?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
ci updates
[git.git] / diff.c
blob3be927b073931127cdf6e08acd2cb1637b89913d
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4
5 #define USE_THE_REPOSITORY_VARIABLE
6
7 #include "git-compat-util.h"
8 #include "abspath.h"
9 #include "base85.h"
10 #include "config.h"
11 #include "convert.h"
12 #include "environment.h"
13 #include "gettext.h"
14 #include "tempfile.h"
15 #include "revision.h"
16 #include "quote.h"
17 #include "diff.h"
18 #include "diffcore.h"
19 #include "delta.h"
20 #include "hex.h"
21 #include "xdiff-interface.h"
22 #include "color.h"
23 #include "run-command.h"
24 #include "utf8.h"
25 #include "object-store-ll.h"
26 #include "userdiff.h"
27 #include "submodule.h"
28 #include "hashmap.h"
29 #include "mem-pool.h"
30 #include "merge-ll.h"
31 #include "string-list.h"
32 #include "strvec.h"
33 #include "tmp-objdir.h"
34 #include "graph.h"
35 #include "oid-array.h"
36 #include "packfile.h"
37 #include "pager.h"
38 #include "parse-options.h"
39 #include "help.h"
40 #include "promisor-remote.h"
41 #include "dir.h"
42 #include "object-file.h"
43 #include "object-name.h"
44 #include "read-cache-ll.h"
45 #include "setup.h"
46 #include "strmap.h"
47 #include "ws.h"
48
49 #ifdef NO_FAST_WORKING_DIRECTORY
50 #define FAST_WORKING_DIRECTORY 0
51 #else
52 #define FAST_WORKING_DIRECTORY 1
53 #endif
54
55 static int diff_detect_rename_default;
56 static int diff_indent_heuristic = 1;
57 static int diff_rename_limit_default = 1000;
58 static int diff_suppress_blank_empty;
59 static int diff_use_color_default = -1;
60 static int diff_color_moved_default;
61 static int diff_color_moved_ws_default;
62 static int diff_context_default = 3;
63 static int diff_interhunk_context_default;
64 static char *diff_word_regex_cfg;
65 static struct external_diff external_diff_cfg;
66 static char *diff_order_file_cfg;
67 int diff_auto_refresh_index = 1;
68 static int diff_mnemonic_prefix;
69 static int diff_no_prefix;
70 static char *diff_src_prefix;
71 static char *diff_dst_prefix;
72 static int diff_relative;
73 static int diff_stat_name_width;
74 static int diff_stat_graph_width;
75 static int diff_dirstat_permille_default = 30;
76 static struct diff_options default_diff_options;
77 static long diff_algorithm;
78 static unsigned ws_error_highlight_default = WSEH_NEW;
79
80 static char diff_colors[][COLOR_MAXLEN] = {
81 GIT_COLOR_RESET,
82 GIT_COLOR_NORMAL, /* CONTEXT */
83 GIT_COLOR_BOLD, /* METAINFO */
84 GIT_COLOR_CYAN, /* FRAGINFO */
85 GIT_COLOR_RED, /* OLD */
86 GIT_COLOR_GREEN, /* NEW */
87 GIT_COLOR_YELLOW, /* COMMIT */
88 GIT_COLOR_BG_RED, /* WHITESPACE */
89 GIT_COLOR_NORMAL, /* FUNCINFO */
90 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
91 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
92 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
93 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
94 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
95 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
96 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
97 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
98 GIT_COLOR_FAINT, /* CONTEXT_DIM */
99 GIT_COLOR_FAINT_RED, /* OLD_DIM */
100 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
101 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
102 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
103 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
104 };
105
106 static const char *color_diff_slots[] = {
107 [DIFF_CONTEXT] = "context",
108 [DIFF_METAINFO] = "meta",
109 [DIFF_FRAGINFO] = "frag",
110 [DIFF_FILE_OLD] = "old",
111 [DIFF_FILE_NEW] = "new",
112 [DIFF_COMMIT] = "commit",
113 [DIFF_WHITESPACE] = "whitespace",
114 [DIFF_FUNCINFO] = "func",
115 [DIFF_FILE_OLD_MOVED] = "oldMoved",
116 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
117 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
118 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
119 [DIFF_FILE_NEW_MOVED] = "newMoved",
120 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
121 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
122 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
123 [DIFF_CONTEXT_DIM] = "contextDimmed",
124 [DIFF_FILE_OLD_DIM] = "oldDimmed",
125 [DIFF_FILE_NEW_DIM] = "newDimmed",
126 [DIFF_CONTEXT_BOLD] = "contextBold",
127 [DIFF_FILE_OLD_BOLD] = "oldBold",
128 [DIFF_FILE_NEW_BOLD] = "newBold",
129 };
130
131 define_list_config_array_extra(color_diff_slots, {"plain"});
132
133 static int parse_diff_color_slot(const char *var)
134 {
135 if (!strcasecmp(var, "plain"))
136 return DIFF_CONTEXT;
137 return LOOKUP_CONFIG(color_diff_slots, var);
138 }
139
140 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
141 struct strbuf *errmsg)
142 {
143 char *params_copy = xstrdup(params_string);
144 struct string_list params = STRING_LIST_INIT_NODUP;
145 int ret = 0;
146 int i;
147
148 if (*params_copy)
149 string_list_split_in_place(&params, params_copy, ",", -1);
150 for (i = 0; i < params.nr; i++) {
151 const char *p = params.items[i].string;
152 if (!strcmp(p, "changes")) {
153 options->flags.dirstat_by_line = 0;
154 options->flags.dirstat_by_file = 0;
155 } else if (!strcmp(p, "lines")) {
156 options->flags.dirstat_by_line = 1;
157 options->flags.dirstat_by_file = 0;
158 } else if (!strcmp(p, "files")) {
159 options->flags.dirstat_by_line = 0;
160 options->flags.dirstat_by_file = 1;
161 } else if (!strcmp(p, "noncumulative")) {
162 options->flags.dirstat_cumulative = 0;
163 } else if (!strcmp(p, "cumulative")) {
164 options->flags.dirstat_cumulative = 1;
165 } else if (isdigit(*p)) {
166 char *end;
167 int permille = strtoul(p, &end, 10) * 10;
168 if (*end == '.' && isdigit(*++end)) {
169 /* only use first digit */
170 permille += *end - '0';
171 /* .. and ignore any further digits */
172 while (isdigit(*++end))
173 ; /* nothing */
174 }
175 if (!*end)
176 options->dirstat_permille = permille;
177 else {
178 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
179 p);
180 ret++;
181 }
182 } else {
183 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
184 ret++;
185 }
186
187 }
188 string_list_clear(&params, 0);
189 free(params_copy);
190 return ret;
191 }
192
193 static int parse_submodule_params(struct diff_options *options, const char *value)
194 {
195 if (!strcmp(value, "log"))
196 options->submodule_format = DIFF_SUBMODULE_LOG;
197 else if (!strcmp(value, "short"))
198 options->submodule_format = DIFF_SUBMODULE_SHORT;
199 else if (!strcmp(value, "diff"))
200 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
201 /*
202 * Please update $__git_diff_submodule_formats in
203 * git-completion.bash when you add new formats.
204 */
205 else
206 return -1;
207 return 0;
208 }
209
210 int git_config_rename(const char *var, const char *value)
211 {
212 if (!value)
213 return DIFF_DETECT_RENAME;
214 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
215 return DIFF_DETECT_COPY;
216 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
217 }
218
219 long parse_algorithm_value(const char *value)
220 {
221 if (!value)
222 return -1;
223 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
224 return 0;
225 else if (!strcasecmp(value, "minimal"))
226 return XDF_NEED_MINIMAL;
227 else if (!strcasecmp(value, "patience"))
228 return XDF_PATIENCE_DIFF;
229 else if (!strcasecmp(value, "histogram"))
230 return XDF_HISTOGRAM_DIFF;
231 /*
232 * Please update $__git_diff_algorithms in git-completion.bash
233 * when you add new algorithms.
234 */
235 return -1;
236 }
237
238 static int parse_one_token(const char **arg, const char *token)
239 {
240 const char *rest;
241 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
242 *arg = rest;
243 return 1;
244 }
245 return 0;
246 }
247
248 static int parse_ws_error_highlight(const char *arg)
249 {
250 const char *orig_arg = arg;
251 unsigned val = 0;
252
253 while (*arg) {
254 if (parse_one_token(&arg, "none"))
255 val = 0;
256 else if (parse_one_token(&arg, "default"))
257 val = WSEH_NEW;
258 else if (parse_one_token(&arg, "all"))
259 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
260 else if (parse_one_token(&arg, "new"))
261 val |= WSEH_NEW;
262 else if (parse_one_token(&arg, "old"))
263 val |= WSEH_OLD;
264 else if (parse_one_token(&arg, "context"))
265 val |= WSEH_CONTEXT;
266 else {
267 return -1 - (int)(arg - orig_arg);
268 }
269 if (*arg)
270 arg++;
271 }
272 return val;
273 }
274
275 /*
276 * These are to give UI layer defaults.
277 * The core-level commands such as git-diff-files should
278 * never be affected by the setting of diff.renames
279 * the user happens to have in the configuration file.
280 */
281 void init_diff_ui_defaults(void)
282 {
283 diff_detect_rename_default = DIFF_DETECT_RENAME;
284 }
285
286 int git_diff_heuristic_config(const char *var, const char *value,
287 void *cb UNUSED)
288 {
289 if (!strcmp(var, "diff.indentheuristic"))
290 diff_indent_heuristic = git_config_bool(var, value);
291 return 0;
292 }
293
294 static int parse_color_moved(const char *arg)
295 {
296 switch (git_parse_maybe_bool(arg)) {
297 case 0:
298 return COLOR_MOVED_NO;
299 case 1:
300 return COLOR_MOVED_DEFAULT;
301 default:
302 break;
303 }
304
305 if (!strcmp(arg, "no"))
306 return COLOR_MOVED_NO;
307 else if (!strcmp(arg, "plain"))
308 return COLOR_MOVED_PLAIN;
309 else if (!strcmp(arg, "blocks"))
310 return COLOR_MOVED_BLOCKS;
311 else if (!strcmp(arg, "zebra"))
312 return COLOR_MOVED_ZEBRA;
313 else if (!strcmp(arg, "default"))
314 return COLOR_MOVED_DEFAULT;
315 else if (!strcmp(arg, "dimmed-zebra"))
316 return COLOR_MOVED_ZEBRA_DIM;
317 else if (!strcmp(arg, "dimmed_zebra"))
318 return COLOR_MOVED_ZEBRA_DIM;
319 else
320 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
321 }
322
323 static unsigned parse_color_moved_ws(const char *arg)
324 {
325 int ret = 0;
326 struct string_list l = STRING_LIST_INIT_DUP;
327 struct string_list_item *i;
328
329 string_list_split(&l, arg, ',', -1);
330
331 for_each_string_list_item(i, &l) {
332 struct strbuf sb = STRBUF_INIT;
333 strbuf_addstr(&sb, i->string);
334 strbuf_trim(&sb);
335
336 if (!strcmp(sb.buf, "no"))
337 ret = 0;
338 else if (!strcmp(sb.buf, "ignore-space-change"))
339 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
340 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
341 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
342 else if (!strcmp(sb.buf, "ignore-all-space"))
343 ret |= XDF_IGNORE_WHITESPACE;
344 else if (!strcmp(sb.buf, "allow-indentation-change"))
345 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
346 else {
347 ret |= COLOR_MOVED_WS_ERROR;
348 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);
349 }
350
351 strbuf_release(&sb);
352 }
353
354 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
355 (ret & XDF_WHITESPACE_FLAGS)) {
356 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
357 ret |= COLOR_MOVED_WS_ERROR;
358 }
359
360 string_list_clear(&l, 0);
361
362 return ret;
363 }
364
365 int git_diff_ui_config(const char *var, const char *value,
366 const struct config_context *ctx, void *cb)
367 {
368 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
369 diff_use_color_default = git_config_colorbool(var, value);
370 return 0;
371 }
372 if (!strcmp(var, "diff.colormoved")) {
373 int cm = parse_color_moved(value);
374 if (cm < 0)
375 return -1;
376 diff_color_moved_default = cm;
377 return 0;
378 }
379 if (!strcmp(var, "diff.colormovedws")) {
380 unsigned cm;
381 if (!value)
382 return config_error_nonbool(var);
383 cm = parse_color_moved_ws(value);
384 if (cm & COLOR_MOVED_WS_ERROR)
385 return -1;
386 diff_color_moved_ws_default = cm;
387 return 0;
388 }
389 if (!strcmp(var, "diff.context")) {
390 diff_context_default = git_config_int(var, value, ctx->kvi);
391 if (diff_context_default < 0)
392 return -1;
393 return 0;
394 }
395 if (!strcmp(var, "diff.interhunkcontext")) {
396 diff_interhunk_context_default = git_config_int(var, value,
397 ctx->kvi);
398 if (diff_interhunk_context_default < 0)
399 return -1;
400 return 0;
401 }
402 if (!strcmp(var, "diff.renames")) {
403 diff_detect_rename_default = git_config_rename(var, value);
404 return 0;
405 }
406 if (!strcmp(var, "diff.autorefreshindex")) {
407 diff_auto_refresh_index = git_config_bool(var, value);
408 return 0;
409 }
410 if (!strcmp(var, "diff.mnemonicprefix")) {
411 diff_mnemonic_prefix = git_config_bool(var, value);
412 return 0;
413 }
414 if (!strcmp(var, "diff.noprefix")) {
415 diff_no_prefix = git_config_bool(var, value);
416 return 0;
417 }
418 if (!strcmp(var, "diff.srcprefix")) {
419 FREE_AND_NULL(diff_src_prefix);
420 return git_config_string(&diff_src_prefix, var, value);
421 }
422 if (!strcmp(var, "diff.dstprefix")) {
423 FREE_AND_NULL(diff_dst_prefix);
424 return git_config_string(&diff_dst_prefix, var, value);
425 }
426 if (!strcmp(var, "diff.relative")) {
427 diff_relative = git_config_bool(var, value);
428 return 0;
429 }
430 if (!strcmp(var, "diff.statnamewidth")) {
431 diff_stat_name_width = git_config_int(var, value, ctx->kvi);
432 return 0;
433 }
434 if (!strcmp(var, "diff.statgraphwidth")) {
435 diff_stat_graph_width = git_config_int(var, value, ctx->kvi);
436 return 0;
437 }
438 if (!strcmp(var, "diff.external"))
439 return git_config_string(&external_diff_cfg.cmd, var, value);
440 if (!strcmp(var, "diff.trustexitcode")) {
441 external_diff_cfg.trust_exit_code = git_config_bool(var, value);
442 return 0;
443 }
444 if (!strcmp(var, "diff.wordregex"))
445 return git_config_string(&diff_word_regex_cfg, var, value);
446 if (!strcmp(var, "diff.orderfile"))
447 return git_config_pathname(&diff_order_file_cfg, var, value);
448
449 if (!strcmp(var, "diff.ignoresubmodules")) {
450 if (!value)
451 return config_error_nonbool(var);
452 handle_ignore_submodules_arg(&default_diff_options, value);
453 }
454
455 if (!strcmp(var, "diff.submodule")) {
456 if (!value)
457 return config_error_nonbool(var);
458 if (parse_submodule_params(&default_diff_options, value))
459 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
460 value);
461 return 0;
462 }
463
464 if (!strcmp(var, "diff.algorithm")) {
465 if (!value)
466 return config_error_nonbool(var);
467 diff_algorithm = parse_algorithm_value(value);
468 if (diff_algorithm < 0)
469 return error(_("unknown value for config '%s': %s"),
470 var, value);
471 return 0;
472 }
473
474 if (git_color_config(var, value, cb) < 0)
475 return -1;
476
477 return git_diff_basic_config(var, value, ctx, cb);
478 }
479
480 int git_diff_basic_config(const char *var, const char *value,
481 const struct config_context *ctx, void *cb)
482 {
483 const char *name;
484
485 if (!strcmp(var, "diff.renamelimit")) {
486 diff_rename_limit_default = git_config_int(var, value, ctx->kvi);
487 return 0;
488 }
489
490 if (userdiff_config(var, value) < 0)
491 return -1;
492
493 if (skip_prefix(var, "diff.color.", &name) ||
494 skip_prefix(var, "color.diff.", &name)) {
495 int slot = parse_diff_color_slot(name);
496 if (slot < 0)
497 return 0;
498 if (!value)
499 return config_error_nonbool(var);
500 return color_parse(value, diff_colors[slot]);
501 }
502
503 if (!strcmp(var, "diff.wserrorhighlight")) {
504 int val;
505 if (!value)
506 return config_error_nonbool(var);
507 val = parse_ws_error_highlight(value);
508 if (val < 0)
509 return error(_("unknown value for config '%s': %s"),
510 var, value);
511 ws_error_highlight_default = val;
512 return 0;
513 }
514
515 /* like GNU diff's --suppress-blank-empty option */
516 if (!strcmp(var, "diff.suppressblankempty") ||
517 /* for backwards compatibility */
518 !strcmp(var, "diff.suppress-blank-empty")) {
519 diff_suppress_blank_empty = git_config_bool(var, value);
520 return 0;
521 }
522
523 if (!strcmp(var, "diff.dirstat")) {
524 struct strbuf errmsg = STRBUF_INIT;
525 if (!value)
526 return config_error_nonbool(var);
527 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
528 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
529 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
530 errmsg.buf);
531 strbuf_release(&errmsg);
532 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
533 return 0;
534 }
535
536 if (git_diff_heuristic_config(var, value, cb) < 0)
537 return -1;
538
539 return git_default_config(var, value, ctx, cb);
540 }
541
542 static char *quote_two(const char *one, const char *two)
543 {
544 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
545 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
546 struct strbuf res = STRBUF_INIT;
547
548 if (need_one + need_two) {
549 strbuf_addch(&res, '"');
550 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
551 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
552 strbuf_addch(&res, '"');
553 } else {
554 strbuf_addstr(&res, one);
555 strbuf_addstr(&res, two);
556 }
557 return strbuf_detach(&res, NULL);
558 }
559
560 static const struct external_diff *external_diff(void)
561 {
562 static struct external_diff external_diff_env, *external_diff_ptr;
563 static int done_preparing = 0;
564
565 if (done_preparing)
566 return external_diff_ptr;
567 external_diff_env.cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
568 if (git_env_bool("GIT_EXTERNAL_DIFF_TRUST_EXIT_CODE", 0))
569 external_diff_env.trust_exit_code = 1;
570 if (external_diff_env.cmd)
571 external_diff_ptr = &external_diff_env;
572 else if (external_diff_cfg.cmd)
573 external_diff_ptr = &external_diff_cfg;
574 done_preparing = 1;
575 return external_diff_ptr;
576 }
577
578 /*
579 * Keep track of files used for diffing. Sometimes such an entry
580 * refers to a temporary file, sometimes to an existing file, and
581 * sometimes to "/dev/null".
582 */
583 static struct diff_tempfile {
584 /*
585 * filename external diff should read from, or NULL if this
586 * entry is currently not in use:
587 */
588 const char *name;
589
590 char hex[GIT_MAX_HEXSZ + 1];
591 char mode[10];
592
593 /*
594 * If this diff_tempfile instance refers to a temporary file,
595 * this tempfile object is used to manage its lifetime.
596 */
597 struct tempfile *tempfile;
598 } diff_temp[2];
599
600 struct emit_callback {
601 int color_diff;
602 unsigned ws_rule;
603 int blank_at_eof_in_preimage;
604 int blank_at_eof_in_postimage;
605 int lno_in_preimage;
606 int lno_in_postimage;
607 const char **label_path;
608 struct diff_words_data *diff_words;
609 struct diff_options *opt;
610 struct strbuf *header;
611 };
612
613 static int count_lines(const char *data, int size)
614 {
615 int count, ch, completely_empty = 1, nl_just_seen = 0;
616 count = 0;
617 while (0 < size--) {
618 ch = *data++;
619 if (ch == '\n') {
620 count++;
621 nl_just_seen = 1;
622 completely_empty = 0;
623 }
624 else {
625 nl_just_seen = 0;
626 completely_empty = 0;
627 }
628 }
629 if (completely_empty)
630 return 0;
631 if (!nl_just_seen)
632 count++; /* no trailing newline */
633 return count;
634 }
635
636 static int fill_mmfile(struct repository *r, mmfile_t *mf,
637 struct diff_filespec *one)
638 {
639 if (!DIFF_FILE_VALID(one)) {
640 mf->ptr = (char *)""; /* does not matter */
641 mf->size = 0;
642 return 0;
643 }
644 else if (diff_populate_filespec(r, one, NULL))
645 return -1;
646
647 mf->ptr = one->data;
648 mf->size = one->size;
649 return 0;
650 }
651
652 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
653 static unsigned long diff_filespec_size(struct repository *r,
654 struct diff_filespec *one)
655 {
656 struct diff_populate_filespec_options dpf_options = {
657 .check_size_only = 1,
658 };
659
660 if (!DIFF_FILE_VALID(one))
661 return 0;
662 diff_populate_filespec(r, one, &dpf_options);
663 return one->size;
664 }
665
666 static int count_trailing_blank(mmfile_t *mf)
667 {
668 char *ptr = mf->ptr;
669 long size = mf->size;
670 int cnt = 0;
671
672 if (!size)
673 return cnt;
674 ptr += size - 1; /* pointing at the very end */
675 if (*ptr != '\n')
676 ; /* incomplete line */
677 else
678 ptr--; /* skip the last LF */
679 while (mf->ptr < ptr) {
680 char *prev_eol;
681 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
682 if (*prev_eol == '\n')
683 break;
684 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
685 break;
686 cnt++;
687 ptr = prev_eol - 1;
688 }
689 return cnt;
690 }
691
692 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
693 struct emit_callback *ecbdata)
694 {
695 int l1, l2, at;
696 l1 = count_trailing_blank(mf1);
697 l2 = count_trailing_blank(mf2);
698 if (l2 <= l1) {
699 ecbdata->blank_at_eof_in_preimage = 0;
700 ecbdata->blank_at_eof_in_postimage = 0;
701 return;
702 }
703 at = count_lines(mf1->ptr, mf1->size);
704 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
705
706 at = count_lines(mf2->ptr, mf2->size);
707 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
708 }
709
710 static void emit_line_0(struct diff_options *o,
711 const char *set_sign, const char *set, unsigned reverse, const char *reset,
712 int first, const char *line, int len)
713 {
714 int has_trailing_newline, has_trailing_carriage_return;
715 int needs_reset = 0; /* at the end of the line */
716 FILE *file = o->file;
717
718 fputs(diff_line_prefix(o), file);
719
720 has_trailing_newline = (len > 0 && line[len-1] == '\n');
721 if (has_trailing_newline)
722 len--;
723
724 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
725 if (has_trailing_carriage_return)
726 len--;
727
728 if (!len && !first)
729 goto end_of_line;
730
731 if (reverse && want_color(o->use_color)) {
732 fputs(GIT_COLOR_REVERSE, file);
733 needs_reset = 1;
734 }
735
736 if (set_sign) {
737 fputs(set_sign, file);
738 needs_reset = 1;
739 }
740
741 if (first)
742 fputc(first, file);
743
744 if (!len)
745 goto end_of_line;
746
747 if (set) {
748 if (set_sign && set != set_sign)
749 fputs(reset, file);
750 fputs(set, file);
751 needs_reset = 1;
752 }
753 fwrite(line, len, 1, file);
754 needs_reset = 1; /* 'line' may contain color codes. */
755
756 end_of_line:
757 if (needs_reset)
758 fputs(reset, file);
759 if (has_trailing_carriage_return)
760 fputc('\r', file);
761 if (has_trailing_newline)
762 fputc('\n', file);
763 }
764
765 static void emit_line(struct diff_options *o, const char *set, const char *reset,
766 const char *line, int len)
767 {
768 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
769 }
770
771 enum diff_symbol {
772 DIFF_SYMBOL_BINARY_DIFF_HEADER,
773 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
774 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
775 DIFF_SYMBOL_BINARY_DIFF_BODY,
776 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
777 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
778 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
779 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
780 DIFF_SYMBOL_STATS_LINE,
781 DIFF_SYMBOL_WORD_DIFF,
782 DIFF_SYMBOL_STAT_SEP,
783 DIFF_SYMBOL_SUMMARY,
784 DIFF_SYMBOL_SUBMODULE_ADD,
785 DIFF_SYMBOL_SUBMODULE_DEL,
786 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
787 DIFF_SYMBOL_SUBMODULE_MODIFIED,
788 DIFF_SYMBOL_SUBMODULE_HEADER,
789 DIFF_SYMBOL_SUBMODULE_ERROR,
790 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
791 DIFF_SYMBOL_REWRITE_DIFF,
792 DIFF_SYMBOL_BINARY_FILES,
793 DIFF_SYMBOL_HEADER,
794 DIFF_SYMBOL_FILEPAIR_PLUS,
795 DIFF_SYMBOL_FILEPAIR_MINUS,
796 DIFF_SYMBOL_WORDS_PORCELAIN,
797 DIFF_SYMBOL_WORDS,
798 DIFF_SYMBOL_CONTEXT,
799 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
800 DIFF_SYMBOL_PLUS,
801 DIFF_SYMBOL_MINUS,
802 DIFF_SYMBOL_NO_LF_EOF,
803 DIFF_SYMBOL_CONTEXT_FRAGINFO,
804 DIFF_SYMBOL_CONTEXT_MARKER,
805 DIFF_SYMBOL_SEPARATOR
806 };
807 /*
808 * Flags for content lines:
809 * 0..12 are whitespace rules
810 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
811 * 16 is marking if the line is blank at EOF
812 */
813 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
814 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
815 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
816 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
817 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
818
819 /*
820 * This struct is used when we need to buffer the output of the diff output.
821 *
822 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
823 * into the pre/post image file. This pointer could be a union with the
824 * line pointer. By storing an offset into the file instead of the literal line,
825 * we can decrease the memory footprint for the buffered output. At first we
826 * may want to only have indirection for the content lines, but we could also
827 * enhance the state for emitting prefabricated lines, e.g. the similarity
828 * score line or hunk/file headers would only need to store a number or path
829 * and then the output can be constructed later on depending on state.
830 */
831 struct emitted_diff_symbol {
832 const char *line;
833 int len;
834 int flags;
835 int indent_off; /* Offset to first non-whitespace character */
836 int indent_width; /* The visual width of the indentation */
837 unsigned id;
838 enum diff_symbol s;
839 };
840 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
841
842 struct emitted_diff_symbols {
843 struct emitted_diff_symbol *buf;
844 int nr, alloc;
845 };
846 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
847
848 static void append_emitted_diff_symbol(struct diff_options *o,
849 struct emitted_diff_symbol *e)
850 {
851 struct emitted_diff_symbol *f;
852
853 ALLOC_GROW(o->emitted_symbols->buf,
854 o->emitted_symbols->nr + 1,
855 o->emitted_symbols->alloc);
856 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
857
858 memcpy(f, e, sizeof(struct emitted_diff_symbol));
859 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
860 }
861
862 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
863 {
864 if (!e)
865 return;
866 free(e->buf);
867 free(e);
868 }
869
870 struct moved_entry {
871 const struct emitted_diff_symbol *es;
872 struct moved_entry *next_line;
873 struct moved_entry *next_match;
874 };
875
876 struct moved_block {
877 struct moved_entry *match;
878 int wsd; /* The whitespace delta of this block */
879 };
880
881 #define INDENT_BLANKLINE INT_MIN
882
883 static void fill_es_indent_data(struct emitted_diff_symbol *es)
884 {
885 unsigned int off = 0, i;
886 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
887 const char *s = es->line;
888 const int len = es->len;
889
890 /* skip any \v \f \r at start of indentation */
891 while (s[off] == '\f' || s[off] == '\v' ||
892 (s[off] == '\r' && off < len - 1))
893 off++;
894
895 /* calculate the visual width of indentation */
896 while(1) {
897 if (s[off] == ' ') {
898 width++;
899 off++;
900 } else if (s[off] == '\t') {
901 width += tab_width - (width % tab_width);
902 while (s[++off] == '\t')
903 width += tab_width;
904 } else {
905 break;
906 }
907 }
908
909 /* check if this line is blank */
910 for (i = off; i < len; i++)
911 if (!isspace(s[i]))
912 break;
913
914 if (i == len) {
915 es->indent_width = INDENT_BLANKLINE;
916 es->indent_off = len;
917 } else {
918 es->indent_off = off;
919 es->indent_width = width;
920 }
921 }
922
923 static int compute_ws_delta(const struct emitted_diff_symbol *a,
924 const struct emitted_diff_symbol *b)
925 {
926 int a_width = a->indent_width,
927 b_width = b->indent_width;
928
929 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
930 return INDENT_BLANKLINE;
931
932 return a_width - b_width;
933 }
934
935 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
936 const struct emitted_diff_symbol *l,
937 struct moved_block *pmb)
938 {
939 int a_width = cur->es->indent_width, b_width = l->indent_width;
940 int delta;
941
942 /* The text of each line must match */
943 if (cur->es->id != l->id)
944 return 1;
945
946 /*
947 * If 'l' and 'cur' are both blank then we don't need to check the
948 * indent. We only need to check cur as we know the strings match.
949 * */
950 if (a_width == INDENT_BLANKLINE)
951 return 0;
952
953 /*
954 * The indent changes of the block are known and stored in pmb->wsd;
955 * however we need to check if the indent changes of the current line
956 * match those of the current block.
957 */
958 delta = b_width - a_width;
959
960 /*
961 * If the previous lines of this block were all blank then set its
962 * whitespace delta.
963 */
964 if (pmb->wsd == INDENT_BLANKLINE)
965 pmb->wsd = delta;
966
967 return delta != pmb->wsd;
968 }
969
970 struct interned_diff_symbol {
971 struct hashmap_entry ent;
972 struct emitted_diff_symbol *es;
973 };
974
975 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
976 const struct hashmap_entry *eptr,
977 const struct hashmap_entry *entry_or_key,
978 const void *keydata UNUSED)
979 {
980 const struct diff_options *diffopt = hashmap_cmp_fn_data;
981 const struct emitted_diff_symbol *a, *b;
982 unsigned flags = diffopt->color_moved_ws_handling
983 & XDF_WHITESPACE_FLAGS;
984
985 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
986 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
987
988 return !xdiff_compare_lines(a->line + a->indent_off,
989 a->len - a->indent_off,
990 b->line + b->indent_off,
991 b->len - b->indent_off, flags);
992 }
993
994 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
995 struct interned_diff_symbol *s)
996 {
997 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
998 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
999 l->len - l->indent_off, flags);
1000
1001 hashmap_entry_init(&s->ent, hash);
1002 s->es = l;
1003 }
1004
1005 struct moved_entry_list {
1006 struct moved_entry *add, *del;
1007 };
1008
1009 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
1010 struct mem_pool *entry_mem_pool)
1011 {
1012 struct moved_entry *prev_line = NULL;
1013 struct mem_pool interned_pool;
1014 struct hashmap interned_map;
1015 struct moved_entry_list *entry_list = NULL;
1016 size_t entry_list_alloc = 0;
1017 unsigned id = 0;
1018 int n;
1019
1020 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
1021 mem_pool_init(&interned_pool, 1024 * 1024);
1022
1023 for (n = 0; n < o->emitted_symbols->nr; n++) {
1024 struct interned_diff_symbol key;
1025 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1026 struct interned_diff_symbol *s;
1027 struct moved_entry *entry;
1028
1029 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
1030 prev_line = NULL;
1031 continue;
1032 }
1033
1034 if (o->color_moved_ws_handling &
1035 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1036 fill_es_indent_data(l);
1037
1038 prepare_entry(o, l, &key);
1039 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
1040 if (s) {
1041 l->id = s->es->id;
1042 } else {
1043 l->id = id;
1044 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
1045 hashmap_add(&interned_map,
1046 memcpy(mem_pool_alloc(&interned_pool,
1047 sizeof(key)),
1048 &key, sizeof(key)));
1049 }
1050 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1051 entry->es = l;
1052 entry->next_line = NULL;
1053 if (prev_line && prev_line->es->s == l->s)
1054 prev_line->next_line = entry;
1055 prev_line = entry;
1056 if (l->s == DIFF_SYMBOL_PLUS) {
1057 entry->next_match = entry_list[l->id].add;
1058 entry_list[l->id].add = entry;
1059 } else {
1060 entry->next_match = entry_list[l->id].del;
1061 entry_list[l->id].del = entry;
1062 }
1063 }
1064
1065 hashmap_clear(&interned_map);
1066 mem_pool_discard(&interned_pool, 0);
1067
1068 return entry_list;
1069 }
1070
1071 static void pmb_advance_or_null(struct diff_options *o,
1072 struct emitted_diff_symbol *l,
1073 struct moved_block *pmb,
1074 int *pmb_nr)
1075 {
1076 int i, j;
1077
1078 for (i = 0, j = 0; i < *pmb_nr; i++) {
1079 int match;
1080 struct moved_entry *prev = pmb[i].match;
1081 struct moved_entry *cur = (prev && prev->next_line) ?
1082 prev->next_line : NULL;
1083
1084 if (o->color_moved_ws_handling &
1085 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1086 match = cur &&
1087 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1088 else
1089 match = cur && cur->es->id == l->id;
1090
1091 if (match) {
1092 pmb[j] = pmb[i];
1093 pmb[j++].match = cur;
1094 }
1095 }
1096 *pmb_nr = j;
1097 }
1098
1099 static void fill_potential_moved_blocks(struct diff_options *o,
1100 struct moved_entry *match,
1101 struct emitted_diff_symbol *l,
1102 struct moved_block **pmb_p,
1103 int *pmb_alloc_p, int *pmb_nr_p)
1104
1105 {
1106 struct moved_block *pmb = *pmb_p;
1107 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1108
1109 /*
1110 * The current line is the start of a new block.
1111 * Setup the set of potential blocks.
1112 */
1113 for (; match; match = match->next_match) {
1114 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1115 if (o->color_moved_ws_handling &
1116 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1117 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1118 else
1119 pmb[pmb_nr].wsd = 0;
1120 pmb[pmb_nr++].match = match;
1121 }
1122
1123 *pmb_p = pmb;
1124 *pmb_alloc_p = pmb_alloc;
1125 *pmb_nr_p = pmb_nr;
1126 }
1127
1128 /*
1129 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1130 *
1131 * Otherwise, if the last block has fewer alphanumeric characters than
1132 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1133 * that block.
1134 *
1135 * The last block consists of the (n - block_length)'th line up to but not
1136 * including the nth line.
1137 *
1138 * Returns 0 if the last block is empty or is unset by this function, non zero
1139 * otherwise.
1140 *
1141 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1142 * Think of a way to unify them.
1143 */
1144 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1145 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1146 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1147 {
1148 int i, alnum_count = 0;
1149 if (o->color_moved == COLOR_MOVED_PLAIN)
1150 return block_length;
1151 for (i = 1; i < block_length + 1; i++) {
1152 const char *c = o->emitted_symbols->buf[n - i].line;
1153 for (; *c; c++) {
1154 if (!isalnum(*c))
1155 continue;
1156 alnum_count++;
1157 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1158 return 1;
1159 }
1160 }
1161 for (i = 1; i < block_length + 1; i++)
1162 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1163 return 0;
1164 }
1165
1166 /* Find blocks of moved code, delegate actual coloring decision to helper */
1167 static void mark_color_as_moved(struct diff_options *o,
1168 struct moved_entry_list *entry_list)
1169 {
1170 struct moved_block *pmb = NULL; /* potentially moved blocks */
1171 int pmb_nr = 0, pmb_alloc = 0;
1172 int n, flipped_block = 0, block_length = 0;
1173 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1174
1175
1176 for (n = 0; n < o->emitted_symbols->nr; n++) {
1177 struct moved_entry *match = NULL;
1178 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1179
1180 switch (l->s) {
1181 case DIFF_SYMBOL_PLUS:
1182 match = entry_list[l->id].del;
1183 break;
1184 case DIFF_SYMBOL_MINUS:
1185 match = entry_list[l->id].add;
1186 break;
1187 default:
1188 flipped_block = 0;
1189 }
1190
1191 if (pmb_nr && (!match || l->s != moved_symbol)) {
1192 if (!adjust_last_block(o, n, block_length) &&
1193 block_length > 1) {
1194 /*
1195 * Rewind in case there is another match
1196 * starting at the second line of the block
1197 */
1198 match = NULL;
1199 n -= block_length;
1200 }
1201 pmb_nr = 0;
1202 block_length = 0;
1203 flipped_block = 0;
1204 }
1205 if (!match) {
1206 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1207 continue;
1208 }
1209
1210 if (o->color_moved == COLOR_MOVED_PLAIN) {
1211 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1212 continue;
1213 }
1214
1215 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1216
1217 if (pmb_nr == 0) {
1218 int contiguous = adjust_last_block(o, n, block_length);
1219
1220 if (!contiguous && block_length > 1)
1221 /*
1222 * Rewind in case there is another match
1223 * starting at the second line of the block
1224 */
1225 n -= block_length;
1226 else
1227 fill_potential_moved_blocks(o, match, l,
1228 &pmb, &pmb_alloc,
1229 &pmb_nr);
1230
1231 if (contiguous && pmb_nr && moved_symbol == l->s)
1232 flipped_block = (flipped_block + 1) % 2;
1233 else
1234 flipped_block = 0;
1235
1236 if (pmb_nr)
1237 moved_symbol = l->s;
1238 else
1239 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1240
1241 block_length = 0;
1242 }
1243
1244 if (pmb_nr) {
1245 block_length++;
1246 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1247 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1248 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1249 }
1250 }
1251 adjust_last_block(o, n, block_length);
1252
1253 free(pmb);
1254 }
1255
1256 static void dim_moved_lines(struct diff_options *o)
1257 {
1258 int n;
1259 for (n = 0; n < o->emitted_symbols->nr; n++) {
1260 struct emitted_diff_symbol *prev = (n != 0) ?
1261 &o->emitted_symbols->buf[n - 1] : NULL;
1262 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1263 struct emitted_diff_symbol *next =
1264 (n < o->emitted_symbols->nr - 1) ?
1265 &o->emitted_symbols->buf[n + 1] : NULL;
1266
1267 /* Not a plus or minus line? */
1268 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1269 continue;
1270
1271 /* Not a moved line? */
1272 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1273 continue;
1274
1275 /*
1276 * If prev or next are not a plus or minus line,
1277 * pretend they don't exist
1278 */
1279 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1280 prev->s != DIFF_SYMBOL_MINUS)
1281 prev = NULL;
1282 if (next && next->s != DIFF_SYMBOL_PLUS &&
1283 next->s != DIFF_SYMBOL_MINUS)
1284 next = NULL;
1285
1286 /* Inside a block? */
1287 if ((prev &&
1288 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1289 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1290 (next &&
1291 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1292 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1293 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1294 continue;
1295 }
1296
1297 /* Check if we are at an interesting bound: */
1298 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1299 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1300 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1301 continue;
1302 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1303 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1304 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1305 continue;
1306
1307 /*
1308 * The boundary to prev and next are not interesting,
1309 * so this line is not interesting as a whole
1310 */
1311 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1312 }
1313 }
1314
1315 static void emit_line_ws_markup(struct diff_options *o,
1316 const char *set_sign, const char *set,
1317 const char *reset,
1318 int sign_index, const char *line, int len,
1319 unsigned ws_rule, int blank_at_eof)
1320 {
1321 const char *ws = NULL;
1322 int sign = o->output_indicators[sign_index];
1323
1324 if (o->ws_error_highlight & ws_rule) {
1325 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1326 if (!*ws)
1327 ws = NULL;
1328 }
1329
1330 if (!ws && !set_sign)
1331 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1332 else if (!ws) {
1333 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1334 } else if (blank_at_eof)
1335 /* Blank line at EOF - paint '+' as well */
1336 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1337 else {
1338 /* Emit just the prefix, then the rest. */
1339 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1340 sign, "", 0);
1341 ws_check_emit(line, len, ws_rule,
1342 o->file, set, reset, ws);
1343 }
1344 }
1345
1346 static void emit_diff_symbol_from_struct(struct diff_options *o,
1347 struct emitted_diff_symbol *eds)
1348 {
1349 static const char *nneof = " No newline at end of file\n";
1350 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1351
1352 enum diff_symbol s = eds->s;
1353 const char *line = eds->line;
1354 int len = eds->len;
1355 unsigned flags = eds->flags;
1356
1357 switch (s) {
1358 case DIFF_SYMBOL_NO_LF_EOF:
1359 context = diff_get_color_opt(o, DIFF_CONTEXT);
1360 reset = diff_get_color_opt(o, DIFF_RESET);
1361 putc('\n', o->file);
1362 emit_line_0(o, context, NULL, 0, reset, '\\',
1363 nneof, strlen(nneof));
1364 break;
1365 case DIFF_SYMBOL_SUBMODULE_HEADER:
1366 case DIFF_SYMBOL_SUBMODULE_ERROR:
1367 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1368 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1369 case DIFF_SYMBOL_SUMMARY:
1370 case DIFF_SYMBOL_STATS_LINE:
1371 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1372 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1373 emit_line(o, "", "", line, len);
1374 break;
1375 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1376 case DIFF_SYMBOL_CONTEXT_MARKER:
1377 context = diff_get_color_opt(o, DIFF_CONTEXT);
1378 reset = diff_get_color_opt(o, DIFF_RESET);
1379 emit_line(o, context, reset, line, len);
1380 break;
1381 case DIFF_SYMBOL_SEPARATOR:
1382 fprintf(o->file, "%s%c",
1383 diff_line_prefix(o),
1384 o->line_termination);
1385 break;
1386 case DIFF_SYMBOL_CONTEXT:
1387 set = diff_get_color_opt(o, DIFF_CONTEXT);
1388 reset = diff_get_color_opt(o, DIFF_RESET);
1389 set_sign = NULL;
1390 if (o->flags.dual_color_diffed_diffs) {
1391 char c = !len ? 0 : line[0];
1392
1393 if (c == '+')
1394 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1395 else if (c == '@')
1396 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1397 else if (c == '-')
1398 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1399 }
1400 emit_line_ws_markup(o, set_sign, set, reset,
1401 OUTPUT_INDICATOR_CONTEXT, line, len,
1402 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1403 break;
1404 case DIFF_SYMBOL_PLUS:
1405 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1406 DIFF_SYMBOL_MOVED_LINE_ALT |
1407 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1408 case DIFF_SYMBOL_MOVED_LINE |
1409 DIFF_SYMBOL_MOVED_LINE_ALT |
1410 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1411 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1412 break;
1413 case DIFF_SYMBOL_MOVED_LINE |
1414 DIFF_SYMBOL_MOVED_LINE_ALT:
1415 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1416 break;
1417 case DIFF_SYMBOL_MOVED_LINE |
1418 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1419 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1420 break;
1421 case DIFF_SYMBOL_MOVED_LINE:
1422 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1423 break;
1424 default:
1425 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1426 }
1427 reset = diff_get_color_opt(o, DIFF_RESET);
1428 if (!o->flags.dual_color_diffed_diffs)
1429 set_sign = NULL;
1430 else {
1431 char c = !len ? 0 : line[0];
1432
1433 set_sign = set;
1434 if (c == '-')
1435 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1436 else if (c == '@')
1437 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1438 else if (c == '+')
1439 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1440 else
1441 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1442 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1443 }
1444 emit_line_ws_markup(o, set_sign, set, reset,
1445 OUTPUT_INDICATOR_NEW, line, len,
1446 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1447 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1448 break;
1449 case DIFF_SYMBOL_MINUS:
1450 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1451 DIFF_SYMBOL_MOVED_LINE_ALT |
1452 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1453 case DIFF_SYMBOL_MOVED_LINE |
1454 DIFF_SYMBOL_MOVED_LINE_ALT |
1455 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1456 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1457 break;
1458 case DIFF_SYMBOL_MOVED_LINE |
1459 DIFF_SYMBOL_MOVED_LINE_ALT:
1460 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1461 break;
1462 case DIFF_SYMBOL_MOVED_LINE |
1463 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1464 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1465 break;
1466 case DIFF_SYMBOL_MOVED_LINE:
1467 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1468 break;
1469 default:
1470 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1471 }
1472 reset = diff_get_color_opt(o, DIFF_RESET);
1473 if (!o->flags.dual_color_diffed_diffs)
1474 set_sign = NULL;
1475 else {
1476 char c = !len ? 0 : line[0];
1477
1478 set_sign = set;
1479 if (c == '+')
1480 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1481 else if (c == '@')
1482 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1483 else if (c == '-')
1484 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1485 else
1486 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1487 }
1488 emit_line_ws_markup(o, set_sign, set, reset,
1489 OUTPUT_INDICATOR_OLD, line, len,
1490 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1491 break;
1492 case DIFF_SYMBOL_WORDS_PORCELAIN:
1493 context = diff_get_color_opt(o, DIFF_CONTEXT);
1494 reset = diff_get_color_opt(o, DIFF_RESET);
1495 emit_line(o, context, reset, line, len);
1496 fputs("~\n", o->file);
1497 break;
1498 case DIFF_SYMBOL_WORDS:
1499 context = diff_get_color_opt(o, DIFF_CONTEXT);
1500 reset = diff_get_color_opt(o, DIFF_RESET);
1501 /*
1502 * Skip the prefix character, if any. With
1503 * diff_suppress_blank_empty, there may be
1504 * none.
1505 */
1506 if (line[0] != '\n') {
1507 line++;
1508 len--;
1509 }
1510 emit_line(o, context, reset, line, len);
1511 break;
1512 case DIFF_SYMBOL_FILEPAIR_PLUS:
1513 meta = diff_get_color_opt(o, DIFF_METAINFO);
1514 reset = diff_get_color_opt(o, DIFF_RESET);
1515 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1516 line, reset,
1517 strchr(line, ' ') ? "\t" : "");
1518 break;
1519 case DIFF_SYMBOL_FILEPAIR_MINUS:
1520 meta = diff_get_color_opt(o, DIFF_METAINFO);
1521 reset = diff_get_color_opt(o, DIFF_RESET);
1522 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1523 line, reset,
1524 strchr(line, ' ') ? "\t" : "");
1525 break;
1526 case DIFF_SYMBOL_BINARY_FILES:
1527 case DIFF_SYMBOL_HEADER:
1528 fprintf(o->file, "%s", line);
1529 break;
1530 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1531 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1532 break;
1533 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1534 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1535 break;
1536 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1537 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1538 break;
1539 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1540 fputs(diff_line_prefix(o), o->file);
1541 fputc('\n', o->file);
1542 break;
1543 case DIFF_SYMBOL_REWRITE_DIFF:
1544 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1545 reset = diff_get_color_opt(o, DIFF_RESET);
1546 emit_line(o, fraginfo, reset, line, len);
1547 break;
1548 case DIFF_SYMBOL_SUBMODULE_ADD:
1549 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1550 reset = diff_get_color_opt(o, DIFF_RESET);
1551 emit_line(o, set, reset, line, len);
1552 break;
1553 case DIFF_SYMBOL_SUBMODULE_DEL:
1554 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1555 reset = diff_get_color_opt(o, DIFF_RESET);
1556 emit_line(o, set, reset, line, len);
1557 break;
1558 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1559 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1560 diff_line_prefix(o), line);
1561 break;
1562 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1563 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1564 diff_line_prefix(o), line);
1565 break;
1566 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1567 emit_line(o, "", "", " 0 files changed\n",
1568 strlen(" 0 files changed\n"));
1569 break;
1570 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1571 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1572 break;
1573 case DIFF_SYMBOL_WORD_DIFF:
1574 fprintf(o->file, "%.*s", len, line);
1575 break;
1576 case DIFF_SYMBOL_STAT_SEP:
1577 fputs(o->stat_sep, o->file);
1578 break;
1579 default:
1580 BUG("unknown diff symbol");
1581 }
1582 }
1583
1584 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1585 const char *line, int len, unsigned flags)
1586 {
1587 struct emitted_diff_symbol e = {
1588 .line = line, .len = len, .flags = flags, .s = s
1589 };
1590
1591 if (o->emitted_symbols)
1592 append_emitted_diff_symbol(o, &e);
1593 else
1594 emit_diff_symbol_from_struct(o, &e);
1595 }
1596
1597 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1598 {
1599 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1600 }
1601
1602 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1603 {
1604 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1605 }
1606
1607 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1608 {
1609 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1610 path, strlen(path), 0);
1611 }
1612
1613 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1614 {
1615 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1616 path, strlen(path), 0);
1617 }
1618
1619 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1620 {
1621 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1622 header, strlen(header), 0);
1623 }
1624
1625 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1626 {
1627 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1628 }
1629
1630 void diff_emit_submodule_pipethrough(struct diff_options *o,
1631 const char *line, int len)
1632 {
1633 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1634 }
1635
1636 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1637 {
1638 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1639 ecbdata->blank_at_eof_in_preimage &&
1640 ecbdata->blank_at_eof_in_postimage &&
1641 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1642 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1643 return 0;
1644 return ws_blank_line(line, len);
1645 }
1646
1647 static void emit_add_line(struct emit_callback *ecbdata,
1648 const char *line, int len)
1649 {
1650 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1651 if (new_blank_line_at_eof(ecbdata, line, len))
1652 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1653
1654 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1655 }
1656
1657 static void emit_del_line(struct emit_callback *ecbdata,
1658 const char *line, int len)
1659 {
1660 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1661 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1662 }
1663
1664 static void emit_context_line(struct emit_callback *ecbdata,
1665 const char *line, int len)
1666 {
1667 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1668 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1669 }
1670
1671 static void emit_hunk_header(struct emit_callback *ecbdata,
1672 const char *line, int len)
1673 {
1674 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1675 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1676 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1677 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1678 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1679 static const char atat[2] = { '@', '@' };
1680 const char *cp, *ep;
1681 struct strbuf msgbuf = STRBUF_INIT;
1682 int org_len = len;
1683 int i = 1;
1684
1685 /*
1686 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1687 * it always is at least 10 bytes long.
1688 */
1689 if (len < 10 ||
1690 memcmp(line, atat, 2) ||
1691 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1692 emit_diff_symbol(ecbdata->opt,
1693 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1694 return;
1695 }
1696 ep += 2; /* skip over @@ */
1697
1698 /* The hunk header in fraginfo color */
1699 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1700 strbuf_addstr(&msgbuf, reverse);
1701 strbuf_addstr(&msgbuf, frag);
1702 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1703 strbuf_add(&msgbuf, atat, sizeof(atat));
1704 else
1705 strbuf_add(&msgbuf, line, ep - line);
1706 strbuf_addstr(&msgbuf, reset);
1707
1708 /*
1709 * trailing "\r\n"
1710 */
1711 for ( ; i < 3; i++)
1712 if (line[len - i] == '\r' || line[len - i] == '\n')
1713 len--;
1714
1715 /* blank before the func header */
1716 for (cp = ep; ep - line < len; ep++)
1717 if (*ep != ' ' && *ep != '\t')
1718 break;
1719 if (ep != cp) {
1720 strbuf_addstr(&msgbuf, context);
1721 strbuf_add(&msgbuf, cp, ep - cp);
1722 strbuf_addstr(&msgbuf, reset);
1723 }
1724
1725 if (ep < line + len) {
1726 strbuf_addstr(&msgbuf, func);
1727 strbuf_add(&msgbuf, ep, line + len - ep);
1728 strbuf_addstr(&msgbuf, reset);
1729 }
1730
1731 strbuf_add(&msgbuf, line + len, org_len - len);
1732 strbuf_complete_line(&msgbuf);
1733 emit_diff_symbol(ecbdata->opt,
1734 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1735 strbuf_release(&msgbuf);
1736 }
1737
1738 static struct diff_tempfile *claim_diff_tempfile(void)
1739 {
1740 int i;
1741 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1742 if (!diff_temp[i].name)
1743 return diff_temp + i;
1744 BUG("diff is failing to clean up its tempfiles");
1745 }
1746
1747 static void remove_tempfile(void)
1748 {
1749 int i;
1750 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1751 if (is_tempfile_active(diff_temp[i].tempfile))
1752 delete_tempfile(&diff_temp[i].tempfile);
1753 diff_temp[i].name = NULL;
1754 }
1755 }
1756
1757 static void add_line_count(struct strbuf *out, int count)
1758 {
1759 switch (count) {
1760 case 0:
1761 strbuf_addstr(out, "0,0");
1762 break;
1763 case 1:
1764 strbuf_addstr(out, "1");
1765 break;
1766 default:
1767 strbuf_addf(out, "1,%d", count);
1768 break;
1769 }
1770 }
1771
1772 static void emit_rewrite_lines(struct emit_callback *ecb,
1773 int prefix, const char *data, int size)
1774 {
1775 const char *endp = NULL;
1776
1777 while (0 < size) {
1778 int len;
1779
1780 endp = memchr(data, '\n', size);
1781 len = endp ? (endp - data + 1) : size;
1782 if (prefix != '+') {
1783 ecb->lno_in_preimage++;
1784 emit_del_line(ecb, data, len);
1785 } else {
1786 ecb->lno_in_postimage++;
1787 emit_add_line(ecb, data, len);
1788 }
1789 size -= len;
1790 data += len;
1791 }
1792 if (!endp)
1793 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1794 }
1795
1796 static void emit_rewrite_diff(const char *name_a,
1797 const char *name_b,
1798 struct diff_filespec *one,
1799 struct diff_filespec *two,
1800 struct userdiff_driver *textconv_one,
1801 struct userdiff_driver *textconv_two,
1802 struct diff_options *o)
1803 {
1804 int lc_a, lc_b;
1805 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1806 const char *a_prefix, *b_prefix;
1807 char *data_one, *data_two;
1808 size_t size_one, size_two;
1809 struct emit_callback ecbdata;
1810 struct strbuf out = STRBUF_INIT;
1811
1812 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1813 a_prefix = o->b_prefix;
1814 b_prefix = o->a_prefix;
1815 } else {
1816 a_prefix = o->a_prefix;
1817 b_prefix = o->b_prefix;
1818 }
1819
1820 name_a += (*name_a == '/');
1821 name_b += (*name_b == '/');
1822
1823 strbuf_reset(&a_name);
1824 strbuf_reset(&b_name);
1825 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1826 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1827
1828 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1829 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1830
1831 memset(&ecbdata, 0, sizeof(ecbdata));
1832 ecbdata.color_diff = want_color(o->use_color);
1833 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1834 ecbdata.opt = o;
1835 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1836 mmfile_t mf1, mf2;
1837 mf1.ptr = (char *)data_one;
1838 mf2.ptr = (char *)data_two;
1839 mf1.size = size_one;
1840 mf2.size = size_two;
1841 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1842 }
1843 ecbdata.lno_in_preimage = 1;
1844 ecbdata.lno_in_postimage = 1;
1845
1846 lc_a = count_lines(data_one, size_one);
1847 lc_b = count_lines(data_two, size_two);
1848
1849 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1850 a_name.buf, a_name.len, 0);
1851 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1852 b_name.buf, b_name.len, 0);
1853
1854 strbuf_addstr(&out, "@@ -");
1855 if (!o->irreversible_delete)
1856 add_line_count(&out, lc_a);
1857 else
1858 strbuf_addstr(&out, "?,?");
1859 strbuf_addstr(&out, " +");
1860 add_line_count(&out, lc_b);
1861 strbuf_addstr(&out, " @@\n");
1862 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1863 strbuf_release(&out);
1864
1865 if (lc_a && !o->irreversible_delete)
1866 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1867 if (lc_b)
1868 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1869 if (textconv_one)
1870 free((char *)data_one);
1871 if (textconv_two)
1872 free((char *)data_two);
1873 }
1874
1875 struct diff_words_buffer {
1876 mmfile_t text;
1877 unsigned long alloc;
1878 struct diff_words_orig {
1879 const char *begin, *end;
1880 } *orig;
1881 int orig_nr, orig_alloc;
1882 };
1883
1884 static void diff_words_append(char *line, unsigned long len,
1885 struct diff_words_buffer *buffer)
1886 {
1887 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1888 line++;
1889 len--;
1890 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1891 buffer->text.size += len;
1892 buffer->text.ptr[buffer->text.size] = '\0';
1893 }
1894
1895 struct diff_words_style_elem {
1896 const char *prefix;
1897 const char *suffix;
1898 const char *color; /* NULL; filled in by the setup code if
1899 * color is enabled */
1900 };
1901
1902 struct diff_words_style {
1903 enum diff_words_type type;
1904 struct diff_words_style_elem new_word, old_word, ctx;
1905 const char *newline;
1906 };
1907
1908 static struct diff_words_style diff_words_styles[] = {
1909 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1910 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1911 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1912 };
1913
1914 struct diff_words_data {
1915 struct diff_words_buffer minus, plus;
1916 const char *current_plus;
1917 int last_minus;
1918 struct diff_options *opt;
1919 regex_t *word_regex;
1920 enum diff_words_type type;
1921 struct diff_words_style *style;
1922 };
1923
1924 static int fn_out_diff_words_write_helper(struct diff_options *o,
1925 struct diff_words_style_elem *st_el,
1926 const char *newline,
1927 size_t count, const char *buf)
1928 {
1929 int print = 0;
1930 struct strbuf sb = STRBUF_INIT;
1931
1932 while (count) {
1933 char *p = memchr(buf, '\n', count);
1934 if (print)
1935 strbuf_addstr(&sb, diff_line_prefix(o));
1936
1937 if (p != buf) {
1938 const char *reset = st_el->color && *st_el->color ?
1939 GIT_COLOR_RESET : NULL;
1940 if (st_el->color && *st_el->color)
1941 strbuf_addstr(&sb, st_el->color);
1942 strbuf_addstr(&sb, st_el->prefix);
1943 strbuf_add(&sb, buf, p ? p - buf : count);
1944 strbuf_addstr(&sb, st_el->suffix);
1945 if (reset)
1946 strbuf_addstr(&sb, reset);
1947 }
1948 if (!p)
1949 goto out;
1950
1951 strbuf_addstr(&sb, newline);
1952 count -= p + 1 - buf;
1953 buf = p + 1;
1954 print = 1;
1955 if (count) {
1956 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1957 sb.buf, sb.len, 0);
1958 strbuf_reset(&sb);
1959 }
1960 }
1961
1962 out:
1963 if (sb.len)
1964 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1965 sb.buf, sb.len, 0);
1966 strbuf_release(&sb);
1967 return 0;
1968 }
1969
1970 /*
1971 * '--color-words' algorithm can be described as:
1972 *
1973 * 1. collect the minus/plus lines of a diff hunk, divided into
1974 * minus-lines and plus-lines;
1975 *
1976 * 2. break both minus-lines and plus-lines into words and
1977 * place them into two mmfile_t with one word for each line;
1978 *
1979 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1980 *
1981 * And for the common parts of the both file, we output the plus side text.
1982 * diff_words->current_plus is used to trace the current position of the plus file
1983 * which printed. diff_words->last_minus is used to trace the last minus word
1984 * printed.
1985 *
1986 * For '--graph' to work with '--color-words', we need to output the graph prefix
1987 * on each line of color words output. Generally, there are two conditions on
1988 * which we should output the prefix.
1989 *
1990 * 1. diff_words->last_minus == 0 &&
1991 * diff_words->current_plus == diff_words->plus.text.ptr
1992 *
1993 * that is: the plus text must start as a new line, and if there is no minus
1994 * word printed, a graph prefix must be printed.
1995 *
1996 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1997 * *(diff_words->current_plus - 1) == '\n'
1998 *
1999 * that is: a graph prefix must be printed following a '\n'
2000 */
2001 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
2002 {
2003 if ((diff_words->last_minus == 0 &&
2004 diff_words->current_plus == diff_words->plus.text.ptr) ||
2005 (diff_words->current_plus > diff_words->plus.text.ptr &&
2006 *(diff_words->current_plus - 1) == '\n')) {
2007 return 1;
2008 } else {
2009 return 0;
2010 }
2011 }
2012
2013 static void fn_out_diff_words_aux(void *priv,
2014 long minus_first, long minus_len,
2015 long plus_first, long plus_len,
2016 const char *func UNUSED, long funclen UNUSED)
2017 {
2018 struct diff_words_data *diff_words = priv;
2019 struct diff_words_style *style = diff_words->style;
2020 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
2021 struct diff_options *opt = diff_words->opt;
2022 const char *line_prefix;
2023
2024 assert(opt);
2025 line_prefix = diff_line_prefix(opt);
2026
2027 /* POSIX requires that first be decremented by one if len == 0... */
2028 if (minus_len) {
2029 minus_begin = diff_words->minus.orig[minus_first].begin;
2030 minus_end =
2031 diff_words->minus.orig[minus_first + minus_len - 1].end;
2032 } else
2033 minus_begin = minus_end =
2034 diff_words->minus.orig[minus_first].end;
2035
2036 if (plus_len) {
2037 plus_begin = diff_words->plus.orig[plus_first].begin;
2038 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2039 } else
2040 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2041
2042 if (color_words_output_graph_prefix(diff_words)) {
2043 fputs(line_prefix, diff_words->opt->file);
2044 }
2045 if (diff_words->current_plus != plus_begin) {
2046 fn_out_diff_words_write_helper(diff_words->opt,
2047 &style->ctx, style->newline,
2048 plus_begin - diff_words->current_plus,
2049 diff_words->current_plus);
2050 }
2051 if (minus_begin != minus_end) {
2052 fn_out_diff_words_write_helper(diff_words->opt,
2053 &style->old_word, style->newline,
2054 minus_end - minus_begin, minus_begin);
2055 }
2056 if (plus_begin != plus_end) {
2057 fn_out_diff_words_write_helper(diff_words->opt,
2058 &style->new_word, style->newline,
2059 plus_end - plus_begin, plus_begin);
2060 }
2061
2062 diff_words->current_plus = plus_end;
2063 diff_words->last_minus = minus_first;
2064 }
2065
2066 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2067 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2068 int *begin, int *end)
2069 {
2070 while (word_regex && *begin < buffer->size) {
2071 regmatch_t match[1];
2072 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2073 buffer->size - *begin, 1, match, 0)) {
2074 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2075 '\n', match[0].rm_eo - match[0].rm_so);
2076 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2077 *begin += match[0].rm_so;
2078 if (*begin == *end)
2079 (*begin)++;
2080 else
2081 return *begin > *end;
2082 } else {
2083 return -1;
2084 }
2085 }
2086
2087 /* find the next word */
2088 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2089 (*begin)++;
2090 if (*begin >= buffer->size)
2091 return -1;
2092
2093 /* find the end of the word */
2094 *end = *begin + 1;
2095 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2096 (*end)++;
2097
2098 return 0;
2099 }
2100
2101 /*
2102 * This function splits the words in buffer->text, stores the list with
2103 * newline separator into out, and saves the offsets of the original words
2104 * in buffer->orig.
2105 */
2106 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2107 regex_t *word_regex)
2108 {
2109 int i, j;
2110 long alloc = 0;
2111
2112 out->size = 0;
2113 out->ptr = NULL;
2114
2115 /* fake an empty "0th" word */
2116 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2117 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2118 buffer->orig_nr = 1;
2119
2120 for (i = 0; i < buffer->text.size; i++) {
2121 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2122 return;
2123
2124 /* store original boundaries */
2125 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2126 buffer->orig_alloc);
2127 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2128 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2129 buffer->orig_nr++;
2130
2131 /* store one word */
2132 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2133 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2134 out->ptr[out->size + j - i] = '\n';
2135 out->size += j - i + 1;
2136
2137 i = j - 1;
2138 }
2139 }
2140
2141 /* this executes the word diff on the accumulated buffers */
2142 static void diff_words_show(struct diff_words_data *diff_words)
2143 {
2144 xpparam_t xpp;
2145 xdemitconf_t xecfg;
2146 mmfile_t minus, plus;
2147 struct diff_words_style *style = diff_words->style;
2148
2149 struct diff_options *opt = diff_words->opt;
2150 const char *line_prefix;
2151
2152 assert(opt);
2153 line_prefix = diff_line_prefix(opt);
2154
2155 /* special case: only removal */
2156 if (!diff_words->plus.text.size) {
2157 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2158 line_prefix, strlen(line_prefix), 0);
2159 fn_out_diff_words_write_helper(diff_words->opt,
2160 &style->old_word, style->newline,
2161 diff_words->minus.text.size,
2162 diff_words->minus.text.ptr);
2163 diff_words->minus.text.size = 0;
2164 return;
2165 }
2166
2167 diff_words->current_plus = diff_words->plus.text.ptr;
2168 diff_words->last_minus = 0;
2169
2170 memset(&xpp, 0, sizeof(xpp));
2171 memset(&xecfg, 0, sizeof(xecfg));
2172 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2173 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2174 xpp.flags = 0;
2175 /* as only the hunk header will be parsed, we need a 0-context */
2176 xecfg.ctxlen = 0;
2177 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2178 diff_words, &xpp, &xecfg))
2179 die("unable to generate word diff");
2180 free(minus.ptr);
2181 free(plus.ptr);
2182 if (diff_words->current_plus != diff_words->plus.text.ptr +
2183 diff_words->plus.text.size) {
2184 if (color_words_output_graph_prefix(diff_words))
2185 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2186 line_prefix, strlen(line_prefix), 0);
2187 fn_out_diff_words_write_helper(diff_words->opt,
2188 &style->ctx, style->newline,
2189 diff_words->plus.text.ptr + diff_words->plus.text.size
2190 - diff_words->current_plus, diff_words->current_plus);
2191 }
2192 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2193 }
2194
2195 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2196 static void diff_words_flush(struct emit_callback *ecbdata)
2197 {
2198 struct diff_options *wo = ecbdata->diff_words->opt;
2199
2200 if (ecbdata->diff_words->minus.text.size ||
2201 ecbdata->diff_words->plus.text.size)
2202 diff_words_show(ecbdata->diff_words);
2203
2204 if (wo->emitted_symbols) {
2205 struct diff_options *o = ecbdata->opt;
2206 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2207 int i;
2208
2209 /*
2210 * NEEDSWORK:
2211 * Instead of appending each, concat all words to a line?
2212 */
2213 for (i = 0; i < wol->nr; i++)
2214 append_emitted_diff_symbol(o, &wol->buf[i]);
2215
2216 for (i = 0; i < wol->nr; i++)
2217 free((void *)wol->buf[i].line);
2218
2219 wol->nr = 0;
2220 }
2221 }
2222
2223 static void diff_filespec_load_driver(struct diff_filespec *one,
2224 struct index_state *istate)
2225 {
2226 /* Use already-loaded driver */
2227 if (one->driver)
2228 return;
2229
2230 if (S_ISREG(one->mode))
2231 one->driver = userdiff_find_by_path(istate, one->path);
2232
2233 /* Fallback to default settings */
2234 if (!one->driver)
2235 one->driver = userdiff_find_by_name("default");
2236 }
2237
2238 static const char *userdiff_word_regex(struct diff_filespec *one,
2239 struct index_state *istate)
2240 {
2241 diff_filespec_load_driver(one, istate);
2242 return one->driver->word_regex;
2243 }
2244
2245 static void init_diff_words_data(struct emit_callback *ecbdata,
2246 struct diff_options *orig_opts,
2247 struct diff_filespec *one,
2248 struct diff_filespec *two)
2249 {
2250 int i;
2251 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2252 memcpy(o, orig_opts, sizeof(struct diff_options));
2253
2254 CALLOC_ARRAY(ecbdata->diff_words, 1);
2255 ecbdata->diff_words->type = o->word_diff;
2256 ecbdata->diff_words->opt = o;
2257
2258 if (orig_opts->emitted_symbols)
2259 CALLOC_ARRAY(o->emitted_symbols, 1);
2260
2261 if (!o->word_regex)
2262 o->word_regex = userdiff_word_regex(one, o->repo->index);
2263 if (!o->word_regex)
2264 o->word_regex = userdiff_word_regex(two, o->repo->index);
2265 if (!o->word_regex)
2266 o->word_regex = diff_word_regex_cfg;
2267 if (o->word_regex) {
2268 ecbdata->diff_words->word_regex = (regex_t *)
2269 xmalloc(sizeof(regex_t));
2270 if (regcomp(ecbdata->diff_words->word_regex,
2271 o->word_regex,
2272 REG_EXTENDED | REG_NEWLINE))
2273 die("invalid regular expression: %s",
2274 o->word_regex);
2275 }
2276 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2277 if (o->word_diff == diff_words_styles[i].type) {
2278 ecbdata->diff_words->style =
2279 &diff_words_styles[i];
2280 break;
2281 }
2282 }
2283 if (want_color(o->use_color)) {
2284 struct diff_words_style *st = ecbdata->diff_words->style;
2285 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2286 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2287 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2288 }
2289 }
2290
2291 static void free_diff_words_data(struct emit_callback *ecbdata)
2292 {
2293 if (ecbdata->diff_words) {
2294 diff_words_flush(ecbdata);
2295 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2296 free (ecbdata->diff_words->opt);
2297 free (ecbdata->diff_words->minus.text.ptr);
2298 free (ecbdata->diff_words->minus.orig);
2299 free (ecbdata->diff_words->plus.text.ptr);
2300 free (ecbdata->diff_words->plus.orig);
2301 if (ecbdata->diff_words->word_regex) {
2302 regfree(ecbdata->diff_words->word_regex);
2303 free(ecbdata->diff_words->word_regex);
2304 }
2305 FREE_AND_NULL(ecbdata->diff_words);
2306 }
2307 }
2308
2309 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2310 {
2311 if (want_color(diff_use_color))
2312 return diff_colors[ix];
2313 return "";
2314 }
2315
2316 const char *diff_line_prefix(struct diff_options *opt)
2317 {
2318 struct strbuf *msgbuf;
2319 if (!opt->output_prefix)
2320 return "";
2321
2322 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2323 return msgbuf->buf;
2324 }
2325
2326 static unsigned long sane_truncate_line(char *line, unsigned long len)
2327 {
2328 const char *cp;
2329 unsigned long allot;
2330 size_t l = len;
2331
2332 cp = line;
2333 allot = l;
2334 while (0 < l) {
2335 (void) utf8_width(&cp, &l);
2336 if (!cp)
2337 break; /* truncated in the middle? */
2338 }
2339 return allot - l;
2340 }
2341
2342 static void find_lno(const char *line, struct emit_callback *ecbdata)
2343 {
2344 const char *p;
2345 ecbdata->lno_in_preimage = 0;
2346 ecbdata->lno_in_postimage = 0;
2347 p = strchr(line, '-');
2348 if (!p)
2349 return; /* cannot happen */
2350 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2351 p = strchr(p, '+');
2352 if (!p)
2353 return; /* cannot happen */
2354 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2355 }
2356
2357 static int fn_out_consume(void *priv, char *line, unsigned long len)
2358 {
2359 struct emit_callback *ecbdata = priv;
2360 struct diff_options *o = ecbdata->opt;
2361
2362 o->found_changes = 1;
2363
2364 if (ecbdata->header) {
2365 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2366 ecbdata->header->buf, ecbdata->header->len, 0);
2367 strbuf_reset(ecbdata->header);
2368 ecbdata->header = NULL;
2369 }
2370
2371 if (ecbdata->label_path[0]) {
2372 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2373 ecbdata->label_path[0],
2374 strlen(ecbdata->label_path[0]), 0);
2375 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2376 ecbdata->label_path[1],
2377 strlen(ecbdata->label_path[1]), 0);
2378 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2379 }
2380
2381 if (diff_suppress_blank_empty
2382 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2383 line[0] = '\n';
2384 len = 1;
2385 }
2386
2387 if (line[0] == '@') {
2388 if (ecbdata->diff_words)
2389 diff_words_flush(ecbdata);
2390 len = sane_truncate_line(line, len);
2391 find_lno(line, ecbdata);
2392 emit_hunk_header(ecbdata, line, len);
2393 return 0;
2394 }
2395
2396 if (ecbdata->diff_words) {
2397 enum diff_symbol s =
2398 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2399 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2400 if (line[0] == '-') {
2401 diff_words_append(line, len,
2402 &ecbdata->diff_words->minus);
2403 return 0;
2404 } else if (line[0] == '+') {
2405 diff_words_append(line, len,
2406 &ecbdata->diff_words->plus);
2407 return 0;
2408 } else if (starts_with(line, "\\ ")) {
2409 /*
2410 * Eat the "no newline at eof" marker as if we
2411 * saw a "+" or "-" line with nothing on it,
2412 * and return without diff_words_flush() to
2413 * defer processing. If this is the end of
2414 * preimage, more "+" lines may come after it.
2415 */
2416 return 0;
2417 }
2418 diff_words_flush(ecbdata);
2419 emit_diff_symbol(o, s, line, len, 0);
2420 return 0;
2421 }
2422
2423 switch (line[0]) {
2424 case '+':
2425 ecbdata->lno_in_postimage++;
2426 emit_add_line(ecbdata, line + 1, len - 1);
2427 break;
2428 case '-':
2429 ecbdata->lno_in_preimage++;
2430 emit_del_line(ecbdata, line + 1, len - 1);
2431 break;
2432 case ' ':
2433 ecbdata->lno_in_postimage++;
2434 ecbdata->lno_in_preimage++;
2435 emit_context_line(ecbdata, line + 1, len - 1);
2436 break;
2437 default:
2438 /* incomplete line at the end */
2439 ecbdata->lno_in_preimage++;
2440 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2441 line, len, 0);
2442 break;
2443 }
2444 return 0;
2445 }
2446
2447 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2448 {
2449 const char *old_name = a;
2450 const char *new_name = b;
2451 int pfx_length, sfx_length;
2452 int pfx_adjust_for_slash;
2453 int len_a = strlen(a);
2454 int len_b = strlen(b);
2455 int a_midlen, b_midlen;
2456 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2457 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2458
2459 if (qlen_a || qlen_b) {
2460 quote_c_style(a, name, NULL, 0);
2461 strbuf_addstr(name, " => ");
2462 quote_c_style(b, name, NULL, 0);
2463 return;
2464 }
2465
2466 /* Find common prefix */
2467 pfx_length = 0;
2468 while (*old_name && *new_name && *old_name == *new_name) {
2469 if (*old_name == '/')
2470 pfx_length = old_name - a + 1;
2471 old_name++;
2472 new_name++;
2473 }
2474
2475 /* Find common suffix */
2476 old_name = a + len_a;
2477 new_name = b + len_b;
2478 sfx_length = 0;
2479 /*
2480 * If there is a common prefix, it must end in a slash. In
2481 * that case we let this loop run 1 into the prefix to see the
2482 * same slash.
2483 *
2484 * If there is no common prefix, we cannot do this as it would
2485 * underrun the input strings.
2486 */
2487 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2488 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2489 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2490 *old_name == *new_name) {
2491 if (*old_name == '/')
2492 sfx_length = len_a - (old_name - a);
2493 old_name--;
2494 new_name--;
2495 }
2496
2497 /*
2498 * pfx{mid-a => mid-b}sfx
2499 * {pfx-a => pfx-b}sfx
2500 * pfx{sfx-a => sfx-b}
2501 * name-a => name-b
2502 */
2503 a_midlen = len_a - pfx_length - sfx_length;
2504 b_midlen = len_b - pfx_length - sfx_length;
2505 if (a_midlen < 0)
2506 a_midlen = 0;
2507 if (b_midlen < 0)
2508 b_midlen = 0;
2509
2510 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2511 if (pfx_length + sfx_length) {
2512 strbuf_add(name, a, pfx_length);
2513 strbuf_addch(name, '{');
2514 }
2515 strbuf_add(name, a + pfx_length, a_midlen);
2516 strbuf_addstr(name, " => ");
2517 strbuf_add(name, b + pfx_length, b_midlen);
2518 if (pfx_length + sfx_length) {
2519 strbuf_addch(name, '}');
2520 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2521 }
2522 }
2523
2524 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2525 const char *name_a,
2526 const char *name_b)
2527 {
2528 struct diffstat_file *x;
2529 CALLOC_ARRAY(x, 1);
2530 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2531 diffstat->files[diffstat->nr++] = x;
2532 if (name_b) {
2533 x->from_name = xstrdup(name_a);
2534 x->name = xstrdup(name_b);
2535 x->is_renamed = 1;
2536 }
2537 else {
2538 x->from_name = NULL;
2539 x->name = xstrdup(name_a);
2540 }
2541 return x;
2542 }
2543
2544 static int diffstat_consume(void *priv, char *line, unsigned long len)
2545 {
2546 struct diffstat_t *diffstat = priv;
2547 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2548
2549 if (!len)
2550 BUG("xdiff fed us an empty line");
2551
2552 if (line[0] == '+')
2553 x->added++;
2554 else if (line[0] == '-')
2555 x->deleted++;
2556 return 0;
2557 }
2558
2559 const char mime_boundary_leader[] = "------------";
2560
2561 static int scale_linear(int it, int width, int max_change)
2562 {
2563 if (!it)
2564 return 0;
2565 /*
2566 * make sure that at least one '-' or '+' is printed if
2567 * there is any change to this path. The easiest way is to
2568 * scale linearly as if the allotted width is one column shorter
2569 * than it is, and then add 1 to the result.
2570 */
2571 return 1 + (it * (width - 1) / max_change);
2572 }
2573
2574 static void show_graph(struct strbuf *out, char ch, int cnt,
2575 const char *set, const char *reset)
2576 {
2577 if (cnt <= 0)
2578 return;
2579 strbuf_addstr(out, set);
2580 strbuf_addchars(out, ch, cnt);
2581 strbuf_addstr(out, reset);
2582 }
2583
2584 static void fill_print_name(struct diffstat_file *file)
2585 {
2586 struct strbuf pname = STRBUF_INIT;
2587
2588 if (file->print_name)
2589 return;
2590
2591 if (file->is_renamed)
2592 pprint_rename(&pname, file->from_name, file->name);
2593 else
2594 quote_c_style(file->name, &pname, NULL, 0);
2595
2596 if (file->comments)
2597 strbuf_addf(&pname, " (%s)", file->comments);
2598
2599 file->print_name = strbuf_detach(&pname, NULL);
2600 }
2601
2602 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2603 int files, int insertions, int deletions)
2604 {
2605 struct strbuf sb = STRBUF_INIT;
2606
2607 if (!files) {
2608 assert(insertions == 0 && deletions == 0);
2609 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2610 NULL, 0, 0);
2611 return;
2612 }
2613
2614 strbuf_addf(&sb,
2615 (files == 1) ? " %d file changed" : " %d files changed",
2616 files);
2617
2618 /*
2619 * For binary diff, the caller may want to print "x files
2620 * changed" with insertions == 0 && deletions == 0.
2621 *
2622 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2623 * is probably less confusing (i.e skip over "2 files changed
2624 * but nothing about added/removed lines? Is this a bug in Git?").
2625 */
2626 if (insertions || deletions == 0) {
2627 strbuf_addf(&sb,
2628 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2629 insertions);
2630 }
2631
2632 if (deletions || insertions == 0) {
2633 strbuf_addf(&sb,
2634 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2635 deletions);
2636 }
2637 strbuf_addch(&sb, '\n');
2638 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2639 sb.buf, sb.len, 0);
2640 strbuf_release(&sb);
2641 }
2642
2643 void print_stat_summary(FILE *fp, int files,
2644 int insertions, int deletions)
2645 {
2646 struct diff_options o;
2647 memset(&o, 0, sizeof(o));
2648 o.file = fp;
2649
2650 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2651 }
2652
2653 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2654 {
2655 int i, len, add, del, adds = 0, dels = 0;
2656 uintmax_t max_change = 0, max_len = 0;
2657 int total_files = data->nr, count;
2658 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2659 const char *reset, *add_c, *del_c;
2660 int extra_shown = 0;
2661 const char *line_prefix = diff_line_prefix(options);
2662 struct strbuf out = STRBUF_INIT;
2663
2664 if (data->nr == 0)
2665 return;
2666
2667 count = options->stat_count ? options->stat_count : data->nr;
2668
2669 reset = diff_get_color_opt(options, DIFF_RESET);
2670 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2671 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2672
2673 /*
2674 * Find the longest filename and max number of changes
2675 */
2676 for (i = 0; (i < count) && (i < data->nr); i++) {
2677 struct diffstat_file *file = data->files[i];
2678 uintmax_t change = file->added + file->deleted;
2679
2680 if (!file->is_interesting && (change == 0)) {
2681 count++; /* not shown == room for one more */
2682 continue;
2683 }
2684 fill_print_name(file);
2685 len = utf8_strwidth(file->print_name);
2686 if (max_len < len)
2687 max_len = len;
2688
2689 if (file->is_unmerged) {
2690 /* "Unmerged" is 8 characters */
2691 bin_width = bin_width < 8 ? 8 : bin_width;
2692 continue;
2693 }
2694 if (file->is_binary) {
2695 /* "Bin XXX -> YYY bytes" */
2696 int w = 14 + decimal_width(file->added)
2697 + decimal_width(file->deleted);
2698 bin_width = bin_width < w ? w : bin_width;
2699 /* Display change counts aligned with "Bin" */
2700 number_width = 3;
2701 continue;
2702 }
2703
2704 if (max_change < change)
2705 max_change = change;
2706 }
2707 count = i; /* where we can stop scanning in data->files[] */
2708
2709 /*
2710 * We have width = stat_width or term_columns() columns total.
2711 * We want a maximum of min(max_len, stat_name_width) for the name part.
2712 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2713 * We also need 1 for " " and 4 + decimal_width(max_change)
2714 * for " | NNNN " and one the empty column at the end, altogether
2715 * 6 + decimal_width(max_change).
2716 *
2717 * If there's not enough space, we will use the smaller of
2718 * stat_name_width (if set) and 5/8*width for the filename,
2719 * and the rest for constant elements + graph part, but no more
2720 * than stat_graph_width for the graph part.
2721 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2722 * for the standard terminal size).
2723 *
2724 * In other words: stat_width limits the maximum width, and
2725 * stat_name_width fixes the maximum width of the filename,
2726 * and is also used to divide available columns if there
2727 * aren't enough.
2728 *
2729 * Binary files are displayed with "Bin XXX -> YYY bytes"
2730 * instead of the change count and graph. This part is treated
2731 * similarly to the graph part, except that it is not
2732 * "scaled". If total width is too small to accommodate the
2733 * guaranteed minimum width of the filename part and the
2734 * separators and this message, this message will "overflow"
2735 * making the line longer than the maximum width.
2736 */
2737
2738 /*
2739 * NEEDSWORK: line_prefix is often used for "log --graph" output
2740 * and contains ANSI-colored string. utf8_strnwidth() should be
2741 * used to correctly count the display width instead of strlen().
2742 */
2743 if (options->stat_width == -1)
2744 width = term_columns() - strlen(line_prefix);
2745 else
2746 width = options->stat_width ? options->stat_width : 80;
2747 number_width = decimal_width(max_change) > number_width ?
2748 decimal_width(max_change) : number_width;
2749
2750 if (options->stat_name_width == -1)
2751 options->stat_name_width = diff_stat_name_width;
2752 if (options->stat_graph_width == -1)
2753 options->stat_graph_width = diff_stat_graph_width;
2754
2755 /*
2756 * Guarantee 3/8*16 == 6 for the graph part
2757 * and 5/8*16 == 10 for the filename part
2758 */
2759 if (width < 16 + 6 + number_width)
2760 width = 16 + 6 + number_width;
2761
2762 /*
2763 * First assign sizes that are wanted, ignoring available width.
2764 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2765 * starting from "XXX" should fit in graph_width.
2766 */
2767 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2768 if (options->stat_graph_width &&
2769 options->stat_graph_width < graph_width)
2770 graph_width = options->stat_graph_width;
2771
2772 name_width = (options->stat_name_width > 0 &&
2773 options->stat_name_width < max_len) ?
2774 options->stat_name_width : max_len;
2775
2776 /*
2777 * Adjust adjustable widths not to exceed maximum width
2778 */
2779 if (name_width + number_width + 6 + graph_width > width) {
2780 if (graph_width > width * 3/8 - number_width - 6) {
2781 graph_width = width * 3/8 - number_width - 6;
2782 if (graph_width < 6)
2783 graph_width = 6;
2784 }
2785
2786 if (options->stat_graph_width &&
2787 graph_width > options->stat_graph_width)
2788 graph_width = options->stat_graph_width;
2789 if (name_width > width - number_width - 6 - graph_width)
2790 name_width = width - number_width - 6 - graph_width;
2791 else
2792 graph_width = width - number_width - 6 - name_width;
2793 }
2794
2795 /*
2796 * From here name_width is the width of the name area,
2797 * and graph_width is the width of the graph area.
2798 * max_change is used to scale graph properly.
2799 */
2800 for (i = 0; i < count; i++) {
2801 const char *prefix = "";
2802 struct diffstat_file *file = data->files[i];
2803 char *name = file->print_name;
2804 uintmax_t added = file->added;
2805 uintmax_t deleted = file->deleted;
2806 int name_len, padding;
2807
2808 if (!file->is_interesting && (added + deleted == 0))
2809 continue;
2810
2811 /*
2812 * "scale" the filename
2813 */
2814 len = name_width;
2815 name_len = utf8_strwidth(name);
2816 if (name_width < name_len) {
2817 char *slash;
2818 prefix = "...";
2819 len -= 3;
2820 /*
2821 * NEEDSWORK: (name_len - len) counts the display
2822 * width, which would be shorter than the byte
2823 * length of the corresponding substring.
2824 * Advancing "name" by that number of bytes does
2825 * *NOT* skip over that many columns, so it is
2826 * very likely that chomping the pathname at the
2827 * slash we will find starting from "name" will
2828 * leave the resulting string still too long.
2829 */
2830 name += name_len - len;
2831 slash = strchr(name, '/');
2832 if (slash)
2833 name = slash;
2834 }
2835 padding = len - utf8_strwidth(name);
2836 if (padding < 0)
2837 padding = 0;
2838
2839 if (file->is_binary) {
2840 strbuf_addf(&out, " %s%s%*s | %*s",
2841 prefix, name, padding, "",
2842 number_width, "Bin");
2843 if (!added && !deleted) {
2844 strbuf_addch(&out, '\n');
2845 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2846 out.buf, out.len, 0);
2847 strbuf_reset(&out);
2848 continue;
2849 }
2850 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2851 del_c, deleted, reset);
2852 strbuf_addstr(&out, " -> ");
2853 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2854 add_c, added, reset);
2855 strbuf_addstr(&out, " bytes\n");
2856 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2857 out.buf, out.len, 0);
2858 strbuf_reset(&out);
2859 continue;
2860 }
2861 else if (file->is_unmerged) {
2862 strbuf_addf(&out, " %s%s%*s | %*s",
2863 prefix, name, padding, "",
2864 number_width, "Unmerged\n");
2865 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2866 out.buf, out.len, 0);
2867 strbuf_reset(&out);
2868 continue;
2869 }
2870
2871 /*
2872 * scale the add/delete
2873 */
2874 add = added;
2875 del = deleted;
2876
2877 if (graph_width <= max_change) {
2878 int total = scale_linear(add + del, graph_width, max_change);
2879 if (total < 2 && add && del)
2880 /* width >= 2 due to the sanity check */
2881 total = 2;
2882 if (add < del) {
2883 add = scale_linear(add, graph_width, max_change);
2884 del = total - add;
2885 } else {
2886 del = scale_linear(del, graph_width, max_change);
2887 add = total - del;
2888 }
2889 }
2890 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2891 prefix, name, padding, "",
2892 number_width, added + deleted,
2893 added + deleted ? " " : "");
2894 show_graph(&out, '+', add, add_c, reset);
2895 show_graph(&out, '-', del, del_c, reset);
2896 strbuf_addch(&out, '\n');
2897 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2898 out.buf, out.len, 0);
2899 strbuf_reset(&out);
2900 }
2901
2902 for (i = 0; i < data->nr; i++) {
2903 struct diffstat_file *file = data->files[i];
2904 uintmax_t added = file->added;
2905 uintmax_t deleted = file->deleted;
2906
2907 if (file->is_unmerged ||
2908 (!file->is_interesting && (added + deleted == 0))) {
2909 total_files--;
2910 continue;
2911 }
2912
2913 if (!file->is_binary) {
2914 adds += added;
2915 dels += deleted;
2916 }
2917 if (i < count)
2918 continue;
2919 if (!extra_shown)
2920 emit_diff_symbol(options,
2921 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2922 NULL, 0, 0);
2923 extra_shown = 1;
2924 }
2925
2926 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2927 strbuf_release(&out);
2928 }
2929
2930 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2931 {
2932 int i, adds = 0, dels = 0, total_files = data->nr;
2933
2934 if (data->nr == 0)
2935 return;
2936
2937 for (i = 0; i < data->nr; i++) {
2938 int added = data->files[i]->added;
2939 int deleted = data->files[i]->deleted;
2940
2941 if (data->files[i]->is_unmerged ||
2942 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2943 total_files--;
2944 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2945 adds += added;
2946 dels += deleted;
2947 }
2948 }
2949 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2950 }
2951
2952 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2953 {
2954 int i;
2955
2956 if (data->nr == 0)
2957 return;
2958
2959 for (i = 0; i < data->nr; i++) {
2960 struct diffstat_file *file = data->files[i];
2961
2962 fprintf(options->file, "%s", diff_line_prefix(options));
2963
2964 if (file->is_binary)
2965 fprintf(options->file, "-\t-\t");
2966 else
2967 fprintf(options->file,
2968 "%"PRIuMAX"\t%"PRIuMAX"\t",
2969 file->added, file->deleted);
2970 if (options->line_termination) {
2971 fill_print_name(file);
2972 if (!file->is_renamed)
2973 write_name_quoted(file->name, options->file,
2974 options->line_termination);
2975 else {
2976 fputs(file->print_name, options->file);
2977 putc(options->line_termination, options->file);
2978 }
2979 } else {
2980 if (file->is_renamed) {
2981 putc('\0', options->file);
2982 write_name_quoted(file->from_name, options->file, '\0');
2983 }
2984 write_name_quoted(file->name, options->file, '\0');
2985 }
2986 }
2987 }
2988
2989 struct dirstat_file {
2990 const char *name;
2991 unsigned long changed;
2992 };
2993
2994 struct dirstat_dir {
2995 struct dirstat_file *files;
2996 int alloc, nr, permille, cumulative;
2997 };
2998
2999 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
3000 unsigned long changed, const char *base, int baselen)
3001 {
3002 unsigned long sum_changes = 0;
3003 unsigned int sources = 0;
3004 const char *line_prefix = diff_line_prefix(opt);
3005
3006 while (dir->nr) {
3007 struct dirstat_file *f = dir->files;
3008 int namelen = strlen(f->name);
3009 unsigned long changes;
3010 char *slash;
3011
3012 if (namelen < baselen)
3013 break;
3014 if (memcmp(f->name, base, baselen))
3015 break;
3016 slash = strchr(f->name + baselen, '/');
3017 if (slash) {
3018 int newbaselen = slash + 1 - f->name;
3019 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
3020 sources++;
3021 } else {
3022 changes = f->changed;
3023 dir->files++;
3024 dir->nr--;
3025 sources += 2;
3026 }
3027 sum_changes += changes;
3028 }
3029
3030 /*
3031 * We don't report dirstat's for
3032 * - the top level
3033 * - or cases where everything came from a single directory
3034 * under this directory (sources == 1).
3035 */
3036 if (baselen && sources != 1) {
3037 if (sum_changes) {
3038 int permille = sum_changes * 1000 / changed;
3039 if (permille >= dir->permille) {
3040 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
3041 permille / 10, permille % 10, baselen, base);
3042 if (!dir->cumulative)
3043 return 0;
3044 }
3045 }
3046 }
3047 return sum_changes;
3048 }
3049
3050 static int dirstat_compare(const void *_a, const void *_b)
3051 {
3052 const struct dirstat_file *a = _a;
3053 const struct dirstat_file *b = _b;
3054 return strcmp(a->name, b->name);
3055 }
3056
3057 static void conclude_dirstat(struct diff_options *options,
3058 struct dirstat_dir *dir,
3059 unsigned long changed)
3060 {
3061 struct dirstat_file *to_free = dir->files;
3062
3063 if (!changed) {
3064 /* This can happen even with many files, if everything was renames */
3065 ;
3066 } else {
3067 /* Show all directories with more than x% of the changes */
3068 QSORT(dir->files, dir->nr, dirstat_compare);
3069 gather_dirstat(options, dir, changed, "", 0);
3070 }
3071
3072 free(to_free);
3073 }
3074
3075 static void show_dirstat(struct diff_options *options)
3076 {
3077 int i;
3078 unsigned long changed;
3079 struct dirstat_dir dir;
3080 struct diff_queue_struct *q = &diff_queued_diff;
3081
3082 dir.files = NULL;
3083 dir.alloc = 0;
3084 dir.nr = 0;
3085 dir.permille = options->dirstat_permille;
3086 dir.cumulative = options->flags.dirstat_cumulative;
3087
3088 changed = 0;
3089 for (i = 0; i < q->nr; i++) {
3090 struct diff_filepair *p = q->queue[i];
3091 const char *name;
3092 unsigned long copied, added, damage;
3093 struct diff_populate_filespec_options dpf_options = {
3094 .check_size_only = 1,
3095 };
3096
3097 name = p->two->path ? p->two->path : p->one->path;
3098
3099 if (p->one->oid_valid && p->two->oid_valid &&
3100 oideq(&p->one->oid, &p->two->oid)) {
3101 /*
3102 * The SHA1 has not changed, so pre-/post-content is
3103 * identical. We can therefore skip looking at the
3104 * file contents altogether.
3105 */
3106 damage = 0;
3107 goto found_damage;
3108 }
3109
3110 if (options->flags.dirstat_by_file) {
3111 /*
3112 * In --dirstat-by-file mode, we don't really need to
3113 * look at the actual file contents at all.
3114 * The fact that the SHA1 changed is enough for us to
3115 * add this file to the list of results
3116 * (with each file contributing equal damage).
3117 */
3118 damage = 1;
3119 goto found_damage;
3120 }
3121
3122 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3123 diff_populate_filespec(options->repo, p->one, NULL);
3124 diff_populate_filespec(options->repo, p->two, NULL);
3125 diffcore_count_changes(options->repo,
3126 p->one, p->two, NULL, NULL,
3127 &copied, &added);
3128 diff_free_filespec_data(p->one);
3129 diff_free_filespec_data(p->two);
3130 } else if (DIFF_FILE_VALID(p->one)) {
3131 diff_populate_filespec(options->repo, p->one, &dpf_options);
3132 copied = added = 0;
3133 diff_free_filespec_data(p->one);
3134 } else if (DIFF_FILE_VALID(p->two)) {
3135 diff_populate_filespec(options->repo, p->two, &dpf_options);
3136 copied = 0;
3137 added = p->two->size;
3138 diff_free_filespec_data(p->two);
3139 } else
3140 continue;
3141
3142 /*
3143 * Original minus copied is the removed material,
3144 * added is the new material. They are both damages
3145 * made to the preimage.
3146 * If the resulting damage is zero, we know that
3147 * diffcore_count_changes() considers the two entries to
3148 * be identical, but since the oid changed, we
3149 * know that there must have been _some_ kind of change,
3150 * so we force all entries to have damage > 0.
3151 */
3152 damage = (p->one->size - copied) + added;
3153 if (!damage)
3154 damage = 1;
3155
3156 found_damage:
3157 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3158 dir.files[dir.nr].name = name;
3159 dir.files[dir.nr].changed = damage;
3160 changed += damage;
3161 dir.nr++;
3162 }
3163
3164 conclude_dirstat(options, &dir, changed);
3165 }
3166
3167 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3168 {
3169 int i;
3170 unsigned long changed;
3171 struct dirstat_dir dir;
3172
3173 if (data->nr == 0)
3174 return;
3175
3176 dir.files = NULL;
3177 dir.alloc = 0;
3178 dir.nr = 0;
3179 dir.permille = options->dirstat_permille;
3180 dir.cumulative = options->flags.dirstat_cumulative;
3181
3182 changed = 0;
3183 for (i = 0; i < data->nr; i++) {
3184 struct diffstat_file *file = data->files[i];
3185 unsigned long damage = file->added + file->deleted;
3186 if (file->is_binary)
3187 /*
3188 * binary files counts bytes, not lines. Must find some
3189 * way to normalize binary bytes vs. textual lines.
3190 * The following heuristic assumes that there are 64
3191 * bytes per "line".
3192 * This is stupid and ugly, but very cheap...
3193 */
3194 damage = DIV_ROUND_UP(damage, 64);
3195 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3196 dir.files[dir.nr].name = file->name;
3197 dir.files[dir.nr].changed = damage;
3198 changed += damage;
3199 dir.nr++;
3200 }
3201
3202 conclude_dirstat(options, &dir, changed);
3203 }
3204
3205 static void free_diffstat_file(struct diffstat_file *f)
3206 {
3207 free(f->print_name);
3208 free(f->name);
3209 free(f->from_name);
3210 free(f);
3211 }
3212
3213 void free_diffstat_info(struct diffstat_t *diffstat)
3214 {
3215 int i;
3216 for (i = 0; i < diffstat->nr; i++)
3217 free_diffstat_file(diffstat->files[i]);
3218 free(diffstat->files);
3219 }
3220
3221 struct checkdiff_t {
3222 const char *filename;
3223 int lineno;
3224 int conflict_marker_size;
3225 struct diff_options *o;
3226 unsigned ws_rule;
3227 unsigned status;
3228 };
3229
3230 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3231 {
3232 char firstchar;
3233 int cnt;
3234
3235 if (len < marker_size + 1)
3236 return 0;
3237 firstchar = line[0];
3238 switch (firstchar) {
3239 case '=': case '>': case '<': case '|':
3240 break;
3241 default:
3242 return 0;
3243 }
3244 for (cnt = 1; cnt < marker_size; cnt++)
3245 if (line[cnt] != firstchar)
3246 return 0;
3247 /* line[1] through line[marker_size-1] are same as firstchar */
3248 if (len < marker_size + 1 || !isspace(line[marker_size]))
3249 return 0;
3250 return 1;
3251 }
3252
3253 static void checkdiff_consume_hunk(void *priv,
3254 long ob UNUSED, long on UNUSED,
3255 long nb, long nn UNUSED,
3256 const char *func UNUSED, long funclen UNUSED)
3257
3258 {
3259 struct checkdiff_t *data = priv;
3260 data->lineno = nb - 1;
3261 }
3262
3263 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3264 {
3265 struct checkdiff_t *data = priv;
3266 int marker_size = data->conflict_marker_size;
3267 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3268 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3269 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3270 char *err;
3271 const char *line_prefix;
3272
3273 assert(data->o);
3274 line_prefix = diff_line_prefix(data->o);
3275
3276 if (line[0] == '+') {
3277 unsigned bad;
3278 data->lineno++;
3279 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3280 data->status |= 1;
3281 fprintf(data->o->file,
3282 "%s%s:%d: leftover conflict marker\n",
3283 line_prefix, data->filename, data->lineno);
3284 }
3285 bad = ws_check(line + 1, len - 1, data->ws_rule);
3286 if (!bad)
3287 return 0;
3288 data->status |= bad;
3289 err = whitespace_error_string(bad);
3290 fprintf(data->o->file, "%s%s:%d: %s.\n",
3291 line_prefix, data->filename, data->lineno, err);
3292 free(err);
3293 emit_line(data->o, set, reset, line, 1);
3294 ws_check_emit(line + 1, len - 1, data->ws_rule,
3295 data->o->file, set, reset, ws);
3296 } else if (line[0] == ' ') {
3297 data->lineno++;
3298 }
3299 return 0;
3300 }
3301
3302 static unsigned char *deflate_it(char *data,
3303 unsigned long size,
3304 unsigned long *result_size)
3305 {
3306 int bound;
3307 unsigned char *deflated;
3308 git_zstream stream;
3309
3310 git_deflate_init(&stream, zlib_compression_level);
3311 bound = git_deflate_bound(&stream, size);
3312 deflated = xmalloc(bound);
3313 stream.next_out = deflated;
3314 stream.avail_out = bound;
3315
3316 stream.next_in = (unsigned char *)data;
3317 stream.avail_in = size;
3318 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3319 ; /* nothing */
3320 git_deflate_end(&stream);
3321 *result_size = stream.total_out;
3322 return deflated;
3323 }
3324
3325 static void emit_binary_diff_body(struct diff_options *o,
3326 mmfile_t *one, mmfile_t *two)
3327 {
3328 void *cp;
3329 void *delta;
3330 void *deflated;
3331 void *data;
3332 unsigned long orig_size;
3333 unsigned long delta_size;
3334 unsigned long deflate_size;
3335 unsigned long data_size;
3336
3337 /* We could do deflated delta, or we could do just deflated two,
3338 * whichever is smaller.
3339 */
3340 delta = NULL;
3341 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3342 if (one->size && two->size) {
3343 delta = diff_delta(one->ptr, one->size,
3344 two->ptr, two->size,
3345 &delta_size, deflate_size);
3346 if (delta) {
3347 void *to_free = delta;
3348 orig_size = delta_size;
3349 delta = deflate_it(delta, delta_size, &delta_size);
3350 free(to_free);
3351 }
3352 }
3353
3354 if (delta && delta_size < deflate_size) {
3355 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3356 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3357 s, strlen(s), 0);
3358 free(s);
3359 free(deflated);
3360 data = delta;
3361 data_size = delta_size;
3362 } else {
3363 char *s = xstrfmt("%lu", two->size);
3364 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3365 s, strlen(s), 0);
3366 free(s);
3367 free(delta);
3368 data = deflated;
3369 data_size = deflate_size;
3370 }
3371
3372 /* emit data encoded in base85 */
3373 cp = data;
3374 while (data_size) {
3375 int len;
3376 int bytes = (52 < data_size) ? 52 : data_size;
3377 char line[71];
3378 data_size -= bytes;
3379 if (bytes <= 26)
3380 line[0] = bytes + 'A' - 1;
3381 else
3382 line[0] = bytes - 26 + 'a' - 1;
3383 encode_85(line + 1, cp, bytes);
3384 cp = (char *) cp + bytes;
3385
3386 len = strlen(line);
3387 line[len++] = '\n';
3388 line[len] = '\0';
3389
3390 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3391 line, len, 0);
3392 }
3393 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3394 free(data);
3395 }
3396
3397 static void emit_binary_diff(struct diff_options *o,
3398 mmfile_t *one, mmfile_t *two)
3399 {
3400 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3401 emit_binary_diff_body(o, one, two);
3402 emit_binary_diff_body(o, two, one);
3403 }
3404
3405 int diff_filespec_is_binary(struct repository *r,
3406 struct diff_filespec *one)
3407 {
3408 struct diff_populate_filespec_options dpf_options = {
3409 .check_binary = 1,
3410 };
3411
3412 if (one->is_binary == -1) {
3413 diff_filespec_load_driver(one, r->index);
3414 if (one->driver->binary != -1)
3415 one->is_binary = one->driver->binary;
3416 else {
3417 if (!one->data && DIFF_FILE_VALID(one))
3418 diff_populate_filespec(r, one, &dpf_options);
3419 if (one->is_binary == -1 && one->data)
3420 one->is_binary = buffer_is_binary(one->data,
3421 one->size);
3422 if (one->is_binary == -1)
3423 one->is_binary = 0;
3424 }
3425 }
3426 return one->is_binary;
3427 }
3428
3429 static const struct userdiff_funcname *
3430 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3431 {
3432 diff_filespec_load_driver(one, o->repo->index);
3433 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3434 }
3435
3436 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3437 {
3438 if (!options->a_prefix)
3439 options->a_prefix = a;
3440 if (!options->b_prefix)
3441 options->b_prefix = b;
3442 }
3443
3444 void diff_set_noprefix(struct diff_options *options)
3445 {
3446 options->a_prefix = options->b_prefix = "";
3447 }
3448
3449 void diff_set_default_prefix(struct diff_options *options)
3450 {
3451 options->a_prefix = diff_src_prefix ? diff_src_prefix : "a/";
3452 options->b_prefix = diff_dst_prefix ? diff_dst_prefix : "b/";
3453 }
3454
3455 struct userdiff_driver *get_textconv(struct repository *r,
3456 struct diff_filespec *one)
3457 {
3458 if (!DIFF_FILE_VALID(one))
3459 return NULL;
3460
3461 diff_filespec_load_driver(one, r->index);
3462 return userdiff_get_textconv(r, one->driver);
3463 }
3464
3465 static struct string_list *additional_headers(struct diff_options *o,
3466 const char *path)
3467 {
3468 if (!o->additional_path_headers)
3469 return NULL;
3470 return strmap_get(o->additional_path_headers, path);
3471 }
3472
3473 static void add_formatted_header(struct strbuf *msg,
3474 const char *header,
3475 const char *line_prefix,
3476 const char *meta,
3477 const char *reset)
3478 {
3479 const char *next, *newline;
3480
3481 for (next = header; *next; next = newline) {
3482 newline = strchrnul(next, '\n');
3483 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3484 (int)(newline - next), next, reset);
3485 if (*newline)
3486 newline++;
3487 }
3488 }
3489
3490 static void add_formatted_headers(struct strbuf *msg,
3491 struct string_list *more_headers,
3492 const char *line_prefix,
3493 const char *meta,
3494 const char *reset)
3495 {
3496 int i;
3497
3498 for (i = 0; i < more_headers->nr; i++)
3499 add_formatted_header(msg, more_headers->items[i].string,
3500 line_prefix, meta, reset);
3501 }
3502
3503 static int diff_filepair_is_phoney(struct diff_filespec *one,
3504 struct diff_filespec *two)
3505 {
3506 /*
3507 * This function specifically looks for pairs injected by
3508 * create_filepairs_for_header_only_notifications(). Such
3509 * pairs are "phoney" in that they do not represent any
3510 * content or even mode difference, but were inserted because
3511 * diff_queued_diff previously had no pair associated with
3512 * that path but we needed some pair to avoid losing the
3513 * "remerge CONFLICT" header associated with the path.
3514 */
3515 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3516 }
3517
3518 static int set_diff_algorithm(struct diff_options *opts,
3519 const char *alg)
3520 {
3521 long value = parse_algorithm_value(alg);
3522
3523 if (value < 0)
3524 return -1;
3525
3526 /* clear out previous settings */
3527 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3528 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3529 opts->xdl_opts |= value;
3530
3531 return 0;
3532 }
3533
3534 static void builtin_diff(const char *name_a,
3535 const char *name_b,
3536 struct diff_filespec *one,
3537 struct diff_filespec *two,
3538 const char *xfrm_msg,
3539 int must_show_header,
3540 struct diff_options *o,
3541 int complete_rewrite)
3542 {
3543 mmfile_t mf1, mf2;
3544 const char *lbl[2];
3545 char *a_one, *b_two;
3546 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3547 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3548 const char *a_prefix, *b_prefix;
3549 struct userdiff_driver *textconv_one = NULL;
3550 struct userdiff_driver *textconv_two = NULL;
3551 struct strbuf header = STRBUF_INIT;
3552 const char *line_prefix = diff_line_prefix(o);
3553
3554 diff_set_mnemonic_prefix(o, "a/", "b/");
3555 if (o->flags.reverse_diff) {
3556 a_prefix = o->b_prefix;
3557 b_prefix = o->a_prefix;
3558 } else {
3559 a_prefix = o->a_prefix;
3560 b_prefix = o->b_prefix;
3561 }
3562
3563 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3564 (!one->mode || S_ISGITLINK(one->mode)) &&
3565 (!two->mode || S_ISGITLINK(two->mode)) &&
3566 (!diff_filepair_is_phoney(one, two))) {
3567 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3568 &one->oid, &two->oid,
3569 two->dirty_submodule);
3570 o->found_changes = 1;
3571 return;
3572 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3573 (!one->mode || S_ISGITLINK(one->mode)) &&
3574 (!two->mode || S_ISGITLINK(two->mode)) &&
3575 (!diff_filepair_is_phoney(one, two))) {
3576 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3577 &one->oid, &two->oid,
3578 two->dirty_submodule);
3579 o->found_changes = 1;
3580 return;
3581 }
3582
3583 if (o->flags.allow_textconv) {
3584 textconv_one = get_textconv(o->repo, one);
3585 textconv_two = get_textconv(o->repo, two);
3586 }
3587
3588 /* Never use a non-valid filename anywhere if at all possible */
3589 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3590 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3591
3592 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3593 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3594 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3595 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3596 if (diff_filepair_is_phoney(one, two)) {
3597 /*
3598 * We should only reach this point for pairs generated from
3599 * create_filepairs_for_header_only_notifications(). For
3600 * these, we want to avoid the "/dev/null" special casing
3601 * above, because we do not want such pairs shown as either
3602 * "new file" or "deleted file" below.
3603 */
3604 lbl[0] = a_one;
3605 lbl[1] = b_two;
3606 }
3607 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3608 if (lbl[0][0] == '/') {
3609 /* /dev/null */
3610 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3611 if (xfrm_msg)
3612 strbuf_addstr(&header, xfrm_msg);
3613 o->found_changes = 1;
3614 must_show_header = 1;
3615 }
3616 else if (lbl[1][0] == '/') {
3617 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3618 if (xfrm_msg)
3619 strbuf_addstr(&header, xfrm_msg);
3620 o->found_changes = 1;
3621 must_show_header = 1;
3622 }
3623 else {
3624 if (one->mode != two->mode) {
3625 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3626 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3627 o->found_changes = 1;
3628 must_show_header = 1;
3629 }
3630 if (xfrm_msg)
3631 strbuf_addstr(&header, xfrm_msg);
3632
3633 /*
3634 * we do not run diff between different kind
3635 * of objects.
3636 */
3637 if ((one->mode ^ two->mode) & S_IFMT)
3638 goto free_ab_and_return;
3639 if (complete_rewrite &&
3640 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3641 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3642 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3643 header.buf, header.len, 0);
3644 strbuf_reset(&header);
3645 emit_rewrite_diff(name_a, name_b, one, two,
3646 textconv_one, textconv_two, o);
3647 o->found_changes = 1;
3648 goto free_ab_and_return;
3649 }
3650 }
3651
3652 if (o->irreversible_delete && lbl[1][0] == '/') {
3653 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3654 header.len, 0);
3655 strbuf_reset(&header);
3656 goto free_ab_and_return;
3657 } else if (!o->flags.text &&
3658 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3659 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3660 struct strbuf sb = STRBUF_INIT;
3661 if (!one->data && !two->data &&
3662 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3663 !o->flags.binary) {
3664 if (oideq(&one->oid, &two->oid)) {
3665 if (must_show_header)
3666 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3667 header.buf, header.len,
3668 0);
3669 goto free_ab_and_return;
3670 }
3671 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3672 header.buf, header.len, 0);
3673 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3674 diff_line_prefix(o), lbl[0], lbl[1]);
3675 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3676 sb.buf, sb.len, 0);
3677 strbuf_release(&sb);
3678 goto free_ab_and_return;
3679 }
3680 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3681 fill_mmfile(o->repo, &mf2, two) < 0)
3682 die("unable to read files to diff");
3683 /* Quite common confusing case */
3684 if (mf1.size == mf2.size &&
3685 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3686 if (must_show_header)
3687 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3688 header.buf, header.len, 0);
3689 goto free_ab_and_return;
3690 }
3691 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3692 strbuf_reset(&header);
3693 if (o->flags.binary)
3694 emit_binary_diff(o, &mf1, &mf2);
3695 else {
3696 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3697 diff_line_prefix(o), lbl[0], lbl[1]);
3698 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3699 sb.buf, sb.len, 0);
3700 strbuf_release(&sb);
3701 }
3702 o->found_changes = 1;
3703 } else {
3704 /* Crazy xdl interfaces.. */
3705 const char *diffopts;
3706 const char *v;
3707 xpparam_t xpp;
3708 xdemitconf_t xecfg;
3709 struct emit_callback ecbdata;
3710 const struct userdiff_funcname *pe;
3711
3712 if (must_show_header) {
3713 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3714 header.buf, header.len, 0);
3715 strbuf_reset(&header);
3716 }
3717
3718 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3719 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3720
3721 pe = diff_funcname_pattern(o, one);
3722 if (!pe)
3723 pe = diff_funcname_pattern(o, two);
3724
3725 memset(&xpp, 0, sizeof(xpp));
3726 memset(&xecfg, 0, sizeof(xecfg));
3727 memset(&ecbdata, 0, sizeof(ecbdata));
3728 if (o->flags.suppress_diff_headers)
3729 lbl[0] = NULL;
3730 ecbdata.label_path = lbl;
3731 ecbdata.color_diff = want_color(o->use_color);
3732 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3733 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3734 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3735 ecbdata.opt = o;
3736 if (header.len && !o->flags.suppress_diff_headers)
3737 ecbdata.header = &header;
3738 xpp.flags = o->xdl_opts;
3739 xpp.ignore_regex = o->ignore_regex;
3740 xpp.ignore_regex_nr = o->ignore_regex_nr;
3741 xpp.anchors = o->anchors;
3742 xpp.anchors_nr = o->anchors_nr;
3743 xecfg.ctxlen = o->context;
3744 xecfg.interhunkctxlen = o->interhunkcontext;
3745 xecfg.flags = XDL_EMIT_FUNCNAMES;
3746 if (o->flags.funccontext)
3747 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3748 if (pe)
3749 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3750
3751 diffopts = getenv("GIT_DIFF_OPTS");
3752 if (!diffopts)
3753 ;
3754 else if (skip_prefix(diffopts, "--unified=", &v))
3755 xecfg.ctxlen = strtoul(v, NULL, 10);
3756 else if (skip_prefix(diffopts, "-u", &v))
3757 xecfg.ctxlen = strtoul(v, NULL, 10);
3758
3759 if (o->word_diff)
3760 init_diff_words_data(&ecbdata, o, one, two);
3761 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3762 &ecbdata, &xpp, &xecfg))
3763 die("unable to generate diff for %s", one->path);
3764 if (o->word_diff)
3765 free_diff_words_data(&ecbdata);
3766 if (textconv_one)
3767 free(mf1.ptr);
3768 if (textconv_two)
3769 free(mf2.ptr);
3770 xdiff_clear_find_func(&xecfg);
3771 }
3772
3773 free_ab_and_return:
3774 strbuf_release(&header);
3775 diff_free_filespec_data(one);
3776 diff_free_filespec_data(two);
3777 free(a_one);
3778 free(b_two);
3779 return;
3780 }
3781
3782 static const char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3783 {
3784 if (!is_renamed) {
3785 if (p->status == DIFF_STATUS_ADDED) {
3786 if (S_ISLNK(p->two->mode))
3787 return "new +l";
3788 else if ((p->two->mode & 0777) == 0755)
3789 return "new +x";
3790 else
3791 return "new";
3792 } else if (p->status == DIFF_STATUS_DELETED)
3793 return "gone";
3794 }
3795 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3796 return "mode -l";
3797 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3798 return "mode +l";
3799 else if ((p->one->mode & 0777) == 0644 &&
3800 (p->two->mode & 0777) == 0755)
3801 return "mode +x";
3802 else if ((p->one->mode & 0777) == 0755 &&
3803 (p->two->mode & 0777) == 0644)
3804 return "mode -x";
3805 return NULL;
3806 }
3807
3808 static void builtin_diffstat(const char *name_a, const char *name_b,
3809 struct diff_filespec *one,
3810 struct diff_filespec *two,
3811 struct diffstat_t *diffstat,
3812 struct diff_options *o,
3813 struct diff_filepair *p)
3814 {
3815 mmfile_t mf1, mf2;
3816 struct diffstat_file *data;
3817 int may_differ;
3818 int complete_rewrite = 0;
3819
3820 if (!DIFF_PAIR_UNMERGED(p)) {
3821 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3822 complete_rewrite = 1;
3823 }
3824
3825 data = diffstat_add(diffstat, name_a, name_b);
3826 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3827 if (o->flags.stat_with_summary)
3828 data->comments = get_compact_summary(p, data->is_renamed);
3829
3830 if (!one || !two) {
3831 data->is_unmerged = 1;
3832 return;
3833 }
3834
3835 /* saves some reads if true, not a guarantee of diff outcome */
3836 may_differ = !(one->oid_valid && two->oid_valid &&
3837 oideq(&one->oid, &two->oid));
3838
3839 if (diff_filespec_is_binary(o->repo, one) ||
3840 diff_filespec_is_binary(o->repo, two)) {
3841 data->is_binary = 1;
3842 if (!may_differ) {
3843 data->added = 0;
3844 data->deleted = 0;
3845 } else {
3846 data->added = diff_filespec_size(o->repo, two);
3847 data->deleted = diff_filespec_size(o->repo, one);
3848 }
3849 }
3850
3851 else if (complete_rewrite) {
3852 diff_populate_filespec(o->repo, one, NULL);
3853 diff_populate_filespec(o->repo, two, NULL);
3854 data->deleted = count_lines(one->data, one->size);
3855 data->added = count_lines(two->data, two->size);
3856 }
3857
3858 else if (may_differ) {
3859 /* Crazy xdl interfaces.. */
3860 xpparam_t xpp;
3861 xdemitconf_t xecfg;
3862
3863 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3864 fill_mmfile(o->repo, &mf2, two) < 0)
3865 die("unable to read files to diff");
3866
3867 memset(&xpp, 0, sizeof(xpp));
3868 memset(&xecfg, 0, sizeof(xecfg));
3869 xpp.flags = o->xdl_opts;
3870 xpp.ignore_regex = o->ignore_regex;
3871 xpp.ignore_regex_nr = o->ignore_regex_nr;
3872 xpp.anchors = o->anchors;
3873 xpp.anchors_nr = o->anchors_nr;
3874 xecfg.ctxlen = o->context;
3875 xecfg.interhunkctxlen = o->interhunkcontext;
3876 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3877 if (xdi_diff_outf(&mf1, &mf2, NULL,
3878 diffstat_consume, diffstat, &xpp, &xecfg))
3879 die("unable to generate diffstat for %s", one->path);
3880
3881 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3882 struct diffstat_file *file =
3883 diffstat->files[diffstat->nr - 1];
3884 /*
3885 * Omit diffstats of modified files where nothing changed.
3886 * Even if may_differ, this might be the case due to
3887 * ignoring whitespace changes, etc.
3888 *
3889 * But note that we special-case additions, deletions,
3890 * renames, and mode changes as adding an empty file,
3891 * for example is still of interest.
3892 */
3893 if ((p->status == DIFF_STATUS_MODIFIED)
3894 && !file->added
3895 && !file->deleted
3896 && one->mode == two->mode) {
3897 free_diffstat_file(file);
3898 diffstat->nr--;
3899 }
3900 }
3901 }
3902
3903 diff_free_filespec_data(one);
3904 diff_free_filespec_data(two);
3905 }
3906
3907 static void builtin_checkdiff(const char *name_a, const char *name_b,
3908 const char *attr_path,
3909 struct diff_filespec *one,
3910 struct diff_filespec *two,
3911 struct diff_options *o)
3912 {
3913 mmfile_t mf1, mf2;
3914 struct checkdiff_t data;
3915
3916 if (!two)
3917 return;
3918
3919 memset(&data, 0, sizeof(data));
3920 data.filename = name_b ? name_b : name_a;
3921 data.lineno = 0;
3922 data.o = o;
3923 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3924 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3925
3926 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3927 fill_mmfile(o->repo, &mf2, two) < 0)
3928 die("unable to read files to diff");
3929
3930 /*
3931 * All the other codepaths check both sides, but not checking
3932 * the "old" side here is deliberate. We are checking the newly
3933 * introduced changes, and as long as the "new" side is text, we
3934 * can and should check what it introduces.
3935 */
3936 if (diff_filespec_is_binary(o->repo, two))
3937 goto free_and_return;
3938 else {
3939 /* Crazy xdl interfaces.. */
3940 xpparam_t xpp;
3941 xdemitconf_t xecfg;
3942
3943 memset(&xpp, 0, sizeof(xpp));
3944 memset(&xecfg, 0, sizeof(xecfg));
3945 xecfg.ctxlen = 1; /* at least one context line */
3946 xpp.flags = 0;
3947 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3948 checkdiff_consume, &data,
3949 &xpp, &xecfg))
3950 die("unable to generate checkdiff for %s", one->path);
3951
3952 if (data.ws_rule & WS_BLANK_AT_EOF) {
3953 struct emit_callback ecbdata;
3954 int blank_at_eof;
3955
3956 ecbdata.ws_rule = data.ws_rule;
3957 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3958 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3959
3960 if (blank_at_eof) {
3961 static char *err;
3962 if (!err)
3963 err = whitespace_error_string(WS_BLANK_AT_EOF);
3964 fprintf(o->file, "%s:%d: %s.\n",
3965 data.filename, blank_at_eof, err);
3966 data.status = 1; /* report errors */
3967 }
3968 }
3969 }
3970 free_and_return:
3971 diff_free_filespec_data(one);
3972 diff_free_filespec_data(two);
3973 if (data.status)
3974 o->flags.check_failed = 1;
3975 }
3976
3977 struct diff_filespec *alloc_filespec(const char *path)
3978 {
3979 struct diff_filespec *spec;
3980
3981 FLEXPTR_ALLOC_STR(spec, path, path);
3982 spec->count = 1;
3983 spec->is_binary = -1;
3984 return spec;
3985 }
3986
3987 void free_filespec(struct diff_filespec *spec)
3988 {
3989 if (!--spec->count) {
3990 diff_free_filespec_data(spec);
3991 free(spec);
3992 }
3993 }
3994
3995 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3996 int oid_valid, unsigned short mode)
3997 {
3998 if (mode) {
3999 spec->mode = canon_mode(mode);
4000 oidcpy(&spec->oid, oid);
4001 spec->oid_valid = oid_valid;
4002 }
4003 }
4004
4005 /*
4006 * Given a name and sha1 pair, if the index tells us the file in
4007 * the work tree has that object contents, return true, so that
4008 * prepare_temp_file() does not have to inflate and extract.
4009 */
4010 static int reuse_worktree_file(struct index_state *istate,
4011 const char *name,
4012 const struct object_id *oid,
4013 int want_file)
4014 {
4015 const struct cache_entry *ce;
4016 struct stat st;
4017 int pos, len;
4018
4019 /*
4020 * We do not read the cache ourselves here, because the
4021 * benchmark with my previous version that always reads cache
4022 * shows that it makes things worse for diff-tree comparing
4023 * two linux-2.6 kernel trees in an already checked out work
4024 * tree. This is because most diff-tree comparisons deal with
4025 * only a small number of files, while reading the cache is
4026 * expensive for a large project, and its cost outweighs the
4027 * savings we get by not inflating the object to a temporary
4028 * file. Practically, this code only helps when we are used
4029 * by diff-cache --cached, which does read the cache before
4030 * calling us.
4031 */
4032 if (!istate->cache)
4033 return 0;
4034
4035 /* We want to avoid the working directory if our caller
4036 * doesn't need the data in a normal file, this system
4037 * is rather slow with its stat/open/mmap/close syscalls,
4038 * and the object is contained in a pack file. The pack
4039 * is probably already open and will be faster to obtain
4040 * the data through than the working directory. Loose
4041 * objects however would tend to be slower as they need
4042 * to be individually opened and inflated.
4043 */
4044 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
4045 return 0;
4046
4047 /*
4048 * Similarly, if we'd have to convert the file contents anyway, that
4049 * makes the optimization not worthwhile.
4050 */
4051 if (!want_file && would_convert_to_git(istate, name))
4052 return 0;
4053
4054 /*
4055 * If this path does not match our sparse-checkout definition,
4056 * then the file will not be in the working directory.
4057 */
4058 if (!path_in_sparse_checkout(name, istate))
4059 return 0;
4060
4061 len = strlen(name);
4062 pos = index_name_pos(istate, name, len);
4063 if (pos < 0)
4064 return 0;
4065 ce = istate->cache[pos];
4066
4067 /*
4068 * This is not the sha1 we are looking for, or
4069 * unreusable because it is not a regular file.
4070 */
4071 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4072 return 0;
4073
4074 /*
4075 * If ce is marked as "assume unchanged", there is no
4076 * guarantee that work tree matches what we are looking for.
4077 */
4078 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4079 return 0;
4080
4081 /*
4082 * If ce matches the file in the work tree, we can reuse it.
4083 */
4084 if (ce_uptodate(ce) ||
4085 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4086 return 1;
4087
4088 return 0;
4089 }
4090
4091 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4092 {
4093 struct strbuf buf = STRBUF_INIT;
4094 const char *dirty = "";
4095
4096 /* Are we looking at the work tree? */
4097 if (s->dirty_submodule)
4098 dirty = "-dirty";
4099
4100 strbuf_addf(&buf, "Subproject commit %s%s\n",
4101 oid_to_hex(&s->oid), dirty);
4102 s->size = buf.len;
4103 if (size_only) {
4104 s->data = NULL;
4105 strbuf_release(&buf);
4106 } else {
4107 s->data = strbuf_detach(&buf, NULL);
4108 s->should_free = 1;
4109 }
4110 return 0;
4111 }
4112
4113 /*
4114 * While doing rename detection and pickaxe operation, we may need to
4115 * grab the data for the blob (or file) for our own in-core comparison.
4116 * diff_filespec has data and size fields for this purpose.
4117 */
4118 int diff_populate_filespec(struct repository *r,
4119 struct diff_filespec *s,
4120 const struct diff_populate_filespec_options *options)
4121 {
4122 int size_only = options ? options->check_size_only : 0;
4123 int check_binary = options ? options->check_binary : 0;
4124 int err = 0;
4125 int conv_flags = global_conv_flags_eol;
4126 /*
4127 * demote FAIL to WARN to allow inspecting the situation
4128 * instead of refusing.
4129 */
4130 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4131 conv_flags = CONV_EOL_RNDTRP_WARN;
4132
4133 if (!DIFF_FILE_VALID(s))
4134 die("internal error: asking to populate invalid file.");
4135 if (S_ISDIR(s->mode))
4136 return -1;
4137
4138 if (s->data)
4139 return 0;
4140
4141 if (size_only && 0 < s->size)
4142 return 0;
4143
4144 if (S_ISGITLINK(s->mode))
4145 return diff_populate_gitlink(s, size_only);
4146
4147 if (!s->oid_valid ||
4148 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4149 struct strbuf buf = STRBUF_INIT;
4150 struct stat st;
4151 int fd;
4152
4153 if (lstat(s->path, &st) < 0) {
4154 err_empty:
4155 err = -1;
4156 empty:
4157 s->data = (char *)"";
4158 s->size = 0;
4159 return err;
4160 }
4161 s->size = xsize_t(st.st_size);
4162 if (!s->size)
4163 goto empty;
4164 if (S_ISLNK(st.st_mode)) {
4165 struct strbuf sb = STRBUF_INIT;
4166
4167 if (strbuf_readlink(&sb, s->path, s->size))
4168 goto err_empty;
4169 s->size = sb.len;
4170 s->data = strbuf_detach(&sb, NULL);
4171 s->should_free = 1;
4172 return 0;
4173 }
4174
4175 /*
4176 * Even if the caller would be happy with getting
4177 * only the size, we cannot return early at this
4178 * point if the path requires us to run the content
4179 * conversion.
4180 */
4181 if (size_only && !would_convert_to_git(r->index, s->path))
4182 return 0;
4183
4184 /*
4185 * Note: this check uses xsize_t(st.st_size) that may
4186 * not be the true size of the blob after it goes
4187 * through convert_to_git(). This may not strictly be
4188 * correct, but the whole point of big_file_threshold
4189 * and is_binary check being that we want to avoid
4190 * opening the file and inspecting the contents, this
4191 * is probably fine.
4192 */
4193 if (check_binary &&
4194 s->size > big_file_threshold && s->is_binary == -1) {
4195 s->is_binary = 1;
4196 return 0;
4197 }
4198 fd = open(s->path, O_RDONLY);
4199 if (fd < 0)
4200 goto err_empty;
4201 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4202 close(fd);
4203 s->should_munmap = 1;
4204
4205 /*
4206 * Convert from working tree format to canonical git format
4207 */
4208 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4209 size_t size = 0;
4210 munmap(s->data, s->size);
4211 s->should_munmap = 0;
4212 s->data = strbuf_detach(&buf, &size);
4213 s->size = size;
4214 s->should_free = 1;
4215 }
4216 }
4217 else {
4218 struct object_info info = {
4219 .sizep = &s->size
4220 };
4221
4222 if (!(size_only || check_binary))
4223 /*
4224 * Set contentp, since there is no chance that merely
4225 * the size is sufficient.
4226 */
4227 info.contentp = &s->data;
4228
4229 if (options && options->missing_object_cb) {
4230 if (!oid_object_info_extended(r, &s->oid, &info,
4231 OBJECT_INFO_LOOKUP_REPLACE |
4232 OBJECT_INFO_SKIP_FETCH_OBJECT))
4233 goto object_read;
4234 options->missing_object_cb(options->missing_object_data);
4235 }
4236 if (oid_object_info_extended(r, &s->oid, &info,
4237 OBJECT_INFO_LOOKUP_REPLACE))
4238 die("unable to read %s", oid_to_hex(&s->oid));
4239
4240 object_read:
4241 if (size_only || check_binary) {
4242 if (size_only)
4243 return 0;
4244 if (s->size > big_file_threshold && s->is_binary == -1) {
4245 s->is_binary = 1;
4246 return 0;
4247 }
4248 }
4249 if (!info.contentp) {
4250 info.contentp = &s->data;
4251 if (oid_object_info_extended(r, &s->oid, &info,
4252 OBJECT_INFO_LOOKUP_REPLACE))
4253 die("unable to read %s", oid_to_hex(&s->oid));
4254 }
4255 s->should_free = 1;
4256 }
4257 return 0;
4258 }
4259
4260 void diff_free_filespec_blob(struct diff_filespec *s)
4261 {
4262 if (s->should_free)
4263 free(s->data);
4264 else if (s->should_munmap)
4265 munmap(s->data, s->size);
4266
4267 if (s->should_free || s->should_munmap) {
4268 s->should_free = s->should_munmap = 0;
4269 s->data = NULL;
4270 }
4271 }
4272
4273 void diff_free_filespec_data(struct diff_filespec *s)
4274 {
4275 if (!s)
4276 return;
4277
4278 diff_free_filespec_blob(s);
4279 FREE_AND_NULL(s->cnt_data);
4280 }
4281
4282 static void prep_temp_blob(struct index_state *istate,
4283 const char *path, struct diff_tempfile *temp,
4284 void *blob,
4285 unsigned long size,
4286 const struct object_id *oid,
4287 int mode)
4288 {
4289 struct strbuf buf = STRBUF_INIT;
4290 char *path_dup = xstrdup(path);
4291 const char *base = basename(path_dup);
4292 struct checkout_metadata meta;
4293
4294 init_checkout_metadata(&meta, NULL, NULL, oid);
4295
4296 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4297 if (!temp->tempfile)
4298 die_errno("unable to create temp-file");
4299 if (convert_to_working_tree(istate, path,
4300 (const char *)blob, (size_t)size, &buf, &meta)) {
4301 blob = buf.buf;
4302 size = buf.len;
4303 }
4304 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4305 close_tempfile_gently(temp->tempfile))
4306 die_errno("unable to write temp-file");
4307 temp->name = get_tempfile_path(temp->tempfile);
4308 oid_to_hex_r(temp->hex, oid);
4309 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4310 strbuf_release(&buf);
4311 free(path_dup);
4312 }
4313
4314 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4315 struct diff_filespec *one)
4316 {
4317 struct diff_tempfile *temp = claim_diff_tempfile();
4318
4319 if (!DIFF_FILE_VALID(one)) {
4320 not_a_valid_file:
4321 /* A '-' entry produces this for file-2, and
4322 * a '+' entry produces this for file-1.
4323 */
4324 temp->name = "/dev/null";
4325 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4326 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4327 return temp;
4328 }
4329
4330 if (!S_ISGITLINK(one->mode) &&
4331 (!one->oid_valid ||
4332 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4333 struct stat st;
4334 if (lstat(one->path, &st) < 0) {
4335 if (errno == ENOENT)
4336 goto not_a_valid_file;
4337 die_errno("stat(%s)", one->path);
4338 }
4339 if (S_ISLNK(st.st_mode)) {
4340 struct strbuf sb = STRBUF_INIT;
4341 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4342 die_errno("readlink(%s)", one->path);
4343 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4344 (one->oid_valid ?
4345 &one->oid : null_oid()),
4346 (one->oid_valid ?
4347 one->mode : S_IFLNK));
4348 strbuf_release(&sb);
4349 }
4350 else {
4351 /* we can borrow from the file in the work tree */
4352 temp->name = one->path;
4353 if (!one->oid_valid)
4354 oid_to_hex_r(temp->hex, null_oid());
4355 else
4356 oid_to_hex_r(temp->hex, &one->oid);
4357 /* Even though we may sometimes borrow the
4358 * contents from the work tree, we always want
4359 * one->mode. mode is trustworthy even when
4360 * !(one->oid_valid), as long as
4361 * DIFF_FILE_VALID(one).
4362 */
4363 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4364 }
4365 return temp;
4366 }
4367 else {
4368 if (diff_populate_filespec(r, one, NULL))
4369 die("cannot read data blob for %s", one->path);
4370 prep_temp_blob(r->index, one->path, temp,
4371 one->data, one->size,
4372 &one->oid, one->mode);
4373 }
4374 return temp;
4375 }
4376
4377 static void add_external_diff_name(struct repository *r,
4378 struct strvec *argv,
4379 struct diff_filespec *df)
4380 {
4381 struct diff_tempfile *temp = prepare_temp_file(r, df);
4382 strvec_push(argv, temp->name);
4383 strvec_push(argv, temp->hex);
4384 strvec_push(argv, temp->mode);
4385 }
4386
4387 /* An external diff command takes:
4388 *
4389 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4390 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4391 *
4392 */
4393 static void run_external_diff(const struct external_diff *pgm,
4394 const char *name,
4395 const char *other,
4396 struct diff_filespec *one,
4397 struct diff_filespec *two,
4398 const char *xfrm_msg,
4399 struct diff_options *o)
4400 {
4401 struct child_process cmd = CHILD_PROCESS_INIT;
4402 struct diff_queue_struct *q = &diff_queued_diff;
4403 int quiet = !(o->output_format & DIFF_FORMAT_PATCH);
4404 int rc;
4405
4406 /*
4407 * Trivial equality is handled by diff_unmodified_pair() before
4408 * we get here. If we don't need to show the diff and the
4409 * external diff program lacks the ability to tell us whether
4410 * it's empty then we consider it non-empty without even asking.
4411 */
4412 if (!pgm->trust_exit_code && quiet) {
4413 o->found_changes = 1;
4414 return;
4415 }
4416
4417 strvec_push(&cmd.args, pgm->cmd);
4418 strvec_push(&cmd.args, name);
4419
4420 if (one && two) {
4421 add_external_diff_name(o->repo, &cmd.args, one);
4422 add_external_diff_name(o->repo, &cmd.args, two);
4423 if (other) {
4424 strvec_push(&cmd.args, other);
4425 if (xfrm_msg)
4426 strvec_push(&cmd.args, xfrm_msg);
4427 }
4428 }
4429
4430 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4431 ++o->diff_path_counter);
4432 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4433
4434 diff_free_filespec_data(one);
4435 diff_free_filespec_data(two);
4436 cmd.use_shell = 1;
4437 cmd.no_stdout = quiet;
4438 rc = run_command(&cmd);
4439 if (!pgm->trust_exit_code && rc == 0)
4440 o->found_changes = 1;
4441 else if (pgm->trust_exit_code && rc == 0)
4442 ; /* nothing */
4443 else if (pgm->trust_exit_code && rc == 1)
4444 o->found_changes = 1;
4445 else
4446 die(_("external diff died, stopping at %s"), name);
4447
4448 remove_tempfile();
4449 }
4450
4451 static int similarity_index(struct diff_filepair *p)
4452 {
4453 return p->score * 100 / MAX_SCORE;
4454 }
4455
4456 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4457 {
4458 if (startup_info->have_repository)
4459 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4460 else {
4461 char *hex = oid_to_hex(oid);
4462 if (abbrev < 0)
4463 abbrev = FALLBACK_DEFAULT_ABBREV;
4464 if (abbrev > the_hash_algo->hexsz)
4465 BUG("oid abbreviation out of range: %d", abbrev);
4466 if (abbrev)
4467 hex[abbrev] = '\0';
4468 return hex;
4469 }
4470 }
4471
4472 static void fill_metainfo(struct strbuf *msg,
4473 const char *name,
4474 const char *other,
4475 struct diff_filespec *one,
4476 struct diff_filespec *two,
4477 struct diff_options *o,
4478 struct diff_filepair *p,
4479 int *must_show_header,
4480 int use_color)
4481 {
4482 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4483 const char *reset = diff_get_color(use_color, DIFF_RESET);
4484 const char *line_prefix = diff_line_prefix(o);
4485 struct string_list *more_headers = NULL;
4486
4487 *must_show_header = 1;
4488 strbuf_init(msg, PATH_MAX * 2 + 300);
4489 switch (p->status) {
4490 case DIFF_STATUS_COPIED:
4491 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4492 line_prefix, set, similarity_index(p));
4493 strbuf_addf(msg, "%s\n%s%scopy from ",
4494 reset, line_prefix, set);
4495 quote_c_style(name, msg, NULL, 0);
4496 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4497 quote_c_style(other, msg, NULL, 0);
4498 strbuf_addf(msg, "%s\n", reset);
4499 break;
4500 case DIFF_STATUS_RENAMED:
4501 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4502 line_prefix, set, similarity_index(p));
4503 strbuf_addf(msg, "%s\n%s%srename from ",
4504 reset, line_prefix, set);
4505 quote_c_style(name, msg, NULL, 0);
4506 strbuf_addf(msg, "%s\n%s%srename to ",
4507 reset, line_prefix, set);
4508 quote_c_style(other, msg, NULL, 0);
4509 strbuf_addf(msg, "%s\n", reset);
4510 break;
4511 case DIFF_STATUS_MODIFIED:
4512 if (p->score) {
4513 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4514 line_prefix,
4515 set, similarity_index(p), reset);
4516 break;
4517 }
4518 /* fallthru */
4519 default:
4520 *must_show_header = 0;
4521 }
4522 if ((more_headers = additional_headers(o, name))) {
4523 add_formatted_headers(msg, more_headers,
4524 line_prefix, set, reset);
4525 *must_show_header = 1;
4526 }
4527 if (one && two && !oideq(&one->oid, &two->oid)) {
4528 const unsigned hexsz = the_hash_algo->hexsz;
4529 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4530
4531 if (o->flags.full_index)
4532 abbrev = hexsz;
4533
4534 if (o->flags.binary) {
4535 mmfile_t mf;
4536 if ((!fill_mmfile(o->repo, &mf, one) &&
4537 diff_filespec_is_binary(o->repo, one)) ||
4538 (!fill_mmfile(o->repo, &mf, two) &&
4539 diff_filespec_is_binary(o->repo, two)))
4540 abbrev = hexsz;
4541 }
4542 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4543 diff_abbrev_oid(&one->oid, abbrev),
4544 diff_abbrev_oid(&two->oid, abbrev));
4545 if (one->mode == two->mode)
4546 strbuf_addf(msg, " %06o", one->mode);
4547 strbuf_addf(msg, "%s\n", reset);
4548 }
4549 }
4550
4551 static void run_diff_cmd(const struct external_diff *pgm,
4552 const char *name,
4553 const char *other,
4554 const char *attr_path,
4555 struct diff_filespec *one,
4556 struct diff_filespec *two,
4557 struct strbuf *msg,
4558 struct diff_options *o,
4559 struct diff_filepair *p)
4560 {
4561 const char *xfrm_msg = NULL;
4562 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4563 int must_show_header = 0;
4564 struct userdiff_driver *drv = NULL;
4565
4566 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4567 drv = userdiff_find_by_path(o->repo->index, attr_path);
4568
4569 if (o->flags.allow_external && drv && drv->external.cmd)
4570 pgm = &drv->external;
4571
4572 if (msg) {
4573 /*
4574 * don't use colors when the header is intended for an
4575 * external diff driver
4576 */
4577 fill_metainfo(msg, name, other, one, two, o, p,
4578 &must_show_header,
4579 want_color(o->use_color) && !pgm);
4580 xfrm_msg = msg->len ? msg->buf : NULL;
4581 }
4582
4583 if (pgm) {
4584 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4585 return;
4586 }
4587 if (one && two) {
4588 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4589 set_diff_algorithm(o, drv->algorithm);
4590
4591 builtin_diff(name, other ? other : name,
4592 one, two, xfrm_msg, must_show_header,
4593 o, complete_rewrite);
4594 if (p->status == DIFF_STATUS_COPIED ||
4595 p->status == DIFF_STATUS_RENAMED)
4596 o->found_changes = 1;
4597 } else {
4598 fprintf(o->file, "* Unmerged path %s\n", name);
4599 o->found_changes = 1;
4600 }
4601 }
4602
4603 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4604 {
4605 if (DIFF_FILE_VALID(one)) {
4606 if (!one->oid_valid) {
4607 struct stat st;
4608 if (one->is_stdin) {
4609 oidclr(&one->oid, the_repository->hash_algo);
4610 return;
4611 }
4612 if (lstat(one->path, &st) < 0)
4613 die_errno("stat '%s'", one->path);
4614 if (index_path(istate, &one->oid, one->path, &st, 0))
4615 die("cannot hash %s", one->path);
4616 }
4617 }
4618 else
4619 oidclr(&one->oid, the_repository->hash_algo);
4620 }
4621
4622 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4623 {
4624 /* Strip the prefix but do not molest /dev/null and absolute paths */
4625 if (*namep && !is_absolute_path(*namep)) {
4626 *namep += prefix_length;
4627 if (**namep == '/')
4628 ++*namep;
4629 }
4630 if (*otherp && !is_absolute_path(*otherp)) {
4631 *otherp += prefix_length;
4632 if (**otherp == '/')
4633 ++*otherp;
4634 }
4635 }
4636
4637 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4638 {
4639 const struct external_diff *pgm = external_diff();
4640 struct strbuf msg;
4641 struct diff_filespec *one = p->one;
4642 struct diff_filespec *two = p->two;
4643 const char *name;
4644 const char *other;
4645 const char *attr_path;
4646
4647 name = one->path;
4648 other = (strcmp(name, two->path) ? two->path : NULL);
4649 attr_path = name;
4650 if (o->prefix_length)
4651 strip_prefix(o->prefix_length, &name, &other);
4652
4653 if (!o->flags.allow_external)
4654 pgm = NULL;
4655
4656 if (DIFF_PAIR_UNMERGED(p)) {
4657 run_diff_cmd(pgm, name, NULL, attr_path,
4658 NULL, NULL, NULL, o, p);
4659 return;
4660 }
4661
4662 diff_fill_oid_info(one, o->repo->index);
4663 diff_fill_oid_info(two, o->repo->index);
4664
4665 if (!pgm &&
4666 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4667 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4668 /*
4669 * a filepair that changes between file and symlink
4670 * needs to be split into deletion and creation.
4671 */
4672 struct diff_filespec *null = alloc_filespec(two->path);
4673 run_diff_cmd(NULL, name, other, attr_path,
4674 one, null, &msg,
4675 o, p);
4676 free(null);
4677 strbuf_release(&msg);
4678
4679 null = alloc_filespec(one->path);
4680 run_diff_cmd(NULL, name, other, attr_path,
4681 null, two, &msg, o, p);
4682 free(null);
4683 }
4684 else
4685 run_diff_cmd(pgm, name, other, attr_path,
4686 one, two, &msg, o, p);
4687
4688 strbuf_release(&msg);
4689 }
4690
4691 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4692 struct diffstat_t *diffstat)
4693 {
4694 const char *name;
4695 const char *other;
4696
4697 if (!o->ignore_driver_algorithm) {
4698 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4699 p->one->path);
4700
4701 if (drv && drv->algorithm)
4702 set_diff_algorithm(o, drv->algorithm);
4703 }
4704
4705 if (DIFF_PAIR_UNMERGED(p)) {
4706 /* unmerged */
4707 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4708 diffstat, o, p);
4709 return;
4710 }
4711
4712 name = p->one->path;
4713 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4714
4715 if (o->prefix_length)
4716 strip_prefix(o->prefix_length, &name, &other);
4717
4718 diff_fill_oid_info(p->one, o->repo->index);
4719 diff_fill_oid_info(p->two, o->repo->index);
4720
4721 builtin_diffstat(name, other, p->one, p->two,
4722 diffstat, o, p);
4723 }
4724
4725 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4726 {
4727 const char *name;
4728 const char *other;
4729 const char *attr_path;
4730
4731 if (DIFF_PAIR_UNMERGED(p)) {
4732 /* unmerged */
4733 return;
4734 }
4735
4736 name = p->one->path;
4737 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4738 attr_path = other ? other : name;
4739
4740 if (o->prefix_length)
4741 strip_prefix(o->prefix_length, &name, &other);
4742
4743 diff_fill_oid_info(p->one, o->repo->index);
4744 diff_fill_oid_info(p->two, o->repo->index);
4745
4746 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4747 }
4748
4749 void repo_diff_setup(struct repository *r, struct diff_options *options)
4750 {
4751 memcpy(options, &default_diff_options, sizeof(*options));
4752
4753 options->file = stdout;
4754 options->repo = r;
4755
4756 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4757 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4758 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4759 options->abbrev = DEFAULT_ABBREV;
4760 options->line_termination = '\n';
4761 options->break_opt = -1;
4762 options->rename_limit = -1;
4763 options->dirstat_permille = diff_dirstat_permille_default;
4764 options->context = diff_context_default;
4765 options->interhunkcontext = diff_interhunk_context_default;
4766 options->ws_error_highlight = ws_error_highlight_default;
4767 options->flags.rename_empty = 1;
4768 options->flags.relative_name = diff_relative;
4769 options->objfind = NULL;
4770
4771 /* pathchange left =NULL by default */
4772 options->change = diff_change;
4773 options->add_remove = diff_addremove;
4774 options->use_color = diff_use_color_default;
4775 options->detect_rename = diff_detect_rename_default;
4776 options->xdl_opts |= diff_algorithm;
4777 if (diff_indent_heuristic)
4778 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4779
4780 options->orderfile = diff_order_file_cfg;
4781
4782 if (!options->flags.ignore_submodule_set)
4783 options->flags.ignore_untracked_in_submodules = 1;
4784
4785 if (diff_no_prefix) {
4786 diff_set_noprefix(options);
4787 } else if (!diff_mnemonic_prefix) {
4788 diff_set_default_prefix(options);
4789 }
4790
4791 options->color_moved = diff_color_moved_default;
4792 options->color_moved_ws_handling = diff_color_moved_ws_default;
4793 }
4794
4795 static const char diff_status_letters[] = {
4796 DIFF_STATUS_ADDED,
4797 DIFF_STATUS_COPIED,
4798 DIFF_STATUS_DELETED,
4799 DIFF_STATUS_MODIFIED,
4800 DIFF_STATUS_RENAMED,
4801 DIFF_STATUS_TYPE_CHANGED,
4802 DIFF_STATUS_UNKNOWN,
4803 DIFF_STATUS_UNMERGED,
4804 DIFF_STATUS_FILTER_AON,
4805 DIFF_STATUS_FILTER_BROKEN,
4806 '\0',
4807 };
4808
4809 static unsigned int filter_bit['Z' + 1];
4810
4811 static void prepare_filter_bits(void)
4812 {
4813 int i;
4814
4815 if (!filter_bit[DIFF_STATUS_ADDED]) {
4816 for (i = 0; diff_status_letters[i]; i++)
4817 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4818 }
4819 }
4820
4821 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4822 {
4823 return opt->filter & filter_bit[(int) status];
4824 }
4825
4826 unsigned diff_filter_bit(char status)
4827 {
4828 prepare_filter_bits();
4829 return filter_bit[(int) status];
4830 }
4831
4832 int diff_check_follow_pathspec(struct pathspec *ps, int die_on_error)
4833 {
4834 unsigned forbidden_magic;
4835
4836 if (ps->nr != 1) {
4837 if (die_on_error)
4838 die(_("--follow requires exactly one pathspec"));
4839 return 0;
4840 }
4841
4842 forbidden_magic = ps->items[0].magic & ~(PATHSPEC_FROMTOP |
4843 PATHSPEC_LITERAL);
4844 if (forbidden_magic) {
4845 if (die_on_error) {
4846 struct strbuf sb = STRBUF_INIT;
4847 pathspec_magic_names(forbidden_magic, &sb);
4848 die(_("pathspec magic not supported by --follow: %s"),
4849 sb.buf);
4850 }
4851 return 0;
4852 }
4853
4854 return 1;
4855 }
4856
4857 void diff_setup_done(struct diff_options *options)
4858 {
4859 unsigned check_mask = DIFF_FORMAT_NAME |
4860 DIFF_FORMAT_NAME_STATUS |
4861 DIFF_FORMAT_CHECKDIFF |
4862 DIFF_FORMAT_NO_OUTPUT;
4863 /*
4864 * This must be signed because we're comparing against a potentially
4865 * negative value.
4866 */
4867 const int hexsz = the_hash_algo->hexsz;
4868
4869 if (options->set_default)
4870 options->set_default(options);
4871
4872 if (HAS_MULTI_BITS(options->output_format & check_mask))
4873 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4874 "--name-only", "--name-status", "--check", "-s");
4875
4876 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4877 die(_("options '%s', '%s', and '%s' cannot be used together"),
4878 "-G", "-S", "--find-object");
4879
4880 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4881 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4882 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4883
4884 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4885 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4886 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4887
4888 /*
4889 * Most of the time we can say "there are changes"
4890 * only by checking if there are changed paths, but
4891 * --ignore-whitespace* options force us to look
4892 * inside contents.
4893 */
4894
4895 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4896 options->ignore_regex_nr)
4897 options->flags.diff_from_contents = 1;
4898 else
4899 options->flags.diff_from_contents = 0;
4900
4901 if (options->flags.find_copies_harder)
4902 options->detect_rename = DIFF_DETECT_COPY;
4903
4904 if (!options->flags.relative_name)
4905 options->prefix = NULL;
4906 if (options->prefix)
4907 options->prefix_length = strlen(options->prefix);
4908 else
4909 options->prefix_length = 0;
4910
4911 /*
4912 * --name-only, --name-status, --checkdiff, and -s
4913 * turn other output format off.
4914 */
4915 if (options->output_format & (DIFF_FORMAT_NAME |
4916 DIFF_FORMAT_NAME_STATUS |
4917 DIFF_FORMAT_CHECKDIFF |
4918 DIFF_FORMAT_NO_OUTPUT))
4919 options->output_format &= ~(DIFF_FORMAT_RAW |
4920 DIFF_FORMAT_NUMSTAT |
4921 DIFF_FORMAT_DIFFSTAT |
4922 DIFF_FORMAT_SHORTSTAT |
4923 DIFF_FORMAT_DIRSTAT |
4924 DIFF_FORMAT_SUMMARY |
4925 DIFF_FORMAT_PATCH);
4926
4927 /*
4928 * These cases always need recursive; we do not drop caller-supplied
4929 * recursive bits for other formats here.
4930 */
4931 if (options->output_format & (DIFF_FORMAT_PATCH |
4932 DIFF_FORMAT_NUMSTAT |
4933 DIFF_FORMAT_DIFFSTAT |
4934 DIFF_FORMAT_SHORTSTAT |
4935 DIFF_FORMAT_DIRSTAT |
4936 DIFF_FORMAT_SUMMARY |
4937 DIFF_FORMAT_CHECKDIFF))
4938 options->flags.recursive = 1;
4939 /*
4940 * Also pickaxe would not work very well if you do not say recursive
4941 */
4942 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4943 options->flags.recursive = 1;
4944 /*
4945 * When patches are generated, submodules diffed against the work tree
4946 * must be checked for dirtiness too so it can be shown in the output
4947 */
4948 if (options->output_format & DIFF_FORMAT_PATCH)
4949 options->flags.dirty_submodules = 1;
4950
4951 if (options->detect_rename && options->rename_limit < 0)
4952 options->rename_limit = diff_rename_limit_default;
4953 if (hexsz < options->abbrev)
4954 options->abbrev = hexsz; /* full */
4955
4956 /*
4957 * It does not make sense to show the first hit we happened
4958 * to have found. It does not make sense not to return with
4959 * exit code in such a case either.
4960 */
4961 if (options->flags.quick) {
4962 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4963 options->flags.exit_with_status = 1;
4964 }
4965
4966 /*
4967 * External diffs could declare non-identical contents equal
4968 * (think diff --ignore-space-change).
4969 */
4970 if (options->flags.allow_external && options->flags.exit_with_status)
4971 options->flags.diff_from_contents = 1;
4972
4973 options->diff_path_counter = 0;
4974
4975 if (options->flags.follow_renames)
4976 diff_check_follow_pathspec(&options->pathspec, 1);
4977
4978 if (!options->use_color ||
4979 (options->flags.allow_external && external_diff()))
4980 options->color_moved = 0;
4981
4982 if (options->filter_not) {
4983 if (!options->filter)
4984 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4985 options->filter &= ~options->filter_not;
4986 }
4987 }
4988
4989 int parse_long_opt(const char *opt, const char **argv,
4990 const char **optarg)
4991 {
4992 const char *arg = argv[0];
4993 if (!skip_prefix(arg, "--", &arg))
4994 return 0;
4995 if (!skip_prefix(arg, opt, &arg))
4996 return 0;
4997 if (*arg == '=') { /* stuck form: --option=value */
4998 *optarg = arg + 1;
4999 return 1;
5000 }
5001 if (*arg != '\0')
5002 return 0;
5003 /* separate form: --option value */
5004 if (!argv[1])
5005 die("Option '--%s' requires a value", opt);
5006 *optarg = argv[1];
5007 return 2;
5008 }
5009
5010 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
5011 {
5012 struct diff_options *options = opt->value;
5013 int width = options->stat_width;
5014 int name_width = options->stat_name_width;
5015 int graph_width = options->stat_graph_width;
5016 int count = options->stat_count;
5017 char *end;
5018
5019 BUG_ON_OPT_NEG(unset);
5020
5021 if (!strcmp(opt->long_name, "stat")) {
5022 if (value) {
5023 width = strtoul(value, &end, 10);
5024 if (*end == ',')
5025 name_width = strtoul(end+1, &end, 10);
5026 if (*end == ',')
5027 count = strtoul(end+1, &end, 10);
5028 if (*end)
5029 return error(_("invalid --stat value: %s"), value);
5030 }
5031 } else if (!strcmp(opt->long_name, "stat-width")) {
5032 width = strtoul(value, &end, 10);
5033 if (*end)
5034 return error(_("%s expects a numerical value"),
5035 opt->long_name);
5036 } else if (!strcmp(opt->long_name, "stat-name-width")) {
5037 name_width = strtoul(value, &end, 10);
5038 if (*end)
5039 return error(_("%s expects a numerical value"),
5040 opt->long_name);
5041 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
5042 graph_width = strtoul(value, &end, 10);
5043 if (*end)
5044 return error(_("%s expects a numerical value"),
5045 opt->long_name);
5046 } else if (!strcmp(opt->long_name, "stat-count")) {
5047 count = strtoul(value, &end, 10);
5048 if (*end)
5049 return error(_("%s expects a numerical value"),
5050 opt->long_name);
5051 } else
5052 BUG("%s should not get here", opt->long_name);
5053
5054 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5055 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5056 options->stat_name_width = name_width;
5057 options->stat_graph_width = graph_width;
5058 options->stat_width = width;
5059 options->stat_count = count;
5060 return 0;
5061 }
5062
5063 static int parse_dirstat_opt(struct diff_options *options, const char *params)
5064 {
5065 struct strbuf errmsg = STRBUF_INIT;
5066 if (parse_dirstat_params(options, params, &errmsg))
5067 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
5068 errmsg.buf);
5069 strbuf_release(&errmsg);
5070 /*
5071 * The caller knows a dirstat-related option is given from the command
5072 * line; allow it to say "return this_function();"
5073 */
5074 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5075 options->output_format |= DIFF_FORMAT_DIRSTAT;
5076 return 1;
5077 }
5078
5079 static int diff_opt_diff_filter(const struct option *option,
5080 const char *optarg, int unset)
5081 {
5082 struct diff_options *opt = option->value;
5083 int i, optch;
5084
5085 BUG_ON_OPT_NEG(unset);
5086 prepare_filter_bits();
5087
5088 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
5089 unsigned int bit;
5090 int negate;
5091
5092 if ('a' <= optch && optch <= 'z') {
5093 negate = 1;
5094 optch = toupper(optch);
5095 } else {
5096 negate = 0;
5097 }
5098
5099 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
5100 if (!bit)
5101 return error(_("unknown change class '%c' in --diff-filter=%s"),
5102 optarg[i], optarg);
5103 if (negate)
5104 opt->filter_not |= bit;
5105 else
5106 opt->filter |= bit;
5107 }
5108 return 0;
5109 }
5110
5111 static void enable_patch_output(int *fmt)
5112 {
5113 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
5114 *fmt |= DIFF_FORMAT_PATCH;
5115 }
5116
5117 static int diff_opt_ws_error_highlight(const struct option *option,
5118 const char *arg, int unset)
5119 {
5120 struct diff_options *opt = option->value;
5121 int val = parse_ws_error_highlight(arg);
5122
5123 BUG_ON_OPT_NEG(unset);
5124 if (val < 0)
5125 return error(_("unknown value after ws-error-highlight=%.*s"),
5126 -1 - val, arg);
5127 opt->ws_error_highlight = val;
5128 return 0;
5129 }
5130
5131 static int diff_opt_find_object(const struct option *option,
5132 const char *arg, int unset)
5133 {
5134 struct diff_options *opt = option->value;
5135 struct object_id oid;
5136
5137 BUG_ON_OPT_NEG(unset);
5138 if (repo_get_oid(the_repository, arg, &oid))
5139 return error(_("unable to resolve '%s'"), arg);
5140
5141 if (!opt->objfind)
5142 CALLOC_ARRAY(opt->objfind, 1);
5143
5144 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5145 opt->flags.recursive = 1;
5146 opt->flags.tree_in_recursive = 1;
5147 oidset_insert(opt->objfind, &oid);
5148 return 0;
5149 }
5150
5151 static int diff_opt_anchored(const struct option *opt,
5152 const char *arg, int unset)
5153 {
5154 struct diff_options *options = opt->value;
5155
5156 BUG_ON_OPT_NEG(unset);
5157 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5158 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5159 options->anchors_alloc);
5160 options->anchors[options->anchors_nr++] = xstrdup(arg);
5161 return 0;
5162 }
5163
5164 static int diff_opt_binary(const struct option *opt,
5165 const char *arg, int unset)
5166 {
5167 struct diff_options *options = opt->value;
5168
5169 BUG_ON_OPT_NEG(unset);
5170 BUG_ON_OPT_ARG(arg);
5171 enable_patch_output(&options->output_format);
5172 options->flags.binary = 1;
5173 return 0;
5174 }
5175
5176 static int diff_opt_break_rewrites(const struct option *opt,
5177 const char *arg, int unset)
5178 {
5179 int *break_opt = opt->value;
5180 int opt1, opt2;
5181
5182 BUG_ON_OPT_NEG(unset);
5183 if (!arg)
5184 arg = "";
5185 opt1 = parse_rename_score(&arg);
5186 if (*arg == 0)
5187 opt2 = 0;
5188 else if (*arg != '/')
5189 return error(_("%s expects <n>/<m> form"), opt->long_name);
5190 else {
5191 arg++;
5192 opt2 = parse_rename_score(&arg);
5193 }
5194 if (*arg != 0)
5195 return error(_("%s expects <n>/<m> form"), opt->long_name);
5196 *break_opt = opt1 | (opt2 << 16);
5197 return 0;
5198 }
5199
5200 static int diff_opt_char(const struct option *opt,
5201 const char *arg, int unset)
5202 {
5203 char *value = opt->value;
5204
5205 BUG_ON_OPT_NEG(unset);
5206 if (arg[1])
5207 return error(_("%s expects a character, got '%s'"),
5208 opt->long_name, arg);
5209 *value = arg[0];
5210 return 0;
5211 }
5212
5213 static int diff_opt_color_moved(const struct option *opt,
5214 const char *arg, int unset)
5215 {
5216 struct diff_options *options = opt->value;
5217
5218 if (unset) {
5219 options->color_moved = COLOR_MOVED_NO;
5220 } else if (!arg) {
5221 if (diff_color_moved_default)
5222 options->color_moved = diff_color_moved_default;
5223 if (options->color_moved == COLOR_MOVED_NO)
5224 options->color_moved = COLOR_MOVED_DEFAULT;
5225 } else {
5226 int cm = parse_color_moved(arg);
5227 if (cm < 0)
5228 return error(_("bad --color-moved argument: %s"), arg);
5229 options->color_moved = cm;
5230 }
5231 return 0;
5232 }
5233
5234 static int diff_opt_color_moved_ws(const struct option *opt,
5235 const char *arg, int unset)
5236 {
5237 struct diff_options *options = opt->value;
5238 unsigned cm;
5239
5240 if (unset) {
5241 options->color_moved_ws_handling = 0;
5242 return 0;
5243 }
5244
5245 cm = parse_color_moved_ws(arg);
5246 if (cm & COLOR_MOVED_WS_ERROR)
5247 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5248 options->color_moved_ws_handling = cm;
5249 return 0;
5250 }
5251
5252 static int diff_opt_color_words(const struct option *opt,
5253 const char *arg, int unset)
5254 {
5255 struct diff_options *options = opt->value;
5256
5257 BUG_ON_OPT_NEG(unset);
5258 options->use_color = 1;
5259 options->word_diff = DIFF_WORDS_COLOR;
5260 options->word_regex = arg;
5261 return 0;
5262 }
5263
5264 static int diff_opt_compact_summary(const struct option *opt,
5265 const char *arg, int unset)
5266 {
5267 struct diff_options *options = opt->value;
5268
5269 BUG_ON_OPT_ARG(arg);
5270 if (unset) {
5271 options->flags.stat_with_summary = 0;
5272 } else {
5273 options->flags.stat_with_summary = 1;
5274 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5275 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5276 }
5277 return 0;
5278 }
5279
5280 static int diff_opt_diff_algorithm(const struct option *opt,
5281 const char *arg, int unset)
5282 {
5283 struct diff_options *options = opt->value;
5284
5285 BUG_ON_OPT_NEG(unset);
5286
5287 if (set_diff_algorithm(options, arg))
5288 return error(_("option diff-algorithm accepts \"myers\", "
5289 "\"minimal\", \"patience\" and \"histogram\""));
5290
5291 options->ignore_driver_algorithm = 1;
5292
5293 return 0;
5294 }
5295
5296 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5297 const char *arg, int unset)
5298 {
5299 struct diff_options *options = opt->value;
5300
5301 BUG_ON_OPT_NEG(unset);
5302 BUG_ON_OPT_ARG(arg);
5303
5304 if (set_diff_algorithm(options, opt->long_name))
5305 BUG("available diff algorithms include \"myers\", "
5306 "\"minimal\", \"patience\" and \"histogram\"");
5307
5308 options->ignore_driver_algorithm = 1;
5309
5310 return 0;
5311 }
5312
5313 static int diff_opt_dirstat(const struct option *opt,
5314 const char *arg, int unset)
5315 {
5316 struct diff_options *options = opt->value;
5317
5318 BUG_ON_OPT_NEG(unset);
5319 if (!strcmp(opt->long_name, "cumulative")) {
5320 if (arg)
5321 BUG("how come --cumulative take a value?");
5322 arg = "cumulative";
5323 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5324 parse_dirstat_opt(options, "files");
5325 parse_dirstat_opt(options, arg ? arg : "");
5326 return 0;
5327 }
5328
5329 static int diff_opt_find_copies(const struct option *opt,
5330 const char *arg, int unset)
5331 {
5332 struct diff_options *options = opt->value;
5333
5334 BUG_ON_OPT_NEG(unset);
5335 if (!arg)
5336 arg = "";
5337 options->rename_score = parse_rename_score(&arg);
5338 if (*arg != 0)
5339 return error(_("invalid argument to %s"), opt->long_name);
5340
5341 if (options->detect_rename == DIFF_DETECT_COPY)
5342 options->flags.find_copies_harder = 1;
5343 else
5344 options->detect_rename = DIFF_DETECT_COPY;
5345
5346 return 0;
5347 }
5348
5349 static int diff_opt_find_renames(const struct option *opt,
5350 const char *arg, int unset)
5351 {
5352 struct diff_options *options = opt->value;
5353
5354 BUG_ON_OPT_NEG(unset);
5355 if (!arg)
5356 arg = "";
5357 options->rename_score = parse_rename_score(&arg);
5358 if (*arg != 0)
5359 return error(_("invalid argument to %s"), opt->long_name);
5360
5361 options->detect_rename = DIFF_DETECT_RENAME;
5362 return 0;
5363 }
5364
5365 static int diff_opt_follow(const struct option *opt,
5366 const char *arg, int unset)
5367 {
5368 struct diff_options *options = opt->value;
5369
5370 BUG_ON_OPT_ARG(arg);
5371 if (unset) {
5372 options->flags.follow_renames = 0;
5373 options->flags.default_follow_renames = 0;
5374 } else {
5375 options->flags.follow_renames = 1;
5376 }
5377 return 0;
5378 }
5379
5380 static int diff_opt_ignore_submodules(const struct option *opt,
5381 const char *arg, int unset)
5382 {
5383 struct diff_options *options = opt->value;
5384
5385 BUG_ON_OPT_NEG(unset);
5386 if (!arg)
5387 arg = "all";
5388 options->flags.override_submodule_config = 1;
5389 handle_ignore_submodules_arg(options, arg);
5390 return 0;
5391 }
5392
5393 static int diff_opt_line_prefix(const struct option *opt,
5394 const char *optarg, int unset)
5395 {
5396 struct diff_options *options = opt->value;
5397
5398 BUG_ON_OPT_NEG(unset);
5399 options->line_prefix = optarg;
5400 options->line_prefix_length = strlen(options->line_prefix);
5401 graph_setup_line_prefix(options);
5402 return 0;
5403 }
5404
5405 static int diff_opt_no_prefix(const struct option *opt,
5406 const char *optarg, int unset)
5407 {
5408 struct diff_options *options = opt->value;
5409
5410 BUG_ON_OPT_NEG(unset);
5411 BUG_ON_OPT_ARG(optarg);
5412 diff_set_noprefix(options);
5413 return 0;
5414 }
5415
5416 static int diff_opt_default_prefix(const struct option *opt,
5417 const char *optarg, int unset)
5418 {
5419 struct diff_options *options = opt->value;
5420
5421 BUG_ON_OPT_NEG(unset);
5422 BUG_ON_OPT_ARG(optarg);
5423 FREE_AND_NULL(diff_src_prefix);
5424 FREE_AND_NULL(diff_dst_prefix);
5425 diff_set_default_prefix(options);
5426 return 0;
5427 }
5428
5429 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5430 const struct option *opt,
5431 const char *arg, int unset)
5432 {
5433 struct diff_options *options = opt->value;
5434 char *path;
5435
5436 BUG_ON_OPT_NEG(unset);
5437 path = prefix_filename(ctx->prefix, arg);
5438 options->file = xfopen(path, "w");
5439 options->close_file = 1;
5440 if (options->use_color != GIT_COLOR_ALWAYS)
5441 options->use_color = GIT_COLOR_NEVER;
5442 free(path);
5443 return 0;
5444 }
5445
5446 static int diff_opt_patience(const struct option *opt,
5447 const char *arg, int unset)
5448 {
5449 struct diff_options *options = opt->value;
5450 int i;
5451
5452 BUG_ON_OPT_NEG(unset);
5453 BUG_ON_OPT_ARG(arg);
5454 /*
5455 * Both --patience and --anchored use PATIENCE_DIFF
5456 * internally, so remove any anchors previously
5457 * specified.
5458 */
5459 for (i = 0; i < options->anchors_nr; i++)
5460 free(options->anchors[i]);
5461 options->anchors_nr = 0;
5462 options->ignore_driver_algorithm = 1;
5463
5464 return set_diff_algorithm(options, "patience");
5465 }
5466
5467 static int diff_opt_ignore_regex(const struct option *opt,
5468 const char *arg, int unset)
5469 {
5470 struct diff_options *options = opt->value;
5471 regex_t *regex;
5472
5473 BUG_ON_OPT_NEG(unset);
5474
5475 regex = xmalloc(sizeof(*regex));
5476 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE)) {
5477 free(regex);
5478 return error(_("invalid regex given to -I: '%s'"), arg);
5479 }
5480
5481 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5482 options->ignore_regex_alloc);
5483 options->ignore_regex[options->ignore_regex_nr++] = regex;
5484 return 0;
5485 }
5486
5487 static int diff_opt_pickaxe_regex(const struct option *opt,
5488 const char *arg, int unset)
5489 {
5490 struct diff_options *options = opt->value;
5491
5492 BUG_ON_OPT_NEG(unset);
5493 options->pickaxe = arg;
5494 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5495 return 0;
5496 }
5497
5498 static int diff_opt_pickaxe_string(const struct option *opt,
5499 const char *arg, int unset)
5500 {
5501 struct diff_options *options = opt->value;
5502
5503 BUG_ON_OPT_NEG(unset);
5504 options->pickaxe = arg;
5505 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5506 return 0;
5507 }
5508
5509 static int diff_opt_relative(const struct option *opt,
5510 const char *arg, int unset)
5511 {
5512 struct diff_options *options = opt->value;
5513
5514 options->flags.relative_name = !unset;
5515 if (arg)
5516 options->prefix = arg;
5517 return 0;
5518 }
5519
5520 static int diff_opt_submodule(const struct option *opt,
5521 const char *arg, int unset)
5522 {
5523 struct diff_options *options = opt->value;
5524
5525 BUG_ON_OPT_NEG(unset);
5526 if (!arg)
5527 arg = "log";
5528 if (parse_submodule_params(options, arg))
5529 return error(_("failed to parse --submodule option parameter: '%s'"),
5530 arg);
5531 return 0;
5532 }
5533
5534 static int diff_opt_textconv(const struct option *opt,
5535 const char *arg, int unset)
5536 {
5537 struct diff_options *options = opt->value;
5538
5539 BUG_ON_OPT_ARG(arg);
5540 if (unset) {
5541 options->flags.allow_textconv = 0;
5542 } else {
5543 options->flags.allow_textconv = 1;
5544 options->flags.textconv_set_via_cmdline = 1;
5545 }
5546 return 0;
5547 }
5548
5549 static int diff_opt_unified(const struct option *opt,
5550 const char *arg, int unset)
5551 {
5552 struct diff_options *options = opt->value;
5553 char *s;
5554
5555 BUG_ON_OPT_NEG(unset);
5556
5557 if (arg) {
5558 options->context = strtol(arg, &s, 10);
5559 if (*s)
5560 return error(_("%s expects a numerical value"), "--unified");
5561 }
5562 enable_patch_output(&options->output_format);
5563
5564 return 0;
5565 }
5566
5567 static int diff_opt_word_diff(const struct option *opt,
5568 const char *arg, int unset)
5569 {
5570 struct diff_options *options = opt->value;
5571
5572 BUG_ON_OPT_NEG(unset);
5573 if (arg) {
5574 if (!strcmp(arg, "plain"))
5575 options->word_diff = DIFF_WORDS_PLAIN;
5576 else if (!strcmp(arg, "color")) {
5577 options->use_color = 1;
5578 options->word_diff = DIFF_WORDS_COLOR;
5579 }
5580 else if (!strcmp(arg, "porcelain"))
5581 options->word_diff = DIFF_WORDS_PORCELAIN;
5582 else if (!strcmp(arg, "none"))
5583 options->word_diff = DIFF_WORDS_NONE;
5584 else
5585 return error(_("bad --word-diff argument: %s"), arg);
5586 } else {
5587 if (options->word_diff == DIFF_WORDS_NONE)
5588 options->word_diff = DIFF_WORDS_PLAIN;
5589 }
5590 return 0;
5591 }
5592
5593 static int diff_opt_word_diff_regex(const struct option *opt,
5594 const char *arg, int unset)
5595 {
5596 struct diff_options *options = opt->value;
5597
5598 BUG_ON_OPT_NEG(unset);
5599 if (options->word_diff == DIFF_WORDS_NONE)
5600 options->word_diff = DIFF_WORDS_PLAIN;
5601 options->word_regex = arg;
5602 return 0;
5603 }
5604
5605 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5606 {
5607 struct diff_options *options = opt->value;
5608
5609 BUG_ON_OPT_NEG(unset);
5610 if (!strcmp(opt->long_name, "skip-to"))
5611 options->skip_instead_of_rotate = 1;
5612 else
5613 options->skip_instead_of_rotate = 0;
5614 options->rotate_to = arg;
5615 return 0;
5616 }
5617
5618 /*
5619 * Consider adding new flags to __git_diff_common_options
5620 * in contrib/completion/git-completion.bash
5621 */
5622 struct option *add_diff_options(const struct option *opts,
5623 struct diff_options *options)
5624 {
5625 struct option parseopts[] = {
5626 OPT_GROUP(N_("Diff output format options")),
5627 OPT_BITOP('p', "patch", &options->output_format,
5628 N_("generate patch"),
5629 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5630 OPT_SET_INT('s', "no-patch", &options->output_format,
5631 N_("suppress diff output"), DIFF_FORMAT_NO_OUTPUT),
5632 OPT_BITOP('u', NULL, &options->output_format,
5633 N_("generate patch"),
5634 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5635 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5636 N_("generate diffs with <n> lines context"),
5637 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5638 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5639 N_("generate diffs with <n> lines context")),
5640 OPT_BITOP(0, "raw", &options->output_format,
5641 N_("generate the diff in raw format"),
5642 DIFF_FORMAT_RAW, DIFF_FORMAT_NO_OUTPUT),
5643 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5644 N_("synonym for '-p --raw'"),
5645 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5646 DIFF_FORMAT_NO_OUTPUT),
5647 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5648 N_("synonym for '-p --stat'"),
5649 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5650 DIFF_FORMAT_NO_OUTPUT),
5651 OPT_BITOP(0, "numstat", &options->output_format,
5652 N_("machine friendly --stat"),
5653 DIFF_FORMAT_NUMSTAT, DIFF_FORMAT_NO_OUTPUT),
5654 OPT_BITOP(0, "shortstat", &options->output_format,
5655 N_("output only the last line of --stat"),
5656 DIFF_FORMAT_SHORTSTAT, DIFF_FORMAT_NO_OUTPUT),
5657 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1>,<param2>..."),
5658 N_("output the distribution of relative amount of changes for each sub-directory"),
5659 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5660 diff_opt_dirstat),
5661 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5662 N_("synonym for --dirstat=cumulative"),
5663 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5664 diff_opt_dirstat),
5665 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1>,<param2>..."),
5666 N_("synonym for --dirstat=files,<param1>,<param2>..."),
5667 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5668 diff_opt_dirstat),
5669 OPT_BIT_F(0, "check", &options->output_format,
5670 N_("warn if changes introduce conflict markers or whitespace errors"),
5671 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5672 OPT_BITOP(0, "summary", &options->output_format,
5673 N_("condensed summary such as creations, renames and mode changes"),
5674 DIFF_FORMAT_SUMMARY, DIFF_FORMAT_NO_OUTPUT),
5675 OPT_BIT_F(0, "name-only", &options->output_format,
5676 N_("show only names of changed files"),
5677 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5678 OPT_BIT_F(0, "name-status", &options->output_format,
5679 N_("show only names and status of changed files"),
5680 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5681 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5682 N_("generate diffstat"),
5683 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5684 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5685 N_("generate diffstat with a given width"),
5686 PARSE_OPT_NONEG, diff_opt_stat),
5687 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5688 N_("generate diffstat with a given name width"),
5689 PARSE_OPT_NONEG, diff_opt_stat),
5690 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5691 N_("generate diffstat with a given graph width"),
5692 PARSE_OPT_NONEG, diff_opt_stat),
5693 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5694 N_("generate diffstat with limited lines"),
5695 PARSE_OPT_NONEG, diff_opt_stat),
5696 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5697 N_("generate compact summary in diffstat"),
5698 PARSE_OPT_NOARG, diff_opt_compact_summary),
5699 OPT_CALLBACK_F(0, "binary", options, NULL,
5700 N_("output a binary diff that can be applied"),
5701 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5702 OPT_BOOL(0, "full-index", &options->flags.full_index,
5703 N_("show full pre- and post-image object names on the \"index\" lines")),
5704 OPT_COLOR_FLAG(0, "color", &options->use_color,
5705 N_("show colored diff")),
5706 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5707 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5708 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5709 OPT_SET_INT('z', NULL, &options->line_termination,
5710 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5711 0),
5712 OPT__ABBREV(&options->abbrev),
5713 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5714 N_("show the given source prefix instead of \"a/\""),
5715 PARSE_OPT_NONEG),
5716 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5717 N_("show the given destination prefix instead of \"b/\""),
5718 PARSE_OPT_NONEG),
5719 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5720 N_("prepend an additional prefix to every line of output"),
5721 PARSE_OPT_NONEG, diff_opt_line_prefix),
5722 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5723 N_("do not show any source or destination prefix"),
5724 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5725 OPT_CALLBACK_F(0, "default-prefix", options, NULL,
5726 N_("use default prefixes a/ and b/"),
5727 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_default_prefix),
5728 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5729 N_("show context between diff hunks up to the specified number of lines"),
5730 PARSE_OPT_NONEG),
5731 OPT_CALLBACK_F(0, "output-indicator-new",
5732 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5733 N_("<char>"),
5734 N_("specify the character to indicate a new line instead of '+'"),
5735 PARSE_OPT_NONEG, diff_opt_char),
5736 OPT_CALLBACK_F(0, "output-indicator-old",
5737 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5738 N_("<char>"),
5739 N_("specify the character to indicate an old line instead of '-'"),
5740 PARSE_OPT_NONEG, diff_opt_char),
5741 OPT_CALLBACK_F(0, "output-indicator-context",
5742 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5743 N_("<char>"),
5744 N_("specify the character to indicate a context instead of ' '"),
5745 PARSE_OPT_NONEG, diff_opt_char),
5746
5747 OPT_GROUP(N_("Diff rename options")),
5748 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5749 N_("break complete rewrite changes into pairs of delete and create"),
5750 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5751 diff_opt_break_rewrites),
5752 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5753 N_("detect renames"),
5754 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5755 diff_opt_find_renames),
5756 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5757 N_("omit the preimage for deletes"),
5758 1, PARSE_OPT_NONEG),
5759 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5760 N_("detect copies"),
5761 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5762 diff_opt_find_copies),
5763 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5764 N_("use unmodified files as source to find copies")),
5765 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5766 N_("disable rename detection"),
5767 0, PARSE_OPT_NONEG),
5768 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5769 N_("use empty blobs as rename source")),
5770 OPT_CALLBACK_F(0, "follow", options, NULL,
5771 N_("continue listing the history of a file beyond renames"),
5772 PARSE_OPT_NOARG, diff_opt_follow),
5773 OPT_INTEGER('l', NULL, &options->rename_limit,
5774 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5775
5776 OPT_GROUP(N_("Diff algorithm options")),
5777 OPT_CALLBACK_F(0, "minimal", options, NULL,
5778 N_("produce the smallest possible diff"),
5779 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5780 diff_opt_diff_algorithm_no_arg),
5781 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5782 N_("ignore whitespace when comparing lines"),
5783 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5784 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5785 N_("ignore changes in amount of whitespace"),
5786 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5787 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5788 N_("ignore changes in whitespace at EOL"),
5789 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5790 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5791 N_("ignore carrier-return at the end of line"),
5792 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5793 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5794 N_("ignore changes whose lines are all blank"),
5795 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5796 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5797 N_("ignore changes whose all lines match <regex>"),
5798 0, diff_opt_ignore_regex),
5799 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5800 N_("heuristic to shift diff hunk boundaries for easy reading"),
5801 XDF_INDENT_HEURISTIC),
5802 OPT_CALLBACK_F(0, "patience", options, NULL,
5803 N_("generate diff using the \"patience diff\" algorithm"),
5804 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5805 diff_opt_patience),
5806 OPT_CALLBACK_F(0, "histogram", options, NULL,
5807 N_("generate diff using the \"histogram diff\" algorithm"),
5808 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5809 diff_opt_diff_algorithm_no_arg),
5810 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5811 N_("choose a diff algorithm"),
5812 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5813 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5814 N_("generate diff using the \"anchored diff\" algorithm"),
5815 PARSE_OPT_NONEG, diff_opt_anchored),
5816 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5817 N_("show word diff, using <mode> to delimit changed words"),
5818 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5819 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5820 N_("use <regex> to decide what a word is"),
5821 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5822 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5823 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5824 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5825 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5826 N_("moved lines of code are colored differently"),
5827 PARSE_OPT_OPTARG, diff_opt_color_moved),
5828 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5829 N_("how white spaces are ignored in --color-moved"),
5830 0, diff_opt_color_moved_ws),
5831
5832 OPT_GROUP(N_("Other diff options")),
5833 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5834 N_("when run from subdir, exclude changes outside and show relative paths"),
5835 PARSE_OPT_OPTARG,
5836 diff_opt_relative),
5837 OPT_BOOL('a', "text", &options->flags.text,
5838 N_("treat all files as text")),
5839 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5840 N_("swap two inputs, reverse the diff")),
5841 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5842 N_("exit with 1 if there were differences, 0 otherwise")),
5843 OPT_BOOL(0, "quiet", &options->flags.quick,
5844 N_("disable all output of the program")),
5845 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5846 N_("allow an external diff helper to be executed")),
5847 OPT_CALLBACK_F(0, "textconv", options, NULL,
5848 N_("run external text conversion filters when comparing binary files"),
5849 PARSE_OPT_NOARG, diff_opt_textconv),
5850 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5851 N_("ignore changes to submodules in the diff generation"),
5852 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5853 diff_opt_ignore_submodules),
5854 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5855 N_("specify how differences in submodules are shown"),
5856 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5857 diff_opt_submodule),
5858 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5859 N_("hide 'git add -N' entries from the index"),
5860 1, PARSE_OPT_NONEG),
5861 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5862 N_("treat 'git add -N' entries as real in the index"),
5863 0, PARSE_OPT_NONEG),
5864 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5865 N_("look for differences that change the number of occurrences of the specified string"),
5866 0, diff_opt_pickaxe_string),
5867 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5868 N_("look for differences that change the number of occurrences of the specified regex"),
5869 0, diff_opt_pickaxe_regex),
5870 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5871 N_("show all changes in the changeset with -S or -G"),
5872 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5873 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5874 N_("treat <string> in -S as extended POSIX regular expression"),
5875 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5876 OPT_FILENAME('O', NULL, &options->orderfile,
5877 N_("control the order in which files appear in the output")),
5878 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5879 N_("show the change in the specified path first"),
5880 PARSE_OPT_NONEG, diff_opt_rotate_to),
5881 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5882 N_("skip the output to the specified path"),
5883 PARSE_OPT_NONEG, diff_opt_rotate_to),
5884 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5885 N_("look for differences that change the number of occurrences of the specified object"),
5886 PARSE_OPT_NONEG, diff_opt_find_object),
5887 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5888 N_("select files by diff type"),
5889 PARSE_OPT_NONEG, diff_opt_diff_filter),
5890 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5891 N_("output to a specific file"),
5892 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5893
5894 OPT_END()
5895 };
5896
5897 return parse_options_concat(opts, parseopts);
5898 }
5899
5900 int diff_opt_parse(struct diff_options *options,
5901 const char **av, int ac, const char *prefix)
5902 {
5903 struct option no_options[] = { OPT_END() };
5904 struct option *parseopts = add_diff_options(no_options, options);
5905
5906 if (!prefix)
5907 prefix = "";
5908
5909 ac = parse_options(ac, av, prefix, parseopts, NULL,
5910 PARSE_OPT_KEEP_DASHDASH |
5911 PARSE_OPT_KEEP_UNKNOWN_OPT |
5912 PARSE_OPT_NO_INTERNAL_HELP |
5913 PARSE_OPT_ONE_SHOT |
5914 PARSE_OPT_STOP_AT_NON_OPTION);
5915 free(parseopts);
5916
5917 return ac;
5918 }
5919
5920 int parse_rename_score(const char **cp_p)
5921 {
5922 unsigned long num, scale;
5923 int ch, dot;
5924 const char *cp = *cp_p;
5925
5926 num = 0;
5927 scale = 1;
5928 dot = 0;
5929 for (;;) {
5930 ch = *cp;
5931 if ( !dot && ch == '.' ) {
5932 scale = 1;
5933 dot = 1;
5934 } else if ( ch == '%' ) {
5935 scale = dot ? scale*100 : 100;
5936 cp++; /* % is always at the end */
5937 break;
5938 } else if ( ch >= '0' && ch <= '9' ) {
5939 if ( scale < 100000 ) {
5940 scale *= 10;
5941 num = (num*10) + (ch-'0');
5942 }
5943 } else {
5944 break;
5945 }
5946 cp++;
5947 }
5948 *cp_p = cp;
5949
5950 /* user says num divided by scale and we say internally that
5951 * is MAX_SCORE * num / scale.
5952 */
5953 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5954 }
5955
5956 struct diff_queue_struct diff_queued_diff;
5957
5958 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5959 {
5960 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5961 queue->queue[queue->nr++] = dp;
5962 }
5963
5964 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5965 struct diff_filespec *one,
5966 struct diff_filespec *two)
5967 {
5968 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5969 dp->one = one;
5970 dp->two = two;
5971 if (queue)
5972 diff_q(queue, dp);
5973 return dp;
5974 }
5975
5976 void diff_free_filepair(struct diff_filepair *p)
5977 {
5978 free_filespec(p->one);
5979 free_filespec(p->two);
5980 free(p);
5981 }
5982
5983 void diff_free_queue(struct diff_queue_struct *q)
5984 {
5985 for (int i = 0; i < q->nr; i++)
5986 diff_free_filepair(q->queue[i]);
5987 free(q->queue);
5988 }
5989
5990 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5991 {
5992 int abblen;
5993 const char *abbrev;
5994
5995 /* Do we want all 40 hex characters? */
5996 if (len == the_hash_algo->hexsz)
5997 return oid_to_hex(oid);
5998
5999 /* An abbreviated value is fine, possibly followed by an ellipsis. */
6000 abbrev = diff_abbrev_oid(oid, len);
6001
6002 if (!print_sha1_ellipsis())
6003 return abbrev;
6004
6005 abblen = strlen(abbrev);
6006
6007 /*
6008 * In well-behaved cases, where the abbreviated result is the
6009 * same as the requested length, append three dots after the
6010 * abbreviation (hence the whole logic is limited to the case
6011 * where abblen < 37); when the actual abbreviated result is a
6012 * bit longer than the requested length, we reduce the number
6013 * of dots so that they match the well-behaved ones. However,
6014 * if the actual abbreviation is longer than the requested
6015 * length by more than three, we give up on aligning, and add
6016 * three dots anyway, to indicate that the output is not the
6017 * full object name. Yes, this may be suboptimal, but this
6018 * appears only in "diff --raw --abbrev" output and it is not
6019 * worth the effort to change it now. Note that this would
6020 * likely to work fine when the automatic sizing of default
6021 * abbreviation length is used--we would be fed -1 in "len" in
6022 * that case, and will end up always appending three-dots, but
6023 * the automatic sizing is supposed to give abblen that ensures
6024 * uniqueness across all objects (statistically speaking).
6025 */
6026 if (abblen < the_hash_algo->hexsz - 3) {
6027 static char hex[GIT_MAX_HEXSZ + 1];
6028 if (len < abblen && abblen <= len + 2)
6029 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
6030 else
6031 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
6032 return hex;
6033 }
6034
6035 return oid_to_hex(oid);
6036 }
6037
6038 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
6039 {
6040 int line_termination = opt->line_termination;
6041 int inter_name_termination = line_termination ? '\t' : '\0';
6042
6043 fprintf(opt->file, "%s", diff_line_prefix(opt));
6044 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
6045 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
6046 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
6047 fprintf(opt->file, "%s ",
6048 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
6049 }
6050 if (p->score) {
6051 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
6052 inter_name_termination);
6053 } else {
6054 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
6055 }
6056
6057 if (p->status == DIFF_STATUS_COPIED ||
6058 p->status == DIFF_STATUS_RENAMED) {
6059 const char *name_a, *name_b;
6060 name_a = p->one->path;
6061 name_b = p->two->path;
6062 strip_prefix(opt->prefix_length, &name_a, &name_b);
6063 write_name_quoted(name_a, opt->file, inter_name_termination);
6064 write_name_quoted(name_b, opt->file, line_termination);
6065 } else {
6066 const char *name_a, *name_b;
6067 name_a = p->one->mode ? p->one->path : p->two->path;
6068 name_b = NULL;
6069 strip_prefix(opt->prefix_length, &name_a, &name_b);
6070 write_name_quoted(name_a, opt->file, line_termination);
6071 }
6072 }
6073
6074 int diff_unmodified_pair(struct diff_filepair *p)
6075 {
6076 /* This function is written stricter than necessary to support
6077 * the currently implemented transformers, but the idea is to
6078 * let transformers to produce diff_filepairs any way they want,
6079 * and filter and clean them up here before producing the output.
6080 */
6081 struct diff_filespec *one = p->one, *two = p->two;
6082
6083 if (DIFF_PAIR_UNMERGED(p))
6084 return 0; /* unmerged is interesting */
6085
6086 /* deletion, addition, mode or type change
6087 * and rename are all interesting.
6088 */
6089 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
6090 DIFF_PAIR_MODE_CHANGED(p) ||
6091 strcmp(one->path, two->path))
6092 return 0;
6093
6094 /* both are valid and point at the same path. that is, we are
6095 * dealing with a change.
6096 */
6097 if (one->oid_valid && two->oid_valid &&
6098 oideq(&one->oid, &two->oid) &&
6099 !one->dirty_submodule && !two->dirty_submodule)
6100 return 1; /* no change */
6101 if (!one->oid_valid && !two->oid_valid)
6102 return 1; /* both look at the same file on the filesystem. */
6103 return 0;
6104 }
6105
6106 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
6107 {
6108 int include_conflict_headers =
6109 (additional_headers(o, p->one->path) &&
6110 !o->pickaxe_opts &&
6111 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6112
6113 /*
6114 * Check if we can return early without showing a diff. Note that
6115 * diff_filepair only stores {oid, path, mode, is_valid}
6116 * information for each path, and thus diff_unmodified_pair() only
6117 * considers those bits of info. However, we do not want pairs
6118 * created by create_filepairs_for_header_only_notifications()
6119 * (which always look like unmodified pairs) to be ignored, so
6120 * return early if both p is unmodified AND we don't want to
6121 * include_conflict_headers.
6122 */
6123 if (diff_unmodified_pair(p) && !include_conflict_headers)
6124 return;
6125
6126 /* Actually, we can also return early to avoid showing tree diffs */
6127 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6128 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6129 return;
6130
6131 run_diff(p, o);
6132 }
6133
6134 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
6135 struct diffstat_t *diffstat)
6136 {
6137 if (diff_unmodified_pair(p))
6138 return;
6139
6140 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6141 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6142 return; /* no useful stat for tree diffs */
6143
6144 run_diffstat(p, o, diffstat);
6145 }
6146
6147 static void diff_flush_checkdiff(struct diff_filepair *p,
6148 struct diff_options *o)
6149 {
6150 if (diff_unmodified_pair(p))
6151 return;
6152
6153 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6154 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6155 return; /* nothing to check in tree diffs */
6156
6157 run_checkdiff(p, o);
6158 }
6159
6160 int diff_queue_is_empty(struct diff_options *o)
6161 {
6162 struct diff_queue_struct *q = &diff_queued_diff;
6163 int i;
6164 int include_conflict_headers =
6165 (o->additional_path_headers &&
6166 strmap_get_size(o->additional_path_headers) &&
6167 !o->pickaxe_opts &&
6168 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6169
6170 if (include_conflict_headers)
6171 return 0;
6172
6173 for (i = 0; i < q->nr; i++)
6174 if (!diff_unmodified_pair(q->queue[i]))
6175 return 0;
6176 return 1;
6177 }
6178
6179 #if DIFF_DEBUG
6180 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6181 {
6182 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6183 x, one ? one : "",
6184 s->path,
6185 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6186 s->mode,
6187 s->oid_valid ? oid_to_hex(&s->oid) : "");
6188 fprintf(stderr, "queue[%d] %s size %lu\n",
6189 x, one ? one : "",
6190 s->size);
6191 }
6192
6193 void diff_debug_filepair(const struct diff_filepair *p, int i)
6194 {
6195 diff_debug_filespec(p->one, i, "one");
6196 diff_debug_filespec(p->two, i, "two");
6197 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6198 p->score, p->status ? p->status : '?',
6199 p->one->rename_used, p->broken_pair);
6200 }
6201
6202 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6203 {
6204 int i;
6205 if (msg)
6206 fprintf(stderr, "%s\n", msg);
6207 fprintf(stderr, "q->nr = %d\n", q->nr);
6208 for (i = 0; i < q->nr; i++) {
6209 struct diff_filepair *p = q->queue[i];
6210 diff_debug_filepair(p, i);
6211 }
6212 }
6213 #endif
6214
6215 static void diff_resolve_rename_copy(void)
6216 {
6217 int i;
6218 struct diff_filepair *p;
6219 struct diff_queue_struct *q = &diff_queued_diff;
6220
6221 diff_debug_queue("resolve-rename-copy", q);
6222
6223 for (i = 0; i < q->nr; i++) {
6224 p = q->queue[i];
6225 p->status = 0; /* undecided */
6226 if (DIFF_PAIR_UNMERGED(p))
6227 p->status = DIFF_STATUS_UNMERGED;
6228 else if (!DIFF_FILE_VALID(p->one))
6229 p->status = DIFF_STATUS_ADDED;
6230 else if (!DIFF_FILE_VALID(p->two))
6231 p->status = DIFF_STATUS_DELETED;
6232 else if (DIFF_PAIR_TYPE_CHANGED(p))
6233 p->status = DIFF_STATUS_TYPE_CHANGED;
6234
6235 /* from this point on, we are dealing with a pair
6236 * whose both sides are valid and of the same type, i.e.
6237 * either in-place edit or rename/copy edit.
6238 */
6239 else if (DIFF_PAIR_RENAME(p)) {
6240 /*
6241 * A rename might have re-connected a broken
6242 * pair up, causing the pathnames to be the
6243 * same again. If so, that's not a rename at
6244 * all, just a modification..
6245 *
6246 * Otherwise, see if this source was used for
6247 * multiple renames, in which case we decrement
6248 * the count, and call it a copy.
6249 */
6250 if (!strcmp(p->one->path, p->two->path))
6251 p->status = DIFF_STATUS_MODIFIED;
6252 else if (--p->one->rename_used > 0)
6253 p->status = DIFF_STATUS_COPIED;
6254 else
6255 p->status = DIFF_STATUS_RENAMED;
6256 }
6257 else if (!oideq(&p->one->oid, &p->two->oid) ||
6258 p->one->mode != p->two->mode ||
6259 p->one->dirty_submodule ||
6260 p->two->dirty_submodule ||
6261 is_null_oid(&p->one->oid))
6262 p->status = DIFF_STATUS_MODIFIED;
6263 else {
6264 /* This is a "no-change" entry and should not
6265 * happen anymore, but prepare for broken callers.
6266 */
6267 error("feeding unmodified %s to diffcore",
6268 p->one->path);
6269 p->status = DIFF_STATUS_UNKNOWN;
6270 }
6271 }
6272 diff_debug_queue("resolve-rename-copy done", q);
6273 }
6274
6275 static int check_pair_status(struct diff_filepair *p)
6276 {
6277 switch (p->status) {
6278 case DIFF_STATUS_UNKNOWN:
6279 return 0;
6280 case 0:
6281 die("internal error in diff-resolve-rename-copy");
6282 default:
6283 return 1;
6284 }
6285 }
6286
6287 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6288 {
6289 int fmt = opt->output_format;
6290
6291 if (fmt & DIFF_FORMAT_CHECKDIFF)
6292 diff_flush_checkdiff(p, opt);
6293 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6294 diff_flush_raw(p, opt);
6295 else if (fmt & DIFF_FORMAT_NAME) {
6296 const char *name_a, *name_b;
6297 name_a = p->two->path;
6298 name_b = NULL;
6299 strip_prefix(opt->prefix_length, &name_a, &name_b);
6300 fprintf(opt->file, "%s", diff_line_prefix(opt));
6301 write_name_quoted(name_a, opt->file, opt->line_termination);
6302 }
6303
6304 opt->found_changes = 1;
6305 }
6306
6307 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6308 {
6309 struct strbuf sb = STRBUF_INIT;
6310 if (fs->mode)
6311 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6312 else
6313 strbuf_addf(&sb, " %s ", newdelete);
6314
6315 quote_c_style(fs->path, &sb, NULL, 0);
6316 strbuf_addch(&sb, '\n');
6317 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6318 sb.buf, sb.len, 0);
6319 strbuf_release(&sb);
6320 }
6321
6322 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6323 int show_name)
6324 {
6325 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6326 struct strbuf sb = STRBUF_INIT;
6327 strbuf_addf(&sb, " mode change %06o => %06o",
6328 p->one->mode, p->two->mode);
6329 if (show_name) {
6330 strbuf_addch(&sb, ' ');
6331 quote_c_style(p->two->path, &sb, NULL, 0);
6332 }
6333 strbuf_addch(&sb, '\n');
6334 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6335 sb.buf, sb.len, 0);
6336 strbuf_release(&sb);
6337 }
6338 }
6339
6340 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6341 struct diff_filepair *p)
6342 {
6343 struct strbuf sb = STRBUF_INIT;
6344 struct strbuf names = STRBUF_INIT;
6345
6346 pprint_rename(&names, p->one->path, p->two->path);
6347 strbuf_addf(&sb, " %s %s (%d%%)\n",
6348 renamecopy, names.buf, similarity_index(p));
6349 strbuf_release(&names);
6350 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6351 sb.buf, sb.len, 0);
6352 show_mode_change(opt, p, 0);
6353 strbuf_release(&sb);
6354 }
6355
6356 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6357 {
6358 switch(p->status) {
6359 case DIFF_STATUS_DELETED:
6360 show_file_mode_name(opt, "delete", p->one);
6361 break;
6362 case DIFF_STATUS_ADDED:
6363 show_file_mode_name(opt, "create", p->two);
6364 break;
6365 case DIFF_STATUS_COPIED:
6366 show_rename_copy(opt, "copy", p);
6367 break;
6368 case DIFF_STATUS_RENAMED:
6369 show_rename_copy(opt, "rename", p);
6370 break;
6371 default:
6372 if (p->score) {
6373 struct strbuf sb = STRBUF_INIT;
6374 strbuf_addstr(&sb, " rewrite ");
6375 quote_c_style(p->two->path, &sb, NULL, 0);
6376 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6377 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6378 sb.buf, sb.len, 0);
6379 strbuf_release(&sb);
6380 }
6381 show_mode_change(opt, p, !p->score);
6382 break;
6383 }
6384 }
6385
6386 struct patch_id_t {
6387 git_hash_ctx *ctx;
6388 int patchlen;
6389 };
6390
6391 static int remove_space(char *line, int len)
6392 {
6393 int i;
6394 char *dst = line;
6395 unsigned char c;
6396
6397 for (i = 0; i < len; i++)
6398 if (!isspace((c = line[i])))
6399 *dst++ = c;
6400
6401 return dst - line;
6402 }
6403
6404 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6405 {
6406 unsigned char hash[GIT_MAX_RAWSZ];
6407 unsigned short carry = 0;
6408 int i;
6409
6410 the_hash_algo->final_fn(hash, ctx);
6411 the_hash_algo->init_fn(ctx);
6412 /* 20-byte sum, with carry */
6413 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6414 carry += result->hash[i] + hash[i];
6415 result->hash[i] = carry;
6416 carry >>= 8;
6417 }
6418 }
6419
6420 static int patch_id_consume(void *priv, char *line, unsigned long len)
6421 {
6422 struct patch_id_t *data = priv;
6423 int new_len;
6424
6425 if (len > 12 && starts_with(line, "\\ "))
6426 return 0;
6427 new_len = remove_space(line, len);
6428
6429 the_hash_algo->update_fn(data->ctx, line, new_len);
6430 data->patchlen += new_len;
6431 return 0;
6432 }
6433
6434 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6435 {
6436 the_hash_algo->update_fn(ctx, str, strlen(str));
6437 }
6438
6439 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6440 {
6441 /* large enough for 2^32 in octal */
6442 char buf[12];
6443 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6444 the_hash_algo->update_fn(ctx, buf, len);
6445 }
6446
6447 /* returns 0 upon success, and writes result into oid */
6448 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6449 {
6450 struct diff_queue_struct *q = &diff_queued_diff;
6451 int i;
6452 git_hash_ctx ctx;
6453 struct patch_id_t data;
6454
6455 the_hash_algo->init_fn(&ctx);
6456 memset(&data, 0, sizeof(struct patch_id_t));
6457 data.ctx = &ctx;
6458 oidclr(oid, the_repository->hash_algo);
6459
6460 for (i = 0; i < q->nr; i++) {
6461 xpparam_t xpp;
6462 xdemitconf_t xecfg;
6463 mmfile_t mf1, mf2;
6464 struct diff_filepair *p = q->queue[i];
6465 int len1, len2;
6466
6467 memset(&xpp, 0, sizeof(xpp));
6468 memset(&xecfg, 0, sizeof(xecfg));
6469 if (p->status == 0)
6470 return error("internal diff status error");
6471 if (p->status == DIFF_STATUS_UNKNOWN)
6472 continue;
6473 if (diff_unmodified_pair(p))
6474 continue;
6475 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6476 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6477 continue;
6478 if (DIFF_PAIR_UNMERGED(p))
6479 continue;
6480
6481 diff_fill_oid_info(p->one, options->repo->index);
6482 diff_fill_oid_info(p->two, options->repo->index);
6483
6484 len1 = remove_space(p->one->path, strlen(p->one->path));
6485 len2 = remove_space(p->two->path, strlen(p->two->path));
6486 patch_id_add_string(&ctx, "diff--git");
6487 patch_id_add_string(&ctx, "a/");
6488 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6489 patch_id_add_string(&ctx, "b/");
6490 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6491
6492 if (p->one->mode == 0) {
6493 patch_id_add_string(&ctx, "newfilemode");
6494 patch_id_add_mode(&ctx, p->two->mode);
6495 } else if (p->two->mode == 0) {
6496 patch_id_add_string(&ctx, "deletedfilemode");
6497 patch_id_add_mode(&ctx, p->one->mode);
6498 } else if (p->one->mode != p->two->mode) {
6499 patch_id_add_string(&ctx, "oldmode");
6500 patch_id_add_mode(&ctx, p->one->mode);
6501 patch_id_add_string(&ctx, "newmode");
6502 patch_id_add_mode(&ctx, p->two->mode);
6503 }
6504
6505 if (diff_header_only) {
6506 /* don't do anything since we're only populating header info */
6507 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6508 diff_filespec_is_binary(options->repo, p->two)) {
6509 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6510 the_hash_algo->hexsz);
6511 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6512 the_hash_algo->hexsz);
6513 } else {
6514 if (p->one->mode == 0) {
6515 patch_id_add_string(&ctx, "---/dev/null");
6516 patch_id_add_string(&ctx, "+++b/");
6517 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6518 } else if (p->two->mode == 0) {
6519 patch_id_add_string(&ctx, "---a/");
6520 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6521 patch_id_add_string(&ctx, "+++/dev/null");
6522 } else {
6523 patch_id_add_string(&ctx, "---a/");
6524 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6525 patch_id_add_string(&ctx, "+++b/");
6526 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6527 }
6528
6529 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6530 fill_mmfile(options->repo, &mf2, p->two) < 0)
6531 return error("unable to read files to diff");
6532 xpp.flags = 0;
6533 xecfg.ctxlen = 3;
6534 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6535 if (xdi_diff_outf(&mf1, &mf2, NULL,
6536 patch_id_consume, &data, &xpp, &xecfg))
6537 return error("unable to generate patch-id diff for %s",
6538 p->one->path);
6539 }
6540 flush_one_hunk(oid, &ctx);
6541 }
6542
6543 return 0;
6544 }
6545
6546 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6547 {
6548 struct diff_queue_struct *q = &diff_queued_diff;
6549 int result = diff_get_patch_id(options, oid, diff_header_only);
6550
6551 diff_free_queue(q);
6552 DIFF_QUEUE_CLEAR(q);
6553
6554 return result;
6555 }
6556
6557 static int is_summary_empty(const struct diff_queue_struct *q)
6558 {
6559 int i;
6560
6561 for (i = 0; i < q->nr; i++) {
6562 const struct diff_filepair *p = q->queue[i];
6563
6564 switch (p->status) {
6565 case DIFF_STATUS_DELETED:
6566 case DIFF_STATUS_ADDED:
6567 case DIFF_STATUS_COPIED:
6568 case DIFF_STATUS_RENAMED:
6569 return 0;
6570 default:
6571 if (p->score)
6572 return 0;
6573 if (p->one->mode && p->two->mode &&
6574 p->one->mode != p->two->mode)
6575 return 0;
6576 break;
6577 }
6578 }
6579 return 1;
6580 }
6581
6582 static const char rename_limit_warning[] =
6583 N_("exhaustive rename detection was skipped due to too many files.");
6584
6585 static const char degrade_cc_to_c_warning[] =
6586 N_("only found copies from modified paths due to too many files.");
6587
6588 static const char rename_limit_advice[] =
6589 N_("you may want to set your %s variable to at least "
6590 "%d and retry the command.");
6591
6592 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6593 {
6594 fflush(stdout);
6595 if (degraded_cc)
6596 warning(_(degrade_cc_to_c_warning));
6597 else if (needed)
6598 warning(_(rename_limit_warning));
6599 else
6600 return;
6601 if (0 < needed)
6602 warning(_(rename_limit_advice), varname, needed);
6603 }
6604
6605 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6606 {
6607 struct strset present;
6608 struct diff_queue_struct *q = &diff_queued_diff;
6609 struct hashmap_iter iter;
6610 struct strmap_entry *e;
6611 int i;
6612
6613 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6614
6615 /*
6616 * Find out which paths exist in diff_queued_diff, preferring
6617 * one->path for any pair that has multiple paths.
6618 */
6619 for (i = 0; i < q->nr; i++) {
6620 struct diff_filepair *p = q->queue[i];
6621 char *path = p->one->path ? p->one->path : p->two->path;
6622
6623 if (strmap_contains(o->additional_path_headers, path))
6624 strset_add(&present, path);
6625 }
6626
6627 /*
6628 * Loop over paths in additional_path_headers; for each NOT already
6629 * in diff_queued_diff, create a synthetic filepair and insert that
6630 * into diff_queued_diff.
6631 */
6632 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6633 if (!strset_contains(&present, e->key)) {
6634 struct diff_filespec *one, *two;
6635 struct diff_filepair *p;
6636
6637 one = alloc_filespec(e->key);
6638 two = alloc_filespec(e->key);
6639 fill_filespec(one, null_oid(), 0, 0);
6640 fill_filespec(two, null_oid(), 0, 0);
6641 p = diff_queue(q, one, two);
6642 p->status = DIFF_STATUS_MODIFIED;
6643 }
6644 }
6645
6646 /* Re-sort the filepairs */
6647 diffcore_fix_diff_index();
6648
6649 /* Cleanup */
6650 strset_clear(&present);
6651 }
6652
6653 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6654 {
6655 int i;
6656 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6657 struct diff_queue_struct *q = &diff_queued_diff;
6658
6659 if (WSEH_NEW & WS_RULE_MASK)
6660 BUG("WS rules bit mask overlaps with diff symbol flags");
6661
6662 if (o->color_moved)
6663 o->emitted_symbols = &esm;
6664
6665 if (o->additional_path_headers)
6666 create_filepairs_for_header_only_notifications(o);
6667
6668 for (i = 0; i < q->nr; i++) {
6669 struct diff_filepair *p = q->queue[i];
6670 if (check_pair_status(p))
6671 diff_flush_patch(p, o);
6672 }
6673
6674 if (o->emitted_symbols) {
6675 if (o->color_moved) {
6676 struct mem_pool entry_pool;
6677 struct moved_entry_list *entry_list;
6678
6679 mem_pool_init(&entry_pool, 1024 * 1024);
6680 entry_list = add_lines_to_move_detection(o,
6681 &entry_pool);
6682 mark_color_as_moved(o, entry_list);
6683 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6684 dim_moved_lines(o);
6685
6686 mem_pool_discard(&entry_pool, 0);
6687 free(entry_list);
6688 }
6689
6690 for (i = 0; i < esm.nr; i++)
6691 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6692
6693 for (i = 0; i < esm.nr; i++)
6694 free((void *)esm.buf[i].line);
6695 esm.nr = 0;
6696
6697 o->emitted_symbols = NULL;
6698 }
6699 }
6700
6701 static void diff_free_file(struct diff_options *options)
6702 {
6703 if (options->close_file && options->file) {
6704 fclose(options->file);
6705 options->file = NULL;
6706 }
6707 }
6708
6709 static void diff_free_ignore_regex(struct diff_options *options)
6710 {
6711 int i;
6712
6713 for (i = 0; i < options->ignore_regex_nr; i++) {
6714 regfree(options->ignore_regex[i]);
6715 free(options->ignore_regex[i]);
6716 }
6717
6718 FREE_AND_NULL(options->ignore_regex);
6719 options->ignore_regex_nr = 0;
6720 }
6721
6722 void diff_free(struct diff_options *options)
6723 {
6724 if (options->no_free)
6725 return;
6726
6727 if (options->objfind) {
6728 oidset_clear(options->objfind);
6729 FREE_AND_NULL(options->objfind);
6730 }
6731
6732 for (size_t i = 0; i < options->anchors_nr; i++)
6733 free(options->anchors[i]);
6734 FREE_AND_NULL(options->anchors);
6735 options->anchors_nr = options->anchors_alloc = 0;
6736
6737 diff_free_file(options);
6738 diff_free_ignore_regex(options);
6739 clear_pathspec(&options->pathspec);
6740 }
6741
6742 void diff_flush(struct diff_options *options)
6743 {
6744 struct diff_queue_struct *q = &diff_queued_diff;
6745 int i, output_format = options->output_format;
6746 int separator = 0;
6747 int dirstat_by_line = 0;
6748
6749 /*
6750 * Order: raw, stat, summary, patch
6751 * or: name/name-status/checkdiff (other bits clear)
6752 */
6753 if (!q->nr && !options->additional_path_headers)
6754 goto free_queue;
6755
6756 if (output_format & (DIFF_FORMAT_RAW |
6757 DIFF_FORMAT_NAME |
6758 DIFF_FORMAT_NAME_STATUS |
6759 DIFF_FORMAT_CHECKDIFF)) {
6760 for (i = 0; i < q->nr; i++) {
6761 struct diff_filepair *p = q->queue[i];
6762 if (check_pair_status(p))
6763 flush_one_pair(p, options);
6764 }
6765 separator++;
6766 }
6767
6768 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6769 dirstat_by_line = 1;
6770
6771 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6772 dirstat_by_line) {
6773 struct diffstat_t diffstat;
6774
6775 compute_diffstat(options, &diffstat, q);
6776 if (output_format & DIFF_FORMAT_NUMSTAT)
6777 show_numstat(&diffstat, options);
6778 if (output_format & DIFF_FORMAT_DIFFSTAT)
6779 show_stats(&diffstat, options);
6780 if (output_format & DIFF_FORMAT_SHORTSTAT)
6781 show_shortstats(&diffstat, options);
6782 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6783 show_dirstat_by_line(&diffstat, options);
6784 free_diffstat_info(&diffstat);
6785 separator++;
6786 }
6787 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6788 show_dirstat(options);
6789
6790 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6791 for (i = 0; i < q->nr; i++) {
6792 diff_summary(options, q->queue[i]);
6793 }
6794 separator++;
6795 }
6796
6797 if (output_format & DIFF_FORMAT_PATCH) {
6798 if (separator) {
6799 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6800 if (options->stat_sep)
6801 /* attach patch instead of inline */
6802 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6803 NULL, 0, 0);
6804 }
6805
6806 diff_flush_patch_all_file_pairs(options);
6807 }
6808
6809 if (output_format & DIFF_FORMAT_CALLBACK)
6810 options->format_callback(q, options, options->format_callback_data);
6811
6812 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6813 options->flags.exit_with_status &&
6814 options->flags.diff_from_contents) {
6815 /*
6816 * run diff_flush_patch for the exit status. setting
6817 * options->file to /dev/null should be safe, because we
6818 * aren't supposed to produce any output anyway.
6819 */
6820 diff_free_file(options);
6821 options->file = xfopen("/dev/null", "w");
6822 options->close_file = 1;
6823 options->color_moved = 0;
6824 for (i = 0; i < q->nr; i++) {
6825 struct diff_filepair *p = q->queue[i];
6826 if (check_pair_status(p))
6827 diff_flush_patch(p, options);
6828 if (options->found_changes)
6829 break;
6830 }
6831 }
6832
6833 free_queue:
6834 diff_free_queue(q);
6835 DIFF_QUEUE_CLEAR(q);
6836 diff_free(options);
6837
6838 /*
6839 * Report the content-level differences with HAS_CHANGES;
6840 * diff_addremove/diff_change does not set the bit when
6841 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6842 */
6843 if (options->flags.diff_from_contents) {
6844 if (options->found_changes)
6845 options->flags.has_changes = 1;
6846 else
6847 options->flags.has_changes = 0;
6848 }
6849 }
6850
6851 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6852 {
6853 return (((p->status == DIFF_STATUS_MODIFIED) &&
6854 ((p->score &&
6855 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6856 (!p->score &&
6857 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6858 ((p->status != DIFF_STATUS_MODIFIED) &&
6859 filter_bit_tst(p->status, options)));
6860 }
6861
6862 static void diffcore_apply_filter(struct diff_options *options)
6863 {
6864 int i;
6865 struct diff_queue_struct *q = &diff_queued_diff;
6866 struct diff_queue_struct outq;
6867
6868 DIFF_QUEUE_CLEAR(&outq);
6869
6870 if (!options->filter)
6871 return;
6872
6873 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6874 int found;
6875 for (i = found = 0; !found && i < q->nr; i++) {
6876 if (match_filter(options, q->queue[i]))
6877 found++;
6878 }
6879 if (found)
6880 return;
6881
6882 /* otherwise we will clear the whole queue
6883 * by copying the empty outq at the end of this
6884 * function, but first clear the current entries
6885 * in the queue.
6886 */
6887 for (i = 0; i < q->nr; i++)
6888 diff_free_filepair(q->queue[i]);
6889 }
6890 else {
6891 /* Only the matching ones */
6892 for (i = 0; i < q->nr; i++) {
6893 struct diff_filepair *p = q->queue[i];
6894 if (match_filter(options, p))
6895 diff_q(&outq, p);
6896 else
6897 diff_free_filepair(p);
6898 }
6899 }
6900 free(q->queue);
6901 *q = outq;
6902 }
6903
6904 /* Check whether two filespecs with the same mode and size are identical */
6905 static int diff_filespec_is_identical(struct repository *r,
6906 struct diff_filespec *one,
6907 struct diff_filespec *two)
6908 {
6909 if (S_ISGITLINK(one->mode))
6910 return 0;
6911 if (diff_populate_filespec(r, one, NULL))
6912 return 0;
6913 if (diff_populate_filespec(r, two, NULL))
6914 return 0;
6915 return !memcmp(one->data, two->data, one->size);
6916 }
6917
6918 static int diff_filespec_check_stat_unmatch(struct repository *r,
6919 struct diff_filepair *p)
6920 {
6921 struct diff_populate_filespec_options dpf_options = {
6922 .check_size_only = 1,
6923 .missing_object_cb = diff_queued_diff_prefetch,
6924 .missing_object_data = r,
6925 };
6926
6927 if (p->done_skip_stat_unmatch)
6928 return p->skip_stat_unmatch_result;
6929
6930 p->done_skip_stat_unmatch = 1;
6931 p->skip_stat_unmatch_result = 0;
6932 /*
6933 * 1. Entries that come from stat info dirtiness
6934 * always have both sides (iow, not create/delete),
6935 * one side of the object name is unknown, with
6936 * the same mode and size. Keep the ones that
6937 * do not match these criteria. They have real
6938 * differences.
6939 *
6940 * 2. At this point, the file is known to be modified,
6941 * with the same mode and size, and the object
6942 * name of one side is unknown. Need to inspect
6943 * the identical contents.
6944 */
6945 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6946 !DIFF_FILE_VALID(p->two) ||
6947 (p->one->oid_valid && p->two->oid_valid) ||
6948 (p->one->mode != p->two->mode) ||
6949 diff_populate_filespec(r, p->one, &dpf_options) ||
6950 diff_populate_filespec(r, p->two, &dpf_options) ||
6951 (p->one->size != p->two->size) ||
6952 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6953 p->skip_stat_unmatch_result = 1;
6954 return p->skip_stat_unmatch_result;
6955 }
6956
6957 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6958 {
6959 int i;
6960 struct diff_queue_struct *q = &diff_queued_diff;
6961 struct diff_queue_struct outq;
6962 DIFF_QUEUE_CLEAR(&outq);
6963
6964 for (i = 0; i < q->nr; i++) {
6965 struct diff_filepair *p = q->queue[i];
6966
6967 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6968 diff_q(&outq, p);
6969 else {
6970 /*
6971 * The caller can subtract 1 from skip_stat_unmatch
6972 * to determine how many paths were dirty only
6973 * due to stat info mismatch.
6974 */
6975 if (!diffopt->flags.no_index)
6976 diffopt->skip_stat_unmatch++;
6977 diff_free_filepair(p);
6978 }
6979 }
6980 free(q->queue);
6981 *q = outq;
6982 }
6983
6984 static int diffnamecmp(const void *a_, const void *b_)
6985 {
6986 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6987 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6988 const char *name_a, *name_b;
6989
6990 name_a = a->one ? a->one->path : a->two->path;
6991 name_b = b->one ? b->one->path : b->two->path;
6992 return strcmp(name_a, name_b);
6993 }
6994
6995 void diffcore_fix_diff_index(void)
6996 {
6997 struct diff_queue_struct *q = &diff_queued_diff;
6998 QSORT(q->queue, q->nr, diffnamecmp);
6999 }
7000
7001 void diff_add_if_missing(struct repository *r,
7002 struct oid_array *to_fetch,
7003 const struct diff_filespec *filespec)
7004 {
7005 if (filespec && filespec->oid_valid &&
7006 !S_ISGITLINK(filespec->mode) &&
7007 oid_object_info_extended(r, &filespec->oid, NULL,
7008 OBJECT_INFO_FOR_PREFETCH))
7009 oid_array_append(to_fetch, &filespec->oid);
7010 }
7011
7012 void diff_queued_diff_prefetch(void *repository)
7013 {
7014 struct repository *repo = repository;
7015 int i;
7016 struct diff_queue_struct *q = &diff_queued_diff;
7017 struct oid_array to_fetch = OID_ARRAY_INIT;
7018
7019 for (i = 0; i < q->nr; i++) {
7020 struct diff_filepair *p = q->queue[i];
7021 diff_add_if_missing(repo, &to_fetch, p->one);
7022 diff_add_if_missing(repo, &to_fetch, p->two);
7023 }
7024
7025 /*
7026 * NEEDSWORK: Consider deduplicating the OIDs sent.
7027 */
7028 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
7029
7030 oid_array_clear(&to_fetch);
7031 }
7032
7033 void init_diffstat_widths(struct diff_options *options)
7034 {
7035 options->stat_width = -1; /* use full terminal width */
7036 options->stat_name_width = -1; /* respect diff.statNameWidth config */
7037 options->stat_graph_width = -1; /* respect diff.statGraphWidth config */
7038 }
7039
7040 void diffcore_std(struct diff_options *options)
7041 {
7042 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
7043 DIFF_FORMAT_NUMSTAT |
7044 DIFF_FORMAT_PATCH |
7045 DIFF_FORMAT_SHORTSTAT |
7046 DIFF_FORMAT_DIRSTAT;
7047
7048 /*
7049 * Check if the user requested a blob-data-requiring diff output and/or
7050 * break-rewrite detection (which requires blob data). If yes, prefetch
7051 * the diff pairs.
7052 *
7053 * If no prefetching occurs, diffcore_rename() will prefetch if it
7054 * decides that it needs inexact rename detection.
7055 */
7056 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
7057 (options->output_format & output_formats_to_prefetch ||
7058 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
7059 diff_queued_diff_prefetch(options->repo);
7060
7061 /* NOTE please keep the following in sync with diff_tree_combined() */
7062 if (options->skip_stat_unmatch)
7063 diffcore_skip_stat_unmatch(options);
7064 if (!options->found_follow) {
7065 /* See try_to_follow_renames() in tree-diff.c */
7066 if (options->break_opt != -1)
7067 diffcore_break(options->repo,
7068 options->break_opt);
7069 if (options->detect_rename)
7070 diffcore_rename(options);
7071 if (options->break_opt != -1)
7072 diffcore_merge_broken();
7073 }
7074 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
7075 diffcore_pickaxe(options);
7076 if (options->orderfile)
7077 diffcore_order(options->orderfile);
7078 if (options->rotate_to)
7079 diffcore_rotate(options);
7080 if (!options->found_follow)
7081 /* See try_to_follow_renames() in tree-diff.c */
7082 diff_resolve_rename_copy();
7083 diffcore_apply_filter(options);
7084
7085 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
7086 options->flags.has_changes = 1;
7087 else
7088 options->flags.has_changes = 0;
7089
7090 options->found_follow = 0;
7091 }
7092
7093 int diff_result_code(struct rev_info *revs)
7094 {
7095 struct diff_options *opt = &revs->diffopt;
7096 int result = 0;
7097
7098 if (revs->remerge_diff) {
7099 tmp_objdir_destroy(revs->remerge_objdir);
7100 revs->remerge_objdir = NULL;
7101 }
7102
7103 diff_warn_rename_limit("diff.renameLimit",
7104 opt->needed_rename_limit,
7105 opt->degraded_cc_to_c);
7106
7107 if (opt->flags.exit_with_status &&
7108 opt->flags.has_changes)
7109 result |= 01;
7110 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
7111 opt->flags.check_failed)
7112 result |= 02;
7113 return result;
7114 }
7115
7116 int diff_can_quit_early(struct diff_options *opt)
7117 {
7118 return (opt->flags.quick &&
7119 !opt->filter &&
7120 opt->flags.has_changes);
7121 }
7122
7123 /*
7124 * Shall changes to this submodule be ignored?
7125 *
7126 * Submodule changes can be configured to be ignored separately for each path,
7127 * but that configuration can be overridden from the command line.
7128 */
7129 static int is_submodule_ignored(const char *path, struct diff_options *options)
7130 {
7131 int ignored = 0;
7132 struct diff_flags orig_flags = options->flags;
7133 if (!options->flags.override_submodule_config)
7134 set_diffopt_flags_from_submodule_config(options, path);
7135 if (options->flags.ignore_submodules)
7136 ignored = 1;
7137 options->flags = orig_flags;
7138 return ignored;
7139 }
7140
7141 void compute_diffstat(struct diff_options *options,
7142 struct diffstat_t *diffstat,
7143 struct diff_queue_struct *q)
7144 {
7145 int i;
7146
7147 memset(diffstat, 0, sizeof(struct diffstat_t));
7148 for (i = 0; i < q->nr; i++) {
7149 struct diff_filepair *p = q->queue[i];
7150 if (check_pair_status(p))
7151 diff_flush_stat(p, options, diffstat);
7152 }
7153 options->found_changes = !!diffstat->nr;
7154 }
7155
7156 void diff_addremove(struct diff_options *options,
7157 int addremove, unsigned mode,
7158 const struct object_id *oid,
7159 int oid_valid,
7160 const char *concatpath, unsigned dirty_submodule)
7161 {
7162 struct diff_filespec *one, *two;
7163
7164 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
7165 return;
7166
7167 /* This may look odd, but it is a preparation for
7168 * feeding "there are unchanged files which should
7169 * not produce diffs, but when you are doing copy
7170 * detection you would need them, so here they are"
7171 * entries to the diff-core. They will be prefixed
7172 * with something like '=' or '*' (I haven't decided
7173 * which but should not make any difference).
7174 * Feeding the same new and old to diff_change()
7175 * also has the same effect.
7176 * Before the final output happens, they are pruned after
7177 * merged into rename/copy pairs as appropriate.
7178 */
7179 if (options->flags.reverse_diff)
7180 addremove = (addremove == '+' ? '-' :
7181 addremove == '-' ? '+' : addremove);
7182
7183 if (options->prefix &&
7184 strncmp(concatpath, options->prefix, options->prefix_length))
7185 return;
7186
7187 one = alloc_filespec(concatpath);
7188 two = alloc_filespec(concatpath);
7189
7190 if (addremove != '+')
7191 fill_filespec(one, oid, oid_valid, mode);
7192 if (addremove != '-') {
7193 fill_filespec(two, oid, oid_valid, mode);
7194 two->dirty_submodule = dirty_submodule;
7195 }
7196
7197 diff_queue(&diff_queued_diff, one, two);
7198 if (!options->flags.diff_from_contents)
7199 options->flags.has_changes = 1;
7200 }
7201
7202 void diff_change(struct diff_options *options,
7203 unsigned old_mode, unsigned new_mode,
7204 const struct object_id *old_oid,
7205 const struct object_id *new_oid,
7206 int old_oid_valid, int new_oid_valid,
7207 const char *concatpath,
7208 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7209 {
7210 struct diff_filespec *one, *two;
7211 struct diff_filepair *p;
7212
7213 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7214 is_submodule_ignored(concatpath, options))
7215 return;
7216
7217 if (options->flags.reverse_diff) {
7218 SWAP(old_mode, new_mode);
7219 SWAP(old_oid, new_oid);
7220 SWAP(old_oid_valid, new_oid_valid);
7221 SWAP(old_dirty_submodule, new_dirty_submodule);
7222 }
7223
7224 if (options->prefix &&
7225 strncmp(concatpath, options->prefix, options->prefix_length))
7226 return;
7227
7228 one = alloc_filespec(concatpath);
7229 two = alloc_filespec(concatpath);
7230 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7231 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7232 one->dirty_submodule = old_dirty_submodule;
7233 two->dirty_submodule = new_dirty_submodule;
7234 p = diff_queue(&diff_queued_diff, one, two);
7235
7236 if (options->flags.diff_from_contents)
7237 return;
7238
7239 if (options->flags.quick && options->skip_stat_unmatch &&
7240 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7241 diff_free_filespec_data(p->one);
7242 diff_free_filespec_data(p->two);
7243 return;
7244 }
7245
7246 options->flags.has_changes = 1;
7247 }
7248
7249 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7250 {
7251 struct diff_filepair *pair;
7252 struct diff_filespec *one, *two;
7253
7254 if (options->prefix &&
7255 strncmp(path, options->prefix, options->prefix_length))
7256 return NULL;
7257
7258 one = alloc_filespec(path);
7259 two = alloc_filespec(path);
7260 pair = diff_queue(&diff_queued_diff, one, two);
7261 pair->is_unmerged = 1;
7262 return pair;
7263 }
7264
7265 static char *run_textconv(struct repository *r,
7266 const char *pgm,
7267 struct diff_filespec *spec,
7268 size_t *outsize)
7269 {
7270 struct diff_tempfile *temp;
7271 struct child_process child = CHILD_PROCESS_INIT;
7272 struct strbuf buf = STRBUF_INIT;
7273 int err = 0;
7274
7275 temp = prepare_temp_file(r, spec);
7276 strvec_push(&child.args, pgm);
7277 strvec_push(&child.args, temp->name);
7278
7279 child.use_shell = 1;
7280 child.out = -1;
7281 if (start_command(&child)) {
7282 remove_tempfile();
7283 return NULL;
7284 }
7285
7286 if (strbuf_read(&buf, child.out, 0) < 0)
7287 err = error("error reading from textconv command '%s'", pgm);
7288 close(child.out);
7289
7290 if (finish_command(&child) || err) {
7291 strbuf_release(&buf);
7292 remove_tempfile();
7293 return NULL;
7294 }
7295 remove_tempfile();
7296
7297 return strbuf_detach(&buf, outsize);
7298 }
7299
7300 size_t fill_textconv(struct repository *r,
7301 struct userdiff_driver *driver,
7302 struct diff_filespec *df,
7303 char **outbuf)
7304 {
7305 size_t size;
7306
7307 if (!driver) {
7308 if (!DIFF_FILE_VALID(df)) {
7309 *outbuf = (char *) "";
7310 return 0;
7311 }
7312 if (diff_populate_filespec(r, df, NULL))
7313 die("unable to read files to diff");
7314 *outbuf = df->data;
7315 return df->size;
7316 }
7317
7318 if (!driver->textconv)
7319 BUG("fill_textconv called with non-textconv driver");
7320
7321 if (driver->textconv_cache && df->oid_valid) {
7322 *outbuf = notes_cache_get(driver->textconv_cache,
7323 &df->oid,
7324 &size);
7325 if (*outbuf)
7326 return size;
7327 }
7328
7329 *outbuf = run_textconv(r, driver->textconv, df, &size);
7330 if (!*outbuf)
7331 die("unable to read files to diff");
7332
7333 if (driver->textconv_cache && df->oid_valid) {
7334 /* ignore errors, as we might be in a readonly repository */
7335 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7336 size);
7337 /*
7338 * we could save up changes and flush them all at the end,
7339 * but we would need an extra call after all diffing is done.
7340 * Since generating a cache entry is the slow path anyway,
7341 * this extra overhead probably isn't a big deal.
7342 */
7343 notes_cache_write(driver->textconv_cache);
7344 }
7345
7346 return size;
7347 }
7348
7349 int textconv_object(struct repository *r,
7350 const char *path,
7351 unsigned mode,
7352 const struct object_id *oid,
7353 int oid_valid,
7354 char **buf,
7355 unsigned long *buf_size)
7356 {
7357 struct diff_filespec *df;
7358 struct userdiff_driver *textconv;
7359
7360 df = alloc_filespec(path);
7361 fill_filespec(df, oid, oid_valid, mode);
7362 textconv = get_textconv(r, df);
7363 if (!textconv) {
7364 free_filespec(df);
7365 return 0;
7366 }
7367
7368 *buf_size = fill_textconv(r, textconv, df, buf);
7369 free_filespec(df);
7370 return 1;
7371 }
7372
7373 void setup_diff_pager(struct diff_options *opt)
7374 {
7375 /*
7376 * If the user asked for our exit code, then either they want --quiet
7377 * or --exit-code. We should definitely not bother with a pager in the
7378 * former case, as we will generate no output. Since we still properly
7379 * report our exit code even when a pager is run, we _could_ run a
7380 * pager with --exit-code. But since we have not done so historically,
7381 * and because it is easy to find people oneline advising "git diff
7382 * --exit-code" in hooks and other scripts, we do not do so.
7383 */
7384 if (!opt->flags.exit_with_status &&
7385 check_pager_config("diff") != 0)
7386 setup_pager();
7387 }
The Best VCS Around