?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
The eighth batch
[alt-git.git] / diff.c
blob0822ae443361f80271c91a863a6f03b0e8c403c3
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4
5 #define USE_THE_REPOSITORY_VARIABLE
6 #define DISABLE_SIGN_COMPARE_WARNINGS
7
8 #include "git-compat-util.h"
9 #include "abspath.h"
10 #include "base85.h"
11 #include "config.h"
12 #include "convert.h"
13 #include "environment.h"
14 #include "gettext.h"
15 #include "tempfile.h"
16 #include "revision.h"
17 #include "quote.h"
18 #include "diff.h"
19 #include "diffcore.h"
20 #include "delta.h"
21 #include "hex.h"
22 #include "xdiff-interface.h"
23 #include "color.h"
24 #include "run-command.h"
25 #include "utf8.h"
26 #include "object-store-ll.h"
27 #include "userdiff.h"
28 #include "submodule.h"
29 #include "hashmap.h"
30 #include "mem-pool.h"
31 #include "merge-ll.h"
32 #include "string-list.h"
33 #include "strvec.h"
34 #include "tmp-objdir.h"
35 #include "graph.h"
36 #include "oid-array.h"
37 #include "packfile.h"
38 #include "pager.h"
39 #include "parse-options.h"
40 #include "help.h"
41 #include "promisor-remote.h"
42 #include "dir.h"
43 #include "object-file.h"
44 #include "object-name.h"
45 #include "read-cache-ll.h"
46 #include "setup.h"
47 #include "strmap.h"
48 #include "ws.h"
49
50 #ifdef NO_FAST_WORKING_DIRECTORY
51 #define FAST_WORKING_DIRECTORY 0
52 #else
53 #define FAST_WORKING_DIRECTORY 1
54 #endif
55
56 static int diff_detect_rename_default;
57 static int diff_indent_heuristic = 1;
58 static int diff_rename_limit_default = 1000;
59 static int diff_suppress_blank_empty;
60 static int diff_use_color_default = -1;
61 static int diff_color_moved_default;
62 static int diff_color_moved_ws_default;
63 static int diff_context_default = 3;
64 static int diff_interhunk_context_default;
65 static char *diff_word_regex_cfg;
66 static struct external_diff external_diff_cfg;
67 static char *diff_order_file_cfg;
68 int diff_auto_refresh_index = 1;
69 static int diff_mnemonic_prefix;
70 static int diff_no_prefix;
71 static char *diff_src_prefix;
72 static char *diff_dst_prefix;
73 static int diff_relative;
74 static int diff_stat_name_width;
75 static int diff_stat_graph_width;
76 static int diff_dirstat_permille_default = 30;
77 static struct diff_options default_diff_options;
78 static long diff_algorithm;
79 static unsigned ws_error_highlight_default = WSEH_NEW;
80
81 static char diff_colors[][COLOR_MAXLEN] = {
82 GIT_COLOR_RESET,
83 GIT_COLOR_NORMAL, /* CONTEXT */
84 GIT_COLOR_BOLD, /* METAINFO */
85 GIT_COLOR_CYAN, /* FRAGINFO */
86 GIT_COLOR_RED, /* OLD */
87 GIT_COLOR_GREEN, /* NEW */
88 GIT_COLOR_YELLOW, /* COMMIT */
89 GIT_COLOR_BG_RED, /* WHITESPACE */
90 GIT_COLOR_NORMAL, /* FUNCINFO */
91 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
92 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
93 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
94 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
95 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
96 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
97 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
98 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
99 GIT_COLOR_FAINT, /* CONTEXT_DIM */
100 GIT_COLOR_FAINT_RED, /* OLD_DIM */
101 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
102 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
103 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
104 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
105 };
106
107 static const char *color_diff_slots[] = {
108 [DIFF_CONTEXT] = "context",
109 [DIFF_METAINFO] = "meta",
110 [DIFF_FRAGINFO] = "frag",
111 [DIFF_FILE_OLD] = "old",
112 [DIFF_FILE_NEW] = "new",
113 [DIFF_COMMIT] = "commit",
114 [DIFF_WHITESPACE] = "whitespace",
115 [DIFF_FUNCINFO] = "func",
116 [DIFF_FILE_OLD_MOVED] = "oldMoved",
117 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
118 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
119 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
120 [DIFF_FILE_NEW_MOVED] = "newMoved",
121 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
122 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
123 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
124 [DIFF_CONTEXT_DIM] = "contextDimmed",
125 [DIFF_FILE_OLD_DIM] = "oldDimmed",
126 [DIFF_FILE_NEW_DIM] = "newDimmed",
127 [DIFF_CONTEXT_BOLD] = "contextBold",
128 [DIFF_FILE_OLD_BOLD] = "oldBold",
129 [DIFF_FILE_NEW_BOLD] = "newBold",
130 };
131
132 define_list_config_array_extra(color_diff_slots, {"plain"});
133
134 static int parse_diff_color_slot(const char *var)
135 {
136 if (!strcasecmp(var, "plain"))
137 return DIFF_CONTEXT;
138 return LOOKUP_CONFIG(color_diff_slots, var);
139 }
140
141 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
142 struct strbuf *errmsg)
143 {
144 char *params_copy = xstrdup(params_string);
145 struct string_list params = STRING_LIST_INIT_NODUP;
146 int ret = 0;
147 int i;
148
149 if (*params_copy)
150 string_list_split_in_place(&params, params_copy, ",", -1);
151 for (i = 0; i < params.nr; i++) {
152 const char *p = params.items[i].string;
153 if (!strcmp(p, "changes")) {
154 options->flags.dirstat_by_line = 0;
155 options->flags.dirstat_by_file = 0;
156 } else if (!strcmp(p, "lines")) {
157 options->flags.dirstat_by_line = 1;
158 options->flags.dirstat_by_file = 0;
159 } else if (!strcmp(p, "files")) {
160 options->flags.dirstat_by_line = 0;
161 options->flags.dirstat_by_file = 1;
162 } else if (!strcmp(p, "noncumulative")) {
163 options->flags.dirstat_cumulative = 0;
164 } else if (!strcmp(p, "cumulative")) {
165 options->flags.dirstat_cumulative = 1;
166 } else if (isdigit(*p)) {
167 char *end;
168 int permille = strtoul(p, &end, 10) * 10;
169 if (*end == '.' && isdigit(*++end)) {
170 /* only use first digit */
171 permille += *end - '0';
172 /* .. and ignore any further digits */
173 while (isdigit(*++end))
174 ; /* nothing */
175 }
176 if (!*end)
177 options->dirstat_permille = permille;
178 else {
179 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
180 p);
181 ret++;
182 }
183 } else {
184 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
185 ret++;
186 }
187
188 }
189 string_list_clear(&params, 0);
190 free(params_copy);
191 return ret;
192 }
193
194 static int parse_submodule_params(struct diff_options *options, const char *value)
195 {
196 if (!strcmp(value, "log"))
197 options->submodule_format = DIFF_SUBMODULE_LOG;
198 else if (!strcmp(value, "short"))
199 options->submodule_format = DIFF_SUBMODULE_SHORT;
200 else if (!strcmp(value, "diff"))
201 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
202 /*
203 * Please update $__git_diff_submodule_formats in
204 * git-completion.bash when you add new formats.
205 */
206 else
207 return -1;
208 return 0;
209 }
210
211 int git_config_rename(const char *var, const char *value)
212 {
213 if (!value)
214 return DIFF_DETECT_RENAME;
215 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
216 return DIFF_DETECT_COPY;
217 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
218 }
219
220 long parse_algorithm_value(const char *value)
221 {
222 if (!value)
223 return -1;
224 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
225 return 0;
226 else if (!strcasecmp(value, "minimal"))
227 return XDF_NEED_MINIMAL;
228 else if (!strcasecmp(value, "patience"))
229 return XDF_PATIENCE_DIFF;
230 else if (!strcasecmp(value, "histogram"))
231 return XDF_HISTOGRAM_DIFF;
232 /*
233 * Please update $__git_diff_algorithms in git-completion.bash
234 * when you add new algorithms.
235 */
236 return -1;
237 }
238
239 static int parse_one_token(const char **arg, const char *token)
240 {
241 const char *rest;
242 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
243 *arg = rest;
244 return 1;
245 }
246 return 0;
247 }
248
249 static int parse_ws_error_highlight(const char *arg)
250 {
251 const char *orig_arg = arg;
252 unsigned val = 0;
253
254 while (*arg) {
255 if (parse_one_token(&arg, "none"))
256 val = 0;
257 else if (parse_one_token(&arg, "default"))
258 val = WSEH_NEW;
259 else if (parse_one_token(&arg, "all"))
260 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
261 else if (parse_one_token(&arg, "new"))
262 val |= WSEH_NEW;
263 else if (parse_one_token(&arg, "old"))
264 val |= WSEH_OLD;
265 else if (parse_one_token(&arg, "context"))
266 val |= WSEH_CONTEXT;
267 else {
268 return -1 - (int)(arg - orig_arg);
269 }
270 if (*arg)
271 arg++;
272 }
273 return val;
274 }
275
276 /*
277 * These are to give UI layer defaults.
278 * The core-level commands such as git-diff-files should
279 * never be affected by the setting of diff.renames
280 * the user happens to have in the configuration file.
281 */
282 void init_diff_ui_defaults(void)
283 {
284 diff_detect_rename_default = DIFF_DETECT_RENAME;
285 }
286
287 int git_diff_heuristic_config(const char *var, const char *value,
288 void *cb UNUSED)
289 {
290 if (!strcmp(var, "diff.indentheuristic"))
291 diff_indent_heuristic = git_config_bool(var, value);
292 return 0;
293 }
294
295 static int parse_color_moved(const char *arg)
296 {
297 switch (git_parse_maybe_bool(arg)) {
298 case 0:
299 return COLOR_MOVED_NO;
300 case 1:
301 return COLOR_MOVED_DEFAULT;
302 default:
303 break;
304 }
305
306 if (!strcmp(arg, "no"))
307 return COLOR_MOVED_NO;
308 else if (!strcmp(arg, "plain"))
309 return COLOR_MOVED_PLAIN;
310 else if (!strcmp(arg, "blocks"))
311 return COLOR_MOVED_BLOCKS;
312 else if (!strcmp(arg, "zebra"))
313 return COLOR_MOVED_ZEBRA;
314 else if (!strcmp(arg, "default"))
315 return COLOR_MOVED_DEFAULT;
316 else if (!strcmp(arg, "dimmed-zebra"))
317 return COLOR_MOVED_ZEBRA_DIM;
318 else if (!strcmp(arg, "dimmed_zebra"))
319 return COLOR_MOVED_ZEBRA_DIM;
320 else
321 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
322 }
323
324 static unsigned parse_color_moved_ws(const char *arg)
325 {
326 int ret = 0;
327 struct string_list l = STRING_LIST_INIT_DUP;
328 struct string_list_item *i;
329
330 string_list_split(&l, arg, ',', -1);
331
332 for_each_string_list_item(i, &l) {
333 struct strbuf sb = STRBUF_INIT;
334 strbuf_addstr(&sb, i->string);
335 strbuf_trim(&sb);
336
337 if (!strcmp(sb.buf, "no"))
338 ret = 0;
339 else if (!strcmp(sb.buf, "ignore-space-change"))
340 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
341 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
342 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
343 else if (!strcmp(sb.buf, "ignore-all-space"))
344 ret |= XDF_IGNORE_WHITESPACE;
345 else if (!strcmp(sb.buf, "allow-indentation-change"))
346 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
347 else {
348 ret |= COLOR_MOVED_WS_ERROR;
349 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);
350 }
351
352 strbuf_release(&sb);
353 }
354
355 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
356 (ret & XDF_WHITESPACE_FLAGS)) {
357 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
358 ret |= COLOR_MOVED_WS_ERROR;
359 }
360
361 string_list_clear(&l, 0);
362
363 return ret;
364 }
365
366 int git_diff_ui_config(const char *var, const char *value,
367 const struct config_context *ctx, void *cb)
368 {
369 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
370 diff_use_color_default = git_config_colorbool(var, value);
371 return 0;
372 }
373 if (!strcmp(var, "diff.colormoved")) {
374 int cm = parse_color_moved(value);
375 if (cm < 0)
376 return -1;
377 diff_color_moved_default = cm;
378 return 0;
379 }
380 if (!strcmp(var, "diff.colormovedws")) {
381 unsigned cm;
382 if (!value)
383 return config_error_nonbool(var);
384 cm = parse_color_moved_ws(value);
385 if (cm & COLOR_MOVED_WS_ERROR)
386 return -1;
387 diff_color_moved_ws_default = cm;
388 return 0;
389 }
390 if (!strcmp(var, "diff.context")) {
391 diff_context_default = git_config_int(var, value, ctx->kvi);
392 if (diff_context_default < 0)
393 return -1;
394 return 0;
395 }
396 if (!strcmp(var, "diff.interhunkcontext")) {
397 diff_interhunk_context_default = git_config_int(var, value,
398 ctx->kvi);
399 if (diff_interhunk_context_default < 0)
400 return -1;
401 return 0;
402 }
403 if (!strcmp(var, "diff.renames")) {
404 diff_detect_rename_default = git_config_rename(var, value);
405 return 0;
406 }
407 if (!strcmp(var, "diff.autorefreshindex")) {
408 diff_auto_refresh_index = git_config_bool(var, value);
409 return 0;
410 }
411 if (!strcmp(var, "diff.mnemonicprefix")) {
412 diff_mnemonic_prefix = git_config_bool(var, value);
413 return 0;
414 }
415 if (!strcmp(var, "diff.noprefix")) {
416 diff_no_prefix = git_config_bool(var, value);
417 return 0;
418 }
419 if (!strcmp(var, "diff.srcprefix")) {
420 FREE_AND_NULL(diff_src_prefix);
421 return git_config_string(&diff_src_prefix, var, value);
422 }
423 if (!strcmp(var, "diff.dstprefix")) {
424 FREE_AND_NULL(diff_dst_prefix);
425 return git_config_string(&diff_dst_prefix, var, value);
426 }
427 if (!strcmp(var, "diff.relative")) {
428 diff_relative = git_config_bool(var, value);
429 return 0;
430 }
431 if (!strcmp(var, "diff.statnamewidth")) {
432 diff_stat_name_width = git_config_int(var, value, ctx->kvi);
433 return 0;
434 }
435 if (!strcmp(var, "diff.statgraphwidth")) {
436 diff_stat_graph_width = git_config_int(var, value, ctx->kvi);
437 return 0;
438 }
439 if (!strcmp(var, "diff.external"))
440 return git_config_string(&external_diff_cfg.cmd, var, value);
441 if (!strcmp(var, "diff.trustexitcode")) {
442 external_diff_cfg.trust_exit_code = git_config_bool(var, value);
443 return 0;
444 }
445 if (!strcmp(var, "diff.wordregex"))
446 return git_config_string(&diff_word_regex_cfg, var, value);
447 if (!strcmp(var, "diff.orderfile")) {
448 FREE_AND_NULL(diff_order_file_cfg);
449 return git_config_pathname(&diff_order_file_cfg, var, value);
450 }
451
452 if (!strcmp(var, "diff.ignoresubmodules")) {
453 if (!value)
454 return config_error_nonbool(var);
455 handle_ignore_submodules_arg(&default_diff_options, value);
456 }
457
458 if (!strcmp(var, "diff.submodule")) {
459 if (!value)
460 return config_error_nonbool(var);
461 if (parse_submodule_params(&default_diff_options, value))
462 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
463 value);
464 return 0;
465 }
466
467 if (!strcmp(var, "diff.algorithm")) {
468 if (!value)
469 return config_error_nonbool(var);
470 diff_algorithm = parse_algorithm_value(value);
471 if (diff_algorithm < 0)
472 return error(_("unknown value for config '%s': %s"),
473 var, value);
474 return 0;
475 }
476
477 if (git_color_config(var, value, cb) < 0)
478 return -1;
479
480 return git_diff_basic_config(var, value, ctx, cb);
481 }
482
483 int git_diff_basic_config(const char *var, const char *value,
484 const struct config_context *ctx, void *cb)
485 {
486 const char *name;
487
488 if (!strcmp(var, "diff.renamelimit")) {
489 diff_rename_limit_default = git_config_int(var, value, ctx->kvi);
490 return 0;
491 }
492
493 if (userdiff_config(var, value) < 0)
494 return -1;
495
496 if (skip_prefix(var, "diff.color.", &name) ||
497 skip_prefix(var, "color.diff.", &name)) {
498 int slot = parse_diff_color_slot(name);
499 if (slot < 0)
500 return 0;
501 if (!value)
502 return config_error_nonbool(var);
503 return color_parse(value, diff_colors[slot]);
504 }
505
506 if (!strcmp(var, "diff.wserrorhighlight")) {
507 int val;
508 if (!value)
509 return config_error_nonbool(var);
510 val = parse_ws_error_highlight(value);
511 if (val < 0)
512 return error(_("unknown value for config '%s': %s"),
513 var, value);
514 ws_error_highlight_default = val;
515 return 0;
516 }
517
518 /* like GNU diff's --suppress-blank-empty option */
519 if (!strcmp(var, "diff.suppressblankempty") ||
520 /* for backwards compatibility */
521 !strcmp(var, "diff.suppress-blank-empty")) {
522 diff_suppress_blank_empty = git_config_bool(var, value);
523 return 0;
524 }
525
526 if (!strcmp(var, "diff.dirstat")) {
527 struct strbuf errmsg = STRBUF_INIT;
528 if (!value)
529 return config_error_nonbool(var);
530 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
531 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
532 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
533 errmsg.buf);
534 strbuf_release(&errmsg);
535 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
536 return 0;
537 }
538
539 if (git_diff_heuristic_config(var, value, cb) < 0)
540 return -1;
541
542 return git_default_config(var, value, ctx, cb);
543 }
544
545 static char *quote_two(const char *one, const char *two)
546 {
547 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
548 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
549 struct strbuf res = STRBUF_INIT;
550
551 if (need_one + need_two) {
552 strbuf_addch(&res, '"');
553 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
554 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
555 strbuf_addch(&res, '"');
556 } else {
557 strbuf_addstr(&res, one);
558 strbuf_addstr(&res, two);
559 }
560 return strbuf_detach(&res, NULL);
561 }
562
563 static const struct external_diff *external_diff(void)
564 {
565 static struct external_diff external_diff_env, *external_diff_ptr;
566 static int done_preparing = 0;
567
568 if (done_preparing)
569 return external_diff_ptr;
570 external_diff_env.cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
571 if (git_env_bool("GIT_EXTERNAL_DIFF_TRUST_EXIT_CODE", 0))
572 external_diff_env.trust_exit_code = 1;
573 if (external_diff_env.cmd)
574 external_diff_ptr = &external_diff_env;
575 else if (external_diff_cfg.cmd)
576 external_diff_ptr = &external_diff_cfg;
577 done_preparing = 1;
578 return external_diff_ptr;
579 }
580
581 /*
582 * Keep track of files used for diffing. Sometimes such an entry
583 * refers to a temporary file, sometimes to an existing file, and
584 * sometimes to "/dev/null".
585 */
586 static struct diff_tempfile {
587 /*
588 * filename external diff should read from, or NULL if this
589 * entry is currently not in use:
590 */
591 const char *name;
592
593 char hex[GIT_MAX_HEXSZ + 1];
594 char mode[10];
595
596 /*
597 * If this diff_tempfile instance refers to a temporary file,
598 * this tempfile object is used to manage its lifetime.
599 */
600 struct tempfile *tempfile;
601 } diff_temp[2];
602
603 struct emit_callback {
604 int color_diff;
605 unsigned ws_rule;
606 int blank_at_eof_in_preimage;
607 int blank_at_eof_in_postimage;
608 int lno_in_preimage;
609 int lno_in_postimage;
610 const char **label_path;
611 struct diff_words_data *diff_words;
612 struct diff_options *opt;
613 struct strbuf *header;
614 };
615
616 static int count_lines(const char *data, int size)
617 {
618 int count, ch, completely_empty = 1, nl_just_seen = 0;
619 count = 0;
620 while (0 < size--) {
621 ch = *data++;
622 if (ch == '\n') {
623 count++;
624 nl_just_seen = 1;
625 completely_empty = 0;
626 }
627 else {
628 nl_just_seen = 0;
629 completely_empty = 0;
630 }
631 }
632 if (completely_empty)
633 return 0;
634 if (!nl_just_seen)
635 count++; /* no trailing newline */
636 return count;
637 }
638
639 static int fill_mmfile(struct repository *r, mmfile_t *mf,
640 struct diff_filespec *one)
641 {
642 if (!DIFF_FILE_VALID(one)) {
643 mf->ptr = (char *)""; /* does not matter */
644 mf->size = 0;
645 return 0;
646 }
647 else if (diff_populate_filespec(r, one, NULL))
648 return -1;
649
650 mf->ptr = one->data;
651 mf->size = one->size;
652 return 0;
653 }
654
655 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
656 static unsigned long diff_filespec_size(struct repository *r,
657 struct diff_filespec *one)
658 {
659 struct diff_populate_filespec_options dpf_options = {
660 .check_size_only = 1,
661 };
662
663 if (!DIFF_FILE_VALID(one))
664 return 0;
665 diff_populate_filespec(r, one, &dpf_options);
666 return one->size;
667 }
668
669 static int count_trailing_blank(mmfile_t *mf)
670 {
671 char *ptr = mf->ptr;
672 long size = mf->size;
673 int cnt = 0;
674
675 if (!size)
676 return cnt;
677 ptr += size - 1; /* pointing at the very end */
678 if (*ptr != '\n')
679 ; /* incomplete line */
680 else
681 ptr--; /* skip the last LF */
682 while (mf->ptr < ptr) {
683 char *prev_eol;
684 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
685 if (*prev_eol == '\n')
686 break;
687 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
688 break;
689 cnt++;
690 ptr = prev_eol - 1;
691 }
692 return cnt;
693 }
694
695 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
696 struct emit_callback *ecbdata)
697 {
698 int l1, l2, at;
699 l1 = count_trailing_blank(mf1);
700 l2 = count_trailing_blank(mf2);
701 if (l2 <= l1) {
702 ecbdata->blank_at_eof_in_preimage = 0;
703 ecbdata->blank_at_eof_in_postimage = 0;
704 return;
705 }
706 at = count_lines(mf1->ptr, mf1->size);
707 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
708
709 at = count_lines(mf2->ptr, mf2->size);
710 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
711 }
712
713 static void emit_line_0(struct diff_options *o,
714 const char *set_sign, const char *set, unsigned reverse, const char *reset,
715 int first, const char *line, int len)
716 {
717 int has_trailing_newline, has_trailing_carriage_return;
718 int needs_reset = 0; /* at the end of the line */
719 FILE *file = o->file;
720
721 fputs(diff_line_prefix(o), file);
722
723 has_trailing_newline = (len > 0 && line[len-1] == '\n');
724 if (has_trailing_newline)
725 len--;
726
727 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
728 if (has_trailing_carriage_return)
729 len--;
730
731 if (!len && !first)
732 goto end_of_line;
733
734 if (reverse && want_color(o->use_color)) {
735 fputs(GIT_COLOR_REVERSE, file);
736 needs_reset = 1;
737 }
738
739 if (set_sign) {
740 fputs(set_sign, file);
741 needs_reset = 1;
742 }
743
744 if (first)
745 fputc(first, file);
746
747 if (!len)
748 goto end_of_line;
749
750 if (set) {
751 if (set_sign && set != set_sign)
752 fputs(reset, file);
753 fputs(set, file);
754 needs_reset = 1;
755 }
756 fwrite(line, len, 1, file);
757 needs_reset = 1; /* 'line' may contain color codes. */
758
759 end_of_line:
760 if (needs_reset)
761 fputs(reset, file);
762 if (has_trailing_carriage_return)
763 fputc('\r', file);
764 if (has_trailing_newline)
765 fputc('\n', file);
766 }
767
768 static void emit_line(struct diff_options *o, const char *set, const char *reset,
769 const char *line, int len)
770 {
771 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
772 }
773
774 enum diff_symbol {
775 DIFF_SYMBOL_BINARY_DIFF_HEADER,
776 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
777 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
778 DIFF_SYMBOL_BINARY_DIFF_BODY,
779 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
780 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
781 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
782 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
783 DIFF_SYMBOL_STATS_LINE,
784 DIFF_SYMBOL_WORD_DIFF,
785 DIFF_SYMBOL_STAT_SEP,
786 DIFF_SYMBOL_SUMMARY,
787 DIFF_SYMBOL_SUBMODULE_ADD,
788 DIFF_SYMBOL_SUBMODULE_DEL,
789 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
790 DIFF_SYMBOL_SUBMODULE_MODIFIED,
791 DIFF_SYMBOL_SUBMODULE_HEADER,
792 DIFF_SYMBOL_SUBMODULE_ERROR,
793 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
794 DIFF_SYMBOL_REWRITE_DIFF,
795 DIFF_SYMBOL_BINARY_FILES,
796 DIFF_SYMBOL_HEADER,
797 DIFF_SYMBOL_FILEPAIR_PLUS,
798 DIFF_SYMBOL_FILEPAIR_MINUS,
799 DIFF_SYMBOL_WORDS_PORCELAIN,
800 DIFF_SYMBOL_WORDS,
801 DIFF_SYMBOL_CONTEXT,
802 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
803 DIFF_SYMBOL_PLUS,
804 DIFF_SYMBOL_MINUS,
805 DIFF_SYMBOL_NO_LF_EOF,
806 DIFF_SYMBOL_CONTEXT_FRAGINFO,
807 DIFF_SYMBOL_CONTEXT_MARKER,
808 DIFF_SYMBOL_SEPARATOR
809 };
810 /*
811 * Flags for content lines:
812 * 0..12 are whitespace rules
813 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
814 * 16 is marking if the line is blank at EOF
815 */
816 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
817 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
818 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
819 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
820 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
821
822 /*
823 * This struct is used when we need to buffer the output of the diff output.
824 *
825 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
826 * into the pre/post image file. This pointer could be a union with the
827 * line pointer. By storing an offset into the file instead of the literal line,
828 * we can decrease the memory footprint for the buffered output. At first we
829 * may want to only have indirection for the content lines, but we could also
830 * enhance the state for emitting prefabricated lines, e.g. the similarity
831 * score line or hunk/file headers would only need to store a number or path
832 * and then the output can be constructed later on depending on state.
833 */
834 struct emitted_diff_symbol {
835 const char *line;
836 int len;
837 int flags;
838 int indent_off; /* Offset to first non-whitespace character */
839 int indent_width; /* The visual width of the indentation */
840 unsigned id;
841 enum diff_symbol s;
842 };
843 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
844
845 struct emitted_diff_symbols {
846 struct emitted_diff_symbol *buf;
847 int nr, alloc;
848 };
849 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
850
851 static void append_emitted_diff_symbol(struct diff_options *o,
852 struct emitted_diff_symbol *e)
853 {
854 struct emitted_diff_symbol *f;
855
856 ALLOC_GROW(o->emitted_symbols->buf,
857 o->emitted_symbols->nr + 1,
858 o->emitted_symbols->alloc);
859 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
860
861 memcpy(f, e, sizeof(struct emitted_diff_symbol));
862 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
863 }
864
865 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
866 {
867 if (!e)
868 return;
869 free(e->buf);
870 free(e);
871 }
872
873 struct moved_entry {
874 const struct emitted_diff_symbol *es;
875 struct moved_entry *next_line;
876 struct moved_entry *next_match;
877 };
878
879 struct moved_block {
880 struct moved_entry *match;
881 int wsd; /* The whitespace delta of this block */
882 };
883
884 #define INDENT_BLANKLINE INT_MIN
885
886 static void fill_es_indent_data(struct emitted_diff_symbol *es)
887 {
888 unsigned int off = 0, i;
889 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
890 const char *s = es->line;
891 const int len = es->len;
892
893 /* skip any \v \f \r at start of indentation */
894 while (s[off] == '\f' || s[off] == '\v' ||
895 (s[off] == '\r' && off < len - 1))
896 off++;
897
898 /* calculate the visual width of indentation */
899 while(1) {
900 if (s[off] == ' ') {
901 width++;
902 off++;
903 } else if (s[off] == '\t') {
904 width += tab_width - (width % tab_width);
905 while (s[++off] == '\t')
906 width += tab_width;
907 } else {
908 break;
909 }
910 }
911
912 /* check if this line is blank */
913 for (i = off; i < len; i++)
914 if (!isspace(s[i]))
915 break;
916
917 if (i == len) {
918 es->indent_width = INDENT_BLANKLINE;
919 es->indent_off = len;
920 } else {
921 es->indent_off = off;
922 es->indent_width = width;
923 }
924 }
925
926 static int compute_ws_delta(const struct emitted_diff_symbol *a,
927 const struct emitted_diff_symbol *b)
928 {
929 int a_width = a->indent_width,
930 b_width = b->indent_width;
931
932 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
933 return INDENT_BLANKLINE;
934
935 return a_width - b_width;
936 }
937
938 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
939 const struct emitted_diff_symbol *l,
940 struct moved_block *pmb)
941 {
942 int a_width = cur->es->indent_width, b_width = l->indent_width;
943 int delta;
944
945 /* The text of each line must match */
946 if (cur->es->id != l->id)
947 return 1;
948
949 /*
950 * If 'l' and 'cur' are both blank then we don't need to check the
951 * indent. We only need to check cur as we know the strings match.
952 * */
953 if (a_width == INDENT_BLANKLINE)
954 return 0;
955
956 /*
957 * The indent changes of the block are known and stored in pmb->wsd;
958 * however we need to check if the indent changes of the current line
959 * match those of the current block.
960 */
961 delta = b_width - a_width;
962
963 /*
964 * If the previous lines of this block were all blank then set its
965 * whitespace delta.
966 */
967 if (pmb->wsd == INDENT_BLANKLINE)
968 pmb->wsd = delta;
969
970 return delta != pmb->wsd;
971 }
972
973 struct interned_diff_symbol {
974 struct hashmap_entry ent;
975 struct emitted_diff_symbol *es;
976 };
977
978 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
979 const struct hashmap_entry *eptr,
980 const struct hashmap_entry *entry_or_key,
981 const void *keydata UNUSED)
982 {
983 const struct diff_options *diffopt = hashmap_cmp_fn_data;
984 const struct emitted_diff_symbol *a, *b;
985 unsigned flags = diffopt->color_moved_ws_handling
986 & XDF_WHITESPACE_FLAGS;
987
988 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
989 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
990
991 return !xdiff_compare_lines(a->line + a->indent_off,
992 a->len - a->indent_off,
993 b->line + b->indent_off,
994 b->len - b->indent_off, flags);
995 }
996
997 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
998 struct interned_diff_symbol *s)
999 {
1000 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
1001 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
1002 l->len - l->indent_off, flags);
1003
1004 hashmap_entry_init(&s->ent, hash);
1005 s->es = l;
1006 }
1007
1008 struct moved_entry_list {
1009 struct moved_entry *add, *del;
1010 };
1011
1012 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
1013 struct mem_pool *entry_mem_pool)
1014 {
1015 struct moved_entry *prev_line = NULL;
1016 struct mem_pool interned_pool;
1017 struct hashmap interned_map;
1018 struct moved_entry_list *entry_list = NULL;
1019 size_t entry_list_alloc = 0;
1020 unsigned id = 0;
1021 int n;
1022
1023 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
1024 mem_pool_init(&interned_pool, 1024 * 1024);
1025
1026 for (n = 0; n < o->emitted_symbols->nr; n++) {
1027 struct interned_diff_symbol key;
1028 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1029 struct interned_diff_symbol *s;
1030 struct moved_entry *entry;
1031
1032 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
1033 prev_line = NULL;
1034 continue;
1035 }
1036
1037 if (o->color_moved_ws_handling &
1038 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1039 fill_es_indent_data(l);
1040
1041 prepare_entry(o, l, &key);
1042 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
1043 if (s) {
1044 l->id = s->es->id;
1045 } else {
1046 l->id = id;
1047 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
1048 hashmap_add(&interned_map,
1049 memcpy(mem_pool_alloc(&interned_pool,
1050 sizeof(key)),
1051 &key, sizeof(key)));
1052 }
1053 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1054 entry->es = l;
1055 entry->next_line = NULL;
1056 if (prev_line && prev_line->es->s == l->s)
1057 prev_line->next_line = entry;
1058 prev_line = entry;
1059 if (l->s == DIFF_SYMBOL_PLUS) {
1060 entry->next_match = entry_list[l->id].add;
1061 entry_list[l->id].add = entry;
1062 } else {
1063 entry->next_match = entry_list[l->id].del;
1064 entry_list[l->id].del = entry;
1065 }
1066 }
1067
1068 hashmap_clear(&interned_map);
1069 mem_pool_discard(&interned_pool, 0);
1070
1071 return entry_list;
1072 }
1073
1074 static void pmb_advance_or_null(struct diff_options *o,
1075 struct emitted_diff_symbol *l,
1076 struct moved_block *pmb,
1077 int *pmb_nr)
1078 {
1079 int i, j;
1080
1081 for (i = 0, j = 0; i < *pmb_nr; i++) {
1082 int match;
1083 struct moved_entry *prev = pmb[i].match;
1084 struct moved_entry *cur = (prev && prev->next_line) ?
1085 prev->next_line : NULL;
1086
1087 if (o->color_moved_ws_handling &
1088 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1089 match = cur &&
1090 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1091 else
1092 match = cur && cur->es->id == l->id;
1093
1094 if (match) {
1095 pmb[j] = pmb[i];
1096 pmb[j++].match = cur;
1097 }
1098 }
1099 *pmb_nr = j;
1100 }
1101
1102 static void fill_potential_moved_blocks(struct diff_options *o,
1103 struct moved_entry *match,
1104 struct emitted_diff_symbol *l,
1105 struct moved_block **pmb_p,
1106 int *pmb_alloc_p, int *pmb_nr_p)
1107
1108 {
1109 struct moved_block *pmb = *pmb_p;
1110 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1111
1112 /*
1113 * The current line is the start of a new block.
1114 * Setup the set of potential blocks.
1115 */
1116 for (; match; match = match->next_match) {
1117 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1118 if (o->color_moved_ws_handling &
1119 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1120 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1121 else
1122 pmb[pmb_nr].wsd = 0;
1123 pmb[pmb_nr++].match = match;
1124 }
1125
1126 *pmb_p = pmb;
1127 *pmb_alloc_p = pmb_alloc;
1128 *pmb_nr_p = pmb_nr;
1129 }
1130
1131 /*
1132 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1133 *
1134 * Otherwise, if the last block has fewer alphanumeric characters than
1135 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1136 * that block.
1137 *
1138 * The last block consists of the (n - block_length)'th line up to but not
1139 * including the nth line.
1140 *
1141 * Returns 0 if the last block is empty or is unset by this function, non zero
1142 * otherwise.
1143 *
1144 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1145 * Think of a way to unify them.
1146 */
1147 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1148 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1149 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1150 {
1151 int i, alnum_count = 0;
1152 if (o->color_moved == COLOR_MOVED_PLAIN)
1153 return block_length;
1154 for (i = 1; i < block_length + 1; i++) {
1155 const char *c = o->emitted_symbols->buf[n - i].line;
1156 for (; *c; c++) {
1157 if (!isalnum(*c))
1158 continue;
1159 alnum_count++;
1160 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1161 return 1;
1162 }
1163 }
1164 for (i = 1; i < block_length + 1; i++)
1165 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1166 return 0;
1167 }
1168
1169 /* Find blocks of moved code, delegate actual coloring decision to helper */
1170 static void mark_color_as_moved(struct diff_options *o,
1171 struct moved_entry_list *entry_list)
1172 {
1173 struct moved_block *pmb = NULL; /* potentially moved blocks */
1174 int pmb_nr = 0, pmb_alloc = 0;
1175 int n, flipped_block = 0, block_length = 0;
1176 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1177
1178
1179 for (n = 0; n < o->emitted_symbols->nr; n++) {
1180 struct moved_entry *match = NULL;
1181 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1182
1183 switch (l->s) {
1184 case DIFF_SYMBOL_PLUS:
1185 match = entry_list[l->id].del;
1186 break;
1187 case DIFF_SYMBOL_MINUS:
1188 match = entry_list[l->id].add;
1189 break;
1190 default:
1191 flipped_block = 0;
1192 }
1193
1194 if (pmb_nr && (!match || l->s != moved_symbol)) {
1195 if (!adjust_last_block(o, n, block_length) &&
1196 block_length > 1) {
1197 /*
1198 * Rewind in case there is another match
1199 * starting at the second line of the block
1200 */
1201 match = NULL;
1202 n -= block_length;
1203 }
1204 pmb_nr = 0;
1205 block_length = 0;
1206 flipped_block = 0;
1207 }
1208 if (!match) {
1209 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1210 continue;
1211 }
1212
1213 if (o->color_moved == COLOR_MOVED_PLAIN) {
1214 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1215 continue;
1216 }
1217
1218 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1219
1220 if (pmb_nr == 0) {
1221 int contiguous = adjust_last_block(o, n, block_length);
1222
1223 if (!contiguous && block_length > 1)
1224 /*
1225 * Rewind in case there is another match
1226 * starting at the second line of the block
1227 */
1228 n -= block_length;
1229 else
1230 fill_potential_moved_blocks(o, match, l,
1231 &pmb, &pmb_alloc,
1232 &pmb_nr);
1233
1234 if (contiguous && pmb_nr && moved_symbol == l->s)
1235 flipped_block = (flipped_block + 1) % 2;
1236 else
1237 flipped_block = 0;
1238
1239 if (pmb_nr)
1240 moved_symbol = l->s;
1241 else
1242 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1243
1244 block_length = 0;
1245 }
1246
1247 if (pmb_nr) {
1248 block_length++;
1249 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1250 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1251 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1252 }
1253 }
1254 adjust_last_block(o, n, block_length);
1255
1256 free(pmb);
1257 }
1258
1259 static void dim_moved_lines(struct diff_options *o)
1260 {
1261 int n;
1262 for (n = 0; n < o->emitted_symbols->nr; n++) {
1263 struct emitted_diff_symbol *prev = (n != 0) ?
1264 &o->emitted_symbols->buf[n - 1] : NULL;
1265 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1266 struct emitted_diff_symbol *next =
1267 (n < o->emitted_symbols->nr - 1) ?
1268 &o->emitted_symbols->buf[n + 1] : NULL;
1269
1270 /* Not a plus or minus line? */
1271 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1272 continue;
1273
1274 /* Not a moved line? */
1275 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1276 continue;
1277
1278 /*
1279 * If prev or next are not a plus or minus line,
1280 * pretend they don't exist
1281 */
1282 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1283 prev->s != DIFF_SYMBOL_MINUS)
1284 prev = NULL;
1285 if (next && next->s != DIFF_SYMBOL_PLUS &&
1286 next->s != DIFF_SYMBOL_MINUS)
1287 next = NULL;
1288
1289 /* Inside a block? */
1290 if ((prev &&
1291 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1292 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1293 (next &&
1294 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1295 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1296 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1297 continue;
1298 }
1299
1300 /* Check if we are at an interesting bound: */
1301 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1302 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1303 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1304 continue;
1305 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1306 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1307 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1308 continue;
1309
1310 /*
1311 * The boundary to prev and next are not interesting,
1312 * so this line is not interesting as a whole
1313 */
1314 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1315 }
1316 }
1317
1318 static void emit_line_ws_markup(struct diff_options *o,
1319 const char *set_sign, const char *set,
1320 const char *reset,
1321 int sign_index, const char *line, int len,
1322 unsigned ws_rule, int blank_at_eof)
1323 {
1324 const char *ws = NULL;
1325 int sign = o->output_indicators[sign_index];
1326
1327 if (o->ws_error_highlight & ws_rule) {
1328 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1329 if (!*ws)
1330 ws = NULL;
1331 }
1332
1333 if (!ws && !set_sign)
1334 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1335 else if (!ws) {
1336 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1337 } else if (blank_at_eof)
1338 /* Blank line at EOF - paint '+' as well */
1339 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1340 else {
1341 /* Emit just the prefix, then the rest. */
1342 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1343 sign, "", 0);
1344 ws_check_emit(line, len, ws_rule,
1345 o->file, set, reset, ws);
1346 }
1347 }
1348
1349 static void emit_diff_symbol_from_struct(struct diff_options *o,
1350 struct emitted_diff_symbol *eds)
1351 {
1352 static const char *nneof = " No newline at end of file\n";
1353 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1354
1355 enum diff_symbol s = eds->s;
1356 const char *line = eds->line;
1357 int len = eds->len;
1358 unsigned flags = eds->flags;
1359
1360 switch (s) {
1361 case DIFF_SYMBOL_NO_LF_EOF:
1362 context = diff_get_color_opt(o, DIFF_CONTEXT);
1363 reset = diff_get_color_opt(o, DIFF_RESET);
1364 putc('\n', o->file);
1365 emit_line_0(o, context, NULL, 0, reset, '\\',
1366 nneof, strlen(nneof));
1367 break;
1368 case DIFF_SYMBOL_SUBMODULE_HEADER:
1369 case DIFF_SYMBOL_SUBMODULE_ERROR:
1370 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1371 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1372 case DIFF_SYMBOL_SUMMARY:
1373 case DIFF_SYMBOL_STATS_LINE:
1374 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1375 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1376 emit_line(o, "", "", line, len);
1377 break;
1378 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1379 case DIFF_SYMBOL_CONTEXT_MARKER:
1380 context = diff_get_color_opt(o, DIFF_CONTEXT);
1381 reset = diff_get_color_opt(o, DIFF_RESET);
1382 emit_line(o, context, reset, line, len);
1383 break;
1384 case DIFF_SYMBOL_SEPARATOR:
1385 fprintf(o->file, "%s%c",
1386 diff_line_prefix(o),
1387 o->line_termination);
1388 break;
1389 case DIFF_SYMBOL_CONTEXT:
1390 set = diff_get_color_opt(o, DIFF_CONTEXT);
1391 reset = diff_get_color_opt(o, DIFF_RESET);
1392 set_sign = NULL;
1393 if (o->flags.dual_color_diffed_diffs) {
1394 char c = !len ? 0 : line[0];
1395
1396 if (c == '+')
1397 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1398 else if (c == '@')
1399 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1400 else if (c == '-')
1401 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1402 }
1403 emit_line_ws_markup(o, set_sign, set, reset,
1404 OUTPUT_INDICATOR_CONTEXT, line, len,
1405 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1406 break;
1407 case DIFF_SYMBOL_PLUS:
1408 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1409 DIFF_SYMBOL_MOVED_LINE_ALT |
1410 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1411 case DIFF_SYMBOL_MOVED_LINE |
1412 DIFF_SYMBOL_MOVED_LINE_ALT |
1413 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1414 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1415 break;
1416 case DIFF_SYMBOL_MOVED_LINE |
1417 DIFF_SYMBOL_MOVED_LINE_ALT:
1418 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1419 break;
1420 case DIFF_SYMBOL_MOVED_LINE |
1421 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1422 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1423 break;
1424 case DIFF_SYMBOL_MOVED_LINE:
1425 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1426 break;
1427 default:
1428 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1429 }
1430 reset = diff_get_color_opt(o, DIFF_RESET);
1431 if (!o->flags.dual_color_diffed_diffs)
1432 set_sign = NULL;
1433 else {
1434 char c = !len ? 0 : line[0];
1435
1436 set_sign = set;
1437 if (c == '-')
1438 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1439 else if (c == '@')
1440 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1441 else if (c == '+')
1442 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1443 else
1444 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1445 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1446 }
1447 emit_line_ws_markup(o, set_sign, set, reset,
1448 OUTPUT_INDICATOR_NEW, line, len,
1449 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1450 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1451 break;
1452 case DIFF_SYMBOL_MINUS:
1453 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1454 DIFF_SYMBOL_MOVED_LINE_ALT |
1455 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1456 case DIFF_SYMBOL_MOVED_LINE |
1457 DIFF_SYMBOL_MOVED_LINE_ALT |
1458 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1459 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1460 break;
1461 case DIFF_SYMBOL_MOVED_LINE |
1462 DIFF_SYMBOL_MOVED_LINE_ALT:
1463 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1464 break;
1465 case DIFF_SYMBOL_MOVED_LINE |
1466 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1467 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1468 break;
1469 case DIFF_SYMBOL_MOVED_LINE:
1470 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1471 break;
1472 default:
1473 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1474 }
1475 reset = diff_get_color_opt(o, DIFF_RESET);
1476 if (!o->flags.dual_color_diffed_diffs)
1477 set_sign = NULL;
1478 else {
1479 char c = !len ? 0 : line[0];
1480
1481 set_sign = set;
1482 if (c == '+')
1483 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1484 else if (c == '@')
1485 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1486 else if (c == '-')
1487 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1488 else
1489 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1490 }
1491 emit_line_ws_markup(o, set_sign, set, reset,
1492 OUTPUT_INDICATOR_OLD, line, len,
1493 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1494 break;
1495 case DIFF_SYMBOL_WORDS_PORCELAIN:
1496 context = diff_get_color_opt(o, DIFF_CONTEXT);
1497 reset = diff_get_color_opt(o, DIFF_RESET);
1498 emit_line(o, context, reset, line, len);
1499 fputs("~\n", o->file);
1500 break;
1501 case DIFF_SYMBOL_WORDS:
1502 context = diff_get_color_opt(o, DIFF_CONTEXT);
1503 reset = diff_get_color_opt(o, DIFF_RESET);
1504 /*
1505 * Skip the prefix character, if any. With
1506 * diff_suppress_blank_empty, there may be
1507 * none.
1508 */
1509 if (line[0] != '\n') {
1510 line++;
1511 len--;
1512 }
1513 emit_line(o, context, reset, line, len);
1514 break;
1515 case DIFF_SYMBOL_FILEPAIR_PLUS:
1516 meta = diff_get_color_opt(o, DIFF_METAINFO);
1517 reset = diff_get_color_opt(o, DIFF_RESET);
1518 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1519 line, reset,
1520 strchr(line, ' ') ? "\t" : "");
1521 break;
1522 case DIFF_SYMBOL_FILEPAIR_MINUS:
1523 meta = diff_get_color_opt(o, DIFF_METAINFO);
1524 reset = diff_get_color_opt(o, DIFF_RESET);
1525 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1526 line, reset,
1527 strchr(line, ' ') ? "\t" : "");
1528 break;
1529 case DIFF_SYMBOL_BINARY_FILES:
1530 case DIFF_SYMBOL_HEADER:
1531 fprintf(o->file, "%s", line);
1532 break;
1533 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1534 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1535 break;
1536 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1537 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1538 break;
1539 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1540 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1541 break;
1542 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1543 fputs(diff_line_prefix(o), o->file);
1544 fputc('\n', o->file);
1545 break;
1546 case DIFF_SYMBOL_REWRITE_DIFF:
1547 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1548 reset = diff_get_color_opt(o, DIFF_RESET);
1549 emit_line(o, fraginfo, reset, line, len);
1550 break;
1551 case DIFF_SYMBOL_SUBMODULE_ADD:
1552 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1553 reset = diff_get_color_opt(o, DIFF_RESET);
1554 emit_line(o, set, reset, line, len);
1555 break;
1556 case DIFF_SYMBOL_SUBMODULE_DEL:
1557 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1558 reset = diff_get_color_opt(o, DIFF_RESET);
1559 emit_line(o, set, reset, line, len);
1560 break;
1561 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1562 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1563 diff_line_prefix(o), line);
1564 break;
1565 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1566 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1567 diff_line_prefix(o), line);
1568 break;
1569 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1570 emit_line(o, "", "", " 0 files changed\n",
1571 strlen(" 0 files changed\n"));
1572 break;
1573 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1574 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1575 break;
1576 case DIFF_SYMBOL_WORD_DIFF:
1577 fprintf(o->file, "%.*s", len, line);
1578 break;
1579 case DIFF_SYMBOL_STAT_SEP:
1580 fputs(o->stat_sep, o->file);
1581 break;
1582 default:
1583 BUG("unknown diff symbol");
1584 }
1585 }
1586
1587 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1588 const char *line, int len, unsigned flags)
1589 {
1590 struct emitted_diff_symbol e = {
1591 .line = line, .len = len, .flags = flags, .s = s
1592 };
1593
1594 if (o->emitted_symbols)
1595 append_emitted_diff_symbol(o, &e);
1596 else
1597 emit_diff_symbol_from_struct(o, &e);
1598 }
1599
1600 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1601 {
1602 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1603 }
1604
1605 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1606 {
1607 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1608 }
1609
1610 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1611 {
1612 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1613 path, strlen(path), 0);
1614 }
1615
1616 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1617 {
1618 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1619 path, strlen(path), 0);
1620 }
1621
1622 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1623 {
1624 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1625 header, strlen(header), 0);
1626 }
1627
1628 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1629 {
1630 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1631 }
1632
1633 void diff_emit_submodule_pipethrough(struct diff_options *o,
1634 const char *line, int len)
1635 {
1636 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1637 }
1638
1639 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1640 {
1641 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1642 ecbdata->blank_at_eof_in_preimage &&
1643 ecbdata->blank_at_eof_in_postimage &&
1644 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1645 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1646 return 0;
1647 return ws_blank_line(line, len);
1648 }
1649
1650 static void emit_add_line(struct emit_callback *ecbdata,
1651 const char *line, int len)
1652 {
1653 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1654 if (new_blank_line_at_eof(ecbdata, line, len))
1655 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1656
1657 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1658 }
1659
1660 static void emit_del_line(struct emit_callback *ecbdata,
1661 const char *line, int len)
1662 {
1663 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1664 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1665 }
1666
1667 static void emit_context_line(struct emit_callback *ecbdata,
1668 const char *line, int len)
1669 {
1670 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1671 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1672 }
1673
1674 static void emit_hunk_header(struct emit_callback *ecbdata,
1675 const char *line, int len)
1676 {
1677 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1678 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1679 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1680 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1681 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1682 static const char atat[2] = { '@', '@' };
1683 const char *cp, *ep;
1684 struct strbuf msgbuf = STRBUF_INIT;
1685 int org_len = len;
1686 int i = 1;
1687
1688 /*
1689 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1690 * it always is at least 10 bytes long.
1691 */
1692 if (len < 10 ||
1693 memcmp(line, atat, 2) ||
1694 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1695 emit_diff_symbol(ecbdata->opt,
1696 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1697 return;
1698 }
1699 ep += 2; /* skip over @@ */
1700
1701 /* The hunk header in fraginfo color */
1702 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1703 strbuf_addstr(&msgbuf, reverse);
1704 strbuf_addstr(&msgbuf, frag);
1705 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1706 strbuf_add(&msgbuf, atat, sizeof(atat));
1707 else
1708 strbuf_add(&msgbuf, line, ep - line);
1709 strbuf_addstr(&msgbuf, reset);
1710
1711 /*
1712 * trailing "\r\n"
1713 */
1714 for ( ; i < 3; i++)
1715 if (line[len - i] == '\r' || line[len - i] == '\n')
1716 len--;
1717
1718 /* blank before the func header */
1719 for (cp = ep; ep - line < len; ep++)
1720 if (*ep != ' ' && *ep != '\t')
1721 break;
1722 if (ep != cp) {
1723 strbuf_addstr(&msgbuf, context);
1724 strbuf_add(&msgbuf, cp, ep - cp);
1725 strbuf_addstr(&msgbuf, reset);
1726 }
1727
1728 if (ep < line + len) {
1729 strbuf_addstr(&msgbuf, func);
1730 strbuf_add(&msgbuf, ep, line + len - ep);
1731 strbuf_addstr(&msgbuf, reset);
1732 }
1733
1734 strbuf_add(&msgbuf, line + len, org_len - len);
1735 strbuf_complete_line(&msgbuf);
1736 emit_diff_symbol(ecbdata->opt,
1737 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1738 strbuf_release(&msgbuf);
1739 }
1740
1741 static struct diff_tempfile *claim_diff_tempfile(void)
1742 {
1743 int i;
1744 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1745 if (!diff_temp[i].name)
1746 return diff_temp + i;
1747 BUG("diff is failing to clean up its tempfiles");
1748 }
1749
1750 static void remove_tempfile(void)
1751 {
1752 int i;
1753 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1754 if (is_tempfile_active(diff_temp[i].tempfile))
1755 delete_tempfile(&diff_temp[i].tempfile);
1756 diff_temp[i].name = NULL;
1757 }
1758 }
1759
1760 static void add_line_count(struct strbuf *out, int count)
1761 {
1762 switch (count) {
1763 case 0:
1764 strbuf_addstr(out, "0,0");
1765 break;
1766 case 1:
1767 strbuf_addstr(out, "1");
1768 break;
1769 default:
1770 strbuf_addf(out, "1,%d", count);
1771 break;
1772 }
1773 }
1774
1775 static void emit_rewrite_lines(struct emit_callback *ecb,
1776 int prefix, const char *data, int size)
1777 {
1778 const char *endp = NULL;
1779
1780 while (0 < size) {
1781 int len;
1782
1783 endp = memchr(data, '\n', size);
1784 len = endp ? (endp - data + 1) : size;
1785 if (prefix != '+') {
1786 ecb->lno_in_preimage++;
1787 emit_del_line(ecb, data, len);
1788 } else {
1789 ecb->lno_in_postimage++;
1790 emit_add_line(ecb, data, len);
1791 }
1792 size -= len;
1793 data += len;
1794 }
1795 if (!endp)
1796 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1797 }
1798
1799 static void emit_rewrite_diff(const char *name_a,
1800 const char *name_b,
1801 struct diff_filespec *one,
1802 struct diff_filespec *two,
1803 struct userdiff_driver *textconv_one,
1804 struct userdiff_driver *textconv_two,
1805 struct diff_options *o)
1806 {
1807 int lc_a, lc_b;
1808 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1809 const char *a_prefix, *b_prefix;
1810 char *data_one, *data_two;
1811 size_t size_one, size_two;
1812 struct emit_callback ecbdata;
1813 struct strbuf out = STRBUF_INIT;
1814
1815 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1816 a_prefix = o->b_prefix;
1817 b_prefix = o->a_prefix;
1818 } else {
1819 a_prefix = o->a_prefix;
1820 b_prefix = o->b_prefix;
1821 }
1822
1823 name_a += (*name_a == '/');
1824 name_b += (*name_b == '/');
1825
1826 strbuf_reset(&a_name);
1827 strbuf_reset(&b_name);
1828 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1829 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1830
1831 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1832 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1833
1834 memset(&ecbdata, 0, sizeof(ecbdata));
1835 ecbdata.color_diff = want_color(o->use_color);
1836 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1837 ecbdata.opt = o;
1838 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1839 mmfile_t mf1, mf2;
1840 mf1.ptr = (char *)data_one;
1841 mf2.ptr = (char *)data_two;
1842 mf1.size = size_one;
1843 mf2.size = size_two;
1844 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1845 }
1846 ecbdata.lno_in_preimage = 1;
1847 ecbdata.lno_in_postimage = 1;
1848
1849 lc_a = count_lines(data_one, size_one);
1850 lc_b = count_lines(data_two, size_two);
1851
1852 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1853 a_name.buf, a_name.len, 0);
1854 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1855 b_name.buf, b_name.len, 0);
1856
1857 strbuf_addstr(&out, "@@ -");
1858 if (!o->irreversible_delete)
1859 add_line_count(&out, lc_a);
1860 else
1861 strbuf_addstr(&out, "?,?");
1862 strbuf_addstr(&out, " +");
1863 add_line_count(&out, lc_b);
1864 strbuf_addstr(&out, " @@\n");
1865 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1866 strbuf_release(&out);
1867
1868 if (lc_a && !o->irreversible_delete)
1869 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1870 if (lc_b)
1871 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1872 if (textconv_one)
1873 free((char *)data_one);
1874 if (textconv_two)
1875 free((char *)data_two);
1876 }
1877
1878 struct diff_words_buffer {
1879 mmfile_t text;
1880 unsigned long alloc;
1881 struct diff_words_orig {
1882 const char *begin, *end;
1883 } *orig;
1884 int orig_nr, orig_alloc;
1885 };
1886
1887 static void diff_words_append(char *line, unsigned long len,
1888 struct diff_words_buffer *buffer)
1889 {
1890 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1891 line++;
1892 len--;
1893 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1894 buffer->text.size += len;
1895 buffer->text.ptr[buffer->text.size] = '\0';
1896 }
1897
1898 struct diff_words_style_elem {
1899 const char *prefix;
1900 const char *suffix;
1901 const char *color; /* NULL; filled in by the setup code if
1902 * color is enabled */
1903 };
1904
1905 struct diff_words_style {
1906 enum diff_words_type type;
1907 struct diff_words_style_elem new_word, old_word, ctx;
1908 const char *newline;
1909 };
1910
1911 static struct diff_words_style diff_words_styles[] = {
1912 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1913 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1914 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1915 };
1916
1917 struct diff_words_data {
1918 struct diff_words_buffer minus, plus;
1919 const char *current_plus;
1920 int last_minus;
1921 struct diff_options *opt;
1922 regex_t *word_regex;
1923 enum diff_words_type type;
1924 struct diff_words_style *style;
1925 };
1926
1927 static int fn_out_diff_words_write_helper(struct diff_options *o,
1928 struct diff_words_style_elem *st_el,
1929 const char *newline,
1930 size_t count, const char *buf)
1931 {
1932 int print = 0;
1933 struct strbuf sb = STRBUF_INIT;
1934
1935 while (count) {
1936 char *p = memchr(buf, '\n', count);
1937 if (print)
1938 strbuf_addstr(&sb, diff_line_prefix(o));
1939
1940 if (p != buf) {
1941 const char *reset = st_el->color && *st_el->color ?
1942 GIT_COLOR_RESET : NULL;
1943 if (st_el->color && *st_el->color)
1944 strbuf_addstr(&sb, st_el->color);
1945 strbuf_addstr(&sb, st_el->prefix);
1946 strbuf_add(&sb, buf, p ? p - buf : count);
1947 strbuf_addstr(&sb, st_el->suffix);
1948 if (reset)
1949 strbuf_addstr(&sb, reset);
1950 }
1951 if (!p)
1952 goto out;
1953
1954 strbuf_addstr(&sb, newline);
1955 count -= p + 1 - buf;
1956 buf = p + 1;
1957 print = 1;
1958 if (count) {
1959 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1960 sb.buf, sb.len, 0);
1961 strbuf_reset(&sb);
1962 }
1963 }
1964
1965 out:
1966 if (sb.len)
1967 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1968 sb.buf, sb.len, 0);
1969 strbuf_release(&sb);
1970 return 0;
1971 }
1972
1973 /*
1974 * '--color-words' algorithm can be described as:
1975 *
1976 * 1. collect the minus/plus lines of a diff hunk, divided into
1977 * minus-lines and plus-lines;
1978 *
1979 * 2. break both minus-lines and plus-lines into words and
1980 * place them into two mmfile_t with one word for each line;
1981 *
1982 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1983 *
1984 * And for the common parts of the both file, we output the plus side text.
1985 * diff_words->current_plus is used to trace the current position of the plus file
1986 * which printed. diff_words->last_minus is used to trace the last minus word
1987 * printed.
1988 *
1989 * For '--graph' to work with '--color-words', we need to output the graph prefix
1990 * on each line of color words output. Generally, there are two conditions on
1991 * which we should output the prefix.
1992 *
1993 * 1. diff_words->last_minus == 0 &&
1994 * diff_words->current_plus == diff_words->plus.text.ptr
1995 *
1996 * that is: the plus text must start as a new line, and if there is no minus
1997 * word printed, a graph prefix must be printed.
1998 *
1999 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
2000 * *(diff_words->current_plus - 1) == '\n'
2001 *
2002 * that is: a graph prefix must be printed following a '\n'
2003 */
2004 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
2005 {
2006 if ((diff_words->last_minus == 0 &&
2007 diff_words->current_plus == diff_words->plus.text.ptr) ||
2008 (diff_words->current_plus > diff_words->plus.text.ptr &&
2009 *(diff_words->current_plus - 1) == '\n')) {
2010 return 1;
2011 } else {
2012 return 0;
2013 }
2014 }
2015
2016 static void fn_out_diff_words_aux(void *priv,
2017 long minus_first, long minus_len,
2018 long plus_first, long plus_len,
2019 const char *func UNUSED, long funclen UNUSED)
2020 {
2021 struct diff_words_data *diff_words = priv;
2022 struct diff_words_style *style = diff_words->style;
2023 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
2024 struct diff_options *opt = diff_words->opt;
2025 const char *line_prefix;
2026
2027 assert(opt);
2028 line_prefix = diff_line_prefix(opt);
2029
2030 /* POSIX requires that first be decremented by one if len == 0... */
2031 if (minus_len) {
2032 minus_begin = diff_words->minus.orig[minus_first].begin;
2033 minus_end =
2034 diff_words->minus.orig[minus_first + minus_len - 1].end;
2035 } else
2036 minus_begin = minus_end =
2037 diff_words->minus.orig[minus_first].end;
2038
2039 if (plus_len) {
2040 plus_begin = diff_words->plus.orig[plus_first].begin;
2041 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2042 } else
2043 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2044
2045 if (color_words_output_graph_prefix(diff_words)) {
2046 fputs(line_prefix, diff_words->opt->file);
2047 }
2048 if (diff_words->current_plus != plus_begin) {
2049 fn_out_diff_words_write_helper(diff_words->opt,
2050 &style->ctx, style->newline,
2051 plus_begin - diff_words->current_plus,
2052 diff_words->current_plus);
2053 }
2054 if (minus_begin != minus_end) {
2055 fn_out_diff_words_write_helper(diff_words->opt,
2056 &style->old_word, style->newline,
2057 minus_end - minus_begin, minus_begin);
2058 }
2059 if (plus_begin != plus_end) {
2060 fn_out_diff_words_write_helper(diff_words->opt,
2061 &style->new_word, style->newline,
2062 plus_end - plus_begin, plus_begin);
2063 }
2064
2065 diff_words->current_plus = plus_end;
2066 diff_words->last_minus = minus_first;
2067 }
2068
2069 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2070 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2071 int *begin, int *end)
2072 {
2073 while (word_regex && *begin < buffer->size) {
2074 regmatch_t match[1];
2075 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2076 buffer->size - *begin, 1, match, 0)) {
2077 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2078 '\n', match[0].rm_eo - match[0].rm_so);
2079 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2080 *begin += match[0].rm_so;
2081 if (*begin == *end)
2082 (*begin)++;
2083 else
2084 return *begin > *end;
2085 } else {
2086 return -1;
2087 }
2088 }
2089
2090 /* find the next word */
2091 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2092 (*begin)++;
2093 if (*begin >= buffer->size)
2094 return -1;
2095
2096 /* find the end of the word */
2097 *end = *begin + 1;
2098 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2099 (*end)++;
2100
2101 return 0;
2102 }
2103
2104 /*
2105 * This function splits the words in buffer->text, stores the list with
2106 * newline separator into out, and saves the offsets of the original words
2107 * in buffer->orig.
2108 */
2109 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2110 regex_t *word_regex)
2111 {
2112 int i, j;
2113 long alloc = 0;
2114
2115 out->size = 0;
2116 out->ptr = NULL;
2117
2118 /* fake an empty "0th" word */
2119 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2120 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2121 buffer->orig_nr = 1;
2122
2123 for (i = 0; i < buffer->text.size; i++) {
2124 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2125 return;
2126
2127 /* store original boundaries */
2128 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2129 buffer->orig_alloc);
2130 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2131 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2132 buffer->orig_nr++;
2133
2134 /* store one word */
2135 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2136 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2137 out->ptr[out->size + j - i] = '\n';
2138 out->size += j - i + 1;
2139
2140 i = j - 1;
2141 }
2142 }
2143
2144 /* this executes the word diff on the accumulated buffers */
2145 static void diff_words_show(struct diff_words_data *diff_words)
2146 {
2147 xpparam_t xpp;
2148 xdemitconf_t xecfg;
2149 mmfile_t minus, plus;
2150 struct diff_words_style *style = diff_words->style;
2151
2152 struct diff_options *opt = diff_words->opt;
2153 const char *line_prefix;
2154
2155 assert(opt);
2156 line_prefix = diff_line_prefix(opt);
2157
2158 /* special case: only removal */
2159 if (!diff_words->plus.text.size) {
2160 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2161 line_prefix, strlen(line_prefix), 0);
2162 fn_out_diff_words_write_helper(diff_words->opt,
2163 &style->old_word, style->newline,
2164 diff_words->minus.text.size,
2165 diff_words->minus.text.ptr);
2166 diff_words->minus.text.size = 0;
2167 return;
2168 }
2169
2170 diff_words->current_plus = diff_words->plus.text.ptr;
2171 diff_words->last_minus = 0;
2172
2173 memset(&xpp, 0, sizeof(xpp));
2174 memset(&xecfg, 0, sizeof(xecfg));
2175 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2176 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2177 xpp.flags = 0;
2178 /* as only the hunk header will be parsed, we need a 0-context */
2179 xecfg.ctxlen = 0;
2180 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2181 diff_words, &xpp, &xecfg))
2182 die("unable to generate word diff");
2183 free(minus.ptr);
2184 free(plus.ptr);
2185 if (diff_words->current_plus != diff_words->plus.text.ptr +
2186 diff_words->plus.text.size) {
2187 if (color_words_output_graph_prefix(diff_words))
2188 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2189 line_prefix, strlen(line_prefix), 0);
2190 fn_out_diff_words_write_helper(diff_words->opt,
2191 &style->ctx, style->newline,
2192 diff_words->plus.text.ptr + diff_words->plus.text.size
2193 - diff_words->current_plus, diff_words->current_plus);
2194 }
2195 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2196 }
2197
2198 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2199 static void diff_words_flush(struct emit_callback *ecbdata)
2200 {
2201 struct diff_options *wo = ecbdata->diff_words->opt;
2202
2203 if (ecbdata->diff_words->minus.text.size ||
2204 ecbdata->diff_words->plus.text.size)
2205 diff_words_show(ecbdata->diff_words);
2206
2207 if (wo->emitted_symbols) {
2208 struct diff_options *o = ecbdata->opt;
2209 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2210 int i;
2211
2212 /*
2213 * NEEDSWORK:
2214 * Instead of appending each, concat all words to a line?
2215 */
2216 for (i = 0; i < wol->nr; i++)
2217 append_emitted_diff_symbol(o, &wol->buf[i]);
2218
2219 for (i = 0; i < wol->nr; i++)
2220 free((void *)wol->buf[i].line);
2221
2222 wol->nr = 0;
2223 }
2224 }
2225
2226 static void diff_filespec_load_driver(struct diff_filespec *one,
2227 struct index_state *istate)
2228 {
2229 /* Use already-loaded driver */
2230 if (one->driver)
2231 return;
2232
2233 if (S_ISREG(one->mode))
2234 one->driver = userdiff_find_by_path(istate, one->path);
2235
2236 /* Fallback to default settings */
2237 if (!one->driver)
2238 one->driver = userdiff_find_by_name("default");
2239 }
2240
2241 static const char *userdiff_word_regex(struct diff_filespec *one,
2242 struct index_state *istate)
2243 {
2244 diff_filespec_load_driver(one, istate);
2245 return one->driver->word_regex;
2246 }
2247
2248 static void init_diff_words_data(struct emit_callback *ecbdata,
2249 struct diff_options *orig_opts,
2250 struct diff_filespec *one,
2251 struct diff_filespec *two)
2252 {
2253 int i;
2254 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2255 memcpy(o, orig_opts, sizeof(struct diff_options));
2256
2257 CALLOC_ARRAY(ecbdata->diff_words, 1);
2258 ecbdata->diff_words->type = o->word_diff;
2259 ecbdata->diff_words->opt = o;
2260
2261 if (orig_opts->emitted_symbols)
2262 CALLOC_ARRAY(o->emitted_symbols, 1);
2263
2264 if (!o->word_regex)
2265 o->word_regex = userdiff_word_regex(one, o->repo->index);
2266 if (!o->word_regex)
2267 o->word_regex = userdiff_word_regex(two, o->repo->index);
2268 if (!o->word_regex)
2269 o->word_regex = diff_word_regex_cfg;
2270 if (o->word_regex) {
2271 ecbdata->diff_words->word_regex = (regex_t *)
2272 xmalloc(sizeof(regex_t));
2273 if (regcomp(ecbdata->diff_words->word_regex,
2274 o->word_regex,
2275 REG_EXTENDED | REG_NEWLINE))
2276 die("invalid regular expression: %s",
2277 o->word_regex);
2278 }
2279 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2280 if (o->word_diff == diff_words_styles[i].type) {
2281 ecbdata->diff_words->style =
2282 &diff_words_styles[i];
2283 break;
2284 }
2285 }
2286 if (want_color(o->use_color)) {
2287 struct diff_words_style *st = ecbdata->diff_words->style;
2288 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2289 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2290 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2291 }
2292 }
2293
2294 static void free_diff_words_data(struct emit_callback *ecbdata)
2295 {
2296 if (ecbdata->diff_words) {
2297 diff_words_flush(ecbdata);
2298 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2299 free (ecbdata->diff_words->opt);
2300 free (ecbdata->diff_words->minus.text.ptr);
2301 free (ecbdata->diff_words->minus.orig);
2302 free (ecbdata->diff_words->plus.text.ptr);
2303 free (ecbdata->diff_words->plus.orig);
2304 if (ecbdata->diff_words->word_regex) {
2305 regfree(ecbdata->diff_words->word_regex);
2306 free(ecbdata->diff_words->word_regex);
2307 }
2308 FREE_AND_NULL(ecbdata->diff_words);
2309 }
2310 }
2311
2312 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2313 {
2314 if (want_color(diff_use_color))
2315 return diff_colors[ix];
2316 return "";
2317 }
2318
2319 const char *diff_line_prefix(struct diff_options *opt)
2320 {
2321 return opt->output_prefix ?
2322 opt->output_prefix(opt, opt->output_prefix_data) :
2323 "";
2324 }
2325
2326 static unsigned long sane_truncate_line(char *line, unsigned long len)
2327 {
2328 const char *cp;
2329 unsigned long allot;
2330 size_t l = len;
2331
2332 cp = line;
2333 allot = l;
2334 while (0 < l) {
2335 (void) utf8_width(&cp, &l);
2336 if (!cp)
2337 break; /* truncated in the middle? */
2338 }
2339 return allot - l;
2340 }
2341
2342 static void find_lno(const char *line, struct emit_callback *ecbdata)
2343 {
2344 const char *p;
2345 ecbdata->lno_in_preimage = 0;
2346 ecbdata->lno_in_postimage = 0;
2347 p = strchr(line, '-');
2348 if (!p)
2349 return; /* cannot happen */
2350 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2351 p = strchr(p, '+');
2352 if (!p)
2353 return; /* cannot happen */
2354 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2355 }
2356
2357 static int fn_out_consume(void *priv, char *line, unsigned long len)
2358 {
2359 struct emit_callback *ecbdata = priv;
2360 struct diff_options *o = ecbdata->opt;
2361
2362 o->found_changes = 1;
2363
2364 if (ecbdata->header) {
2365 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2366 ecbdata->header->buf, ecbdata->header->len, 0);
2367 strbuf_reset(ecbdata->header);
2368 ecbdata->header = NULL;
2369 }
2370
2371 if (ecbdata->label_path[0]) {
2372 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2373 ecbdata->label_path[0],
2374 strlen(ecbdata->label_path[0]), 0);
2375 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2376 ecbdata->label_path[1],
2377 strlen(ecbdata->label_path[1]), 0);
2378 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2379 }
2380
2381 if (diff_suppress_blank_empty
2382 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2383 line[0] = '\n';
2384 len = 1;
2385 }
2386
2387 if (line[0] == '@') {
2388 if (ecbdata->diff_words)
2389 diff_words_flush(ecbdata);
2390 len = sane_truncate_line(line, len);
2391 find_lno(line, ecbdata);
2392 emit_hunk_header(ecbdata, line, len);
2393 return 0;
2394 }
2395
2396 if (ecbdata->diff_words) {
2397 enum diff_symbol s =
2398 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2399 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2400 if (line[0] == '-') {
2401 diff_words_append(line, len,
2402 &ecbdata->diff_words->minus);
2403 return 0;
2404 } else if (line[0] == '+') {
2405 diff_words_append(line, len,
2406 &ecbdata->diff_words->plus);
2407 return 0;
2408 } else if (starts_with(line, "\\ ")) {
2409 /*
2410 * Eat the "no newline at eof" marker as if we
2411 * saw a "+" or "-" line with nothing on it,
2412 * and return without diff_words_flush() to
2413 * defer processing. If this is the end of
2414 * preimage, more "+" lines may come after it.
2415 */
2416 return 0;
2417 }
2418 diff_words_flush(ecbdata);
2419 emit_diff_symbol(o, s, line, len, 0);
2420 return 0;
2421 }
2422
2423 switch (line[0]) {
2424 case '+':
2425 ecbdata->lno_in_postimage++;
2426 emit_add_line(ecbdata, line + 1, len - 1);
2427 break;
2428 case '-':
2429 ecbdata->lno_in_preimage++;
2430 emit_del_line(ecbdata, line + 1, len - 1);
2431 break;
2432 case ' ':
2433 ecbdata->lno_in_postimage++;
2434 ecbdata->lno_in_preimage++;
2435 emit_context_line(ecbdata, line + 1, len - 1);
2436 break;
2437 default:
2438 /* incomplete line at the end */
2439 ecbdata->lno_in_preimage++;
2440 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2441 line, len, 0);
2442 break;
2443 }
2444 return 0;
2445 }
2446
2447 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2448 {
2449 const char *old_name = a;
2450 const char *new_name = b;
2451 int pfx_length, sfx_length;
2452 int pfx_adjust_for_slash;
2453 int len_a = strlen(a);
2454 int len_b = strlen(b);
2455 int a_midlen, b_midlen;
2456 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2457 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2458
2459 if (qlen_a || qlen_b) {
2460 quote_c_style(a, name, NULL, 0);
2461 strbuf_addstr(name, " => ");
2462 quote_c_style(b, name, NULL, 0);
2463 return;
2464 }
2465
2466 /* Find common prefix */
2467 pfx_length = 0;
2468 while (*old_name && *new_name && *old_name == *new_name) {
2469 if (*old_name == '/')
2470 pfx_length = old_name - a + 1;
2471 old_name++;
2472 new_name++;
2473 }
2474
2475 /* Find common suffix */
2476 old_name = a + len_a;
2477 new_name = b + len_b;
2478 sfx_length = 0;
2479 /*
2480 * If there is a common prefix, it must end in a slash. In
2481 * that case we let this loop run 1 into the prefix to see the
2482 * same slash.
2483 *
2484 * If there is no common prefix, we cannot do this as it would
2485 * underrun the input strings.
2486 */
2487 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2488 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2489 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2490 *old_name == *new_name) {
2491 if (*old_name == '/')
2492 sfx_length = len_a - (old_name - a);
2493 old_name--;
2494 new_name--;
2495 }
2496
2497 /*
2498 * pfx{mid-a => mid-b}sfx
2499 * {pfx-a => pfx-b}sfx
2500 * pfx{sfx-a => sfx-b}
2501 * name-a => name-b
2502 */
2503 a_midlen = len_a - pfx_length - sfx_length;
2504 b_midlen = len_b - pfx_length - sfx_length;
2505 if (a_midlen < 0)
2506 a_midlen = 0;
2507 if (b_midlen < 0)
2508 b_midlen = 0;
2509
2510 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2511 if (pfx_length + sfx_length) {
2512 strbuf_add(name, a, pfx_length);
2513 strbuf_addch(name, '{');
2514 }
2515 strbuf_add(name, a + pfx_length, a_midlen);
2516 strbuf_addstr(name, " => ");
2517 strbuf_add(name, b + pfx_length, b_midlen);
2518 if (pfx_length + sfx_length) {
2519 strbuf_addch(name, '}');
2520 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2521 }
2522 }
2523
2524 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2525 const char *name_a,
2526 const char *name_b)
2527 {
2528 struct diffstat_file *x;
2529 CALLOC_ARRAY(x, 1);
2530 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2531 diffstat->files[diffstat->nr++] = x;
2532 if (name_b) {
2533 x->from_name = xstrdup(name_a);
2534 x->name = xstrdup(name_b);
2535 x->is_renamed = 1;
2536 }
2537 else {
2538 x->from_name = NULL;
2539 x->name = xstrdup(name_a);
2540 }
2541 return x;
2542 }
2543
2544 static int diffstat_consume(void *priv, char *line, unsigned long len)
2545 {
2546 struct diffstat_t *diffstat = priv;
2547 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2548
2549 if (!len)
2550 BUG("xdiff fed us an empty line");
2551
2552 if (line[0] == '+')
2553 x->added++;
2554 else if (line[0] == '-')
2555 x->deleted++;
2556 return 0;
2557 }
2558
2559 const char mime_boundary_leader[] = "------------";
2560
2561 static int scale_linear(int it, int width, int max_change)
2562 {
2563 if (!it)
2564 return 0;
2565 /*
2566 * make sure that at least one '-' or '+' is printed if
2567 * there is any change to this path. The easiest way is to
2568 * scale linearly as if the allotted width is one column shorter
2569 * than it is, and then add 1 to the result.
2570 */
2571 return 1 + (it * (width - 1) / max_change);
2572 }
2573
2574 static void show_graph(struct strbuf *out, char ch, int cnt,
2575 const char *set, const char *reset)
2576 {
2577 if (cnt <= 0)
2578 return;
2579 strbuf_addstr(out, set);
2580 strbuf_addchars(out, ch, cnt);
2581 strbuf_addstr(out, reset);
2582 }
2583
2584 static void fill_print_name(struct diffstat_file *file)
2585 {
2586 struct strbuf pname = STRBUF_INIT;
2587
2588 if (file->print_name)
2589 return;
2590
2591 if (file->is_renamed)
2592 pprint_rename(&pname, file->from_name, file->name);
2593 else
2594 quote_c_style(file->name, &pname, NULL, 0);
2595
2596 if (file->comments)
2597 strbuf_addf(&pname, " (%s)", file->comments);
2598
2599 file->print_name = strbuf_detach(&pname, NULL);
2600 }
2601
2602 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2603 int files, int insertions, int deletions)
2604 {
2605 struct strbuf sb = STRBUF_INIT;
2606
2607 if (!files) {
2608 assert(insertions == 0 && deletions == 0);
2609 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2610 NULL, 0, 0);
2611 return;
2612 }
2613
2614 strbuf_addf(&sb,
2615 (files == 1) ? " %d file changed" : " %d files changed",
2616 files);
2617
2618 /*
2619 * For binary diff, the caller may want to print "x files
2620 * changed" with insertions == 0 && deletions == 0.
2621 *
2622 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2623 * is probably less confusing (i.e skip over "2 files changed
2624 * but nothing about added/removed lines? Is this a bug in Git?").
2625 */
2626 if (insertions || deletions == 0) {
2627 strbuf_addf(&sb,
2628 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2629 insertions);
2630 }
2631
2632 if (deletions || insertions == 0) {
2633 strbuf_addf(&sb,
2634 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2635 deletions);
2636 }
2637 strbuf_addch(&sb, '\n');
2638 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2639 sb.buf, sb.len, 0);
2640 strbuf_release(&sb);
2641 }
2642
2643 void print_stat_summary(FILE *fp, int files,
2644 int insertions, int deletions)
2645 {
2646 struct diff_options o;
2647 memset(&o, 0, sizeof(o));
2648 o.file = fp;
2649
2650 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2651 }
2652
2653 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2654 {
2655 int i, len, add, del, adds = 0, dels = 0;
2656 uintmax_t max_change = 0, max_len = 0;
2657 int total_files = data->nr, count;
2658 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2659 const char *reset, *add_c, *del_c;
2660 int extra_shown = 0;
2661 const char *line_prefix = diff_line_prefix(options);
2662 struct strbuf out = STRBUF_INIT;
2663
2664 if (data->nr == 0)
2665 return;
2666
2667 count = options->stat_count ? options->stat_count : data->nr;
2668
2669 reset = diff_get_color_opt(options, DIFF_RESET);
2670 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2671 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2672
2673 /*
2674 * Find the longest filename and max number of changes
2675 */
2676 for (i = 0; (i < count) && (i < data->nr); i++) {
2677 struct diffstat_file *file = data->files[i];
2678 uintmax_t change = file->added + file->deleted;
2679
2680 if (!file->is_interesting && (change == 0)) {
2681 count++; /* not shown == room for one more */
2682 continue;
2683 }
2684 fill_print_name(file);
2685 len = utf8_strwidth(file->print_name);
2686 if (max_len < len)
2687 max_len = len;
2688
2689 if (file->is_unmerged) {
2690 /* "Unmerged" is 8 characters */
2691 bin_width = bin_width < 8 ? 8 : bin_width;
2692 continue;
2693 }
2694 if (file->is_binary) {
2695 /* "Bin XXX -> YYY bytes" */
2696 int w = 14 + decimal_width(file->added)
2697 + decimal_width(file->deleted);
2698 bin_width = bin_width < w ? w : bin_width;
2699 /* Display change counts aligned with "Bin" */
2700 number_width = 3;
2701 continue;
2702 }
2703
2704 if (max_change < change)
2705 max_change = change;
2706 }
2707 count = i; /* where we can stop scanning in data->files[] */
2708
2709 /*
2710 * We have width = stat_width or term_columns() columns total.
2711 * We want a maximum of min(max_len, stat_name_width) for the name part.
2712 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2713 * We also need 1 for " " and 4 + decimal_width(max_change)
2714 * for " | NNNN " and one the empty column at the end, altogether
2715 * 6 + decimal_width(max_change).
2716 *
2717 * If there's not enough space, we will use the smaller of
2718 * stat_name_width (if set) and 5/8*width for the filename,
2719 * and the rest for constant elements + graph part, but no more
2720 * than stat_graph_width for the graph part.
2721 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2722 * for the standard terminal size).
2723 *
2724 * In other words: stat_width limits the maximum width, and
2725 * stat_name_width fixes the maximum width of the filename,
2726 * and is also used to divide available columns if there
2727 * aren't enough.
2728 *
2729 * Binary files are displayed with "Bin XXX -> YYY bytes"
2730 * instead of the change count and graph. This part is treated
2731 * similarly to the graph part, except that it is not
2732 * "scaled". If total width is too small to accommodate the
2733 * guaranteed minimum width of the filename part and the
2734 * separators and this message, this message will "overflow"
2735 * making the line longer than the maximum width.
2736 */
2737
2738 /*
2739 * NEEDSWORK: line_prefix is often used for "log --graph" output
2740 * and contains ANSI-colored string. utf8_strnwidth() should be
2741 * used to correctly count the display width instead of strlen().
2742 */
2743 if (options->stat_width == -1)
2744 width = term_columns() - strlen(line_prefix);
2745 else
2746 width = options->stat_width ? options->stat_width : 80;
2747 number_width = decimal_width(max_change) > number_width ?
2748 decimal_width(max_change) : number_width;
2749
2750 if (options->stat_name_width == -1)
2751 options->stat_name_width = diff_stat_name_width;
2752 if (options->stat_graph_width == -1)
2753 options->stat_graph_width = diff_stat_graph_width;
2754
2755 /*
2756 * Guarantee 3/8*16 == 6 for the graph part
2757 * and 5/8*16 == 10 for the filename part
2758 */
2759 if (width < 16 + 6 + number_width)
2760 width = 16 + 6 + number_width;
2761
2762 /*
2763 * First assign sizes that are wanted, ignoring available width.
2764 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2765 * starting from "XXX" should fit in graph_width.
2766 */
2767 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2768 if (options->stat_graph_width &&
2769 options->stat_graph_width < graph_width)
2770 graph_width = options->stat_graph_width;
2771
2772 name_width = (options->stat_name_width > 0 &&
2773 options->stat_name_width < max_len) ?
2774 options->stat_name_width : max_len;
2775
2776 /*
2777 * Adjust adjustable widths not to exceed maximum width
2778 */
2779 if (name_width + number_width + 6 + graph_width > width) {
2780 if (graph_width > width * 3/8 - number_width - 6) {
2781 graph_width = width * 3/8 - number_width - 6;
2782 if (graph_width < 6)
2783 graph_width = 6;
2784 }
2785
2786 if (options->stat_graph_width &&
2787 graph_width > options->stat_graph_width)
2788 graph_width = options->stat_graph_width;
2789 if (name_width > width - number_width - 6 - graph_width)
2790 name_width = width - number_width - 6 - graph_width;
2791 else
2792 graph_width = width - number_width - 6 - name_width;
2793 }
2794
2795 /*
2796 * From here name_width is the width of the name area,
2797 * and graph_width is the width of the graph area.
2798 * max_change is used to scale graph properly.
2799 */
2800 for (i = 0; i < count; i++) {
2801 const char *prefix = "";
2802 struct diffstat_file *file = data->files[i];
2803 char *name = file->print_name;
2804 uintmax_t added = file->added;
2805 uintmax_t deleted = file->deleted;
2806 int name_len, padding;
2807
2808 if (!file->is_interesting && (added + deleted == 0))
2809 continue;
2810
2811 /*
2812 * "scale" the filename
2813 */
2814 len = name_width;
2815 name_len = utf8_strwidth(name);
2816 if (name_width < name_len) {
2817 char *slash;
2818 prefix = "...";
2819 len -= 3;
2820 /*
2821 * NEEDSWORK: (name_len - len) counts the display
2822 * width, which would be shorter than the byte
2823 * length of the corresponding substring.
2824 * Advancing "name" by that number of bytes does
2825 * *NOT* skip over that many columns, so it is
2826 * very likely that chomping the pathname at the
2827 * slash we will find starting from "name" will
2828 * leave the resulting string still too long.
2829 */
2830 name += name_len - len;
2831 slash = strchr(name, '/');
2832 if (slash)
2833 name = slash;
2834 }
2835 padding = len - utf8_strwidth(name);
2836 if (padding < 0)
2837 padding = 0;
2838
2839 if (file->is_binary) {
2840 strbuf_addf(&out, " %s%s%*s | %*s",
2841 prefix, name, padding, "",
2842 number_width, "Bin");
2843 if (!added && !deleted) {
2844 strbuf_addch(&out, '\n');
2845 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2846 out.buf, out.len, 0);
2847 strbuf_reset(&out);
2848 continue;
2849 }
2850 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2851 del_c, deleted, reset);
2852 strbuf_addstr(&out, " -> ");
2853 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2854 add_c, added, reset);
2855 strbuf_addstr(&out, " bytes\n");
2856 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2857 out.buf, out.len, 0);
2858 strbuf_reset(&out);
2859 continue;
2860 }
2861 else if (file->is_unmerged) {
2862 strbuf_addf(&out, " %s%s%*s | %*s",
2863 prefix, name, padding, "",
2864 number_width, "Unmerged\n");
2865 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2866 out.buf, out.len, 0);
2867 strbuf_reset(&out);
2868 continue;
2869 }
2870
2871 /*
2872 * scale the add/delete
2873 */
2874 add = added;
2875 del = deleted;
2876
2877 if (graph_width <= max_change) {
2878 int total = scale_linear(add + del, graph_width, max_change);
2879 if (total < 2 && add && del)
2880 /* width >= 2 due to the sanity check */
2881 total = 2;
2882 if (add < del) {
2883 add = scale_linear(add, graph_width, max_change);
2884 del = total - add;
2885 } else {
2886 del = scale_linear(del, graph_width, max_change);
2887 add = total - del;
2888 }
2889 }
2890 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2891 prefix, name, padding, "",
2892 number_width, added + deleted,
2893 added + deleted ? " " : "");
2894 show_graph(&out, '+', add, add_c, reset);
2895 show_graph(&out, '-', del, del_c, reset);
2896 strbuf_addch(&out, '\n');
2897 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2898 out.buf, out.len, 0);
2899 strbuf_reset(&out);
2900 }
2901
2902 for (i = 0; i < data->nr; i++) {
2903 struct diffstat_file *file = data->files[i];
2904 uintmax_t added = file->added;
2905 uintmax_t deleted = file->deleted;
2906
2907 if (file->is_unmerged ||
2908 (!file->is_interesting && (added + deleted == 0))) {
2909 total_files--;
2910 continue;
2911 }
2912
2913 if (!file->is_binary) {
2914 adds += added;
2915 dels += deleted;
2916 }
2917 if (i < count)
2918 continue;
2919 if (!extra_shown)
2920 emit_diff_symbol(options,
2921 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2922 NULL, 0, 0);
2923 extra_shown = 1;
2924 }
2925
2926 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2927 strbuf_release(&out);
2928 }
2929
2930 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2931 {
2932 int i, adds = 0, dels = 0, total_files = data->nr;
2933
2934 if (data->nr == 0)
2935 return;
2936
2937 for (i = 0; i < data->nr; i++) {
2938 int added = data->files[i]->added;
2939 int deleted = data->files[i]->deleted;
2940
2941 if (data->files[i]->is_unmerged ||
2942 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2943 total_files--;
2944 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2945 adds += added;
2946 dels += deleted;
2947 }
2948 }
2949 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2950 }
2951
2952 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2953 {
2954 int i;
2955
2956 if (data->nr == 0)
2957 return;
2958
2959 for (i = 0; i < data->nr; i++) {
2960 struct diffstat_file *file = data->files[i];
2961
2962 fprintf(options->file, "%s", diff_line_prefix(options));
2963
2964 if (file->is_binary)
2965 fprintf(options->file, "-\t-\t");
2966 else
2967 fprintf(options->file,
2968 "%"PRIuMAX"\t%"PRIuMAX"\t",
2969 file->added, file->deleted);
2970 if (options->line_termination) {
2971 fill_print_name(file);
2972 if (!file->is_renamed)
2973 write_name_quoted(file->name, options->file,
2974 options->line_termination);
2975 else {
2976 fputs(file->print_name, options->file);
2977 putc(options->line_termination, options->file);
2978 }
2979 } else {
2980 if (file->is_renamed) {
2981 putc('\0', options->file);
2982 write_name_quoted(file->from_name, options->file, '\0');
2983 }
2984 write_name_quoted(file->name, options->file, '\0');
2985 }
2986 }
2987 }
2988
2989 struct dirstat_file {
2990 const char *name;
2991 unsigned long changed;
2992 };
2993
2994 struct dirstat_dir {
2995 struct dirstat_file *files;
2996 int alloc, nr, permille, cumulative;
2997 };
2998
2999 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
3000 unsigned long changed, const char *base, int baselen)
3001 {
3002 unsigned long sum_changes = 0;
3003 unsigned int sources = 0;
3004 const char *line_prefix = diff_line_prefix(opt);
3005
3006 while (dir->nr) {
3007 struct dirstat_file *f = dir->files;
3008 int namelen = strlen(f->name);
3009 unsigned long changes;
3010 char *slash;
3011
3012 if (namelen < baselen)
3013 break;
3014 if (memcmp(f->name, base, baselen))
3015 break;
3016 slash = strchr(f->name + baselen, '/');
3017 if (slash) {
3018 int newbaselen = slash + 1 - f->name;
3019 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
3020 sources++;
3021 } else {
3022 changes = f->changed;
3023 dir->files++;
3024 dir->nr--;
3025 sources += 2;
3026 }
3027 sum_changes += changes;
3028 }
3029
3030 /*
3031 * We don't report dirstat's for
3032 * - the top level
3033 * - or cases where everything came from a single directory
3034 * under this directory (sources == 1).
3035 */
3036 if (baselen && sources != 1) {
3037 if (sum_changes) {
3038 int permille = sum_changes * 1000 / changed;
3039 if (permille >= dir->permille) {
3040 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
3041 permille / 10, permille % 10, baselen, base);
3042 if (!dir->cumulative)
3043 return 0;
3044 }
3045 }
3046 }
3047 return sum_changes;
3048 }
3049
3050 static int dirstat_compare(const void *_a, const void *_b)
3051 {
3052 const struct dirstat_file *a = _a;
3053 const struct dirstat_file *b = _b;
3054 return strcmp(a->name, b->name);
3055 }
3056
3057 static void conclude_dirstat(struct diff_options *options,
3058 struct dirstat_dir *dir,
3059 unsigned long changed)
3060 {
3061 struct dirstat_file *to_free = dir->files;
3062
3063 if (!changed) {
3064 /* This can happen even with many files, if everything was renames */
3065 ;
3066 } else {
3067 /* Show all directories with more than x% of the changes */
3068 QSORT(dir->files, dir->nr, dirstat_compare);
3069 gather_dirstat(options, dir, changed, "", 0);
3070 }
3071
3072 free(to_free);
3073 }
3074
3075 static void show_dirstat(struct diff_options *options)
3076 {
3077 int i;
3078 unsigned long changed;
3079 struct dirstat_dir dir;
3080 struct diff_queue_struct *q = &diff_queued_diff;
3081
3082 dir.files = NULL;
3083 dir.alloc = 0;
3084 dir.nr = 0;
3085 dir.permille = options->dirstat_permille;
3086 dir.cumulative = options->flags.dirstat_cumulative;
3087
3088 changed = 0;
3089 for (i = 0; i < q->nr; i++) {
3090 struct diff_filepair *p = q->queue[i];
3091 const char *name;
3092 unsigned long copied, added, damage;
3093 struct diff_populate_filespec_options dpf_options = {
3094 .check_size_only = 1,
3095 };
3096
3097 name = p->two->path ? p->two->path : p->one->path;
3098
3099 if (p->one->oid_valid && p->two->oid_valid &&
3100 oideq(&p->one->oid, &p->two->oid)) {
3101 /*
3102 * The SHA1 has not changed, so pre-/post-content is
3103 * identical. We can therefore skip looking at the
3104 * file contents altogether.
3105 */
3106 damage = 0;
3107 goto found_damage;
3108 }
3109
3110 if (options->flags.dirstat_by_file) {
3111 /*
3112 * In --dirstat-by-file mode, we don't really need to
3113 * look at the actual file contents at all.
3114 * The fact that the SHA1 changed is enough for us to
3115 * add this file to the list of results
3116 * (with each file contributing equal damage).
3117 */
3118 damage = 1;
3119 goto found_damage;
3120 }
3121
3122 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3123 diff_populate_filespec(options->repo, p->one, NULL);
3124 diff_populate_filespec(options->repo, p->two, NULL);
3125 diffcore_count_changes(options->repo,
3126 p->one, p->two, NULL, NULL,
3127 &copied, &added);
3128 diff_free_filespec_data(p->one);
3129 diff_free_filespec_data(p->two);
3130 } else if (DIFF_FILE_VALID(p->one)) {
3131 diff_populate_filespec(options->repo, p->one, &dpf_options);
3132 copied = added = 0;
3133 diff_free_filespec_data(p->one);
3134 } else if (DIFF_FILE_VALID(p->two)) {
3135 diff_populate_filespec(options->repo, p->two, &dpf_options);
3136 copied = 0;
3137 added = p->two->size;
3138 diff_free_filespec_data(p->two);
3139 } else
3140 continue;
3141
3142 /*
3143 * Original minus copied is the removed material,
3144 * added is the new material. They are both damages
3145 * made to the preimage.
3146 * If the resulting damage is zero, we know that
3147 * diffcore_count_changes() considers the two entries to
3148 * be identical, but since the oid changed, we
3149 * know that there must have been _some_ kind of change,
3150 * so we force all entries to have damage > 0.
3151 */
3152 damage = (p->one->size - copied) + added;
3153 if (!damage)
3154 damage = 1;
3155
3156 found_damage:
3157 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3158 dir.files[dir.nr].name = name;
3159 dir.files[dir.nr].changed = damage;
3160 changed += damage;
3161 dir.nr++;
3162 }
3163
3164 conclude_dirstat(options, &dir, changed);
3165 }
3166
3167 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3168 {
3169 int i;
3170 unsigned long changed;
3171 struct dirstat_dir dir;
3172
3173 if (data->nr == 0)
3174 return;
3175
3176 dir.files = NULL;
3177 dir.alloc = 0;
3178 dir.nr = 0;
3179 dir.permille = options->dirstat_permille;
3180 dir.cumulative = options->flags.dirstat_cumulative;
3181
3182 changed = 0;
3183 for (i = 0; i < data->nr; i++) {
3184 struct diffstat_file *file = data->files[i];
3185 unsigned long damage = file->added + file->deleted;
3186 if (file->is_binary)
3187 /*
3188 * binary files counts bytes, not lines. Must find some
3189 * way to normalize binary bytes vs. textual lines.
3190 * The following heuristic assumes that there are 64
3191 * bytes per "line".
3192 * This is stupid and ugly, but very cheap...
3193 */
3194 damage = DIV_ROUND_UP(damage, 64);
3195 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3196 dir.files[dir.nr].name = file->name;
3197 dir.files[dir.nr].changed = damage;
3198 changed += damage;
3199 dir.nr++;
3200 }
3201
3202 conclude_dirstat(options, &dir, changed);
3203 }
3204
3205 static void free_diffstat_file(struct diffstat_file *f)
3206 {
3207 free(f->print_name);
3208 free(f->name);
3209 free(f->from_name);
3210 free(f);
3211 }
3212
3213 void free_diffstat_info(struct diffstat_t *diffstat)
3214 {
3215 int i;
3216 for (i = 0; i < diffstat->nr; i++)
3217 free_diffstat_file(diffstat->files[i]);
3218 free(diffstat->files);
3219 }
3220
3221 struct checkdiff_t {
3222 const char *filename;
3223 int lineno;
3224 int conflict_marker_size;
3225 struct diff_options *o;
3226 unsigned ws_rule;
3227 unsigned status;
3228 };
3229
3230 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3231 {
3232 char firstchar;
3233 int cnt;
3234
3235 if (len < marker_size + 1)
3236 return 0;
3237 firstchar = line[0];
3238 switch (firstchar) {
3239 case '=': case '>': case '<': case '|':
3240 break;
3241 default:
3242 return 0;
3243 }
3244 for (cnt = 1; cnt < marker_size; cnt++)
3245 if (line[cnt] != firstchar)
3246 return 0;
3247 /* line[1] through line[marker_size-1] are same as firstchar */
3248 if (len < marker_size + 1 || !isspace(line[marker_size]))
3249 return 0;
3250 return 1;
3251 }
3252
3253 static void checkdiff_consume_hunk(void *priv,
3254 long ob UNUSED, long on UNUSED,
3255 long nb, long nn UNUSED,
3256 const char *func UNUSED, long funclen UNUSED)
3257
3258 {
3259 struct checkdiff_t *data = priv;
3260 data->lineno = nb - 1;
3261 }
3262
3263 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3264 {
3265 struct checkdiff_t *data = priv;
3266 int marker_size = data->conflict_marker_size;
3267 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3268 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3269 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3270 char *err;
3271 const char *line_prefix;
3272
3273 assert(data->o);
3274 line_prefix = diff_line_prefix(data->o);
3275
3276 if (line[0] == '+') {
3277 unsigned bad;
3278 data->lineno++;
3279 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3280 data->status |= 1;
3281 fprintf(data->o->file,
3282 "%s%s:%d: leftover conflict marker\n",
3283 line_prefix, data->filename, data->lineno);
3284 }
3285 bad = ws_check(line + 1, len - 1, data->ws_rule);
3286 if (!bad)
3287 return 0;
3288 data->status |= bad;
3289 err = whitespace_error_string(bad);
3290 fprintf(data->o->file, "%s%s:%d: %s.\n",
3291 line_prefix, data->filename, data->lineno, err);
3292 free(err);
3293 emit_line(data->o, set, reset, line, 1);
3294 ws_check_emit(line + 1, len - 1, data->ws_rule,
3295 data->o->file, set, reset, ws);
3296 } else if (line[0] == ' ') {
3297 data->lineno++;
3298 }
3299 return 0;
3300 }
3301
3302 static unsigned char *deflate_it(char *data,
3303 unsigned long size,
3304 unsigned long *result_size)
3305 {
3306 int bound;
3307 unsigned char *deflated;
3308 git_zstream stream;
3309
3310 git_deflate_init(&stream, zlib_compression_level);
3311 bound = git_deflate_bound(&stream, size);
3312 deflated = xmalloc(bound);
3313 stream.next_out = deflated;
3314 stream.avail_out = bound;
3315
3316 stream.next_in = (unsigned char *)data;
3317 stream.avail_in = size;
3318 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3319 ; /* nothing */
3320 git_deflate_end(&stream);
3321 *result_size = stream.total_out;
3322 return deflated;
3323 }
3324
3325 static void emit_binary_diff_body(struct diff_options *o,
3326 mmfile_t *one, mmfile_t *two)
3327 {
3328 void *cp;
3329 void *delta;
3330 void *deflated;
3331 void *data;
3332 unsigned long orig_size;
3333 unsigned long delta_size;
3334 unsigned long deflate_size;
3335 unsigned long data_size;
3336
3337 /* We could do deflated delta, or we could do just deflated two,
3338 * whichever is smaller.
3339 */
3340 delta = NULL;
3341 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3342 if (one->size && two->size) {
3343 delta = diff_delta(one->ptr, one->size,
3344 two->ptr, two->size,
3345 &delta_size, deflate_size);
3346 if (delta) {
3347 void *to_free = delta;
3348 orig_size = delta_size;
3349 delta = deflate_it(delta, delta_size, &delta_size);
3350 free(to_free);
3351 }
3352 }
3353
3354 if (delta && delta_size < deflate_size) {
3355 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3356 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3357 s, strlen(s), 0);
3358 free(s);
3359 free(deflated);
3360 data = delta;
3361 data_size = delta_size;
3362 } else {
3363 char *s = xstrfmt("%lu", two->size);
3364 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3365 s, strlen(s), 0);
3366 free(s);
3367 free(delta);
3368 data = deflated;
3369 data_size = deflate_size;
3370 }
3371
3372 /* emit data encoded in base85 */
3373 cp = data;
3374 while (data_size) {
3375 int len;
3376 int bytes = (52 < data_size) ? 52 : data_size;
3377 char line[71];
3378 data_size -= bytes;
3379 if (bytes <= 26)
3380 line[0] = bytes + 'A' - 1;
3381 else
3382 line[0] = bytes - 26 + 'a' - 1;
3383 encode_85(line + 1, cp, bytes);
3384 cp = (char *) cp + bytes;
3385
3386 len = strlen(line);
3387 line[len++] = '\n';
3388 line[len] = '\0';
3389
3390 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3391 line, len, 0);
3392 }
3393 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3394 free(data);
3395 }
3396
3397 static void emit_binary_diff(struct diff_options *o,
3398 mmfile_t *one, mmfile_t *two)
3399 {
3400 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3401 emit_binary_diff_body(o, one, two);
3402 emit_binary_diff_body(o, two, one);
3403 }
3404
3405 int diff_filespec_is_binary(struct repository *r,
3406 struct diff_filespec *one)
3407 {
3408 struct diff_populate_filespec_options dpf_options = {
3409 .check_binary = 1,
3410 };
3411
3412 if (one->is_binary == -1) {
3413 diff_filespec_load_driver(one, r->index);
3414 if (one->driver->binary != -1)
3415 one->is_binary = one->driver->binary;
3416 else {
3417 if (!one->data && DIFF_FILE_VALID(one))
3418 diff_populate_filespec(r, one, &dpf_options);
3419 if (one->is_binary == -1 && one->data)
3420 one->is_binary = buffer_is_binary(one->data,
3421 one->size);
3422 if (one->is_binary == -1)
3423 one->is_binary = 0;
3424 }
3425 }
3426 return one->is_binary;
3427 }
3428
3429 static const struct userdiff_funcname *
3430 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3431 {
3432 diff_filespec_load_driver(one, o->repo->index);
3433 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3434 }
3435
3436 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3437 {
3438 if (!options->a_prefix)
3439 options->a_prefix = a;
3440 if (!options->b_prefix)
3441 options->b_prefix = b;
3442 }
3443
3444 void diff_set_noprefix(struct diff_options *options)
3445 {
3446 options->a_prefix = options->b_prefix = "";
3447 }
3448
3449 void diff_set_default_prefix(struct diff_options *options)
3450 {
3451 options->a_prefix = diff_src_prefix ? diff_src_prefix : "a/";
3452 options->b_prefix = diff_dst_prefix ? diff_dst_prefix : "b/";
3453 }
3454
3455 struct userdiff_driver *get_textconv(struct repository *r,
3456 struct diff_filespec *one)
3457 {
3458 if (!DIFF_FILE_VALID(one))
3459 return NULL;
3460
3461 diff_filespec_load_driver(one, r->index);
3462 return userdiff_get_textconv(r, one->driver);
3463 }
3464
3465 static struct string_list *additional_headers(struct diff_options *o,
3466 const char *path)
3467 {
3468 if (!o->additional_path_headers)
3469 return NULL;
3470 return strmap_get(o->additional_path_headers, path);
3471 }
3472
3473 static void add_formatted_header(struct strbuf *msg,
3474 const char *header,
3475 const char *line_prefix,
3476 const char *meta,
3477 const char *reset)
3478 {
3479 const char *next, *newline;
3480
3481 for (next = header; *next; next = newline) {
3482 newline = strchrnul(next, '\n');
3483 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3484 (int)(newline - next), next, reset);
3485 if (*newline)
3486 newline++;
3487 }
3488 }
3489
3490 static void add_formatted_headers(struct strbuf *msg,
3491 struct string_list *more_headers,
3492 const char *line_prefix,
3493 const char *meta,
3494 const char *reset)
3495 {
3496 int i;
3497
3498 for (i = 0; i < more_headers->nr; i++)
3499 add_formatted_header(msg, more_headers->items[i].string,
3500 line_prefix, meta, reset);
3501 }
3502
3503 static int diff_filepair_is_phoney(struct diff_filespec *one,
3504 struct diff_filespec *two)
3505 {
3506 /*
3507 * This function specifically looks for pairs injected by
3508 * create_filepairs_for_header_only_notifications(). Such
3509 * pairs are "phoney" in that they do not represent any
3510 * content or even mode difference, but were inserted because
3511 * diff_queued_diff previously had no pair associated with
3512 * that path but we needed some pair to avoid losing the
3513 * "remerge CONFLICT" header associated with the path.
3514 */
3515 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3516 }
3517
3518 static int set_diff_algorithm(struct diff_options *opts,
3519 const char *alg)
3520 {
3521 long value = parse_algorithm_value(alg);
3522
3523 if (value < 0)
3524 return -1;
3525
3526 /* clear out previous settings */
3527 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3528 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3529 opts->xdl_opts |= value;
3530
3531 return 0;
3532 }
3533
3534 static void builtin_diff(const char *name_a,
3535 const char *name_b,
3536 struct diff_filespec *one,
3537 struct diff_filespec *two,
3538 const char *xfrm_msg,
3539 int must_show_header,
3540 struct diff_options *o,
3541 int complete_rewrite)
3542 {
3543 mmfile_t mf1, mf2;
3544 const char *lbl[2];
3545 char *a_one, *b_two;
3546 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3547 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3548 const char *a_prefix, *b_prefix;
3549 struct userdiff_driver *textconv_one = NULL;
3550 struct userdiff_driver *textconv_two = NULL;
3551 struct strbuf header = STRBUF_INIT;
3552 const char *line_prefix = diff_line_prefix(o);
3553
3554 diff_set_mnemonic_prefix(o, "a/", "b/");
3555 if (o->flags.reverse_diff) {
3556 a_prefix = o->b_prefix;
3557 b_prefix = o->a_prefix;
3558 } else {
3559 a_prefix = o->a_prefix;
3560 b_prefix = o->b_prefix;
3561 }
3562
3563 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3564 (!one->mode || S_ISGITLINK(one->mode)) &&
3565 (!two->mode || S_ISGITLINK(two->mode)) &&
3566 (!diff_filepair_is_phoney(one, two))) {
3567 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3568 &one->oid, &two->oid,
3569 two->dirty_submodule);
3570 o->found_changes = 1;
3571 return;
3572 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3573 (!one->mode || S_ISGITLINK(one->mode)) &&
3574 (!two->mode || S_ISGITLINK(two->mode)) &&
3575 (!diff_filepair_is_phoney(one, two))) {
3576 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3577 &one->oid, &two->oid,
3578 two->dirty_submodule);
3579 o->found_changes = 1;
3580 return;
3581 }
3582
3583 if (o->flags.allow_textconv) {
3584 textconv_one = get_textconv(o->repo, one);
3585 textconv_two = get_textconv(o->repo, two);
3586 }
3587
3588 /* Never use a non-valid filename anywhere if at all possible */
3589 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3590 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3591
3592 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3593 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3594 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3595 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3596 if (diff_filepair_is_phoney(one, two)) {
3597 /*
3598 * We should only reach this point for pairs generated from
3599 * create_filepairs_for_header_only_notifications(). For
3600 * these, we want to avoid the "/dev/null" special casing
3601 * above, because we do not want such pairs shown as either
3602 * "new file" or "deleted file" below.
3603 */
3604 lbl[0] = a_one;
3605 lbl[1] = b_two;
3606 }
3607 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3608 if (lbl[0][0] == '/') {
3609 /* /dev/null */
3610 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3611 if (xfrm_msg)
3612 strbuf_addstr(&header, xfrm_msg);
3613 o->found_changes = 1;
3614 must_show_header = 1;
3615 }
3616 else if (lbl[1][0] == '/') {
3617 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3618 if (xfrm_msg)
3619 strbuf_addstr(&header, xfrm_msg);
3620 o->found_changes = 1;
3621 must_show_header = 1;
3622 }
3623 else {
3624 if (one->mode != two->mode) {
3625 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3626 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3627 o->found_changes = 1;
3628 must_show_header = 1;
3629 }
3630 if (xfrm_msg)
3631 strbuf_addstr(&header, xfrm_msg);
3632
3633 /*
3634 * we do not run diff between different kind
3635 * of objects.
3636 */
3637 if ((one->mode ^ two->mode) & S_IFMT)
3638 goto free_ab_and_return;
3639 if (complete_rewrite &&
3640 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3641 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3642 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3643 header.buf, header.len, 0);
3644 strbuf_reset(&header);
3645 emit_rewrite_diff(name_a, name_b, one, two,
3646 textconv_one, textconv_two, o);
3647 o->found_changes = 1;
3648 goto free_ab_and_return;
3649 }
3650 }
3651
3652 if (o->irreversible_delete && lbl[1][0] == '/') {
3653 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3654 header.len, 0);
3655 strbuf_reset(&header);
3656 goto free_ab_and_return;
3657 } else if (!o->flags.text &&
3658 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3659 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3660 struct strbuf sb = STRBUF_INIT;
3661 if (!one->data && !two->data &&
3662 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3663 !o->flags.binary) {
3664 if (oideq(&one->oid, &two->oid)) {
3665 if (must_show_header)
3666 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3667 header.buf, header.len,
3668 0);
3669 goto free_ab_and_return;
3670 }
3671 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3672 header.buf, header.len, 0);
3673 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3674 diff_line_prefix(o), lbl[0], lbl[1]);
3675 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3676 sb.buf, sb.len, 0);
3677 strbuf_release(&sb);
3678 o->found_changes = 1;
3679 goto free_ab_and_return;
3680 }
3681 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3682 fill_mmfile(o->repo, &mf2, two) < 0)
3683 die("unable to read files to diff");
3684 /* Quite common confusing case */
3685 if (mf1.size == mf2.size &&
3686 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3687 if (must_show_header)
3688 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3689 header.buf, header.len, 0);
3690 goto free_ab_and_return;
3691 }
3692 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3693 strbuf_reset(&header);
3694 if (o->flags.binary)
3695 emit_binary_diff(o, &mf1, &mf2);
3696 else {
3697 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3698 diff_line_prefix(o), lbl[0], lbl[1]);
3699 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3700 sb.buf, sb.len, 0);
3701 strbuf_release(&sb);
3702 }
3703 o->found_changes = 1;
3704 } else {
3705 /* Crazy xdl interfaces.. */
3706 const char *diffopts;
3707 const char *v;
3708 xpparam_t xpp;
3709 xdemitconf_t xecfg;
3710 struct emit_callback ecbdata;
3711 const struct userdiff_funcname *pe;
3712
3713 if (must_show_header) {
3714 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3715 header.buf, header.len, 0);
3716 strbuf_reset(&header);
3717 }
3718
3719 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3720 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3721
3722 pe = diff_funcname_pattern(o, one);
3723 if (!pe)
3724 pe = diff_funcname_pattern(o, two);
3725
3726 memset(&xpp, 0, sizeof(xpp));
3727 memset(&xecfg, 0, sizeof(xecfg));
3728 memset(&ecbdata, 0, sizeof(ecbdata));
3729 if (o->flags.suppress_diff_headers)
3730 lbl[0] = NULL;
3731 ecbdata.label_path = lbl;
3732 ecbdata.color_diff = want_color(o->use_color);
3733 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3734 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3735 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3736 ecbdata.opt = o;
3737 if (header.len && !o->flags.suppress_diff_headers)
3738 ecbdata.header = &header;
3739 xpp.flags = o->xdl_opts;
3740 xpp.ignore_regex = o->ignore_regex;
3741 xpp.ignore_regex_nr = o->ignore_regex_nr;
3742 xpp.anchors = o->anchors;
3743 xpp.anchors_nr = o->anchors_nr;
3744 xecfg.ctxlen = o->context;
3745 xecfg.interhunkctxlen = o->interhunkcontext;
3746 xecfg.flags = XDL_EMIT_FUNCNAMES;
3747 if (o->flags.funccontext)
3748 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3749 if (pe)
3750 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3751
3752 diffopts = getenv("GIT_DIFF_OPTS");
3753 if (!diffopts)
3754 ;
3755 else if (skip_prefix(diffopts, "--unified=", &v))
3756 xecfg.ctxlen = strtoul(v, NULL, 10);
3757 else if (skip_prefix(diffopts, "-u", &v))
3758 xecfg.ctxlen = strtoul(v, NULL, 10);
3759
3760 if (o->word_diff)
3761 init_diff_words_data(&ecbdata, o, one, two);
3762 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3763 &ecbdata, &xpp, &xecfg))
3764 die("unable to generate diff for %s", one->path);
3765 if (o->word_diff)
3766 free_diff_words_data(&ecbdata);
3767 if (textconv_one)
3768 free(mf1.ptr);
3769 if (textconv_two)
3770 free(mf2.ptr);
3771 xdiff_clear_find_func(&xecfg);
3772 }
3773
3774 free_ab_and_return:
3775 strbuf_release(&header);
3776 diff_free_filespec_data(one);
3777 diff_free_filespec_data(two);
3778 free(a_one);
3779 free(b_two);
3780 return;
3781 }
3782
3783 static const char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3784 {
3785 if (!is_renamed) {
3786 if (p->status == DIFF_STATUS_ADDED) {
3787 if (S_ISLNK(p->two->mode))
3788 return "new +l";
3789 else if ((p->two->mode & 0777) == 0755)
3790 return "new +x";
3791 else
3792 return "new";
3793 } else if (p->status == DIFF_STATUS_DELETED)
3794 return "gone";
3795 }
3796 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3797 return "mode -l";
3798 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3799 return "mode +l";
3800 else if ((p->one->mode & 0777) == 0644 &&
3801 (p->two->mode & 0777) == 0755)
3802 return "mode +x";
3803 else if ((p->one->mode & 0777) == 0755 &&
3804 (p->two->mode & 0777) == 0644)
3805 return "mode -x";
3806 return NULL;
3807 }
3808
3809 static void builtin_diffstat(const char *name_a, const char *name_b,
3810 struct diff_filespec *one,
3811 struct diff_filespec *two,
3812 struct diffstat_t *diffstat,
3813 struct diff_options *o,
3814 struct diff_filepair *p)
3815 {
3816 mmfile_t mf1, mf2;
3817 struct diffstat_file *data;
3818 int may_differ;
3819 int complete_rewrite = 0;
3820
3821 if (!DIFF_PAIR_UNMERGED(p)) {
3822 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3823 complete_rewrite = 1;
3824 }
3825
3826 data = diffstat_add(diffstat, name_a, name_b);
3827 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3828 if (o->flags.stat_with_summary)
3829 data->comments = get_compact_summary(p, data->is_renamed);
3830
3831 if (!one || !two) {
3832 data->is_unmerged = 1;
3833 return;
3834 }
3835
3836 /* saves some reads if true, not a guarantee of diff outcome */
3837 may_differ = !(one->oid_valid && two->oid_valid &&
3838 oideq(&one->oid, &two->oid));
3839
3840 if (diff_filespec_is_binary(o->repo, one) ||
3841 diff_filespec_is_binary(o->repo, two)) {
3842 data->is_binary = 1;
3843 if (!may_differ) {
3844 data->added = 0;
3845 data->deleted = 0;
3846 } else {
3847 data->added = diff_filespec_size(o->repo, two);
3848 data->deleted = diff_filespec_size(o->repo, one);
3849 }
3850 }
3851
3852 else if (complete_rewrite) {
3853 diff_populate_filespec(o->repo, one, NULL);
3854 diff_populate_filespec(o->repo, two, NULL);
3855 data->deleted = count_lines(one->data, one->size);
3856 data->added = count_lines(two->data, two->size);
3857 }
3858
3859 else if (may_differ) {
3860 /* Crazy xdl interfaces.. */
3861 xpparam_t xpp;
3862 xdemitconf_t xecfg;
3863
3864 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3865 fill_mmfile(o->repo, &mf2, two) < 0)
3866 die("unable to read files to diff");
3867
3868 memset(&xpp, 0, sizeof(xpp));
3869 memset(&xecfg, 0, sizeof(xecfg));
3870 xpp.flags = o->xdl_opts;
3871 xpp.ignore_regex = o->ignore_regex;
3872 xpp.ignore_regex_nr = o->ignore_regex_nr;
3873 xpp.anchors = o->anchors;
3874 xpp.anchors_nr = o->anchors_nr;
3875 xecfg.ctxlen = o->context;
3876 xecfg.interhunkctxlen = o->interhunkcontext;
3877 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3878 if (xdi_diff_outf(&mf1, &mf2, NULL,
3879 diffstat_consume, diffstat, &xpp, &xecfg))
3880 die("unable to generate diffstat for %s", one->path);
3881
3882 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3883 struct diffstat_file *file =
3884 diffstat->files[diffstat->nr - 1];
3885 /*
3886 * Omit diffstats of modified files where nothing changed.
3887 * Even if may_differ, this might be the case due to
3888 * ignoring whitespace changes, etc.
3889 *
3890 * But note that we special-case additions, deletions,
3891 * renames, and mode changes as adding an empty file,
3892 * for example is still of interest.
3893 */
3894 if ((p->status == DIFF_STATUS_MODIFIED)
3895 && !file->added
3896 && !file->deleted
3897 && one->mode == two->mode) {
3898 free_diffstat_file(file);
3899 diffstat->nr--;
3900 }
3901 }
3902 }
3903
3904 diff_free_filespec_data(one);
3905 diff_free_filespec_data(two);
3906 }
3907
3908 static void builtin_checkdiff(const char *name_a, const char *name_b,
3909 const char *attr_path,
3910 struct diff_filespec *one,
3911 struct diff_filespec *two,
3912 struct diff_options *o)
3913 {
3914 mmfile_t mf1, mf2;
3915 struct checkdiff_t data;
3916
3917 if (!two)
3918 return;
3919
3920 memset(&data, 0, sizeof(data));
3921 data.filename = name_b ? name_b : name_a;
3922 data.lineno = 0;
3923 data.o = o;
3924 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3925 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3926
3927 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3928 fill_mmfile(o->repo, &mf2, two) < 0)
3929 die("unable to read files to diff");
3930
3931 /*
3932 * All the other codepaths check both sides, but not checking
3933 * the "old" side here is deliberate. We are checking the newly
3934 * introduced changes, and as long as the "new" side is text, we
3935 * can and should check what it introduces.
3936 */
3937 if (diff_filespec_is_binary(o->repo, two))
3938 goto free_and_return;
3939 else {
3940 /* Crazy xdl interfaces.. */
3941 xpparam_t xpp;
3942 xdemitconf_t xecfg;
3943
3944 memset(&xpp, 0, sizeof(xpp));
3945 memset(&xecfg, 0, sizeof(xecfg));
3946 xecfg.ctxlen = 1; /* at least one context line */
3947 xpp.flags = 0;
3948 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3949 checkdiff_consume, &data,
3950 &xpp, &xecfg))
3951 die("unable to generate checkdiff for %s", one->path);
3952
3953 if (data.ws_rule & WS_BLANK_AT_EOF) {
3954 struct emit_callback ecbdata;
3955 int blank_at_eof;
3956
3957 ecbdata.ws_rule = data.ws_rule;
3958 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3959 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3960
3961 if (blank_at_eof) {
3962 static char *err;
3963 if (!err)
3964 err = whitespace_error_string(WS_BLANK_AT_EOF);
3965 fprintf(o->file, "%s:%d: %s.\n",
3966 data.filename, blank_at_eof, err);
3967 data.status = 1; /* report errors */
3968 }
3969 }
3970 }
3971 free_and_return:
3972 diff_free_filespec_data(one);
3973 diff_free_filespec_data(two);
3974 if (data.status)
3975 o->flags.check_failed = 1;
3976 }
3977
3978 struct diff_filespec *alloc_filespec(const char *path)
3979 {
3980 struct diff_filespec *spec;
3981
3982 FLEXPTR_ALLOC_STR(spec, path, path);
3983 spec->count = 1;
3984 spec->is_binary = -1;
3985 return spec;
3986 }
3987
3988 void free_filespec(struct diff_filespec *spec)
3989 {
3990 if (!--spec->count) {
3991 diff_free_filespec_data(spec);
3992 free(spec);
3993 }
3994 }
3995
3996 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3997 int oid_valid, unsigned short mode)
3998 {
3999 if (mode) {
4000 spec->mode = canon_mode(mode);
4001 oidcpy(&spec->oid, oid);
4002 spec->oid_valid = oid_valid;
4003 }
4004 }
4005
4006 /*
4007 * Given a name and sha1 pair, if the index tells us the file in
4008 * the work tree has that object contents, return true, so that
4009 * prepare_temp_file() does not have to inflate and extract.
4010 */
4011 static int reuse_worktree_file(struct index_state *istate,
4012 const char *name,
4013 const struct object_id *oid,
4014 int want_file)
4015 {
4016 const struct cache_entry *ce;
4017 struct stat st;
4018 int pos, len;
4019
4020 /*
4021 * We do not read the cache ourselves here, because the
4022 * benchmark with my previous version that always reads cache
4023 * shows that it makes things worse for diff-tree comparing
4024 * two linux-2.6 kernel trees in an already checked out work
4025 * tree. This is because most diff-tree comparisons deal with
4026 * only a small number of files, while reading the cache is
4027 * expensive for a large project, and its cost outweighs the
4028 * savings we get by not inflating the object to a temporary
4029 * file. Practically, this code only helps when we are used
4030 * by diff-cache --cached, which does read the cache before
4031 * calling us.
4032 */
4033 if (!istate->cache)
4034 return 0;
4035
4036 /* We want to avoid the working directory if our caller
4037 * doesn't need the data in a normal file, this system
4038 * is rather slow with its stat/open/mmap/close syscalls,
4039 * and the object is contained in a pack file. The pack
4040 * is probably already open and will be faster to obtain
4041 * the data through than the working directory. Loose
4042 * objects however would tend to be slower as they need
4043 * to be individually opened and inflated.
4044 */
4045 if (!FAST_WORKING_DIRECTORY && !want_file &&
4046 has_object_pack(istate->repo, oid))
4047 return 0;
4048
4049 /*
4050 * Similarly, if we'd have to convert the file contents anyway, that
4051 * makes the optimization not worthwhile.
4052 */
4053 if (!want_file && would_convert_to_git(istate, name))
4054 return 0;
4055
4056 /*
4057 * If this path does not match our sparse-checkout definition,
4058 * then the file will not be in the working directory.
4059 */
4060 if (!path_in_sparse_checkout(name, istate))
4061 return 0;
4062
4063 len = strlen(name);
4064 pos = index_name_pos(istate, name, len);
4065 if (pos < 0)
4066 return 0;
4067 ce = istate->cache[pos];
4068
4069 /*
4070 * This is not the sha1 we are looking for, or
4071 * unreusable because it is not a regular file.
4072 */
4073 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4074 return 0;
4075
4076 /*
4077 * If ce is marked as "assume unchanged", there is no
4078 * guarantee that work tree matches what we are looking for.
4079 */
4080 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4081 return 0;
4082
4083 /*
4084 * If ce matches the file in the work tree, we can reuse it.
4085 */
4086 if (ce_uptodate(ce) ||
4087 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4088 return 1;
4089
4090 return 0;
4091 }
4092
4093 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4094 {
4095 struct strbuf buf = STRBUF_INIT;
4096 const char *dirty = "";
4097
4098 /* Are we looking at the work tree? */
4099 if (s->dirty_submodule)
4100 dirty = "-dirty";
4101
4102 strbuf_addf(&buf, "Subproject commit %s%s\n",
4103 oid_to_hex(&s->oid), dirty);
4104 s->size = buf.len;
4105 if (size_only) {
4106 s->data = NULL;
4107 strbuf_release(&buf);
4108 } else {
4109 s->data = strbuf_detach(&buf, NULL);
4110 s->should_free = 1;
4111 }
4112 return 0;
4113 }
4114
4115 /*
4116 * While doing rename detection and pickaxe operation, we may need to
4117 * grab the data for the blob (or file) for our own in-core comparison.
4118 * diff_filespec has data and size fields for this purpose.
4119 */
4120 int diff_populate_filespec(struct repository *r,
4121 struct diff_filespec *s,
4122 const struct diff_populate_filespec_options *options)
4123 {
4124 int size_only = options ? options->check_size_only : 0;
4125 int check_binary = options ? options->check_binary : 0;
4126 int err = 0;
4127 int conv_flags = global_conv_flags_eol;
4128 /*
4129 * demote FAIL to WARN to allow inspecting the situation
4130 * instead of refusing.
4131 */
4132 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4133 conv_flags = CONV_EOL_RNDTRP_WARN;
4134
4135 if (!DIFF_FILE_VALID(s))
4136 die("internal error: asking to populate invalid file.");
4137 if (S_ISDIR(s->mode))
4138 return -1;
4139
4140 if (s->data)
4141 return 0;
4142
4143 if (size_only && 0 < s->size)
4144 return 0;
4145
4146 if (S_ISGITLINK(s->mode))
4147 return diff_populate_gitlink(s, size_only);
4148
4149 if (!s->oid_valid ||
4150 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4151 struct strbuf buf = STRBUF_INIT;
4152 struct stat st;
4153 int fd;
4154
4155 if (lstat(s->path, &st) < 0) {
4156 err_empty:
4157 err = -1;
4158 empty:
4159 s->data = (char *)"";
4160 s->size = 0;
4161 return err;
4162 }
4163 s->size = xsize_t(st.st_size);
4164 if (!s->size)
4165 goto empty;
4166 if (S_ISLNK(st.st_mode)) {
4167 struct strbuf sb = STRBUF_INIT;
4168
4169 if (strbuf_readlink(&sb, s->path, s->size))
4170 goto err_empty;
4171 s->size = sb.len;
4172 s->data = strbuf_detach(&sb, NULL);
4173 s->should_free = 1;
4174 return 0;
4175 }
4176
4177 /*
4178 * Even if the caller would be happy with getting
4179 * only the size, we cannot return early at this
4180 * point if the path requires us to run the content
4181 * conversion.
4182 */
4183 if (size_only && !would_convert_to_git(r->index, s->path))
4184 return 0;
4185
4186 /*
4187 * Note: this check uses xsize_t(st.st_size) that may
4188 * not be the true size of the blob after it goes
4189 * through convert_to_git(). This may not strictly be
4190 * correct, but the whole point of big_file_threshold
4191 * and is_binary check being that we want to avoid
4192 * opening the file and inspecting the contents, this
4193 * is probably fine.
4194 */
4195 if (check_binary &&
4196 s->size > big_file_threshold && s->is_binary == -1) {
4197 s->is_binary = 1;
4198 return 0;
4199 }
4200 fd = open(s->path, O_RDONLY);
4201 if (fd < 0)
4202 goto err_empty;
4203 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4204 close(fd);
4205 s->should_munmap = 1;
4206
4207 /*
4208 * Convert from working tree format to canonical git format
4209 */
4210 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4211 size_t size = 0;
4212 munmap(s->data, s->size);
4213 s->should_munmap = 0;
4214 s->data = strbuf_detach(&buf, &size);
4215 s->size = size;
4216 s->should_free = 1;
4217 }
4218 }
4219 else {
4220 struct object_info info = {
4221 .sizep = &s->size
4222 };
4223
4224 if (!(size_only || check_binary))
4225 /*
4226 * Set contentp, since there is no chance that merely
4227 * the size is sufficient.
4228 */
4229 info.contentp = &s->data;
4230
4231 if (options && options->missing_object_cb) {
4232 if (!oid_object_info_extended(r, &s->oid, &info,
4233 OBJECT_INFO_LOOKUP_REPLACE |
4234 OBJECT_INFO_SKIP_FETCH_OBJECT))
4235 goto object_read;
4236 options->missing_object_cb(options->missing_object_data);
4237 }
4238 if (oid_object_info_extended(r, &s->oid, &info,
4239 OBJECT_INFO_LOOKUP_REPLACE))
4240 die("unable to read %s", oid_to_hex(&s->oid));
4241
4242 object_read:
4243 if (size_only || check_binary) {
4244 if (size_only)
4245 return 0;
4246 if (s->size > big_file_threshold && s->is_binary == -1) {
4247 s->is_binary = 1;
4248 return 0;
4249 }
4250 }
4251 if (!info.contentp) {
4252 info.contentp = &s->data;
4253 if (oid_object_info_extended(r, &s->oid, &info,
4254 OBJECT_INFO_LOOKUP_REPLACE))
4255 die("unable to read %s", oid_to_hex(&s->oid));
4256 }
4257 s->should_free = 1;
4258 }
4259 return 0;
4260 }
4261
4262 void diff_free_filespec_blob(struct diff_filespec *s)
4263 {
4264 if (s->should_free)
4265 free(s->data);
4266 else if (s->should_munmap)
4267 munmap(s->data, s->size);
4268
4269 if (s->should_free || s->should_munmap) {
4270 s->should_free = s->should_munmap = 0;
4271 s->data = NULL;
4272 }
4273 }
4274
4275 void diff_free_filespec_data(struct diff_filespec *s)
4276 {
4277 if (!s)
4278 return;
4279
4280 diff_free_filespec_blob(s);
4281 FREE_AND_NULL(s->cnt_data);
4282 }
4283
4284 static void prep_temp_blob(struct index_state *istate,
4285 const char *path, struct diff_tempfile *temp,
4286 void *blob,
4287 unsigned long size,
4288 const struct object_id *oid,
4289 int mode)
4290 {
4291 struct strbuf buf = STRBUF_INIT;
4292 char *path_dup = xstrdup(path);
4293 const char *base = basename(path_dup);
4294 struct checkout_metadata meta;
4295
4296 init_checkout_metadata(&meta, NULL, NULL, oid);
4297
4298 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4299 if (!temp->tempfile)
4300 die_errno("unable to create temp-file");
4301 if (convert_to_working_tree(istate, path,
4302 (const char *)blob, (size_t)size, &buf, &meta)) {
4303 blob = buf.buf;
4304 size = buf.len;
4305 }
4306 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4307 close_tempfile_gently(temp->tempfile))
4308 die_errno("unable to write temp-file");
4309 temp->name = get_tempfile_path(temp->tempfile);
4310 oid_to_hex_r(temp->hex, oid);
4311 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4312 strbuf_release(&buf);
4313 free(path_dup);
4314 }
4315
4316 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4317 struct diff_filespec *one)
4318 {
4319 struct diff_tempfile *temp = claim_diff_tempfile();
4320
4321 if (!DIFF_FILE_VALID(one)) {
4322 not_a_valid_file:
4323 /* A '-' entry produces this for file-2, and
4324 * a '+' entry produces this for file-1.
4325 */
4326 temp->name = "/dev/null";
4327 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4328 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4329 return temp;
4330 }
4331
4332 if (!S_ISGITLINK(one->mode) &&
4333 (!one->oid_valid ||
4334 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4335 struct stat st;
4336 if (lstat(one->path, &st) < 0) {
4337 if (errno == ENOENT)
4338 goto not_a_valid_file;
4339 die_errno("stat(%s)", one->path);
4340 }
4341 if (S_ISLNK(st.st_mode)) {
4342 struct strbuf sb = STRBUF_INIT;
4343 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4344 die_errno("readlink(%s)", one->path);
4345 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4346 (one->oid_valid ?
4347 &one->oid : null_oid()),
4348 (one->oid_valid ?
4349 one->mode : S_IFLNK));
4350 strbuf_release(&sb);
4351 }
4352 else {
4353 /* we can borrow from the file in the work tree */
4354 temp->name = one->path;
4355 if (!one->oid_valid)
4356 oid_to_hex_r(temp->hex, null_oid());
4357 else
4358 oid_to_hex_r(temp->hex, &one->oid);
4359 /* Even though we may sometimes borrow the
4360 * contents from the work tree, we always want
4361 * one->mode. mode is trustworthy even when
4362 * !(one->oid_valid), as long as
4363 * DIFF_FILE_VALID(one).
4364 */
4365 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4366 }
4367 return temp;
4368 }
4369 else {
4370 if (diff_populate_filespec(r, one, NULL))
4371 die("cannot read data blob for %s", one->path);
4372 prep_temp_blob(r->index, one->path, temp,
4373 one->data, one->size,
4374 &one->oid, one->mode);
4375 }
4376 return temp;
4377 }
4378
4379 static void add_external_diff_name(struct repository *r,
4380 struct strvec *argv,
4381 struct diff_filespec *df)
4382 {
4383 struct diff_tempfile *temp = prepare_temp_file(r, df);
4384 strvec_push(argv, temp->name);
4385 strvec_push(argv, temp->hex);
4386 strvec_push(argv, temp->mode);
4387 }
4388
4389 /* An external diff command takes:
4390 *
4391 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4392 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4393 *
4394 */
4395 static void run_external_diff(const struct external_diff *pgm,
4396 const char *name,
4397 const char *other,
4398 struct diff_filespec *one,
4399 struct diff_filespec *two,
4400 const char *xfrm_msg,
4401 struct diff_options *o)
4402 {
4403 struct child_process cmd = CHILD_PROCESS_INIT;
4404 struct diff_queue_struct *q = &diff_queued_diff;
4405 int quiet = !(o->output_format & DIFF_FORMAT_PATCH);
4406 int rc;
4407
4408 /*
4409 * Trivial equality is handled by diff_unmodified_pair() before
4410 * we get here. If we don't need to show the diff and the
4411 * external diff program lacks the ability to tell us whether
4412 * it's empty then we consider it non-empty without even asking.
4413 */
4414 if (!pgm->trust_exit_code && quiet) {
4415 o->found_changes = 1;
4416 return;
4417 }
4418
4419 strvec_push(&cmd.args, pgm->cmd);
4420 strvec_push(&cmd.args, name);
4421
4422 if (one && two) {
4423 add_external_diff_name(o->repo, &cmd.args, one);
4424 add_external_diff_name(o->repo, &cmd.args, two);
4425 if (other) {
4426 strvec_push(&cmd.args, other);
4427 if (xfrm_msg)
4428 strvec_push(&cmd.args, xfrm_msg);
4429 }
4430 }
4431
4432 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4433 ++o->diff_path_counter);
4434 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4435
4436 diff_free_filespec_data(one);
4437 diff_free_filespec_data(two);
4438 cmd.use_shell = 1;
4439 cmd.no_stdout = quiet;
4440 rc = run_command(&cmd);
4441 if (!pgm->trust_exit_code && rc == 0)
4442 o->found_changes = 1;
4443 else if (pgm->trust_exit_code && rc == 0)
4444 ; /* nothing */
4445 else if (pgm->trust_exit_code && rc == 1)
4446 o->found_changes = 1;
4447 else
4448 die(_("external diff died, stopping at %s"), name);
4449
4450 remove_tempfile();
4451 }
4452
4453 static int similarity_index(struct diff_filepair *p)
4454 {
4455 return p->score * 100 / MAX_SCORE;
4456 }
4457
4458 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4459 {
4460 if (startup_info->have_repository)
4461 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4462 else {
4463 char *hex = oid_to_hex(oid);
4464 if (abbrev < 0)
4465 abbrev = FALLBACK_DEFAULT_ABBREV;
4466 if (abbrev > the_hash_algo->hexsz)
4467 BUG("oid abbreviation out of range: %d", abbrev);
4468 if (abbrev)
4469 hex[abbrev] = '\0';
4470 return hex;
4471 }
4472 }
4473
4474 static void fill_metainfo(struct strbuf *msg,
4475 const char *name,
4476 const char *other,
4477 struct diff_filespec *one,
4478 struct diff_filespec *two,
4479 struct diff_options *o,
4480 struct diff_filepair *p,
4481 int *must_show_header,
4482 int use_color)
4483 {
4484 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4485 const char *reset = diff_get_color(use_color, DIFF_RESET);
4486 const char *line_prefix = diff_line_prefix(o);
4487 struct string_list *more_headers = NULL;
4488
4489 *must_show_header = 1;
4490 strbuf_init(msg, PATH_MAX * 2 + 300);
4491 switch (p->status) {
4492 case DIFF_STATUS_COPIED:
4493 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4494 line_prefix, set, similarity_index(p));
4495 strbuf_addf(msg, "%s\n%s%scopy from ",
4496 reset, line_prefix, set);
4497 quote_c_style(name, msg, NULL, 0);
4498 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4499 quote_c_style(other, msg, NULL, 0);
4500 strbuf_addf(msg, "%s\n", reset);
4501 break;
4502 case DIFF_STATUS_RENAMED:
4503 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4504 line_prefix, set, similarity_index(p));
4505 strbuf_addf(msg, "%s\n%s%srename from ",
4506 reset, line_prefix, set);
4507 quote_c_style(name, msg, NULL, 0);
4508 strbuf_addf(msg, "%s\n%s%srename to ",
4509 reset, line_prefix, set);
4510 quote_c_style(other, msg, NULL, 0);
4511 strbuf_addf(msg, "%s\n", reset);
4512 break;
4513 case DIFF_STATUS_MODIFIED:
4514 if (p->score) {
4515 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4516 line_prefix,
4517 set, similarity_index(p), reset);
4518 break;
4519 }
4520 /* fallthru */
4521 default:
4522 *must_show_header = 0;
4523 }
4524 if ((more_headers = additional_headers(o, name))) {
4525 add_formatted_headers(msg, more_headers,
4526 line_prefix, set, reset);
4527 *must_show_header = 1;
4528 }
4529 if (one && two && !oideq(&one->oid, &two->oid)) {
4530 const unsigned hexsz = the_hash_algo->hexsz;
4531 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4532
4533 if (o->flags.full_index)
4534 abbrev = hexsz;
4535
4536 if (o->flags.binary) {
4537 mmfile_t mf;
4538 if ((!fill_mmfile(o->repo, &mf, one) &&
4539 diff_filespec_is_binary(o->repo, one)) ||
4540 (!fill_mmfile(o->repo, &mf, two) &&
4541 diff_filespec_is_binary(o->repo, two)))
4542 abbrev = hexsz;
4543 }
4544 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4545 diff_abbrev_oid(&one->oid, abbrev),
4546 diff_abbrev_oid(&two->oid, abbrev));
4547 if (one->mode == two->mode)
4548 strbuf_addf(msg, " %06o", one->mode);
4549 strbuf_addf(msg, "%s\n", reset);
4550 }
4551 }
4552
4553 static void run_diff_cmd(const struct external_diff *pgm,
4554 const char *name,
4555 const char *other,
4556 const char *attr_path,
4557 struct diff_filespec *one,
4558 struct diff_filespec *two,
4559 struct strbuf *msg,
4560 struct diff_options *o,
4561 struct diff_filepair *p)
4562 {
4563 const char *xfrm_msg = NULL;
4564 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4565 int must_show_header = 0;
4566 struct userdiff_driver *drv = NULL;
4567
4568 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4569 drv = userdiff_find_by_path(o->repo->index, attr_path);
4570
4571 if (o->flags.allow_external && drv && drv->external.cmd)
4572 pgm = &drv->external;
4573
4574 if (msg) {
4575 /*
4576 * don't use colors when the header is intended for an
4577 * external diff driver
4578 */
4579 fill_metainfo(msg, name, other, one, two, o, p,
4580 &must_show_header,
4581 want_color(o->use_color) && !pgm);
4582 xfrm_msg = msg->len ? msg->buf : NULL;
4583 }
4584
4585 if (pgm) {
4586 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4587 return;
4588 }
4589 if (one && two) {
4590 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4591 set_diff_algorithm(o, drv->algorithm);
4592
4593 builtin_diff(name, other ? other : name,
4594 one, two, xfrm_msg, must_show_header,
4595 o, complete_rewrite);
4596 if (p->status == DIFF_STATUS_COPIED ||
4597 p->status == DIFF_STATUS_RENAMED)
4598 o->found_changes = 1;
4599 } else {
4600 fprintf(o->file, "* Unmerged path %s\n", name);
4601 o->found_changes = 1;
4602 }
4603 }
4604
4605 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4606 {
4607 if (DIFF_FILE_VALID(one)) {
4608 if (!one->oid_valid) {
4609 struct stat st;
4610 if (one->is_stdin) {
4611 oidclr(&one->oid, the_repository->hash_algo);
4612 return;
4613 }
4614 if (lstat(one->path, &st) < 0)
4615 die_errno("stat '%s'", one->path);
4616 if (index_path(istate, &one->oid, one->path, &st, 0))
4617 die("cannot hash %s", one->path);
4618 }
4619 }
4620 else
4621 oidclr(&one->oid, the_repository->hash_algo);
4622 }
4623
4624 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4625 {
4626 /* Strip the prefix but do not molest /dev/null and absolute paths */
4627 if (*namep && !is_absolute_path(*namep)) {
4628 *namep += prefix_length;
4629 if (**namep == '/')
4630 ++*namep;
4631 }
4632 if (*otherp && !is_absolute_path(*otherp)) {
4633 *otherp += prefix_length;
4634 if (**otherp == '/')
4635 ++*otherp;
4636 }
4637 }
4638
4639 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4640 {
4641 const struct external_diff *pgm = external_diff();
4642 struct strbuf msg;
4643 struct diff_filespec *one = p->one;
4644 struct diff_filespec *two = p->two;
4645 const char *name;
4646 const char *other;
4647 const char *attr_path;
4648
4649 name = one->path;
4650 other = (strcmp(name, two->path) ? two->path : NULL);
4651 attr_path = name;
4652 if (o->prefix_length)
4653 strip_prefix(o->prefix_length, &name, &other);
4654
4655 if (!o->flags.allow_external)
4656 pgm = NULL;
4657
4658 if (DIFF_PAIR_UNMERGED(p)) {
4659 run_diff_cmd(pgm, name, NULL, attr_path,
4660 NULL, NULL, NULL, o, p);
4661 return;
4662 }
4663
4664 diff_fill_oid_info(one, o->repo->index);
4665 diff_fill_oid_info(two, o->repo->index);
4666
4667 if (!pgm &&
4668 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4669 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4670 /*
4671 * a filepair that changes between file and symlink
4672 * needs to be split into deletion and creation.
4673 */
4674 struct diff_filespec *null = alloc_filespec(two->path);
4675 run_diff_cmd(NULL, name, other, attr_path,
4676 one, null, &msg,
4677 o, p);
4678 free(null);
4679 strbuf_release(&msg);
4680
4681 null = alloc_filespec(one->path);
4682 run_diff_cmd(NULL, name, other, attr_path,
4683 null, two, &msg, o, p);
4684 free(null);
4685 }
4686 else
4687 run_diff_cmd(pgm, name, other, attr_path,
4688 one, two, &msg, o, p);
4689
4690 strbuf_release(&msg);
4691 }
4692
4693 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4694 struct diffstat_t *diffstat)
4695 {
4696 const char *name;
4697 const char *other;
4698
4699 if (!o->ignore_driver_algorithm) {
4700 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4701 p->one->path);
4702
4703 if (drv && drv->algorithm)
4704 set_diff_algorithm(o, drv->algorithm);
4705 }
4706
4707 if (DIFF_PAIR_UNMERGED(p)) {
4708 /* unmerged */
4709 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4710 diffstat, o, p);
4711 return;
4712 }
4713
4714 name = p->one->path;
4715 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4716
4717 if (o->prefix_length)
4718 strip_prefix(o->prefix_length, &name, &other);
4719
4720 diff_fill_oid_info(p->one, o->repo->index);
4721 diff_fill_oid_info(p->two, o->repo->index);
4722
4723 builtin_diffstat(name, other, p->one, p->two,
4724 diffstat, o, p);
4725 }
4726
4727 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4728 {
4729 const char *name;
4730 const char *other;
4731 const char *attr_path;
4732
4733 if (DIFF_PAIR_UNMERGED(p)) {
4734 /* unmerged */
4735 return;
4736 }
4737
4738 name = p->one->path;
4739 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4740 attr_path = other ? other : name;
4741
4742 if (o->prefix_length)
4743 strip_prefix(o->prefix_length, &name, &other);
4744
4745 diff_fill_oid_info(p->one, o->repo->index);
4746 diff_fill_oid_info(p->two, o->repo->index);
4747
4748 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4749 }
4750
4751 void repo_diff_setup(struct repository *r, struct diff_options *options)
4752 {
4753 memcpy(options, &default_diff_options, sizeof(*options));
4754
4755 options->file = stdout;
4756 options->repo = r;
4757
4758 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4759 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4760 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4761 options->abbrev = DEFAULT_ABBREV;
4762 options->line_termination = '\n';
4763 options->break_opt = -1;
4764 options->rename_limit = -1;
4765 options->dirstat_permille = diff_dirstat_permille_default;
4766 options->context = diff_context_default;
4767 options->interhunkcontext = diff_interhunk_context_default;
4768 options->ws_error_highlight = ws_error_highlight_default;
4769 options->flags.rename_empty = 1;
4770 options->flags.relative_name = diff_relative;
4771 options->objfind = NULL;
4772
4773 /* pathchange left =NULL by default */
4774 options->change = diff_change;
4775 options->add_remove = diff_addremove;
4776 options->use_color = diff_use_color_default;
4777 options->detect_rename = diff_detect_rename_default;
4778 options->xdl_opts |= diff_algorithm;
4779 if (diff_indent_heuristic)
4780 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4781
4782 options->orderfile = xstrdup_or_null(diff_order_file_cfg);
4783
4784 if (!options->flags.ignore_submodule_set)
4785 options->flags.ignore_untracked_in_submodules = 1;
4786
4787 if (diff_no_prefix) {
4788 diff_set_noprefix(options);
4789 } else if (!diff_mnemonic_prefix) {
4790 diff_set_default_prefix(options);
4791 }
4792
4793 options->color_moved = diff_color_moved_default;
4794 options->color_moved_ws_handling = diff_color_moved_ws_default;
4795 }
4796
4797 static const char diff_status_letters[] = {
4798 DIFF_STATUS_ADDED,
4799 DIFF_STATUS_COPIED,
4800 DIFF_STATUS_DELETED,
4801 DIFF_STATUS_MODIFIED,
4802 DIFF_STATUS_RENAMED,
4803 DIFF_STATUS_TYPE_CHANGED,
4804 DIFF_STATUS_UNKNOWN,
4805 DIFF_STATUS_UNMERGED,
4806 DIFF_STATUS_FILTER_AON,
4807 DIFF_STATUS_FILTER_BROKEN,
4808 '\0',
4809 };
4810
4811 static unsigned int filter_bit['Z' + 1];
4812
4813 static void prepare_filter_bits(void)
4814 {
4815 int i;
4816
4817 if (!filter_bit[DIFF_STATUS_ADDED]) {
4818 for (i = 0; diff_status_letters[i]; i++)
4819 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4820 }
4821 }
4822
4823 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4824 {
4825 return opt->filter & filter_bit[(int) status];
4826 }
4827
4828 unsigned diff_filter_bit(char status)
4829 {
4830 prepare_filter_bits();
4831 return filter_bit[(int) status];
4832 }
4833
4834 int diff_check_follow_pathspec(struct pathspec *ps, int die_on_error)
4835 {
4836 unsigned forbidden_magic;
4837
4838 if (ps->nr != 1) {
4839 if (die_on_error)
4840 die(_("--follow requires exactly one pathspec"));
4841 return 0;
4842 }
4843
4844 forbidden_magic = ps->items[0].magic & ~(PATHSPEC_FROMTOP |
4845 PATHSPEC_LITERAL);
4846 if (forbidden_magic) {
4847 if (die_on_error) {
4848 struct strbuf sb = STRBUF_INIT;
4849 pathspec_magic_names(forbidden_magic, &sb);
4850 die(_("pathspec magic not supported by --follow: %s"),
4851 sb.buf);
4852 }
4853 return 0;
4854 }
4855
4856 return 1;
4857 }
4858
4859 void diff_setup_done(struct diff_options *options)
4860 {
4861 unsigned check_mask = DIFF_FORMAT_NAME |
4862 DIFF_FORMAT_NAME_STATUS |
4863 DIFF_FORMAT_CHECKDIFF |
4864 DIFF_FORMAT_NO_OUTPUT;
4865 /*
4866 * This must be signed because we're comparing against a potentially
4867 * negative value.
4868 */
4869 const int hexsz = the_hash_algo->hexsz;
4870
4871 if (options->set_default)
4872 options->set_default(options);
4873
4874 if (HAS_MULTI_BITS(options->output_format & check_mask))
4875 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4876 "--name-only", "--name-status", "--check", "-s");
4877
4878 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4879 die(_("options '%s', '%s', and '%s' cannot be used together"),
4880 "-G", "-S", "--find-object");
4881
4882 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4883 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4884 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4885
4886 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4887 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4888 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4889
4890 /*
4891 * Most of the time we can say "there are changes"
4892 * only by checking if there are changed paths, but
4893 * --ignore-whitespace* options force us to look
4894 * inside contents.
4895 */
4896
4897 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4898 options->ignore_regex_nr)
4899 options->flags.diff_from_contents = 1;
4900 else
4901 options->flags.diff_from_contents = 0;
4902
4903 if (options->flags.find_copies_harder)
4904 options->detect_rename = DIFF_DETECT_COPY;
4905
4906 if (!options->flags.relative_name)
4907 options->prefix = NULL;
4908 if (options->prefix)
4909 options->prefix_length = strlen(options->prefix);
4910 else
4911 options->prefix_length = 0;
4912
4913 /*
4914 * --name-only, --name-status, --checkdiff, and -s
4915 * turn other output format off.
4916 */
4917 if (options->output_format & (DIFF_FORMAT_NAME |
4918 DIFF_FORMAT_NAME_STATUS |
4919 DIFF_FORMAT_CHECKDIFF |
4920 DIFF_FORMAT_NO_OUTPUT))
4921 options->output_format &= ~(DIFF_FORMAT_RAW |
4922 DIFF_FORMAT_NUMSTAT |
4923 DIFF_FORMAT_DIFFSTAT |
4924 DIFF_FORMAT_SHORTSTAT |
4925 DIFF_FORMAT_DIRSTAT |
4926 DIFF_FORMAT_SUMMARY |
4927 DIFF_FORMAT_PATCH);
4928
4929 /*
4930 * These cases always need recursive; we do not drop caller-supplied
4931 * recursive bits for other formats here.
4932 */
4933 if (options->output_format & (DIFF_FORMAT_PATCH |
4934 DIFF_FORMAT_NUMSTAT |
4935 DIFF_FORMAT_DIFFSTAT |
4936 DIFF_FORMAT_SHORTSTAT |
4937 DIFF_FORMAT_DIRSTAT |
4938 DIFF_FORMAT_SUMMARY |
4939 DIFF_FORMAT_CHECKDIFF))
4940 options->flags.recursive = 1;
4941 /*
4942 * Also pickaxe would not work very well if you do not say recursive
4943 */
4944 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4945 options->flags.recursive = 1;
4946 /*
4947 * When patches are generated, submodules diffed against the work tree
4948 * must be checked for dirtiness too so it can be shown in the output
4949 */
4950 if (options->output_format & DIFF_FORMAT_PATCH)
4951 options->flags.dirty_submodules = 1;
4952
4953 if (options->detect_rename && options->rename_limit < 0)
4954 options->rename_limit = diff_rename_limit_default;
4955 if (hexsz < options->abbrev)
4956 options->abbrev = hexsz; /* full */
4957
4958 /*
4959 * It does not make sense to show the first hit we happened
4960 * to have found. It does not make sense not to return with
4961 * exit code in such a case either.
4962 */
4963 if (options->flags.quick) {
4964 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4965 options->flags.exit_with_status = 1;
4966 }
4967
4968 /*
4969 * External diffs could declare non-identical contents equal
4970 * (think diff --ignore-space-change).
4971 */
4972 if (options->flags.allow_external && options->flags.exit_with_status)
4973 options->flags.diff_from_contents = 1;
4974
4975 options->diff_path_counter = 0;
4976
4977 if (options->flags.follow_renames)
4978 diff_check_follow_pathspec(&options->pathspec, 1);
4979
4980 if (!options->use_color ||
4981 (options->flags.allow_external && external_diff()))
4982 options->color_moved = 0;
4983
4984 if (options->filter_not) {
4985 if (!options->filter)
4986 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4987 options->filter &= ~options->filter_not;
4988 }
4989 }
4990
4991 int parse_long_opt(const char *opt, const char **argv,
4992 const char **optarg)
4993 {
4994 const char *arg = argv[0];
4995 if (!skip_prefix(arg, "--", &arg))
4996 return 0;
4997 if (!skip_prefix(arg, opt, &arg))
4998 return 0;
4999 if (*arg == '=') { /* stuck form: --option=value */
5000 *optarg = arg + 1;
5001 return 1;
5002 }
5003 if (*arg != '\0')
5004 return 0;
5005 /* separate form: --option value */
5006 if (!argv[1])
5007 die("Option '--%s' requires a value", opt);
5008 *optarg = argv[1];
5009 return 2;
5010 }
5011
5012 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
5013 {
5014 struct diff_options *options = opt->value;
5015 int width = options->stat_width;
5016 int name_width = options->stat_name_width;
5017 int graph_width = options->stat_graph_width;
5018 int count = options->stat_count;
5019 char *end;
5020
5021 BUG_ON_OPT_NEG(unset);
5022
5023 if (!strcmp(opt->long_name, "stat")) {
5024 if (value) {
5025 width = strtoul(value, &end, 10);
5026 if (*end == ',')
5027 name_width = strtoul(end+1, &end, 10);
5028 if (*end == ',')
5029 count = strtoul(end+1, &end, 10);
5030 if (*end)
5031 return error(_("invalid --stat value: %s"), value);
5032 }
5033 } else if (!strcmp(opt->long_name, "stat-width")) {
5034 width = strtoul(value, &end, 10);
5035 if (*end)
5036 return error(_("%s expects a numerical value"),
5037 opt->long_name);
5038 } else if (!strcmp(opt->long_name, "stat-name-width")) {
5039 name_width = strtoul(value, &end, 10);
5040 if (*end)
5041 return error(_("%s expects a numerical value"),
5042 opt->long_name);
5043 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
5044 graph_width = strtoul(value, &end, 10);
5045 if (*end)
5046 return error(_("%s expects a numerical value"),
5047 opt->long_name);
5048 } else if (!strcmp(opt->long_name, "stat-count")) {
5049 count = strtoul(value, &end, 10);
5050 if (*end)
5051 return error(_("%s expects a numerical value"),
5052 opt->long_name);
5053 } else
5054 BUG("%s should not get here", opt->long_name);
5055
5056 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5057 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5058 options->stat_name_width = name_width;
5059 options->stat_graph_width = graph_width;
5060 options->stat_width = width;
5061 options->stat_count = count;
5062 return 0;
5063 }
5064
5065 static int parse_dirstat_opt(struct diff_options *options, const char *params)
5066 {
5067 struct strbuf errmsg = STRBUF_INIT;
5068 if (parse_dirstat_params(options, params, &errmsg))
5069 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
5070 errmsg.buf);
5071 strbuf_release(&errmsg);
5072 /*
5073 * The caller knows a dirstat-related option is given from the command
5074 * line; allow it to say "return this_function();"
5075 */
5076 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5077 options->output_format |= DIFF_FORMAT_DIRSTAT;
5078 return 1;
5079 }
5080
5081 static int diff_opt_diff_filter(const struct option *option,
5082 const char *optarg, int unset)
5083 {
5084 struct diff_options *opt = option->value;
5085 int i, optch;
5086
5087 BUG_ON_OPT_NEG(unset);
5088 prepare_filter_bits();
5089
5090 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
5091 unsigned int bit;
5092 int negate;
5093
5094 if ('a' <= optch && optch <= 'z') {
5095 negate = 1;
5096 optch = toupper(optch);
5097 } else {
5098 negate = 0;
5099 }
5100
5101 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
5102 if (!bit)
5103 return error(_("unknown change class '%c' in --diff-filter=%s"),
5104 optarg[i], optarg);
5105 if (negate)
5106 opt->filter_not |= bit;
5107 else
5108 opt->filter |= bit;
5109 }
5110 return 0;
5111 }
5112
5113 static void enable_patch_output(int *fmt)
5114 {
5115 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
5116 *fmt |= DIFF_FORMAT_PATCH;
5117 }
5118
5119 static int diff_opt_ws_error_highlight(const struct option *option,
5120 const char *arg, int unset)
5121 {
5122 struct diff_options *opt = option->value;
5123 int val = parse_ws_error_highlight(arg);
5124
5125 BUG_ON_OPT_NEG(unset);
5126 if (val < 0)
5127 return error(_("unknown value after ws-error-highlight=%.*s"),
5128 -1 - val, arg);
5129 opt->ws_error_highlight = val;
5130 return 0;
5131 }
5132
5133 static int diff_opt_find_object(const struct option *option,
5134 const char *arg, int unset)
5135 {
5136 struct diff_options *opt = option->value;
5137 struct object_id oid;
5138
5139 BUG_ON_OPT_NEG(unset);
5140 if (repo_get_oid(the_repository, arg, &oid))
5141 return error(_("unable to resolve '%s'"), arg);
5142
5143 if (!opt->objfind)
5144 CALLOC_ARRAY(opt->objfind, 1);
5145
5146 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5147 opt->flags.recursive = 1;
5148 opt->flags.tree_in_recursive = 1;
5149 oidset_insert(opt->objfind, &oid);
5150 return 0;
5151 }
5152
5153 static int diff_opt_anchored(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 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5160 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5161 options->anchors_alloc);
5162 options->anchors[options->anchors_nr++] = xstrdup(arg);
5163 return 0;
5164 }
5165
5166 static int diff_opt_binary(const struct option *opt,
5167 const char *arg, int unset)
5168 {
5169 struct diff_options *options = opt->value;
5170
5171 BUG_ON_OPT_NEG(unset);
5172 BUG_ON_OPT_ARG(arg);
5173 enable_patch_output(&options->output_format);
5174 options->flags.binary = 1;
5175 return 0;
5176 }
5177
5178 static int diff_opt_break_rewrites(const struct option *opt,
5179 const char *arg, int unset)
5180 {
5181 int *break_opt = opt->value;
5182 int opt1, opt2;
5183
5184 BUG_ON_OPT_NEG(unset);
5185 if (!arg)
5186 arg = "";
5187 opt1 = parse_rename_score(&arg);
5188 if (*arg == 0)
5189 opt2 = 0;
5190 else if (*arg != '/')
5191 return error(_("%s expects <n>/<m> form"), opt->long_name);
5192 else {
5193 arg++;
5194 opt2 = parse_rename_score(&arg);
5195 }
5196 if (*arg != 0)
5197 return error(_("%s expects <n>/<m> form"), opt->long_name);
5198 *break_opt = opt1 | (opt2 << 16);
5199 return 0;
5200 }
5201
5202 static int diff_opt_char(const struct option *opt,
5203 const char *arg, int unset)
5204 {
5205 char *value = opt->value;
5206
5207 BUG_ON_OPT_NEG(unset);
5208 if (arg[1])
5209 return error(_("%s expects a character, got '%s'"),
5210 opt->long_name, arg);
5211 *value = arg[0];
5212 return 0;
5213 }
5214
5215 static int diff_opt_color_moved(const struct option *opt,
5216 const char *arg, int unset)
5217 {
5218 struct diff_options *options = opt->value;
5219
5220 if (unset) {
5221 options->color_moved = COLOR_MOVED_NO;
5222 } else if (!arg) {
5223 if (diff_color_moved_default)
5224 options->color_moved = diff_color_moved_default;
5225 if (options->color_moved == COLOR_MOVED_NO)
5226 options->color_moved = COLOR_MOVED_DEFAULT;
5227 } else {
5228 int cm = parse_color_moved(arg);
5229 if (cm < 0)
5230 return error(_("bad --color-moved argument: %s"), arg);
5231 options->color_moved = cm;
5232 }
5233 return 0;
5234 }
5235
5236 static int diff_opt_color_moved_ws(const struct option *opt,
5237 const char *arg, int unset)
5238 {
5239 struct diff_options *options = opt->value;
5240 unsigned cm;
5241
5242 if (unset) {
5243 options->color_moved_ws_handling = 0;
5244 return 0;
5245 }
5246
5247 cm = parse_color_moved_ws(arg);
5248 if (cm & COLOR_MOVED_WS_ERROR)
5249 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5250 options->color_moved_ws_handling = cm;
5251 return 0;
5252 }
5253
5254 static int diff_opt_color_words(const struct option *opt,
5255 const char *arg, int unset)
5256 {
5257 struct diff_options *options = opt->value;
5258
5259 BUG_ON_OPT_NEG(unset);
5260 options->use_color = 1;
5261 options->word_diff = DIFF_WORDS_COLOR;
5262 options->word_regex = arg;
5263 return 0;
5264 }
5265
5266 static int diff_opt_compact_summary(const struct option *opt,
5267 const char *arg, int unset)
5268 {
5269 struct diff_options *options = opt->value;
5270
5271 BUG_ON_OPT_ARG(arg);
5272 if (unset) {
5273 options->flags.stat_with_summary = 0;
5274 } else {
5275 options->flags.stat_with_summary = 1;
5276 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5277 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5278 }
5279 return 0;
5280 }
5281
5282 static int diff_opt_diff_algorithm(const struct option *opt,
5283 const char *arg, int unset)
5284 {
5285 struct diff_options *options = opt->value;
5286
5287 BUG_ON_OPT_NEG(unset);
5288
5289 if (set_diff_algorithm(options, arg))
5290 return error(_("option diff-algorithm accepts \"myers\", "
5291 "\"minimal\", \"patience\" and \"histogram\""));
5292
5293 options->ignore_driver_algorithm = 1;
5294
5295 return 0;
5296 }
5297
5298 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5299 const char *arg, int unset)
5300 {
5301 struct diff_options *options = opt->value;
5302
5303 BUG_ON_OPT_NEG(unset);
5304 BUG_ON_OPT_ARG(arg);
5305
5306 if (set_diff_algorithm(options, opt->long_name))
5307 BUG("available diff algorithms include \"myers\", "
5308 "\"minimal\", \"patience\" and \"histogram\"");
5309
5310 options->ignore_driver_algorithm = 1;
5311
5312 return 0;
5313 }
5314
5315 static int diff_opt_dirstat(const struct option *opt,
5316 const char *arg, int unset)
5317 {
5318 struct diff_options *options = opt->value;
5319
5320 BUG_ON_OPT_NEG(unset);
5321 if (!strcmp(opt->long_name, "cumulative")) {
5322 if (arg)
5323 BUG("how come --cumulative take a value?");
5324 arg = "cumulative";
5325 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5326 parse_dirstat_opt(options, "files");
5327 parse_dirstat_opt(options, arg ? arg : "");
5328 return 0;
5329 }
5330
5331 static int diff_opt_find_copies(const struct option *opt,
5332 const char *arg, int unset)
5333 {
5334 struct diff_options *options = opt->value;
5335
5336 BUG_ON_OPT_NEG(unset);
5337 if (!arg)
5338 arg = "";
5339 options->rename_score = parse_rename_score(&arg);
5340 if (*arg != 0)
5341 return error(_("invalid argument to %s"), opt->long_name);
5342
5343 if (options->detect_rename == DIFF_DETECT_COPY)
5344 options->flags.find_copies_harder = 1;
5345 else
5346 options->detect_rename = DIFF_DETECT_COPY;
5347
5348 return 0;
5349 }
5350
5351 static int diff_opt_find_renames(const struct option *opt,
5352 const char *arg, int unset)
5353 {
5354 struct diff_options *options = opt->value;
5355
5356 BUG_ON_OPT_NEG(unset);
5357 if (!arg)
5358 arg = "";
5359 options->rename_score = parse_rename_score(&arg);
5360 if (*arg != 0)
5361 return error(_("invalid argument to %s"), opt->long_name);
5362
5363 options->detect_rename = DIFF_DETECT_RENAME;
5364 return 0;
5365 }
5366
5367 static int diff_opt_follow(const struct option *opt,
5368 const char *arg, int unset)
5369 {
5370 struct diff_options *options = opt->value;
5371
5372 BUG_ON_OPT_ARG(arg);
5373 if (unset) {
5374 options->flags.follow_renames = 0;
5375 options->flags.default_follow_renames = 0;
5376 } else {
5377 options->flags.follow_renames = 1;
5378 }
5379 return 0;
5380 }
5381
5382 static int diff_opt_ignore_submodules(const struct option *opt,
5383 const char *arg, int unset)
5384 {
5385 struct diff_options *options = opt->value;
5386
5387 BUG_ON_OPT_NEG(unset);
5388 if (!arg)
5389 arg = "all";
5390 options->flags.override_submodule_config = 1;
5391 handle_ignore_submodules_arg(options, arg);
5392 return 0;
5393 }
5394
5395 static int diff_opt_line_prefix(const struct option *opt,
5396 const char *optarg, int unset)
5397 {
5398 struct diff_options *options = opt->value;
5399
5400 BUG_ON_OPT_NEG(unset);
5401 options->line_prefix = optarg;
5402 graph_setup_line_prefix(options);
5403 return 0;
5404 }
5405
5406 static int diff_opt_no_prefix(const struct option *opt,
5407 const char *optarg, int unset)
5408 {
5409 struct diff_options *options = opt->value;
5410
5411 BUG_ON_OPT_NEG(unset);
5412 BUG_ON_OPT_ARG(optarg);
5413 diff_set_noprefix(options);
5414 return 0;
5415 }
5416
5417 static int diff_opt_default_prefix(const struct option *opt,
5418 const char *optarg, int unset)
5419 {
5420 struct diff_options *options = opt->value;
5421
5422 BUG_ON_OPT_NEG(unset);
5423 BUG_ON_OPT_ARG(optarg);
5424 FREE_AND_NULL(diff_src_prefix);
5425 FREE_AND_NULL(diff_dst_prefix);
5426 diff_set_default_prefix(options);
5427 return 0;
5428 }
5429
5430 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5431 const struct option *opt,
5432 const char *arg, int unset)
5433 {
5434 struct diff_options *options = opt->value;
5435 char *path;
5436
5437 BUG_ON_OPT_NEG(unset);
5438 path = prefix_filename(ctx->prefix, arg);
5439 options->file = xfopen(path, "w");
5440 options->close_file = 1;
5441 if (options->use_color != GIT_COLOR_ALWAYS)
5442 options->use_color = GIT_COLOR_NEVER;
5443 free(path);
5444 return 0;
5445 }
5446
5447 static int diff_opt_patience(const struct option *opt,
5448 const char *arg, int unset)
5449 {
5450 struct diff_options *options = opt->value;
5451 int i;
5452
5453 BUG_ON_OPT_NEG(unset);
5454 BUG_ON_OPT_ARG(arg);
5455 /*
5456 * Both --patience and --anchored use PATIENCE_DIFF
5457 * internally, so remove any anchors previously
5458 * specified.
5459 */
5460 for (i = 0; i < options->anchors_nr; i++)
5461 free(options->anchors[i]);
5462 options->anchors_nr = 0;
5463 options->ignore_driver_algorithm = 1;
5464
5465 return set_diff_algorithm(options, "patience");
5466 }
5467
5468 static int diff_opt_ignore_regex(const struct option *opt,
5469 const char *arg, int unset)
5470 {
5471 struct diff_options *options = opt->value;
5472 regex_t *regex;
5473
5474 BUG_ON_OPT_NEG(unset);
5475
5476 regex = xmalloc(sizeof(*regex));
5477 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE)) {
5478 free(regex);
5479 return error(_("invalid regex given to -I: '%s'"), arg);
5480 }
5481
5482 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5483 options->ignore_regex_alloc);
5484 options->ignore_regex[options->ignore_regex_nr++] = regex;
5485 return 0;
5486 }
5487
5488 static int diff_opt_pickaxe_regex(const struct option *opt,
5489 const char *arg, int unset)
5490 {
5491 struct diff_options *options = opt->value;
5492
5493 BUG_ON_OPT_NEG(unset);
5494 options->pickaxe = arg;
5495 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5496 return 0;
5497 }
5498
5499 static int diff_opt_pickaxe_string(const struct option *opt,
5500 const char *arg, int unset)
5501 {
5502 struct diff_options *options = opt->value;
5503
5504 BUG_ON_OPT_NEG(unset);
5505 options->pickaxe = arg;
5506 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5507 return 0;
5508 }
5509
5510 static int diff_opt_relative(const struct option *opt,
5511 const char *arg, int unset)
5512 {
5513 struct diff_options *options = opt->value;
5514
5515 options->flags.relative_name = !unset;
5516 if (arg)
5517 options->prefix = arg;
5518 return 0;
5519 }
5520
5521 static int diff_opt_submodule(const struct option *opt,
5522 const char *arg, int unset)
5523 {
5524 struct diff_options *options = opt->value;
5525
5526 BUG_ON_OPT_NEG(unset);
5527 if (!arg)
5528 arg = "log";
5529 if (parse_submodule_params(options, arg))
5530 return error(_("failed to parse --submodule option parameter: '%s'"),
5531 arg);
5532 return 0;
5533 }
5534
5535 static int diff_opt_textconv(const struct option *opt,
5536 const char *arg, int unset)
5537 {
5538 struct diff_options *options = opt->value;
5539
5540 BUG_ON_OPT_ARG(arg);
5541 if (unset) {
5542 options->flags.allow_textconv = 0;
5543 } else {
5544 options->flags.allow_textconv = 1;
5545 options->flags.textconv_set_via_cmdline = 1;
5546 }
5547 return 0;
5548 }
5549
5550 static int diff_opt_unified(const struct option *opt,
5551 const char *arg, int unset)
5552 {
5553 struct diff_options *options = opt->value;
5554 char *s;
5555
5556 BUG_ON_OPT_NEG(unset);
5557
5558 if (arg) {
5559 options->context = strtol(arg, &s, 10);
5560 if (*s)
5561 return error(_("%s expects a numerical value"), "--unified");
5562 }
5563 enable_patch_output(&options->output_format);
5564
5565 return 0;
5566 }
5567
5568 static int diff_opt_word_diff(const struct option *opt,
5569 const char *arg, int unset)
5570 {
5571 struct diff_options *options = opt->value;
5572
5573 BUG_ON_OPT_NEG(unset);
5574 if (arg) {
5575 if (!strcmp(arg, "plain"))
5576 options->word_diff = DIFF_WORDS_PLAIN;
5577 else if (!strcmp(arg, "color")) {
5578 options->use_color = 1;
5579 options->word_diff = DIFF_WORDS_COLOR;
5580 }
5581 else if (!strcmp(arg, "porcelain"))
5582 options->word_diff = DIFF_WORDS_PORCELAIN;
5583 else if (!strcmp(arg, "none"))
5584 options->word_diff = DIFF_WORDS_NONE;
5585 else
5586 return error(_("bad --word-diff argument: %s"), arg);
5587 } else {
5588 if (options->word_diff == DIFF_WORDS_NONE)
5589 options->word_diff = DIFF_WORDS_PLAIN;
5590 }
5591 return 0;
5592 }
5593
5594 static int diff_opt_word_diff_regex(const struct option *opt,
5595 const char *arg, int unset)
5596 {
5597 struct diff_options *options = opt->value;
5598
5599 BUG_ON_OPT_NEG(unset);
5600 if (options->word_diff == DIFF_WORDS_NONE)
5601 options->word_diff = DIFF_WORDS_PLAIN;
5602 options->word_regex = arg;
5603 return 0;
5604 }
5605
5606 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5607 {
5608 struct diff_options *options = opt->value;
5609
5610 BUG_ON_OPT_NEG(unset);
5611 if (!strcmp(opt->long_name, "skip-to"))
5612 options->skip_instead_of_rotate = 1;
5613 else
5614 options->skip_instead_of_rotate = 0;
5615 options->rotate_to = arg;
5616 return 0;
5617 }
5618
5619 /*
5620 * Consider adding new flags to __git_diff_common_options
5621 * in contrib/completion/git-completion.bash
5622 */
5623 struct option *add_diff_options(const struct option *opts,
5624 struct diff_options *options)
5625 {
5626 struct option parseopts[] = {
5627 OPT_GROUP(N_("Diff output format options")),
5628 OPT_BITOP('p', "patch", &options->output_format,
5629 N_("generate patch"),
5630 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5631 OPT_SET_INT('s', "no-patch", &options->output_format,
5632 N_("suppress diff output"), DIFF_FORMAT_NO_OUTPUT),
5633 OPT_BITOP('u', NULL, &options->output_format,
5634 N_("generate patch"),
5635 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5636 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5637 N_("generate diffs with <n> lines context"),
5638 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5639 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5640 N_("generate diffs with <n> lines context")),
5641 OPT_BITOP(0, "raw", &options->output_format,
5642 N_("generate the diff in raw format"),
5643 DIFF_FORMAT_RAW, DIFF_FORMAT_NO_OUTPUT),
5644 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5645 N_("synonym for '-p --raw'"),
5646 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5647 DIFF_FORMAT_NO_OUTPUT),
5648 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5649 N_("synonym for '-p --stat'"),
5650 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5651 DIFF_FORMAT_NO_OUTPUT),
5652 OPT_BITOP(0, "numstat", &options->output_format,
5653 N_("machine friendly --stat"),
5654 DIFF_FORMAT_NUMSTAT, DIFF_FORMAT_NO_OUTPUT),
5655 OPT_BITOP(0, "shortstat", &options->output_format,
5656 N_("output only the last line of --stat"),
5657 DIFF_FORMAT_SHORTSTAT, DIFF_FORMAT_NO_OUTPUT),
5658 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1>,<param2>..."),
5659 N_("output the distribution of relative amount of changes for each sub-directory"),
5660 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5661 diff_opt_dirstat),
5662 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5663 N_("synonym for --dirstat=cumulative"),
5664 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5665 diff_opt_dirstat),
5666 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1>,<param2>..."),
5667 N_("synonym for --dirstat=files,<param1>,<param2>..."),
5668 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5669 diff_opt_dirstat),
5670 OPT_BIT_F(0, "check", &options->output_format,
5671 N_("warn if changes introduce conflict markers or whitespace errors"),
5672 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5673 OPT_BITOP(0, "summary", &options->output_format,
5674 N_("condensed summary such as creations, renames and mode changes"),
5675 DIFF_FORMAT_SUMMARY, DIFF_FORMAT_NO_OUTPUT),
5676 OPT_BIT_F(0, "name-only", &options->output_format,
5677 N_("show only names of changed files"),
5678 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5679 OPT_BIT_F(0, "name-status", &options->output_format,
5680 N_("show only names and status of changed files"),
5681 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5682 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5683 N_("generate diffstat"),
5684 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5685 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5686 N_("generate diffstat with a given width"),
5687 PARSE_OPT_NONEG, diff_opt_stat),
5688 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5689 N_("generate diffstat with a given name width"),
5690 PARSE_OPT_NONEG, diff_opt_stat),
5691 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5692 N_("generate diffstat with a given graph width"),
5693 PARSE_OPT_NONEG, diff_opt_stat),
5694 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5695 N_("generate diffstat with limited lines"),
5696 PARSE_OPT_NONEG, diff_opt_stat),
5697 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5698 N_("generate compact summary in diffstat"),
5699 PARSE_OPT_NOARG, diff_opt_compact_summary),
5700 OPT_CALLBACK_F(0, "binary", options, NULL,
5701 N_("output a binary diff that can be applied"),
5702 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5703 OPT_BOOL(0, "full-index", &options->flags.full_index,
5704 N_("show full pre- and post-image object names on the \"index\" lines")),
5705 OPT_COLOR_FLAG(0, "color", &options->use_color,
5706 N_("show colored diff")),
5707 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5708 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5709 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5710 OPT_SET_INT('z', NULL, &options->line_termination,
5711 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5712 0),
5713 OPT__ABBREV(&options->abbrev),
5714 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5715 N_("show the given source prefix instead of \"a/\""),
5716 PARSE_OPT_NONEG),
5717 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5718 N_("show the given destination prefix instead of \"b/\""),
5719 PARSE_OPT_NONEG),
5720 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5721 N_("prepend an additional prefix to every line of output"),
5722 PARSE_OPT_NONEG, diff_opt_line_prefix),
5723 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5724 N_("do not show any source or destination prefix"),
5725 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5726 OPT_CALLBACK_F(0, "default-prefix", options, NULL,
5727 N_("use default prefixes a/ and b/"),
5728 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_default_prefix),
5729 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5730 N_("show context between diff hunks up to the specified number of lines"),
5731 PARSE_OPT_NONEG),
5732 OPT_CALLBACK_F(0, "output-indicator-new",
5733 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5734 N_("<char>"),
5735 N_("specify the character to indicate a new line instead of '+'"),
5736 PARSE_OPT_NONEG, diff_opt_char),
5737 OPT_CALLBACK_F(0, "output-indicator-old",
5738 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5739 N_("<char>"),
5740 N_("specify the character to indicate an old line instead of '-'"),
5741 PARSE_OPT_NONEG, diff_opt_char),
5742 OPT_CALLBACK_F(0, "output-indicator-context",
5743 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5744 N_("<char>"),
5745 N_("specify the character to indicate a context instead of ' '"),
5746 PARSE_OPT_NONEG, diff_opt_char),
5747
5748 OPT_GROUP(N_("Diff rename options")),
5749 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5750 N_("break complete rewrite changes into pairs of delete and create"),
5751 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5752 diff_opt_break_rewrites),
5753 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5754 N_("detect renames"),
5755 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5756 diff_opt_find_renames),
5757 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5758 N_("omit the preimage for deletes"),
5759 1, PARSE_OPT_NONEG),
5760 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5761 N_("detect copies"),
5762 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5763 diff_opt_find_copies),
5764 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5765 N_("use unmodified files as source to find copies")),
5766 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5767 N_("disable rename detection"),
5768 0, PARSE_OPT_NONEG),
5769 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5770 N_("use empty blobs as rename source")),
5771 OPT_CALLBACK_F(0, "follow", options, NULL,
5772 N_("continue listing the history of a file beyond renames"),
5773 PARSE_OPT_NOARG, diff_opt_follow),
5774 OPT_INTEGER('l', NULL, &options->rename_limit,
5775 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5776
5777 OPT_GROUP(N_("Diff algorithm options")),
5778 OPT_CALLBACK_F(0, "minimal", options, NULL,
5779 N_("produce the smallest possible diff"),
5780 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5781 diff_opt_diff_algorithm_no_arg),
5782 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5783 N_("ignore whitespace when comparing lines"),
5784 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5785 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5786 N_("ignore changes in amount of whitespace"),
5787 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5788 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5789 N_("ignore changes in whitespace at EOL"),
5790 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5791 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5792 N_("ignore carrier-return at the end of line"),
5793 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5794 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5795 N_("ignore changes whose lines are all blank"),
5796 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5797 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5798 N_("ignore changes whose all lines match <regex>"),
5799 0, diff_opt_ignore_regex),
5800 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5801 N_("heuristic to shift diff hunk boundaries for easy reading"),
5802 XDF_INDENT_HEURISTIC),
5803 OPT_CALLBACK_F(0, "patience", options, NULL,
5804 N_("generate diff using the \"patience diff\" algorithm"),
5805 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5806 diff_opt_patience),
5807 OPT_CALLBACK_F(0, "histogram", options, NULL,
5808 N_("generate diff using the \"histogram diff\" algorithm"),
5809 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5810 diff_opt_diff_algorithm_no_arg),
5811 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5812 N_("choose a diff algorithm"),
5813 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5814 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5815 N_("generate diff using the \"anchored diff\" algorithm"),
5816 PARSE_OPT_NONEG, diff_opt_anchored),
5817 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5818 N_("show word diff, using <mode> to delimit changed words"),
5819 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5820 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5821 N_("use <regex> to decide what a word is"),
5822 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5823 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5824 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5825 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5826 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5827 N_("moved lines of code are colored differently"),
5828 PARSE_OPT_OPTARG, diff_opt_color_moved),
5829 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5830 N_("how white spaces are ignored in --color-moved"),
5831 0, diff_opt_color_moved_ws),
5832
5833 OPT_GROUP(N_("Other diff options")),
5834 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5835 N_("when run from subdir, exclude changes outside and show relative paths"),
5836 PARSE_OPT_OPTARG,
5837 diff_opt_relative),
5838 OPT_BOOL('a', "text", &options->flags.text,
5839 N_("treat all files as text")),
5840 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5841 N_("swap two inputs, reverse the diff")),
5842 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5843 N_("exit with 1 if there were differences, 0 otherwise")),
5844 OPT_BOOL(0, "quiet", &options->flags.quick,
5845 N_("disable all output of the program")),
5846 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5847 N_("allow an external diff helper to be executed")),
5848 OPT_CALLBACK_F(0, "textconv", options, NULL,
5849 N_("run external text conversion filters when comparing binary files"),
5850 PARSE_OPT_NOARG, diff_opt_textconv),
5851 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5852 N_("ignore changes to submodules in the diff generation"),
5853 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5854 diff_opt_ignore_submodules),
5855 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5856 N_("specify how differences in submodules are shown"),
5857 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5858 diff_opt_submodule),
5859 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5860 N_("hide 'git add -N' entries from the index"),
5861 1, PARSE_OPT_NONEG),
5862 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5863 N_("treat 'git add -N' entries as real in the index"),
5864 0, PARSE_OPT_NONEG),
5865 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5866 N_("look for differences that change the number of occurrences of the specified string"),
5867 0, diff_opt_pickaxe_string),
5868 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5869 N_("look for differences that change the number of occurrences of the specified regex"),
5870 0, diff_opt_pickaxe_regex),
5871 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5872 N_("show all changes in the changeset with -S or -G"),
5873 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5874 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5875 N_("treat <string> in -S as extended POSIX regular expression"),
5876 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5877 OPT_FILENAME('O', NULL, &options->orderfile,
5878 N_("control the order in which files appear in the output")),
5879 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5880 N_("show the change in the specified path first"),
5881 PARSE_OPT_NONEG, diff_opt_rotate_to),
5882 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5883 N_("skip the output to the specified path"),
5884 PARSE_OPT_NONEG, diff_opt_rotate_to),
5885 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5886 N_("look for differences that change the number of occurrences of the specified object"),
5887 PARSE_OPT_NONEG, diff_opt_find_object),
5888 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5889 N_("select files by diff type"),
5890 PARSE_OPT_NONEG, diff_opt_diff_filter),
5891 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5892 N_("output to a specific file"),
5893 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5894
5895 OPT_END()
5896 };
5897
5898 return parse_options_concat(opts, parseopts);
5899 }
5900
5901 int diff_opt_parse(struct diff_options *options,
5902 const char **av, int ac, const char *prefix)
5903 {
5904 struct option no_options[] = { OPT_END() };
5905 struct option *parseopts = add_diff_options(no_options, options);
5906
5907 if (!prefix)
5908 prefix = "";
5909
5910 ac = parse_options(ac, av, prefix, parseopts, NULL,
5911 PARSE_OPT_KEEP_DASHDASH |
5912 PARSE_OPT_KEEP_UNKNOWN_OPT |
5913 PARSE_OPT_NO_INTERNAL_HELP |
5914 PARSE_OPT_ONE_SHOT |
5915 PARSE_OPT_STOP_AT_NON_OPTION);
5916 free(parseopts);
5917
5918 return ac;
5919 }
5920
5921 int parse_rename_score(const char **cp_p)
5922 {
5923 unsigned long num, scale;
5924 int ch, dot;
5925 const char *cp = *cp_p;
5926
5927 num = 0;
5928 scale = 1;
5929 dot = 0;
5930 for (;;) {
5931 ch = *cp;
5932 if ( !dot && ch == '.' ) {
5933 scale = 1;
5934 dot = 1;
5935 } else if ( ch == '%' ) {
5936 scale = dot ? scale*100 : 100;
5937 cp++; /* % is always at the end */
5938 break;
5939 } else if ( ch >= '0' && ch <= '9' ) {
5940 if ( scale < 100000 ) {
5941 scale *= 10;
5942 num = (num*10) + (ch-'0');
5943 }
5944 } else {
5945 break;
5946 }
5947 cp++;
5948 }
5949 *cp_p = cp;
5950
5951 /* user says num divided by scale and we say internally that
5952 * is MAX_SCORE * num / scale.
5953 */
5954 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5955 }
5956
5957 struct diff_queue_struct diff_queued_diff;
5958
5959 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5960 {
5961 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5962 queue->queue[queue->nr++] = dp;
5963 }
5964
5965 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5966 struct diff_filespec *one,
5967 struct diff_filespec *two)
5968 {
5969 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5970 dp->one = one;
5971 dp->two = two;
5972 if (queue)
5973 diff_q(queue, dp);
5974 return dp;
5975 }
5976
5977 void diff_free_filepair(struct diff_filepair *p)
5978 {
5979 free_filespec(p->one);
5980 free_filespec(p->two);
5981 free(p);
5982 }
5983
5984 void diff_queue_init(struct diff_queue_struct *q)
5985 {
5986 struct diff_queue_struct blank = DIFF_QUEUE_INIT;
5987 memcpy(q, &blank, sizeof(*q));
5988 }
5989
5990 void diff_queue_clear(struct diff_queue_struct *q)
5991 {
5992 for (int i = 0; i < q->nr; i++)
5993 diff_free_filepair(q->queue[i]);
5994 free(q->queue);
5995 diff_queue_init(q);
5996 }
5997
5998 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5999 {
6000 int abblen;
6001 const char *abbrev;
6002
6003 /* Do we want all 40 hex characters? */
6004 if (len == the_hash_algo->hexsz)
6005 return oid_to_hex(oid);
6006
6007 /* An abbreviated value is fine, possibly followed by an ellipsis. */
6008 abbrev = diff_abbrev_oid(oid, len);
6009
6010 if (!print_sha1_ellipsis())
6011 return abbrev;
6012
6013 abblen = strlen(abbrev);
6014
6015 /*
6016 * In well-behaved cases, where the abbreviated result is the
6017 * same as the requested length, append three dots after the
6018 * abbreviation (hence the whole logic is limited to the case
6019 * where abblen < 37); when the actual abbreviated result is a
6020 * bit longer than the requested length, we reduce the number
6021 * of dots so that they match the well-behaved ones. However,
6022 * if the actual abbreviation is longer than the requested
6023 * length by more than three, we give up on aligning, and add
6024 * three dots anyway, to indicate that the output is not the
6025 * full object name. Yes, this may be suboptimal, but this
6026 * appears only in "diff --raw --abbrev" output and it is not
6027 * worth the effort to change it now. Note that this would
6028 * likely to work fine when the automatic sizing of default
6029 * abbreviation length is used--we would be fed -1 in "len" in
6030 * that case, and will end up always appending three-dots, but
6031 * the automatic sizing is supposed to give abblen that ensures
6032 * uniqueness across all objects (statistically speaking).
6033 */
6034 if (abblen < the_hash_algo->hexsz - 3) {
6035 static char hex[GIT_MAX_HEXSZ + 1];
6036 if (len < abblen && abblen <= len + 2)
6037 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
6038 else
6039 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
6040 return hex;
6041 }
6042
6043 return oid_to_hex(oid);
6044 }
6045
6046 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
6047 {
6048 int line_termination = opt->line_termination;
6049 int inter_name_termination = line_termination ? '\t' : '\0';
6050
6051 fprintf(opt->file, "%s", diff_line_prefix(opt));
6052 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
6053 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
6054 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
6055 fprintf(opt->file, "%s ",
6056 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
6057 }
6058 if (p->score) {
6059 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
6060 inter_name_termination);
6061 } else {
6062 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
6063 }
6064
6065 if (p->status == DIFF_STATUS_COPIED ||
6066 p->status == DIFF_STATUS_RENAMED) {
6067 const char *name_a, *name_b;
6068 name_a = p->one->path;
6069 name_b = p->two->path;
6070 strip_prefix(opt->prefix_length, &name_a, &name_b);
6071 write_name_quoted(name_a, opt->file, inter_name_termination);
6072 write_name_quoted(name_b, opt->file, line_termination);
6073 } else {
6074 const char *name_a, *name_b;
6075 name_a = p->one->mode ? p->one->path : p->two->path;
6076 name_b = NULL;
6077 strip_prefix(opt->prefix_length, &name_a, &name_b);
6078 write_name_quoted(name_a, opt->file, line_termination);
6079 }
6080 }
6081
6082 int diff_unmodified_pair(struct diff_filepair *p)
6083 {
6084 /* This function is written stricter than necessary to support
6085 * the currently implemented transformers, but the idea is to
6086 * let transformers to produce diff_filepairs any way they want,
6087 * and filter and clean them up here before producing the output.
6088 */
6089 struct diff_filespec *one = p->one, *two = p->two;
6090
6091 if (DIFF_PAIR_UNMERGED(p))
6092 return 0; /* unmerged is interesting */
6093
6094 /* deletion, addition, mode or type change
6095 * and rename are all interesting.
6096 */
6097 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
6098 DIFF_PAIR_MODE_CHANGED(p) ||
6099 strcmp(one->path, two->path))
6100 return 0;
6101
6102 /* both are valid and point at the same path. that is, we are
6103 * dealing with a change.
6104 */
6105 if (one->oid_valid && two->oid_valid &&
6106 oideq(&one->oid, &two->oid) &&
6107 !one->dirty_submodule && !two->dirty_submodule)
6108 return 1; /* no change */
6109 if (!one->oid_valid && !two->oid_valid)
6110 return 1; /* both look at the same file on the filesystem. */
6111 return 0;
6112 }
6113
6114 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
6115 {
6116 int include_conflict_headers =
6117 (additional_headers(o, p->one->path) &&
6118 !o->pickaxe_opts &&
6119 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6120
6121 /*
6122 * Check if we can return early without showing a diff. Note that
6123 * diff_filepair only stores {oid, path, mode, is_valid}
6124 * information for each path, and thus diff_unmodified_pair() only
6125 * considers those bits of info. However, we do not want pairs
6126 * created by create_filepairs_for_header_only_notifications()
6127 * (which always look like unmodified pairs) to be ignored, so
6128 * return early if both p is unmodified AND we don't want to
6129 * include_conflict_headers.
6130 */
6131 if (diff_unmodified_pair(p) && !include_conflict_headers)
6132 return;
6133
6134 /* Actually, we can also return early to avoid showing tree diffs */
6135 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6136 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6137 return;
6138
6139 run_diff(p, o);
6140 }
6141
6142 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
6143 struct diffstat_t *diffstat)
6144 {
6145 if (diff_unmodified_pair(p))
6146 return;
6147
6148 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6149 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6150 return; /* no useful stat for tree diffs */
6151
6152 run_diffstat(p, o, diffstat);
6153 }
6154
6155 static void diff_flush_checkdiff(struct diff_filepair *p,
6156 struct diff_options *o)
6157 {
6158 if (diff_unmodified_pair(p))
6159 return;
6160
6161 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6162 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6163 return; /* nothing to check in tree diffs */
6164
6165 run_checkdiff(p, o);
6166 }
6167
6168 int diff_queue_is_empty(struct diff_options *o)
6169 {
6170 struct diff_queue_struct *q = &diff_queued_diff;
6171 int i;
6172 int include_conflict_headers =
6173 (o->additional_path_headers &&
6174 strmap_get_size(o->additional_path_headers) &&
6175 !o->pickaxe_opts &&
6176 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6177
6178 if (include_conflict_headers)
6179 return 0;
6180
6181 for (i = 0; i < q->nr; i++)
6182 if (!diff_unmodified_pair(q->queue[i]))
6183 return 0;
6184 return 1;
6185 }
6186
6187 #if DIFF_DEBUG
6188 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6189 {
6190 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6191 x, one ? one : "",
6192 s->path,
6193 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6194 s->mode,
6195 s->oid_valid ? oid_to_hex(&s->oid) : "");
6196 fprintf(stderr, "queue[%d] %s size %lu\n",
6197 x, one ? one : "",
6198 s->size);
6199 }
6200
6201 void diff_debug_filepair(const struct diff_filepair *p, int i)
6202 {
6203 diff_debug_filespec(p->one, i, "one");
6204 diff_debug_filespec(p->two, i, "two");
6205 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6206 p->score, p->status ? p->status : '?',
6207 p->one->rename_used, p->broken_pair);
6208 }
6209
6210 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6211 {
6212 int i;
6213 if (msg)
6214 fprintf(stderr, "%s\n", msg);
6215 fprintf(stderr, "q->nr = %d\n", q->nr);
6216 for (i = 0; i < q->nr; i++) {
6217 struct diff_filepair *p = q->queue[i];
6218 diff_debug_filepair(p, i);
6219 }
6220 }
6221 #endif
6222
6223 static void diff_resolve_rename_copy(void)
6224 {
6225 int i;
6226 struct diff_filepair *p;
6227 struct diff_queue_struct *q = &diff_queued_diff;
6228
6229 diff_debug_queue("resolve-rename-copy", q);
6230
6231 for (i = 0; i < q->nr; i++) {
6232 p = q->queue[i];
6233 p->status = 0; /* undecided */
6234 if (DIFF_PAIR_UNMERGED(p))
6235 p->status = DIFF_STATUS_UNMERGED;
6236 else if (!DIFF_FILE_VALID(p->one))
6237 p->status = DIFF_STATUS_ADDED;
6238 else if (!DIFF_FILE_VALID(p->two))
6239 p->status = DIFF_STATUS_DELETED;
6240 else if (DIFF_PAIR_TYPE_CHANGED(p))
6241 p->status = DIFF_STATUS_TYPE_CHANGED;
6242
6243 /* from this point on, we are dealing with a pair
6244 * whose both sides are valid and of the same type, i.e.
6245 * either in-place edit or rename/copy edit.
6246 */
6247 else if (DIFF_PAIR_RENAME(p)) {
6248 /*
6249 * A rename might have re-connected a broken
6250 * pair up, causing the pathnames to be the
6251 * same again. If so, that's not a rename at
6252 * all, just a modification..
6253 *
6254 * Otherwise, see if this source was used for
6255 * multiple renames, in which case we decrement
6256 * the count, and call it a copy.
6257 */
6258 if (!strcmp(p->one->path, p->two->path))
6259 p->status = DIFF_STATUS_MODIFIED;
6260 else if (--p->one->rename_used > 0)
6261 p->status = DIFF_STATUS_COPIED;
6262 else
6263 p->status = DIFF_STATUS_RENAMED;
6264 }
6265 else if (!oideq(&p->one->oid, &p->two->oid) ||
6266 p->one->mode != p->two->mode ||
6267 p->one->dirty_submodule ||
6268 p->two->dirty_submodule ||
6269 is_null_oid(&p->one->oid))
6270 p->status = DIFF_STATUS_MODIFIED;
6271 else {
6272 /* This is a "no-change" entry and should not
6273 * happen anymore, but prepare for broken callers.
6274 */
6275 error("feeding unmodified %s to diffcore",
6276 p->one->path);
6277 p->status = DIFF_STATUS_UNKNOWN;
6278 }
6279 }
6280 diff_debug_queue("resolve-rename-copy done", q);
6281 }
6282
6283 static int check_pair_status(struct diff_filepair *p)
6284 {
6285 switch (p->status) {
6286 case DIFF_STATUS_UNKNOWN:
6287 return 0;
6288 case 0:
6289 die("internal error in diff-resolve-rename-copy");
6290 default:
6291 return 1;
6292 }
6293 }
6294
6295 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6296 {
6297 int fmt = opt->output_format;
6298
6299 if (fmt & DIFF_FORMAT_CHECKDIFF)
6300 diff_flush_checkdiff(p, opt);
6301 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6302 diff_flush_raw(p, opt);
6303 else if (fmt & DIFF_FORMAT_NAME) {
6304 const char *name_a, *name_b;
6305 name_a = p->two->path;
6306 name_b = NULL;
6307 strip_prefix(opt->prefix_length, &name_a, &name_b);
6308 fprintf(opt->file, "%s", diff_line_prefix(opt));
6309 write_name_quoted(name_a, opt->file, opt->line_termination);
6310 }
6311
6312 opt->found_changes = 1;
6313 }
6314
6315 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6316 {
6317 struct strbuf sb = STRBUF_INIT;
6318 if (fs->mode)
6319 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6320 else
6321 strbuf_addf(&sb, " %s ", newdelete);
6322
6323 quote_c_style(fs->path, &sb, NULL, 0);
6324 strbuf_addch(&sb, '\n');
6325 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6326 sb.buf, sb.len, 0);
6327 strbuf_release(&sb);
6328 }
6329
6330 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6331 int show_name)
6332 {
6333 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6334 struct strbuf sb = STRBUF_INIT;
6335 strbuf_addf(&sb, " mode change %06o => %06o",
6336 p->one->mode, p->two->mode);
6337 if (show_name) {
6338 strbuf_addch(&sb, ' ');
6339 quote_c_style(p->two->path, &sb, NULL, 0);
6340 }
6341 strbuf_addch(&sb, '\n');
6342 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6343 sb.buf, sb.len, 0);
6344 strbuf_release(&sb);
6345 }
6346 }
6347
6348 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6349 struct diff_filepair *p)
6350 {
6351 struct strbuf sb = STRBUF_INIT;
6352 struct strbuf names = STRBUF_INIT;
6353
6354 pprint_rename(&names, p->one->path, p->two->path);
6355 strbuf_addf(&sb, " %s %s (%d%%)\n",
6356 renamecopy, names.buf, similarity_index(p));
6357 strbuf_release(&names);
6358 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6359 sb.buf, sb.len, 0);
6360 show_mode_change(opt, p, 0);
6361 strbuf_release(&sb);
6362 }
6363
6364 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6365 {
6366 switch(p->status) {
6367 case DIFF_STATUS_DELETED:
6368 show_file_mode_name(opt, "delete", p->one);
6369 break;
6370 case DIFF_STATUS_ADDED:
6371 show_file_mode_name(opt, "create", p->two);
6372 break;
6373 case DIFF_STATUS_COPIED:
6374 show_rename_copy(opt, "copy", p);
6375 break;
6376 case DIFF_STATUS_RENAMED:
6377 show_rename_copy(opt, "rename", p);
6378 break;
6379 default:
6380 if (p->score) {
6381 struct strbuf sb = STRBUF_INIT;
6382 strbuf_addstr(&sb, " rewrite ");
6383 quote_c_style(p->two->path, &sb, NULL, 0);
6384 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6385 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6386 sb.buf, sb.len, 0);
6387 strbuf_release(&sb);
6388 }
6389 show_mode_change(opt, p, !p->score);
6390 break;
6391 }
6392 }
6393
6394 struct patch_id_t {
6395 git_hash_ctx *ctx;
6396 int patchlen;
6397 };
6398
6399 static int remove_space(char *line, int len)
6400 {
6401 int i;
6402 char *dst = line;
6403 unsigned char c;
6404
6405 for (i = 0; i < len; i++)
6406 if (!isspace((c = line[i])))
6407 *dst++ = c;
6408
6409 return dst - line;
6410 }
6411
6412 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6413 {
6414 unsigned char hash[GIT_MAX_RAWSZ];
6415 unsigned short carry = 0;
6416 int i;
6417
6418 the_hash_algo->final_fn(hash, ctx);
6419 the_hash_algo->init_fn(ctx);
6420 /* 20-byte sum, with carry */
6421 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6422 carry += result->hash[i] + hash[i];
6423 result->hash[i] = carry;
6424 carry >>= 8;
6425 }
6426 }
6427
6428 static int patch_id_consume(void *priv, char *line, unsigned long len)
6429 {
6430 struct patch_id_t *data = priv;
6431 int new_len;
6432
6433 if (len > 12 && starts_with(line, "\\ "))
6434 return 0;
6435 new_len = remove_space(line, len);
6436
6437 the_hash_algo->update_fn(data->ctx, line, new_len);
6438 data->patchlen += new_len;
6439 return 0;
6440 }
6441
6442 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6443 {
6444 the_hash_algo->update_fn(ctx, str, strlen(str));
6445 }
6446
6447 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6448 {
6449 /* large enough for 2^32 in octal */
6450 char buf[12];
6451 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6452 the_hash_algo->update_fn(ctx, buf, len);
6453 }
6454
6455 /* returns 0 upon success, and writes result into oid */
6456 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6457 {
6458 struct diff_queue_struct *q = &diff_queued_diff;
6459 int i;
6460 git_hash_ctx ctx;
6461 struct patch_id_t data;
6462
6463 the_hash_algo->init_fn(&ctx);
6464 memset(&data, 0, sizeof(struct patch_id_t));
6465 data.ctx = &ctx;
6466 oidclr(oid, the_repository->hash_algo);
6467
6468 for (i = 0; i < q->nr; i++) {
6469 xpparam_t xpp;
6470 xdemitconf_t xecfg;
6471 mmfile_t mf1, mf2;
6472 struct diff_filepair *p = q->queue[i];
6473 int len1, len2;
6474
6475 memset(&xpp, 0, sizeof(xpp));
6476 memset(&xecfg, 0, sizeof(xecfg));
6477 if (p->status == 0)
6478 return error("internal diff status error");
6479 if (p->status == DIFF_STATUS_UNKNOWN)
6480 continue;
6481 if (diff_unmodified_pair(p))
6482 continue;
6483 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6484 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6485 continue;
6486 if (DIFF_PAIR_UNMERGED(p))
6487 continue;
6488
6489 diff_fill_oid_info(p->one, options->repo->index);
6490 diff_fill_oid_info(p->two, options->repo->index);
6491
6492 len1 = remove_space(p->one->path, strlen(p->one->path));
6493 len2 = remove_space(p->two->path, strlen(p->two->path));
6494 patch_id_add_string(&ctx, "diff--git");
6495 patch_id_add_string(&ctx, "a/");
6496 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6497 patch_id_add_string(&ctx, "b/");
6498 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6499
6500 if (p->one->mode == 0) {
6501 patch_id_add_string(&ctx, "newfilemode");
6502 patch_id_add_mode(&ctx, p->two->mode);
6503 } else if (p->two->mode == 0) {
6504 patch_id_add_string(&ctx, "deletedfilemode");
6505 patch_id_add_mode(&ctx, p->one->mode);
6506 } else if (p->one->mode != p->two->mode) {
6507 patch_id_add_string(&ctx, "oldmode");
6508 patch_id_add_mode(&ctx, p->one->mode);
6509 patch_id_add_string(&ctx, "newmode");
6510 patch_id_add_mode(&ctx, p->two->mode);
6511 }
6512
6513 if (diff_header_only) {
6514 /* don't do anything since we're only populating header info */
6515 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6516 diff_filespec_is_binary(options->repo, p->two)) {
6517 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6518 the_hash_algo->hexsz);
6519 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6520 the_hash_algo->hexsz);
6521 } else {
6522 if (p->one->mode == 0) {
6523 patch_id_add_string(&ctx, "---/dev/null");
6524 patch_id_add_string(&ctx, "+++b/");
6525 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6526 } else if (p->two->mode == 0) {
6527 patch_id_add_string(&ctx, "---a/");
6528 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6529 patch_id_add_string(&ctx, "+++/dev/null");
6530 } else {
6531 patch_id_add_string(&ctx, "---a/");
6532 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6533 patch_id_add_string(&ctx, "+++b/");
6534 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6535 }
6536
6537 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6538 fill_mmfile(options->repo, &mf2, p->two) < 0)
6539 return error("unable to read files to diff");
6540 xpp.flags = 0;
6541 xecfg.ctxlen = 3;
6542 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6543 if (xdi_diff_outf(&mf1, &mf2, NULL,
6544 patch_id_consume, &data, &xpp, &xecfg))
6545 return error("unable to generate patch-id diff for %s",
6546 p->one->path);
6547 }
6548 flush_one_hunk(oid, &ctx);
6549 }
6550
6551 return 0;
6552 }
6553
6554 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6555 {
6556 struct diff_queue_struct *q = &diff_queued_diff;
6557 int result = diff_get_patch_id(options, oid, diff_header_only);
6558
6559 diff_queue_clear(q);
6560
6561 return result;
6562 }
6563
6564 static int is_summary_empty(const struct diff_queue_struct *q)
6565 {
6566 int i;
6567
6568 for (i = 0; i < q->nr; i++) {
6569 const struct diff_filepair *p = q->queue[i];
6570
6571 switch (p->status) {
6572 case DIFF_STATUS_DELETED:
6573 case DIFF_STATUS_ADDED:
6574 case DIFF_STATUS_COPIED:
6575 case DIFF_STATUS_RENAMED:
6576 return 0;
6577 default:
6578 if (p->score)
6579 return 0;
6580 if (p->one->mode && p->two->mode &&
6581 p->one->mode != p->two->mode)
6582 return 0;
6583 break;
6584 }
6585 }
6586 return 1;
6587 }
6588
6589 static const char rename_limit_warning[] =
6590 N_("exhaustive rename detection was skipped due to too many files.");
6591
6592 static const char degrade_cc_to_c_warning[] =
6593 N_("only found copies from modified paths due to too many files.");
6594
6595 static const char rename_limit_advice[] =
6596 N_("you may want to set your %s variable to at least "
6597 "%d and retry the command.");
6598
6599 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6600 {
6601 fflush(stdout);
6602 if (degraded_cc)
6603 warning(_(degrade_cc_to_c_warning));
6604 else if (needed)
6605 warning(_(rename_limit_warning));
6606 else
6607 return;
6608 if (0 < needed)
6609 warning(_(rename_limit_advice), varname, needed);
6610 }
6611
6612 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6613 {
6614 struct strset present;
6615 struct diff_queue_struct *q = &diff_queued_diff;
6616 struct hashmap_iter iter;
6617 struct strmap_entry *e;
6618 int i;
6619
6620 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6621
6622 /*
6623 * Find out which paths exist in diff_queued_diff, preferring
6624 * one->path for any pair that has multiple paths.
6625 */
6626 for (i = 0; i < q->nr; i++) {
6627 struct diff_filepair *p = q->queue[i];
6628 char *path = p->one->path ? p->one->path : p->two->path;
6629
6630 if (strmap_contains(o->additional_path_headers, path))
6631 strset_add(&present, path);
6632 }
6633
6634 /*
6635 * Loop over paths in additional_path_headers; for each NOT already
6636 * in diff_queued_diff, create a synthetic filepair and insert that
6637 * into diff_queued_diff.
6638 */
6639 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6640 if (!strset_contains(&present, e->key)) {
6641 struct diff_filespec *one, *two;
6642 struct diff_filepair *p;
6643
6644 one = alloc_filespec(e->key);
6645 two = alloc_filespec(e->key);
6646 fill_filespec(one, null_oid(), 0, 0);
6647 fill_filespec(two, null_oid(), 0, 0);
6648 p = diff_queue(q, one, two);
6649 p->status = DIFF_STATUS_MODIFIED;
6650 }
6651 }
6652
6653 /* Re-sort the filepairs */
6654 diffcore_fix_diff_index();
6655
6656 /* Cleanup */
6657 strset_clear(&present);
6658 }
6659
6660 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6661 {
6662 int i;
6663 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6664 struct diff_queue_struct *q = &diff_queued_diff;
6665
6666 if (WSEH_NEW & WS_RULE_MASK)
6667 BUG("WS rules bit mask overlaps with diff symbol flags");
6668
6669 if (o->color_moved)
6670 o->emitted_symbols = &esm;
6671
6672 if (o->additional_path_headers)
6673 create_filepairs_for_header_only_notifications(o);
6674
6675 for (i = 0; i < q->nr; i++) {
6676 struct diff_filepair *p = q->queue[i];
6677 if (check_pair_status(p))
6678 diff_flush_patch(p, o);
6679 }
6680
6681 if (o->emitted_symbols) {
6682 if (o->color_moved) {
6683 struct mem_pool entry_pool;
6684 struct moved_entry_list *entry_list;
6685
6686 mem_pool_init(&entry_pool, 1024 * 1024);
6687 entry_list = add_lines_to_move_detection(o,
6688 &entry_pool);
6689 mark_color_as_moved(o, entry_list);
6690 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6691 dim_moved_lines(o);
6692
6693 mem_pool_discard(&entry_pool, 0);
6694 free(entry_list);
6695 }
6696
6697 for (i = 0; i < esm.nr; i++)
6698 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6699
6700 for (i = 0; i < esm.nr; i++)
6701 free((void *)esm.buf[i].line);
6702 esm.nr = 0;
6703
6704 o->emitted_symbols = NULL;
6705 }
6706 }
6707
6708 static void diff_free_file(struct diff_options *options)
6709 {
6710 if (options->close_file && options->file) {
6711 fclose(options->file);
6712 options->file = NULL;
6713 }
6714 }
6715
6716 static void diff_free_ignore_regex(struct diff_options *options)
6717 {
6718 int i;
6719
6720 for (i = 0; i < options->ignore_regex_nr; i++) {
6721 regfree(options->ignore_regex[i]);
6722 free(options->ignore_regex[i]);
6723 }
6724
6725 FREE_AND_NULL(options->ignore_regex);
6726 options->ignore_regex_nr = 0;
6727 }
6728
6729 void diff_free(struct diff_options *options)
6730 {
6731 if (options->no_free)
6732 return;
6733
6734 if (options->objfind) {
6735 oidset_clear(options->objfind);
6736 FREE_AND_NULL(options->objfind);
6737 }
6738
6739 FREE_AND_NULL(options->orderfile);
6740 for (size_t i = 0; i < options->anchors_nr; i++)
6741 free(options->anchors[i]);
6742 FREE_AND_NULL(options->anchors);
6743 options->anchors_nr = options->anchors_alloc = 0;
6744
6745 diff_free_file(options);
6746 diff_free_ignore_regex(options);
6747 clear_pathspec(&options->pathspec);
6748 }
6749
6750 void diff_flush(struct diff_options *options)
6751 {
6752 struct diff_queue_struct *q = &diff_queued_diff;
6753 int i, output_format = options->output_format;
6754 int separator = 0;
6755 int dirstat_by_line = 0;
6756
6757 /*
6758 * Order: raw, stat, summary, patch
6759 * or: name/name-status/checkdiff (other bits clear)
6760 */
6761 if (!q->nr && !options->additional_path_headers)
6762 goto free_queue;
6763
6764 if (output_format & (DIFF_FORMAT_RAW |
6765 DIFF_FORMAT_NAME |
6766 DIFF_FORMAT_NAME_STATUS |
6767 DIFF_FORMAT_CHECKDIFF)) {
6768 for (i = 0; i < q->nr; i++) {
6769 struct diff_filepair *p = q->queue[i];
6770 if (check_pair_status(p))
6771 flush_one_pair(p, options);
6772 }
6773 separator++;
6774 }
6775
6776 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6777 dirstat_by_line = 1;
6778
6779 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6780 dirstat_by_line) {
6781 struct diffstat_t diffstat;
6782
6783 compute_diffstat(options, &diffstat, q);
6784 if (output_format & DIFF_FORMAT_NUMSTAT)
6785 show_numstat(&diffstat, options);
6786 if (output_format & DIFF_FORMAT_DIFFSTAT)
6787 show_stats(&diffstat, options);
6788 if (output_format & DIFF_FORMAT_SHORTSTAT)
6789 show_shortstats(&diffstat, options);
6790 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6791 show_dirstat_by_line(&diffstat, options);
6792 free_diffstat_info(&diffstat);
6793 separator++;
6794 }
6795 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6796 show_dirstat(options);
6797
6798 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6799 for (i = 0; i < q->nr; i++) {
6800 diff_summary(options, q->queue[i]);
6801 }
6802 separator++;
6803 }
6804
6805 if (output_format & DIFF_FORMAT_PATCH) {
6806 if (separator) {
6807 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6808 if (options->stat_sep)
6809 /* attach patch instead of inline */
6810 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6811 NULL, 0, 0);
6812 }
6813
6814 diff_flush_patch_all_file_pairs(options);
6815 }
6816
6817 if (output_format & DIFF_FORMAT_CALLBACK)
6818 options->format_callback(q, options, options->format_callback_data);
6819
6820 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6821 options->flags.exit_with_status &&
6822 options->flags.diff_from_contents) {
6823 /*
6824 * run diff_flush_patch for the exit status. setting
6825 * options->file to /dev/null should be safe, because we
6826 * aren't supposed to produce any output anyway.
6827 */
6828 diff_free_file(options);
6829 options->file = xfopen("/dev/null", "w");
6830 options->close_file = 1;
6831 options->color_moved = 0;
6832 for (i = 0; i < q->nr; i++) {
6833 struct diff_filepair *p = q->queue[i];
6834 if (check_pair_status(p))
6835 diff_flush_patch(p, options);
6836 if (options->found_changes)
6837 break;
6838 }
6839 }
6840
6841 free_queue:
6842 diff_queue_clear(q);
6843 diff_free(options);
6844
6845 /*
6846 * Report the content-level differences with HAS_CHANGES;
6847 * diff_addremove/diff_change does not set the bit when
6848 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6849 */
6850 if (options->flags.diff_from_contents) {
6851 if (options->found_changes)
6852 options->flags.has_changes = 1;
6853 else
6854 options->flags.has_changes = 0;
6855 }
6856 }
6857
6858 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6859 {
6860 return (((p->status == DIFF_STATUS_MODIFIED) &&
6861 ((p->score &&
6862 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6863 (!p->score &&
6864 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6865 ((p->status != DIFF_STATUS_MODIFIED) &&
6866 filter_bit_tst(p->status, options)));
6867 }
6868
6869 static void diffcore_apply_filter(struct diff_options *options)
6870 {
6871 int i;
6872 struct diff_queue_struct *q = &diff_queued_diff;
6873 struct diff_queue_struct outq = DIFF_QUEUE_INIT;
6874
6875 if (!options->filter)
6876 return;
6877
6878 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6879 int found;
6880 for (i = found = 0; !found && i < q->nr; i++) {
6881 if (match_filter(options, q->queue[i]))
6882 found++;
6883 }
6884 if (found)
6885 return;
6886
6887 /* otherwise we will clear the whole queue
6888 * by copying the empty outq at the end of this
6889 * function, but first clear the current entries
6890 * in the queue.
6891 */
6892 for (i = 0; i < q->nr; i++)
6893 diff_free_filepair(q->queue[i]);
6894 }
6895 else {
6896 /* Only the matching ones */
6897 for (i = 0; i < q->nr; i++) {
6898 struct diff_filepair *p = q->queue[i];
6899 if (match_filter(options, p))
6900 diff_q(&outq, p);
6901 else
6902 diff_free_filepair(p);
6903 }
6904 }
6905 free(q->queue);
6906 *q = outq;
6907 }
6908
6909 /* Check whether two filespecs with the same mode and size are identical */
6910 static int diff_filespec_is_identical(struct repository *r,
6911 struct diff_filespec *one,
6912 struct diff_filespec *two)
6913 {
6914 if (S_ISGITLINK(one->mode))
6915 return 0;
6916 if (diff_populate_filespec(r, one, NULL))
6917 return 0;
6918 if (diff_populate_filespec(r, two, NULL))
6919 return 0;
6920 return !memcmp(one->data, two->data, one->size);
6921 }
6922
6923 static int diff_filespec_check_stat_unmatch(struct repository *r,
6924 struct diff_filepair *p)
6925 {
6926 struct diff_populate_filespec_options dpf_options = {
6927 .check_size_only = 1,
6928 .missing_object_cb = diff_queued_diff_prefetch,
6929 .missing_object_data = r,
6930 };
6931
6932 if (p->done_skip_stat_unmatch)
6933 return p->skip_stat_unmatch_result;
6934
6935 p->done_skip_stat_unmatch = 1;
6936 p->skip_stat_unmatch_result = 0;
6937 /*
6938 * 1. Entries that come from stat info dirtiness
6939 * always have both sides (iow, not create/delete),
6940 * one side of the object name is unknown, with
6941 * the same mode and size. Keep the ones that
6942 * do not match these criteria. They have real
6943 * differences.
6944 *
6945 * 2. At this point, the file is known to be modified,
6946 * with the same mode and size, and the object
6947 * name of one side is unknown. Need to inspect
6948 * the identical contents.
6949 */
6950 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6951 !DIFF_FILE_VALID(p->two) ||
6952 (p->one->oid_valid && p->two->oid_valid) ||
6953 (p->one->mode != p->two->mode) ||
6954 diff_populate_filespec(r, p->one, &dpf_options) ||
6955 diff_populate_filespec(r, p->two, &dpf_options) ||
6956 (p->one->size != p->two->size) ||
6957 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6958 p->skip_stat_unmatch_result = 1;
6959 return p->skip_stat_unmatch_result;
6960 }
6961
6962 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6963 {
6964 int i;
6965 struct diff_queue_struct *q = &diff_queued_diff;
6966 struct diff_queue_struct outq = DIFF_QUEUE_INIT;
6967
6968 for (i = 0; i < q->nr; i++) {
6969 struct diff_filepair *p = q->queue[i];
6970
6971 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6972 diff_q(&outq, p);
6973 else {
6974 /*
6975 * The caller can subtract 1 from skip_stat_unmatch
6976 * to determine how many paths were dirty only
6977 * due to stat info mismatch.
6978 */
6979 if (!diffopt->flags.no_index)
6980 diffopt->skip_stat_unmatch++;
6981 diff_free_filepair(p);
6982 }
6983 }
6984 free(q->queue);
6985 *q = outq;
6986 }
6987
6988 static int diffnamecmp(const void *a_, const void *b_)
6989 {
6990 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6991 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6992 const char *name_a, *name_b;
6993
6994 name_a = a->one ? a->one->path : a->two->path;
6995 name_b = b->one ? b->one->path : b->two->path;
6996 return strcmp(name_a, name_b);
6997 }
6998
6999 void diffcore_fix_diff_index(void)
7000 {
7001 struct diff_queue_struct *q = &diff_queued_diff;
7002 QSORT(q->queue, q->nr, diffnamecmp);
7003 }
7004
7005 void diff_add_if_missing(struct repository *r,
7006 struct oid_array *to_fetch,
7007 const struct diff_filespec *filespec)
7008 {
7009 if (filespec && filespec->oid_valid &&
7010 !S_ISGITLINK(filespec->mode) &&
7011 oid_object_info_extended(r, &filespec->oid, NULL,
7012 OBJECT_INFO_FOR_PREFETCH))
7013 oid_array_append(to_fetch, &filespec->oid);
7014 }
7015
7016 void diff_queued_diff_prefetch(void *repository)
7017 {
7018 struct repository *repo = repository;
7019 int i;
7020 struct diff_queue_struct *q = &diff_queued_diff;
7021 struct oid_array to_fetch = OID_ARRAY_INIT;
7022
7023 for (i = 0; i < q->nr; i++) {
7024 struct diff_filepair *p = q->queue[i];
7025 diff_add_if_missing(repo, &to_fetch, p->one);
7026 diff_add_if_missing(repo, &to_fetch, p->two);
7027 }
7028
7029 /*
7030 * NEEDSWORK: Consider deduplicating the OIDs sent.
7031 */
7032 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
7033
7034 oid_array_clear(&to_fetch);
7035 }
7036
7037 void init_diffstat_widths(struct diff_options *options)
7038 {
7039 options->stat_width = -1; /* use full terminal width */
7040 options->stat_name_width = -1; /* respect diff.statNameWidth config */
7041 options->stat_graph_width = -1; /* respect diff.statGraphWidth config */
7042 }
7043
7044 void diffcore_std(struct diff_options *options)
7045 {
7046 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
7047 DIFF_FORMAT_NUMSTAT |
7048 DIFF_FORMAT_PATCH |
7049 DIFF_FORMAT_SHORTSTAT |
7050 DIFF_FORMAT_DIRSTAT;
7051
7052 /*
7053 * Check if the user requested a blob-data-requiring diff output and/or
7054 * break-rewrite detection (which requires blob data). If yes, prefetch
7055 * the diff pairs.
7056 *
7057 * If no prefetching occurs, diffcore_rename() will prefetch if it
7058 * decides that it needs inexact rename detection.
7059 */
7060 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
7061 (options->output_format & output_formats_to_prefetch ||
7062 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
7063 diff_queued_diff_prefetch(options->repo);
7064
7065 /* NOTE please keep the following in sync with diff_tree_combined() */
7066 if (options->skip_stat_unmatch)
7067 diffcore_skip_stat_unmatch(options);
7068 if (!options->found_follow) {
7069 /* See try_to_follow_renames() in tree-diff.c */
7070 if (options->break_opt != -1)
7071 diffcore_break(options->repo,
7072 options->break_opt);
7073 if (options->detect_rename)
7074 diffcore_rename(options);
7075 if (options->break_opt != -1)
7076 diffcore_merge_broken();
7077 }
7078 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
7079 diffcore_pickaxe(options);
7080 if (options->orderfile)
7081 diffcore_order(options->orderfile);
7082 if (options->rotate_to)
7083 diffcore_rotate(options);
7084 if (!options->found_follow)
7085 /* See try_to_follow_renames() in tree-diff.c */
7086 diff_resolve_rename_copy();
7087 diffcore_apply_filter(options);
7088
7089 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
7090 options->flags.has_changes = 1;
7091 else
7092 options->flags.has_changes = 0;
7093
7094 options->found_follow = 0;
7095 }
7096
7097 int diff_result_code(struct rev_info *revs)
7098 {
7099 struct diff_options *opt = &revs->diffopt;
7100 int result = 0;
7101
7102 if (revs->remerge_diff) {
7103 tmp_objdir_destroy(revs->remerge_objdir);
7104 revs->remerge_objdir = NULL;
7105 }
7106
7107 diff_warn_rename_limit("diff.renameLimit",
7108 opt->needed_rename_limit,
7109 opt->degraded_cc_to_c);
7110
7111 if (opt->flags.exit_with_status &&
7112 opt->flags.has_changes)
7113 result |= 01;
7114 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
7115 opt->flags.check_failed)
7116 result |= 02;
7117 return result;
7118 }
7119
7120 int diff_can_quit_early(struct diff_options *opt)
7121 {
7122 return (opt->flags.quick &&
7123 !opt->filter &&
7124 opt->flags.has_changes);
7125 }
7126
7127 /*
7128 * Shall changes to this submodule be ignored?
7129 *
7130 * Submodule changes can be configured to be ignored separately for each path,
7131 * but that configuration can be overridden from the command line.
7132 */
7133 static int is_submodule_ignored(const char *path, struct diff_options *options)
7134 {
7135 int ignored = 0;
7136 struct diff_flags orig_flags = options->flags;
7137 if (!options->flags.override_submodule_config)
7138 set_diffopt_flags_from_submodule_config(options, path);
7139 if (options->flags.ignore_submodules)
7140 ignored = 1;
7141 options->flags = orig_flags;
7142 return ignored;
7143 }
7144
7145 void compute_diffstat(struct diff_options *options,
7146 struct diffstat_t *diffstat,
7147 struct diff_queue_struct *q)
7148 {
7149 int i;
7150
7151 memset(diffstat, 0, sizeof(struct diffstat_t));
7152 for (i = 0; i < q->nr; i++) {
7153 struct diff_filepair *p = q->queue[i];
7154 if (check_pair_status(p))
7155 diff_flush_stat(p, options, diffstat);
7156 }
7157 options->found_changes = !!diffstat->nr;
7158 }
7159
7160 void diff_addremove(struct diff_options *options,
7161 int addremove, unsigned mode,
7162 const struct object_id *oid,
7163 int oid_valid,
7164 const char *concatpath, unsigned dirty_submodule)
7165 {
7166 struct diff_filespec *one, *two;
7167
7168 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
7169 return;
7170
7171 /* This may look odd, but it is a preparation for
7172 * feeding "there are unchanged files which should
7173 * not produce diffs, but when you are doing copy
7174 * detection you would need them, so here they are"
7175 * entries to the diff-core. They will be prefixed
7176 * with something like '=' or '*' (I haven't decided
7177 * which but should not make any difference).
7178 * Feeding the same new and old to diff_change()
7179 * also has the same effect.
7180 * Before the final output happens, they are pruned after
7181 * merged into rename/copy pairs as appropriate.
7182 */
7183 if (options->flags.reverse_diff)
7184 addremove = (addremove == '+' ? '-' :
7185 addremove == '-' ? '+' : addremove);
7186
7187 if (options->prefix &&
7188 strncmp(concatpath, options->prefix, options->prefix_length))
7189 return;
7190
7191 one = alloc_filespec(concatpath);
7192 two = alloc_filespec(concatpath);
7193
7194 if (addremove != '+')
7195 fill_filespec(one, oid, oid_valid, mode);
7196 if (addremove != '-') {
7197 fill_filespec(two, oid, oid_valid, mode);
7198 two->dirty_submodule = dirty_submodule;
7199 }
7200
7201 diff_queue(&diff_queued_diff, one, two);
7202 if (!options->flags.diff_from_contents)
7203 options->flags.has_changes = 1;
7204 }
7205
7206 void diff_change(struct diff_options *options,
7207 unsigned old_mode, unsigned new_mode,
7208 const struct object_id *old_oid,
7209 const struct object_id *new_oid,
7210 int old_oid_valid, int new_oid_valid,
7211 const char *concatpath,
7212 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7213 {
7214 struct diff_filespec *one, *two;
7215 struct diff_filepair *p;
7216
7217 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7218 is_submodule_ignored(concatpath, options))
7219 return;
7220
7221 if (options->flags.reverse_diff) {
7222 SWAP(old_mode, new_mode);
7223 SWAP(old_oid, new_oid);
7224 SWAP(old_oid_valid, new_oid_valid);
7225 SWAP(old_dirty_submodule, new_dirty_submodule);
7226 }
7227
7228 if (options->prefix &&
7229 strncmp(concatpath, options->prefix, options->prefix_length))
7230 return;
7231
7232 one = alloc_filespec(concatpath);
7233 two = alloc_filespec(concatpath);
7234 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7235 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7236 one->dirty_submodule = old_dirty_submodule;
7237 two->dirty_submodule = new_dirty_submodule;
7238 p = diff_queue(&diff_queued_diff, one, two);
7239
7240 if (options->flags.diff_from_contents)
7241 return;
7242
7243 if (options->flags.quick && options->skip_stat_unmatch &&
7244 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7245 diff_free_filespec_data(p->one);
7246 diff_free_filespec_data(p->two);
7247 return;
7248 }
7249
7250 options->flags.has_changes = 1;
7251 }
7252
7253 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7254 {
7255 struct diff_filepair *pair;
7256 struct diff_filespec *one, *two;
7257
7258 if (options->prefix &&
7259 strncmp(path, options->prefix, options->prefix_length))
7260 return NULL;
7261
7262 one = alloc_filespec(path);
7263 two = alloc_filespec(path);
7264 pair = diff_queue(&diff_queued_diff, one, two);
7265 pair->is_unmerged = 1;
7266 return pair;
7267 }
7268
7269 static char *run_textconv(struct repository *r,
7270 const char *pgm,
7271 struct diff_filespec *spec,
7272 size_t *outsize)
7273 {
7274 struct diff_tempfile *temp;
7275 struct child_process child = CHILD_PROCESS_INIT;
7276 struct strbuf buf = STRBUF_INIT;
7277 int err = 0;
7278
7279 temp = prepare_temp_file(r, spec);
7280 strvec_push(&child.args, pgm);
7281 strvec_push(&child.args, temp->name);
7282
7283 child.use_shell = 1;
7284 child.out = -1;
7285 if (start_command(&child)) {
7286 remove_tempfile();
7287 return NULL;
7288 }
7289
7290 if (strbuf_read(&buf, child.out, 0) < 0)
7291 err = error("error reading from textconv command '%s'", pgm);
7292 close(child.out);
7293
7294 if (finish_command(&child) || err) {
7295 strbuf_release(&buf);
7296 remove_tempfile();
7297 return NULL;
7298 }
7299 remove_tempfile();
7300
7301 return strbuf_detach(&buf, outsize);
7302 }
7303
7304 size_t fill_textconv(struct repository *r,
7305 struct userdiff_driver *driver,
7306 struct diff_filespec *df,
7307 char **outbuf)
7308 {
7309 size_t size;
7310
7311 if (!driver) {
7312 if (!DIFF_FILE_VALID(df)) {
7313 *outbuf = (char *) "";
7314 return 0;
7315 }
7316 if (diff_populate_filespec(r, df, NULL))
7317 die("unable to read files to diff");
7318 *outbuf = df->data;
7319 return df->size;
7320 }
7321
7322 if (!driver->textconv)
7323 BUG("fill_textconv called with non-textconv driver");
7324
7325 if (driver->textconv_cache && df->oid_valid) {
7326 *outbuf = notes_cache_get(driver->textconv_cache,
7327 &df->oid,
7328 &size);
7329 if (*outbuf)
7330 return size;
7331 }
7332
7333 *outbuf = run_textconv(r, driver->textconv, df, &size);
7334 if (!*outbuf)
7335 die("unable to read files to diff");
7336
7337 if (driver->textconv_cache && df->oid_valid) {
7338 /* ignore errors, as we might be in a readonly repository */
7339 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7340 size);
7341 /*
7342 * we could save up changes and flush them all at the end,
7343 * but we would need an extra call after all diffing is done.
7344 * Since generating a cache entry is the slow path anyway,
7345 * this extra overhead probably isn't a big deal.
7346 */
7347 notes_cache_write(driver->textconv_cache);
7348 }
7349
7350 return size;
7351 }
7352
7353 int textconv_object(struct repository *r,
7354 const char *path,
7355 unsigned mode,
7356 const struct object_id *oid,
7357 int oid_valid,
7358 char **buf,
7359 unsigned long *buf_size)
7360 {
7361 struct diff_filespec *df;
7362 struct userdiff_driver *textconv;
7363
7364 df = alloc_filespec(path);
7365 fill_filespec(df, oid, oid_valid, mode);
7366 textconv = get_textconv(r, df);
7367 if (!textconv) {
7368 free_filespec(df);
7369 return 0;
7370 }
7371
7372 *buf_size = fill_textconv(r, textconv, df, buf);
7373 free_filespec(df);
7374 return 1;
7375 }
7376
7377 void setup_diff_pager(struct diff_options *opt)
7378 {
7379 /*
7380 * If the user asked for our exit code, then either they want --quiet
7381 * or --exit-code. We should definitely not bother with a pager in the
7382 * former case, as we will generate no output. Since we still properly
7383 * report our exit code even when a pager is run, we _could_ run a
7384 * pager with --exit-code. But since we have not done so historically,
7385 * and because it is easy to find people oneline advising "git diff
7386 * --exit-code" in hooks and other scripts, we do not do so.
7387 */
7388 if (!opt->flags.exit_with_status &&
7389 check_pager_config(the_repository, "diff") != 0)
7390 setup_pager(the_repository);
7391 }
Alternate source for your git.git needs