?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
The fifteenth batch
[git.git] / diff.c
blob6e432cb8fc4fe0025c5c7e73591983f5e0b421ab
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "git-compat-util.h"
5 #include "abspath.h"
6 #include "base85.h"
7 #include "config.h"
8 #include "convert.h"
9 #include "environment.h"
10 #include "gettext.h"
11 #include "tempfile.h"
12 #include "quote.h"
13 #include "diff.h"
14 #include "diffcore.h"
15 #include "delta.h"
16 #include "hex.h"
17 #include "xdiff-interface.h"
18 #include "color.h"
19 #include "run-command.h"
20 #include "utf8.h"
21 #include "object-store-ll.h"
22 #include "userdiff.h"
23 #include "submodule.h"
24 #include "hashmap.h"
25 #include "mem-pool.h"
26 #include "merge-ll.h"
27 #include "string-list.h"
28 #include "strvec.h"
29 #include "graph.h"
30 #include "oid-array.h"
31 #include "packfile.h"
32 #include "pager.h"
33 #include "parse-options.h"
34 #include "help.h"
35 #include "promisor-remote.h"
36 #include "dir.h"
37 #include "object-file.h"
38 #include "object-name.h"
39 #include "read-cache-ll.h"
40 #include "setup.h"
41 #include "strmap.h"
42 #include "ws.h"
43
44 #ifdef NO_FAST_WORKING_DIRECTORY
45 #define FAST_WORKING_DIRECTORY 0
46 #else
47 #define FAST_WORKING_DIRECTORY 1
48 #endif
49
50 static int diff_detect_rename_default;
51 static int diff_indent_heuristic = 1;
52 static int diff_rename_limit_default = 1000;
53 static int diff_suppress_blank_empty;
54 static int diff_use_color_default = -1;
55 static int diff_color_moved_default;
56 static int diff_color_moved_ws_default;
57 static int diff_context_default = 3;
58 static int diff_interhunk_context_default;
59 static char *diff_word_regex_cfg;
60 static struct external_diff external_diff_cfg;
61 static char *diff_order_file_cfg;
62 int diff_auto_refresh_index = 1;
63 static int diff_mnemonic_prefix;
64 static int diff_no_prefix;
65 static char *diff_src_prefix;
66 static char *diff_dst_prefix;
67 static int diff_relative;
68 static int diff_stat_name_width;
69 static int diff_stat_graph_width;
70 static int diff_dirstat_permille_default = 30;
71 static struct diff_options default_diff_options;
72 static long diff_algorithm;
73 static unsigned ws_error_highlight_default = WSEH_NEW;
74
75 static char diff_colors[][COLOR_MAXLEN] = {
76 GIT_COLOR_RESET,
77 GIT_COLOR_NORMAL, /* CONTEXT */
78 GIT_COLOR_BOLD, /* METAINFO */
79 GIT_COLOR_CYAN, /* FRAGINFO */
80 GIT_COLOR_RED, /* OLD */
81 GIT_COLOR_GREEN, /* NEW */
82 GIT_COLOR_YELLOW, /* COMMIT */
83 GIT_COLOR_BG_RED, /* WHITESPACE */
84 GIT_COLOR_NORMAL, /* FUNCINFO */
85 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
86 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
87 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
88 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
89 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
90 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
91 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
92 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
93 GIT_COLOR_FAINT, /* CONTEXT_DIM */
94 GIT_COLOR_FAINT_RED, /* OLD_DIM */
95 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
96 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
97 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
98 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
99 };
100
101 static const char *color_diff_slots[] = {
102 [DIFF_CONTEXT] = "context",
103 [DIFF_METAINFO] = "meta",
104 [DIFF_FRAGINFO] = "frag",
105 [DIFF_FILE_OLD] = "old",
106 [DIFF_FILE_NEW] = "new",
107 [DIFF_COMMIT] = "commit",
108 [DIFF_WHITESPACE] = "whitespace",
109 [DIFF_FUNCINFO] = "func",
110 [DIFF_FILE_OLD_MOVED] = "oldMoved",
111 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
112 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
113 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
114 [DIFF_FILE_NEW_MOVED] = "newMoved",
115 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
116 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
117 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
118 [DIFF_CONTEXT_DIM] = "contextDimmed",
119 [DIFF_FILE_OLD_DIM] = "oldDimmed",
120 [DIFF_FILE_NEW_DIM] = "newDimmed",
121 [DIFF_CONTEXT_BOLD] = "contextBold",
122 [DIFF_FILE_OLD_BOLD] = "oldBold",
123 [DIFF_FILE_NEW_BOLD] = "newBold",
124 };
125
126 define_list_config_array_extra(color_diff_slots, {"plain"});
127
128 static int parse_diff_color_slot(const char *var)
129 {
130 if (!strcasecmp(var, "plain"))
131 return DIFF_CONTEXT;
132 return LOOKUP_CONFIG(color_diff_slots, var);
133 }
134
135 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
136 struct strbuf *errmsg)
137 {
138 char *params_copy = xstrdup(params_string);
139 struct string_list params = STRING_LIST_INIT_NODUP;
140 int ret = 0;
141 int i;
142
143 if (*params_copy)
144 string_list_split_in_place(&params, params_copy, ",", -1);
145 for (i = 0; i < params.nr; i++) {
146 const char *p = params.items[i].string;
147 if (!strcmp(p, "changes")) {
148 options->flags.dirstat_by_line = 0;
149 options->flags.dirstat_by_file = 0;
150 } else if (!strcmp(p, "lines")) {
151 options->flags.dirstat_by_line = 1;
152 options->flags.dirstat_by_file = 0;
153 } else if (!strcmp(p, "files")) {
154 options->flags.dirstat_by_line = 0;
155 options->flags.dirstat_by_file = 1;
156 } else if (!strcmp(p, "noncumulative")) {
157 options->flags.dirstat_cumulative = 0;
158 } else if (!strcmp(p, "cumulative")) {
159 options->flags.dirstat_cumulative = 1;
160 } else if (isdigit(*p)) {
161 char *end;
162 int permille = strtoul(p, &end, 10) * 10;
163 if (*end == '.' && isdigit(*++end)) {
164 /* only use first digit */
165 permille += *end - '0';
166 /* .. and ignore any further digits */
167 while (isdigit(*++end))
168 ; /* nothing */
169 }
170 if (!*end)
171 options->dirstat_permille = permille;
172 else {
173 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
174 p);
175 ret++;
176 }
177 } else {
178 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
179 ret++;
180 }
181
182 }
183 string_list_clear(&params, 0);
184 free(params_copy);
185 return ret;
186 }
187
188 static int parse_submodule_params(struct diff_options *options, const char *value)
189 {
190 if (!strcmp(value, "log"))
191 options->submodule_format = DIFF_SUBMODULE_LOG;
192 else if (!strcmp(value, "short"))
193 options->submodule_format = DIFF_SUBMODULE_SHORT;
194 else if (!strcmp(value, "diff"))
195 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
196 /*
197 * Please update $__git_diff_submodule_formats in
198 * git-completion.bash when you add new formats.
199 */
200 else
201 return -1;
202 return 0;
203 }
204
205 int git_config_rename(const char *var, const char *value)
206 {
207 if (!value)
208 return DIFF_DETECT_RENAME;
209 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
210 return DIFF_DETECT_COPY;
211 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
212 }
213
214 long parse_algorithm_value(const char *value)
215 {
216 if (!value)
217 return -1;
218 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
219 return 0;
220 else if (!strcasecmp(value, "minimal"))
221 return XDF_NEED_MINIMAL;
222 else if (!strcasecmp(value, "patience"))
223 return XDF_PATIENCE_DIFF;
224 else if (!strcasecmp(value, "histogram"))
225 return XDF_HISTOGRAM_DIFF;
226 /*
227 * Please update $__git_diff_algorithms in git-completion.bash
228 * when you add new algorithms.
229 */
230 return -1;
231 }
232
233 static int parse_one_token(const char **arg, const char *token)
234 {
235 const char *rest;
236 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
237 *arg = rest;
238 return 1;
239 }
240 return 0;
241 }
242
243 static int parse_ws_error_highlight(const char *arg)
244 {
245 const char *orig_arg = arg;
246 unsigned val = 0;
247
248 while (*arg) {
249 if (parse_one_token(&arg, "none"))
250 val = 0;
251 else if (parse_one_token(&arg, "default"))
252 val = WSEH_NEW;
253 else if (parse_one_token(&arg, "all"))
254 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
255 else if (parse_one_token(&arg, "new"))
256 val |= WSEH_NEW;
257 else if (parse_one_token(&arg, "old"))
258 val |= WSEH_OLD;
259 else if (parse_one_token(&arg, "context"))
260 val |= WSEH_CONTEXT;
261 else {
262 return -1 - (int)(arg - orig_arg);
263 }
264 if (*arg)
265 arg++;
266 }
267 return val;
268 }
269
270 /*
271 * These are to give UI layer defaults.
272 * The core-level commands such as git-diff-files should
273 * never be affected by the setting of diff.renames
274 * the user happens to have in the configuration file.
275 */
276 void init_diff_ui_defaults(void)
277 {
278 diff_detect_rename_default = DIFF_DETECT_RENAME;
279 }
280
281 int git_diff_heuristic_config(const char *var, const char *value,
282 void *cb UNUSED)
283 {
284 if (!strcmp(var, "diff.indentheuristic"))
285 diff_indent_heuristic = git_config_bool(var, value);
286 return 0;
287 }
288
289 static int parse_color_moved(const char *arg)
290 {
291 switch (git_parse_maybe_bool(arg)) {
292 case 0:
293 return COLOR_MOVED_NO;
294 case 1:
295 return COLOR_MOVED_DEFAULT;
296 default:
297 break;
298 }
299
300 if (!strcmp(arg, "no"))
301 return COLOR_MOVED_NO;
302 else if (!strcmp(arg, "plain"))
303 return COLOR_MOVED_PLAIN;
304 else if (!strcmp(arg, "blocks"))
305 return COLOR_MOVED_BLOCKS;
306 else if (!strcmp(arg, "zebra"))
307 return COLOR_MOVED_ZEBRA;
308 else if (!strcmp(arg, "default"))
309 return COLOR_MOVED_DEFAULT;
310 else if (!strcmp(arg, "dimmed-zebra"))
311 return COLOR_MOVED_ZEBRA_DIM;
312 else if (!strcmp(arg, "dimmed_zebra"))
313 return COLOR_MOVED_ZEBRA_DIM;
314 else
315 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
316 }
317
318 static unsigned parse_color_moved_ws(const char *arg)
319 {
320 int ret = 0;
321 struct string_list l = STRING_LIST_INIT_DUP;
322 struct string_list_item *i;
323
324 string_list_split(&l, arg, ',', -1);
325
326 for_each_string_list_item(i, &l) {
327 struct strbuf sb = STRBUF_INIT;
328 strbuf_addstr(&sb, i->string);
329 strbuf_trim(&sb);
330
331 if (!strcmp(sb.buf, "no"))
332 ret = 0;
333 else if (!strcmp(sb.buf, "ignore-space-change"))
334 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
335 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
336 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
337 else if (!strcmp(sb.buf, "ignore-all-space"))
338 ret |= XDF_IGNORE_WHITESPACE;
339 else if (!strcmp(sb.buf, "allow-indentation-change"))
340 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
341 else {
342 ret |= COLOR_MOVED_WS_ERROR;
343 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);
344 }
345
346 strbuf_release(&sb);
347 }
348
349 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
350 (ret & XDF_WHITESPACE_FLAGS)) {
351 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
352 ret |= COLOR_MOVED_WS_ERROR;
353 }
354
355 string_list_clear(&l, 0);
356
357 return ret;
358 }
359
360 int git_diff_ui_config(const char *var, const char *value,
361 const struct config_context *ctx, void *cb)
362 {
363 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
364 diff_use_color_default = git_config_colorbool(var, value);
365 return 0;
366 }
367 if (!strcmp(var, "diff.colormoved")) {
368 int cm = parse_color_moved(value);
369 if (cm < 0)
370 return -1;
371 diff_color_moved_default = cm;
372 return 0;
373 }
374 if (!strcmp(var, "diff.colormovedws")) {
375 unsigned cm;
376 if (!value)
377 return config_error_nonbool(var);
378 cm = parse_color_moved_ws(value);
379 if (cm & COLOR_MOVED_WS_ERROR)
380 return -1;
381 diff_color_moved_ws_default = cm;
382 return 0;
383 }
384 if (!strcmp(var, "diff.context")) {
385 diff_context_default = git_config_int(var, value, ctx->kvi);
386 if (diff_context_default < 0)
387 return -1;
388 return 0;
389 }
390 if (!strcmp(var, "diff.interhunkcontext")) {
391 diff_interhunk_context_default = git_config_int(var, value,
392 ctx->kvi);
393 if (diff_interhunk_context_default < 0)
394 return -1;
395 return 0;
396 }
397 if (!strcmp(var, "diff.renames")) {
398 diff_detect_rename_default = git_config_rename(var, value);
399 return 0;
400 }
401 if (!strcmp(var, "diff.autorefreshindex")) {
402 diff_auto_refresh_index = git_config_bool(var, value);
403 return 0;
404 }
405 if (!strcmp(var, "diff.mnemonicprefix")) {
406 diff_mnemonic_prefix = git_config_bool(var, value);
407 return 0;
408 }
409 if (!strcmp(var, "diff.noprefix")) {
410 diff_no_prefix = git_config_bool(var, value);
411 return 0;
412 }
413 if (!strcmp(var, "diff.srcprefix")) {
414 FREE_AND_NULL(diff_src_prefix);
415 return git_config_string(&diff_src_prefix, var, value);
416 }
417 if (!strcmp(var, "diff.dstprefix")) {
418 FREE_AND_NULL(diff_dst_prefix);
419 return git_config_string(&diff_dst_prefix, var, value);
420 }
421 if (!strcmp(var, "diff.relative")) {
422 diff_relative = git_config_bool(var, value);
423 return 0;
424 }
425 if (!strcmp(var, "diff.statnamewidth")) {
426 diff_stat_name_width = git_config_int(var, value, ctx->kvi);
427 return 0;
428 }
429 if (!strcmp(var, "diff.statgraphwidth")) {
430 diff_stat_graph_width = git_config_int(var, value, ctx->kvi);
431 return 0;
432 }
433 if (!strcmp(var, "diff.external"))
434 return git_config_string(&external_diff_cfg.cmd, var, value);
435 if (!strcmp(var, "diff.trustexitcode")) {
436 external_diff_cfg.trust_exit_code = git_config_bool(var, value);
437 return 0;
438 }
439 if (!strcmp(var, "diff.wordregex"))
440 return git_config_string(&diff_word_regex_cfg, var, value);
441 if (!strcmp(var, "diff.orderfile"))
442 return git_config_pathname(&diff_order_file_cfg, var, value);
443
444 if (!strcmp(var, "diff.ignoresubmodules")) {
445 if (!value)
446 return config_error_nonbool(var);
447 handle_ignore_submodules_arg(&default_diff_options, value);
448 }
449
450 if (!strcmp(var, "diff.submodule")) {
451 if (!value)
452 return config_error_nonbool(var);
453 if (parse_submodule_params(&default_diff_options, value))
454 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
455 value);
456 return 0;
457 }
458
459 if (!strcmp(var, "diff.algorithm")) {
460 if (!value)
461 return config_error_nonbool(var);
462 diff_algorithm = parse_algorithm_value(value);
463 if (diff_algorithm < 0)
464 return error(_("unknown value for config '%s': %s"),
465 var, value);
466 return 0;
467 }
468
469 if (git_color_config(var, value, cb) < 0)
470 return -1;
471
472 return git_diff_basic_config(var, value, ctx, cb);
473 }
474
475 int git_diff_basic_config(const char *var, const char *value,
476 const struct config_context *ctx, void *cb)
477 {
478 const char *name;
479
480 if (!strcmp(var, "diff.renamelimit")) {
481 diff_rename_limit_default = git_config_int(var, value, ctx->kvi);
482 return 0;
483 }
484
485 if (userdiff_config(var, value) < 0)
486 return -1;
487
488 if (skip_prefix(var, "diff.color.", &name) ||
489 skip_prefix(var, "color.diff.", &name)) {
490 int slot = parse_diff_color_slot(name);
491 if (slot < 0)
492 return 0;
493 if (!value)
494 return config_error_nonbool(var);
495 return color_parse(value, diff_colors[slot]);
496 }
497
498 if (!strcmp(var, "diff.wserrorhighlight")) {
499 int val;
500 if (!value)
501 return config_error_nonbool(var);
502 val = parse_ws_error_highlight(value);
503 if (val < 0)
504 return error(_("unknown value for config '%s': %s"),
505 var, value);
506 ws_error_highlight_default = val;
507 return 0;
508 }
509
510 /* like GNU diff's --suppress-blank-empty option */
511 if (!strcmp(var, "diff.suppressblankempty") ||
512 /* for backwards compatibility */
513 !strcmp(var, "diff.suppress-blank-empty")) {
514 diff_suppress_blank_empty = git_config_bool(var, value);
515 return 0;
516 }
517
518 if (!strcmp(var, "diff.dirstat")) {
519 struct strbuf errmsg = STRBUF_INIT;
520 if (!value)
521 return config_error_nonbool(var);
522 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
523 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
524 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
525 errmsg.buf);
526 strbuf_release(&errmsg);
527 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
528 return 0;
529 }
530
531 if (git_diff_heuristic_config(var, value, cb) < 0)
532 return -1;
533
534 return git_default_config(var, value, ctx, cb);
535 }
536
537 static char *quote_two(const char *one, const char *two)
538 {
539 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
540 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
541 struct strbuf res = STRBUF_INIT;
542
543 if (need_one + need_two) {
544 strbuf_addch(&res, '"');
545 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
546 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
547 strbuf_addch(&res, '"');
548 } else {
549 strbuf_addstr(&res, one);
550 strbuf_addstr(&res, two);
551 }
552 return strbuf_detach(&res, NULL);
553 }
554
555 static const struct external_diff *external_diff(void)
556 {
557 static struct external_diff external_diff_env, *external_diff_ptr;
558 static int done_preparing = 0;
559
560 if (done_preparing)
561 return external_diff_ptr;
562 external_diff_env.cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
563 if (git_env_bool("GIT_EXTERNAL_DIFF_TRUST_EXIT_CODE", 0))
564 external_diff_env.trust_exit_code = 1;
565 if (external_diff_env.cmd)
566 external_diff_ptr = &external_diff_env;
567 else if (external_diff_cfg.cmd)
568 external_diff_ptr = &external_diff_cfg;
569 done_preparing = 1;
570 return external_diff_ptr;
571 }
572
573 /*
574 * Keep track of files used for diffing. Sometimes such an entry
575 * refers to a temporary file, sometimes to an existing file, and
576 * sometimes to "/dev/null".
577 */
578 static struct diff_tempfile {
579 /*
580 * filename external diff should read from, or NULL if this
581 * entry is currently not in use:
582 */
583 const char *name;
584
585 char hex[GIT_MAX_HEXSZ + 1];
586 char mode[10];
587
588 /*
589 * If this diff_tempfile instance refers to a temporary file,
590 * this tempfile object is used to manage its lifetime.
591 */
592 struct tempfile *tempfile;
593 } diff_temp[2];
594
595 struct emit_callback {
596 int color_diff;
597 unsigned ws_rule;
598 int blank_at_eof_in_preimage;
599 int blank_at_eof_in_postimage;
600 int lno_in_preimage;
601 int lno_in_postimage;
602 const char **label_path;
603 struct diff_words_data *diff_words;
604 struct diff_options *opt;
605 struct strbuf *header;
606 };
607
608 static int count_lines(const char *data, int size)
609 {
610 int count, ch, completely_empty = 1, nl_just_seen = 0;
611 count = 0;
612 while (0 < size--) {
613 ch = *data++;
614 if (ch == '\n') {
615 count++;
616 nl_just_seen = 1;
617 completely_empty = 0;
618 }
619 else {
620 nl_just_seen = 0;
621 completely_empty = 0;
622 }
623 }
624 if (completely_empty)
625 return 0;
626 if (!nl_just_seen)
627 count++; /* no trailing newline */
628 return count;
629 }
630
631 static int fill_mmfile(struct repository *r, mmfile_t *mf,
632 struct diff_filespec *one)
633 {
634 if (!DIFF_FILE_VALID(one)) {
635 mf->ptr = (char *)""; /* does not matter */
636 mf->size = 0;
637 return 0;
638 }
639 else if (diff_populate_filespec(r, one, NULL))
640 return -1;
641
642 mf->ptr = one->data;
643 mf->size = one->size;
644 return 0;
645 }
646
647 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
648 static unsigned long diff_filespec_size(struct repository *r,
649 struct diff_filespec *one)
650 {
651 struct diff_populate_filespec_options dpf_options = {
652 .check_size_only = 1,
653 };
654
655 if (!DIFF_FILE_VALID(one))
656 return 0;
657 diff_populate_filespec(r, one, &dpf_options);
658 return one->size;
659 }
660
661 static int count_trailing_blank(mmfile_t *mf)
662 {
663 char *ptr = mf->ptr;
664 long size = mf->size;
665 int cnt = 0;
666
667 if (!size)
668 return cnt;
669 ptr += size - 1; /* pointing at the very end */
670 if (*ptr != '\n')
671 ; /* incomplete line */
672 else
673 ptr--; /* skip the last LF */
674 while (mf->ptr < ptr) {
675 char *prev_eol;
676 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
677 if (*prev_eol == '\n')
678 break;
679 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
680 break;
681 cnt++;
682 ptr = prev_eol - 1;
683 }
684 return cnt;
685 }
686
687 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
688 struct emit_callback *ecbdata)
689 {
690 int l1, l2, at;
691 l1 = count_trailing_blank(mf1);
692 l2 = count_trailing_blank(mf2);
693 if (l2 <= l1) {
694 ecbdata->blank_at_eof_in_preimage = 0;
695 ecbdata->blank_at_eof_in_postimage = 0;
696 return;
697 }
698 at = count_lines(mf1->ptr, mf1->size);
699 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
700
701 at = count_lines(mf2->ptr, mf2->size);
702 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
703 }
704
705 static void emit_line_0(struct diff_options *o,
706 const char *set_sign, const char *set, unsigned reverse, const char *reset,
707 int first, const char *line, int len)
708 {
709 int has_trailing_newline, has_trailing_carriage_return;
710 int needs_reset = 0; /* at the end of the line */
711 FILE *file = o->file;
712
713 fputs(diff_line_prefix(o), file);
714
715 has_trailing_newline = (len > 0 && line[len-1] == '\n');
716 if (has_trailing_newline)
717 len--;
718
719 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
720 if (has_trailing_carriage_return)
721 len--;
722
723 if (!len && !first)
724 goto end_of_line;
725
726 if (reverse && want_color(o->use_color)) {
727 fputs(GIT_COLOR_REVERSE, file);
728 needs_reset = 1;
729 }
730
731 if (set_sign) {
732 fputs(set_sign, file);
733 needs_reset = 1;
734 }
735
736 if (first)
737 fputc(first, file);
738
739 if (!len)
740 goto end_of_line;
741
742 if (set) {
743 if (set_sign && set != set_sign)
744 fputs(reset, file);
745 fputs(set, file);
746 needs_reset = 1;
747 }
748 fwrite(line, len, 1, file);
749 needs_reset = 1; /* 'line' may contain color codes. */
750
751 end_of_line:
752 if (needs_reset)
753 fputs(reset, file);
754 if (has_trailing_carriage_return)
755 fputc('\r', file);
756 if (has_trailing_newline)
757 fputc('\n', file);
758 }
759
760 static void emit_line(struct diff_options *o, const char *set, const char *reset,
761 const char *line, int len)
762 {
763 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
764 }
765
766 enum diff_symbol {
767 DIFF_SYMBOL_BINARY_DIFF_HEADER,
768 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
769 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
770 DIFF_SYMBOL_BINARY_DIFF_BODY,
771 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
772 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
773 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
774 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
775 DIFF_SYMBOL_STATS_LINE,
776 DIFF_SYMBOL_WORD_DIFF,
777 DIFF_SYMBOL_STAT_SEP,
778 DIFF_SYMBOL_SUMMARY,
779 DIFF_SYMBOL_SUBMODULE_ADD,
780 DIFF_SYMBOL_SUBMODULE_DEL,
781 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
782 DIFF_SYMBOL_SUBMODULE_MODIFIED,
783 DIFF_SYMBOL_SUBMODULE_HEADER,
784 DIFF_SYMBOL_SUBMODULE_ERROR,
785 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
786 DIFF_SYMBOL_REWRITE_DIFF,
787 DIFF_SYMBOL_BINARY_FILES,
788 DIFF_SYMBOL_HEADER,
789 DIFF_SYMBOL_FILEPAIR_PLUS,
790 DIFF_SYMBOL_FILEPAIR_MINUS,
791 DIFF_SYMBOL_WORDS_PORCELAIN,
792 DIFF_SYMBOL_WORDS,
793 DIFF_SYMBOL_CONTEXT,
794 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
795 DIFF_SYMBOL_PLUS,
796 DIFF_SYMBOL_MINUS,
797 DIFF_SYMBOL_NO_LF_EOF,
798 DIFF_SYMBOL_CONTEXT_FRAGINFO,
799 DIFF_SYMBOL_CONTEXT_MARKER,
800 DIFF_SYMBOL_SEPARATOR
801 };
802 /*
803 * Flags for content lines:
804 * 0..12 are whitespace rules
805 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
806 * 16 is marking if the line is blank at EOF
807 */
808 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
809 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
810 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
811 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
812 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
813
814 /*
815 * This struct is used when we need to buffer the output of the diff output.
816 *
817 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
818 * into the pre/post image file. This pointer could be a union with the
819 * line pointer. By storing an offset into the file instead of the literal line,
820 * we can decrease the memory footprint for the buffered output. At first we
821 * may want to only have indirection for the content lines, but we could also
822 * enhance the state for emitting prefabricated lines, e.g. the similarity
823 * score line or hunk/file headers would only need to store a number or path
824 * and then the output can be constructed later on depending on state.
825 */
826 struct emitted_diff_symbol {
827 const char *line;
828 int len;
829 int flags;
830 int indent_off; /* Offset to first non-whitespace character */
831 int indent_width; /* The visual width of the indentation */
832 unsigned id;
833 enum diff_symbol s;
834 };
835 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
836
837 struct emitted_diff_symbols {
838 struct emitted_diff_symbol *buf;
839 int nr, alloc;
840 };
841 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
842
843 static void append_emitted_diff_symbol(struct diff_options *o,
844 struct emitted_diff_symbol *e)
845 {
846 struct emitted_diff_symbol *f;
847
848 ALLOC_GROW(o->emitted_symbols->buf,
849 o->emitted_symbols->nr + 1,
850 o->emitted_symbols->alloc);
851 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
852
853 memcpy(f, e, sizeof(struct emitted_diff_symbol));
854 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
855 }
856
857 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
858 {
859 if (!e)
860 return;
861 free(e->buf);
862 free(e);
863 }
864
865 struct moved_entry {
866 const struct emitted_diff_symbol *es;
867 struct moved_entry *next_line;
868 struct moved_entry *next_match;
869 };
870
871 struct moved_block {
872 struct moved_entry *match;
873 int wsd; /* The whitespace delta of this block */
874 };
875
876 #define INDENT_BLANKLINE INT_MIN
877
878 static void fill_es_indent_data(struct emitted_diff_symbol *es)
879 {
880 unsigned int off = 0, i;
881 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
882 const char *s = es->line;
883 const int len = es->len;
884
885 /* skip any \v \f \r at start of indentation */
886 while (s[off] == '\f' || s[off] == '\v' ||
887 (s[off] == '\r' && off < len - 1))
888 off++;
889
890 /* calculate the visual width of indentation */
891 while(1) {
892 if (s[off] == ' ') {
893 width++;
894 off++;
895 } else if (s[off] == '\t') {
896 width += tab_width - (width % tab_width);
897 while (s[++off] == '\t')
898 width += tab_width;
899 } else {
900 break;
901 }
902 }
903
904 /* check if this line is blank */
905 for (i = off; i < len; i++)
906 if (!isspace(s[i]))
907 break;
908
909 if (i == len) {
910 es->indent_width = INDENT_BLANKLINE;
911 es->indent_off = len;
912 } else {
913 es->indent_off = off;
914 es->indent_width = width;
915 }
916 }
917
918 static int compute_ws_delta(const struct emitted_diff_symbol *a,
919 const struct emitted_diff_symbol *b)
920 {
921 int a_width = a->indent_width,
922 b_width = b->indent_width;
923
924 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
925 return INDENT_BLANKLINE;
926
927 return a_width - b_width;
928 }
929
930 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
931 const struct emitted_diff_symbol *l,
932 struct moved_block *pmb)
933 {
934 int a_width = cur->es->indent_width, b_width = l->indent_width;
935 int delta;
936
937 /* The text of each line must match */
938 if (cur->es->id != l->id)
939 return 1;
940
941 /*
942 * If 'l' and 'cur' are both blank then we don't need to check the
943 * indent. We only need to check cur as we know the strings match.
944 * */
945 if (a_width == INDENT_BLANKLINE)
946 return 0;
947
948 /*
949 * The indent changes of the block are known and stored in pmb->wsd;
950 * however we need to check if the indent changes of the current line
951 * match those of the current block.
952 */
953 delta = b_width - a_width;
954
955 /*
956 * If the previous lines of this block were all blank then set its
957 * whitespace delta.
958 */
959 if (pmb->wsd == INDENT_BLANKLINE)
960 pmb->wsd = delta;
961
962 return delta != pmb->wsd;
963 }
964
965 struct interned_diff_symbol {
966 struct hashmap_entry ent;
967 struct emitted_diff_symbol *es;
968 };
969
970 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
971 const struct hashmap_entry *eptr,
972 const struct hashmap_entry *entry_or_key,
973 const void *keydata UNUSED)
974 {
975 const struct diff_options *diffopt = hashmap_cmp_fn_data;
976 const struct emitted_diff_symbol *a, *b;
977 unsigned flags = diffopt->color_moved_ws_handling
978 & XDF_WHITESPACE_FLAGS;
979
980 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
981 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
982
983 return !xdiff_compare_lines(a->line + a->indent_off,
984 a->len - a->indent_off,
985 b->line + b->indent_off,
986 b->len - b->indent_off, flags);
987 }
988
989 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
990 struct interned_diff_symbol *s)
991 {
992 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
993 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
994 l->len - l->indent_off, flags);
995
996 hashmap_entry_init(&s->ent, hash);
997 s->es = l;
998 }
999
1000 struct moved_entry_list {
1001 struct moved_entry *add, *del;
1002 };
1003
1004 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
1005 struct mem_pool *entry_mem_pool)
1006 {
1007 struct moved_entry *prev_line = NULL;
1008 struct mem_pool interned_pool;
1009 struct hashmap interned_map;
1010 struct moved_entry_list *entry_list = NULL;
1011 size_t entry_list_alloc = 0;
1012 unsigned id = 0;
1013 int n;
1014
1015 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
1016 mem_pool_init(&interned_pool, 1024 * 1024);
1017
1018 for (n = 0; n < o->emitted_symbols->nr; n++) {
1019 struct interned_diff_symbol key;
1020 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1021 struct interned_diff_symbol *s;
1022 struct moved_entry *entry;
1023
1024 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
1025 prev_line = NULL;
1026 continue;
1027 }
1028
1029 if (o->color_moved_ws_handling &
1030 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1031 fill_es_indent_data(l);
1032
1033 prepare_entry(o, l, &key);
1034 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
1035 if (s) {
1036 l->id = s->es->id;
1037 } else {
1038 l->id = id;
1039 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
1040 hashmap_add(&interned_map,
1041 memcpy(mem_pool_alloc(&interned_pool,
1042 sizeof(key)),
1043 &key, sizeof(key)));
1044 }
1045 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1046 entry->es = l;
1047 entry->next_line = NULL;
1048 if (prev_line && prev_line->es->s == l->s)
1049 prev_line->next_line = entry;
1050 prev_line = entry;
1051 if (l->s == DIFF_SYMBOL_PLUS) {
1052 entry->next_match = entry_list[l->id].add;
1053 entry_list[l->id].add = entry;
1054 } else {
1055 entry->next_match = entry_list[l->id].del;
1056 entry_list[l->id].del = entry;
1057 }
1058 }
1059
1060 hashmap_clear(&interned_map);
1061 mem_pool_discard(&interned_pool, 0);
1062
1063 return entry_list;
1064 }
1065
1066 static void pmb_advance_or_null(struct diff_options *o,
1067 struct emitted_diff_symbol *l,
1068 struct moved_block *pmb,
1069 int *pmb_nr)
1070 {
1071 int i, j;
1072
1073 for (i = 0, j = 0; i < *pmb_nr; i++) {
1074 int match;
1075 struct moved_entry *prev = pmb[i].match;
1076 struct moved_entry *cur = (prev && prev->next_line) ?
1077 prev->next_line : NULL;
1078
1079 if (o->color_moved_ws_handling &
1080 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1081 match = cur &&
1082 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1083 else
1084 match = cur && cur->es->id == l->id;
1085
1086 if (match) {
1087 pmb[j] = pmb[i];
1088 pmb[j++].match = cur;
1089 }
1090 }
1091 *pmb_nr = j;
1092 }
1093
1094 static void fill_potential_moved_blocks(struct diff_options *o,
1095 struct moved_entry *match,
1096 struct emitted_diff_symbol *l,
1097 struct moved_block **pmb_p,
1098 int *pmb_alloc_p, int *pmb_nr_p)
1099
1100 {
1101 struct moved_block *pmb = *pmb_p;
1102 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1103
1104 /*
1105 * The current line is the start of a new block.
1106 * Setup the set of potential blocks.
1107 */
1108 for (; match; match = match->next_match) {
1109 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1110 if (o->color_moved_ws_handling &
1111 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1112 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1113 else
1114 pmb[pmb_nr].wsd = 0;
1115 pmb[pmb_nr++].match = match;
1116 }
1117
1118 *pmb_p = pmb;
1119 *pmb_alloc_p = pmb_alloc;
1120 *pmb_nr_p = pmb_nr;
1121 }
1122
1123 /*
1124 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1125 *
1126 * Otherwise, if the last block has fewer alphanumeric characters than
1127 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1128 * that block.
1129 *
1130 * The last block consists of the (n - block_length)'th line up to but not
1131 * including the nth line.
1132 *
1133 * Returns 0 if the last block is empty or is unset by this function, non zero
1134 * otherwise.
1135 *
1136 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1137 * Think of a way to unify them.
1138 */
1139 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1140 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1141 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1142 {
1143 int i, alnum_count = 0;
1144 if (o->color_moved == COLOR_MOVED_PLAIN)
1145 return block_length;
1146 for (i = 1; i < block_length + 1; i++) {
1147 const char *c = o->emitted_symbols->buf[n - i].line;
1148 for (; *c; c++) {
1149 if (!isalnum(*c))
1150 continue;
1151 alnum_count++;
1152 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1153 return 1;
1154 }
1155 }
1156 for (i = 1; i < block_length + 1; i++)
1157 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1158 return 0;
1159 }
1160
1161 /* Find blocks of moved code, delegate actual coloring decision to helper */
1162 static void mark_color_as_moved(struct diff_options *o,
1163 struct moved_entry_list *entry_list)
1164 {
1165 struct moved_block *pmb = NULL; /* potentially moved blocks */
1166 int pmb_nr = 0, pmb_alloc = 0;
1167 int n, flipped_block = 0, block_length = 0;
1168 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1169
1170
1171 for (n = 0; n < o->emitted_symbols->nr; n++) {
1172 struct moved_entry *match = NULL;
1173 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1174
1175 switch (l->s) {
1176 case DIFF_SYMBOL_PLUS:
1177 match = entry_list[l->id].del;
1178 break;
1179 case DIFF_SYMBOL_MINUS:
1180 match = entry_list[l->id].add;
1181 break;
1182 default:
1183 flipped_block = 0;
1184 }
1185
1186 if (pmb_nr && (!match || l->s != moved_symbol)) {
1187 if (!adjust_last_block(o, n, block_length) &&
1188 block_length > 1) {
1189 /*
1190 * Rewind in case there is another match
1191 * starting at the second line of the block
1192 */
1193 match = NULL;
1194 n -= block_length;
1195 }
1196 pmb_nr = 0;
1197 block_length = 0;
1198 flipped_block = 0;
1199 }
1200 if (!match) {
1201 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1202 continue;
1203 }
1204
1205 if (o->color_moved == COLOR_MOVED_PLAIN) {
1206 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1207 continue;
1208 }
1209
1210 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1211
1212 if (pmb_nr == 0) {
1213 int contiguous = adjust_last_block(o, n, block_length);
1214
1215 if (!contiguous && block_length > 1)
1216 /*
1217 * Rewind in case there is another match
1218 * starting at the second line of the block
1219 */
1220 n -= block_length;
1221 else
1222 fill_potential_moved_blocks(o, match, l,
1223 &pmb, &pmb_alloc,
1224 &pmb_nr);
1225
1226 if (contiguous && pmb_nr && moved_symbol == l->s)
1227 flipped_block = (flipped_block + 1) % 2;
1228 else
1229 flipped_block = 0;
1230
1231 if (pmb_nr)
1232 moved_symbol = l->s;
1233 else
1234 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1235
1236 block_length = 0;
1237 }
1238
1239 if (pmb_nr) {
1240 block_length++;
1241 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1242 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1243 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1244 }
1245 }
1246 adjust_last_block(o, n, block_length);
1247
1248 free(pmb);
1249 }
1250
1251 static void dim_moved_lines(struct diff_options *o)
1252 {
1253 int n;
1254 for (n = 0; n < o->emitted_symbols->nr; n++) {
1255 struct emitted_diff_symbol *prev = (n != 0) ?
1256 &o->emitted_symbols->buf[n - 1] : NULL;
1257 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1258 struct emitted_diff_symbol *next =
1259 (n < o->emitted_symbols->nr - 1) ?
1260 &o->emitted_symbols->buf[n + 1] : NULL;
1261
1262 /* Not a plus or minus line? */
1263 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1264 continue;
1265
1266 /* Not a moved line? */
1267 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1268 continue;
1269
1270 /*
1271 * If prev or next are not a plus or minus line,
1272 * pretend they don't exist
1273 */
1274 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1275 prev->s != DIFF_SYMBOL_MINUS)
1276 prev = NULL;
1277 if (next && next->s != DIFF_SYMBOL_PLUS &&
1278 next->s != DIFF_SYMBOL_MINUS)
1279 next = NULL;
1280
1281 /* Inside a block? */
1282 if ((prev &&
1283 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1284 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1285 (next &&
1286 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1287 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1288 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1289 continue;
1290 }
1291
1292 /* Check if we are at an interesting bound: */
1293 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1294 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1295 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1296 continue;
1297 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1298 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1299 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1300 continue;
1301
1302 /*
1303 * The boundary to prev and next are not interesting,
1304 * so this line is not interesting as a whole
1305 */
1306 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1307 }
1308 }
1309
1310 static void emit_line_ws_markup(struct diff_options *o,
1311 const char *set_sign, const char *set,
1312 const char *reset,
1313 int sign_index, const char *line, int len,
1314 unsigned ws_rule, int blank_at_eof)
1315 {
1316 const char *ws = NULL;
1317 int sign = o->output_indicators[sign_index];
1318
1319 if (o->ws_error_highlight & ws_rule) {
1320 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1321 if (!*ws)
1322 ws = NULL;
1323 }
1324
1325 if (!ws && !set_sign)
1326 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1327 else if (!ws) {
1328 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1329 } else if (blank_at_eof)
1330 /* Blank line at EOF - paint '+' as well */
1331 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1332 else {
1333 /* Emit just the prefix, then the rest. */
1334 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1335 sign, "", 0);
1336 ws_check_emit(line, len, ws_rule,
1337 o->file, set, reset, ws);
1338 }
1339 }
1340
1341 static void emit_diff_symbol_from_struct(struct diff_options *o,
1342 struct emitted_diff_symbol *eds)
1343 {
1344 static const char *nneof = " No newline at end of file\n";
1345 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1346
1347 enum diff_symbol s = eds->s;
1348 const char *line = eds->line;
1349 int len = eds->len;
1350 unsigned flags = eds->flags;
1351
1352 switch (s) {
1353 case DIFF_SYMBOL_NO_LF_EOF:
1354 context = diff_get_color_opt(o, DIFF_CONTEXT);
1355 reset = diff_get_color_opt(o, DIFF_RESET);
1356 putc('\n', o->file);
1357 emit_line_0(o, context, NULL, 0, reset, '\\',
1358 nneof, strlen(nneof));
1359 break;
1360 case DIFF_SYMBOL_SUBMODULE_HEADER:
1361 case DIFF_SYMBOL_SUBMODULE_ERROR:
1362 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1363 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1364 case DIFF_SYMBOL_SUMMARY:
1365 case DIFF_SYMBOL_STATS_LINE:
1366 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1367 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1368 emit_line(o, "", "", line, len);
1369 break;
1370 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1371 case DIFF_SYMBOL_CONTEXT_MARKER:
1372 context = diff_get_color_opt(o, DIFF_CONTEXT);
1373 reset = diff_get_color_opt(o, DIFF_RESET);
1374 emit_line(o, context, reset, line, len);
1375 break;
1376 case DIFF_SYMBOL_SEPARATOR:
1377 fprintf(o->file, "%s%c",
1378 diff_line_prefix(o),
1379 o->line_termination);
1380 break;
1381 case DIFF_SYMBOL_CONTEXT:
1382 set = diff_get_color_opt(o, DIFF_CONTEXT);
1383 reset = diff_get_color_opt(o, DIFF_RESET);
1384 set_sign = NULL;
1385 if (o->flags.dual_color_diffed_diffs) {
1386 char c = !len ? 0 : line[0];
1387
1388 if (c == '+')
1389 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1390 else if (c == '@')
1391 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1392 else if (c == '-')
1393 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1394 }
1395 emit_line_ws_markup(o, set_sign, set, reset,
1396 OUTPUT_INDICATOR_CONTEXT, line, len,
1397 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1398 break;
1399 case DIFF_SYMBOL_PLUS:
1400 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1401 DIFF_SYMBOL_MOVED_LINE_ALT |
1402 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1403 case DIFF_SYMBOL_MOVED_LINE |
1404 DIFF_SYMBOL_MOVED_LINE_ALT |
1405 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1406 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1407 break;
1408 case DIFF_SYMBOL_MOVED_LINE |
1409 DIFF_SYMBOL_MOVED_LINE_ALT:
1410 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1411 break;
1412 case DIFF_SYMBOL_MOVED_LINE |
1413 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1414 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1415 break;
1416 case DIFF_SYMBOL_MOVED_LINE:
1417 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1418 break;
1419 default:
1420 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1421 }
1422 reset = diff_get_color_opt(o, DIFF_RESET);
1423 if (!o->flags.dual_color_diffed_diffs)
1424 set_sign = NULL;
1425 else {
1426 char c = !len ? 0 : line[0];
1427
1428 set_sign = set;
1429 if (c == '-')
1430 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1431 else if (c == '@')
1432 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1433 else if (c == '+')
1434 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1435 else
1436 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1437 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1438 }
1439 emit_line_ws_markup(o, set_sign, set, reset,
1440 OUTPUT_INDICATOR_NEW, line, len,
1441 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1442 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1443 break;
1444 case DIFF_SYMBOL_MINUS:
1445 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1446 DIFF_SYMBOL_MOVED_LINE_ALT |
1447 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1448 case DIFF_SYMBOL_MOVED_LINE |
1449 DIFF_SYMBOL_MOVED_LINE_ALT |
1450 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1451 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1452 break;
1453 case DIFF_SYMBOL_MOVED_LINE |
1454 DIFF_SYMBOL_MOVED_LINE_ALT:
1455 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1456 break;
1457 case DIFF_SYMBOL_MOVED_LINE |
1458 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1459 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1460 break;
1461 case DIFF_SYMBOL_MOVED_LINE:
1462 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1463 break;
1464 default:
1465 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1466 }
1467 reset = diff_get_color_opt(o, DIFF_RESET);
1468 if (!o->flags.dual_color_diffed_diffs)
1469 set_sign = NULL;
1470 else {
1471 char c = !len ? 0 : line[0];
1472
1473 set_sign = set;
1474 if (c == '+')
1475 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1476 else if (c == '@')
1477 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1478 else if (c == '-')
1479 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1480 else
1481 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1482 }
1483 emit_line_ws_markup(o, set_sign, set, reset,
1484 OUTPUT_INDICATOR_OLD, line, len,
1485 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1486 break;
1487 case DIFF_SYMBOL_WORDS_PORCELAIN:
1488 context = diff_get_color_opt(o, DIFF_CONTEXT);
1489 reset = diff_get_color_opt(o, DIFF_RESET);
1490 emit_line(o, context, reset, line, len);
1491 fputs("~\n", o->file);
1492 break;
1493 case DIFF_SYMBOL_WORDS:
1494 context = diff_get_color_opt(o, DIFF_CONTEXT);
1495 reset = diff_get_color_opt(o, DIFF_RESET);
1496 /*
1497 * Skip the prefix character, if any. With
1498 * diff_suppress_blank_empty, there may be
1499 * none.
1500 */
1501 if (line[0] != '\n') {
1502 line++;
1503 len--;
1504 }
1505 emit_line(o, context, reset, line, len);
1506 break;
1507 case DIFF_SYMBOL_FILEPAIR_PLUS:
1508 meta = diff_get_color_opt(o, DIFF_METAINFO);
1509 reset = diff_get_color_opt(o, DIFF_RESET);
1510 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1511 line, reset,
1512 strchr(line, ' ') ? "\t" : "");
1513 break;
1514 case DIFF_SYMBOL_FILEPAIR_MINUS:
1515 meta = diff_get_color_opt(o, DIFF_METAINFO);
1516 reset = diff_get_color_opt(o, DIFF_RESET);
1517 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1518 line, reset,
1519 strchr(line, ' ') ? "\t" : "");
1520 break;
1521 case DIFF_SYMBOL_BINARY_FILES:
1522 case DIFF_SYMBOL_HEADER:
1523 fprintf(o->file, "%s", line);
1524 break;
1525 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1526 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1527 break;
1528 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1529 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1530 break;
1531 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1532 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1533 break;
1534 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1535 fputs(diff_line_prefix(o), o->file);
1536 fputc('\n', o->file);
1537 break;
1538 case DIFF_SYMBOL_REWRITE_DIFF:
1539 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1540 reset = diff_get_color_opt(o, DIFF_RESET);
1541 emit_line(o, fraginfo, reset, line, len);
1542 break;
1543 case DIFF_SYMBOL_SUBMODULE_ADD:
1544 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1545 reset = diff_get_color_opt(o, DIFF_RESET);
1546 emit_line(o, set, reset, line, len);
1547 break;
1548 case DIFF_SYMBOL_SUBMODULE_DEL:
1549 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1550 reset = diff_get_color_opt(o, DIFF_RESET);
1551 emit_line(o, set, reset, line, len);
1552 break;
1553 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1554 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1555 diff_line_prefix(o), line);
1556 break;
1557 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1558 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1559 diff_line_prefix(o), line);
1560 break;
1561 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1562 emit_line(o, "", "", " 0 files changed\n",
1563 strlen(" 0 files changed\n"));
1564 break;
1565 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1566 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1567 break;
1568 case DIFF_SYMBOL_WORD_DIFF:
1569 fprintf(o->file, "%.*s", len, line);
1570 break;
1571 case DIFF_SYMBOL_STAT_SEP:
1572 fputs(o->stat_sep, o->file);
1573 break;
1574 default:
1575 BUG("unknown diff symbol");
1576 }
1577 }
1578
1579 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1580 const char *line, int len, unsigned flags)
1581 {
1582 struct emitted_diff_symbol e = {
1583 .line = line, .len = len, .flags = flags, .s = s
1584 };
1585
1586 if (o->emitted_symbols)
1587 append_emitted_diff_symbol(o, &e);
1588 else
1589 emit_diff_symbol_from_struct(o, &e);
1590 }
1591
1592 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1593 {
1594 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1595 }
1596
1597 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1598 {
1599 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1600 }
1601
1602 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1603 {
1604 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1605 path, strlen(path), 0);
1606 }
1607
1608 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1609 {
1610 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1611 path, strlen(path), 0);
1612 }
1613
1614 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1615 {
1616 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1617 header, strlen(header), 0);
1618 }
1619
1620 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1621 {
1622 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1623 }
1624
1625 void diff_emit_submodule_pipethrough(struct diff_options *o,
1626 const char *line, int len)
1627 {
1628 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1629 }
1630
1631 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1632 {
1633 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1634 ecbdata->blank_at_eof_in_preimage &&
1635 ecbdata->blank_at_eof_in_postimage &&
1636 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1637 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1638 return 0;
1639 return ws_blank_line(line, len);
1640 }
1641
1642 static void emit_add_line(struct emit_callback *ecbdata,
1643 const char *line, int len)
1644 {
1645 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1646 if (new_blank_line_at_eof(ecbdata, line, len))
1647 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1648
1649 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1650 }
1651
1652 static void emit_del_line(struct emit_callback *ecbdata,
1653 const char *line, int len)
1654 {
1655 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1656 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1657 }
1658
1659 static void emit_context_line(struct emit_callback *ecbdata,
1660 const char *line, int len)
1661 {
1662 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1663 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1664 }
1665
1666 static void emit_hunk_header(struct emit_callback *ecbdata,
1667 const char *line, int len)
1668 {
1669 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1670 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1671 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1672 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1673 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1674 static const char atat[2] = { '@', '@' };
1675 const char *cp, *ep;
1676 struct strbuf msgbuf = STRBUF_INIT;
1677 int org_len = len;
1678 int i = 1;
1679
1680 /*
1681 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1682 * it always is at least 10 bytes long.
1683 */
1684 if (len < 10 ||
1685 memcmp(line, atat, 2) ||
1686 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1687 emit_diff_symbol(ecbdata->opt,
1688 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1689 return;
1690 }
1691 ep += 2; /* skip over @@ */
1692
1693 /* The hunk header in fraginfo color */
1694 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1695 strbuf_addstr(&msgbuf, reverse);
1696 strbuf_addstr(&msgbuf, frag);
1697 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1698 strbuf_add(&msgbuf, atat, sizeof(atat));
1699 else
1700 strbuf_add(&msgbuf, line, ep - line);
1701 strbuf_addstr(&msgbuf, reset);
1702
1703 /*
1704 * trailing "\r\n"
1705 */
1706 for ( ; i < 3; i++)
1707 if (line[len - i] == '\r' || line[len - i] == '\n')
1708 len--;
1709
1710 /* blank before the func header */
1711 for (cp = ep; ep - line < len; ep++)
1712 if (*ep != ' ' && *ep != '\t')
1713 break;
1714 if (ep != cp) {
1715 strbuf_addstr(&msgbuf, context);
1716 strbuf_add(&msgbuf, cp, ep - cp);
1717 strbuf_addstr(&msgbuf, reset);
1718 }
1719
1720 if (ep < line + len) {
1721 strbuf_addstr(&msgbuf, func);
1722 strbuf_add(&msgbuf, ep, line + len - ep);
1723 strbuf_addstr(&msgbuf, reset);
1724 }
1725
1726 strbuf_add(&msgbuf, line + len, org_len - len);
1727 strbuf_complete_line(&msgbuf);
1728 emit_diff_symbol(ecbdata->opt,
1729 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1730 strbuf_release(&msgbuf);
1731 }
1732
1733 static struct diff_tempfile *claim_diff_tempfile(void)
1734 {
1735 int i;
1736 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1737 if (!diff_temp[i].name)
1738 return diff_temp + i;
1739 BUG("diff is failing to clean up its tempfiles");
1740 }
1741
1742 static void remove_tempfile(void)
1743 {
1744 int i;
1745 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1746 if (is_tempfile_active(diff_temp[i].tempfile))
1747 delete_tempfile(&diff_temp[i].tempfile);
1748 diff_temp[i].name = NULL;
1749 }
1750 }
1751
1752 static void add_line_count(struct strbuf *out, int count)
1753 {
1754 switch (count) {
1755 case 0:
1756 strbuf_addstr(out, "0,0");
1757 break;
1758 case 1:
1759 strbuf_addstr(out, "1");
1760 break;
1761 default:
1762 strbuf_addf(out, "1,%d", count);
1763 break;
1764 }
1765 }
1766
1767 static void emit_rewrite_lines(struct emit_callback *ecb,
1768 int prefix, const char *data, int size)
1769 {
1770 const char *endp = NULL;
1771
1772 while (0 < size) {
1773 int len;
1774
1775 endp = memchr(data, '\n', size);
1776 len = endp ? (endp - data + 1) : size;
1777 if (prefix != '+') {
1778 ecb->lno_in_preimage++;
1779 emit_del_line(ecb, data, len);
1780 } else {
1781 ecb->lno_in_postimage++;
1782 emit_add_line(ecb, data, len);
1783 }
1784 size -= len;
1785 data += len;
1786 }
1787 if (!endp)
1788 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1789 }
1790
1791 static void emit_rewrite_diff(const char *name_a,
1792 const char *name_b,
1793 struct diff_filespec *one,
1794 struct diff_filespec *two,
1795 struct userdiff_driver *textconv_one,
1796 struct userdiff_driver *textconv_two,
1797 struct diff_options *o)
1798 {
1799 int lc_a, lc_b;
1800 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1801 const char *a_prefix, *b_prefix;
1802 char *data_one, *data_two;
1803 size_t size_one, size_two;
1804 struct emit_callback ecbdata;
1805 struct strbuf out = STRBUF_INIT;
1806
1807 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1808 a_prefix = o->b_prefix;
1809 b_prefix = o->a_prefix;
1810 } else {
1811 a_prefix = o->a_prefix;
1812 b_prefix = o->b_prefix;
1813 }
1814
1815 name_a += (*name_a == '/');
1816 name_b += (*name_b == '/');
1817
1818 strbuf_reset(&a_name);
1819 strbuf_reset(&b_name);
1820 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1821 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1822
1823 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1824 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1825
1826 memset(&ecbdata, 0, sizeof(ecbdata));
1827 ecbdata.color_diff = want_color(o->use_color);
1828 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1829 ecbdata.opt = o;
1830 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1831 mmfile_t mf1, mf2;
1832 mf1.ptr = (char *)data_one;
1833 mf2.ptr = (char *)data_two;
1834 mf1.size = size_one;
1835 mf2.size = size_two;
1836 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1837 }
1838 ecbdata.lno_in_preimage = 1;
1839 ecbdata.lno_in_postimage = 1;
1840
1841 lc_a = count_lines(data_one, size_one);
1842 lc_b = count_lines(data_two, size_two);
1843
1844 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1845 a_name.buf, a_name.len, 0);
1846 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1847 b_name.buf, b_name.len, 0);
1848
1849 strbuf_addstr(&out, "@@ -");
1850 if (!o->irreversible_delete)
1851 add_line_count(&out, lc_a);
1852 else
1853 strbuf_addstr(&out, "?,?");
1854 strbuf_addstr(&out, " +");
1855 add_line_count(&out, lc_b);
1856 strbuf_addstr(&out, " @@\n");
1857 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1858 strbuf_release(&out);
1859
1860 if (lc_a && !o->irreversible_delete)
1861 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1862 if (lc_b)
1863 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1864 if (textconv_one)
1865 free((char *)data_one);
1866 if (textconv_two)
1867 free((char *)data_two);
1868 }
1869
1870 struct diff_words_buffer {
1871 mmfile_t text;
1872 unsigned long alloc;
1873 struct diff_words_orig {
1874 const char *begin, *end;
1875 } *orig;
1876 int orig_nr, orig_alloc;
1877 };
1878
1879 static void diff_words_append(char *line, unsigned long len,
1880 struct diff_words_buffer *buffer)
1881 {
1882 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1883 line++;
1884 len--;
1885 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1886 buffer->text.size += len;
1887 buffer->text.ptr[buffer->text.size] = '\0';
1888 }
1889
1890 struct diff_words_style_elem {
1891 const char *prefix;
1892 const char *suffix;
1893 const char *color; /* NULL; filled in by the setup code if
1894 * color is enabled */
1895 };
1896
1897 struct diff_words_style {
1898 enum diff_words_type type;
1899 struct diff_words_style_elem new_word, old_word, ctx;
1900 const char *newline;
1901 };
1902
1903 static struct diff_words_style diff_words_styles[] = {
1904 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1905 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1906 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1907 };
1908
1909 struct diff_words_data {
1910 struct diff_words_buffer minus, plus;
1911 const char *current_plus;
1912 int last_minus;
1913 struct diff_options *opt;
1914 regex_t *word_regex;
1915 enum diff_words_type type;
1916 struct diff_words_style *style;
1917 };
1918
1919 static int fn_out_diff_words_write_helper(struct diff_options *o,
1920 struct diff_words_style_elem *st_el,
1921 const char *newline,
1922 size_t count, const char *buf)
1923 {
1924 int print = 0;
1925 struct strbuf sb = STRBUF_INIT;
1926
1927 while (count) {
1928 char *p = memchr(buf, '\n', count);
1929 if (print)
1930 strbuf_addstr(&sb, diff_line_prefix(o));
1931
1932 if (p != buf) {
1933 const char *reset = st_el->color && *st_el->color ?
1934 GIT_COLOR_RESET : NULL;
1935 if (st_el->color && *st_el->color)
1936 strbuf_addstr(&sb, st_el->color);
1937 strbuf_addstr(&sb, st_el->prefix);
1938 strbuf_add(&sb, buf, p ? p - buf : count);
1939 strbuf_addstr(&sb, st_el->suffix);
1940 if (reset)
1941 strbuf_addstr(&sb, reset);
1942 }
1943 if (!p)
1944 goto out;
1945
1946 strbuf_addstr(&sb, newline);
1947 count -= p + 1 - buf;
1948 buf = p + 1;
1949 print = 1;
1950 if (count) {
1951 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1952 sb.buf, sb.len, 0);
1953 strbuf_reset(&sb);
1954 }
1955 }
1956
1957 out:
1958 if (sb.len)
1959 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1960 sb.buf, sb.len, 0);
1961 strbuf_release(&sb);
1962 return 0;
1963 }
1964
1965 /*
1966 * '--color-words' algorithm can be described as:
1967 *
1968 * 1. collect the minus/plus lines of a diff hunk, divided into
1969 * minus-lines and plus-lines;
1970 *
1971 * 2. break both minus-lines and plus-lines into words and
1972 * place them into two mmfile_t with one word for each line;
1973 *
1974 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1975 *
1976 * And for the common parts of the both file, we output the plus side text.
1977 * diff_words->current_plus is used to trace the current position of the plus file
1978 * which printed. diff_words->last_minus is used to trace the last minus word
1979 * printed.
1980 *
1981 * For '--graph' to work with '--color-words', we need to output the graph prefix
1982 * on each line of color words output. Generally, there are two conditions on
1983 * which we should output the prefix.
1984 *
1985 * 1. diff_words->last_minus == 0 &&
1986 * diff_words->current_plus == diff_words->plus.text.ptr
1987 *
1988 * that is: the plus text must start as a new line, and if there is no minus
1989 * word printed, a graph prefix must be printed.
1990 *
1991 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1992 * *(diff_words->current_plus - 1) == '\n'
1993 *
1994 * that is: a graph prefix must be printed following a '\n'
1995 */
1996 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1997 {
1998 if ((diff_words->last_minus == 0 &&
1999 diff_words->current_plus == diff_words->plus.text.ptr) ||
2000 (diff_words->current_plus > diff_words->plus.text.ptr &&
2001 *(diff_words->current_plus - 1) == '\n')) {
2002 return 1;
2003 } else {
2004 return 0;
2005 }
2006 }
2007
2008 static void fn_out_diff_words_aux(void *priv,
2009 long minus_first, long minus_len,
2010 long plus_first, long plus_len,
2011 const char *func UNUSED, long funclen UNUSED)
2012 {
2013 struct diff_words_data *diff_words = priv;
2014 struct diff_words_style *style = diff_words->style;
2015 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
2016 struct diff_options *opt = diff_words->opt;
2017 const char *line_prefix;
2018
2019 assert(opt);
2020 line_prefix = diff_line_prefix(opt);
2021
2022 /* POSIX requires that first be decremented by one if len == 0... */
2023 if (minus_len) {
2024 minus_begin = diff_words->minus.orig[minus_first].begin;
2025 minus_end =
2026 diff_words->minus.orig[minus_first + minus_len - 1].end;
2027 } else
2028 minus_begin = minus_end =
2029 diff_words->minus.orig[minus_first].end;
2030
2031 if (plus_len) {
2032 plus_begin = diff_words->plus.orig[plus_first].begin;
2033 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2034 } else
2035 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2036
2037 if (color_words_output_graph_prefix(diff_words)) {
2038 fputs(line_prefix, diff_words->opt->file);
2039 }
2040 if (diff_words->current_plus != plus_begin) {
2041 fn_out_diff_words_write_helper(diff_words->opt,
2042 &style->ctx, style->newline,
2043 plus_begin - diff_words->current_plus,
2044 diff_words->current_plus);
2045 }
2046 if (minus_begin != minus_end) {
2047 fn_out_diff_words_write_helper(diff_words->opt,
2048 &style->old_word, style->newline,
2049 minus_end - minus_begin, minus_begin);
2050 }
2051 if (plus_begin != plus_end) {
2052 fn_out_diff_words_write_helper(diff_words->opt,
2053 &style->new_word, style->newline,
2054 plus_end - plus_begin, plus_begin);
2055 }
2056
2057 diff_words->current_plus = plus_end;
2058 diff_words->last_minus = minus_first;
2059 }
2060
2061 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2062 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2063 int *begin, int *end)
2064 {
2065 while (word_regex && *begin < buffer->size) {
2066 regmatch_t match[1];
2067 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2068 buffer->size - *begin, 1, match, 0)) {
2069 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2070 '\n', match[0].rm_eo - match[0].rm_so);
2071 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2072 *begin += match[0].rm_so;
2073 if (*begin == *end)
2074 (*begin)++;
2075 else
2076 return *begin > *end;
2077 } else {
2078 return -1;
2079 }
2080 }
2081
2082 /* find the next word */
2083 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2084 (*begin)++;
2085 if (*begin >= buffer->size)
2086 return -1;
2087
2088 /* find the end of the word */
2089 *end = *begin + 1;
2090 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2091 (*end)++;
2092
2093 return 0;
2094 }
2095
2096 /*
2097 * This function splits the words in buffer->text, stores the list with
2098 * newline separator into out, and saves the offsets of the original words
2099 * in buffer->orig.
2100 */
2101 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2102 regex_t *word_regex)
2103 {
2104 int i, j;
2105 long alloc = 0;
2106
2107 out->size = 0;
2108 out->ptr = NULL;
2109
2110 /* fake an empty "0th" word */
2111 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2112 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2113 buffer->orig_nr = 1;
2114
2115 for (i = 0; i < buffer->text.size; i++) {
2116 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2117 return;
2118
2119 /* store original boundaries */
2120 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2121 buffer->orig_alloc);
2122 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2123 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2124 buffer->orig_nr++;
2125
2126 /* store one word */
2127 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2128 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2129 out->ptr[out->size + j - i] = '\n';
2130 out->size += j - i + 1;
2131
2132 i = j - 1;
2133 }
2134 }
2135
2136 /* this executes the word diff on the accumulated buffers */
2137 static void diff_words_show(struct diff_words_data *diff_words)
2138 {
2139 xpparam_t xpp;
2140 xdemitconf_t xecfg;
2141 mmfile_t minus, plus;
2142 struct diff_words_style *style = diff_words->style;
2143
2144 struct diff_options *opt = diff_words->opt;
2145 const char *line_prefix;
2146
2147 assert(opt);
2148 line_prefix = diff_line_prefix(opt);
2149
2150 /* special case: only removal */
2151 if (!diff_words->plus.text.size) {
2152 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2153 line_prefix, strlen(line_prefix), 0);
2154 fn_out_diff_words_write_helper(diff_words->opt,
2155 &style->old_word, style->newline,
2156 diff_words->minus.text.size,
2157 diff_words->minus.text.ptr);
2158 diff_words->minus.text.size = 0;
2159 return;
2160 }
2161
2162 diff_words->current_plus = diff_words->plus.text.ptr;
2163 diff_words->last_minus = 0;
2164
2165 memset(&xpp, 0, sizeof(xpp));
2166 memset(&xecfg, 0, sizeof(xecfg));
2167 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2168 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2169 xpp.flags = 0;
2170 /* as only the hunk header will be parsed, we need a 0-context */
2171 xecfg.ctxlen = 0;
2172 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2173 diff_words, &xpp, &xecfg))
2174 die("unable to generate word diff");
2175 free(minus.ptr);
2176 free(plus.ptr);
2177 if (diff_words->current_plus != diff_words->plus.text.ptr +
2178 diff_words->plus.text.size) {
2179 if (color_words_output_graph_prefix(diff_words))
2180 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2181 line_prefix, strlen(line_prefix), 0);
2182 fn_out_diff_words_write_helper(diff_words->opt,
2183 &style->ctx, style->newline,
2184 diff_words->plus.text.ptr + diff_words->plus.text.size
2185 - diff_words->current_plus, diff_words->current_plus);
2186 }
2187 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2188 }
2189
2190 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2191 static void diff_words_flush(struct emit_callback *ecbdata)
2192 {
2193 struct diff_options *wo = ecbdata->diff_words->opt;
2194
2195 if (ecbdata->diff_words->minus.text.size ||
2196 ecbdata->diff_words->plus.text.size)
2197 diff_words_show(ecbdata->diff_words);
2198
2199 if (wo->emitted_symbols) {
2200 struct diff_options *o = ecbdata->opt;
2201 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2202 int i;
2203
2204 /*
2205 * NEEDSWORK:
2206 * Instead of appending each, concat all words to a line?
2207 */
2208 for (i = 0; i < wol->nr; i++)
2209 append_emitted_diff_symbol(o, &wol->buf[i]);
2210
2211 for (i = 0; i < wol->nr; i++)
2212 free((void *)wol->buf[i].line);
2213
2214 wol->nr = 0;
2215 }
2216 }
2217
2218 static void diff_filespec_load_driver(struct diff_filespec *one,
2219 struct index_state *istate)
2220 {
2221 /* Use already-loaded driver */
2222 if (one->driver)
2223 return;
2224
2225 if (S_ISREG(one->mode))
2226 one->driver = userdiff_find_by_path(istate, one->path);
2227
2228 /* Fallback to default settings */
2229 if (!one->driver)
2230 one->driver = userdiff_find_by_name("default");
2231 }
2232
2233 static const char *userdiff_word_regex(struct diff_filespec *one,
2234 struct index_state *istate)
2235 {
2236 diff_filespec_load_driver(one, istate);
2237 return one->driver->word_regex;
2238 }
2239
2240 static void init_diff_words_data(struct emit_callback *ecbdata,
2241 struct diff_options *orig_opts,
2242 struct diff_filespec *one,
2243 struct diff_filespec *two)
2244 {
2245 int i;
2246 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2247 memcpy(o, orig_opts, sizeof(struct diff_options));
2248
2249 CALLOC_ARRAY(ecbdata->diff_words, 1);
2250 ecbdata->diff_words->type = o->word_diff;
2251 ecbdata->diff_words->opt = o;
2252
2253 if (orig_opts->emitted_symbols)
2254 CALLOC_ARRAY(o->emitted_symbols, 1);
2255
2256 if (!o->word_regex)
2257 o->word_regex = userdiff_word_regex(one, o->repo->index);
2258 if (!o->word_regex)
2259 o->word_regex = userdiff_word_regex(two, o->repo->index);
2260 if (!o->word_regex)
2261 o->word_regex = diff_word_regex_cfg;
2262 if (o->word_regex) {
2263 ecbdata->diff_words->word_regex = (regex_t *)
2264 xmalloc(sizeof(regex_t));
2265 if (regcomp(ecbdata->diff_words->word_regex,
2266 o->word_regex,
2267 REG_EXTENDED | REG_NEWLINE))
2268 die("invalid regular expression: %s",
2269 o->word_regex);
2270 }
2271 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2272 if (o->word_diff == diff_words_styles[i].type) {
2273 ecbdata->diff_words->style =
2274 &diff_words_styles[i];
2275 break;
2276 }
2277 }
2278 if (want_color(o->use_color)) {
2279 struct diff_words_style *st = ecbdata->diff_words->style;
2280 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2281 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2282 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2283 }
2284 }
2285
2286 static void free_diff_words_data(struct emit_callback *ecbdata)
2287 {
2288 if (ecbdata->diff_words) {
2289 diff_words_flush(ecbdata);
2290 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2291 free (ecbdata->diff_words->opt);
2292 free (ecbdata->diff_words->minus.text.ptr);
2293 free (ecbdata->diff_words->minus.orig);
2294 free (ecbdata->diff_words->plus.text.ptr);
2295 free (ecbdata->diff_words->plus.orig);
2296 if (ecbdata->diff_words->word_regex) {
2297 regfree(ecbdata->diff_words->word_regex);
2298 free(ecbdata->diff_words->word_regex);
2299 }
2300 FREE_AND_NULL(ecbdata->diff_words);
2301 }
2302 }
2303
2304 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2305 {
2306 if (want_color(diff_use_color))
2307 return diff_colors[ix];
2308 return "";
2309 }
2310
2311 const char *diff_line_prefix(struct diff_options *opt)
2312 {
2313 struct strbuf *msgbuf;
2314 if (!opt->output_prefix)
2315 return "";
2316
2317 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2318 return msgbuf->buf;
2319 }
2320
2321 static unsigned long sane_truncate_line(char *line, unsigned long len)
2322 {
2323 const char *cp;
2324 unsigned long allot;
2325 size_t l = len;
2326
2327 cp = line;
2328 allot = l;
2329 while (0 < l) {
2330 (void) utf8_width(&cp, &l);
2331 if (!cp)
2332 break; /* truncated in the middle? */
2333 }
2334 return allot - l;
2335 }
2336
2337 static void find_lno(const char *line, struct emit_callback *ecbdata)
2338 {
2339 const char *p;
2340 ecbdata->lno_in_preimage = 0;
2341 ecbdata->lno_in_postimage = 0;
2342 p = strchr(line, '-');
2343 if (!p)
2344 return; /* cannot happen */
2345 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2346 p = strchr(p, '+');
2347 if (!p)
2348 return; /* cannot happen */
2349 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2350 }
2351
2352 static int fn_out_consume(void *priv, char *line, unsigned long len)
2353 {
2354 struct emit_callback *ecbdata = priv;
2355 struct diff_options *o = ecbdata->opt;
2356
2357 o->found_changes = 1;
2358
2359 if (ecbdata->header) {
2360 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2361 ecbdata->header->buf, ecbdata->header->len, 0);
2362 strbuf_reset(ecbdata->header);
2363 ecbdata->header = NULL;
2364 }
2365
2366 if (ecbdata->label_path[0]) {
2367 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2368 ecbdata->label_path[0],
2369 strlen(ecbdata->label_path[0]), 0);
2370 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2371 ecbdata->label_path[1],
2372 strlen(ecbdata->label_path[1]), 0);
2373 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2374 }
2375
2376 if (diff_suppress_blank_empty
2377 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2378 line[0] = '\n';
2379 len = 1;
2380 }
2381
2382 if (line[0] == '@') {
2383 if (ecbdata->diff_words)
2384 diff_words_flush(ecbdata);
2385 len = sane_truncate_line(line, len);
2386 find_lno(line, ecbdata);
2387 emit_hunk_header(ecbdata, line, len);
2388 return 0;
2389 }
2390
2391 if (ecbdata->diff_words) {
2392 enum diff_symbol s =
2393 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2394 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2395 if (line[0] == '-') {
2396 diff_words_append(line, len,
2397 &ecbdata->diff_words->minus);
2398 return 0;
2399 } else if (line[0] == '+') {
2400 diff_words_append(line, len,
2401 &ecbdata->diff_words->plus);
2402 return 0;
2403 } else if (starts_with(line, "\\ ")) {
2404 /*
2405 * Eat the "no newline at eof" marker as if we
2406 * saw a "+" or "-" line with nothing on it,
2407 * and return without diff_words_flush() to
2408 * defer processing. If this is the end of
2409 * preimage, more "+" lines may come after it.
2410 */
2411 return 0;
2412 }
2413 diff_words_flush(ecbdata);
2414 emit_diff_symbol(o, s, line, len, 0);
2415 return 0;
2416 }
2417
2418 switch (line[0]) {
2419 case '+':
2420 ecbdata->lno_in_postimage++;
2421 emit_add_line(ecbdata, line + 1, len - 1);
2422 break;
2423 case '-':
2424 ecbdata->lno_in_preimage++;
2425 emit_del_line(ecbdata, line + 1, len - 1);
2426 break;
2427 case ' ':
2428 ecbdata->lno_in_postimage++;
2429 ecbdata->lno_in_preimage++;
2430 emit_context_line(ecbdata, line + 1, len - 1);
2431 break;
2432 default:
2433 /* incomplete line at the end */
2434 ecbdata->lno_in_preimage++;
2435 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2436 line, len, 0);
2437 break;
2438 }
2439 return 0;
2440 }
2441
2442 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2443 {
2444 const char *old_name = a;
2445 const char *new_name = b;
2446 int pfx_length, sfx_length;
2447 int pfx_adjust_for_slash;
2448 int len_a = strlen(a);
2449 int len_b = strlen(b);
2450 int a_midlen, b_midlen;
2451 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2452 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2453
2454 if (qlen_a || qlen_b) {
2455 quote_c_style(a, name, NULL, 0);
2456 strbuf_addstr(name, " => ");
2457 quote_c_style(b, name, NULL, 0);
2458 return;
2459 }
2460
2461 /* Find common prefix */
2462 pfx_length = 0;
2463 while (*old_name && *new_name && *old_name == *new_name) {
2464 if (*old_name == '/')
2465 pfx_length = old_name - a + 1;
2466 old_name++;
2467 new_name++;
2468 }
2469
2470 /* Find common suffix */
2471 old_name = a + len_a;
2472 new_name = b + len_b;
2473 sfx_length = 0;
2474 /*
2475 * If there is a common prefix, it must end in a slash. In
2476 * that case we let this loop run 1 into the prefix to see the
2477 * same slash.
2478 *
2479 * If there is no common prefix, we cannot do this as it would
2480 * underrun the input strings.
2481 */
2482 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2483 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2484 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2485 *old_name == *new_name) {
2486 if (*old_name == '/')
2487 sfx_length = len_a - (old_name - a);
2488 old_name--;
2489 new_name--;
2490 }
2491
2492 /*
2493 * pfx{mid-a => mid-b}sfx
2494 * {pfx-a => pfx-b}sfx
2495 * pfx{sfx-a => sfx-b}
2496 * name-a => name-b
2497 */
2498 a_midlen = len_a - pfx_length - sfx_length;
2499 b_midlen = len_b - pfx_length - sfx_length;
2500 if (a_midlen < 0)
2501 a_midlen = 0;
2502 if (b_midlen < 0)
2503 b_midlen = 0;
2504
2505 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2506 if (pfx_length + sfx_length) {
2507 strbuf_add(name, a, pfx_length);
2508 strbuf_addch(name, '{');
2509 }
2510 strbuf_add(name, a + pfx_length, a_midlen);
2511 strbuf_addstr(name, " => ");
2512 strbuf_add(name, b + pfx_length, b_midlen);
2513 if (pfx_length + sfx_length) {
2514 strbuf_addch(name, '}');
2515 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2516 }
2517 }
2518
2519 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2520 const char *name_a,
2521 const char *name_b)
2522 {
2523 struct diffstat_file *x;
2524 CALLOC_ARRAY(x, 1);
2525 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2526 diffstat->files[diffstat->nr++] = x;
2527 if (name_b) {
2528 x->from_name = xstrdup(name_a);
2529 x->name = xstrdup(name_b);
2530 x->is_renamed = 1;
2531 }
2532 else {
2533 x->from_name = NULL;
2534 x->name = xstrdup(name_a);
2535 }
2536 return x;
2537 }
2538
2539 static int diffstat_consume(void *priv, char *line, unsigned long len)
2540 {
2541 struct diffstat_t *diffstat = priv;
2542 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2543
2544 if (!len)
2545 BUG("xdiff fed us an empty line");
2546
2547 if (line[0] == '+')
2548 x->added++;
2549 else if (line[0] == '-')
2550 x->deleted++;
2551 return 0;
2552 }
2553
2554 const char mime_boundary_leader[] = "------------";
2555
2556 static int scale_linear(int it, int width, int max_change)
2557 {
2558 if (!it)
2559 return 0;
2560 /*
2561 * make sure that at least one '-' or '+' is printed if
2562 * there is any change to this path. The easiest way is to
2563 * scale linearly as if the allotted width is one column shorter
2564 * than it is, and then add 1 to the result.
2565 */
2566 return 1 + (it * (width - 1) / max_change);
2567 }
2568
2569 static void show_graph(struct strbuf *out, char ch, int cnt,
2570 const char *set, const char *reset)
2571 {
2572 if (cnt <= 0)
2573 return;
2574 strbuf_addstr(out, set);
2575 strbuf_addchars(out, ch, cnt);
2576 strbuf_addstr(out, reset);
2577 }
2578
2579 static void fill_print_name(struct diffstat_file *file)
2580 {
2581 struct strbuf pname = STRBUF_INIT;
2582
2583 if (file->print_name)
2584 return;
2585
2586 if (file->is_renamed)
2587 pprint_rename(&pname, file->from_name, file->name);
2588 else
2589 quote_c_style(file->name, &pname, NULL, 0);
2590
2591 if (file->comments)
2592 strbuf_addf(&pname, " (%s)", file->comments);
2593
2594 file->print_name = strbuf_detach(&pname, NULL);
2595 }
2596
2597 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2598 int files, int insertions, int deletions)
2599 {
2600 struct strbuf sb = STRBUF_INIT;
2601
2602 if (!files) {
2603 assert(insertions == 0 && deletions == 0);
2604 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2605 NULL, 0, 0);
2606 return;
2607 }
2608
2609 strbuf_addf(&sb,
2610 (files == 1) ? " %d file changed" : " %d files changed",
2611 files);
2612
2613 /*
2614 * For binary diff, the caller may want to print "x files
2615 * changed" with insertions == 0 && deletions == 0.
2616 *
2617 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2618 * is probably less confusing (i.e skip over "2 files changed
2619 * but nothing about added/removed lines? Is this a bug in Git?").
2620 */
2621 if (insertions || deletions == 0) {
2622 strbuf_addf(&sb,
2623 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2624 insertions);
2625 }
2626
2627 if (deletions || insertions == 0) {
2628 strbuf_addf(&sb,
2629 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2630 deletions);
2631 }
2632 strbuf_addch(&sb, '\n');
2633 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2634 sb.buf, sb.len, 0);
2635 strbuf_release(&sb);
2636 }
2637
2638 void print_stat_summary(FILE *fp, int files,
2639 int insertions, int deletions)
2640 {
2641 struct diff_options o;
2642 memset(&o, 0, sizeof(o));
2643 o.file = fp;
2644
2645 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2646 }
2647
2648 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2649 {
2650 int i, len, add, del, adds = 0, dels = 0;
2651 uintmax_t max_change = 0, max_len = 0;
2652 int total_files = data->nr, count;
2653 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2654 const char *reset, *add_c, *del_c;
2655 int extra_shown = 0;
2656 const char *line_prefix = diff_line_prefix(options);
2657 struct strbuf out = STRBUF_INIT;
2658
2659 if (data->nr == 0)
2660 return;
2661
2662 count = options->stat_count ? options->stat_count : data->nr;
2663
2664 reset = diff_get_color_opt(options, DIFF_RESET);
2665 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2666 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2667
2668 /*
2669 * Find the longest filename and max number of changes
2670 */
2671 for (i = 0; (i < count) && (i < data->nr); i++) {
2672 struct diffstat_file *file = data->files[i];
2673 uintmax_t change = file->added + file->deleted;
2674
2675 if (!file->is_interesting && (change == 0)) {
2676 count++; /* not shown == room for one more */
2677 continue;
2678 }
2679 fill_print_name(file);
2680 len = utf8_strwidth(file->print_name);
2681 if (max_len < len)
2682 max_len = len;
2683
2684 if (file->is_unmerged) {
2685 /* "Unmerged" is 8 characters */
2686 bin_width = bin_width < 8 ? 8 : bin_width;
2687 continue;
2688 }
2689 if (file->is_binary) {
2690 /* "Bin XXX -> YYY bytes" */
2691 int w = 14 + decimal_width(file->added)
2692 + decimal_width(file->deleted);
2693 bin_width = bin_width < w ? w : bin_width;
2694 /* Display change counts aligned with "Bin" */
2695 number_width = 3;
2696 continue;
2697 }
2698
2699 if (max_change < change)
2700 max_change = change;
2701 }
2702 count = i; /* where we can stop scanning in data->files[] */
2703
2704 /*
2705 * We have width = stat_width or term_columns() columns total.
2706 * We want a maximum of min(max_len, stat_name_width) for the name part.
2707 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2708 * We also need 1 for " " and 4 + decimal_width(max_change)
2709 * for " | NNNN " and one the empty column at the end, altogether
2710 * 6 + decimal_width(max_change).
2711 *
2712 * If there's not enough space, we will use the smaller of
2713 * stat_name_width (if set) and 5/8*width for the filename,
2714 * and the rest for constant elements + graph part, but no more
2715 * than stat_graph_width for the graph part.
2716 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2717 * for the standard terminal size).
2718 *
2719 * In other words: stat_width limits the maximum width, and
2720 * stat_name_width fixes the maximum width of the filename,
2721 * and is also used to divide available columns if there
2722 * aren't enough.
2723 *
2724 * Binary files are displayed with "Bin XXX -> YYY bytes"
2725 * instead of the change count and graph. This part is treated
2726 * similarly to the graph part, except that it is not
2727 * "scaled". If total width is too small to accommodate the
2728 * guaranteed minimum width of the filename part and the
2729 * separators and this message, this message will "overflow"
2730 * making the line longer than the maximum width.
2731 */
2732
2733 /*
2734 * NEEDSWORK: line_prefix is often used for "log --graph" output
2735 * and contains ANSI-colored string. utf8_strnwidth() should be
2736 * used to correctly count the display width instead of strlen().
2737 */
2738 if (options->stat_width == -1)
2739 width = term_columns() - strlen(line_prefix);
2740 else
2741 width = options->stat_width ? options->stat_width : 80;
2742 number_width = decimal_width(max_change) > number_width ?
2743 decimal_width(max_change) : number_width;
2744
2745 if (options->stat_name_width == -1)
2746 options->stat_name_width = diff_stat_name_width;
2747 if (options->stat_graph_width == -1)
2748 options->stat_graph_width = diff_stat_graph_width;
2749
2750 /*
2751 * Guarantee 3/8*16 == 6 for the graph part
2752 * and 5/8*16 == 10 for the filename part
2753 */
2754 if (width < 16 + 6 + number_width)
2755 width = 16 + 6 + number_width;
2756
2757 /*
2758 * First assign sizes that are wanted, ignoring available width.
2759 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2760 * starting from "XXX" should fit in graph_width.
2761 */
2762 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2763 if (options->stat_graph_width &&
2764 options->stat_graph_width < graph_width)
2765 graph_width = options->stat_graph_width;
2766
2767 name_width = (options->stat_name_width > 0 &&
2768 options->stat_name_width < max_len) ?
2769 options->stat_name_width : max_len;
2770
2771 /*
2772 * Adjust adjustable widths not to exceed maximum width
2773 */
2774 if (name_width + number_width + 6 + graph_width > width) {
2775 if (graph_width > width * 3/8 - number_width - 6) {
2776 graph_width = width * 3/8 - number_width - 6;
2777 if (graph_width < 6)
2778 graph_width = 6;
2779 }
2780
2781 if (options->stat_graph_width &&
2782 graph_width > options->stat_graph_width)
2783 graph_width = options->stat_graph_width;
2784 if (name_width > width - number_width - 6 - graph_width)
2785 name_width = width - number_width - 6 - graph_width;
2786 else
2787 graph_width = width - number_width - 6 - name_width;
2788 }
2789
2790 /*
2791 * From here name_width is the width of the name area,
2792 * and graph_width is the width of the graph area.
2793 * max_change is used to scale graph properly.
2794 */
2795 for (i = 0; i < count; i++) {
2796 const char *prefix = "";
2797 struct diffstat_file *file = data->files[i];
2798 char *name = file->print_name;
2799 uintmax_t added = file->added;
2800 uintmax_t deleted = file->deleted;
2801 int name_len, padding;
2802
2803 if (!file->is_interesting && (added + deleted == 0))
2804 continue;
2805
2806 /*
2807 * "scale" the filename
2808 */
2809 len = name_width;
2810 name_len = utf8_strwidth(name);
2811 if (name_width < name_len) {
2812 char *slash;
2813 prefix = "...";
2814 len -= 3;
2815 /*
2816 * NEEDSWORK: (name_len - len) counts the display
2817 * width, which would be shorter than the byte
2818 * length of the corresponding substring.
2819 * Advancing "name" by that number of bytes does
2820 * *NOT* skip over that many columns, so it is
2821 * very likely that chomping the pathname at the
2822 * slash we will find starting from "name" will
2823 * leave the resulting string still too long.
2824 */
2825 name += name_len - len;
2826 slash = strchr(name, '/');
2827 if (slash)
2828 name = slash;
2829 }
2830 padding = len - utf8_strwidth(name);
2831 if (padding < 0)
2832 padding = 0;
2833
2834 if (file->is_binary) {
2835 strbuf_addf(&out, " %s%s%*s | %*s",
2836 prefix, name, padding, "",
2837 number_width, "Bin");
2838 if (!added && !deleted) {
2839 strbuf_addch(&out, '\n');
2840 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2841 out.buf, out.len, 0);
2842 strbuf_reset(&out);
2843 continue;
2844 }
2845 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2846 del_c, deleted, reset);
2847 strbuf_addstr(&out, " -> ");
2848 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2849 add_c, added, reset);
2850 strbuf_addstr(&out, " bytes\n");
2851 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2852 out.buf, out.len, 0);
2853 strbuf_reset(&out);
2854 continue;
2855 }
2856 else if (file->is_unmerged) {
2857 strbuf_addf(&out, " %s%s%*s | %*s",
2858 prefix, name, padding, "",
2859 number_width, "Unmerged\n");
2860 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2861 out.buf, out.len, 0);
2862 strbuf_reset(&out);
2863 continue;
2864 }
2865
2866 /*
2867 * scale the add/delete
2868 */
2869 add = added;
2870 del = deleted;
2871
2872 if (graph_width <= max_change) {
2873 int total = scale_linear(add + del, graph_width, max_change);
2874 if (total < 2 && add && del)
2875 /* width >= 2 due to the sanity check */
2876 total = 2;
2877 if (add < del) {
2878 add = scale_linear(add, graph_width, max_change);
2879 del = total - add;
2880 } else {
2881 del = scale_linear(del, graph_width, max_change);
2882 add = total - del;
2883 }
2884 }
2885 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2886 prefix, name, padding, "",
2887 number_width, added + deleted,
2888 added + deleted ? " " : "");
2889 show_graph(&out, '+', add, add_c, reset);
2890 show_graph(&out, '-', del, del_c, reset);
2891 strbuf_addch(&out, '\n');
2892 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2893 out.buf, out.len, 0);
2894 strbuf_reset(&out);
2895 }
2896
2897 for (i = 0; i < data->nr; i++) {
2898 struct diffstat_file *file = data->files[i];
2899 uintmax_t added = file->added;
2900 uintmax_t deleted = file->deleted;
2901
2902 if (file->is_unmerged ||
2903 (!file->is_interesting && (added + deleted == 0))) {
2904 total_files--;
2905 continue;
2906 }
2907
2908 if (!file->is_binary) {
2909 adds += added;
2910 dels += deleted;
2911 }
2912 if (i < count)
2913 continue;
2914 if (!extra_shown)
2915 emit_diff_symbol(options,
2916 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2917 NULL, 0, 0);
2918 extra_shown = 1;
2919 }
2920
2921 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2922 strbuf_release(&out);
2923 }
2924
2925 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2926 {
2927 int i, adds = 0, dels = 0, total_files = data->nr;
2928
2929 if (data->nr == 0)
2930 return;
2931
2932 for (i = 0; i < data->nr; i++) {
2933 int added = data->files[i]->added;
2934 int deleted = data->files[i]->deleted;
2935
2936 if (data->files[i]->is_unmerged ||
2937 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2938 total_files--;
2939 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2940 adds += added;
2941 dels += deleted;
2942 }
2943 }
2944 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2945 }
2946
2947 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2948 {
2949 int i;
2950
2951 if (data->nr == 0)
2952 return;
2953
2954 for (i = 0; i < data->nr; i++) {
2955 struct diffstat_file *file = data->files[i];
2956
2957 fprintf(options->file, "%s", diff_line_prefix(options));
2958
2959 if (file->is_binary)
2960 fprintf(options->file, "-\t-\t");
2961 else
2962 fprintf(options->file,
2963 "%"PRIuMAX"\t%"PRIuMAX"\t",
2964 file->added, file->deleted);
2965 if (options->line_termination) {
2966 fill_print_name(file);
2967 if (!file->is_renamed)
2968 write_name_quoted(file->name, options->file,
2969 options->line_termination);
2970 else {
2971 fputs(file->print_name, options->file);
2972 putc(options->line_termination, options->file);
2973 }
2974 } else {
2975 if (file->is_renamed) {
2976 putc('\0', options->file);
2977 write_name_quoted(file->from_name, options->file, '\0');
2978 }
2979 write_name_quoted(file->name, options->file, '\0');
2980 }
2981 }
2982 }
2983
2984 struct dirstat_file {
2985 const char *name;
2986 unsigned long changed;
2987 };
2988
2989 struct dirstat_dir {
2990 struct dirstat_file *files;
2991 int alloc, nr, permille, cumulative;
2992 };
2993
2994 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2995 unsigned long changed, const char *base, int baselen)
2996 {
2997 unsigned long sum_changes = 0;
2998 unsigned int sources = 0;
2999 const char *line_prefix = diff_line_prefix(opt);
3000
3001 while (dir->nr) {
3002 struct dirstat_file *f = dir->files;
3003 int namelen = strlen(f->name);
3004 unsigned long changes;
3005 char *slash;
3006
3007 if (namelen < baselen)
3008 break;
3009 if (memcmp(f->name, base, baselen))
3010 break;
3011 slash = strchr(f->name + baselen, '/');
3012 if (slash) {
3013 int newbaselen = slash + 1 - f->name;
3014 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
3015 sources++;
3016 } else {
3017 changes = f->changed;
3018 dir->files++;
3019 dir->nr--;
3020 sources += 2;
3021 }
3022 sum_changes += changes;
3023 }
3024
3025 /*
3026 * We don't report dirstat's for
3027 * - the top level
3028 * - or cases where everything came from a single directory
3029 * under this directory (sources == 1).
3030 */
3031 if (baselen && sources != 1) {
3032 if (sum_changes) {
3033 int permille = sum_changes * 1000 / changed;
3034 if (permille >= dir->permille) {
3035 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
3036 permille / 10, permille % 10, baselen, base);
3037 if (!dir->cumulative)
3038 return 0;
3039 }
3040 }
3041 }
3042 return sum_changes;
3043 }
3044
3045 static int dirstat_compare(const void *_a, const void *_b)
3046 {
3047 const struct dirstat_file *a = _a;
3048 const struct dirstat_file *b = _b;
3049 return strcmp(a->name, b->name);
3050 }
3051
3052 static void conclude_dirstat(struct diff_options *options,
3053 struct dirstat_dir *dir,
3054 unsigned long changed)
3055 {
3056 struct dirstat_file *to_free = dir->files;
3057
3058 if (!changed) {
3059 /* This can happen even with many files, if everything was renames */
3060 ;
3061 } else {
3062 /* Show all directories with more than x% of the changes */
3063 QSORT(dir->files, dir->nr, dirstat_compare);
3064 gather_dirstat(options, dir, changed, "", 0);
3065 }
3066
3067 free(to_free);
3068 }
3069
3070 static void show_dirstat(struct diff_options *options)
3071 {
3072 int i;
3073 unsigned long changed;
3074 struct dirstat_dir dir;
3075 struct diff_queue_struct *q = &diff_queued_diff;
3076
3077 dir.files = NULL;
3078 dir.alloc = 0;
3079 dir.nr = 0;
3080 dir.permille = options->dirstat_permille;
3081 dir.cumulative = options->flags.dirstat_cumulative;
3082
3083 changed = 0;
3084 for (i = 0; i < q->nr; i++) {
3085 struct diff_filepair *p = q->queue[i];
3086 const char *name;
3087 unsigned long copied, added, damage;
3088 struct diff_populate_filespec_options dpf_options = {
3089 .check_size_only = 1,
3090 };
3091
3092 name = p->two->path ? p->two->path : p->one->path;
3093
3094 if (p->one->oid_valid && p->two->oid_valid &&
3095 oideq(&p->one->oid, &p->two->oid)) {
3096 /*
3097 * The SHA1 has not changed, so pre-/post-content is
3098 * identical. We can therefore skip looking at the
3099 * file contents altogether.
3100 */
3101 damage = 0;
3102 goto found_damage;
3103 }
3104
3105 if (options->flags.dirstat_by_file) {
3106 /*
3107 * In --dirstat-by-file mode, we don't really need to
3108 * look at the actual file contents at all.
3109 * The fact that the SHA1 changed is enough for us to
3110 * add this file to the list of results
3111 * (with each file contributing equal damage).
3112 */
3113 damage = 1;
3114 goto found_damage;
3115 }
3116
3117 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3118 diff_populate_filespec(options->repo, p->one, NULL);
3119 diff_populate_filespec(options->repo, p->two, NULL);
3120 diffcore_count_changes(options->repo,
3121 p->one, p->two, NULL, NULL,
3122 &copied, &added);
3123 diff_free_filespec_data(p->one);
3124 diff_free_filespec_data(p->two);
3125 } else if (DIFF_FILE_VALID(p->one)) {
3126 diff_populate_filespec(options->repo, p->one, &dpf_options);
3127 copied = added = 0;
3128 diff_free_filespec_data(p->one);
3129 } else if (DIFF_FILE_VALID(p->two)) {
3130 diff_populate_filespec(options->repo, p->two, &dpf_options);
3131 copied = 0;
3132 added = p->two->size;
3133 diff_free_filespec_data(p->two);
3134 } else
3135 continue;
3136
3137 /*
3138 * Original minus copied is the removed material,
3139 * added is the new material. They are both damages
3140 * made to the preimage.
3141 * If the resulting damage is zero, we know that
3142 * diffcore_count_changes() considers the two entries to
3143 * be identical, but since the oid changed, we
3144 * know that there must have been _some_ kind of change,
3145 * so we force all entries to have damage > 0.
3146 */
3147 damage = (p->one->size - copied) + added;
3148 if (!damage)
3149 damage = 1;
3150
3151 found_damage:
3152 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3153 dir.files[dir.nr].name = name;
3154 dir.files[dir.nr].changed = damage;
3155 changed += damage;
3156 dir.nr++;
3157 }
3158
3159 conclude_dirstat(options, &dir, changed);
3160 }
3161
3162 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3163 {
3164 int i;
3165 unsigned long changed;
3166 struct dirstat_dir dir;
3167
3168 if (data->nr == 0)
3169 return;
3170
3171 dir.files = NULL;
3172 dir.alloc = 0;
3173 dir.nr = 0;
3174 dir.permille = options->dirstat_permille;
3175 dir.cumulative = options->flags.dirstat_cumulative;
3176
3177 changed = 0;
3178 for (i = 0; i < data->nr; i++) {
3179 struct diffstat_file *file = data->files[i];
3180 unsigned long damage = file->added + file->deleted;
3181 if (file->is_binary)
3182 /*
3183 * binary files counts bytes, not lines. Must find some
3184 * way to normalize binary bytes vs. textual lines.
3185 * The following heuristic assumes that there are 64
3186 * bytes per "line".
3187 * This is stupid and ugly, but very cheap...
3188 */
3189 damage = DIV_ROUND_UP(damage, 64);
3190 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3191 dir.files[dir.nr].name = file->name;
3192 dir.files[dir.nr].changed = damage;
3193 changed += damage;
3194 dir.nr++;
3195 }
3196
3197 conclude_dirstat(options, &dir, changed);
3198 }
3199
3200 static void free_diffstat_file(struct diffstat_file *f)
3201 {
3202 free(f->print_name);
3203 free(f->name);
3204 free(f->from_name);
3205 free(f);
3206 }
3207
3208 void free_diffstat_info(struct diffstat_t *diffstat)
3209 {
3210 int i;
3211 for (i = 0; i < diffstat->nr; i++)
3212 free_diffstat_file(diffstat->files[i]);
3213 free(diffstat->files);
3214 }
3215
3216 struct checkdiff_t {
3217 const char *filename;
3218 int lineno;
3219 int conflict_marker_size;
3220 struct diff_options *o;
3221 unsigned ws_rule;
3222 unsigned status;
3223 };
3224
3225 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3226 {
3227 char firstchar;
3228 int cnt;
3229
3230 if (len < marker_size + 1)
3231 return 0;
3232 firstchar = line[0];
3233 switch (firstchar) {
3234 case '=': case '>': case '<': case '|':
3235 break;
3236 default:
3237 return 0;
3238 }
3239 for (cnt = 1; cnt < marker_size; cnt++)
3240 if (line[cnt] != firstchar)
3241 return 0;
3242 /* line[1] through line[marker_size-1] are same as firstchar */
3243 if (len < marker_size + 1 || !isspace(line[marker_size]))
3244 return 0;
3245 return 1;
3246 }
3247
3248 static void checkdiff_consume_hunk(void *priv,
3249 long ob UNUSED, long on UNUSED,
3250 long nb, long nn UNUSED,
3251 const char *func UNUSED, long funclen UNUSED)
3252
3253 {
3254 struct checkdiff_t *data = priv;
3255 data->lineno = nb - 1;
3256 }
3257
3258 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3259 {
3260 struct checkdiff_t *data = priv;
3261 int marker_size = data->conflict_marker_size;
3262 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3263 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3264 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3265 char *err;
3266 const char *line_prefix;
3267
3268 assert(data->o);
3269 line_prefix = diff_line_prefix(data->o);
3270
3271 if (line[0] == '+') {
3272 unsigned bad;
3273 data->lineno++;
3274 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3275 data->status |= 1;
3276 fprintf(data->o->file,
3277 "%s%s:%d: leftover conflict marker\n",
3278 line_prefix, data->filename, data->lineno);
3279 }
3280 bad = ws_check(line + 1, len - 1, data->ws_rule);
3281 if (!bad)
3282 return 0;
3283 data->status |= bad;
3284 err = whitespace_error_string(bad);
3285 fprintf(data->o->file, "%s%s:%d: %s.\n",
3286 line_prefix, data->filename, data->lineno, err);
3287 free(err);
3288 emit_line(data->o, set, reset, line, 1);
3289 ws_check_emit(line + 1, len - 1, data->ws_rule,
3290 data->o->file, set, reset, ws);
3291 } else if (line[0] == ' ') {
3292 data->lineno++;
3293 }
3294 return 0;
3295 }
3296
3297 static unsigned char *deflate_it(char *data,
3298 unsigned long size,
3299 unsigned long *result_size)
3300 {
3301 int bound;
3302 unsigned char *deflated;
3303 git_zstream stream;
3304
3305 git_deflate_init(&stream, zlib_compression_level);
3306 bound = git_deflate_bound(&stream, size);
3307 deflated = xmalloc(bound);
3308 stream.next_out = deflated;
3309 stream.avail_out = bound;
3310
3311 stream.next_in = (unsigned char *)data;
3312 stream.avail_in = size;
3313 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3314 ; /* nothing */
3315 git_deflate_end(&stream);
3316 *result_size = stream.total_out;
3317 return deflated;
3318 }
3319
3320 static void emit_binary_diff_body(struct diff_options *o,
3321 mmfile_t *one, mmfile_t *two)
3322 {
3323 void *cp;
3324 void *delta;
3325 void *deflated;
3326 void *data;
3327 unsigned long orig_size;
3328 unsigned long delta_size;
3329 unsigned long deflate_size;
3330 unsigned long data_size;
3331
3332 /* We could do deflated delta, or we could do just deflated two,
3333 * whichever is smaller.
3334 */
3335 delta = NULL;
3336 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3337 if (one->size && two->size) {
3338 delta = diff_delta(one->ptr, one->size,
3339 two->ptr, two->size,
3340 &delta_size, deflate_size);
3341 if (delta) {
3342 void *to_free = delta;
3343 orig_size = delta_size;
3344 delta = deflate_it(delta, delta_size, &delta_size);
3345 free(to_free);
3346 }
3347 }
3348
3349 if (delta && delta_size < deflate_size) {
3350 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3351 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3352 s, strlen(s), 0);
3353 free(s);
3354 free(deflated);
3355 data = delta;
3356 data_size = delta_size;
3357 } else {
3358 char *s = xstrfmt("%lu", two->size);
3359 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3360 s, strlen(s), 0);
3361 free(s);
3362 free(delta);
3363 data = deflated;
3364 data_size = deflate_size;
3365 }
3366
3367 /* emit data encoded in base85 */
3368 cp = data;
3369 while (data_size) {
3370 int len;
3371 int bytes = (52 < data_size) ? 52 : data_size;
3372 char line[71];
3373 data_size -= bytes;
3374 if (bytes <= 26)
3375 line[0] = bytes + 'A' - 1;
3376 else
3377 line[0] = bytes - 26 + 'a' - 1;
3378 encode_85(line + 1, cp, bytes);
3379 cp = (char *) cp + bytes;
3380
3381 len = strlen(line);
3382 line[len++] = '\n';
3383 line[len] = '\0';
3384
3385 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3386 line, len, 0);
3387 }
3388 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3389 free(data);
3390 }
3391
3392 static void emit_binary_diff(struct diff_options *o,
3393 mmfile_t *one, mmfile_t *two)
3394 {
3395 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3396 emit_binary_diff_body(o, one, two);
3397 emit_binary_diff_body(o, two, one);
3398 }
3399
3400 int diff_filespec_is_binary(struct repository *r,
3401 struct diff_filespec *one)
3402 {
3403 struct diff_populate_filespec_options dpf_options = {
3404 .check_binary = 1,
3405 };
3406
3407 if (one->is_binary == -1) {
3408 diff_filespec_load_driver(one, r->index);
3409 if (one->driver->binary != -1)
3410 one->is_binary = one->driver->binary;
3411 else {
3412 if (!one->data && DIFF_FILE_VALID(one))
3413 diff_populate_filespec(r, one, &dpf_options);
3414 if (one->is_binary == -1 && one->data)
3415 one->is_binary = buffer_is_binary(one->data,
3416 one->size);
3417 if (one->is_binary == -1)
3418 one->is_binary = 0;
3419 }
3420 }
3421 return one->is_binary;
3422 }
3423
3424 static const struct userdiff_funcname *
3425 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3426 {
3427 diff_filespec_load_driver(one, o->repo->index);
3428 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3429 }
3430
3431 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3432 {
3433 if (!options->a_prefix)
3434 options->a_prefix = a;
3435 if (!options->b_prefix)
3436 options->b_prefix = b;
3437 }
3438
3439 void diff_set_noprefix(struct diff_options *options)
3440 {
3441 options->a_prefix = options->b_prefix = "";
3442 }
3443
3444 void diff_set_default_prefix(struct diff_options *options)
3445 {
3446 options->a_prefix = diff_src_prefix ? diff_src_prefix : "a/";
3447 options->b_prefix = diff_dst_prefix ? diff_dst_prefix : "b/";
3448 }
3449
3450 struct userdiff_driver *get_textconv(struct repository *r,
3451 struct diff_filespec *one)
3452 {
3453 if (!DIFF_FILE_VALID(one))
3454 return NULL;
3455
3456 diff_filespec_load_driver(one, r->index);
3457 return userdiff_get_textconv(r, one->driver);
3458 }
3459
3460 static struct string_list *additional_headers(struct diff_options *o,
3461 const char *path)
3462 {
3463 if (!o->additional_path_headers)
3464 return NULL;
3465 return strmap_get(o->additional_path_headers, path);
3466 }
3467
3468 static void add_formatted_header(struct strbuf *msg,
3469 const char *header,
3470 const char *line_prefix,
3471 const char *meta,
3472 const char *reset)
3473 {
3474 const char *next, *newline;
3475
3476 for (next = header; *next; next = newline) {
3477 newline = strchrnul(next, '\n');
3478 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3479 (int)(newline - next), next, reset);
3480 if (*newline)
3481 newline++;
3482 }
3483 }
3484
3485 static void add_formatted_headers(struct strbuf *msg,
3486 struct string_list *more_headers,
3487 const char *line_prefix,
3488 const char *meta,
3489 const char *reset)
3490 {
3491 int i;
3492
3493 for (i = 0; i < more_headers->nr; i++)
3494 add_formatted_header(msg, more_headers->items[i].string,
3495 line_prefix, meta, reset);
3496 }
3497
3498 static int diff_filepair_is_phoney(struct diff_filespec *one,
3499 struct diff_filespec *two)
3500 {
3501 /*
3502 * This function specifically looks for pairs injected by
3503 * create_filepairs_for_header_only_notifications(). Such
3504 * pairs are "phoney" in that they do not represent any
3505 * content or even mode difference, but were inserted because
3506 * diff_queued_diff previously had no pair associated with
3507 * that path but we needed some pair to avoid losing the
3508 * "remerge CONFLICT" header associated with the path.
3509 */
3510 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3511 }
3512
3513 static int set_diff_algorithm(struct diff_options *opts,
3514 const char *alg)
3515 {
3516 long value = parse_algorithm_value(alg);
3517
3518 if (value < 0)
3519 return -1;
3520
3521 /* clear out previous settings */
3522 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3523 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3524 opts->xdl_opts |= value;
3525
3526 return 0;
3527 }
3528
3529 static void builtin_diff(const char *name_a,
3530 const char *name_b,
3531 struct diff_filespec *one,
3532 struct diff_filespec *two,
3533 const char *xfrm_msg,
3534 int must_show_header,
3535 struct diff_options *o,
3536 int complete_rewrite)
3537 {
3538 mmfile_t mf1, mf2;
3539 const char *lbl[2];
3540 char *a_one, *b_two;
3541 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3542 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3543 const char *a_prefix, *b_prefix;
3544 struct userdiff_driver *textconv_one = NULL;
3545 struct userdiff_driver *textconv_two = NULL;
3546 struct strbuf header = STRBUF_INIT;
3547 const char *line_prefix = diff_line_prefix(o);
3548
3549 diff_set_mnemonic_prefix(o, "a/", "b/");
3550 if (o->flags.reverse_diff) {
3551 a_prefix = o->b_prefix;
3552 b_prefix = o->a_prefix;
3553 } else {
3554 a_prefix = o->a_prefix;
3555 b_prefix = o->b_prefix;
3556 }
3557
3558 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3559 (!one->mode || S_ISGITLINK(one->mode)) &&
3560 (!two->mode || S_ISGITLINK(two->mode)) &&
3561 (!diff_filepair_is_phoney(one, two))) {
3562 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3563 &one->oid, &two->oid,
3564 two->dirty_submodule);
3565 return;
3566 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3567 (!one->mode || S_ISGITLINK(one->mode)) &&
3568 (!two->mode || S_ISGITLINK(two->mode)) &&
3569 (!diff_filepair_is_phoney(one, two))) {
3570 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3571 &one->oid, &two->oid,
3572 two->dirty_submodule);
3573 return;
3574 }
3575
3576 if (o->flags.allow_textconv) {
3577 textconv_one = get_textconv(o->repo, one);
3578 textconv_two = get_textconv(o->repo, two);
3579 }
3580
3581 /* Never use a non-valid filename anywhere if at all possible */
3582 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3583 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3584
3585 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3586 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3587 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3588 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3589 if (diff_filepair_is_phoney(one, two)) {
3590 /*
3591 * We should only reach this point for pairs generated from
3592 * create_filepairs_for_header_only_notifications(). For
3593 * these, we want to avoid the "/dev/null" special casing
3594 * above, because we do not want such pairs shown as either
3595 * "new file" or "deleted file" below.
3596 */
3597 lbl[0] = a_one;
3598 lbl[1] = b_two;
3599 }
3600 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3601 if (lbl[0][0] == '/') {
3602 /* /dev/null */
3603 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3604 if (xfrm_msg)
3605 strbuf_addstr(&header, xfrm_msg);
3606 o->found_changes = 1;
3607 must_show_header = 1;
3608 }
3609 else if (lbl[1][0] == '/') {
3610 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->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 {
3617 if (one->mode != two->mode) {
3618 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3619 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3620 o->found_changes = 1;
3621 must_show_header = 1;
3622 }
3623 if (xfrm_msg)
3624 strbuf_addstr(&header, xfrm_msg);
3625
3626 /*
3627 * we do not run diff between different kind
3628 * of objects.
3629 */
3630 if ((one->mode ^ two->mode) & S_IFMT)
3631 goto free_ab_and_return;
3632 if (complete_rewrite &&
3633 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3634 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3635 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3636 header.buf, header.len, 0);
3637 strbuf_reset(&header);
3638 emit_rewrite_diff(name_a, name_b, one, two,
3639 textconv_one, textconv_two, o);
3640 o->found_changes = 1;
3641 goto free_ab_and_return;
3642 }
3643 }
3644
3645 if (o->irreversible_delete && lbl[1][0] == '/') {
3646 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3647 header.len, 0);
3648 strbuf_reset(&header);
3649 goto free_ab_and_return;
3650 } else if (!o->flags.text &&
3651 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3652 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3653 struct strbuf sb = STRBUF_INIT;
3654 if (!one->data && !two->data &&
3655 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3656 !o->flags.binary) {
3657 if (oideq(&one->oid, &two->oid)) {
3658 if (must_show_header)
3659 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3660 header.buf, header.len,
3661 0);
3662 goto free_ab_and_return;
3663 }
3664 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3665 header.buf, header.len, 0);
3666 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3667 diff_line_prefix(o), lbl[0], lbl[1]);
3668 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3669 sb.buf, sb.len, 0);
3670 strbuf_release(&sb);
3671 goto free_ab_and_return;
3672 }
3673 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3674 fill_mmfile(o->repo, &mf2, two) < 0)
3675 die("unable to read files to diff");
3676 /* Quite common confusing case */
3677 if (mf1.size == mf2.size &&
3678 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3679 if (must_show_header)
3680 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3681 header.buf, header.len, 0);
3682 goto free_ab_and_return;
3683 }
3684 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3685 strbuf_reset(&header);
3686 if (o->flags.binary)
3687 emit_binary_diff(o, &mf1, &mf2);
3688 else {
3689 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3690 diff_line_prefix(o), lbl[0], lbl[1]);
3691 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3692 sb.buf, sb.len, 0);
3693 strbuf_release(&sb);
3694 }
3695 o->found_changes = 1;
3696 } else {
3697 /* Crazy xdl interfaces.. */
3698 const char *diffopts;
3699 const char *v;
3700 xpparam_t xpp;
3701 xdemitconf_t xecfg;
3702 struct emit_callback ecbdata;
3703 const struct userdiff_funcname *pe;
3704
3705 if (must_show_header) {
3706 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3707 header.buf, header.len, 0);
3708 strbuf_reset(&header);
3709 }
3710
3711 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3712 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3713
3714 pe = diff_funcname_pattern(o, one);
3715 if (!pe)
3716 pe = diff_funcname_pattern(o, two);
3717
3718 memset(&xpp, 0, sizeof(xpp));
3719 memset(&xecfg, 0, sizeof(xecfg));
3720 memset(&ecbdata, 0, sizeof(ecbdata));
3721 if (o->flags.suppress_diff_headers)
3722 lbl[0] = NULL;
3723 ecbdata.label_path = lbl;
3724 ecbdata.color_diff = want_color(o->use_color);
3725 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3726 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3727 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3728 ecbdata.opt = o;
3729 if (header.len && !o->flags.suppress_diff_headers)
3730 ecbdata.header = &header;
3731 xpp.flags = o->xdl_opts;
3732 xpp.ignore_regex = o->ignore_regex;
3733 xpp.ignore_regex_nr = o->ignore_regex_nr;
3734 xpp.anchors = o->anchors;
3735 xpp.anchors_nr = o->anchors_nr;
3736 xecfg.ctxlen = o->context;
3737 xecfg.interhunkctxlen = o->interhunkcontext;
3738 xecfg.flags = XDL_EMIT_FUNCNAMES;
3739 if (o->flags.funccontext)
3740 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3741 if (pe)
3742 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3743
3744 diffopts = getenv("GIT_DIFF_OPTS");
3745 if (!diffopts)
3746 ;
3747 else if (skip_prefix(diffopts, "--unified=", &v))
3748 xecfg.ctxlen = strtoul(v, NULL, 10);
3749 else if (skip_prefix(diffopts, "-u", &v))
3750 xecfg.ctxlen = strtoul(v, NULL, 10);
3751
3752 if (o->word_diff)
3753 init_diff_words_data(&ecbdata, o, one, two);
3754 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3755 &ecbdata, &xpp, &xecfg))
3756 die("unable to generate diff for %s", one->path);
3757 if (o->word_diff)
3758 free_diff_words_data(&ecbdata);
3759 if (textconv_one)
3760 free(mf1.ptr);
3761 if (textconv_two)
3762 free(mf2.ptr);
3763 xdiff_clear_find_func(&xecfg);
3764 }
3765
3766 free_ab_and_return:
3767 strbuf_release(&header);
3768 diff_free_filespec_data(one);
3769 diff_free_filespec_data(two);
3770 free(a_one);
3771 free(b_two);
3772 return;
3773 }
3774
3775 static const char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3776 {
3777 if (!is_renamed) {
3778 if (p->status == DIFF_STATUS_ADDED) {
3779 if (S_ISLNK(p->two->mode))
3780 return "new +l";
3781 else if ((p->two->mode & 0777) == 0755)
3782 return "new +x";
3783 else
3784 return "new";
3785 } else if (p->status == DIFF_STATUS_DELETED)
3786 return "gone";
3787 }
3788 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3789 return "mode -l";
3790 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3791 return "mode +l";
3792 else if ((p->one->mode & 0777) == 0644 &&
3793 (p->two->mode & 0777) == 0755)
3794 return "mode +x";
3795 else if ((p->one->mode & 0777) == 0755 &&
3796 (p->two->mode & 0777) == 0644)
3797 return "mode -x";
3798 return NULL;
3799 }
3800
3801 static void builtin_diffstat(const char *name_a, const char *name_b,
3802 struct diff_filespec *one,
3803 struct diff_filespec *two,
3804 struct diffstat_t *diffstat,
3805 struct diff_options *o,
3806 struct diff_filepair *p)
3807 {
3808 mmfile_t mf1, mf2;
3809 struct diffstat_file *data;
3810 int may_differ;
3811 int complete_rewrite = 0;
3812
3813 if (!DIFF_PAIR_UNMERGED(p)) {
3814 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3815 complete_rewrite = 1;
3816 }
3817
3818 data = diffstat_add(diffstat, name_a, name_b);
3819 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3820 if (o->flags.stat_with_summary)
3821 data->comments = get_compact_summary(p, data->is_renamed);
3822
3823 if (!one || !two) {
3824 data->is_unmerged = 1;
3825 return;
3826 }
3827
3828 /* saves some reads if true, not a guarantee of diff outcome */
3829 may_differ = !(one->oid_valid && two->oid_valid &&
3830 oideq(&one->oid, &two->oid));
3831
3832 if (diff_filespec_is_binary(o->repo, one) ||
3833 diff_filespec_is_binary(o->repo, two)) {
3834 data->is_binary = 1;
3835 if (!may_differ) {
3836 data->added = 0;
3837 data->deleted = 0;
3838 } else {
3839 data->added = diff_filespec_size(o->repo, two);
3840 data->deleted = diff_filespec_size(o->repo, one);
3841 }
3842 }
3843
3844 else if (complete_rewrite) {
3845 diff_populate_filespec(o->repo, one, NULL);
3846 diff_populate_filespec(o->repo, two, NULL);
3847 data->deleted = count_lines(one->data, one->size);
3848 data->added = count_lines(two->data, two->size);
3849 }
3850
3851 else if (may_differ) {
3852 /* Crazy xdl interfaces.. */
3853 xpparam_t xpp;
3854 xdemitconf_t xecfg;
3855
3856 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3857 fill_mmfile(o->repo, &mf2, two) < 0)
3858 die("unable to read files to diff");
3859
3860 memset(&xpp, 0, sizeof(xpp));
3861 memset(&xecfg, 0, sizeof(xecfg));
3862 xpp.flags = o->xdl_opts;
3863 xpp.ignore_regex = o->ignore_regex;
3864 xpp.ignore_regex_nr = o->ignore_regex_nr;
3865 xpp.anchors = o->anchors;
3866 xpp.anchors_nr = o->anchors_nr;
3867 xecfg.ctxlen = o->context;
3868 xecfg.interhunkctxlen = o->interhunkcontext;
3869 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3870 if (xdi_diff_outf(&mf1, &mf2, NULL,
3871 diffstat_consume, diffstat, &xpp, &xecfg))
3872 die("unable to generate diffstat for %s", one->path);
3873
3874 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3875 struct diffstat_file *file =
3876 diffstat->files[diffstat->nr - 1];
3877 /*
3878 * Omit diffstats of modified files where nothing changed.
3879 * Even if may_differ, this might be the case due to
3880 * ignoring whitespace changes, etc.
3881 *
3882 * But note that we special-case additions, deletions,
3883 * renames, and mode changes as adding an empty file,
3884 * for example is still of interest.
3885 */
3886 if ((p->status == DIFF_STATUS_MODIFIED)
3887 && !file->added
3888 && !file->deleted
3889 && one->mode == two->mode) {
3890 free_diffstat_file(file);
3891 diffstat->nr--;
3892 }
3893 }
3894 }
3895
3896 diff_free_filespec_data(one);
3897 diff_free_filespec_data(two);
3898 }
3899
3900 static void builtin_checkdiff(const char *name_a, const char *name_b,
3901 const char *attr_path,
3902 struct diff_filespec *one,
3903 struct diff_filespec *two,
3904 struct diff_options *o)
3905 {
3906 mmfile_t mf1, mf2;
3907 struct checkdiff_t data;
3908
3909 if (!two)
3910 return;
3911
3912 memset(&data, 0, sizeof(data));
3913 data.filename = name_b ? name_b : name_a;
3914 data.lineno = 0;
3915 data.o = o;
3916 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3917 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3918
3919 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3920 fill_mmfile(o->repo, &mf2, two) < 0)
3921 die("unable to read files to diff");
3922
3923 /*
3924 * All the other codepaths check both sides, but not checking
3925 * the "old" side here is deliberate. We are checking the newly
3926 * introduced changes, and as long as the "new" side is text, we
3927 * can and should check what it introduces.
3928 */
3929 if (diff_filespec_is_binary(o->repo, two))
3930 goto free_and_return;
3931 else {
3932 /* Crazy xdl interfaces.. */
3933 xpparam_t xpp;
3934 xdemitconf_t xecfg;
3935
3936 memset(&xpp, 0, sizeof(xpp));
3937 memset(&xecfg, 0, sizeof(xecfg));
3938 xecfg.ctxlen = 1; /* at least one context line */
3939 xpp.flags = 0;
3940 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3941 checkdiff_consume, &data,
3942 &xpp, &xecfg))
3943 die("unable to generate checkdiff for %s", one->path);
3944
3945 if (data.ws_rule & WS_BLANK_AT_EOF) {
3946 struct emit_callback ecbdata;
3947 int blank_at_eof;
3948
3949 ecbdata.ws_rule = data.ws_rule;
3950 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3951 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3952
3953 if (blank_at_eof) {
3954 static char *err;
3955 if (!err)
3956 err = whitespace_error_string(WS_BLANK_AT_EOF);
3957 fprintf(o->file, "%s:%d: %s.\n",
3958 data.filename, blank_at_eof, err);
3959 data.status = 1; /* report errors */
3960 }
3961 }
3962 }
3963 free_and_return:
3964 diff_free_filespec_data(one);
3965 diff_free_filespec_data(two);
3966 if (data.status)
3967 o->flags.check_failed = 1;
3968 }
3969
3970 struct diff_filespec *alloc_filespec(const char *path)
3971 {
3972 struct diff_filespec *spec;
3973
3974 FLEXPTR_ALLOC_STR(spec, path, path);
3975 spec->count = 1;
3976 spec->is_binary = -1;
3977 return spec;
3978 }
3979
3980 void free_filespec(struct diff_filespec *spec)
3981 {
3982 if (!--spec->count) {
3983 diff_free_filespec_data(spec);
3984 free(spec);
3985 }
3986 }
3987
3988 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3989 int oid_valid, unsigned short mode)
3990 {
3991 if (mode) {
3992 spec->mode = canon_mode(mode);
3993 oidcpy(&spec->oid, oid);
3994 spec->oid_valid = oid_valid;
3995 }
3996 }
3997
3998 /*
3999 * Given a name and sha1 pair, if the index tells us the file in
4000 * the work tree has that object contents, return true, so that
4001 * prepare_temp_file() does not have to inflate and extract.
4002 */
4003 static int reuse_worktree_file(struct index_state *istate,
4004 const char *name,
4005 const struct object_id *oid,
4006 int want_file)
4007 {
4008 const struct cache_entry *ce;
4009 struct stat st;
4010 int pos, len;
4011
4012 /*
4013 * We do not read the cache ourselves here, because the
4014 * benchmark with my previous version that always reads cache
4015 * shows that it makes things worse for diff-tree comparing
4016 * two linux-2.6 kernel trees in an already checked out work
4017 * tree. This is because most diff-tree comparisons deal with
4018 * only a small number of files, while reading the cache is
4019 * expensive for a large project, and its cost outweighs the
4020 * savings we get by not inflating the object to a temporary
4021 * file. Practically, this code only helps when we are used
4022 * by diff-cache --cached, which does read the cache before
4023 * calling us.
4024 */
4025 if (!istate->cache)
4026 return 0;
4027
4028 /* We want to avoid the working directory if our caller
4029 * doesn't need the data in a normal file, this system
4030 * is rather slow with its stat/open/mmap/close syscalls,
4031 * and the object is contained in a pack file. The pack
4032 * is probably already open and will be faster to obtain
4033 * the data through than the working directory. Loose
4034 * objects however would tend to be slower as they need
4035 * to be individually opened and inflated.
4036 */
4037 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
4038 return 0;
4039
4040 /*
4041 * Similarly, if we'd have to convert the file contents anyway, that
4042 * makes the optimization not worthwhile.
4043 */
4044 if (!want_file && would_convert_to_git(istate, name))
4045 return 0;
4046
4047 /*
4048 * If this path does not match our sparse-checkout definition,
4049 * then the file will not be in the working directory.
4050 */
4051 if (!path_in_sparse_checkout(name, istate))
4052 return 0;
4053
4054 len = strlen(name);
4055 pos = index_name_pos(istate, name, len);
4056 if (pos < 0)
4057 return 0;
4058 ce = istate->cache[pos];
4059
4060 /*
4061 * This is not the sha1 we are looking for, or
4062 * unreusable because it is not a regular file.
4063 */
4064 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4065 return 0;
4066
4067 /*
4068 * If ce is marked as "assume unchanged", there is no
4069 * guarantee that work tree matches what we are looking for.
4070 */
4071 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4072 return 0;
4073
4074 /*
4075 * If ce matches the file in the work tree, we can reuse it.
4076 */
4077 if (ce_uptodate(ce) ||
4078 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4079 return 1;
4080
4081 return 0;
4082 }
4083
4084 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4085 {
4086 struct strbuf buf = STRBUF_INIT;
4087 const char *dirty = "";
4088
4089 /* Are we looking at the work tree? */
4090 if (s->dirty_submodule)
4091 dirty = "-dirty";
4092
4093 strbuf_addf(&buf, "Subproject commit %s%s\n",
4094 oid_to_hex(&s->oid), dirty);
4095 s->size = buf.len;
4096 if (size_only) {
4097 s->data = NULL;
4098 strbuf_release(&buf);
4099 } else {
4100 s->data = strbuf_detach(&buf, NULL);
4101 s->should_free = 1;
4102 }
4103 return 0;
4104 }
4105
4106 /*
4107 * While doing rename detection and pickaxe operation, we may need to
4108 * grab the data for the blob (or file) for our own in-core comparison.
4109 * diff_filespec has data and size fields for this purpose.
4110 */
4111 int diff_populate_filespec(struct repository *r,
4112 struct diff_filespec *s,
4113 const struct diff_populate_filespec_options *options)
4114 {
4115 int size_only = options ? options->check_size_only : 0;
4116 int check_binary = options ? options->check_binary : 0;
4117 int err = 0;
4118 int conv_flags = global_conv_flags_eol;
4119 /*
4120 * demote FAIL to WARN to allow inspecting the situation
4121 * instead of refusing.
4122 */
4123 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4124 conv_flags = CONV_EOL_RNDTRP_WARN;
4125
4126 if (!DIFF_FILE_VALID(s))
4127 die("internal error: asking to populate invalid file.");
4128 if (S_ISDIR(s->mode))
4129 return -1;
4130
4131 if (s->data)
4132 return 0;
4133
4134 if (size_only && 0 < s->size)
4135 return 0;
4136
4137 if (S_ISGITLINK(s->mode))
4138 return diff_populate_gitlink(s, size_only);
4139
4140 if (!s->oid_valid ||
4141 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4142 struct strbuf buf = STRBUF_INIT;
4143 struct stat st;
4144 int fd;
4145
4146 if (lstat(s->path, &st) < 0) {
4147 err_empty:
4148 err = -1;
4149 empty:
4150 s->data = (char *)"";
4151 s->size = 0;
4152 return err;
4153 }
4154 s->size = xsize_t(st.st_size);
4155 if (!s->size)
4156 goto empty;
4157 if (S_ISLNK(st.st_mode)) {
4158 struct strbuf sb = STRBUF_INIT;
4159
4160 if (strbuf_readlink(&sb, s->path, s->size))
4161 goto err_empty;
4162 s->size = sb.len;
4163 s->data = strbuf_detach(&sb, NULL);
4164 s->should_free = 1;
4165 return 0;
4166 }
4167
4168 /*
4169 * Even if the caller would be happy with getting
4170 * only the size, we cannot return early at this
4171 * point if the path requires us to run the content
4172 * conversion.
4173 */
4174 if (size_only && !would_convert_to_git(r->index, s->path))
4175 return 0;
4176
4177 /*
4178 * Note: this check uses xsize_t(st.st_size) that may
4179 * not be the true size of the blob after it goes
4180 * through convert_to_git(). This may not strictly be
4181 * correct, but the whole point of big_file_threshold
4182 * and is_binary check being that we want to avoid
4183 * opening the file and inspecting the contents, this
4184 * is probably fine.
4185 */
4186 if (check_binary &&
4187 s->size > big_file_threshold && s->is_binary == -1) {
4188 s->is_binary = 1;
4189 return 0;
4190 }
4191 fd = open(s->path, O_RDONLY);
4192 if (fd < 0)
4193 goto err_empty;
4194 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4195 close(fd);
4196 s->should_munmap = 1;
4197
4198 /*
4199 * Convert from working tree format to canonical git format
4200 */
4201 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4202 size_t size = 0;
4203 munmap(s->data, s->size);
4204 s->should_munmap = 0;
4205 s->data = strbuf_detach(&buf, &size);
4206 s->size = size;
4207 s->should_free = 1;
4208 }
4209 }
4210 else {
4211 struct object_info info = {
4212 .sizep = &s->size
4213 };
4214
4215 if (!(size_only || check_binary))
4216 /*
4217 * Set contentp, since there is no chance that merely
4218 * the size is sufficient.
4219 */
4220 info.contentp = &s->data;
4221
4222 if (options && options->missing_object_cb) {
4223 if (!oid_object_info_extended(r, &s->oid, &info,
4224 OBJECT_INFO_LOOKUP_REPLACE |
4225 OBJECT_INFO_SKIP_FETCH_OBJECT))
4226 goto object_read;
4227 options->missing_object_cb(options->missing_object_data);
4228 }
4229 if (oid_object_info_extended(r, &s->oid, &info,
4230 OBJECT_INFO_LOOKUP_REPLACE))
4231 die("unable to read %s", oid_to_hex(&s->oid));
4232
4233 object_read:
4234 if (size_only || check_binary) {
4235 if (size_only)
4236 return 0;
4237 if (s->size > big_file_threshold && s->is_binary == -1) {
4238 s->is_binary = 1;
4239 return 0;
4240 }
4241 }
4242 if (!info.contentp) {
4243 info.contentp = &s->data;
4244 if (oid_object_info_extended(r, &s->oid, &info,
4245 OBJECT_INFO_LOOKUP_REPLACE))
4246 die("unable to read %s", oid_to_hex(&s->oid));
4247 }
4248 s->should_free = 1;
4249 }
4250 return 0;
4251 }
4252
4253 void diff_free_filespec_blob(struct diff_filespec *s)
4254 {
4255 if (s->should_free)
4256 free(s->data);
4257 else if (s->should_munmap)
4258 munmap(s->data, s->size);
4259
4260 if (s->should_free || s->should_munmap) {
4261 s->should_free = s->should_munmap = 0;
4262 s->data = NULL;
4263 }
4264 }
4265
4266 void diff_free_filespec_data(struct diff_filespec *s)
4267 {
4268 if (!s)
4269 return;
4270
4271 diff_free_filespec_blob(s);
4272 FREE_AND_NULL(s->cnt_data);
4273 }
4274
4275 static void prep_temp_blob(struct index_state *istate,
4276 const char *path, struct diff_tempfile *temp,
4277 void *blob,
4278 unsigned long size,
4279 const struct object_id *oid,
4280 int mode)
4281 {
4282 struct strbuf buf = STRBUF_INIT;
4283 char *path_dup = xstrdup(path);
4284 const char *base = basename(path_dup);
4285 struct checkout_metadata meta;
4286
4287 init_checkout_metadata(&meta, NULL, NULL, oid);
4288
4289 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4290 if (!temp->tempfile)
4291 die_errno("unable to create temp-file");
4292 if (convert_to_working_tree(istate, path,
4293 (const char *)blob, (size_t)size, &buf, &meta)) {
4294 blob = buf.buf;
4295 size = buf.len;
4296 }
4297 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4298 close_tempfile_gently(temp->tempfile))
4299 die_errno("unable to write temp-file");
4300 temp->name = get_tempfile_path(temp->tempfile);
4301 oid_to_hex_r(temp->hex, oid);
4302 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4303 strbuf_release(&buf);
4304 free(path_dup);
4305 }
4306
4307 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4308 struct diff_filespec *one)
4309 {
4310 struct diff_tempfile *temp = claim_diff_tempfile();
4311
4312 if (!DIFF_FILE_VALID(one)) {
4313 not_a_valid_file:
4314 /* A '-' entry produces this for file-2, and
4315 * a '+' entry produces this for file-1.
4316 */
4317 temp->name = "/dev/null";
4318 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4319 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4320 return temp;
4321 }
4322
4323 if (!S_ISGITLINK(one->mode) &&
4324 (!one->oid_valid ||
4325 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4326 struct stat st;
4327 if (lstat(one->path, &st) < 0) {
4328 if (errno == ENOENT)
4329 goto not_a_valid_file;
4330 die_errno("stat(%s)", one->path);
4331 }
4332 if (S_ISLNK(st.st_mode)) {
4333 struct strbuf sb = STRBUF_INIT;
4334 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4335 die_errno("readlink(%s)", one->path);
4336 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4337 (one->oid_valid ?
4338 &one->oid : null_oid()),
4339 (one->oid_valid ?
4340 one->mode : S_IFLNK));
4341 strbuf_release(&sb);
4342 }
4343 else {
4344 /* we can borrow from the file in the work tree */
4345 temp->name = one->path;
4346 if (!one->oid_valid)
4347 oid_to_hex_r(temp->hex, null_oid());
4348 else
4349 oid_to_hex_r(temp->hex, &one->oid);
4350 /* Even though we may sometimes borrow the
4351 * contents from the work tree, we always want
4352 * one->mode. mode is trustworthy even when
4353 * !(one->oid_valid), as long as
4354 * DIFF_FILE_VALID(one).
4355 */
4356 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4357 }
4358 return temp;
4359 }
4360 else {
4361 if (diff_populate_filespec(r, one, NULL))
4362 die("cannot read data blob for %s", one->path);
4363 prep_temp_blob(r->index, one->path, temp,
4364 one->data, one->size,
4365 &one->oid, one->mode);
4366 }
4367 return temp;
4368 }
4369
4370 static void add_external_diff_name(struct repository *r,
4371 struct strvec *argv,
4372 struct diff_filespec *df)
4373 {
4374 struct diff_tempfile *temp = prepare_temp_file(r, df);
4375 strvec_push(argv, temp->name);
4376 strvec_push(argv, temp->hex);
4377 strvec_push(argv, temp->mode);
4378 }
4379
4380 /* An external diff command takes:
4381 *
4382 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4383 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4384 *
4385 */
4386 static void run_external_diff(const struct external_diff *pgm,
4387 const char *name,
4388 const char *other,
4389 struct diff_filespec *one,
4390 struct diff_filespec *two,
4391 const char *xfrm_msg,
4392 struct diff_options *o)
4393 {
4394 struct child_process cmd = CHILD_PROCESS_INIT;
4395 struct diff_queue_struct *q = &diff_queued_diff;
4396 int quiet = !(o->output_format & DIFF_FORMAT_PATCH);
4397 int rc;
4398
4399 /*
4400 * Trivial equality is handled by diff_unmodified_pair() before
4401 * we get here. If we don't need to show the diff and the
4402 * external diff program lacks the ability to tell us whether
4403 * it's empty then we consider it non-empty without even asking.
4404 */
4405 if (!pgm->trust_exit_code && quiet) {
4406 o->found_changes = 1;
4407 return;
4408 }
4409
4410 strvec_push(&cmd.args, pgm->cmd);
4411 strvec_push(&cmd.args, name);
4412
4413 if (one && two) {
4414 add_external_diff_name(o->repo, &cmd.args, one);
4415 add_external_diff_name(o->repo, &cmd.args, two);
4416 if (other) {
4417 strvec_push(&cmd.args, other);
4418 if (xfrm_msg)
4419 strvec_push(&cmd.args, xfrm_msg);
4420 }
4421 }
4422
4423 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4424 ++o->diff_path_counter);
4425 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4426
4427 diff_free_filespec_data(one);
4428 diff_free_filespec_data(two);
4429 cmd.use_shell = 1;
4430 cmd.no_stdout = quiet;
4431 rc = run_command(&cmd);
4432 if (!pgm->trust_exit_code && rc == 0)
4433 o->found_changes = 1;
4434 else if (pgm->trust_exit_code && rc == 0)
4435 ; /* nothing */
4436 else if (pgm->trust_exit_code && rc == 1)
4437 o->found_changes = 1;
4438 else
4439 die(_("external diff died, stopping at %s"), name);
4440
4441 remove_tempfile();
4442 }
4443
4444 static int similarity_index(struct diff_filepair *p)
4445 {
4446 return p->score * 100 / MAX_SCORE;
4447 }
4448
4449 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4450 {
4451 if (startup_info->have_repository)
4452 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4453 else {
4454 char *hex = oid_to_hex(oid);
4455 if (abbrev < 0)
4456 abbrev = FALLBACK_DEFAULT_ABBREV;
4457 if (abbrev > the_hash_algo->hexsz)
4458 BUG("oid abbreviation out of range: %d", abbrev);
4459 if (abbrev)
4460 hex[abbrev] = '\0';
4461 return hex;
4462 }
4463 }
4464
4465 static void fill_metainfo(struct strbuf *msg,
4466 const char *name,
4467 const char *other,
4468 struct diff_filespec *one,
4469 struct diff_filespec *two,
4470 struct diff_options *o,
4471 struct diff_filepair *p,
4472 int *must_show_header,
4473 int use_color)
4474 {
4475 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4476 const char *reset = diff_get_color(use_color, DIFF_RESET);
4477 const char *line_prefix = diff_line_prefix(o);
4478 struct string_list *more_headers = NULL;
4479
4480 *must_show_header = 1;
4481 strbuf_init(msg, PATH_MAX * 2 + 300);
4482 switch (p->status) {
4483 case DIFF_STATUS_COPIED:
4484 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4485 line_prefix, set, similarity_index(p));
4486 strbuf_addf(msg, "%s\n%s%scopy from ",
4487 reset, line_prefix, set);
4488 quote_c_style(name, msg, NULL, 0);
4489 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4490 quote_c_style(other, msg, NULL, 0);
4491 strbuf_addf(msg, "%s\n", reset);
4492 break;
4493 case DIFF_STATUS_RENAMED:
4494 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4495 line_prefix, set, similarity_index(p));
4496 strbuf_addf(msg, "%s\n%s%srename from ",
4497 reset, line_prefix, set);
4498 quote_c_style(name, msg, NULL, 0);
4499 strbuf_addf(msg, "%s\n%s%srename to ",
4500 reset, line_prefix, set);
4501 quote_c_style(other, msg, NULL, 0);
4502 strbuf_addf(msg, "%s\n", reset);
4503 break;
4504 case DIFF_STATUS_MODIFIED:
4505 if (p->score) {
4506 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4507 line_prefix,
4508 set, similarity_index(p), reset);
4509 break;
4510 }
4511 /* fallthru */
4512 default:
4513 *must_show_header = 0;
4514 }
4515 if ((more_headers = additional_headers(o, name))) {
4516 add_formatted_headers(msg, more_headers,
4517 line_prefix, set, reset);
4518 *must_show_header = 1;
4519 }
4520 if (one && two && !oideq(&one->oid, &two->oid)) {
4521 const unsigned hexsz = the_hash_algo->hexsz;
4522 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4523
4524 if (o->flags.full_index)
4525 abbrev = hexsz;
4526
4527 if (o->flags.binary) {
4528 mmfile_t mf;
4529 if ((!fill_mmfile(o->repo, &mf, one) &&
4530 diff_filespec_is_binary(o->repo, one)) ||
4531 (!fill_mmfile(o->repo, &mf, two) &&
4532 diff_filespec_is_binary(o->repo, two)))
4533 abbrev = hexsz;
4534 }
4535 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4536 diff_abbrev_oid(&one->oid, abbrev),
4537 diff_abbrev_oid(&two->oid, abbrev));
4538 if (one->mode == two->mode)
4539 strbuf_addf(msg, " %06o", one->mode);
4540 strbuf_addf(msg, "%s\n", reset);
4541 }
4542 }
4543
4544 static void run_diff_cmd(const struct external_diff *pgm,
4545 const char *name,
4546 const char *other,
4547 const char *attr_path,
4548 struct diff_filespec *one,
4549 struct diff_filespec *two,
4550 struct strbuf *msg,
4551 struct diff_options *o,
4552 struct diff_filepair *p)
4553 {
4554 const char *xfrm_msg = NULL;
4555 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4556 int must_show_header = 0;
4557 struct userdiff_driver *drv = NULL;
4558
4559 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4560 drv = userdiff_find_by_path(o->repo->index, attr_path);
4561
4562 if (o->flags.allow_external && drv && drv->external.cmd)
4563 pgm = &drv->external;
4564
4565 if (msg) {
4566 /*
4567 * don't use colors when the header is intended for an
4568 * external diff driver
4569 */
4570 fill_metainfo(msg, name, other, one, two, o, p,
4571 &must_show_header,
4572 want_color(o->use_color) && !pgm);
4573 xfrm_msg = msg->len ? msg->buf : NULL;
4574 }
4575
4576 if (pgm) {
4577 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4578 return;
4579 }
4580 if (one && two) {
4581 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4582 set_diff_algorithm(o, drv->algorithm);
4583
4584 builtin_diff(name, other ? other : name,
4585 one, two, xfrm_msg, must_show_header,
4586 o, complete_rewrite);
4587 } else {
4588 fprintf(o->file, "* Unmerged path %s\n", name);
4589 o->found_changes = 1;
4590 }
4591 }
4592
4593 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4594 {
4595 if (DIFF_FILE_VALID(one)) {
4596 if (!one->oid_valid) {
4597 struct stat st;
4598 if (one->is_stdin) {
4599 oidclr(&one->oid);
4600 return;
4601 }
4602 if (lstat(one->path, &st) < 0)
4603 die_errno("stat '%s'", one->path);
4604 if (index_path(istate, &one->oid, one->path, &st, 0))
4605 die("cannot hash %s", one->path);
4606 }
4607 }
4608 else
4609 oidclr(&one->oid);
4610 }
4611
4612 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4613 {
4614 /* Strip the prefix but do not molest /dev/null and absolute paths */
4615 if (*namep && !is_absolute_path(*namep)) {
4616 *namep += prefix_length;
4617 if (**namep == '/')
4618 ++*namep;
4619 }
4620 if (*otherp && !is_absolute_path(*otherp)) {
4621 *otherp += prefix_length;
4622 if (**otherp == '/')
4623 ++*otherp;
4624 }
4625 }
4626
4627 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4628 {
4629 const struct external_diff *pgm = external_diff();
4630 struct strbuf msg;
4631 struct diff_filespec *one = p->one;
4632 struct diff_filespec *two = p->two;
4633 const char *name;
4634 const char *other;
4635 const char *attr_path;
4636
4637 name = one->path;
4638 other = (strcmp(name, two->path) ? two->path : NULL);
4639 attr_path = name;
4640 if (o->prefix_length)
4641 strip_prefix(o->prefix_length, &name, &other);
4642
4643 if (!o->flags.allow_external)
4644 pgm = NULL;
4645
4646 if (DIFF_PAIR_UNMERGED(p)) {
4647 run_diff_cmd(pgm, name, NULL, attr_path,
4648 NULL, NULL, NULL, o, p);
4649 return;
4650 }
4651
4652 diff_fill_oid_info(one, o->repo->index);
4653 diff_fill_oid_info(two, o->repo->index);
4654
4655 if (!pgm &&
4656 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4657 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4658 /*
4659 * a filepair that changes between file and symlink
4660 * needs to be split into deletion and creation.
4661 */
4662 struct diff_filespec *null = alloc_filespec(two->path);
4663 run_diff_cmd(NULL, name, other, attr_path,
4664 one, null, &msg,
4665 o, p);
4666 free(null);
4667 strbuf_release(&msg);
4668
4669 null = alloc_filespec(one->path);
4670 run_diff_cmd(NULL, name, other, attr_path,
4671 null, two, &msg, o, p);
4672 free(null);
4673 }
4674 else
4675 run_diff_cmd(pgm, name, other, attr_path,
4676 one, two, &msg, o, p);
4677
4678 strbuf_release(&msg);
4679 }
4680
4681 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4682 struct diffstat_t *diffstat)
4683 {
4684 const char *name;
4685 const char *other;
4686
4687 if (!o->ignore_driver_algorithm) {
4688 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4689 p->one->path);
4690
4691 if (drv && drv->algorithm)
4692 set_diff_algorithm(o, drv->algorithm);
4693 }
4694
4695 if (DIFF_PAIR_UNMERGED(p)) {
4696 /* unmerged */
4697 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4698 diffstat, o, p);
4699 return;
4700 }
4701
4702 name = p->one->path;
4703 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4704
4705 if (o->prefix_length)
4706 strip_prefix(o->prefix_length, &name, &other);
4707
4708 diff_fill_oid_info(p->one, o->repo->index);
4709 diff_fill_oid_info(p->two, o->repo->index);
4710
4711 builtin_diffstat(name, other, p->one, p->two,
4712 diffstat, o, p);
4713 }
4714
4715 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4716 {
4717 const char *name;
4718 const char *other;
4719 const char *attr_path;
4720
4721 if (DIFF_PAIR_UNMERGED(p)) {
4722 /* unmerged */
4723 return;
4724 }
4725
4726 name = p->one->path;
4727 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4728 attr_path = other ? other : name;
4729
4730 if (o->prefix_length)
4731 strip_prefix(o->prefix_length, &name, &other);
4732
4733 diff_fill_oid_info(p->one, o->repo->index);
4734 diff_fill_oid_info(p->two, o->repo->index);
4735
4736 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4737 }
4738
4739 void repo_diff_setup(struct repository *r, struct diff_options *options)
4740 {
4741 memcpy(options, &default_diff_options, sizeof(*options));
4742
4743 options->file = stdout;
4744 options->repo = r;
4745
4746 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4747 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4748 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4749 options->abbrev = DEFAULT_ABBREV;
4750 options->line_termination = '\n';
4751 options->break_opt = -1;
4752 options->rename_limit = -1;
4753 options->dirstat_permille = diff_dirstat_permille_default;
4754 options->context = diff_context_default;
4755 options->interhunkcontext = diff_interhunk_context_default;
4756 options->ws_error_highlight = ws_error_highlight_default;
4757 options->flags.rename_empty = 1;
4758 options->flags.relative_name = diff_relative;
4759 options->objfind = NULL;
4760
4761 /* pathchange left =NULL by default */
4762 options->change = diff_change;
4763 options->add_remove = diff_addremove;
4764 options->use_color = diff_use_color_default;
4765 options->detect_rename = diff_detect_rename_default;
4766 options->xdl_opts |= diff_algorithm;
4767 if (diff_indent_heuristic)
4768 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4769
4770 options->orderfile = diff_order_file_cfg;
4771
4772 if (!options->flags.ignore_submodule_set)
4773 options->flags.ignore_untracked_in_submodules = 1;
4774
4775 if (diff_no_prefix) {
4776 diff_set_noprefix(options);
4777 } else if (!diff_mnemonic_prefix) {
4778 diff_set_default_prefix(options);
4779 }
4780
4781 options->color_moved = diff_color_moved_default;
4782 options->color_moved_ws_handling = diff_color_moved_ws_default;
4783 }
4784
4785 static const char diff_status_letters[] = {
4786 DIFF_STATUS_ADDED,
4787 DIFF_STATUS_COPIED,
4788 DIFF_STATUS_DELETED,
4789 DIFF_STATUS_MODIFIED,
4790 DIFF_STATUS_RENAMED,
4791 DIFF_STATUS_TYPE_CHANGED,
4792 DIFF_STATUS_UNKNOWN,
4793 DIFF_STATUS_UNMERGED,
4794 DIFF_STATUS_FILTER_AON,
4795 DIFF_STATUS_FILTER_BROKEN,
4796 '\0',
4797 };
4798
4799 static unsigned int filter_bit['Z' + 1];
4800
4801 static void prepare_filter_bits(void)
4802 {
4803 int i;
4804
4805 if (!filter_bit[DIFF_STATUS_ADDED]) {
4806 for (i = 0; diff_status_letters[i]; i++)
4807 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4808 }
4809 }
4810
4811 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4812 {
4813 return opt->filter & filter_bit[(int) status];
4814 }
4815
4816 unsigned diff_filter_bit(char status)
4817 {
4818 prepare_filter_bits();
4819 return filter_bit[(int) status];
4820 }
4821
4822 int diff_check_follow_pathspec(struct pathspec *ps, int die_on_error)
4823 {
4824 unsigned forbidden_magic;
4825
4826 if (ps->nr != 1) {
4827 if (die_on_error)
4828 die(_("--follow requires exactly one pathspec"));
4829 return 0;
4830 }
4831
4832 forbidden_magic = ps->items[0].magic & ~(PATHSPEC_FROMTOP |
4833 PATHSPEC_LITERAL);
4834 if (forbidden_magic) {
4835 if (die_on_error) {
4836 struct strbuf sb = STRBUF_INIT;
4837 pathspec_magic_names(forbidden_magic, &sb);
4838 die(_("pathspec magic not supported by --follow: %s"),
4839 sb.buf);
4840 }
4841 return 0;
4842 }
4843
4844 return 1;
4845 }
4846
4847 void diff_setup_done(struct diff_options *options)
4848 {
4849 unsigned check_mask = DIFF_FORMAT_NAME |
4850 DIFF_FORMAT_NAME_STATUS |
4851 DIFF_FORMAT_CHECKDIFF |
4852 DIFF_FORMAT_NO_OUTPUT;
4853 /*
4854 * This must be signed because we're comparing against a potentially
4855 * negative value.
4856 */
4857 const int hexsz = the_hash_algo->hexsz;
4858
4859 if (options->set_default)
4860 options->set_default(options);
4861
4862 if (HAS_MULTI_BITS(options->output_format & check_mask))
4863 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4864 "--name-only", "--name-status", "--check", "-s");
4865
4866 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4867 die(_("options '%s', '%s', and '%s' cannot be used together"),
4868 "-G", "-S", "--find-object");
4869
4870 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4871 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4872 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4873
4874 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4875 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4876 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4877
4878 /*
4879 * Most of the time we can say "there are changes"
4880 * only by checking if there are changed paths, but
4881 * --ignore-whitespace* options force us to look
4882 * inside contents.
4883 */
4884
4885 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4886 options->ignore_regex_nr)
4887 options->flags.diff_from_contents = 1;
4888 else
4889 options->flags.diff_from_contents = 0;
4890
4891 if (options->flags.find_copies_harder)
4892 options->detect_rename = DIFF_DETECT_COPY;
4893
4894 if (!options->flags.relative_name)
4895 options->prefix = NULL;
4896 if (options->prefix)
4897 options->prefix_length = strlen(options->prefix);
4898 else
4899 options->prefix_length = 0;
4900
4901 /*
4902 * --name-only, --name-status, --checkdiff, and -s
4903 * turn other output format off.
4904 */
4905 if (options->output_format & (DIFF_FORMAT_NAME |
4906 DIFF_FORMAT_NAME_STATUS |
4907 DIFF_FORMAT_CHECKDIFF |
4908 DIFF_FORMAT_NO_OUTPUT))
4909 options->output_format &= ~(DIFF_FORMAT_RAW |
4910 DIFF_FORMAT_NUMSTAT |
4911 DIFF_FORMAT_DIFFSTAT |
4912 DIFF_FORMAT_SHORTSTAT |
4913 DIFF_FORMAT_DIRSTAT |
4914 DIFF_FORMAT_SUMMARY |
4915 DIFF_FORMAT_PATCH);
4916
4917 /*
4918 * These cases always need recursive; we do not drop caller-supplied
4919 * recursive bits for other formats here.
4920 */
4921 if (options->output_format & (DIFF_FORMAT_PATCH |
4922 DIFF_FORMAT_NUMSTAT |
4923 DIFF_FORMAT_DIFFSTAT |
4924 DIFF_FORMAT_SHORTSTAT |
4925 DIFF_FORMAT_DIRSTAT |
4926 DIFF_FORMAT_SUMMARY |
4927 DIFF_FORMAT_CHECKDIFF))
4928 options->flags.recursive = 1;
4929 /*
4930 * Also pickaxe would not work very well if you do not say recursive
4931 */
4932 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4933 options->flags.recursive = 1;
4934 /*
4935 * When patches are generated, submodules diffed against the work tree
4936 * must be checked for dirtiness too so it can be shown in the output
4937 */
4938 if (options->output_format & DIFF_FORMAT_PATCH)
4939 options->flags.dirty_submodules = 1;
4940
4941 if (options->detect_rename && options->rename_limit < 0)
4942 options->rename_limit = diff_rename_limit_default;
4943 if (hexsz < options->abbrev)
4944 options->abbrev = hexsz; /* full */
4945
4946 /*
4947 * It does not make sense to show the first hit we happened
4948 * to have found. It does not make sense not to return with
4949 * exit code in such a case either.
4950 */
4951 if (options->flags.quick) {
4952 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4953 options->flags.exit_with_status = 1;
4954 }
4955
4956 /*
4957 * External diffs could declare non-identical contents equal
4958 * (think diff --ignore-space-change).
4959 */
4960 if (options->flags.allow_external && options->flags.exit_with_status)
4961 options->flags.diff_from_contents = 1;
4962
4963 options->diff_path_counter = 0;
4964
4965 if (options->flags.follow_renames)
4966 diff_check_follow_pathspec(&options->pathspec, 1);
4967
4968 if (!options->use_color || external_diff())
4969 options->color_moved = 0;
4970
4971 if (options->filter_not) {
4972 if (!options->filter)
4973 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4974 options->filter &= ~options->filter_not;
4975 }
4976 }
4977
4978 int parse_long_opt(const char *opt, const char **argv,
4979 const char **optarg)
4980 {
4981 const char *arg = argv[0];
4982 if (!skip_prefix(arg, "--", &arg))
4983 return 0;
4984 if (!skip_prefix(arg, opt, &arg))
4985 return 0;
4986 if (*arg == '=') { /* stuck form: --option=value */
4987 *optarg = arg + 1;
4988 return 1;
4989 }
4990 if (*arg != '\0')
4991 return 0;
4992 /* separate form: --option value */
4993 if (!argv[1])
4994 die("Option '--%s' requires a value", opt);
4995 *optarg = argv[1];
4996 return 2;
4997 }
4998
4999 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
5000 {
5001 struct diff_options *options = opt->value;
5002 int width = options->stat_width;
5003 int name_width = options->stat_name_width;
5004 int graph_width = options->stat_graph_width;
5005 int count = options->stat_count;
5006 char *end;
5007
5008 BUG_ON_OPT_NEG(unset);
5009
5010 if (!strcmp(opt->long_name, "stat")) {
5011 if (value) {
5012 width = strtoul(value, &end, 10);
5013 if (*end == ',')
5014 name_width = strtoul(end+1, &end, 10);
5015 if (*end == ',')
5016 count = strtoul(end+1, &end, 10);
5017 if (*end)
5018 return error(_("invalid --stat value: %s"), value);
5019 }
5020 } else if (!strcmp(opt->long_name, "stat-width")) {
5021 width = strtoul(value, &end, 10);
5022 if (*end)
5023 return error(_("%s expects a numerical value"),
5024 opt->long_name);
5025 } else if (!strcmp(opt->long_name, "stat-name-width")) {
5026 name_width = strtoul(value, &end, 10);
5027 if (*end)
5028 return error(_("%s expects a numerical value"),
5029 opt->long_name);
5030 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
5031 graph_width = strtoul(value, &end, 10);
5032 if (*end)
5033 return error(_("%s expects a numerical value"),
5034 opt->long_name);
5035 } else if (!strcmp(opt->long_name, "stat-count")) {
5036 count = strtoul(value, &end, 10);
5037 if (*end)
5038 return error(_("%s expects a numerical value"),
5039 opt->long_name);
5040 } else
5041 BUG("%s should not get here", opt->long_name);
5042
5043 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5044 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5045 options->stat_name_width = name_width;
5046 options->stat_graph_width = graph_width;
5047 options->stat_width = width;
5048 options->stat_count = count;
5049 return 0;
5050 }
5051
5052 static int parse_dirstat_opt(struct diff_options *options, const char *params)
5053 {
5054 struct strbuf errmsg = STRBUF_INIT;
5055 if (parse_dirstat_params(options, params, &errmsg))
5056 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
5057 errmsg.buf);
5058 strbuf_release(&errmsg);
5059 /*
5060 * The caller knows a dirstat-related option is given from the command
5061 * line; allow it to say "return this_function();"
5062 */
5063 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5064 options->output_format |= DIFF_FORMAT_DIRSTAT;
5065 return 1;
5066 }
5067
5068 static int diff_opt_diff_filter(const struct option *option,
5069 const char *optarg, int unset)
5070 {
5071 struct diff_options *opt = option->value;
5072 int i, optch;
5073
5074 BUG_ON_OPT_NEG(unset);
5075 prepare_filter_bits();
5076
5077 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
5078 unsigned int bit;
5079 int negate;
5080
5081 if ('a' <= optch && optch <= 'z') {
5082 negate = 1;
5083 optch = toupper(optch);
5084 } else {
5085 negate = 0;
5086 }
5087
5088 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
5089 if (!bit)
5090 return error(_("unknown change class '%c' in --diff-filter=%s"),
5091 optarg[i], optarg);
5092 if (negate)
5093 opt->filter_not |= bit;
5094 else
5095 opt->filter |= bit;
5096 }
5097 return 0;
5098 }
5099
5100 static void enable_patch_output(int *fmt)
5101 {
5102 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
5103 *fmt |= DIFF_FORMAT_PATCH;
5104 }
5105
5106 static int diff_opt_ws_error_highlight(const struct option *option,
5107 const char *arg, int unset)
5108 {
5109 struct diff_options *opt = option->value;
5110 int val = parse_ws_error_highlight(arg);
5111
5112 BUG_ON_OPT_NEG(unset);
5113 if (val < 0)
5114 return error(_("unknown value after ws-error-highlight=%.*s"),
5115 -1 - val, arg);
5116 opt->ws_error_highlight = val;
5117 return 0;
5118 }
5119
5120 static int diff_opt_find_object(const struct option *option,
5121 const char *arg, int unset)
5122 {
5123 struct diff_options *opt = option->value;
5124 struct object_id oid;
5125
5126 BUG_ON_OPT_NEG(unset);
5127 if (repo_get_oid(the_repository, arg, &oid))
5128 return error(_("unable to resolve '%s'"), arg);
5129
5130 if (!opt->objfind)
5131 CALLOC_ARRAY(opt->objfind, 1);
5132
5133 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5134 opt->flags.recursive = 1;
5135 opt->flags.tree_in_recursive = 1;
5136 oidset_insert(opt->objfind, &oid);
5137 return 0;
5138 }
5139
5140 static int diff_opt_anchored(const struct option *opt,
5141 const char *arg, int unset)
5142 {
5143 struct diff_options *options = opt->value;
5144
5145 BUG_ON_OPT_NEG(unset);
5146 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5147 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5148 options->anchors_alloc);
5149 options->anchors[options->anchors_nr++] = xstrdup(arg);
5150 return 0;
5151 }
5152
5153 static int diff_opt_binary(const struct option *opt,
5154 const char *arg, int unset)
5155 {
5156 struct diff_options *options = opt->value;
5157
5158 BUG_ON_OPT_NEG(unset);
5159 BUG_ON_OPT_ARG(arg);
5160 enable_patch_output(&options->output_format);
5161 options->flags.binary = 1;
5162 return 0;
5163 }
5164
5165 static int diff_opt_break_rewrites(const struct option *opt,
5166 const char *arg, int unset)
5167 {
5168 int *break_opt = opt->value;
5169 int opt1, opt2;
5170
5171 BUG_ON_OPT_NEG(unset);
5172 if (!arg)
5173 arg = "";
5174 opt1 = parse_rename_score(&arg);
5175 if (*arg == 0)
5176 opt2 = 0;
5177 else if (*arg != '/')
5178 return error(_("%s expects <n>/<m> form"), opt->long_name);
5179 else {
5180 arg++;
5181 opt2 = parse_rename_score(&arg);
5182 }
5183 if (*arg != 0)
5184 return error(_("%s expects <n>/<m> form"), opt->long_name);
5185 *break_opt = opt1 | (opt2 << 16);
5186 return 0;
5187 }
5188
5189 static int diff_opt_char(const struct option *opt,
5190 const char *arg, int unset)
5191 {
5192 char *value = opt->value;
5193
5194 BUG_ON_OPT_NEG(unset);
5195 if (arg[1])
5196 return error(_("%s expects a character, got '%s'"),
5197 opt->long_name, arg);
5198 *value = arg[0];
5199 return 0;
5200 }
5201
5202 static int diff_opt_color_moved(const struct option *opt,
5203 const char *arg, int unset)
5204 {
5205 struct diff_options *options = opt->value;
5206
5207 if (unset) {
5208 options->color_moved = COLOR_MOVED_NO;
5209 } else if (!arg) {
5210 if (diff_color_moved_default)
5211 options->color_moved = diff_color_moved_default;
5212 if (options->color_moved == COLOR_MOVED_NO)
5213 options->color_moved = COLOR_MOVED_DEFAULT;
5214 } else {
5215 int cm = parse_color_moved(arg);
5216 if (cm < 0)
5217 return error(_("bad --color-moved argument: %s"), arg);
5218 options->color_moved = cm;
5219 }
5220 return 0;
5221 }
5222
5223 static int diff_opt_color_moved_ws(const struct option *opt,
5224 const char *arg, int unset)
5225 {
5226 struct diff_options *options = opt->value;
5227 unsigned cm;
5228
5229 if (unset) {
5230 options->color_moved_ws_handling = 0;
5231 return 0;
5232 }
5233
5234 cm = parse_color_moved_ws(arg);
5235 if (cm & COLOR_MOVED_WS_ERROR)
5236 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5237 options->color_moved_ws_handling = cm;
5238 return 0;
5239 }
5240
5241 static int diff_opt_color_words(const struct option *opt,
5242 const char *arg, int unset)
5243 {
5244 struct diff_options *options = opt->value;
5245
5246 BUG_ON_OPT_NEG(unset);
5247 options->use_color = 1;
5248 options->word_diff = DIFF_WORDS_COLOR;
5249 options->word_regex = arg;
5250 return 0;
5251 }
5252
5253 static int diff_opt_compact_summary(const struct option *opt,
5254 const char *arg, int unset)
5255 {
5256 struct diff_options *options = opt->value;
5257
5258 BUG_ON_OPT_ARG(arg);
5259 if (unset) {
5260 options->flags.stat_with_summary = 0;
5261 } else {
5262 options->flags.stat_with_summary = 1;
5263 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5264 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5265 }
5266 return 0;
5267 }
5268
5269 static int diff_opt_diff_algorithm(const struct option *opt,
5270 const char *arg, int unset)
5271 {
5272 struct diff_options *options = opt->value;
5273
5274 BUG_ON_OPT_NEG(unset);
5275
5276 if (set_diff_algorithm(options, arg))
5277 return error(_("option diff-algorithm accepts \"myers\", "
5278 "\"minimal\", \"patience\" and \"histogram\""));
5279
5280 options->ignore_driver_algorithm = 1;
5281
5282 return 0;
5283 }
5284
5285 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5286 const char *arg, int unset)
5287 {
5288 struct diff_options *options = opt->value;
5289
5290 BUG_ON_OPT_NEG(unset);
5291 BUG_ON_OPT_ARG(arg);
5292
5293 if (set_diff_algorithm(options, opt->long_name))
5294 BUG("available diff algorithms include \"myers\", "
5295 "\"minimal\", \"patience\" and \"histogram\"");
5296
5297 options->ignore_driver_algorithm = 1;
5298
5299 return 0;
5300 }
5301
5302 static int diff_opt_dirstat(const struct option *opt,
5303 const char *arg, int unset)
5304 {
5305 struct diff_options *options = opt->value;
5306
5307 BUG_ON_OPT_NEG(unset);
5308 if (!strcmp(opt->long_name, "cumulative")) {
5309 if (arg)
5310 BUG("how come --cumulative take a value?");
5311 arg = "cumulative";
5312 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5313 parse_dirstat_opt(options, "files");
5314 parse_dirstat_opt(options, arg ? arg : "");
5315 return 0;
5316 }
5317
5318 static int diff_opt_find_copies(const struct option *opt,
5319 const char *arg, int unset)
5320 {
5321 struct diff_options *options = opt->value;
5322
5323 BUG_ON_OPT_NEG(unset);
5324 if (!arg)
5325 arg = "";
5326 options->rename_score = parse_rename_score(&arg);
5327 if (*arg != 0)
5328 return error(_("invalid argument to %s"), opt->long_name);
5329
5330 if (options->detect_rename == DIFF_DETECT_COPY)
5331 options->flags.find_copies_harder = 1;
5332 else
5333 options->detect_rename = DIFF_DETECT_COPY;
5334
5335 return 0;
5336 }
5337
5338 static int diff_opt_find_renames(const struct option *opt,
5339 const char *arg, int unset)
5340 {
5341 struct diff_options *options = opt->value;
5342
5343 BUG_ON_OPT_NEG(unset);
5344 if (!arg)
5345 arg = "";
5346 options->rename_score = parse_rename_score(&arg);
5347 if (*arg != 0)
5348 return error(_("invalid argument to %s"), opt->long_name);
5349
5350 options->detect_rename = DIFF_DETECT_RENAME;
5351 return 0;
5352 }
5353
5354 static int diff_opt_follow(const struct option *opt,
5355 const char *arg, int unset)
5356 {
5357 struct diff_options *options = opt->value;
5358
5359 BUG_ON_OPT_ARG(arg);
5360 if (unset) {
5361 options->flags.follow_renames = 0;
5362 options->flags.default_follow_renames = 0;
5363 } else {
5364 options->flags.follow_renames = 1;
5365 }
5366 return 0;
5367 }
5368
5369 static int diff_opt_ignore_submodules(const struct option *opt,
5370 const char *arg, int unset)
5371 {
5372 struct diff_options *options = opt->value;
5373
5374 BUG_ON_OPT_NEG(unset);
5375 if (!arg)
5376 arg = "all";
5377 options->flags.override_submodule_config = 1;
5378 handle_ignore_submodules_arg(options, arg);
5379 return 0;
5380 }
5381
5382 static int diff_opt_line_prefix(const struct option *opt,
5383 const char *optarg, int unset)
5384 {
5385 struct diff_options *options = opt->value;
5386
5387 BUG_ON_OPT_NEG(unset);
5388 options->line_prefix = optarg;
5389 options->line_prefix_length = strlen(options->line_prefix);
5390 graph_setup_line_prefix(options);
5391 return 0;
5392 }
5393
5394 static int diff_opt_no_prefix(const struct option *opt,
5395 const char *optarg, int unset)
5396 {
5397 struct diff_options *options = opt->value;
5398
5399 BUG_ON_OPT_NEG(unset);
5400 BUG_ON_OPT_ARG(optarg);
5401 diff_set_noprefix(options);
5402 return 0;
5403 }
5404
5405 static int diff_opt_default_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 FREE_AND_NULL(diff_src_prefix);
5413 FREE_AND_NULL(diff_dst_prefix);
5414 diff_set_default_prefix(options);
5415 return 0;
5416 }
5417
5418 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5419 const struct option *opt,
5420 const char *arg, int unset)
5421 {
5422 struct diff_options *options = opt->value;
5423 char *path;
5424
5425 BUG_ON_OPT_NEG(unset);
5426 path = prefix_filename(ctx->prefix, arg);
5427 options->file = xfopen(path, "w");
5428 options->close_file = 1;
5429 if (options->use_color != GIT_COLOR_ALWAYS)
5430 options->use_color = GIT_COLOR_NEVER;
5431 free(path);
5432 return 0;
5433 }
5434
5435 static int diff_opt_patience(const struct option *opt,
5436 const char *arg, int unset)
5437 {
5438 struct diff_options *options = opt->value;
5439 int i;
5440
5441 BUG_ON_OPT_NEG(unset);
5442 BUG_ON_OPT_ARG(arg);
5443 /*
5444 * Both --patience and --anchored use PATIENCE_DIFF
5445 * internally, so remove any anchors previously
5446 * specified.
5447 */
5448 for (i = 0; i < options->anchors_nr; i++)
5449 free(options->anchors[i]);
5450 options->anchors_nr = 0;
5451 options->ignore_driver_algorithm = 1;
5452
5453 return set_diff_algorithm(options, "patience");
5454 }
5455
5456 static int diff_opt_ignore_regex(const struct option *opt,
5457 const char *arg, int unset)
5458 {
5459 struct diff_options *options = opt->value;
5460 regex_t *regex;
5461
5462 BUG_ON_OPT_NEG(unset);
5463 regex = xmalloc(sizeof(*regex));
5464 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5465 return error(_("invalid regex given to -I: '%s'"), arg);
5466 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5467 options->ignore_regex_alloc);
5468 options->ignore_regex[options->ignore_regex_nr++] = regex;
5469 return 0;
5470 }
5471
5472 static int diff_opt_pickaxe_regex(const struct option *opt,
5473 const char *arg, int unset)
5474 {
5475 struct diff_options *options = opt->value;
5476
5477 BUG_ON_OPT_NEG(unset);
5478 options->pickaxe = arg;
5479 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5480 return 0;
5481 }
5482
5483 static int diff_opt_pickaxe_string(const struct option *opt,
5484 const char *arg, int unset)
5485 {
5486 struct diff_options *options = opt->value;
5487
5488 BUG_ON_OPT_NEG(unset);
5489 options->pickaxe = arg;
5490 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5491 return 0;
5492 }
5493
5494 static int diff_opt_relative(const struct option *opt,
5495 const char *arg, int unset)
5496 {
5497 struct diff_options *options = opt->value;
5498
5499 options->flags.relative_name = !unset;
5500 if (arg)
5501 options->prefix = arg;
5502 return 0;
5503 }
5504
5505 static int diff_opt_submodule(const struct option *opt,
5506 const char *arg, int unset)
5507 {
5508 struct diff_options *options = opt->value;
5509
5510 BUG_ON_OPT_NEG(unset);
5511 if (!arg)
5512 arg = "log";
5513 if (parse_submodule_params(options, arg))
5514 return error(_("failed to parse --submodule option parameter: '%s'"),
5515 arg);
5516 return 0;
5517 }
5518
5519 static int diff_opt_textconv(const struct option *opt,
5520 const char *arg, int unset)
5521 {
5522 struct diff_options *options = opt->value;
5523
5524 BUG_ON_OPT_ARG(arg);
5525 if (unset) {
5526 options->flags.allow_textconv = 0;
5527 } else {
5528 options->flags.allow_textconv = 1;
5529 options->flags.textconv_set_via_cmdline = 1;
5530 }
5531 return 0;
5532 }
5533
5534 static int diff_opt_unified(const struct option *opt,
5535 const char *arg, int unset)
5536 {
5537 struct diff_options *options = opt->value;
5538 char *s;
5539
5540 BUG_ON_OPT_NEG(unset);
5541
5542 if (arg) {
5543 options->context = strtol(arg, &s, 10);
5544 if (*s)
5545 return error(_("%s expects a numerical value"), "--unified");
5546 }
5547 enable_patch_output(&options->output_format);
5548
5549 return 0;
5550 }
5551
5552 static int diff_opt_word_diff(const struct option *opt,
5553 const char *arg, int unset)
5554 {
5555 struct diff_options *options = opt->value;
5556
5557 BUG_ON_OPT_NEG(unset);
5558 if (arg) {
5559 if (!strcmp(arg, "plain"))
5560 options->word_diff = DIFF_WORDS_PLAIN;
5561 else if (!strcmp(arg, "color")) {
5562 options->use_color = 1;
5563 options->word_diff = DIFF_WORDS_COLOR;
5564 }
5565 else if (!strcmp(arg, "porcelain"))
5566 options->word_diff = DIFF_WORDS_PORCELAIN;
5567 else if (!strcmp(arg, "none"))
5568 options->word_diff = DIFF_WORDS_NONE;
5569 else
5570 return error(_("bad --word-diff argument: %s"), arg);
5571 } else {
5572 if (options->word_diff == DIFF_WORDS_NONE)
5573 options->word_diff = DIFF_WORDS_PLAIN;
5574 }
5575 return 0;
5576 }
5577
5578 static int diff_opt_word_diff_regex(const struct option *opt,
5579 const char *arg, int unset)
5580 {
5581 struct diff_options *options = opt->value;
5582
5583 BUG_ON_OPT_NEG(unset);
5584 if (options->word_diff == DIFF_WORDS_NONE)
5585 options->word_diff = DIFF_WORDS_PLAIN;
5586 options->word_regex = arg;
5587 return 0;
5588 }
5589
5590 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5591 {
5592 struct diff_options *options = opt->value;
5593
5594 BUG_ON_OPT_NEG(unset);
5595 if (!strcmp(opt->long_name, "skip-to"))
5596 options->skip_instead_of_rotate = 1;
5597 else
5598 options->skip_instead_of_rotate = 0;
5599 options->rotate_to = arg;
5600 return 0;
5601 }
5602
5603 /*
5604 * Consider adding new flags to __git_diff_common_options
5605 * in contrib/completion/git-completion.bash
5606 */
5607 struct option *add_diff_options(const struct option *opts,
5608 struct diff_options *options)
5609 {
5610 struct option parseopts[] = {
5611 OPT_GROUP(N_("Diff output format options")),
5612 OPT_BITOP('p', "patch", &options->output_format,
5613 N_("generate patch"),
5614 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5615 OPT_SET_INT('s', "no-patch", &options->output_format,
5616 N_("suppress diff output"), DIFF_FORMAT_NO_OUTPUT),
5617 OPT_BITOP('u', NULL, &options->output_format,
5618 N_("generate patch"),
5619 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5620 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5621 N_("generate diffs with <n> lines context"),
5622 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5623 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5624 N_("generate diffs with <n> lines context")),
5625 OPT_BITOP(0, "raw", &options->output_format,
5626 N_("generate the diff in raw format"),
5627 DIFF_FORMAT_RAW, DIFF_FORMAT_NO_OUTPUT),
5628 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5629 N_("synonym for '-p --raw'"),
5630 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5631 DIFF_FORMAT_NO_OUTPUT),
5632 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5633 N_("synonym for '-p --stat'"),
5634 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5635 DIFF_FORMAT_NO_OUTPUT),
5636 OPT_BITOP(0, "numstat", &options->output_format,
5637 N_("machine friendly --stat"),
5638 DIFF_FORMAT_NUMSTAT, DIFF_FORMAT_NO_OUTPUT),
5639 OPT_BITOP(0, "shortstat", &options->output_format,
5640 N_("output only the last line of --stat"),
5641 DIFF_FORMAT_SHORTSTAT, DIFF_FORMAT_NO_OUTPUT),
5642 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1>,<param2>..."),
5643 N_("output the distribution of relative amount of changes for each sub-directory"),
5644 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5645 diff_opt_dirstat),
5646 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5647 N_("synonym for --dirstat=cumulative"),
5648 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5649 diff_opt_dirstat),
5650 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1>,<param2>..."),
5651 N_("synonym for --dirstat=files,<param1>,<param2>..."),
5652 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5653 diff_opt_dirstat),
5654 OPT_BIT_F(0, "check", &options->output_format,
5655 N_("warn if changes introduce conflict markers or whitespace errors"),
5656 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5657 OPT_BITOP(0, "summary", &options->output_format,
5658 N_("condensed summary such as creations, renames and mode changes"),
5659 DIFF_FORMAT_SUMMARY, DIFF_FORMAT_NO_OUTPUT),
5660 OPT_BIT_F(0, "name-only", &options->output_format,
5661 N_("show only names of changed files"),
5662 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5663 OPT_BIT_F(0, "name-status", &options->output_format,
5664 N_("show only names and status of changed files"),
5665 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5666 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5667 N_("generate diffstat"),
5668 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5669 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5670 N_("generate diffstat with a given width"),
5671 PARSE_OPT_NONEG, diff_opt_stat),
5672 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5673 N_("generate diffstat with a given name width"),
5674 PARSE_OPT_NONEG, diff_opt_stat),
5675 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5676 N_("generate diffstat with a given graph width"),
5677 PARSE_OPT_NONEG, diff_opt_stat),
5678 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5679 N_("generate diffstat with limited lines"),
5680 PARSE_OPT_NONEG, diff_opt_stat),
5681 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5682 N_("generate compact summary in diffstat"),
5683 PARSE_OPT_NOARG, diff_opt_compact_summary),
5684 OPT_CALLBACK_F(0, "binary", options, NULL,
5685 N_("output a binary diff that can be applied"),
5686 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5687 OPT_BOOL(0, "full-index", &options->flags.full_index,
5688 N_("show full pre- and post-image object names on the \"index\" lines")),
5689 OPT_COLOR_FLAG(0, "color", &options->use_color,
5690 N_("show colored diff")),
5691 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5692 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5693 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5694 OPT_SET_INT('z', NULL, &options->line_termination,
5695 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5696 0),
5697 OPT__ABBREV(&options->abbrev),
5698 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5699 N_("show the given source prefix instead of \"a/\""),
5700 PARSE_OPT_NONEG),
5701 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5702 N_("show the given destination prefix instead of \"b/\""),
5703 PARSE_OPT_NONEG),
5704 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5705 N_("prepend an additional prefix to every line of output"),
5706 PARSE_OPT_NONEG, diff_opt_line_prefix),
5707 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5708 N_("do not show any source or destination prefix"),
5709 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5710 OPT_CALLBACK_F(0, "default-prefix", options, NULL,
5711 N_("use default prefixes a/ and b/"),
5712 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_default_prefix),
5713 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5714 N_("show context between diff hunks up to the specified number of lines"),
5715 PARSE_OPT_NONEG),
5716 OPT_CALLBACK_F(0, "output-indicator-new",
5717 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5718 N_("<char>"),
5719 N_("specify the character to indicate a new line instead of '+'"),
5720 PARSE_OPT_NONEG, diff_opt_char),
5721 OPT_CALLBACK_F(0, "output-indicator-old",
5722 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5723 N_("<char>"),
5724 N_("specify the character to indicate an old line instead of '-'"),
5725 PARSE_OPT_NONEG, diff_opt_char),
5726 OPT_CALLBACK_F(0, "output-indicator-context",
5727 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5728 N_("<char>"),
5729 N_("specify the character to indicate a context instead of ' '"),
5730 PARSE_OPT_NONEG, diff_opt_char),
5731
5732 OPT_GROUP(N_("Diff rename options")),
5733 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5734 N_("break complete rewrite changes into pairs of delete and create"),
5735 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5736 diff_opt_break_rewrites),
5737 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5738 N_("detect renames"),
5739 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5740 diff_opt_find_renames),
5741 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5742 N_("omit the preimage for deletes"),
5743 1, PARSE_OPT_NONEG),
5744 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5745 N_("detect copies"),
5746 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5747 diff_opt_find_copies),
5748 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5749 N_("use unmodified files as source to find copies")),
5750 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5751 N_("disable rename detection"),
5752 0, PARSE_OPT_NONEG),
5753 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5754 N_("use empty blobs as rename source")),
5755 OPT_CALLBACK_F(0, "follow", options, NULL,
5756 N_("continue listing the history of a file beyond renames"),
5757 PARSE_OPT_NOARG, diff_opt_follow),
5758 OPT_INTEGER('l', NULL, &options->rename_limit,
5759 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5760
5761 OPT_GROUP(N_("Diff algorithm options")),
5762 OPT_CALLBACK_F(0, "minimal", options, NULL,
5763 N_("produce the smallest possible diff"),
5764 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5765 diff_opt_diff_algorithm_no_arg),
5766 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5767 N_("ignore whitespace when comparing lines"),
5768 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5769 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5770 N_("ignore changes in amount of whitespace"),
5771 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5772 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5773 N_("ignore changes in whitespace at EOL"),
5774 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5775 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5776 N_("ignore carrier-return at the end of line"),
5777 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5778 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5779 N_("ignore changes whose lines are all blank"),
5780 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5781 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5782 N_("ignore changes whose all lines match <regex>"),
5783 0, diff_opt_ignore_regex),
5784 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5785 N_("heuristic to shift diff hunk boundaries for easy reading"),
5786 XDF_INDENT_HEURISTIC),
5787 OPT_CALLBACK_F(0, "patience", options, NULL,
5788 N_("generate diff using the \"patience diff\" algorithm"),
5789 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5790 diff_opt_patience),
5791 OPT_CALLBACK_F(0, "histogram", options, NULL,
5792 N_("generate diff using the \"histogram diff\" algorithm"),
5793 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5794 diff_opt_diff_algorithm_no_arg),
5795 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5796 N_("choose a diff algorithm"),
5797 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5798 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5799 N_("generate diff using the \"anchored diff\" algorithm"),
5800 PARSE_OPT_NONEG, diff_opt_anchored),
5801 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5802 N_("show word diff, using <mode> to delimit changed words"),
5803 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5804 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5805 N_("use <regex> to decide what a word is"),
5806 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5807 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5808 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5809 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5810 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5811 N_("moved lines of code are colored differently"),
5812 PARSE_OPT_OPTARG, diff_opt_color_moved),
5813 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5814 N_("how white spaces are ignored in --color-moved"),
5815 0, diff_opt_color_moved_ws),
5816
5817 OPT_GROUP(N_("Other diff options")),
5818 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5819 N_("when run from subdir, exclude changes outside and show relative paths"),
5820 PARSE_OPT_OPTARG,
5821 diff_opt_relative),
5822 OPT_BOOL('a', "text", &options->flags.text,
5823 N_("treat all files as text")),
5824 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5825 N_("swap two inputs, reverse the diff")),
5826 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5827 N_("exit with 1 if there were differences, 0 otherwise")),
5828 OPT_BOOL(0, "quiet", &options->flags.quick,
5829 N_("disable all output of the program")),
5830 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5831 N_("allow an external diff helper to be executed")),
5832 OPT_CALLBACK_F(0, "textconv", options, NULL,
5833 N_("run external text conversion filters when comparing binary files"),
5834 PARSE_OPT_NOARG, diff_opt_textconv),
5835 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5836 N_("ignore changes to submodules in the diff generation"),
5837 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5838 diff_opt_ignore_submodules),
5839 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5840 N_("specify how differences in submodules are shown"),
5841 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5842 diff_opt_submodule),
5843 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5844 N_("hide 'git add -N' entries from the index"),
5845 1, PARSE_OPT_NONEG),
5846 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5847 N_("treat 'git add -N' entries as real in the index"),
5848 0, PARSE_OPT_NONEG),
5849 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5850 N_("look for differences that change the number of occurrences of the specified string"),
5851 0, diff_opt_pickaxe_string),
5852 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5853 N_("look for differences that change the number of occurrences of the specified regex"),
5854 0, diff_opt_pickaxe_regex),
5855 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5856 N_("show all changes in the changeset with -S or -G"),
5857 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5858 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5859 N_("treat <string> in -S as extended POSIX regular expression"),
5860 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5861 OPT_FILENAME('O', NULL, &options->orderfile,
5862 N_("control the order in which files appear in the output")),
5863 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5864 N_("show the change in the specified path first"),
5865 PARSE_OPT_NONEG, diff_opt_rotate_to),
5866 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5867 N_("skip the output to the specified path"),
5868 PARSE_OPT_NONEG, diff_opt_rotate_to),
5869 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5870 N_("look for differences that change the number of occurrences of the specified object"),
5871 PARSE_OPT_NONEG, diff_opt_find_object),
5872 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5873 N_("select files by diff type"),
5874 PARSE_OPT_NONEG, diff_opt_diff_filter),
5875 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5876 N_("output to a specific file"),
5877 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5878
5879 OPT_END()
5880 };
5881
5882 return parse_options_concat(opts, parseopts);
5883 }
5884
5885 int diff_opt_parse(struct diff_options *options,
5886 const char **av, int ac, const char *prefix)
5887 {
5888 struct option no_options[] = { OPT_END() };
5889 struct option *parseopts = add_diff_options(no_options, options);
5890
5891 if (!prefix)
5892 prefix = "";
5893
5894 ac = parse_options(ac, av, prefix, parseopts, NULL,
5895 PARSE_OPT_KEEP_DASHDASH |
5896 PARSE_OPT_KEEP_UNKNOWN_OPT |
5897 PARSE_OPT_NO_INTERNAL_HELP |
5898 PARSE_OPT_ONE_SHOT |
5899 PARSE_OPT_STOP_AT_NON_OPTION);
5900 free(parseopts);
5901
5902 return ac;
5903 }
5904
5905 int parse_rename_score(const char **cp_p)
5906 {
5907 unsigned long num, scale;
5908 int ch, dot;
5909 const char *cp = *cp_p;
5910
5911 num = 0;
5912 scale = 1;
5913 dot = 0;
5914 for (;;) {
5915 ch = *cp;
5916 if ( !dot && ch == '.' ) {
5917 scale = 1;
5918 dot = 1;
5919 } else if ( ch == '%' ) {
5920 scale = dot ? scale*100 : 100;
5921 cp++; /* % is always at the end */
5922 break;
5923 } else if ( ch >= '0' && ch <= '9' ) {
5924 if ( scale < 100000 ) {
5925 scale *= 10;
5926 num = (num*10) + (ch-'0');
5927 }
5928 } else {
5929 break;
5930 }
5931 cp++;
5932 }
5933 *cp_p = cp;
5934
5935 /* user says num divided by scale and we say internally that
5936 * is MAX_SCORE * num / scale.
5937 */
5938 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5939 }
5940
5941 struct diff_queue_struct diff_queued_diff;
5942
5943 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5944 {
5945 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5946 queue->queue[queue->nr++] = dp;
5947 }
5948
5949 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5950 struct diff_filespec *one,
5951 struct diff_filespec *two)
5952 {
5953 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5954 dp->one = one;
5955 dp->two = two;
5956 if (queue)
5957 diff_q(queue, dp);
5958 return dp;
5959 }
5960
5961 void diff_free_filepair(struct diff_filepair *p)
5962 {
5963 free_filespec(p->one);
5964 free_filespec(p->two);
5965 free(p);
5966 }
5967
5968 void diff_free_queue(struct diff_queue_struct *q)
5969 {
5970 for (int i = 0; i < q->nr; i++)
5971 diff_free_filepair(q->queue[i]);
5972 free(q->queue);
5973 }
5974
5975 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5976 {
5977 int abblen;
5978 const char *abbrev;
5979
5980 /* Do we want all 40 hex characters? */
5981 if (len == the_hash_algo->hexsz)
5982 return oid_to_hex(oid);
5983
5984 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5985 abbrev = diff_abbrev_oid(oid, len);
5986
5987 if (!print_sha1_ellipsis())
5988 return abbrev;
5989
5990 abblen = strlen(abbrev);
5991
5992 /*
5993 * In well-behaved cases, where the abbreviated result is the
5994 * same as the requested length, append three dots after the
5995 * abbreviation (hence the whole logic is limited to the case
5996 * where abblen < 37); when the actual abbreviated result is a
5997 * bit longer than the requested length, we reduce the number
5998 * of dots so that they match the well-behaved ones. However,
5999 * if the actual abbreviation is longer than the requested
6000 * length by more than three, we give up on aligning, and add
6001 * three dots anyway, to indicate that the output is not the
6002 * full object name. Yes, this may be suboptimal, but this
6003 * appears only in "diff --raw --abbrev" output and it is not
6004 * worth the effort to change it now. Note that this would
6005 * likely to work fine when the automatic sizing of default
6006 * abbreviation length is used--we would be fed -1 in "len" in
6007 * that case, and will end up always appending three-dots, but
6008 * the automatic sizing is supposed to give abblen that ensures
6009 * uniqueness across all objects (statistically speaking).
6010 */
6011 if (abblen < the_hash_algo->hexsz - 3) {
6012 static char hex[GIT_MAX_HEXSZ + 1];
6013 if (len < abblen && abblen <= len + 2)
6014 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
6015 else
6016 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
6017 return hex;
6018 }
6019
6020 return oid_to_hex(oid);
6021 }
6022
6023 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
6024 {
6025 int line_termination = opt->line_termination;
6026 int inter_name_termination = line_termination ? '\t' : '\0';
6027
6028 fprintf(opt->file, "%s", diff_line_prefix(opt));
6029 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
6030 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
6031 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
6032 fprintf(opt->file, "%s ",
6033 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
6034 }
6035 if (p->score) {
6036 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
6037 inter_name_termination);
6038 } else {
6039 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
6040 }
6041
6042 if (p->status == DIFF_STATUS_COPIED ||
6043 p->status == DIFF_STATUS_RENAMED) {
6044 const char *name_a, *name_b;
6045 name_a = p->one->path;
6046 name_b = p->two->path;
6047 strip_prefix(opt->prefix_length, &name_a, &name_b);
6048 write_name_quoted(name_a, opt->file, inter_name_termination);
6049 write_name_quoted(name_b, opt->file, line_termination);
6050 } else {
6051 const char *name_a, *name_b;
6052 name_a = p->one->mode ? p->one->path : p->two->path;
6053 name_b = NULL;
6054 strip_prefix(opt->prefix_length, &name_a, &name_b);
6055 write_name_quoted(name_a, opt->file, line_termination);
6056 }
6057 }
6058
6059 int diff_unmodified_pair(struct diff_filepair *p)
6060 {
6061 /* This function is written stricter than necessary to support
6062 * the currently implemented transformers, but the idea is to
6063 * let transformers to produce diff_filepairs any way they want,
6064 * and filter and clean them up here before producing the output.
6065 */
6066 struct diff_filespec *one = p->one, *two = p->two;
6067
6068 if (DIFF_PAIR_UNMERGED(p))
6069 return 0; /* unmerged is interesting */
6070
6071 /* deletion, addition, mode or type change
6072 * and rename are all interesting.
6073 */
6074 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
6075 DIFF_PAIR_MODE_CHANGED(p) ||
6076 strcmp(one->path, two->path))
6077 return 0;
6078
6079 /* both are valid and point at the same path. that is, we are
6080 * dealing with a change.
6081 */
6082 if (one->oid_valid && two->oid_valid &&
6083 oideq(&one->oid, &two->oid) &&
6084 !one->dirty_submodule && !two->dirty_submodule)
6085 return 1; /* no change */
6086 if (!one->oid_valid && !two->oid_valid)
6087 return 1; /* both look at the same file on the filesystem. */
6088 return 0;
6089 }
6090
6091 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
6092 {
6093 int include_conflict_headers =
6094 (additional_headers(o, p->one->path) &&
6095 !o->pickaxe_opts &&
6096 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6097
6098 /*
6099 * Check if we can return early without showing a diff. Note that
6100 * diff_filepair only stores {oid, path, mode, is_valid}
6101 * information for each path, and thus diff_unmodified_pair() only
6102 * considers those bits of info. However, we do not want pairs
6103 * created by create_filepairs_for_header_only_notifications()
6104 * (which always look like unmodified pairs) to be ignored, so
6105 * return early if both p is unmodified AND we don't want to
6106 * include_conflict_headers.
6107 */
6108 if (diff_unmodified_pair(p) && !include_conflict_headers)
6109 return;
6110
6111 /* Actually, we can also return early to avoid showing tree diffs */
6112 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6113 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6114 return;
6115
6116 run_diff(p, o);
6117 }
6118
6119 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
6120 struct diffstat_t *diffstat)
6121 {
6122 if (diff_unmodified_pair(p))
6123 return;
6124
6125 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6126 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6127 return; /* no useful stat for tree diffs */
6128
6129 run_diffstat(p, o, diffstat);
6130 }
6131
6132 static void diff_flush_checkdiff(struct diff_filepair *p,
6133 struct diff_options *o)
6134 {
6135 if (diff_unmodified_pair(p))
6136 return;
6137
6138 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6139 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6140 return; /* nothing to check in tree diffs */
6141
6142 run_checkdiff(p, o);
6143 }
6144
6145 int diff_queue_is_empty(struct diff_options *o)
6146 {
6147 struct diff_queue_struct *q = &diff_queued_diff;
6148 int i;
6149 int include_conflict_headers =
6150 (o->additional_path_headers &&
6151 strmap_get_size(o->additional_path_headers) &&
6152 !o->pickaxe_opts &&
6153 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6154
6155 if (include_conflict_headers)
6156 return 0;
6157
6158 for (i = 0; i < q->nr; i++)
6159 if (!diff_unmodified_pair(q->queue[i]))
6160 return 0;
6161 return 1;
6162 }
6163
6164 #if DIFF_DEBUG
6165 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6166 {
6167 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6168 x, one ? one : "",
6169 s->path,
6170 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6171 s->mode,
6172 s->oid_valid ? oid_to_hex(&s->oid) : "");
6173 fprintf(stderr, "queue[%d] %s size %lu\n",
6174 x, one ? one : "",
6175 s->size);
6176 }
6177
6178 void diff_debug_filepair(const struct diff_filepair *p, int i)
6179 {
6180 diff_debug_filespec(p->one, i, "one");
6181 diff_debug_filespec(p->two, i, "two");
6182 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6183 p->score, p->status ? p->status : '?',
6184 p->one->rename_used, p->broken_pair);
6185 }
6186
6187 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6188 {
6189 int i;
6190 if (msg)
6191 fprintf(stderr, "%s\n", msg);
6192 fprintf(stderr, "q->nr = %d\n", q->nr);
6193 for (i = 0; i < q->nr; i++) {
6194 struct diff_filepair *p = q->queue[i];
6195 diff_debug_filepair(p, i);
6196 }
6197 }
6198 #endif
6199
6200 static void diff_resolve_rename_copy(void)
6201 {
6202 int i;
6203 struct diff_filepair *p;
6204 struct diff_queue_struct *q = &diff_queued_diff;
6205
6206 diff_debug_queue("resolve-rename-copy", q);
6207
6208 for (i = 0; i < q->nr; i++) {
6209 p = q->queue[i];
6210 p->status = 0; /* undecided */
6211 if (DIFF_PAIR_UNMERGED(p))
6212 p->status = DIFF_STATUS_UNMERGED;
6213 else if (!DIFF_FILE_VALID(p->one))
6214 p->status = DIFF_STATUS_ADDED;
6215 else if (!DIFF_FILE_VALID(p->two))
6216 p->status = DIFF_STATUS_DELETED;
6217 else if (DIFF_PAIR_TYPE_CHANGED(p))
6218 p->status = DIFF_STATUS_TYPE_CHANGED;
6219
6220 /* from this point on, we are dealing with a pair
6221 * whose both sides are valid and of the same type, i.e.
6222 * either in-place edit or rename/copy edit.
6223 */
6224 else if (DIFF_PAIR_RENAME(p)) {
6225 /*
6226 * A rename might have re-connected a broken
6227 * pair up, causing the pathnames to be the
6228 * same again. If so, that's not a rename at
6229 * all, just a modification..
6230 *
6231 * Otherwise, see if this source was used for
6232 * multiple renames, in which case we decrement
6233 * the count, and call it a copy.
6234 */
6235 if (!strcmp(p->one->path, p->two->path))
6236 p->status = DIFF_STATUS_MODIFIED;
6237 else if (--p->one->rename_used > 0)
6238 p->status = DIFF_STATUS_COPIED;
6239 else
6240 p->status = DIFF_STATUS_RENAMED;
6241 }
6242 else if (!oideq(&p->one->oid, &p->two->oid) ||
6243 p->one->mode != p->two->mode ||
6244 p->one->dirty_submodule ||
6245 p->two->dirty_submodule ||
6246 is_null_oid(&p->one->oid))
6247 p->status = DIFF_STATUS_MODIFIED;
6248 else {
6249 /* This is a "no-change" entry and should not
6250 * happen anymore, but prepare for broken callers.
6251 */
6252 error("feeding unmodified %s to diffcore",
6253 p->one->path);
6254 p->status = DIFF_STATUS_UNKNOWN;
6255 }
6256 }
6257 diff_debug_queue("resolve-rename-copy done", q);
6258 }
6259
6260 static int check_pair_status(struct diff_filepair *p)
6261 {
6262 switch (p->status) {
6263 case DIFF_STATUS_UNKNOWN:
6264 return 0;
6265 case 0:
6266 die("internal error in diff-resolve-rename-copy");
6267 default:
6268 return 1;
6269 }
6270 }
6271
6272 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6273 {
6274 int fmt = opt->output_format;
6275
6276 if (fmt & DIFF_FORMAT_CHECKDIFF)
6277 diff_flush_checkdiff(p, opt);
6278 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6279 diff_flush_raw(p, opt);
6280 else if (fmt & DIFF_FORMAT_NAME) {
6281 const char *name_a, *name_b;
6282 name_a = p->two->path;
6283 name_b = NULL;
6284 strip_prefix(opt->prefix_length, &name_a, &name_b);
6285 fprintf(opt->file, "%s", diff_line_prefix(opt));
6286 write_name_quoted(name_a, opt->file, opt->line_termination);
6287 }
6288
6289 opt->found_changes = 1;
6290 }
6291
6292 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6293 {
6294 struct strbuf sb = STRBUF_INIT;
6295 if (fs->mode)
6296 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6297 else
6298 strbuf_addf(&sb, " %s ", newdelete);
6299
6300 quote_c_style(fs->path, &sb, NULL, 0);
6301 strbuf_addch(&sb, '\n');
6302 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6303 sb.buf, sb.len, 0);
6304 strbuf_release(&sb);
6305 }
6306
6307 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6308 int show_name)
6309 {
6310 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6311 struct strbuf sb = STRBUF_INIT;
6312 strbuf_addf(&sb, " mode change %06o => %06o",
6313 p->one->mode, p->two->mode);
6314 if (show_name) {
6315 strbuf_addch(&sb, ' ');
6316 quote_c_style(p->two->path, &sb, NULL, 0);
6317 }
6318 strbuf_addch(&sb, '\n');
6319 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6320 sb.buf, sb.len, 0);
6321 strbuf_release(&sb);
6322 }
6323 }
6324
6325 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6326 struct diff_filepair *p)
6327 {
6328 struct strbuf sb = STRBUF_INIT;
6329 struct strbuf names = STRBUF_INIT;
6330
6331 pprint_rename(&names, p->one->path, p->two->path);
6332 strbuf_addf(&sb, " %s %s (%d%%)\n",
6333 renamecopy, names.buf, similarity_index(p));
6334 strbuf_release(&names);
6335 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6336 sb.buf, sb.len, 0);
6337 show_mode_change(opt, p, 0);
6338 strbuf_release(&sb);
6339 }
6340
6341 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6342 {
6343 switch(p->status) {
6344 case DIFF_STATUS_DELETED:
6345 show_file_mode_name(opt, "delete", p->one);
6346 break;
6347 case DIFF_STATUS_ADDED:
6348 show_file_mode_name(opt, "create", p->two);
6349 break;
6350 case DIFF_STATUS_COPIED:
6351 show_rename_copy(opt, "copy", p);
6352 break;
6353 case DIFF_STATUS_RENAMED:
6354 show_rename_copy(opt, "rename", p);
6355 break;
6356 default:
6357 if (p->score) {
6358 struct strbuf sb = STRBUF_INIT;
6359 strbuf_addstr(&sb, " rewrite ");
6360 quote_c_style(p->two->path, &sb, NULL, 0);
6361 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6362 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6363 sb.buf, sb.len, 0);
6364 strbuf_release(&sb);
6365 }
6366 show_mode_change(opt, p, !p->score);
6367 break;
6368 }
6369 }
6370
6371 struct patch_id_t {
6372 git_hash_ctx *ctx;
6373 int patchlen;
6374 };
6375
6376 static int remove_space(char *line, int len)
6377 {
6378 int i;
6379 char *dst = line;
6380 unsigned char c;
6381
6382 for (i = 0; i < len; i++)
6383 if (!isspace((c = line[i])))
6384 *dst++ = c;
6385
6386 return dst - line;
6387 }
6388
6389 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6390 {
6391 unsigned char hash[GIT_MAX_RAWSZ];
6392 unsigned short carry = 0;
6393 int i;
6394
6395 the_hash_algo->final_fn(hash, ctx);
6396 the_hash_algo->init_fn(ctx);
6397 /* 20-byte sum, with carry */
6398 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6399 carry += result->hash[i] + hash[i];
6400 result->hash[i] = carry;
6401 carry >>= 8;
6402 }
6403 }
6404
6405 static int patch_id_consume(void *priv, char *line, unsigned long len)
6406 {
6407 struct patch_id_t *data = priv;
6408 int new_len;
6409
6410 if (len > 12 && starts_with(line, "\\ "))
6411 return 0;
6412 new_len = remove_space(line, len);
6413
6414 the_hash_algo->update_fn(data->ctx, line, new_len);
6415 data->patchlen += new_len;
6416 return 0;
6417 }
6418
6419 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6420 {
6421 the_hash_algo->update_fn(ctx, str, strlen(str));
6422 }
6423
6424 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6425 {
6426 /* large enough for 2^32 in octal */
6427 char buf[12];
6428 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6429 the_hash_algo->update_fn(ctx, buf, len);
6430 }
6431
6432 /* returns 0 upon success, and writes result into oid */
6433 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6434 {
6435 struct diff_queue_struct *q = &diff_queued_diff;
6436 int i;
6437 git_hash_ctx ctx;
6438 struct patch_id_t data;
6439
6440 the_hash_algo->init_fn(&ctx);
6441 memset(&data, 0, sizeof(struct patch_id_t));
6442 data.ctx = &ctx;
6443 oidclr(oid);
6444
6445 for (i = 0; i < q->nr; i++) {
6446 xpparam_t xpp;
6447 xdemitconf_t xecfg;
6448 mmfile_t mf1, mf2;
6449 struct diff_filepair *p = q->queue[i];
6450 int len1, len2;
6451
6452 memset(&xpp, 0, sizeof(xpp));
6453 memset(&xecfg, 0, sizeof(xecfg));
6454 if (p->status == 0)
6455 return error("internal diff status error");
6456 if (p->status == DIFF_STATUS_UNKNOWN)
6457 continue;
6458 if (diff_unmodified_pair(p))
6459 continue;
6460 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6461 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6462 continue;
6463 if (DIFF_PAIR_UNMERGED(p))
6464 continue;
6465
6466 diff_fill_oid_info(p->one, options->repo->index);
6467 diff_fill_oid_info(p->two, options->repo->index);
6468
6469 len1 = remove_space(p->one->path, strlen(p->one->path));
6470 len2 = remove_space(p->two->path, strlen(p->two->path));
6471 patch_id_add_string(&ctx, "diff--git");
6472 patch_id_add_string(&ctx, "a/");
6473 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6474 patch_id_add_string(&ctx, "b/");
6475 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6476
6477 if (p->one->mode == 0) {
6478 patch_id_add_string(&ctx, "newfilemode");
6479 patch_id_add_mode(&ctx, p->two->mode);
6480 } else if (p->two->mode == 0) {
6481 patch_id_add_string(&ctx, "deletedfilemode");
6482 patch_id_add_mode(&ctx, p->one->mode);
6483 } else if (p->one->mode != p->two->mode) {
6484 patch_id_add_string(&ctx, "oldmode");
6485 patch_id_add_mode(&ctx, p->one->mode);
6486 patch_id_add_string(&ctx, "newmode");
6487 patch_id_add_mode(&ctx, p->two->mode);
6488 }
6489
6490 if (diff_header_only) {
6491 /* don't do anything since we're only populating header info */
6492 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6493 diff_filespec_is_binary(options->repo, p->two)) {
6494 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6495 the_hash_algo->hexsz);
6496 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6497 the_hash_algo->hexsz);
6498 } else {
6499 if (p->one->mode == 0) {
6500 patch_id_add_string(&ctx, "---/dev/null");
6501 patch_id_add_string(&ctx, "+++b/");
6502 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6503 } else if (p->two->mode == 0) {
6504 patch_id_add_string(&ctx, "---a/");
6505 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6506 patch_id_add_string(&ctx, "+++/dev/null");
6507 } else {
6508 patch_id_add_string(&ctx, "---a/");
6509 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6510 patch_id_add_string(&ctx, "+++b/");
6511 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6512 }
6513
6514 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6515 fill_mmfile(options->repo, &mf2, p->two) < 0)
6516 return error("unable to read files to diff");
6517 xpp.flags = 0;
6518 xecfg.ctxlen = 3;
6519 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6520 if (xdi_diff_outf(&mf1, &mf2, NULL,
6521 patch_id_consume, &data, &xpp, &xecfg))
6522 return error("unable to generate patch-id diff for %s",
6523 p->one->path);
6524 }
6525 flush_one_hunk(oid, &ctx);
6526 }
6527
6528 return 0;
6529 }
6530
6531 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6532 {
6533 struct diff_queue_struct *q = &diff_queued_diff;
6534 int result = diff_get_patch_id(options, oid, diff_header_only);
6535
6536 diff_free_queue(q);
6537 DIFF_QUEUE_CLEAR(q);
6538
6539 return result;
6540 }
6541
6542 static int is_summary_empty(const struct diff_queue_struct *q)
6543 {
6544 int i;
6545
6546 for (i = 0; i < q->nr; i++) {
6547 const struct diff_filepair *p = q->queue[i];
6548
6549 switch (p->status) {
6550 case DIFF_STATUS_DELETED:
6551 case DIFF_STATUS_ADDED:
6552 case DIFF_STATUS_COPIED:
6553 case DIFF_STATUS_RENAMED:
6554 return 0;
6555 default:
6556 if (p->score)
6557 return 0;
6558 if (p->one->mode && p->two->mode &&
6559 p->one->mode != p->two->mode)
6560 return 0;
6561 break;
6562 }
6563 }
6564 return 1;
6565 }
6566
6567 static const char rename_limit_warning[] =
6568 N_("exhaustive rename detection was skipped due to too many files.");
6569
6570 static const char degrade_cc_to_c_warning[] =
6571 N_("only found copies from modified paths due to too many files.");
6572
6573 static const char rename_limit_advice[] =
6574 N_("you may want to set your %s variable to at least "
6575 "%d and retry the command.");
6576
6577 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6578 {
6579 fflush(stdout);
6580 if (degraded_cc)
6581 warning(_(degrade_cc_to_c_warning));
6582 else if (needed)
6583 warning(_(rename_limit_warning));
6584 else
6585 return;
6586 if (0 < needed)
6587 warning(_(rename_limit_advice), varname, needed);
6588 }
6589
6590 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6591 {
6592 struct strset present;
6593 struct diff_queue_struct *q = &diff_queued_diff;
6594 struct hashmap_iter iter;
6595 struct strmap_entry *e;
6596 int i;
6597
6598 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6599
6600 /*
6601 * Find out which paths exist in diff_queued_diff, preferring
6602 * one->path for any pair that has multiple paths.
6603 */
6604 for (i = 0; i < q->nr; i++) {
6605 struct diff_filepair *p = q->queue[i];
6606 char *path = p->one->path ? p->one->path : p->two->path;
6607
6608 if (strmap_contains(o->additional_path_headers, path))
6609 strset_add(&present, path);
6610 }
6611
6612 /*
6613 * Loop over paths in additional_path_headers; for each NOT already
6614 * in diff_queued_diff, create a synthetic filepair and insert that
6615 * into diff_queued_diff.
6616 */
6617 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6618 if (!strset_contains(&present, e->key)) {
6619 struct diff_filespec *one, *two;
6620 struct diff_filepair *p;
6621
6622 one = alloc_filespec(e->key);
6623 two = alloc_filespec(e->key);
6624 fill_filespec(one, null_oid(), 0, 0);
6625 fill_filespec(two, null_oid(), 0, 0);
6626 p = diff_queue(q, one, two);
6627 p->status = DIFF_STATUS_MODIFIED;
6628 }
6629 }
6630
6631 /* Re-sort the filepairs */
6632 diffcore_fix_diff_index();
6633
6634 /* Cleanup */
6635 strset_clear(&present);
6636 }
6637
6638 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6639 {
6640 int i;
6641 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6642 struct diff_queue_struct *q = &diff_queued_diff;
6643
6644 if (WSEH_NEW & WS_RULE_MASK)
6645 BUG("WS rules bit mask overlaps with diff symbol flags");
6646
6647 if (o->color_moved)
6648 o->emitted_symbols = &esm;
6649
6650 if (o->additional_path_headers)
6651 create_filepairs_for_header_only_notifications(o);
6652
6653 for (i = 0; i < q->nr; i++) {
6654 struct diff_filepair *p = q->queue[i];
6655 if (check_pair_status(p))
6656 diff_flush_patch(p, o);
6657 }
6658
6659 if (o->emitted_symbols) {
6660 if (o->color_moved) {
6661 struct mem_pool entry_pool;
6662 struct moved_entry_list *entry_list;
6663
6664 mem_pool_init(&entry_pool, 1024 * 1024);
6665 entry_list = add_lines_to_move_detection(o,
6666 &entry_pool);
6667 mark_color_as_moved(o, entry_list);
6668 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6669 dim_moved_lines(o);
6670
6671 mem_pool_discard(&entry_pool, 0);
6672 free(entry_list);
6673 }
6674
6675 for (i = 0; i < esm.nr; i++)
6676 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6677
6678 for (i = 0; i < esm.nr; i++)
6679 free((void *)esm.buf[i].line);
6680 esm.nr = 0;
6681
6682 o->emitted_symbols = NULL;
6683 }
6684 }
6685
6686 static void diff_free_file(struct diff_options *options)
6687 {
6688 if (options->close_file)
6689 fclose(options->file);
6690 }
6691
6692 static void diff_free_ignore_regex(struct diff_options *options)
6693 {
6694 int i;
6695
6696 for (i = 0; i < options->ignore_regex_nr; i++) {
6697 regfree(options->ignore_regex[i]);
6698 free(options->ignore_regex[i]);
6699 }
6700 free(options->ignore_regex);
6701 }
6702
6703 void diff_free(struct diff_options *options)
6704 {
6705 if (options->no_free)
6706 return;
6707
6708 diff_free_file(options);
6709 diff_free_ignore_regex(options);
6710 clear_pathspec(&options->pathspec);
6711 }
6712
6713 void diff_flush(struct diff_options *options)
6714 {
6715 struct diff_queue_struct *q = &diff_queued_diff;
6716 int i, output_format = options->output_format;
6717 int separator = 0;
6718 int dirstat_by_line = 0;
6719
6720 /*
6721 * Order: raw, stat, summary, patch
6722 * or: name/name-status/checkdiff (other bits clear)
6723 */
6724 if (!q->nr && !options->additional_path_headers)
6725 goto free_queue;
6726
6727 if (output_format & (DIFF_FORMAT_RAW |
6728 DIFF_FORMAT_NAME |
6729 DIFF_FORMAT_NAME_STATUS |
6730 DIFF_FORMAT_CHECKDIFF)) {
6731 for (i = 0; i < q->nr; i++) {
6732 struct diff_filepair *p = q->queue[i];
6733 if (check_pair_status(p))
6734 flush_one_pair(p, options);
6735 }
6736 separator++;
6737 }
6738
6739 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6740 dirstat_by_line = 1;
6741
6742 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6743 dirstat_by_line) {
6744 struct diffstat_t diffstat;
6745
6746 compute_diffstat(options, &diffstat, q);
6747 if (output_format & DIFF_FORMAT_NUMSTAT)
6748 show_numstat(&diffstat, options);
6749 if (output_format & DIFF_FORMAT_DIFFSTAT)
6750 show_stats(&diffstat, options);
6751 if (output_format & DIFF_FORMAT_SHORTSTAT)
6752 show_shortstats(&diffstat, options);
6753 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6754 show_dirstat_by_line(&diffstat, options);
6755 free_diffstat_info(&diffstat);
6756 separator++;
6757 }
6758 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6759 show_dirstat(options);
6760
6761 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6762 for (i = 0; i < q->nr; i++) {
6763 diff_summary(options, q->queue[i]);
6764 }
6765 separator++;
6766 }
6767
6768 if (output_format & DIFF_FORMAT_PATCH) {
6769 if (separator) {
6770 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6771 if (options->stat_sep)
6772 /* attach patch instead of inline */
6773 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6774 NULL, 0, 0);
6775 }
6776
6777 diff_flush_patch_all_file_pairs(options);
6778 }
6779
6780 if (output_format & DIFF_FORMAT_CALLBACK)
6781 options->format_callback(q, options, options->format_callback_data);
6782
6783 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6784 options->flags.exit_with_status &&
6785 options->flags.diff_from_contents) {
6786 /*
6787 * run diff_flush_patch for the exit status. setting
6788 * options->file to /dev/null should be safe, because we
6789 * aren't supposed to produce any output anyway.
6790 */
6791 diff_free_file(options);
6792 options->file = xfopen("/dev/null", "w");
6793 options->close_file = 1;
6794 options->color_moved = 0;
6795 for (i = 0; i < q->nr; i++) {
6796 struct diff_filepair *p = q->queue[i];
6797 if (check_pair_status(p))
6798 diff_flush_patch(p, options);
6799 if (options->found_changes)
6800 break;
6801 }
6802 }
6803
6804 free_queue:
6805 diff_free_queue(q);
6806 DIFF_QUEUE_CLEAR(q);
6807 diff_free(options);
6808
6809 /*
6810 * Report the content-level differences with HAS_CHANGES;
6811 * diff_addremove/diff_change does not set the bit when
6812 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6813 */
6814 if (options->flags.diff_from_contents) {
6815 if (options->found_changes)
6816 options->flags.has_changes = 1;
6817 else
6818 options->flags.has_changes = 0;
6819 }
6820 }
6821
6822 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6823 {
6824 return (((p->status == DIFF_STATUS_MODIFIED) &&
6825 ((p->score &&
6826 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6827 (!p->score &&
6828 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6829 ((p->status != DIFF_STATUS_MODIFIED) &&
6830 filter_bit_tst(p->status, options)));
6831 }
6832
6833 static void diffcore_apply_filter(struct diff_options *options)
6834 {
6835 int i;
6836 struct diff_queue_struct *q = &diff_queued_diff;
6837 struct diff_queue_struct outq;
6838
6839 DIFF_QUEUE_CLEAR(&outq);
6840
6841 if (!options->filter)
6842 return;
6843
6844 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6845 int found;
6846 for (i = found = 0; !found && i < q->nr; i++) {
6847 if (match_filter(options, q->queue[i]))
6848 found++;
6849 }
6850 if (found)
6851 return;
6852
6853 /* otherwise we will clear the whole queue
6854 * by copying the empty outq at the end of this
6855 * function, but first clear the current entries
6856 * in the queue.
6857 */
6858 for (i = 0; i < q->nr; i++)
6859 diff_free_filepair(q->queue[i]);
6860 }
6861 else {
6862 /* Only the matching ones */
6863 for (i = 0; i < q->nr; i++) {
6864 struct diff_filepair *p = q->queue[i];
6865 if (match_filter(options, p))
6866 diff_q(&outq, p);
6867 else
6868 diff_free_filepair(p);
6869 }
6870 }
6871 free(q->queue);
6872 *q = outq;
6873 }
6874
6875 /* Check whether two filespecs with the same mode and size are identical */
6876 static int diff_filespec_is_identical(struct repository *r,
6877 struct diff_filespec *one,
6878 struct diff_filespec *two)
6879 {
6880 if (S_ISGITLINK(one->mode))
6881 return 0;
6882 if (diff_populate_filespec(r, one, NULL))
6883 return 0;
6884 if (diff_populate_filespec(r, two, NULL))
6885 return 0;
6886 return !memcmp(one->data, two->data, one->size);
6887 }
6888
6889 static int diff_filespec_check_stat_unmatch(struct repository *r,
6890 struct diff_filepair *p)
6891 {
6892 struct diff_populate_filespec_options dpf_options = {
6893 .check_size_only = 1,
6894 .missing_object_cb = diff_queued_diff_prefetch,
6895 .missing_object_data = r,
6896 };
6897
6898 if (p->done_skip_stat_unmatch)
6899 return p->skip_stat_unmatch_result;
6900
6901 p->done_skip_stat_unmatch = 1;
6902 p->skip_stat_unmatch_result = 0;
6903 /*
6904 * 1. Entries that come from stat info dirtiness
6905 * always have both sides (iow, not create/delete),
6906 * one side of the object name is unknown, with
6907 * the same mode and size. Keep the ones that
6908 * do not match these criteria. They have real
6909 * differences.
6910 *
6911 * 2. At this point, the file is known to be modified,
6912 * with the same mode and size, and the object
6913 * name of one side is unknown. Need to inspect
6914 * the identical contents.
6915 */
6916 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6917 !DIFF_FILE_VALID(p->two) ||
6918 (p->one->oid_valid && p->two->oid_valid) ||
6919 (p->one->mode != p->two->mode) ||
6920 diff_populate_filespec(r, p->one, &dpf_options) ||
6921 diff_populate_filespec(r, p->two, &dpf_options) ||
6922 (p->one->size != p->two->size) ||
6923 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6924 p->skip_stat_unmatch_result = 1;
6925 return p->skip_stat_unmatch_result;
6926 }
6927
6928 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6929 {
6930 int i;
6931 struct diff_queue_struct *q = &diff_queued_diff;
6932 struct diff_queue_struct outq;
6933 DIFF_QUEUE_CLEAR(&outq);
6934
6935 for (i = 0; i < q->nr; i++) {
6936 struct diff_filepair *p = q->queue[i];
6937
6938 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6939 diff_q(&outq, p);
6940 else {
6941 /*
6942 * The caller can subtract 1 from skip_stat_unmatch
6943 * to determine how many paths were dirty only
6944 * due to stat info mismatch.
6945 */
6946 if (!diffopt->flags.no_index)
6947 diffopt->skip_stat_unmatch++;
6948 diff_free_filepair(p);
6949 }
6950 }
6951 free(q->queue);
6952 *q = outq;
6953 }
6954
6955 static int diffnamecmp(const void *a_, const void *b_)
6956 {
6957 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6958 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6959 const char *name_a, *name_b;
6960
6961 name_a = a->one ? a->one->path : a->two->path;
6962 name_b = b->one ? b->one->path : b->two->path;
6963 return strcmp(name_a, name_b);
6964 }
6965
6966 void diffcore_fix_diff_index(void)
6967 {
6968 struct diff_queue_struct *q = &diff_queued_diff;
6969 QSORT(q->queue, q->nr, diffnamecmp);
6970 }
6971
6972 void diff_add_if_missing(struct repository *r,
6973 struct oid_array *to_fetch,
6974 const struct diff_filespec *filespec)
6975 {
6976 if (filespec && filespec->oid_valid &&
6977 !S_ISGITLINK(filespec->mode) &&
6978 oid_object_info_extended(r, &filespec->oid, NULL,
6979 OBJECT_INFO_FOR_PREFETCH))
6980 oid_array_append(to_fetch, &filespec->oid);
6981 }
6982
6983 void diff_queued_diff_prefetch(void *repository)
6984 {
6985 struct repository *repo = repository;
6986 int i;
6987 struct diff_queue_struct *q = &diff_queued_diff;
6988 struct oid_array to_fetch = OID_ARRAY_INIT;
6989
6990 for (i = 0; i < q->nr; i++) {
6991 struct diff_filepair *p = q->queue[i];
6992 diff_add_if_missing(repo, &to_fetch, p->one);
6993 diff_add_if_missing(repo, &to_fetch, p->two);
6994 }
6995
6996 /*
6997 * NEEDSWORK: Consider deduplicating the OIDs sent.
6998 */
6999 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
7000
7001 oid_array_clear(&to_fetch);
7002 }
7003
7004 void init_diffstat_widths(struct diff_options *options)
7005 {
7006 options->stat_width = -1; /* use full terminal width */
7007 options->stat_name_width = -1; /* respect diff.statNameWidth config */
7008 options->stat_graph_width = -1; /* respect diff.statGraphWidth config */
7009 }
7010
7011 void diffcore_std(struct diff_options *options)
7012 {
7013 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
7014 DIFF_FORMAT_NUMSTAT |
7015 DIFF_FORMAT_PATCH |
7016 DIFF_FORMAT_SHORTSTAT |
7017 DIFF_FORMAT_DIRSTAT;
7018
7019 /*
7020 * Check if the user requested a blob-data-requiring diff output and/or
7021 * break-rewrite detection (which requires blob data). If yes, prefetch
7022 * the diff pairs.
7023 *
7024 * If no prefetching occurs, diffcore_rename() will prefetch if it
7025 * decides that it needs inexact rename detection.
7026 */
7027 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
7028 (options->output_format & output_formats_to_prefetch ||
7029 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
7030 diff_queued_diff_prefetch(options->repo);
7031
7032 /* NOTE please keep the following in sync with diff_tree_combined() */
7033 if (options->skip_stat_unmatch)
7034 diffcore_skip_stat_unmatch(options);
7035 if (!options->found_follow) {
7036 /* See try_to_follow_renames() in tree-diff.c */
7037 if (options->break_opt != -1)
7038 diffcore_break(options->repo,
7039 options->break_opt);
7040 if (options->detect_rename)
7041 diffcore_rename(options);
7042 if (options->break_opt != -1)
7043 diffcore_merge_broken();
7044 }
7045 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
7046 diffcore_pickaxe(options);
7047 if (options->orderfile)
7048 diffcore_order(options->orderfile);
7049 if (options->rotate_to)
7050 diffcore_rotate(options);
7051 if (!options->found_follow)
7052 /* See try_to_follow_renames() in tree-diff.c */
7053 diff_resolve_rename_copy();
7054 diffcore_apply_filter(options);
7055
7056 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
7057 options->flags.has_changes = 1;
7058 else
7059 options->flags.has_changes = 0;
7060
7061 options->found_follow = 0;
7062 }
7063
7064 int diff_result_code(struct diff_options *opt)
7065 {
7066 int result = 0;
7067
7068 diff_warn_rename_limit("diff.renameLimit",
7069 opt->needed_rename_limit,
7070 opt->degraded_cc_to_c);
7071
7072 if (opt->flags.exit_with_status &&
7073 opt->flags.has_changes)
7074 result |= 01;
7075 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
7076 opt->flags.check_failed)
7077 result |= 02;
7078 return result;
7079 }
7080
7081 int diff_can_quit_early(struct diff_options *opt)
7082 {
7083 return (opt->flags.quick &&
7084 !opt->filter &&
7085 opt->flags.has_changes);
7086 }
7087
7088 /*
7089 * Shall changes to this submodule be ignored?
7090 *
7091 * Submodule changes can be configured to be ignored separately for each path,
7092 * but that configuration can be overridden from the command line.
7093 */
7094 static int is_submodule_ignored(const char *path, struct diff_options *options)
7095 {
7096 int ignored = 0;
7097 struct diff_flags orig_flags = options->flags;
7098 if (!options->flags.override_submodule_config)
7099 set_diffopt_flags_from_submodule_config(options, path);
7100 if (options->flags.ignore_submodules)
7101 ignored = 1;
7102 options->flags = orig_flags;
7103 return ignored;
7104 }
7105
7106 void compute_diffstat(struct diff_options *options,
7107 struct diffstat_t *diffstat,
7108 struct diff_queue_struct *q)
7109 {
7110 int i;
7111
7112 memset(diffstat, 0, sizeof(struct diffstat_t));
7113 for (i = 0; i < q->nr; i++) {
7114 struct diff_filepair *p = q->queue[i];
7115 if (check_pair_status(p))
7116 diff_flush_stat(p, options, diffstat);
7117 }
7118 options->found_changes = !!diffstat->nr;
7119 }
7120
7121 void diff_addremove(struct diff_options *options,
7122 int addremove, unsigned mode,
7123 const struct object_id *oid,
7124 int oid_valid,
7125 const char *concatpath, unsigned dirty_submodule)
7126 {
7127 struct diff_filespec *one, *two;
7128
7129 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
7130 return;
7131
7132 /* This may look odd, but it is a preparation for
7133 * feeding "there are unchanged files which should
7134 * not produce diffs, but when you are doing copy
7135 * detection you would need them, so here they are"
7136 * entries to the diff-core. They will be prefixed
7137 * with something like '=' or '*' (I haven't decided
7138 * which but should not make any difference).
7139 * Feeding the same new and old to diff_change()
7140 * also has the same effect.
7141 * Before the final output happens, they are pruned after
7142 * merged into rename/copy pairs as appropriate.
7143 */
7144 if (options->flags.reverse_diff)
7145 addremove = (addremove == '+' ? '-' :
7146 addremove == '-' ? '+' : addremove);
7147
7148 if (options->prefix &&
7149 strncmp(concatpath, options->prefix, options->prefix_length))
7150 return;
7151
7152 one = alloc_filespec(concatpath);
7153 two = alloc_filespec(concatpath);
7154
7155 if (addremove != '+')
7156 fill_filespec(one, oid, oid_valid, mode);
7157 if (addremove != '-') {
7158 fill_filespec(two, oid, oid_valid, mode);
7159 two->dirty_submodule = dirty_submodule;
7160 }
7161
7162 diff_queue(&diff_queued_diff, one, two);
7163 if (!options->flags.diff_from_contents)
7164 options->flags.has_changes = 1;
7165 }
7166
7167 void diff_change(struct diff_options *options,
7168 unsigned old_mode, unsigned new_mode,
7169 const struct object_id *old_oid,
7170 const struct object_id *new_oid,
7171 int old_oid_valid, int new_oid_valid,
7172 const char *concatpath,
7173 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7174 {
7175 struct diff_filespec *one, *two;
7176 struct diff_filepair *p;
7177
7178 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7179 is_submodule_ignored(concatpath, options))
7180 return;
7181
7182 if (options->flags.reverse_diff) {
7183 SWAP(old_mode, new_mode);
7184 SWAP(old_oid, new_oid);
7185 SWAP(old_oid_valid, new_oid_valid);
7186 SWAP(old_dirty_submodule, new_dirty_submodule);
7187 }
7188
7189 if (options->prefix &&
7190 strncmp(concatpath, options->prefix, options->prefix_length))
7191 return;
7192
7193 one = alloc_filespec(concatpath);
7194 two = alloc_filespec(concatpath);
7195 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7196 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7197 one->dirty_submodule = old_dirty_submodule;
7198 two->dirty_submodule = new_dirty_submodule;
7199 p = diff_queue(&diff_queued_diff, one, two);
7200
7201 if (options->flags.diff_from_contents)
7202 return;
7203
7204 if (options->flags.quick && options->skip_stat_unmatch &&
7205 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7206 diff_free_filespec_data(p->one);
7207 diff_free_filespec_data(p->two);
7208 return;
7209 }
7210
7211 options->flags.has_changes = 1;
7212 }
7213
7214 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7215 {
7216 struct diff_filepair *pair;
7217 struct diff_filespec *one, *two;
7218
7219 if (options->prefix &&
7220 strncmp(path, options->prefix, options->prefix_length))
7221 return NULL;
7222
7223 one = alloc_filespec(path);
7224 two = alloc_filespec(path);
7225 pair = diff_queue(&diff_queued_diff, one, two);
7226 pair->is_unmerged = 1;
7227 return pair;
7228 }
7229
7230 static char *run_textconv(struct repository *r,
7231 const char *pgm,
7232 struct diff_filespec *spec,
7233 size_t *outsize)
7234 {
7235 struct diff_tempfile *temp;
7236 struct child_process child = CHILD_PROCESS_INIT;
7237 struct strbuf buf = STRBUF_INIT;
7238 int err = 0;
7239
7240 temp = prepare_temp_file(r, spec);
7241 strvec_push(&child.args, pgm);
7242 strvec_push(&child.args, temp->name);
7243
7244 child.use_shell = 1;
7245 child.out = -1;
7246 if (start_command(&child)) {
7247 remove_tempfile();
7248 return NULL;
7249 }
7250
7251 if (strbuf_read(&buf, child.out, 0) < 0)
7252 err = error("error reading from textconv command '%s'", pgm);
7253 close(child.out);
7254
7255 if (finish_command(&child) || err) {
7256 strbuf_release(&buf);
7257 remove_tempfile();
7258 return NULL;
7259 }
7260 remove_tempfile();
7261
7262 return strbuf_detach(&buf, outsize);
7263 }
7264
7265 size_t fill_textconv(struct repository *r,
7266 struct userdiff_driver *driver,
7267 struct diff_filespec *df,
7268 char **outbuf)
7269 {
7270 size_t size;
7271
7272 if (!driver) {
7273 if (!DIFF_FILE_VALID(df)) {
7274 *outbuf = (char *) "";
7275 return 0;
7276 }
7277 if (diff_populate_filespec(r, df, NULL))
7278 die("unable to read files to diff");
7279 *outbuf = df->data;
7280 return df->size;
7281 }
7282
7283 if (!driver->textconv)
7284 BUG("fill_textconv called with non-textconv driver");
7285
7286 if (driver->textconv_cache && df->oid_valid) {
7287 *outbuf = notes_cache_get(driver->textconv_cache,
7288 &df->oid,
7289 &size);
7290 if (*outbuf)
7291 return size;
7292 }
7293
7294 *outbuf = run_textconv(r, driver->textconv, df, &size);
7295 if (!*outbuf)
7296 die("unable to read files to diff");
7297
7298 if (driver->textconv_cache && df->oid_valid) {
7299 /* ignore errors, as we might be in a readonly repository */
7300 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7301 size);
7302 /*
7303 * we could save up changes and flush them all at the end,
7304 * but we would need an extra call after all diffing is done.
7305 * Since generating a cache entry is the slow path anyway,
7306 * this extra overhead probably isn't a big deal.
7307 */
7308 notes_cache_write(driver->textconv_cache);
7309 }
7310
7311 return size;
7312 }
7313
7314 int textconv_object(struct repository *r,
7315 const char *path,
7316 unsigned mode,
7317 const struct object_id *oid,
7318 int oid_valid,
7319 char **buf,
7320 unsigned long *buf_size)
7321 {
7322 struct diff_filespec *df;
7323 struct userdiff_driver *textconv;
7324
7325 df = alloc_filespec(path);
7326 fill_filespec(df, oid, oid_valid, mode);
7327 textconv = get_textconv(r, df);
7328 if (!textconv) {
7329 free_filespec(df);
7330 return 0;
7331 }
7332
7333 *buf_size = fill_textconv(r, textconv, df, buf);
7334 free_filespec(df);
7335 return 1;
7336 }
7337
7338 void setup_diff_pager(struct diff_options *opt)
7339 {
7340 /*
7341 * If the user asked for our exit code, then either they want --quiet
7342 * or --exit-code. We should definitely not bother with a pager in the
7343 * former case, as we will generate no output. Since we still properly
7344 * report our exit code even when a pager is run, we _could_ run a
7345 * pager with --exit-code. But since we have not done so historically,
7346 * and because it is easy to find people oneline advising "git diff
7347 * --exit-code" in hooks and other scripts, we do not do so.
7348 */
7349 if (!opt->flags.exit_with_status &&
7350 check_pager_config("diff") != 0)
7351 setup_pager();
7352 }
The Best VCS Around