?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Git 2.44
[git.git] / diff.c
blobe50def45383eba4af74300802a161bdeeb6c4e8f
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "git-compat-util.h"
5 #include "abspath.h"
6 #include "base85.h"
7 #include "config.h"
8 #include "convert.h"
9 #include "environment.h"
10 #include "gettext.h"
11 #include "tempfile.h"
12 #include "quote.h"
13 #include "diff.h"
14 #include "diffcore.h"
15 #include "delta.h"
16 #include "hex.h"
17 #include "xdiff-interface.h"
18 #include "color.h"
19 #include "run-command.h"
20 #include "utf8.h"
21 #include "object-store-ll.h"
22 #include "userdiff.h"
23 #include "submodule.h"
24 #include "hashmap.h"
25 #include "mem-pool.h"
26 #include "merge-ll.h"
27 #include "string-list.h"
28 #include "strvec.h"
29 #include "graph.h"
30 #include "oid-array.h"
31 #include "packfile.h"
32 #include "pager.h"
33 #include "parse-options.h"
34 #include "help.h"
35 #include "promisor-remote.h"
36 #include "dir.h"
37 #include "object-file.h"
38 #include "object-name.h"
39 #include "read-cache-ll.h"
40 #include "setup.h"
41 #include "strmap.h"
42 #include "ws.h"
43
44 #ifdef NO_FAST_WORKING_DIRECTORY
45 #define FAST_WORKING_DIRECTORY 0
46 #else
47 #define FAST_WORKING_DIRECTORY 1
48 #endif
49
50 static int diff_detect_rename_default;
51 static int diff_indent_heuristic = 1;
52 static int diff_rename_limit_default = 1000;
53 static int diff_suppress_blank_empty;
54 static int diff_use_color_default = -1;
55 static int diff_color_moved_default;
56 static int diff_color_moved_ws_default;
57 static int diff_context_default = 3;
58 static int diff_interhunk_context_default;
59 static const char *diff_word_regex_cfg;
60 static const char *external_diff_cmd_cfg;
61 static const char *diff_order_file_cfg;
62 int diff_auto_refresh_index = 1;
63 static int diff_mnemonic_prefix;
64 static int diff_no_prefix;
65 static int diff_relative;
66 static int diff_stat_name_width;
67 static int diff_stat_graph_width;
68 static int diff_dirstat_permille_default = 30;
69 static struct diff_options default_diff_options;
70 static long diff_algorithm;
71 static unsigned ws_error_highlight_default = WSEH_NEW;
72
73 static char diff_colors[][COLOR_MAXLEN] = {
74 GIT_COLOR_RESET,
75 GIT_COLOR_NORMAL, /* CONTEXT */
76 GIT_COLOR_BOLD, /* METAINFO */
77 GIT_COLOR_CYAN, /* FRAGINFO */
78 GIT_COLOR_RED, /* OLD */
79 GIT_COLOR_GREEN, /* NEW */
80 GIT_COLOR_YELLOW, /* COMMIT */
81 GIT_COLOR_BG_RED, /* WHITESPACE */
82 GIT_COLOR_NORMAL, /* FUNCINFO */
83 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
84 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
85 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
86 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
87 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
88 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
89 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
90 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
91 GIT_COLOR_FAINT, /* CONTEXT_DIM */
92 GIT_COLOR_FAINT_RED, /* OLD_DIM */
93 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
94 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
95 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
96 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
97 };
98
99 static const char *color_diff_slots[] = {
100 [DIFF_CONTEXT] = "context",
101 [DIFF_METAINFO] = "meta",
102 [DIFF_FRAGINFO] = "frag",
103 [DIFF_FILE_OLD] = "old",
104 [DIFF_FILE_NEW] = "new",
105 [DIFF_COMMIT] = "commit",
106 [DIFF_WHITESPACE] = "whitespace",
107 [DIFF_FUNCINFO] = "func",
108 [DIFF_FILE_OLD_MOVED] = "oldMoved",
109 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
110 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
111 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
112 [DIFF_FILE_NEW_MOVED] = "newMoved",
113 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
114 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
115 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
116 [DIFF_CONTEXT_DIM] = "contextDimmed",
117 [DIFF_FILE_OLD_DIM] = "oldDimmed",
118 [DIFF_FILE_NEW_DIM] = "newDimmed",
119 [DIFF_CONTEXT_BOLD] = "contextBold",
120 [DIFF_FILE_OLD_BOLD] = "oldBold",
121 [DIFF_FILE_NEW_BOLD] = "newBold",
122 };
123
124 define_list_config_array_extra(color_diff_slots, {"plain"});
125
126 static int parse_diff_color_slot(const char *var)
127 {
128 if (!strcasecmp(var, "plain"))
129 return DIFF_CONTEXT;
130 return LOOKUP_CONFIG(color_diff_slots, var);
131 }
132
133 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
134 struct strbuf *errmsg)
135 {
136 char *params_copy = xstrdup(params_string);
137 struct string_list params = STRING_LIST_INIT_NODUP;
138 int ret = 0;
139 int i;
140
141 if (*params_copy)
142 string_list_split_in_place(&params, params_copy, ",", -1);
143 for (i = 0; i < params.nr; i++) {
144 const char *p = params.items[i].string;
145 if (!strcmp(p, "changes")) {
146 options->flags.dirstat_by_line = 0;
147 options->flags.dirstat_by_file = 0;
148 } else if (!strcmp(p, "lines")) {
149 options->flags.dirstat_by_line = 1;
150 options->flags.dirstat_by_file = 0;
151 } else if (!strcmp(p, "files")) {
152 options->flags.dirstat_by_line = 0;
153 options->flags.dirstat_by_file = 1;
154 } else if (!strcmp(p, "noncumulative")) {
155 options->flags.dirstat_cumulative = 0;
156 } else if (!strcmp(p, "cumulative")) {
157 options->flags.dirstat_cumulative = 1;
158 } else if (isdigit(*p)) {
159 char *end;
160 int permille = strtoul(p, &end, 10) * 10;
161 if (*end == '.' && isdigit(*++end)) {
162 /* only use first digit */
163 permille += *end - '0';
164 /* .. and ignore any further digits */
165 while (isdigit(*++end))
166 ; /* nothing */
167 }
168 if (!*end)
169 options->dirstat_permille = permille;
170 else {
171 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
172 p);
173 ret++;
174 }
175 } else {
176 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
177 ret++;
178 }
179
180 }
181 string_list_clear(&params, 0);
182 free(params_copy);
183 return ret;
184 }
185
186 static int parse_submodule_params(struct diff_options *options, const char *value)
187 {
188 if (!strcmp(value, "log"))
189 options->submodule_format = DIFF_SUBMODULE_LOG;
190 else if (!strcmp(value, "short"))
191 options->submodule_format = DIFF_SUBMODULE_SHORT;
192 else if (!strcmp(value, "diff"))
193 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
194 /*
195 * Please update $__git_diff_submodule_formats in
196 * git-completion.bash when you add new formats.
197 */
198 else
199 return -1;
200 return 0;
201 }
202
203 int git_config_rename(const char *var, const char *value)
204 {
205 if (!value)
206 return DIFF_DETECT_RENAME;
207 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
208 return DIFF_DETECT_COPY;
209 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
210 }
211
212 long parse_algorithm_value(const char *value)
213 {
214 if (!value)
215 return -1;
216 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
217 return 0;
218 else if (!strcasecmp(value, "minimal"))
219 return XDF_NEED_MINIMAL;
220 else if (!strcasecmp(value, "patience"))
221 return XDF_PATIENCE_DIFF;
222 else if (!strcasecmp(value, "histogram"))
223 return XDF_HISTOGRAM_DIFF;
224 /*
225 * Please update $__git_diff_algorithms in git-completion.bash
226 * when you add new algorithms.
227 */
228 return -1;
229 }
230
231 static int parse_one_token(const char **arg, const char *token)
232 {
233 const char *rest;
234 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
235 *arg = rest;
236 return 1;
237 }
238 return 0;
239 }
240
241 static int parse_ws_error_highlight(const char *arg)
242 {
243 const char *orig_arg = arg;
244 unsigned val = 0;
245
246 while (*arg) {
247 if (parse_one_token(&arg, "none"))
248 val = 0;
249 else if (parse_one_token(&arg, "default"))
250 val = WSEH_NEW;
251 else if (parse_one_token(&arg, "all"))
252 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
253 else if (parse_one_token(&arg, "new"))
254 val |= WSEH_NEW;
255 else if (parse_one_token(&arg, "old"))
256 val |= WSEH_OLD;
257 else if (parse_one_token(&arg, "context"))
258 val |= WSEH_CONTEXT;
259 else {
260 return -1 - (int)(arg - orig_arg);
261 }
262 if (*arg)
263 arg++;
264 }
265 return val;
266 }
267
268 /*
269 * These are to give UI layer defaults.
270 * The core-level commands such as git-diff-files should
271 * never be affected by the setting of diff.renames
272 * the user happens to have in the configuration file.
273 */
274 void init_diff_ui_defaults(void)
275 {
276 diff_detect_rename_default = DIFF_DETECT_RENAME;
277 }
278
279 int git_diff_heuristic_config(const char *var, const char *value,
280 void *cb UNUSED)
281 {
282 if (!strcmp(var, "diff.indentheuristic"))
283 diff_indent_heuristic = git_config_bool(var, value);
284 return 0;
285 }
286
287 static int parse_color_moved(const char *arg)
288 {
289 switch (git_parse_maybe_bool(arg)) {
290 case 0:
291 return COLOR_MOVED_NO;
292 case 1:
293 return COLOR_MOVED_DEFAULT;
294 default:
295 break;
296 }
297
298 if (!strcmp(arg, "no"))
299 return COLOR_MOVED_NO;
300 else if (!strcmp(arg, "plain"))
301 return COLOR_MOVED_PLAIN;
302 else if (!strcmp(arg, "blocks"))
303 return COLOR_MOVED_BLOCKS;
304 else if (!strcmp(arg, "zebra"))
305 return COLOR_MOVED_ZEBRA;
306 else if (!strcmp(arg, "default"))
307 return COLOR_MOVED_DEFAULT;
308 else if (!strcmp(arg, "dimmed-zebra"))
309 return COLOR_MOVED_ZEBRA_DIM;
310 else if (!strcmp(arg, "dimmed_zebra"))
311 return COLOR_MOVED_ZEBRA_DIM;
312 else
313 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
314 }
315
316 static unsigned parse_color_moved_ws(const char *arg)
317 {
318 int ret = 0;
319 struct string_list l = STRING_LIST_INIT_DUP;
320 struct string_list_item *i;
321
322 string_list_split(&l, arg, ',', -1);
323
324 for_each_string_list_item(i, &l) {
325 struct strbuf sb = STRBUF_INIT;
326 strbuf_addstr(&sb, i->string);
327 strbuf_trim(&sb);
328
329 if (!strcmp(sb.buf, "no"))
330 ret = 0;
331 else if (!strcmp(sb.buf, "ignore-space-change"))
332 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
333 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
334 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
335 else if (!strcmp(sb.buf, "ignore-all-space"))
336 ret |= XDF_IGNORE_WHITESPACE;
337 else if (!strcmp(sb.buf, "allow-indentation-change"))
338 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
339 else {
340 ret |= COLOR_MOVED_WS_ERROR;
341 error(_("unknown color-moved-ws mode '%s', possible values are 'ignore-space-change', 'ignore-space-at-eol', 'ignore-all-space', 'allow-indentation-change'"), sb.buf);
342 }
343
344 strbuf_release(&sb);
345 }
346
347 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
348 (ret & XDF_WHITESPACE_FLAGS)) {
349 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
350 ret |= COLOR_MOVED_WS_ERROR;
351 }
352
353 string_list_clear(&l, 0);
354
355 return ret;
356 }
357
358 int git_diff_ui_config(const char *var, const char *value,
359 const struct config_context *ctx, void *cb)
360 {
361 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
362 diff_use_color_default = git_config_colorbool(var, value);
363 return 0;
364 }
365 if (!strcmp(var, "diff.colormoved")) {
366 int cm = parse_color_moved(value);
367 if (cm < 0)
368 return -1;
369 diff_color_moved_default = cm;
370 return 0;
371 }
372 if (!strcmp(var, "diff.colormovedws")) {
373 unsigned cm;
374 if (!value)
375 return config_error_nonbool(var);
376 cm = parse_color_moved_ws(value);
377 if (cm & COLOR_MOVED_WS_ERROR)
378 return -1;
379 diff_color_moved_ws_default = cm;
380 return 0;
381 }
382 if (!strcmp(var, "diff.context")) {
383 diff_context_default = git_config_int(var, value, ctx->kvi);
384 if (diff_context_default < 0)
385 return -1;
386 return 0;
387 }
388 if (!strcmp(var, "diff.interhunkcontext")) {
389 diff_interhunk_context_default = git_config_int(var, value,
390 ctx->kvi);
391 if (diff_interhunk_context_default < 0)
392 return -1;
393 return 0;
394 }
395 if (!strcmp(var, "diff.renames")) {
396 diff_detect_rename_default = git_config_rename(var, value);
397 return 0;
398 }
399 if (!strcmp(var, "diff.autorefreshindex")) {
400 diff_auto_refresh_index = git_config_bool(var, value);
401 return 0;
402 }
403 if (!strcmp(var, "diff.mnemonicprefix")) {
404 diff_mnemonic_prefix = git_config_bool(var, value);
405 return 0;
406 }
407 if (!strcmp(var, "diff.noprefix")) {
408 diff_no_prefix = git_config_bool(var, value);
409 return 0;
410 }
411 if (!strcmp(var, "diff.relative")) {
412 diff_relative = git_config_bool(var, value);
413 return 0;
414 }
415 if (!strcmp(var, "diff.statnamewidth")) {
416 diff_stat_name_width = git_config_int(var, value, ctx->kvi);
417 return 0;
418 }
419 if (!strcmp(var, "diff.statgraphwidth")) {
420 diff_stat_graph_width = git_config_int(var, value, ctx->kvi);
421 return 0;
422 }
423 if (!strcmp(var, "diff.external"))
424 return git_config_string(&external_diff_cmd_cfg, var, value);
425 if (!strcmp(var, "diff.wordregex"))
426 return git_config_string(&diff_word_regex_cfg, var, value);
427 if (!strcmp(var, "diff.orderfile"))
428 return git_config_pathname(&diff_order_file_cfg, var, value);
429
430 if (!strcmp(var, "diff.ignoresubmodules")) {
431 if (!value)
432 return config_error_nonbool(var);
433 handle_ignore_submodules_arg(&default_diff_options, value);
434 }
435
436 if (!strcmp(var, "diff.submodule")) {
437 if (!value)
438 return config_error_nonbool(var);
439 if (parse_submodule_params(&default_diff_options, value))
440 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
441 value);
442 return 0;
443 }
444
445 if (!strcmp(var, "diff.algorithm")) {
446 if (!value)
447 return config_error_nonbool(var);
448 diff_algorithm = parse_algorithm_value(value);
449 if (diff_algorithm < 0)
450 return error(_("unknown value for config '%s': %s"),
451 var, value);
452 return 0;
453 }
454
455 if (git_color_config(var, value, cb) < 0)
456 return -1;
457
458 return git_diff_basic_config(var, value, ctx, cb);
459 }
460
461 int git_diff_basic_config(const char *var, const char *value,
462 const struct config_context *ctx, void *cb)
463 {
464 const char *name;
465
466 if (!strcmp(var, "diff.renamelimit")) {
467 diff_rename_limit_default = git_config_int(var, value, ctx->kvi);
468 return 0;
469 }
470
471 if (userdiff_config(var, value) < 0)
472 return -1;
473
474 if (skip_prefix(var, "diff.color.", &name) ||
475 skip_prefix(var, "color.diff.", &name)) {
476 int slot = parse_diff_color_slot(name);
477 if (slot < 0)
478 return 0;
479 if (!value)
480 return config_error_nonbool(var);
481 return color_parse(value, diff_colors[slot]);
482 }
483
484 if (!strcmp(var, "diff.wserrorhighlight")) {
485 int val;
486 if (!value)
487 return config_error_nonbool(var);
488 val = parse_ws_error_highlight(value);
489 if (val < 0)
490 return error(_("unknown value for config '%s': %s"),
491 var, value);
492 ws_error_highlight_default = val;
493 return 0;
494 }
495
496 /* like GNU diff's --suppress-blank-empty option */
497 if (!strcmp(var, "diff.suppressblankempty") ||
498 /* for backwards compatibility */
499 !strcmp(var, "diff.suppress-blank-empty")) {
500 diff_suppress_blank_empty = git_config_bool(var, value);
501 return 0;
502 }
503
504 if (!strcmp(var, "diff.dirstat")) {
505 struct strbuf errmsg = STRBUF_INIT;
506 if (!value)
507 return config_error_nonbool(var);
508 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
509 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
510 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
511 errmsg.buf);
512 strbuf_release(&errmsg);
513 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
514 return 0;
515 }
516
517 if (git_diff_heuristic_config(var, value, cb) < 0)
518 return -1;
519
520 return git_default_config(var, value, ctx, cb);
521 }
522
523 static char *quote_two(const char *one, const char *two)
524 {
525 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
526 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
527 struct strbuf res = STRBUF_INIT;
528
529 if (need_one + need_two) {
530 strbuf_addch(&res, '"');
531 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
532 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
533 strbuf_addch(&res, '"');
534 } else {
535 strbuf_addstr(&res, one);
536 strbuf_addstr(&res, two);
537 }
538 return strbuf_detach(&res, NULL);
539 }
540
541 static const char *external_diff(void)
542 {
543 static const char *external_diff_cmd = NULL;
544 static int done_preparing = 0;
545
546 if (done_preparing)
547 return external_diff_cmd;
548 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
549 if (!external_diff_cmd)
550 external_diff_cmd = external_diff_cmd_cfg;
551 done_preparing = 1;
552 return external_diff_cmd;
553 }
554
555 /*
556 * Keep track of files used for diffing. Sometimes such an entry
557 * refers to a temporary file, sometimes to an existing file, and
558 * sometimes to "/dev/null".
559 */
560 static struct diff_tempfile {
561 /*
562 * filename external diff should read from, or NULL if this
563 * entry is currently not in use:
564 */
565 const char *name;
566
567 char hex[GIT_MAX_HEXSZ + 1];
568 char mode[10];
569
570 /*
571 * If this diff_tempfile instance refers to a temporary file,
572 * this tempfile object is used to manage its lifetime.
573 */
574 struct tempfile *tempfile;
575 } diff_temp[2];
576
577 struct emit_callback {
578 int color_diff;
579 unsigned ws_rule;
580 int blank_at_eof_in_preimage;
581 int blank_at_eof_in_postimage;
582 int lno_in_preimage;
583 int lno_in_postimage;
584 const char **label_path;
585 struct diff_words_data *diff_words;
586 struct diff_options *opt;
587 struct strbuf *header;
588 };
589
590 static int count_lines(const char *data, int size)
591 {
592 int count, ch, completely_empty = 1, nl_just_seen = 0;
593 count = 0;
594 while (0 < size--) {
595 ch = *data++;
596 if (ch == '\n') {
597 count++;
598 nl_just_seen = 1;
599 completely_empty = 0;
600 }
601 else {
602 nl_just_seen = 0;
603 completely_empty = 0;
604 }
605 }
606 if (completely_empty)
607 return 0;
608 if (!nl_just_seen)
609 count++; /* no trailing newline */
610 return count;
611 }
612
613 static int fill_mmfile(struct repository *r, mmfile_t *mf,
614 struct diff_filespec *one)
615 {
616 if (!DIFF_FILE_VALID(one)) {
617 mf->ptr = (char *)""; /* does not matter */
618 mf->size = 0;
619 return 0;
620 }
621 else if (diff_populate_filespec(r, one, NULL))
622 return -1;
623
624 mf->ptr = one->data;
625 mf->size = one->size;
626 return 0;
627 }
628
629 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
630 static unsigned long diff_filespec_size(struct repository *r,
631 struct diff_filespec *one)
632 {
633 struct diff_populate_filespec_options dpf_options = {
634 .check_size_only = 1,
635 };
636
637 if (!DIFF_FILE_VALID(one))
638 return 0;
639 diff_populate_filespec(r, one, &dpf_options);
640 return one->size;
641 }
642
643 static int count_trailing_blank(mmfile_t *mf)
644 {
645 char *ptr = mf->ptr;
646 long size = mf->size;
647 int cnt = 0;
648
649 if (!size)
650 return cnt;
651 ptr += size - 1; /* pointing at the very end */
652 if (*ptr != '\n')
653 ; /* incomplete line */
654 else
655 ptr--; /* skip the last LF */
656 while (mf->ptr < ptr) {
657 char *prev_eol;
658 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
659 if (*prev_eol == '\n')
660 break;
661 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
662 break;
663 cnt++;
664 ptr = prev_eol - 1;
665 }
666 return cnt;
667 }
668
669 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
670 struct emit_callback *ecbdata)
671 {
672 int l1, l2, at;
673 l1 = count_trailing_blank(mf1);
674 l2 = count_trailing_blank(mf2);
675 if (l2 <= l1) {
676 ecbdata->blank_at_eof_in_preimage = 0;
677 ecbdata->blank_at_eof_in_postimage = 0;
678 return;
679 }
680 at = count_lines(mf1->ptr, mf1->size);
681 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
682
683 at = count_lines(mf2->ptr, mf2->size);
684 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
685 }
686
687 static void emit_line_0(struct diff_options *o,
688 const char *set_sign, const char *set, unsigned reverse, const char *reset,
689 int first, const char *line, int len)
690 {
691 int has_trailing_newline, has_trailing_carriage_return;
692 int needs_reset = 0; /* at the end of the line */
693 FILE *file = o->file;
694
695 fputs(diff_line_prefix(o), file);
696
697 has_trailing_newline = (len > 0 && line[len-1] == '\n');
698 if (has_trailing_newline)
699 len--;
700
701 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
702 if (has_trailing_carriage_return)
703 len--;
704
705 if (!len && !first)
706 goto end_of_line;
707
708 if (reverse && want_color(o->use_color)) {
709 fputs(GIT_COLOR_REVERSE, file);
710 needs_reset = 1;
711 }
712
713 if (set_sign) {
714 fputs(set_sign, file);
715 needs_reset = 1;
716 }
717
718 if (first)
719 fputc(first, file);
720
721 if (!len)
722 goto end_of_line;
723
724 if (set) {
725 if (set_sign && set != set_sign)
726 fputs(reset, file);
727 fputs(set, file);
728 needs_reset = 1;
729 }
730 fwrite(line, len, 1, file);
731 needs_reset = 1; /* 'line' may contain color codes. */
732
733 end_of_line:
734 if (needs_reset)
735 fputs(reset, file);
736 if (has_trailing_carriage_return)
737 fputc('\r', file);
738 if (has_trailing_newline)
739 fputc('\n', file);
740 }
741
742 static void emit_line(struct diff_options *o, const char *set, const char *reset,
743 const char *line, int len)
744 {
745 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
746 }
747
748 enum diff_symbol {
749 DIFF_SYMBOL_BINARY_DIFF_HEADER,
750 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
751 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
752 DIFF_SYMBOL_BINARY_DIFF_BODY,
753 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
754 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
755 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
756 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
757 DIFF_SYMBOL_STATS_LINE,
758 DIFF_SYMBOL_WORD_DIFF,
759 DIFF_SYMBOL_STAT_SEP,
760 DIFF_SYMBOL_SUMMARY,
761 DIFF_SYMBOL_SUBMODULE_ADD,
762 DIFF_SYMBOL_SUBMODULE_DEL,
763 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
764 DIFF_SYMBOL_SUBMODULE_MODIFIED,
765 DIFF_SYMBOL_SUBMODULE_HEADER,
766 DIFF_SYMBOL_SUBMODULE_ERROR,
767 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
768 DIFF_SYMBOL_REWRITE_DIFF,
769 DIFF_SYMBOL_BINARY_FILES,
770 DIFF_SYMBOL_HEADER,
771 DIFF_SYMBOL_FILEPAIR_PLUS,
772 DIFF_SYMBOL_FILEPAIR_MINUS,
773 DIFF_SYMBOL_WORDS_PORCELAIN,
774 DIFF_SYMBOL_WORDS,
775 DIFF_SYMBOL_CONTEXT,
776 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
777 DIFF_SYMBOL_PLUS,
778 DIFF_SYMBOL_MINUS,
779 DIFF_SYMBOL_NO_LF_EOF,
780 DIFF_SYMBOL_CONTEXT_FRAGINFO,
781 DIFF_SYMBOL_CONTEXT_MARKER,
782 DIFF_SYMBOL_SEPARATOR
783 };
784 /*
785 * Flags for content lines:
786 * 0..12 are whitespace rules
787 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
788 * 16 is marking if the line is blank at EOF
789 */
790 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
791 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
792 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
793 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
794 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
795
796 /*
797 * This struct is used when we need to buffer the output of the diff output.
798 *
799 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
800 * into the pre/post image file. This pointer could be a union with the
801 * line pointer. By storing an offset into the file instead of the literal line,
802 * we can decrease the memory footprint for the buffered output. At first we
803 * may want to only have indirection for the content lines, but we could also
804 * enhance the state for emitting prefabricated lines, e.g. the similarity
805 * score line or hunk/file headers would only need to store a number or path
806 * and then the output can be constructed later on depending on state.
807 */
808 struct emitted_diff_symbol {
809 const char *line;
810 int len;
811 int flags;
812 int indent_off; /* Offset to first non-whitespace character */
813 int indent_width; /* The visual width of the indentation */
814 unsigned id;
815 enum diff_symbol s;
816 };
817 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
818
819 struct emitted_diff_symbols {
820 struct emitted_diff_symbol *buf;
821 int nr, alloc;
822 };
823 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
824
825 static void append_emitted_diff_symbol(struct diff_options *o,
826 struct emitted_diff_symbol *e)
827 {
828 struct emitted_diff_symbol *f;
829
830 ALLOC_GROW(o->emitted_symbols->buf,
831 o->emitted_symbols->nr + 1,
832 o->emitted_symbols->alloc);
833 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
834
835 memcpy(f, e, sizeof(struct emitted_diff_symbol));
836 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
837 }
838
839 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
840 {
841 if (!e)
842 return;
843 free(e->buf);
844 free(e);
845 }
846
847 struct moved_entry {
848 const struct emitted_diff_symbol *es;
849 struct moved_entry *next_line;
850 struct moved_entry *next_match;
851 };
852
853 struct moved_block {
854 struct moved_entry *match;
855 int wsd; /* The whitespace delta of this block */
856 };
857
858 #define INDENT_BLANKLINE INT_MIN
859
860 static void fill_es_indent_data(struct emitted_diff_symbol *es)
861 {
862 unsigned int off = 0, i;
863 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
864 const char *s = es->line;
865 const int len = es->len;
866
867 /* skip any \v \f \r at start of indentation */
868 while (s[off] == '\f' || s[off] == '\v' ||
869 (s[off] == '\r' && off < len - 1))
870 off++;
871
872 /* calculate the visual width of indentation */
873 while(1) {
874 if (s[off] == ' ') {
875 width++;
876 off++;
877 } else if (s[off] == '\t') {
878 width += tab_width - (width % tab_width);
879 while (s[++off] == '\t')
880 width += tab_width;
881 } else {
882 break;
883 }
884 }
885
886 /* check if this line is blank */
887 for (i = off; i < len; i++)
888 if (!isspace(s[i]))
889 break;
890
891 if (i == len) {
892 es->indent_width = INDENT_BLANKLINE;
893 es->indent_off = len;
894 } else {
895 es->indent_off = off;
896 es->indent_width = width;
897 }
898 }
899
900 static int compute_ws_delta(const struct emitted_diff_symbol *a,
901 const struct emitted_diff_symbol *b)
902 {
903 int a_width = a->indent_width,
904 b_width = b->indent_width;
905
906 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
907 return INDENT_BLANKLINE;
908
909 return a_width - b_width;
910 }
911
912 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
913 const struct emitted_diff_symbol *l,
914 struct moved_block *pmb)
915 {
916 int a_width = cur->es->indent_width, b_width = l->indent_width;
917 int delta;
918
919 /* The text of each line must match */
920 if (cur->es->id != l->id)
921 return 1;
922
923 /*
924 * If 'l' and 'cur' are both blank then we don't need to check the
925 * indent. We only need to check cur as we know the strings match.
926 * */
927 if (a_width == INDENT_BLANKLINE)
928 return 0;
929
930 /*
931 * The indent changes of the block are known and stored in pmb->wsd;
932 * however we need to check if the indent changes of the current line
933 * match those of the current block.
934 */
935 delta = b_width - a_width;
936
937 /*
938 * If the previous lines of this block were all blank then set its
939 * whitespace delta.
940 */
941 if (pmb->wsd == INDENT_BLANKLINE)
942 pmb->wsd = delta;
943
944 return delta != pmb->wsd;
945 }
946
947 struct interned_diff_symbol {
948 struct hashmap_entry ent;
949 struct emitted_diff_symbol *es;
950 };
951
952 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
953 const struct hashmap_entry *eptr,
954 const struct hashmap_entry *entry_or_key,
955 const void *keydata UNUSED)
956 {
957 const struct diff_options *diffopt = hashmap_cmp_fn_data;
958 const struct emitted_diff_symbol *a, *b;
959 unsigned flags = diffopt->color_moved_ws_handling
960 & XDF_WHITESPACE_FLAGS;
961
962 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
963 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
964
965 return !xdiff_compare_lines(a->line + a->indent_off,
966 a->len - a->indent_off,
967 b->line + b->indent_off,
968 b->len - b->indent_off, flags);
969 }
970
971 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
972 struct interned_diff_symbol *s)
973 {
974 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
975 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
976 l->len - l->indent_off, flags);
977
978 hashmap_entry_init(&s->ent, hash);
979 s->es = l;
980 }
981
982 struct moved_entry_list {
983 struct moved_entry *add, *del;
984 };
985
986 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
987 struct mem_pool *entry_mem_pool)
988 {
989 struct moved_entry *prev_line = NULL;
990 struct mem_pool interned_pool;
991 struct hashmap interned_map;
992 struct moved_entry_list *entry_list = NULL;
993 size_t entry_list_alloc = 0;
994 unsigned id = 0;
995 int n;
996
997 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
998 mem_pool_init(&interned_pool, 1024 * 1024);
999
1000 for (n = 0; n < o->emitted_symbols->nr; n++) {
1001 struct interned_diff_symbol key;
1002 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1003 struct interned_diff_symbol *s;
1004 struct moved_entry *entry;
1005
1006 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
1007 prev_line = NULL;
1008 continue;
1009 }
1010
1011 if (o->color_moved_ws_handling &
1012 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1013 fill_es_indent_data(l);
1014
1015 prepare_entry(o, l, &key);
1016 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
1017 if (s) {
1018 l->id = s->es->id;
1019 } else {
1020 l->id = id;
1021 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
1022 hashmap_add(&interned_map,
1023 memcpy(mem_pool_alloc(&interned_pool,
1024 sizeof(key)),
1025 &key, sizeof(key)));
1026 }
1027 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1028 entry->es = l;
1029 entry->next_line = NULL;
1030 if (prev_line && prev_line->es->s == l->s)
1031 prev_line->next_line = entry;
1032 prev_line = entry;
1033 if (l->s == DIFF_SYMBOL_PLUS) {
1034 entry->next_match = entry_list[l->id].add;
1035 entry_list[l->id].add = entry;
1036 } else {
1037 entry->next_match = entry_list[l->id].del;
1038 entry_list[l->id].del = entry;
1039 }
1040 }
1041
1042 hashmap_clear(&interned_map);
1043 mem_pool_discard(&interned_pool, 0);
1044
1045 return entry_list;
1046 }
1047
1048 static void pmb_advance_or_null(struct diff_options *o,
1049 struct emitted_diff_symbol *l,
1050 struct moved_block *pmb,
1051 int *pmb_nr)
1052 {
1053 int i, j;
1054
1055 for (i = 0, j = 0; i < *pmb_nr; i++) {
1056 int match;
1057 struct moved_entry *prev = pmb[i].match;
1058 struct moved_entry *cur = (prev && prev->next_line) ?
1059 prev->next_line : NULL;
1060
1061 if (o->color_moved_ws_handling &
1062 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1063 match = cur &&
1064 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1065 else
1066 match = cur && cur->es->id == l->id;
1067
1068 if (match) {
1069 pmb[j] = pmb[i];
1070 pmb[j++].match = cur;
1071 }
1072 }
1073 *pmb_nr = j;
1074 }
1075
1076 static void fill_potential_moved_blocks(struct diff_options *o,
1077 struct moved_entry *match,
1078 struct emitted_diff_symbol *l,
1079 struct moved_block **pmb_p,
1080 int *pmb_alloc_p, int *pmb_nr_p)
1081
1082 {
1083 struct moved_block *pmb = *pmb_p;
1084 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1085
1086 /*
1087 * The current line is the start of a new block.
1088 * Setup the set of potential blocks.
1089 */
1090 for (; match; match = match->next_match) {
1091 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1092 if (o->color_moved_ws_handling &
1093 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1094 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1095 else
1096 pmb[pmb_nr].wsd = 0;
1097 pmb[pmb_nr++].match = match;
1098 }
1099
1100 *pmb_p = pmb;
1101 *pmb_alloc_p = pmb_alloc;
1102 *pmb_nr_p = pmb_nr;
1103 }
1104
1105 /*
1106 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1107 *
1108 * Otherwise, if the last block has fewer alphanumeric characters than
1109 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1110 * that block.
1111 *
1112 * The last block consists of the (n - block_length)'th line up to but not
1113 * including the nth line.
1114 *
1115 * Returns 0 if the last block is empty or is unset by this function, non zero
1116 * otherwise.
1117 *
1118 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1119 * Think of a way to unify them.
1120 */
1121 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1122 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1123 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1124 {
1125 int i, alnum_count = 0;
1126 if (o->color_moved == COLOR_MOVED_PLAIN)
1127 return block_length;
1128 for (i = 1; i < block_length + 1; i++) {
1129 const char *c = o->emitted_symbols->buf[n - i].line;
1130 for (; *c; c++) {
1131 if (!isalnum(*c))
1132 continue;
1133 alnum_count++;
1134 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1135 return 1;
1136 }
1137 }
1138 for (i = 1; i < block_length + 1; i++)
1139 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1140 return 0;
1141 }
1142
1143 /* Find blocks of moved code, delegate actual coloring decision to helper */
1144 static void mark_color_as_moved(struct diff_options *o,
1145 struct moved_entry_list *entry_list)
1146 {
1147 struct moved_block *pmb = NULL; /* potentially moved blocks */
1148 int pmb_nr = 0, pmb_alloc = 0;
1149 int n, flipped_block = 0, block_length = 0;
1150 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1151
1152
1153 for (n = 0; n < o->emitted_symbols->nr; n++) {
1154 struct moved_entry *match = NULL;
1155 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1156
1157 switch (l->s) {
1158 case DIFF_SYMBOL_PLUS:
1159 match = entry_list[l->id].del;
1160 break;
1161 case DIFF_SYMBOL_MINUS:
1162 match = entry_list[l->id].add;
1163 break;
1164 default:
1165 flipped_block = 0;
1166 }
1167
1168 if (pmb_nr && (!match || l->s != moved_symbol)) {
1169 if (!adjust_last_block(o, n, block_length) &&
1170 block_length > 1) {
1171 /*
1172 * Rewind in case there is another match
1173 * starting at the second line of the block
1174 */
1175 match = NULL;
1176 n -= block_length;
1177 }
1178 pmb_nr = 0;
1179 block_length = 0;
1180 flipped_block = 0;
1181 }
1182 if (!match) {
1183 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1184 continue;
1185 }
1186
1187 if (o->color_moved == COLOR_MOVED_PLAIN) {
1188 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1189 continue;
1190 }
1191
1192 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1193
1194 if (pmb_nr == 0) {
1195 int contiguous = adjust_last_block(o, n, block_length);
1196
1197 if (!contiguous && block_length > 1)
1198 /*
1199 * Rewind in case there is another match
1200 * starting at the second line of the block
1201 */
1202 n -= block_length;
1203 else
1204 fill_potential_moved_blocks(o, match, l,
1205 &pmb, &pmb_alloc,
1206 &pmb_nr);
1207
1208 if (contiguous && pmb_nr && moved_symbol == l->s)
1209 flipped_block = (flipped_block + 1) % 2;
1210 else
1211 flipped_block = 0;
1212
1213 if (pmb_nr)
1214 moved_symbol = l->s;
1215 else
1216 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1217
1218 block_length = 0;
1219 }
1220
1221 if (pmb_nr) {
1222 block_length++;
1223 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1224 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1225 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1226 }
1227 }
1228 adjust_last_block(o, n, block_length);
1229
1230 free(pmb);
1231 }
1232
1233 static void dim_moved_lines(struct diff_options *o)
1234 {
1235 int n;
1236 for (n = 0; n < o->emitted_symbols->nr; n++) {
1237 struct emitted_diff_symbol *prev = (n != 0) ?
1238 &o->emitted_symbols->buf[n - 1] : NULL;
1239 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1240 struct emitted_diff_symbol *next =
1241 (n < o->emitted_symbols->nr - 1) ?
1242 &o->emitted_symbols->buf[n + 1] : NULL;
1243
1244 /* Not a plus or minus line? */
1245 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1246 continue;
1247
1248 /* Not a moved line? */
1249 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1250 continue;
1251
1252 /*
1253 * If prev or next are not a plus or minus line,
1254 * pretend they don't exist
1255 */
1256 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1257 prev->s != DIFF_SYMBOL_MINUS)
1258 prev = NULL;
1259 if (next && next->s != DIFF_SYMBOL_PLUS &&
1260 next->s != DIFF_SYMBOL_MINUS)
1261 next = NULL;
1262
1263 /* Inside a block? */
1264 if ((prev &&
1265 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1266 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1267 (next &&
1268 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1269 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1270 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1271 continue;
1272 }
1273
1274 /* Check if we are at an interesting bound: */
1275 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1276 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1277 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1278 continue;
1279 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1280 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1281 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1282 continue;
1283
1284 /*
1285 * The boundary to prev and next are not interesting,
1286 * so this line is not interesting as a whole
1287 */
1288 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1289 }
1290 }
1291
1292 static void emit_line_ws_markup(struct diff_options *o,
1293 const char *set_sign, const char *set,
1294 const char *reset,
1295 int sign_index, const char *line, int len,
1296 unsigned ws_rule, int blank_at_eof)
1297 {
1298 const char *ws = NULL;
1299 int sign = o->output_indicators[sign_index];
1300
1301 if (o->ws_error_highlight & ws_rule) {
1302 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1303 if (!*ws)
1304 ws = NULL;
1305 }
1306
1307 if (!ws && !set_sign)
1308 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1309 else if (!ws) {
1310 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1311 } else if (blank_at_eof)
1312 /* Blank line at EOF - paint '+' as well */
1313 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1314 else {
1315 /* Emit just the prefix, then the rest. */
1316 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1317 sign, "", 0);
1318 ws_check_emit(line, len, ws_rule,
1319 o->file, set, reset, ws);
1320 }
1321 }
1322
1323 static void emit_diff_symbol_from_struct(struct diff_options *o,
1324 struct emitted_diff_symbol *eds)
1325 {
1326 static const char *nneof = " No newline at end of file\n";
1327 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1328
1329 enum diff_symbol s = eds->s;
1330 const char *line = eds->line;
1331 int len = eds->len;
1332 unsigned flags = eds->flags;
1333
1334 switch (s) {
1335 case DIFF_SYMBOL_NO_LF_EOF:
1336 context = diff_get_color_opt(o, DIFF_CONTEXT);
1337 reset = diff_get_color_opt(o, DIFF_RESET);
1338 putc('\n', o->file);
1339 emit_line_0(o, context, NULL, 0, reset, '\\',
1340 nneof, strlen(nneof));
1341 break;
1342 case DIFF_SYMBOL_SUBMODULE_HEADER:
1343 case DIFF_SYMBOL_SUBMODULE_ERROR:
1344 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1345 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1346 case DIFF_SYMBOL_SUMMARY:
1347 case DIFF_SYMBOL_STATS_LINE:
1348 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1349 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1350 emit_line(o, "", "", line, len);
1351 break;
1352 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1353 case DIFF_SYMBOL_CONTEXT_MARKER:
1354 context = diff_get_color_opt(o, DIFF_CONTEXT);
1355 reset = diff_get_color_opt(o, DIFF_RESET);
1356 emit_line(o, context, reset, line, len);
1357 break;
1358 case DIFF_SYMBOL_SEPARATOR:
1359 fprintf(o->file, "%s%c",
1360 diff_line_prefix(o),
1361 o->line_termination);
1362 break;
1363 case DIFF_SYMBOL_CONTEXT:
1364 set = diff_get_color_opt(o, DIFF_CONTEXT);
1365 reset = diff_get_color_opt(o, DIFF_RESET);
1366 set_sign = NULL;
1367 if (o->flags.dual_color_diffed_diffs) {
1368 char c = !len ? 0 : line[0];
1369
1370 if (c == '+')
1371 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1372 else if (c == '@')
1373 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1374 else if (c == '-')
1375 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1376 }
1377 emit_line_ws_markup(o, set_sign, set, reset,
1378 OUTPUT_INDICATOR_CONTEXT, line, len,
1379 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1380 break;
1381 case DIFF_SYMBOL_PLUS:
1382 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1383 DIFF_SYMBOL_MOVED_LINE_ALT |
1384 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1385 case DIFF_SYMBOL_MOVED_LINE |
1386 DIFF_SYMBOL_MOVED_LINE_ALT |
1387 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1388 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1389 break;
1390 case DIFF_SYMBOL_MOVED_LINE |
1391 DIFF_SYMBOL_MOVED_LINE_ALT:
1392 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1393 break;
1394 case DIFF_SYMBOL_MOVED_LINE |
1395 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1396 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1397 break;
1398 case DIFF_SYMBOL_MOVED_LINE:
1399 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1400 break;
1401 default:
1402 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1403 }
1404 reset = diff_get_color_opt(o, DIFF_RESET);
1405 if (!o->flags.dual_color_diffed_diffs)
1406 set_sign = NULL;
1407 else {
1408 char c = !len ? 0 : line[0];
1409
1410 set_sign = set;
1411 if (c == '-')
1412 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1413 else if (c == '@')
1414 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1415 else if (c == '+')
1416 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1417 else
1418 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1419 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1420 }
1421 emit_line_ws_markup(o, set_sign, set, reset,
1422 OUTPUT_INDICATOR_NEW, line, len,
1423 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1424 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1425 break;
1426 case DIFF_SYMBOL_MINUS:
1427 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1428 DIFF_SYMBOL_MOVED_LINE_ALT |
1429 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1430 case DIFF_SYMBOL_MOVED_LINE |
1431 DIFF_SYMBOL_MOVED_LINE_ALT |
1432 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1433 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1434 break;
1435 case DIFF_SYMBOL_MOVED_LINE |
1436 DIFF_SYMBOL_MOVED_LINE_ALT:
1437 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1438 break;
1439 case DIFF_SYMBOL_MOVED_LINE |
1440 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1441 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1442 break;
1443 case DIFF_SYMBOL_MOVED_LINE:
1444 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1445 break;
1446 default:
1447 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1448 }
1449 reset = diff_get_color_opt(o, DIFF_RESET);
1450 if (!o->flags.dual_color_diffed_diffs)
1451 set_sign = NULL;
1452 else {
1453 char c = !len ? 0 : line[0];
1454
1455 set_sign = set;
1456 if (c == '+')
1457 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1458 else if (c == '@')
1459 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1460 else if (c == '-')
1461 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1462 else
1463 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1464 }
1465 emit_line_ws_markup(o, set_sign, set, reset,
1466 OUTPUT_INDICATOR_OLD, line, len,
1467 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1468 break;
1469 case DIFF_SYMBOL_WORDS_PORCELAIN:
1470 context = diff_get_color_opt(o, DIFF_CONTEXT);
1471 reset = diff_get_color_opt(o, DIFF_RESET);
1472 emit_line(o, context, reset, line, len);
1473 fputs("~\n", o->file);
1474 break;
1475 case DIFF_SYMBOL_WORDS:
1476 context = diff_get_color_opt(o, DIFF_CONTEXT);
1477 reset = diff_get_color_opt(o, DIFF_RESET);
1478 /*
1479 * Skip the prefix character, if any. With
1480 * diff_suppress_blank_empty, there may be
1481 * none.
1482 */
1483 if (line[0] != '\n') {
1484 line++;
1485 len--;
1486 }
1487 emit_line(o, context, reset, line, len);
1488 break;
1489 case DIFF_SYMBOL_FILEPAIR_PLUS:
1490 meta = diff_get_color_opt(o, DIFF_METAINFO);
1491 reset = diff_get_color_opt(o, DIFF_RESET);
1492 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1493 line, reset,
1494 strchr(line, ' ') ? "\t" : "");
1495 break;
1496 case DIFF_SYMBOL_FILEPAIR_MINUS:
1497 meta = diff_get_color_opt(o, DIFF_METAINFO);
1498 reset = diff_get_color_opt(o, DIFF_RESET);
1499 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1500 line, reset,
1501 strchr(line, ' ') ? "\t" : "");
1502 break;
1503 case DIFF_SYMBOL_BINARY_FILES:
1504 case DIFF_SYMBOL_HEADER:
1505 fprintf(o->file, "%s", line);
1506 break;
1507 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1508 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1509 break;
1510 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1511 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1512 break;
1513 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1514 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1515 break;
1516 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1517 fputs(diff_line_prefix(o), o->file);
1518 fputc('\n', o->file);
1519 break;
1520 case DIFF_SYMBOL_REWRITE_DIFF:
1521 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1522 reset = diff_get_color_opt(o, DIFF_RESET);
1523 emit_line(o, fraginfo, reset, line, len);
1524 break;
1525 case DIFF_SYMBOL_SUBMODULE_ADD:
1526 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1527 reset = diff_get_color_opt(o, DIFF_RESET);
1528 emit_line(o, set, reset, line, len);
1529 break;
1530 case DIFF_SYMBOL_SUBMODULE_DEL:
1531 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1532 reset = diff_get_color_opt(o, DIFF_RESET);
1533 emit_line(o, set, reset, line, len);
1534 break;
1535 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1536 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1537 diff_line_prefix(o), line);
1538 break;
1539 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1540 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1541 diff_line_prefix(o), line);
1542 break;
1543 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1544 emit_line(o, "", "", " 0 files changed\n",
1545 strlen(" 0 files changed\n"));
1546 break;
1547 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1548 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1549 break;
1550 case DIFF_SYMBOL_WORD_DIFF:
1551 fprintf(o->file, "%.*s", len, line);
1552 break;
1553 case DIFF_SYMBOL_STAT_SEP:
1554 fputs(o->stat_sep, o->file);
1555 break;
1556 default:
1557 BUG("unknown diff symbol");
1558 }
1559 }
1560
1561 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1562 const char *line, int len, unsigned flags)
1563 {
1564 struct emitted_diff_symbol e = {
1565 .line = line, .len = len, .flags = flags, .s = s
1566 };
1567
1568 if (o->emitted_symbols)
1569 append_emitted_diff_symbol(o, &e);
1570 else
1571 emit_diff_symbol_from_struct(o, &e);
1572 }
1573
1574 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1575 {
1576 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1577 }
1578
1579 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1580 {
1581 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1582 }
1583
1584 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1585 {
1586 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1587 path, strlen(path), 0);
1588 }
1589
1590 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1591 {
1592 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1593 path, strlen(path), 0);
1594 }
1595
1596 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1597 {
1598 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1599 header, strlen(header), 0);
1600 }
1601
1602 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1603 {
1604 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1605 }
1606
1607 void diff_emit_submodule_pipethrough(struct diff_options *o,
1608 const char *line, int len)
1609 {
1610 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1611 }
1612
1613 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1614 {
1615 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1616 ecbdata->blank_at_eof_in_preimage &&
1617 ecbdata->blank_at_eof_in_postimage &&
1618 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1619 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1620 return 0;
1621 return ws_blank_line(line, len);
1622 }
1623
1624 static void emit_add_line(struct emit_callback *ecbdata,
1625 const char *line, int len)
1626 {
1627 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1628 if (new_blank_line_at_eof(ecbdata, line, len))
1629 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1630
1631 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1632 }
1633
1634 static void emit_del_line(struct emit_callback *ecbdata,
1635 const char *line, int len)
1636 {
1637 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1638 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1639 }
1640
1641 static void emit_context_line(struct emit_callback *ecbdata,
1642 const char *line, int len)
1643 {
1644 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1645 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1646 }
1647
1648 static void emit_hunk_header(struct emit_callback *ecbdata,
1649 const char *line, int len)
1650 {
1651 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1652 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1653 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1654 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1655 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1656 static const char atat[2] = { '@', '@' };
1657 const char *cp, *ep;
1658 struct strbuf msgbuf = STRBUF_INIT;
1659 int org_len = len;
1660 int i = 1;
1661
1662 /*
1663 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1664 * it always is at least 10 bytes long.
1665 */
1666 if (len < 10 ||
1667 memcmp(line, atat, 2) ||
1668 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1669 emit_diff_symbol(ecbdata->opt,
1670 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1671 return;
1672 }
1673 ep += 2; /* skip over @@ */
1674
1675 /* The hunk header in fraginfo color */
1676 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1677 strbuf_addstr(&msgbuf, reverse);
1678 strbuf_addstr(&msgbuf, frag);
1679 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1680 strbuf_add(&msgbuf, atat, sizeof(atat));
1681 else
1682 strbuf_add(&msgbuf, line, ep - line);
1683 strbuf_addstr(&msgbuf, reset);
1684
1685 /*
1686 * trailing "\r\n"
1687 */
1688 for ( ; i < 3; i++)
1689 if (line[len - i] == '\r' || line[len - i] == '\n')
1690 len--;
1691
1692 /* blank before the func header */
1693 for (cp = ep; ep - line < len; ep++)
1694 if (*ep != ' ' && *ep != '\t')
1695 break;
1696 if (ep != cp) {
1697 strbuf_addstr(&msgbuf, context);
1698 strbuf_add(&msgbuf, cp, ep - cp);
1699 strbuf_addstr(&msgbuf, reset);
1700 }
1701
1702 if (ep < line + len) {
1703 strbuf_addstr(&msgbuf, func);
1704 strbuf_add(&msgbuf, ep, line + len - ep);
1705 strbuf_addstr(&msgbuf, reset);
1706 }
1707
1708 strbuf_add(&msgbuf, line + len, org_len - len);
1709 strbuf_complete_line(&msgbuf);
1710 emit_diff_symbol(ecbdata->opt,
1711 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1712 strbuf_release(&msgbuf);
1713 }
1714
1715 static struct diff_tempfile *claim_diff_tempfile(void)
1716 {
1717 int i;
1718 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1719 if (!diff_temp[i].name)
1720 return diff_temp + i;
1721 BUG("diff is failing to clean up its tempfiles");
1722 }
1723
1724 static void remove_tempfile(void)
1725 {
1726 int i;
1727 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1728 if (is_tempfile_active(diff_temp[i].tempfile))
1729 delete_tempfile(&diff_temp[i].tempfile);
1730 diff_temp[i].name = NULL;
1731 }
1732 }
1733
1734 static void add_line_count(struct strbuf *out, int count)
1735 {
1736 switch (count) {
1737 case 0:
1738 strbuf_addstr(out, "0,0");
1739 break;
1740 case 1:
1741 strbuf_addstr(out, "1");
1742 break;
1743 default:
1744 strbuf_addf(out, "1,%d", count);
1745 break;
1746 }
1747 }
1748
1749 static void emit_rewrite_lines(struct emit_callback *ecb,
1750 int prefix, const char *data, int size)
1751 {
1752 const char *endp = NULL;
1753
1754 while (0 < size) {
1755 int len;
1756
1757 endp = memchr(data, '\n', size);
1758 len = endp ? (endp - data + 1) : size;
1759 if (prefix != '+') {
1760 ecb->lno_in_preimage++;
1761 emit_del_line(ecb, data, len);
1762 } else {
1763 ecb->lno_in_postimage++;
1764 emit_add_line(ecb, data, len);
1765 }
1766 size -= len;
1767 data += len;
1768 }
1769 if (!endp)
1770 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1771 }
1772
1773 static void emit_rewrite_diff(const char *name_a,
1774 const char *name_b,
1775 struct diff_filespec *one,
1776 struct diff_filespec *two,
1777 struct userdiff_driver *textconv_one,
1778 struct userdiff_driver *textconv_two,
1779 struct diff_options *o)
1780 {
1781 int lc_a, lc_b;
1782 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1783 const char *a_prefix, *b_prefix;
1784 char *data_one, *data_two;
1785 size_t size_one, size_two;
1786 struct emit_callback ecbdata;
1787 struct strbuf out = STRBUF_INIT;
1788
1789 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1790 a_prefix = o->b_prefix;
1791 b_prefix = o->a_prefix;
1792 } else {
1793 a_prefix = o->a_prefix;
1794 b_prefix = o->b_prefix;
1795 }
1796
1797 name_a += (*name_a == '/');
1798 name_b += (*name_b == '/');
1799
1800 strbuf_reset(&a_name);
1801 strbuf_reset(&b_name);
1802 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1803 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1804
1805 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1806 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1807
1808 memset(&ecbdata, 0, sizeof(ecbdata));
1809 ecbdata.color_diff = want_color(o->use_color);
1810 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1811 ecbdata.opt = o;
1812 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1813 mmfile_t mf1, mf2;
1814 mf1.ptr = (char *)data_one;
1815 mf2.ptr = (char *)data_two;
1816 mf1.size = size_one;
1817 mf2.size = size_two;
1818 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1819 }
1820 ecbdata.lno_in_preimage = 1;
1821 ecbdata.lno_in_postimage = 1;
1822
1823 lc_a = count_lines(data_one, size_one);
1824 lc_b = count_lines(data_two, size_two);
1825
1826 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1827 a_name.buf, a_name.len, 0);
1828 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1829 b_name.buf, b_name.len, 0);
1830
1831 strbuf_addstr(&out, "@@ -");
1832 if (!o->irreversible_delete)
1833 add_line_count(&out, lc_a);
1834 else
1835 strbuf_addstr(&out, "?,?");
1836 strbuf_addstr(&out, " +");
1837 add_line_count(&out, lc_b);
1838 strbuf_addstr(&out, " @@\n");
1839 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1840 strbuf_release(&out);
1841
1842 if (lc_a && !o->irreversible_delete)
1843 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1844 if (lc_b)
1845 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1846 if (textconv_one)
1847 free((char *)data_one);
1848 if (textconv_two)
1849 free((char *)data_two);
1850 }
1851
1852 struct diff_words_buffer {
1853 mmfile_t text;
1854 unsigned long alloc;
1855 struct diff_words_orig {
1856 const char *begin, *end;
1857 } *orig;
1858 int orig_nr, orig_alloc;
1859 };
1860
1861 static void diff_words_append(char *line, unsigned long len,
1862 struct diff_words_buffer *buffer)
1863 {
1864 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1865 line++;
1866 len--;
1867 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1868 buffer->text.size += len;
1869 buffer->text.ptr[buffer->text.size] = '\0';
1870 }
1871
1872 struct diff_words_style_elem {
1873 const char *prefix;
1874 const char *suffix;
1875 const char *color; /* NULL; filled in by the setup code if
1876 * color is enabled */
1877 };
1878
1879 struct diff_words_style {
1880 enum diff_words_type type;
1881 struct diff_words_style_elem new_word, old_word, ctx;
1882 const char *newline;
1883 };
1884
1885 static struct diff_words_style diff_words_styles[] = {
1886 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1887 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1888 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1889 };
1890
1891 struct diff_words_data {
1892 struct diff_words_buffer minus, plus;
1893 const char *current_plus;
1894 int last_minus;
1895 struct diff_options *opt;
1896 regex_t *word_regex;
1897 enum diff_words_type type;
1898 struct diff_words_style *style;
1899 };
1900
1901 static int fn_out_diff_words_write_helper(struct diff_options *o,
1902 struct diff_words_style_elem *st_el,
1903 const char *newline,
1904 size_t count, const char *buf)
1905 {
1906 int print = 0;
1907 struct strbuf sb = STRBUF_INIT;
1908
1909 while (count) {
1910 char *p = memchr(buf, '\n', count);
1911 if (print)
1912 strbuf_addstr(&sb, diff_line_prefix(o));
1913
1914 if (p != buf) {
1915 const char *reset = st_el->color && *st_el->color ?
1916 GIT_COLOR_RESET : NULL;
1917 if (st_el->color && *st_el->color)
1918 strbuf_addstr(&sb, st_el->color);
1919 strbuf_addstr(&sb, st_el->prefix);
1920 strbuf_add(&sb, buf, p ? p - buf : count);
1921 strbuf_addstr(&sb, st_el->suffix);
1922 if (reset)
1923 strbuf_addstr(&sb, reset);
1924 }
1925 if (!p)
1926 goto out;
1927
1928 strbuf_addstr(&sb, newline);
1929 count -= p + 1 - buf;
1930 buf = p + 1;
1931 print = 1;
1932 if (count) {
1933 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1934 sb.buf, sb.len, 0);
1935 strbuf_reset(&sb);
1936 }
1937 }
1938
1939 out:
1940 if (sb.len)
1941 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1942 sb.buf, sb.len, 0);
1943 strbuf_release(&sb);
1944 return 0;
1945 }
1946
1947 /*
1948 * '--color-words' algorithm can be described as:
1949 *
1950 * 1. collect the minus/plus lines of a diff hunk, divided into
1951 * minus-lines and plus-lines;
1952 *
1953 * 2. break both minus-lines and plus-lines into words and
1954 * place them into two mmfile_t with one word for each line;
1955 *
1956 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1957 *
1958 * And for the common parts of the both file, we output the plus side text.
1959 * diff_words->current_plus is used to trace the current position of the plus file
1960 * which printed. diff_words->last_minus is used to trace the last minus word
1961 * printed.
1962 *
1963 * For '--graph' to work with '--color-words', we need to output the graph prefix
1964 * on each line of color words output. Generally, there are two conditions on
1965 * which we should output the prefix.
1966 *
1967 * 1. diff_words->last_minus == 0 &&
1968 * diff_words->current_plus == diff_words->plus.text.ptr
1969 *
1970 * that is: the plus text must start as a new line, and if there is no minus
1971 * word printed, a graph prefix must be printed.
1972 *
1973 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1974 * *(diff_words->current_plus - 1) == '\n'
1975 *
1976 * that is: a graph prefix must be printed following a '\n'
1977 */
1978 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1979 {
1980 if ((diff_words->last_minus == 0 &&
1981 diff_words->current_plus == diff_words->plus.text.ptr) ||
1982 (diff_words->current_plus > diff_words->plus.text.ptr &&
1983 *(diff_words->current_plus - 1) == '\n')) {
1984 return 1;
1985 } else {
1986 return 0;
1987 }
1988 }
1989
1990 static void fn_out_diff_words_aux(void *priv,
1991 long minus_first, long minus_len,
1992 long plus_first, long plus_len,
1993 const char *func UNUSED, long funclen UNUSED)
1994 {
1995 struct diff_words_data *diff_words = priv;
1996 struct diff_words_style *style = diff_words->style;
1997 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1998 struct diff_options *opt = diff_words->opt;
1999 const char *line_prefix;
2000
2001 assert(opt);
2002 line_prefix = diff_line_prefix(opt);
2003
2004 /* POSIX requires that first be decremented by one if len == 0... */
2005 if (minus_len) {
2006 minus_begin = diff_words->minus.orig[minus_first].begin;
2007 minus_end =
2008 diff_words->minus.orig[minus_first + minus_len - 1].end;
2009 } else
2010 minus_begin = minus_end =
2011 diff_words->minus.orig[minus_first].end;
2012
2013 if (plus_len) {
2014 plus_begin = diff_words->plus.orig[plus_first].begin;
2015 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2016 } else
2017 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2018
2019 if (color_words_output_graph_prefix(diff_words)) {
2020 fputs(line_prefix, diff_words->opt->file);
2021 }
2022 if (diff_words->current_plus != plus_begin) {
2023 fn_out_diff_words_write_helper(diff_words->opt,
2024 &style->ctx, style->newline,
2025 plus_begin - diff_words->current_plus,
2026 diff_words->current_plus);
2027 }
2028 if (minus_begin != minus_end) {
2029 fn_out_diff_words_write_helper(diff_words->opt,
2030 &style->old_word, style->newline,
2031 minus_end - minus_begin, minus_begin);
2032 }
2033 if (plus_begin != plus_end) {
2034 fn_out_diff_words_write_helper(diff_words->opt,
2035 &style->new_word, style->newline,
2036 plus_end - plus_begin, plus_begin);
2037 }
2038
2039 diff_words->current_plus = plus_end;
2040 diff_words->last_minus = minus_first;
2041 }
2042
2043 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2044 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2045 int *begin, int *end)
2046 {
2047 while (word_regex && *begin < buffer->size) {
2048 regmatch_t match[1];
2049 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2050 buffer->size - *begin, 1, match, 0)) {
2051 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2052 '\n', match[0].rm_eo - match[0].rm_so);
2053 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2054 *begin += match[0].rm_so;
2055 if (*begin == *end)
2056 (*begin)++;
2057 else
2058 return *begin > *end;
2059 } else {
2060 return -1;
2061 }
2062 }
2063
2064 /* find the next word */
2065 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2066 (*begin)++;
2067 if (*begin >= buffer->size)
2068 return -1;
2069
2070 /* find the end of the word */
2071 *end = *begin + 1;
2072 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2073 (*end)++;
2074
2075 return 0;
2076 }
2077
2078 /*
2079 * This function splits the words in buffer->text, stores the list with
2080 * newline separator into out, and saves the offsets of the original words
2081 * in buffer->orig.
2082 */
2083 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2084 regex_t *word_regex)
2085 {
2086 int i, j;
2087 long alloc = 0;
2088
2089 out->size = 0;
2090 out->ptr = NULL;
2091
2092 /* fake an empty "0th" word */
2093 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2094 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2095 buffer->orig_nr = 1;
2096
2097 for (i = 0; i < buffer->text.size; i++) {
2098 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2099 return;
2100
2101 /* store original boundaries */
2102 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2103 buffer->orig_alloc);
2104 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2105 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2106 buffer->orig_nr++;
2107
2108 /* store one word */
2109 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2110 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2111 out->ptr[out->size + j - i] = '\n';
2112 out->size += j - i + 1;
2113
2114 i = j - 1;
2115 }
2116 }
2117
2118 /* this executes the word diff on the accumulated buffers */
2119 static void diff_words_show(struct diff_words_data *diff_words)
2120 {
2121 xpparam_t xpp;
2122 xdemitconf_t xecfg;
2123 mmfile_t minus, plus;
2124 struct diff_words_style *style = diff_words->style;
2125
2126 struct diff_options *opt = diff_words->opt;
2127 const char *line_prefix;
2128
2129 assert(opt);
2130 line_prefix = diff_line_prefix(opt);
2131
2132 /* special case: only removal */
2133 if (!diff_words->plus.text.size) {
2134 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2135 line_prefix, strlen(line_prefix), 0);
2136 fn_out_diff_words_write_helper(diff_words->opt,
2137 &style->old_word, style->newline,
2138 diff_words->minus.text.size,
2139 diff_words->minus.text.ptr);
2140 diff_words->minus.text.size = 0;
2141 return;
2142 }
2143
2144 diff_words->current_plus = diff_words->plus.text.ptr;
2145 diff_words->last_minus = 0;
2146
2147 memset(&xpp, 0, sizeof(xpp));
2148 memset(&xecfg, 0, sizeof(xecfg));
2149 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2150 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2151 xpp.flags = 0;
2152 /* as only the hunk header will be parsed, we need a 0-context */
2153 xecfg.ctxlen = 0;
2154 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2155 diff_words, &xpp, &xecfg))
2156 die("unable to generate word diff");
2157 free(minus.ptr);
2158 free(plus.ptr);
2159 if (diff_words->current_plus != diff_words->plus.text.ptr +
2160 diff_words->plus.text.size) {
2161 if (color_words_output_graph_prefix(diff_words))
2162 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2163 line_prefix, strlen(line_prefix), 0);
2164 fn_out_diff_words_write_helper(diff_words->opt,
2165 &style->ctx, style->newline,
2166 diff_words->plus.text.ptr + diff_words->plus.text.size
2167 - diff_words->current_plus, diff_words->current_plus);
2168 }
2169 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2170 }
2171
2172 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2173 static void diff_words_flush(struct emit_callback *ecbdata)
2174 {
2175 struct diff_options *wo = ecbdata->diff_words->opt;
2176
2177 if (ecbdata->diff_words->minus.text.size ||
2178 ecbdata->diff_words->plus.text.size)
2179 diff_words_show(ecbdata->diff_words);
2180
2181 if (wo->emitted_symbols) {
2182 struct diff_options *o = ecbdata->opt;
2183 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2184 int i;
2185
2186 /*
2187 * NEEDSWORK:
2188 * Instead of appending each, concat all words to a line?
2189 */
2190 for (i = 0; i < wol->nr; i++)
2191 append_emitted_diff_symbol(o, &wol->buf[i]);
2192
2193 for (i = 0; i < wol->nr; i++)
2194 free((void *)wol->buf[i].line);
2195
2196 wol->nr = 0;
2197 }
2198 }
2199
2200 static void diff_filespec_load_driver(struct diff_filespec *one,
2201 struct index_state *istate)
2202 {
2203 /* Use already-loaded driver */
2204 if (one->driver)
2205 return;
2206
2207 if (S_ISREG(one->mode))
2208 one->driver = userdiff_find_by_path(istate, one->path);
2209
2210 /* Fallback to default settings */
2211 if (!one->driver)
2212 one->driver = userdiff_find_by_name("default");
2213 }
2214
2215 static const char *userdiff_word_regex(struct diff_filespec *one,
2216 struct index_state *istate)
2217 {
2218 diff_filespec_load_driver(one, istate);
2219 return one->driver->word_regex;
2220 }
2221
2222 static void init_diff_words_data(struct emit_callback *ecbdata,
2223 struct diff_options *orig_opts,
2224 struct diff_filespec *one,
2225 struct diff_filespec *two)
2226 {
2227 int i;
2228 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2229 memcpy(o, orig_opts, sizeof(struct diff_options));
2230
2231 CALLOC_ARRAY(ecbdata->diff_words, 1);
2232 ecbdata->diff_words->type = o->word_diff;
2233 ecbdata->diff_words->opt = o;
2234
2235 if (orig_opts->emitted_symbols)
2236 CALLOC_ARRAY(o->emitted_symbols, 1);
2237
2238 if (!o->word_regex)
2239 o->word_regex = userdiff_word_regex(one, o->repo->index);
2240 if (!o->word_regex)
2241 o->word_regex = userdiff_word_regex(two, o->repo->index);
2242 if (!o->word_regex)
2243 o->word_regex = diff_word_regex_cfg;
2244 if (o->word_regex) {
2245 ecbdata->diff_words->word_regex = (regex_t *)
2246 xmalloc(sizeof(regex_t));
2247 if (regcomp(ecbdata->diff_words->word_regex,
2248 o->word_regex,
2249 REG_EXTENDED | REG_NEWLINE))
2250 die("invalid regular expression: %s",
2251 o->word_regex);
2252 }
2253 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2254 if (o->word_diff == diff_words_styles[i].type) {
2255 ecbdata->diff_words->style =
2256 &diff_words_styles[i];
2257 break;
2258 }
2259 }
2260 if (want_color(o->use_color)) {
2261 struct diff_words_style *st = ecbdata->diff_words->style;
2262 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2263 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2264 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2265 }
2266 }
2267
2268 static void free_diff_words_data(struct emit_callback *ecbdata)
2269 {
2270 if (ecbdata->diff_words) {
2271 diff_words_flush(ecbdata);
2272 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2273 free (ecbdata->diff_words->opt);
2274 free (ecbdata->diff_words->minus.text.ptr);
2275 free (ecbdata->diff_words->minus.orig);
2276 free (ecbdata->diff_words->plus.text.ptr);
2277 free (ecbdata->diff_words->plus.orig);
2278 if (ecbdata->diff_words->word_regex) {
2279 regfree(ecbdata->diff_words->word_regex);
2280 free(ecbdata->diff_words->word_regex);
2281 }
2282 FREE_AND_NULL(ecbdata->diff_words);
2283 }
2284 }
2285
2286 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2287 {
2288 if (want_color(diff_use_color))
2289 return diff_colors[ix];
2290 return "";
2291 }
2292
2293 const char *diff_line_prefix(struct diff_options *opt)
2294 {
2295 struct strbuf *msgbuf;
2296 if (!opt->output_prefix)
2297 return "";
2298
2299 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2300 return msgbuf->buf;
2301 }
2302
2303 static unsigned long sane_truncate_line(char *line, unsigned long len)
2304 {
2305 const char *cp;
2306 unsigned long allot;
2307 size_t l = len;
2308
2309 cp = line;
2310 allot = l;
2311 while (0 < l) {
2312 (void) utf8_width(&cp, &l);
2313 if (!cp)
2314 break; /* truncated in the middle? */
2315 }
2316 return allot - l;
2317 }
2318
2319 static void find_lno(const char *line, struct emit_callback *ecbdata)
2320 {
2321 const char *p;
2322 ecbdata->lno_in_preimage = 0;
2323 ecbdata->lno_in_postimage = 0;
2324 p = strchr(line, '-');
2325 if (!p)
2326 return; /* cannot happen */
2327 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2328 p = strchr(p, '+');
2329 if (!p)
2330 return; /* cannot happen */
2331 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2332 }
2333
2334 static int fn_out_consume(void *priv, char *line, unsigned long len)
2335 {
2336 struct emit_callback *ecbdata = priv;
2337 struct diff_options *o = ecbdata->opt;
2338
2339 o->found_changes = 1;
2340
2341 if (ecbdata->header) {
2342 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2343 ecbdata->header->buf, ecbdata->header->len, 0);
2344 strbuf_reset(ecbdata->header);
2345 ecbdata->header = NULL;
2346 }
2347
2348 if (ecbdata->label_path[0]) {
2349 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2350 ecbdata->label_path[0],
2351 strlen(ecbdata->label_path[0]), 0);
2352 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2353 ecbdata->label_path[1],
2354 strlen(ecbdata->label_path[1]), 0);
2355 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2356 }
2357
2358 if (diff_suppress_blank_empty
2359 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2360 line[0] = '\n';
2361 len = 1;
2362 }
2363
2364 if (line[0] == '@') {
2365 if (ecbdata->diff_words)
2366 diff_words_flush(ecbdata);
2367 len = sane_truncate_line(line, len);
2368 find_lno(line, ecbdata);
2369 emit_hunk_header(ecbdata, line, len);
2370 return 0;
2371 }
2372
2373 if (ecbdata->diff_words) {
2374 enum diff_symbol s =
2375 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2376 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2377 if (line[0] == '-') {
2378 diff_words_append(line, len,
2379 &ecbdata->diff_words->minus);
2380 return 0;
2381 } else if (line[0] == '+') {
2382 diff_words_append(line, len,
2383 &ecbdata->diff_words->plus);
2384 return 0;
2385 } else if (starts_with(line, "\\ ")) {
2386 /*
2387 * Eat the "no newline at eof" marker as if we
2388 * saw a "+" or "-" line with nothing on it,
2389 * and return without diff_words_flush() to
2390 * defer processing. If this is the end of
2391 * preimage, more "+" lines may come after it.
2392 */
2393 return 0;
2394 }
2395 diff_words_flush(ecbdata);
2396 emit_diff_symbol(o, s, line, len, 0);
2397 return 0;
2398 }
2399
2400 switch (line[0]) {
2401 case '+':
2402 ecbdata->lno_in_postimage++;
2403 emit_add_line(ecbdata, line + 1, len - 1);
2404 break;
2405 case '-':
2406 ecbdata->lno_in_preimage++;
2407 emit_del_line(ecbdata, line + 1, len - 1);
2408 break;
2409 case ' ':
2410 ecbdata->lno_in_postimage++;
2411 ecbdata->lno_in_preimage++;
2412 emit_context_line(ecbdata, line + 1, len - 1);
2413 break;
2414 default:
2415 /* incomplete line at the end */
2416 ecbdata->lno_in_preimage++;
2417 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2418 line, len, 0);
2419 break;
2420 }
2421 return 0;
2422 }
2423
2424 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2425 {
2426 const char *old_name = a;
2427 const char *new_name = b;
2428 int pfx_length, sfx_length;
2429 int pfx_adjust_for_slash;
2430 int len_a = strlen(a);
2431 int len_b = strlen(b);
2432 int a_midlen, b_midlen;
2433 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2434 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2435
2436 if (qlen_a || qlen_b) {
2437 quote_c_style(a, name, NULL, 0);
2438 strbuf_addstr(name, " => ");
2439 quote_c_style(b, name, NULL, 0);
2440 return;
2441 }
2442
2443 /* Find common prefix */
2444 pfx_length = 0;
2445 while (*old_name && *new_name && *old_name == *new_name) {
2446 if (*old_name == '/')
2447 pfx_length = old_name - a + 1;
2448 old_name++;
2449 new_name++;
2450 }
2451
2452 /* Find common suffix */
2453 old_name = a + len_a;
2454 new_name = b + len_b;
2455 sfx_length = 0;
2456 /*
2457 * If there is a common prefix, it must end in a slash. In
2458 * that case we let this loop run 1 into the prefix to see the
2459 * same slash.
2460 *
2461 * If there is no common prefix, we cannot do this as it would
2462 * underrun the input strings.
2463 */
2464 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2465 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2466 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2467 *old_name == *new_name) {
2468 if (*old_name == '/')
2469 sfx_length = len_a - (old_name - a);
2470 old_name--;
2471 new_name--;
2472 }
2473
2474 /*
2475 * pfx{mid-a => mid-b}sfx
2476 * {pfx-a => pfx-b}sfx
2477 * pfx{sfx-a => sfx-b}
2478 * name-a => name-b
2479 */
2480 a_midlen = len_a - pfx_length - sfx_length;
2481 b_midlen = len_b - pfx_length - sfx_length;
2482 if (a_midlen < 0)
2483 a_midlen = 0;
2484 if (b_midlen < 0)
2485 b_midlen = 0;
2486
2487 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2488 if (pfx_length + sfx_length) {
2489 strbuf_add(name, a, pfx_length);
2490 strbuf_addch(name, '{');
2491 }
2492 strbuf_add(name, a + pfx_length, a_midlen);
2493 strbuf_addstr(name, " => ");
2494 strbuf_add(name, b + pfx_length, b_midlen);
2495 if (pfx_length + sfx_length) {
2496 strbuf_addch(name, '}');
2497 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2498 }
2499 }
2500
2501 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2502 const char *name_a,
2503 const char *name_b)
2504 {
2505 struct diffstat_file *x;
2506 CALLOC_ARRAY(x, 1);
2507 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2508 diffstat->files[diffstat->nr++] = x;
2509 if (name_b) {
2510 x->from_name = xstrdup(name_a);
2511 x->name = xstrdup(name_b);
2512 x->is_renamed = 1;
2513 }
2514 else {
2515 x->from_name = NULL;
2516 x->name = xstrdup(name_a);
2517 }
2518 return x;
2519 }
2520
2521 static int diffstat_consume(void *priv, char *line, unsigned long len)
2522 {
2523 struct diffstat_t *diffstat = priv;
2524 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2525
2526 if (!len)
2527 BUG("xdiff fed us an empty line");
2528
2529 if (line[0] == '+')
2530 x->added++;
2531 else if (line[0] == '-')
2532 x->deleted++;
2533 return 0;
2534 }
2535
2536 const char mime_boundary_leader[] = "------------";
2537
2538 static int scale_linear(int it, int width, int max_change)
2539 {
2540 if (!it)
2541 return 0;
2542 /*
2543 * make sure that at least one '-' or '+' is printed if
2544 * there is any change to this path. The easiest way is to
2545 * scale linearly as if the allotted width is one column shorter
2546 * than it is, and then add 1 to the result.
2547 */
2548 return 1 + (it * (width - 1) / max_change);
2549 }
2550
2551 static void show_graph(struct strbuf *out, char ch, int cnt,
2552 const char *set, const char *reset)
2553 {
2554 if (cnt <= 0)
2555 return;
2556 strbuf_addstr(out, set);
2557 strbuf_addchars(out, ch, cnt);
2558 strbuf_addstr(out, reset);
2559 }
2560
2561 static void fill_print_name(struct diffstat_file *file)
2562 {
2563 struct strbuf pname = STRBUF_INIT;
2564
2565 if (file->print_name)
2566 return;
2567
2568 if (file->is_renamed)
2569 pprint_rename(&pname, file->from_name, file->name);
2570 else
2571 quote_c_style(file->name, &pname, NULL, 0);
2572
2573 if (file->comments)
2574 strbuf_addf(&pname, " (%s)", file->comments);
2575
2576 file->print_name = strbuf_detach(&pname, NULL);
2577 }
2578
2579 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2580 int files, int insertions, int deletions)
2581 {
2582 struct strbuf sb = STRBUF_INIT;
2583
2584 if (!files) {
2585 assert(insertions == 0 && deletions == 0);
2586 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2587 NULL, 0, 0);
2588 return;
2589 }
2590
2591 strbuf_addf(&sb,
2592 (files == 1) ? " %d file changed" : " %d files changed",
2593 files);
2594
2595 /*
2596 * For binary diff, the caller may want to print "x files
2597 * changed" with insertions == 0 && deletions == 0.
2598 *
2599 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2600 * is probably less confusing (i.e skip over "2 files changed
2601 * but nothing about added/removed lines? Is this a bug in Git?").
2602 */
2603 if (insertions || deletions == 0) {
2604 strbuf_addf(&sb,
2605 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2606 insertions);
2607 }
2608
2609 if (deletions || insertions == 0) {
2610 strbuf_addf(&sb,
2611 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2612 deletions);
2613 }
2614 strbuf_addch(&sb, '\n');
2615 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2616 sb.buf, sb.len, 0);
2617 strbuf_release(&sb);
2618 }
2619
2620 void print_stat_summary(FILE *fp, int files,
2621 int insertions, int deletions)
2622 {
2623 struct diff_options o;
2624 memset(&o, 0, sizeof(o));
2625 o.file = fp;
2626
2627 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2628 }
2629
2630 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2631 {
2632 int i, len, add, del, adds = 0, dels = 0;
2633 uintmax_t max_change = 0, max_len = 0;
2634 int total_files = data->nr, count;
2635 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2636 const char *reset, *add_c, *del_c;
2637 int extra_shown = 0;
2638 const char *line_prefix = diff_line_prefix(options);
2639 struct strbuf out = STRBUF_INIT;
2640
2641 if (data->nr == 0)
2642 return;
2643
2644 count = options->stat_count ? options->stat_count : data->nr;
2645
2646 reset = diff_get_color_opt(options, DIFF_RESET);
2647 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2648 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2649
2650 /*
2651 * Find the longest filename and max number of changes
2652 */
2653 for (i = 0; (i < count) && (i < data->nr); i++) {
2654 struct diffstat_file *file = data->files[i];
2655 uintmax_t change = file->added + file->deleted;
2656
2657 if (!file->is_interesting && (change == 0)) {
2658 count++; /* not shown == room for one more */
2659 continue;
2660 }
2661 fill_print_name(file);
2662 len = utf8_strwidth(file->print_name);
2663 if (max_len < len)
2664 max_len = len;
2665
2666 if (file->is_unmerged) {
2667 /* "Unmerged" is 8 characters */
2668 bin_width = bin_width < 8 ? 8 : bin_width;
2669 continue;
2670 }
2671 if (file->is_binary) {
2672 /* "Bin XXX -> YYY bytes" */
2673 int w = 14 + decimal_width(file->added)
2674 + decimal_width(file->deleted);
2675 bin_width = bin_width < w ? w : bin_width;
2676 /* Display change counts aligned with "Bin" */
2677 number_width = 3;
2678 continue;
2679 }
2680
2681 if (max_change < change)
2682 max_change = change;
2683 }
2684 count = i; /* where we can stop scanning in data->files[] */
2685
2686 /*
2687 * We have width = stat_width or term_columns() columns total.
2688 * We want a maximum of min(max_len, stat_name_width) for the name part.
2689 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2690 * We also need 1 for " " and 4 + decimal_width(max_change)
2691 * for " | NNNN " and one the empty column at the end, altogether
2692 * 6 + decimal_width(max_change).
2693 *
2694 * If there's not enough space, we will use the smaller of
2695 * stat_name_width (if set) and 5/8*width for the filename,
2696 * and the rest for constant elements + graph part, but no more
2697 * than stat_graph_width for the graph part.
2698 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2699 * for the standard terminal size).
2700 *
2701 * In other words: stat_width limits the maximum width, and
2702 * stat_name_width fixes the maximum width of the filename,
2703 * and is also used to divide available columns if there
2704 * aren't enough.
2705 *
2706 * Binary files are displayed with "Bin XXX -> YYY bytes"
2707 * instead of the change count and graph. This part is treated
2708 * similarly to the graph part, except that it is not
2709 * "scaled". If total width is too small to accommodate the
2710 * guaranteed minimum width of the filename part and the
2711 * separators and this message, this message will "overflow"
2712 * making the line longer than the maximum width.
2713 */
2714
2715 /*
2716 * NEEDSWORK: line_prefix is often used for "log --graph" output
2717 * and contains ANSI-colored string. utf8_strnwidth() should be
2718 * used to correctly count the display width instead of strlen().
2719 */
2720 if (options->stat_width == -1)
2721 width = term_columns() - strlen(line_prefix);
2722 else
2723 width = options->stat_width ? options->stat_width : 80;
2724 number_width = decimal_width(max_change) > number_width ?
2725 decimal_width(max_change) : number_width;
2726
2727 if (options->stat_name_width == -1)
2728 options->stat_name_width = diff_stat_name_width;
2729 if (options->stat_graph_width == -1)
2730 options->stat_graph_width = diff_stat_graph_width;
2731
2732 /*
2733 * Guarantee 3/8*16 == 6 for the graph part
2734 * and 5/8*16 == 10 for the filename part
2735 */
2736 if (width < 16 + 6 + number_width)
2737 width = 16 + 6 + number_width;
2738
2739 /*
2740 * First assign sizes that are wanted, ignoring available width.
2741 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2742 * starting from "XXX" should fit in graph_width.
2743 */
2744 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2745 if (options->stat_graph_width &&
2746 options->stat_graph_width < graph_width)
2747 graph_width = options->stat_graph_width;
2748
2749 name_width = (options->stat_name_width > 0 &&
2750 options->stat_name_width < max_len) ?
2751 options->stat_name_width : max_len;
2752
2753 /*
2754 * Adjust adjustable widths not to exceed maximum width
2755 */
2756 if (name_width + number_width + 6 + graph_width > width) {
2757 if (graph_width > width * 3/8 - number_width - 6) {
2758 graph_width = width * 3/8 - number_width - 6;
2759 if (graph_width < 6)
2760 graph_width = 6;
2761 }
2762
2763 if (options->stat_graph_width &&
2764 graph_width > options->stat_graph_width)
2765 graph_width = options->stat_graph_width;
2766 if (name_width > width - number_width - 6 - graph_width)
2767 name_width = width - number_width - 6 - graph_width;
2768 else
2769 graph_width = width - number_width - 6 - name_width;
2770 }
2771
2772 /*
2773 * From here name_width is the width of the name area,
2774 * and graph_width is the width of the graph area.
2775 * max_change is used to scale graph properly.
2776 */
2777 for (i = 0; i < count; i++) {
2778 const char *prefix = "";
2779 struct diffstat_file *file = data->files[i];
2780 char *name = file->print_name;
2781 uintmax_t added = file->added;
2782 uintmax_t deleted = file->deleted;
2783 int name_len, padding;
2784
2785 if (!file->is_interesting && (added + deleted == 0))
2786 continue;
2787
2788 /*
2789 * "scale" the filename
2790 */
2791 len = name_width;
2792 name_len = utf8_strwidth(name);
2793 if (name_width < name_len) {
2794 char *slash;
2795 prefix = "...";
2796 len -= 3;
2797 /*
2798 * NEEDSWORK: (name_len - len) counts the display
2799 * width, which would be shorter than the byte
2800 * length of the corresponding substring.
2801 * Advancing "name" by that number of bytes does
2802 * *NOT* skip over that many columns, so it is
2803 * very likely that chomping the pathname at the
2804 * slash we will find starting from "name" will
2805 * leave the resulting string still too long.
2806 */
2807 name += name_len - len;
2808 slash = strchr(name, '/');
2809 if (slash)
2810 name = slash;
2811 }
2812 padding = len - utf8_strwidth(name);
2813 if (padding < 0)
2814 padding = 0;
2815
2816 if (file->is_binary) {
2817 strbuf_addf(&out, " %s%s%*s | %*s",
2818 prefix, name, padding, "",
2819 number_width, "Bin");
2820 if (!added && !deleted) {
2821 strbuf_addch(&out, '\n');
2822 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2823 out.buf, out.len, 0);
2824 strbuf_reset(&out);
2825 continue;
2826 }
2827 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2828 del_c, deleted, reset);
2829 strbuf_addstr(&out, " -> ");
2830 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2831 add_c, added, reset);
2832 strbuf_addstr(&out, " bytes\n");
2833 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2834 out.buf, out.len, 0);
2835 strbuf_reset(&out);
2836 continue;
2837 }
2838 else if (file->is_unmerged) {
2839 strbuf_addf(&out, " %s%s%*s | %*s",
2840 prefix, name, padding, "",
2841 number_width, "Unmerged\n");
2842 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2843 out.buf, out.len, 0);
2844 strbuf_reset(&out);
2845 continue;
2846 }
2847
2848 /*
2849 * scale the add/delete
2850 */
2851 add = added;
2852 del = deleted;
2853
2854 if (graph_width <= max_change) {
2855 int total = scale_linear(add + del, graph_width, max_change);
2856 if (total < 2 && add && del)
2857 /* width >= 2 due to the sanity check */
2858 total = 2;
2859 if (add < del) {
2860 add = scale_linear(add, graph_width, max_change);
2861 del = total - add;
2862 } else {
2863 del = scale_linear(del, graph_width, max_change);
2864 add = total - del;
2865 }
2866 }
2867 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2868 prefix, name, padding, "",
2869 number_width, added + deleted,
2870 added + deleted ? " " : "");
2871 show_graph(&out, '+', add, add_c, reset);
2872 show_graph(&out, '-', del, del_c, reset);
2873 strbuf_addch(&out, '\n');
2874 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2875 out.buf, out.len, 0);
2876 strbuf_reset(&out);
2877 }
2878
2879 for (i = 0; i < data->nr; i++) {
2880 struct diffstat_file *file = data->files[i];
2881 uintmax_t added = file->added;
2882 uintmax_t deleted = file->deleted;
2883
2884 if (file->is_unmerged ||
2885 (!file->is_interesting && (added + deleted == 0))) {
2886 total_files--;
2887 continue;
2888 }
2889
2890 if (!file->is_binary) {
2891 adds += added;
2892 dels += deleted;
2893 }
2894 if (i < count)
2895 continue;
2896 if (!extra_shown)
2897 emit_diff_symbol(options,
2898 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2899 NULL, 0, 0);
2900 extra_shown = 1;
2901 }
2902
2903 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2904 strbuf_release(&out);
2905 }
2906
2907 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2908 {
2909 int i, adds = 0, dels = 0, total_files = data->nr;
2910
2911 if (data->nr == 0)
2912 return;
2913
2914 for (i = 0; i < data->nr; i++) {
2915 int added = data->files[i]->added;
2916 int deleted = data->files[i]->deleted;
2917
2918 if (data->files[i]->is_unmerged ||
2919 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2920 total_files--;
2921 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2922 adds += added;
2923 dels += deleted;
2924 }
2925 }
2926 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2927 }
2928
2929 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2930 {
2931 int i;
2932
2933 if (data->nr == 0)
2934 return;
2935
2936 for (i = 0; i < data->nr; i++) {
2937 struct diffstat_file *file = data->files[i];
2938
2939 fprintf(options->file, "%s", diff_line_prefix(options));
2940
2941 if (file->is_binary)
2942 fprintf(options->file, "-\t-\t");
2943 else
2944 fprintf(options->file,
2945 "%"PRIuMAX"\t%"PRIuMAX"\t",
2946 file->added, file->deleted);
2947 if (options->line_termination) {
2948 fill_print_name(file);
2949 if (!file->is_renamed)
2950 write_name_quoted(file->name, options->file,
2951 options->line_termination);
2952 else {
2953 fputs(file->print_name, options->file);
2954 putc(options->line_termination, options->file);
2955 }
2956 } else {
2957 if (file->is_renamed) {
2958 putc('\0', options->file);
2959 write_name_quoted(file->from_name, options->file, '\0');
2960 }
2961 write_name_quoted(file->name, options->file, '\0');
2962 }
2963 }
2964 }
2965
2966 struct dirstat_file {
2967 const char *name;
2968 unsigned long changed;
2969 };
2970
2971 struct dirstat_dir {
2972 struct dirstat_file *files;
2973 int alloc, nr, permille, cumulative;
2974 };
2975
2976 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2977 unsigned long changed, const char *base, int baselen)
2978 {
2979 unsigned long sum_changes = 0;
2980 unsigned int sources = 0;
2981 const char *line_prefix = diff_line_prefix(opt);
2982
2983 while (dir->nr) {
2984 struct dirstat_file *f = dir->files;
2985 int namelen = strlen(f->name);
2986 unsigned long changes;
2987 char *slash;
2988
2989 if (namelen < baselen)
2990 break;
2991 if (memcmp(f->name, base, baselen))
2992 break;
2993 slash = strchr(f->name + baselen, '/');
2994 if (slash) {
2995 int newbaselen = slash + 1 - f->name;
2996 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2997 sources++;
2998 } else {
2999 changes = f->changed;
3000 dir->files++;
3001 dir->nr--;
3002 sources += 2;
3003 }
3004 sum_changes += changes;
3005 }
3006
3007 /*
3008 * We don't report dirstat's for
3009 * - the top level
3010 * - or cases where everything came from a single directory
3011 * under this directory (sources == 1).
3012 */
3013 if (baselen && sources != 1) {
3014 if (sum_changes) {
3015 int permille = sum_changes * 1000 / changed;
3016 if (permille >= dir->permille) {
3017 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
3018 permille / 10, permille % 10, baselen, base);
3019 if (!dir->cumulative)
3020 return 0;
3021 }
3022 }
3023 }
3024 return sum_changes;
3025 }
3026
3027 static int dirstat_compare(const void *_a, const void *_b)
3028 {
3029 const struct dirstat_file *a = _a;
3030 const struct dirstat_file *b = _b;
3031 return strcmp(a->name, b->name);
3032 }
3033
3034 static void conclude_dirstat(struct diff_options *options,
3035 struct dirstat_dir *dir,
3036 unsigned long changed)
3037 {
3038 struct dirstat_file *to_free = dir->files;
3039
3040 if (!changed) {
3041 /* This can happen even with many files, if everything was renames */
3042 ;
3043 } else {
3044 /* Show all directories with more than x% of the changes */
3045 QSORT(dir->files, dir->nr, dirstat_compare);
3046 gather_dirstat(options, dir, changed, "", 0);
3047 }
3048
3049 free(to_free);
3050 }
3051
3052 static void show_dirstat(struct diff_options *options)
3053 {
3054 int i;
3055 unsigned long changed;
3056 struct dirstat_dir dir;
3057 struct diff_queue_struct *q = &diff_queued_diff;
3058
3059 dir.files = NULL;
3060 dir.alloc = 0;
3061 dir.nr = 0;
3062 dir.permille = options->dirstat_permille;
3063 dir.cumulative = options->flags.dirstat_cumulative;
3064
3065 changed = 0;
3066 for (i = 0; i < q->nr; i++) {
3067 struct diff_filepair *p = q->queue[i];
3068 const char *name;
3069 unsigned long copied, added, damage;
3070 struct diff_populate_filespec_options dpf_options = {
3071 .check_size_only = 1,
3072 };
3073
3074 name = p->two->path ? p->two->path : p->one->path;
3075
3076 if (p->one->oid_valid && p->two->oid_valid &&
3077 oideq(&p->one->oid, &p->two->oid)) {
3078 /*
3079 * The SHA1 has not changed, so pre-/post-content is
3080 * identical. We can therefore skip looking at the
3081 * file contents altogether.
3082 */
3083 damage = 0;
3084 goto found_damage;
3085 }
3086
3087 if (options->flags.dirstat_by_file) {
3088 /*
3089 * In --dirstat-by-file mode, we don't really need to
3090 * look at the actual file contents at all.
3091 * The fact that the SHA1 changed is enough for us to
3092 * add this file to the list of results
3093 * (with each file contributing equal damage).
3094 */
3095 damage = 1;
3096 goto found_damage;
3097 }
3098
3099 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3100 diff_populate_filespec(options->repo, p->one, NULL);
3101 diff_populate_filespec(options->repo, p->two, NULL);
3102 diffcore_count_changes(options->repo,
3103 p->one, p->two, NULL, NULL,
3104 &copied, &added);
3105 diff_free_filespec_data(p->one);
3106 diff_free_filespec_data(p->two);
3107 } else if (DIFF_FILE_VALID(p->one)) {
3108 diff_populate_filespec(options->repo, p->one, &dpf_options);
3109 copied = added = 0;
3110 diff_free_filespec_data(p->one);
3111 } else if (DIFF_FILE_VALID(p->two)) {
3112 diff_populate_filespec(options->repo, p->two, &dpf_options);
3113 copied = 0;
3114 added = p->two->size;
3115 diff_free_filespec_data(p->two);
3116 } else
3117 continue;
3118
3119 /*
3120 * Original minus copied is the removed material,
3121 * added is the new material. They are both damages
3122 * made to the preimage.
3123 * If the resulting damage is zero, we know that
3124 * diffcore_count_changes() considers the two entries to
3125 * be identical, but since the oid changed, we
3126 * know that there must have been _some_ kind of change,
3127 * so we force all entries to have damage > 0.
3128 */
3129 damage = (p->one->size - copied) + added;
3130 if (!damage)
3131 damage = 1;
3132
3133 found_damage:
3134 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3135 dir.files[dir.nr].name = name;
3136 dir.files[dir.nr].changed = damage;
3137 changed += damage;
3138 dir.nr++;
3139 }
3140
3141 conclude_dirstat(options, &dir, changed);
3142 }
3143
3144 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3145 {
3146 int i;
3147 unsigned long changed;
3148 struct dirstat_dir dir;
3149
3150 if (data->nr == 0)
3151 return;
3152
3153 dir.files = NULL;
3154 dir.alloc = 0;
3155 dir.nr = 0;
3156 dir.permille = options->dirstat_permille;
3157 dir.cumulative = options->flags.dirstat_cumulative;
3158
3159 changed = 0;
3160 for (i = 0; i < data->nr; i++) {
3161 struct diffstat_file *file = data->files[i];
3162 unsigned long damage = file->added + file->deleted;
3163 if (file->is_binary)
3164 /*
3165 * binary files counts bytes, not lines. Must find some
3166 * way to normalize binary bytes vs. textual lines.
3167 * The following heuristic assumes that there are 64
3168 * bytes per "line".
3169 * This is stupid and ugly, but very cheap...
3170 */
3171 damage = DIV_ROUND_UP(damage, 64);
3172 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3173 dir.files[dir.nr].name = file->name;
3174 dir.files[dir.nr].changed = damage;
3175 changed += damage;
3176 dir.nr++;
3177 }
3178
3179 conclude_dirstat(options, &dir, changed);
3180 }
3181
3182 static void free_diffstat_file(struct diffstat_file *f)
3183 {
3184 free(f->print_name);
3185 free(f->name);
3186 free(f->from_name);
3187 free(f);
3188 }
3189
3190 void free_diffstat_info(struct diffstat_t *diffstat)
3191 {
3192 int i;
3193 for (i = 0; i < diffstat->nr; i++)
3194 free_diffstat_file(diffstat->files[i]);
3195 free(diffstat->files);
3196 }
3197
3198 struct checkdiff_t {
3199 const char *filename;
3200 int lineno;
3201 int conflict_marker_size;
3202 struct diff_options *o;
3203 unsigned ws_rule;
3204 unsigned status;
3205 };
3206
3207 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3208 {
3209 char firstchar;
3210 int cnt;
3211
3212 if (len < marker_size + 1)
3213 return 0;
3214 firstchar = line[0];
3215 switch (firstchar) {
3216 case '=': case '>': case '<': case '|':
3217 break;
3218 default:
3219 return 0;
3220 }
3221 for (cnt = 1; cnt < marker_size; cnt++)
3222 if (line[cnt] != firstchar)
3223 return 0;
3224 /* line[1] through line[marker_size-1] are same as firstchar */
3225 if (len < marker_size + 1 || !isspace(line[marker_size]))
3226 return 0;
3227 return 1;
3228 }
3229
3230 static void checkdiff_consume_hunk(void *priv,
3231 long ob UNUSED, long on UNUSED,
3232 long nb, long nn UNUSED,
3233 const char *func UNUSED, long funclen UNUSED)
3234
3235 {
3236 struct checkdiff_t *data = priv;
3237 data->lineno = nb - 1;
3238 }
3239
3240 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3241 {
3242 struct checkdiff_t *data = priv;
3243 int marker_size = data->conflict_marker_size;
3244 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3245 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3246 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3247 char *err;
3248 const char *line_prefix;
3249
3250 assert(data->o);
3251 line_prefix = diff_line_prefix(data->o);
3252
3253 if (line[0] == '+') {
3254 unsigned bad;
3255 data->lineno++;
3256 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3257 data->status |= 1;
3258 fprintf(data->o->file,
3259 "%s%s:%d: leftover conflict marker\n",
3260 line_prefix, data->filename, data->lineno);
3261 }
3262 bad = ws_check(line + 1, len - 1, data->ws_rule);
3263 if (!bad)
3264 return 0;
3265 data->status |= bad;
3266 err = whitespace_error_string(bad);
3267 fprintf(data->o->file, "%s%s:%d: %s.\n",
3268 line_prefix, data->filename, data->lineno, err);
3269 free(err);
3270 emit_line(data->o, set, reset, line, 1);
3271 ws_check_emit(line + 1, len - 1, data->ws_rule,
3272 data->o->file, set, reset, ws);
3273 } else if (line[0] == ' ') {
3274 data->lineno++;
3275 }
3276 return 0;
3277 }
3278
3279 static unsigned char *deflate_it(char *data,
3280 unsigned long size,
3281 unsigned long *result_size)
3282 {
3283 int bound;
3284 unsigned char *deflated;
3285 git_zstream stream;
3286
3287 git_deflate_init(&stream, zlib_compression_level);
3288 bound = git_deflate_bound(&stream, size);
3289 deflated = xmalloc(bound);
3290 stream.next_out = deflated;
3291 stream.avail_out = bound;
3292
3293 stream.next_in = (unsigned char *)data;
3294 stream.avail_in = size;
3295 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3296 ; /* nothing */
3297 git_deflate_end(&stream);
3298 *result_size = stream.total_out;
3299 return deflated;
3300 }
3301
3302 static void emit_binary_diff_body(struct diff_options *o,
3303 mmfile_t *one, mmfile_t *two)
3304 {
3305 void *cp;
3306 void *delta;
3307 void *deflated;
3308 void *data;
3309 unsigned long orig_size;
3310 unsigned long delta_size;
3311 unsigned long deflate_size;
3312 unsigned long data_size;
3313
3314 /* We could do deflated delta, or we could do just deflated two,
3315 * whichever is smaller.
3316 */
3317 delta = NULL;
3318 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3319 if (one->size && two->size) {
3320 delta = diff_delta(one->ptr, one->size,
3321 two->ptr, two->size,
3322 &delta_size, deflate_size);
3323 if (delta) {
3324 void *to_free = delta;
3325 orig_size = delta_size;
3326 delta = deflate_it(delta, delta_size, &delta_size);
3327 free(to_free);
3328 }
3329 }
3330
3331 if (delta && delta_size < deflate_size) {
3332 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3333 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3334 s, strlen(s), 0);
3335 free(s);
3336 free(deflated);
3337 data = delta;
3338 data_size = delta_size;
3339 } else {
3340 char *s = xstrfmt("%lu", two->size);
3341 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3342 s, strlen(s), 0);
3343 free(s);
3344 free(delta);
3345 data = deflated;
3346 data_size = deflate_size;
3347 }
3348
3349 /* emit data encoded in base85 */
3350 cp = data;
3351 while (data_size) {
3352 int len;
3353 int bytes = (52 < data_size) ? 52 : data_size;
3354 char line[71];
3355 data_size -= bytes;
3356 if (bytes <= 26)
3357 line[0] = bytes + 'A' - 1;
3358 else
3359 line[0] = bytes - 26 + 'a' - 1;
3360 encode_85(line + 1, cp, bytes);
3361 cp = (char *) cp + bytes;
3362
3363 len = strlen(line);
3364 line[len++] = '\n';
3365 line[len] = '\0';
3366
3367 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3368 line, len, 0);
3369 }
3370 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3371 free(data);
3372 }
3373
3374 static void emit_binary_diff(struct diff_options *o,
3375 mmfile_t *one, mmfile_t *two)
3376 {
3377 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3378 emit_binary_diff_body(o, one, two);
3379 emit_binary_diff_body(o, two, one);
3380 }
3381
3382 int diff_filespec_is_binary(struct repository *r,
3383 struct diff_filespec *one)
3384 {
3385 struct diff_populate_filespec_options dpf_options = {
3386 .check_binary = 1,
3387 };
3388
3389 if (one->is_binary == -1) {
3390 diff_filespec_load_driver(one, r->index);
3391 if (one->driver->binary != -1)
3392 one->is_binary = one->driver->binary;
3393 else {
3394 if (!one->data && DIFF_FILE_VALID(one))
3395 diff_populate_filespec(r, one, &dpf_options);
3396 if (one->is_binary == -1 && one->data)
3397 one->is_binary = buffer_is_binary(one->data,
3398 one->size);
3399 if (one->is_binary == -1)
3400 one->is_binary = 0;
3401 }
3402 }
3403 return one->is_binary;
3404 }
3405
3406 static const struct userdiff_funcname *
3407 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3408 {
3409 diff_filespec_load_driver(one, o->repo->index);
3410 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3411 }
3412
3413 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3414 {
3415 if (!options->a_prefix)
3416 options->a_prefix = a;
3417 if (!options->b_prefix)
3418 options->b_prefix = b;
3419 }
3420
3421 void diff_set_noprefix(struct diff_options *options)
3422 {
3423 options->a_prefix = options->b_prefix = "";
3424 }
3425
3426 void diff_set_default_prefix(struct diff_options *options)
3427 {
3428 options->a_prefix = "a/";
3429 options->b_prefix = "b/";
3430 }
3431
3432 struct userdiff_driver *get_textconv(struct repository *r,
3433 struct diff_filespec *one)
3434 {
3435 if (!DIFF_FILE_VALID(one))
3436 return NULL;
3437
3438 diff_filespec_load_driver(one, r->index);
3439 return userdiff_get_textconv(r, one->driver);
3440 }
3441
3442 static struct string_list *additional_headers(struct diff_options *o,
3443 const char *path)
3444 {
3445 if (!o->additional_path_headers)
3446 return NULL;
3447 return strmap_get(o->additional_path_headers, path);
3448 }
3449
3450 static void add_formatted_header(struct strbuf *msg,
3451 const char *header,
3452 const char *line_prefix,
3453 const char *meta,
3454 const char *reset)
3455 {
3456 const char *next, *newline;
3457
3458 for (next = header; *next; next = newline) {
3459 newline = strchrnul(next, '\n');
3460 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3461 (int)(newline - next), next, reset);
3462 if (*newline)
3463 newline++;
3464 }
3465 }
3466
3467 static void add_formatted_headers(struct strbuf *msg,
3468 struct string_list *more_headers,
3469 const char *line_prefix,
3470 const char *meta,
3471 const char *reset)
3472 {
3473 int i;
3474
3475 for (i = 0; i < more_headers->nr; i++)
3476 add_formatted_header(msg, more_headers->items[i].string,
3477 line_prefix, meta, reset);
3478 }
3479
3480 static int diff_filepair_is_phoney(struct diff_filespec *one,
3481 struct diff_filespec *two)
3482 {
3483 /*
3484 * This function specifically looks for pairs injected by
3485 * create_filepairs_for_header_only_notifications(). Such
3486 * pairs are "phoney" in that they do not represent any
3487 * content or even mode difference, but were inserted because
3488 * diff_queued_diff previously had no pair associated with
3489 * that path but we needed some pair to avoid losing the
3490 * "remerge CONFLICT" header associated with the path.
3491 */
3492 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3493 }
3494
3495 static int set_diff_algorithm(struct diff_options *opts,
3496 const char *alg)
3497 {
3498 long value = parse_algorithm_value(alg);
3499
3500 if (value < 0)
3501 return -1;
3502
3503 /* clear out previous settings */
3504 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3505 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3506 opts->xdl_opts |= value;
3507
3508 return 0;
3509 }
3510
3511 static void builtin_diff(const char *name_a,
3512 const char *name_b,
3513 struct diff_filespec *one,
3514 struct diff_filespec *two,
3515 const char *xfrm_msg,
3516 int must_show_header,
3517 struct diff_options *o,
3518 int complete_rewrite)
3519 {
3520 mmfile_t mf1, mf2;
3521 const char *lbl[2];
3522 char *a_one, *b_two;
3523 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3524 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3525 const char *a_prefix, *b_prefix;
3526 struct userdiff_driver *textconv_one = NULL;
3527 struct userdiff_driver *textconv_two = NULL;
3528 struct strbuf header = STRBUF_INIT;
3529 const char *line_prefix = diff_line_prefix(o);
3530
3531 diff_set_mnemonic_prefix(o, "a/", "b/");
3532 if (o->flags.reverse_diff) {
3533 a_prefix = o->b_prefix;
3534 b_prefix = o->a_prefix;
3535 } else {
3536 a_prefix = o->a_prefix;
3537 b_prefix = o->b_prefix;
3538 }
3539
3540 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3541 (!one->mode || S_ISGITLINK(one->mode)) &&
3542 (!two->mode || S_ISGITLINK(two->mode)) &&
3543 (!diff_filepair_is_phoney(one, two))) {
3544 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3545 &one->oid, &two->oid,
3546 two->dirty_submodule);
3547 return;
3548 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3549 (!one->mode || S_ISGITLINK(one->mode)) &&
3550 (!two->mode || S_ISGITLINK(two->mode)) &&
3551 (!diff_filepair_is_phoney(one, two))) {
3552 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3553 &one->oid, &two->oid,
3554 two->dirty_submodule);
3555 return;
3556 }
3557
3558 if (o->flags.allow_textconv) {
3559 textconv_one = get_textconv(o->repo, one);
3560 textconv_two = get_textconv(o->repo, two);
3561 }
3562
3563 /* Never use a non-valid filename anywhere if at all possible */
3564 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3565 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3566
3567 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3568 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3569 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3570 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3571 if (diff_filepair_is_phoney(one, two)) {
3572 /*
3573 * We should only reach this point for pairs generated from
3574 * create_filepairs_for_header_only_notifications(). For
3575 * these, we want to avoid the "/dev/null" special casing
3576 * above, because we do not want such pairs shown as either
3577 * "new file" or "deleted file" below.
3578 */
3579 lbl[0] = a_one;
3580 lbl[1] = b_two;
3581 }
3582 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3583 if (lbl[0][0] == '/') {
3584 /* /dev/null */
3585 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3586 if (xfrm_msg)
3587 strbuf_addstr(&header, xfrm_msg);
3588 o->found_changes = 1;
3589 must_show_header = 1;
3590 }
3591 else if (lbl[1][0] == '/') {
3592 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3593 if (xfrm_msg)
3594 strbuf_addstr(&header, xfrm_msg);
3595 o->found_changes = 1;
3596 must_show_header = 1;
3597 }
3598 else {
3599 if (one->mode != two->mode) {
3600 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3601 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3602 o->found_changes = 1;
3603 must_show_header = 1;
3604 }
3605 if (xfrm_msg)
3606 strbuf_addstr(&header, xfrm_msg);
3607
3608 /*
3609 * we do not run diff between different kind
3610 * of objects.
3611 */
3612 if ((one->mode ^ two->mode) & S_IFMT)
3613 goto free_ab_and_return;
3614 if (complete_rewrite &&
3615 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3616 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3617 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3618 header.buf, header.len, 0);
3619 strbuf_reset(&header);
3620 emit_rewrite_diff(name_a, name_b, one, two,
3621 textconv_one, textconv_two, o);
3622 o->found_changes = 1;
3623 goto free_ab_and_return;
3624 }
3625 }
3626
3627 if (o->irreversible_delete && lbl[1][0] == '/') {
3628 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3629 header.len, 0);
3630 strbuf_reset(&header);
3631 goto free_ab_and_return;
3632 } else if (!o->flags.text &&
3633 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3634 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3635 struct strbuf sb = STRBUF_INIT;
3636 if (!one->data && !two->data &&
3637 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3638 !o->flags.binary) {
3639 if (oideq(&one->oid, &two->oid)) {
3640 if (must_show_header)
3641 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3642 header.buf, header.len,
3643 0);
3644 goto free_ab_and_return;
3645 }
3646 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3647 header.buf, header.len, 0);
3648 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3649 diff_line_prefix(o), lbl[0], lbl[1]);
3650 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3651 sb.buf, sb.len, 0);
3652 strbuf_release(&sb);
3653 goto free_ab_and_return;
3654 }
3655 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3656 fill_mmfile(o->repo, &mf2, two) < 0)
3657 die("unable to read files to diff");
3658 /* Quite common confusing case */
3659 if (mf1.size == mf2.size &&
3660 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3661 if (must_show_header)
3662 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3663 header.buf, header.len, 0);
3664 goto free_ab_and_return;
3665 }
3666 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3667 strbuf_reset(&header);
3668 if (o->flags.binary)
3669 emit_binary_diff(o, &mf1, &mf2);
3670 else {
3671 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3672 diff_line_prefix(o), lbl[0], lbl[1]);
3673 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3674 sb.buf, sb.len, 0);
3675 strbuf_release(&sb);
3676 }
3677 o->found_changes = 1;
3678 } else {
3679 /* Crazy xdl interfaces.. */
3680 const char *diffopts;
3681 const char *v;
3682 xpparam_t xpp;
3683 xdemitconf_t xecfg;
3684 struct emit_callback ecbdata;
3685 const struct userdiff_funcname *pe;
3686
3687 if (must_show_header) {
3688 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3689 header.buf, header.len, 0);
3690 strbuf_reset(&header);
3691 }
3692
3693 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3694 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3695
3696 pe = diff_funcname_pattern(o, one);
3697 if (!pe)
3698 pe = diff_funcname_pattern(o, two);
3699
3700 memset(&xpp, 0, sizeof(xpp));
3701 memset(&xecfg, 0, sizeof(xecfg));
3702 memset(&ecbdata, 0, sizeof(ecbdata));
3703 if (o->flags.suppress_diff_headers)
3704 lbl[0] = NULL;
3705 ecbdata.label_path = lbl;
3706 ecbdata.color_diff = want_color(o->use_color);
3707 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3708 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3709 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3710 ecbdata.opt = o;
3711 if (header.len && !o->flags.suppress_diff_headers)
3712 ecbdata.header = &header;
3713 xpp.flags = o->xdl_opts;
3714 xpp.ignore_regex = o->ignore_regex;
3715 xpp.ignore_regex_nr = o->ignore_regex_nr;
3716 xpp.anchors = o->anchors;
3717 xpp.anchors_nr = o->anchors_nr;
3718 xecfg.ctxlen = o->context;
3719 xecfg.interhunkctxlen = o->interhunkcontext;
3720 xecfg.flags = XDL_EMIT_FUNCNAMES;
3721 if (o->flags.funccontext)
3722 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3723 if (pe)
3724 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3725
3726 diffopts = getenv("GIT_DIFF_OPTS");
3727 if (!diffopts)
3728 ;
3729 else if (skip_prefix(diffopts, "--unified=", &v))
3730 xecfg.ctxlen = strtoul(v, NULL, 10);
3731 else if (skip_prefix(diffopts, "-u", &v))
3732 xecfg.ctxlen = strtoul(v, NULL, 10);
3733
3734 if (o->word_diff)
3735 init_diff_words_data(&ecbdata, o, one, two);
3736 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3737 &ecbdata, &xpp, &xecfg))
3738 die("unable to generate diff for %s", one->path);
3739 if (o->word_diff)
3740 free_diff_words_data(&ecbdata);
3741 if (textconv_one)
3742 free(mf1.ptr);
3743 if (textconv_two)
3744 free(mf2.ptr);
3745 xdiff_clear_find_func(&xecfg);
3746 }
3747
3748 free_ab_and_return:
3749 strbuf_release(&header);
3750 diff_free_filespec_data(one);
3751 diff_free_filespec_data(two);
3752 free(a_one);
3753 free(b_two);
3754 return;
3755 }
3756
3757 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3758 {
3759 if (!is_renamed) {
3760 if (p->status == DIFF_STATUS_ADDED) {
3761 if (S_ISLNK(p->two->mode))
3762 return "new +l";
3763 else if ((p->two->mode & 0777) == 0755)
3764 return "new +x";
3765 else
3766 return "new";
3767 } else if (p->status == DIFF_STATUS_DELETED)
3768 return "gone";
3769 }
3770 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3771 return "mode -l";
3772 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3773 return "mode +l";
3774 else if ((p->one->mode & 0777) == 0644 &&
3775 (p->two->mode & 0777) == 0755)
3776 return "mode +x";
3777 else if ((p->one->mode & 0777) == 0755 &&
3778 (p->two->mode & 0777) == 0644)
3779 return "mode -x";
3780 return NULL;
3781 }
3782
3783 static void builtin_diffstat(const char *name_a, const char *name_b,
3784 struct diff_filespec *one,
3785 struct diff_filespec *two,
3786 struct diffstat_t *diffstat,
3787 struct diff_options *o,
3788 struct diff_filepair *p)
3789 {
3790 mmfile_t mf1, mf2;
3791 struct diffstat_file *data;
3792 int may_differ;
3793 int complete_rewrite = 0;
3794
3795 if (!DIFF_PAIR_UNMERGED(p)) {
3796 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3797 complete_rewrite = 1;
3798 }
3799
3800 data = diffstat_add(diffstat, name_a, name_b);
3801 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3802 if (o->flags.stat_with_summary)
3803 data->comments = get_compact_summary(p, data->is_renamed);
3804
3805 if (!one || !two) {
3806 data->is_unmerged = 1;
3807 return;
3808 }
3809
3810 /* saves some reads if true, not a guarantee of diff outcome */
3811 may_differ = !(one->oid_valid && two->oid_valid &&
3812 oideq(&one->oid, &two->oid));
3813
3814 if (diff_filespec_is_binary(o->repo, one) ||
3815 diff_filespec_is_binary(o->repo, two)) {
3816 data->is_binary = 1;
3817 if (!may_differ) {
3818 data->added = 0;
3819 data->deleted = 0;
3820 } else {
3821 data->added = diff_filespec_size(o->repo, two);
3822 data->deleted = diff_filespec_size(o->repo, one);
3823 }
3824 }
3825
3826 else if (complete_rewrite) {
3827 diff_populate_filespec(o->repo, one, NULL);
3828 diff_populate_filespec(o->repo, two, NULL);
3829 data->deleted = count_lines(one->data, one->size);
3830 data->added = count_lines(two->data, two->size);
3831 }
3832
3833 else if (may_differ) {
3834 /* Crazy xdl interfaces.. */
3835 xpparam_t xpp;
3836 xdemitconf_t xecfg;
3837
3838 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3839 fill_mmfile(o->repo, &mf2, two) < 0)
3840 die("unable to read files to diff");
3841
3842 memset(&xpp, 0, sizeof(xpp));
3843 memset(&xecfg, 0, sizeof(xecfg));
3844 xpp.flags = o->xdl_opts;
3845 xpp.ignore_regex = o->ignore_regex;
3846 xpp.ignore_regex_nr = o->ignore_regex_nr;
3847 xpp.anchors = o->anchors;
3848 xpp.anchors_nr = o->anchors_nr;
3849 xecfg.ctxlen = o->context;
3850 xecfg.interhunkctxlen = o->interhunkcontext;
3851 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3852 if (xdi_diff_outf(&mf1, &mf2, NULL,
3853 diffstat_consume, diffstat, &xpp, &xecfg))
3854 die("unable to generate diffstat for %s", one->path);
3855
3856 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3857 struct diffstat_file *file =
3858 diffstat->files[diffstat->nr - 1];
3859 /*
3860 * Omit diffstats of modified files where nothing changed.
3861 * Even if may_differ, this might be the case due to
3862 * ignoring whitespace changes, etc.
3863 *
3864 * But note that we special-case additions, deletions,
3865 * renames, and mode changes as adding an empty file,
3866 * for example is still of interest.
3867 */
3868 if ((p->status == DIFF_STATUS_MODIFIED)
3869 && !file->added
3870 && !file->deleted
3871 && one->mode == two->mode) {
3872 free_diffstat_file(file);
3873 diffstat->nr--;
3874 }
3875 }
3876 }
3877
3878 diff_free_filespec_data(one);
3879 diff_free_filespec_data(two);
3880 }
3881
3882 static void builtin_checkdiff(const char *name_a, const char *name_b,
3883 const char *attr_path,
3884 struct diff_filespec *one,
3885 struct diff_filespec *two,
3886 struct diff_options *o)
3887 {
3888 mmfile_t mf1, mf2;
3889 struct checkdiff_t data;
3890
3891 if (!two)
3892 return;
3893
3894 memset(&data, 0, sizeof(data));
3895 data.filename = name_b ? name_b : name_a;
3896 data.lineno = 0;
3897 data.o = o;
3898 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3899 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3900
3901 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3902 fill_mmfile(o->repo, &mf2, two) < 0)
3903 die("unable to read files to diff");
3904
3905 /*
3906 * All the other codepaths check both sides, but not checking
3907 * the "old" side here is deliberate. We are checking the newly
3908 * introduced changes, and as long as the "new" side is text, we
3909 * can and should check what it introduces.
3910 */
3911 if (diff_filespec_is_binary(o->repo, two))
3912 goto free_and_return;
3913 else {
3914 /* Crazy xdl interfaces.. */
3915 xpparam_t xpp;
3916 xdemitconf_t xecfg;
3917
3918 memset(&xpp, 0, sizeof(xpp));
3919 memset(&xecfg, 0, sizeof(xecfg));
3920 xecfg.ctxlen = 1; /* at least one context line */
3921 xpp.flags = 0;
3922 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3923 checkdiff_consume, &data,
3924 &xpp, &xecfg))
3925 die("unable to generate checkdiff for %s", one->path);
3926
3927 if (data.ws_rule & WS_BLANK_AT_EOF) {
3928 struct emit_callback ecbdata;
3929 int blank_at_eof;
3930
3931 ecbdata.ws_rule = data.ws_rule;
3932 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3933 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3934
3935 if (blank_at_eof) {
3936 static char *err;
3937 if (!err)
3938 err = whitespace_error_string(WS_BLANK_AT_EOF);
3939 fprintf(o->file, "%s:%d: %s.\n",
3940 data.filename, blank_at_eof, err);
3941 data.status = 1; /* report errors */
3942 }
3943 }
3944 }
3945 free_and_return:
3946 diff_free_filespec_data(one);
3947 diff_free_filespec_data(two);
3948 if (data.status)
3949 o->flags.check_failed = 1;
3950 }
3951
3952 struct diff_filespec *alloc_filespec(const char *path)
3953 {
3954 struct diff_filespec *spec;
3955
3956 FLEXPTR_ALLOC_STR(spec, path, path);
3957 spec->count = 1;
3958 spec->is_binary = -1;
3959 return spec;
3960 }
3961
3962 void free_filespec(struct diff_filespec *spec)
3963 {
3964 if (!--spec->count) {
3965 diff_free_filespec_data(spec);
3966 free(spec);
3967 }
3968 }
3969
3970 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3971 int oid_valid, unsigned short mode)
3972 {
3973 if (mode) {
3974 spec->mode = canon_mode(mode);
3975 oidcpy(&spec->oid, oid);
3976 spec->oid_valid = oid_valid;
3977 }
3978 }
3979
3980 /*
3981 * Given a name and sha1 pair, if the index tells us the file in
3982 * the work tree has that object contents, return true, so that
3983 * prepare_temp_file() does not have to inflate and extract.
3984 */
3985 static int reuse_worktree_file(struct index_state *istate,
3986 const char *name,
3987 const struct object_id *oid,
3988 int want_file)
3989 {
3990 const struct cache_entry *ce;
3991 struct stat st;
3992 int pos, len;
3993
3994 /*
3995 * We do not read the cache ourselves here, because the
3996 * benchmark with my previous version that always reads cache
3997 * shows that it makes things worse for diff-tree comparing
3998 * two linux-2.6 kernel trees in an already checked out work
3999 * tree. This is because most diff-tree comparisons deal with
4000 * only a small number of files, while reading the cache is
4001 * expensive for a large project, and its cost outweighs the
4002 * savings we get by not inflating the object to a temporary
4003 * file. Practically, this code only helps when we are used
4004 * by diff-cache --cached, which does read the cache before
4005 * calling us.
4006 */
4007 if (!istate->cache)
4008 return 0;
4009
4010 /* We want to avoid the working directory if our caller
4011 * doesn't need the data in a normal file, this system
4012 * is rather slow with its stat/open/mmap/close syscalls,
4013 * and the object is contained in a pack file. The pack
4014 * is probably already open and will be faster to obtain
4015 * the data through than the working directory. Loose
4016 * objects however would tend to be slower as they need
4017 * to be individually opened and inflated.
4018 */
4019 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
4020 return 0;
4021
4022 /*
4023 * Similarly, if we'd have to convert the file contents anyway, that
4024 * makes the optimization not worthwhile.
4025 */
4026 if (!want_file && would_convert_to_git(istate, name))
4027 return 0;
4028
4029 /*
4030 * If this path does not match our sparse-checkout definition,
4031 * then the file will not be in the working directory.
4032 */
4033 if (!path_in_sparse_checkout(name, istate))
4034 return 0;
4035
4036 len = strlen(name);
4037 pos = index_name_pos(istate, name, len);
4038 if (pos < 0)
4039 return 0;
4040 ce = istate->cache[pos];
4041
4042 /*
4043 * This is not the sha1 we are looking for, or
4044 * unreusable because it is not a regular file.
4045 */
4046 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4047 return 0;
4048
4049 /*
4050 * If ce is marked as "assume unchanged", there is no
4051 * guarantee that work tree matches what we are looking for.
4052 */
4053 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4054 return 0;
4055
4056 /*
4057 * If ce matches the file in the work tree, we can reuse it.
4058 */
4059 if (ce_uptodate(ce) ||
4060 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4061 return 1;
4062
4063 return 0;
4064 }
4065
4066 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4067 {
4068 struct strbuf buf = STRBUF_INIT;
4069 char *dirty = "";
4070
4071 /* Are we looking at the work tree? */
4072 if (s->dirty_submodule)
4073 dirty = "-dirty";
4074
4075 strbuf_addf(&buf, "Subproject commit %s%s\n",
4076 oid_to_hex(&s->oid), dirty);
4077 s->size = buf.len;
4078 if (size_only) {
4079 s->data = NULL;
4080 strbuf_release(&buf);
4081 } else {
4082 s->data = strbuf_detach(&buf, NULL);
4083 s->should_free = 1;
4084 }
4085 return 0;
4086 }
4087
4088 /*
4089 * While doing rename detection and pickaxe operation, we may need to
4090 * grab the data for the blob (or file) for our own in-core comparison.
4091 * diff_filespec has data and size fields for this purpose.
4092 */
4093 int diff_populate_filespec(struct repository *r,
4094 struct diff_filespec *s,
4095 const struct diff_populate_filespec_options *options)
4096 {
4097 int size_only = options ? options->check_size_only : 0;
4098 int check_binary = options ? options->check_binary : 0;
4099 int err = 0;
4100 int conv_flags = global_conv_flags_eol;
4101 /*
4102 * demote FAIL to WARN to allow inspecting the situation
4103 * instead of refusing.
4104 */
4105 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4106 conv_flags = CONV_EOL_RNDTRP_WARN;
4107
4108 if (!DIFF_FILE_VALID(s))
4109 die("internal error: asking to populate invalid file.");
4110 if (S_ISDIR(s->mode))
4111 return -1;
4112
4113 if (s->data)
4114 return 0;
4115
4116 if (size_only && 0 < s->size)
4117 return 0;
4118
4119 if (S_ISGITLINK(s->mode))
4120 return diff_populate_gitlink(s, size_only);
4121
4122 if (!s->oid_valid ||
4123 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4124 struct strbuf buf = STRBUF_INIT;
4125 struct stat st;
4126 int fd;
4127
4128 if (lstat(s->path, &st) < 0) {
4129 err_empty:
4130 err = -1;
4131 empty:
4132 s->data = (char *)"";
4133 s->size = 0;
4134 return err;
4135 }
4136 s->size = xsize_t(st.st_size);
4137 if (!s->size)
4138 goto empty;
4139 if (S_ISLNK(st.st_mode)) {
4140 struct strbuf sb = STRBUF_INIT;
4141
4142 if (strbuf_readlink(&sb, s->path, s->size))
4143 goto err_empty;
4144 s->size = sb.len;
4145 s->data = strbuf_detach(&sb, NULL);
4146 s->should_free = 1;
4147 return 0;
4148 }
4149
4150 /*
4151 * Even if the caller would be happy with getting
4152 * only the size, we cannot return early at this
4153 * point if the path requires us to run the content
4154 * conversion.
4155 */
4156 if (size_only && !would_convert_to_git(r->index, s->path))
4157 return 0;
4158
4159 /*
4160 * Note: this check uses xsize_t(st.st_size) that may
4161 * not be the true size of the blob after it goes
4162 * through convert_to_git(). This may not strictly be
4163 * correct, but the whole point of big_file_threshold
4164 * and is_binary check being that we want to avoid
4165 * opening the file and inspecting the contents, this
4166 * is probably fine.
4167 */
4168 if (check_binary &&
4169 s->size > big_file_threshold && s->is_binary == -1) {
4170 s->is_binary = 1;
4171 return 0;
4172 }
4173 fd = open(s->path, O_RDONLY);
4174 if (fd < 0)
4175 goto err_empty;
4176 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4177 close(fd);
4178 s->should_munmap = 1;
4179
4180 /*
4181 * Convert from working tree format to canonical git format
4182 */
4183 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4184 size_t size = 0;
4185 munmap(s->data, s->size);
4186 s->should_munmap = 0;
4187 s->data = strbuf_detach(&buf, &size);
4188 s->size = size;
4189 s->should_free = 1;
4190 }
4191 }
4192 else {
4193 struct object_info info = {
4194 .sizep = &s->size
4195 };
4196
4197 if (!(size_only || check_binary))
4198 /*
4199 * Set contentp, since there is no chance that merely
4200 * the size is sufficient.
4201 */
4202 info.contentp = &s->data;
4203
4204 if (options && options->missing_object_cb) {
4205 if (!oid_object_info_extended(r, &s->oid, &info,
4206 OBJECT_INFO_LOOKUP_REPLACE |
4207 OBJECT_INFO_SKIP_FETCH_OBJECT))
4208 goto object_read;
4209 options->missing_object_cb(options->missing_object_data);
4210 }
4211 if (oid_object_info_extended(r, &s->oid, &info,
4212 OBJECT_INFO_LOOKUP_REPLACE))
4213 die("unable to read %s", oid_to_hex(&s->oid));
4214
4215 object_read:
4216 if (size_only || check_binary) {
4217 if (size_only)
4218 return 0;
4219 if (s->size > big_file_threshold && s->is_binary == -1) {
4220 s->is_binary = 1;
4221 return 0;
4222 }
4223 }
4224 if (!info.contentp) {
4225 info.contentp = &s->data;
4226 if (oid_object_info_extended(r, &s->oid, &info,
4227 OBJECT_INFO_LOOKUP_REPLACE))
4228 die("unable to read %s", oid_to_hex(&s->oid));
4229 }
4230 s->should_free = 1;
4231 }
4232 return 0;
4233 }
4234
4235 void diff_free_filespec_blob(struct diff_filespec *s)
4236 {
4237 if (s->should_free)
4238 free(s->data);
4239 else if (s->should_munmap)
4240 munmap(s->data, s->size);
4241
4242 if (s->should_free || s->should_munmap) {
4243 s->should_free = s->should_munmap = 0;
4244 s->data = NULL;
4245 }
4246 }
4247
4248 void diff_free_filespec_data(struct diff_filespec *s)
4249 {
4250 if (!s)
4251 return;
4252
4253 diff_free_filespec_blob(s);
4254 FREE_AND_NULL(s->cnt_data);
4255 }
4256
4257 static void prep_temp_blob(struct index_state *istate,
4258 const char *path, struct diff_tempfile *temp,
4259 void *blob,
4260 unsigned long size,
4261 const struct object_id *oid,
4262 int mode)
4263 {
4264 struct strbuf buf = STRBUF_INIT;
4265 char *path_dup = xstrdup(path);
4266 const char *base = basename(path_dup);
4267 struct checkout_metadata meta;
4268
4269 init_checkout_metadata(&meta, NULL, NULL, oid);
4270
4271 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4272 if (!temp->tempfile)
4273 die_errno("unable to create temp-file");
4274 if (convert_to_working_tree(istate, path,
4275 (const char *)blob, (size_t)size, &buf, &meta)) {
4276 blob = buf.buf;
4277 size = buf.len;
4278 }
4279 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4280 close_tempfile_gently(temp->tempfile))
4281 die_errno("unable to write temp-file");
4282 temp->name = get_tempfile_path(temp->tempfile);
4283 oid_to_hex_r(temp->hex, oid);
4284 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4285 strbuf_release(&buf);
4286 free(path_dup);
4287 }
4288
4289 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4290 struct diff_filespec *one)
4291 {
4292 struct diff_tempfile *temp = claim_diff_tempfile();
4293
4294 if (!DIFF_FILE_VALID(one)) {
4295 not_a_valid_file:
4296 /* A '-' entry produces this for file-2, and
4297 * a '+' entry produces this for file-1.
4298 */
4299 temp->name = "/dev/null";
4300 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4301 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4302 return temp;
4303 }
4304
4305 if (!S_ISGITLINK(one->mode) &&
4306 (!one->oid_valid ||
4307 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4308 struct stat st;
4309 if (lstat(one->path, &st) < 0) {
4310 if (errno == ENOENT)
4311 goto not_a_valid_file;
4312 die_errno("stat(%s)", one->path);
4313 }
4314 if (S_ISLNK(st.st_mode)) {
4315 struct strbuf sb = STRBUF_INIT;
4316 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4317 die_errno("readlink(%s)", one->path);
4318 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4319 (one->oid_valid ?
4320 &one->oid : null_oid()),
4321 (one->oid_valid ?
4322 one->mode : S_IFLNK));
4323 strbuf_release(&sb);
4324 }
4325 else {
4326 /* we can borrow from the file in the work tree */
4327 temp->name = one->path;
4328 if (!one->oid_valid)
4329 oid_to_hex_r(temp->hex, null_oid());
4330 else
4331 oid_to_hex_r(temp->hex, &one->oid);
4332 /* Even though we may sometimes borrow the
4333 * contents from the work tree, we always want
4334 * one->mode. mode is trustworthy even when
4335 * !(one->oid_valid), as long as
4336 * DIFF_FILE_VALID(one).
4337 */
4338 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4339 }
4340 return temp;
4341 }
4342 else {
4343 if (diff_populate_filespec(r, one, NULL))
4344 die("cannot read data blob for %s", one->path);
4345 prep_temp_blob(r->index, one->path, temp,
4346 one->data, one->size,
4347 &one->oid, one->mode);
4348 }
4349 return temp;
4350 }
4351
4352 static void add_external_diff_name(struct repository *r,
4353 struct strvec *argv,
4354 struct diff_filespec *df)
4355 {
4356 struct diff_tempfile *temp = prepare_temp_file(r, df);
4357 strvec_push(argv, temp->name);
4358 strvec_push(argv, temp->hex);
4359 strvec_push(argv, temp->mode);
4360 }
4361
4362 /* An external diff command takes:
4363 *
4364 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4365 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4366 *
4367 */
4368 static void run_external_diff(const char *pgm,
4369 const char *name,
4370 const char *other,
4371 struct diff_filespec *one,
4372 struct diff_filespec *two,
4373 const char *xfrm_msg,
4374 struct diff_options *o)
4375 {
4376 struct child_process cmd = CHILD_PROCESS_INIT;
4377 struct diff_queue_struct *q = &diff_queued_diff;
4378
4379 strvec_push(&cmd.args, pgm);
4380 strvec_push(&cmd.args, name);
4381
4382 if (one && two) {
4383 add_external_diff_name(o->repo, &cmd.args, one);
4384 add_external_diff_name(o->repo, &cmd.args, two);
4385 if (other) {
4386 strvec_push(&cmd.args, other);
4387 if (xfrm_msg)
4388 strvec_push(&cmd.args, xfrm_msg);
4389 }
4390 }
4391
4392 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4393 ++o->diff_path_counter);
4394 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4395
4396 diff_free_filespec_data(one);
4397 diff_free_filespec_data(two);
4398 cmd.use_shell = 1;
4399 if (run_command(&cmd))
4400 die(_("external diff died, stopping at %s"), name);
4401
4402 remove_tempfile();
4403 }
4404
4405 static int similarity_index(struct diff_filepair *p)
4406 {
4407 return p->score * 100 / MAX_SCORE;
4408 }
4409
4410 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4411 {
4412 if (startup_info->have_repository)
4413 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4414 else {
4415 char *hex = oid_to_hex(oid);
4416 if (abbrev < 0)
4417 abbrev = FALLBACK_DEFAULT_ABBREV;
4418 if (abbrev > the_hash_algo->hexsz)
4419 BUG("oid abbreviation out of range: %d", abbrev);
4420 if (abbrev)
4421 hex[abbrev] = '\0';
4422 return hex;
4423 }
4424 }
4425
4426 static void fill_metainfo(struct strbuf *msg,
4427 const char *name,
4428 const char *other,
4429 struct diff_filespec *one,
4430 struct diff_filespec *two,
4431 struct diff_options *o,
4432 struct diff_filepair *p,
4433 int *must_show_header,
4434 int use_color)
4435 {
4436 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4437 const char *reset = diff_get_color(use_color, DIFF_RESET);
4438 const char *line_prefix = diff_line_prefix(o);
4439 struct string_list *more_headers = NULL;
4440
4441 *must_show_header = 1;
4442 strbuf_init(msg, PATH_MAX * 2 + 300);
4443 switch (p->status) {
4444 case DIFF_STATUS_COPIED:
4445 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4446 line_prefix, set, similarity_index(p));
4447 strbuf_addf(msg, "%s\n%s%scopy from ",
4448 reset, line_prefix, set);
4449 quote_c_style(name, msg, NULL, 0);
4450 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4451 quote_c_style(other, msg, NULL, 0);
4452 strbuf_addf(msg, "%s\n", reset);
4453 break;
4454 case DIFF_STATUS_RENAMED:
4455 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4456 line_prefix, set, similarity_index(p));
4457 strbuf_addf(msg, "%s\n%s%srename from ",
4458 reset, line_prefix, set);
4459 quote_c_style(name, msg, NULL, 0);
4460 strbuf_addf(msg, "%s\n%s%srename to ",
4461 reset, line_prefix, set);
4462 quote_c_style(other, msg, NULL, 0);
4463 strbuf_addf(msg, "%s\n", reset);
4464 break;
4465 case DIFF_STATUS_MODIFIED:
4466 if (p->score) {
4467 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4468 line_prefix,
4469 set, similarity_index(p), reset);
4470 break;
4471 }
4472 /* fallthru */
4473 default:
4474 *must_show_header = 0;
4475 }
4476 if ((more_headers = additional_headers(o, name))) {
4477 add_formatted_headers(msg, more_headers,
4478 line_prefix, set, reset);
4479 *must_show_header = 1;
4480 }
4481 if (one && two && !oideq(&one->oid, &two->oid)) {
4482 const unsigned hexsz = the_hash_algo->hexsz;
4483 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4484
4485 if (o->flags.full_index)
4486 abbrev = hexsz;
4487
4488 if (o->flags.binary) {
4489 mmfile_t mf;
4490 if ((!fill_mmfile(o->repo, &mf, one) &&
4491 diff_filespec_is_binary(o->repo, one)) ||
4492 (!fill_mmfile(o->repo, &mf, two) &&
4493 diff_filespec_is_binary(o->repo, two)))
4494 abbrev = hexsz;
4495 }
4496 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4497 diff_abbrev_oid(&one->oid, abbrev),
4498 diff_abbrev_oid(&two->oid, abbrev));
4499 if (one->mode == two->mode)
4500 strbuf_addf(msg, " %06o", one->mode);
4501 strbuf_addf(msg, "%s\n", reset);
4502 }
4503 }
4504
4505 static void run_diff_cmd(const char *pgm,
4506 const char *name,
4507 const char *other,
4508 const char *attr_path,
4509 struct diff_filespec *one,
4510 struct diff_filespec *two,
4511 struct strbuf *msg,
4512 struct diff_options *o,
4513 struct diff_filepair *p)
4514 {
4515 const char *xfrm_msg = NULL;
4516 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4517 int must_show_header = 0;
4518 struct userdiff_driver *drv = NULL;
4519
4520 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4521 drv = userdiff_find_by_path(o->repo->index, attr_path);
4522
4523 if (o->flags.allow_external && drv && drv->external)
4524 pgm = drv->external;
4525
4526 if (msg) {
4527 /*
4528 * don't use colors when the header is intended for an
4529 * external diff driver
4530 */
4531 fill_metainfo(msg, name, other, one, two, o, p,
4532 &must_show_header,
4533 want_color(o->use_color) && !pgm);
4534 xfrm_msg = msg->len ? msg->buf : NULL;
4535 }
4536
4537 if (pgm) {
4538 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4539 return;
4540 }
4541 if (one && two) {
4542 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4543 set_diff_algorithm(o, drv->algorithm);
4544
4545 builtin_diff(name, other ? other : name,
4546 one, two, xfrm_msg, must_show_header,
4547 o, complete_rewrite);
4548 } else {
4549 fprintf(o->file, "* Unmerged path %s\n", name);
4550 }
4551 }
4552
4553 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4554 {
4555 if (DIFF_FILE_VALID(one)) {
4556 if (!one->oid_valid) {
4557 struct stat st;
4558 if (one->is_stdin) {
4559 oidclr(&one->oid);
4560 return;
4561 }
4562 if (lstat(one->path, &st) < 0)
4563 die_errno("stat '%s'", one->path);
4564 if (index_path(istate, &one->oid, one->path, &st, 0))
4565 die("cannot hash %s", one->path);
4566 }
4567 }
4568 else
4569 oidclr(&one->oid);
4570 }
4571
4572 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4573 {
4574 /* Strip the prefix but do not molest /dev/null and absolute paths */
4575 if (*namep && !is_absolute_path(*namep)) {
4576 *namep += prefix_length;
4577 if (**namep == '/')
4578 ++*namep;
4579 }
4580 if (*otherp && !is_absolute_path(*otherp)) {
4581 *otherp += prefix_length;
4582 if (**otherp == '/')
4583 ++*otherp;
4584 }
4585 }
4586
4587 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4588 {
4589 const char *pgm = external_diff();
4590 struct strbuf msg;
4591 struct diff_filespec *one = p->one;
4592 struct diff_filespec *two = p->two;
4593 const char *name;
4594 const char *other;
4595 const char *attr_path;
4596
4597 name = one->path;
4598 other = (strcmp(name, two->path) ? two->path : NULL);
4599 attr_path = name;
4600 if (o->prefix_length)
4601 strip_prefix(o->prefix_length, &name, &other);
4602
4603 if (!o->flags.allow_external)
4604 pgm = NULL;
4605
4606 if (DIFF_PAIR_UNMERGED(p)) {
4607 run_diff_cmd(pgm, name, NULL, attr_path,
4608 NULL, NULL, NULL, o, p);
4609 return;
4610 }
4611
4612 diff_fill_oid_info(one, o->repo->index);
4613 diff_fill_oid_info(two, o->repo->index);
4614
4615 if (!pgm &&
4616 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4617 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4618 /*
4619 * a filepair that changes between file and symlink
4620 * needs to be split into deletion and creation.
4621 */
4622 struct diff_filespec *null = alloc_filespec(two->path);
4623 run_diff_cmd(NULL, name, other, attr_path,
4624 one, null, &msg,
4625 o, p);
4626 free(null);
4627 strbuf_release(&msg);
4628
4629 null = alloc_filespec(one->path);
4630 run_diff_cmd(NULL, name, other, attr_path,
4631 null, two, &msg, o, p);
4632 free(null);
4633 }
4634 else
4635 run_diff_cmd(pgm, name, other, attr_path,
4636 one, two, &msg, o, p);
4637
4638 strbuf_release(&msg);
4639 }
4640
4641 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4642 struct diffstat_t *diffstat)
4643 {
4644 const char *name;
4645 const char *other;
4646
4647 if (!o->ignore_driver_algorithm) {
4648 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4649 p->one->path);
4650
4651 if (drv && drv->algorithm)
4652 set_diff_algorithm(o, drv->algorithm);
4653 }
4654
4655 if (DIFF_PAIR_UNMERGED(p)) {
4656 /* unmerged */
4657 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4658 diffstat, o, p);
4659 return;
4660 }
4661
4662 name = p->one->path;
4663 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4664
4665 if (o->prefix_length)
4666 strip_prefix(o->prefix_length, &name, &other);
4667
4668 diff_fill_oid_info(p->one, o->repo->index);
4669 diff_fill_oid_info(p->two, o->repo->index);
4670
4671 builtin_diffstat(name, other, p->one, p->two,
4672 diffstat, o, p);
4673 }
4674
4675 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4676 {
4677 const char *name;
4678 const char *other;
4679 const char *attr_path;
4680
4681 if (DIFF_PAIR_UNMERGED(p)) {
4682 /* unmerged */
4683 return;
4684 }
4685
4686 name = p->one->path;
4687 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4688 attr_path = other ? other : name;
4689
4690 if (o->prefix_length)
4691 strip_prefix(o->prefix_length, &name, &other);
4692
4693 diff_fill_oid_info(p->one, o->repo->index);
4694 diff_fill_oid_info(p->two, o->repo->index);
4695
4696 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4697 }
4698
4699 void repo_diff_setup(struct repository *r, struct diff_options *options)
4700 {
4701 memcpy(options, &default_diff_options, sizeof(*options));
4702
4703 options->file = stdout;
4704 options->repo = r;
4705
4706 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4707 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4708 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4709 options->abbrev = DEFAULT_ABBREV;
4710 options->line_termination = '\n';
4711 options->break_opt = -1;
4712 options->rename_limit = -1;
4713 options->dirstat_permille = diff_dirstat_permille_default;
4714 options->context = diff_context_default;
4715 options->interhunkcontext = diff_interhunk_context_default;
4716 options->ws_error_highlight = ws_error_highlight_default;
4717 options->flags.rename_empty = 1;
4718 options->flags.relative_name = diff_relative;
4719 options->objfind = NULL;
4720
4721 /* pathchange left =NULL by default */
4722 options->change = diff_change;
4723 options->add_remove = diff_addremove;
4724 options->use_color = diff_use_color_default;
4725 options->detect_rename = diff_detect_rename_default;
4726 options->xdl_opts |= diff_algorithm;
4727 if (diff_indent_heuristic)
4728 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4729
4730 options->orderfile = diff_order_file_cfg;
4731
4732 if (!options->flags.ignore_submodule_set)
4733 options->flags.ignore_untracked_in_submodules = 1;
4734
4735 if (diff_no_prefix) {
4736 diff_set_noprefix(options);
4737 } else if (!diff_mnemonic_prefix) {
4738 diff_set_default_prefix(options);
4739 }
4740
4741 options->color_moved = diff_color_moved_default;
4742 options->color_moved_ws_handling = diff_color_moved_ws_default;
4743 }
4744
4745 static const char diff_status_letters[] = {
4746 DIFF_STATUS_ADDED,
4747 DIFF_STATUS_COPIED,
4748 DIFF_STATUS_DELETED,
4749 DIFF_STATUS_MODIFIED,
4750 DIFF_STATUS_RENAMED,
4751 DIFF_STATUS_TYPE_CHANGED,
4752 DIFF_STATUS_UNKNOWN,
4753 DIFF_STATUS_UNMERGED,
4754 DIFF_STATUS_FILTER_AON,
4755 DIFF_STATUS_FILTER_BROKEN,
4756 '\0',
4757 };
4758
4759 static unsigned int filter_bit['Z' + 1];
4760
4761 static void prepare_filter_bits(void)
4762 {
4763 int i;
4764
4765 if (!filter_bit[DIFF_STATUS_ADDED]) {
4766 for (i = 0; diff_status_letters[i]; i++)
4767 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4768 }
4769 }
4770
4771 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4772 {
4773 return opt->filter & filter_bit[(int) status];
4774 }
4775
4776 unsigned diff_filter_bit(char status)
4777 {
4778 prepare_filter_bits();
4779 return filter_bit[(int) status];
4780 }
4781
4782 int diff_check_follow_pathspec(struct pathspec *ps, int die_on_error)
4783 {
4784 unsigned forbidden_magic;
4785
4786 if (ps->nr != 1) {
4787 if (die_on_error)
4788 die(_("--follow requires exactly one pathspec"));
4789 return 0;
4790 }
4791
4792 forbidden_magic = ps->items[0].magic & ~(PATHSPEC_FROMTOP |
4793 PATHSPEC_LITERAL);
4794 if (forbidden_magic) {
4795 if (die_on_error) {
4796 struct strbuf sb = STRBUF_INIT;
4797 pathspec_magic_names(forbidden_magic, &sb);
4798 die(_("pathspec magic not supported by --follow: %s"),
4799 sb.buf);
4800 }
4801 return 0;
4802 }
4803
4804 return 1;
4805 }
4806
4807 void diff_setup_done(struct diff_options *options)
4808 {
4809 unsigned check_mask = DIFF_FORMAT_NAME |
4810 DIFF_FORMAT_NAME_STATUS |
4811 DIFF_FORMAT_CHECKDIFF |
4812 DIFF_FORMAT_NO_OUTPUT;
4813 /*
4814 * This must be signed because we're comparing against a potentially
4815 * negative value.
4816 */
4817 const int hexsz = the_hash_algo->hexsz;
4818
4819 if (options->set_default)
4820 options->set_default(options);
4821
4822 if (HAS_MULTI_BITS(options->output_format & check_mask))
4823 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4824 "--name-only", "--name-status", "--check", "-s");
4825
4826 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4827 die(_("options '%s', '%s', and '%s' cannot be used together"),
4828 "-G", "-S", "--find-object");
4829
4830 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4831 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4832 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4833
4834 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4835 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4836 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4837
4838 /*
4839 * Most of the time we can say "there are changes"
4840 * only by checking if there are changed paths, but
4841 * --ignore-whitespace* options force us to look
4842 * inside contents.
4843 */
4844
4845 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4846 options->ignore_regex_nr)
4847 options->flags.diff_from_contents = 1;
4848 else
4849 options->flags.diff_from_contents = 0;
4850
4851 if (options->flags.find_copies_harder)
4852 options->detect_rename = DIFF_DETECT_COPY;
4853
4854 if (!options->flags.relative_name)
4855 options->prefix = NULL;
4856 if (options->prefix)
4857 options->prefix_length = strlen(options->prefix);
4858 else
4859 options->prefix_length = 0;
4860
4861 /*
4862 * --name-only, --name-status, --checkdiff, and -s
4863 * turn other output format off.
4864 */
4865 if (options->output_format & (DIFF_FORMAT_NAME |
4866 DIFF_FORMAT_NAME_STATUS |
4867 DIFF_FORMAT_CHECKDIFF |
4868 DIFF_FORMAT_NO_OUTPUT))
4869 options->output_format &= ~(DIFF_FORMAT_RAW |
4870 DIFF_FORMAT_NUMSTAT |
4871 DIFF_FORMAT_DIFFSTAT |
4872 DIFF_FORMAT_SHORTSTAT |
4873 DIFF_FORMAT_DIRSTAT |
4874 DIFF_FORMAT_SUMMARY |
4875 DIFF_FORMAT_PATCH);
4876
4877 /*
4878 * These cases always need recursive; we do not drop caller-supplied
4879 * recursive bits for other formats here.
4880 */
4881 if (options->output_format & (DIFF_FORMAT_PATCH |
4882 DIFF_FORMAT_NUMSTAT |
4883 DIFF_FORMAT_DIFFSTAT |
4884 DIFF_FORMAT_SHORTSTAT |
4885 DIFF_FORMAT_DIRSTAT |
4886 DIFF_FORMAT_SUMMARY |
4887 DIFF_FORMAT_CHECKDIFF))
4888 options->flags.recursive = 1;
4889 /*
4890 * Also pickaxe would not work very well if you do not say recursive
4891 */
4892 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4893 options->flags.recursive = 1;
4894 /*
4895 * When patches are generated, submodules diffed against the work tree
4896 * must be checked for dirtiness too so it can be shown in the output
4897 */
4898 if (options->output_format & DIFF_FORMAT_PATCH)
4899 options->flags.dirty_submodules = 1;
4900
4901 if (options->detect_rename && options->rename_limit < 0)
4902 options->rename_limit = diff_rename_limit_default;
4903 if (hexsz < options->abbrev)
4904 options->abbrev = hexsz; /* full */
4905
4906 /*
4907 * It does not make sense to show the first hit we happened
4908 * to have found. It does not make sense not to return with
4909 * exit code in such a case either.
4910 */
4911 if (options->flags.quick) {
4912 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4913 options->flags.exit_with_status = 1;
4914 }
4915
4916 options->diff_path_counter = 0;
4917
4918 if (options->flags.follow_renames)
4919 diff_check_follow_pathspec(&options->pathspec, 1);
4920
4921 if (!options->use_color || external_diff())
4922 options->color_moved = 0;
4923
4924 if (options->filter_not) {
4925 if (!options->filter)
4926 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4927 options->filter &= ~options->filter_not;
4928 }
4929 }
4930
4931 int parse_long_opt(const char *opt, const char **argv,
4932 const char **optarg)
4933 {
4934 const char *arg = argv[0];
4935 if (!skip_prefix(arg, "--", &arg))
4936 return 0;
4937 if (!skip_prefix(arg, opt, &arg))
4938 return 0;
4939 if (*arg == '=') { /* stuck form: --option=value */
4940 *optarg = arg + 1;
4941 return 1;
4942 }
4943 if (*arg != '\0')
4944 return 0;
4945 /* separate form: --option value */
4946 if (!argv[1])
4947 die("Option '--%s' requires a value", opt);
4948 *optarg = argv[1];
4949 return 2;
4950 }
4951
4952 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4953 {
4954 struct diff_options *options = opt->value;
4955 int width = options->stat_width;
4956 int name_width = options->stat_name_width;
4957 int graph_width = options->stat_graph_width;
4958 int count = options->stat_count;
4959 char *end;
4960
4961 BUG_ON_OPT_NEG(unset);
4962
4963 if (!strcmp(opt->long_name, "stat")) {
4964 if (value) {
4965 width = strtoul(value, &end, 10);
4966 if (*end == ',')
4967 name_width = strtoul(end+1, &end, 10);
4968 if (*end == ',')
4969 count = strtoul(end+1, &end, 10);
4970 if (*end)
4971 return error(_("invalid --stat value: %s"), value);
4972 }
4973 } else if (!strcmp(opt->long_name, "stat-width")) {
4974 width = strtoul(value, &end, 10);
4975 if (*end)
4976 return error(_("%s expects a numerical value"),
4977 opt->long_name);
4978 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4979 name_width = strtoul(value, &end, 10);
4980 if (*end)
4981 return error(_("%s expects a numerical value"),
4982 opt->long_name);
4983 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4984 graph_width = strtoul(value, &end, 10);
4985 if (*end)
4986 return error(_("%s expects a numerical value"),
4987 opt->long_name);
4988 } else if (!strcmp(opt->long_name, "stat-count")) {
4989 count = strtoul(value, &end, 10);
4990 if (*end)
4991 return error(_("%s expects a numerical value"),
4992 opt->long_name);
4993 } else
4994 BUG("%s should not get here", opt->long_name);
4995
4996 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
4997 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4998 options->stat_name_width = name_width;
4999 options->stat_graph_width = graph_width;
5000 options->stat_width = width;
5001 options->stat_count = count;
5002 return 0;
5003 }
5004
5005 static int parse_dirstat_opt(struct diff_options *options, const char *params)
5006 {
5007 struct strbuf errmsg = STRBUF_INIT;
5008 if (parse_dirstat_params(options, params, &errmsg))
5009 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
5010 errmsg.buf);
5011 strbuf_release(&errmsg);
5012 /*
5013 * The caller knows a dirstat-related option is given from the command
5014 * line; allow it to say "return this_function();"
5015 */
5016 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5017 options->output_format |= DIFF_FORMAT_DIRSTAT;
5018 return 1;
5019 }
5020
5021 static int diff_opt_diff_filter(const struct option *option,
5022 const char *optarg, int unset)
5023 {
5024 struct diff_options *opt = option->value;
5025 int i, optch;
5026
5027 BUG_ON_OPT_NEG(unset);
5028 prepare_filter_bits();
5029
5030 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
5031 unsigned int bit;
5032 int negate;
5033
5034 if ('a' <= optch && optch <= 'z') {
5035 negate = 1;
5036 optch = toupper(optch);
5037 } else {
5038 negate = 0;
5039 }
5040
5041 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
5042 if (!bit)
5043 return error(_("unknown change class '%c' in --diff-filter=%s"),
5044 optarg[i], optarg);
5045 if (negate)
5046 opt->filter_not |= bit;
5047 else
5048 opt->filter |= bit;
5049 }
5050 return 0;
5051 }
5052
5053 static void enable_patch_output(int *fmt)
5054 {
5055 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
5056 *fmt |= DIFF_FORMAT_PATCH;
5057 }
5058
5059 static int diff_opt_ws_error_highlight(const struct option *option,
5060 const char *arg, int unset)
5061 {
5062 struct diff_options *opt = option->value;
5063 int val = parse_ws_error_highlight(arg);
5064
5065 BUG_ON_OPT_NEG(unset);
5066 if (val < 0)
5067 return error(_("unknown value after ws-error-highlight=%.*s"),
5068 -1 - val, arg);
5069 opt->ws_error_highlight = val;
5070 return 0;
5071 }
5072
5073 static int diff_opt_find_object(const struct option *option,
5074 const char *arg, int unset)
5075 {
5076 struct diff_options *opt = option->value;
5077 struct object_id oid;
5078
5079 BUG_ON_OPT_NEG(unset);
5080 if (repo_get_oid(the_repository, arg, &oid))
5081 return error(_("unable to resolve '%s'"), arg);
5082
5083 if (!opt->objfind)
5084 CALLOC_ARRAY(opt->objfind, 1);
5085
5086 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5087 opt->flags.recursive = 1;
5088 opt->flags.tree_in_recursive = 1;
5089 oidset_insert(opt->objfind, &oid);
5090 return 0;
5091 }
5092
5093 static int diff_opt_anchored(const struct option *opt,
5094 const char *arg, int unset)
5095 {
5096 struct diff_options *options = opt->value;
5097
5098 BUG_ON_OPT_NEG(unset);
5099 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5100 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5101 options->anchors_alloc);
5102 options->anchors[options->anchors_nr++] = xstrdup(arg);
5103 return 0;
5104 }
5105
5106 static int diff_opt_binary(const struct option *opt,
5107 const char *arg, int unset)
5108 {
5109 struct diff_options *options = opt->value;
5110
5111 BUG_ON_OPT_NEG(unset);
5112 BUG_ON_OPT_ARG(arg);
5113 enable_patch_output(&options->output_format);
5114 options->flags.binary = 1;
5115 return 0;
5116 }
5117
5118 static int diff_opt_break_rewrites(const struct option *opt,
5119 const char *arg, int unset)
5120 {
5121 int *break_opt = opt->value;
5122 int opt1, opt2;
5123
5124 BUG_ON_OPT_NEG(unset);
5125 if (!arg)
5126 arg = "";
5127 opt1 = parse_rename_score(&arg);
5128 if (*arg == 0)
5129 opt2 = 0;
5130 else if (*arg != '/')
5131 return error(_("%s expects <n>/<m> form"), opt->long_name);
5132 else {
5133 arg++;
5134 opt2 = parse_rename_score(&arg);
5135 }
5136 if (*arg != 0)
5137 return error(_("%s expects <n>/<m> form"), opt->long_name);
5138 *break_opt = opt1 | (opt2 << 16);
5139 return 0;
5140 }
5141
5142 static int diff_opt_char(const struct option *opt,
5143 const char *arg, int unset)
5144 {
5145 char *value = opt->value;
5146
5147 BUG_ON_OPT_NEG(unset);
5148 if (arg[1])
5149 return error(_("%s expects a character, got '%s'"),
5150 opt->long_name, arg);
5151 *value = arg[0];
5152 return 0;
5153 }
5154
5155 static int diff_opt_color_moved(const struct option *opt,
5156 const char *arg, int unset)
5157 {
5158 struct diff_options *options = opt->value;
5159
5160 if (unset) {
5161 options->color_moved = COLOR_MOVED_NO;
5162 } else if (!arg) {
5163 if (diff_color_moved_default)
5164 options->color_moved = diff_color_moved_default;
5165 if (options->color_moved == COLOR_MOVED_NO)
5166 options->color_moved = COLOR_MOVED_DEFAULT;
5167 } else {
5168 int cm = parse_color_moved(arg);
5169 if (cm < 0)
5170 return error(_("bad --color-moved argument: %s"), arg);
5171 options->color_moved = cm;
5172 }
5173 return 0;
5174 }
5175
5176 static int diff_opt_color_moved_ws(const struct option *opt,
5177 const char *arg, int unset)
5178 {
5179 struct diff_options *options = opt->value;
5180 unsigned cm;
5181
5182 if (unset) {
5183 options->color_moved_ws_handling = 0;
5184 return 0;
5185 }
5186
5187 cm = parse_color_moved_ws(arg);
5188 if (cm & COLOR_MOVED_WS_ERROR)
5189 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5190 options->color_moved_ws_handling = cm;
5191 return 0;
5192 }
5193
5194 static int diff_opt_color_words(const struct option *opt,
5195 const char *arg, int unset)
5196 {
5197 struct diff_options *options = opt->value;
5198
5199 BUG_ON_OPT_NEG(unset);
5200 options->use_color = 1;
5201 options->word_diff = DIFF_WORDS_COLOR;
5202 options->word_regex = arg;
5203 return 0;
5204 }
5205
5206 static int diff_opt_compact_summary(const struct option *opt,
5207 const char *arg, int unset)
5208 {
5209 struct diff_options *options = opt->value;
5210
5211 BUG_ON_OPT_ARG(arg);
5212 if (unset) {
5213 options->flags.stat_with_summary = 0;
5214 } else {
5215 options->flags.stat_with_summary = 1;
5216 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5217 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5218 }
5219 return 0;
5220 }
5221
5222 static int diff_opt_diff_algorithm(const struct option *opt,
5223 const char *arg, int unset)
5224 {
5225 struct diff_options *options = opt->value;
5226
5227 BUG_ON_OPT_NEG(unset);
5228
5229 if (set_diff_algorithm(options, arg))
5230 return error(_("option diff-algorithm accepts \"myers\", "
5231 "\"minimal\", \"patience\" and \"histogram\""));
5232
5233 options->ignore_driver_algorithm = 1;
5234
5235 return 0;
5236 }
5237
5238 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5239 const char *arg, int unset)
5240 {
5241 struct diff_options *options = opt->value;
5242
5243 BUG_ON_OPT_NEG(unset);
5244 BUG_ON_OPT_ARG(arg);
5245
5246 if (set_diff_algorithm(options, opt->long_name))
5247 BUG("available diff algorithms include \"myers\", "
5248 "\"minimal\", \"patience\" and \"histogram\"");
5249
5250 options->ignore_driver_algorithm = 1;
5251
5252 return 0;
5253 }
5254
5255 static int diff_opt_dirstat(const struct option *opt,
5256 const char *arg, int unset)
5257 {
5258 struct diff_options *options = opt->value;
5259
5260 BUG_ON_OPT_NEG(unset);
5261 if (!strcmp(opt->long_name, "cumulative")) {
5262 if (arg)
5263 BUG("how come --cumulative take a value?");
5264 arg = "cumulative";
5265 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5266 parse_dirstat_opt(options, "files");
5267 parse_dirstat_opt(options, arg ? arg : "");
5268 return 0;
5269 }
5270
5271 static int diff_opt_find_copies(const struct option *opt,
5272 const char *arg, int unset)
5273 {
5274 struct diff_options *options = opt->value;
5275
5276 BUG_ON_OPT_NEG(unset);
5277 if (!arg)
5278 arg = "";
5279 options->rename_score = parse_rename_score(&arg);
5280 if (*arg != 0)
5281 return error(_("invalid argument to %s"), opt->long_name);
5282
5283 if (options->detect_rename == DIFF_DETECT_COPY)
5284 options->flags.find_copies_harder = 1;
5285 else
5286 options->detect_rename = DIFF_DETECT_COPY;
5287
5288 return 0;
5289 }
5290
5291 static int diff_opt_find_renames(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 if (!arg)
5298 arg = "";
5299 options->rename_score = parse_rename_score(&arg);
5300 if (*arg != 0)
5301 return error(_("invalid argument to %s"), opt->long_name);
5302
5303 options->detect_rename = DIFF_DETECT_RENAME;
5304 return 0;
5305 }
5306
5307 static int diff_opt_follow(const struct option *opt,
5308 const char *arg, int unset)
5309 {
5310 struct diff_options *options = opt->value;
5311
5312 BUG_ON_OPT_ARG(arg);
5313 if (unset) {
5314 options->flags.follow_renames = 0;
5315 options->flags.default_follow_renames = 0;
5316 } else {
5317 options->flags.follow_renames = 1;
5318 }
5319 return 0;
5320 }
5321
5322 static int diff_opt_ignore_submodules(const struct option *opt,
5323 const char *arg, int unset)
5324 {
5325 struct diff_options *options = opt->value;
5326
5327 BUG_ON_OPT_NEG(unset);
5328 if (!arg)
5329 arg = "all";
5330 options->flags.override_submodule_config = 1;
5331 handle_ignore_submodules_arg(options, arg);
5332 return 0;
5333 }
5334
5335 static int diff_opt_line_prefix(const struct option *opt,
5336 const char *optarg, int unset)
5337 {
5338 struct diff_options *options = opt->value;
5339
5340 BUG_ON_OPT_NEG(unset);
5341 options->line_prefix = optarg;
5342 options->line_prefix_length = strlen(options->line_prefix);
5343 graph_setup_line_prefix(options);
5344 return 0;
5345 }
5346
5347 static int diff_opt_no_prefix(const struct option *opt,
5348 const char *optarg, int unset)
5349 {
5350 struct diff_options *options = opt->value;
5351
5352 BUG_ON_OPT_NEG(unset);
5353 BUG_ON_OPT_ARG(optarg);
5354 diff_set_noprefix(options);
5355 return 0;
5356 }
5357
5358 static int diff_opt_default_prefix(const struct option *opt,
5359 const char *optarg, int unset)
5360 {
5361 struct diff_options *options = opt->value;
5362
5363 BUG_ON_OPT_NEG(unset);
5364 BUG_ON_OPT_ARG(optarg);
5365 diff_set_default_prefix(options);
5366 return 0;
5367 }
5368
5369 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5370 const struct option *opt,
5371 const char *arg, int unset)
5372 {
5373 struct diff_options *options = opt->value;
5374 char *path;
5375
5376 BUG_ON_OPT_NEG(unset);
5377 path = prefix_filename(ctx->prefix, arg);
5378 options->file = xfopen(path, "w");
5379 options->close_file = 1;
5380 if (options->use_color != GIT_COLOR_ALWAYS)
5381 options->use_color = GIT_COLOR_NEVER;
5382 free(path);
5383 return 0;
5384 }
5385
5386 static int diff_opt_patience(const struct option *opt,
5387 const char *arg, int unset)
5388 {
5389 struct diff_options *options = opt->value;
5390 int i;
5391
5392 BUG_ON_OPT_NEG(unset);
5393 BUG_ON_OPT_ARG(arg);
5394 /*
5395 * Both --patience and --anchored use PATIENCE_DIFF
5396 * internally, so remove any anchors previously
5397 * specified.
5398 */
5399 for (i = 0; i < options->anchors_nr; i++)
5400 free(options->anchors[i]);
5401 options->anchors_nr = 0;
5402 options->ignore_driver_algorithm = 1;
5403
5404 return set_diff_algorithm(options, "patience");
5405 }
5406
5407 static int diff_opt_ignore_regex(const struct option *opt,
5408 const char *arg, int unset)
5409 {
5410 struct diff_options *options = opt->value;
5411 regex_t *regex;
5412
5413 BUG_ON_OPT_NEG(unset);
5414 regex = xmalloc(sizeof(*regex));
5415 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5416 return error(_("invalid regex given to -I: '%s'"), arg);
5417 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5418 options->ignore_regex_alloc);
5419 options->ignore_regex[options->ignore_regex_nr++] = regex;
5420 return 0;
5421 }
5422
5423 static int diff_opt_pickaxe_regex(const struct option *opt,
5424 const char *arg, int unset)
5425 {
5426 struct diff_options *options = opt->value;
5427
5428 BUG_ON_OPT_NEG(unset);
5429 options->pickaxe = arg;
5430 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5431 return 0;
5432 }
5433
5434 static int diff_opt_pickaxe_string(const struct option *opt,
5435 const char *arg, int unset)
5436 {
5437 struct diff_options *options = opt->value;
5438
5439 BUG_ON_OPT_NEG(unset);
5440 options->pickaxe = arg;
5441 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5442 return 0;
5443 }
5444
5445 static int diff_opt_relative(const struct option *opt,
5446 const char *arg, int unset)
5447 {
5448 struct diff_options *options = opt->value;
5449
5450 options->flags.relative_name = !unset;
5451 if (arg)
5452 options->prefix = arg;
5453 return 0;
5454 }
5455
5456 static int diff_opt_submodule(const struct option *opt,
5457 const char *arg, int unset)
5458 {
5459 struct diff_options *options = opt->value;
5460
5461 BUG_ON_OPT_NEG(unset);
5462 if (!arg)
5463 arg = "log";
5464 if (parse_submodule_params(options, arg))
5465 return error(_("failed to parse --submodule option parameter: '%s'"),
5466 arg);
5467 return 0;
5468 }
5469
5470 static int diff_opt_textconv(const struct option *opt,
5471 const char *arg, int unset)
5472 {
5473 struct diff_options *options = opt->value;
5474
5475 BUG_ON_OPT_ARG(arg);
5476 if (unset) {
5477 options->flags.allow_textconv = 0;
5478 } else {
5479 options->flags.allow_textconv = 1;
5480 options->flags.textconv_set_via_cmdline = 1;
5481 }
5482 return 0;
5483 }
5484
5485 static int diff_opt_unified(const struct option *opt,
5486 const char *arg, int unset)
5487 {
5488 struct diff_options *options = opt->value;
5489 char *s;
5490
5491 BUG_ON_OPT_NEG(unset);
5492
5493 if (arg) {
5494 options->context = strtol(arg, &s, 10);
5495 if (*s)
5496 return error(_("%s expects a numerical value"), "--unified");
5497 }
5498 enable_patch_output(&options->output_format);
5499
5500 return 0;
5501 }
5502
5503 static int diff_opt_word_diff(const struct option *opt,
5504 const char *arg, int unset)
5505 {
5506 struct diff_options *options = opt->value;
5507
5508 BUG_ON_OPT_NEG(unset);
5509 if (arg) {
5510 if (!strcmp(arg, "plain"))
5511 options->word_diff = DIFF_WORDS_PLAIN;
5512 else if (!strcmp(arg, "color")) {
5513 options->use_color = 1;
5514 options->word_diff = DIFF_WORDS_COLOR;
5515 }
5516 else if (!strcmp(arg, "porcelain"))
5517 options->word_diff = DIFF_WORDS_PORCELAIN;
5518 else if (!strcmp(arg, "none"))
5519 options->word_diff = DIFF_WORDS_NONE;
5520 else
5521 return error(_("bad --word-diff argument: %s"), arg);
5522 } else {
5523 if (options->word_diff == DIFF_WORDS_NONE)
5524 options->word_diff = DIFF_WORDS_PLAIN;
5525 }
5526 return 0;
5527 }
5528
5529 static int diff_opt_word_diff_regex(const struct option *opt,
5530 const char *arg, int unset)
5531 {
5532 struct diff_options *options = opt->value;
5533
5534 BUG_ON_OPT_NEG(unset);
5535 if (options->word_diff == DIFF_WORDS_NONE)
5536 options->word_diff = DIFF_WORDS_PLAIN;
5537 options->word_regex = arg;
5538 return 0;
5539 }
5540
5541 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5542 {
5543 struct diff_options *options = opt->value;
5544
5545 BUG_ON_OPT_NEG(unset);
5546 if (!strcmp(opt->long_name, "skip-to"))
5547 options->skip_instead_of_rotate = 1;
5548 else
5549 options->skip_instead_of_rotate = 0;
5550 options->rotate_to = arg;
5551 return 0;
5552 }
5553
5554 /*
5555 * Consider adding new flags to __git_diff_common_options
5556 * in contrib/completion/git-completion.bash
5557 */
5558 struct option *add_diff_options(const struct option *opts,
5559 struct diff_options *options)
5560 {
5561 struct option parseopts[] = {
5562 OPT_GROUP(N_("Diff output format options")),
5563 OPT_BITOP('p', "patch", &options->output_format,
5564 N_("generate patch"),
5565 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5566 OPT_SET_INT('s', "no-patch", &options->output_format,
5567 N_("suppress diff output"), DIFF_FORMAT_NO_OUTPUT),
5568 OPT_BITOP('u', NULL, &options->output_format,
5569 N_("generate patch"),
5570 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5571 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5572 N_("generate diffs with <n> lines context"),
5573 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5574 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5575 N_("generate diffs with <n> lines context")),
5576 OPT_BITOP(0, "raw", &options->output_format,
5577 N_("generate the diff in raw format"),
5578 DIFF_FORMAT_RAW, DIFF_FORMAT_NO_OUTPUT),
5579 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5580 N_("synonym for '-p --raw'"),
5581 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5582 DIFF_FORMAT_NO_OUTPUT),
5583 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5584 N_("synonym for '-p --stat'"),
5585 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5586 DIFF_FORMAT_NO_OUTPUT),
5587 OPT_BITOP(0, "numstat", &options->output_format,
5588 N_("machine friendly --stat"),
5589 DIFF_FORMAT_NUMSTAT, DIFF_FORMAT_NO_OUTPUT),
5590 OPT_BITOP(0, "shortstat", &options->output_format,
5591 N_("output only the last line of --stat"),
5592 DIFF_FORMAT_SHORTSTAT, DIFF_FORMAT_NO_OUTPUT),
5593 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1>,<param2>..."),
5594 N_("output the distribution of relative amount of changes for each sub-directory"),
5595 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5596 diff_opt_dirstat),
5597 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5598 N_("synonym for --dirstat=cumulative"),
5599 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5600 diff_opt_dirstat),
5601 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1>,<param2>..."),
5602 N_("synonym for --dirstat=files,<param1>,<param2>..."),
5603 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5604 diff_opt_dirstat),
5605 OPT_BIT_F(0, "check", &options->output_format,
5606 N_("warn if changes introduce conflict markers or whitespace errors"),
5607 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5608 OPT_BITOP(0, "summary", &options->output_format,
5609 N_("condensed summary such as creations, renames and mode changes"),
5610 DIFF_FORMAT_SUMMARY, DIFF_FORMAT_NO_OUTPUT),
5611 OPT_BIT_F(0, "name-only", &options->output_format,
5612 N_("show only names of changed files"),
5613 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5614 OPT_BIT_F(0, "name-status", &options->output_format,
5615 N_("show only names and status of changed files"),
5616 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5617 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5618 N_("generate diffstat"),
5619 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5620 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5621 N_("generate diffstat with a given width"),
5622 PARSE_OPT_NONEG, diff_opt_stat),
5623 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5624 N_("generate diffstat with a given name width"),
5625 PARSE_OPT_NONEG, diff_opt_stat),
5626 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5627 N_("generate diffstat with a given graph width"),
5628 PARSE_OPT_NONEG, diff_opt_stat),
5629 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5630 N_("generate diffstat with limited lines"),
5631 PARSE_OPT_NONEG, diff_opt_stat),
5632 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5633 N_("generate compact summary in diffstat"),
5634 PARSE_OPT_NOARG, diff_opt_compact_summary),
5635 OPT_CALLBACK_F(0, "binary", options, NULL,
5636 N_("output a binary diff that can be applied"),
5637 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5638 OPT_BOOL(0, "full-index", &options->flags.full_index,
5639 N_("show full pre- and post-image object names on the \"index\" lines")),
5640 OPT_COLOR_FLAG(0, "color", &options->use_color,
5641 N_("show colored diff")),
5642 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5643 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5644 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5645 OPT_SET_INT('z', NULL, &options->line_termination,
5646 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5647 0),
5648 OPT__ABBREV(&options->abbrev),
5649 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5650 N_("show the given source prefix instead of \"a/\""),
5651 PARSE_OPT_NONEG),
5652 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5653 N_("show the given destination prefix instead of \"b/\""),
5654 PARSE_OPT_NONEG),
5655 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5656 N_("prepend an additional prefix to every line of output"),
5657 PARSE_OPT_NONEG, diff_opt_line_prefix),
5658 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5659 N_("do not show any source or destination prefix"),
5660 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5661 OPT_CALLBACK_F(0, "default-prefix", options, NULL,
5662 N_("use default prefixes a/ and b/"),
5663 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_default_prefix),
5664 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5665 N_("show context between diff hunks up to the specified number of lines"),
5666 PARSE_OPT_NONEG),
5667 OPT_CALLBACK_F(0, "output-indicator-new",
5668 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5669 N_("<char>"),
5670 N_("specify the character to indicate a new line instead of '+'"),
5671 PARSE_OPT_NONEG, diff_opt_char),
5672 OPT_CALLBACK_F(0, "output-indicator-old",
5673 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5674 N_("<char>"),
5675 N_("specify the character to indicate an old line instead of '-'"),
5676 PARSE_OPT_NONEG, diff_opt_char),
5677 OPT_CALLBACK_F(0, "output-indicator-context",
5678 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5679 N_("<char>"),
5680 N_("specify the character to indicate a context instead of ' '"),
5681 PARSE_OPT_NONEG, diff_opt_char),
5682
5683 OPT_GROUP(N_("Diff rename options")),
5684 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5685 N_("break complete rewrite changes into pairs of delete and create"),
5686 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5687 diff_opt_break_rewrites),
5688 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5689 N_("detect renames"),
5690 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5691 diff_opt_find_renames),
5692 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5693 N_("omit the preimage for deletes"),
5694 1, PARSE_OPT_NONEG),
5695 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5696 N_("detect copies"),
5697 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5698 diff_opt_find_copies),
5699 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5700 N_("use unmodified files as source to find copies")),
5701 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5702 N_("disable rename detection"),
5703 0, PARSE_OPT_NONEG),
5704 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5705 N_("use empty blobs as rename source")),
5706 OPT_CALLBACK_F(0, "follow", options, NULL,
5707 N_("continue listing the history of a file beyond renames"),
5708 PARSE_OPT_NOARG, diff_opt_follow),
5709 OPT_INTEGER('l', NULL, &options->rename_limit,
5710 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5711
5712 OPT_GROUP(N_("Diff algorithm options")),
5713 OPT_CALLBACK_F(0, "minimal", options, NULL,
5714 N_("produce the smallest possible diff"),
5715 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5716 diff_opt_diff_algorithm_no_arg),
5717 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5718 N_("ignore whitespace when comparing lines"),
5719 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5720 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5721 N_("ignore changes in amount of whitespace"),
5722 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5723 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5724 N_("ignore changes in whitespace at EOL"),
5725 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5726 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5727 N_("ignore carrier-return at the end of line"),
5728 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5729 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5730 N_("ignore changes whose lines are all blank"),
5731 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5732 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5733 N_("ignore changes whose all lines match <regex>"),
5734 0, diff_opt_ignore_regex),
5735 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5736 N_("heuristic to shift diff hunk boundaries for easy reading"),
5737 XDF_INDENT_HEURISTIC),
5738 OPT_CALLBACK_F(0, "patience", options, NULL,
5739 N_("generate diff using the \"patience diff\" algorithm"),
5740 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5741 diff_opt_patience),
5742 OPT_CALLBACK_F(0, "histogram", options, NULL,
5743 N_("generate diff using the \"histogram diff\" algorithm"),
5744 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5745 diff_opt_diff_algorithm_no_arg),
5746 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5747 N_("choose a diff algorithm"),
5748 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5749 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5750 N_("generate diff using the \"anchored diff\" algorithm"),
5751 PARSE_OPT_NONEG, diff_opt_anchored),
5752 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5753 N_("show word diff, using <mode> to delimit changed words"),
5754 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5755 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5756 N_("use <regex> to decide what a word is"),
5757 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5758 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5759 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5760 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5761 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5762 N_("moved lines of code are colored differently"),
5763 PARSE_OPT_OPTARG, diff_opt_color_moved),
5764 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5765 N_("how white spaces are ignored in --color-moved"),
5766 0, diff_opt_color_moved_ws),
5767
5768 OPT_GROUP(N_("Other diff options")),
5769 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5770 N_("when run from subdir, exclude changes outside and show relative paths"),
5771 PARSE_OPT_OPTARG,
5772 diff_opt_relative),
5773 OPT_BOOL('a', "text", &options->flags.text,
5774 N_("treat all files as text")),
5775 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5776 N_("swap two inputs, reverse the diff")),
5777 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5778 N_("exit with 1 if there were differences, 0 otherwise")),
5779 OPT_BOOL(0, "quiet", &options->flags.quick,
5780 N_("disable all output of the program")),
5781 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5782 N_("allow an external diff helper to be executed")),
5783 OPT_CALLBACK_F(0, "textconv", options, NULL,
5784 N_("run external text conversion filters when comparing binary files"),
5785 PARSE_OPT_NOARG, diff_opt_textconv),
5786 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5787 N_("ignore changes to submodules in the diff generation"),
5788 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5789 diff_opt_ignore_submodules),
5790 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5791 N_("specify how differences in submodules are shown"),
5792 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5793 diff_opt_submodule),
5794 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5795 N_("hide 'git add -N' entries from the index"),
5796 1, PARSE_OPT_NONEG),
5797 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5798 N_("treat 'git add -N' entries as real in the index"),
5799 0, PARSE_OPT_NONEG),
5800 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5801 N_("look for differences that change the number of occurrences of the specified string"),
5802 0, diff_opt_pickaxe_string),
5803 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5804 N_("look for differences that change the number of occurrences of the specified regex"),
5805 0, diff_opt_pickaxe_regex),
5806 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5807 N_("show all changes in the changeset with -S or -G"),
5808 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5809 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5810 N_("treat <string> in -S as extended POSIX regular expression"),
5811 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5812 OPT_FILENAME('O', NULL, &options->orderfile,
5813 N_("control the order in which files appear in the output")),
5814 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5815 N_("show the change in the specified path first"),
5816 PARSE_OPT_NONEG, diff_opt_rotate_to),
5817 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5818 N_("skip the output to the specified path"),
5819 PARSE_OPT_NONEG, diff_opt_rotate_to),
5820 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5821 N_("look for differences that change the number of occurrences of the specified object"),
5822 PARSE_OPT_NONEG, diff_opt_find_object),
5823 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5824 N_("select files by diff type"),
5825 PARSE_OPT_NONEG, diff_opt_diff_filter),
5826 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5827 N_("output to a specific file"),
5828 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5829
5830 OPT_END()
5831 };
5832
5833 return parse_options_concat(opts, parseopts);
5834 }
5835
5836 int diff_opt_parse(struct diff_options *options,
5837 const char **av, int ac, const char *prefix)
5838 {
5839 struct option no_options[] = { OPT_END() };
5840 struct option *parseopts = add_diff_options(no_options, options);
5841
5842 if (!prefix)
5843 prefix = "";
5844
5845 ac = parse_options(ac, av, prefix, parseopts, NULL,
5846 PARSE_OPT_KEEP_DASHDASH |
5847 PARSE_OPT_KEEP_UNKNOWN_OPT |
5848 PARSE_OPT_NO_INTERNAL_HELP |
5849 PARSE_OPT_ONE_SHOT |
5850 PARSE_OPT_STOP_AT_NON_OPTION);
5851 free(parseopts);
5852
5853 return ac;
5854 }
5855
5856 int parse_rename_score(const char **cp_p)
5857 {
5858 unsigned long num, scale;
5859 int ch, dot;
5860 const char *cp = *cp_p;
5861
5862 num = 0;
5863 scale = 1;
5864 dot = 0;
5865 for (;;) {
5866 ch = *cp;
5867 if ( !dot && ch == '.' ) {
5868 scale = 1;
5869 dot = 1;
5870 } else if ( ch == '%' ) {
5871 scale = dot ? scale*100 : 100;
5872 cp++; /* % is always at the end */
5873 break;
5874 } else if ( ch >= '0' && ch <= '9' ) {
5875 if ( scale < 100000 ) {
5876 scale *= 10;
5877 num = (num*10) + (ch-'0');
5878 }
5879 } else {
5880 break;
5881 }
5882 cp++;
5883 }
5884 *cp_p = cp;
5885
5886 /* user says num divided by scale and we say internally that
5887 * is MAX_SCORE * num / scale.
5888 */
5889 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5890 }
5891
5892 struct diff_queue_struct diff_queued_diff;
5893
5894 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5895 {
5896 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5897 queue->queue[queue->nr++] = dp;
5898 }
5899
5900 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5901 struct diff_filespec *one,
5902 struct diff_filespec *two)
5903 {
5904 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5905 dp->one = one;
5906 dp->two = two;
5907 if (queue)
5908 diff_q(queue, dp);
5909 return dp;
5910 }
5911
5912 void diff_free_filepair(struct diff_filepair *p)
5913 {
5914 free_filespec(p->one);
5915 free_filespec(p->two);
5916 free(p);
5917 }
5918
5919 void diff_free_queue(struct diff_queue_struct *q)
5920 {
5921 for (int i = 0; i < q->nr; i++)
5922 diff_free_filepair(q->queue[i]);
5923 free(q->queue);
5924 }
5925
5926 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5927 {
5928 int abblen;
5929 const char *abbrev;
5930
5931 /* Do we want all 40 hex characters? */
5932 if (len == the_hash_algo->hexsz)
5933 return oid_to_hex(oid);
5934
5935 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5936 abbrev = diff_abbrev_oid(oid, len);
5937
5938 if (!print_sha1_ellipsis())
5939 return abbrev;
5940
5941 abblen = strlen(abbrev);
5942
5943 /*
5944 * In well-behaved cases, where the abbreviated result is the
5945 * same as the requested length, append three dots after the
5946 * abbreviation (hence the whole logic is limited to the case
5947 * where abblen < 37); when the actual abbreviated result is a
5948 * bit longer than the requested length, we reduce the number
5949 * of dots so that they match the well-behaved ones. However,
5950 * if the actual abbreviation is longer than the requested
5951 * length by more than three, we give up on aligning, and add
5952 * three dots anyway, to indicate that the output is not the
5953 * full object name. Yes, this may be suboptimal, but this
5954 * appears only in "diff --raw --abbrev" output and it is not
5955 * worth the effort to change it now. Note that this would
5956 * likely to work fine when the automatic sizing of default
5957 * abbreviation length is used--we would be fed -1 in "len" in
5958 * that case, and will end up always appending three-dots, but
5959 * the automatic sizing is supposed to give abblen that ensures
5960 * uniqueness across all objects (statistically speaking).
5961 */
5962 if (abblen < the_hash_algo->hexsz - 3) {
5963 static char hex[GIT_MAX_HEXSZ + 1];
5964 if (len < abblen && abblen <= len + 2)
5965 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5966 else
5967 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5968 return hex;
5969 }
5970
5971 return oid_to_hex(oid);
5972 }
5973
5974 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5975 {
5976 int line_termination = opt->line_termination;
5977 int inter_name_termination = line_termination ? '\t' : '\0';
5978
5979 fprintf(opt->file, "%s", diff_line_prefix(opt));
5980 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5981 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5982 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5983 fprintf(opt->file, "%s ",
5984 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5985 }
5986 if (p->score) {
5987 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5988 inter_name_termination);
5989 } else {
5990 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5991 }
5992
5993 if (p->status == DIFF_STATUS_COPIED ||
5994 p->status == DIFF_STATUS_RENAMED) {
5995 const char *name_a, *name_b;
5996 name_a = p->one->path;
5997 name_b = p->two->path;
5998 strip_prefix(opt->prefix_length, &name_a, &name_b);
5999 write_name_quoted(name_a, opt->file, inter_name_termination);
6000 write_name_quoted(name_b, opt->file, line_termination);
6001 } else {
6002 const char *name_a, *name_b;
6003 name_a = p->one->mode ? p->one->path : p->two->path;
6004 name_b = NULL;
6005 strip_prefix(opt->prefix_length, &name_a, &name_b);
6006 write_name_quoted(name_a, opt->file, line_termination);
6007 }
6008 }
6009
6010 int diff_unmodified_pair(struct diff_filepair *p)
6011 {
6012 /* This function is written stricter than necessary to support
6013 * the currently implemented transformers, but the idea is to
6014 * let transformers to produce diff_filepairs any way they want,
6015 * and filter and clean them up here before producing the output.
6016 */
6017 struct diff_filespec *one = p->one, *two = p->two;
6018
6019 if (DIFF_PAIR_UNMERGED(p))
6020 return 0; /* unmerged is interesting */
6021
6022 /* deletion, addition, mode or type change
6023 * and rename are all interesting.
6024 */
6025 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
6026 DIFF_PAIR_MODE_CHANGED(p) ||
6027 strcmp(one->path, two->path))
6028 return 0;
6029
6030 /* both are valid and point at the same path. that is, we are
6031 * dealing with a change.
6032 */
6033 if (one->oid_valid && two->oid_valid &&
6034 oideq(&one->oid, &two->oid) &&
6035 !one->dirty_submodule && !two->dirty_submodule)
6036 return 1; /* no change */
6037 if (!one->oid_valid && !two->oid_valid)
6038 return 1; /* both look at the same file on the filesystem. */
6039 return 0;
6040 }
6041
6042 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
6043 {
6044 int include_conflict_headers =
6045 (additional_headers(o, p->one->path) &&
6046 !o->pickaxe_opts &&
6047 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6048
6049 /*
6050 * Check if we can return early without showing a diff. Note that
6051 * diff_filepair only stores {oid, path, mode, is_valid}
6052 * information for each path, and thus diff_unmodified_pair() only
6053 * considers those bits of info. However, we do not want pairs
6054 * created by create_filepairs_for_header_only_notifications()
6055 * (which always look like unmodified pairs) to be ignored, so
6056 * return early if both p is unmodified AND we don't want to
6057 * include_conflict_headers.
6058 */
6059 if (diff_unmodified_pair(p) && !include_conflict_headers)
6060 return;
6061
6062 /* Actually, we can also return early to avoid showing tree diffs */
6063 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6064 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6065 return;
6066
6067 run_diff(p, o);
6068 }
6069
6070 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
6071 struct diffstat_t *diffstat)
6072 {
6073 if (diff_unmodified_pair(p))
6074 return;
6075
6076 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6077 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6078 return; /* no useful stat for tree diffs */
6079
6080 run_diffstat(p, o, diffstat);
6081 }
6082
6083 static void diff_flush_checkdiff(struct diff_filepair *p,
6084 struct diff_options *o)
6085 {
6086 if (diff_unmodified_pair(p))
6087 return;
6088
6089 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6090 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6091 return; /* nothing to check in tree diffs */
6092
6093 run_checkdiff(p, o);
6094 }
6095
6096 int diff_queue_is_empty(struct diff_options *o)
6097 {
6098 struct diff_queue_struct *q = &diff_queued_diff;
6099 int i;
6100 int include_conflict_headers =
6101 (o->additional_path_headers &&
6102 strmap_get_size(o->additional_path_headers) &&
6103 !o->pickaxe_opts &&
6104 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6105
6106 if (include_conflict_headers)
6107 return 0;
6108
6109 for (i = 0; i < q->nr; i++)
6110 if (!diff_unmodified_pair(q->queue[i]))
6111 return 0;
6112 return 1;
6113 }
6114
6115 #if DIFF_DEBUG
6116 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6117 {
6118 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6119 x, one ? one : "",
6120 s->path,
6121 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6122 s->mode,
6123 s->oid_valid ? oid_to_hex(&s->oid) : "");
6124 fprintf(stderr, "queue[%d] %s size %lu\n",
6125 x, one ? one : "",
6126 s->size);
6127 }
6128
6129 void diff_debug_filepair(const struct diff_filepair *p, int i)
6130 {
6131 diff_debug_filespec(p->one, i, "one");
6132 diff_debug_filespec(p->two, i, "two");
6133 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6134 p->score, p->status ? p->status : '?',
6135 p->one->rename_used, p->broken_pair);
6136 }
6137
6138 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6139 {
6140 int i;
6141 if (msg)
6142 fprintf(stderr, "%s\n", msg);
6143 fprintf(stderr, "q->nr = %d\n", q->nr);
6144 for (i = 0; i < q->nr; i++) {
6145 struct diff_filepair *p = q->queue[i];
6146 diff_debug_filepair(p, i);
6147 }
6148 }
6149 #endif
6150
6151 static void diff_resolve_rename_copy(void)
6152 {
6153 int i;
6154 struct diff_filepair *p;
6155 struct diff_queue_struct *q = &diff_queued_diff;
6156
6157 diff_debug_queue("resolve-rename-copy", q);
6158
6159 for (i = 0; i < q->nr; i++) {
6160 p = q->queue[i];
6161 p->status = 0; /* undecided */
6162 if (DIFF_PAIR_UNMERGED(p))
6163 p->status = DIFF_STATUS_UNMERGED;
6164 else if (!DIFF_FILE_VALID(p->one))
6165 p->status = DIFF_STATUS_ADDED;
6166 else if (!DIFF_FILE_VALID(p->two))
6167 p->status = DIFF_STATUS_DELETED;
6168 else if (DIFF_PAIR_TYPE_CHANGED(p))
6169 p->status = DIFF_STATUS_TYPE_CHANGED;
6170
6171 /* from this point on, we are dealing with a pair
6172 * whose both sides are valid and of the same type, i.e.
6173 * either in-place edit or rename/copy edit.
6174 */
6175 else if (DIFF_PAIR_RENAME(p)) {
6176 /*
6177 * A rename might have re-connected a broken
6178 * pair up, causing the pathnames to be the
6179 * same again. If so, that's not a rename at
6180 * all, just a modification..
6181 *
6182 * Otherwise, see if this source was used for
6183 * multiple renames, in which case we decrement
6184 * the count, and call it a copy.
6185 */
6186 if (!strcmp(p->one->path, p->two->path))
6187 p->status = DIFF_STATUS_MODIFIED;
6188 else if (--p->one->rename_used > 0)
6189 p->status = DIFF_STATUS_COPIED;
6190 else
6191 p->status = DIFF_STATUS_RENAMED;
6192 }
6193 else if (!oideq(&p->one->oid, &p->two->oid) ||
6194 p->one->mode != p->two->mode ||
6195 p->one->dirty_submodule ||
6196 p->two->dirty_submodule ||
6197 is_null_oid(&p->one->oid))
6198 p->status = DIFF_STATUS_MODIFIED;
6199 else {
6200 /* This is a "no-change" entry and should not
6201 * happen anymore, but prepare for broken callers.
6202 */
6203 error("feeding unmodified %s to diffcore",
6204 p->one->path);
6205 p->status = DIFF_STATUS_UNKNOWN;
6206 }
6207 }
6208 diff_debug_queue("resolve-rename-copy done", q);
6209 }
6210
6211 static int check_pair_status(struct diff_filepair *p)
6212 {
6213 switch (p->status) {
6214 case DIFF_STATUS_UNKNOWN:
6215 return 0;
6216 case 0:
6217 die("internal error in diff-resolve-rename-copy");
6218 default:
6219 return 1;
6220 }
6221 }
6222
6223 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6224 {
6225 int fmt = opt->output_format;
6226
6227 if (fmt & DIFF_FORMAT_CHECKDIFF)
6228 diff_flush_checkdiff(p, opt);
6229 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6230 diff_flush_raw(p, opt);
6231 else if (fmt & DIFF_FORMAT_NAME) {
6232 const char *name_a, *name_b;
6233 name_a = p->two->path;
6234 name_b = NULL;
6235 strip_prefix(opt->prefix_length, &name_a, &name_b);
6236 fprintf(opt->file, "%s", diff_line_prefix(opt));
6237 write_name_quoted(name_a, opt->file, opt->line_termination);
6238 }
6239
6240 opt->found_changes = 1;
6241 }
6242
6243 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6244 {
6245 struct strbuf sb = STRBUF_INIT;
6246 if (fs->mode)
6247 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6248 else
6249 strbuf_addf(&sb, " %s ", newdelete);
6250
6251 quote_c_style(fs->path, &sb, NULL, 0);
6252 strbuf_addch(&sb, '\n');
6253 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6254 sb.buf, sb.len, 0);
6255 strbuf_release(&sb);
6256 }
6257
6258 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6259 int show_name)
6260 {
6261 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6262 struct strbuf sb = STRBUF_INIT;
6263 strbuf_addf(&sb, " mode change %06o => %06o",
6264 p->one->mode, p->two->mode);
6265 if (show_name) {
6266 strbuf_addch(&sb, ' ');
6267 quote_c_style(p->two->path, &sb, NULL, 0);
6268 }
6269 strbuf_addch(&sb, '\n');
6270 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6271 sb.buf, sb.len, 0);
6272 strbuf_release(&sb);
6273 }
6274 }
6275
6276 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6277 struct diff_filepair *p)
6278 {
6279 struct strbuf sb = STRBUF_INIT;
6280 struct strbuf names = STRBUF_INIT;
6281
6282 pprint_rename(&names, p->one->path, p->two->path);
6283 strbuf_addf(&sb, " %s %s (%d%%)\n",
6284 renamecopy, names.buf, similarity_index(p));
6285 strbuf_release(&names);
6286 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6287 sb.buf, sb.len, 0);
6288 show_mode_change(opt, p, 0);
6289 strbuf_release(&sb);
6290 }
6291
6292 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6293 {
6294 switch(p->status) {
6295 case DIFF_STATUS_DELETED:
6296 show_file_mode_name(opt, "delete", p->one);
6297 break;
6298 case DIFF_STATUS_ADDED:
6299 show_file_mode_name(opt, "create", p->two);
6300 break;
6301 case DIFF_STATUS_COPIED:
6302 show_rename_copy(opt, "copy", p);
6303 break;
6304 case DIFF_STATUS_RENAMED:
6305 show_rename_copy(opt, "rename", p);
6306 break;
6307 default:
6308 if (p->score) {
6309 struct strbuf sb = STRBUF_INIT;
6310 strbuf_addstr(&sb, " rewrite ");
6311 quote_c_style(p->two->path, &sb, NULL, 0);
6312 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6313 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6314 sb.buf, sb.len, 0);
6315 strbuf_release(&sb);
6316 }
6317 show_mode_change(opt, p, !p->score);
6318 break;
6319 }
6320 }
6321
6322 struct patch_id_t {
6323 git_hash_ctx *ctx;
6324 int patchlen;
6325 };
6326
6327 static int remove_space(char *line, int len)
6328 {
6329 int i;
6330 char *dst = line;
6331 unsigned char c;
6332
6333 for (i = 0; i < len; i++)
6334 if (!isspace((c = line[i])))
6335 *dst++ = c;
6336
6337 return dst - line;
6338 }
6339
6340 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6341 {
6342 unsigned char hash[GIT_MAX_RAWSZ];
6343 unsigned short carry = 0;
6344 int i;
6345
6346 the_hash_algo->final_fn(hash, ctx);
6347 the_hash_algo->init_fn(ctx);
6348 /* 20-byte sum, with carry */
6349 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6350 carry += result->hash[i] + hash[i];
6351 result->hash[i] = carry;
6352 carry >>= 8;
6353 }
6354 }
6355
6356 static int patch_id_consume(void *priv, char *line, unsigned long len)
6357 {
6358 struct patch_id_t *data = priv;
6359 int new_len;
6360
6361 if (len > 12 && starts_with(line, "\\ "))
6362 return 0;
6363 new_len = remove_space(line, len);
6364
6365 the_hash_algo->update_fn(data->ctx, line, new_len);
6366 data->patchlen += new_len;
6367 return 0;
6368 }
6369
6370 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6371 {
6372 the_hash_algo->update_fn(ctx, str, strlen(str));
6373 }
6374
6375 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6376 {
6377 /* large enough for 2^32 in octal */
6378 char buf[12];
6379 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6380 the_hash_algo->update_fn(ctx, buf, len);
6381 }
6382
6383 /* returns 0 upon success, and writes result into oid */
6384 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6385 {
6386 struct diff_queue_struct *q = &diff_queued_diff;
6387 int i;
6388 git_hash_ctx ctx;
6389 struct patch_id_t data;
6390
6391 the_hash_algo->init_fn(&ctx);
6392 memset(&data, 0, sizeof(struct patch_id_t));
6393 data.ctx = &ctx;
6394 oidclr(oid);
6395
6396 for (i = 0; i < q->nr; i++) {
6397 xpparam_t xpp;
6398 xdemitconf_t xecfg;
6399 mmfile_t mf1, mf2;
6400 struct diff_filepair *p = q->queue[i];
6401 int len1, len2;
6402
6403 memset(&xpp, 0, sizeof(xpp));
6404 memset(&xecfg, 0, sizeof(xecfg));
6405 if (p->status == 0)
6406 return error("internal diff status error");
6407 if (p->status == DIFF_STATUS_UNKNOWN)
6408 continue;
6409 if (diff_unmodified_pair(p))
6410 continue;
6411 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6412 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6413 continue;
6414 if (DIFF_PAIR_UNMERGED(p))
6415 continue;
6416
6417 diff_fill_oid_info(p->one, options->repo->index);
6418 diff_fill_oid_info(p->two, options->repo->index);
6419
6420 len1 = remove_space(p->one->path, strlen(p->one->path));
6421 len2 = remove_space(p->two->path, strlen(p->two->path));
6422 patch_id_add_string(&ctx, "diff--git");
6423 patch_id_add_string(&ctx, "a/");
6424 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6425 patch_id_add_string(&ctx, "b/");
6426 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6427
6428 if (p->one->mode == 0) {
6429 patch_id_add_string(&ctx, "newfilemode");
6430 patch_id_add_mode(&ctx, p->two->mode);
6431 } else if (p->two->mode == 0) {
6432 patch_id_add_string(&ctx, "deletedfilemode");
6433 patch_id_add_mode(&ctx, p->one->mode);
6434 } else if (p->one->mode != p->two->mode) {
6435 patch_id_add_string(&ctx, "oldmode");
6436 patch_id_add_mode(&ctx, p->one->mode);
6437 patch_id_add_string(&ctx, "newmode");
6438 patch_id_add_mode(&ctx, p->two->mode);
6439 }
6440
6441 if (diff_header_only) {
6442 /* don't do anything since we're only populating header info */
6443 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6444 diff_filespec_is_binary(options->repo, p->two)) {
6445 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6446 the_hash_algo->hexsz);
6447 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6448 the_hash_algo->hexsz);
6449 } else {
6450 if (p->one->mode == 0) {
6451 patch_id_add_string(&ctx, "---/dev/null");
6452 patch_id_add_string(&ctx, "+++b/");
6453 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6454 } else if (p->two->mode == 0) {
6455 patch_id_add_string(&ctx, "---a/");
6456 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6457 patch_id_add_string(&ctx, "+++/dev/null");
6458 } else {
6459 patch_id_add_string(&ctx, "---a/");
6460 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6461 patch_id_add_string(&ctx, "+++b/");
6462 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6463 }
6464
6465 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6466 fill_mmfile(options->repo, &mf2, p->two) < 0)
6467 return error("unable to read files to diff");
6468 xpp.flags = 0;
6469 xecfg.ctxlen = 3;
6470 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6471 if (xdi_diff_outf(&mf1, &mf2, NULL,
6472 patch_id_consume, &data, &xpp, &xecfg))
6473 return error("unable to generate patch-id diff for %s",
6474 p->one->path);
6475 }
6476 flush_one_hunk(oid, &ctx);
6477 }
6478
6479 return 0;
6480 }
6481
6482 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6483 {
6484 struct diff_queue_struct *q = &diff_queued_diff;
6485 int result = diff_get_patch_id(options, oid, diff_header_only);
6486
6487 diff_free_queue(q);
6488 DIFF_QUEUE_CLEAR(q);
6489
6490 return result;
6491 }
6492
6493 static int is_summary_empty(const struct diff_queue_struct *q)
6494 {
6495 int i;
6496
6497 for (i = 0; i < q->nr; i++) {
6498 const struct diff_filepair *p = q->queue[i];
6499
6500 switch (p->status) {
6501 case DIFF_STATUS_DELETED:
6502 case DIFF_STATUS_ADDED:
6503 case DIFF_STATUS_COPIED:
6504 case DIFF_STATUS_RENAMED:
6505 return 0;
6506 default:
6507 if (p->score)
6508 return 0;
6509 if (p->one->mode && p->two->mode &&
6510 p->one->mode != p->two->mode)
6511 return 0;
6512 break;
6513 }
6514 }
6515 return 1;
6516 }
6517
6518 static const char rename_limit_warning[] =
6519 N_("exhaustive rename detection was skipped due to too many files.");
6520
6521 static const char degrade_cc_to_c_warning[] =
6522 N_("only found copies from modified paths due to too many files.");
6523
6524 static const char rename_limit_advice[] =
6525 N_("you may want to set your %s variable to at least "
6526 "%d and retry the command.");
6527
6528 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6529 {
6530 fflush(stdout);
6531 if (degraded_cc)
6532 warning(_(degrade_cc_to_c_warning));
6533 else if (needed)
6534 warning(_(rename_limit_warning));
6535 else
6536 return;
6537 if (0 < needed)
6538 warning(_(rename_limit_advice), varname, needed);
6539 }
6540
6541 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6542 {
6543 struct strset present;
6544 struct diff_queue_struct *q = &diff_queued_diff;
6545 struct hashmap_iter iter;
6546 struct strmap_entry *e;
6547 int i;
6548
6549 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6550
6551 /*
6552 * Find out which paths exist in diff_queued_diff, preferring
6553 * one->path for any pair that has multiple paths.
6554 */
6555 for (i = 0; i < q->nr; i++) {
6556 struct diff_filepair *p = q->queue[i];
6557 char *path = p->one->path ? p->one->path : p->two->path;
6558
6559 if (strmap_contains(o->additional_path_headers, path))
6560 strset_add(&present, path);
6561 }
6562
6563 /*
6564 * Loop over paths in additional_path_headers; for each NOT already
6565 * in diff_queued_diff, create a synthetic filepair and insert that
6566 * into diff_queued_diff.
6567 */
6568 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6569 if (!strset_contains(&present, e->key)) {
6570 struct diff_filespec *one, *two;
6571 struct diff_filepair *p;
6572
6573 one = alloc_filespec(e->key);
6574 two = alloc_filespec(e->key);
6575 fill_filespec(one, null_oid(), 0, 0);
6576 fill_filespec(two, null_oid(), 0, 0);
6577 p = diff_queue(q, one, two);
6578 p->status = DIFF_STATUS_MODIFIED;
6579 }
6580 }
6581
6582 /* Re-sort the filepairs */
6583 diffcore_fix_diff_index();
6584
6585 /* Cleanup */
6586 strset_clear(&present);
6587 }
6588
6589 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6590 {
6591 int i;
6592 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6593 struct diff_queue_struct *q = &diff_queued_diff;
6594
6595 if (WSEH_NEW & WS_RULE_MASK)
6596 BUG("WS rules bit mask overlaps with diff symbol flags");
6597
6598 if (o->color_moved)
6599 o->emitted_symbols = &esm;
6600
6601 if (o->additional_path_headers)
6602 create_filepairs_for_header_only_notifications(o);
6603
6604 for (i = 0; i < q->nr; i++) {
6605 struct diff_filepair *p = q->queue[i];
6606 if (check_pair_status(p))
6607 diff_flush_patch(p, o);
6608 }
6609
6610 if (o->emitted_symbols) {
6611 if (o->color_moved) {
6612 struct mem_pool entry_pool;
6613 struct moved_entry_list *entry_list;
6614
6615 mem_pool_init(&entry_pool, 1024 * 1024);
6616 entry_list = add_lines_to_move_detection(o,
6617 &entry_pool);
6618 mark_color_as_moved(o, entry_list);
6619 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6620 dim_moved_lines(o);
6621
6622 mem_pool_discard(&entry_pool, 0);
6623 free(entry_list);
6624 }
6625
6626 for (i = 0; i < esm.nr; i++)
6627 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6628
6629 for (i = 0; i < esm.nr; i++)
6630 free((void *)esm.buf[i].line);
6631 esm.nr = 0;
6632
6633 o->emitted_symbols = NULL;
6634 }
6635 }
6636
6637 static void diff_free_file(struct diff_options *options)
6638 {
6639 if (options->close_file)
6640 fclose(options->file);
6641 }
6642
6643 static void diff_free_ignore_regex(struct diff_options *options)
6644 {
6645 int i;
6646
6647 for (i = 0; i < options->ignore_regex_nr; i++) {
6648 regfree(options->ignore_regex[i]);
6649 free(options->ignore_regex[i]);
6650 }
6651 free(options->ignore_regex);
6652 }
6653
6654 void diff_free(struct diff_options *options)
6655 {
6656 if (options->no_free)
6657 return;
6658
6659 diff_free_file(options);
6660 diff_free_ignore_regex(options);
6661 clear_pathspec(&options->pathspec);
6662 }
6663
6664 void diff_flush(struct diff_options *options)
6665 {
6666 struct diff_queue_struct *q = &diff_queued_diff;
6667 int i, output_format = options->output_format;
6668 int separator = 0;
6669 int dirstat_by_line = 0;
6670
6671 /*
6672 * Order: raw, stat, summary, patch
6673 * or: name/name-status/checkdiff (other bits clear)
6674 */
6675 if (!q->nr && !options->additional_path_headers)
6676 goto free_queue;
6677
6678 if (output_format & (DIFF_FORMAT_RAW |
6679 DIFF_FORMAT_NAME |
6680 DIFF_FORMAT_NAME_STATUS |
6681 DIFF_FORMAT_CHECKDIFF)) {
6682 for (i = 0; i < q->nr; i++) {
6683 struct diff_filepair *p = q->queue[i];
6684 if (check_pair_status(p))
6685 flush_one_pair(p, options);
6686 }
6687 separator++;
6688 }
6689
6690 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6691 dirstat_by_line = 1;
6692
6693 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6694 dirstat_by_line) {
6695 struct diffstat_t diffstat;
6696
6697 compute_diffstat(options, &diffstat, q);
6698 if (output_format & DIFF_FORMAT_NUMSTAT)
6699 show_numstat(&diffstat, options);
6700 if (output_format & DIFF_FORMAT_DIFFSTAT)
6701 show_stats(&diffstat, options);
6702 if (output_format & DIFF_FORMAT_SHORTSTAT)
6703 show_shortstats(&diffstat, options);
6704 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6705 show_dirstat_by_line(&diffstat, options);
6706 free_diffstat_info(&diffstat);
6707 separator++;
6708 }
6709 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6710 show_dirstat(options);
6711
6712 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6713 for (i = 0; i < q->nr; i++) {
6714 diff_summary(options, q->queue[i]);
6715 }
6716 separator++;
6717 }
6718
6719 if (output_format & DIFF_FORMAT_PATCH) {
6720 if (separator) {
6721 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6722 if (options->stat_sep)
6723 /* attach patch instead of inline */
6724 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6725 NULL, 0, 0);
6726 }
6727
6728 diff_flush_patch_all_file_pairs(options);
6729 }
6730
6731 if (output_format & DIFF_FORMAT_CALLBACK)
6732 options->format_callback(q, options, options->format_callback_data);
6733
6734 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6735 options->flags.exit_with_status &&
6736 options->flags.diff_from_contents) {
6737 /*
6738 * run diff_flush_patch for the exit status. setting
6739 * options->file to /dev/null should be safe, because we
6740 * aren't supposed to produce any output anyway.
6741 */
6742 diff_free_file(options);
6743 options->file = xfopen("/dev/null", "w");
6744 options->close_file = 1;
6745 options->color_moved = 0;
6746 for (i = 0; i < q->nr; i++) {
6747 struct diff_filepair *p = q->queue[i];
6748 if (check_pair_status(p))
6749 diff_flush_patch(p, options);
6750 if (options->found_changes)
6751 break;
6752 }
6753 }
6754
6755 free_queue:
6756 diff_free_queue(q);
6757 DIFF_QUEUE_CLEAR(q);
6758 diff_free(options);
6759
6760 /*
6761 * Report the content-level differences with HAS_CHANGES;
6762 * diff_addremove/diff_change does not set the bit when
6763 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6764 */
6765 if (options->flags.diff_from_contents) {
6766 if (options->found_changes)
6767 options->flags.has_changes = 1;
6768 else
6769 options->flags.has_changes = 0;
6770 }
6771 }
6772
6773 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6774 {
6775 return (((p->status == DIFF_STATUS_MODIFIED) &&
6776 ((p->score &&
6777 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6778 (!p->score &&
6779 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6780 ((p->status != DIFF_STATUS_MODIFIED) &&
6781 filter_bit_tst(p->status, options)));
6782 }
6783
6784 static void diffcore_apply_filter(struct diff_options *options)
6785 {
6786 int i;
6787 struct diff_queue_struct *q = &diff_queued_diff;
6788 struct diff_queue_struct outq;
6789
6790 DIFF_QUEUE_CLEAR(&outq);
6791
6792 if (!options->filter)
6793 return;
6794
6795 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6796 int found;
6797 for (i = found = 0; !found && i < q->nr; i++) {
6798 if (match_filter(options, q->queue[i]))
6799 found++;
6800 }
6801 if (found)
6802 return;
6803
6804 /* otherwise we will clear the whole queue
6805 * by copying the empty outq at the end of this
6806 * function, but first clear the current entries
6807 * in the queue.
6808 */
6809 for (i = 0; i < q->nr; i++)
6810 diff_free_filepair(q->queue[i]);
6811 }
6812 else {
6813 /* Only the matching ones */
6814 for (i = 0; i < q->nr; i++) {
6815 struct diff_filepair *p = q->queue[i];
6816 if (match_filter(options, p))
6817 diff_q(&outq, p);
6818 else
6819 diff_free_filepair(p);
6820 }
6821 }
6822 free(q->queue);
6823 *q = outq;
6824 }
6825
6826 /* Check whether two filespecs with the same mode and size are identical */
6827 static int diff_filespec_is_identical(struct repository *r,
6828 struct diff_filespec *one,
6829 struct diff_filespec *two)
6830 {
6831 if (S_ISGITLINK(one->mode))
6832 return 0;
6833 if (diff_populate_filespec(r, one, NULL))
6834 return 0;
6835 if (diff_populate_filespec(r, two, NULL))
6836 return 0;
6837 return !memcmp(one->data, two->data, one->size);
6838 }
6839
6840 static int diff_filespec_check_stat_unmatch(struct repository *r,
6841 struct diff_filepair *p)
6842 {
6843 struct diff_populate_filespec_options dpf_options = {
6844 .check_size_only = 1,
6845 .missing_object_cb = diff_queued_diff_prefetch,
6846 .missing_object_data = r,
6847 };
6848
6849 if (p->done_skip_stat_unmatch)
6850 return p->skip_stat_unmatch_result;
6851
6852 p->done_skip_stat_unmatch = 1;
6853 p->skip_stat_unmatch_result = 0;
6854 /*
6855 * 1. Entries that come from stat info dirtiness
6856 * always have both sides (iow, not create/delete),
6857 * one side of the object name is unknown, with
6858 * the same mode and size. Keep the ones that
6859 * do not match these criteria. They have real
6860 * differences.
6861 *
6862 * 2. At this point, the file is known to be modified,
6863 * with the same mode and size, and the object
6864 * name of one side is unknown. Need to inspect
6865 * the identical contents.
6866 */
6867 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6868 !DIFF_FILE_VALID(p->two) ||
6869 (p->one->oid_valid && p->two->oid_valid) ||
6870 (p->one->mode != p->two->mode) ||
6871 diff_populate_filespec(r, p->one, &dpf_options) ||
6872 diff_populate_filespec(r, p->two, &dpf_options) ||
6873 (p->one->size != p->two->size) ||
6874 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6875 p->skip_stat_unmatch_result = 1;
6876 return p->skip_stat_unmatch_result;
6877 }
6878
6879 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6880 {
6881 int i;
6882 struct diff_queue_struct *q = &diff_queued_diff;
6883 struct diff_queue_struct outq;
6884 DIFF_QUEUE_CLEAR(&outq);
6885
6886 for (i = 0; i < q->nr; i++) {
6887 struct diff_filepair *p = q->queue[i];
6888
6889 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6890 diff_q(&outq, p);
6891 else {
6892 /*
6893 * The caller can subtract 1 from skip_stat_unmatch
6894 * to determine how many paths were dirty only
6895 * due to stat info mismatch.
6896 */
6897 if (!diffopt->flags.no_index)
6898 diffopt->skip_stat_unmatch++;
6899 diff_free_filepair(p);
6900 }
6901 }
6902 free(q->queue);
6903 *q = outq;
6904 }
6905
6906 static int diffnamecmp(const void *a_, const void *b_)
6907 {
6908 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6909 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6910 const char *name_a, *name_b;
6911
6912 name_a = a->one ? a->one->path : a->two->path;
6913 name_b = b->one ? b->one->path : b->two->path;
6914 return strcmp(name_a, name_b);
6915 }
6916
6917 void diffcore_fix_diff_index(void)
6918 {
6919 struct diff_queue_struct *q = &diff_queued_diff;
6920 QSORT(q->queue, q->nr, diffnamecmp);
6921 }
6922
6923 void diff_add_if_missing(struct repository *r,
6924 struct oid_array *to_fetch,
6925 const struct diff_filespec *filespec)
6926 {
6927 if (filespec && filespec->oid_valid &&
6928 !S_ISGITLINK(filespec->mode) &&
6929 oid_object_info_extended(r, &filespec->oid, NULL,
6930 OBJECT_INFO_FOR_PREFETCH))
6931 oid_array_append(to_fetch, &filespec->oid);
6932 }
6933
6934 void diff_queued_diff_prefetch(void *repository)
6935 {
6936 struct repository *repo = repository;
6937 int i;
6938 struct diff_queue_struct *q = &diff_queued_diff;
6939 struct oid_array to_fetch = OID_ARRAY_INIT;
6940
6941 for (i = 0; i < q->nr; i++) {
6942 struct diff_filepair *p = q->queue[i];
6943 diff_add_if_missing(repo, &to_fetch, p->one);
6944 diff_add_if_missing(repo, &to_fetch, p->two);
6945 }
6946
6947 /*
6948 * NEEDSWORK: Consider deduplicating the OIDs sent.
6949 */
6950 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6951
6952 oid_array_clear(&to_fetch);
6953 }
6954
6955 void init_diffstat_widths(struct diff_options *options)
6956 {
6957 options->stat_width = -1; /* use full terminal width */
6958 options->stat_name_width = -1; /* respect diff.statNameWidth config */
6959 options->stat_graph_width = -1; /* respect diff.statGraphWidth config */
6960 }
6961
6962 void diffcore_std(struct diff_options *options)
6963 {
6964 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6965 DIFF_FORMAT_NUMSTAT |
6966 DIFF_FORMAT_PATCH |
6967 DIFF_FORMAT_SHORTSTAT |
6968 DIFF_FORMAT_DIRSTAT;
6969
6970 /*
6971 * Check if the user requested a blob-data-requiring diff output and/or
6972 * break-rewrite detection (which requires blob data). If yes, prefetch
6973 * the diff pairs.
6974 *
6975 * If no prefetching occurs, diffcore_rename() will prefetch if it
6976 * decides that it needs inexact rename detection.
6977 */
6978 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
6979 (options->output_format & output_formats_to_prefetch ||
6980 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6981 diff_queued_diff_prefetch(options->repo);
6982
6983 /* NOTE please keep the following in sync with diff_tree_combined() */
6984 if (options->skip_stat_unmatch)
6985 diffcore_skip_stat_unmatch(options);
6986 if (!options->found_follow) {
6987 /* See try_to_follow_renames() in tree-diff.c */
6988 if (options->break_opt != -1)
6989 diffcore_break(options->repo,
6990 options->break_opt);
6991 if (options->detect_rename)
6992 diffcore_rename(options);
6993 if (options->break_opt != -1)
6994 diffcore_merge_broken();
6995 }
6996 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6997 diffcore_pickaxe(options);
6998 if (options->orderfile)
6999 diffcore_order(options->orderfile);
7000 if (options->rotate_to)
7001 diffcore_rotate(options);
7002 if (!options->found_follow)
7003 /* See try_to_follow_renames() in tree-diff.c */
7004 diff_resolve_rename_copy();
7005 diffcore_apply_filter(options);
7006
7007 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
7008 options->flags.has_changes = 1;
7009 else
7010 options->flags.has_changes = 0;
7011
7012 options->found_follow = 0;
7013 }
7014
7015 int diff_result_code(struct diff_options *opt)
7016 {
7017 int result = 0;
7018
7019 diff_warn_rename_limit("diff.renameLimit",
7020 opt->needed_rename_limit,
7021 opt->degraded_cc_to_c);
7022
7023 if (opt->flags.exit_with_status &&
7024 opt->flags.has_changes)
7025 result |= 01;
7026 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
7027 opt->flags.check_failed)
7028 result |= 02;
7029 return result;
7030 }
7031
7032 int diff_can_quit_early(struct diff_options *opt)
7033 {
7034 return (opt->flags.quick &&
7035 !opt->filter &&
7036 opt->flags.has_changes);
7037 }
7038
7039 /*
7040 * Shall changes to this submodule be ignored?
7041 *
7042 * Submodule changes can be configured to be ignored separately for each path,
7043 * but that configuration can be overridden from the command line.
7044 */
7045 static int is_submodule_ignored(const char *path, struct diff_options *options)
7046 {
7047 int ignored = 0;
7048 struct diff_flags orig_flags = options->flags;
7049 if (!options->flags.override_submodule_config)
7050 set_diffopt_flags_from_submodule_config(options, path);
7051 if (options->flags.ignore_submodules)
7052 ignored = 1;
7053 options->flags = orig_flags;
7054 return ignored;
7055 }
7056
7057 void compute_diffstat(struct diff_options *options,
7058 struct diffstat_t *diffstat,
7059 struct diff_queue_struct *q)
7060 {
7061 int i;
7062
7063 memset(diffstat, 0, sizeof(struct diffstat_t));
7064 for (i = 0; i < q->nr; i++) {
7065 struct diff_filepair *p = q->queue[i];
7066 if (check_pair_status(p))
7067 diff_flush_stat(p, options, diffstat);
7068 }
7069 options->found_changes = !!diffstat->nr;
7070 }
7071
7072 void diff_addremove(struct diff_options *options,
7073 int addremove, unsigned mode,
7074 const struct object_id *oid,
7075 int oid_valid,
7076 const char *concatpath, unsigned dirty_submodule)
7077 {
7078 struct diff_filespec *one, *two;
7079
7080 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
7081 return;
7082
7083 /* This may look odd, but it is a preparation for
7084 * feeding "there are unchanged files which should
7085 * not produce diffs, but when you are doing copy
7086 * detection you would need them, so here they are"
7087 * entries to the diff-core. They will be prefixed
7088 * with something like '=' or '*' (I haven't decided
7089 * which but should not make any difference).
7090 * Feeding the same new and old to diff_change()
7091 * also has the same effect.
7092 * Before the final output happens, they are pruned after
7093 * merged into rename/copy pairs as appropriate.
7094 */
7095 if (options->flags.reverse_diff)
7096 addremove = (addremove == '+' ? '-' :
7097 addremove == '-' ? '+' : addremove);
7098
7099 if (options->prefix &&
7100 strncmp(concatpath, options->prefix, options->prefix_length))
7101 return;
7102
7103 one = alloc_filespec(concatpath);
7104 two = alloc_filespec(concatpath);
7105
7106 if (addremove != '+')
7107 fill_filespec(one, oid, oid_valid, mode);
7108 if (addremove != '-') {
7109 fill_filespec(two, oid, oid_valid, mode);
7110 two->dirty_submodule = dirty_submodule;
7111 }
7112
7113 diff_queue(&diff_queued_diff, one, two);
7114 if (!options->flags.diff_from_contents)
7115 options->flags.has_changes = 1;
7116 }
7117
7118 void diff_change(struct diff_options *options,
7119 unsigned old_mode, unsigned new_mode,
7120 const struct object_id *old_oid,
7121 const struct object_id *new_oid,
7122 int old_oid_valid, int new_oid_valid,
7123 const char *concatpath,
7124 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7125 {
7126 struct diff_filespec *one, *two;
7127 struct diff_filepair *p;
7128
7129 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7130 is_submodule_ignored(concatpath, options))
7131 return;
7132
7133 if (options->flags.reverse_diff) {
7134 SWAP(old_mode, new_mode);
7135 SWAP(old_oid, new_oid);
7136 SWAP(old_oid_valid, new_oid_valid);
7137 SWAP(old_dirty_submodule, new_dirty_submodule);
7138 }
7139
7140 if (options->prefix &&
7141 strncmp(concatpath, options->prefix, options->prefix_length))
7142 return;
7143
7144 one = alloc_filespec(concatpath);
7145 two = alloc_filespec(concatpath);
7146 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7147 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7148 one->dirty_submodule = old_dirty_submodule;
7149 two->dirty_submodule = new_dirty_submodule;
7150 p = diff_queue(&diff_queued_diff, one, two);
7151
7152 if (options->flags.diff_from_contents)
7153 return;
7154
7155 if (options->flags.quick && options->skip_stat_unmatch &&
7156 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7157 diff_free_filespec_data(p->one);
7158 diff_free_filespec_data(p->two);
7159 return;
7160 }
7161
7162 options->flags.has_changes = 1;
7163 }
7164
7165 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7166 {
7167 struct diff_filepair *pair;
7168 struct diff_filespec *one, *two;
7169
7170 if (options->prefix &&
7171 strncmp(path, options->prefix, options->prefix_length))
7172 return NULL;
7173
7174 one = alloc_filespec(path);
7175 two = alloc_filespec(path);
7176 pair = diff_queue(&diff_queued_diff, one, two);
7177 pair->is_unmerged = 1;
7178 return pair;
7179 }
7180
7181 static char *run_textconv(struct repository *r,
7182 const char *pgm,
7183 struct diff_filespec *spec,
7184 size_t *outsize)
7185 {
7186 struct diff_tempfile *temp;
7187 struct child_process child = CHILD_PROCESS_INIT;
7188 struct strbuf buf = STRBUF_INIT;
7189 int err = 0;
7190
7191 temp = prepare_temp_file(r, spec);
7192 strvec_push(&child.args, pgm);
7193 strvec_push(&child.args, temp->name);
7194
7195 child.use_shell = 1;
7196 child.out = -1;
7197 if (start_command(&child)) {
7198 remove_tempfile();
7199 return NULL;
7200 }
7201
7202 if (strbuf_read(&buf, child.out, 0) < 0)
7203 err = error("error reading from textconv command '%s'", pgm);
7204 close(child.out);
7205
7206 if (finish_command(&child) || err) {
7207 strbuf_release(&buf);
7208 remove_tempfile();
7209 return NULL;
7210 }
7211 remove_tempfile();
7212
7213 return strbuf_detach(&buf, outsize);
7214 }
7215
7216 size_t fill_textconv(struct repository *r,
7217 struct userdiff_driver *driver,
7218 struct diff_filespec *df,
7219 char **outbuf)
7220 {
7221 size_t size;
7222
7223 if (!driver) {
7224 if (!DIFF_FILE_VALID(df)) {
7225 *outbuf = "";
7226 return 0;
7227 }
7228 if (diff_populate_filespec(r, df, NULL))
7229 die("unable to read files to diff");
7230 *outbuf = df->data;
7231 return df->size;
7232 }
7233
7234 if (!driver->textconv)
7235 BUG("fill_textconv called with non-textconv driver");
7236
7237 if (driver->textconv_cache && df->oid_valid) {
7238 *outbuf = notes_cache_get(driver->textconv_cache,
7239 &df->oid,
7240 &size);
7241 if (*outbuf)
7242 return size;
7243 }
7244
7245 *outbuf = run_textconv(r, driver->textconv, df, &size);
7246 if (!*outbuf)
7247 die("unable to read files to diff");
7248
7249 if (driver->textconv_cache && df->oid_valid) {
7250 /* ignore errors, as we might be in a readonly repository */
7251 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7252 size);
7253 /*
7254 * we could save up changes and flush them all at the end,
7255 * but we would need an extra call after all diffing is done.
7256 * Since generating a cache entry is the slow path anyway,
7257 * this extra overhead probably isn't a big deal.
7258 */
7259 notes_cache_write(driver->textconv_cache);
7260 }
7261
7262 return size;
7263 }
7264
7265 int textconv_object(struct repository *r,
7266 const char *path,
7267 unsigned mode,
7268 const struct object_id *oid,
7269 int oid_valid,
7270 char **buf,
7271 unsigned long *buf_size)
7272 {
7273 struct diff_filespec *df;
7274 struct userdiff_driver *textconv;
7275
7276 df = alloc_filespec(path);
7277 fill_filespec(df, oid, oid_valid, mode);
7278 textconv = get_textconv(r, df);
7279 if (!textconv) {
7280 free_filespec(df);
7281 return 0;
7282 }
7283
7284 *buf_size = fill_textconv(r, textconv, df, buf);
7285 free_filespec(df);
7286 return 1;
7287 }
7288
7289 void setup_diff_pager(struct diff_options *opt)
7290 {
7291 /*
7292 * If the user asked for our exit code, then either they want --quiet
7293 * or --exit-code. We should definitely not bother with a pager in the
7294 * former case, as we will generate no output. Since we still properly
7295 * report our exit code even when a pager is run, we _could_ run a
7296 * pager with --exit-code. But since we have not done so historically,
7297 * and because it is easy to find people oneline advising "git diff
7298 * --exit-code" in hooks and other scripts, we do not do so.
7299 */
7300 if (!opt->flags.exit_with_status &&
7301 check_pager_config("diff") != 0)
7302 setup_pager();
7303 }
The Best VCS Around