?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
read-cache API & users: make discard_index() return void
[git.git] / diff.c
blob9f9a92ec9d29701e601362a4b3f2d8d2ed35d394
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, unsigned ws_rule)
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, ws_rule))
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 unsigned ws_rule = ecbdata->ws_rule;
638 l1 = count_trailing_blank(mf1, ws_rule);
639 l2 = count_trailing_blank(mf2, ws_rule);
640 if (l2 <= l1) {
641 ecbdata->blank_at_eof_in_preimage = 0;
642 ecbdata->blank_at_eof_in_postimage = 0;
643 return;
644 }
645 at = count_lines(mf1->ptr, mf1->size);
646 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
647
648 at = count_lines(mf2->ptr, mf2->size);
649 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
650 }
651
652 static void emit_line_0(struct diff_options *o,
653 const char *set_sign, const char *set, unsigned reverse, const char *reset,
654 int first, const char *line, int len)
655 {
656 int has_trailing_newline, has_trailing_carriage_return;
657 int needs_reset = 0; /* at the end of the line */
658 FILE *file = o->file;
659
660 fputs(diff_line_prefix(o), file);
661
662 has_trailing_newline = (len > 0 && line[len-1] == '\n');
663 if (has_trailing_newline)
664 len--;
665
666 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
667 if (has_trailing_carriage_return)
668 len--;
669
670 if (!len && !first)
671 goto end_of_line;
672
673 if (reverse && want_color(o->use_color)) {
674 fputs(GIT_COLOR_REVERSE, file);
675 needs_reset = 1;
676 }
677
678 if (set_sign) {
679 fputs(set_sign, file);
680 needs_reset = 1;
681 }
682
683 if (first)
684 fputc(first, file);
685
686 if (!len)
687 goto end_of_line;
688
689 if (set) {
690 if (set_sign && set != set_sign)
691 fputs(reset, file);
692 fputs(set, file);
693 needs_reset = 1;
694 }
695 fwrite(line, len, 1, file);
696 needs_reset = 1; /* 'line' may contain color codes. */
697
698 end_of_line:
699 if (needs_reset)
700 fputs(reset, file);
701 if (has_trailing_carriage_return)
702 fputc('\r', file);
703 if (has_trailing_newline)
704 fputc('\n', file);
705 }
706
707 static void emit_line(struct diff_options *o, const char *set, const char *reset,
708 const char *line, int len)
709 {
710 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
711 }
712
713 enum diff_symbol {
714 DIFF_SYMBOL_BINARY_DIFF_HEADER,
715 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
716 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
717 DIFF_SYMBOL_BINARY_DIFF_BODY,
718 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
719 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
720 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
721 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
722 DIFF_SYMBOL_STATS_LINE,
723 DIFF_SYMBOL_WORD_DIFF,
724 DIFF_SYMBOL_STAT_SEP,
725 DIFF_SYMBOL_SUMMARY,
726 DIFF_SYMBOL_SUBMODULE_ADD,
727 DIFF_SYMBOL_SUBMODULE_DEL,
728 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
729 DIFF_SYMBOL_SUBMODULE_MODIFIED,
730 DIFF_SYMBOL_SUBMODULE_HEADER,
731 DIFF_SYMBOL_SUBMODULE_ERROR,
732 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
733 DIFF_SYMBOL_REWRITE_DIFF,
734 DIFF_SYMBOL_BINARY_FILES,
735 DIFF_SYMBOL_HEADER,
736 DIFF_SYMBOL_FILEPAIR_PLUS,
737 DIFF_SYMBOL_FILEPAIR_MINUS,
738 DIFF_SYMBOL_WORDS_PORCELAIN,
739 DIFF_SYMBOL_WORDS,
740 DIFF_SYMBOL_CONTEXT,
741 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
742 DIFF_SYMBOL_PLUS,
743 DIFF_SYMBOL_MINUS,
744 DIFF_SYMBOL_NO_LF_EOF,
745 DIFF_SYMBOL_CONTEXT_FRAGINFO,
746 DIFF_SYMBOL_CONTEXT_MARKER,
747 DIFF_SYMBOL_SEPARATOR
748 };
749 /*
750 * Flags for content lines:
751 * 0..12 are whitespace rules
752 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
753 * 16 is marking if the line is blank at EOF
754 */
755 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
756 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
757 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
758 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
759 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
760
761 /*
762 * This struct is used when we need to buffer the output of the diff output.
763 *
764 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
765 * into the pre/post image file. This pointer could be a union with the
766 * line pointer. By storing an offset into the file instead of the literal line,
767 * we can decrease the memory footprint for the buffered output. At first we
768 * may want to only have indirection for the content lines, but we could also
769 * enhance the state for emitting prefabricated lines, e.g. the similarity
770 * score line or hunk/file headers would only need to store a number or path
771 * and then the output can be constructed later on depending on state.
772 */
773 struct emitted_diff_symbol {
774 const char *line;
775 int len;
776 int flags;
777 int indent_off; /* Offset to first non-whitespace character */
778 int indent_width; /* The visual width of the indentation */
779 unsigned id;
780 enum diff_symbol s;
781 };
782 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
783
784 struct emitted_diff_symbols {
785 struct emitted_diff_symbol *buf;
786 int nr, alloc;
787 };
788 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
789
790 static void append_emitted_diff_symbol(struct diff_options *o,
791 struct emitted_diff_symbol *e)
792 {
793 struct emitted_diff_symbol *f;
794
795 ALLOC_GROW(o->emitted_symbols->buf,
796 o->emitted_symbols->nr + 1,
797 o->emitted_symbols->alloc);
798 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
799
800 memcpy(f, e, sizeof(struct emitted_diff_symbol));
801 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
802 }
803
804 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
805 {
806 if (!e)
807 return;
808 free(e->buf);
809 free(e);
810 }
811
812 struct moved_entry {
813 const struct emitted_diff_symbol *es;
814 struct moved_entry *next_line;
815 struct moved_entry *next_match;
816 };
817
818 struct moved_block {
819 struct moved_entry *match;
820 int wsd; /* The whitespace delta of this block */
821 };
822
823 #define INDENT_BLANKLINE INT_MIN
824
825 static void fill_es_indent_data(struct emitted_diff_symbol *es)
826 {
827 unsigned int off = 0, i;
828 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
829 const char *s = es->line;
830 const int len = es->len;
831
832 /* skip any \v \f \r at start of indentation */
833 while (s[off] == '\f' || s[off] == '\v' ||
834 (s[off] == '\r' && off < len - 1))
835 off++;
836
837 /* calculate the visual width of indentation */
838 while(1) {
839 if (s[off] == ' ') {
840 width++;
841 off++;
842 } else if (s[off] == '\t') {
843 width += tab_width - (width % tab_width);
844 while (s[++off] == '\t')
845 width += tab_width;
846 } else {
847 break;
848 }
849 }
850
851 /* check if this line is blank */
852 for (i = off; i < len; i++)
853 if (!isspace(s[i]))
854 break;
855
856 if (i == len) {
857 es->indent_width = INDENT_BLANKLINE;
858 es->indent_off = len;
859 } else {
860 es->indent_off = off;
861 es->indent_width = width;
862 }
863 }
864
865 static int compute_ws_delta(const struct emitted_diff_symbol *a,
866 const struct emitted_diff_symbol *b)
867 {
868 int a_width = a->indent_width,
869 b_width = b->indent_width;
870
871 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
872 return INDENT_BLANKLINE;
873
874 return a_width - b_width;
875 }
876
877 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
878 const struct emitted_diff_symbol *l,
879 struct moved_block *pmb)
880 {
881 int a_width = cur->es->indent_width, b_width = l->indent_width;
882 int delta;
883
884 /* The text of each line must match */
885 if (cur->es->id != l->id)
886 return 1;
887
888 /*
889 * If 'l' and 'cur' are both blank then we don't need to check the
890 * indent. We only need to check cur as we know the strings match.
891 * */
892 if (a_width == INDENT_BLANKLINE)
893 return 0;
894
895 /*
896 * The indent changes of the block are known and stored in pmb->wsd;
897 * however we need to check if the indent changes of the current line
898 * match those of the current block.
899 */
900 delta = b_width - a_width;
901
902 /*
903 * If the previous lines of this block were all blank then set its
904 * whitespace delta.
905 */
906 if (pmb->wsd == INDENT_BLANKLINE)
907 pmb->wsd = delta;
908
909 return delta != pmb->wsd;
910 }
911
912 struct interned_diff_symbol {
913 struct hashmap_entry ent;
914 struct emitted_diff_symbol *es;
915 };
916
917 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
918 const struct hashmap_entry *eptr,
919 const struct hashmap_entry *entry_or_key,
920 const void *keydata UNUSED)
921 {
922 const struct diff_options *diffopt = hashmap_cmp_fn_data;
923 const struct emitted_diff_symbol *a, *b;
924 unsigned flags = diffopt->color_moved_ws_handling
925 & XDF_WHITESPACE_FLAGS;
926
927 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
928 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
929
930 return !xdiff_compare_lines(a->line + a->indent_off,
931 a->len - a->indent_off,
932 b->line + b->indent_off,
933 b->len - b->indent_off, flags);
934 }
935
936 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
937 struct interned_diff_symbol *s)
938 {
939 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
940 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
941 l->len - l->indent_off, flags);
942
943 hashmap_entry_init(&s->ent, hash);
944 s->es = l;
945 }
946
947 struct moved_entry_list {
948 struct moved_entry *add, *del;
949 };
950
951 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
952 struct mem_pool *entry_mem_pool)
953 {
954 struct moved_entry *prev_line = NULL;
955 struct mem_pool interned_pool;
956 struct hashmap interned_map;
957 struct moved_entry_list *entry_list = NULL;
958 size_t entry_list_alloc = 0;
959 unsigned id = 0;
960 int n;
961
962 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
963 mem_pool_init(&interned_pool, 1024 * 1024);
964
965 for (n = 0; n < o->emitted_symbols->nr; n++) {
966 struct interned_diff_symbol key;
967 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
968 struct interned_diff_symbol *s;
969 struct moved_entry *entry;
970
971 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
972 prev_line = NULL;
973 continue;
974 }
975
976 if (o->color_moved_ws_handling &
977 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
978 fill_es_indent_data(l);
979
980 prepare_entry(o, l, &key);
981 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
982 if (s) {
983 l->id = s->es->id;
984 } else {
985 l->id = id;
986 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
987 hashmap_add(&interned_map,
988 memcpy(mem_pool_alloc(&interned_pool,
989 sizeof(key)),
990 &key, sizeof(key)));
991 }
992 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
993 entry->es = l;
994 entry->next_line = NULL;
995 if (prev_line && prev_line->es->s == l->s)
996 prev_line->next_line = entry;
997 prev_line = entry;
998 if (l->s == DIFF_SYMBOL_PLUS) {
999 entry->next_match = entry_list[l->id].add;
1000 entry_list[l->id].add = entry;
1001 } else {
1002 entry->next_match = entry_list[l->id].del;
1003 entry_list[l->id].del = entry;
1004 }
1005 }
1006
1007 hashmap_clear(&interned_map);
1008 mem_pool_discard(&interned_pool, 0);
1009
1010 return entry_list;
1011 }
1012
1013 static void pmb_advance_or_null(struct diff_options *o,
1014 struct emitted_diff_symbol *l,
1015 struct moved_block *pmb,
1016 int *pmb_nr)
1017 {
1018 int i, j;
1019
1020 for (i = 0, j = 0; i < *pmb_nr; i++) {
1021 int match;
1022 struct moved_entry *prev = pmb[i].match;
1023 struct moved_entry *cur = (prev && prev->next_line) ?
1024 prev->next_line : NULL;
1025
1026 if (o->color_moved_ws_handling &
1027 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1028 match = cur &&
1029 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1030 else
1031 match = cur && cur->es->id == l->id;
1032
1033 if (match) {
1034 pmb[j] = pmb[i];
1035 pmb[j++].match = cur;
1036 }
1037 }
1038 *pmb_nr = j;
1039 }
1040
1041 static void fill_potential_moved_blocks(struct diff_options *o,
1042 struct moved_entry *match,
1043 struct emitted_diff_symbol *l,
1044 struct moved_block **pmb_p,
1045 int *pmb_alloc_p, int *pmb_nr_p)
1046
1047 {
1048 struct moved_block *pmb = *pmb_p;
1049 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1050
1051 /*
1052 * The current line is the start of a new block.
1053 * Setup the set of potential blocks.
1054 */
1055 for (; match; match = match->next_match) {
1056 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1057 if (o->color_moved_ws_handling &
1058 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1059 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1060 else
1061 pmb[pmb_nr].wsd = 0;
1062 pmb[pmb_nr++].match = match;
1063 }
1064
1065 *pmb_p = pmb;
1066 *pmb_alloc_p = pmb_alloc;
1067 *pmb_nr_p = pmb_nr;
1068 }
1069
1070 /*
1071 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1072 *
1073 * Otherwise, if the last block has fewer alphanumeric characters than
1074 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1075 * that block.
1076 *
1077 * The last block consists of the (n - block_length)'th line up to but not
1078 * including the nth line.
1079 *
1080 * Returns 0 if the last block is empty or is unset by this function, non zero
1081 * otherwise.
1082 *
1083 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1084 * Think of a way to unify them.
1085 */
1086 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1087 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1088 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1089 {
1090 int i, alnum_count = 0;
1091 if (o->color_moved == COLOR_MOVED_PLAIN)
1092 return block_length;
1093 for (i = 1; i < block_length + 1; i++) {
1094 const char *c = o->emitted_symbols->buf[n - i].line;
1095 for (; *c; c++) {
1096 if (!isalnum(*c))
1097 continue;
1098 alnum_count++;
1099 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1100 return 1;
1101 }
1102 }
1103 for (i = 1; i < block_length + 1; i++)
1104 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1105 return 0;
1106 }
1107
1108 /* Find blocks of moved code, delegate actual coloring decision to helper */
1109 static void mark_color_as_moved(struct diff_options *o,
1110 struct moved_entry_list *entry_list)
1111 {
1112 struct moved_block *pmb = NULL; /* potentially moved blocks */
1113 int pmb_nr = 0, pmb_alloc = 0;
1114 int n, flipped_block = 0, block_length = 0;
1115 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1116
1117
1118 for (n = 0; n < o->emitted_symbols->nr; n++) {
1119 struct moved_entry *match = NULL;
1120 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1121
1122 switch (l->s) {
1123 case DIFF_SYMBOL_PLUS:
1124 match = entry_list[l->id].del;
1125 break;
1126 case DIFF_SYMBOL_MINUS:
1127 match = entry_list[l->id].add;
1128 break;
1129 default:
1130 flipped_block = 0;
1131 }
1132
1133 if (pmb_nr && (!match || l->s != moved_symbol)) {
1134 if (!adjust_last_block(o, n, block_length) &&
1135 block_length > 1) {
1136 /*
1137 * Rewind in case there is another match
1138 * starting at the second line of the block
1139 */
1140 match = NULL;
1141 n -= block_length;
1142 }
1143 pmb_nr = 0;
1144 block_length = 0;
1145 flipped_block = 0;
1146 }
1147 if (!match) {
1148 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1149 continue;
1150 }
1151
1152 if (o->color_moved == COLOR_MOVED_PLAIN) {
1153 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1154 continue;
1155 }
1156
1157 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1158
1159 if (pmb_nr == 0) {
1160 int contiguous = adjust_last_block(o, n, block_length);
1161
1162 if (!contiguous && block_length > 1)
1163 /*
1164 * Rewind in case there is another match
1165 * starting at the second line of the block
1166 */
1167 n -= block_length;
1168 else
1169 fill_potential_moved_blocks(o, match, l,
1170 &pmb, &pmb_alloc,
1171 &pmb_nr);
1172
1173 if (contiguous && pmb_nr && moved_symbol == l->s)
1174 flipped_block = (flipped_block + 1) % 2;
1175 else
1176 flipped_block = 0;
1177
1178 if (pmb_nr)
1179 moved_symbol = l->s;
1180 else
1181 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1182
1183 block_length = 0;
1184 }
1185
1186 if (pmb_nr) {
1187 block_length++;
1188 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1189 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1190 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1191 }
1192 }
1193 adjust_last_block(o, n, block_length);
1194
1195 free(pmb);
1196 }
1197
1198 static void dim_moved_lines(struct diff_options *o)
1199 {
1200 int n;
1201 for (n = 0; n < o->emitted_symbols->nr; n++) {
1202 struct emitted_diff_symbol *prev = (n != 0) ?
1203 &o->emitted_symbols->buf[n - 1] : NULL;
1204 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1205 struct emitted_diff_symbol *next =
1206 (n < o->emitted_symbols->nr - 1) ?
1207 &o->emitted_symbols->buf[n + 1] : NULL;
1208
1209 /* Not a plus or minus line? */
1210 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1211 continue;
1212
1213 /* Not a moved line? */
1214 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1215 continue;
1216
1217 /*
1218 * If prev or next are not a plus or minus line,
1219 * pretend they don't exist
1220 */
1221 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1222 prev->s != DIFF_SYMBOL_MINUS)
1223 prev = NULL;
1224 if (next && next->s != DIFF_SYMBOL_PLUS &&
1225 next->s != DIFF_SYMBOL_MINUS)
1226 next = NULL;
1227
1228 /* Inside a block? */
1229 if ((prev &&
1230 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1231 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1232 (next &&
1233 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1234 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1235 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1236 continue;
1237 }
1238
1239 /* Check if we are at an interesting bound: */
1240 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1241 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1242 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1243 continue;
1244 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1245 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1246 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1247 continue;
1248
1249 /*
1250 * The boundary to prev and next are not interesting,
1251 * so this line is not interesting as a whole
1252 */
1253 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1254 }
1255 }
1256
1257 static void emit_line_ws_markup(struct diff_options *o,
1258 const char *set_sign, const char *set,
1259 const char *reset,
1260 int sign_index, const char *line, int len,
1261 unsigned ws_rule, int blank_at_eof)
1262 {
1263 const char *ws = NULL;
1264 int sign = o->output_indicators[sign_index];
1265
1266 if (o->ws_error_highlight & ws_rule) {
1267 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1268 if (!*ws)
1269 ws = NULL;
1270 }
1271
1272 if (!ws && !set_sign)
1273 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1274 else if (!ws) {
1275 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1276 } else if (blank_at_eof)
1277 /* Blank line at EOF - paint '+' as well */
1278 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1279 else {
1280 /* Emit just the prefix, then the rest. */
1281 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1282 sign, "", 0);
1283 ws_check_emit(line, len, ws_rule,
1284 o->file, set, reset, ws);
1285 }
1286 }
1287
1288 static void emit_diff_symbol_from_struct(struct diff_options *o,
1289 struct emitted_diff_symbol *eds)
1290 {
1291 static const char *nneof = " No newline at end of file\n";
1292 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1293
1294 enum diff_symbol s = eds->s;
1295 const char *line = eds->line;
1296 int len = eds->len;
1297 unsigned flags = eds->flags;
1298
1299 switch (s) {
1300 case DIFF_SYMBOL_NO_LF_EOF:
1301 context = diff_get_color_opt(o, DIFF_CONTEXT);
1302 reset = diff_get_color_opt(o, DIFF_RESET);
1303 putc('\n', o->file);
1304 emit_line_0(o, context, NULL, 0, reset, '\\',
1305 nneof, strlen(nneof));
1306 break;
1307 case DIFF_SYMBOL_SUBMODULE_HEADER:
1308 case DIFF_SYMBOL_SUBMODULE_ERROR:
1309 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1310 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1311 case DIFF_SYMBOL_SUMMARY:
1312 case DIFF_SYMBOL_STATS_LINE:
1313 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1314 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1315 emit_line(o, "", "", line, len);
1316 break;
1317 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1318 case DIFF_SYMBOL_CONTEXT_MARKER:
1319 context = diff_get_color_opt(o, DIFF_CONTEXT);
1320 reset = diff_get_color_opt(o, DIFF_RESET);
1321 emit_line(o, context, reset, line, len);
1322 break;
1323 case DIFF_SYMBOL_SEPARATOR:
1324 fprintf(o->file, "%s%c",
1325 diff_line_prefix(o),
1326 o->line_termination);
1327 break;
1328 case DIFF_SYMBOL_CONTEXT:
1329 set = diff_get_color_opt(o, DIFF_CONTEXT);
1330 reset = diff_get_color_opt(o, DIFF_RESET);
1331 set_sign = NULL;
1332 if (o->flags.dual_color_diffed_diffs) {
1333 char c = !len ? 0 : line[0];
1334
1335 if (c == '+')
1336 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1337 else if (c == '@')
1338 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1339 else if (c == '-')
1340 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1341 }
1342 emit_line_ws_markup(o, set_sign, set, reset,
1343 OUTPUT_INDICATOR_CONTEXT, line, len,
1344 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1345 break;
1346 case DIFF_SYMBOL_PLUS:
1347 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1348 DIFF_SYMBOL_MOVED_LINE_ALT |
1349 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1350 case DIFF_SYMBOL_MOVED_LINE |
1351 DIFF_SYMBOL_MOVED_LINE_ALT |
1352 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1353 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1354 break;
1355 case DIFF_SYMBOL_MOVED_LINE |
1356 DIFF_SYMBOL_MOVED_LINE_ALT:
1357 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1358 break;
1359 case DIFF_SYMBOL_MOVED_LINE |
1360 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1361 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1362 break;
1363 case DIFF_SYMBOL_MOVED_LINE:
1364 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1365 break;
1366 default:
1367 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1368 }
1369 reset = diff_get_color_opt(o, DIFF_RESET);
1370 if (!o->flags.dual_color_diffed_diffs)
1371 set_sign = NULL;
1372 else {
1373 char c = !len ? 0 : line[0];
1374
1375 set_sign = set;
1376 if (c == '-')
1377 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1378 else if (c == '@')
1379 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1380 else if (c == '+')
1381 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1382 else
1383 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1384 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1385 }
1386 emit_line_ws_markup(o, set_sign, set, reset,
1387 OUTPUT_INDICATOR_NEW, line, len,
1388 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1389 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1390 break;
1391 case DIFF_SYMBOL_MINUS:
1392 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1393 DIFF_SYMBOL_MOVED_LINE_ALT |
1394 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1395 case DIFF_SYMBOL_MOVED_LINE |
1396 DIFF_SYMBOL_MOVED_LINE_ALT |
1397 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1398 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1399 break;
1400 case DIFF_SYMBOL_MOVED_LINE |
1401 DIFF_SYMBOL_MOVED_LINE_ALT:
1402 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1403 break;
1404 case DIFF_SYMBOL_MOVED_LINE |
1405 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1406 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1407 break;
1408 case DIFF_SYMBOL_MOVED_LINE:
1409 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1410 break;
1411 default:
1412 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1413 }
1414 reset = diff_get_color_opt(o, DIFF_RESET);
1415 if (!o->flags.dual_color_diffed_diffs)
1416 set_sign = NULL;
1417 else {
1418 char c = !len ? 0 : line[0];
1419
1420 set_sign = set;
1421 if (c == '+')
1422 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1423 else if (c == '@')
1424 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1425 else if (c == '-')
1426 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1427 else
1428 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1429 }
1430 emit_line_ws_markup(o, set_sign, set, reset,
1431 OUTPUT_INDICATOR_OLD, line, len,
1432 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1433 break;
1434 case DIFF_SYMBOL_WORDS_PORCELAIN:
1435 context = diff_get_color_opt(o, DIFF_CONTEXT);
1436 reset = diff_get_color_opt(o, DIFF_RESET);
1437 emit_line(o, context, reset, line, len);
1438 fputs("~\n", o->file);
1439 break;
1440 case DIFF_SYMBOL_WORDS:
1441 context = diff_get_color_opt(o, DIFF_CONTEXT);
1442 reset = diff_get_color_opt(o, DIFF_RESET);
1443 /*
1444 * Skip the prefix character, if any. With
1445 * diff_suppress_blank_empty, there may be
1446 * none.
1447 */
1448 if (line[0] != '\n') {
1449 line++;
1450 len--;
1451 }
1452 emit_line(o, context, reset, line, len);
1453 break;
1454 case DIFF_SYMBOL_FILEPAIR_PLUS:
1455 meta = diff_get_color_opt(o, DIFF_METAINFO);
1456 reset = diff_get_color_opt(o, DIFF_RESET);
1457 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1458 line, reset,
1459 strchr(line, ' ') ? "\t" : "");
1460 break;
1461 case DIFF_SYMBOL_FILEPAIR_MINUS:
1462 meta = diff_get_color_opt(o, DIFF_METAINFO);
1463 reset = diff_get_color_opt(o, DIFF_RESET);
1464 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1465 line, reset,
1466 strchr(line, ' ') ? "\t" : "");
1467 break;
1468 case DIFF_SYMBOL_BINARY_FILES:
1469 case DIFF_SYMBOL_HEADER:
1470 fprintf(o->file, "%s", line);
1471 break;
1472 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1473 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1474 break;
1475 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1476 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1477 break;
1478 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1479 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1480 break;
1481 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1482 fputs(diff_line_prefix(o), o->file);
1483 fputc('\n', o->file);
1484 break;
1485 case DIFF_SYMBOL_REWRITE_DIFF:
1486 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1487 reset = diff_get_color_opt(o, DIFF_RESET);
1488 emit_line(o, fraginfo, reset, line, len);
1489 break;
1490 case DIFF_SYMBOL_SUBMODULE_ADD:
1491 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1492 reset = diff_get_color_opt(o, DIFF_RESET);
1493 emit_line(o, set, reset, line, len);
1494 break;
1495 case DIFF_SYMBOL_SUBMODULE_DEL:
1496 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1497 reset = diff_get_color_opt(o, DIFF_RESET);
1498 emit_line(o, set, reset, line, len);
1499 break;
1500 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1501 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1502 diff_line_prefix(o), line);
1503 break;
1504 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1505 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1506 diff_line_prefix(o), line);
1507 break;
1508 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1509 emit_line(o, "", "", " 0 files changed\n",
1510 strlen(" 0 files changed\n"));
1511 break;
1512 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1513 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1514 break;
1515 case DIFF_SYMBOL_WORD_DIFF:
1516 fprintf(o->file, "%.*s", len, line);
1517 break;
1518 case DIFF_SYMBOL_STAT_SEP:
1519 fputs(o->stat_sep, o->file);
1520 break;
1521 default:
1522 BUG("unknown diff symbol");
1523 }
1524 }
1525
1526 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1527 const char *line, int len, unsigned flags)
1528 {
1529 struct emitted_diff_symbol e = {
1530 .line = line, .len = len, .flags = flags, .s = s
1531 };
1532
1533 if (o->emitted_symbols)
1534 append_emitted_diff_symbol(o, &e);
1535 else
1536 emit_diff_symbol_from_struct(o, &e);
1537 }
1538
1539 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1540 {
1541 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1542 }
1543
1544 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1545 {
1546 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1547 }
1548
1549 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1550 {
1551 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1552 path, strlen(path), 0);
1553 }
1554
1555 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1556 {
1557 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1558 path, strlen(path), 0);
1559 }
1560
1561 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1562 {
1563 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1564 header, strlen(header), 0);
1565 }
1566
1567 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1568 {
1569 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1570 }
1571
1572 void diff_emit_submodule_pipethrough(struct diff_options *o,
1573 const char *line, int len)
1574 {
1575 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1576 }
1577
1578 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1579 {
1580 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1581 ecbdata->blank_at_eof_in_preimage &&
1582 ecbdata->blank_at_eof_in_postimage &&
1583 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1584 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1585 return 0;
1586 return ws_blank_line(line, len, ecbdata->ws_rule);
1587 }
1588
1589 static void emit_add_line(struct emit_callback *ecbdata,
1590 const char *line, int len)
1591 {
1592 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1593 if (new_blank_line_at_eof(ecbdata, line, len))
1594 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1595
1596 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1597 }
1598
1599 static void emit_del_line(struct emit_callback *ecbdata,
1600 const char *line, int len)
1601 {
1602 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1603 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1604 }
1605
1606 static void emit_context_line(struct emit_callback *ecbdata,
1607 const char *line, int len)
1608 {
1609 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1610 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1611 }
1612
1613 static void emit_hunk_header(struct emit_callback *ecbdata,
1614 const char *line, int len)
1615 {
1616 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1617 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1618 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1619 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1620 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1621 static const char atat[2] = { '@', '@' };
1622 const char *cp, *ep;
1623 struct strbuf msgbuf = STRBUF_INIT;
1624 int org_len = len;
1625 int i = 1;
1626
1627 /*
1628 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1629 * it always is at least 10 bytes long.
1630 */
1631 if (len < 10 ||
1632 memcmp(line, atat, 2) ||
1633 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1634 emit_diff_symbol(ecbdata->opt,
1635 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1636 return;
1637 }
1638 ep += 2; /* skip over @@ */
1639
1640 /* The hunk header in fraginfo color */
1641 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1642 strbuf_addstr(&msgbuf, reverse);
1643 strbuf_addstr(&msgbuf, frag);
1644 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1645 strbuf_add(&msgbuf, atat, sizeof(atat));
1646 else
1647 strbuf_add(&msgbuf, line, ep - line);
1648 strbuf_addstr(&msgbuf, reset);
1649
1650 /*
1651 * trailing "\r\n"
1652 */
1653 for ( ; i < 3; i++)
1654 if (line[len - i] == '\r' || line[len - i] == '\n')
1655 len--;
1656
1657 /* blank before the func header */
1658 for (cp = ep; ep - line < len; ep++)
1659 if (*ep != ' ' && *ep != '\t')
1660 break;
1661 if (ep != cp) {
1662 strbuf_addstr(&msgbuf, context);
1663 strbuf_add(&msgbuf, cp, ep - cp);
1664 strbuf_addstr(&msgbuf, reset);
1665 }
1666
1667 if (ep < line + len) {
1668 strbuf_addstr(&msgbuf, func);
1669 strbuf_add(&msgbuf, ep, line + len - ep);
1670 strbuf_addstr(&msgbuf, reset);
1671 }
1672
1673 strbuf_add(&msgbuf, line + len, org_len - len);
1674 strbuf_complete_line(&msgbuf);
1675 emit_diff_symbol(ecbdata->opt,
1676 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1677 strbuf_release(&msgbuf);
1678 }
1679
1680 static struct diff_tempfile *claim_diff_tempfile(void)
1681 {
1682 int i;
1683 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1684 if (!diff_temp[i].name)
1685 return diff_temp + i;
1686 BUG("diff is failing to clean up its tempfiles");
1687 }
1688
1689 static void remove_tempfile(void)
1690 {
1691 int i;
1692 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1693 if (is_tempfile_active(diff_temp[i].tempfile))
1694 delete_tempfile(&diff_temp[i].tempfile);
1695 diff_temp[i].name = NULL;
1696 }
1697 }
1698
1699 static void add_line_count(struct strbuf *out, int count)
1700 {
1701 switch (count) {
1702 case 0:
1703 strbuf_addstr(out, "0,0");
1704 break;
1705 case 1:
1706 strbuf_addstr(out, "1");
1707 break;
1708 default:
1709 strbuf_addf(out, "1,%d", count);
1710 break;
1711 }
1712 }
1713
1714 static void emit_rewrite_lines(struct emit_callback *ecb,
1715 int prefix, const char *data, int size)
1716 {
1717 const char *endp = NULL;
1718
1719 while (0 < size) {
1720 int len;
1721
1722 endp = memchr(data, '\n', size);
1723 len = endp ? (endp - data + 1) : size;
1724 if (prefix != '+') {
1725 ecb->lno_in_preimage++;
1726 emit_del_line(ecb, data, len);
1727 } else {
1728 ecb->lno_in_postimage++;
1729 emit_add_line(ecb, data, len);
1730 }
1731 size -= len;
1732 data += len;
1733 }
1734 if (!endp)
1735 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1736 }
1737
1738 static void emit_rewrite_diff(const char *name_a,
1739 const char *name_b,
1740 struct diff_filespec *one,
1741 struct diff_filespec *two,
1742 struct userdiff_driver *textconv_one,
1743 struct userdiff_driver *textconv_two,
1744 struct diff_options *o)
1745 {
1746 int lc_a, lc_b;
1747 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1748 const char *a_prefix, *b_prefix;
1749 char *data_one, *data_two;
1750 size_t size_one, size_two;
1751 struct emit_callback ecbdata;
1752 struct strbuf out = STRBUF_INIT;
1753
1754 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1755 a_prefix = o->b_prefix;
1756 b_prefix = o->a_prefix;
1757 } else {
1758 a_prefix = o->a_prefix;
1759 b_prefix = o->b_prefix;
1760 }
1761
1762 name_a += (*name_a == '/');
1763 name_b += (*name_b == '/');
1764
1765 strbuf_reset(&a_name);
1766 strbuf_reset(&b_name);
1767 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1768 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1769
1770 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1771 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1772
1773 memset(&ecbdata, 0, sizeof(ecbdata));
1774 ecbdata.color_diff = want_color(o->use_color);
1775 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1776 ecbdata.opt = o;
1777 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1778 mmfile_t mf1, mf2;
1779 mf1.ptr = (char *)data_one;
1780 mf2.ptr = (char *)data_two;
1781 mf1.size = size_one;
1782 mf2.size = size_two;
1783 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1784 }
1785 ecbdata.lno_in_preimage = 1;
1786 ecbdata.lno_in_postimage = 1;
1787
1788 lc_a = count_lines(data_one, size_one);
1789 lc_b = count_lines(data_two, size_two);
1790
1791 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1792 a_name.buf, a_name.len, 0);
1793 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1794 b_name.buf, b_name.len, 0);
1795
1796 strbuf_addstr(&out, "@@ -");
1797 if (!o->irreversible_delete)
1798 add_line_count(&out, lc_a);
1799 else
1800 strbuf_addstr(&out, "?,?");
1801 strbuf_addstr(&out, " +");
1802 add_line_count(&out, lc_b);
1803 strbuf_addstr(&out, " @@\n");
1804 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1805 strbuf_release(&out);
1806
1807 if (lc_a && !o->irreversible_delete)
1808 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1809 if (lc_b)
1810 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1811 if (textconv_one)
1812 free((char *)data_one);
1813 if (textconv_two)
1814 free((char *)data_two);
1815 }
1816
1817 struct diff_words_buffer {
1818 mmfile_t text;
1819 unsigned long alloc;
1820 struct diff_words_orig {
1821 const char *begin, *end;
1822 } *orig;
1823 int orig_nr, orig_alloc;
1824 };
1825
1826 static void diff_words_append(char *line, unsigned long len,
1827 struct diff_words_buffer *buffer)
1828 {
1829 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1830 line++;
1831 len--;
1832 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1833 buffer->text.size += len;
1834 buffer->text.ptr[buffer->text.size] = '\0';
1835 }
1836
1837 struct diff_words_style_elem {
1838 const char *prefix;
1839 const char *suffix;
1840 const char *color; /* NULL; filled in by the setup code if
1841 * color is enabled */
1842 };
1843
1844 struct diff_words_style {
1845 enum diff_words_type type;
1846 struct diff_words_style_elem new_word, old_word, ctx;
1847 const char *newline;
1848 };
1849
1850 static struct diff_words_style diff_words_styles[] = {
1851 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1852 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1853 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1854 };
1855
1856 struct diff_words_data {
1857 struct diff_words_buffer minus, plus;
1858 const char *current_plus;
1859 int last_minus;
1860 struct diff_options *opt;
1861 regex_t *word_regex;
1862 enum diff_words_type type;
1863 struct diff_words_style *style;
1864 };
1865
1866 static int fn_out_diff_words_write_helper(struct diff_options *o,
1867 struct diff_words_style_elem *st_el,
1868 const char *newline,
1869 size_t count, const char *buf)
1870 {
1871 int print = 0;
1872 struct strbuf sb = STRBUF_INIT;
1873
1874 while (count) {
1875 char *p = memchr(buf, '\n', count);
1876 if (print)
1877 strbuf_addstr(&sb, diff_line_prefix(o));
1878
1879 if (p != buf) {
1880 const char *reset = st_el->color && *st_el->color ?
1881 GIT_COLOR_RESET : NULL;
1882 if (st_el->color && *st_el->color)
1883 strbuf_addstr(&sb, st_el->color);
1884 strbuf_addstr(&sb, st_el->prefix);
1885 strbuf_add(&sb, buf, p ? p - buf : count);
1886 strbuf_addstr(&sb, st_el->suffix);
1887 if (reset)
1888 strbuf_addstr(&sb, reset);
1889 }
1890 if (!p)
1891 goto out;
1892
1893 strbuf_addstr(&sb, newline);
1894 count -= p + 1 - buf;
1895 buf = p + 1;
1896 print = 1;
1897 if (count) {
1898 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1899 sb.buf, sb.len, 0);
1900 strbuf_reset(&sb);
1901 }
1902 }
1903
1904 out:
1905 if (sb.len)
1906 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1907 sb.buf, sb.len, 0);
1908 strbuf_release(&sb);
1909 return 0;
1910 }
1911
1912 /*
1913 * '--color-words' algorithm can be described as:
1914 *
1915 * 1. collect the minus/plus lines of a diff hunk, divided into
1916 * minus-lines and plus-lines;
1917 *
1918 * 2. break both minus-lines and plus-lines into words and
1919 * place them into two mmfile_t with one word for each line;
1920 *
1921 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1922 *
1923 * And for the common parts of the both file, we output the plus side text.
1924 * diff_words->current_plus is used to trace the current position of the plus file
1925 * which printed. diff_words->last_minus is used to trace the last minus word
1926 * printed.
1927 *
1928 * For '--graph' to work with '--color-words', we need to output the graph prefix
1929 * on each line of color words output. Generally, there are two conditions on
1930 * which we should output the prefix.
1931 *
1932 * 1. diff_words->last_minus == 0 &&
1933 * diff_words->current_plus == diff_words->plus.text.ptr
1934 *
1935 * that is: the plus text must start as a new line, and if there is no minus
1936 * word printed, a graph prefix must be printed.
1937 *
1938 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1939 * *(diff_words->current_plus - 1) == '\n'
1940 *
1941 * that is: a graph prefix must be printed following a '\n'
1942 */
1943 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1944 {
1945 if ((diff_words->last_minus == 0 &&
1946 diff_words->current_plus == diff_words->plus.text.ptr) ||
1947 (diff_words->current_plus > diff_words->plus.text.ptr &&
1948 *(diff_words->current_plus - 1) == '\n')) {
1949 return 1;
1950 } else {
1951 return 0;
1952 }
1953 }
1954
1955 static void fn_out_diff_words_aux(void *priv,
1956 long minus_first, long minus_len,
1957 long plus_first, long plus_len,
1958 const char *func, long funclen)
1959 {
1960 struct diff_words_data *diff_words = priv;
1961 struct diff_words_style *style = diff_words->style;
1962 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1963 struct diff_options *opt = diff_words->opt;
1964 const char *line_prefix;
1965
1966 assert(opt);
1967 line_prefix = diff_line_prefix(opt);
1968
1969 /* POSIX requires that first be decremented by one if len == 0... */
1970 if (minus_len) {
1971 minus_begin = diff_words->minus.orig[minus_first].begin;
1972 minus_end =
1973 diff_words->minus.orig[minus_first + minus_len - 1].end;
1974 } else
1975 minus_begin = minus_end =
1976 diff_words->minus.orig[minus_first].end;
1977
1978 if (plus_len) {
1979 plus_begin = diff_words->plus.orig[plus_first].begin;
1980 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1981 } else
1982 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1983
1984 if (color_words_output_graph_prefix(diff_words)) {
1985 fputs(line_prefix, diff_words->opt->file);
1986 }
1987 if (diff_words->current_plus != plus_begin) {
1988 fn_out_diff_words_write_helper(diff_words->opt,
1989 &style->ctx, style->newline,
1990 plus_begin - diff_words->current_plus,
1991 diff_words->current_plus);
1992 }
1993 if (minus_begin != minus_end) {
1994 fn_out_diff_words_write_helper(diff_words->opt,
1995 &style->old_word, style->newline,
1996 minus_end - minus_begin, minus_begin);
1997 }
1998 if (plus_begin != plus_end) {
1999 fn_out_diff_words_write_helper(diff_words->opt,
2000 &style->new_word, style->newline,
2001 plus_end - plus_begin, plus_begin);
2002 }
2003
2004 diff_words->current_plus = plus_end;
2005 diff_words->last_minus = minus_first;
2006 }
2007
2008 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2009 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2010 int *begin, int *end)
2011 {
2012 while (word_regex && *begin < buffer->size) {
2013 regmatch_t match[1];
2014 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2015 buffer->size - *begin, 1, match, 0)) {
2016 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2017 '\n', match[0].rm_eo - match[0].rm_so);
2018 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2019 *begin += match[0].rm_so;
2020 if (*begin == *end)
2021 (*begin)++;
2022 else
2023 return *begin > *end;
2024 } else {
2025 return -1;
2026 }
2027 }
2028
2029 /* find the next word */
2030 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2031 (*begin)++;
2032 if (*begin >= buffer->size)
2033 return -1;
2034
2035 /* find the end of the word */
2036 *end = *begin + 1;
2037 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2038 (*end)++;
2039
2040 return 0;
2041 }
2042
2043 /*
2044 * This function splits the words in buffer->text, stores the list with
2045 * newline separator into out, and saves the offsets of the original words
2046 * in buffer->orig.
2047 */
2048 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2049 regex_t *word_regex)
2050 {
2051 int i, j;
2052 long alloc = 0;
2053
2054 out->size = 0;
2055 out->ptr = NULL;
2056
2057 /* fake an empty "0th" word */
2058 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2059 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2060 buffer->orig_nr = 1;
2061
2062 for (i = 0; i < buffer->text.size; i++) {
2063 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2064 return;
2065
2066 /* store original boundaries */
2067 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2068 buffer->orig_alloc);
2069 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2070 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2071 buffer->orig_nr++;
2072
2073 /* store one word */
2074 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2075 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2076 out->ptr[out->size + j - i] = '\n';
2077 out->size += j - i + 1;
2078
2079 i = j - 1;
2080 }
2081 }
2082
2083 /* this executes the word diff on the accumulated buffers */
2084 static void diff_words_show(struct diff_words_data *diff_words)
2085 {
2086 xpparam_t xpp;
2087 xdemitconf_t xecfg;
2088 mmfile_t minus, plus;
2089 struct diff_words_style *style = diff_words->style;
2090
2091 struct diff_options *opt = diff_words->opt;
2092 const char *line_prefix;
2093
2094 assert(opt);
2095 line_prefix = diff_line_prefix(opt);
2096
2097 /* special case: only removal */
2098 if (!diff_words->plus.text.size) {
2099 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2100 line_prefix, strlen(line_prefix), 0);
2101 fn_out_diff_words_write_helper(diff_words->opt,
2102 &style->old_word, style->newline,
2103 diff_words->minus.text.size,
2104 diff_words->minus.text.ptr);
2105 diff_words->minus.text.size = 0;
2106 return;
2107 }
2108
2109 diff_words->current_plus = diff_words->plus.text.ptr;
2110 diff_words->last_minus = 0;
2111
2112 memset(&xpp, 0, sizeof(xpp));
2113 memset(&xecfg, 0, sizeof(xecfg));
2114 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2115 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2116 xpp.flags = 0;
2117 /* as only the hunk header will be parsed, we need a 0-context */
2118 xecfg.ctxlen = 0;
2119 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2120 diff_words, &xpp, &xecfg))
2121 die("unable to generate word diff");
2122 free(minus.ptr);
2123 free(plus.ptr);
2124 if (diff_words->current_plus != diff_words->plus.text.ptr +
2125 diff_words->plus.text.size) {
2126 if (color_words_output_graph_prefix(diff_words))
2127 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2128 line_prefix, strlen(line_prefix), 0);
2129 fn_out_diff_words_write_helper(diff_words->opt,
2130 &style->ctx, style->newline,
2131 diff_words->plus.text.ptr + diff_words->plus.text.size
2132 - diff_words->current_plus, diff_words->current_plus);
2133 }
2134 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2135 }
2136
2137 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2138 static void diff_words_flush(struct emit_callback *ecbdata)
2139 {
2140 struct diff_options *wo = ecbdata->diff_words->opt;
2141
2142 if (ecbdata->diff_words->minus.text.size ||
2143 ecbdata->diff_words->plus.text.size)
2144 diff_words_show(ecbdata->diff_words);
2145
2146 if (wo->emitted_symbols) {
2147 struct diff_options *o = ecbdata->opt;
2148 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2149 int i;
2150
2151 /*
2152 * NEEDSWORK:
2153 * Instead of appending each, concat all words to a line?
2154 */
2155 for (i = 0; i < wol->nr; i++)
2156 append_emitted_diff_symbol(o, &wol->buf[i]);
2157
2158 for (i = 0; i < wol->nr; i++)
2159 free((void *)wol->buf[i].line);
2160
2161 wol->nr = 0;
2162 }
2163 }
2164
2165 static void diff_filespec_load_driver(struct diff_filespec *one,
2166 struct index_state *istate)
2167 {
2168 /* Use already-loaded driver */
2169 if (one->driver)
2170 return;
2171
2172 if (S_ISREG(one->mode))
2173 one->driver = userdiff_find_by_path(istate, one->path);
2174
2175 /* Fallback to default settings */
2176 if (!one->driver)
2177 one->driver = userdiff_find_by_name("default");
2178 }
2179
2180 static const char *userdiff_word_regex(struct diff_filespec *one,
2181 struct index_state *istate)
2182 {
2183 diff_filespec_load_driver(one, istate);
2184 return one->driver->word_regex;
2185 }
2186
2187 static void init_diff_words_data(struct emit_callback *ecbdata,
2188 struct diff_options *orig_opts,
2189 struct diff_filespec *one,
2190 struct diff_filespec *two)
2191 {
2192 int i;
2193 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2194 memcpy(o, orig_opts, sizeof(struct diff_options));
2195
2196 CALLOC_ARRAY(ecbdata->diff_words, 1);
2197 ecbdata->diff_words->type = o->word_diff;
2198 ecbdata->diff_words->opt = o;
2199
2200 if (orig_opts->emitted_symbols)
2201 CALLOC_ARRAY(o->emitted_symbols, 1);
2202
2203 if (!o->word_regex)
2204 o->word_regex = userdiff_word_regex(one, o->repo->index);
2205 if (!o->word_regex)
2206 o->word_regex = userdiff_word_regex(two, o->repo->index);
2207 if (!o->word_regex)
2208 o->word_regex = diff_word_regex_cfg;
2209 if (o->word_regex) {
2210 ecbdata->diff_words->word_regex = (regex_t *)
2211 xmalloc(sizeof(regex_t));
2212 if (regcomp(ecbdata->diff_words->word_regex,
2213 o->word_regex,
2214 REG_EXTENDED | REG_NEWLINE))
2215 die("invalid regular expression: %s",
2216 o->word_regex);
2217 }
2218 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2219 if (o->word_diff == diff_words_styles[i].type) {
2220 ecbdata->diff_words->style =
2221 &diff_words_styles[i];
2222 break;
2223 }
2224 }
2225 if (want_color(o->use_color)) {
2226 struct diff_words_style *st = ecbdata->diff_words->style;
2227 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2228 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2229 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2230 }
2231 }
2232
2233 static void free_diff_words_data(struct emit_callback *ecbdata)
2234 {
2235 if (ecbdata->diff_words) {
2236 diff_words_flush(ecbdata);
2237 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2238 free (ecbdata->diff_words->opt);
2239 free (ecbdata->diff_words->minus.text.ptr);
2240 free (ecbdata->diff_words->minus.orig);
2241 free (ecbdata->diff_words->plus.text.ptr);
2242 free (ecbdata->diff_words->plus.orig);
2243 if (ecbdata->diff_words->word_regex) {
2244 regfree(ecbdata->diff_words->word_regex);
2245 free(ecbdata->diff_words->word_regex);
2246 }
2247 FREE_AND_NULL(ecbdata->diff_words);
2248 }
2249 }
2250
2251 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2252 {
2253 if (want_color(diff_use_color))
2254 return diff_colors[ix];
2255 return "";
2256 }
2257
2258 const char *diff_line_prefix(struct diff_options *opt)
2259 {
2260 struct strbuf *msgbuf;
2261 if (!opt->output_prefix)
2262 return "";
2263
2264 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2265 return msgbuf->buf;
2266 }
2267
2268 static unsigned long sane_truncate_line(char *line, unsigned long len)
2269 {
2270 const char *cp;
2271 unsigned long allot;
2272 size_t l = len;
2273
2274 cp = line;
2275 allot = l;
2276 while (0 < l) {
2277 (void) utf8_width(&cp, &l);
2278 if (!cp)
2279 break; /* truncated in the middle? */
2280 }
2281 return allot - l;
2282 }
2283
2284 static void find_lno(const char *line, struct emit_callback *ecbdata)
2285 {
2286 const char *p;
2287 ecbdata->lno_in_preimage = 0;
2288 ecbdata->lno_in_postimage = 0;
2289 p = strchr(line, '-');
2290 if (!p)
2291 return; /* cannot happen */
2292 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2293 p = strchr(p, '+');
2294 if (!p)
2295 return; /* cannot happen */
2296 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2297 }
2298
2299 static int fn_out_consume(void *priv, char *line, unsigned long len)
2300 {
2301 struct emit_callback *ecbdata = priv;
2302 struct diff_options *o = ecbdata->opt;
2303
2304 o->found_changes = 1;
2305
2306 if (ecbdata->header) {
2307 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2308 ecbdata->header->buf, ecbdata->header->len, 0);
2309 strbuf_reset(ecbdata->header);
2310 ecbdata->header = NULL;
2311 }
2312
2313 if (ecbdata->label_path[0]) {
2314 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2315 ecbdata->label_path[0],
2316 strlen(ecbdata->label_path[0]), 0);
2317 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2318 ecbdata->label_path[1],
2319 strlen(ecbdata->label_path[1]), 0);
2320 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2321 }
2322
2323 if (diff_suppress_blank_empty
2324 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2325 line[0] = '\n';
2326 len = 1;
2327 }
2328
2329 if (line[0] == '@') {
2330 if (ecbdata->diff_words)
2331 diff_words_flush(ecbdata);
2332 len = sane_truncate_line(line, len);
2333 find_lno(line, ecbdata);
2334 emit_hunk_header(ecbdata, line, len);
2335 return 0;
2336 }
2337
2338 if (ecbdata->diff_words) {
2339 enum diff_symbol s =
2340 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2341 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2342 if (line[0] == '-') {
2343 diff_words_append(line, len,
2344 &ecbdata->diff_words->minus);
2345 return 0;
2346 } else if (line[0] == '+') {
2347 diff_words_append(line, len,
2348 &ecbdata->diff_words->plus);
2349 return 0;
2350 } else if (starts_with(line, "\\ ")) {
2351 /*
2352 * Eat the "no newline at eof" marker as if we
2353 * saw a "+" or "-" line with nothing on it,
2354 * and return without diff_words_flush() to
2355 * defer processing. If this is the end of
2356 * preimage, more "+" lines may come after it.
2357 */
2358 return 0;
2359 }
2360 diff_words_flush(ecbdata);
2361 emit_diff_symbol(o, s, line, len, 0);
2362 return 0;
2363 }
2364
2365 switch (line[0]) {
2366 case '+':
2367 ecbdata->lno_in_postimage++;
2368 emit_add_line(ecbdata, line + 1, len - 1);
2369 break;
2370 case '-':
2371 ecbdata->lno_in_preimage++;
2372 emit_del_line(ecbdata, line + 1, len - 1);
2373 break;
2374 case ' ':
2375 ecbdata->lno_in_postimage++;
2376 ecbdata->lno_in_preimage++;
2377 emit_context_line(ecbdata, line + 1, len - 1);
2378 break;
2379 default:
2380 /* incomplete line at the end */
2381 ecbdata->lno_in_preimage++;
2382 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2383 line, len, 0);
2384 break;
2385 }
2386 return 0;
2387 }
2388
2389 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2390 {
2391 const char *old_name = a;
2392 const char *new_name = b;
2393 int pfx_length, sfx_length;
2394 int pfx_adjust_for_slash;
2395 int len_a = strlen(a);
2396 int len_b = strlen(b);
2397 int a_midlen, b_midlen;
2398 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2399 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2400
2401 if (qlen_a || qlen_b) {
2402 quote_c_style(a, name, NULL, 0);
2403 strbuf_addstr(name, " => ");
2404 quote_c_style(b, name, NULL, 0);
2405 return;
2406 }
2407
2408 /* Find common prefix */
2409 pfx_length = 0;
2410 while (*old_name && *new_name && *old_name == *new_name) {
2411 if (*old_name == '/')
2412 pfx_length = old_name - a + 1;
2413 old_name++;
2414 new_name++;
2415 }
2416
2417 /* Find common suffix */
2418 old_name = a + len_a;
2419 new_name = b + len_b;
2420 sfx_length = 0;
2421 /*
2422 * If there is a common prefix, it must end in a slash. In
2423 * that case we let this loop run 1 into the prefix to see the
2424 * same slash.
2425 *
2426 * If there is no common prefix, we cannot do this as it would
2427 * underrun the input strings.
2428 */
2429 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2430 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2431 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2432 *old_name == *new_name) {
2433 if (*old_name == '/')
2434 sfx_length = len_a - (old_name - a);
2435 old_name--;
2436 new_name--;
2437 }
2438
2439 /*
2440 * pfx{mid-a => mid-b}sfx
2441 * {pfx-a => pfx-b}sfx
2442 * pfx{sfx-a => sfx-b}
2443 * name-a => name-b
2444 */
2445 a_midlen = len_a - pfx_length - sfx_length;
2446 b_midlen = len_b - pfx_length - sfx_length;
2447 if (a_midlen < 0)
2448 a_midlen = 0;
2449 if (b_midlen < 0)
2450 b_midlen = 0;
2451
2452 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2453 if (pfx_length + sfx_length) {
2454 strbuf_add(name, a, pfx_length);
2455 strbuf_addch(name, '{');
2456 }
2457 strbuf_add(name, a + pfx_length, a_midlen);
2458 strbuf_addstr(name, " => ");
2459 strbuf_add(name, b + pfx_length, b_midlen);
2460 if (pfx_length + sfx_length) {
2461 strbuf_addch(name, '}');
2462 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2463 }
2464 }
2465
2466 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2467 const char *name_a,
2468 const char *name_b)
2469 {
2470 struct diffstat_file *x;
2471 CALLOC_ARRAY(x, 1);
2472 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2473 diffstat->files[diffstat->nr++] = x;
2474 if (name_b) {
2475 x->from_name = xstrdup(name_a);
2476 x->name = xstrdup(name_b);
2477 x->is_renamed = 1;
2478 }
2479 else {
2480 x->from_name = NULL;
2481 x->name = xstrdup(name_a);
2482 }
2483 return x;
2484 }
2485
2486 static int diffstat_consume(void *priv, char *line, unsigned long len)
2487 {
2488 struct diffstat_t *diffstat = priv;
2489 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2490
2491 if (!len)
2492 BUG("xdiff fed us an empty line");
2493
2494 if (line[0] == '+')
2495 x->added++;
2496 else if (line[0] == '-')
2497 x->deleted++;
2498 return 0;
2499 }
2500
2501 const char mime_boundary_leader[] = "------------";
2502
2503 static int scale_linear(int it, int width, int max_change)
2504 {
2505 if (!it)
2506 return 0;
2507 /*
2508 * make sure that at least one '-' or '+' is printed if
2509 * there is any change to this path. The easiest way is to
2510 * scale linearly as if the allotted width is one column shorter
2511 * than it is, and then add 1 to the result.
2512 */
2513 return 1 + (it * (width - 1) / max_change);
2514 }
2515
2516 static void show_graph(struct strbuf *out, char ch, int cnt,
2517 const char *set, const char *reset)
2518 {
2519 if (cnt <= 0)
2520 return;
2521 strbuf_addstr(out, set);
2522 strbuf_addchars(out, ch, cnt);
2523 strbuf_addstr(out, reset);
2524 }
2525
2526 static void fill_print_name(struct diffstat_file *file)
2527 {
2528 struct strbuf pname = STRBUF_INIT;
2529
2530 if (file->print_name)
2531 return;
2532
2533 if (file->is_renamed)
2534 pprint_rename(&pname, file->from_name, file->name);
2535 else
2536 quote_c_style(file->name, &pname, NULL, 0);
2537
2538 if (file->comments)
2539 strbuf_addf(&pname, " (%s)", file->comments);
2540
2541 file->print_name = strbuf_detach(&pname, NULL);
2542 }
2543
2544 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2545 int files, int insertions, int deletions)
2546 {
2547 struct strbuf sb = STRBUF_INIT;
2548
2549 if (!files) {
2550 assert(insertions == 0 && deletions == 0);
2551 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2552 NULL, 0, 0);
2553 return;
2554 }
2555
2556 strbuf_addf(&sb,
2557 (files == 1) ? " %d file changed" : " %d files changed",
2558 files);
2559
2560 /*
2561 * For binary diff, the caller may want to print "x files
2562 * changed" with insertions == 0 && deletions == 0.
2563 *
2564 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2565 * is probably less confusing (i.e skip over "2 files changed
2566 * but nothing about added/removed lines? Is this a bug in Git?").
2567 */
2568 if (insertions || deletions == 0) {
2569 strbuf_addf(&sb,
2570 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2571 insertions);
2572 }
2573
2574 if (deletions || insertions == 0) {
2575 strbuf_addf(&sb,
2576 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2577 deletions);
2578 }
2579 strbuf_addch(&sb, '\n');
2580 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2581 sb.buf, sb.len, 0);
2582 strbuf_release(&sb);
2583 }
2584
2585 void print_stat_summary(FILE *fp, int files,
2586 int insertions, int deletions)
2587 {
2588 struct diff_options o;
2589 memset(&o, 0, sizeof(o));
2590 o.file = fp;
2591
2592 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2593 }
2594
2595 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2596 {
2597 int i, len, add, del, adds = 0, dels = 0;
2598 uintmax_t max_change = 0, max_len = 0;
2599 int total_files = data->nr, count;
2600 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2601 const char *reset, *add_c, *del_c;
2602 int extra_shown = 0;
2603 const char *line_prefix = diff_line_prefix(options);
2604 struct strbuf out = STRBUF_INIT;
2605
2606 if (data->nr == 0)
2607 return;
2608
2609 count = options->stat_count ? options->stat_count : data->nr;
2610
2611 reset = diff_get_color_opt(options, DIFF_RESET);
2612 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2613 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2614
2615 /*
2616 * Find the longest filename and max number of changes
2617 */
2618 for (i = 0; (i < count) && (i < data->nr); i++) {
2619 struct diffstat_file *file = data->files[i];
2620 uintmax_t change = file->added + file->deleted;
2621
2622 if (!file->is_interesting && (change == 0)) {
2623 count++; /* not shown == room for one more */
2624 continue;
2625 }
2626 fill_print_name(file);
2627 len = utf8_strwidth(file->print_name);
2628 if (max_len < len)
2629 max_len = len;
2630
2631 if (file->is_unmerged) {
2632 /* "Unmerged" is 8 characters */
2633 bin_width = bin_width < 8 ? 8 : bin_width;
2634 continue;
2635 }
2636 if (file->is_binary) {
2637 /* "Bin XXX -> YYY bytes" */
2638 int w = 14 + decimal_width(file->added)
2639 + decimal_width(file->deleted);
2640 bin_width = bin_width < w ? w : bin_width;
2641 /* Display change counts aligned with "Bin" */
2642 number_width = 3;
2643 continue;
2644 }
2645
2646 if (max_change < change)
2647 max_change = change;
2648 }
2649 count = i; /* where we can stop scanning in data->files[] */
2650
2651 /*
2652 * We have width = stat_width or term_columns() columns total.
2653 * We want a maximum of min(max_len, stat_name_width) for the name part.
2654 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2655 * We also need 1 for " " and 4 + decimal_width(max_change)
2656 * for " | NNNN " and one the empty column at the end, altogether
2657 * 6 + decimal_width(max_change).
2658 *
2659 * If there's not enough space, we will use the smaller of
2660 * stat_name_width (if set) and 5/8*width for the filename,
2661 * and the rest for constant elements + graph part, but no more
2662 * than stat_graph_width for the graph part.
2663 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2664 * for the standard terminal size).
2665 *
2666 * In other words: stat_width limits the maximum width, and
2667 * stat_name_width fixes the maximum width of the filename,
2668 * and is also used to divide available columns if there
2669 * aren't enough.
2670 *
2671 * Binary files are displayed with "Bin XXX -> YYY bytes"
2672 * instead of the change count and graph. This part is treated
2673 * similarly to the graph part, except that it is not
2674 * "scaled". If total width is too small to accommodate the
2675 * guaranteed minimum width of the filename part and the
2676 * separators and this message, this message will "overflow"
2677 * making the line longer than the maximum width.
2678 */
2679
2680 /*
2681 * NEEDSWORK: line_prefix is often used for "log --graph" output
2682 * and contains ANSI-colored string. utf8_strnwidth() should be
2683 * used to correctly count the display width instead of strlen().
2684 */
2685 if (options->stat_width == -1)
2686 width = term_columns() - strlen(line_prefix);
2687 else
2688 width = options->stat_width ? options->stat_width : 80;
2689 number_width = decimal_width(max_change) > number_width ?
2690 decimal_width(max_change) : number_width;
2691
2692 if (options->stat_graph_width == -1)
2693 options->stat_graph_width = diff_stat_graph_width;
2694
2695 /*
2696 * Guarantee 3/8*16==6 for the graph part
2697 * and 5/8*16==10 for the filename part
2698 */
2699 if (width < 16 + 6 + number_width)
2700 width = 16 + 6 + number_width;
2701
2702 /*
2703 * First assign sizes that are wanted, ignoring available width.
2704 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2705 * starting from "XXX" should fit in graph_width.
2706 */
2707 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2708 if (options->stat_graph_width &&
2709 options->stat_graph_width < graph_width)
2710 graph_width = options->stat_graph_width;
2711
2712 name_width = (options->stat_name_width > 0 &&
2713 options->stat_name_width < max_len) ?
2714 options->stat_name_width : max_len;
2715
2716 /*
2717 * Adjust adjustable widths not to exceed maximum width
2718 */
2719 if (name_width + number_width + 6 + graph_width > width) {
2720 if (graph_width > width * 3/8 - number_width - 6) {
2721 graph_width = width * 3/8 - number_width - 6;
2722 if (graph_width < 6)
2723 graph_width = 6;
2724 }
2725
2726 if (options->stat_graph_width &&
2727 graph_width > options->stat_graph_width)
2728 graph_width = options->stat_graph_width;
2729 if (name_width > width - number_width - 6 - graph_width)
2730 name_width = width - number_width - 6 - graph_width;
2731 else
2732 graph_width = width - number_width - 6 - name_width;
2733 }
2734
2735 /*
2736 * From here name_width is the width of the name area,
2737 * and graph_width is the width of the graph area.
2738 * max_change is used to scale graph properly.
2739 */
2740 for (i = 0; i < count; i++) {
2741 const char *prefix = "";
2742 struct diffstat_file *file = data->files[i];
2743 char *name = file->print_name;
2744 uintmax_t added = file->added;
2745 uintmax_t deleted = file->deleted;
2746 int name_len, padding;
2747
2748 if (!file->is_interesting && (added + deleted == 0))
2749 continue;
2750
2751 /*
2752 * "scale" the filename
2753 */
2754 len = name_width;
2755 name_len = utf8_strwidth(name);
2756 if (name_width < name_len) {
2757 char *slash;
2758 prefix = "...";
2759 len -= 3;
2760 /*
2761 * NEEDSWORK: (name_len - len) counts the display
2762 * width, which would be shorter than the byte
2763 * length of the corresponding substring.
2764 * Advancing "name" by that number of bytes does
2765 * *NOT* skip over that many columns, so it is
2766 * very likely that chomping the pathname at the
2767 * slash we will find starting from "name" will
2768 * leave the resulting string still too long.
2769 */
2770 name += name_len - len;
2771 slash = strchr(name, '/');
2772 if (slash)
2773 name = slash;
2774 }
2775 padding = len - utf8_strwidth(name);
2776 if (padding < 0)
2777 padding = 0;
2778
2779 if (file->is_binary) {
2780 strbuf_addf(&out, " %s%s%*s | %*s",
2781 prefix, name, padding, "",
2782 number_width, "Bin");
2783 if (!added && !deleted) {
2784 strbuf_addch(&out, '\n');
2785 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2786 out.buf, out.len, 0);
2787 strbuf_reset(&out);
2788 continue;
2789 }
2790 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2791 del_c, deleted, reset);
2792 strbuf_addstr(&out, " -> ");
2793 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2794 add_c, added, reset);
2795 strbuf_addstr(&out, " bytes\n");
2796 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2797 out.buf, out.len, 0);
2798 strbuf_reset(&out);
2799 continue;
2800 }
2801 else if (file->is_unmerged) {
2802 strbuf_addf(&out, " %s%s%*s | %*s",
2803 prefix, name, padding, "",
2804 number_width, "Unmerged");
2805 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2806 out.buf, out.len, 0);
2807 strbuf_reset(&out);
2808 continue;
2809 }
2810
2811 /*
2812 * scale the add/delete
2813 */
2814 add = added;
2815 del = deleted;
2816
2817 if (graph_width <= max_change) {
2818 int total = scale_linear(add + del, graph_width, max_change);
2819 if (total < 2 && add && del)
2820 /* width >= 2 due to the sanity check */
2821 total = 2;
2822 if (add < del) {
2823 add = scale_linear(add, graph_width, max_change);
2824 del = total - add;
2825 } else {
2826 del = scale_linear(del, graph_width, max_change);
2827 add = total - del;
2828 }
2829 }
2830 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2831 prefix, name, padding, "",
2832 number_width, added + deleted,
2833 added + deleted ? " " : "");
2834 show_graph(&out, '+', add, add_c, reset);
2835 show_graph(&out, '-', del, del_c, reset);
2836 strbuf_addch(&out, '\n');
2837 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2838 out.buf, out.len, 0);
2839 strbuf_reset(&out);
2840 }
2841
2842 for (i = 0; i < data->nr; i++) {
2843 struct diffstat_file *file = data->files[i];
2844 uintmax_t added = file->added;
2845 uintmax_t deleted = file->deleted;
2846
2847 if (file->is_unmerged ||
2848 (!file->is_interesting && (added + deleted == 0))) {
2849 total_files--;
2850 continue;
2851 }
2852
2853 if (!file->is_binary) {
2854 adds += added;
2855 dels += deleted;
2856 }
2857 if (i < count)
2858 continue;
2859 if (!extra_shown)
2860 emit_diff_symbol(options,
2861 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2862 NULL, 0, 0);
2863 extra_shown = 1;
2864 }
2865
2866 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2867 strbuf_release(&out);
2868 }
2869
2870 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2871 {
2872 int i, adds = 0, dels = 0, total_files = data->nr;
2873
2874 if (data->nr == 0)
2875 return;
2876
2877 for (i = 0; i < data->nr; i++) {
2878 int added = data->files[i]->added;
2879 int deleted = data->files[i]->deleted;
2880
2881 if (data->files[i]->is_unmerged ||
2882 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2883 total_files--;
2884 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2885 adds += added;
2886 dels += deleted;
2887 }
2888 }
2889 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2890 }
2891
2892 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2893 {
2894 int i;
2895
2896 if (data->nr == 0)
2897 return;
2898
2899 for (i = 0; i < data->nr; i++) {
2900 struct diffstat_file *file = data->files[i];
2901
2902 fprintf(options->file, "%s", diff_line_prefix(options));
2903
2904 if (file->is_binary)
2905 fprintf(options->file, "-\t-\t");
2906 else
2907 fprintf(options->file,
2908 "%"PRIuMAX"\t%"PRIuMAX"\t",
2909 file->added, file->deleted);
2910 if (options->line_termination) {
2911 fill_print_name(file);
2912 if (!file->is_renamed)
2913 write_name_quoted(file->name, options->file,
2914 options->line_termination);
2915 else {
2916 fputs(file->print_name, options->file);
2917 putc(options->line_termination, options->file);
2918 }
2919 } else {
2920 if (file->is_renamed) {
2921 putc('\0', options->file);
2922 write_name_quoted(file->from_name, options->file, '\0');
2923 }
2924 write_name_quoted(file->name, options->file, '\0');
2925 }
2926 }
2927 }
2928
2929 struct dirstat_file {
2930 const char *name;
2931 unsigned long changed;
2932 };
2933
2934 struct dirstat_dir {
2935 struct dirstat_file *files;
2936 int alloc, nr, permille, cumulative;
2937 };
2938
2939 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2940 unsigned long changed, const char *base, int baselen)
2941 {
2942 unsigned long sum_changes = 0;
2943 unsigned int sources = 0;
2944 const char *line_prefix = diff_line_prefix(opt);
2945
2946 while (dir->nr) {
2947 struct dirstat_file *f = dir->files;
2948 int namelen = strlen(f->name);
2949 unsigned long changes;
2950 char *slash;
2951
2952 if (namelen < baselen)
2953 break;
2954 if (memcmp(f->name, base, baselen))
2955 break;
2956 slash = strchr(f->name + baselen, '/');
2957 if (slash) {
2958 int newbaselen = slash + 1 - f->name;
2959 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2960 sources++;
2961 } else {
2962 changes = f->changed;
2963 dir->files++;
2964 dir->nr--;
2965 sources += 2;
2966 }
2967 sum_changes += changes;
2968 }
2969
2970 /*
2971 * We don't report dirstat's for
2972 * - the top level
2973 * - or cases where everything came from a single directory
2974 * under this directory (sources == 1).
2975 */
2976 if (baselen && sources != 1) {
2977 if (sum_changes) {
2978 int permille = sum_changes * 1000 / changed;
2979 if (permille >= dir->permille) {
2980 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2981 permille / 10, permille % 10, baselen, base);
2982 if (!dir->cumulative)
2983 return 0;
2984 }
2985 }
2986 }
2987 return sum_changes;
2988 }
2989
2990 static int dirstat_compare(const void *_a, const void *_b)
2991 {
2992 const struct dirstat_file *a = _a;
2993 const struct dirstat_file *b = _b;
2994 return strcmp(a->name, b->name);
2995 }
2996
2997 static void show_dirstat(struct diff_options *options)
2998 {
2999 int i;
3000 unsigned long changed;
3001 struct dirstat_dir dir;
3002 struct diff_queue_struct *q = &diff_queued_diff;
3003
3004 dir.files = NULL;
3005 dir.alloc = 0;
3006 dir.nr = 0;
3007 dir.permille = options->dirstat_permille;
3008 dir.cumulative = options->flags.dirstat_cumulative;
3009
3010 changed = 0;
3011 for (i = 0; i < q->nr; i++) {
3012 struct diff_filepair *p = q->queue[i];
3013 const char *name;
3014 unsigned long copied, added, damage;
3015 struct diff_populate_filespec_options dpf_options = {
3016 .check_size_only = 1,
3017 };
3018
3019 name = p->two->path ? p->two->path : p->one->path;
3020
3021 if (p->one->oid_valid && p->two->oid_valid &&
3022 oideq(&p->one->oid, &p->two->oid)) {
3023 /*
3024 * The SHA1 has not changed, so pre-/post-content is
3025 * identical. We can therefore skip looking at the
3026 * file contents altogether.
3027 */
3028 damage = 0;
3029 goto found_damage;
3030 }
3031
3032 if (options->flags.dirstat_by_file) {
3033 /*
3034 * In --dirstat-by-file mode, we don't really need to
3035 * look at the actual file contents at all.
3036 * The fact that the SHA1 changed is enough for us to
3037 * add this file to the list of results
3038 * (with each file contributing equal damage).
3039 */
3040 damage = 1;
3041 goto found_damage;
3042 }
3043
3044 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3045 diff_populate_filespec(options->repo, p->one, NULL);
3046 diff_populate_filespec(options->repo, p->two, NULL);
3047 diffcore_count_changes(options->repo,
3048 p->one, p->two, NULL, NULL,
3049 &copied, &added);
3050 diff_free_filespec_data(p->one);
3051 diff_free_filespec_data(p->two);
3052 } else if (DIFF_FILE_VALID(p->one)) {
3053 diff_populate_filespec(options->repo, p->one, &dpf_options);
3054 copied = added = 0;
3055 diff_free_filespec_data(p->one);
3056 } else if (DIFF_FILE_VALID(p->two)) {
3057 diff_populate_filespec(options->repo, p->two, &dpf_options);
3058 copied = 0;
3059 added = p->two->size;
3060 diff_free_filespec_data(p->two);
3061 } else
3062 continue;
3063
3064 /*
3065 * Original minus copied is the removed material,
3066 * added is the new material. They are both damages
3067 * made to the preimage.
3068 * If the resulting damage is zero, we know that
3069 * diffcore_count_changes() considers the two entries to
3070 * be identical, but since the oid changed, we
3071 * know that there must have been _some_ kind of change,
3072 * so we force all entries to have damage > 0.
3073 */
3074 damage = (p->one->size - copied) + added;
3075 if (!damage)
3076 damage = 1;
3077
3078 found_damage:
3079 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3080 dir.files[dir.nr].name = name;
3081 dir.files[dir.nr].changed = damage;
3082 changed += damage;
3083 dir.nr++;
3084 }
3085
3086 /* This can happen even with many files, if everything was renames */
3087 if (!changed)
3088 return;
3089
3090 /* Show all directories with more than x% of the changes */
3091 QSORT(dir.files, dir.nr, dirstat_compare);
3092 gather_dirstat(options, &dir, changed, "", 0);
3093 }
3094
3095 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3096 {
3097 int i;
3098 unsigned long changed;
3099 struct dirstat_dir dir;
3100
3101 if (data->nr == 0)
3102 return;
3103
3104 dir.files = NULL;
3105 dir.alloc = 0;
3106 dir.nr = 0;
3107 dir.permille = options->dirstat_permille;
3108 dir.cumulative = options->flags.dirstat_cumulative;
3109
3110 changed = 0;
3111 for (i = 0; i < data->nr; i++) {
3112 struct diffstat_file *file = data->files[i];
3113 unsigned long damage = file->added + file->deleted;
3114 if (file->is_binary)
3115 /*
3116 * binary files counts bytes, not lines. Must find some
3117 * way to normalize binary bytes vs. textual lines.
3118 * The following heuristic assumes that there are 64
3119 * bytes per "line".
3120 * This is stupid and ugly, but very cheap...
3121 */
3122 damage = DIV_ROUND_UP(damage, 64);
3123 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3124 dir.files[dir.nr].name = file->name;
3125 dir.files[dir.nr].changed = damage;
3126 changed += damage;
3127 dir.nr++;
3128 }
3129
3130 /* This can happen even with many files, if everything was renames */
3131 if (!changed)
3132 return;
3133
3134 /* Show all directories with more than x% of the changes */
3135 QSORT(dir.files, dir.nr, dirstat_compare);
3136 gather_dirstat(options, &dir, changed, "", 0);
3137 }
3138
3139 static void free_diffstat_file(struct diffstat_file *f)
3140 {
3141 free(f->print_name);
3142 free(f->name);
3143 free(f->from_name);
3144 free(f);
3145 }
3146
3147 void free_diffstat_info(struct diffstat_t *diffstat)
3148 {
3149 int i;
3150 for (i = 0; i < diffstat->nr; i++)
3151 free_diffstat_file(diffstat->files[i]);
3152 free(diffstat->files);
3153 }
3154
3155 struct checkdiff_t {
3156 const char *filename;
3157 int lineno;
3158 int conflict_marker_size;
3159 struct diff_options *o;
3160 unsigned ws_rule;
3161 unsigned status;
3162 };
3163
3164 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3165 {
3166 char firstchar;
3167 int cnt;
3168
3169 if (len < marker_size + 1)
3170 return 0;
3171 firstchar = line[0];
3172 switch (firstchar) {
3173 case '=': case '>': case '<': case '|':
3174 break;
3175 default:
3176 return 0;
3177 }
3178 for (cnt = 1; cnt < marker_size; cnt++)
3179 if (line[cnt] != firstchar)
3180 return 0;
3181 /* line[1] through line[marker_size-1] are same as firstchar */
3182 if (len < marker_size + 1 || !isspace(line[marker_size]))
3183 return 0;
3184 return 1;
3185 }
3186
3187 static void checkdiff_consume_hunk(void *priv,
3188 long ob, long on, long nb, long nn,
3189 const char *func, long funclen)
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 void builtin_diff(const char *name_a,
3441 const char *name_b,
3442 struct diff_filespec *one,
3443 struct diff_filespec *two,
3444 const char *xfrm_msg,
3445 int must_show_header,
3446 struct diff_options *o,
3447 int complete_rewrite)
3448 {
3449 mmfile_t mf1, mf2;
3450 const char *lbl[2];
3451 char *a_one, *b_two;
3452 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3453 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3454 const char *a_prefix, *b_prefix;
3455 struct userdiff_driver *textconv_one = NULL;
3456 struct userdiff_driver *textconv_two = NULL;
3457 struct strbuf header = STRBUF_INIT;
3458 const char *line_prefix = diff_line_prefix(o);
3459
3460 diff_set_mnemonic_prefix(o, "a/", "b/");
3461 if (o->flags.reverse_diff) {
3462 a_prefix = o->b_prefix;
3463 b_prefix = o->a_prefix;
3464 } else {
3465 a_prefix = o->a_prefix;
3466 b_prefix = o->b_prefix;
3467 }
3468
3469 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3470 (!one->mode || S_ISGITLINK(one->mode)) &&
3471 (!two->mode || S_ISGITLINK(two->mode)) &&
3472 (!diff_filepair_is_phoney(one, two))) {
3473 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3474 &one->oid, &two->oid,
3475 two->dirty_submodule);
3476 return;
3477 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3478 (!one->mode || S_ISGITLINK(one->mode)) &&
3479 (!two->mode || S_ISGITLINK(two->mode)) &&
3480 (!diff_filepair_is_phoney(one, two))) {
3481 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3482 &one->oid, &two->oid,
3483 two->dirty_submodule);
3484 return;
3485 }
3486
3487 if (o->flags.allow_textconv) {
3488 textconv_one = get_textconv(o->repo, one);
3489 textconv_two = get_textconv(o->repo, two);
3490 }
3491
3492 /* Never use a non-valid filename anywhere if at all possible */
3493 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3494 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3495
3496 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3497 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3498 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3499 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3500 if (diff_filepair_is_phoney(one, two)) {
3501 /*
3502 * We should only reach this point for pairs generated from
3503 * create_filepairs_for_header_only_notifications(). For
3504 * these, we want to avoid the "/dev/null" special casing
3505 * above, because we do not want such pairs shown as either
3506 * "new file" or "deleted file" below.
3507 */
3508 lbl[0] = a_one;
3509 lbl[1] = b_two;
3510 }
3511 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3512 if (lbl[0][0] == '/') {
3513 /* /dev/null */
3514 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3515 if (xfrm_msg)
3516 strbuf_addstr(&header, xfrm_msg);
3517 must_show_header = 1;
3518 }
3519 else if (lbl[1][0] == '/') {
3520 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3521 if (xfrm_msg)
3522 strbuf_addstr(&header, xfrm_msg);
3523 must_show_header = 1;
3524 }
3525 else {
3526 if (one->mode != two->mode) {
3527 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3528 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3529 must_show_header = 1;
3530 }
3531 if (xfrm_msg)
3532 strbuf_addstr(&header, xfrm_msg);
3533
3534 /*
3535 * we do not run diff between different kind
3536 * of objects.
3537 */
3538 if ((one->mode ^ two->mode) & S_IFMT)
3539 goto free_ab_and_return;
3540 if (complete_rewrite &&
3541 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3542 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3543 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3544 header.buf, header.len, 0);
3545 strbuf_reset(&header);
3546 emit_rewrite_diff(name_a, name_b, one, two,
3547 textconv_one, textconv_two, o);
3548 o->found_changes = 1;
3549 goto free_ab_and_return;
3550 }
3551 }
3552
3553 if (o->irreversible_delete && lbl[1][0] == '/') {
3554 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3555 header.len, 0);
3556 strbuf_reset(&header);
3557 goto free_ab_and_return;
3558 } else if (!o->flags.text &&
3559 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3560 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3561 struct strbuf sb = STRBUF_INIT;
3562 if (!one->data && !two->data &&
3563 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3564 !o->flags.binary) {
3565 if (oideq(&one->oid, &two->oid)) {
3566 if (must_show_header)
3567 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3568 header.buf, header.len,
3569 0);
3570 goto free_ab_and_return;
3571 }
3572 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3573 header.buf, header.len, 0);
3574 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3575 diff_line_prefix(o), lbl[0], lbl[1]);
3576 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3577 sb.buf, sb.len, 0);
3578 strbuf_release(&sb);
3579 goto free_ab_and_return;
3580 }
3581 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3582 fill_mmfile(o->repo, &mf2, two) < 0)
3583 die("unable to read files to diff");
3584 /* Quite common confusing case */
3585 if (mf1.size == mf2.size &&
3586 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3587 if (must_show_header)
3588 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3589 header.buf, header.len, 0);
3590 goto free_ab_and_return;
3591 }
3592 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3593 strbuf_reset(&header);
3594 if (o->flags.binary)
3595 emit_binary_diff(o, &mf1, &mf2);
3596 else {
3597 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3598 diff_line_prefix(o), lbl[0], lbl[1]);
3599 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3600 sb.buf, sb.len, 0);
3601 strbuf_release(&sb);
3602 }
3603 o->found_changes = 1;
3604 } else {
3605 /* Crazy xdl interfaces.. */
3606 const char *diffopts;
3607 const char *v;
3608 xpparam_t xpp;
3609 xdemitconf_t xecfg;
3610 struct emit_callback ecbdata;
3611 const struct userdiff_funcname *pe;
3612
3613 if (must_show_header) {
3614 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3615 header.buf, header.len, 0);
3616 strbuf_reset(&header);
3617 }
3618
3619 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3620 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3621
3622 pe = diff_funcname_pattern(o, one);
3623 if (!pe)
3624 pe = diff_funcname_pattern(o, two);
3625
3626 memset(&xpp, 0, sizeof(xpp));
3627 memset(&xecfg, 0, sizeof(xecfg));
3628 memset(&ecbdata, 0, sizeof(ecbdata));
3629 if (o->flags.suppress_diff_headers)
3630 lbl[0] = NULL;
3631 ecbdata.label_path = lbl;
3632 ecbdata.color_diff = want_color(o->use_color);
3633 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3634 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3635 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3636 ecbdata.opt = o;
3637 if (header.len && !o->flags.suppress_diff_headers)
3638 ecbdata.header = &header;
3639 xpp.flags = o->xdl_opts;
3640 xpp.ignore_regex = o->ignore_regex;
3641 xpp.ignore_regex_nr = o->ignore_regex_nr;
3642 xpp.anchors = o->anchors;
3643 xpp.anchors_nr = o->anchors_nr;
3644 xecfg.ctxlen = o->context;
3645 xecfg.interhunkctxlen = o->interhunkcontext;
3646 xecfg.flags = XDL_EMIT_FUNCNAMES;
3647 if (o->flags.funccontext)
3648 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3649 if (pe)
3650 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3651
3652 diffopts = getenv("GIT_DIFF_OPTS");
3653 if (!diffopts)
3654 ;
3655 else if (skip_prefix(diffopts, "--unified=", &v))
3656 xecfg.ctxlen = strtoul(v, NULL, 10);
3657 else if (skip_prefix(diffopts, "-u", &v))
3658 xecfg.ctxlen = strtoul(v, NULL, 10);
3659
3660 if (o->word_diff)
3661 init_diff_words_data(&ecbdata, o, one, two);
3662 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3663 &ecbdata, &xpp, &xecfg))
3664 die("unable to generate diff for %s", one->path);
3665 if (o->word_diff)
3666 free_diff_words_data(&ecbdata);
3667 if (textconv_one)
3668 free(mf1.ptr);
3669 if (textconv_two)
3670 free(mf2.ptr);
3671 xdiff_clear_find_func(&xecfg);
3672 }
3673
3674 free_ab_and_return:
3675 strbuf_release(&header);
3676 diff_free_filespec_data(one);
3677 diff_free_filespec_data(two);
3678 free(a_one);
3679 free(b_two);
3680 return;
3681 }
3682
3683 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3684 {
3685 if (!is_renamed) {
3686 if (p->status == DIFF_STATUS_ADDED) {
3687 if (S_ISLNK(p->two->mode))
3688 return "new +l";
3689 else if ((p->two->mode & 0777) == 0755)
3690 return "new +x";
3691 else
3692 return "new";
3693 } else if (p->status == DIFF_STATUS_DELETED)
3694 return "gone";
3695 }
3696 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3697 return "mode -l";
3698 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3699 return "mode +l";
3700 else if ((p->one->mode & 0777) == 0644 &&
3701 (p->two->mode & 0777) == 0755)
3702 return "mode +x";
3703 else if ((p->one->mode & 0777) == 0755 &&
3704 (p->two->mode & 0777) == 0644)
3705 return "mode -x";
3706 return NULL;
3707 }
3708
3709 static void builtin_diffstat(const char *name_a, const char *name_b,
3710 struct diff_filespec *one,
3711 struct diff_filespec *two,
3712 struct diffstat_t *diffstat,
3713 struct diff_options *o,
3714 struct diff_filepair *p)
3715 {
3716 mmfile_t mf1, mf2;
3717 struct diffstat_file *data;
3718 int may_differ;
3719 int complete_rewrite = 0;
3720
3721 if (!DIFF_PAIR_UNMERGED(p)) {
3722 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3723 complete_rewrite = 1;
3724 }
3725
3726 data = diffstat_add(diffstat, name_a, name_b);
3727 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3728 if (o->flags.stat_with_summary)
3729 data->comments = get_compact_summary(p, data->is_renamed);
3730
3731 if (!one || !two) {
3732 data->is_unmerged = 1;
3733 return;
3734 }
3735
3736 /* saves some reads if true, not a guarantee of diff outcome */
3737 may_differ = !(one->oid_valid && two->oid_valid &&
3738 oideq(&one->oid, &two->oid));
3739
3740 if (diff_filespec_is_binary(o->repo, one) ||
3741 diff_filespec_is_binary(o->repo, two)) {
3742 data->is_binary = 1;
3743 if (!may_differ) {
3744 data->added = 0;
3745 data->deleted = 0;
3746 } else {
3747 data->added = diff_filespec_size(o->repo, two);
3748 data->deleted = diff_filespec_size(o->repo, one);
3749 }
3750 }
3751
3752 else if (complete_rewrite) {
3753 diff_populate_filespec(o->repo, one, NULL);
3754 diff_populate_filespec(o->repo, two, NULL);
3755 data->deleted = count_lines(one->data, one->size);
3756 data->added = count_lines(two->data, two->size);
3757 }
3758
3759 else if (may_differ) {
3760 /* Crazy xdl interfaces.. */
3761 xpparam_t xpp;
3762 xdemitconf_t xecfg;
3763
3764 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3765 fill_mmfile(o->repo, &mf2, two) < 0)
3766 die("unable to read files to diff");
3767
3768 memset(&xpp, 0, sizeof(xpp));
3769 memset(&xecfg, 0, sizeof(xecfg));
3770 xpp.flags = o->xdl_opts;
3771 xpp.ignore_regex = o->ignore_regex;
3772 xpp.ignore_regex_nr = o->ignore_regex_nr;
3773 xpp.anchors = o->anchors;
3774 xpp.anchors_nr = o->anchors_nr;
3775 xecfg.ctxlen = o->context;
3776 xecfg.interhunkctxlen = o->interhunkcontext;
3777 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3778 if (xdi_diff_outf(&mf1, &mf2, NULL,
3779 diffstat_consume, diffstat, &xpp, &xecfg))
3780 die("unable to generate diffstat for %s", one->path);
3781
3782 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3783 struct diffstat_file *file =
3784 diffstat->files[diffstat->nr - 1];
3785 /*
3786 * Omit diffstats of modified files where nothing changed.
3787 * Even if may_differ, this might be the case due to
3788 * ignoring whitespace changes, etc.
3789 *
3790 * But note that we special-case additions, deletions,
3791 * renames, and mode changes as adding an empty file,
3792 * for example is still of interest.
3793 */
3794 if ((p->status == DIFF_STATUS_MODIFIED)
3795 && !file->added
3796 && !file->deleted
3797 && one->mode == two->mode) {
3798 free_diffstat_file(file);
3799 diffstat->nr--;
3800 }
3801 }
3802 }
3803
3804 diff_free_filespec_data(one);
3805 diff_free_filespec_data(two);
3806 }
3807
3808 static void builtin_checkdiff(const char *name_a, const char *name_b,
3809 const char *attr_path,
3810 struct diff_filespec *one,
3811 struct diff_filespec *two,
3812 struct diff_options *o)
3813 {
3814 mmfile_t mf1, mf2;
3815 struct checkdiff_t data;
3816
3817 if (!two)
3818 return;
3819
3820 memset(&data, 0, sizeof(data));
3821 data.filename = name_b ? name_b : name_a;
3822 data.lineno = 0;
3823 data.o = o;
3824 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3825 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3826
3827 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3828 fill_mmfile(o->repo, &mf2, two) < 0)
3829 die("unable to read files to diff");
3830
3831 /*
3832 * All the other codepaths check both sides, but not checking
3833 * the "old" side here is deliberate. We are checking the newly
3834 * introduced changes, and as long as the "new" side is text, we
3835 * can and should check what it introduces.
3836 */
3837 if (diff_filespec_is_binary(o->repo, two))
3838 goto free_and_return;
3839 else {
3840 /* Crazy xdl interfaces.. */
3841 xpparam_t xpp;
3842 xdemitconf_t xecfg;
3843
3844 memset(&xpp, 0, sizeof(xpp));
3845 memset(&xecfg, 0, sizeof(xecfg));
3846 xecfg.ctxlen = 1; /* at least one context line */
3847 xpp.flags = 0;
3848 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3849 checkdiff_consume, &data,
3850 &xpp, &xecfg))
3851 die("unable to generate checkdiff for %s", one->path);
3852
3853 if (data.ws_rule & WS_BLANK_AT_EOF) {
3854 struct emit_callback ecbdata;
3855 int blank_at_eof;
3856
3857 ecbdata.ws_rule = data.ws_rule;
3858 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3859 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3860
3861 if (blank_at_eof) {
3862 static char *err;
3863 if (!err)
3864 err = whitespace_error_string(WS_BLANK_AT_EOF);
3865 fprintf(o->file, "%s:%d: %s.\n",
3866 data.filename, blank_at_eof, err);
3867 data.status = 1; /* report errors */
3868 }
3869 }
3870 }
3871 free_and_return:
3872 diff_free_filespec_data(one);
3873 diff_free_filespec_data(two);
3874 if (data.status)
3875 o->flags.check_failed = 1;
3876 }
3877
3878 struct diff_filespec *alloc_filespec(const char *path)
3879 {
3880 struct diff_filespec *spec;
3881
3882 FLEXPTR_ALLOC_STR(spec, path, path);
3883 spec->count = 1;
3884 spec->is_binary = -1;
3885 return spec;
3886 }
3887
3888 void free_filespec(struct diff_filespec *spec)
3889 {
3890 if (!--spec->count) {
3891 diff_free_filespec_data(spec);
3892 free(spec);
3893 }
3894 }
3895
3896 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3897 int oid_valid, unsigned short mode)
3898 {
3899 if (mode) {
3900 spec->mode = canon_mode(mode);
3901 oidcpy(&spec->oid, oid);
3902 spec->oid_valid = oid_valid;
3903 }
3904 }
3905
3906 /*
3907 * Given a name and sha1 pair, if the index tells us the file in
3908 * the work tree has that object contents, return true, so that
3909 * prepare_temp_file() does not have to inflate and extract.
3910 */
3911 static int reuse_worktree_file(struct index_state *istate,
3912 const char *name,
3913 const struct object_id *oid,
3914 int want_file)
3915 {
3916 const struct cache_entry *ce;
3917 struct stat st;
3918 int pos, len;
3919
3920 /*
3921 * We do not read the cache ourselves here, because the
3922 * benchmark with my previous version that always reads cache
3923 * shows that it makes things worse for diff-tree comparing
3924 * two linux-2.6 kernel trees in an already checked out work
3925 * tree. This is because most diff-tree comparisons deal with
3926 * only a small number of files, while reading the cache is
3927 * expensive for a large project, and its cost outweighs the
3928 * savings we get by not inflating the object to a temporary
3929 * file. Practically, this code only helps when we are used
3930 * by diff-cache --cached, which does read the cache before
3931 * calling us.
3932 */
3933 if (!istate->cache)
3934 return 0;
3935
3936 /* We want to avoid the working directory if our caller
3937 * doesn't need the data in a normal file, this system
3938 * is rather slow with its stat/open/mmap/close syscalls,
3939 * and the object is contained in a pack file. The pack
3940 * is probably already open and will be faster to obtain
3941 * the data through than the working directory. Loose
3942 * objects however would tend to be slower as they need
3943 * to be individually opened and inflated.
3944 */
3945 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3946 return 0;
3947
3948 /*
3949 * Similarly, if we'd have to convert the file contents anyway, that
3950 * makes the optimization not worthwhile.
3951 */
3952 if (!want_file && would_convert_to_git(istate, name))
3953 return 0;
3954
3955 /*
3956 * If this path does not match our sparse-checkout definition,
3957 * then the file will not be in the working directory.
3958 */
3959 if (!path_in_sparse_checkout(name, istate))
3960 return 0;
3961
3962 len = strlen(name);
3963 pos = index_name_pos(istate, name, len);
3964 if (pos < 0)
3965 return 0;
3966 ce = istate->cache[pos];
3967
3968 /*
3969 * This is not the sha1 we are looking for, or
3970 * unreusable because it is not a regular file.
3971 */
3972 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3973 return 0;
3974
3975 /*
3976 * If ce is marked as "assume unchanged", there is no
3977 * guarantee that work tree matches what we are looking for.
3978 */
3979 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3980 return 0;
3981
3982 /*
3983 * If ce matches the file in the work tree, we can reuse it.
3984 */
3985 if (ce_uptodate(ce) ||
3986 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
3987 return 1;
3988
3989 return 0;
3990 }
3991
3992 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3993 {
3994 struct strbuf buf = STRBUF_INIT;
3995 char *dirty = "";
3996
3997 /* Are we looking at the work tree? */
3998 if (s->dirty_submodule)
3999 dirty = "-dirty";
4000
4001 strbuf_addf(&buf, "Subproject commit %s%s\n",
4002 oid_to_hex(&s->oid), dirty);
4003 s->size = buf.len;
4004 if (size_only) {
4005 s->data = NULL;
4006 strbuf_release(&buf);
4007 } else {
4008 s->data = strbuf_detach(&buf, NULL);
4009 s->should_free = 1;
4010 }
4011 return 0;
4012 }
4013
4014 /*
4015 * While doing rename detection and pickaxe operation, we may need to
4016 * grab the data for the blob (or file) for our own in-core comparison.
4017 * diff_filespec has data and size fields for this purpose.
4018 */
4019 int diff_populate_filespec(struct repository *r,
4020 struct diff_filespec *s,
4021 const struct diff_populate_filespec_options *options)
4022 {
4023 int size_only = options ? options->check_size_only : 0;
4024 int check_binary = options ? options->check_binary : 0;
4025 int err = 0;
4026 int conv_flags = global_conv_flags_eol;
4027 /*
4028 * demote FAIL to WARN to allow inspecting the situation
4029 * instead of refusing.
4030 */
4031 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4032 conv_flags = CONV_EOL_RNDTRP_WARN;
4033
4034 if (!DIFF_FILE_VALID(s))
4035 die("internal error: asking to populate invalid file.");
4036 if (S_ISDIR(s->mode))
4037 return -1;
4038
4039 if (s->data)
4040 return 0;
4041
4042 if (size_only && 0 < s->size)
4043 return 0;
4044
4045 if (S_ISGITLINK(s->mode))
4046 return diff_populate_gitlink(s, size_only);
4047
4048 if (!s->oid_valid ||
4049 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4050 struct strbuf buf = STRBUF_INIT;
4051 struct stat st;
4052 int fd;
4053
4054 if (lstat(s->path, &st) < 0) {
4055 err_empty:
4056 err = -1;
4057 empty:
4058 s->data = (char *)"";
4059 s->size = 0;
4060 return err;
4061 }
4062 s->size = xsize_t(st.st_size);
4063 if (!s->size)
4064 goto empty;
4065 if (S_ISLNK(st.st_mode)) {
4066 struct strbuf sb = STRBUF_INIT;
4067
4068 if (strbuf_readlink(&sb, s->path, s->size))
4069 goto err_empty;
4070 s->size = sb.len;
4071 s->data = strbuf_detach(&sb, NULL);
4072 s->should_free = 1;
4073 return 0;
4074 }
4075
4076 /*
4077 * Even if the caller would be happy with getting
4078 * only the size, we cannot return early at this
4079 * point if the path requires us to run the content
4080 * conversion.
4081 */
4082 if (size_only && !would_convert_to_git(r->index, s->path))
4083 return 0;
4084
4085 /*
4086 * Note: this check uses xsize_t(st.st_size) that may
4087 * not be the true size of the blob after it goes
4088 * through convert_to_git(). This may not strictly be
4089 * correct, but the whole point of big_file_threshold
4090 * and is_binary check being that we want to avoid
4091 * opening the file and inspecting the contents, this
4092 * is probably fine.
4093 */
4094 if (check_binary &&
4095 s->size > big_file_threshold && s->is_binary == -1) {
4096 s->is_binary = 1;
4097 return 0;
4098 }
4099 fd = open(s->path, O_RDONLY);
4100 if (fd < 0)
4101 goto err_empty;
4102 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4103 close(fd);
4104 s->should_munmap = 1;
4105
4106 /*
4107 * Convert from working tree format to canonical git format
4108 */
4109 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4110 size_t size = 0;
4111 munmap(s->data, s->size);
4112 s->should_munmap = 0;
4113 s->data = strbuf_detach(&buf, &size);
4114 s->size = size;
4115 s->should_free = 1;
4116 }
4117 }
4118 else {
4119 struct object_info info = {
4120 .sizep = &s->size
4121 };
4122
4123 if (!(size_only || check_binary))
4124 /*
4125 * Set contentp, since there is no chance that merely
4126 * the size is sufficient.
4127 */
4128 info.contentp = &s->data;
4129
4130 if (options && options->missing_object_cb) {
4131 if (!oid_object_info_extended(r, &s->oid, &info,
4132 OBJECT_INFO_LOOKUP_REPLACE |
4133 OBJECT_INFO_SKIP_FETCH_OBJECT))
4134 goto object_read;
4135 options->missing_object_cb(options->missing_object_data);
4136 }
4137 if (oid_object_info_extended(r, &s->oid, &info,
4138 OBJECT_INFO_LOOKUP_REPLACE))
4139 die("unable to read %s", oid_to_hex(&s->oid));
4140
4141 object_read:
4142 if (size_only || check_binary) {
4143 if (size_only)
4144 return 0;
4145 if (s->size > big_file_threshold && s->is_binary == -1) {
4146 s->is_binary = 1;
4147 return 0;
4148 }
4149 }
4150 if (!info.contentp) {
4151 info.contentp = &s->data;
4152 if (oid_object_info_extended(r, &s->oid, &info,
4153 OBJECT_INFO_LOOKUP_REPLACE))
4154 die("unable to read %s", oid_to_hex(&s->oid));
4155 }
4156 s->should_free = 1;
4157 }
4158 return 0;
4159 }
4160
4161 void diff_free_filespec_blob(struct diff_filespec *s)
4162 {
4163 if (s->should_free)
4164 free(s->data);
4165 else if (s->should_munmap)
4166 munmap(s->data, s->size);
4167
4168 if (s->should_free || s->should_munmap) {
4169 s->should_free = s->should_munmap = 0;
4170 s->data = NULL;
4171 }
4172 }
4173
4174 void diff_free_filespec_data(struct diff_filespec *s)
4175 {
4176 if (!s)
4177 return;
4178
4179 diff_free_filespec_blob(s);
4180 FREE_AND_NULL(s->cnt_data);
4181 }
4182
4183 static void prep_temp_blob(struct index_state *istate,
4184 const char *path, struct diff_tempfile *temp,
4185 void *blob,
4186 unsigned long size,
4187 const struct object_id *oid,
4188 int mode)
4189 {
4190 struct strbuf buf = STRBUF_INIT;
4191 char *path_dup = xstrdup(path);
4192 const char *base = basename(path_dup);
4193 struct checkout_metadata meta;
4194
4195 init_checkout_metadata(&meta, NULL, NULL, oid);
4196
4197 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4198 if (!temp->tempfile)
4199 die_errno("unable to create temp-file");
4200 if (convert_to_working_tree(istate, path,
4201 (const char *)blob, (size_t)size, &buf, &meta)) {
4202 blob = buf.buf;
4203 size = buf.len;
4204 }
4205 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4206 close_tempfile_gently(temp->tempfile))
4207 die_errno("unable to write temp-file");
4208 temp->name = get_tempfile_path(temp->tempfile);
4209 oid_to_hex_r(temp->hex, oid);
4210 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4211 strbuf_release(&buf);
4212 free(path_dup);
4213 }
4214
4215 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4216 const char *name,
4217 struct diff_filespec *one)
4218 {
4219 struct diff_tempfile *temp = claim_diff_tempfile();
4220
4221 if (!DIFF_FILE_VALID(one)) {
4222 not_a_valid_file:
4223 /* A '-' entry produces this for file-2, and
4224 * a '+' entry produces this for file-1.
4225 */
4226 temp->name = "/dev/null";
4227 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4228 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4229 return temp;
4230 }
4231
4232 if (!S_ISGITLINK(one->mode) &&
4233 (!one->oid_valid ||
4234 reuse_worktree_file(r->index, name, &one->oid, 1))) {
4235 struct stat st;
4236 if (lstat(name, &st) < 0) {
4237 if (errno == ENOENT)
4238 goto not_a_valid_file;
4239 die_errno("stat(%s)", name);
4240 }
4241 if (S_ISLNK(st.st_mode)) {
4242 struct strbuf sb = STRBUF_INIT;
4243 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4244 die_errno("readlink(%s)", name);
4245 prep_temp_blob(r->index, name, temp, sb.buf, sb.len,
4246 (one->oid_valid ?
4247 &one->oid : null_oid()),
4248 (one->oid_valid ?
4249 one->mode : S_IFLNK));
4250 strbuf_release(&sb);
4251 }
4252 else {
4253 /* we can borrow from the file in the work tree */
4254 temp->name = name;
4255 if (!one->oid_valid)
4256 oid_to_hex_r(temp->hex, null_oid());
4257 else
4258 oid_to_hex_r(temp->hex, &one->oid);
4259 /* Even though we may sometimes borrow the
4260 * contents from the work tree, we always want
4261 * one->mode. mode is trustworthy even when
4262 * !(one->oid_valid), as long as
4263 * DIFF_FILE_VALID(one).
4264 */
4265 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4266 }
4267 return temp;
4268 }
4269 else {
4270 if (diff_populate_filespec(r, one, NULL))
4271 die("cannot read data blob for %s", one->path);
4272 prep_temp_blob(r->index, name, temp,
4273 one->data, one->size,
4274 &one->oid, one->mode);
4275 }
4276 return temp;
4277 }
4278
4279 static void add_external_diff_name(struct repository *r,
4280 struct strvec *argv,
4281 const char *name,
4282 struct diff_filespec *df)
4283 {
4284 struct diff_tempfile *temp = prepare_temp_file(r, name, df);
4285 strvec_push(argv, temp->name);
4286 strvec_push(argv, temp->hex);
4287 strvec_push(argv, temp->mode);
4288 }
4289
4290 /* An external diff command takes:
4291 *
4292 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4293 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4294 *
4295 */
4296 static void run_external_diff(const char *pgm,
4297 const char *name,
4298 const char *other,
4299 struct diff_filespec *one,
4300 struct diff_filespec *two,
4301 const char *xfrm_msg,
4302 struct diff_options *o)
4303 {
4304 struct child_process cmd = CHILD_PROCESS_INIT;
4305 struct diff_queue_struct *q = &diff_queued_diff;
4306
4307 strvec_push(&cmd.args, pgm);
4308 strvec_push(&cmd.args, name);
4309
4310 if (one && two) {
4311 add_external_diff_name(o->repo, &cmd.args, name, one);
4312 if (!other)
4313 add_external_diff_name(o->repo, &cmd.args, name, two);
4314 else {
4315 add_external_diff_name(o->repo, &cmd.args, other, two);
4316 strvec_push(&cmd.args, other);
4317 strvec_push(&cmd.args, xfrm_msg);
4318 }
4319 }
4320
4321 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4322 ++o->diff_path_counter);
4323 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4324
4325 diff_free_filespec_data(one);
4326 diff_free_filespec_data(two);
4327 cmd.use_shell = 1;
4328 if (run_command(&cmd))
4329 die(_("external diff died, stopping at %s"), name);
4330
4331 remove_tempfile();
4332 }
4333
4334 static int similarity_index(struct diff_filepair *p)
4335 {
4336 return p->score * 100 / MAX_SCORE;
4337 }
4338
4339 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4340 {
4341 if (startup_info->have_repository)
4342 return find_unique_abbrev(oid, abbrev);
4343 else {
4344 char *hex = oid_to_hex(oid);
4345 if (abbrev < 0)
4346 abbrev = FALLBACK_DEFAULT_ABBREV;
4347 if (abbrev > the_hash_algo->hexsz)
4348 BUG("oid abbreviation out of range: %d", abbrev);
4349 if (abbrev)
4350 hex[abbrev] = '\0';
4351 return hex;
4352 }
4353 }
4354
4355 static void fill_metainfo(struct strbuf *msg,
4356 const char *name,
4357 const char *other,
4358 struct diff_filespec *one,
4359 struct diff_filespec *two,
4360 struct diff_options *o,
4361 struct diff_filepair *p,
4362 int *must_show_header,
4363 int use_color)
4364 {
4365 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4366 const char *reset = diff_get_color(use_color, DIFF_RESET);
4367 const char *line_prefix = diff_line_prefix(o);
4368 struct string_list *more_headers = NULL;
4369
4370 *must_show_header = 1;
4371 strbuf_init(msg, PATH_MAX * 2 + 300);
4372 switch (p->status) {
4373 case DIFF_STATUS_COPIED:
4374 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4375 line_prefix, set, similarity_index(p));
4376 strbuf_addf(msg, "%s\n%s%scopy from ",
4377 reset, line_prefix, set);
4378 quote_c_style(name, msg, NULL, 0);
4379 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4380 quote_c_style(other, msg, NULL, 0);
4381 strbuf_addf(msg, "%s\n", reset);
4382 break;
4383 case DIFF_STATUS_RENAMED:
4384 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4385 line_prefix, set, similarity_index(p));
4386 strbuf_addf(msg, "%s\n%s%srename from ",
4387 reset, line_prefix, set);
4388 quote_c_style(name, msg, NULL, 0);
4389 strbuf_addf(msg, "%s\n%s%srename to ",
4390 reset, line_prefix, set);
4391 quote_c_style(other, msg, NULL, 0);
4392 strbuf_addf(msg, "%s\n", reset);
4393 break;
4394 case DIFF_STATUS_MODIFIED:
4395 if (p->score) {
4396 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4397 line_prefix,
4398 set, similarity_index(p), reset);
4399 break;
4400 }
4401 /* fallthru */
4402 default:
4403 *must_show_header = 0;
4404 }
4405 if ((more_headers = additional_headers(o, name))) {
4406 add_formatted_headers(msg, more_headers,
4407 line_prefix, set, reset);
4408 *must_show_header = 1;
4409 }
4410 if (one && two && !oideq(&one->oid, &two->oid)) {
4411 const unsigned hexsz = the_hash_algo->hexsz;
4412 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4413
4414 if (o->flags.full_index)
4415 abbrev = hexsz;
4416
4417 if (o->flags.binary) {
4418 mmfile_t mf;
4419 if ((!fill_mmfile(o->repo, &mf, one) &&
4420 diff_filespec_is_binary(o->repo, one)) ||
4421 (!fill_mmfile(o->repo, &mf, two) &&
4422 diff_filespec_is_binary(o->repo, two)))
4423 abbrev = hexsz;
4424 }
4425 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4426 diff_abbrev_oid(&one->oid, abbrev),
4427 diff_abbrev_oid(&two->oid, abbrev));
4428 if (one->mode == two->mode)
4429 strbuf_addf(msg, " %06o", one->mode);
4430 strbuf_addf(msg, "%s\n", reset);
4431 }
4432 }
4433
4434 static void run_diff_cmd(const char *pgm,
4435 const char *name,
4436 const char *other,
4437 const char *attr_path,
4438 struct diff_filespec *one,
4439 struct diff_filespec *two,
4440 struct strbuf *msg,
4441 struct diff_options *o,
4442 struct diff_filepair *p)
4443 {
4444 const char *xfrm_msg = NULL;
4445 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4446 int must_show_header = 0;
4447
4448
4449 if (o->flags.allow_external) {
4450 struct userdiff_driver *drv;
4451
4452 drv = userdiff_find_by_path(o->repo->index, attr_path);
4453 if (drv && drv->external)
4454 pgm = drv->external;
4455 }
4456
4457 if (msg) {
4458 /*
4459 * don't use colors when the header is intended for an
4460 * external diff driver
4461 */
4462 fill_metainfo(msg, name, other, one, two, o, p,
4463 &must_show_header,
4464 want_color(o->use_color) && !pgm);
4465 xfrm_msg = msg->len ? msg->buf : NULL;
4466 }
4467
4468 if (pgm) {
4469 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4470 return;
4471 }
4472 if (one && two)
4473 builtin_diff(name, other ? other : name,
4474 one, two, xfrm_msg, must_show_header,
4475 o, complete_rewrite);
4476 else
4477 fprintf(o->file, "* Unmerged path %s\n", name);
4478 }
4479
4480 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4481 {
4482 if (DIFF_FILE_VALID(one)) {
4483 if (!one->oid_valid) {
4484 struct stat st;
4485 if (one->is_stdin) {
4486 oidclr(&one->oid);
4487 return;
4488 }
4489 if (lstat(one->path, &st) < 0)
4490 die_errno("stat '%s'", one->path);
4491 if (index_path(istate, &one->oid, one->path, &st, 0))
4492 die("cannot hash %s", one->path);
4493 }
4494 }
4495 else
4496 oidclr(&one->oid);
4497 }
4498
4499 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4500 {
4501 /* Strip the prefix but do not molest /dev/null and absolute paths */
4502 if (*namep && !is_absolute_path(*namep)) {
4503 *namep += prefix_length;
4504 if (**namep == '/')
4505 ++*namep;
4506 }
4507 if (*otherp && !is_absolute_path(*otherp)) {
4508 *otherp += prefix_length;
4509 if (**otherp == '/')
4510 ++*otherp;
4511 }
4512 }
4513
4514 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4515 {
4516 const char *pgm = external_diff();
4517 struct strbuf msg;
4518 struct diff_filespec *one = p->one;
4519 struct diff_filespec *two = p->two;
4520 const char *name;
4521 const char *other;
4522 const char *attr_path;
4523
4524 name = one->path;
4525 other = (strcmp(name, two->path) ? two->path : NULL);
4526 attr_path = name;
4527 if (o->prefix_length)
4528 strip_prefix(o->prefix_length, &name, &other);
4529
4530 if (!o->flags.allow_external)
4531 pgm = NULL;
4532
4533 if (DIFF_PAIR_UNMERGED(p)) {
4534 run_diff_cmd(pgm, name, NULL, attr_path,
4535 NULL, NULL, NULL, o, p);
4536 return;
4537 }
4538
4539 diff_fill_oid_info(one, o->repo->index);
4540 diff_fill_oid_info(two, o->repo->index);
4541
4542 if (!pgm &&
4543 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4544 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4545 /*
4546 * a filepair that changes between file and symlink
4547 * needs to be split into deletion and creation.
4548 */
4549 struct diff_filespec *null = alloc_filespec(two->path);
4550 run_diff_cmd(NULL, name, other, attr_path,
4551 one, null, &msg,
4552 o, p);
4553 free(null);
4554 strbuf_release(&msg);
4555
4556 null = alloc_filespec(one->path);
4557 run_diff_cmd(NULL, name, other, attr_path,
4558 null, two, &msg, o, p);
4559 free(null);
4560 }
4561 else
4562 run_diff_cmd(pgm, name, other, attr_path,
4563 one, two, &msg, o, p);
4564
4565 strbuf_release(&msg);
4566 }
4567
4568 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4569 struct diffstat_t *diffstat)
4570 {
4571 const char *name;
4572 const char *other;
4573
4574 if (DIFF_PAIR_UNMERGED(p)) {
4575 /* unmerged */
4576 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4577 diffstat, o, p);
4578 return;
4579 }
4580
4581 name = p->one->path;
4582 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4583
4584 if (o->prefix_length)
4585 strip_prefix(o->prefix_length, &name, &other);
4586
4587 diff_fill_oid_info(p->one, o->repo->index);
4588 diff_fill_oid_info(p->two, o->repo->index);
4589
4590 builtin_diffstat(name, other, p->one, p->two,
4591 diffstat, o, p);
4592 }
4593
4594 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4595 {
4596 const char *name;
4597 const char *other;
4598 const char *attr_path;
4599
4600 if (DIFF_PAIR_UNMERGED(p)) {
4601 /* unmerged */
4602 return;
4603 }
4604
4605 name = p->one->path;
4606 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4607 attr_path = other ? other : name;
4608
4609 if (o->prefix_length)
4610 strip_prefix(o->prefix_length, &name, &other);
4611
4612 diff_fill_oid_info(p->one, o->repo->index);
4613 diff_fill_oid_info(p->two, o->repo->index);
4614
4615 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4616 }
4617
4618 static void prep_parse_options(struct diff_options *options);
4619
4620 void repo_diff_setup(struct repository *r, struct diff_options *options)
4621 {
4622 memcpy(options, &default_diff_options, sizeof(*options));
4623
4624 options->file = stdout;
4625 options->repo = r;
4626
4627 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4628 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4629 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4630 options->abbrev = DEFAULT_ABBREV;
4631 options->line_termination = '\n';
4632 options->break_opt = -1;
4633 options->rename_limit = -1;
4634 options->dirstat_permille = diff_dirstat_permille_default;
4635 options->context = diff_context_default;
4636 options->interhunkcontext = diff_interhunk_context_default;
4637 options->ws_error_highlight = ws_error_highlight_default;
4638 options->flags.rename_empty = 1;
4639 options->flags.relative_name = diff_relative;
4640 options->objfind = NULL;
4641
4642 /* pathchange left =NULL by default */
4643 options->change = diff_change;
4644 options->add_remove = diff_addremove;
4645 options->use_color = diff_use_color_default;
4646 options->detect_rename = diff_detect_rename_default;
4647 options->xdl_opts |= diff_algorithm;
4648 if (diff_indent_heuristic)
4649 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4650
4651 options->orderfile = diff_order_file_cfg;
4652
4653 if (!options->flags.ignore_submodule_set)
4654 options->flags.ignore_untracked_in_submodules = 1;
4655
4656 if (diff_no_prefix) {
4657 options->a_prefix = options->b_prefix = "";
4658 } else if (!diff_mnemonic_prefix) {
4659 options->a_prefix = "a/";
4660 options->b_prefix = "b/";
4661 }
4662
4663 options->color_moved = diff_color_moved_default;
4664 options->color_moved_ws_handling = diff_color_moved_ws_default;
4665
4666 prep_parse_options(options);
4667 }
4668
4669 static const char diff_status_letters[] = {
4670 DIFF_STATUS_ADDED,
4671 DIFF_STATUS_COPIED,
4672 DIFF_STATUS_DELETED,
4673 DIFF_STATUS_MODIFIED,
4674 DIFF_STATUS_RENAMED,
4675 DIFF_STATUS_TYPE_CHANGED,
4676 DIFF_STATUS_UNKNOWN,
4677 DIFF_STATUS_UNMERGED,
4678 DIFF_STATUS_FILTER_AON,
4679 DIFF_STATUS_FILTER_BROKEN,
4680 '\0',
4681 };
4682
4683 static unsigned int filter_bit['Z' + 1];
4684
4685 static void prepare_filter_bits(void)
4686 {
4687 int i;
4688
4689 if (!filter_bit[DIFF_STATUS_ADDED]) {
4690 for (i = 0; diff_status_letters[i]; i++)
4691 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4692 }
4693 }
4694
4695 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4696 {
4697 return opt->filter & filter_bit[(int) status];
4698 }
4699
4700 unsigned diff_filter_bit(char status)
4701 {
4702 prepare_filter_bits();
4703 return filter_bit[(int) status];
4704 }
4705
4706 void diff_setup_done(struct diff_options *options)
4707 {
4708 unsigned check_mask = DIFF_FORMAT_NAME |
4709 DIFF_FORMAT_NAME_STATUS |
4710 DIFF_FORMAT_CHECKDIFF |
4711 DIFF_FORMAT_NO_OUTPUT;
4712 /*
4713 * This must be signed because we're comparing against a potentially
4714 * negative value.
4715 */
4716 const int hexsz = the_hash_algo->hexsz;
4717
4718 if (options->set_default)
4719 options->set_default(options);
4720
4721 if (HAS_MULTI_BITS(options->output_format & check_mask))
4722 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4723 "--name-only", "--name-status", "--check", "-s");
4724
4725 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4726 die(_("options '%s', '%s', and '%s' cannot be used together"),
4727 "-G", "-S", "--find-object");
4728
4729 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4730 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4731 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4732
4733 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4734 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4735 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4736
4737 /*
4738 * Most of the time we can say "there are changes"
4739 * only by checking if there are changed paths, but
4740 * --ignore-whitespace* options force us to look
4741 * inside contents.
4742 */
4743
4744 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4745 options->ignore_regex_nr)
4746 options->flags.diff_from_contents = 1;
4747 else
4748 options->flags.diff_from_contents = 0;
4749
4750 if (options->flags.find_copies_harder)
4751 options->detect_rename = DIFF_DETECT_COPY;
4752
4753 if (!options->flags.relative_name)
4754 options->prefix = NULL;
4755 if (options->prefix)
4756 options->prefix_length = strlen(options->prefix);
4757 else
4758 options->prefix_length = 0;
4759
4760 if (options->output_format & (DIFF_FORMAT_NAME |
4761 DIFF_FORMAT_NAME_STATUS |
4762 DIFF_FORMAT_CHECKDIFF |
4763 DIFF_FORMAT_NO_OUTPUT))
4764 options->output_format &= ~(DIFF_FORMAT_RAW |
4765 DIFF_FORMAT_NUMSTAT |
4766 DIFF_FORMAT_DIFFSTAT |
4767 DIFF_FORMAT_SHORTSTAT |
4768 DIFF_FORMAT_DIRSTAT |
4769 DIFF_FORMAT_SUMMARY |
4770 DIFF_FORMAT_PATCH);
4771
4772 /*
4773 * These cases always need recursive; we do not drop caller-supplied
4774 * recursive bits for other formats here.
4775 */
4776 if (options->output_format & (DIFF_FORMAT_PATCH |
4777 DIFF_FORMAT_NUMSTAT |
4778 DIFF_FORMAT_DIFFSTAT |
4779 DIFF_FORMAT_SHORTSTAT |
4780 DIFF_FORMAT_DIRSTAT |
4781 DIFF_FORMAT_SUMMARY |
4782 DIFF_FORMAT_CHECKDIFF))
4783 options->flags.recursive = 1;
4784 /*
4785 * Also pickaxe would not work very well if you do not say recursive
4786 */
4787 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4788 options->flags.recursive = 1;
4789 /*
4790 * When patches are generated, submodules diffed against the work tree
4791 * must be checked for dirtiness too so it can be shown in the output
4792 */
4793 if (options->output_format & DIFF_FORMAT_PATCH)
4794 options->flags.dirty_submodules = 1;
4795
4796 if (options->detect_rename && options->rename_limit < 0)
4797 options->rename_limit = diff_rename_limit_default;
4798 if (hexsz < options->abbrev)
4799 options->abbrev = hexsz; /* full */
4800
4801 /*
4802 * It does not make sense to show the first hit we happened
4803 * to have found. It does not make sense not to return with
4804 * exit code in such a case either.
4805 */
4806 if (options->flags.quick) {
4807 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4808 options->flags.exit_with_status = 1;
4809 }
4810
4811 options->diff_path_counter = 0;
4812
4813 if (options->flags.follow_renames && options->pathspec.nr != 1)
4814 die(_("--follow requires exactly one pathspec"));
4815
4816 if (!options->use_color || external_diff())
4817 options->color_moved = 0;
4818
4819 if (options->filter_not) {
4820 if (!options->filter)
4821 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4822 options->filter &= ~options->filter_not;
4823 }
4824
4825 FREE_AND_NULL(options->parseopts);
4826 }
4827
4828 int parse_long_opt(const char *opt, const char **argv,
4829 const char **optarg)
4830 {
4831 const char *arg = argv[0];
4832 if (!skip_prefix(arg, "--", &arg))
4833 return 0;
4834 if (!skip_prefix(arg, opt, &arg))
4835 return 0;
4836 if (*arg == '=') { /* stuck form: --option=value */
4837 *optarg = arg + 1;
4838 return 1;
4839 }
4840 if (*arg != '\0')
4841 return 0;
4842 /* separate form: --option value */
4843 if (!argv[1])
4844 die("Option '--%s' requires a value", opt);
4845 *optarg = argv[1];
4846 return 2;
4847 }
4848
4849 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4850 {
4851 struct diff_options *options = opt->value;
4852 int width = options->stat_width;
4853 int name_width = options->stat_name_width;
4854 int graph_width = options->stat_graph_width;
4855 int count = options->stat_count;
4856 char *end;
4857
4858 BUG_ON_OPT_NEG(unset);
4859
4860 if (!strcmp(opt->long_name, "stat")) {
4861 if (value) {
4862 width = strtoul(value, &end, 10);
4863 if (*end == ',')
4864 name_width = strtoul(end+1, &end, 10);
4865 if (*end == ',')
4866 count = strtoul(end+1, &end, 10);
4867 if (*end)
4868 return error(_("invalid --stat value: %s"), value);
4869 }
4870 } else if (!strcmp(opt->long_name, "stat-width")) {
4871 width = strtoul(value, &end, 10);
4872 if (*end)
4873 return error(_("%s expects a numerical value"),
4874 opt->long_name);
4875 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4876 name_width = strtoul(value, &end, 10);
4877 if (*end)
4878 return error(_("%s expects a numerical value"),
4879 opt->long_name);
4880 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4881 graph_width = strtoul(value, &end, 10);
4882 if (*end)
4883 return error(_("%s expects a numerical value"),
4884 opt->long_name);
4885 } else if (!strcmp(opt->long_name, "stat-count")) {
4886 count = strtoul(value, &end, 10);
4887 if (*end)
4888 return error(_("%s expects a numerical value"),
4889 opt->long_name);
4890 } else
4891 BUG("%s should not get here", opt->long_name);
4892
4893 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4894 options->stat_name_width = name_width;
4895 options->stat_graph_width = graph_width;
4896 options->stat_width = width;
4897 options->stat_count = count;
4898 return 0;
4899 }
4900
4901 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4902 {
4903 struct strbuf errmsg = STRBUF_INIT;
4904 if (parse_dirstat_params(options, params, &errmsg))
4905 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4906 errmsg.buf);
4907 strbuf_release(&errmsg);
4908 /*
4909 * The caller knows a dirstat-related option is given from the command
4910 * line; allow it to say "return this_function();"
4911 */
4912 options->output_format |= DIFF_FORMAT_DIRSTAT;
4913 return 1;
4914 }
4915
4916 static int diff_opt_diff_filter(const struct option *option,
4917 const char *optarg, int unset)
4918 {
4919 struct diff_options *opt = option->value;
4920 int i, optch;
4921
4922 BUG_ON_OPT_NEG(unset);
4923 prepare_filter_bits();
4924
4925 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4926 unsigned int bit;
4927 int negate;
4928
4929 if ('a' <= optch && optch <= 'z') {
4930 negate = 1;
4931 optch = toupper(optch);
4932 } else {
4933 negate = 0;
4934 }
4935
4936 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4937 if (!bit)
4938 return error(_("unknown change class '%c' in --diff-filter=%s"),
4939 optarg[i], optarg);
4940 if (negate)
4941 opt->filter_not |= bit;
4942 else
4943 opt->filter |= bit;
4944 }
4945 return 0;
4946 }
4947
4948 static void enable_patch_output(int *fmt)
4949 {
4950 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4951 *fmt |= DIFF_FORMAT_PATCH;
4952 }
4953
4954 static int diff_opt_ws_error_highlight(const struct option *option,
4955 const char *arg, int unset)
4956 {
4957 struct diff_options *opt = option->value;
4958 int val = parse_ws_error_highlight(arg);
4959
4960 BUG_ON_OPT_NEG(unset);
4961 if (val < 0)
4962 return error(_("unknown value after ws-error-highlight=%.*s"),
4963 -1 - val, arg);
4964 opt->ws_error_highlight = val;
4965 return 0;
4966 }
4967
4968 static int diff_opt_find_object(const struct option *option,
4969 const char *arg, int unset)
4970 {
4971 struct diff_options *opt = option->value;
4972 struct object_id oid;
4973
4974 BUG_ON_OPT_NEG(unset);
4975 if (get_oid(arg, &oid))
4976 return error(_("unable to resolve '%s'"), arg);
4977
4978 if (!opt->objfind)
4979 CALLOC_ARRAY(opt->objfind, 1);
4980
4981 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4982 opt->flags.recursive = 1;
4983 opt->flags.tree_in_recursive = 1;
4984 oidset_insert(opt->objfind, &oid);
4985 return 0;
4986 }
4987
4988 static int diff_opt_anchored(const struct option *opt,
4989 const char *arg, int unset)
4990 {
4991 struct diff_options *options = opt->value;
4992
4993 BUG_ON_OPT_NEG(unset);
4994 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4995 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
4996 options->anchors_alloc);
4997 options->anchors[options->anchors_nr++] = xstrdup(arg);
4998 return 0;
4999 }
5000
5001 static int diff_opt_binary(const struct option *opt,
5002 const char *arg, int unset)
5003 {
5004 struct diff_options *options = opt->value;
5005
5006 BUG_ON_OPT_NEG(unset);
5007 BUG_ON_OPT_ARG(arg);
5008 enable_patch_output(&options->output_format);
5009 options->flags.binary = 1;
5010 return 0;
5011 }
5012
5013 static int diff_opt_break_rewrites(const struct option *opt,
5014 const char *arg, int unset)
5015 {
5016 int *break_opt = opt->value;
5017 int opt1, opt2;
5018
5019 BUG_ON_OPT_NEG(unset);
5020 if (!arg)
5021 arg = "";
5022 opt1 = parse_rename_score(&arg);
5023 if (*arg == 0)
5024 opt2 = 0;
5025 else if (*arg != '/')
5026 return error(_("%s expects <n>/<m> form"), opt->long_name);
5027 else {
5028 arg++;
5029 opt2 = parse_rename_score(&arg);
5030 }
5031 if (*arg != 0)
5032 return error(_("%s expects <n>/<m> form"), opt->long_name);
5033 *break_opt = opt1 | (opt2 << 16);
5034 return 0;
5035 }
5036
5037 static int diff_opt_char(const struct option *opt,
5038 const char *arg, int unset)
5039 {
5040 char *value = opt->value;
5041
5042 BUG_ON_OPT_NEG(unset);
5043 if (arg[1])
5044 return error(_("%s expects a character, got '%s'"),
5045 opt->long_name, arg);
5046 *value = arg[0];
5047 return 0;
5048 }
5049
5050 static int diff_opt_color_moved(const struct option *opt,
5051 const char *arg, int unset)
5052 {
5053 struct diff_options *options = opt->value;
5054
5055 if (unset) {
5056 options->color_moved = COLOR_MOVED_NO;
5057 } else if (!arg) {
5058 if (diff_color_moved_default)
5059 options->color_moved = diff_color_moved_default;
5060 if (options->color_moved == COLOR_MOVED_NO)
5061 options->color_moved = COLOR_MOVED_DEFAULT;
5062 } else {
5063 int cm = parse_color_moved(arg);
5064 if (cm < 0)
5065 return error(_("bad --color-moved argument: %s"), arg);
5066 options->color_moved = cm;
5067 }
5068 return 0;
5069 }
5070
5071 static int diff_opt_color_moved_ws(const struct option *opt,
5072 const char *arg, int unset)
5073 {
5074 struct diff_options *options = opt->value;
5075 unsigned cm;
5076
5077 if (unset) {
5078 options->color_moved_ws_handling = 0;
5079 return 0;
5080 }
5081
5082 cm = parse_color_moved_ws(arg);
5083 if (cm & COLOR_MOVED_WS_ERROR)
5084 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5085 options->color_moved_ws_handling = cm;
5086 return 0;
5087 }
5088
5089 static int diff_opt_color_words(const struct option *opt,
5090 const char *arg, int unset)
5091 {
5092 struct diff_options *options = opt->value;
5093
5094 BUG_ON_OPT_NEG(unset);
5095 options->use_color = 1;
5096 options->word_diff = DIFF_WORDS_COLOR;
5097 options->word_regex = arg;
5098 return 0;
5099 }
5100
5101 static int diff_opt_compact_summary(const struct option *opt,
5102 const char *arg, int unset)
5103 {
5104 struct diff_options *options = opt->value;
5105
5106 BUG_ON_OPT_ARG(arg);
5107 if (unset) {
5108 options->flags.stat_with_summary = 0;
5109 } else {
5110 options->flags.stat_with_summary = 1;
5111 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5112 }
5113 return 0;
5114 }
5115
5116 static int diff_opt_diff_algorithm(const struct option *opt,
5117 const char *arg, int unset)
5118 {
5119 struct diff_options *options = opt->value;
5120 long value = parse_algorithm_value(arg);
5121
5122 BUG_ON_OPT_NEG(unset);
5123 if (value < 0)
5124 return error(_("option diff-algorithm accepts \"myers\", "
5125 "\"minimal\", \"patience\" and \"histogram\""));
5126
5127 /* clear out previous settings */
5128 DIFF_XDL_CLR(options, NEED_MINIMAL);
5129 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
5130 options->xdl_opts |= value;
5131 return 0;
5132 }
5133
5134 static int diff_opt_dirstat(const struct option *opt,
5135 const char *arg, int unset)
5136 {
5137 struct diff_options *options = opt->value;
5138
5139 BUG_ON_OPT_NEG(unset);
5140 if (!strcmp(opt->long_name, "cumulative")) {
5141 if (arg)
5142 BUG("how come --cumulative take a value?");
5143 arg = "cumulative";
5144 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5145 parse_dirstat_opt(options, "files");
5146 parse_dirstat_opt(options, arg ? arg : "");
5147 return 0;
5148 }
5149
5150 static int diff_opt_find_copies(const struct option *opt,
5151 const char *arg, int unset)
5152 {
5153 struct diff_options *options = opt->value;
5154
5155 BUG_ON_OPT_NEG(unset);
5156 if (!arg)
5157 arg = "";
5158 options->rename_score = parse_rename_score(&arg);
5159 if (*arg != 0)
5160 return error(_("invalid argument to %s"), opt->long_name);
5161
5162 if (options->detect_rename == DIFF_DETECT_COPY)
5163 options->flags.find_copies_harder = 1;
5164 else
5165 options->detect_rename = DIFF_DETECT_COPY;
5166
5167 return 0;
5168 }
5169
5170 static int diff_opt_find_renames(const struct option *opt,
5171 const char *arg, int unset)
5172 {
5173 struct diff_options *options = opt->value;
5174
5175 BUG_ON_OPT_NEG(unset);
5176 if (!arg)
5177 arg = "";
5178 options->rename_score = parse_rename_score(&arg);
5179 if (*arg != 0)
5180 return error(_("invalid argument to %s"), opt->long_name);
5181
5182 options->detect_rename = DIFF_DETECT_RENAME;
5183 return 0;
5184 }
5185
5186 static int diff_opt_follow(const struct option *opt,
5187 const char *arg, int unset)
5188 {
5189 struct diff_options *options = opt->value;
5190
5191 BUG_ON_OPT_ARG(arg);
5192 if (unset) {
5193 options->flags.follow_renames = 0;
5194 options->flags.default_follow_renames = 0;
5195 } else {
5196 options->flags.follow_renames = 1;
5197 }
5198 return 0;
5199 }
5200
5201 static int diff_opt_ignore_submodules(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 = "all";
5209 options->flags.override_submodule_config = 1;
5210 handle_ignore_submodules_arg(options, arg);
5211 return 0;
5212 }
5213
5214 static int diff_opt_line_prefix(const struct option *opt,
5215 const char *optarg, int unset)
5216 {
5217 struct diff_options *options = opt->value;
5218
5219 BUG_ON_OPT_NEG(unset);
5220 options->line_prefix = optarg;
5221 options->line_prefix_length = strlen(options->line_prefix);
5222 graph_setup_line_prefix(options);
5223 return 0;
5224 }
5225
5226 static int diff_opt_no_prefix(const struct option *opt,
5227 const char *optarg, int unset)
5228 {
5229 struct diff_options *options = opt->value;
5230
5231 BUG_ON_OPT_NEG(unset);
5232 BUG_ON_OPT_ARG(optarg);
5233 options->a_prefix = "";
5234 options->b_prefix = "";
5235 return 0;
5236 }
5237
5238 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5239 const struct option *opt,
5240 const char *arg, int unset)
5241 {
5242 struct diff_options *options = opt->value;
5243 char *path;
5244
5245 BUG_ON_OPT_NEG(unset);
5246 path = prefix_filename(ctx->prefix, arg);
5247 options->file = xfopen(path, "w");
5248 options->close_file = 1;
5249 if (options->use_color != GIT_COLOR_ALWAYS)
5250 options->use_color = GIT_COLOR_NEVER;
5251 free(path);
5252 return 0;
5253 }
5254
5255 static int diff_opt_patience(const struct option *opt,
5256 const char *arg, int unset)
5257 {
5258 struct diff_options *options = opt->value;
5259 int i;
5260
5261 BUG_ON_OPT_NEG(unset);
5262 BUG_ON_OPT_ARG(arg);
5263 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5264 /*
5265 * Both --patience and --anchored use PATIENCE_DIFF
5266 * internally, so remove any anchors previously
5267 * specified.
5268 */
5269 for (i = 0; i < options->anchors_nr; i++)
5270 free(options->anchors[i]);
5271 options->anchors_nr = 0;
5272 return 0;
5273 }
5274
5275 static int diff_opt_ignore_regex(const struct option *opt,
5276 const char *arg, int unset)
5277 {
5278 struct diff_options *options = opt->value;
5279 regex_t *regex;
5280
5281 BUG_ON_OPT_NEG(unset);
5282 regex = xmalloc(sizeof(*regex));
5283 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5284 return error(_("invalid regex given to -I: '%s'"), arg);
5285 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5286 options->ignore_regex_alloc);
5287 options->ignore_regex[options->ignore_regex_nr++] = regex;
5288 return 0;
5289 }
5290
5291 static int diff_opt_pickaxe_regex(const struct option *opt,
5292 const char *arg, int unset)
5293 {
5294 struct diff_options *options = opt->value;
5295
5296 BUG_ON_OPT_NEG(unset);
5297 options->pickaxe = arg;
5298 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5299 return 0;
5300 }
5301
5302 static int diff_opt_pickaxe_string(const struct option *opt,
5303 const char *arg, int unset)
5304 {
5305 struct diff_options *options = opt->value;
5306
5307 BUG_ON_OPT_NEG(unset);
5308 options->pickaxe = arg;
5309 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5310 return 0;
5311 }
5312
5313 static int diff_opt_relative(const struct option *opt,
5314 const char *arg, int unset)
5315 {
5316 struct diff_options *options = opt->value;
5317
5318 options->flags.relative_name = !unset;
5319 if (arg)
5320 options->prefix = arg;
5321 return 0;
5322 }
5323
5324 static int diff_opt_submodule(const struct option *opt,
5325 const char *arg, int unset)
5326 {
5327 struct diff_options *options = opt->value;
5328
5329 BUG_ON_OPT_NEG(unset);
5330 if (!arg)
5331 arg = "log";
5332 if (parse_submodule_params(options, arg))
5333 return error(_("failed to parse --submodule option parameter: '%s'"),
5334 arg);
5335 return 0;
5336 }
5337
5338 static int diff_opt_textconv(const struct option *opt,
5339 const char *arg, int unset)
5340 {
5341 struct diff_options *options = opt->value;
5342
5343 BUG_ON_OPT_ARG(arg);
5344 if (unset) {
5345 options->flags.allow_textconv = 0;
5346 } else {
5347 options->flags.allow_textconv = 1;
5348 options->flags.textconv_set_via_cmdline = 1;
5349 }
5350 return 0;
5351 }
5352
5353 static int diff_opt_unified(const struct option *opt,
5354 const char *arg, int unset)
5355 {
5356 struct diff_options *options = opt->value;
5357 char *s;
5358
5359 BUG_ON_OPT_NEG(unset);
5360
5361 if (arg) {
5362 options->context = strtol(arg, &s, 10);
5363 if (*s)
5364 return error(_("%s expects a numerical value"), "--unified");
5365 }
5366 enable_patch_output(&options->output_format);
5367
5368 return 0;
5369 }
5370
5371 static int diff_opt_word_diff(const struct option *opt,
5372 const char *arg, int unset)
5373 {
5374 struct diff_options *options = opt->value;
5375
5376 BUG_ON_OPT_NEG(unset);
5377 if (arg) {
5378 if (!strcmp(arg, "plain"))
5379 options->word_diff = DIFF_WORDS_PLAIN;
5380 else if (!strcmp(arg, "color")) {
5381 options->use_color = 1;
5382 options->word_diff = DIFF_WORDS_COLOR;
5383 }
5384 else if (!strcmp(arg, "porcelain"))
5385 options->word_diff = DIFF_WORDS_PORCELAIN;
5386 else if (!strcmp(arg, "none"))
5387 options->word_diff = DIFF_WORDS_NONE;
5388 else
5389 return error(_("bad --word-diff argument: %s"), arg);
5390 } else {
5391 if (options->word_diff == DIFF_WORDS_NONE)
5392 options->word_diff = DIFF_WORDS_PLAIN;
5393 }
5394 return 0;
5395 }
5396
5397 static int diff_opt_word_diff_regex(const struct option *opt,
5398 const char *arg, int unset)
5399 {
5400 struct diff_options *options = opt->value;
5401
5402 BUG_ON_OPT_NEG(unset);
5403 if (options->word_diff == DIFF_WORDS_NONE)
5404 options->word_diff = DIFF_WORDS_PLAIN;
5405 options->word_regex = arg;
5406 return 0;
5407 }
5408
5409 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5410 {
5411 struct diff_options *options = opt->value;
5412
5413 BUG_ON_OPT_NEG(unset);
5414 if (!strcmp(opt->long_name, "skip-to"))
5415 options->skip_instead_of_rotate = 1;
5416 else
5417 options->skip_instead_of_rotate = 0;
5418 options->rotate_to = arg;
5419 return 0;
5420 }
5421
5422 static void prep_parse_options(struct diff_options *options)
5423 {
5424 struct option parseopts[] = {
5425 OPT_GROUP(N_("Diff output format options")),
5426 OPT_BITOP('p', "patch", &options->output_format,
5427 N_("generate patch"),
5428 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5429 OPT_BIT_F('s', "no-patch", &options->output_format,
5430 N_("suppress diff output"),
5431 DIFF_FORMAT_NO_OUTPUT, PARSE_OPT_NONEG),
5432 OPT_BITOP('u', NULL, &options->output_format,
5433 N_("generate patch"),
5434 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5435 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5436 N_("generate diffs with <n> lines context"),
5437 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5438 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5439 N_("generate diffs with <n> lines context")),
5440 OPT_BIT_F(0, "raw", &options->output_format,
5441 N_("generate the diff in raw format"),
5442 DIFF_FORMAT_RAW, PARSE_OPT_NONEG),
5443 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5444 N_("synonym for '-p --raw'"),
5445 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5446 DIFF_FORMAT_NO_OUTPUT),
5447 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5448 N_("synonym for '-p --stat'"),
5449 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5450 DIFF_FORMAT_NO_OUTPUT),
5451 OPT_BIT_F(0, "numstat", &options->output_format,
5452 N_("machine friendly --stat"),
5453 DIFF_FORMAT_NUMSTAT, PARSE_OPT_NONEG),
5454 OPT_BIT_F(0, "shortstat", &options->output_format,
5455 N_("output only the last line of --stat"),
5456 DIFF_FORMAT_SHORTSTAT, PARSE_OPT_NONEG),
5457 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1,param2>..."),
5458 N_("output the distribution of relative amount of changes for each sub-directory"),
5459 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5460 diff_opt_dirstat),
5461 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5462 N_("synonym for --dirstat=cumulative"),
5463 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5464 diff_opt_dirstat),
5465 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1,param2>..."),
5466 N_("synonym for --dirstat=files,param1,param2..."),
5467 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5468 diff_opt_dirstat),
5469 OPT_BIT_F(0, "check", &options->output_format,
5470 N_("warn if changes introduce conflict markers or whitespace errors"),
5471 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5472 OPT_BIT_F(0, "summary", &options->output_format,
5473 N_("condensed summary such as creations, renames and mode changes"),
5474 DIFF_FORMAT_SUMMARY, PARSE_OPT_NONEG),
5475 OPT_BIT_F(0, "name-only", &options->output_format,
5476 N_("show only names of changed files"),
5477 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5478 OPT_BIT_F(0, "name-status", &options->output_format,
5479 N_("show only names and status of changed files"),
5480 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5481 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5482 N_("generate diffstat"),
5483 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5484 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5485 N_("generate diffstat with a given width"),
5486 PARSE_OPT_NONEG, diff_opt_stat),
5487 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5488 N_("generate diffstat with a given name width"),
5489 PARSE_OPT_NONEG, diff_opt_stat),
5490 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5491 N_("generate diffstat with a given graph width"),
5492 PARSE_OPT_NONEG, diff_opt_stat),
5493 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5494 N_("generate diffstat with limited lines"),
5495 PARSE_OPT_NONEG, diff_opt_stat),
5496 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5497 N_("generate compact summary in diffstat"),
5498 PARSE_OPT_NOARG, diff_opt_compact_summary),
5499 OPT_CALLBACK_F(0, "binary", options, NULL,
5500 N_("output a binary diff that can be applied"),
5501 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5502 OPT_BOOL(0, "full-index", &options->flags.full_index,
5503 N_("show full pre- and post-image object names on the \"index\" lines")),
5504 OPT_COLOR_FLAG(0, "color", &options->use_color,
5505 N_("show colored diff")),
5506 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5507 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5508 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5509 OPT_SET_INT('z', NULL, &options->line_termination,
5510 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5511 0),
5512 OPT__ABBREV(&options->abbrev),
5513 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5514 N_("show the given source prefix instead of \"a/\""),
5515 PARSE_OPT_NONEG),
5516 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5517 N_("show the given destination prefix instead of \"b/\""),
5518 PARSE_OPT_NONEG),
5519 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5520 N_("prepend an additional prefix to every line of output"),
5521 PARSE_OPT_NONEG, diff_opt_line_prefix),
5522 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5523 N_("do not show any source or destination prefix"),
5524 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5525 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5526 N_("show context between diff hunks up to the specified number of lines"),
5527 PARSE_OPT_NONEG),
5528 OPT_CALLBACK_F(0, "output-indicator-new",
5529 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5530 N_("<char>"),
5531 N_("specify the character to indicate a new line instead of '+'"),
5532 PARSE_OPT_NONEG, diff_opt_char),
5533 OPT_CALLBACK_F(0, "output-indicator-old",
5534 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5535 N_("<char>"),
5536 N_("specify the character to indicate an old line instead of '-'"),
5537 PARSE_OPT_NONEG, diff_opt_char),
5538 OPT_CALLBACK_F(0, "output-indicator-context",
5539 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5540 N_("<char>"),
5541 N_("specify the character to indicate a context instead of ' '"),
5542 PARSE_OPT_NONEG, diff_opt_char),
5543
5544 OPT_GROUP(N_("Diff rename options")),
5545 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5546 N_("break complete rewrite changes into pairs of delete and create"),
5547 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5548 diff_opt_break_rewrites),
5549 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5550 N_("detect renames"),
5551 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5552 diff_opt_find_renames),
5553 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5554 N_("omit the preimage for deletes"),
5555 1, PARSE_OPT_NONEG),
5556 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5557 N_("detect copies"),
5558 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5559 diff_opt_find_copies),
5560 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5561 N_("use unmodified files as source to find copies")),
5562 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5563 N_("disable rename detection"),
5564 0, PARSE_OPT_NONEG),
5565 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5566 N_("use empty blobs as rename source")),
5567 OPT_CALLBACK_F(0, "follow", options, NULL,
5568 N_("continue listing the history of a file beyond renames"),
5569 PARSE_OPT_NOARG, diff_opt_follow),
5570 OPT_INTEGER('l', NULL, &options->rename_limit,
5571 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5572
5573 OPT_GROUP(N_("Diff algorithm options")),
5574 OPT_BIT(0, "minimal", &options->xdl_opts,
5575 N_("produce the smallest possible diff"),
5576 XDF_NEED_MINIMAL),
5577 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5578 N_("ignore whitespace when comparing lines"),
5579 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5580 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5581 N_("ignore changes in amount of whitespace"),
5582 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5583 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5584 N_("ignore changes in whitespace at EOL"),
5585 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5586 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5587 N_("ignore carrier-return at the end of line"),
5588 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5589 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5590 N_("ignore changes whose lines are all blank"),
5591 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5592 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5593 N_("ignore changes whose all lines match <regex>"),
5594 0, diff_opt_ignore_regex),
5595 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5596 N_("heuristic to shift diff hunk boundaries for easy reading"),
5597 XDF_INDENT_HEURISTIC),
5598 OPT_CALLBACK_F(0, "patience", options, NULL,
5599 N_("generate diff using the \"patience diff\" algorithm"),
5600 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5601 diff_opt_patience),
5602 OPT_BITOP(0, "histogram", &options->xdl_opts,
5603 N_("generate diff using the \"histogram diff\" algorithm"),
5604 XDF_HISTOGRAM_DIFF, XDF_DIFF_ALGORITHM_MASK),
5605 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5606 N_("choose a diff algorithm"),
5607 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5608 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5609 N_("generate diff using the \"anchored diff\" algorithm"),
5610 PARSE_OPT_NONEG, diff_opt_anchored),
5611 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5612 N_("show word diff, using <mode> to delimit changed words"),
5613 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5614 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5615 N_("use <regex> to decide what a word is"),
5616 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5617 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5618 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5619 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5620 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5621 N_("moved lines of code are colored differently"),
5622 PARSE_OPT_OPTARG, diff_opt_color_moved),
5623 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5624 N_("how white spaces are ignored in --color-moved"),
5625 0, diff_opt_color_moved_ws),
5626
5627 OPT_GROUP(N_("Other diff options")),
5628 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5629 N_("when run from subdir, exclude changes outside and show relative paths"),
5630 PARSE_OPT_OPTARG,
5631 diff_opt_relative),
5632 OPT_BOOL('a', "text", &options->flags.text,
5633 N_("treat all files as text")),
5634 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5635 N_("swap two inputs, reverse the diff")),
5636 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5637 N_("exit with 1 if there were differences, 0 otherwise")),
5638 OPT_BOOL(0, "quiet", &options->flags.quick,
5639 N_("disable all output of the program")),
5640 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5641 N_("allow an external diff helper to be executed")),
5642 OPT_CALLBACK_F(0, "textconv", options, NULL,
5643 N_("run external text conversion filters when comparing binary files"),
5644 PARSE_OPT_NOARG, diff_opt_textconv),
5645 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5646 N_("ignore changes to submodules in the diff generation"),
5647 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5648 diff_opt_ignore_submodules),
5649 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5650 N_("specify how differences in submodules are shown"),
5651 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5652 diff_opt_submodule),
5653 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5654 N_("hide 'git add -N' entries from the index"),
5655 1, PARSE_OPT_NONEG),
5656 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5657 N_("treat 'git add -N' entries as real in the index"),
5658 0, PARSE_OPT_NONEG),
5659 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5660 N_("look for differences that change the number of occurrences of the specified string"),
5661 0, diff_opt_pickaxe_string),
5662 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5663 N_("look for differences that change the number of occurrences of the specified regex"),
5664 0, diff_opt_pickaxe_regex),
5665 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5666 N_("show all changes in the changeset with -S or -G"),
5667 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5668 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5669 N_("treat <string> in -S as extended POSIX regular expression"),
5670 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5671 OPT_FILENAME('O', NULL, &options->orderfile,
5672 N_("control the order in which files appear in the output")),
5673 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5674 N_("show the change in the specified path first"),
5675 PARSE_OPT_NONEG, diff_opt_rotate_to),
5676 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5677 N_("skip the output to the specified path"),
5678 PARSE_OPT_NONEG, diff_opt_rotate_to),
5679 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5680 N_("look for differences that change the number of occurrences of the specified object"),
5681 PARSE_OPT_NONEG, diff_opt_find_object),
5682 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5683 N_("select files by diff type"),
5684 PARSE_OPT_NONEG, diff_opt_diff_filter),
5685 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5686 N_("output to a specific file"),
5687 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5688
5689 OPT_END()
5690 };
5691
5692 ALLOC_ARRAY(options->parseopts, ARRAY_SIZE(parseopts));
5693 memcpy(options->parseopts, parseopts, sizeof(parseopts));
5694 }
5695
5696 int diff_opt_parse(struct diff_options *options,
5697 const char **av, int ac, const char *prefix)
5698 {
5699 if (!prefix)
5700 prefix = "";
5701
5702 ac = parse_options(ac, av, prefix, options->parseopts, NULL,
5703 PARSE_OPT_KEEP_DASHDASH |
5704 PARSE_OPT_KEEP_UNKNOWN_OPT |
5705 PARSE_OPT_NO_INTERNAL_HELP |
5706 PARSE_OPT_ONE_SHOT |
5707 PARSE_OPT_STOP_AT_NON_OPTION);
5708
5709 return ac;
5710 }
5711
5712 int parse_rename_score(const char **cp_p)
5713 {
5714 unsigned long num, scale;
5715 int ch, dot;
5716 const char *cp = *cp_p;
5717
5718 num = 0;
5719 scale = 1;
5720 dot = 0;
5721 for (;;) {
5722 ch = *cp;
5723 if ( !dot && ch == '.' ) {
5724 scale = 1;
5725 dot = 1;
5726 } else if ( ch == '%' ) {
5727 scale = dot ? scale*100 : 100;
5728 cp++; /* % is always at the end */
5729 break;
5730 } else if ( ch >= '0' && ch <= '9' ) {
5731 if ( scale < 100000 ) {
5732 scale *= 10;
5733 num = (num*10) + (ch-'0');
5734 }
5735 } else {
5736 break;
5737 }
5738 cp++;
5739 }
5740 *cp_p = cp;
5741
5742 /* user says num divided by scale and we say internally that
5743 * is MAX_SCORE * num / scale.
5744 */
5745 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5746 }
5747
5748 struct diff_queue_struct diff_queued_diff;
5749
5750 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5751 {
5752 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5753 queue->queue[queue->nr++] = dp;
5754 }
5755
5756 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5757 struct diff_filespec *one,
5758 struct diff_filespec *two)
5759 {
5760 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5761 dp->one = one;
5762 dp->two = two;
5763 if (queue)
5764 diff_q(queue, dp);
5765 return dp;
5766 }
5767
5768 void diff_free_filepair(struct diff_filepair *p)
5769 {
5770 free_filespec(p->one);
5771 free_filespec(p->two);
5772 free(p);
5773 }
5774
5775 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5776 {
5777 int abblen;
5778 const char *abbrev;
5779
5780 /* Do we want all 40 hex characters? */
5781 if (len == the_hash_algo->hexsz)
5782 return oid_to_hex(oid);
5783
5784 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5785 abbrev = diff_abbrev_oid(oid, len);
5786
5787 if (!print_sha1_ellipsis())
5788 return abbrev;
5789
5790 abblen = strlen(abbrev);
5791
5792 /*
5793 * In well-behaved cases, where the abbreviated result is the
5794 * same as the requested length, append three dots after the
5795 * abbreviation (hence the whole logic is limited to the case
5796 * where abblen < 37); when the actual abbreviated result is a
5797 * bit longer than the requested length, we reduce the number
5798 * of dots so that they match the well-behaved ones. However,
5799 * if the actual abbreviation is longer than the requested
5800 * length by more than three, we give up on aligning, and add
5801 * three dots anyway, to indicate that the output is not the
5802 * full object name. Yes, this may be suboptimal, but this
5803 * appears only in "diff --raw --abbrev" output and it is not
5804 * worth the effort to change it now. Note that this would
5805 * likely to work fine when the automatic sizing of default
5806 * abbreviation length is used--we would be fed -1 in "len" in
5807 * that case, and will end up always appending three-dots, but
5808 * the automatic sizing is supposed to give abblen that ensures
5809 * uniqueness across all objects (statistically speaking).
5810 */
5811 if (abblen < the_hash_algo->hexsz - 3) {
5812 static char hex[GIT_MAX_HEXSZ + 1];
5813 if (len < abblen && abblen <= len + 2)
5814 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5815 else
5816 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5817 return hex;
5818 }
5819
5820 return oid_to_hex(oid);
5821 }
5822
5823 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5824 {
5825 int line_termination = opt->line_termination;
5826 int inter_name_termination = line_termination ? '\t' : '\0';
5827
5828 fprintf(opt->file, "%s", diff_line_prefix(opt));
5829 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5830 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5831 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5832 fprintf(opt->file, "%s ",
5833 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5834 }
5835 if (p->score) {
5836 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5837 inter_name_termination);
5838 } else {
5839 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5840 }
5841
5842 if (p->status == DIFF_STATUS_COPIED ||
5843 p->status == DIFF_STATUS_RENAMED) {
5844 const char *name_a, *name_b;
5845 name_a = p->one->path;
5846 name_b = p->two->path;
5847 strip_prefix(opt->prefix_length, &name_a, &name_b);
5848 write_name_quoted(name_a, opt->file, inter_name_termination);
5849 write_name_quoted(name_b, opt->file, line_termination);
5850 } else {
5851 const char *name_a, *name_b;
5852 name_a = p->one->mode ? p->one->path : p->two->path;
5853 name_b = NULL;
5854 strip_prefix(opt->prefix_length, &name_a, &name_b);
5855 write_name_quoted(name_a, opt->file, line_termination);
5856 }
5857 }
5858
5859 int diff_unmodified_pair(struct diff_filepair *p)
5860 {
5861 /* This function is written stricter than necessary to support
5862 * the currently implemented transformers, but the idea is to
5863 * let transformers to produce diff_filepairs any way they want,
5864 * and filter and clean them up here before producing the output.
5865 */
5866 struct diff_filespec *one = p->one, *two = p->two;
5867
5868 if (DIFF_PAIR_UNMERGED(p))
5869 return 0; /* unmerged is interesting */
5870
5871 /* deletion, addition, mode or type change
5872 * and rename are all interesting.
5873 */
5874 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5875 DIFF_PAIR_MODE_CHANGED(p) ||
5876 strcmp(one->path, two->path))
5877 return 0;
5878
5879 /* both are valid and point at the same path. that is, we are
5880 * dealing with a change.
5881 */
5882 if (one->oid_valid && two->oid_valid &&
5883 oideq(&one->oid, &two->oid) &&
5884 !one->dirty_submodule && !two->dirty_submodule)
5885 return 1; /* no change */
5886 if (!one->oid_valid && !two->oid_valid)
5887 return 1; /* both look at the same file on the filesystem. */
5888 return 0;
5889 }
5890
5891 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5892 {
5893 int include_conflict_headers =
5894 (additional_headers(o, p->one->path) &&
5895 !o->pickaxe_opts &&
5896 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
5897
5898 /*
5899 * Check if we can return early without showing a diff. Note that
5900 * diff_filepair only stores {oid, path, mode, is_valid}
5901 * information for each path, and thus diff_unmodified_pair() only
5902 * considers those bits of info. However, we do not want pairs
5903 * created by create_filepairs_for_header_only_notifications()
5904 * (which always look like unmodified pairs) to be ignored, so
5905 * return early if both p is unmodified AND we don't want to
5906 * include_conflict_headers.
5907 */
5908 if (diff_unmodified_pair(p) && !include_conflict_headers)
5909 return;
5910
5911 /* Actually, we can also return early to avoid showing tree diffs */
5912 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5913 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5914 return;
5915
5916 run_diff(p, o);
5917 }
5918
5919 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5920 struct diffstat_t *diffstat)
5921 {
5922 if (diff_unmodified_pair(p))
5923 return;
5924
5925 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5926 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5927 return; /* no useful stat for tree diffs */
5928
5929 run_diffstat(p, o, diffstat);
5930 }
5931
5932 static void diff_flush_checkdiff(struct diff_filepair *p,
5933 struct diff_options *o)
5934 {
5935 if (diff_unmodified_pair(p))
5936 return;
5937
5938 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5939 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5940 return; /* nothing to check in tree diffs */
5941
5942 run_checkdiff(p, o);
5943 }
5944
5945 int diff_queue_is_empty(struct diff_options *o)
5946 {
5947 struct diff_queue_struct *q = &diff_queued_diff;
5948 int i;
5949 int include_conflict_headers =
5950 (o->additional_path_headers &&
5951 strmap_get_size(o->additional_path_headers) &&
5952 !o->pickaxe_opts &&
5953 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
5954
5955 if (include_conflict_headers)
5956 return 0;
5957
5958 for (i = 0; i < q->nr; i++)
5959 if (!diff_unmodified_pair(q->queue[i]))
5960 return 0;
5961 return 1;
5962 }
5963
5964 #if DIFF_DEBUG
5965 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5966 {
5967 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5968 x, one ? one : "",
5969 s->path,
5970 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5971 s->mode,
5972 s->oid_valid ? oid_to_hex(&s->oid) : "");
5973 fprintf(stderr, "queue[%d] %s size %lu\n",
5974 x, one ? one : "",
5975 s->size);
5976 }
5977
5978 void diff_debug_filepair(const struct diff_filepair *p, int i)
5979 {
5980 diff_debug_filespec(p->one, i, "one");
5981 diff_debug_filespec(p->two, i, "two");
5982 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5983 p->score, p->status ? p->status : '?',
5984 p->one->rename_used, p->broken_pair);
5985 }
5986
5987 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5988 {
5989 int i;
5990 if (msg)
5991 fprintf(stderr, "%s\n", msg);
5992 fprintf(stderr, "q->nr = %d\n", q->nr);
5993 for (i = 0; i < q->nr; i++) {
5994 struct diff_filepair *p = q->queue[i];
5995 diff_debug_filepair(p, i);
5996 }
5997 }
5998 #endif
5999
6000 static void diff_resolve_rename_copy(void)
6001 {
6002 int i;
6003 struct diff_filepair *p;
6004 struct diff_queue_struct *q = &diff_queued_diff;
6005
6006 diff_debug_queue("resolve-rename-copy", q);
6007
6008 for (i = 0; i < q->nr; i++) {
6009 p = q->queue[i];
6010 p->status = 0; /* undecided */
6011 if (DIFF_PAIR_UNMERGED(p))
6012 p->status = DIFF_STATUS_UNMERGED;
6013 else if (!DIFF_FILE_VALID(p->one))
6014 p->status = DIFF_STATUS_ADDED;
6015 else if (!DIFF_FILE_VALID(p->two))
6016 p->status = DIFF_STATUS_DELETED;
6017 else if (DIFF_PAIR_TYPE_CHANGED(p))
6018 p->status = DIFF_STATUS_TYPE_CHANGED;
6019
6020 /* from this point on, we are dealing with a pair
6021 * whose both sides are valid and of the same type, i.e.
6022 * either in-place edit or rename/copy edit.
6023 */
6024 else if (DIFF_PAIR_RENAME(p)) {
6025 /*
6026 * A rename might have re-connected a broken
6027 * pair up, causing the pathnames to be the
6028 * same again. If so, that's not a rename at
6029 * all, just a modification..
6030 *
6031 * Otherwise, see if this source was used for
6032 * multiple renames, in which case we decrement
6033 * the count, and call it a copy.
6034 */
6035 if (!strcmp(p->one->path, p->two->path))
6036 p->status = DIFF_STATUS_MODIFIED;
6037 else if (--p->one->rename_used > 0)
6038 p->status = DIFF_STATUS_COPIED;
6039 else
6040 p->status = DIFF_STATUS_RENAMED;
6041 }
6042 else if (!oideq(&p->one->oid, &p->two->oid) ||
6043 p->one->mode != p->two->mode ||
6044 p->one->dirty_submodule ||
6045 p->two->dirty_submodule ||
6046 is_null_oid(&p->one->oid))
6047 p->status = DIFF_STATUS_MODIFIED;
6048 else {
6049 /* This is a "no-change" entry and should not
6050 * happen anymore, but prepare for broken callers.
6051 */
6052 error("feeding unmodified %s to diffcore",
6053 p->one->path);
6054 p->status = DIFF_STATUS_UNKNOWN;
6055 }
6056 }
6057 diff_debug_queue("resolve-rename-copy done", q);
6058 }
6059
6060 static int check_pair_status(struct diff_filepair *p)
6061 {
6062 switch (p->status) {
6063 case DIFF_STATUS_UNKNOWN:
6064 return 0;
6065 case 0:
6066 die("internal error in diff-resolve-rename-copy");
6067 default:
6068 return 1;
6069 }
6070 }
6071
6072 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6073 {
6074 int fmt = opt->output_format;
6075
6076 if (fmt & DIFF_FORMAT_CHECKDIFF)
6077 diff_flush_checkdiff(p, opt);
6078 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6079 diff_flush_raw(p, opt);
6080 else if (fmt & DIFF_FORMAT_NAME) {
6081 const char *name_a, *name_b;
6082 name_a = p->two->path;
6083 name_b = NULL;
6084 strip_prefix(opt->prefix_length, &name_a, &name_b);
6085 fprintf(opt->file, "%s", diff_line_prefix(opt));
6086 write_name_quoted(name_a, opt->file, opt->line_termination);
6087 }
6088 }
6089
6090 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6091 {
6092 struct strbuf sb = STRBUF_INIT;
6093 if (fs->mode)
6094 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6095 else
6096 strbuf_addf(&sb, " %s ", newdelete);
6097
6098 quote_c_style(fs->path, &sb, NULL, 0);
6099 strbuf_addch(&sb, '\n');
6100 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6101 sb.buf, sb.len, 0);
6102 strbuf_release(&sb);
6103 }
6104
6105 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6106 int show_name)
6107 {
6108 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6109 struct strbuf sb = STRBUF_INIT;
6110 strbuf_addf(&sb, " mode change %06o => %06o",
6111 p->one->mode, p->two->mode);
6112 if (show_name) {
6113 strbuf_addch(&sb, ' ');
6114 quote_c_style(p->two->path, &sb, NULL, 0);
6115 }
6116 strbuf_addch(&sb, '\n');
6117 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6118 sb.buf, sb.len, 0);
6119 strbuf_release(&sb);
6120 }
6121 }
6122
6123 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6124 struct diff_filepair *p)
6125 {
6126 struct strbuf sb = STRBUF_INIT;
6127 struct strbuf names = STRBUF_INIT;
6128
6129 pprint_rename(&names, p->one->path, p->two->path);
6130 strbuf_addf(&sb, " %s %s (%d%%)\n",
6131 renamecopy, names.buf, similarity_index(p));
6132 strbuf_release(&names);
6133 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6134 sb.buf, sb.len, 0);
6135 show_mode_change(opt, p, 0);
6136 strbuf_release(&sb);
6137 }
6138
6139 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6140 {
6141 switch(p->status) {
6142 case DIFF_STATUS_DELETED:
6143 show_file_mode_name(opt, "delete", p->one);
6144 break;
6145 case DIFF_STATUS_ADDED:
6146 show_file_mode_name(opt, "create", p->two);
6147 break;
6148 case DIFF_STATUS_COPIED:
6149 show_rename_copy(opt, "copy", p);
6150 break;
6151 case DIFF_STATUS_RENAMED:
6152 show_rename_copy(opt, "rename", p);
6153 break;
6154 default:
6155 if (p->score) {
6156 struct strbuf sb = STRBUF_INIT;
6157 strbuf_addstr(&sb, " rewrite ");
6158 quote_c_style(p->two->path, &sb, NULL, 0);
6159 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6160 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6161 sb.buf, sb.len, 0);
6162 strbuf_release(&sb);
6163 }
6164 show_mode_change(opt, p, !p->score);
6165 break;
6166 }
6167 }
6168
6169 struct patch_id_t {
6170 git_hash_ctx *ctx;
6171 int patchlen;
6172 };
6173
6174 static int remove_space(char *line, int len)
6175 {
6176 int i;
6177 char *dst = line;
6178 unsigned char c;
6179
6180 for (i = 0; i < len; i++)
6181 if (!isspace((c = line[i])))
6182 *dst++ = c;
6183
6184 return dst - line;
6185 }
6186
6187 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6188 {
6189 unsigned char hash[GIT_MAX_RAWSZ];
6190 unsigned short carry = 0;
6191 int i;
6192
6193 the_hash_algo->final_fn(hash, ctx);
6194 the_hash_algo->init_fn(ctx);
6195 /* 20-byte sum, with carry */
6196 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6197 carry += result->hash[i] + hash[i];
6198 result->hash[i] = carry;
6199 carry >>= 8;
6200 }
6201 }
6202
6203 static int patch_id_consume(void *priv, char *line, unsigned long len)
6204 {
6205 struct patch_id_t *data = priv;
6206 int new_len;
6207
6208 if (len > 12 && starts_with(line, "\\ "))
6209 return 0;
6210 new_len = remove_space(line, len);
6211
6212 the_hash_algo->update_fn(data->ctx, line, new_len);
6213 data->patchlen += new_len;
6214 return 0;
6215 }
6216
6217 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6218 {
6219 the_hash_algo->update_fn(ctx, str, strlen(str));
6220 }
6221
6222 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6223 {
6224 /* large enough for 2^32 in octal */
6225 char buf[12];
6226 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6227 the_hash_algo->update_fn(ctx, buf, len);
6228 }
6229
6230 /* returns 0 upon success, and writes result into oid */
6231 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6232 {
6233 struct diff_queue_struct *q = &diff_queued_diff;
6234 int i;
6235 git_hash_ctx ctx;
6236 struct patch_id_t data;
6237
6238 the_hash_algo->init_fn(&ctx);
6239 memset(&data, 0, sizeof(struct patch_id_t));
6240 data.ctx = &ctx;
6241 oidclr(oid);
6242
6243 for (i = 0; i < q->nr; i++) {
6244 xpparam_t xpp;
6245 xdemitconf_t xecfg;
6246 mmfile_t mf1, mf2;
6247 struct diff_filepair *p = q->queue[i];
6248 int len1, len2;
6249
6250 memset(&xpp, 0, sizeof(xpp));
6251 memset(&xecfg, 0, sizeof(xecfg));
6252 if (p->status == 0)
6253 return error("internal diff status error");
6254 if (p->status == DIFF_STATUS_UNKNOWN)
6255 continue;
6256 if (diff_unmodified_pair(p))
6257 continue;
6258 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6259 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6260 continue;
6261 if (DIFF_PAIR_UNMERGED(p))
6262 continue;
6263
6264 diff_fill_oid_info(p->one, options->repo->index);
6265 diff_fill_oid_info(p->two, options->repo->index);
6266
6267 len1 = remove_space(p->one->path, strlen(p->one->path));
6268 len2 = remove_space(p->two->path, strlen(p->two->path));
6269 patch_id_add_string(&ctx, "diff--git");
6270 patch_id_add_string(&ctx, "a/");
6271 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6272 patch_id_add_string(&ctx, "b/");
6273 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6274
6275 if (p->one->mode == 0) {
6276 patch_id_add_string(&ctx, "newfilemode");
6277 patch_id_add_mode(&ctx, p->two->mode);
6278 } else if (p->two->mode == 0) {
6279 patch_id_add_string(&ctx, "deletedfilemode");
6280 patch_id_add_mode(&ctx, p->one->mode);
6281 } else if (p->one->mode != p->two->mode) {
6282 patch_id_add_string(&ctx, "oldmode");
6283 patch_id_add_mode(&ctx, p->one->mode);
6284 patch_id_add_string(&ctx, "newmode");
6285 patch_id_add_mode(&ctx, p->two->mode);
6286 }
6287
6288 if (diff_header_only) {
6289 /* don't do anything since we're only populating header info */
6290 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6291 diff_filespec_is_binary(options->repo, p->two)) {
6292 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6293 the_hash_algo->hexsz);
6294 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6295 the_hash_algo->hexsz);
6296 } else {
6297 if (p->one->mode == 0) {
6298 patch_id_add_string(&ctx, "---/dev/null");
6299 patch_id_add_string(&ctx, "+++b/");
6300 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6301 } else if (p->two->mode == 0) {
6302 patch_id_add_string(&ctx, "---a/");
6303 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6304 patch_id_add_string(&ctx, "+++/dev/null");
6305 } else {
6306 patch_id_add_string(&ctx, "---a/");
6307 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6308 patch_id_add_string(&ctx, "+++b/");
6309 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6310 }
6311
6312 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6313 fill_mmfile(options->repo, &mf2, p->two) < 0)
6314 return error("unable to read files to diff");
6315 xpp.flags = 0;
6316 xecfg.ctxlen = 3;
6317 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6318 if (xdi_diff_outf(&mf1, &mf2, NULL,
6319 patch_id_consume, &data, &xpp, &xecfg))
6320 return error("unable to generate patch-id diff for %s",
6321 p->one->path);
6322 }
6323 flush_one_hunk(oid, &ctx);
6324 }
6325
6326 return 0;
6327 }
6328
6329 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6330 {
6331 struct diff_queue_struct *q = &diff_queued_diff;
6332 int i;
6333 int result = diff_get_patch_id(options, oid, diff_header_only);
6334
6335 for (i = 0; i < q->nr; i++)
6336 diff_free_filepair(q->queue[i]);
6337
6338 free(q->queue);
6339 DIFF_QUEUE_CLEAR(q);
6340
6341 return result;
6342 }
6343
6344 static int is_summary_empty(const struct diff_queue_struct *q)
6345 {
6346 int i;
6347
6348 for (i = 0; i < q->nr; i++) {
6349 const struct diff_filepair *p = q->queue[i];
6350
6351 switch (p->status) {
6352 case DIFF_STATUS_DELETED:
6353 case DIFF_STATUS_ADDED:
6354 case DIFF_STATUS_COPIED:
6355 case DIFF_STATUS_RENAMED:
6356 return 0;
6357 default:
6358 if (p->score)
6359 return 0;
6360 if (p->one->mode && p->two->mode &&
6361 p->one->mode != p->two->mode)
6362 return 0;
6363 break;
6364 }
6365 }
6366 return 1;
6367 }
6368
6369 static const char rename_limit_warning[] =
6370 N_("exhaustive rename detection was skipped due to too many files.");
6371
6372 static const char degrade_cc_to_c_warning[] =
6373 N_("only found copies from modified paths due to too many files.");
6374
6375 static const char rename_limit_advice[] =
6376 N_("you may want to set your %s variable to at least "
6377 "%d and retry the command.");
6378
6379 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6380 {
6381 fflush(stdout);
6382 if (degraded_cc)
6383 warning(_(degrade_cc_to_c_warning));
6384 else if (needed)
6385 warning(_(rename_limit_warning));
6386 else
6387 return;
6388 if (0 < needed)
6389 warning(_(rename_limit_advice), varname, needed);
6390 }
6391
6392 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6393 {
6394 struct strset present;
6395 struct diff_queue_struct *q = &diff_queued_diff;
6396 struct hashmap_iter iter;
6397 struct strmap_entry *e;
6398 int i;
6399
6400 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6401
6402 /*
6403 * Find out which paths exist in diff_queued_diff, preferring
6404 * one->path for any pair that has multiple paths.
6405 */
6406 for (i = 0; i < q->nr; i++) {
6407 struct diff_filepair *p = q->queue[i];
6408 char *path = p->one->path ? p->one->path : p->two->path;
6409
6410 if (strmap_contains(o->additional_path_headers, path))
6411 strset_add(&present, path);
6412 }
6413
6414 /*
6415 * Loop over paths in additional_path_headers; for each NOT already
6416 * in diff_queued_diff, create a synthetic filepair and insert that
6417 * into diff_queued_diff.
6418 */
6419 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6420 if (!strset_contains(&present, e->key)) {
6421 struct diff_filespec *one, *two;
6422 struct diff_filepair *p;
6423
6424 one = alloc_filespec(e->key);
6425 two = alloc_filespec(e->key);
6426 fill_filespec(one, null_oid(), 0, 0);
6427 fill_filespec(two, null_oid(), 0, 0);
6428 p = diff_queue(q, one, two);
6429 p->status = DIFF_STATUS_MODIFIED;
6430 }
6431 }
6432
6433 /* Re-sort the filepairs */
6434 diffcore_fix_diff_index();
6435
6436 /* Cleanup */
6437 strset_clear(&present);
6438 }
6439
6440 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6441 {
6442 int i;
6443 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6444 struct diff_queue_struct *q = &diff_queued_diff;
6445
6446 if (WSEH_NEW & WS_RULE_MASK)
6447 BUG("WS rules bit mask overlaps with diff symbol flags");
6448
6449 if (o->color_moved)
6450 o->emitted_symbols = &esm;
6451
6452 if (o->additional_path_headers)
6453 create_filepairs_for_header_only_notifications(o);
6454
6455 for (i = 0; i < q->nr; i++) {
6456 struct diff_filepair *p = q->queue[i];
6457 if (check_pair_status(p))
6458 diff_flush_patch(p, o);
6459 }
6460
6461 if (o->emitted_symbols) {
6462 if (o->color_moved) {
6463 struct mem_pool entry_pool;
6464 struct moved_entry_list *entry_list;
6465
6466 mem_pool_init(&entry_pool, 1024 * 1024);
6467 entry_list = add_lines_to_move_detection(o,
6468 &entry_pool);
6469 mark_color_as_moved(o, entry_list);
6470 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6471 dim_moved_lines(o);
6472
6473 mem_pool_discard(&entry_pool, 0);
6474 free(entry_list);
6475 }
6476
6477 for (i = 0; i < esm.nr; i++)
6478 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6479
6480 for (i = 0; i < esm.nr; i++)
6481 free((void *)esm.buf[i].line);
6482 esm.nr = 0;
6483
6484 o->emitted_symbols = NULL;
6485 }
6486 }
6487
6488 static void diff_free_file(struct diff_options *options)
6489 {
6490 if (options->close_file)
6491 fclose(options->file);
6492 }
6493
6494 static void diff_free_ignore_regex(struct diff_options *options)
6495 {
6496 int i;
6497
6498 for (i = 0; i < options->ignore_regex_nr; i++) {
6499 regfree(options->ignore_regex[i]);
6500 free(options->ignore_regex[i]);
6501 }
6502 free(options->ignore_regex);
6503 }
6504
6505 void diff_free(struct diff_options *options)
6506 {
6507 if (options->no_free)
6508 return;
6509
6510 diff_free_file(options);
6511 diff_free_ignore_regex(options);
6512 clear_pathspec(&options->pathspec);
6513 FREE_AND_NULL(options->parseopts);
6514 }
6515
6516 void diff_flush(struct diff_options *options)
6517 {
6518 struct diff_queue_struct *q = &diff_queued_diff;
6519 int i, output_format = options->output_format;
6520 int separator = 0;
6521 int dirstat_by_line = 0;
6522
6523 /*
6524 * Order: raw, stat, summary, patch
6525 * or: name/name-status/checkdiff (other bits clear)
6526 */
6527 if (!q->nr && !options->additional_path_headers)
6528 goto free_queue;
6529
6530 if (output_format & (DIFF_FORMAT_RAW |
6531 DIFF_FORMAT_NAME |
6532 DIFF_FORMAT_NAME_STATUS |
6533 DIFF_FORMAT_CHECKDIFF)) {
6534 for (i = 0; i < q->nr; i++) {
6535 struct diff_filepair *p = q->queue[i];
6536 if (check_pair_status(p))
6537 flush_one_pair(p, options);
6538 }
6539 separator++;
6540 }
6541
6542 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6543 dirstat_by_line = 1;
6544
6545 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6546 dirstat_by_line) {
6547 struct diffstat_t diffstat;
6548
6549 compute_diffstat(options, &diffstat, q);
6550 if (output_format & DIFF_FORMAT_NUMSTAT)
6551 show_numstat(&diffstat, options);
6552 if (output_format & DIFF_FORMAT_DIFFSTAT)
6553 show_stats(&diffstat, options);
6554 if (output_format & DIFF_FORMAT_SHORTSTAT)
6555 show_shortstats(&diffstat, options);
6556 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6557 show_dirstat_by_line(&diffstat, options);
6558 free_diffstat_info(&diffstat);
6559 separator++;
6560 }
6561 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6562 show_dirstat(options);
6563
6564 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6565 for (i = 0; i < q->nr; i++) {
6566 diff_summary(options, q->queue[i]);
6567 }
6568 separator++;
6569 }
6570
6571 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6572 options->flags.exit_with_status &&
6573 options->flags.diff_from_contents) {
6574 /*
6575 * run diff_flush_patch for the exit status. setting
6576 * options->file to /dev/null should be safe, because we
6577 * aren't supposed to produce any output anyway.
6578 */
6579 diff_free_file(options);
6580 options->file = xfopen("/dev/null", "w");
6581 options->close_file = 1;
6582 options->color_moved = 0;
6583 for (i = 0; i < q->nr; i++) {
6584 struct diff_filepair *p = q->queue[i];
6585 if (check_pair_status(p))
6586 diff_flush_patch(p, options);
6587 if (options->found_changes)
6588 break;
6589 }
6590 }
6591
6592 if (output_format & DIFF_FORMAT_PATCH) {
6593 if (separator) {
6594 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6595 if (options->stat_sep)
6596 /* attach patch instead of inline */
6597 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6598 NULL, 0, 0);
6599 }
6600
6601 diff_flush_patch_all_file_pairs(options);
6602 }
6603
6604 if (output_format & DIFF_FORMAT_CALLBACK)
6605 options->format_callback(q, options, options->format_callback_data);
6606
6607 for (i = 0; i < q->nr; i++)
6608 diff_free_filepair(q->queue[i]);
6609 free_queue:
6610 free(q->queue);
6611 DIFF_QUEUE_CLEAR(q);
6612 diff_free(options);
6613
6614 /*
6615 * Report the content-level differences with HAS_CHANGES;
6616 * diff_addremove/diff_change does not set the bit when
6617 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6618 */
6619 if (options->flags.diff_from_contents) {
6620 if (options->found_changes)
6621 options->flags.has_changes = 1;
6622 else
6623 options->flags.has_changes = 0;
6624 }
6625 }
6626
6627 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6628 {
6629 return (((p->status == DIFF_STATUS_MODIFIED) &&
6630 ((p->score &&
6631 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6632 (!p->score &&
6633 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6634 ((p->status != DIFF_STATUS_MODIFIED) &&
6635 filter_bit_tst(p->status, options)));
6636 }
6637
6638 static void diffcore_apply_filter(struct diff_options *options)
6639 {
6640 int i;
6641 struct diff_queue_struct *q = &diff_queued_diff;
6642 struct diff_queue_struct outq;
6643
6644 DIFF_QUEUE_CLEAR(&outq);
6645
6646 if (!options->filter)
6647 return;
6648
6649 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6650 int found;
6651 for (i = found = 0; !found && i < q->nr; i++) {
6652 if (match_filter(options, q->queue[i]))
6653 found++;
6654 }
6655 if (found)
6656 return;
6657
6658 /* otherwise we will clear the whole queue
6659 * by copying the empty outq at the end of this
6660 * function, but first clear the current entries
6661 * in the queue.
6662 */
6663 for (i = 0; i < q->nr; i++)
6664 diff_free_filepair(q->queue[i]);
6665 }
6666 else {
6667 /* Only the matching ones */
6668 for (i = 0; i < q->nr; i++) {
6669 struct diff_filepair *p = q->queue[i];
6670 if (match_filter(options, p))
6671 diff_q(&outq, p);
6672 else
6673 diff_free_filepair(p);
6674 }
6675 }
6676 free(q->queue);
6677 *q = outq;
6678 }
6679
6680 /* Check whether two filespecs with the same mode and size are identical */
6681 static int diff_filespec_is_identical(struct repository *r,
6682 struct diff_filespec *one,
6683 struct diff_filespec *two)
6684 {
6685 if (S_ISGITLINK(one->mode))
6686 return 0;
6687 if (diff_populate_filespec(r, one, NULL))
6688 return 0;
6689 if (diff_populate_filespec(r, two, NULL))
6690 return 0;
6691 return !memcmp(one->data, two->data, one->size);
6692 }
6693
6694 static int diff_filespec_check_stat_unmatch(struct repository *r,
6695 struct diff_filepair *p)
6696 {
6697 struct diff_populate_filespec_options dpf_options = {
6698 .check_size_only = 1,
6699 .missing_object_cb = diff_queued_diff_prefetch,
6700 .missing_object_data = r,
6701 };
6702
6703 if (p->done_skip_stat_unmatch)
6704 return p->skip_stat_unmatch_result;
6705
6706 p->done_skip_stat_unmatch = 1;
6707 p->skip_stat_unmatch_result = 0;
6708 /*
6709 * 1. Entries that come from stat info dirtiness
6710 * always have both sides (iow, not create/delete),
6711 * one side of the object name is unknown, with
6712 * the same mode and size. Keep the ones that
6713 * do not match these criteria. They have real
6714 * differences.
6715 *
6716 * 2. At this point, the file is known to be modified,
6717 * with the same mode and size, and the object
6718 * name of one side is unknown. Need to inspect
6719 * the identical contents.
6720 */
6721 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6722 !DIFF_FILE_VALID(p->two) ||
6723 (p->one->oid_valid && p->two->oid_valid) ||
6724 (p->one->mode != p->two->mode) ||
6725 diff_populate_filespec(r, p->one, &dpf_options) ||
6726 diff_populate_filespec(r, p->two, &dpf_options) ||
6727 (p->one->size != p->two->size) ||
6728 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6729 p->skip_stat_unmatch_result = 1;
6730 return p->skip_stat_unmatch_result;
6731 }
6732
6733 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6734 {
6735 int i;
6736 struct diff_queue_struct *q = &diff_queued_diff;
6737 struct diff_queue_struct outq;
6738 DIFF_QUEUE_CLEAR(&outq);
6739
6740 for (i = 0; i < q->nr; i++) {
6741 struct diff_filepair *p = q->queue[i];
6742
6743 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6744 diff_q(&outq, p);
6745 else {
6746 /*
6747 * The caller can subtract 1 from skip_stat_unmatch
6748 * to determine how many paths were dirty only
6749 * due to stat info mismatch.
6750 */
6751 if (!diffopt->flags.no_index)
6752 diffopt->skip_stat_unmatch++;
6753 diff_free_filepair(p);
6754 }
6755 }
6756 free(q->queue);
6757 *q = outq;
6758 }
6759
6760 static int diffnamecmp(const void *a_, const void *b_)
6761 {
6762 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6763 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6764 const char *name_a, *name_b;
6765
6766 name_a = a->one ? a->one->path : a->two->path;
6767 name_b = b->one ? b->one->path : b->two->path;
6768 return strcmp(name_a, name_b);
6769 }
6770
6771 void diffcore_fix_diff_index(void)
6772 {
6773 struct diff_queue_struct *q = &diff_queued_diff;
6774 QSORT(q->queue, q->nr, diffnamecmp);
6775 }
6776
6777 void diff_add_if_missing(struct repository *r,
6778 struct oid_array *to_fetch,
6779 const struct diff_filespec *filespec)
6780 {
6781 if (filespec && filespec->oid_valid &&
6782 !S_ISGITLINK(filespec->mode) &&
6783 oid_object_info_extended(r, &filespec->oid, NULL,
6784 OBJECT_INFO_FOR_PREFETCH))
6785 oid_array_append(to_fetch, &filespec->oid);
6786 }
6787
6788 void diff_queued_diff_prefetch(void *repository)
6789 {
6790 struct repository *repo = repository;
6791 int i;
6792 struct diff_queue_struct *q = &diff_queued_diff;
6793 struct oid_array to_fetch = OID_ARRAY_INIT;
6794
6795 for (i = 0; i < q->nr; i++) {
6796 struct diff_filepair *p = q->queue[i];
6797 diff_add_if_missing(repo, &to_fetch, p->one);
6798 diff_add_if_missing(repo, &to_fetch, p->two);
6799 }
6800
6801 /*
6802 * NEEDSWORK: Consider deduplicating the OIDs sent.
6803 */
6804 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6805
6806 oid_array_clear(&to_fetch);
6807 }
6808
6809 void diffcore_std(struct diff_options *options)
6810 {
6811 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6812 DIFF_FORMAT_NUMSTAT |
6813 DIFF_FORMAT_PATCH |
6814 DIFF_FORMAT_SHORTSTAT |
6815 DIFF_FORMAT_DIRSTAT;
6816
6817 /*
6818 * Check if the user requested a blob-data-requiring diff output and/or
6819 * break-rewrite detection (which requires blob data). If yes, prefetch
6820 * the diff pairs.
6821 *
6822 * If no prefetching occurs, diffcore_rename() will prefetch if it
6823 * decides that it needs inexact rename detection.
6824 */
6825 if (options->repo == the_repository && has_promisor_remote() &&
6826 (options->output_format & output_formats_to_prefetch ||
6827 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6828 diff_queued_diff_prefetch(options->repo);
6829
6830 /* NOTE please keep the following in sync with diff_tree_combined() */
6831 if (options->skip_stat_unmatch)
6832 diffcore_skip_stat_unmatch(options);
6833 if (!options->found_follow) {
6834 /* See try_to_follow_renames() in tree-diff.c */
6835 if (options->break_opt != -1)
6836 diffcore_break(options->repo,
6837 options->break_opt);
6838 if (options->detect_rename)
6839 diffcore_rename(options);
6840 if (options->break_opt != -1)
6841 diffcore_merge_broken();
6842 }
6843 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6844 diffcore_pickaxe(options);
6845 if (options->orderfile)
6846 diffcore_order(options->orderfile);
6847 if (options->rotate_to)
6848 diffcore_rotate(options);
6849 if (!options->found_follow)
6850 /* See try_to_follow_renames() in tree-diff.c */
6851 diff_resolve_rename_copy();
6852 diffcore_apply_filter(options);
6853
6854 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6855 options->flags.has_changes = 1;
6856 else
6857 options->flags.has_changes = 0;
6858
6859 options->found_follow = 0;
6860 }
6861
6862 int diff_result_code(struct diff_options *opt, int status)
6863 {
6864 int result = 0;
6865
6866 diff_warn_rename_limit("diff.renameLimit",
6867 opt->needed_rename_limit,
6868 opt->degraded_cc_to_c);
6869 if (!opt->flags.exit_with_status &&
6870 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6871 return status;
6872 if (opt->flags.exit_with_status &&
6873 opt->flags.has_changes)
6874 result |= 01;
6875 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6876 opt->flags.check_failed)
6877 result |= 02;
6878 return result;
6879 }
6880
6881 int diff_can_quit_early(struct diff_options *opt)
6882 {
6883 return (opt->flags.quick &&
6884 !opt->filter &&
6885 opt->flags.has_changes);
6886 }
6887
6888 /*
6889 * Shall changes to this submodule be ignored?
6890 *
6891 * Submodule changes can be configured to be ignored separately for each path,
6892 * but that configuration can be overridden from the command line.
6893 */
6894 static int is_submodule_ignored(const char *path, struct diff_options *options)
6895 {
6896 int ignored = 0;
6897 struct diff_flags orig_flags = options->flags;
6898 if (!options->flags.override_submodule_config)
6899 set_diffopt_flags_from_submodule_config(options, path);
6900 if (options->flags.ignore_submodules)
6901 ignored = 1;
6902 options->flags = orig_flags;
6903 return ignored;
6904 }
6905
6906 void compute_diffstat(struct diff_options *options,
6907 struct diffstat_t *diffstat,
6908 struct diff_queue_struct *q)
6909 {
6910 int i;
6911
6912 memset(diffstat, 0, sizeof(struct diffstat_t));
6913 for (i = 0; i < q->nr; i++) {
6914 struct diff_filepair *p = q->queue[i];
6915 if (check_pair_status(p))
6916 diff_flush_stat(p, options, diffstat);
6917 }
6918 }
6919
6920 void diff_addremove(struct diff_options *options,
6921 int addremove, unsigned mode,
6922 const struct object_id *oid,
6923 int oid_valid,
6924 const char *concatpath, unsigned dirty_submodule)
6925 {
6926 struct diff_filespec *one, *two;
6927
6928 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6929 return;
6930
6931 /* This may look odd, but it is a preparation for
6932 * feeding "there are unchanged files which should
6933 * not produce diffs, but when you are doing copy
6934 * detection you would need them, so here they are"
6935 * entries to the diff-core. They will be prefixed
6936 * with something like '=' or '*' (I haven't decided
6937 * which but should not make any difference).
6938 * Feeding the same new and old to diff_change()
6939 * also has the same effect.
6940 * Before the final output happens, they are pruned after
6941 * merged into rename/copy pairs as appropriate.
6942 */
6943 if (options->flags.reverse_diff)
6944 addremove = (addremove == '+' ? '-' :
6945 addremove == '-' ? '+' : addremove);
6946
6947 if (options->prefix &&
6948 strncmp(concatpath, options->prefix, options->prefix_length))
6949 return;
6950
6951 one = alloc_filespec(concatpath);
6952 two = alloc_filespec(concatpath);
6953
6954 if (addremove != '+')
6955 fill_filespec(one, oid, oid_valid, mode);
6956 if (addremove != '-') {
6957 fill_filespec(two, oid, oid_valid, mode);
6958 two->dirty_submodule = dirty_submodule;
6959 }
6960
6961 diff_queue(&diff_queued_diff, one, two);
6962 if (!options->flags.diff_from_contents)
6963 options->flags.has_changes = 1;
6964 }
6965
6966 void diff_change(struct diff_options *options,
6967 unsigned old_mode, unsigned new_mode,
6968 const struct object_id *old_oid,
6969 const struct object_id *new_oid,
6970 int old_oid_valid, int new_oid_valid,
6971 const char *concatpath,
6972 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6973 {
6974 struct diff_filespec *one, *two;
6975 struct diff_filepair *p;
6976
6977 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6978 is_submodule_ignored(concatpath, options))
6979 return;
6980
6981 if (options->flags.reverse_diff) {
6982 SWAP(old_mode, new_mode);
6983 SWAP(old_oid, new_oid);
6984 SWAP(old_oid_valid, new_oid_valid);
6985 SWAP(old_dirty_submodule, new_dirty_submodule);
6986 }
6987
6988 if (options->prefix &&
6989 strncmp(concatpath, options->prefix, options->prefix_length))
6990 return;
6991
6992 one = alloc_filespec(concatpath);
6993 two = alloc_filespec(concatpath);
6994 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6995 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6996 one->dirty_submodule = old_dirty_submodule;
6997 two->dirty_submodule = new_dirty_submodule;
6998 p = diff_queue(&diff_queued_diff, one, two);
6999
7000 if (options->flags.diff_from_contents)
7001 return;
7002
7003 if (options->flags.quick && options->skip_stat_unmatch &&
7004 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7005 diff_free_filespec_data(p->one);
7006 diff_free_filespec_data(p->two);
7007 return;
7008 }
7009
7010 options->flags.has_changes = 1;
7011 }
7012
7013 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7014 {
7015 struct diff_filepair *pair;
7016 struct diff_filespec *one, *two;
7017
7018 if (options->prefix &&
7019 strncmp(path, options->prefix, options->prefix_length))
7020 return NULL;
7021
7022 one = alloc_filespec(path);
7023 two = alloc_filespec(path);
7024 pair = diff_queue(&diff_queued_diff, one, two);
7025 pair->is_unmerged = 1;
7026 return pair;
7027 }
7028
7029 static char *run_textconv(struct repository *r,
7030 const char *pgm,
7031 struct diff_filespec *spec,
7032 size_t *outsize)
7033 {
7034 struct diff_tempfile *temp;
7035 struct child_process child = CHILD_PROCESS_INIT;
7036 struct strbuf buf = STRBUF_INIT;
7037 int err = 0;
7038
7039 temp = prepare_temp_file(r, spec->path, spec);
7040 strvec_push(&child.args, pgm);
7041 strvec_push(&child.args, temp->name);
7042
7043 child.use_shell = 1;
7044 child.out = -1;
7045 if (start_command(&child)) {
7046 remove_tempfile();
7047 return NULL;
7048 }
7049
7050 if (strbuf_read(&buf, child.out, 0) < 0)
7051 err = error("error reading from textconv command '%s'", pgm);
7052 close(child.out);
7053
7054 if (finish_command(&child) || err) {
7055 strbuf_release(&buf);
7056 remove_tempfile();
7057 return NULL;
7058 }
7059 remove_tempfile();
7060
7061 return strbuf_detach(&buf, outsize);
7062 }
7063
7064 size_t fill_textconv(struct repository *r,
7065 struct userdiff_driver *driver,
7066 struct diff_filespec *df,
7067 char **outbuf)
7068 {
7069 size_t size;
7070
7071 if (!driver) {
7072 if (!DIFF_FILE_VALID(df)) {
7073 *outbuf = "";
7074 return 0;
7075 }
7076 if (diff_populate_filespec(r, df, NULL))
7077 die("unable to read files to diff");
7078 *outbuf = df->data;
7079 return df->size;
7080 }
7081
7082 if (!driver->textconv)
7083 BUG("fill_textconv called with non-textconv driver");
7084
7085 if (driver->textconv_cache && df->oid_valid) {
7086 *outbuf = notes_cache_get(driver->textconv_cache,
7087 &df->oid,
7088 &size);
7089 if (*outbuf)
7090 return size;
7091 }
7092
7093 *outbuf = run_textconv(r, driver->textconv, df, &size);
7094 if (!*outbuf)
7095 die("unable to read files to diff");
7096
7097 if (driver->textconv_cache && df->oid_valid) {
7098 /* ignore errors, as we might be in a readonly repository */
7099 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7100 size);
7101 /*
7102 * we could save up changes and flush them all at the end,
7103 * but we would need an extra call after all diffing is done.
7104 * Since generating a cache entry is the slow path anyway,
7105 * this extra overhead probably isn't a big deal.
7106 */
7107 notes_cache_write(driver->textconv_cache);
7108 }
7109
7110 return size;
7111 }
7112
7113 int textconv_object(struct repository *r,
7114 const char *path,
7115 unsigned mode,
7116 const struct object_id *oid,
7117 int oid_valid,
7118 char **buf,
7119 unsigned long *buf_size)
7120 {
7121 struct diff_filespec *df;
7122 struct userdiff_driver *textconv;
7123
7124 df = alloc_filespec(path);
7125 fill_filespec(df, oid, oid_valid, mode);
7126 textconv = get_textconv(r, df);
7127 if (!textconv) {
7128 free_filespec(df);
7129 return 0;
7130 }
7131
7132 *buf_size = fill_textconv(r, textconv, df, buf);
7133 free_filespec(df);
7134 return 1;
7135 }
7136
7137 void setup_diff_pager(struct diff_options *opt)
7138 {
7139 /*
7140 * If the user asked for our exit code, then either they want --quiet
7141 * or --exit-code. We should definitely not bother with a pager in the
7142 * former case, as we will generate no output. Since we still properly
7143 * report our exit code even when a pager is run, we _could_ run a
7144 * pager with --exit-code. But since we have not done so historically,
7145 * and because it is easy to find people oneline advising "git diff
7146 * --exit-code" in hooks and other scripts, we do not do so.
7147 */
7148 if (!opt->flags.exit_with_status &&
7149 check_pager_config("diff") != 0)
7150 setup_pager();
7151 }
The Best VCS Around