?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
unicode: update the width tables to Unicode 15.1
[git.git] / diff.c
blob469e18aed20ed0128bfa85774cb617d455287eaa
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "config.h"
6 #include "tempfile.h"
7 #include "quote.h"
8 #include "diff.h"
9 #include "diffcore.h"
10 #include "delta.h"
11 #include "xdiff-interface.h"
12 #include "color.h"
13 #include "attr.h"
14 #include "run-command.h"
15 #include "utf8.h"
16 #include "object-store.h"
17 #include "userdiff.h"
18 #include "submodule-config.h"
19 #include "submodule.h"
20 #include "hashmap.h"
21 #include "mem-pool.h"
22 #include "ll-merge.h"
23 #include "string-list.h"
24 #include "strvec.h"
25 #include "graph.h"
26 #include "packfile.h"
27 #include "parse-options.h"
28 #include "help.h"
29 #include "promisor-remote.h"
30 #include "dir.h"
31 #include "strmap.h"
32
33 #ifdef NO_FAST_WORKING_DIRECTORY
34 #define FAST_WORKING_DIRECTORY 0
35 #else
36 #define FAST_WORKING_DIRECTORY 1
37 #endif
38
39 static int diff_detect_rename_default;
40 static int diff_indent_heuristic = 1;
41 static int diff_rename_limit_default = 1000;
42 static int diff_suppress_blank_empty;
43 static int diff_use_color_default = -1;
44 static int diff_color_moved_default;
45 static int diff_color_moved_ws_default;
46 static int diff_context_default = 3;
47 static int diff_interhunk_context_default;
48 static const char *diff_word_regex_cfg;
49 static const char *external_diff_cmd_cfg;
50 static const char *diff_order_file_cfg;
51 int diff_auto_refresh_index = 1;
52 static int diff_mnemonic_prefix;
53 static int diff_no_prefix;
54 static int diff_relative;
55 static int diff_stat_graph_width;
56 static int diff_dirstat_permille_default = 30;
57 static struct diff_options default_diff_options;
58 static long diff_algorithm;
59 static unsigned ws_error_highlight_default = WSEH_NEW;
60
61 static char diff_colors[][COLOR_MAXLEN] = {
62 GIT_COLOR_RESET,
63 GIT_COLOR_NORMAL, /* CONTEXT */
64 GIT_COLOR_BOLD, /* METAINFO */
65 GIT_COLOR_CYAN, /* FRAGINFO */
66 GIT_COLOR_RED, /* OLD */
67 GIT_COLOR_GREEN, /* NEW */
68 GIT_COLOR_YELLOW, /* COMMIT */
69 GIT_COLOR_BG_RED, /* WHITESPACE */
70 GIT_COLOR_NORMAL, /* FUNCINFO */
71 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
72 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
73 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
74 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
75 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
76 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
77 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
78 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
79 GIT_COLOR_FAINT, /* CONTEXT_DIM */
80 GIT_COLOR_FAINT_RED, /* OLD_DIM */
81 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
82 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
83 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
84 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
85 };
86
87 static const char *color_diff_slots[] = {
88 [DIFF_CONTEXT] = "context",
89 [DIFF_METAINFO] = "meta",
90 [DIFF_FRAGINFO] = "frag",
91 [DIFF_FILE_OLD] = "old",
92 [DIFF_FILE_NEW] = "new",
93 [DIFF_COMMIT] = "commit",
94 [DIFF_WHITESPACE] = "whitespace",
95 [DIFF_FUNCINFO] = "func",
96 [DIFF_FILE_OLD_MOVED] = "oldMoved",
97 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
98 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
99 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
100 [DIFF_FILE_NEW_MOVED] = "newMoved",
101 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
102 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
103 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
104 [DIFF_CONTEXT_DIM] = "contextDimmed",
105 [DIFF_FILE_OLD_DIM] = "oldDimmed",
106 [DIFF_FILE_NEW_DIM] = "newDimmed",
107 [DIFF_CONTEXT_BOLD] = "contextBold",
108 [DIFF_FILE_OLD_BOLD] = "oldBold",
109 [DIFF_FILE_NEW_BOLD] = "newBold",
110 };
111
112 define_list_config_array_extra(color_diff_slots, {"plain"});
113
114 static int parse_diff_color_slot(const char *var)
115 {
116 if (!strcasecmp(var, "plain"))
117 return DIFF_CONTEXT;
118 return LOOKUP_CONFIG(color_diff_slots, var);
119 }
120
121 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
122 struct strbuf *errmsg)
123 {
124 char *params_copy = xstrdup(params_string);
125 struct string_list params = STRING_LIST_INIT_NODUP;
126 int ret = 0;
127 int i;
128
129 if (*params_copy)
130 string_list_split_in_place(&params, params_copy, ',', -1);
131 for (i = 0; i < params.nr; i++) {
132 const char *p = params.items[i].string;
133 if (!strcmp(p, "changes")) {
134 options->flags.dirstat_by_line = 0;
135 options->flags.dirstat_by_file = 0;
136 } else if (!strcmp(p, "lines")) {
137 options->flags.dirstat_by_line = 1;
138 options->flags.dirstat_by_file = 0;
139 } else if (!strcmp(p, "files")) {
140 options->flags.dirstat_by_line = 0;
141 options->flags.dirstat_by_file = 1;
142 } else if (!strcmp(p, "noncumulative")) {
143 options->flags.dirstat_cumulative = 0;
144 } else if (!strcmp(p, "cumulative")) {
145 options->flags.dirstat_cumulative = 1;
146 } else if (isdigit(*p)) {
147 char *end;
148 int permille = strtoul(p, &end, 10) * 10;
149 if (*end == '.' && isdigit(*++end)) {
150 /* only use first digit */
151 permille += *end - '0';
152 /* .. and ignore any further digits */
153 while (isdigit(*++end))
154 ; /* nothing */
155 }
156 if (!*end)
157 options->dirstat_permille = permille;
158 else {
159 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
160 p);
161 ret++;
162 }
163 } else {
164 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
165 ret++;
166 }
167
168 }
169 string_list_clear(&params, 0);
170 free(params_copy);
171 return ret;
172 }
173
174 static int parse_submodule_params(struct diff_options *options, const char *value)
175 {
176 if (!strcmp(value, "log"))
177 options->submodule_format = DIFF_SUBMODULE_LOG;
178 else if (!strcmp(value, "short"))
179 options->submodule_format = DIFF_SUBMODULE_SHORT;
180 else if (!strcmp(value, "diff"))
181 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
182 /*
183 * Please update $__git_diff_submodule_formats in
184 * git-completion.bash when you add new formats.
185 */
186 else
187 return -1;
188 return 0;
189 }
190
191 int git_config_rename(const char *var, const char *value)
192 {
193 if (!value)
194 return DIFF_DETECT_RENAME;
195 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
196 return DIFF_DETECT_COPY;
197 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
198 }
199
200 long parse_algorithm_value(const char *value)
201 {
202 if (!value)
203 return -1;
204 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
205 return 0;
206 else if (!strcasecmp(value, "minimal"))
207 return XDF_NEED_MINIMAL;
208 else if (!strcasecmp(value, "patience"))
209 return XDF_PATIENCE_DIFF;
210 else if (!strcasecmp(value, "histogram"))
211 return XDF_HISTOGRAM_DIFF;
212 /*
213 * Please update $__git_diff_algorithms in git-completion.bash
214 * when you add new algorithms.
215 */
216 return -1;
217 }
218
219 static int parse_one_token(const char **arg, const char *token)
220 {
221 const char *rest;
222 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
223 *arg = rest;
224 return 1;
225 }
226 return 0;
227 }
228
229 static int parse_ws_error_highlight(const char *arg)
230 {
231 const char *orig_arg = arg;
232 unsigned val = 0;
233
234 while (*arg) {
235 if (parse_one_token(&arg, "none"))
236 val = 0;
237 else if (parse_one_token(&arg, "default"))
238 val = WSEH_NEW;
239 else if (parse_one_token(&arg, "all"))
240 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
241 else if (parse_one_token(&arg, "new"))
242 val |= WSEH_NEW;
243 else if (parse_one_token(&arg, "old"))
244 val |= WSEH_OLD;
245 else if (parse_one_token(&arg, "context"))
246 val |= WSEH_CONTEXT;
247 else {
248 return -1 - (int)(arg - orig_arg);
249 }
250 if (*arg)
251 arg++;
252 }
253 return val;
254 }
255
256 /*
257 * These are to give UI layer defaults.
258 * The core-level commands such as git-diff-files should
259 * never be affected by the setting of diff.renames
260 * the user happens to have in the configuration file.
261 */
262 void init_diff_ui_defaults(void)
263 {
264 diff_detect_rename_default = DIFF_DETECT_RENAME;
265 }
266
267 int git_diff_heuristic_config(const char *var, const char *value,
268 void *cb UNUSED)
269 {
270 if (!strcmp(var, "diff.indentheuristic"))
271 diff_indent_heuristic = git_config_bool(var, value);
272 return 0;
273 }
274
275 static int parse_color_moved(const char *arg)
276 {
277 switch (git_parse_maybe_bool(arg)) {
278 case 0:
279 return COLOR_MOVED_NO;
280 case 1:
281 return COLOR_MOVED_DEFAULT;
282 default:
283 break;
284 }
285
286 if (!strcmp(arg, "no"))
287 return COLOR_MOVED_NO;
288 else if (!strcmp(arg, "plain"))
289 return COLOR_MOVED_PLAIN;
290 else if (!strcmp(arg, "blocks"))
291 return COLOR_MOVED_BLOCKS;
292 else if (!strcmp(arg, "zebra"))
293 return COLOR_MOVED_ZEBRA;
294 else if (!strcmp(arg, "default"))
295 return COLOR_MOVED_DEFAULT;
296 else if (!strcmp(arg, "dimmed-zebra"))
297 return COLOR_MOVED_ZEBRA_DIM;
298 else if (!strcmp(arg, "dimmed_zebra"))
299 return COLOR_MOVED_ZEBRA_DIM;
300 else
301 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
302 }
303
304 static unsigned parse_color_moved_ws(const char *arg)
305 {
306 int ret = 0;
307 struct string_list l = STRING_LIST_INIT_DUP;
308 struct string_list_item *i;
309
310 string_list_split(&l, arg, ',', -1);
311
312 for_each_string_list_item(i, &l) {
313 struct strbuf sb = STRBUF_INIT;
314 strbuf_addstr(&sb, i->string);
315 strbuf_trim(&sb);
316
317 if (!strcmp(sb.buf, "no"))
318 ret = 0;
319 else if (!strcmp(sb.buf, "ignore-space-change"))
320 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
321 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
322 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
323 else if (!strcmp(sb.buf, "ignore-all-space"))
324 ret |= XDF_IGNORE_WHITESPACE;
325 else if (!strcmp(sb.buf, "allow-indentation-change"))
326 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
327 else {
328 ret |= COLOR_MOVED_WS_ERROR;
329 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);
330 }
331
332 strbuf_release(&sb);
333 }
334
335 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
336 (ret & XDF_WHITESPACE_FLAGS)) {
337 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
338 ret |= COLOR_MOVED_WS_ERROR;
339 }
340
341 string_list_clear(&l, 0);
342
343 return ret;
344 }
345
346 int git_diff_ui_config(const char *var, const char *value, void *cb)
347 {
348 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
349 diff_use_color_default = git_config_colorbool(var, value);
350 return 0;
351 }
352 if (!strcmp(var, "diff.colormoved")) {
353 int cm = parse_color_moved(value);
354 if (cm < 0)
355 return -1;
356 diff_color_moved_default = cm;
357 return 0;
358 }
359 if (!strcmp(var, "diff.colormovedws")) {
360 unsigned cm = parse_color_moved_ws(value);
361 if (cm & COLOR_MOVED_WS_ERROR)
362 return -1;
363 diff_color_moved_ws_default = cm;
364 return 0;
365 }
366 if (!strcmp(var, "diff.context")) {
367 diff_context_default = git_config_int(var, value);
368 if (diff_context_default < 0)
369 return -1;
370 return 0;
371 }
372 if (!strcmp(var, "diff.interhunkcontext")) {
373 diff_interhunk_context_default = git_config_int(var, value);
374 if (diff_interhunk_context_default < 0)
375 return -1;
376 return 0;
377 }
378 if (!strcmp(var, "diff.renames")) {
379 diff_detect_rename_default = git_config_rename(var, value);
380 return 0;
381 }
382 if (!strcmp(var, "diff.autorefreshindex")) {
383 diff_auto_refresh_index = git_config_bool(var, value);
384 return 0;
385 }
386 if (!strcmp(var, "diff.mnemonicprefix")) {
387 diff_mnemonic_prefix = git_config_bool(var, value);
388 return 0;
389 }
390 if (!strcmp(var, "diff.noprefix")) {
391 diff_no_prefix = git_config_bool(var, value);
392 return 0;
393 }
394 if (!strcmp(var, "diff.relative")) {
395 diff_relative = git_config_bool(var, value);
396 return 0;
397 }
398 if (!strcmp(var, "diff.statgraphwidth")) {
399 diff_stat_graph_width = git_config_int(var, value);
400 return 0;
401 }
402 if (!strcmp(var, "diff.external"))
403 return git_config_string(&external_diff_cmd_cfg, var, value);
404 if (!strcmp(var, "diff.wordregex"))
405 return git_config_string(&diff_word_regex_cfg, var, value);
406 if (!strcmp(var, "diff.orderfile"))
407 return git_config_pathname(&diff_order_file_cfg, var, value);
408
409 if (!strcmp(var, "diff.ignoresubmodules"))
410 handle_ignore_submodules_arg(&default_diff_options, value);
411
412 if (!strcmp(var, "diff.submodule")) {
413 if (parse_submodule_params(&default_diff_options, value))
414 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
415 value);
416 return 0;
417 }
418
419 if (!strcmp(var, "diff.algorithm")) {
420 diff_algorithm = parse_algorithm_value(value);
421 if (diff_algorithm < 0)
422 return -1;
423 return 0;
424 }
425
426 if (git_color_config(var, value, cb) < 0)
427 return -1;
428
429 return git_diff_basic_config(var, value, cb);
430 }
431
432 int git_diff_basic_config(const char *var, const char *value, void *cb)
433 {
434 const char *name;
435
436 if (!strcmp(var, "diff.renamelimit")) {
437 diff_rename_limit_default = git_config_int(var, value);
438 return 0;
439 }
440
441 if (userdiff_config(var, value) < 0)
442 return -1;
443
444 if (skip_prefix(var, "diff.color.", &name) ||
445 skip_prefix(var, "color.diff.", &name)) {
446 int slot = parse_diff_color_slot(name);
447 if (slot < 0)
448 return 0;
449 if (!value)
450 return config_error_nonbool(var);
451 return color_parse(value, diff_colors[slot]);
452 }
453
454 if (!strcmp(var, "diff.wserrorhighlight")) {
455 int val = parse_ws_error_highlight(value);
456 if (val < 0)
457 return -1;
458 ws_error_highlight_default = val;
459 return 0;
460 }
461
462 /* like GNU diff's --suppress-blank-empty option */
463 if (!strcmp(var, "diff.suppressblankempty") ||
464 /* for backwards compatibility */
465 !strcmp(var, "diff.suppress-blank-empty")) {
466 diff_suppress_blank_empty = git_config_bool(var, value);
467 return 0;
468 }
469
470 if (!strcmp(var, "diff.dirstat")) {
471 struct strbuf errmsg = STRBUF_INIT;
472 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
473 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
474 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
475 errmsg.buf);
476 strbuf_release(&errmsg);
477 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
478 return 0;
479 }
480
481 if (git_diff_heuristic_config(var, value, cb) < 0)
482 return -1;
483
484 return git_default_config(var, value, cb);
485 }
486
487 static char *quote_two(const char *one, const char *two)
488 {
489 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
490 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
491 struct strbuf res = STRBUF_INIT;
492
493 if (need_one + need_two) {
494 strbuf_addch(&res, '"');
495 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
496 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
497 strbuf_addch(&res, '"');
498 } else {
499 strbuf_addstr(&res, one);
500 strbuf_addstr(&res, two);
501 }
502 return strbuf_detach(&res, NULL);
503 }
504
505 static const char *external_diff(void)
506 {
507 static const char *external_diff_cmd = NULL;
508 static int done_preparing = 0;
509
510 if (done_preparing)
511 return external_diff_cmd;
512 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
513 if (!external_diff_cmd)
514 external_diff_cmd = external_diff_cmd_cfg;
515 done_preparing = 1;
516 return external_diff_cmd;
517 }
518
519 /*
520 * Keep track of files used for diffing. Sometimes such an entry
521 * refers to a temporary file, sometimes to an existing file, and
522 * sometimes to "/dev/null".
523 */
524 static struct diff_tempfile {
525 /*
526 * filename external diff should read from, or NULL if this
527 * entry is currently not in use:
528 */
529 const char *name;
530
531 char hex[GIT_MAX_HEXSZ + 1];
532 char mode[10];
533
534 /*
535 * If this diff_tempfile instance refers to a temporary file,
536 * this tempfile object is used to manage its lifetime.
537 */
538 struct tempfile *tempfile;
539 } diff_temp[2];
540
541 struct emit_callback {
542 int color_diff;
543 unsigned ws_rule;
544 int blank_at_eof_in_preimage;
545 int blank_at_eof_in_postimage;
546 int lno_in_preimage;
547 int lno_in_postimage;
548 const char **label_path;
549 struct diff_words_data *diff_words;
550 struct diff_options *opt;
551 struct strbuf *header;
552 };
553
554 static int count_lines(const char *data, int size)
555 {
556 int count, ch, completely_empty = 1, nl_just_seen = 0;
557 count = 0;
558 while (0 < size--) {
559 ch = *data++;
560 if (ch == '\n') {
561 count++;
562 nl_just_seen = 1;
563 completely_empty = 0;
564 }
565 else {
566 nl_just_seen = 0;
567 completely_empty = 0;
568 }
569 }
570 if (completely_empty)
571 return 0;
572 if (!nl_just_seen)
573 count++; /* no trailing newline */
574 return count;
575 }
576
577 static int fill_mmfile(struct repository *r, mmfile_t *mf,
578 struct diff_filespec *one)
579 {
580 if (!DIFF_FILE_VALID(one)) {
581 mf->ptr = (char *)""; /* does not matter */
582 mf->size = 0;
583 return 0;
584 }
585 else if (diff_populate_filespec(r, one, NULL))
586 return -1;
587
588 mf->ptr = one->data;
589 mf->size = one->size;
590 return 0;
591 }
592
593 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
594 static unsigned long diff_filespec_size(struct repository *r,
595 struct diff_filespec *one)
596 {
597 struct diff_populate_filespec_options dpf_options = {
598 .check_size_only = 1,
599 };
600
601 if (!DIFF_FILE_VALID(one))
602 return 0;
603 diff_populate_filespec(r, one, &dpf_options);
604 return one->size;
605 }
606
607 static int count_trailing_blank(mmfile_t *mf)
608 {
609 char *ptr = mf->ptr;
610 long size = mf->size;
611 int cnt = 0;
612
613 if (!size)
614 return cnt;
615 ptr += size - 1; /* pointing at the very end */
616 if (*ptr != '\n')
617 ; /* incomplete line */
618 else
619 ptr--; /* skip the last LF */
620 while (mf->ptr < ptr) {
621 char *prev_eol;
622 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
623 if (*prev_eol == '\n')
624 break;
625 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
626 break;
627 cnt++;
628 ptr = prev_eol - 1;
629 }
630 return cnt;
631 }
632
633 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
634 struct emit_callback *ecbdata)
635 {
636 int l1, l2, at;
637 l1 = count_trailing_blank(mf1);
638 l2 = count_trailing_blank(mf2);
639 if (l2 <= l1) {
640 ecbdata->blank_at_eof_in_preimage = 0;
641 ecbdata->blank_at_eof_in_postimage = 0;
642 return;
643 }
644 at = count_lines(mf1->ptr, mf1->size);
645 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
646
647 at = count_lines(mf2->ptr, mf2->size);
648 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
649 }
650
651 static void emit_line_0(struct diff_options *o,
652 const char *set_sign, const char *set, unsigned reverse, const char *reset,
653 int first, const char *line, int len)
654 {
655 int has_trailing_newline, has_trailing_carriage_return;
656 int needs_reset = 0; /* at the end of the line */
657 FILE *file = o->file;
658
659 fputs(diff_line_prefix(o), file);
660
661 has_trailing_newline = (len > 0 && line[len-1] == '\n');
662 if (has_trailing_newline)
663 len--;
664
665 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
666 if (has_trailing_carriage_return)
667 len--;
668
669 if (!len && !first)
670 goto end_of_line;
671
672 if (reverse && want_color(o->use_color)) {
673 fputs(GIT_COLOR_REVERSE, file);
674 needs_reset = 1;
675 }
676
677 if (set_sign) {
678 fputs(set_sign, file);
679 needs_reset = 1;
680 }
681
682 if (first)
683 fputc(first, file);
684
685 if (!len)
686 goto end_of_line;
687
688 if (set) {
689 if (set_sign && set != set_sign)
690 fputs(reset, file);
691 fputs(set, file);
692 needs_reset = 1;
693 }
694 fwrite(line, len, 1, file);
695 needs_reset = 1; /* 'line' may contain color codes. */
696
697 end_of_line:
698 if (needs_reset)
699 fputs(reset, file);
700 if (has_trailing_carriage_return)
701 fputc('\r', file);
702 if (has_trailing_newline)
703 fputc('\n', file);
704 }
705
706 static void emit_line(struct diff_options *o, const char *set, const char *reset,
707 const char *line, int len)
708 {
709 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
710 }
711
712 enum diff_symbol {
713 DIFF_SYMBOL_BINARY_DIFF_HEADER,
714 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
715 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
716 DIFF_SYMBOL_BINARY_DIFF_BODY,
717 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
718 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
719 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
720 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
721 DIFF_SYMBOL_STATS_LINE,
722 DIFF_SYMBOL_WORD_DIFF,
723 DIFF_SYMBOL_STAT_SEP,
724 DIFF_SYMBOL_SUMMARY,
725 DIFF_SYMBOL_SUBMODULE_ADD,
726 DIFF_SYMBOL_SUBMODULE_DEL,
727 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
728 DIFF_SYMBOL_SUBMODULE_MODIFIED,
729 DIFF_SYMBOL_SUBMODULE_HEADER,
730 DIFF_SYMBOL_SUBMODULE_ERROR,
731 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
732 DIFF_SYMBOL_REWRITE_DIFF,
733 DIFF_SYMBOL_BINARY_FILES,
734 DIFF_SYMBOL_HEADER,
735 DIFF_SYMBOL_FILEPAIR_PLUS,
736 DIFF_SYMBOL_FILEPAIR_MINUS,
737 DIFF_SYMBOL_WORDS_PORCELAIN,
738 DIFF_SYMBOL_WORDS,
739 DIFF_SYMBOL_CONTEXT,
740 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
741 DIFF_SYMBOL_PLUS,
742 DIFF_SYMBOL_MINUS,
743 DIFF_SYMBOL_NO_LF_EOF,
744 DIFF_SYMBOL_CONTEXT_FRAGINFO,
745 DIFF_SYMBOL_CONTEXT_MARKER,
746 DIFF_SYMBOL_SEPARATOR
747 };
748 /*
749 * Flags for content lines:
750 * 0..12 are whitespace rules
751 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
752 * 16 is marking if the line is blank at EOF
753 */
754 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
755 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
756 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
757 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
758 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
759
760 /*
761 * This struct is used when we need to buffer the output of the diff output.
762 *
763 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
764 * into the pre/post image file. This pointer could be a union with the
765 * line pointer. By storing an offset into the file instead of the literal line,
766 * we can decrease the memory footprint for the buffered output. At first we
767 * may want to only have indirection for the content lines, but we could also
768 * enhance the state for emitting prefabricated lines, e.g. the similarity
769 * score line or hunk/file headers would only need to store a number or path
770 * and then the output can be constructed later on depending on state.
771 */
772 struct emitted_diff_symbol {
773 const char *line;
774 int len;
775 int flags;
776 int indent_off; /* Offset to first non-whitespace character */
777 int indent_width; /* The visual width of the indentation */
778 unsigned id;
779 enum diff_symbol s;
780 };
781 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
782
783 struct emitted_diff_symbols {
784 struct emitted_diff_symbol *buf;
785 int nr, alloc;
786 };
787 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
788
789 static void append_emitted_diff_symbol(struct diff_options *o,
790 struct emitted_diff_symbol *e)
791 {
792 struct emitted_diff_symbol *f;
793
794 ALLOC_GROW(o->emitted_symbols->buf,
795 o->emitted_symbols->nr + 1,
796 o->emitted_symbols->alloc);
797 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
798
799 memcpy(f, e, sizeof(struct emitted_diff_symbol));
800 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
801 }
802
803 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
804 {
805 if (!e)
806 return;
807 free(e->buf);
808 free(e);
809 }
810
811 struct moved_entry {
812 const struct emitted_diff_symbol *es;
813 struct moved_entry *next_line;
814 struct moved_entry *next_match;
815 };
816
817 struct moved_block {
818 struct moved_entry *match;
819 int wsd; /* The whitespace delta of this block */
820 };
821
822 #define INDENT_BLANKLINE INT_MIN
823
824 static void fill_es_indent_data(struct emitted_diff_symbol *es)
825 {
826 unsigned int off = 0, i;
827 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
828 const char *s = es->line;
829 const int len = es->len;
830
831 /* skip any \v \f \r at start of indentation */
832 while (s[off] == '\f' || s[off] == '\v' ||
833 (s[off] == '\r' && off < len - 1))
834 off++;
835
836 /* calculate the visual width of indentation */
837 while(1) {
838 if (s[off] == ' ') {
839 width++;
840 off++;
841 } else if (s[off] == '\t') {
842 width += tab_width - (width % tab_width);
843 while (s[++off] == '\t')
844 width += tab_width;
845 } else {
846 break;
847 }
848 }
849
850 /* check if this line is blank */
851 for (i = off; i < len; i++)
852 if (!isspace(s[i]))
853 break;
854
855 if (i == len) {
856 es->indent_width = INDENT_BLANKLINE;
857 es->indent_off = len;
858 } else {
859 es->indent_off = off;
860 es->indent_width = width;
861 }
862 }
863
864 static int compute_ws_delta(const struct emitted_diff_symbol *a,
865 const struct emitted_diff_symbol *b)
866 {
867 int a_width = a->indent_width,
868 b_width = b->indent_width;
869
870 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
871 return INDENT_BLANKLINE;
872
873 return a_width - b_width;
874 }
875
876 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
877 const struct emitted_diff_symbol *l,
878 struct moved_block *pmb)
879 {
880 int a_width = cur->es->indent_width, b_width = l->indent_width;
881 int delta;
882
883 /* The text of each line must match */
884 if (cur->es->id != l->id)
885 return 1;
886
887 /*
888 * If 'l' and 'cur' are both blank then we don't need to check the
889 * indent. We only need to check cur as we know the strings match.
890 * */
891 if (a_width == INDENT_BLANKLINE)
892 return 0;
893
894 /*
895 * The indent changes of the block are known and stored in pmb->wsd;
896 * however we need to check if the indent changes of the current line
897 * match those of the current block.
898 */
899 delta = b_width - a_width;
900
901 /*
902 * If the previous lines of this block were all blank then set its
903 * whitespace delta.
904 */
905 if (pmb->wsd == INDENT_BLANKLINE)
906 pmb->wsd = delta;
907
908 return delta != pmb->wsd;
909 }
910
911 struct interned_diff_symbol {
912 struct hashmap_entry ent;
913 struct emitted_diff_symbol *es;
914 };
915
916 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
917 const struct hashmap_entry *eptr,
918 const struct hashmap_entry *entry_or_key,
919 const void *keydata UNUSED)
920 {
921 const struct diff_options *diffopt = hashmap_cmp_fn_data;
922 const struct emitted_diff_symbol *a, *b;
923 unsigned flags = diffopt->color_moved_ws_handling
924 & XDF_WHITESPACE_FLAGS;
925
926 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
927 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
928
929 return !xdiff_compare_lines(a->line + a->indent_off,
930 a->len - a->indent_off,
931 b->line + b->indent_off,
932 b->len - b->indent_off, flags);
933 }
934
935 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
936 struct interned_diff_symbol *s)
937 {
938 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
939 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
940 l->len - l->indent_off, flags);
941
942 hashmap_entry_init(&s->ent, hash);
943 s->es = l;
944 }
945
946 struct moved_entry_list {
947 struct moved_entry *add, *del;
948 };
949
950 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
951 struct mem_pool *entry_mem_pool)
952 {
953 struct moved_entry *prev_line = NULL;
954 struct mem_pool interned_pool;
955 struct hashmap interned_map;
956 struct moved_entry_list *entry_list = NULL;
957 size_t entry_list_alloc = 0;
958 unsigned id = 0;
959 int n;
960
961 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
962 mem_pool_init(&interned_pool, 1024 * 1024);
963
964 for (n = 0; n < o->emitted_symbols->nr; n++) {
965 struct interned_diff_symbol key;
966 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
967 struct interned_diff_symbol *s;
968 struct moved_entry *entry;
969
970 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
971 prev_line = NULL;
972 continue;
973 }
974
975 if (o->color_moved_ws_handling &
976 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
977 fill_es_indent_data(l);
978
979 prepare_entry(o, l, &key);
980 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
981 if (s) {
982 l->id = s->es->id;
983 } else {
984 l->id = id;
985 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
986 hashmap_add(&interned_map,
987 memcpy(mem_pool_alloc(&interned_pool,
988 sizeof(key)),
989 &key, sizeof(key)));
990 }
991 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
992 entry->es = l;
993 entry->next_line = NULL;
994 if (prev_line && prev_line->es->s == l->s)
995 prev_line->next_line = entry;
996 prev_line = entry;
997 if (l->s == DIFF_SYMBOL_PLUS) {
998 entry->next_match = entry_list[l->id].add;
999 entry_list[l->id].add = entry;
1000 } else {
1001 entry->next_match = entry_list[l->id].del;
1002 entry_list[l->id].del = entry;
1003 }
1004 }
1005
1006 hashmap_clear(&interned_map);
1007 mem_pool_discard(&interned_pool, 0);
1008
1009 return entry_list;
1010 }
1011
1012 static void pmb_advance_or_null(struct diff_options *o,
1013 struct emitted_diff_symbol *l,
1014 struct moved_block *pmb,
1015 int *pmb_nr)
1016 {
1017 int i, j;
1018
1019 for (i = 0, j = 0; i < *pmb_nr; i++) {
1020 int match;
1021 struct moved_entry *prev = pmb[i].match;
1022 struct moved_entry *cur = (prev && prev->next_line) ?
1023 prev->next_line : NULL;
1024
1025 if (o->color_moved_ws_handling &
1026 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1027 match = cur &&
1028 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1029 else
1030 match = cur && cur->es->id == l->id;
1031
1032 if (match) {
1033 pmb[j] = pmb[i];
1034 pmb[j++].match = cur;
1035 }
1036 }
1037 *pmb_nr = j;
1038 }
1039
1040 static void fill_potential_moved_blocks(struct diff_options *o,
1041 struct moved_entry *match,
1042 struct emitted_diff_symbol *l,
1043 struct moved_block **pmb_p,
1044 int *pmb_alloc_p, int *pmb_nr_p)
1045
1046 {
1047 struct moved_block *pmb = *pmb_p;
1048 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1049
1050 /*
1051 * The current line is the start of a new block.
1052 * Setup the set of potential blocks.
1053 */
1054 for (; match; match = match->next_match) {
1055 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1056 if (o->color_moved_ws_handling &
1057 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1058 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1059 else
1060 pmb[pmb_nr].wsd = 0;
1061 pmb[pmb_nr++].match = match;
1062 }
1063
1064 *pmb_p = pmb;
1065 *pmb_alloc_p = pmb_alloc;
1066 *pmb_nr_p = pmb_nr;
1067 }
1068
1069 /*
1070 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1071 *
1072 * Otherwise, if the last block has fewer alphanumeric characters than
1073 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1074 * that block.
1075 *
1076 * The last block consists of the (n - block_length)'th line up to but not
1077 * including the nth line.
1078 *
1079 * Returns 0 if the last block is empty or is unset by this function, non zero
1080 * otherwise.
1081 *
1082 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1083 * Think of a way to unify them.
1084 */
1085 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1086 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1087 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1088 {
1089 int i, alnum_count = 0;
1090 if (o->color_moved == COLOR_MOVED_PLAIN)
1091 return block_length;
1092 for (i = 1; i < block_length + 1; i++) {
1093 const char *c = o->emitted_symbols->buf[n - i].line;
1094 for (; *c; c++) {
1095 if (!isalnum(*c))
1096 continue;
1097 alnum_count++;
1098 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1099 return 1;
1100 }
1101 }
1102 for (i = 1; i < block_length + 1; i++)
1103 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1104 return 0;
1105 }
1106
1107 /* Find blocks of moved code, delegate actual coloring decision to helper */
1108 static void mark_color_as_moved(struct diff_options *o,
1109 struct moved_entry_list *entry_list)
1110 {
1111 struct moved_block *pmb = NULL; /* potentially moved blocks */
1112 int pmb_nr = 0, pmb_alloc = 0;
1113 int n, flipped_block = 0, block_length = 0;
1114 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1115
1116
1117 for (n = 0; n < o->emitted_symbols->nr; n++) {
1118 struct moved_entry *match = NULL;
1119 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1120
1121 switch (l->s) {
1122 case DIFF_SYMBOL_PLUS:
1123 match = entry_list[l->id].del;
1124 break;
1125 case DIFF_SYMBOL_MINUS:
1126 match = entry_list[l->id].add;
1127 break;
1128 default:
1129 flipped_block = 0;
1130 }
1131
1132 if (pmb_nr && (!match || l->s != moved_symbol)) {
1133 if (!adjust_last_block(o, n, block_length) &&
1134 block_length > 1) {
1135 /*
1136 * Rewind in case there is another match
1137 * starting at the second line of the block
1138 */
1139 match = NULL;
1140 n -= block_length;
1141 }
1142 pmb_nr = 0;
1143 block_length = 0;
1144 flipped_block = 0;
1145 }
1146 if (!match) {
1147 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1148 continue;
1149 }
1150
1151 if (o->color_moved == COLOR_MOVED_PLAIN) {
1152 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1153 continue;
1154 }
1155
1156 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1157
1158 if (pmb_nr == 0) {
1159 int contiguous = adjust_last_block(o, n, block_length);
1160
1161 if (!contiguous && block_length > 1)
1162 /*
1163 * Rewind in case there is another match
1164 * starting at the second line of the block
1165 */
1166 n -= block_length;
1167 else
1168 fill_potential_moved_blocks(o, match, l,
1169 &pmb, &pmb_alloc,
1170 &pmb_nr);
1171
1172 if (contiguous && pmb_nr && moved_symbol == l->s)
1173 flipped_block = (flipped_block + 1) % 2;
1174 else
1175 flipped_block = 0;
1176
1177 if (pmb_nr)
1178 moved_symbol = l->s;
1179 else
1180 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1181
1182 block_length = 0;
1183 }
1184
1185 if (pmb_nr) {
1186 block_length++;
1187 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1188 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1189 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1190 }
1191 }
1192 adjust_last_block(o, n, block_length);
1193
1194 free(pmb);
1195 }
1196
1197 static void dim_moved_lines(struct diff_options *o)
1198 {
1199 int n;
1200 for (n = 0; n < o->emitted_symbols->nr; n++) {
1201 struct emitted_diff_symbol *prev = (n != 0) ?
1202 &o->emitted_symbols->buf[n - 1] : NULL;
1203 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1204 struct emitted_diff_symbol *next =
1205 (n < o->emitted_symbols->nr - 1) ?
1206 &o->emitted_symbols->buf[n + 1] : NULL;
1207
1208 /* Not a plus or minus line? */
1209 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1210 continue;
1211
1212 /* Not a moved line? */
1213 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1214 continue;
1215
1216 /*
1217 * If prev or next are not a plus or minus line,
1218 * pretend they don't exist
1219 */
1220 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1221 prev->s != DIFF_SYMBOL_MINUS)
1222 prev = NULL;
1223 if (next && next->s != DIFF_SYMBOL_PLUS &&
1224 next->s != DIFF_SYMBOL_MINUS)
1225 next = NULL;
1226
1227 /* Inside a block? */
1228 if ((prev &&
1229 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1230 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1231 (next &&
1232 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1233 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1234 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1235 continue;
1236 }
1237
1238 /* Check if we are at an interesting bound: */
1239 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1240 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1241 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1242 continue;
1243 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1244 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1245 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1246 continue;
1247
1248 /*
1249 * The boundary to prev and next are not interesting,
1250 * so this line is not interesting as a whole
1251 */
1252 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1253 }
1254 }
1255
1256 static void emit_line_ws_markup(struct diff_options *o,
1257 const char *set_sign, const char *set,
1258 const char *reset,
1259 int sign_index, const char *line, int len,
1260 unsigned ws_rule, int blank_at_eof)
1261 {
1262 const char *ws = NULL;
1263 int sign = o->output_indicators[sign_index];
1264
1265 if (o->ws_error_highlight & ws_rule) {
1266 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1267 if (!*ws)
1268 ws = NULL;
1269 }
1270
1271 if (!ws && !set_sign)
1272 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1273 else if (!ws) {
1274 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1275 } else if (blank_at_eof)
1276 /* Blank line at EOF - paint '+' as well */
1277 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1278 else {
1279 /* Emit just the prefix, then the rest. */
1280 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1281 sign, "", 0);
1282 ws_check_emit(line, len, ws_rule,
1283 o->file, set, reset, ws);
1284 }
1285 }
1286
1287 static void emit_diff_symbol_from_struct(struct diff_options *o,
1288 struct emitted_diff_symbol *eds)
1289 {
1290 static const char *nneof = " No newline at end of file\n";
1291 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1292
1293 enum diff_symbol s = eds->s;
1294 const char *line = eds->line;
1295 int len = eds->len;
1296 unsigned flags = eds->flags;
1297
1298 switch (s) {
1299 case DIFF_SYMBOL_NO_LF_EOF:
1300 context = diff_get_color_opt(o, DIFF_CONTEXT);
1301 reset = diff_get_color_opt(o, DIFF_RESET);
1302 putc('\n', o->file);
1303 emit_line_0(o, context, NULL, 0, reset, '\\',
1304 nneof, strlen(nneof));
1305 break;
1306 case DIFF_SYMBOL_SUBMODULE_HEADER:
1307 case DIFF_SYMBOL_SUBMODULE_ERROR:
1308 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1309 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1310 case DIFF_SYMBOL_SUMMARY:
1311 case DIFF_SYMBOL_STATS_LINE:
1312 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1313 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1314 emit_line(o, "", "", line, len);
1315 break;
1316 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1317 case DIFF_SYMBOL_CONTEXT_MARKER:
1318 context = diff_get_color_opt(o, DIFF_CONTEXT);
1319 reset = diff_get_color_opt(o, DIFF_RESET);
1320 emit_line(o, context, reset, line, len);
1321 break;
1322 case DIFF_SYMBOL_SEPARATOR:
1323 fprintf(o->file, "%s%c",
1324 diff_line_prefix(o),
1325 o->line_termination);
1326 break;
1327 case DIFF_SYMBOL_CONTEXT:
1328 set = diff_get_color_opt(o, DIFF_CONTEXT);
1329 reset = diff_get_color_opt(o, DIFF_RESET);
1330 set_sign = NULL;
1331 if (o->flags.dual_color_diffed_diffs) {
1332 char c = !len ? 0 : line[0];
1333
1334 if (c == '+')
1335 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1336 else if (c == '@')
1337 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1338 else if (c == '-')
1339 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1340 }
1341 emit_line_ws_markup(o, set_sign, set, reset,
1342 OUTPUT_INDICATOR_CONTEXT, line, len,
1343 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1344 break;
1345 case DIFF_SYMBOL_PLUS:
1346 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1347 DIFF_SYMBOL_MOVED_LINE_ALT |
1348 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1349 case DIFF_SYMBOL_MOVED_LINE |
1350 DIFF_SYMBOL_MOVED_LINE_ALT |
1351 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1352 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1353 break;
1354 case DIFF_SYMBOL_MOVED_LINE |
1355 DIFF_SYMBOL_MOVED_LINE_ALT:
1356 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1357 break;
1358 case DIFF_SYMBOL_MOVED_LINE |
1359 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1360 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1361 break;
1362 case DIFF_SYMBOL_MOVED_LINE:
1363 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1364 break;
1365 default:
1366 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1367 }
1368 reset = diff_get_color_opt(o, DIFF_RESET);
1369 if (!o->flags.dual_color_diffed_diffs)
1370 set_sign = NULL;
1371 else {
1372 char c = !len ? 0 : line[0];
1373
1374 set_sign = set;
1375 if (c == '-')
1376 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1377 else if (c == '@')
1378 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1379 else if (c == '+')
1380 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1381 else
1382 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1383 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1384 }
1385 emit_line_ws_markup(o, set_sign, set, reset,
1386 OUTPUT_INDICATOR_NEW, line, len,
1387 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1388 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1389 break;
1390 case DIFF_SYMBOL_MINUS:
1391 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1392 DIFF_SYMBOL_MOVED_LINE_ALT |
1393 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1394 case DIFF_SYMBOL_MOVED_LINE |
1395 DIFF_SYMBOL_MOVED_LINE_ALT |
1396 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1397 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1398 break;
1399 case DIFF_SYMBOL_MOVED_LINE |
1400 DIFF_SYMBOL_MOVED_LINE_ALT:
1401 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1402 break;
1403 case DIFF_SYMBOL_MOVED_LINE |
1404 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1405 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1406 break;
1407 case DIFF_SYMBOL_MOVED_LINE:
1408 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1409 break;
1410 default:
1411 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1412 }
1413 reset = diff_get_color_opt(o, DIFF_RESET);
1414 if (!o->flags.dual_color_diffed_diffs)
1415 set_sign = NULL;
1416 else {
1417 char c = !len ? 0 : line[0];
1418
1419 set_sign = set;
1420 if (c == '+')
1421 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1422 else if (c == '@')
1423 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1424 else if (c == '-')
1425 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1426 else
1427 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1428 }
1429 emit_line_ws_markup(o, set_sign, set, reset,
1430 OUTPUT_INDICATOR_OLD, line, len,
1431 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1432 break;
1433 case DIFF_SYMBOL_WORDS_PORCELAIN:
1434 context = diff_get_color_opt(o, DIFF_CONTEXT);
1435 reset = diff_get_color_opt(o, DIFF_RESET);
1436 emit_line(o, context, reset, line, len);
1437 fputs("~\n", o->file);
1438 break;
1439 case DIFF_SYMBOL_WORDS:
1440 context = diff_get_color_opt(o, DIFF_CONTEXT);
1441 reset = diff_get_color_opt(o, DIFF_RESET);
1442 /*
1443 * Skip the prefix character, if any. With
1444 * diff_suppress_blank_empty, there may be
1445 * none.
1446 */
1447 if (line[0] != '\n') {
1448 line++;
1449 len--;
1450 }
1451 emit_line(o, context, reset, line, len);
1452 break;
1453 case DIFF_SYMBOL_FILEPAIR_PLUS:
1454 meta = diff_get_color_opt(o, DIFF_METAINFO);
1455 reset = diff_get_color_opt(o, DIFF_RESET);
1456 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1457 line, reset,
1458 strchr(line, ' ') ? "\t" : "");
1459 break;
1460 case DIFF_SYMBOL_FILEPAIR_MINUS:
1461 meta = diff_get_color_opt(o, DIFF_METAINFO);
1462 reset = diff_get_color_opt(o, DIFF_RESET);
1463 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1464 line, reset,
1465 strchr(line, ' ') ? "\t" : "");
1466 break;
1467 case DIFF_SYMBOL_BINARY_FILES:
1468 case DIFF_SYMBOL_HEADER:
1469 fprintf(o->file, "%s", line);
1470 break;
1471 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1472 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1473 break;
1474 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1475 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1476 break;
1477 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1478 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1479 break;
1480 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1481 fputs(diff_line_prefix(o), o->file);
1482 fputc('\n', o->file);
1483 break;
1484 case DIFF_SYMBOL_REWRITE_DIFF:
1485 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1486 reset = diff_get_color_opt(o, DIFF_RESET);
1487 emit_line(o, fraginfo, reset, line, len);
1488 break;
1489 case DIFF_SYMBOL_SUBMODULE_ADD:
1490 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1491 reset = diff_get_color_opt(o, DIFF_RESET);
1492 emit_line(o, set, reset, line, len);
1493 break;
1494 case DIFF_SYMBOL_SUBMODULE_DEL:
1495 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1496 reset = diff_get_color_opt(o, DIFF_RESET);
1497 emit_line(o, set, reset, line, len);
1498 break;
1499 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1500 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1501 diff_line_prefix(o), line);
1502 break;
1503 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1504 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1505 diff_line_prefix(o), line);
1506 break;
1507 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1508 emit_line(o, "", "", " 0 files changed\n",
1509 strlen(" 0 files changed\n"));
1510 break;
1511 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1512 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1513 break;
1514 case DIFF_SYMBOL_WORD_DIFF:
1515 fprintf(o->file, "%.*s", len, line);
1516 break;
1517 case DIFF_SYMBOL_STAT_SEP:
1518 fputs(o->stat_sep, o->file);
1519 break;
1520 default:
1521 BUG("unknown diff symbol");
1522 }
1523 }
1524
1525 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1526 const char *line, int len, unsigned flags)
1527 {
1528 struct emitted_diff_symbol e = {
1529 .line = line, .len = len, .flags = flags, .s = s
1530 };
1531
1532 if (o->emitted_symbols)
1533 append_emitted_diff_symbol(o, &e);
1534 else
1535 emit_diff_symbol_from_struct(o, &e);
1536 }
1537
1538 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1539 {
1540 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1541 }
1542
1543 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1544 {
1545 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1546 }
1547
1548 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1549 {
1550 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1551 path, strlen(path), 0);
1552 }
1553
1554 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1555 {
1556 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1557 path, strlen(path), 0);
1558 }
1559
1560 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1561 {
1562 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1563 header, strlen(header), 0);
1564 }
1565
1566 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1567 {
1568 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1569 }
1570
1571 void diff_emit_submodule_pipethrough(struct diff_options *o,
1572 const char *line, int len)
1573 {
1574 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1575 }
1576
1577 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1578 {
1579 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1580 ecbdata->blank_at_eof_in_preimage &&
1581 ecbdata->blank_at_eof_in_postimage &&
1582 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1583 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1584 return 0;
1585 return ws_blank_line(line, len);
1586 }
1587
1588 static void emit_add_line(struct emit_callback *ecbdata,
1589 const char *line, int len)
1590 {
1591 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1592 if (new_blank_line_at_eof(ecbdata, line, len))
1593 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1594
1595 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1596 }
1597
1598 static void emit_del_line(struct emit_callback *ecbdata,
1599 const char *line, int len)
1600 {
1601 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1602 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1603 }
1604
1605 static void emit_context_line(struct emit_callback *ecbdata,
1606 const char *line, int len)
1607 {
1608 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1609 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1610 }
1611
1612 static void emit_hunk_header(struct emit_callback *ecbdata,
1613 const char *line, int len)
1614 {
1615 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1616 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1617 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1618 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1619 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1620 static const char atat[2] = { '@', '@' };
1621 const char *cp, *ep;
1622 struct strbuf msgbuf = STRBUF_INIT;
1623 int org_len = len;
1624 int i = 1;
1625
1626 /*
1627 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1628 * it always is at least 10 bytes long.
1629 */
1630 if (len < 10 ||
1631 memcmp(line, atat, 2) ||
1632 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1633 emit_diff_symbol(ecbdata->opt,
1634 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1635 return;
1636 }
1637 ep += 2; /* skip over @@ */
1638
1639 /* The hunk header in fraginfo color */
1640 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1641 strbuf_addstr(&msgbuf, reverse);
1642 strbuf_addstr(&msgbuf, frag);
1643 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1644 strbuf_add(&msgbuf, atat, sizeof(atat));
1645 else
1646 strbuf_add(&msgbuf, line, ep - line);
1647 strbuf_addstr(&msgbuf, reset);
1648
1649 /*
1650 * trailing "\r\n"
1651 */
1652 for ( ; i < 3; i++)
1653 if (line[len - i] == '\r' || line[len - i] == '\n')
1654 len--;
1655
1656 /* blank before the func header */
1657 for (cp = ep; ep - line < len; ep++)
1658 if (*ep != ' ' && *ep != '\t')
1659 break;
1660 if (ep != cp) {
1661 strbuf_addstr(&msgbuf, context);
1662 strbuf_add(&msgbuf, cp, ep - cp);
1663 strbuf_addstr(&msgbuf, reset);
1664 }
1665
1666 if (ep < line + len) {
1667 strbuf_addstr(&msgbuf, func);
1668 strbuf_add(&msgbuf, ep, line + len - ep);
1669 strbuf_addstr(&msgbuf, reset);
1670 }
1671
1672 strbuf_add(&msgbuf, line + len, org_len - len);
1673 strbuf_complete_line(&msgbuf);
1674 emit_diff_symbol(ecbdata->opt,
1675 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1676 strbuf_release(&msgbuf);
1677 }
1678
1679 static struct diff_tempfile *claim_diff_tempfile(void)
1680 {
1681 int i;
1682 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1683 if (!diff_temp[i].name)
1684 return diff_temp + i;
1685 BUG("diff is failing to clean up its tempfiles");
1686 }
1687
1688 static void remove_tempfile(void)
1689 {
1690 int i;
1691 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1692 if (is_tempfile_active(diff_temp[i].tempfile))
1693 delete_tempfile(&diff_temp[i].tempfile);
1694 diff_temp[i].name = NULL;
1695 }
1696 }
1697
1698 static void add_line_count(struct strbuf *out, int count)
1699 {
1700 switch (count) {
1701 case 0:
1702 strbuf_addstr(out, "0,0");
1703 break;
1704 case 1:
1705 strbuf_addstr(out, "1");
1706 break;
1707 default:
1708 strbuf_addf(out, "1,%d", count);
1709 break;
1710 }
1711 }
1712
1713 static void emit_rewrite_lines(struct emit_callback *ecb,
1714 int prefix, const char *data, int size)
1715 {
1716 const char *endp = NULL;
1717
1718 while (0 < size) {
1719 int len;
1720
1721 endp = memchr(data, '\n', size);
1722 len = endp ? (endp - data + 1) : size;
1723 if (prefix != '+') {
1724 ecb->lno_in_preimage++;
1725 emit_del_line(ecb, data, len);
1726 } else {
1727 ecb->lno_in_postimage++;
1728 emit_add_line(ecb, data, len);
1729 }
1730 size -= len;
1731 data += len;
1732 }
1733 if (!endp)
1734 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1735 }
1736
1737 static void emit_rewrite_diff(const char *name_a,
1738 const char *name_b,
1739 struct diff_filespec *one,
1740 struct diff_filespec *two,
1741 struct userdiff_driver *textconv_one,
1742 struct userdiff_driver *textconv_two,
1743 struct diff_options *o)
1744 {
1745 int lc_a, lc_b;
1746 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1747 const char *a_prefix, *b_prefix;
1748 char *data_one, *data_two;
1749 size_t size_one, size_two;
1750 struct emit_callback ecbdata;
1751 struct strbuf out = STRBUF_INIT;
1752
1753 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1754 a_prefix = o->b_prefix;
1755 b_prefix = o->a_prefix;
1756 } else {
1757 a_prefix = o->a_prefix;
1758 b_prefix = o->b_prefix;
1759 }
1760
1761 name_a += (*name_a == '/');
1762 name_b += (*name_b == '/');
1763
1764 strbuf_reset(&a_name);
1765 strbuf_reset(&b_name);
1766 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1767 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1768
1769 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1770 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1771
1772 memset(&ecbdata, 0, sizeof(ecbdata));
1773 ecbdata.color_diff = want_color(o->use_color);
1774 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1775 ecbdata.opt = o;
1776 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1777 mmfile_t mf1, mf2;
1778 mf1.ptr = (char *)data_one;
1779 mf2.ptr = (char *)data_two;
1780 mf1.size = size_one;
1781 mf2.size = size_two;
1782 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1783 }
1784 ecbdata.lno_in_preimage = 1;
1785 ecbdata.lno_in_postimage = 1;
1786
1787 lc_a = count_lines(data_one, size_one);
1788 lc_b = count_lines(data_two, size_two);
1789
1790 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1791 a_name.buf, a_name.len, 0);
1792 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1793 b_name.buf, b_name.len, 0);
1794
1795 strbuf_addstr(&out, "@@ -");
1796 if (!o->irreversible_delete)
1797 add_line_count(&out, lc_a);
1798 else
1799 strbuf_addstr(&out, "?,?");
1800 strbuf_addstr(&out, " +");
1801 add_line_count(&out, lc_b);
1802 strbuf_addstr(&out, " @@\n");
1803 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1804 strbuf_release(&out);
1805
1806 if (lc_a && !o->irreversible_delete)
1807 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1808 if (lc_b)
1809 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1810 if (textconv_one)
1811 free((char *)data_one);
1812 if (textconv_two)
1813 free((char *)data_two);
1814 }
1815
1816 struct diff_words_buffer {
1817 mmfile_t text;
1818 unsigned long alloc;
1819 struct diff_words_orig {
1820 const char *begin, *end;
1821 } *orig;
1822 int orig_nr, orig_alloc;
1823 };
1824
1825 static void diff_words_append(char *line, unsigned long len,
1826 struct diff_words_buffer *buffer)
1827 {
1828 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1829 line++;
1830 len--;
1831 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1832 buffer->text.size += len;
1833 buffer->text.ptr[buffer->text.size] = '\0';
1834 }
1835
1836 struct diff_words_style_elem {
1837 const char *prefix;
1838 const char *suffix;
1839 const char *color; /* NULL; filled in by the setup code if
1840 * color is enabled */
1841 };
1842
1843 struct diff_words_style {
1844 enum diff_words_type type;
1845 struct diff_words_style_elem new_word, old_word, ctx;
1846 const char *newline;
1847 };
1848
1849 static struct diff_words_style diff_words_styles[] = {
1850 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1851 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1852 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1853 };
1854
1855 struct diff_words_data {
1856 struct diff_words_buffer minus, plus;
1857 const char *current_plus;
1858 int last_minus;
1859 struct diff_options *opt;
1860 regex_t *word_regex;
1861 enum diff_words_type type;
1862 struct diff_words_style *style;
1863 };
1864
1865 static int fn_out_diff_words_write_helper(struct diff_options *o,
1866 struct diff_words_style_elem *st_el,
1867 const char *newline,
1868 size_t count, const char *buf)
1869 {
1870 int print = 0;
1871 struct strbuf sb = STRBUF_INIT;
1872
1873 while (count) {
1874 char *p = memchr(buf, '\n', count);
1875 if (print)
1876 strbuf_addstr(&sb, diff_line_prefix(o));
1877
1878 if (p != buf) {
1879 const char *reset = st_el->color && *st_el->color ?
1880 GIT_COLOR_RESET : NULL;
1881 if (st_el->color && *st_el->color)
1882 strbuf_addstr(&sb, st_el->color);
1883 strbuf_addstr(&sb, st_el->prefix);
1884 strbuf_add(&sb, buf, p ? p - buf : count);
1885 strbuf_addstr(&sb, st_el->suffix);
1886 if (reset)
1887 strbuf_addstr(&sb, reset);
1888 }
1889 if (!p)
1890 goto out;
1891
1892 strbuf_addstr(&sb, newline);
1893 count -= p + 1 - buf;
1894 buf = p + 1;
1895 print = 1;
1896 if (count) {
1897 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1898 sb.buf, sb.len, 0);
1899 strbuf_reset(&sb);
1900 }
1901 }
1902
1903 out:
1904 if (sb.len)
1905 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1906 sb.buf, sb.len, 0);
1907 strbuf_release(&sb);
1908 return 0;
1909 }
1910
1911 /*
1912 * '--color-words' algorithm can be described as:
1913 *
1914 * 1. collect the minus/plus lines of a diff hunk, divided into
1915 * minus-lines and plus-lines;
1916 *
1917 * 2. break both minus-lines and plus-lines into words and
1918 * place them into two mmfile_t with one word for each line;
1919 *
1920 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1921 *
1922 * And for the common parts of the both file, we output the plus side text.
1923 * diff_words->current_plus is used to trace the current position of the plus file
1924 * which printed. diff_words->last_minus is used to trace the last minus word
1925 * printed.
1926 *
1927 * For '--graph' to work with '--color-words', we need to output the graph prefix
1928 * on each line of color words output. Generally, there are two conditions on
1929 * which we should output the prefix.
1930 *
1931 * 1. diff_words->last_minus == 0 &&
1932 * diff_words->current_plus == diff_words->plus.text.ptr
1933 *
1934 * that is: the plus text must start as a new line, and if there is no minus
1935 * word printed, a graph prefix must be printed.
1936 *
1937 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1938 * *(diff_words->current_plus - 1) == '\n'
1939 *
1940 * that is: a graph prefix must be printed following a '\n'
1941 */
1942 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1943 {
1944 if ((diff_words->last_minus == 0 &&
1945 diff_words->current_plus == diff_words->plus.text.ptr) ||
1946 (diff_words->current_plus > diff_words->plus.text.ptr &&
1947 *(diff_words->current_plus - 1) == '\n')) {
1948 return 1;
1949 } else {
1950 return 0;
1951 }
1952 }
1953
1954 static void fn_out_diff_words_aux(void *priv,
1955 long minus_first, long minus_len,
1956 long plus_first, long plus_len,
1957 const char *func UNUSED, long funclen UNUSED)
1958 {
1959 struct diff_words_data *diff_words = priv;
1960 struct diff_words_style *style = diff_words->style;
1961 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1962 struct diff_options *opt = diff_words->opt;
1963 const char *line_prefix;
1964
1965 assert(opt);
1966 line_prefix = diff_line_prefix(opt);
1967
1968 /* POSIX requires that first be decremented by one if len == 0... */
1969 if (minus_len) {
1970 minus_begin = diff_words->minus.orig[minus_first].begin;
1971 minus_end =
1972 diff_words->minus.orig[minus_first + minus_len - 1].end;
1973 } else
1974 minus_begin = minus_end =
1975 diff_words->minus.orig[minus_first].end;
1976
1977 if (plus_len) {
1978 plus_begin = diff_words->plus.orig[plus_first].begin;
1979 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1980 } else
1981 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1982
1983 if (color_words_output_graph_prefix(diff_words)) {
1984 fputs(line_prefix, diff_words->opt->file);
1985 }
1986 if (diff_words->current_plus != plus_begin) {
1987 fn_out_diff_words_write_helper(diff_words->opt,
1988 &style->ctx, style->newline,
1989 plus_begin - diff_words->current_plus,
1990 diff_words->current_plus);
1991 }
1992 if (minus_begin != minus_end) {
1993 fn_out_diff_words_write_helper(diff_words->opt,
1994 &style->old_word, style->newline,
1995 minus_end - minus_begin, minus_begin);
1996 }
1997 if (plus_begin != plus_end) {
1998 fn_out_diff_words_write_helper(diff_words->opt,
1999 &style->new_word, style->newline,
2000 plus_end - plus_begin, plus_begin);
2001 }
2002
2003 diff_words->current_plus = plus_end;
2004 diff_words->last_minus = minus_first;
2005 }
2006
2007 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2008 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2009 int *begin, int *end)
2010 {
2011 while (word_regex && *begin < buffer->size) {
2012 regmatch_t match[1];
2013 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2014 buffer->size - *begin, 1, match, 0)) {
2015 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2016 '\n', match[0].rm_eo - match[0].rm_so);
2017 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2018 *begin += match[0].rm_so;
2019 if (*begin == *end)
2020 (*begin)++;
2021 else
2022 return *begin > *end;
2023 } else {
2024 return -1;
2025 }
2026 }
2027
2028 /* find the next word */
2029 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2030 (*begin)++;
2031 if (*begin >= buffer->size)
2032 return -1;
2033
2034 /* find the end of the word */
2035 *end = *begin + 1;
2036 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2037 (*end)++;
2038
2039 return 0;
2040 }
2041
2042 /*
2043 * This function splits the words in buffer->text, stores the list with
2044 * newline separator into out, and saves the offsets of the original words
2045 * in buffer->orig.
2046 */
2047 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2048 regex_t *word_regex)
2049 {
2050 int i, j;
2051 long alloc = 0;
2052
2053 out->size = 0;
2054 out->ptr = NULL;
2055
2056 /* fake an empty "0th" word */
2057 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2058 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2059 buffer->orig_nr = 1;
2060
2061 for (i = 0; i < buffer->text.size; i++) {
2062 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2063 return;
2064
2065 /* store original boundaries */
2066 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2067 buffer->orig_alloc);
2068 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2069 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2070 buffer->orig_nr++;
2071
2072 /* store one word */
2073 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2074 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2075 out->ptr[out->size + j - i] = '\n';
2076 out->size += j - i + 1;
2077
2078 i = j - 1;
2079 }
2080 }
2081
2082 /* this executes the word diff on the accumulated buffers */
2083 static void diff_words_show(struct diff_words_data *diff_words)
2084 {
2085 xpparam_t xpp;
2086 xdemitconf_t xecfg;
2087 mmfile_t minus, plus;
2088 struct diff_words_style *style = diff_words->style;
2089
2090 struct diff_options *opt = diff_words->opt;
2091 const char *line_prefix;
2092
2093 assert(opt);
2094 line_prefix = diff_line_prefix(opt);
2095
2096 /* special case: only removal */
2097 if (!diff_words->plus.text.size) {
2098 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2099 line_prefix, strlen(line_prefix), 0);
2100 fn_out_diff_words_write_helper(diff_words->opt,
2101 &style->old_word, style->newline,
2102 diff_words->minus.text.size,
2103 diff_words->minus.text.ptr);
2104 diff_words->minus.text.size = 0;
2105 return;
2106 }
2107
2108 diff_words->current_plus = diff_words->plus.text.ptr;
2109 diff_words->last_minus = 0;
2110
2111 memset(&xpp, 0, sizeof(xpp));
2112 memset(&xecfg, 0, sizeof(xecfg));
2113 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2114 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2115 xpp.flags = 0;
2116 /* as only the hunk header will be parsed, we need a 0-context */
2117 xecfg.ctxlen = 0;
2118 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2119 diff_words, &xpp, &xecfg))
2120 die("unable to generate word diff");
2121 free(minus.ptr);
2122 free(plus.ptr);
2123 if (diff_words->current_plus != diff_words->plus.text.ptr +
2124 diff_words->plus.text.size) {
2125 if (color_words_output_graph_prefix(diff_words))
2126 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2127 line_prefix, strlen(line_prefix), 0);
2128 fn_out_diff_words_write_helper(diff_words->opt,
2129 &style->ctx, style->newline,
2130 diff_words->plus.text.ptr + diff_words->plus.text.size
2131 - diff_words->current_plus, diff_words->current_plus);
2132 }
2133 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2134 }
2135
2136 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2137 static void diff_words_flush(struct emit_callback *ecbdata)
2138 {
2139 struct diff_options *wo = ecbdata->diff_words->opt;
2140
2141 if (ecbdata->diff_words->minus.text.size ||
2142 ecbdata->diff_words->plus.text.size)
2143 diff_words_show(ecbdata->diff_words);
2144
2145 if (wo->emitted_symbols) {
2146 struct diff_options *o = ecbdata->opt;
2147 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2148 int i;
2149
2150 /*
2151 * NEEDSWORK:
2152 * Instead of appending each, concat all words to a line?
2153 */
2154 for (i = 0; i < wol->nr; i++)
2155 append_emitted_diff_symbol(o, &wol->buf[i]);
2156
2157 for (i = 0; i < wol->nr; i++)
2158 free((void *)wol->buf[i].line);
2159
2160 wol->nr = 0;
2161 }
2162 }
2163
2164 static void diff_filespec_load_driver(struct diff_filespec *one,
2165 struct index_state *istate)
2166 {
2167 /* Use already-loaded driver */
2168 if (one->driver)
2169 return;
2170
2171 if (S_ISREG(one->mode))
2172 one->driver = userdiff_find_by_path(istate, one->path);
2173
2174 /* Fallback to default settings */
2175 if (!one->driver)
2176 one->driver = userdiff_find_by_name("default");
2177 }
2178
2179 static const char *userdiff_word_regex(struct diff_filespec *one,
2180 struct index_state *istate)
2181 {
2182 diff_filespec_load_driver(one, istate);
2183 return one->driver->word_regex;
2184 }
2185
2186 static void init_diff_words_data(struct emit_callback *ecbdata,
2187 struct diff_options *orig_opts,
2188 struct diff_filespec *one,
2189 struct diff_filespec *two)
2190 {
2191 int i;
2192 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2193 memcpy(o, orig_opts, sizeof(struct diff_options));
2194
2195 CALLOC_ARRAY(ecbdata->diff_words, 1);
2196 ecbdata->diff_words->type = o->word_diff;
2197 ecbdata->diff_words->opt = o;
2198
2199 if (orig_opts->emitted_symbols)
2200 CALLOC_ARRAY(o->emitted_symbols, 1);
2201
2202 if (!o->word_regex)
2203 o->word_regex = userdiff_word_regex(one, o->repo->index);
2204 if (!o->word_regex)
2205 o->word_regex = userdiff_word_regex(two, o->repo->index);
2206 if (!o->word_regex)
2207 o->word_regex = diff_word_regex_cfg;
2208 if (o->word_regex) {
2209 ecbdata->diff_words->word_regex = (regex_t *)
2210 xmalloc(sizeof(regex_t));
2211 if (regcomp(ecbdata->diff_words->word_regex,
2212 o->word_regex,
2213 REG_EXTENDED | REG_NEWLINE))
2214 die("invalid regular expression: %s",
2215 o->word_regex);
2216 }
2217 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2218 if (o->word_diff == diff_words_styles[i].type) {
2219 ecbdata->diff_words->style =
2220 &diff_words_styles[i];
2221 break;
2222 }
2223 }
2224 if (want_color(o->use_color)) {
2225 struct diff_words_style *st = ecbdata->diff_words->style;
2226 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2227 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2228 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2229 }
2230 }
2231
2232 static void free_diff_words_data(struct emit_callback *ecbdata)
2233 {
2234 if (ecbdata->diff_words) {
2235 diff_words_flush(ecbdata);
2236 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2237 free (ecbdata->diff_words->opt);
2238 free (ecbdata->diff_words->minus.text.ptr);
2239 free (ecbdata->diff_words->minus.orig);
2240 free (ecbdata->diff_words->plus.text.ptr);
2241 free (ecbdata->diff_words->plus.orig);
2242 if (ecbdata->diff_words->word_regex) {
2243 regfree(ecbdata->diff_words->word_regex);
2244 free(ecbdata->diff_words->word_regex);
2245 }
2246 FREE_AND_NULL(ecbdata->diff_words);
2247 }
2248 }
2249
2250 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2251 {
2252 if (want_color(diff_use_color))
2253 return diff_colors[ix];
2254 return "";
2255 }
2256
2257 const char *diff_line_prefix(struct diff_options *opt)
2258 {
2259 struct strbuf *msgbuf;
2260 if (!opt->output_prefix)
2261 return "";
2262
2263 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2264 return msgbuf->buf;
2265 }
2266
2267 static unsigned long sane_truncate_line(char *line, unsigned long len)
2268 {
2269 const char *cp;
2270 unsigned long allot;
2271 size_t l = len;
2272
2273 cp = line;
2274 allot = l;
2275 while (0 < l) {
2276 (void) utf8_width(&cp, &l);
2277 if (!cp)
2278 break; /* truncated in the middle? */
2279 }
2280 return allot - l;
2281 }
2282
2283 static void find_lno(const char *line, struct emit_callback *ecbdata)
2284 {
2285 const char *p;
2286 ecbdata->lno_in_preimage = 0;
2287 ecbdata->lno_in_postimage = 0;
2288 p = strchr(line, '-');
2289 if (!p)
2290 return; /* cannot happen */
2291 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2292 p = strchr(p, '+');
2293 if (!p)
2294 return; /* cannot happen */
2295 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2296 }
2297
2298 static int fn_out_consume(void *priv, char *line, unsigned long len)
2299 {
2300 struct emit_callback *ecbdata = priv;
2301 struct diff_options *o = ecbdata->opt;
2302
2303 o->found_changes = 1;
2304
2305 if (ecbdata->header) {
2306 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2307 ecbdata->header->buf, ecbdata->header->len, 0);
2308 strbuf_reset(ecbdata->header);
2309 ecbdata->header = NULL;
2310 }
2311
2312 if (ecbdata->label_path[0]) {
2313 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2314 ecbdata->label_path[0],
2315 strlen(ecbdata->label_path[0]), 0);
2316 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2317 ecbdata->label_path[1],
2318 strlen(ecbdata->label_path[1]), 0);
2319 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2320 }
2321
2322 if (diff_suppress_blank_empty
2323 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2324 line[0] = '\n';
2325 len = 1;
2326 }
2327
2328 if (line[0] == '@') {
2329 if (ecbdata->diff_words)
2330 diff_words_flush(ecbdata);
2331 len = sane_truncate_line(line, len);
2332 find_lno(line, ecbdata);
2333 emit_hunk_header(ecbdata, line, len);
2334 return 0;
2335 }
2336
2337 if (ecbdata->diff_words) {
2338 enum diff_symbol s =
2339 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2340 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2341 if (line[0] == '-') {
2342 diff_words_append(line, len,
2343 &ecbdata->diff_words->minus);
2344 return 0;
2345 } else if (line[0] == '+') {
2346 diff_words_append(line, len,
2347 &ecbdata->diff_words->plus);
2348 return 0;
2349 } else if (starts_with(line, "\\ ")) {
2350 /*
2351 * Eat the "no newline at eof" marker as if we
2352 * saw a "+" or "-" line with nothing on it,
2353 * and return without diff_words_flush() to
2354 * defer processing. If this is the end of
2355 * preimage, more "+" lines may come after it.
2356 */
2357 return 0;
2358 }
2359 diff_words_flush(ecbdata);
2360 emit_diff_symbol(o, s, line, len, 0);
2361 return 0;
2362 }
2363
2364 switch (line[0]) {
2365 case '+':
2366 ecbdata->lno_in_postimage++;
2367 emit_add_line(ecbdata, line + 1, len - 1);
2368 break;
2369 case '-':
2370 ecbdata->lno_in_preimage++;
2371 emit_del_line(ecbdata, line + 1, len - 1);
2372 break;
2373 case ' ':
2374 ecbdata->lno_in_postimage++;
2375 ecbdata->lno_in_preimage++;
2376 emit_context_line(ecbdata, line + 1, len - 1);
2377 break;
2378 default:
2379 /* incomplete line at the end */
2380 ecbdata->lno_in_preimage++;
2381 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2382 line, len, 0);
2383 break;
2384 }
2385 return 0;
2386 }
2387
2388 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2389 {
2390 const char *old_name = a;
2391 const char *new_name = b;
2392 int pfx_length, sfx_length;
2393 int pfx_adjust_for_slash;
2394 int len_a = strlen(a);
2395 int len_b = strlen(b);
2396 int a_midlen, b_midlen;
2397 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2398 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2399
2400 if (qlen_a || qlen_b) {
2401 quote_c_style(a, name, NULL, 0);
2402 strbuf_addstr(name, " => ");
2403 quote_c_style(b, name, NULL, 0);
2404 return;
2405 }
2406
2407 /* Find common prefix */
2408 pfx_length = 0;
2409 while (*old_name && *new_name && *old_name == *new_name) {
2410 if (*old_name == '/')
2411 pfx_length = old_name - a + 1;
2412 old_name++;
2413 new_name++;
2414 }
2415
2416 /* Find common suffix */
2417 old_name = a + len_a;
2418 new_name = b + len_b;
2419 sfx_length = 0;
2420 /*
2421 * If there is a common prefix, it must end in a slash. In
2422 * that case we let this loop run 1 into the prefix to see the
2423 * same slash.
2424 *
2425 * If there is no common prefix, we cannot do this as it would
2426 * underrun the input strings.
2427 */
2428 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2429 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2430 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2431 *old_name == *new_name) {
2432 if (*old_name == '/')
2433 sfx_length = len_a - (old_name - a);
2434 old_name--;
2435 new_name--;
2436 }
2437
2438 /*
2439 * pfx{mid-a => mid-b}sfx
2440 * {pfx-a => pfx-b}sfx
2441 * pfx{sfx-a => sfx-b}
2442 * name-a => name-b
2443 */
2444 a_midlen = len_a - pfx_length - sfx_length;
2445 b_midlen = len_b - pfx_length - sfx_length;
2446 if (a_midlen < 0)
2447 a_midlen = 0;
2448 if (b_midlen < 0)
2449 b_midlen = 0;
2450
2451 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2452 if (pfx_length + sfx_length) {
2453 strbuf_add(name, a, pfx_length);
2454 strbuf_addch(name, '{');
2455 }
2456 strbuf_add(name, a + pfx_length, a_midlen);
2457 strbuf_addstr(name, " => ");
2458 strbuf_add(name, b + pfx_length, b_midlen);
2459 if (pfx_length + sfx_length) {
2460 strbuf_addch(name, '}');
2461 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2462 }
2463 }
2464
2465 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2466 const char *name_a,
2467 const char *name_b)
2468 {
2469 struct diffstat_file *x;
2470 CALLOC_ARRAY(x, 1);
2471 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2472 diffstat->files[diffstat->nr++] = x;
2473 if (name_b) {
2474 x->from_name = xstrdup(name_a);
2475 x->name = xstrdup(name_b);
2476 x->is_renamed = 1;
2477 }
2478 else {
2479 x->from_name = NULL;
2480 x->name = xstrdup(name_a);
2481 }
2482 return x;
2483 }
2484
2485 static int diffstat_consume(void *priv, char *line, unsigned long len)
2486 {
2487 struct diffstat_t *diffstat = priv;
2488 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2489
2490 if (!len)
2491 BUG("xdiff fed us an empty line");
2492
2493 if (line[0] == '+')
2494 x->added++;
2495 else if (line[0] == '-')
2496 x->deleted++;
2497 return 0;
2498 }
2499
2500 const char mime_boundary_leader[] = "------------";
2501
2502 static int scale_linear(int it, int width, int max_change)
2503 {
2504 if (!it)
2505 return 0;
2506 /*
2507 * make sure that at least one '-' or '+' is printed if
2508 * there is any change to this path. The easiest way is to
2509 * scale linearly as if the allotted width is one column shorter
2510 * than it is, and then add 1 to the result.
2511 */
2512 return 1 + (it * (width - 1) / max_change);
2513 }
2514
2515 static void show_graph(struct strbuf *out, char ch, int cnt,
2516 const char *set, const char *reset)
2517 {
2518 if (cnt <= 0)
2519 return;
2520 strbuf_addstr(out, set);
2521 strbuf_addchars(out, ch, cnt);
2522 strbuf_addstr(out, reset);
2523 }
2524
2525 static void fill_print_name(struct diffstat_file *file)
2526 {
2527 struct strbuf pname = STRBUF_INIT;
2528
2529 if (file->print_name)
2530 return;
2531
2532 if (file->is_renamed)
2533 pprint_rename(&pname, file->from_name, file->name);
2534 else
2535 quote_c_style(file->name, &pname, NULL, 0);
2536
2537 if (file->comments)
2538 strbuf_addf(&pname, " (%s)", file->comments);
2539
2540 file->print_name = strbuf_detach(&pname, NULL);
2541 }
2542
2543 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2544 int files, int insertions, int deletions)
2545 {
2546 struct strbuf sb = STRBUF_INIT;
2547
2548 if (!files) {
2549 assert(insertions == 0 && deletions == 0);
2550 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2551 NULL, 0, 0);
2552 return;
2553 }
2554
2555 strbuf_addf(&sb,
2556 (files == 1) ? " %d file changed" : " %d files changed",
2557 files);
2558
2559 /*
2560 * For binary diff, the caller may want to print "x files
2561 * changed" with insertions == 0 && deletions == 0.
2562 *
2563 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2564 * is probably less confusing (i.e skip over "2 files changed
2565 * but nothing about added/removed lines? Is this a bug in Git?").
2566 */
2567 if (insertions || deletions == 0) {
2568 strbuf_addf(&sb,
2569 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2570 insertions);
2571 }
2572
2573 if (deletions || insertions == 0) {
2574 strbuf_addf(&sb,
2575 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2576 deletions);
2577 }
2578 strbuf_addch(&sb, '\n');
2579 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2580 sb.buf, sb.len, 0);
2581 strbuf_release(&sb);
2582 }
2583
2584 void print_stat_summary(FILE *fp, int files,
2585 int insertions, int deletions)
2586 {
2587 struct diff_options o;
2588 memset(&o, 0, sizeof(o));
2589 o.file = fp;
2590
2591 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2592 }
2593
2594 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2595 {
2596 int i, len, add, del, adds = 0, dels = 0;
2597 uintmax_t max_change = 0, max_len = 0;
2598 int total_files = data->nr, count;
2599 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2600 const char *reset, *add_c, *del_c;
2601 int extra_shown = 0;
2602 const char *line_prefix = diff_line_prefix(options);
2603 struct strbuf out = STRBUF_INIT;
2604
2605 if (data->nr == 0)
2606 return;
2607
2608 count = options->stat_count ? options->stat_count : data->nr;
2609
2610 reset = diff_get_color_opt(options, DIFF_RESET);
2611 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2612 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2613
2614 /*
2615 * Find the longest filename and max number of changes
2616 */
2617 for (i = 0; (i < count) && (i < data->nr); i++) {
2618 struct diffstat_file *file = data->files[i];
2619 uintmax_t change = file->added + file->deleted;
2620
2621 if (!file->is_interesting && (change == 0)) {
2622 count++; /* not shown == room for one more */
2623 continue;
2624 }
2625 fill_print_name(file);
2626 len = utf8_strwidth(file->print_name);
2627 if (max_len < len)
2628 max_len = len;
2629
2630 if (file->is_unmerged) {
2631 /* "Unmerged" is 8 characters */
2632 bin_width = bin_width < 8 ? 8 : bin_width;
2633 continue;
2634 }
2635 if (file->is_binary) {
2636 /* "Bin XXX -> YYY bytes" */
2637 int w = 14 + decimal_width(file->added)
2638 + decimal_width(file->deleted);
2639 bin_width = bin_width < w ? w : bin_width;
2640 /* Display change counts aligned with "Bin" */
2641 number_width = 3;
2642 continue;
2643 }
2644
2645 if (max_change < change)
2646 max_change = change;
2647 }
2648 count = i; /* where we can stop scanning in data->files[] */
2649
2650 /*
2651 * We have width = stat_width or term_columns() columns total.
2652 * We want a maximum of min(max_len, stat_name_width) for the name part.
2653 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2654 * We also need 1 for " " and 4 + decimal_width(max_change)
2655 * for " | NNNN " and one the empty column at the end, altogether
2656 * 6 + decimal_width(max_change).
2657 *
2658 * If there's not enough space, we will use the smaller of
2659 * stat_name_width (if set) and 5/8*width for the filename,
2660 * and the rest for constant elements + graph part, but no more
2661 * than stat_graph_width for the graph part.
2662 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2663 * for the standard terminal size).
2664 *
2665 * In other words: stat_width limits the maximum width, and
2666 * stat_name_width fixes the maximum width of the filename,
2667 * and is also used to divide available columns if there
2668 * aren't enough.
2669 *
2670 * Binary files are displayed with "Bin XXX -> YYY bytes"
2671 * instead of the change count and graph. This part is treated
2672 * similarly to the graph part, except that it is not
2673 * "scaled". If total width is too small to accommodate the
2674 * guaranteed minimum width of the filename part and the
2675 * separators and this message, this message will "overflow"
2676 * making the line longer than the maximum width.
2677 */
2678
2679 /*
2680 * NEEDSWORK: line_prefix is often used for "log --graph" output
2681 * and contains ANSI-colored string. utf8_strnwidth() should be
2682 * used to correctly count the display width instead of strlen().
2683 */
2684 if (options->stat_width == -1)
2685 width = term_columns() - strlen(line_prefix);
2686 else
2687 width = options->stat_width ? options->stat_width : 80;
2688 number_width = decimal_width(max_change) > number_width ?
2689 decimal_width(max_change) : number_width;
2690
2691 if (options->stat_graph_width == -1)
2692 options->stat_graph_width = diff_stat_graph_width;
2693
2694 /*
2695 * Guarantee 3/8*16==6 for the graph part
2696 * and 5/8*16==10 for the filename part
2697 */
2698 if (width < 16 + 6 + number_width)
2699 width = 16 + 6 + number_width;
2700
2701 /*
2702 * First assign sizes that are wanted, ignoring available width.
2703 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2704 * starting from "XXX" should fit in graph_width.
2705 */
2706 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2707 if (options->stat_graph_width &&
2708 options->stat_graph_width < graph_width)
2709 graph_width = options->stat_graph_width;
2710
2711 name_width = (options->stat_name_width > 0 &&
2712 options->stat_name_width < max_len) ?
2713 options->stat_name_width : max_len;
2714
2715 /*
2716 * Adjust adjustable widths not to exceed maximum width
2717 */
2718 if (name_width + number_width + 6 + graph_width > width) {
2719 if (graph_width > width * 3/8 - number_width - 6) {
2720 graph_width = width * 3/8 - number_width - 6;
2721 if (graph_width < 6)
2722 graph_width = 6;
2723 }
2724
2725 if (options->stat_graph_width &&
2726 graph_width > options->stat_graph_width)
2727 graph_width = options->stat_graph_width;
2728 if (name_width > width - number_width - 6 - graph_width)
2729 name_width = width - number_width - 6 - graph_width;
2730 else
2731 graph_width = width - number_width - 6 - name_width;
2732 }
2733
2734 /*
2735 * From here name_width is the width of the name area,
2736 * and graph_width is the width of the graph area.
2737 * max_change is used to scale graph properly.
2738 */
2739 for (i = 0; i < count; i++) {
2740 const char *prefix = "";
2741 struct diffstat_file *file = data->files[i];
2742 char *name = file->print_name;
2743 uintmax_t added = file->added;
2744 uintmax_t deleted = file->deleted;
2745 int name_len, padding;
2746
2747 if (!file->is_interesting && (added + deleted == 0))
2748 continue;
2749
2750 /*
2751 * "scale" the filename
2752 */
2753 len = name_width;
2754 name_len = utf8_strwidth(name);
2755 if (name_width < name_len) {
2756 char *slash;
2757 prefix = "...";
2758 len -= 3;
2759 /*
2760 * NEEDSWORK: (name_len - len) counts the display
2761 * width, which would be shorter than the byte
2762 * length of the corresponding substring.
2763 * Advancing "name" by that number of bytes does
2764 * *NOT* skip over that many columns, so it is
2765 * very likely that chomping the pathname at the
2766 * slash we will find starting from "name" will
2767 * leave the resulting string still too long.
2768 */
2769 name += name_len - len;
2770 slash = strchr(name, '/');
2771 if (slash)
2772 name = slash;
2773 }
2774 padding = len - utf8_strwidth(name);
2775 if (padding < 0)
2776 padding = 0;
2777
2778 if (file->is_binary) {
2779 strbuf_addf(&out, " %s%s%*s | %*s",
2780 prefix, name, padding, "",
2781 number_width, "Bin");
2782 if (!added && !deleted) {
2783 strbuf_addch(&out, '\n');
2784 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2785 out.buf, out.len, 0);
2786 strbuf_reset(&out);
2787 continue;
2788 }
2789 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2790 del_c, deleted, reset);
2791 strbuf_addstr(&out, " -> ");
2792 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2793 add_c, added, reset);
2794 strbuf_addstr(&out, " bytes\n");
2795 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2796 out.buf, out.len, 0);
2797 strbuf_reset(&out);
2798 continue;
2799 }
2800 else if (file->is_unmerged) {
2801 strbuf_addf(&out, " %s%s%*s | %*s",
2802 prefix, name, padding, "",
2803 number_width, "Unmerged\n");
2804 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2805 out.buf, out.len, 0);
2806 strbuf_reset(&out);
2807 continue;
2808 }
2809
2810 /*
2811 * scale the add/delete
2812 */
2813 add = added;
2814 del = deleted;
2815
2816 if (graph_width <= max_change) {
2817 int total = scale_linear(add + del, graph_width, max_change);
2818 if (total < 2 && add && del)
2819 /* width >= 2 due to the sanity check */
2820 total = 2;
2821 if (add < del) {
2822 add = scale_linear(add, graph_width, max_change);
2823 del = total - add;
2824 } else {
2825 del = scale_linear(del, graph_width, max_change);
2826 add = total - del;
2827 }
2828 }
2829 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2830 prefix, name, padding, "",
2831 number_width, added + deleted,
2832 added + deleted ? " " : "");
2833 show_graph(&out, '+', add, add_c, reset);
2834 show_graph(&out, '-', del, del_c, reset);
2835 strbuf_addch(&out, '\n');
2836 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2837 out.buf, out.len, 0);
2838 strbuf_reset(&out);
2839 }
2840
2841 for (i = 0; i < data->nr; i++) {
2842 struct diffstat_file *file = data->files[i];
2843 uintmax_t added = file->added;
2844 uintmax_t deleted = file->deleted;
2845
2846 if (file->is_unmerged ||
2847 (!file->is_interesting && (added + deleted == 0))) {
2848 total_files--;
2849 continue;
2850 }
2851
2852 if (!file->is_binary) {
2853 adds += added;
2854 dels += deleted;
2855 }
2856 if (i < count)
2857 continue;
2858 if (!extra_shown)
2859 emit_diff_symbol(options,
2860 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2861 NULL, 0, 0);
2862 extra_shown = 1;
2863 }
2864
2865 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2866 strbuf_release(&out);
2867 }
2868
2869 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2870 {
2871 int i, adds = 0, dels = 0, total_files = data->nr;
2872
2873 if (data->nr == 0)
2874 return;
2875
2876 for (i = 0; i < data->nr; i++) {
2877 int added = data->files[i]->added;
2878 int deleted = data->files[i]->deleted;
2879
2880 if (data->files[i]->is_unmerged ||
2881 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2882 total_files--;
2883 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2884 adds += added;
2885 dels += deleted;
2886 }
2887 }
2888 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2889 }
2890
2891 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2892 {
2893 int i;
2894
2895 if (data->nr == 0)
2896 return;
2897
2898 for (i = 0; i < data->nr; i++) {
2899 struct diffstat_file *file = data->files[i];
2900
2901 fprintf(options->file, "%s", diff_line_prefix(options));
2902
2903 if (file->is_binary)
2904 fprintf(options->file, "-\t-\t");
2905 else
2906 fprintf(options->file,
2907 "%"PRIuMAX"\t%"PRIuMAX"\t",
2908 file->added, file->deleted);
2909 if (options->line_termination) {
2910 fill_print_name(file);
2911 if (!file->is_renamed)
2912 write_name_quoted(file->name, options->file,
2913 options->line_termination);
2914 else {
2915 fputs(file->print_name, options->file);
2916 putc(options->line_termination, options->file);
2917 }
2918 } else {
2919 if (file->is_renamed) {
2920 putc('\0', options->file);
2921 write_name_quoted(file->from_name, options->file, '\0');
2922 }
2923 write_name_quoted(file->name, options->file, '\0');
2924 }
2925 }
2926 }
2927
2928 struct dirstat_file {
2929 const char *name;
2930 unsigned long changed;
2931 };
2932
2933 struct dirstat_dir {
2934 struct dirstat_file *files;
2935 int alloc, nr, permille, cumulative;
2936 };
2937
2938 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2939 unsigned long changed, const char *base, int baselen)
2940 {
2941 unsigned long sum_changes = 0;
2942 unsigned int sources = 0;
2943 const char *line_prefix = diff_line_prefix(opt);
2944
2945 while (dir->nr) {
2946 struct dirstat_file *f = dir->files;
2947 int namelen = strlen(f->name);
2948 unsigned long changes;
2949 char *slash;
2950
2951 if (namelen < baselen)
2952 break;
2953 if (memcmp(f->name, base, baselen))
2954 break;
2955 slash = strchr(f->name + baselen, '/');
2956 if (slash) {
2957 int newbaselen = slash + 1 - f->name;
2958 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2959 sources++;
2960 } else {
2961 changes = f->changed;
2962 dir->files++;
2963 dir->nr--;
2964 sources += 2;
2965 }
2966 sum_changes += changes;
2967 }
2968
2969 /*
2970 * We don't report dirstat's for
2971 * - the top level
2972 * - or cases where everything came from a single directory
2973 * under this directory (sources == 1).
2974 */
2975 if (baselen && sources != 1) {
2976 if (sum_changes) {
2977 int permille = sum_changes * 1000 / changed;
2978 if (permille >= dir->permille) {
2979 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2980 permille / 10, permille % 10, baselen, base);
2981 if (!dir->cumulative)
2982 return 0;
2983 }
2984 }
2985 }
2986 return sum_changes;
2987 }
2988
2989 static int dirstat_compare(const void *_a, const void *_b)
2990 {
2991 const struct dirstat_file *a = _a;
2992 const struct dirstat_file *b = _b;
2993 return strcmp(a->name, b->name);
2994 }
2995
2996 static void show_dirstat(struct diff_options *options)
2997 {
2998 int i;
2999 unsigned long changed;
3000 struct dirstat_dir dir;
3001 struct diff_queue_struct *q = &diff_queued_diff;
3002
3003 dir.files = NULL;
3004 dir.alloc = 0;
3005 dir.nr = 0;
3006 dir.permille = options->dirstat_permille;
3007 dir.cumulative = options->flags.dirstat_cumulative;
3008
3009 changed = 0;
3010 for (i = 0; i < q->nr; i++) {
3011 struct diff_filepair *p = q->queue[i];
3012 const char *name;
3013 unsigned long copied, added, damage;
3014 struct diff_populate_filespec_options dpf_options = {
3015 .check_size_only = 1,
3016 };
3017
3018 name = p->two->path ? p->two->path : p->one->path;
3019
3020 if (p->one->oid_valid && p->two->oid_valid &&
3021 oideq(&p->one->oid, &p->two->oid)) {
3022 /*
3023 * The SHA1 has not changed, so pre-/post-content is
3024 * identical. We can therefore skip looking at the
3025 * file contents altogether.
3026 */
3027 damage = 0;
3028 goto found_damage;
3029 }
3030
3031 if (options->flags.dirstat_by_file) {
3032 /*
3033 * In --dirstat-by-file mode, we don't really need to
3034 * look at the actual file contents at all.
3035 * The fact that the SHA1 changed is enough for us to
3036 * add this file to the list of results
3037 * (with each file contributing equal damage).
3038 */
3039 damage = 1;
3040 goto found_damage;
3041 }
3042
3043 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3044 diff_populate_filespec(options->repo, p->one, NULL);
3045 diff_populate_filespec(options->repo, p->two, NULL);
3046 diffcore_count_changes(options->repo,
3047 p->one, p->two, NULL, NULL,
3048 &copied, &added);
3049 diff_free_filespec_data(p->one);
3050 diff_free_filespec_data(p->two);
3051 } else if (DIFF_FILE_VALID(p->one)) {
3052 diff_populate_filespec(options->repo, p->one, &dpf_options);
3053 copied = added = 0;
3054 diff_free_filespec_data(p->one);
3055 } else if (DIFF_FILE_VALID(p->two)) {
3056 diff_populate_filespec(options->repo, p->two, &dpf_options);
3057 copied = 0;
3058 added = p->two->size;
3059 diff_free_filespec_data(p->two);
3060 } else
3061 continue;
3062
3063 /*
3064 * Original minus copied is the removed material,
3065 * added is the new material. They are both damages
3066 * made to the preimage.
3067 * If the resulting damage is zero, we know that
3068 * diffcore_count_changes() considers the two entries to
3069 * be identical, but since the oid changed, we
3070 * know that there must have been _some_ kind of change,
3071 * so we force all entries to have damage > 0.
3072 */
3073 damage = (p->one->size - copied) + added;
3074 if (!damage)
3075 damage = 1;
3076
3077 found_damage:
3078 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3079 dir.files[dir.nr].name = name;
3080 dir.files[dir.nr].changed = damage;
3081 changed += damage;
3082 dir.nr++;
3083 }
3084
3085 /* This can happen even with many files, if everything was renames */
3086 if (!changed)
3087 return;
3088
3089 /* Show all directories with more than x% of the changes */
3090 QSORT(dir.files, dir.nr, dirstat_compare);
3091 gather_dirstat(options, &dir, changed, "", 0);
3092 }
3093
3094 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3095 {
3096 int i;
3097 unsigned long changed;
3098 struct dirstat_dir dir;
3099
3100 if (data->nr == 0)
3101 return;
3102
3103 dir.files = NULL;
3104 dir.alloc = 0;
3105 dir.nr = 0;
3106 dir.permille = options->dirstat_permille;
3107 dir.cumulative = options->flags.dirstat_cumulative;
3108
3109 changed = 0;
3110 for (i = 0; i < data->nr; i++) {
3111 struct diffstat_file *file = data->files[i];
3112 unsigned long damage = file->added + file->deleted;
3113 if (file->is_binary)
3114 /*
3115 * binary files counts bytes, not lines. Must find some
3116 * way to normalize binary bytes vs. textual lines.
3117 * The following heuristic assumes that there are 64
3118 * bytes per "line".
3119 * This is stupid and ugly, but very cheap...
3120 */
3121 damage = DIV_ROUND_UP(damage, 64);
3122 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3123 dir.files[dir.nr].name = file->name;
3124 dir.files[dir.nr].changed = damage;
3125 changed += damage;
3126 dir.nr++;
3127 }
3128
3129 /* This can happen even with many files, if everything was renames */
3130 if (!changed)
3131 return;
3132
3133 /* Show all directories with more than x% of the changes */
3134 QSORT(dir.files, dir.nr, dirstat_compare);
3135 gather_dirstat(options, &dir, changed, "", 0);
3136 }
3137
3138 static void free_diffstat_file(struct diffstat_file *f)
3139 {
3140 free(f->print_name);
3141 free(f->name);
3142 free(f->from_name);
3143 free(f);
3144 }
3145
3146 void free_diffstat_info(struct diffstat_t *diffstat)
3147 {
3148 int i;
3149 for (i = 0; i < diffstat->nr; i++)
3150 free_diffstat_file(diffstat->files[i]);
3151 free(diffstat->files);
3152 }
3153
3154 struct checkdiff_t {
3155 const char *filename;
3156 int lineno;
3157 int conflict_marker_size;
3158 struct diff_options *o;
3159 unsigned ws_rule;
3160 unsigned status;
3161 };
3162
3163 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3164 {
3165 char firstchar;
3166 int cnt;
3167
3168 if (len < marker_size + 1)
3169 return 0;
3170 firstchar = line[0];
3171 switch (firstchar) {
3172 case '=': case '>': case '<': case '|':
3173 break;
3174 default:
3175 return 0;
3176 }
3177 for (cnt = 1; cnt < marker_size; cnt++)
3178 if (line[cnt] != firstchar)
3179 return 0;
3180 /* line[1] through line[marker_size-1] are same as firstchar */
3181 if (len < marker_size + 1 || !isspace(line[marker_size]))
3182 return 0;
3183 return 1;
3184 }
3185
3186 static void checkdiff_consume_hunk(void *priv,
3187 long ob UNUSED, long on UNUSED,
3188 long nb, long nn UNUSED,
3189 const char *func UNUSED, long funclen UNUSED)
3190
3191 {
3192 struct checkdiff_t *data = priv;
3193 data->lineno = nb - 1;
3194 }
3195
3196 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3197 {
3198 struct checkdiff_t *data = priv;
3199 int marker_size = data->conflict_marker_size;
3200 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3201 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3202 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3203 char *err;
3204 const char *line_prefix;
3205
3206 assert(data->o);
3207 line_prefix = diff_line_prefix(data->o);
3208
3209 if (line[0] == '+') {
3210 unsigned bad;
3211 data->lineno++;
3212 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3213 data->status |= 1;
3214 fprintf(data->o->file,
3215 "%s%s:%d: leftover conflict marker\n",
3216 line_prefix, data->filename, data->lineno);
3217 }
3218 bad = ws_check(line + 1, len - 1, data->ws_rule);
3219 if (!bad)
3220 return 0;
3221 data->status |= bad;
3222 err = whitespace_error_string(bad);
3223 fprintf(data->o->file, "%s%s:%d: %s.\n",
3224 line_prefix, data->filename, data->lineno, err);
3225 free(err);
3226 emit_line(data->o, set, reset, line, 1);
3227 ws_check_emit(line + 1, len - 1, data->ws_rule,
3228 data->o->file, set, reset, ws);
3229 } else if (line[0] == ' ') {
3230 data->lineno++;
3231 }
3232 return 0;
3233 }
3234
3235 static unsigned char *deflate_it(char *data,
3236 unsigned long size,
3237 unsigned long *result_size)
3238 {
3239 int bound;
3240 unsigned char *deflated;
3241 git_zstream stream;
3242
3243 git_deflate_init(&stream, zlib_compression_level);
3244 bound = git_deflate_bound(&stream, size);
3245 deflated = xmalloc(bound);
3246 stream.next_out = deflated;
3247 stream.avail_out = bound;
3248
3249 stream.next_in = (unsigned char *)data;
3250 stream.avail_in = size;
3251 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3252 ; /* nothing */
3253 git_deflate_end(&stream);
3254 *result_size = stream.total_out;
3255 return deflated;
3256 }
3257
3258 static void emit_binary_diff_body(struct diff_options *o,
3259 mmfile_t *one, mmfile_t *two)
3260 {
3261 void *cp;
3262 void *delta;
3263 void *deflated;
3264 void *data;
3265 unsigned long orig_size;
3266 unsigned long delta_size;
3267 unsigned long deflate_size;
3268 unsigned long data_size;
3269
3270 /* We could do deflated delta, or we could do just deflated two,
3271 * whichever is smaller.
3272 */
3273 delta = NULL;
3274 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3275 if (one->size && two->size) {
3276 delta = diff_delta(one->ptr, one->size,
3277 two->ptr, two->size,
3278 &delta_size, deflate_size);
3279 if (delta) {
3280 void *to_free = delta;
3281 orig_size = delta_size;
3282 delta = deflate_it(delta, delta_size, &delta_size);
3283 free(to_free);
3284 }
3285 }
3286
3287 if (delta && delta_size < deflate_size) {
3288 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3289 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3290 s, strlen(s), 0);
3291 free(s);
3292 free(deflated);
3293 data = delta;
3294 data_size = delta_size;
3295 } else {
3296 char *s = xstrfmt("%lu", two->size);
3297 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3298 s, strlen(s), 0);
3299 free(s);
3300 free(delta);
3301 data = deflated;
3302 data_size = deflate_size;
3303 }
3304
3305 /* emit data encoded in base85 */
3306 cp = data;
3307 while (data_size) {
3308 int len;
3309 int bytes = (52 < data_size) ? 52 : data_size;
3310 char line[71];
3311 data_size -= bytes;
3312 if (bytes <= 26)
3313 line[0] = bytes + 'A' - 1;
3314 else
3315 line[0] = bytes - 26 + 'a' - 1;
3316 encode_85(line + 1, cp, bytes);
3317 cp = (char *) cp + bytes;
3318
3319 len = strlen(line);
3320 line[len++] = '\n';
3321 line[len] = '\0';
3322
3323 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3324 line, len, 0);
3325 }
3326 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3327 free(data);
3328 }
3329
3330 static void emit_binary_diff(struct diff_options *o,
3331 mmfile_t *one, mmfile_t *two)
3332 {
3333 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3334 emit_binary_diff_body(o, one, two);
3335 emit_binary_diff_body(o, two, one);
3336 }
3337
3338 int diff_filespec_is_binary(struct repository *r,
3339 struct diff_filespec *one)
3340 {
3341 struct diff_populate_filespec_options dpf_options = {
3342 .check_binary = 1,
3343 };
3344
3345 if (one->is_binary == -1) {
3346 diff_filespec_load_driver(one, r->index);
3347 if (one->driver->binary != -1)
3348 one->is_binary = one->driver->binary;
3349 else {
3350 if (!one->data && DIFF_FILE_VALID(one))
3351 diff_populate_filespec(r, one, &dpf_options);
3352 if (one->is_binary == -1 && one->data)
3353 one->is_binary = buffer_is_binary(one->data,
3354 one->size);
3355 if (one->is_binary == -1)
3356 one->is_binary = 0;
3357 }
3358 }
3359 return one->is_binary;
3360 }
3361
3362 static const struct userdiff_funcname *
3363 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3364 {
3365 diff_filespec_load_driver(one, o->repo->index);
3366 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3367 }
3368
3369 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3370 {
3371 if (!options->a_prefix)
3372 options->a_prefix = a;
3373 if (!options->b_prefix)
3374 options->b_prefix = b;
3375 }
3376
3377 struct userdiff_driver *get_textconv(struct repository *r,
3378 struct diff_filespec *one)
3379 {
3380 if (!DIFF_FILE_VALID(one))
3381 return NULL;
3382
3383 diff_filespec_load_driver(one, r->index);
3384 return userdiff_get_textconv(r, one->driver);
3385 }
3386
3387 static struct string_list *additional_headers(struct diff_options *o,
3388 const char *path)
3389 {
3390 if (!o->additional_path_headers)
3391 return NULL;
3392 return strmap_get(o->additional_path_headers, path);
3393 }
3394
3395 static void add_formatted_header(struct strbuf *msg,
3396 const char *header,
3397 const char *line_prefix,
3398 const char *meta,
3399 const char *reset)
3400 {
3401 const char *next, *newline;
3402
3403 for (next = header; *next; next = newline) {
3404 newline = strchrnul(next, '\n');
3405 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3406 (int)(newline - next), next, reset);
3407 if (*newline)
3408 newline++;
3409 }
3410 }
3411
3412 static void add_formatted_headers(struct strbuf *msg,
3413 struct string_list *more_headers,
3414 const char *line_prefix,
3415 const char *meta,
3416 const char *reset)
3417 {
3418 int i;
3419
3420 for (i = 0; i < more_headers->nr; i++)
3421 add_formatted_header(msg, more_headers->items[i].string,
3422 line_prefix, meta, reset);
3423 }
3424
3425 static int diff_filepair_is_phoney(struct diff_filespec *one,
3426 struct diff_filespec *two)
3427 {
3428 /*
3429 * This function specifically looks for pairs injected by
3430 * create_filepairs_for_header_only_notifications(). Such
3431 * pairs are "phoney" in that they do not represent any
3432 * content or even mode difference, but were inserted because
3433 * diff_queued_diff previously had no pair associated with
3434 * that path but we needed some pair to avoid losing the
3435 * "remerge CONFLICT" header associated with the path.
3436 */
3437 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3438 }
3439
3440 static int set_diff_algorithm(struct diff_options *opts,
3441 const char *alg)
3442 {
3443 long value = parse_algorithm_value(alg);
3444
3445 if (value < 0)
3446 return -1;
3447
3448 /* clear out previous settings */
3449 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3450 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3451 opts->xdl_opts |= value;
3452
3453 return 0;
3454 }
3455
3456 static void builtin_diff(const char *name_a,
3457 const char *name_b,
3458 struct diff_filespec *one,
3459 struct diff_filespec *two,
3460 const char *xfrm_msg,
3461 int must_show_header,
3462 struct diff_options *o,
3463 int complete_rewrite)
3464 {
3465 mmfile_t mf1, mf2;
3466 const char *lbl[2];
3467 char *a_one, *b_two;
3468 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3469 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3470 const char *a_prefix, *b_prefix;
3471 struct userdiff_driver *textconv_one = NULL;
3472 struct userdiff_driver *textconv_two = NULL;
3473 struct strbuf header = STRBUF_INIT;
3474 const char *line_prefix = diff_line_prefix(o);
3475
3476 diff_set_mnemonic_prefix(o, "a/", "b/");
3477 if (o->flags.reverse_diff) {
3478 a_prefix = o->b_prefix;
3479 b_prefix = o->a_prefix;
3480 } else {
3481 a_prefix = o->a_prefix;
3482 b_prefix = o->b_prefix;
3483 }
3484
3485 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3486 (!one->mode || S_ISGITLINK(one->mode)) &&
3487 (!two->mode || S_ISGITLINK(two->mode)) &&
3488 (!diff_filepair_is_phoney(one, two))) {
3489 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3490 &one->oid, &two->oid,
3491 two->dirty_submodule);
3492 return;
3493 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3494 (!one->mode || S_ISGITLINK(one->mode)) &&
3495 (!two->mode || S_ISGITLINK(two->mode)) &&
3496 (!diff_filepair_is_phoney(one, two))) {
3497 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3498 &one->oid, &two->oid,
3499 two->dirty_submodule);
3500 return;
3501 }
3502
3503 if (o->flags.allow_textconv) {
3504 textconv_one = get_textconv(o->repo, one);
3505 textconv_two = get_textconv(o->repo, two);
3506 }
3507
3508 /* Never use a non-valid filename anywhere if at all possible */
3509 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3510 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3511
3512 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3513 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3514 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3515 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3516 if (diff_filepair_is_phoney(one, two)) {
3517 /*
3518 * We should only reach this point for pairs generated from
3519 * create_filepairs_for_header_only_notifications(). For
3520 * these, we want to avoid the "/dev/null" special casing
3521 * above, because we do not want such pairs shown as either
3522 * "new file" or "deleted file" below.
3523 */
3524 lbl[0] = a_one;
3525 lbl[1] = b_two;
3526 }
3527 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3528 if (lbl[0][0] == '/') {
3529 /* /dev/null */
3530 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3531 if (xfrm_msg)
3532 strbuf_addstr(&header, xfrm_msg);
3533 must_show_header = 1;
3534 }
3535 else if (lbl[1][0] == '/') {
3536 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3537 if (xfrm_msg)
3538 strbuf_addstr(&header, xfrm_msg);
3539 must_show_header = 1;
3540 }
3541 else {
3542 if (one->mode != two->mode) {
3543 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3544 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3545 must_show_header = 1;
3546 }
3547 if (xfrm_msg)
3548 strbuf_addstr(&header, xfrm_msg);
3549
3550 /*
3551 * we do not run diff between different kind
3552 * of objects.
3553 */
3554 if ((one->mode ^ two->mode) & S_IFMT)
3555 goto free_ab_and_return;
3556 if (complete_rewrite &&
3557 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3558 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3559 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3560 header.buf, header.len, 0);
3561 strbuf_reset(&header);
3562 emit_rewrite_diff(name_a, name_b, one, two,
3563 textconv_one, textconv_two, o);
3564 o->found_changes = 1;
3565 goto free_ab_and_return;
3566 }
3567 }
3568
3569 if (o->irreversible_delete && lbl[1][0] == '/') {
3570 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3571 header.len, 0);
3572 strbuf_reset(&header);
3573 goto free_ab_and_return;
3574 } else if (!o->flags.text &&
3575 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3576 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3577 struct strbuf sb = STRBUF_INIT;
3578 if (!one->data && !two->data &&
3579 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3580 !o->flags.binary) {
3581 if (oideq(&one->oid, &two->oid)) {
3582 if (must_show_header)
3583 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3584 header.buf, header.len,
3585 0);
3586 goto free_ab_and_return;
3587 }
3588 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3589 header.buf, header.len, 0);
3590 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3591 diff_line_prefix(o), lbl[0], lbl[1]);
3592 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3593 sb.buf, sb.len, 0);
3594 strbuf_release(&sb);
3595 goto free_ab_and_return;
3596 }
3597 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3598 fill_mmfile(o->repo, &mf2, two) < 0)
3599 die("unable to read files to diff");
3600 /* Quite common confusing case */
3601 if (mf1.size == mf2.size &&
3602 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3603 if (must_show_header)
3604 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3605 header.buf, header.len, 0);
3606 goto free_ab_and_return;
3607 }
3608 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3609 strbuf_reset(&header);
3610 if (o->flags.binary)
3611 emit_binary_diff(o, &mf1, &mf2);
3612 else {
3613 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3614 diff_line_prefix(o), lbl[0], lbl[1]);
3615 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3616 sb.buf, sb.len, 0);
3617 strbuf_release(&sb);
3618 }
3619 o->found_changes = 1;
3620 } else {
3621 /* Crazy xdl interfaces.. */
3622 const char *diffopts;
3623 const char *v;
3624 xpparam_t xpp;
3625 xdemitconf_t xecfg;
3626 struct emit_callback ecbdata;
3627 const struct userdiff_funcname *pe;
3628
3629 if (must_show_header) {
3630 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3631 header.buf, header.len, 0);
3632 strbuf_reset(&header);
3633 }
3634
3635 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3636 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3637
3638 pe = diff_funcname_pattern(o, one);
3639 if (!pe)
3640 pe = diff_funcname_pattern(o, two);
3641
3642 memset(&xpp, 0, sizeof(xpp));
3643 memset(&xecfg, 0, sizeof(xecfg));
3644 memset(&ecbdata, 0, sizeof(ecbdata));
3645 if (o->flags.suppress_diff_headers)
3646 lbl[0] = NULL;
3647 ecbdata.label_path = lbl;
3648 ecbdata.color_diff = want_color(o->use_color);
3649 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3650 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3651 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3652 ecbdata.opt = o;
3653 if (header.len && !o->flags.suppress_diff_headers)
3654 ecbdata.header = &header;
3655 xpp.flags = o->xdl_opts;
3656 xpp.ignore_regex = o->ignore_regex;
3657 xpp.ignore_regex_nr = o->ignore_regex_nr;
3658 xpp.anchors = o->anchors;
3659 xpp.anchors_nr = o->anchors_nr;
3660 xecfg.ctxlen = o->context;
3661 xecfg.interhunkctxlen = o->interhunkcontext;
3662 xecfg.flags = XDL_EMIT_FUNCNAMES;
3663 if (o->flags.funccontext)
3664 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3665 if (pe)
3666 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3667
3668 diffopts = getenv("GIT_DIFF_OPTS");
3669 if (!diffopts)
3670 ;
3671 else if (skip_prefix(diffopts, "--unified=", &v))
3672 xecfg.ctxlen = strtoul(v, NULL, 10);
3673 else if (skip_prefix(diffopts, "-u", &v))
3674 xecfg.ctxlen = strtoul(v, NULL, 10);
3675
3676 if (o->word_diff)
3677 init_diff_words_data(&ecbdata, o, one, two);
3678 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3679 &ecbdata, &xpp, &xecfg))
3680 die("unable to generate diff for %s", one->path);
3681 if (o->word_diff)
3682 free_diff_words_data(&ecbdata);
3683 if (textconv_one)
3684 free(mf1.ptr);
3685 if (textconv_two)
3686 free(mf2.ptr);
3687 xdiff_clear_find_func(&xecfg);
3688 }
3689
3690 free_ab_and_return:
3691 strbuf_release(&header);
3692 diff_free_filespec_data(one);
3693 diff_free_filespec_data(two);
3694 free(a_one);
3695 free(b_two);
3696 return;
3697 }
3698
3699 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3700 {
3701 if (!is_renamed) {
3702 if (p->status == DIFF_STATUS_ADDED) {
3703 if (S_ISLNK(p->two->mode))
3704 return "new +l";
3705 else if ((p->two->mode & 0777) == 0755)
3706 return "new +x";
3707 else
3708 return "new";
3709 } else if (p->status == DIFF_STATUS_DELETED)
3710 return "gone";
3711 }
3712 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3713 return "mode -l";
3714 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3715 return "mode +l";
3716 else if ((p->one->mode & 0777) == 0644 &&
3717 (p->two->mode & 0777) == 0755)
3718 return "mode +x";
3719 else if ((p->one->mode & 0777) == 0755 &&
3720 (p->two->mode & 0777) == 0644)
3721 return "mode -x";
3722 return NULL;
3723 }
3724
3725 static void builtin_diffstat(const char *name_a, const char *name_b,
3726 struct diff_filespec *one,
3727 struct diff_filespec *two,
3728 struct diffstat_t *diffstat,
3729 struct diff_options *o,
3730 struct diff_filepair *p)
3731 {
3732 mmfile_t mf1, mf2;
3733 struct diffstat_file *data;
3734 int may_differ;
3735 int complete_rewrite = 0;
3736
3737 if (!DIFF_PAIR_UNMERGED(p)) {
3738 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3739 complete_rewrite = 1;
3740 }
3741
3742 data = diffstat_add(diffstat, name_a, name_b);
3743 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3744 if (o->flags.stat_with_summary)
3745 data->comments = get_compact_summary(p, data->is_renamed);
3746
3747 if (!one || !two) {
3748 data->is_unmerged = 1;
3749 return;
3750 }
3751
3752 /* saves some reads if true, not a guarantee of diff outcome */
3753 may_differ = !(one->oid_valid && two->oid_valid &&
3754 oideq(&one->oid, &two->oid));
3755
3756 if (diff_filespec_is_binary(o->repo, one) ||
3757 diff_filespec_is_binary(o->repo, two)) {
3758 data->is_binary = 1;
3759 if (!may_differ) {
3760 data->added = 0;
3761 data->deleted = 0;
3762 } else {
3763 data->added = diff_filespec_size(o->repo, two);
3764 data->deleted = diff_filespec_size(o->repo, one);
3765 }
3766 }
3767
3768 else if (complete_rewrite) {
3769 diff_populate_filespec(o->repo, one, NULL);
3770 diff_populate_filespec(o->repo, two, NULL);
3771 data->deleted = count_lines(one->data, one->size);
3772 data->added = count_lines(two->data, two->size);
3773 }
3774
3775 else if (may_differ) {
3776 /* Crazy xdl interfaces.. */
3777 xpparam_t xpp;
3778 xdemitconf_t xecfg;
3779
3780 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3781 fill_mmfile(o->repo, &mf2, two) < 0)
3782 die("unable to read files to diff");
3783
3784 memset(&xpp, 0, sizeof(xpp));
3785 memset(&xecfg, 0, sizeof(xecfg));
3786 xpp.flags = o->xdl_opts;
3787 xpp.ignore_regex = o->ignore_regex;
3788 xpp.ignore_regex_nr = o->ignore_regex_nr;
3789 xpp.anchors = o->anchors;
3790 xpp.anchors_nr = o->anchors_nr;
3791 xecfg.ctxlen = o->context;
3792 xecfg.interhunkctxlen = o->interhunkcontext;
3793 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3794 if (xdi_diff_outf(&mf1, &mf2, NULL,
3795 diffstat_consume, diffstat, &xpp, &xecfg))
3796 die("unable to generate diffstat for %s", one->path);
3797
3798 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3799 struct diffstat_file *file =
3800 diffstat->files[diffstat->nr - 1];
3801 /*
3802 * Omit diffstats of modified files where nothing changed.
3803 * Even if may_differ, this might be the case due to
3804 * ignoring whitespace changes, etc.
3805 *
3806 * But note that we special-case additions, deletions,
3807 * renames, and mode changes as adding an empty file,
3808 * for example is still of interest.
3809 */
3810 if ((p->status == DIFF_STATUS_MODIFIED)
3811 && !file->added
3812 && !file->deleted
3813 && one->mode == two->mode) {
3814 free_diffstat_file(file);
3815 diffstat->nr--;
3816 }
3817 }
3818 }
3819
3820 diff_free_filespec_data(one);
3821 diff_free_filespec_data(two);
3822 }
3823
3824 static void builtin_checkdiff(const char *name_a, const char *name_b,
3825 const char *attr_path,
3826 struct diff_filespec *one,
3827 struct diff_filespec *two,
3828 struct diff_options *o)
3829 {
3830 mmfile_t mf1, mf2;
3831 struct checkdiff_t data;
3832
3833 if (!two)
3834 return;
3835
3836 memset(&data, 0, sizeof(data));
3837 data.filename = name_b ? name_b : name_a;
3838 data.lineno = 0;
3839 data.o = o;
3840 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3841 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3842
3843 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3844 fill_mmfile(o->repo, &mf2, two) < 0)
3845 die("unable to read files to diff");
3846
3847 /*
3848 * All the other codepaths check both sides, but not checking
3849 * the "old" side here is deliberate. We are checking the newly
3850 * introduced changes, and as long as the "new" side is text, we
3851 * can and should check what it introduces.
3852 */
3853 if (diff_filespec_is_binary(o->repo, two))
3854 goto free_and_return;
3855 else {
3856 /* Crazy xdl interfaces.. */
3857 xpparam_t xpp;
3858 xdemitconf_t xecfg;
3859
3860 memset(&xpp, 0, sizeof(xpp));
3861 memset(&xecfg, 0, sizeof(xecfg));
3862 xecfg.ctxlen = 1; /* at least one context line */
3863 xpp.flags = 0;
3864 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3865 checkdiff_consume, &data,
3866 &xpp, &xecfg))
3867 die("unable to generate checkdiff for %s", one->path);
3868
3869 if (data.ws_rule & WS_BLANK_AT_EOF) {
3870 struct emit_callback ecbdata;
3871 int blank_at_eof;
3872
3873 ecbdata.ws_rule = data.ws_rule;
3874 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3875 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3876
3877 if (blank_at_eof) {
3878 static char *err;
3879 if (!err)
3880 err = whitespace_error_string(WS_BLANK_AT_EOF);
3881 fprintf(o->file, "%s:%d: %s.\n",
3882 data.filename, blank_at_eof, err);
3883 data.status = 1; /* report errors */
3884 }
3885 }
3886 }
3887 free_and_return:
3888 diff_free_filespec_data(one);
3889 diff_free_filespec_data(two);
3890 if (data.status)
3891 o->flags.check_failed = 1;
3892 }
3893
3894 struct diff_filespec *alloc_filespec(const char *path)
3895 {
3896 struct diff_filespec *spec;
3897
3898 FLEXPTR_ALLOC_STR(spec, path, path);
3899 spec->count = 1;
3900 spec->is_binary = -1;
3901 return spec;
3902 }
3903
3904 void free_filespec(struct diff_filespec *spec)
3905 {
3906 if (!--spec->count) {
3907 diff_free_filespec_data(spec);
3908 free(spec);
3909 }
3910 }
3911
3912 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3913 int oid_valid, unsigned short mode)
3914 {
3915 if (mode) {
3916 spec->mode = canon_mode(mode);
3917 oidcpy(&spec->oid, oid);
3918 spec->oid_valid = oid_valid;
3919 }
3920 }
3921
3922 /*
3923 * Given a name and sha1 pair, if the index tells us the file in
3924 * the work tree has that object contents, return true, so that
3925 * prepare_temp_file() does not have to inflate and extract.
3926 */
3927 static int reuse_worktree_file(struct index_state *istate,
3928 const char *name,
3929 const struct object_id *oid,
3930 int want_file)
3931 {
3932 const struct cache_entry *ce;
3933 struct stat st;
3934 int pos, len;
3935
3936 /*
3937 * We do not read the cache ourselves here, because the
3938 * benchmark with my previous version that always reads cache
3939 * shows that it makes things worse for diff-tree comparing
3940 * two linux-2.6 kernel trees in an already checked out work
3941 * tree. This is because most diff-tree comparisons deal with
3942 * only a small number of files, while reading the cache is
3943 * expensive for a large project, and its cost outweighs the
3944 * savings we get by not inflating the object to a temporary
3945 * file. Practically, this code only helps when we are used
3946 * by diff-cache --cached, which does read the cache before
3947 * calling us.
3948 */
3949 if (!istate->cache)
3950 return 0;
3951
3952 /* We want to avoid the working directory if our caller
3953 * doesn't need the data in a normal file, this system
3954 * is rather slow with its stat/open/mmap/close syscalls,
3955 * and the object is contained in a pack file. The pack
3956 * is probably already open and will be faster to obtain
3957 * the data through than the working directory. Loose
3958 * objects however would tend to be slower as they need
3959 * to be individually opened and inflated.
3960 */
3961 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3962 return 0;
3963
3964 /*
3965 * Similarly, if we'd have to convert the file contents anyway, that
3966 * makes the optimization not worthwhile.
3967 */
3968 if (!want_file && would_convert_to_git(istate, name))
3969 return 0;
3970
3971 /*
3972 * If this path does not match our sparse-checkout definition,
3973 * then the file will not be in the working directory.
3974 */
3975 if (!path_in_sparse_checkout(name, istate))
3976 return 0;
3977
3978 len = strlen(name);
3979 pos = index_name_pos(istate, name, len);
3980 if (pos < 0)
3981 return 0;
3982 ce = istate->cache[pos];
3983
3984 /*
3985 * This is not the sha1 we are looking for, or
3986 * unreusable because it is not a regular file.
3987 */
3988 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3989 return 0;
3990
3991 /*
3992 * If ce is marked as "assume unchanged", there is no
3993 * guarantee that work tree matches what we are looking for.
3994 */
3995 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3996 return 0;
3997
3998 /*
3999 * If ce matches the file in the work tree, we can reuse it.
4000 */
4001 if (ce_uptodate(ce) ||
4002 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4003 return 1;
4004
4005 return 0;
4006 }
4007
4008 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4009 {
4010 struct strbuf buf = STRBUF_INIT;
4011 char *dirty = "";
4012
4013 /* Are we looking at the work tree? */
4014 if (s->dirty_submodule)
4015 dirty = "-dirty";
4016
4017 strbuf_addf(&buf, "Subproject commit %s%s\n",
4018 oid_to_hex(&s->oid), dirty);
4019 s->size = buf.len;
4020 if (size_only) {
4021 s->data = NULL;
4022 strbuf_release(&buf);
4023 } else {
4024 s->data = strbuf_detach(&buf, NULL);
4025 s->should_free = 1;
4026 }
4027 return 0;
4028 }
4029
4030 /*
4031 * While doing rename detection and pickaxe operation, we may need to
4032 * grab the data for the blob (or file) for our own in-core comparison.
4033 * diff_filespec has data and size fields for this purpose.
4034 */
4035 int diff_populate_filespec(struct repository *r,
4036 struct diff_filespec *s,
4037 const struct diff_populate_filespec_options *options)
4038 {
4039 int size_only = options ? options->check_size_only : 0;
4040 int check_binary = options ? options->check_binary : 0;
4041 int err = 0;
4042 int conv_flags = global_conv_flags_eol;
4043 /*
4044 * demote FAIL to WARN to allow inspecting the situation
4045 * instead of refusing.
4046 */
4047 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4048 conv_flags = CONV_EOL_RNDTRP_WARN;
4049
4050 if (!DIFF_FILE_VALID(s))
4051 die("internal error: asking to populate invalid file.");
4052 if (S_ISDIR(s->mode))
4053 return -1;
4054
4055 if (s->data)
4056 return 0;
4057
4058 if (size_only && 0 < s->size)
4059 return 0;
4060
4061 if (S_ISGITLINK(s->mode))
4062 return diff_populate_gitlink(s, size_only);
4063
4064 if (!s->oid_valid ||
4065 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4066 struct strbuf buf = STRBUF_INIT;
4067 struct stat st;
4068 int fd;
4069
4070 if (lstat(s->path, &st) < 0) {
4071 err_empty:
4072 err = -1;
4073 empty:
4074 s->data = (char *)"";
4075 s->size = 0;
4076 return err;
4077 }
4078 s->size = xsize_t(st.st_size);
4079 if (!s->size)
4080 goto empty;
4081 if (S_ISLNK(st.st_mode)) {
4082 struct strbuf sb = STRBUF_INIT;
4083
4084 if (strbuf_readlink(&sb, s->path, s->size))
4085 goto err_empty;
4086 s->size = sb.len;
4087 s->data = strbuf_detach(&sb, NULL);
4088 s->should_free = 1;
4089 return 0;
4090 }
4091
4092 /*
4093 * Even if the caller would be happy with getting
4094 * only the size, we cannot return early at this
4095 * point if the path requires us to run the content
4096 * conversion.
4097 */
4098 if (size_only && !would_convert_to_git(r->index, s->path))
4099 return 0;
4100
4101 /*
4102 * Note: this check uses xsize_t(st.st_size) that may
4103 * not be the true size of the blob after it goes
4104 * through convert_to_git(). This may not strictly be
4105 * correct, but the whole point of big_file_threshold
4106 * and is_binary check being that we want to avoid
4107 * opening the file and inspecting the contents, this
4108 * is probably fine.
4109 */
4110 if (check_binary &&
4111 s->size > big_file_threshold && s->is_binary == -1) {
4112 s->is_binary = 1;
4113 return 0;
4114 }
4115 fd = open(s->path, O_RDONLY);
4116 if (fd < 0)
4117 goto err_empty;
4118 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4119 close(fd);
4120 s->should_munmap = 1;
4121
4122 /*
4123 * Convert from working tree format to canonical git format
4124 */
4125 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4126 size_t size = 0;
4127 munmap(s->data, s->size);
4128 s->should_munmap = 0;
4129 s->data = strbuf_detach(&buf, &size);
4130 s->size = size;
4131 s->should_free = 1;
4132 }
4133 }
4134 else {
4135 struct object_info info = {
4136 .sizep = &s->size
4137 };
4138
4139 if (!(size_only || check_binary))
4140 /*
4141 * Set contentp, since there is no chance that merely
4142 * the size is sufficient.
4143 */
4144 info.contentp = &s->data;
4145
4146 if (options && options->missing_object_cb) {
4147 if (!oid_object_info_extended(r, &s->oid, &info,
4148 OBJECT_INFO_LOOKUP_REPLACE |
4149 OBJECT_INFO_SKIP_FETCH_OBJECT))
4150 goto object_read;
4151 options->missing_object_cb(options->missing_object_data);
4152 }
4153 if (oid_object_info_extended(r, &s->oid, &info,
4154 OBJECT_INFO_LOOKUP_REPLACE))
4155 die("unable to read %s", oid_to_hex(&s->oid));
4156
4157 object_read:
4158 if (size_only || check_binary) {
4159 if (size_only)
4160 return 0;
4161 if (s->size > big_file_threshold && s->is_binary == -1) {
4162 s->is_binary = 1;
4163 return 0;
4164 }
4165 }
4166 if (!info.contentp) {
4167 info.contentp = &s->data;
4168 if (oid_object_info_extended(r, &s->oid, &info,
4169 OBJECT_INFO_LOOKUP_REPLACE))
4170 die("unable to read %s", oid_to_hex(&s->oid));
4171 }
4172 s->should_free = 1;
4173 }
4174 return 0;
4175 }
4176
4177 void diff_free_filespec_blob(struct diff_filespec *s)
4178 {
4179 if (s->should_free)
4180 free(s->data);
4181 else if (s->should_munmap)
4182 munmap(s->data, s->size);
4183
4184 if (s->should_free || s->should_munmap) {
4185 s->should_free = s->should_munmap = 0;
4186 s->data = NULL;
4187 }
4188 }
4189
4190 void diff_free_filespec_data(struct diff_filespec *s)
4191 {
4192 if (!s)
4193 return;
4194
4195 diff_free_filespec_blob(s);
4196 FREE_AND_NULL(s->cnt_data);
4197 }
4198
4199 static void prep_temp_blob(struct index_state *istate,
4200 const char *path, struct diff_tempfile *temp,
4201 void *blob,
4202 unsigned long size,
4203 const struct object_id *oid,
4204 int mode)
4205 {
4206 struct strbuf buf = STRBUF_INIT;
4207 char *path_dup = xstrdup(path);
4208 const char *base = basename(path_dup);
4209 struct checkout_metadata meta;
4210
4211 init_checkout_metadata(&meta, NULL, NULL, oid);
4212
4213 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4214 if (!temp->tempfile)
4215 die_errno("unable to create temp-file");
4216 if (convert_to_working_tree(istate, path,
4217 (const char *)blob, (size_t)size, &buf, &meta)) {
4218 blob = buf.buf;
4219 size = buf.len;
4220 }
4221 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4222 close_tempfile_gently(temp->tempfile))
4223 die_errno("unable to write temp-file");
4224 temp->name = get_tempfile_path(temp->tempfile);
4225 oid_to_hex_r(temp->hex, oid);
4226 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4227 strbuf_release(&buf);
4228 free(path_dup);
4229 }
4230
4231 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4232 struct diff_filespec *one)
4233 {
4234 struct diff_tempfile *temp = claim_diff_tempfile();
4235
4236 if (!DIFF_FILE_VALID(one)) {
4237 not_a_valid_file:
4238 /* A '-' entry produces this for file-2, and
4239 * a '+' entry produces this for file-1.
4240 */
4241 temp->name = "/dev/null";
4242 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4243 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4244 return temp;
4245 }
4246
4247 if (!S_ISGITLINK(one->mode) &&
4248 (!one->oid_valid ||
4249 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4250 struct stat st;
4251 if (lstat(one->path, &st) < 0) {
4252 if (errno == ENOENT)
4253 goto not_a_valid_file;
4254 die_errno("stat(%s)", one->path);
4255 }
4256 if (S_ISLNK(st.st_mode)) {
4257 struct strbuf sb = STRBUF_INIT;
4258 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4259 die_errno("readlink(%s)", one->path);
4260 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4261 (one->oid_valid ?
4262 &one->oid : null_oid()),
4263 (one->oid_valid ?
4264 one->mode : S_IFLNK));
4265 strbuf_release(&sb);
4266 }
4267 else {
4268 /* we can borrow from the file in the work tree */
4269 temp->name = one->path;
4270 if (!one->oid_valid)
4271 oid_to_hex_r(temp->hex, null_oid());
4272 else
4273 oid_to_hex_r(temp->hex, &one->oid);
4274 /* Even though we may sometimes borrow the
4275 * contents from the work tree, we always want
4276 * one->mode. mode is trustworthy even when
4277 * !(one->oid_valid), as long as
4278 * DIFF_FILE_VALID(one).
4279 */
4280 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4281 }
4282 return temp;
4283 }
4284 else {
4285 if (diff_populate_filespec(r, one, NULL))
4286 die("cannot read data blob for %s", one->path);
4287 prep_temp_blob(r->index, one->path, temp,
4288 one->data, one->size,
4289 &one->oid, one->mode);
4290 }
4291 return temp;
4292 }
4293
4294 static void add_external_diff_name(struct repository *r,
4295 struct strvec *argv,
4296 struct diff_filespec *df)
4297 {
4298 struct diff_tempfile *temp = prepare_temp_file(r, df);
4299 strvec_push(argv, temp->name);
4300 strvec_push(argv, temp->hex);
4301 strvec_push(argv, temp->mode);
4302 }
4303
4304 /* An external diff command takes:
4305 *
4306 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4307 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4308 *
4309 */
4310 static void run_external_diff(const char *pgm,
4311 const char *name,
4312 const char *other,
4313 struct diff_filespec *one,
4314 struct diff_filespec *two,
4315 const char *xfrm_msg,
4316 struct diff_options *o)
4317 {
4318 struct child_process cmd = CHILD_PROCESS_INIT;
4319 struct diff_queue_struct *q = &diff_queued_diff;
4320
4321 strvec_push(&cmd.args, pgm);
4322 strvec_push(&cmd.args, name);
4323
4324 if (one && two) {
4325 add_external_diff_name(o->repo, &cmd.args, one);
4326 add_external_diff_name(o->repo, &cmd.args, two);
4327 if (other) {
4328 strvec_push(&cmd.args, other);
4329 strvec_push(&cmd.args, xfrm_msg);
4330 }
4331 }
4332
4333 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4334 ++o->diff_path_counter);
4335 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4336
4337 diff_free_filespec_data(one);
4338 diff_free_filespec_data(two);
4339 cmd.use_shell = 1;
4340 if (run_command(&cmd))
4341 die(_("external diff died, stopping at %s"), name);
4342
4343 remove_tempfile();
4344 }
4345
4346 static int similarity_index(struct diff_filepair *p)
4347 {
4348 return p->score * 100 / MAX_SCORE;
4349 }
4350
4351 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4352 {
4353 if (startup_info->have_repository)
4354 return find_unique_abbrev(oid, abbrev);
4355 else {
4356 char *hex = oid_to_hex(oid);
4357 if (abbrev < 0)
4358 abbrev = FALLBACK_DEFAULT_ABBREV;
4359 if (abbrev > the_hash_algo->hexsz)
4360 BUG("oid abbreviation out of range: %d", abbrev);
4361 if (abbrev)
4362 hex[abbrev] = '\0';
4363 return hex;
4364 }
4365 }
4366
4367 static void fill_metainfo(struct strbuf *msg,
4368 const char *name,
4369 const char *other,
4370 struct diff_filespec *one,
4371 struct diff_filespec *two,
4372 struct diff_options *o,
4373 struct diff_filepair *p,
4374 int *must_show_header,
4375 int use_color)
4376 {
4377 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4378 const char *reset = diff_get_color(use_color, DIFF_RESET);
4379 const char *line_prefix = diff_line_prefix(o);
4380 struct string_list *more_headers = NULL;
4381
4382 *must_show_header = 1;
4383 strbuf_init(msg, PATH_MAX * 2 + 300);
4384 switch (p->status) {
4385 case DIFF_STATUS_COPIED:
4386 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4387 line_prefix, set, similarity_index(p));
4388 strbuf_addf(msg, "%s\n%s%scopy from ",
4389 reset, line_prefix, set);
4390 quote_c_style(name, msg, NULL, 0);
4391 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4392 quote_c_style(other, msg, NULL, 0);
4393 strbuf_addf(msg, "%s\n", reset);
4394 break;
4395 case DIFF_STATUS_RENAMED:
4396 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4397 line_prefix, set, similarity_index(p));
4398 strbuf_addf(msg, "%s\n%s%srename from ",
4399 reset, line_prefix, set);
4400 quote_c_style(name, msg, NULL, 0);
4401 strbuf_addf(msg, "%s\n%s%srename to ",
4402 reset, line_prefix, set);
4403 quote_c_style(other, msg, NULL, 0);
4404 strbuf_addf(msg, "%s\n", reset);
4405 break;
4406 case DIFF_STATUS_MODIFIED:
4407 if (p->score) {
4408 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4409 line_prefix,
4410 set, similarity_index(p), reset);
4411 break;
4412 }
4413 /* fallthru */
4414 default:
4415 *must_show_header = 0;
4416 }
4417 if ((more_headers = additional_headers(o, name))) {
4418 add_formatted_headers(msg, more_headers,
4419 line_prefix, set, reset);
4420 *must_show_header = 1;
4421 }
4422 if (one && two && !oideq(&one->oid, &two->oid)) {
4423 const unsigned hexsz = the_hash_algo->hexsz;
4424 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4425
4426 if (o->flags.full_index)
4427 abbrev = hexsz;
4428
4429 if (o->flags.binary) {
4430 mmfile_t mf;
4431 if ((!fill_mmfile(o->repo, &mf, one) &&
4432 diff_filespec_is_binary(o->repo, one)) ||
4433 (!fill_mmfile(o->repo, &mf, two) &&
4434 diff_filespec_is_binary(o->repo, two)))
4435 abbrev = hexsz;
4436 }
4437 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4438 diff_abbrev_oid(&one->oid, abbrev),
4439 diff_abbrev_oid(&two->oid, abbrev));
4440 if (one->mode == two->mode)
4441 strbuf_addf(msg, " %06o", one->mode);
4442 strbuf_addf(msg, "%s\n", reset);
4443 }
4444 }
4445
4446 static void run_diff_cmd(const char *pgm,
4447 const char *name,
4448 const char *other,
4449 const char *attr_path,
4450 struct diff_filespec *one,
4451 struct diff_filespec *two,
4452 struct strbuf *msg,
4453 struct diff_options *o,
4454 struct diff_filepair *p)
4455 {
4456 const char *xfrm_msg = NULL;
4457 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4458 int must_show_header = 0;
4459 struct userdiff_driver *drv = NULL;
4460
4461 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4462 drv = userdiff_find_by_path(o->repo->index, attr_path);
4463
4464 if (o->flags.allow_external && drv && drv->external)
4465 pgm = drv->external;
4466
4467 if (msg) {
4468 /*
4469 * don't use colors when the header is intended for an
4470 * external diff driver
4471 */
4472 fill_metainfo(msg, name, other, one, two, o, p,
4473 &must_show_header,
4474 want_color(o->use_color) && !pgm);
4475 xfrm_msg = msg->len ? msg->buf : NULL;
4476 }
4477
4478 if (pgm) {
4479 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4480 return;
4481 }
4482 if (one && two) {
4483 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4484 set_diff_algorithm(o, drv->algorithm);
4485
4486 builtin_diff(name, other ? other : name,
4487 one, two, xfrm_msg, must_show_header,
4488 o, complete_rewrite);
4489 } else {
4490 fprintf(o->file, "* Unmerged path %s\n", name);
4491 }
4492 }
4493
4494 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4495 {
4496 if (DIFF_FILE_VALID(one)) {
4497 if (!one->oid_valid) {
4498 struct stat st;
4499 if (one->is_stdin) {
4500 oidclr(&one->oid);
4501 return;
4502 }
4503 if (lstat(one->path, &st) < 0)
4504 die_errno("stat '%s'", one->path);
4505 if (index_path(istate, &one->oid, one->path, &st, 0))
4506 die("cannot hash %s", one->path);
4507 }
4508 }
4509 else
4510 oidclr(&one->oid);
4511 }
4512
4513 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4514 {
4515 /* Strip the prefix but do not molest /dev/null and absolute paths */
4516 if (*namep && !is_absolute_path(*namep)) {
4517 *namep += prefix_length;
4518 if (**namep == '/')
4519 ++*namep;
4520 }
4521 if (*otherp && !is_absolute_path(*otherp)) {
4522 *otherp += prefix_length;
4523 if (**otherp == '/')
4524 ++*otherp;
4525 }
4526 }
4527
4528 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4529 {
4530 const char *pgm = external_diff();
4531 struct strbuf msg;
4532 struct diff_filespec *one = p->one;
4533 struct diff_filespec *two = p->two;
4534 const char *name;
4535 const char *other;
4536 const char *attr_path;
4537
4538 name = one->path;
4539 other = (strcmp(name, two->path) ? two->path : NULL);
4540 attr_path = name;
4541 if (o->prefix_length)
4542 strip_prefix(o->prefix_length, &name, &other);
4543
4544 if (!o->flags.allow_external)
4545 pgm = NULL;
4546
4547 if (DIFF_PAIR_UNMERGED(p)) {
4548 run_diff_cmd(pgm, name, NULL, attr_path,
4549 NULL, NULL, NULL, o, p);
4550 return;
4551 }
4552
4553 diff_fill_oid_info(one, o->repo->index);
4554 diff_fill_oid_info(two, o->repo->index);
4555
4556 if (!pgm &&
4557 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4558 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4559 /*
4560 * a filepair that changes between file and symlink
4561 * needs to be split into deletion and creation.
4562 */
4563 struct diff_filespec *null = alloc_filespec(two->path);
4564 run_diff_cmd(NULL, name, other, attr_path,
4565 one, null, &msg,
4566 o, p);
4567 free(null);
4568 strbuf_release(&msg);
4569
4570 null = alloc_filespec(one->path);
4571 run_diff_cmd(NULL, name, other, attr_path,
4572 null, two, &msg, o, p);
4573 free(null);
4574 }
4575 else
4576 run_diff_cmd(pgm, name, other, attr_path,
4577 one, two, &msg, o, p);
4578
4579 strbuf_release(&msg);
4580 }
4581
4582 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4583 struct diffstat_t *diffstat)
4584 {
4585 const char *name;
4586 const char *other;
4587
4588 if (!o->ignore_driver_algorithm) {
4589 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4590 p->one->path);
4591
4592 if (drv && drv->algorithm)
4593 set_diff_algorithm(o, drv->algorithm);
4594 }
4595
4596 if (DIFF_PAIR_UNMERGED(p)) {
4597 /* unmerged */
4598 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4599 diffstat, o, p);
4600 return;
4601 }
4602
4603 name = p->one->path;
4604 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4605
4606 if (o->prefix_length)
4607 strip_prefix(o->prefix_length, &name, &other);
4608
4609 diff_fill_oid_info(p->one, o->repo->index);
4610 diff_fill_oid_info(p->two, o->repo->index);
4611
4612 builtin_diffstat(name, other, p->one, p->two,
4613 diffstat, o, p);
4614 }
4615
4616 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4617 {
4618 const char *name;
4619 const char *other;
4620 const char *attr_path;
4621
4622 if (DIFF_PAIR_UNMERGED(p)) {
4623 /* unmerged */
4624 return;
4625 }
4626
4627 name = p->one->path;
4628 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4629 attr_path = other ? other : name;
4630
4631 if (o->prefix_length)
4632 strip_prefix(o->prefix_length, &name, &other);
4633
4634 diff_fill_oid_info(p->one, o->repo->index);
4635 diff_fill_oid_info(p->two, o->repo->index);
4636
4637 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4638 }
4639
4640 void repo_diff_setup(struct repository *r, struct diff_options *options)
4641 {
4642 memcpy(options, &default_diff_options, sizeof(*options));
4643
4644 options->file = stdout;
4645 options->repo = r;
4646
4647 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4648 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4649 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4650 options->abbrev = DEFAULT_ABBREV;
4651 options->line_termination = '\n';
4652 options->break_opt = -1;
4653 options->rename_limit = -1;
4654 options->dirstat_permille = diff_dirstat_permille_default;
4655 options->context = diff_context_default;
4656 options->interhunkcontext = diff_interhunk_context_default;
4657 options->ws_error_highlight = ws_error_highlight_default;
4658 options->flags.rename_empty = 1;
4659 options->flags.relative_name = diff_relative;
4660 options->objfind = NULL;
4661
4662 /* pathchange left =NULL by default */
4663 options->change = diff_change;
4664 options->add_remove = diff_addremove;
4665 options->use_color = diff_use_color_default;
4666 options->detect_rename = diff_detect_rename_default;
4667 options->xdl_opts |= diff_algorithm;
4668 if (diff_indent_heuristic)
4669 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4670
4671 options->orderfile = diff_order_file_cfg;
4672
4673 if (!options->flags.ignore_submodule_set)
4674 options->flags.ignore_untracked_in_submodules = 1;
4675
4676 if (diff_no_prefix) {
4677 options->a_prefix = options->b_prefix = "";
4678 } else if (!diff_mnemonic_prefix) {
4679 options->a_prefix = "a/";
4680 options->b_prefix = "b/";
4681 }
4682
4683 options->color_moved = diff_color_moved_default;
4684 options->color_moved_ws_handling = diff_color_moved_ws_default;
4685 }
4686
4687 static const char diff_status_letters[] = {
4688 DIFF_STATUS_ADDED,
4689 DIFF_STATUS_COPIED,
4690 DIFF_STATUS_DELETED,
4691 DIFF_STATUS_MODIFIED,
4692 DIFF_STATUS_RENAMED,
4693 DIFF_STATUS_TYPE_CHANGED,
4694 DIFF_STATUS_UNKNOWN,
4695 DIFF_STATUS_UNMERGED,
4696 DIFF_STATUS_FILTER_AON,
4697 DIFF_STATUS_FILTER_BROKEN,
4698 '\0',
4699 };
4700
4701 static unsigned int filter_bit['Z' + 1];
4702
4703 static void prepare_filter_bits(void)
4704 {
4705 int i;
4706
4707 if (!filter_bit[DIFF_STATUS_ADDED]) {
4708 for (i = 0; diff_status_letters[i]; i++)
4709 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4710 }
4711 }
4712
4713 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4714 {
4715 return opt->filter & filter_bit[(int) status];
4716 }
4717
4718 unsigned diff_filter_bit(char status)
4719 {
4720 prepare_filter_bits();
4721 return filter_bit[(int) status];
4722 }
4723
4724 void diff_setup_done(struct diff_options *options)
4725 {
4726 unsigned check_mask = DIFF_FORMAT_NAME |
4727 DIFF_FORMAT_NAME_STATUS |
4728 DIFF_FORMAT_CHECKDIFF |
4729 DIFF_FORMAT_NO_OUTPUT;
4730 /*
4731 * This must be signed because we're comparing against a potentially
4732 * negative value.
4733 */
4734 const int hexsz = the_hash_algo->hexsz;
4735
4736 if (options->set_default)
4737 options->set_default(options);
4738
4739 if (HAS_MULTI_BITS(options->output_format & check_mask))
4740 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4741 "--name-only", "--name-status", "--check", "-s");
4742
4743 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4744 die(_("options '%s', '%s', and '%s' cannot be used together"),
4745 "-G", "-S", "--find-object");
4746
4747 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4748 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4749 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4750
4751 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4752 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4753 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4754
4755 /*
4756 * Most of the time we can say "there are changes"
4757 * only by checking if there are changed paths, but
4758 * --ignore-whitespace* options force us to look
4759 * inside contents.
4760 */
4761
4762 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4763 options->ignore_regex_nr)
4764 options->flags.diff_from_contents = 1;
4765 else
4766 options->flags.diff_from_contents = 0;
4767
4768 if (options->flags.find_copies_harder)
4769 options->detect_rename = DIFF_DETECT_COPY;
4770
4771 if (!options->flags.relative_name)
4772 options->prefix = NULL;
4773 if (options->prefix)
4774 options->prefix_length = strlen(options->prefix);
4775 else
4776 options->prefix_length = 0;
4777
4778 if (options->output_format & (DIFF_FORMAT_NAME |
4779 DIFF_FORMAT_NAME_STATUS |
4780 DIFF_FORMAT_CHECKDIFF |
4781 DIFF_FORMAT_NO_OUTPUT))
4782 options->output_format &= ~(DIFF_FORMAT_RAW |
4783 DIFF_FORMAT_NUMSTAT |
4784 DIFF_FORMAT_DIFFSTAT |
4785 DIFF_FORMAT_SHORTSTAT |
4786 DIFF_FORMAT_DIRSTAT |
4787 DIFF_FORMAT_SUMMARY |
4788 DIFF_FORMAT_PATCH);
4789
4790 /*
4791 * These cases always need recursive; we do not drop caller-supplied
4792 * recursive bits for other formats here.
4793 */
4794 if (options->output_format & (DIFF_FORMAT_PATCH |
4795 DIFF_FORMAT_NUMSTAT |
4796 DIFF_FORMAT_DIFFSTAT |
4797 DIFF_FORMAT_SHORTSTAT |
4798 DIFF_FORMAT_DIRSTAT |
4799 DIFF_FORMAT_SUMMARY |
4800 DIFF_FORMAT_CHECKDIFF))
4801 options->flags.recursive = 1;
4802 /*
4803 * Also pickaxe would not work very well if you do not say recursive
4804 */
4805 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4806 options->flags.recursive = 1;
4807 /*
4808 * When patches are generated, submodules diffed against the work tree
4809 * must be checked for dirtiness too so it can be shown in the output
4810 */
4811 if (options->output_format & DIFF_FORMAT_PATCH)
4812 options->flags.dirty_submodules = 1;
4813
4814 if (options->detect_rename && options->rename_limit < 0)
4815 options->rename_limit = diff_rename_limit_default;
4816 if (hexsz < options->abbrev)
4817 options->abbrev = hexsz; /* full */
4818
4819 /*
4820 * It does not make sense to show the first hit we happened
4821 * to have found. It does not make sense not to return with
4822 * exit code in such a case either.
4823 */
4824 if (options->flags.quick) {
4825 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4826 options->flags.exit_with_status = 1;
4827 }
4828
4829 options->diff_path_counter = 0;
4830
4831 if (options->flags.follow_renames && options->pathspec.nr != 1)
4832 die(_("--follow requires exactly one pathspec"));
4833
4834 if (!options->use_color || external_diff())
4835 options->color_moved = 0;
4836
4837 if (options->filter_not) {
4838 if (!options->filter)
4839 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4840 options->filter &= ~options->filter_not;
4841 }
4842 }
4843
4844 int parse_long_opt(const char *opt, const char **argv,
4845 const char **optarg)
4846 {
4847 const char *arg = argv[0];
4848 if (!skip_prefix(arg, "--", &arg))
4849 return 0;
4850 if (!skip_prefix(arg, opt, &arg))
4851 return 0;
4852 if (*arg == '=') { /* stuck form: --option=value */
4853 *optarg = arg + 1;
4854 return 1;
4855 }
4856 if (*arg != '\0')
4857 return 0;
4858 /* separate form: --option value */
4859 if (!argv[1])
4860 die("Option '--%s' requires a value", opt);
4861 *optarg = argv[1];
4862 return 2;
4863 }
4864
4865 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4866 {
4867 struct diff_options *options = opt->value;
4868 int width = options->stat_width;
4869 int name_width = options->stat_name_width;
4870 int graph_width = options->stat_graph_width;
4871 int count = options->stat_count;
4872 char *end;
4873
4874 BUG_ON_OPT_NEG(unset);
4875
4876 if (!strcmp(opt->long_name, "stat")) {
4877 if (value) {
4878 width = strtoul(value, &end, 10);
4879 if (*end == ',')
4880 name_width = strtoul(end+1, &end, 10);
4881 if (*end == ',')
4882 count = strtoul(end+1, &end, 10);
4883 if (*end)
4884 return error(_("invalid --stat value: %s"), value);
4885 }
4886 } else if (!strcmp(opt->long_name, "stat-width")) {
4887 width = strtoul(value, &end, 10);
4888 if (*end)
4889 return error(_("%s expects a numerical value"),
4890 opt->long_name);
4891 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4892 name_width = strtoul(value, &end, 10);
4893 if (*end)
4894 return error(_("%s expects a numerical value"),
4895 opt->long_name);
4896 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4897 graph_width = strtoul(value, &end, 10);
4898 if (*end)
4899 return error(_("%s expects a numerical value"),
4900 opt->long_name);
4901 } else if (!strcmp(opt->long_name, "stat-count")) {
4902 count = strtoul(value, &end, 10);
4903 if (*end)
4904 return error(_("%s expects a numerical value"),
4905 opt->long_name);
4906 } else
4907 BUG("%s should not get here", opt->long_name);
4908
4909 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4910 options->stat_name_width = name_width;
4911 options->stat_graph_width = graph_width;
4912 options->stat_width = width;
4913 options->stat_count = count;
4914 return 0;
4915 }
4916
4917 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4918 {
4919 struct strbuf errmsg = STRBUF_INIT;
4920 if (parse_dirstat_params(options, params, &errmsg))
4921 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4922 errmsg.buf);
4923 strbuf_release(&errmsg);
4924 /*
4925 * The caller knows a dirstat-related option is given from the command
4926 * line; allow it to say "return this_function();"
4927 */
4928 options->output_format |= DIFF_FORMAT_DIRSTAT;
4929 return 1;
4930 }
4931
4932 static int diff_opt_diff_filter(const struct option *option,
4933 const char *optarg, int unset)
4934 {
4935 struct diff_options *opt = option->value;
4936 int i, optch;
4937
4938 BUG_ON_OPT_NEG(unset);
4939 prepare_filter_bits();
4940
4941 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4942 unsigned int bit;
4943 int negate;
4944
4945 if ('a' <= optch && optch <= 'z') {
4946 negate = 1;
4947 optch = toupper(optch);
4948 } else {
4949 negate = 0;
4950 }
4951
4952 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4953 if (!bit)
4954 return error(_("unknown change class '%c' in --diff-filter=%s"),
4955 optarg[i], optarg);
4956 if (negate)
4957 opt->filter_not |= bit;
4958 else
4959 opt->filter |= bit;
4960 }
4961 return 0;
4962 }
4963
4964 static void enable_patch_output(int *fmt)
4965 {
4966 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4967 *fmt |= DIFF_FORMAT_PATCH;
4968 }
4969
4970 static int diff_opt_ws_error_highlight(const struct option *option,
4971 const char *arg, int unset)
4972 {
4973 struct diff_options *opt = option->value;
4974 int val = parse_ws_error_highlight(arg);
4975
4976 BUG_ON_OPT_NEG(unset);
4977 if (val < 0)
4978 return error(_("unknown value after ws-error-highlight=%.*s"),
4979 -1 - val, arg);
4980 opt->ws_error_highlight = val;
4981 return 0;
4982 }
4983
4984 static int diff_opt_find_object(const struct option *option,
4985 const char *arg, int unset)
4986 {
4987 struct diff_options *opt = option->value;
4988 struct object_id oid;
4989
4990 BUG_ON_OPT_NEG(unset);
4991 if (get_oid(arg, &oid))
4992 return error(_("unable to resolve '%s'"), arg);
4993
4994 if (!opt->objfind)
4995 CALLOC_ARRAY(opt->objfind, 1);
4996
4997 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4998 opt->flags.recursive = 1;
4999 opt->flags.tree_in_recursive = 1;
5000 oidset_insert(opt->objfind, &oid);
5001 return 0;
5002 }
5003
5004 static int diff_opt_anchored(const struct option *opt,
5005 const char *arg, int unset)
5006 {
5007 struct diff_options *options = opt->value;
5008
5009 BUG_ON_OPT_NEG(unset);
5010 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5011 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5012 options->anchors_alloc);
5013 options->anchors[options->anchors_nr++] = xstrdup(arg);
5014 return 0;
5015 }
5016
5017 static int diff_opt_binary(const struct option *opt,
5018 const char *arg, int unset)
5019 {
5020 struct diff_options *options = opt->value;
5021
5022 BUG_ON_OPT_NEG(unset);
5023 BUG_ON_OPT_ARG(arg);
5024 enable_patch_output(&options->output_format);
5025 options->flags.binary = 1;
5026 return 0;
5027 }
5028
5029 static int diff_opt_break_rewrites(const struct option *opt,
5030 const char *arg, int unset)
5031 {
5032 int *break_opt = opt->value;
5033 int opt1, opt2;
5034
5035 BUG_ON_OPT_NEG(unset);
5036 if (!arg)
5037 arg = "";
5038 opt1 = parse_rename_score(&arg);
5039 if (*arg == 0)
5040 opt2 = 0;
5041 else if (*arg != '/')
5042 return error(_("%s expects <n>/<m> form"), opt->long_name);
5043 else {
5044 arg++;
5045 opt2 = parse_rename_score(&arg);
5046 }
5047 if (*arg != 0)
5048 return error(_("%s expects <n>/<m> form"), opt->long_name);
5049 *break_opt = opt1 | (opt2 << 16);
5050 return 0;
5051 }
5052
5053 static int diff_opt_char(const struct option *opt,
5054 const char *arg, int unset)
5055 {
5056 char *value = opt->value;
5057
5058 BUG_ON_OPT_NEG(unset);
5059 if (arg[1])
5060 return error(_("%s expects a character, got '%s'"),
5061 opt->long_name, arg);
5062 *value = arg[0];
5063 return 0;
5064 }
5065
5066 static int diff_opt_color_moved(const struct option *opt,
5067 const char *arg, int unset)
5068 {
5069 struct diff_options *options = opt->value;
5070
5071 if (unset) {
5072 options->color_moved = COLOR_MOVED_NO;
5073 } else if (!arg) {
5074 if (diff_color_moved_default)
5075 options->color_moved = diff_color_moved_default;
5076 if (options->color_moved == COLOR_MOVED_NO)
5077 options->color_moved = COLOR_MOVED_DEFAULT;
5078 } else {
5079 int cm = parse_color_moved(arg);
5080 if (cm < 0)
5081 return error(_("bad --color-moved argument: %s"), arg);
5082 options->color_moved = cm;
5083 }
5084 return 0;
5085 }
5086
5087 static int diff_opt_color_moved_ws(const struct option *opt,
5088 const char *arg, int unset)
5089 {
5090 struct diff_options *options = opt->value;
5091 unsigned cm;
5092
5093 if (unset) {
5094 options->color_moved_ws_handling = 0;
5095 return 0;
5096 }
5097
5098 cm = parse_color_moved_ws(arg);
5099 if (cm & COLOR_MOVED_WS_ERROR)
5100 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5101 options->color_moved_ws_handling = cm;
5102 return 0;
5103 }
5104
5105 static int diff_opt_color_words(const struct option *opt,
5106 const char *arg, int unset)
5107 {
5108 struct diff_options *options = opt->value;
5109
5110 BUG_ON_OPT_NEG(unset);
5111 options->use_color = 1;
5112 options->word_diff = DIFF_WORDS_COLOR;
5113 options->word_regex = arg;
5114 return 0;
5115 }
5116
5117 static int diff_opt_compact_summary(const struct option *opt,
5118 const char *arg, int unset)
5119 {
5120 struct diff_options *options = opt->value;
5121
5122 BUG_ON_OPT_ARG(arg);
5123 if (unset) {
5124 options->flags.stat_with_summary = 0;
5125 } else {
5126 options->flags.stat_with_summary = 1;
5127 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5128 }
5129 return 0;
5130 }
5131
5132 static int diff_opt_diff_algorithm(const struct option *opt,
5133 const char *arg, int unset)
5134 {
5135 struct diff_options *options = opt->value;
5136
5137 BUG_ON_OPT_NEG(unset);
5138
5139 if (set_diff_algorithm(options, arg))
5140 return error(_("option diff-algorithm accepts \"myers\", "
5141 "\"minimal\", \"patience\" and \"histogram\""));
5142
5143 options->ignore_driver_algorithm = 1;
5144
5145 return 0;
5146 }
5147
5148 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5149 const char *arg, int unset)
5150 {
5151 struct diff_options *options = opt->value;
5152
5153 BUG_ON_OPT_NEG(unset);
5154 BUG_ON_OPT_ARG(arg);
5155
5156 if (set_diff_algorithm(options, opt->long_name))
5157 BUG("available diff algorithms include \"myers\", "
5158 "\"minimal\", \"patience\" and \"histogram\"");
5159
5160 options->ignore_driver_algorithm = 1;
5161
5162 return 0;
5163 }
5164
5165 static int diff_opt_dirstat(const struct option *opt,
5166 const char *arg, int unset)
5167 {
5168 struct diff_options *options = opt->value;
5169
5170 BUG_ON_OPT_NEG(unset);
5171 if (!strcmp(opt->long_name, "cumulative")) {
5172 if (arg)
5173 BUG("how come --cumulative take a value?");
5174 arg = "cumulative";
5175 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5176 parse_dirstat_opt(options, "files");
5177 parse_dirstat_opt(options, arg ? arg : "");
5178 return 0;
5179 }
5180
5181 static int diff_opt_find_copies(const struct option *opt,
5182 const char *arg, int unset)
5183 {
5184 struct diff_options *options = opt->value;
5185
5186 BUG_ON_OPT_NEG(unset);
5187 if (!arg)
5188 arg = "";
5189 options->rename_score = parse_rename_score(&arg);
5190 if (*arg != 0)
5191 return error(_("invalid argument to %s"), opt->long_name);
5192
5193 if (options->detect_rename == DIFF_DETECT_COPY)
5194 options->flags.find_copies_harder = 1;
5195 else
5196 options->detect_rename = DIFF_DETECT_COPY;
5197
5198 return 0;
5199 }
5200
5201 static int diff_opt_find_renames(const struct option *opt,
5202 const char *arg, int unset)
5203 {
5204 struct diff_options *options = opt->value;
5205
5206 BUG_ON_OPT_NEG(unset);
5207 if (!arg)
5208 arg = "";
5209 options->rename_score = parse_rename_score(&arg);
5210 if (*arg != 0)
5211 return error(_("invalid argument to %s"), opt->long_name);
5212
5213 options->detect_rename = DIFF_DETECT_RENAME;
5214 return 0;
5215 }
5216
5217 static int diff_opt_follow(const struct option *opt,
5218 const char *arg, int unset)
5219 {
5220 struct diff_options *options = opt->value;
5221
5222 BUG_ON_OPT_ARG(arg);
5223 if (unset) {
5224 options->flags.follow_renames = 0;
5225 options->flags.default_follow_renames = 0;
5226 } else {
5227 options->flags.follow_renames = 1;
5228 }
5229 return 0;
5230 }
5231
5232 static int diff_opt_ignore_submodules(const struct option *opt,
5233 const char *arg, int unset)
5234 {
5235 struct diff_options *options = opt->value;
5236
5237 BUG_ON_OPT_NEG(unset);
5238 if (!arg)
5239 arg = "all";
5240 options->flags.override_submodule_config = 1;
5241 handle_ignore_submodules_arg(options, arg);
5242 return 0;
5243 }
5244
5245 static int diff_opt_line_prefix(const struct option *opt,
5246 const char *optarg, int unset)
5247 {
5248 struct diff_options *options = opt->value;
5249
5250 BUG_ON_OPT_NEG(unset);
5251 options->line_prefix = optarg;
5252 options->line_prefix_length = strlen(options->line_prefix);
5253 graph_setup_line_prefix(options);
5254 return 0;
5255 }
5256
5257 static int diff_opt_no_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 BUG_ON_OPT_ARG(optarg);
5264 options->a_prefix = "";
5265 options->b_prefix = "";
5266 return 0;
5267 }
5268
5269 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5270 const struct option *opt,
5271 const char *arg, int unset)
5272 {
5273 struct diff_options *options = opt->value;
5274 char *path;
5275
5276 BUG_ON_OPT_NEG(unset);
5277 path = prefix_filename(ctx->prefix, arg);
5278 options->file = xfopen(path, "w");
5279 options->close_file = 1;
5280 if (options->use_color != GIT_COLOR_ALWAYS)
5281 options->use_color = GIT_COLOR_NEVER;
5282 free(path);
5283 return 0;
5284 }
5285
5286 static int diff_opt_patience(const struct option *opt,
5287 const char *arg, int unset)
5288 {
5289 struct diff_options *options = opt->value;
5290 int i;
5291
5292 BUG_ON_OPT_NEG(unset);
5293 BUG_ON_OPT_ARG(arg);
5294 /*
5295 * Both --patience and --anchored use PATIENCE_DIFF
5296 * internally, so remove any anchors previously
5297 * specified.
5298 */
5299 for (i = 0; i < options->anchors_nr; i++)
5300 free(options->anchors[i]);
5301 options->anchors_nr = 0;
5302 options->ignore_driver_algorithm = 1;
5303
5304 return set_diff_algorithm(options, "patience");
5305 }
5306
5307 static int diff_opt_ignore_regex(const struct option *opt,
5308 const char *arg, int unset)
5309 {
5310 struct diff_options *options = opt->value;
5311 regex_t *regex;
5312
5313 BUG_ON_OPT_NEG(unset);
5314 regex = xmalloc(sizeof(*regex));
5315 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5316 return error(_("invalid regex given to -I: '%s'"), arg);
5317 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5318 options->ignore_regex_alloc);
5319 options->ignore_regex[options->ignore_regex_nr++] = regex;
5320 return 0;
5321 }
5322
5323 static int diff_opt_pickaxe_regex(const struct option *opt,
5324 const char *arg, int unset)
5325 {
5326 struct diff_options *options = opt->value;
5327
5328 BUG_ON_OPT_NEG(unset);
5329 options->pickaxe = arg;
5330 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5331 return 0;
5332 }
5333
5334 static int diff_opt_pickaxe_string(const struct option *opt,
5335 const char *arg, int unset)
5336 {
5337 struct diff_options *options = opt->value;
5338
5339 BUG_ON_OPT_NEG(unset);
5340 options->pickaxe = arg;
5341 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5342 return 0;
5343 }
5344
5345 static int diff_opt_relative(const struct option *opt,
5346 const char *arg, int unset)
5347 {
5348 struct diff_options *options = opt->value;
5349
5350 options->flags.relative_name = !unset;
5351 if (arg)
5352 options->prefix = arg;
5353 return 0;
5354 }
5355
5356 static int diff_opt_submodule(const struct option *opt,
5357 const char *arg, int unset)
5358 {
5359 struct diff_options *options = opt->value;
5360
5361 BUG_ON_OPT_NEG(unset);
5362 if (!arg)
5363 arg = "log";
5364 if (parse_submodule_params(options, arg))
5365 return error(_("failed to parse --submodule option parameter: '%s'"),
5366 arg);
5367 return 0;
5368 }
5369
5370 static int diff_opt_textconv(const struct option *opt,
5371 const char *arg, int unset)
5372 {
5373 struct diff_options *options = opt->value;
5374
5375 BUG_ON_OPT_ARG(arg);
5376 if (unset) {
5377 options->flags.allow_textconv = 0;
5378 } else {
5379 options->flags.allow_textconv = 1;
5380 options->flags.textconv_set_via_cmdline = 1;
5381 }
5382 return 0;
5383 }
5384
5385 static int diff_opt_unified(const struct option *opt,
5386 const char *arg, int unset)
5387 {
5388 struct diff_options *options = opt->value;
5389 char *s;
5390
5391 BUG_ON_OPT_NEG(unset);
5392
5393 if (arg) {
5394 options->context = strtol(arg, &s, 10);
5395 if (*s)
5396 return error(_("%s expects a numerical value"), "--unified");
5397 }
5398 enable_patch_output(&options->output_format);
5399
5400 return 0;
5401 }
5402
5403 static int diff_opt_word_diff(const struct option *opt,
5404 const char *arg, int unset)
5405 {
5406 struct diff_options *options = opt->value;
5407
5408 BUG_ON_OPT_NEG(unset);
5409 if (arg) {
5410 if (!strcmp(arg, "plain"))
5411 options->word_diff = DIFF_WORDS_PLAIN;
5412 else if (!strcmp(arg, "color")) {
5413 options->use_color = 1;
5414 options->word_diff = DIFF_WORDS_COLOR;
5415 }
5416 else if (!strcmp(arg, "porcelain"))
5417 options->word_diff = DIFF_WORDS_PORCELAIN;
5418 else if (!strcmp(arg, "none"))
5419 options->word_diff = DIFF_WORDS_NONE;
5420 else
5421 return error(_("bad --word-diff argument: %s"), arg);
5422 } else {
5423 if (options->word_diff == DIFF_WORDS_NONE)
5424 options->word_diff = DIFF_WORDS_PLAIN;
5425 }
5426 return 0;
5427 }
5428
5429 static int diff_opt_word_diff_regex(const struct option *opt,
5430 const char *arg, int unset)
5431 {
5432 struct diff_options *options = opt->value;
5433
5434 BUG_ON_OPT_NEG(unset);
5435 if (options->word_diff == DIFF_WORDS_NONE)
5436 options->word_diff = DIFF_WORDS_PLAIN;
5437 options->word_regex = arg;
5438 return 0;
5439 }
5440
5441 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5442 {
5443 struct diff_options *options = opt->value;
5444
5445 BUG_ON_OPT_NEG(unset);
5446 if (!strcmp(opt->long_name, "skip-to"))
5447 options->skip_instead_of_rotate = 1;
5448 else
5449 options->skip_instead_of_rotate = 0;
5450 options->rotate_to = arg;
5451 return 0;
5452 }
5453
5454 struct option *add_diff_options(const struct option *opts,
5455 struct diff_options *options)
5456 {
5457 struct option parseopts[] = {
5458 OPT_GROUP(N_("Diff output format options")),
5459 OPT_BITOP('p', "patch", &options->output_format,
5460 N_("generate patch"),
5461 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5462 OPT_BIT_F('s', "no-patch", &options->output_format,
5463 N_("suppress diff output"),
5464 DIFF_FORMAT_NO_OUTPUT, PARSE_OPT_NONEG),
5465 OPT_BITOP('u', NULL, &options->output_format,
5466 N_("generate patch"),
5467 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5468 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5469 N_("generate diffs with <n> lines context"),
5470 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5471 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5472 N_("generate diffs with <n> lines context")),
5473 OPT_BIT_F(0, "raw", &options->output_format,
5474 N_("generate the diff in raw format"),
5475 DIFF_FORMAT_RAW, PARSE_OPT_NONEG),
5476 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5477 N_("synonym for '-p --raw'"),
5478 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5479 DIFF_FORMAT_NO_OUTPUT),
5480 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5481 N_("synonym for '-p --stat'"),
5482 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5483 DIFF_FORMAT_NO_OUTPUT),
5484 OPT_BIT_F(0, "numstat", &options->output_format,
5485 N_("machine friendly --stat"),
5486 DIFF_FORMAT_NUMSTAT, PARSE_OPT_NONEG),
5487 OPT_BIT_F(0, "shortstat", &options->output_format,
5488 N_("output only the last line of --stat"),
5489 DIFF_FORMAT_SHORTSTAT, PARSE_OPT_NONEG),
5490 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1,param2>..."),
5491 N_("output the distribution of relative amount of changes for each sub-directory"),
5492 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5493 diff_opt_dirstat),
5494 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5495 N_("synonym for --dirstat=cumulative"),
5496 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5497 diff_opt_dirstat),
5498 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1,param2>..."),
5499 N_("synonym for --dirstat=files,param1,param2..."),
5500 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5501 diff_opt_dirstat),
5502 OPT_BIT_F(0, "check", &options->output_format,
5503 N_("warn if changes introduce conflict markers or whitespace errors"),
5504 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5505 OPT_BIT_F(0, "summary", &options->output_format,
5506 N_("condensed summary such as creations, renames and mode changes"),
5507 DIFF_FORMAT_SUMMARY, PARSE_OPT_NONEG),
5508 OPT_BIT_F(0, "name-only", &options->output_format,
5509 N_("show only names of changed files"),
5510 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5511 OPT_BIT_F(0, "name-status", &options->output_format,
5512 N_("show only names and status of changed files"),
5513 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5514 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5515 N_("generate diffstat"),
5516 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5517 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5518 N_("generate diffstat with a given width"),
5519 PARSE_OPT_NONEG, diff_opt_stat),
5520 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5521 N_("generate diffstat with a given name width"),
5522 PARSE_OPT_NONEG, diff_opt_stat),
5523 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5524 N_("generate diffstat with a given graph width"),
5525 PARSE_OPT_NONEG, diff_opt_stat),
5526 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5527 N_("generate diffstat with limited lines"),
5528 PARSE_OPT_NONEG, diff_opt_stat),
5529 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5530 N_("generate compact summary in diffstat"),
5531 PARSE_OPT_NOARG, diff_opt_compact_summary),
5532 OPT_CALLBACK_F(0, "binary", options, NULL,
5533 N_("output a binary diff that can be applied"),
5534 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5535 OPT_BOOL(0, "full-index", &options->flags.full_index,
5536 N_("show full pre- and post-image object names on the \"index\" lines")),
5537 OPT_COLOR_FLAG(0, "color", &options->use_color,
5538 N_("show colored diff")),
5539 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5540 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5541 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5542 OPT_SET_INT('z', NULL, &options->line_termination,
5543 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5544 0),
5545 OPT__ABBREV(&options->abbrev),
5546 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5547 N_("show the given source prefix instead of \"a/\""),
5548 PARSE_OPT_NONEG),
5549 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5550 N_("show the given destination prefix instead of \"b/\""),
5551 PARSE_OPT_NONEG),
5552 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5553 N_("prepend an additional prefix to every line of output"),
5554 PARSE_OPT_NONEG, diff_opt_line_prefix),
5555 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5556 N_("do not show any source or destination prefix"),
5557 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5558 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5559 N_("show context between diff hunks up to the specified number of lines"),
5560 PARSE_OPT_NONEG),
5561 OPT_CALLBACK_F(0, "output-indicator-new",
5562 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5563 N_("<char>"),
5564 N_("specify the character to indicate a new line instead of '+'"),
5565 PARSE_OPT_NONEG, diff_opt_char),
5566 OPT_CALLBACK_F(0, "output-indicator-old",
5567 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5568 N_("<char>"),
5569 N_("specify the character to indicate an old line instead of '-'"),
5570 PARSE_OPT_NONEG, diff_opt_char),
5571 OPT_CALLBACK_F(0, "output-indicator-context",
5572 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5573 N_("<char>"),
5574 N_("specify the character to indicate a context instead of ' '"),
5575 PARSE_OPT_NONEG, diff_opt_char),
5576
5577 OPT_GROUP(N_("Diff rename options")),
5578 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5579 N_("break complete rewrite changes into pairs of delete and create"),
5580 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5581 diff_opt_break_rewrites),
5582 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5583 N_("detect renames"),
5584 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5585 diff_opt_find_renames),
5586 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5587 N_("omit the preimage for deletes"),
5588 1, PARSE_OPT_NONEG),
5589 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5590 N_("detect copies"),
5591 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5592 diff_opt_find_copies),
5593 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5594 N_("use unmodified files as source to find copies")),
5595 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5596 N_("disable rename detection"),
5597 0, PARSE_OPT_NONEG),
5598 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5599 N_("use empty blobs as rename source")),
5600 OPT_CALLBACK_F(0, "follow", options, NULL,
5601 N_("continue listing the history of a file beyond renames"),
5602 PARSE_OPT_NOARG, diff_opt_follow),
5603 OPT_INTEGER('l', NULL, &options->rename_limit,
5604 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5605
5606 OPT_GROUP(N_("Diff algorithm options")),
5607 OPT_CALLBACK_F(0, "minimal", options, NULL,
5608 N_("produce the smallest possible diff"),
5609 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5610 diff_opt_diff_algorithm_no_arg),
5611 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5612 N_("ignore whitespace when comparing lines"),
5613 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5614 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5615 N_("ignore changes in amount of whitespace"),
5616 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5617 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5618 N_("ignore changes in whitespace at EOL"),
5619 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5620 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5621 N_("ignore carrier-return at the end of line"),
5622 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5623 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5624 N_("ignore changes whose lines are all blank"),
5625 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5626 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5627 N_("ignore changes whose all lines match <regex>"),
5628 0, diff_opt_ignore_regex),
5629 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5630 N_("heuristic to shift diff hunk boundaries for easy reading"),
5631 XDF_INDENT_HEURISTIC),
5632 OPT_CALLBACK_F(0, "patience", options, NULL,
5633 N_("generate diff using the \"patience diff\" algorithm"),
5634 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5635 diff_opt_patience),
5636 OPT_CALLBACK_F(0, "histogram", options, NULL,
5637 N_("generate diff using the \"histogram diff\" algorithm"),
5638 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5639 diff_opt_diff_algorithm_no_arg),
5640 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5641 N_("choose a diff algorithm"),
5642 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5643 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5644 N_("generate diff using the \"anchored diff\" algorithm"),
5645 PARSE_OPT_NONEG, diff_opt_anchored),
5646 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5647 N_("show word diff, using <mode> to delimit changed words"),
5648 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5649 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5650 N_("use <regex> to decide what a word is"),
5651 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5652 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5653 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5654 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5655 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5656 N_("moved lines of code are colored differently"),
5657 PARSE_OPT_OPTARG, diff_opt_color_moved),
5658 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5659 N_("how white spaces are ignored in --color-moved"),
5660 0, diff_opt_color_moved_ws),
5661
5662 OPT_GROUP(N_("Other diff options")),
5663 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5664 N_("when run from subdir, exclude changes outside and show relative paths"),
5665 PARSE_OPT_OPTARG,
5666 diff_opt_relative),
5667 OPT_BOOL('a', "text", &options->flags.text,
5668 N_("treat all files as text")),
5669 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5670 N_("swap two inputs, reverse the diff")),
5671 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5672 N_("exit with 1 if there were differences, 0 otherwise")),
5673 OPT_BOOL(0, "quiet", &options->flags.quick,
5674 N_("disable all output of the program")),
5675 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5676 N_("allow an external diff helper to be executed")),
5677 OPT_CALLBACK_F(0, "textconv", options, NULL,
5678 N_("run external text conversion filters when comparing binary files"),
5679 PARSE_OPT_NOARG, diff_opt_textconv),
5680 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5681 N_("ignore changes to submodules in the diff generation"),
5682 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5683 diff_opt_ignore_submodules),
5684 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5685 N_("specify how differences in submodules are shown"),
5686 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5687 diff_opt_submodule),
5688 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5689 N_("hide 'git add -N' entries from the index"),
5690 1, PARSE_OPT_NONEG),
5691 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5692 N_("treat 'git add -N' entries as real in the index"),
5693 0, PARSE_OPT_NONEG),
5694 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5695 N_("look for differences that change the number of occurrences of the specified string"),
5696 0, diff_opt_pickaxe_string),
5697 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5698 N_("look for differences that change the number of occurrences of the specified regex"),
5699 0, diff_opt_pickaxe_regex),
5700 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5701 N_("show all changes in the changeset with -S or -G"),
5702 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5703 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5704 N_("treat <string> in -S as extended POSIX regular expression"),
5705 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5706 OPT_FILENAME('O', NULL, &options->orderfile,
5707 N_("control the order in which files appear in the output")),
5708 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5709 N_("show the change in the specified path first"),
5710 PARSE_OPT_NONEG, diff_opt_rotate_to),
5711 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5712 N_("skip the output to the specified path"),
5713 PARSE_OPT_NONEG, diff_opt_rotate_to),
5714 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5715 N_("look for differences that change the number of occurrences of the specified object"),
5716 PARSE_OPT_NONEG, diff_opt_find_object),
5717 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5718 N_("select files by diff type"),
5719 PARSE_OPT_NONEG, diff_opt_diff_filter),
5720 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5721 N_("output to a specific file"),
5722 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5723
5724 OPT_END()
5725 };
5726
5727 return parse_options_concat(opts, parseopts);
5728 }
5729
5730 int diff_opt_parse(struct diff_options *options,
5731 const char **av, int ac, const char *prefix)
5732 {
5733 struct option no_options[] = { OPT_END() };
5734 struct option *parseopts = add_diff_options(no_options, options);
5735
5736 if (!prefix)
5737 prefix = "";
5738
5739 ac = parse_options(ac, av, prefix, parseopts, NULL,
5740 PARSE_OPT_KEEP_DASHDASH |
5741 PARSE_OPT_KEEP_UNKNOWN_OPT |
5742 PARSE_OPT_NO_INTERNAL_HELP |
5743 PARSE_OPT_ONE_SHOT |
5744 PARSE_OPT_STOP_AT_NON_OPTION);
5745 free(parseopts);
5746
5747 return ac;
5748 }
5749
5750 int parse_rename_score(const char **cp_p)
5751 {
5752 unsigned long num, scale;
5753 int ch, dot;
5754 const char *cp = *cp_p;
5755
5756 num = 0;
5757 scale = 1;
5758 dot = 0;
5759 for (;;) {
5760 ch = *cp;
5761 if ( !dot && ch == '.' ) {
5762 scale = 1;
5763 dot = 1;
5764 } else if ( ch == '%' ) {
5765 scale = dot ? scale*100 : 100;
5766 cp++; /* % is always at the end */
5767 break;
5768 } else if ( ch >= '0' && ch <= '9' ) {
5769 if ( scale < 100000 ) {
5770 scale *= 10;
5771 num = (num*10) + (ch-'0');
5772 }
5773 } else {
5774 break;
5775 }
5776 cp++;
5777 }
5778 *cp_p = cp;
5779
5780 /* user says num divided by scale and we say internally that
5781 * is MAX_SCORE * num / scale.
5782 */
5783 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5784 }
5785
5786 struct diff_queue_struct diff_queued_diff;
5787
5788 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5789 {
5790 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5791 queue->queue[queue->nr++] = dp;
5792 }
5793
5794 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5795 struct diff_filespec *one,
5796 struct diff_filespec *two)
5797 {
5798 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5799 dp->one = one;
5800 dp->two = two;
5801 if (queue)
5802 diff_q(queue, dp);
5803 return dp;
5804 }
5805
5806 void diff_free_filepair(struct diff_filepair *p)
5807 {
5808 free_filespec(p->one);
5809 free_filespec(p->two);
5810 free(p);
5811 }
5812
5813 void diff_free_queue(struct diff_queue_struct *q)
5814 {
5815 for (int i = 0; i < q->nr; i++)
5816 diff_free_filepair(q->queue[i]);
5817 free(q->queue);
5818 }
5819
5820 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5821 {
5822 int abblen;
5823 const char *abbrev;
5824
5825 /* Do we want all 40 hex characters? */
5826 if (len == the_hash_algo->hexsz)
5827 return oid_to_hex(oid);
5828
5829 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5830 abbrev = diff_abbrev_oid(oid, len);
5831
5832 if (!print_sha1_ellipsis())
5833 return abbrev;
5834
5835 abblen = strlen(abbrev);
5836
5837 /*
5838 * In well-behaved cases, where the abbreviated result is the
5839 * same as the requested length, append three dots after the
5840 * abbreviation (hence the whole logic is limited to the case
5841 * where abblen < 37); when the actual abbreviated result is a
5842 * bit longer than the requested length, we reduce the number
5843 * of dots so that they match the well-behaved ones. However,
5844 * if the actual abbreviation is longer than the requested
5845 * length by more than three, we give up on aligning, and add
5846 * three dots anyway, to indicate that the output is not the
5847 * full object name. Yes, this may be suboptimal, but this
5848 * appears only in "diff --raw --abbrev" output and it is not
5849 * worth the effort to change it now. Note that this would
5850 * likely to work fine when the automatic sizing of default
5851 * abbreviation length is used--we would be fed -1 in "len" in
5852 * that case, and will end up always appending three-dots, but
5853 * the automatic sizing is supposed to give abblen that ensures
5854 * uniqueness across all objects (statistically speaking).
5855 */
5856 if (abblen < the_hash_algo->hexsz - 3) {
5857 static char hex[GIT_MAX_HEXSZ + 1];
5858 if (len < abblen && abblen <= len + 2)
5859 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5860 else
5861 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5862 return hex;
5863 }
5864
5865 return oid_to_hex(oid);
5866 }
5867
5868 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5869 {
5870 int line_termination = opt->line_termination;
5871 int inter_name_termination = line_termination ? '\t' : '\0';
5872
5873 fprintf(opt->file, "%s", diff_line_prefix(opt));
5874 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5875 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5876 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5877 fprintf(opt->file, "%s ",
5878 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5879 }
5880 if (p->score) {
5881 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5882 inter_name_termination);
5883 } else {
5884 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5885 }
5886
5887 if (p->status == DIFF_STATUS_COPIED ||
5888 p->status == DIFF_STATUS_RENAMED) {
5889 const char *name_a, *name_b;
5890 name_a = p->one->path;
5891 name_b = p->two->path;
5892 strip_prefix(opt->prefix_length, &name_a, &name_b);
5893 write_name_quoted(name_a, opt->file, inter_name_termination);
5894 write_name_quoted(name_b, opt->file, line_termination);
5895 } else {
5896 const char *name_a, *name_b;
5897 name_a = p->one->mode ? p->one->path : p->two->path;
5898 name_b = NULL;
5899 strip_prefix(opt->prefix_length, &name_a, &name_b);
5900 write_name_quoted(name_a, opt->file, line_termination);
5901 }
5902 }
5903
5904 int diff_unmodified_pair(struct diff_filepair *p)
5905 {
5906 /* This function is written stricter than necessary to support
5907 * the currently implemented transformers, but the idea is to
5908 * let transformers to produce diff_filepairs any way they want,
5909 * and filter and clean them up here before producing the output.
5910 */
5911 struct diff_filespec *one = p->one, *two = p->two;
5912
5913 if (DIFF_PAIR_UNMERGED(p))
5914 return 0; /* unmerged is interesting */
5915
5916 /* deletion, addition, mode or type change
5917 * and rename are all interesting.
5918 */
5919 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5920 DIFF_PAIR_MODE_CHANGED(p) ||
5921 strcmp(one->path, two->path))
5922 return 0;
5923
5924 /* both are valid and point at the same path. that is, we are
5925 * dealing with a change.
5926 */
5927 if (one->oid_valid && two->oid_valid &&
5928 oideq(&one->oid, &two->oid) &&
5929 !one->dirty_submodule && !two->dirty_submodule)
5930 return 1; /* no change */
5931 if (!one->oid_valid && !two->oid_valid)
5932 return 1; /* both look at the same file on the filesystem. */
5933 return 0;
5934 }
5935
5936 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5937 {
5938 int include_conflict_headers =
5939 (additional_headers(o, p->one->path) &&
5940 !o->pickaxe_opts &&
5941 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
5942
5943 /*
5944 * Check if we can return early without showing a diff. Note that
5945 * diff_filepair only stores {oid, path, mode, is_valid}
5946 * information for each path, and thus diff_unmodified_pair() only
5947 * considers those bits of info. However, we do not want pairs
5948 * created by create_filepairs_for_header_only_notifications()
5949 * (which always look like unmodified pairs) to be ignored, so
5950 * return early if both p is unmodified AND we don't want to
5951 * include_conflict_headers.
5952 */
5953 if (diff_unmodified_pair(p) && !include_conflict_headers)
5954 return;
5955
5956 /* Actually, we can also return early to avoid showing tree diffs */
5957 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5958 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5959 return;
5960
5961 run_diff(p, o);
5962 }
5963
5964 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5965 struct diffstat_t *diffstat)
5966 {
5967 if (diff_unmodified_pair(p))
5968 return;
5969
5970 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5971 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5972 return; /* no useful stat for tree diffs */
5973
5974 run_diffstat(p, o, diffstat);
5975 }
5976
5977 static void diff_flush_checkdiff(struct diff_filepair *p,
5978 struct diff_options *o)
5979 {
5980 if (diff_unmodified_pair(p))
5981 return;
5982
5983 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5984 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5985 return; /* nothing to check in tree diffs */
5986
5987 run_checkdiff(p, o);
5988 }
5989
5990 int diff_queue_is_empty(struct diff_options *o)
5991 {
5992 struct diff_queue_struct *q = &diff_queued_diff;
5993 int i;
5994 int include_conflict_headers =
5995 (o->additional_path_headers &&
5996 strmap_get_size(o->additional_path_headers) &&
5997 !o->pickaxe_opts &&
5998 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
5999
6000 if (include_conflict_headers)
6001 return 0;
6002
6003 for (i = 0; i < q->nr; i++)
6004 if (!diff_unmodified_pair(q->queue[i]))
6005 return 0;
6006 return 1;
6007 }
6008
6009 #if DIFF_DEBUG
6010 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6011 {
6012 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6013 x, one ? one : "",
6014 s->path,
6015 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6016 s->mode,
6017 s->oid_valid ? oid_to_hex(&s->oid) : "");
6018 fprintf(stderr, "queue[%d] %s size %lu\n",
6019 x, one ? one : "",
6020 s->size);
6021 }
6022
6023 void diff_debug_filepair(const struct diff_filepair *p, int i)
6024 {
6025 diff_debug_filespec(p->one, i, "one");
6026 diff_debug_filespec(p->two, i, "two");
6027 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6028 p->score, p->status ? p->status : '?',
6029 p->one->rename_used, p->broken_pair);
6030 }
6031
6032 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6033 {
6034 int i;
6035 if (msg)
6036 fprintf(stderr, "%s\n", msg);
6037 fprintf(stderr, "q->nr = %d\n", q->nr);
6038 for (i = 0; i < q->nr; i++) {
6039 struct diff_filepair *p = q->queue[i];
6040 diff_debug_filepair(p, i);
6041 }
6042 }
6043 #endif
6044
6045 static void diff_resolve_rename_copy(void)
6046 {
6047 int i;
6048 struct diff_filepair *p;
6049 struct diff_queue_struct *q = &diff_queued_diff;
6050
6051 diff_debug_queue("resolve-rename-copy", q);
6052
6053 for (i = 0; i < q->nr; i++) {
6054 p = q->queue[i];
6055 p->status = 0; /* undecided */
6056 if (DIFF_PAIR_UNMERGED(p))
6057 p->status = DIFF_STATUS_UNMERGED;
6058 else if (!DIFF_FILE_VALID(p->one))
6059 p->status = DIFF_STATUS_ADDED;
6060 else if (!DIFF_FILE_VALID(p->two))
6061 p->status = DIFF_STATUS_DELETED;
6062 else if (DIFF_PAIR_TYPE_CHANGED(p))
6063 p->status = DIFF_STATUS_TYPE_CHANGED;
6064
6065 /* from this point on, we are dealing with a pair
6066 * whose both sides are valid and of the same type, i.e.
6067 * either in-place edit or rename/copy edit.
6068 */
6069 else if (DIFF_PAIR_RENAME(p)) {
6070 /*
6071 * A rename might have re-connected a broken
6072 * pair up, causing the pathnames to be the
6073 * same again. If so, that's not a rename at
6074 * all, just a modification..
6075 *
6076 * Otherwise, see if this source was used for
6077 * multiple renames, in which case we decrement
6078 * the count, and call it a copy.
6079 */
6080 if (!strcmp(p->one->path, p->two->path))
6081 p->status = DIFF_STATUS_MODIFIED;
6082 else if (--p->one->rename_used > 0)
6083 p->status = DIFF_STATUS_COPIED;
6084 else
6085 p->status = DIFF_STATUS_RENAMED;
6086 }
6087 else if (!oideq(&p->one->oid, &p->two->oid) ||
6088 p->one->mode != p->two->mode ||
6089 p->one->dirty_submodule ||
6090 p->two->dirty_submodule ||
6091 is_null_oid(&p->one->oid))
6092 p->status = DIFF_STATUS_MODIFIED;
6093 else {
6094 /* This is a "no-change" entry and should not
6095 * happen anymore, but prepare for broken callers.
6096 */
6097 error("feeding unmodified %s to diffcore",
6098 p->one->path);
6099 p->status = DIFF_STATUS_UNKNOWN;
6100 }
6101 }
6102 diff_debug_queue("resolve-rename-copy done", q);
6103 }
6104
6105 static int check_pair_status(struct diff_filepair *p)
6106 {
6107 switch (p->status) {
6108 case DIFF_STATUS_UNKNOWN:
6109 return 0;
6110 case 0:
6111 die("internal error in diff-resolve-rename-copy");
6112 default:
6113 return 1;
6114 }
6115 }
6116
6117 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6118 {
6119 int fmt = opt->output_format;
6120
6121 if (fmt & DIFF_FORMAT_CHECKDIFF)
6122 diff_flush_checkdiff(p, opt);
6123 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6124 diff_flush_raw(p, opt);
6125 else if (fmt & DIFF_FORMAT_NAME) {
6126 const char *name_a, *name_b;
6127 name_a = p->two->path;
6128 name_b = NULL;
6129 strip_prefix(opt->prefix_length, &name_a, &name_b);
6130 fprintf(opt->file, "%s", diff_line_prefix(opt));
6131 write_name_quoted(name_a, opt->file, opt->line_termination);
6132 }
6133 }
6134
6135 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6136 {
6137 struct strbuf sb = STRBUF_INIT;
6138 if (fs->mode)
6139 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6140 else
6141 strbuf_addf(&sb, " %s ", newdelete);
6142
6143 quote_c_style(fs->path, &sb, NULL, 0);
6144 strbuf_addch(&sb, '\n');
6145 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6146 sb.buf, sb.len, 0);
6147 strbuf_release(&sb);
6148 }
6149
6150 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6151 int show_name)
6152 {
6153 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6154 struct strbuf sb = STRBUF_INIT;
6155 strbuf_addf(&sb, " mode change %06o => %06o",
6156 p->one->mode, p->two->mode);
6157 if (show_name) {
6158 strbuf_addch(&sb, ' ');
6159 quote_c_style(p->two->path, &sb, NULL, 0);
6160 }
6161 strbuf_addch(&sb, '\n');
6162 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6163 sb.buf, sb.len, 0);
6164 strbuf_release(&sb);
6165 }
6166 }
6167
6168 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6169 struct diff_filepair *p)
6170 {
6171 struct strbuf sb = STRBUF_INIT;
6172 struct strbuf names = STRBUF_INIT;
6173
6174 pprint_rename(&names, p->one->path, p->two->path);
6175 strbuf_addf(&sb, " %s %s (%d%%)\n",
6176 renamecopy, names.buf, similarity_index(p));
6177 strbuf_release(&names);
6178 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6179 sb.buf, sb.len, 0);
6180 show_mode_change(opt, p, 0);
6181 strbuf_release(&sb);
6182 }
6183
6184 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6185 {
6186 switch(p->status) {
6187 case DIFF_STATUS_DELETED:
6188 show_file_mode_name(opt, "delete", p->one);
6189 break;
6190 case DIFF_STATUS_ADDED:
6191 show_file_mode_name(opt, "create", p->two);
6192 break;
6193 case DIFF_STATUS_COPIED:
6194 show_rename_copy(opt, "copy", p);
6195 break;
6196 case DIFF_STATUS_RENAMED:
6197 show_rename_copy(opt, "rename", p);
6198 break;
6199 default:
6200 if (p->score) {
6201 struct strbuf sb = STRBUF_INIT;
6202 strbuf_addstr(&sb, " rewrite ");
6203 quote_c_style(p->two->path, &sb, NULL, 0);
6204 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6205 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6206 sb.buf, sb.len, 0);
6207 strbuf_release(&sb);
6208 }
6209 show_mode_change(opt, p, !p->score);
6210 break;
6211 }
6212 }
6213
6214 struct patch_id_t {
6215 git_hash_ctx *ctx;
6216 int patchlen;
6217 };
6218
6219 static int remove_space(char *line, int len)
6220 {
6221 int i;
6222 char *dst = line;
6223 unsigned char c;
6224
6225 for (i = 0; i < len; i++)
6226 if (!isspace((c = line[i])))
6227 *dst++ = c;
6228
6229 return dst - line;
6230 }
6231
6232 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6233 {
6234 unsigned char hash[GIT_MAX_RAWSZ];
6235 unsigned short carry = 0;
6236 int i;
6237
6238 the_hash_algo->final_fn(hash, ctx);
6239 the_hash_algo->init_fn(ctx);
6240 /* 20-byte sum, with carry */
6241 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6242 carry += result->hash[i] + hash[i];
6243 result->hash[i] = carry;
6244 carry >>= 8;
6245 }
6246 }
6247
6248 static int patch_id_consume(void *priv, char *line, unsigned long len)
6249 {
6250 struct patch_id_t *data = priv;
6251 int new_len;
6252
6253 if (len > 12 && starts_with(line, "\\ "))
6254 return 0;
6255 new_len = remove_space(line, len);
6256
6257 the_hash_algo->update_fn(data->ctx, line, new_len);
6258 data->patchlen += new_len;
6259 return 0;
6260 }
6261
6262 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6263 {
6264 the_hash_algo->update_fn(ctx, str, strlen(str));
6265 }
6266
6267 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6268 {
6269 /* large enough for 2^32 in octal */
6270 char buf[12];
6271 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6272 the_hash_algo->update_fn(ctx, buf, len);
6273 }
6274
6275 /* returns 0 upon success, and writes result into oid */
6276 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6277 {
6278 struct diff_queue_struct *q = &diff_queued_diff;
6279 int i;
6280 git_hash_ctx ctx;
6281 struct patch_id_t data;
6282
6283 the_hash_algo->init_fn(&ctx);
6284 memset(&data, 0, sizeof(struct patch_id_t));
6285 data.ctx = &ctx;
6286 oidclr(oid);
6287
6288 for (i = 0; i < q->nr; i++) {
6289 xpparam_t xpp;
6290 xdemitconf_t xecfg;
6291 mmfile_t mf1, mf2;
6292 struct diff_filepair *p = q->queue[i];
6293 int len1, len2;
6294
6295 memset(&xpp, 0, sizeof(xpp));
6296 memset(&xecfg, 0, sizeof(xecfg));
6297 if (p->status == 0)
6298 return error("internal diff status error");
6299 if (p->status == DIFF_STATUS_UNKNOWN)
6300 continue;
6301 if (diff_unmodified_pair(p))
6302 continue;
6303 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6304 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6305 continue;
6306 if (DIFF_PAIR_UNMERGED(p))
6307 continue;
6308
6309 diff_fill_oid_info(p->one, options->repo->index);
6310 diff_fill_oid_info(p->two, options->repo->index);
6311
6312 len1 = remove_space(p->one->path, strlen(p->one->path));
6313 len2 = remove_space(p->two->path, strlen(p->two->path));
6314 patch_id_add_string(&ctx, "diff--git");
6315 patch_id_add_string(&ctx, "a/");
6316 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6317 patch_id_add_string(&ctx, "b/");
6318 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6319
6320 if (p->one->mode == 0) {
6321 patch_id_add_string(&ctx, "newfilemode");
6322 patch_id_add_mode(&ctx, p->two->mode);
6323 } else if (p->two->mode == 0) {
6324 patch_id_add_string(&ctx, "deletedfilemode");
6325 patch_id_add_mode(&ctx, p->one->mode);
6326 } else if (p->one->mode != p->two->mode) {
6327 patch_id_add_string(&ctx, "oldmode");
6328 patch_id_add_mode(&ctx, p->one->mode);
6329 patch_id_add_string(&ctx, "newmode");
6330 patch_id_add_mode(&ctx, p->two->mode);
6331 }
6332
6333 if (diff_header_only) {
6334 /* don't do anything since we're only populating header info */
6335 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6336 diff_filespec_is_binary(options->repo, p->two)) {
6337 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6338 the_hash_algo->hexsz);
6339 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6340 the_hash_algo->hexsz);
6341 } else {
6342 if (p->one->mode == 0) {
6343 patch_id_add_string(&ctx, "---/dev/null");
6344 patch_id_add_string(&ctx, "+++b/");
6345 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6346 } else if (p->two->mode == 0) {
6347 patch_id_add_string(&ctx, "---a/");
6348 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6349 patch_id_add_string(&ctx, "+++/dev/null");
6350 } else {
6351 patch_id_add_string(&ctx, "---a/");
6352 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6353 patch_id_add_string(&ctx, "+++b/");
6354 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6355 }
6356
6357 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6358 fill_mmfile(options->repo, &mf2, p->two) < 0)
6359 return error("unable to read files to diff");
6360 xpp.flags = 0;
6361 xecfg.ctxlen = 3;
6362 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6363 if (xdi_diff_outf(&mf1, &mf2, NULL,
6364 patch_id_consume, &data, &xpp, &xecfg))
6365 return error("unable to generate patch-id diff for %s",
6366 p->one->path);
6367 }
6368 flush_one_hunk(oid, &ctx);
6369 }
6370
6371 return 0;
6372 }
6373
6374 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6375 {
6376 struct diff_queue_struct *q = &diff_queued_diff;
6377 int result = diff_get_patch_id(options, oid, diff_header_only);
6378
6379 diff_free_queue(q);
6380 DIFF_QUEUE_CLEAR(q);
6381
6382 return result;
6383 }
6384
6385 static int is_summary_empty(const struct diff_queue_struct *q)
6386 {
6387 int i;
6388
6389 for (i = 0; i < q->nr; i++) {
6390 const struct diff_filepair *p = q->queue[i];
6391
6392 switch (p->status) {
6393 case DIFF_STATUS_DELETED:
6394 case DIFF_STATUS_ADDED:
6395 case DIFF_STATUS_COPIED:
6396 case DIFF_STATUS_RENAMED:
6397 return 0;
6398 default:
6399 if (p->score)
6400 return 0;
6401 if (p->one->mode && p->two->mode &&
6402 p->one->mode != p->two->mode)
6403 return 0;
6404 break;
6405 }
6406 }
6407 return 1;
6408 }
6409
6410 static const char rename_limit_warning[] =
6411 N_("exhaustive rename detection was skipped due to too many files.");
6412
6413 static const char degrade_cc_to_c_warning[] =
6414 N_("only found copies from modified paths due to too many files.");
6415
6416 static const char rename_limit_advice[] =
6417 N_("you may want to set your %s variable to at least "
6418 "%d and retry the command.");
6419
6420 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6421 {
6422 fflush(stdout);
6423 if (degraded_cc)
6424 warning(_(degrade_cc_to_c_warning));
6425 else if (needed)
6426 warning(_(rename_limit_warning));
6427 else
6428 return;
6429 if (0 < needed)
6430 warning(_(rename_limit_advice), varname, needed);
6431 }
6432
6433 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6434 {
6435 struct strset present;
6436 struct diff_queue_struct *q = &diff_queued_diff;
6437 struct hashmap_iter iter;
6438 struct strmap_entry *e;
6439 int i;
6440
6441 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6442
6443 /*
6444 * Find out which paths exist in diff_queued_diff, preferring
6445 * one->path for any pair that has multiple paths.
6446 */
6447 for (i = 0; i < q->nr; i++) {
6448 struct diff_filepair *p = q->queue[i];
6449 char *path = p->one->path ? p->one->path : p->two->path;
6450
6451 if (strmap_contains(o->additional_path_headers, path))
6452 strset_add(&present, path);
6453 }
6454
6455 /*
6456 * Loop over paths in additional_path_headers; for each NOT already
6457 * in diff_queued_diff, create a synthetic filepair and insert that
6458 * into diff_queued_diff.
6459 */
6460 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6461 if (!strset_contains(&present, e->key)) {
6462 struct diff_filespec *one, *two;
6463 struct diff_filepair *p;
6464
6465 one = alloc_filespec(e->key);
6466 two = alloc_filespec(e->key);
6467 fill_filespec(one, null_oid(), 0, 0);
6468 fill_filespec(two, null_oid(), 0, 0);
6469 p = diff_queue(q, one, two);
6470 p->status = DIFF_STATUS_MODIFIED;
6471 }
6472 }
6473
6474 /* Re-sort the filepairs */
6475 diffcore_fix_diff_index();
6476
6477 /* Cleanup */
6478 strset_clear(&present);
6479 }
6480
6481 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6482 {
6483 int i;
6484 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6485 struct diff_queue_struct *q = &diff_queued_diff;
6486
6487 if (WSEH_NEW & WS_RULE_MASK)
6488 BUG("WS rules bit mask overlaps with diff symbol flags");
6489
6490 if (o->color_moved)
6491 o->emitted_symbols = &esm;
6492
6493 if (o->additional_path_headers)
6494 create_filepairs_for_header_only_notifications(o);
6495
6496 for (i = 0; i < q->nr; i++) {
6497 struct diff_filepair *p = q->queue[i];
6498 if (check_pair_status(p))
6499 diff_flush_patch(p, o);
6500 }
6501
6502 if (o->emitted_symbols) {
6503 if (o->color_moved) {
6504 struct mem_pool entry_pool;
6505 struct moved_entry_list *entry_list;
6506
6507 mem_pool_init(&entry_pool, 1024 * 1024);
6508 entry_list = add_lines_to_move_detection(o,
6509 &entry_pool);
6510 mark_color_as_moved(o, entry_list);
6511 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6512 dim_moved_lines(o);
6513
6514 mem_pool_discard(&entry_pool, 0);
6515 free(entry_list);
6516 }
6517
6518 for (i = 0; i < esm.nr; i++)
6519 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6520
6521 for (i = 0; i < esm.nr; i++)
6522 free((void *)esm.buf[i].line);
6523 esm.nr = 0;
6524
6525 o->emitted_symbols = NULL;
6526 }
6527 }
6528
6529 static void diff_free_file(struct diff_options *options)
6530 {
6531 if (options->close_file)
6532 fclose(options->file);
6533 }
6534
6535 static void diff_free_ignore_regex(struct diff_options *options)
6536 {
6537 int i;
6538
6539 for (i = 0; i < options->ignore_regex_nr; i++) {
6540 regfree(options->ignore_regex[i]);
6541 free(options->ignore_regex[i]);
6542 }
6543 free(options->ignore_regex);
6544 }
6545
6546 void diff_free(struct diff_options *options)
6547 {
6548 if (options->no_free)
6549 return;
6550
6551 diff_free_file(options);
6552 diff_free_ignore_regex(options);
6553 clear_pathspec(&options->pathspec);
6554 }
6555
6556 void diff_flush(struct diff_options *options)
6557 {
6558 struct diff_queue_struct *q = &diff_queued_diff;
6559 int i, output_format = options->output_format;
6560 int separator = 0;
6561 int dirstat_by_line = 0;
6562
6563 /*
6564 * Order: raw, stat, summary, patch
6565 * or: name/name-status/checkdiff (other bits clear)
6566 */
6567 if (!q->nr && !options->additional_path_headers)
6568 goto free_queue;
6569
6570 if (output_format & (DIFF_FORMAT_RAW |
6571 DIFF_FORMAT_NAME |
6572 DIFF_FORMAT_NAME_STATUS |
6573 DIFF_FORMAT_CHECKDIFF)) {
6574 for (i = 0; i < q->nr; i++) {
6575 struct diff_filepair *p = q->queue[i];
6576 if (check_pair_status(p))
6577 flush_one_pair(p, options);
6578 }
6579 separator++;
6580 }
6581
6582 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6583 dirstat_by_line = 1;
6584
6585 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6586 dirstat_by_line) {
6587 struct diffstat_t diffstat;
6588
6589 compute_diffstat(options, &diffstat, q);
6590 if (output_format & DIFF_FORMAT_NUMSTAT)
6591 show_numstat(&diffstat, options);
6592 if (output_format & DIFF_FORMAT_DIFFSTAT)
6593 show_stats(&diffstat, options);
6594 if (output_format & DIFF_FORMAT_SHORTSTAT)
6595 show_shortstats(&diffstat, options);
6596 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6597 show_dirstat_by_line(&diffstat, options);
6598 free_diffstat_info(&diffstat);
6599 separator++;
6600 }
6601 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6602 show_dirstat(options);
6603
6604 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6605 for (i = 0; i < q->nr; i++) {
6606 diff_summary(options, q->queue[i]);
6607 }
6608 separator++;
6609 }
6610
6611 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6612 options->flags.exit_with_status &&
6613 options->flags.diff_from_contents) {
6614 /*
6615 * run diff_flush_patch for the exit status. setting
6616 * options->file to /dev/null should be safe, because we
6617 * aren't supposed to produce any output anyway.
6618 */
6619 diff_free_file(options);
6620 options->file = xfopen("/dev/null", "w");
6621 options->close_file = 1;
6622 options->color_moved = 0;
6623 for (i = 0; i < q->nr; i++) {
6624 struct diff_filepair *p = q->queue[i];
6625 if (check_pair_status(p))
6626 diff_flush_patch(p, options);
6627 if (options->found_changes)
6628 break;
6629 }
6630 }
6631
6632 if (output_format & DIFF_FORMAT_PATCH) {
6633 if (separator) {
6634 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6635 if (options->stat_sep)
6636 /* attach patch instead of inline */
6637 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6638 NULL, 0, 0);
6639 }
6640
6641 diff_flush_patch_all_file_pairs(options);
6642 }
6643
6644 if (output_format & DIFF_FORMAT_CALLBACK)
6645 options->format_callback(q, options, options->format_callback_data);
6646
6647 free_queue:
6648 diff_free_queue(q);
6649 DIFF_QUEUE_CLEAR(q);
6650 diff_free(options);
6651
6652 /*
6653 * Report the content-level differences with HAS_CHANGES;
6654 * diff_addremove/diff_change does not set the bit when
6655 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6656 */
6657 if (options->flags.diff_from_contents) {
6658 if (options->found_changes)
6659 options->flags.has_changes = 1;
6660 else
6661 options->flags.has_changes = 0;
6662 }
6663 }
6664
6665 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6666 {
6667 return (((p->status == DIFF_STATUS_MODIFIED) &&
6668 ((p->score &&
6669 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6670 (!p->score &&
6671 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6672 ((p->status != DIFF_STATUS_MODIFIED) &&
6673 filter_bit_tst(p->status, options)));
6674 }
6675
6676 static void diffcore_apply_filter(struct diff_options *options)
6677 {
6678 int i;
6679 struct diff_queue_struct *q = &diff_queued_diff;
6680 struct diff_queue_struct outq;
6681
6682 DIFF_QUEUE_CLEAR(&outq);
6683
6684 if (!options->filter)
6685 return;
6686
6687 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6688 int found;
6689 for (i = found = 0; !found && i < q->nr; i++) {
6690 if (match_filter(options, q->queue[i]))
6691 found++;
6692 }
6693 if (found)
6694 return;
6695
6696 /* otherwise we will clear the whole queue
6697 * by copying the empty outq at the end of this
6698 * function, but first clear the current entries
6699 * in the queue.
6700 */
6701 for (i = 0; i < q->nr; i++)
6702 diff_free_filepair(q->queue[i]);
6703 }
6704 else {
6705 /* Only the matching ones */
6706 for (i = 0; i < q->nr; i++) {
6707 struct diff_filepair *p = q->queue[i];
6708 if (match_filter(options, p))
6709 diff_q(&outq, p);
6710 else
6711 diff_free_filepair(p);
6712 }
6713 }
6714 free(q->queue);
6715 *q = outq;
6716 }
6717
6718 /* Check whether two filespecs with the same mode and size are identical */
6719 static int diff_filespec_is_identical(struct repository *r,
6720 struct diff_filespec *one,
6721 struct diff_filespec *two)
6722 {
6723 if (S_ISGITLINK(one->mode))
6724 return 0;
6725 if (diff_populate_filespec(r, one, NULL))
6726 return 0;
6727 if (diff_populate_filespec(r, two, NULL))
6728 return 0;
6729 return !memcmp(one->data, two->data, one->size);
6730 }
6731
6732 static int diff_filespec_check_stat_unmatch(struct repository *r,
6733 struct diff_filepair *p)
6734 {
6735 struct diff_populate_filespec_options dpf_options = {
6736 .check_size_only = 1,
6737 .missing_object_cb = diff_queued_diff_prefetch,
6738 .missing_object_data = r,
6739 };
6740
6741 if (p->done_skip_stat_unmatch)
6742 return p->skip_stat_unmatch_result;
6743
6744 p->done_skip_stat_unmatch = 1;
6745 p->skip_stat_unmatch_result = 0;
6746 /*
6747 * 1. Entries that come from stat info dirtiness
6748 * always have both sides (iow, not create/delete),
6749 * one side of the object name is unknown, with
6750 * the same mode and size. Keep the ones that
6751 * do not match these criteria. They have real
6752 * differences.
6753 *
6754 * 2. At this point, the file is known to be modified,
6755 * with the same mode and size, and the object
6756 * name of one side is unknown. Need to inspect
6757 * the identical contents.
6758 */
6759 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6760 !DIFF_FILE_VALID(p->two) ||
6761 (p->one->oid_valid && p->two->oid_valid) ||
6762 (p->one->mode != p->two->mode) ||
6763 diff_populate_filespec(r, p->one, &dpf_options) ||
6764 diff_populate_filespec(r, p->two, &dpf_options) ||
6765 (p->one->size != p->two->size) ||
6766 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6767 p->skip_stat_unmatch_result = 1;
6768 return p->skip_stat_unmatch_result;
6769 }
6770
6771 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6772 {
6773 int i;
6774 struct diff_queue_struct *q = &diff_queued_diff;
6775 struct diff_queue_struct outq;
6776 DIFF_QUEUE_CLEAR(&outq);
6777
6778 for (i = 0; i < q->nr; i++) {
6779 struct diff_filepair *p = q->queue[i];
6780
6781 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6782 diff_q(&outq, p);
6783 else {
6784 /*
6785 * The caller can subtract 1 from skip_stat_unmatch
6786 * to determine how many paths were dirty only
6787 * due to stat info mismatch.
6788 */
6789 if (!diffopt->flags.no_index)
6790 diffopt->skip_stat_unmatch++;
6791 diff_free_filepair(p);
6792 }
6793 }
6794 free(q->queue);
6795 *q = outq;
6796 }
6797
6798 static int diffnamecmp(const void *a_, const void *b_)
6799 {
6800 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6801 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6802 const char *name_a, *name_b;
6803
6804 name_a = a->one ? a->one->path : a->two->path;
6805 name_b = b->one ? b->one->path : b->two->path;
6806 return strcmp(name_a, name_b);
6807 }
6808
6809 void diffcore_fix_diff_index(void)
6810 {
6811 struct diff_queue_struct *q = &diff_queued_diff;
6812 QSORT(q->queue, q->nr, diffnamecmp);
6813 }
6814
6815 void diff_add_if_missing(struct repository *r,
6816 struct oid_array *to_fetch,
6817 const struct diff_filespec *filespec)
6818 {
6819 if (filespec && filespec->oid_valid &&
6820 !S_ISGITLINK(filespec->mode) &&
6821 oid_object_info_extended(r, &filespec->oid, NULL,
6822 OBJECT_INFO_FOR_PREFETCH))
6823 oid_array_append(to_fetch, &filespec->oid);
6824 }
6825
6826 void diff_queued_diff_prefetch(void *repository)
6827 {
6828 struct repository *repo = repository;
6829 int i;
6830 struct diff_queue_struct *q = &diff_queued_diff;
6831 struct oid_array to_fetch = OID_ARRAY_INIT;
6832
6833 for (i = 0; i < q->nr; i++) {
6834 struct diff_filepair *p = q->queue[i];
6835 diff_add_if_missing(repo, &to_fetch, p->one);
6836 diff_add_if_missing(repo, &to_fetch, p->two);
6837 }
6838
6839 /*
6840 * NEEDSWORK: Consider deduplicating the OIDs sent.
6841 */
6842 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6843
6844 oid_array_clear(&to_fetch);
6845 }
6846
6847 void diffcore_std(struct diff_options *options)
6848 {
6849 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6850 DIFF_FORMAT_NUMSTAT |
6851 DIFF_FORMAT_PATCH |
6852 DIFF_FORMAT_SHORTSTAT |
6853 DIFF_FORMAT_DIRSTAT;
6854
6855 /*
6856 * Check if the user requested a blob-data-requiring diff output and/or
6857 * break-rewrite detection (which requires blob data). If yes, prefetch
6858 * the diff pairs.
6859 *
6860 * If no prefetching occurs, diffcore_rename() will prefetch if it
6861 * decides that it needs inexact rename detection.
6862 */
6863 if (options->repo == the_repository && has_promisor_remote() &&
6864 (options->output_format & output_formats_to_prefetch ||
6865 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6866 diff_queued_diff_prefetch(options->repo);
6867
6868 /* NOTE please keep the following in sync with diff_tree_combined() */
6869 if (options->skip_stat_unmatch)
6870 diffcore_skip_stat_unmatch(options);
6871 if (!options->found_follow) {
6872 /* See try_to_follow_renames() in tree-diff.c */
6873 if (options->break_opt != -1)
6874 diffcore_break(options->repo,
6875 options->break_opt);
6876 if (options->detect_rename)
6877 diffcore_rename(options);
6878 if (options->break_opt != -1)
6879 diffcore_merge_broken();
6880 }
6881 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6882 diffcore_pickaxe(options);
6883 if (options->orderfile)
6884 diffcore_order(options->orderfile);
6885 if (options->rotate_to)
6886 diffcore_rotate(options);
6887 if (!options->found_follow)
6888 /* See try_to_follow_renames() in tree-diff.c */
6889 diff_resolve_rename_copy();
6890 diffcore_apply_filter(options);
6891
6892 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6893 options->flags.has_changes = 1;
6894 else
6895 options->flags.has_changes = 0;
6896
6897 options->found_follow = 0;
6898 }
6899
6900 int diff_result_code(struct diff_options *opt, int status)
6901 {
6902 int result = 0;
6903
6904 diff_warn_rename_limit("diff.renameLimit",
6905 opt->needed_rename_limit,
6906 opt->degraded_cc_to_c);
6907 if (!opt->flags.exit_with_status &&
6908 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6909 return status;
6910 if (opt->flags.exit_with_status &&
6911 opt->flags.has_changes)
6912 result |= 01;
6913 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6914 opt->flags.check_failed)
6915 result |= 02;
6916 return result;
6917 }
6918
6919 int diff_can_quit_early(struct diff_options *opt)
6920 {
6921 return (opt->flags.quick &&
6922 !opt->filter &&
6923 opt->flags.has_changes);
6924 }
6925
6926 /*
6927 * Shall changes to this submodule be ignored?
6928 *
6929 * Submodule changes can be configured to be ignored separately for each path,
6930 * but that configuration can be overridden from the command line.
6931 */
6932 static int is_submodule_ignored(const char *path, struct diff_options *options)
6933 {
6934 int ignored = 0;
6935 struct diff_flags orig_flags = options->flags;
6936 if (!options->flags.override_submodule_config)
6937 set_diffopt_flags_from_submodule_config(options, path);
6938 if (options->flags.ignore_submodules)
6939 ignored = 1;
6940 options->flags = orig_flags;
6941 return ignored;
6942 }
6943
6944 void compute_diffstat(struct diff_options *options,
6945 struct diffstat_t *diffstat,
6946 struct diff_queue_struct *q)
6947 {
6948 int i;
6949
6950 memset(diffstat, 0, sizeof(struct diffstat_t));
6951 for (i = 0; i < q->nr; i++) {
6952 struct diff_filepair *p = q->queue[i];
6953 if (check_pair_status(p))
6954 diff_flush_stat(p, options, diffstat);
6955 }
6956 }
6957
6958 void diff_addremove(struct diff_options *options,
6959 int addremove, unsigned mode,
6960 const struct object_id *oid,
6961 int oid_valid,
6962 const char *concatpath, unsigned dirty_submodule)
6963 {
6964 struct diff_filespec *one, *two;
6965
6966 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6967 return;
6968
6969 /* This may look odd, but it is a preparation for
6970 * feeding "there are unchanged files which should
6971 * not produce diffs, but when you are doing copy
6972 * detection you would need them, so here they are"
6973 * entries to the diff-core. They will be prefixed
6974 * with something like '=' or '*' (I haven't decided
6975 * which but should not make any difference).
6976 * Feeding the same new and old to diff_change()
6977 * also has the same effect.
6978 * Before the final output happens, they are pruned after
6979 * merged into rename/copy pairs as appropriate.
6980 */
6981 if (options->flags.reverse_diff)
6982 addremove = (addremove == '+' ? '-' :
6983 addremove == '-' ? '+' : addremove);
6984
6985 if (options->prefix &&
6986 strncmp(concatpath, options->prefix, options->prefix_length))
6987 return;
6988
6989 one = alloc_filespec(concatpath);
6990 two = alloc_filespec(concatpath);
6991
6992 if (addremove != '+')
6993 fill_filespec(one, oid, oid_valid, mode);
6994 if (addremove != '-') {
6995 fill_filespec(two, oid, oid_valid, mode);
6996 two->dirty_submodule = dirty_submodule;
6997 }
6998
6999 diff_queue(&diff_queued_diff, one, two);
7000 if (!options->flags.diff_from_contents)
7001 options->flags.has_changes = 1;
7002 }
7003
7004 void diff_change(struct diff_options *options,
7005 unsigned old_mode, unsigned new_mode,
7006 const struct object_id *old_oid,
7007 const struct object_id *new_oid,
7008 int old_oid_valid, int new_oid_valid,
7009 const char *concatpath,
7010 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7011 {
7012 struct diff_filespec *one, *two;
7013 struct diff_filepair *p;
7014
7015 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7016 is_submodule_ignored(concatpath, options))
7017 return;
7018
7019 if (options->flags.reverse_diff) {
7020 SWAP(old_mode, new_mode);
7021 SWAP(old_oid, new_oid);
7022 SWAP(old_oid_valid, new_oid_valid);
7023 SWAP(old_dirty_submodule, new_dirty_submodule);
7024 }
7025
7026 if (options->prefix &&
7027 strncmp(concatpath, options->prefix, options->prefix_length))
7028 return;
7029
7030 one = alloc_filespec(concatpath);
7031 two = alloc_filespec(concatpath);
7032 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7033 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7034 one->dirty_submodule = old_dirty_submodule;
7035 two->dirty_submodule = new_dirty_submodule;
7036 p = diff_queue(&diff_queued_diff, one, two);
7037
7038 if (options->flags.diff_from_contents)
7039 return;
7040
7041 if (options->flags.quick && options->skip_stat_unmatch &&
7042 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7043 diff_free_filespec_data(p->one);
7044 diff_free_filespec_data(p->two);
7045 return;
7046 }
7047
7048 options->flags.has_changes = 1;
7049 }
7050
7051 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7052 {
7053 struct diff_filepair *pair;
7054 struct diff_filespec *one, *two;
7055
7056 if (options->prefix &&
7057 strncmp(path, options->prefix, options->prefix_length))
7058 return NULL;
7059
7060 one = alloc_filespec(path);
7061 two = alloc_filespec(path);
7062 pair = diff_queue(&diff_queued_diff, one, two);
7063 pair->is_unmerged = 1;
7064 return pair;
7065 }
7066
7067 static char *run_textconv(struct repository *r,
7068 const char *pgm,
7069 struct diff_filespec *spec,
7070 size_t *outsize)
7071 {
7072 struct diff_tempfile *temp;
7073 struct child_process child = CHILD_PROCESS_INIT;
7074 struct strbuf buf = STRBUF_INIT;
7075 int err = 0;
7076
7077 temp = prepare_temp_file(r, spec);
7078 strvec_push(&child.args, pgm);
7079 strvec_push(&child.args, temp->name);
7080
7081 child.use_shell = 1;
7082 child.out = -1;
7083 if (start_command(&child)) {
7084 remove_tempfile();
7085 return NULL;
7086 }
7087
7088 if (strbuf_read(&buf, child.out, 0) < 0)
7089 err = error("error reading from textconv command '%s'", pgm);
7090 close(child.out);
7091
7092 if (finish_command(&child) || err) {
7093 strbuf_release(&buf);
7094 remove_tempfile();
7095 return NULL;
7096 }
7097 remove_tempfile();
7098
7099 return strbuf_detach(&buf, outsize);
7100 }
7101
7102 size_t fill_textconv(struct repository *r,
7103 struct userdiff_driver *driver,
7104 struct diff_filespec *df,
7105 char **outbuf)
7106 {
7107 size_t size;
7108
7109 if (!driver) {
7110 if (!DIFF_FILE_VALID(df)) {
7111 *outbuf = "";
7112 return 0;
7113 }
7114 if (diff_populate_filespec(r, df, NULL))
7115 die("unable to read files to diff");
7116 *outbuf = df->data;
7117 return df->size;
7118 }
7119
7120 if (!driver->textconv)
7121 BUG("fill_textconv called with non-textconv driver");
7122
7123 if (driver->textconv_cache && df->oid_valid) {
7124 *outbuf = notes_cache_get(driver->textconv_cache,
7125 &df->oid,
7126 &size);
7127 if (*outbuf)
7128 return size;
7129 }
7130
7131 *outbuf = run_textconv(r, driver->textconv, df, &size);
7132 if (!*outbuf)
7133 die("unable to read files to diff");
7134
7135 if (driver->textconv_cache && df->oid_valid) {
7136 /* ignore errors, as we might be in a readonly repository */
7137 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7138 size);
7139 /*
7140 * we could save up changes and flush them all at the end,
7141 * but we would need an extra call after all diffing is done.
7142 * Since generating a cache entry is the slow path anyway,
7143 * this extra overhead probably isn't a big deal.
7144 */
7145 notes_cache_write(driver->textconv_cache);
7146 }
7147
7148 return size;
7149 }
7150
7151 int textconv_object(struct repository *r,
7152 const char *path,
7153 unsigned mode,
7154 const struct object_id *oid,
7155 int oid_valid,
7156 char **buf,
7157 unsigned long *buf_size)
7158 {
7159 struct diff_filespec *df;
7160 struct userdiff_driver *textconv;
7161
7162 df = alloc_filespec(path);
7163 fill_filespec(df, oid, oid_valid, mode);
7164 textconv = get_textconv(r, df);
7165 if (!textconv) {
7166 free_filespec(df);
7167 return 0;
7168 }
7169
7170 *buf_size = fill_textconv(r, textconv, df, buf);
7171 free_filespec(df);
7172 return 1;
7173 }
7174
7175 void setup_diff_pager(struct diff_options *opt)
7176 {
7177 /*
7178 * If the user asked for our exit code, then either they want --quiet
7179 * or --exit-code. We should definitely not bother with a pager in the
7180 * former case, as we will generate no output. Since we still properly
7181 * report our exit code even when a pager is run, we _could_ run a
7182 * pager with --exit-code. But since we have not done so historically,
7183 * and because it is easy to find people oneline advising "git diff
7184 * --exit-code" in hooks and other scripts, we do not do so.
7185 */
7186 if (!opt->flags.exit_with_status &&
7187 check_pager_config("diff") != 0)
7188 setup_pager();
7189 }
The Best VCS Around