "))
6266 return 0;
6267 new_len = remove_space(line, len);
6268
6269 the_hash_algo->update_fn(data->ctx, line, new_len);
6270 data->patchlen += new_len;
6271 return 0;
6272 }
6273
6274 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6275 {
6276 the_hash_algo->update_fn(ctx, str, strlen(str));
6277 }
6278
6279 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6280 {
6281 /* large enough for 2^32 in octal */
6282 char buf[12];
6283 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6284 the_hash_algo->update_fn(ctx, buf, len);
6285 }
6286
6287 /* returns 0 upon success, and writes result into oid */
6288 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6289 {
6290 struct diff_queue_struct *q = &diff_queued_diff;
6291 int i;
6292 git_hash_ctx ctx;
6293 struct patch_id_t data;
6294
6295 the_hash_algo->init_fn(&ctx);
6296 memset(&data, 0, sizeof(struct patch_id_t));
6297 data.ctx = &ctx;
6298 oidclr(oid);
6299
6300 for (i = 0; i < q->nr; i++) {
6301 xpparam_t xpp;
6302 xdemitconf_t xecfg;
6303 mmfile_t mf1, mf2;
6304 struct diff_filepair *p = q->queue[i];
6305 int len1, len2;
6306
6307 memset(&xpp, 0, sizeof(xpp));
6308 memset(&xecfg, 0, sizeof(xecfg));
6309 if (p->status == 0)
6310 return error("internal diff status error");
6311 if (p->status == DIFF_STATUS_UNKNOWN)
6312 continue;
6313 if (diff_unmodified_pair(p))
6314 continue;
6315 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6316 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6317 continue;
6318 if (DIFF_PAIR_UNMERGED(p))
6319 continue;
6320
6321 diff_fill_oid_info(p->one, options->repo->index);
6322 diff_fill_oid_info(p->two, options->repo->index);
6323
6324 len1 = remove_space(p->one->path, strlen(p->one->path));
6325 len2 = remove_space(p->two->path, strlen(p->two->path));
6326 patch_id_add_string(&ctx, "diff--git");
6327 patch_id_add_string(&ctx, "a/");
6328 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6329 patch_id_add_string(&ctx, "b/");
6330 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6331
6332 if (p->one->mode == 0) {
6333 patch_id_add_string(&ctx, "newfilemode");
6334 patch_id_add_mode(&ctx, p->two->mode);
6335 } else if (p->two->mode == 0) {
6336 patch_id_add_string(&ctx, "deletedfilemode");
6337 patch_id_add_mode(&ctx, p->one->mode);
6338 } else if (p->one->mode != p->two->mode) {
6339 patch_id_add_string(&ctx, "oldmode");
6340 patch_id_add_mode(&ctx, p->one->mode);
6341 patch_id_add_string(&ctx, "newmode");
6342 patch_id_add_mode(&ctx, p->two->mode);
6343 }
6344
6345 if (diff_header_only) {
6346 /* don't do anything since we're only populating header info */
6347 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6348 diff_filespec_is_binary(options->repo, p->two)) {
6349 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6350 the_hash_algo->hexsz);
6351 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6352 the_hash_algo->hexsz);
6353 } else {
6354 if (p->one->mode == 0) {
6355 patch_id_add_string(&ctx, "---/dev/null");
6356 patch_id_add_string(&ctx, "+++b/");
6357 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6358 } else if (p->two->mode == 0) {
6359 patch_id_add_string(&ctx, "---a/");
6360 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6361 patch_id_add_string(&ctx, "+++/dev/null");
6362 } else {
6363 patch_id_add_string(&ctx, "---a/");
6364 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6365 patch_id_add_string(&ctx, "+++b/");
6366 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6367 }
6368
6369 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6370 fill_mmfile(options->repo, &mf2, p->two) < 0)
6371 return error("unable to read files to diff");
6372 xpp.flags = 0;
6373 xecfg.ctxlen = 3;
6374 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6375 if (xdi_diff_outf(&mf1, &mf2, NULL,
6376 patch_id_consume, &data, &xpp, &xecfg))
6377 return error("unable to generate patch-id diff for %s",
6378 p->one->path);
6379 }
6380 flush_one_hunk(oid, &ctx);
6381 }
6382
6383 return 0;
6384 }
6385
6386 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6387 {
6388 struct diff_queue_struct *q = &diff_queued_diff;
6389 int result = diff_get_patch_id(options, oid, diff_header_only);
6390
6391 diff_free_queue(q);
6392 DIFF_QUEUE_CLEAR(q);
6393
6394 return result;
6395 }
6396
6397 static int is_summary_empty(const struct diff_queue_struct *q)
6398 {
6399 int i;
6400
6401 for (i = 0; i < q->nr; i++) {
6402 const struct diff_filepair *p = q->queue[i];
6403
6404 switch (p->status) {
6405 case DIFF_STATUS_DELETED:
6406 case DIFF_STATUS_ADDED:
6407 case DIFF_STATUS_COPIED:
6408 case DIFF_STATUS_RENAMED:
6409 return 0;
6410 default:
6411 if (p->score)
6412 return 0;
6413 if (p->one->mode && p->two->mode &&
6414 p->one->mode != p->two->mode)
6415 return 0;
6416 break;
6417 }
6418 }
6419 return 1;
6420 }
6421
6422 static const char rename_limit_warning[] =
6423 N_("exhaustive rename detection was skipped due to too many files.");
6424
6425 static const char degrade_cc_to_c_warning[] =
6426 N_("only found copies from modified paths due to too many files.");
6427
6428 static const char rename_limit_advice[] =
6429 N_("you may want to set your %s variable to at least "
6430 "%d and retry the command.");
6431
6432 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6433 {
6434 fflush(stdout);
6435 if (degraded_cc)
6436 warning(_(degrade_cc_to_c_warning));
6437 else if (needed)
6438 warning(_(rename_limit_warning));
6439 else
6440 return;
6441 if (0 < needed)
6442 warning(_(rename_limit_advice), varname, needed);
6443 }
6444
6445 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6446 {
6447 struct strset present;
6448 struct diff_queue_struct *q = &diff_queued_diff;
6449 struct hashmap_iter iter;
6450 struct strmap_entry *e;
6451 int i;
6452
6453 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6454
6455 /*
6456 * Find out which paths exist in diff_queued_diff, preferring
6457 * one->path for any pair that has multiple paths.
6458 */
6459 for (i = 0; i < q->nr; i++) {
6460 struct diff_filepair *p = q->queue[i];
6461 char *path = p->one->path ? p->one->path : p->two->path;
6462
6463 if (strmap_contains(o->additional_path_headers, path))
6464 strset_add(&present, path);
6465 }
6466
6467 /*
6468 * Loop over paths in additional_path_headers; for each NOT already
6469 * in diff_queued_diff, create a synthetic filepair and insert that
6470 * into diff_queued_diff.
6471 */
6472 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6473 if (!strset_contains(&present, e->key)) {
6474 struct diff_filespec *one, *two;
6475 struct diff_filepair *p;
6476
6477 one = alloc_filespec(e->key);
6478 two = alloc_filespec(e->key);
6479 fill_filespec(one, null_oid(), 0, 0);
6480 fill_filespec(two, null_oid(), 0, 0);
6481 p = diff_queue(q, one, two);
6482 p->status = DIFF_STATUS_MODIFIED;
6483 }
6484 }
6485
6486 /* Re-sort the filepairs */
6487 diffcore_fix_diff_index();
6488
6489 /* Cleanup */
6490 strset_clear(&present);
6491 }
6492
6493 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6494 {
6495 int i;
6496 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6497 struct diff_queue_struct *q = &diff_queued_diff;
6498
6499 if (WSEH_NEW & WS_RULE_MASK)
6500 BUG("WS rules bit mask overlaps with diff symbol flags");
6501
6502 if (o->color_moved)
6503 o->emitted_symbols = &esm;
6504
6505 if (o->additional_path_headers)
6506 create_filepairs_for_header_only_notifications(o);
6507
6508 for (i = 0; i < q->nr; i++) {
6509 struct diff_filepair *p = q->queue[i];
6510 if (check_pair_status(p))
6511 diff_flush_patch(p, o);
6512 }
6513
6514 if (o->emitted_symbols) {
6515 if (o->color_moved) {
6516 struct mem_pool entry_pool;
6517 struct moved_entry_list *entry_list;
6518
6519 mem_pool_init(&entry_pool, 1024 * 1024);
6520 entry_list = add_lines_to_move_detection(o,
6521 &entry_pool);
6522 mark_color_as_moved(o, entry_list);
6523 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6524 dim_moved_lines(o);
6525
6526 mem_pool_discard(&entry_pool, 0);
6527 free(entry_list);
6528 }
6529
6530 for (i = 0; i < esm.nr; i++)
6531 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6532
6533 for (i = 0; i < esm.nr; i++)
6534 free((void *)esm.buf[i].line);
6535 esm.nr = 0;
6536
6537 o->emitted_symbols = NULL;
6538 }
6539 }
6540
6541 static void diff_free_file(struct diff_options *options)
6542 {
6543 if (options->close_file)
6544 fclose(options->file);
6545 }
6546
6547 static void diff_free_ignore_regex(struct diff_options *options)
6548 {
6549 int i;
6550
6551 for (i = 0; i < options->ignore_regex_nr; i++) {
6552 regfree(options->ignore_regex[i]);
6553 free(options->ignore_regex[i]);
6554 }
6555 free(options->ignore_regex);
6556 }
6557
6558 void diff_free(struct diff_options *options)
6559 {
6560 if (options->no_free)
6561 return;
6562
6563 diff_free_file(options);
6564 diff_free_ignore_regex(options);
6565 clear_pathspec(&options->pathspec);
6566 }
6567
6568 void diff_flush(struct diff_options *options)
6569 {
6570 struct diff_queue_struct *q = &diff_queued_diff;
6571 int i, output_format = options->output_format;
6572 int separator = 0;
6573 int dirstat_by_line = 0;
6574
6575 /*
6576 * Order: raw, stat, summary, patch
6577 * or: name/name-status/checkdiff (other bits clear)
6578 */
6579 if (!q->nr && !options->additional_path_headers)
6580 goto free_queue;
6581
6582 if (output_format & (DIFF_FORMAT_RAW |
6583 DIFF_FORMAT_NAME |
6584 DIFF_FORMAT_NAME_STATUS |
6585 DIFF_FORMAT_CHECKDIFF)) {
6586 for (i = 0; i < q->nr; i++) {
6587 struct diff_filepair *p = q->queue[i];
6588 if (check_pair_status(p))
6589 flush_one_pair(p, options);
6590 }
6591 separator++;
6592 }
6593
6594 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6595 dirstat_by_line = 1;
6596
6597 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6598 dirstat_by_line) {
6599 struct diffstat_t diffstat;
6600
6601 compute_diffstat(options, &diffstat, q);
6602 if (output_format & DIFF_FORMAT_NUMSTAT)
6603 show_numstat(&diffstat, options);
6604 if (output_format & DIFF_FORMAT_DIFFSTAT)
6605 show_stats(&diffstat, options);
6606 if (output_format & DIFF_FORMAT_SHORTSTAT)
6607 show_shortstats(&diffstat, options);
6608 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6609 show_dirstat_by_line(&diffstat, options);
6610 free_diffstat_info(&diffstat);
6611 separator++;
6612 }
6613 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6614 show_dirstat(options);
6615
6616 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6617 for (i = 0; i < q->nr; i++) {
6618 diff_summary(options, q->queue[i]);
6619 }
6620 separator++;
6621 }
6622
6623 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6624 options->flags.exit_with_status &&
6625 options->flags.diff_from_contents) {
6626 /*
6627 * run diff_flush_patch for the exit status. setting
6628 * options->file to /dev/null should be safe, because we
6629 * aren't supposed to produce any output anyway.
6630 */
6631 diff_free_file(options);
6632 options->file = xfopen("/dev/null", "w");
6633 options->close_file = 1;
6634 options->color_moved = 0;
6635 for (i = 0; i < q->nr; i++) {
6636 struct diff_filepair *p = q->queue[i];
6637 if (check_pair_status(p))
6638 diff_flush_patch(p, options);
6639 if (options->found_changes)
6640 break;
6641 }
6642 }
6643
6644 if (output_format & DIFF_FORMAT_PATCH) {
6645 if (separator) {
6646 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6647 if (options->stat_sep)
6648 /* attach patch instead of inline */
6649 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6650 NULL, 0, 0);
6651 }
6652
6653 diff_flush_patch_all_file_pairs(options);
6654 }
6655
6656 if (output_format & DIFF_FORMAT_CALLBACK)
6657 options->format_callback(q, options, options->format_callback_data);
6658
6659 free_queue:
6660 diff_free_queue(q);
6661 DIFF_QUEUE_CLEAR(q);
6662 diff_free(options);
6663
6664 /*
6665 * Report the content-level differences with HAS_CHANGES;
6666 * diff_addremove/diff_change does not set the bit when
6667 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6668 */
6669 if (options->flags.diff_from_contents) {
6670 if (options->found_changes)
6671 options->flags.has_changes = 1;
6672 else
6673 options->flags.has_changes = 0;
6674 }
6675 }
6676
6677 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6678 {
6679 return (((p->status == DIFF_STATUS_MODIFIED) &&
6680 ((p->score &&
6681 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6682 (!p->score &&
6683 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6684 ((p->status != DIFF_STATUS_MODIFIED) &&
6685 filter_bit_tst(p->status, options)));
6686 }
6687
6688 static void diffcore_apply_filter(struct diff_options *options)
6689 {
6690 int i;
6691 struct diff_queue_struct *q = &diff_queued_diff;
6692 struct diff_queue_struct outq;
6693
6694 DIFF_QUEUE_CLEAR(&outq);
6695
6696 if (!options->filter)
6697 return;
6698
6699 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6700 int found;
6701 for (i = found = 0; !found && i < q->nr; i++) {
6702 if (match_filter(options, q->queue[i]))
6703 found++;
6704 }
6705 if (found)
6706 return;
6707
6708 /* otherwise we will clear the whole queue
6709 * by copying the empty outq at the end of this
6710 * function, but first clear the current entries
6711 * in the queue.
6712 */
6713 for (i = 0; i < q->nr; i++)
6714 diff_free_filepair(q->queue[i]);
6715 }
6716 else {
6717 /* Only the matching ones */
6718 for (i = 0; i < q->nr; i++) {
6719 struct diff_filepair *p = q->queue[i];
6720 if (match_filter(options, p))
6721 diff_q(&outq, p);
6722 else
6723 diff_free_filepair(p);
6724 }
6725 }
6726 free(q->queue);
6727 *q = outq;
6728 }
6729
6730 /* Check whether two filespecs with the same mode and size are identical */
6731 static int diff_filespec_is_identical(struct repository *r,
6732 struct diff_filespec *one,
6733 struct diff_filespec *two)
6734 {
6735 if (S_ISGITLINK(one->mode))
6736 return 0;
6737 if (diff_populate_filespec(r, one, NULL))
6738 return 0;
6739 if (diff_populate_filespec(r, two, NULL))
6740 return 0;
6741 return !memcmp(one->data, two->data, one->size);
6742 }
6743
6744 static int diff_filespec_check_stat_unmatch(struct repository *r,
6745 struct diff_filepair *p)
6746 {
6747 struct diff_populate_filespec_options dpf_options = {
6748 .check_size_only = 1,
6749 .missing_object_cb = diff_queued_diff_prefetch,
6750 .missing_object_data = r,
6751 };
6752
6753 if (p->done_skip_stat_unmatch)
6754 return p->skip_stat_unmatch_result;
6755
6756 p->done_skip_stat_unmatch = 1;
6757 p->skip_stat_unmatch_result = 0;
6758 /*
6759 * 1. Entries that come from stat info dirtiness
6760 * always have both sides (iow, not create/delete),
6761 * one side of the object name is unknown, with
6762 * the same mode and size. Keep the ones that
6763 * do not match these criteria. They have real
6764 * differences.
6765 *
6766 * 2. At this point, the file is known to be modified,
6767 * with the same mode and size, and the object
6768 * name of one side is unknown. Need to inspect
6769 * the identical contents.
6770 */
6771 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6772 !DIFF_FILE_VALID(p->two) ||
6773 (p->one->oid_valid && p->two->oid_valid) ||
6774 (p->one->mode != p->two->mode) ||
6775 diff_populate_filespec(r, p->one, &dpf_options) ||
6776 diff_populate_filespec(r, p->two, &dpf_options) ||
6777 (p->one->size != p->two->size) ||
6778 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6779 p->skip_stat_unmatch_result = 1;
6780 return p->skip_stat_unmatch_result;
6781 }
6782
6783 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6784 {
6785 int i;
6786 struct diff_queue_struct *q = &diff_queued_diff;
6787 struct diff_queue_struct outq;
6788 DIFF_QUEUE_CLEAR(&outq);
6789
6790 for (i = 0; i < q->nr; i++) {
6791 struct diff_filepair *p = q->queue[i];
6792
6793 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6794 diff_q(&outq, p);
6795 else {
6796 /*
6797 * The caller can subtract 1 from skip_stat_unmatch
6798 * to determine how many paths were dirty only
6799 * due to stat info mismatch.
6800 */
6801 if (!diffopt->flags.no_index)
6802 diffopt->skip_stat_unmatch++;
6803 diff_free_filepair(p);
6804 }
6805 }
6806 free(q->queue);
6807 *q = outq;
6808 }
6809
6810 static int diffnamecmp(const void *a_, const void *b_)
6811 {
6812 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6813 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6814 const char *name_a, *name_b;
6815
6816 name_a = a->one ? a->one->path : a->two->path;
6817 name_b = b->one ? b->one->path : b->two->path;
6818 return strcmp(name_a, name_b);
6819 }
6820
6821 void diffcore_fix_diff_index(void)
6822 {
6823 struct diff_queue_struct *q = &diff_queued_diff;
6824 QSORT(q->queue, q->nr, diffnamecmp);
6825 }
6826
6827 void diff_add_if_missing(struct repository *r,
6828 struct oid_array *to_fetch,
6829 const struct diff_filespec *filespec)
6830 {
6831 if (filespec && filespec->oid_valid &&
6832 !S_ISGITLINK(filespec->mode) &&
6833 oid_object_info_extended(r, &filespec->oid, NULL,
6834 OBJECT_INFO_FOR_PREFETCH))
6835 oid_array_append(to_fetch, &filespec->oid);
6836 }
6837
6838 void diff_queued_diff_prefetch(void *repository)
6839 {
6840 struct repository *repo = repository;
6841 int i;
6842 struct diff_queue_struct *q = &diff_queued_diff;
6843 struct oid_array to_fetch = OID_ARRAY_INIT;
6844
6845 for (i = 0; i < q->nr; i++) {
6846 struct diff_filepair *p = q->queue[i];
6847 diff_add_if_missing(repo, &to_fetch, p->one);
6848 diff_add_if_missing(repo, &to_fetch, p->two);
6849 }
6850
6851 /*
6852 * NEEDSWORK: Consider deduplicating the OIDs sent.
6853 */
6854 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6855
6856 oid_array_clear(&to_fetch);
6857 }
6858
6859 void diffcore_std(struct diff_options *options)
6860 {
6861 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6862 DIFF_FORMAT_NUMSTAT |
6863 DIFF_FORMAT_PATCH |
6864 DIFF_FORMAT_SHORTSTAT |
6865 DIFF_FORMAT_DIRSTAT;
6866
6867 /*
6868 * Check if the user requested a blob-data-requiring diff output and/or
6869 * break-rewrite detection (which requires blob data). If yes, prefetch
6870 * the diff pairs.
6871 *
6872 * If no prefetching occurs, diffcore_rename() will prefetch if it
6873 * decides that it needs inexact rename detection.
6874 */
6875 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
6876 (options->output_format & output_formats_to_prefetch ||
6877 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6878 diff_queued_diff_prefetch(options->repo);
6879
6880 /* NOTE please keep the following in sync with diff_tree_combined() */
6881 if (options->skip_stat_unmatch)
6882 diffcore_skip_stat_unmatch(options);
6883 if (!options->found_follow) {
6884 /* See try_to_follow_renames() in tree-diff.c */
6885 if (options->break_opt != -1)
6886 diffcore_break(options->repo,
6887 options->break_opt);
6888 if (options->detect_rename)
6889 diffcore_rename(options);
6890 if (options->break_opt != -1)
6891 diffcore_merge_broken();
6892 }
6893 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6894 diffcore_pickaxe(options);
6895 if (options->orderfile)
6896 diffcore_order(options->orderfile);
6897 if (options->rotate_to)
6898 diffcore_rotate(options);
6899 if (!options->found_follow)
6900 /* See try_to_follow_renames() in tree-diff.c */
6901 diff_resolve_rename_copy();
6902 diffcore_apply_filter(options);
6903
6904 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6905 options->flags.has_changes = 1;
6906 else
6907 options->flags.has_changes = 0;
6908
6909 options->found_follow = 0;
6910 }
6911
6912 int diff_result_code(struct diff_options *opt, int status)
6913 {
6914 int result = 0;
6915
6916 diff_warn_rename_limit("diff.renameLimit",
6917 opt->needed_rename_limit,
6918 opt->degraded_cc_to_c);
6919 if (!opt->flags.exit_with_status &&
6920 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6921 return status;
6922 if (opt->flags.exit_with_status &&
6923 opt->flags.has_changes)
6924 result |= 01;
6925 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6926 opt->flags.check_failed)
6927 result |= 02;
6928 return result;
6929 }
6930
6931 int diff_can_quit_early(struct diff_options *opt)
6932 {
6933 return (opt->flags.quick &&
6934 !opt->filter &&
6935 opt->flags.has_changes);
6936 }
6937
6938 /*
6939 * Shall changes to this submodule be ignored?
6940 *
6941 * Submodule changes can be configured to be ignored separately for each path,
6942 * but that configuration can be overridden from the command line.
6943 */
6944 static int is_submodule_ignored(const char *path, struct diff_options *options)
6945 {
6946 int ignored = 0;
6947 struct diff_flags orig_flags = options->flags;
6948 if (!options->flags.override_submodule_config)
6949 set_diffopt_flags_from_submodule_config(options, path);
6950 if (options->flags.ignore_submodules)
6951 ignored = 1;
6952 options->flags = orig_flags;
6953 return ignored;
6954 }
6955
6956 void compute_diffstat(struct diff_options *options,
6957 struct diffstat_t *diffstat,
6958 struct diff_queue_struct *q)
6959 {
6960 int i;
6961
6962 memset(diffstat, 0, sizeof(struct diffstat_t));
6963 for (i = 0; i < q->nr; i++) {
6964 struct diff_filepair *p = q->queue[i];
6965 if (check_pair_status(p))
6966 diff_flush_stat(p, options, diffstat);
6967 }
6968 }
6969
6970 void diff_addremove(struct diff_options *options,
6971 int addremove, unsigned mode,
6972 const struct object_id *oid,
6973 int oid_valid,
6974 const char *concatpath, unsigned dirty_submodule)
6975 {
6976 struct diff_filespec *one, *two;
6977
6978 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6979 return;
6980
6981 /* This may look odd, but it is a preparation for
6982 * feeding "there are unchanged files which should
6983 * not produce diffs, but when you are doing copy
6984 * detection you would need them, so here they are"
6985 * entries to the diff-core. They will be prefixed
6986 * with something like '=' or '*' (I haven't decided
6987 * which but should not make any difference).
6988 * Feeding the same new and old to diff_change()
6989 * also has the same effect.
6990 * Before the final output happens, they are pruned after
6991 * merged into rename/copy pairs as appropriate.
6992 */
6993 if (options->flags.reverse_diff)
6994 addremove = (addremove == '+' ? '-' :
6995 addremove == '-' ? '+' : addremove);
6996
6997 if (options->prefix &&
6998 strncmp(concatpath, options->prefix, options->prefix_length))
6999 return;
7000
7001 one = alloc_filespec(concatpath);
7002 two = alloc_filespec(concatpath);
7003
7004 if (addremove != '+')
7005 fill_filespec(one, oid, oid_valid, mode);
7006 if (addremove != '-') {
7007 fill_filespec(two, oid, oid_valid, mode);
7008 two->dirty_submodule = dirty_submodule;
7009 }
7010
7011 diff_queue(&diff_queued_diff, one, two);
7012 if (!options->flags.diff_from_contents)
7013 options->flags.has_changes = 1;
7014 }
7015
7016 void diff_change(struct diff_options *options,
7017 unsigned old_mode, unsigned new_mode,
7018 const struct object_id *old_oid,
7019 const struct object_id *new_oid,
7020 int old_oid_valid, int new_oid_valid,
7021 const char *concatpath,
7022 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7023 {
7024 struct diff_filespec *one, *two;
7025 struct diff_filepair *p;
7026
7027 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7028 is_submodule_ignored(concatpath, options))
7029 return;
7030
7031 if (options->flags.reverse_diff) {
7032 SWAP(old_mode, new_mode);
7033 SWAP(old_oid, new_oid);
7034 SWAP(old_oid_valid, new_oid_valid);
7035 SWAP(old_dirty_submodule, new_dirty_submodule);
7036 }
7037
7038 if (options->prefix &&
7039 strncmp(concatpath, options->prefix, options->prefix_length))
7040 return;
7041
7042 one = alloc_filespec(concatpath);
7043 two = alloc_filespec(concatpath);
7044 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7045 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7046 one->dirty_submodule = old_dirty_submodule;
7047 two->dirty_submodule = new_dirty_submodule;
7048 p = diff_queue(&diff_queued_diff, one, two);
7049
7050 if (options->flags.diff_from_contents)
7051 return;
7052
7053 if (options->flags.quick && options->skip_stat_unmatch &&
7054 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7055 diff_free_filespec_data(p->one);
7056 diff_free_filespec_data(p->two);
7057 return;
7058 }
7059
7060 options->flags.has_changes = 1;
7061 }
7062
7063 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7064 {
7065 struct diff_filepair *pair;
7066 struct diff_filespec *one, *two;
7067
7068 if (options->prefix &&
7069 strncmp(path, options->prefix, options->prefix_length))
7070 return NULL;
7071
7072 one = alloc_filespec(path);
7073 two = alloc_filespec(path);
7074 pair = diff_queue(&diff_queued_diff, one, two);
7075 pair->is_unmerged = 1;
7076 return pair;
7077 }
7078
7079 static char *run_textconv(struct repository *r,
7080 const char *pgm,
7081 struct diff_filespec *spec,
7082 size_t *outsize)
7083 {
7084 struct diff_tempfile *temp;
7085 struct child_process child = CHILD_PROCESS_INIT;
7086 struct strbuf buf = STRBUF_INIT;
7087 int err = 0;
7088
7089 temp = prepare_temp_file(r, spec);
7090 strvec_push(&child.args, pgm);
7091 strvec_push(&child.args, temp->name);
7092
7093 child.use_shell = 1;
7094 child.out = -1;
7095 if (start_command(&child)) {
7096 remove_tempfile();
7097 return NULL;
7098 }
7099
7100 if (strbuf_read(&buf, child.out, 0) < 0)
7101 err = error("error reading from textconv command '%s'", pgm);
7102 close(child.out);
7103
7104 if (finish_command(&child) || err) {
7105 strbuf_release(&buf);
7106 remove_tempfile();
7107 return NULL;
7108 }
7109 remove_tempfile();
7110
7111 return strbuf_detach(&buf, outsize);
7112 }
7113
7114 size_t fill_textconv(struct repository *r,
7115 struct userdiff_driver *driver,
7116 struct diff_filespec *df,
7117 char **outbuf)
7118 {
7119 size_t size;
7120
7121 if (!driver) {
7122 if (!DIFF_FILE_VALID(df)) {
7123 *outbuf = "";
7124 return 0;
7125 }
7126 if (diff_populate_filespec(r, df, NULL))
7127 die("unable to read files to diff");
7128 *outbuf = df->data;
7129 return df->size;
7130 }
7131
7132 if (!driver->textconv)
7133 BUG("fill_textconv called with non-textconv driver");
7134
7135 if (driver->textconv_cache && df->oid_valid) {
7136 *outbuf = notes_cache_get(driver->textconv_cache,
7137 &df->oid,
7138 &size);
7139 if (*outbuf)
7140 return size;
7141 }
7142
7143 *outbuf = run_textconv(r, driver->textconv, df, &size);
7144 if (!*outbuf)
7145 die("unable to read files to diff");
7146
7147 if (driver->textconv_cache && df->oid_valid) {
7148 /* ignore errors, as we might be in a readonly repository */
7149 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7150 size);
7151 /*
7152 * we could save up changes and flush them all at the end,
7153 * but we would need an extra call after all diffing is done.
7154 * Since generating a cache entry is the slow path anyway,
7155 * this extra overhead probably isn't a big deal.
7156 */
7157 notes_cache_write(driver->textconv_cache);
7158 }
7159
7160 return size;
7161 }
7162
7163 int textconv_object(struct repository *r,
7164 const char *path,
7165 unsigned mode,
7166 const struct object_id *oid,
7167 int oid_valid,
7168 char **buf,
7169 unsigned long *buf_size)
7170 {
7171 struct diff_filespec *df;
7172 struct userdiff_driver *textconv;
7173
7174 df = alloc_filespec(path);
7175 fill_filespec(df, oid, oid_valid, mode);
7176 textconv = get_textconv(r, df);
7177 if (!textconv) {
7178 free_filespec(df);
7179 return 0;
7180 }
7181
7182 *buf_size = fill_textconv(r, textconv, df, buf);
7183 free_filespec(df);
7184 return 1;
7185 }
7186
7187 void setup_diff_pager(struct diff_options *opt)
7188 {
7189 /*
7190 * If the user asked for our exit code, then either they want --quiet
7191 * or --exit-code. We should definitely not bother with a pager in the
7192 * former case, as we will generate no output. Since we still properly
7193 * report our exit code even when a pager is run, we _could_ run a
7194 * pager with --exit-code. But since we have not done so historically,
7195 * and because it is easy to find people oneline advising "git diff
7196 * --exit-code" in hooks and other scripts, we do not do so.
7197 */
7198 if (!opt->flags.exit_with_status &&
7199 check_pager_config("diff") != 0)
7200 setup_pager();
7201 }
The Best VCS Around