?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge tag 'v2.45.1' into debian-sid
[git/debian.git] / diff.c
blob108c1875775df223c4ca4f9dec27194033a7cfaa
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "git-compat-util.h"
5 #include "abspath.h"
6 #include "base85.h"
7 #include "config.h"
8 #include "convert.h"
9 #include "environment.h"
10 #include "gettext.h"
11 #include "tempfile.h"
12 #include "quote.h"
13 #include "diff.h"
14 #include "diffcore.h"
15 #include "delta.h"
16 #include "hex.h"
17 #include "xdiff-interface.h"
18 #include "color.h"
19 #include "run-command.h"
20 #include "utf8.h"
21 #include "object-store-ll.h"
22 #include "userdiff.h"
23 #include "submodule.h"
24 #include "hashmap.h"
25 #include "mem-pool.h"
26 #include "merge-ll.h"
27 #include "string-list.h"
28 #include "strvec.h"
29 #include "graph.h"
30 #include "oid-array.h"
31 #include "packfile.h"
32 #include "pager.h"
33 #include "parse-options.h"
34 #include "help.h"
35 #include "promisor-remote.h"
36 #include "dir.h"
37 #include "object-file.h"
38 #include "object-name.h"
39 #include "read-cache-ll.h"
40 #include "setup.h"
41 #include "strmap.h"
42 #include "ws.h"
43
44 #ifdef NO_FAST_WORKING_DIRECTORY
45 #define FAST_WORKING_DIRECTORY 0
46 #else
47 #define FAST_WORKING_DIRECTORY 1
48 #endif
49
50 static int diff_detect_rename_default;
51 static int diff_indent_heuristic = 1;
52 static int diff_rename_limit_default = 1000;
53 static int diff_suppress_blank_empty;
54 static int diff_use_color_default = -1;
55 static int diff_color_moved_default;
56 static int diff_color_moved_ws_default;
57 static int diff_context_default = 3;
58 static int diff_interhunk_context_default;
59 static const char *diff_word_regex_cfg;
60 static const char *external_diff_cmd_cfg;
61 static const char *diff_order_file_cfg;
62 int diff_auto_refresh_index = 1;
63 static int diff_mnemonic_prefix;
64 static int diff_no_prefix;
65 static const char *diff_src_prefix = "a/";
66 static const char *diff_dst_prefix = "b/";
67 static int diff_relative;
68 static int diff_stat_name_width;
69 static int diff_stat_graph_width;
70 static int diff_dirstat_permille_default = 30;
71 static struct diff_options default_diff_options;
72 static long diff_algorithm;
73 static unsigned ws_error_highlight_default = WSEH_NEW;
74
75 static char diff_colors[][COLOR_MAXLEN] = {
76 GIT_COLOR_RESET,
77 GIT_COLOR_NORMAL, /* CONTEXT */
78 GIT_COLOR_BOLD, /* METAINFO */
79 GIT_COLOR_CYAN, /* FRAGINFO */
80 GIT_COLOR_RED, /* OLD */
81 GIT_COLOR_GREEN, /* NEW */
82 GIT_COLOR_YELLOW, /* COMMIT */
83 GIT_COLOR_BG_RED, /* WHITESPACE */
84 GIT_COLOR_NORMAL, /* FUNCINFO */
85 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
86 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
87 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
88 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
89 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
90 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
91 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
92 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
93 GIT_COLOR_FAINT, /* CONTEXT_DIM */
94 GIT_COLOR_FAINT_RED, /* OLD_DIM */
95 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
96 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
97 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
98 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
99 };
100
101 static const char *color_diff_slots[] = {
102 [DIFF_CONTEXT] = "context",
103 [DIFF_METAINFO] = "meta",
104 [DIFF_FRAGINFO] = "frag",
105 [DIFF_FILE_OLD] = "old",
106 [DIFF_FILE_NEW] = "new",
107 [DIFF_COMMIT] = "commit",
108 [DIFF_WHITESPACE] = "whitespace",
109 [DIFF_FUNCINFO] = "func",
110 [DIFF_FILE_OLD_MOVED] = "oldMoved",
111 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
112 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
113 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
114 [DIFF_FILE_NEW_MOVED] = "newMoved",
115 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
116 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
117 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
118 [DIFF_CONTEXT_DIM] = "contextDimmed",
119 [DIFF_FILE_OLD_DIM] = "oldDimmed",
120 [DIFF_FILE_NEW_DIM] = "newDimmed",
121 [DIFF_CONTEXT_BOLD] = "contextBold",
122 [DIFF_FILE_OLD_BOLD] = "oldBold",
123 [DIFF_FILE_NEW_BOLD] = "newBold",
124 };
125
126 define_list_config_array_extra(color_diff_slots, {"plain"});
127
128 static int parse_diff_color_slot(const char *var)
129 {
130 if (!strcasecmp(var, "plain"))
131 return DIFF_CONTEXT;
132 return LOOKUP_CONFIG(color_diff_slots, var);
133 }
134
135 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
136 struct strbuf *errmsg)
137 {
138 char *params_copy = xstrdup(params_string);
139 struct string_list params = STRING_LIST_INIT_NODUP;
140 int ret = 0;
141 int i;
142
143 if (*params_copy)
144 string_list_split_in_place(&params, params_copy, ",", -1);
145 for (i = 0; i < params.nr; i++) {
146 const char *p = params.items[i].string;
147 if (!strcmp(p, "changes")) {
148 options->flags.dirstat_by_line = 0;
149 options->flags.dirstat_by_file = 0;
150 } else if (!strcmp(p, "lines")) {
151 options->flags.dirstat_by_line = 1;
152 options->flags.dirstat_by_file = 0;
153 } else if (!strcmp(p, "files")) {
154 options->flags.dirstat_by_line = 0;
155 options->flags.dirstat_by_file = 1;
156 } else if (!strcmp(p, "noncumulative")) {
157 options->flags.dirstat_cumulative = 0;
158 } else if (!strcmp(p, "cumulative")) {
159 options->flags.dirstat_cumulative = 1;
160 } else if (isdigit(*p)) {
161 char *end;
162 int permille = strtoul(p, &end, 10) * 10;
163 if (*end == '.' && isdigit(*++end)) {
164 /* only use first digit */
165 permille += *end - '0';
166 /* .. and ignore any further digits */
167 while (isdigit(*++end))
168 ; /* nothing */
169 }
170 if (!*end)
171 options->dirstat_permille = permille;
172 else {
173 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
174 p);
175 ret++;
176 }
177 } else {
178 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
179 ret++;
180 }
181
182 }
183 string_list_clear(&params, 0);
184 free(params_copy);
185 return ret;
186 }
187
188 static int parse_submodule_params(struct diff_options *options, const char *value)
189 {
190 if (!strcmp(value, "log"))
191 options->submodule_format = DIFF_SUBMODULE_LOG;
192 else if (!strcmp(value, "short"))
193 options->submodule_format = DIFF_SUBMODULE_SHORT;
194 else if (!strcmp(value, "diff"))
195 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
196 /*
197 * Please update $__git_diff_submodule_formats in
198 * git-completion.bash when you add new formats.
199 */
200 else
201 return -1;
202 return 0;
203 }
204
205 int git_config_rename(const char *var, const char *value)
206 {
207 if (!value)
208 return DIFF_DETECT_RENAME;
209 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
210 return DIFF_DETECT_COPY;
211 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
212 }
213
214 long parse_algorithm_value(const char *value)
215 {
216 if (!value)
217 return -1;
218 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
219 return 0;
220 else if (!strcasecmp(value, "minimal"))
221 return XDF_NEED_MINIMAL;
222 else if (!strcasecmp(value, "patience"))
223 return XDF_PATIENCE_DIFF;
224 else if (!strcasecmp(value, "histogram"))
225 return XDF_HISTOGRAM_DIFF;
226 /*
227 * Please update $__git_diff_algorithms in git-completion.bash
228 * when you add new algorithms.
229 */
230 return -1;
231 }
232
233 static int parse_one_token(const char **arg, const char *token)
234 {
235 const char *rest;
236 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
237 *arg = rest;
238 return 1;
239 }
240 return 0;
241 }
242
243 static int parse_ws_error_highlight(const char *arg)
244 {
245 const char *orig_arg = arg;
246 unsigned val = 0;
247
248 while (*arg) {
249 if (parse_one_token(&arg, "none"))
250 val = 0;
251 else if (parse_one_token(&arg, "default"))
252 val = WSEH_NEW;
253 else if (parse_one_token(&arg, "all"))
254 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
255 else if (parse_one_token(&arg, "new"))
256 val |= WSEH_NEW;
257 else if (parse_one_token(&arg, "old"))
258 val |= WSEH_OLD;
259 else if (parse_one_token(&arg, "context"))
260 val |= WSEH_CONTEXT;
261 else {
262 return -1 - (int)(arg - orig_arg);
263 }
264 if (*arg)
265 arg++;
266 }
267 return val;
268 }
269
270 /*
271 * These are to give UI layer defaults.
272 * The core-level commands such as git-diff-files should
273 * never be affected by the setting of diff.renames
274 * the user happens to have in the configuration file.
275 */
276 void init_diff_ui_defaults(void)
277 {
278 diff_detect_rename_default = DIFF_DETECT_RENAME;
279 }
280
281 int git_diff_heuristic_config(const char *var, const char *value,
282 void *cb UNUSED)
283 {
284 if (!strcmp(var, "diff.indentheuristic"))
285 diff_indent_heuristic = git_config_bool(var, value);
286 return 0;
287 }
288
289 static int parse_color_moved(const char *arg)
290 {
291 switch (git_parse_maybe_bool(arg)) {
292 case 0:
293 return COLOR_MOVED_NO;
294 case 1:
295 return COLOR_MOVED_DEFAULT;
296 default:
297 break;
298 }
299
300 if (!strcmp(arg, "no"))
301 return COLOR_MOVED_NO;
302 else if (!strcmp(arg, "plain"))
303 return COLOR_MOVED_PLAIN;
304 else if (!strcmp(arg, "blocks"))
305 return COLOR_MOVED_BLOCKS;
306 else if (!strcmp(arg, "zebra"))
307 return COLOR_MOVED_ZEBRA;
308 else if (!strcmp(arg, "default"))
309 return COLOR_MOVED_DEFAULT;
310 else if (!strcmp(arg, "dimmed-zebra"))
311 return COLOR_MOVED_ZEBRA_DIM;
312 else if (!strcmp(arg, "dimmed_zebra"))
313 return COLOR_MOVED_ZEBRA_DIM;
314 else
315 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
316 }
317
318 static unsigned parse_color_moved_ws(const char *arg)
319 {
320 int ret = 0;
321 struct string_list l = STRING_LIST_INIT_DUP;
322 struct string_list_item *i;
323
324 string_list_split(&l, arg, ',', -1);
325
326 for_each_string_list_item(i, &l) {
327 struct strbuf sb = STRBUF_INIT;
328 strbuf_addstr(&sb, i->string);
329 strbuf_trim(&sb);
330
331 if (!strcmp(sb.buf, "no"))
332 ret = 0;
333 else if (!strcmp(sb.buf, "ignore-space-change"))
334 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
335 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
336 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
337 else if (!strcmp(sb.buf, "ignore-all-space"))
338 ret |= XDF_IGNORE_WHITESPACE;
339 else if (!strcmp(sb.buf, "allow-indentation-change"))
340 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
341 else {
342 ret |= COLOR_MOVED_WS_ERROR;
343 error(_("unknown color-moved-ws mode '%s', possible values are 'ignore-space-change', 'ignore-space-at-eol', 'ignore-all-space', 'allow-indentation-change'"), sb.buf);
344 }
345
346 strbuf_release(&sb);
347 }
348
349 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
350 (ret & XDF_WHITESPACE_FLAGS)) {
351 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
352 ret |= COLOR_MOVED_WS_ERROR;
353 }
354
355 string_list_clear(&l, 0);
356
357 return ret;
358 }
359
360 int git_diff_ui_config(const char *var, const char *value,
361 const struct config_context *ctx, void *cb)
362 {
363 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
364 diff_use_color_default = git_config_colorbool(var, value);
365 return 0;
366 }
367 if (!strcmp(var, "diff.colormoved")) {
368 int cm = parse_color_moved(value);
369 if (cm < 0)
370 return -1;
371 diff_color_moved_default = cm;
372 return 0;
373 }
374 if (!strcmp(var, "diff.colormovedws")) {
375 unsigned cm;
376 if (!value)
377 return config_error_nonbool(var);
378 cm = parse_color_moved_ws(value);
379 if (cm & COLOR_MOVED_WS_ERROR)
380 return -1;
381 diff_color_moved_ws_default = cm;
382 return 0;
383 }
384 if (!strcmp(var, "diff.context")) {
385 diff_context_default = git_config_int(var, value, ctx->kvi);
386 if (diff_context_default < 0)
387 return -1;
388 return 0;
389 }
390 if (!strcmp(var, "diff.interhunkcontext")) {
391 diff_interhunk_context_default = git_config_int(var, value,
392 ctx->kvi);
393 if (diff_interhunk_context_default < 0)
394 return -1;
395 return 0;
396 }
397 if (!strcmp(var, "diff.renames")) {
398 diff_detect_rename_default = git_config_rename(var, value);
399 return 0;
400 }
401 if (!strcmp(var, "diff.autorefreshindex")) {
402 diff_auto_refresh_index = git_config_bool(var, value);
403 return 0;
404 }
405 if (!strcmp(var, "diff.mnemonicprefix")) {
406 diff_mnemonic_prefix = git_config_bool(var, value);
407 return 0;
408 }
409 if (!strcmp(var, "diff.noprefix")) {
410 diff_no_prefix = git_config_bool(var, value);
411 return 0;
412 }
413 if (!strcmp(var, "diff.srcprefix")) {
414 return git_config_string(&diff_src_prefix, var, value);
415 }
416 if (!strcmp(var, "diff.dstprefix")) {
417 return git_config_string(&diff_dst_prefix, var, value);
418 }
419 if (!strcmp(var, "diff.relative")) {
420 diff_relative = git_config_bool(var, value);
421 return 0;
422 }
423 if (!strcmp(var, "diff.statnamewidth")) {
424 diff_stat_name_width = git_config_int(var, value, ctx->kvi);
425 return 0;
426 }
427 if (!strcmp(var, "diff.statgraphwidth")) {
428 diff_stat_graph_width = git_config_int(var, value, ctx->kvi);
429 return 0;
430 }
431 if (!strcmp(var, "diff.external"))
432 return git_config_string(&external_diff_cmd_cfg, var, value);
433 if (!strcmp(var, "diff.wordregex"))
434 return git_config_string(&diff_word_regex_cfg, var, value);
435 if (!strcmp(var, "diff.orderfile"))
436 return git_config_pathname(&diff_order_file_cfg, var, value);
437
438 if (!strcmp(var, "diff.ignoresubmodules")) {
439 if (!value)
440 return config_error_nonbool(var);
441 handle_ignore_submodules_arg(&default_diff_options, value);
442 }
443
444 if (!strcmp(var, "diff.submodule")) {
445 if (!value)
446 return config_error_nonbool(var);
447 if (parse_submodule_params(&default_diff_options, value))
448 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
449 value);
450 return 0;
451 }
452
453 if (!strcmp(var, "diff.algorithm")) {
454 if (!value)
455 return config_error_nonbool(var);
456 diff_algorithm = parse_algorithm_value(value);
457 if (diff_algorithm < 0)
458 return error(_("unknown value for config '%s': %s"),
459 var, value);
460 return 0;
461 }
462
463 if (git_color_config(var, value, cb) < 0)
464 return -1;
465
466 return git_diff_basic_config(var, value, ctx, cb);
467 }
468
469 int git_diff_basic_config(const char *var, const char *value,
470 const struct config_context *ctx, void *cb)
471 {
472 const char *name;
473
474 if (!strcmp(var, "diff.renamelimit")) {
475 diff_rename_limit_default = git_config_int(var, value, ctx->kvi);
476 return 0;
477 }
478
479 if (userdiff_config(var, value) < 0)
480 return -1;
481
482 if (skip_prefix(var, "diff.color.", &name) ||
483 skip_prefix(var, "color.diff.", &name)) {
484 int slot = parse_diff_color_slot(name);
485 if (slot < 0)
486 return 0;
487 if (!value)
488 return config_error_nonbool(var);
489 return color_parse(value, diff_colors[slot]);
490 }
491
492 if (!strcmp(var, "diff.wserrorhighlight")) {
493 int val;
494 if (!value)
495 return config_error_nonbool(var);
496 val = parse_ws_error_highlight(value);
497 if (val < 0)
498 return error(_("unknown value for config '%s': %s"),
499 var, value);
500 ws_error_highlight_default = val;
501 return 0;
502 }
503
504 /* like GNU diff's --suppress-blank-empty option */
505 if (!strcmp(var, "diff.suppressblankempty") ||
506 /* for backwards compatibility */
507 !strcmp(var, "diff.suppress-blank-empty")) {
508 diff_suppress_blank_empty = git_config_bool(var, value);
509 return 0;
510 }
511
512 if (!strcmp(var, "diff.dirstat")) {
513 struct strbuf errmsg = STRBUF_INIT;
514 if (!value)
515 return config_error_nonbool(var);
516 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
517 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
518 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
519 errmsg.buf);
520 strbuf_release(&errmsg);
521 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
522 return 0;
523 }
524
525 if (git_diff_heuristic_config(var, value, cb) < 0)
526 return -1;
527
528 return git_default_config(var, value, ctx, cb);
529 }
530
531 static char *quote_two(const char *one, const char *two)
532 {
533 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
534 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
535 struct strbuf res = STRBUF_INIT;
536
537 if (need_one + need_two) {
538 strbuf_addch(&res, '"');
539 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
540 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
541 strbuf_addch(&res, '"');
542 } else {
543 strbuf_addstr(&res, one);
544 strbuf_addstr(&res, two);
545 }
546 return strbuf_detach(&res, NULL);
547 }
548
549 static const char *external_diff(void)
550 {
551 static const char *external_diff_cmd = NULL;
552 static int done_preparing = 0;
553
554 if (done_preparing)
555 return external_diff_cmd;
556 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
557 if (!external_diff_cmd)
558 external_diff_cmd = external_diff_cmd_cfg;
559 done_preparing = 1;
560 return external_diff_cmd;
561 }
562
563 /*
564 * Keep track of files used for diffing. Sometimes such an entry
565 * refers to a temporary file, sometimes to an existing file, and
566 * sometimes to "/dev/null".
567 */
568 static struct diff_tempfile {
569 /*
570 * filename external diff should read from, or NULL if this
571 * entry is currently not in use:
572 */
573 const char *name;
574
575 char hex[GIT_MAX_HEXSZ + 1];
576 char mode[10];
577
578 /*
579 * If this diff_tempfile instance refers to a temporary file,
580 * this tempfile object is used to manage its lifetime.
581 */
582 struct tempfile *tempfile;
583 } diff_temp[2];
584
585 struct emit_callback {
586 int color_diff;
587 unsigned ws_rule;
588 int blank_at_eof_in_preimage;
589 int blank_at_eof_in_postimage;
590 int lno_in_preimage;
591 int lno_in_postimage;
592 const char **label_path;
593 struct diff_words_data *diff_words;
594 struct diff_options *opt;
595 struct strbuf *header;
596 };
597
598 static int count_lines(const char *data, int size)
599 {
600 int count, ch, completely_empty = 1, nl_just_seen = 0;
601 count = 0;
602 while (0 < size--) {
603 ch = *data++;
604 if (ch == '\n') {
605 count++;
606 nl_just_seen = 1;
607 completely_empty = 0;
608 }
609 else {
610 nl_just_seen = 0;
611 completely_empty = 0;
612 }
613 }
614 if (completely_empty)
615 return 0;
616 if (!nl_just_seen)
617 count++; /* no trailing newline */
618 return count;
619 }
620
621 static int fill_mmfile(struct repository *r, mmfile_t *mf,
622 struct diff_filespec *one)
623 {
624 if (!DIFF_FILE_VALID(one)) {
625 mf->ptr = (char *)""; /* does not matter */
626 mf->size = 0;
627 return 0;
628 }
629 else if (diff_populate_filespec(r, one, NULL))
630 return -1;
631
632 mf->ptr = one->data;
633 mf->size = one->size;
634 return 0;
635 }
636
637 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
638 static unsigned long diff_filespec_size(struct repository *r,
639 struct diff_filespec *one)
640 {
641 struct diff_populate_filespec_options dpf_options = {
642 .check_size_only = 1,
643 };
644
645 if (!DIFF_FILE_VALID(one))
646 return 0;
647 diff_populate_filespec(r, one, &dpf_options);
648 return one->size;
649 }
650
651 static int count_trailing_blank(mmfile_t *mf)
652 {
653 char *ptr = mf->ptr;
654 long size = mf->size;
655 int cnt = 0;
656
657 if (!size)
658 return cnt;
659 ptr += size - 1; /* pointing at the very end */
660 if (*ptr != '\n')
661 ; /* incomplete line */
662 else
663 ptr--; /* skip the last LF */
664 while (mf->ptr < ptr) {
665 char *prev_eol;
666 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
667 if (*prev_eol == '\n')
668 break;
669 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol))
670 break;
671 cnt++;
672 ptr = prev_eol - 1;
673 }
674 return cnt;
675 }
676
677 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
678 struct emit_callback *ecbdata)
679 {
680 int l1, l2, at;
681 l1 = count_trailing_blank(mf1);
682 l2 = count_trailing_blank(mf2);
683 if (l2 <= l1) {
684 ecbdata->blank_at_eof_in_preimage = 0;
685 ecbdata->blank_at_eof_in_postimage = 0;
686 return;
687 }
688 at = count_lines(mf1->ptr, mf1->size);
689 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
690
691 at = count_lines(mf2->ptr, mf2->size);
692 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
693 }
694
695 static void emit_line_0(struct diff_options *o,
696 const char *set_sign, const char *set, unsigned reverse, const char *reset,
697 int first, const char *line, int len)
698 {
699 int has_trailing_newline, has_trailing_carriage_return;
700 int needs_reset = 0; /* at the end of the line */
701 FILE *file = o->file;
702
703 fputs(diff_line_prefix(o), file);
704
705 has_trailing_newline = (len > 0 && line[len-1] == '\n');
706 if (has_trailing_newline)
707 len--;
708
709 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
710 if (has_trailing_carriage_return)
711 len--;
712
713 if (!len && !first)
714 goto end_of_line;
715
716 if (reverse && want_color(o->use_color)) {
717 fputs(GIT_COLOR_REVERSE, file);
718 needs_reset = 1;
719 }
720
721 if (set_sign) {
722 fputs(set_sign, file);
723 needs_reset = 1;
724 }
725
726 if (first)
727 fputc(first, file);
728
729 if (!len)
730 goto end_of_line;
731
732 if (set) {
733 if (set_sign && set != set_sign)
734 fputs(reset, file);
735 fputs(set, file);
736 needs_reset = 1;
737 }
738 fwrite(line, len, 1, file);
739 needs_reset = 1; /* 'line' may contain color codes. */
740
741 end_of_line:
742 if (needs_reset)
743 fputs(reset, file);
744 if (has_trailing_carriage_return)
745 fputc('\r', file);
746 if (has_trailing_newline)
747 fputc('\n', file);
748 }
749
750 static void emit_line(struct diff_options *o, const char *set, const char *reset,
751 const char *line, int len)
752 {
753 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
754 }
755
756 enum diff_symbol {
757 DIFF_SYMBOL_BINARY_DIFF_HEADER,
758 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
759 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
760 DIFF_SYMBOL_BINARY_DIFF_BODY,
761 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
762 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
763 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
764 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
765 DIFF_SYMBOL_STATS_LINE,
766 DIFF_SYMBOL_WORD_DIFF,
767 DIFF_SYMBOL_STAT_SEP,
768 DIFF_SYMBOL_SUMMARY,
769 DIFF_SYMBOL_SUBMODULE_ADD,
770 DIFF_SYMBOL_SUBMODULE_DEL,
771 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
772 DIFF_SYMBOL_SUBMODULE_MODIFIED,
773 DIFF_SYMBOL_SUBMODULE_HEADER,
774 DIFF_SYMBOL_SUBMODULE_ERROR,
775 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
776 DIFF_SYMBOL_REWRITE_DIFF,
777 DIFF_SYMBOL_BINARY_FILES,
778 DIFF_SYMBOL_HEADER,
779 DIFF_SYMBOL_FILEPAIR_PLUS,
780 DIFF_SYMBOL_FILEPAIR_MINUS,
781 DIFF_SYMBOL_WORDS_PORCELAIN,
782 DIFF_SYMBOL_WORDS,
783 DIFF_SYMBOL_CONTEXT,
784 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
785 DIFF_SYMBOL_PLUS,
786 DIFF_SYMBOL_MINUS,
787 DIFF_SYMBOL_NO_LF_EOF,
788 DIFF_SYMBOL_CONTEXT_FRAGINFO,
789 DIFF_SYMBOL_CONTEXT_MARKER,
790 DIFF_SYMBOL_SEPARATOR
791 };
792 /*
793 * Flags for content lines:
794 * 0..12 are whitespace rules
795 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
796 * 16 is marking if the line is blank at EOF
797 */
798 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
799 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
800 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
801 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
802 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
803
804 /*
805 * This struct is used when we need to buffer the output of the diff output.
806 *
807 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
808 * into the pre/post image file. This pointer could be a union with the
809 * line pointer. By storing an offset into the file instead of the literal line,
810 * we can decrease the memory footprint for the buffered output. At first we
811 * may want to only have indirection for the content lines, but we could also
812 * enhance the state for emitting prefabricated lines, e.g. the similarity
813 * score line or hunk/file headers would only need to store a number or path
814 * and then the output can be constructed later on depending on state.
815 */
816 struct emitted_diff_symbol {
817 const char *line;
818 int len;
819 int flags;
820 int indent_off; /* Offset to first non-whitespace character */
821 int indent_width; /* The visual width of the indentation */
822 unsigned id;
823 enum diff_symbol s;
824 };
825 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
826
827 struct emitted_diff_symbols {
828 struct emitted_diff_symbol *buf;
829 int nr, alloc;
830 };
831 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
832
833 static void append_emitted_diff_symbol(struct diff_options *o,
834 struct emitted_diff_symbol *e)
835 {
836 struct emitted_diff_symbol *f;
837
838 ALLOC_GROW(o->emitted_symbols->buf,
839 o->emitted_symbols->nr + 1,
840 o->emitted_symbols->alloc);
841 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
842
843 memcpy(f, e, sizeof(struct emitted_diff_symbol));
844 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
845 }
846
847 static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)
848 {
849 if (!e)
850 return;
851 free(e->buf);
852 free(e);
853 }
854
855 struct moved_entry {
856 const struct emitted_diff_symbol *es;
857 struct moved_entry *next_line;
858 struct moved_entry *next_match;
859 };
860
861 struct moved_block {
862 struct moved_entry *match;
863 int wsd; /* The whitespace delta of this block */
864 };
865
866 #define INDENT_BLANKLINE INT_MIN
867
868 static void fill_es_indent_data(struct emitted_diff_symbol *es)
869 {
870 unsigned int off = 0, i;
871 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
872 const char *s = es->line;
873 const int len = es->len;
874
875 /* skip any \v \f \r at start of indentation */
876 while (s[off] == '\f' || s[off] == '\v' ||
877 (s[off] == '\r' && off < len - 1))
878 off++;
879
880 /* calculate the visual width of indentation */
881 while(1) {
882 if (s[off] == ' ') {
883 width++;
884 off++;
885 } else if (s[off] == '\t') {
886 width += tab_width - (width % tab_width);
887 while (s[++off] == '\t')
888 width += tab_width;
889 } else {
890 break;
891 }
892 }
893
894 /* check if this line is blank */
895 for (i = off; i < len; i++)
896 if (!isspace(s[i]))
897 break;
898
899 if (i == len) {
900 es->indent_width = INDENT_BLANKLINE;
901 es->indent_off = len;
902 } else {
903 es->indent_off = off;
904 es->indent_width = width;
905 }
906 }
907
908 static int compute_ws_delta(const struct emitted_diff_symbol *a,
909 const struct emitted_diff_symbol *b)
910 {
911 int a_width = a->indent_width,
912 b_width = b->indent_width;
913
914 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
915 return INDENT_BLANKLINE;
916
917 return a_width - b_width;
918 }
919
920 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
921 const struct emitted_diff_symbol *l,
922 struct moved_block *pmb)
923 {
924 int a_width = cur->es->indent_width, b_width = l->indent_width;
925 int delta;
926
927 /* The text of each line must match */
928 if (cur->es->id != l->id)
929 return 1;
930
931 /*
932 * If 'l' and 'cur' are both blank then we don't need to check the
933 * indent. We only need to check cur as we know the strings match.
934 * */
935 if (a_width == INDENT_BLANKLINE)
936 return 0;
937
938 /*
939 * The indent changes of the block are known and stored in pmb->wsd;
940 * however we need to check if the indent changes of the current line
941 * match those of the current block.
942 */
943 delta = b_width - a_width;
944
945 /*
946 * If the previous lines of this block were all blank then set its
947 * whitespace delta.
948 */
949 if (pmb->wsd == INDENT_BLANKLINE)
950 pmb->wsd = delta;
951
952 return delta != pmb->wsd;
953 }
954
955 struct interned_diff_symbol {
956 struct hashmap_entry ent;
957 struct emitted_diff_symbol *es;
958 };
959
960 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
961 const struct hashmap_entry *eptr,
962 const struct hashmap_entry *entry_or_key,
963 const void *keydata UNUSED)
964 {
965 const struct diff_options *diffopt = hashmap_cmp_fn_data;
966 const struct emitted_diff_symbol *a, *b;
967 unsigned flags = diffopt->color_moved_ws_handling
968 & XDF_WHITESPACE_FLAGS;
969
970 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
971 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
972
973 return !xdiff_compare_lines(a->line + a->indent_off,
974 a->len - a->indent_off,
975 b->line + b->indent_off,
976 b->len - b->indent_off, flags);
977 }
978
979 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
980 struct interned_diff_symbol *s)
981 {
982 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
983 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
984 l->len - l->indent_off, flags);
985
986 hashmap_entry_init(&s->ent, hash);
987 s->es = l;
988 }
989
990 struct moved_entry_list {
991 struct moved_entry *add, *del;
992 };
993
994 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
995 struct mem_pool *entry_mem_pool)
996 {
997 struct moved_entry *prev_line = NULL;
998 struct mem_pool interned_pool;
999 struct hashmap interned_map;
1000 struct moved_entry_list *entry_list = NULL;
1001 size_t entry_list_alloc = 0;
1002 unsigned id = 0;
1003 int n;
1004
1005 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
1006 mem_pool_init(&interned_pool, 1024 * 1024);
1007
1008 for (n = 0; n < o->emitted_symbols->nr; n++) {
1009 struct interned_diff_symbol key;
1010 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1011 struct interned_diff_symbol *s;
1012 struct moved_entry *entry;
1013
1014 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
1015 prev_line = NULL;
1016 continue;
1017 }
1018
1019 if (o->color_moved_ws_handling &
1020 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1021 fill_es_indent_data(l);
1022
1023 prepare_entry(o, l, &key);
1024 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
1025 if (s) {
1026 l->id = s->es->id;
1027 } else {
1028 l->id = id;
1029 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
1030 hashmap_add(&interned_map,
1031 memcpy(mem_pool_alloc(&interned_pool,
1032 sizeof(key)),
1033 &key, sizeof(key)));
1034 }
1035 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
1036 entry->es = l;
1037 entry->next_line = NULL;
1038 if (prev_line && prev_line->es->s == l->s)
1039 prev_line->next_line = entry;
1040 prev_line = entry;
1041 if (l->s == DIFF_SYMBOL_PLUS) {
1042 entry->next_match = entry_list[l->id].add;
1043 entry_list[l->id].add = entry;
1044 } else {
1045 entry->next_match = entry_list[l->id].del;
1046 entry_list[l->id].del = entry;
1047 }
1048 }
1049
1050 hashmap_clear(&interned_map);
1051 mem_pool_discard(&interned_pool, 0);
1052
1053 return entry_list;
1054 }
1055
1056 static void pmb_advance_or_null(struct diff_options *o,
1057 struct emitted_diff_symbol *l,
1058 struct moved_block *pmb,
1059 int *pmb_nr)
1060 {
1061 int i, j;
1062
1063 for (i = 0, j = 0; i < *pmb_nr; i++) {
1064 int match;
1065 struct moved_entry *prev = pmb[i].match;
1066 struct moved_entry *cur = (prev && prev->next_line) ?
1067 prev->next_line : NULL;
1068
1069 if (o->color_moved_ws_handling &
1070 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1071 match = cur &&
1072 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1073 else
1074 match = cur && cur->es->id == l->id;
1075
1076 if (match) {
1077 pmb[j] = pmb[i];
1078 pmb[j++].match = cur;
1079 }
1080 }
1081 *pmb_nr = j;
1082 }
1083
1084 static void fill_potential_moved_blocks(struct diff_options *o,
1085 struct moved_entry *match,
1086 struct emitted_diff_symbol *l,
1087 struct moved_block **pmb_p,
1088 int *pmb_alloc_p, int *pmb_nr_p)
1089
1090 {
1091 struct moved_block *pmb = *pmb_p;
1092 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1093
1094 /*
1095 * The current line is the start of a new block.
1096 * Setup the set of potential blocks.
1097 */
1098 for (; match; match = match->next_match) {
1099 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1100 if (o->color_moved_ws_handling &
1101 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1102 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1103 else
1104 pmb[pmb_nr].wsd = 0;
1105 pmb[pmb_nr++].match = match;
1106 }
1107
1108 *pmb_p = pmb;
1109 *pmb_alloc_p = pmb_alloc;
1110 *pmb_nr_p = pmb_nr;
1111 }
1112
1113 /*
1114 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1115 *
1116 * Otherwise, if the last block has fewer alphanumeric characters than
1117 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1118 * that block.
1119 *
1120 * The last block consists of the (n - block_length)'th line up to but not
1121 * including the nth line.
1122 *
1123 * Returns 0 if the last block is empty or is unset by this function, non zero
1124 * otherwise.
1125 *
1126 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1127 * Think of a way to unify them.
1128 */
1129 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1130 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1131 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1132 {
1133 int i, alnum_count = 0;
1134 if (o->color_moved == COLOR_MOVED_PLAIN)
1135 return block_length;
1136 for (i = 1; i < block_length + 1; i++) {
1137 const char *c = o->emitted_symbols->buf[n - i].line;
1138 for (; *c; c++) {
1139 if (!isalnum(*c))
1140 continue;
1141 alnum_count++;
1142 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1143 return 1;
1144 }
1145 }
1146 for (i = 1; i < block_length + 1; i++)
1147 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1148 return 0;
1149 }
1150
1151 /* Find blocks of moved code, delegate actual coloring decision to helper */
1152 static void mark_color_as_moved(struct diff_options *o,
1153 struct moved_entry_list *entry_list)
1154 {
1155 struct moved_block *pmb = NULL; /* potentially moved blocks */
1156 int pmb_nr = 0, pmb_alloc = 0;
1157 int n, flipped_block = 0, block_length = 0;
1158 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1159
1160
1161 for (n = 0; n < o->emitted_symbols->nr; n++) {
1162 struct moved_entry *match = NULL;
1163 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1164
1165 switch (l->s) {
1166 case DIFF_SYMBOL_PLUS:
1167 match = entry_list[l->id].del;
1168 break;
1169 case DIFF_SYMBOL_MINUS:
1170 match = entry_list[l->id].add;
1171 break;
1172 default:
1173 flipped_block = 0;
1174 }
1175
1176 if (pmb_nr && (!match || l->s != moved_symbol)) {
1177 if (!adjust_last_block(o, n, block_length) &&
1178 block_length > 1) {
1179 /*
1180 * Rewind in case there is another match
1181 * starting at the second line of the block
1182 */
1183 match = NULL;
1184 n -= block_length;
1185 }
1186 pmb_nr = 0;
1187 block_length = 0;
1188 flipped_block = 0;
1189 }
1190 if (!match) {
1191 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1192 continue;
1193 }
1194
1195 if (o->color_moved == COLOR_MOVED_PLAIN) {
1196 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1197 continue;
1198 }
1199
1200 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1201
1202 if (pmb_nr == 0) {
1203 int contiguous = adjust_last_block(o, n, block_length);
1204
1205 if (!contiguous && block_length > 1)
1206 /*
1207 * Rewind in case there is another match
1208 * starting at the second line of the block
1209 */
1210 n -= block_length;
1211 else
1212 fill_potential_moved_blocks(o, match, l,
1213 &pmb, &pmb_alloc,
1214 &pmb_nr);
1215
1216 if (contiguous && pmb_nr && moved_symbol == l->s)
1217 flipped_block = (flipped_block + 1) % 2;
1218 else
1219 flipped_block = 0;
1220
1221 if (pmb_nr)
1222 moved_symbol = l->s;
1223 else
1224 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1225
1226 block_length = 0;
1227 }
1228
1229 if (pmb_nr) {
1230 block_length++;
1231 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1232 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1233 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1234 }
1235 }
1236 adjust_last_block(o, n, block_length);
1237
1238 free(pmb);
1239 }
1240
1241 static void dim_moved_lines(struct diff_options *o)
1242 {
1243 int n;
1244 for (n = 0; n < o->emitted_symbols->nr; n++) {
1245 struct emitted_diff_symbol *prev = (n != 0) ?
1246 &o->emitted_symbols->buf[n - 1] : NULL;
1247 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1248 struct emitted_diff_symbol *next =
1249 (n < o->emitted_symbols->nr - 1) ?
1250 &o->emitted_symbols->buf[n + 1] : NULL;
1251
1252 /* Not a plus or minus line? */
1253 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1254 continue;
1255
1256 /* Not a moved line? */
1257 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1258 continue;
1259
1260 /*
1261 * If prev or next are not a plus or minus line,
1262 * pretend they don't exist
1263 */
1264 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1265 prev->s != DIFF_SYMBOL_MINUS)
1266 prev = NULL;
1267 if (next && next->s != DIFF_SYMBOL_PLUS &&
1268 next->s != DIFF_SYMBOL_MINUS)
1269 next = NULL;
1270
1271 /* Inside a block? */
1272 if ((prev &&
1273 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1274 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1275 (next &&
1276 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1277 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1278 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1279 continue;
1280 }
1281
1282 /* Check if we are at an interesting bound: */
1283 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1284 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1285 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1286 continue;
1287 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1288 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1289 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1290 continue;
1291
1292 /*
1293 * The boundary to prev and next are not interesting,
1294 * so this line is not interesting as a whole
1295 */
1296 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1297 }
1298 }
1299
1300 static void emit_line_ws_markup(struct diff_options *o,
1301 const char *set_sign, const char *set,
1302 const char *reset,
1303 int sign_index, const char *line, int len,
1304 unsigned ws_rule, int blank_at_eof)
1305 {
1306 const char *ws = NULL;
1307 int sign = o->output_indicators[sign_index];
1308
1309 if (o->ws_error_highlight & ws_rule) {
1310 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1311 if (!*ws)
1312 ws = NULL;
1313 }
1314
1315 if (!ws && !set_sign)
1316 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1317 else if (!ws) {
1318 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1319 } else if (blank_at_eof)
1320 /* Blank line at EOF - paint '+' as well */
1321 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1322 else {
1323 /* Emit just the prefix, then the rest. */
1324 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1325 sign, "", 0);
1326 ws_check_emit(line, len, ws_rule,
1327 o->file, set, reset, ws);
1328 }
1329 }
1330
1331 static void emit_diff_symbol_from_struct(struct diff_options *o,
1332 struct emitted_diff_symbol *eds)
1333 {
1334 static const char *nneof = " No newline at end of file\n";
1335 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1336
1337 enum diff_symbol s = eds->s;
1338 const char *line = eds->line;
1339 int len = eds->len;
1340 unsigned flags = eds->flags;
1341
1342 switch (s) {
1343 case DIFF_SYMBOL_NO_LF_EOF:
1344 context = diff_get_color_opt(o, DIFF_CONTEXT);
1345 reset = diff_get_color_opt(o, DIFF_RESET);
1346 putc('\n', o->file);
1347 emit_line_0(o, context, NULL, 0, reset, '\\',
1348 nneof, strlen(nneof));
1349 break;
1350 case DIFF_SYMBOL_SUBMODULE_HEADER:
1351 case DIFF_SYMBOL_SUBMODULE_ERROR:
1352 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1353 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1354 case DIFF_SYMBOL_SUMMARY:
1355 case DIFF_SYMBOL_STATS_LINE:
1356 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1357 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1358 emit_line(o, "", "", line, len);
1359 break;
1360 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1361 case DIFF_SYMBOL_CONTEXT_MARKER:
1362 context = diff_get_color_opt(o, DIFF_CONTEXT);
1363 reset = diff_get_color_opt(o, DIFF_RESET);
1364 emit_line(o, context, reset, line, len);
1365 break;
1366 case DIFF_SYMBOL_SEPARATOR:
1367 fprintf(o->file, "%s%c",
1368 diff_line_prefix(o),
1369 o->line_termination);
1370 break;
1371 case DIFF_SYMBOL_CONTEXT:
1372 set = diff_get_color_opt(o, DIFF_CONTEXT);
1373 reset = diff_get_color_opt(o, DIFF_RESET);
1374 set_sign = NULL;
1375 if (o->flags.dual_color_diffed_diffs) {
1376 char c = !len ? 0 : line[0];
1377
1378 if (c == '+')
1379 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1380 else if (c == '@')
1381 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1382 else if (c == '-')
1383 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1384 }
1385 emit_line_ws_markup(o, set_sign, set, reset,
1386 OUTPUT_INDICATOR_CONTEXT, line, len,
1387 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1388 break;
1389 case DIFF_SYMBOL_PLUS:
1390 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1391 DIFF_SYMBOL_MOVED_LINE_ALT |
1392 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1393 case DIFF_SYMBOL_MOVED_LINE |
1394 DIFF_SYMBOL_MOVED_LINE_ALT |
1395 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1396 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1397 break;
1398 case DIFF_SYMBOL_MOVED_LINE |
1399 DIFF_SYMBOL_MOVED_LINE_ALT:
1400 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1401 break;
1402 case DIFF_SYMBOL_MOVED_LINE |
1403 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1404 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1405 break;
1406 case DIFF_SYMBOL_MOVED_LINE:
1407 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1408 break;
1409 default:
1410 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1411 }
1412 reset = diff_get_color_opt(o, DIFF_RESET);
1413 if (!o->flags.dual_color_diffed_diffs)
1414 set_sign = NULL;
1415 else {
1416 char c = !len ? 0 : line[0];
1417
1418 set_sign = set;
1419 if (c == '-')
1420 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1421 else if (c == '@')
1422 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1423 else if (c == '+')
1424 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1425 else
1426 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1427 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1428 }
1429 emit_line_ws_markup(o, set_sign, set, reset,
1430 OUTPUT_INDICATOR_NEW, line, len,
1431 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1432 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1433 break;
1434 case DIFF_SYMBOL_MINUS:
1435 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1436 DIFF_SYMBOL_MOVED_LINE_ALT |
1437 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1438 case DIFF_SYMBOL_MOVED_LINE |
1439 DIFF_SYMBOL_MOVED_LINE_ALT |
1440 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1441 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1442 break;
1443 case DIFF_SYMBOL_MOVED_LINE |
1444 DIFF_SYMBOL_MOVED_LINE_ALT:
1445 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1446 break;
1447 case DIFF_SYMBOL_MOVED_LINE |
1448 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1449 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1450 break;
1451 case DIFF_SYMBOL_MOVED_LINE:
1452 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1453 break;
1454 default:
1455 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1456 }
1457 reset = diff_get_color_opt(o, DIFF_RESET);
1458 if (!o->flags.dual_color_diffed_diffs)
1459 set_sign = NULL;
1460 else {
1461 char c = !len ? 0 : line[0];
1462
1463 set_sign = set;
1464 if (c == '+')
1465 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1466 else if (c == '@')
1467 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1468 else if (c == '-')
1469 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1470 else
1471 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1472 }
1473 emit_line_ws_markup(o, set_sign, set, reset,
1474 OUTPUT_INDICATOR_OLD, line, len,
1475 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1476 break;
1477 case DIFF_SYMBOL_WORDS_PORCELAIN:
1478 context = diff_get_color_opt(o, DIFF_CONTEXT);
1479 reset = diff_get_color_opt(o, DIFF_RESET);
1480 emit_line(o, context, reset, line, len);
1481 fputs("~\n", o->file);
1482 break;
1483 case DIFF_SYMBOL_WORDS:
1484 context = diff_get_color_opt(o, DIFF_CONTEXT);
1485 reset = diff_get_color_opt(o, DIFF_RESET);
1486 /*
1487 * Skip the prefix character, if any. With
1488 * diff_suppress_blank_empty, there may be
1489 * none.
1490 */
1491 if (line[0] != '\n') {
1492 line++;
1493 len--;
1494 }
1495 emit_line(o, context, reset, line, len);
1496 break;
1497 case DIFF_SYMBOL_FILEPAIR_PLUS:
1498 meta = diff_get_color_opt(o, DIFF_METAINFO);
1499 reset = diff_get_color_opt(o, DIFF_RESET);
1500 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1501 line, reset,
1502 strchr(line, ' ') ? "\t" : "");
1503 break;
1504 case DIFF_SYMBOL_FILEPAIR_MINUS:
1505 meta = diff_get_color_opt(o, DIFF_METAINFO);
1506 reset = diff_get_color_opt(o, DIFF_RESET);
1507 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1508 line, reset,
1509 strchr(line, ' ') ? "\t" : "");
1510 break;
1511 case DIFF_SYMBOL_BINARY_FILES:
1512 case DIFF_SYMBOL_HEADER:
1513 fprintf(o->file, "%s", line);
1514 break;
1515 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1516 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1517 break;
1518 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1519 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1520 break;
1521 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1522 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1523 break;
1524 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1525 fputs(diff_line_prefix(o), o->file);
1526 fputc('\n', o->file);
1527 break;
1528 case DIFF_SYMBOL_REWRITE_DIFF:
1529 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1530 reset = diff_get_color_opt(o, DIFF_RESET);
1531 emit_line(o, fraginfo, reset, line, len);
1532 break;
1533 case DIFF_SYMBOL_SUBMODULE_ADD:
1534 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1535 reset = diff_get_color_opt(o, DIFF_RESET);
1536 emit_line(o, set, reset, line, len);
1537 break;
1538 case DIFF_SYMBOL_SUBMODULE_DEL:
1539 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1540 reset = diff_get_color_opt(o, DIFF_RESET);
1541 emit_line(o, set, reset, line, len);
1542 break;
1543 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1544 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1545 diff_line_prefix(o), line);
1546 break;
1547 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1548 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1549 diff_line_prefix(o), line);
1550 break;
1551 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1552 emit_line(o, "", "", " 0 files changed\n",
1553 strlen(" 0 files changed\n"));
1554 break;
1555 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1556 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1557 break;
1558 case DIFF_SYMBOL_WORD_DIFF:
1559 fprintf(o->file, "%.*s", len, line);
1560 break;
1561 case DIFF_SYMBOL_STAT_SEP:
1562 fputs(o->stat_sep, o->file);
1563 break;
1564 default:
1565 BUG("unknown diff symbol");
1566 }
1567 }
1568
1569 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1570 const char *line, int len, unsigned flags)
1571 {
1572 struct emitted_diff_symbol e = {
1573 .line = line, .len = len, .flags = flags, .s = s
1574 };
1575
1576 if (o->emitted_symbols)
1577 append_emitted_diff_symbol(o, &e);
1578 else
1579 emit_diff_symbol_from_struct(o, &e);
1580 }
1581
1582 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1583 {
1584 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1585 }
1586
1587 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1588 {
1589 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1590 }
1591
1592 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1593 {
1594 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1595 path, strlen(path), 0);
1596 }
1597
1598 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1599 {
1600 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1601 path, strlen(path), 0);
1602 }
1603
1604 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1605 {
1606 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1607 header, strlen(header), 0);
1608 }
1609
1610 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1611 {
1612 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1613 }
1614
1615 void diff_emit_submodule_pipethrough(struct diff_options *o,
1616 const char *line, int len)
1617 {
1618 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1619 }
1620
1621 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1622 {
1623 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1624 ecbdata->blank_at_eof_in_preimage &&
1625 ecbdata->blank_at_eof_in_postimage &&
1626 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1627 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1628 return 0;
1629 return ws_blank_line(line, len);
1630 }
1631
1632 static void emit_add_line(struct emit_callback *ecbdata,
1633 const char *line, int len)
1634 {
1635 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1636 if (new_blank_line_at_eof(ecbdata, line, len))
1637 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1638
1639 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1640 }
1641
1642 static void emit_del_line(struct emit_callback *ecbdata,
1643 const char *line, int len)
1644 {
1645 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1646 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1647 }
1648
1649 static void emit_context_line(struct emit_callback *ecbdata,
1650 const char *line, int len)
1651 {
1652 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1653 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1654 }
1655
1656 static void emit_hunk_header(struct emit_callback *ecbdata,
1657 const char *line, int len)
1658 {
1659 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1660 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1661 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1662 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1663 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1664 static const char atat[2] = { '@', '@' };
1665 const char *cp, *ep;
1666 struct strbuf msgbuf = STRBUF_INIT;
1667 int org_len = len;
1668 int i = 1;
1669
1670 /*
1671 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1672 * it always is at least 10 bytes long.
1673 */
1674 if (len < 10 ||
1675 memcmp(line, atat, 2) ||
1676 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1677 emit_diff_symbol(ecbdata->opt,
1678 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1679 return;
1680 }
1681 ep += 2; /* skip over @@ */
1682
1683 /* The hunk header in fraginfo color */
1684 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1685 strbuf_addstr(&msgbuf, reverse);
1686 strbuf_addstr(&msgbuf, frag);
1687 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1688 strbuf_add(&msgbuf, atat, sizeof(atat));
1689 else
1690 strbuf_add(&msgbuf, line, ep - line);
1691 strbuf_addstr(&msgbuf, reset);
1692
1693 /*
1694 * trailing "\r\n"
1695 */
1696 for ( ; i < 3; i++)
1697 if (line[len - i] == '\r' || line[len - i] == '\n')
1698 len--;
1699
1700 /* blank before the func header */
1701 for (cp = ep; ep - line < len; ep++)
1702 if (*ep != ' ' && *ep != '\t')
1703 break;
1704 if (ep != cp) {
1705 strbuf_addstr(&msgbuf, context);
1706 strbuf_add(&msgbuf, cp, ep - cp);
1707 strbuf_addstr(&msgbuf, reset);
1708 }
1709
1710 if (ep < line + len) {
1711 strbuf_addstr(&msgbuf, func);
1712 strbuf_add(&msgbuf, ep, line + len - ep);
1713 strbuf_addstr(&msgbuf, reset);
1714 }
1715
1716 strbuf_add(&msgbuf, line + len, org_len - len);
1717 strbuf_complete_line(&msgbuf);
1718 emit_diff_symbol(ecbdata->opt,
1719 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1720 strbuf_release(&msgbuf);
1721 }
1722
1723 static struct diff_tempfile *claim_diff_tempfile(void)
1724 {
1725 int i;
1726 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1727 if (!diff_temp[i].name)
1728 return diff_temp + i;
1729 BUG("diff is failing to clean up its tempfiles");
1730 }
1731
1732 static void remove_tempfile(void)
1733 {
1734 int i;
1735 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1736 if (is_tempfile_active(diff_temp[i].tempfile))
1737 delete_tempfile(&diff_temp[i].tempfile);
1738 diff_temp[i].name = NULL;
1739 }
1740 }
1741
1742 static void add_line_count(struct strbuf *out, int count)
1743 {
1744 switch (count) {
1745 case 0:
1746 strbuf_addstr(out, "0,0");
1747 break;
1748 case 1:
1749 strbuf_addstr(out, "1");
1750 break;
1751 default:
1752 strbuf_addf(out, "1,%d", count);
1753 break;
1754 }
1755 }
1756
1757 static void emit_rewrite_lines(struct emit_callback *ecb,
1758 int prefix, const char *data, int size)
1759 {
1760 const char *endp = NULL;
1761
1762 while (0 < size) {
1763 int len;
1764
1765 endp = memchr(data, '\n', size);
1766 len = endp ? (endp - data + 1) : size;
1767 if (prefix != '+') {
1768 ecb->lno_in_preimage++;
1769 emit_del_line(ecb, data, len);
1770 } else {
1771 ecb->lno_in_postimage++;
1772 emit_add_line(ecb, data, len);
1773 }
1774 size -= len;
1775 data += len;
1776 }
1777 if (!endp)
1778 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1779 }
1780
1781 static void emit_rewrite_diff(const char *name_a,
1782 const char *name_b,
1783 struct diff_filespec *one,
1784 struct diff_filespec *two,
1785 struct userdiff_driver *textconv_one,
1786 struct userdiff_driver *textconv_two,
1787 struct diff_options *o)
1788 {
1789 int lc_a, lc_b;
1790 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1791 const char *a_prefix, *b_prefix;
1792 char *data_one, *data_two;
1793 size_t size_one, size_two;
1794 struct emit_callback ecbdata;
1795 struct strbuf out = STRBUF_INIT;
1796
1797 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1798 a_prefix = o->b_prefix;
1799 b_prefix = o->a_prefix;
1800 } else {
1801 a_prefix = o->a_prefix;
1802 b_prefix = o->b_prefix;
1803 }
1804
1805 name_a += (*name_a == '/');
1806 name_b += (*name_b == '/');
1807
1808 strbuf_reset(&a_name);
1809 strbuf_reset(&b_name);
1810 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1811 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1812
1813 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1814 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1815
1816 memset(&ecbdata, 0, sizeof(ecbdata));
1817 ecbdata.color_diff = want_color(o->use_color);
1818 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1819 ecbdata.opt = o;
1820 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1821 mmfile_t mf1, mf2;
1822 mf1.ptr = (char *)data_one;
1823 mf2.ptr = (char *)data_two;
1824 mf1.size = size_one;
1825 mf2.size = size_two;
1826 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1827 }
1828 ecbdata.lno_in_preimage = 1;
1829 ecbdata.lno_in_postimage = 1;
1830
1831 lc_a = count_lines(data_one, size_one);
1832 lc_b = count_lines(data_two, size_two);
1833
1834 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1835 a_name.buf, a_name.len, 0);
1836 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1837 b_name.buf, b_name.len, 0);
1838
1839 strbuf_addstr(&out, "@@ -");
1840 if (!o->irreversible_delete)
1841 add_line_count(&out, lc_a);
1842 else
1843 strbuf_addstr(&out, "?,?");
1844 strbuf_addstr(&out, " +");
1845 add_line_count(&out, lc_b);
1846 strbuf_addstr(&out, " @@\n");
1847 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1848 strbuf_release(&out);
1849
1850 if (lc_a && !o->irreversible_delete)
1851 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1852 if (lc_b)
1853 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1854 if (textconv_one)
1855 free((char *)data_one);
1856 if (textconv_two)
1857 free((char *)data_two);
1858 }
1859
1860 struct diff_words_buffer {
1861 mmfile_t text;
1862 unsigned long alloc;
1863 struct diff_words_orig {
1864 const char *begin, *end;
1865 } *orig;
1866 int orig_nr, orig_alloc;
1867 };
1868
1869 static void diff_words_append(char *line, unsigned long len,
1870 struct diff_words_buffer *buffer)
1871 {
1872 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1873 line++;
1874 len--;
1875 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1876 buffer->text.size += len;
1877 buffer->text.ptr[buffer->text.size] = '\0';
1878 }
1879
1880 struct diff_words_style_elem {
1881 const char *prefix;
1882 const char *suffix;
1883 const char *color; /* NULL; filled in by the setup code if
1884 * color is enabled */
1885 };
1886
1887 struct diff_words_style {
1888 enum diff_words_type type;
1889 struct diff_words_style_elem new_word, old_word, ctx;
1890 const char *newline;
1891 };
1892
1893 static struct diff_words_style diff_words_styles[] = {
1894 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1895 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1896 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1897 };
1898
1899 struct diff_words_data {
1900 struct diff_words_buffer minus, plus;
1901 const char *current_plus;
1902 int last_minus;
1903 struct diff_options *opt;
1904 regex_t *word_regex;
1905 enum diff_words_type type;
1906 struct diff_words_style *style;
1907 };
1908
1909 static int fn_out_diff_words_write_helper(struct diff_options *o,
1910 struct diff_words_style_elem *st_el,
1911 const char *newline,
1912 size_t count, const char *buf)
1913 {
1914 int print = 0;
1915 struct strbuf sb = STRBUF_INIT;
1916
1917 while (count) {
1918 char *p = memchr(buf, '\n', count);
1919 if (print)
1920 strbuf_addstr(&sb, diff_line_prefix(o));
1921
1922 if (p != buf) {
1923 const char *reset = st_el->color && *st_el->color ?
1924 GIT_COLOR_RESET : NULL;
1925 if (st_el->color && *st_el->color)
1926 strbuf_addstr(&sb, st_el->color);
1927 strbuf_addstr(&sb, st_el->prefix);
1928 strbuf_add(&sb, buf, p ? p - buf : count);
1929 strbuf_addstr(&sb, st_el->suffix);
1930 if (reset)
1931 strbuf_addstr(&sb, reset);
1932 }
1933 if (!p)
1934 goto out;
1935
1936 strbuf_addstr(&sb, newline);
1937 count -= p + 1 - buf;
1938 buf = p + 1;
1939 print = 1;
1940 if (count) {
1941 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1942 sb.buf, sb.len, 0);
1943 strbuf_reset(&sb);
1944 }
1945 }
1946
1947 out:
1948 if (sb.len)
1949 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1950 sb.buf, sb.len, 0);
1951 strbuf_release(&sb);
1952 return 0;
1953 }
1954
1955 /*
1956 * '--color-words' algorithm can be described as:
1957 *
1958 * 1. collect the minus/plus lines of a diff hunk, divided into
1959 * minus-lines and plus-lines;
1960 *
1961 * 2. break both minus-lines and plus-lines into words and
1962 * place them into two mmfile_t with one word for each line;
1963 *
1964 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1965 *
1966 * And for the common parts of the both file, we output the plus side text.
1967 * diff_words->current_plus is used to trace the current position of the plus file
1968 * which printed. diff_words->last_minus is used to trace the last minus word
1969 * printed.
1970 *
1971 * For '--graph' to work with '--color-words', we need to output the graph prefix
1972 * on each line of color words output. Generally, there are two conditions on
1973 * which we should output the prefix.
1974 *
1975 * 1. diff_words->last_minus == 0 &&
1976 * diff_words->current_plus == diff_words->plus.text.ptr
1977 *
1978 * that is: the plus text must start as a new line, and if there is no minus
1979 * word printed, a graph prefix must be printed.
1980 *
1981 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1982 * *(diff_words->current_plus - 1) == '\n'
1983 *
1984 * that is: a graph prefix must be printed following a '\n'
1985 */
1986 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1987 {
1988 if ((diff_words->last_minus == 0 &&
1989 diff_words->current_plus == diff_words->plus.text.ptr) ||
1990 (diff_words->current_plus > diff_words->plus.text.ptr &&
1991 *(diff_words->current_plus - 1) == '\n')) {
1992 return 1;
1993 } else {
1994 return 0;
1995 }
1996 }
1997
1998 static void fn_out_diff_words_aux(void *priv,
1999 long minus_first, long minus_len,
2000 long plus_first, long plus_len,
2001 const char *func UNUSED, long funclen UNUSED)
2002 {
2003 struct diff_words_data *diff_words = priv;
2004 struct diff_words_style *style = diff_words->style;
2005 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
2006 struct diff_options *opt = diff_words->opt;
2007 const char *line_prefix;
2008
2009 assert(opt);
2010 line_prefix = diff_line_prefix(opt);
2011
2012 /* POSIX requires that first be decremented by one if len == 0... */
2013 if (minus_len) {
2014 minus_begin = diff_words->minus.orig[minus_first].begin;
2015 minus_end =
2016 diff_words->minus.orig[minus_first + minus_len - 1].end;
2017 } else
2018 minus_begin = minus_end =
2019 diff_words->minus.orig[minus_first].end;
2020
2021 if (plus_len) {
2022 plus_begin = diff_words->plus.orig[plus_first].begin;
2023 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
2024 } else
2025 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
2026
2027 if (color_words_output_graph_prefix(diff_words)) {
2028 fputs(line_prefix, diff_words->opt->file);
2029 }
2030 if (diff_words->current_plus != plus_begin) {
2031 fn_out_diff_words_write_helper(diff_words->opt,
2032 &style->ctx, style->newline,
2033 plus_begin - diff_words->current_plus,
2034 diff_words->current_plus);
2035 }
2036 if (minus_begin != minus_end) {
2037 fn_out_diff_words_write_helper(diff_words->opt,
2038 &style->old_word, style->newline,
2039 minus_end - minus_begin, minus_begin);
2040 }
2041 if (plus_begin != plus_end) {
2042 fn_out_diff_words_write_helper(diff_words->opt,
2043 &style->new_word, style->newline,
2044 plus_end - plus_begin, plus_begin);
2045 }
2046
2047 diff_words->current_plus = plus_end;
2048 diff_words->last_minus = minus_first;
2049 }
2050
2051 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2052 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2053 int *begin, int *end)
2054 {
2055 while (word_regex && *begin < buffer->size) {
2056 regmatch_t match[1];
2057 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2058 buffer->size - *begin, 1, match, 0)) {
2059 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2060 '\n', match[0].rm_eo - match[0].rm_so);
2061 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2062 *begin += match[0].rm_so;
2063 if (*begin == *end)
2064 (*begin)++;
2065 else
2066 return *begin > *end;
2067 } else {
2068 return -1;
2069 }
2070 }
2071
2072 /* find the next word */
2073 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2074 (*begin)++;
2075 if (*begin >= buffer->size)
2076 return -1;
2077
2078 /* find the end of the word */
2079 *end = *begin + 1;
2080 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2081 (*end)++;
2082
2083 return 0;
2084 }
2085
2086 /*
2087 * This function splits the words in buffer->text, stores the list with
2088 * newline separator into out, and saves the offsets of the original words
2089 * in buffer->orig.
2090 */
2091 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2092 regex_t *word_regex)
2093 {
2094 int i, j;
2095 long alloc = 0;
2096
2097 out->size = 0;
2098 out->ptr = NULL;
2099
2100 /* fake an empty "0th" word */
2101 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2102 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2103 buffer->orig_nr = 1;
2104
2105 for (i = 0; i < buffer->text.size; i++) {
2106 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2107 return;
2108
2109 /* store original boundaries */
2110 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2111 buffer->orig_alloc);
2112 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2113 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2114 buffer->orig_nr++;
2115
2116 /* store one word */
2117 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2118 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2119 out->ptr[out->size + j - i] = '\n';
2120 out->size += j - i + 1;
2121
2122 i = j - 1;
2123 }
2124 }
2125
2126 /* this executes the word diff on the accumulated buffers */
2127 static void diff_words_show(struct diff_words_data *diff_words)
2128 {
2129 xpparam_t xpp;
2130 xdemitconf_t xecfg;
2131 mmfile_t minus, plus;
2132 struct diff_words_style *style = diff_words->style;
2133
2134 struct diff_options *opt = diff_words->opt;
2135 const char *line_prefix;
2136
2137 assert(opt);
2138 line_prefix = diff_line_prefix(opt);
2139
2140 /* special case: only removal */
2141 if (!diff_words->plus.text.size) {
2142 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2143 line_prefix, strlen(line_prefix), 0);
2144 fn_out_diff_words_write_helper(diff_words->opt,
2145 &style->old_word, style->newline,
2146 diff_words->minus.text.size,
2147 diff_words->minus.text.ptr);
2148 diff_words->minus.text.size = 0;
2149 return;
2150 }
2151
2152 diff_words->current_plus = diff_words->plus.text.ptr;
2153 diff_words->last_minus = 0;
2154
2155 memset(&xpp, 0, sizeof(xpp));
2156 memset(&xecfg, 0, sizeof(xecfg));
2157 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2158 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2159 xpp.flags = 0;
2160 /* as only the hunk header will be parsed, we need a 0-context */
2161 xecfg.ctxlen = 0;
2162 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2163 diff_words, &xpp, &xecfg))
2164 die("unable to generate word diff");
2165 free(minus.ptr);
2166 free(plus.ptr);
2167 if (diff_words->current_plus != diff_words->plus.text.ptr +
2168 diff_words->plus.text.size) {
2169 if (color_words_output_graph_prefix(diff_words))
2170 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2171 line_prefix, strlen(line_prefix), 0);
2172 fn_out_diff_words_write_helper(diff_words->opt,
2173 &style->ctx, style->newline,
2174 diff_words->plus.text.ptr + diff_words->plus.text.size
2175 - diff_words->current_plus, diff_words->current_plus);
2176 }
2177 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2178 }
2179
2180 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2181 static void diff_words_flush(struct emit_callback *ecbdata)
2182 {
2183 struct diff_options *wo = ecbdata->diff_words->opt;
2184
2185 if (ecbdata->diff_words->minus.text.size ||
2186 ecbdata->diff_words->plus.text.size)
2187 diff_words_show(ecbdata->diff_words);
2188
2189 if (wo->emitted_symbols) {
2190 struct diff_options *o = ecbdata->opt;
2191 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2192 int i;
2193
2194 /*
2195 * NEEDSWORK:
2196 * Instead of appending each, concat all words to a line?
2197 */
2198 for (i = 0; i < wol->nr; i++)
2199 append_emitted_diff_symbol(o, &wol->buf[i]);
2200
2201 for (i = 0; i < wol->nr; i++)
2202 free((void *)wol->buf[i].line);
2203
2204 wol->nr = 0;
2205 }
2206 }
2207
2208 static void diff_filespec_load_driver(struct diff_filespec *one,
2209 struct index_state *istate)
2210 {
2211 /* Use already-loaded driver */
2212 if (one->driver)
2213 return;
2214
2215 if (S_ISREG(one->mode))
2216 one->driver = userdiff_find_by_path(istate, one->path);
2217
2218 /* Fallback to default settings */
2219 if (!one->driver)
2220 one->driver = userdiff_find_by_name("default");
2221 }
2222
2223 static const char *userdiff_word_regex(struct diff_filespec *one,
2224 struct index_state *istate)
2225 {
2226 diff_filespec_load_driver(one, istate);
2227 return one->driver->word_regex;
2228 }
2229
2230 static void init_diff_words_data(struct emit_callback *ecbdata,
2231 struct diff_options *orig_opts,
2232 struct diff_filespec *one,
2233 struct diff_filespec *two)
2234 {
2235 int i;
2236 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2237 memcpy(o, orig_opts, sizeof(struct diff_options));
2238
2239 CALLOC_ARRAY(ecbdata->diff_words, 1);
2240 ecbdata->diff_words->type = o->word_diff;
2241 ecbdata->diff_words->opt = o;
2242
2243 if (orig_opts->emitted_symbols)
2244 CALLOC_ARRAY(o->emitted_symbols, 1);
2245
2246 if (!o->word_regex)
2247 o->word_regex = userdiff_word_regex(one, o->repo->index);
2248 if (!o->word_regex)
2249 o->word_regex = userdiff_word_regex(two, o->repo->index);
2250 if (!o->word_regex)
2251 o->word_regex = diff_word_regex_cfg;
2252 if (o->word_regex) {
2253 ecbdata->diff_words->word_regex = (regex_t *)
2254 xmalloc(sizeof(regex_t));
2255 if (regcomp(ecbdata->diff_words->word_regex,
2256 o->word_regex,
2257 REG_EXTENDED | REG_NEWLINE))
2258 die("invalid regular expression: %s",
2259 o->word_regex);
2260 }
2261 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2262 if (o->word_diff == diff_words_styles[i].type) {
2263 ecbdata->diff_words->style =
2264 &diff_words_styles[i];
2265 break;
2266 }
2267 }
2268 if (want_color(o->use_color)) {
2269 struct diff_words_style *st = ecbdata->diff_words->style;
2270 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2271 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2272 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2273 }
2274 }
2275
2276 static void free_diff_words_data(struct emit_callback *ecbdata)
2277 {
2278 if (ecbdata->diff_words) {
2279 diff_words_flush(ecbdata);
2280 free_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);
2281 free (ecbdata->diff_words->opt);
2282 free (ecbdata->diff_words->minus.text.ptr);
2283 free (ecbdata->diff_words->minus.orig);
2284 free (ecbdata->diff_words->plus.text.ptr);
2285 free (ecbdata->diff_words->plus.orig);
2286 if (ecbdata->diff_words->word_regex) {
2287 regfree(ecbdata->diff_words->word_regex);
2288 free(ecbdata->diff_words->word_regex);
2289 }
2290 FREE_AND_NULL(ecbdata->diff_words);
2291 }
2292 }
2293
2294 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2295 {
2296 if (want_color(diff_use_color))
2297 return diff_colors[ix];
2298 return "";
2299 }
2300
2301 const char *diff_line_prefix(struct diff_options *opt)
2302 {
2303 struct strbuf *msgbuf;
2304 if (!opt->output_prefix)
2305 return "";
2306
2307 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2308 return msgbuf->buf;
2309 }
2310
2311 static unsigned long sane_truncate_line(char *line, unsigned long len)
2312 {
2313 const char *cp;
2314 unsigned long allot;
2315 size_t l = len;
2316
2317 cp = line;
2318 allot = l;
2319 while (0 < l) {
2320 (void) utf8_width(&cp, &l);
2321 if (!cp)
2322 break; /* truncated in the middle? */
2323 }
2324 return allot - l;
2325 }
2326
2327 static void find_lno(const char *line, struct emit_callback *ecbdata)
2328 {
2329 const char *p;
2330 ecbdata->lno_in_preimage = 0;
2331 ecbdata->lno_in_postimage = 0;
2332 p = strchr(line, '-');
2333 if (!p)
2334 return; /* cannot happen */
2335 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2336 p = strchr(p, '+');
2337 if (!p)
2338 return; /* cannot happen */
2339 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2340 }
2341
2342 static int fn_out_consume(void *priv, char *line, unsigned long len)
2343 {
2344 struct emit_callback *ecbdata = priv;
2345 struct diff_options *o = ecbdata->opt;
2346
2347 o->found_changes = 1;
2348
2349 if (ecbdata->header) {
2350 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2351 ecbdata->header->buf, ecbdata->header->len, 0);
2352 strbuf_reset(ecbdata->header);
2353 ecbdata->header = NULL;
2354 }
2355
2356 if (ecbdata->label_path[0]) {
2357 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2358 ecbdata->label_path[0],
2359 strlen(ecbdata->label_path[0]), 0);
2360 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2361 ecbdata->label_path[1],
2362 strlen(ecbdata->label_path[1]), 0);
2363 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2364 }
2365
2366 if (diff_suppress_blank_empty
2367 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2368 line[0] = '\n';
2369 len = 1;
2370 }
2371
2372 if (line[0] == '@') {
2373 if (ecbdata->diff_words)
2374 diff_words_flush(ecbdata);
2375 len = sane_truncate_line(line, len);
2376 find_lno(line, ecbdata);
2377 emit_hunk_header(ecbdata, line, len);
2378 return 0;
2379 }
2380
2381 if (ecbdata->diff_words) {
2382 enum diff_symbol s =
2383 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2384 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2385 if (line[0] == '-') {
2386 diff_words_append(line, len,
2387 &ecbdata->diff_words->minus);
2388 return 0;
2389 } else if (line[0] == '+') {
2390 diff_words_append(line, len,
2391 &ecbdata->diff_words->plus);
2392 return 0;
2393 } else if (starts_with(line, "\\ ")) {
2394 /*
2395 * Eat the "no newline at eof" marker as if we
2396 * saw a "+" or "-" line with nothing on it,
2397 * and return without diff_words_flush() to
2398 * defer processing. If this is the end of
2399 * preimage, more "+" lines may come after it.
2400 */
2401 return 0;
2402 }
2403 diff_words_flush(ecbdata);
2404 emit_diff_symbol(o, s, line, len, 0);
2405 return 0;
2406 }
2407
2408 switch (line[0]) {
2409 case '+':
2410 ecbdata->lno_in_postimage++;
2411 emit_add_line(ecbdata, line + 1, len - 1);
2412 break;
2413 case '-':
2414 ecbdata->lno_in_preimage++;
2415 emit_del_line(ecbdata, line + 1, len - 1);
2416 break;
2417 case ' ':
2418 ecbdata->lno_in_postimage++;
2419 ecbdata->lno_in_preimage++;
2420 emit_context_line(ecbdata, line + 1, len - 1);
2421 break;
2422 default:
2423 /* incomplete line at the end */
2424 ecbdata->lno_in_preimage++;
2425 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2426 line, len, 0);
2427 break;
2428 }
2429 return 0;
2430 }
2431
2432 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2433 {
2434 const char *old_name = a;
2435 const char *new_name = b;
2436 int pfx_length, sfx_length;
2437 int pfx_adjust_for_slash;
2438 int len_a = strlen(a);
2439 int len_b = strlen(b);
2440 int a_midlen, b_midlen;
2441 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2442 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2443
2444 if (qlen_a || qlen_b) {
2445 quote_c_style(a, name, NULL, 0);
2446 strbuf_addstr(name, " => ");
2447 quote_c_style(b, name, NULL, 0);
2448 return;
2449 }
2450
2451 /* Find common prefix */
2452 pfx_length = 0;
2453 while (*old_name && *new_name && *old_name == *new_name) {
2454 if (*old_name == '/')
2455 pfx_length = old_name - a + 1;
2456 old_name++;
2457 new_name++;
2458 }
2459
2460 /* Find common suffix */
2461 old_name = a + len_a;
2462 new_name = b + len_b;
2463 sfx_length = 0;
2464 /*
2465 * If there is a common prefix, it must end in a slash. In
2466 * that case we let this loop run 1 into the prefix to see the
2467 * same slash.
2468 *
2469 * If there is no common prefix, we cannot do this as it would
2470 * underrun the input strings.
2471 */
2472 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2473 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2474 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2475 *old_name == *new_name) {
2476 if (*old_name == '/')
2477 sfx_length = len_a - (old_name - a);
2478 old_name--;
2479 new_name--;
2480 }
2481
2482 /*
2483 * pfx{mid-a => mid-b}sfx
2484 * {pfx-a => pfx-b}sfx
2485 * pfx{sfx-a => sfx-b}
2486 * name-a => name-b
2487 */
2488 a_midlen = len_a - pfx_length - sfx_length;
2489 b_midlen = len_b - pfx_length - sfx_length;
2490 if (a_midlen < 0)
2491 a_midlen = 0;
2492 if (b_midlen < 0)
2493 b_midlen = 0;
2494
2495 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2496 if (pfx_length + sfx_length) {
2497 strbuf_add(name, a, pfx_length);
2498 strbuf_addch(name, '{');
2499 }
2500 strbuf_add(name, a + pfx_length, a_midlen);
2501 strbuf_addstr(name, " => ");
2502 strbuf_add(name, b + pfx_length, b_midlen);
2503 if (pfx_length + sfx_length) {
2504 strbuf_addch(name, '}');
2505 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2506 }
2507 }
2508
2509 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2510 const char *name_a,
2511 const char *name_b)
2512 {
2513 struct diffstat_file *x;
2514 CALLOC_ARRAY(x, 1);
2515 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2516 diffstat->files[diffstat->nr++] = x;
2517 if (name_b) {
2518 x->from_name = xstrdup(name_a);
2519 x->name = xstrdup(name_b);
2520 x->is_renamed = 1;
2521 }
2522 else {
2523 x->from_name = NULL;
2524 x->name = xstrdup(name_a);
2525 }
2526 return x;
2527 }
2528
2529 static int diffstat_consume(void *priv, char *line, unsigned long len)
2530 {
2531 struct diffstat_t *diffstat = priv;
2532 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2533
2534 if (!len)
2535 BUG("xdiff fed us an empty line");
2536
2537 if (line[0] == '+')
2538 x->added++;
2539 else if (line[0] == '-')
2540 x->deleted++;
2541 return 0;
2542 }
2543
2544 const char mime_boundary_leader[] = "------------";
2545
2546 static int scale_linear(int it, int width, int max_change)
2547 {
2548 if (!it)
2549 return 0;
2550 /*
2551 * make sure that at least one '-' or '+' is printed if
2552 * there is any change to this path. The easiest way is to
2553 * scale linearly as if the allotted width is one column shorter
2554 * than it is, and then add 1 to the result.
2555 */
2556 return 1 + (it * (width - 1) / max_change);
2557 }
2558
2559 static void show_graph(struct strbuf *out, char ch, int cnt,
2560 const char *set, const char *reset)
2561 {
2562 if (cnt <= 0)
2563 return;
2564 strbuf_addstr(out, set);
2565 strbuf_addchars(out, ch, cnt);
2566 strbuf_addstr(out, reset);
2567 }
2568
2569 static void fill_print_name(struct diffstat_file *file)
2570 {
2571 struct strbuf pname = STRBUF_INIT;
2572
2573 if (file->print_name)
2574 return;
2575
2576 if (file->is_renamed)
2577 pprint_rename(&pname, file->from_name, file->name);
2578 else
2579 quote_c_style(file->name, &pname, NULL, 0);
2580
2581 if (file->comments)
2582 strbuf_addf(&pname, " (%s)", file->comments);
2583
2584 file->print_name = strbuf_detach(&pname, NULL);
2585 }
2586
2587 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2588 int files, int insertions, int deletions)
2589 {
2590 struct strbuf sb = STRBUF_INIT;
2591
2592 if (!files) {
2593 assert(insertions == 0 && deletions == 0);
2594 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2595 NULL, 0, 0);
2596 return;
2597 }
2598
2599 strbuf_addf(&sb,
2600 (files == 1) ? " %d file changed" : " %d files changed",
2601 files);
2602
2603 /*
2604 * For binary diff, the caller may want to print "x files
2605 * changed" with insertions == 0 && deletions == 0.
2606 *
2607 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2608 * is probably less confusing (i.e skip over "2 files changed
2609 * but nothing about added/removed lines? Is this a bug in Git?").
2610 */
2611 if (insertions || deletions == 0) {
2612 strbuf_addf(&sb,
2613 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2614 insertions);
2615 }
2616
2617 if (deletions || insertions == 0) {
2618 strbuf_addf(&sb,
2619 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2620 deletions);
2621 }
2622 strbuf_addch(&sb, '\n');
2623 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2624 sb.buf, sb.len, 0);
2625 strbuf_release(&sb);
2626 }
2627
2628 void print_stat_summary(FILE *fp, int files,
2629 int insertions, int deletions)
2630 {
2631 struct diff_options o;
2632 memset(&o, 0, sizeof(o));
2633 o.file = fp;
2634
2635 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2636 }
2637
2638 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2639 {
2640 int i, len, add, del, adds = 0, dels = 0;
2641 uintmax_t max_change = 0, max_len = 0;
2642 int total_files = data->nr, count;
2643 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2644 const char *reset, *add_c, *del_c;
2645 int extra_shown = 0;
2646 const char *line_prefix = diff_line_prefix(options);
2647 struct strbuf out = STRBUF_INIT;
2648
2649 if (data->nr == 0)
2650 return;
2651
2652 count = options->stat_count ? options->stat_count : data->nr;
2653
2654 reset = diff_get_color_opt(options, DIFF_RESET);
2655 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2656 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2657
2658 /*
2659 * Find the longest filename and max number of changes
2660 */
2661 for (i = 0; (i < count) && (i < data->nr); i++) {
2662 struct diffstat_file *file = data->files[i];
2663 uintmax_t change = file->added + file->deleted;
2664
2665 if (!file->is_interesting && (change == 0)) {
2666 count++; /* not shown == room for one more */
2667 continue;
2668 }
2669 fill_print_name(file);
2670 len = utf8_strwidth(file->print_name);
2671 if (max_len < len)
2672 max_len = len;
2673
2674 if (file->is_unmerged) {
2675 /* "Unmerged" is 8 characters */
2676 bin_width = bin_width < 8 ? 8 : bin_width;
2677 continue;
2678 }
2679 if (file->is_binary) {
2680 /* "Bin XXX -> YYY bytes" */
2681 int w = 14 + decimal_width(file->added)
2682 + decimal_width(file->deleted);
2683 bin_width = bin_width < w ? w : bin_width;
2684 /* Display change counts aligned with "Bin" */
2685 number_width = 3;
2686 continue;
2687 }
2688
2689 if (max_change < change)
2690 max_change = change;
2691 }
2692 count = i; /* where we can stop scanning in data->files[] */
2693
2694 /*
2695 * We have width = stat_width or term_columns() columns total.
2696 * We want a maximum of min(max_len, stat_name_width) for the name part.
2697 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2698 * We also need 1 for " " and 4 + decimal_width(max_change)
2699 * for " | NNNN " and one the empty column at the end, altogether
2700 * 6 + decimal_width(max_change).
2701 *
2702 * If there's not enough space, we will use the smaller of
2703 * stat_name_width (if set) and 5/8*width for the filename,
2704 * and the rest for constant elements + graph part, but no more
2705 * than stat_graph_width for the graph part.
2706 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2707 * for the standard terminal size).
2708 *
2709 * In other words: stat_width limits the maximum width, and
2710 * stat_name_width fixes the maximum width of the filename,
2711 * and is also used to divide available columns if there
2712 * aren't enough.
2713 *
2714 * Binary files are displayed with "Bin XXX -> YYY bytes"
2715 * instead of the change count and graph. This part is treated
2716 * similarly to the graph part, except that it is not
2717 * "scaled". If total width is too small to accommodate the
2718 * guaranteed minimum width of the filename part and the
2719 * separators and this message, this message will "overflow"
2720 * making the line longer than the maximum width.
2721 */
2722
2723 /*
2724 * NEEDSWORK: line_prefix is often used for "log --graph" output
2725 * and contains ANSI-colored string. utf8_strnwidth() should be
2726 * used to correctly count the display width instead of strlen().
2727 */
2728 if (options->stat_width == -1)
2729 width = term_columns() - strlen(line_prefix);
2730 else
2731 width = options->stat_width ? options->stat_width : 80;
2732 number_width = decimal_width(max_change) > number_width ?
2733 decimal_width(max_change) : number_width;
2734
2735 if (options->stat_name_width == -1)
2736 options->stat_name_width = diff_stat_name_width;
2737 if (options->stat_graph_width == -1)
2738 options->stat_graph_width = diff_stat_graph_width;
2739
2740 /*
2741 * Guarantee 3/8*16 == 6 for the graph part
2742 * and 5/8*16 == 10 for the filename part
2743 */
2744 if (width < 16 + 6 + number_width)
2745 width = 16 + 6 + number_width;
2746
2747 /*
2748 * First assign sizes that are wanted, ignoring available width.
2749 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2750 * starting from "XXX" should fit in graph_width.
2751 */
2752 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2753 if (options->stat_graph_width &&
2754 options->stat_graph_width < graph_width)
2755 graph_width = options->stat_graph_width;
2756
2757 name_width = (options->stat_name_width > 0 &&
2758 options->stat_name_width < max_len) ?
2759 options->stat_name_width : max_len;
2760
2761 /*
2762 * Adjust adjustable widths not to exceed maximum width
2763 */
2764 if (name_width + number_width + 6 + graph_width > width) {
2765 if (graph_width > width * 3/8 - number_width - 6) {
2766 graph_width = width * 3/8 - number_width - 6;
2767 if (graph_width < 6)
2768 graph_width = 6;
2769 }
2770
2771 if (options->stat_graph_width &&
2772 graph_width > options->stat_graph_width)
2773 graph_width = options->stat_graph_width;
2774 if (name_width > width - number_width - 6 - graph_width)
2775 name_width = width - number_width - 6 - graph_width;
2776 else
2777 graph_width = width - number_width - 6 - name_width;
2778 }
2779
2780 /*
2781 * From here name_width is the width of the name area,
2782 * and graph_width is the width of the graph area.
2783 * max_change is used to scale graph properly.
2784 */
2785 for (i = 0; i < count; i++) {
2786 const char *prefix = "";
2787 struct diffstat_file *file = data->files[i];
2788 char *name = file->print_name;
2789 uintmax_t added = file->added;
2790 uintmax_t deleted = file->deleted;
2791 int name_len, padding;
2792
2793 if (!file->is_interesting && (added + deleted == 0))
2794 continue;
2795
2796 /*
2797 * "scale" the filename
2798 */
2799 len = name_width;
2800 name_len = utf8_strwidth(name);
2801 if (name_width < name_len) {
2802 char *slash;
2803 prefix = "...";
2804 len -= 3;
2805 /*
2806 * NEEDSWORK: (name_len - len) counts the display
2807 * width, which would be shorter than the byte
2808 * length of the corresponding substring.
2809 * Advancing "name" by that number of bytes does
2810 * *NOT* skip over that many columns, so it is
2811 * very likely that chomping the pathname at the
2812 * slash we will find starting from "name" will
2813 * leave the resulting string still too long.
2814 */
2815 name += name_len - len;
2816 slash = strchr(name, '/');
2817 if (slash)
2818 name = slash;
2819 }
2820 padding = len - utf8_strwidth(name);
2821 if (padding < 0)
2822 padding = 0;
2823
2824 if (file->is_binary) {
2825 strbuf_addf(&out, " %s%s%*s | %*s",
2826 prefix, name, padding, "",
2827 number_width, "Bin");
2828 if (!added && !deleted) {
2829 strbuf_addch(&out, '\n');
2830 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2831 out.buf, out.len, 0);
2832 strbuf_reset(&out);
2833 continue;
2834 }
2835 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2836 del_c, deleted, reset);
2837 strbuf_addstr(&out, " -> ");
2838 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2839 add_c, added, reset);
2840 strbuf_addstr(&out, " bytes\n");
2841 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2842 out.buf, out.len, 0);
2843 strbuf_reset(&out);
2844 continue;
2845 }
2846 else if (file->is_unmerged) {
2847 strbuf_addf(&out, " %s%s%*s | %*s",
2848 prefix, name, padding, "",
2849 number_width, "Unmerged\n");
2850 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2851 out.buf, out.len, 0);
2852 strbuf_reset(&out);
2853 continue;
2854 }
2855
2856 /*
2857 * scale the add/delete
2858 */
2859 add = added;
2860 del = deleted;
2861
2862 if (graph_width <= max_change) {
2863 int total = scale_linear(add + del, graph_width, max_change);
2864 if (total < 2 && add && del)
2865 /* width >= 2 due to the sanity check */
2866 total = 2;
2867 if (add < del) {
2868 add = scale_linear(add, graph_width, max_change);
2869 del = total - add;
2870 } else {
2871 del = scale_linear(del, graph_width, max_change);
2872 add = total - del;
2873 }
2874 }
2875 strbuf_addf(&out, " %s%s%*s | %*"PRIuMAX"%s",
2876 prefix, name, padding, "",
2877 number_width, added + deleted,
2878 added + deleted ? " " : "");
2879 show_graph(&out, '+', add, add_c, reset);
2880 show_graph(&out, '-', del, del_c, reset);
2881 strbuf_addch(&out, '\n');
2882 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2883 out.buf, out.len, 0);
2884 strbuf_reset(&out);
2885 }
2886
2887 for (i = 0; i < data->nr; i++) {
2888 struct diffstat_file *file = data->files[i];
2889 uintmax_t added = file->added;
2890 uintmax_t deleted = file->deleted;
2891
2892 if (file->is_unmerged ||
2893 (!file->is_interesting && (added + deleted == 0))) {
2894 total_files--;
2895 continue;
2896 }
2897
2898 if (!file->is_binary) {
2899 adds += added;
2900 dels += deleted;
2901 }
2902 if (i < count)
2903 continue;
2904 if (!extra_shown)
2905 emit_diff_symbol(options,
2906 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2907 NULL, 0, 0);
2908 extra_shown = 1;
2909 }
2910
2911 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2912 strbuf_release(&out);
2913 }
2914
2915 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2916 {
2917 int i, adds = 0, dels = 0, total_files = data->nr;
2918
2919 if (data->nr == 0)
2920 return;
2921
2922 for (i = 0; i < data->nr; i++) {
2923 int added = data->files[i]->added;
2924 int deleted = data->files[i]->deleted;
2925
2926 if (data->files[i]->is_unmerged ||
2927 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2928 total_files--;
2929 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2930 adds += added;
2931 dels += deleted;
2932 }
2933 }
2934 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2935 }
2936
2937 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2938 {
2939 int i;
2940
2941 if (data->nr == 0)
2942 return;
2943
2944 for (i = 0; i < data->nr; i++) {
2945 struct diffstat_file *file = data->files[i];
2946
2947 fprintf(options->file, "%s", diff_line_prefix(options));
2948
2949 if (file->is_binary)
2950 fprintf(options->file, "-\t-\t");
2951 else
2952 fprintf(options->file,
2953 "%"PRIuMAX"\t%"PRIuMAX"\t",
2954 file->added, file->deleted);
2955 if (options->line_termination) {
2956 fill_print_name(file);
2957 if (!file->is_renamed)
2958 write_name_quoted(file->name, options->file,
2959 options->line_termination);
2960 else {
2961 fputs(file->print_name, options->file);
2962 putc(options->line_termination, options->file);
2963 }
2964 } else {
2965 if (file->is_renamed) {
2966 putc('\0', options->file);
2967 write_name_quoted(file->from_name, options->file, '\0');
2968 }
2969 write_name_quoted(file->name, options->file, '\0');
2970 }
2971 }
2972 }
2973
2974 struct dirstat_file {
2975 const char *name;
2976 unsigned long changed;
2977 };
2978
2979 struct dirstat_dir {
2980 struct dirstat_file *files;
2981 int alloc, nr, permille, cumulative;
2982 };
2983
2984 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2985 unsigned long changed, const char *base, int baselen)
2986 {
2987 unsigned long sum_changes = 0;
2988 unsigned int sources = 0;
2989 const char *line_prefix = diff_line_prefix(opt);
2990
2991 while (dir->nr) {
2992 struct dirstat_file *f = dir->files;
2993 int namelen = strlen(f->name);
2994 unsigned long changes;
2995 char *slash;
2996
2997 if (namelen < baselen)
2998 break;
2999 if (memcmp(f->name, base, baselen))
3000 break;
3001 slash = strchr(f->name + baselen, '/');
3002 if (slash) {
3003 int newbaselen = slash + 1 - f->name;
3004 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
3005 sources++;
3006 } else {
3007 changes = f->changed;
3008 dir->files++;
3009 dir->nr--;
3010 sources += 2;
3011 }
3012 sum_changes += changes;
3013 }
3014
3015 /*
3016 * We don't report dirstat's for
3017 * - the top level
3018 * - or cases where everything came from a single directory
3019 * under this directory (sources == 1).
3020 */
3021 if (baselen && sources != 1) {
3022 if (sum_changes) {
3023 int permille = sum_changes * 1000 / changed;
3024 if (permille >= dir->permille) {
3025 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
3026 permille / 10, permille % 10, baselen, base);
3027 if (!dir->cumulative)
3028 return 0;
3029 }
3030 }
3031 }
3032 return sum_changes;
3033 }
3034
3035 static int dirstat_compare(const void *_a, const void *_b)
3036 {
3037 const struct dirstat_file *a = _a;
3038 const struct dirstat_file *b = _b;
3039 return strcmp(a->name, b->name);
3040 }
3041
3042 static void conclude_dirstat(struct diff_options *options,
3043 struct dirstat_dir *dir,
3044 unsigned long changed)
3045 {
3046 struct dirstat_file *to_free = dir->files;
3047
3048 if (!changed) {
3049 /* This can happen even with many files, if everything was renames */
3050 ;
3051 } else {
3052 /* Show all directories with more than x% of the changes */
3053 QSORT(dir->files, dir->nr, dirstat_compare);
3054 gather_dirstat(options, dir, changed, "", 0);
3055 }
3056
3057 free(to_free);
3058 }
3059
3060 static void show_dirstat(struct diff_options *options)
3061 {
3062 int i;
3063 unsigned long changed;
3064 struct dirstat_dir dir;
3065 struct diff_queue_struct *q = &diff_queued_diff;
3066
3067 dir.files = NULL;
3068 dir.alloc = 0;
3069 dir.nr = 0;
3070 dir.permille = options->dirstat_permille;
3071 dir.cumulative = options->flags.dirstat_cumulative;
3072
3073 changed = 0;
3074 for (i = 0; i < q->nr; i++) {
3075 struct diff_filepair *p = q->queue[i];
3076 const char *name;
3077 unsigned long copied, added, damage;
3078 struct diff_populate_filespec_options dpf_options = {
3079 .check_size_only = 1,
3080 };
3081
3082 name = p->two->path ? p->two->path : p->one->path;
3083
3084 if (p->one->oid_valid && p->two->oid_valid &&
3085 oideq(&p->one->oid, &p->two->oid)) {
3086 /*
3087 * The SHA1 has not changed, so pre-/post-content is
3088 * identical. We can therefore skip looking at the
3089 * file contents altogether.
3090 */
3091 damage = 0;
3092 goto found_damage;
3093 }
3094
3095 if (options->flags.dirstat_by_file) {
3096 /*
3097 * In --dirstat-by-file mode, we don't really need to
3098 * look at the actual file contents at all.
3099 * The fact that the SHA1 changed is enough for us to
3100 * add this file to the list of results
3101 * (with each file contributing equal damage).
3102 */
3103 damage = 1;
3104 goto found_damage;
3105 }
3106
3107 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3108 diff_populate_filespec(options->repo, p->one, NULL);
3109 diff_populate_filespec(options->repo, p->two, NULL);
3110 diffcore_count_changes(options->repo,
3111 p->one, p->two, NULL, NULL,
3112 &copied, &added);
3113 diff_free_filespec_data(p->one);
3114 diff_free_filespec_data(p->two);
3115 } else if (DIFF_FILE_VALID(p->one)) {
3116 diff_populate_filespec(options->repo, p->one, &dpf_options);
3117 copied = added = 0;
3118 diff_free_filespec_data(p->one);
3119 } else if (DIFF_FILE_VALID(p->two)) {
3120 diff_populate_filespec(options->repo, p->two, &dpf_options);
3121 copied = 0;
3122 added = p->two->size;
3123 diff_free_filespec_data(p->two);
3124 } else
3125 continue;
3126
3127 /*
3128 * Original minus copied is the removed material,
3129 * added is the new material. They are both damages
3130 * made to the preimage.
3131 * If the resulting damage is zero, we know that
3132 * diffcore_count_changes() considers the two entries to
3133 * be identical, but since the oid changed, we
3134 * know that there must have been _some_ kind of change,
3135 * so we force all entries to have damage > 0.
3136 */
3137 damage = (p->one->size - copied) + added;
3138 if (!damage)
3139 damage = 1;
3140
3141 found_damage:
3142 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3143 dir.files[dir.nr].name = name;
3144 dir.files[dir.nr].changed = damage;
3145 changed += damage;
3146 dir.nr++;
3147 }
3148
3149 conclude_dirstat(options, &dir, changed);
3150 }
3151
3152 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3153 {
3154 int i;
3155 unsigned long changed;
3156 struct dirstat_dir dir;
3157
3158 if (data->nr == 0)
3159 return;
3160
3161 dir.files = NULL;
3162 dir.alloc = 0;
3163 dir.nr = 0;
3164 dir.permille = options->dirstat_permille;
3165 dir.cumulative = options->flags.dirstat_cumulative;
3166
3167 changed = 0;
3168 for (i = 0; i < data->nr; i++) {
3169 struct diffstat_file *file = data->files[i];
3170 unsigned long damage = file->added + file->deleted;
3171 if (file->is_binary)
3172 /*
3173 * binary files counts bytes, not lines. Must find some
3174 * way to normalize binary bytes vs. textual lines.
3175 * The following heuristic assumes that there are 64
3176 * bytes per "line".
3177 * This is stupid and ugly, but very cheap...
3178 */
3179 damage = DIV_ROUND_UP(damage, 64);
3180 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3181 dir.files[dir.nr].name = file->name;
3182 dir.files[dir.nr].changed = damage;
3183 changed += damage;
3184 dir.nr++;
3185 }
3186
3187 conclude_dirstat(options, &dir, changed);
3188 }
3189
3190 static void free_diffstat_file(struct diffstat_file *f)
3191 {
3192 free(f->print_name);
3193 free(f->name);
3194 free(f->from_name);
3195 free(f);
3196 }
3197
3198 void free_diffstat_info(struct diffstat_t *diffstat)
3199 {
3200 int i;
3201 for (i = 0; i < diffstat->nr; i++)
3202 free_diffstat_file(diffstat->files[i]);
3203 free(diffstat->files);
3204 }
3205
3206 struct checkdiff_t {
3207 const char *filename;
3208 int lineno;
3209 int conflict_marker_size;
3210 struct diff_options *o;
3211 unsigned ws_rule;
3212 unsigned status;
3213 };
3214
3215 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3216 {
3217 char firstchar;
3218 int cnt;
3219
3220 if (len < marker_size + 1)
3221 return 0;
3222 firstchar = line[0];
3223 switch (firstchar) {
3224 case '=': case '>': case '<': case '|':
3225 break;
3226 default:
3227 return 0;
3228 }
3229 for (cnt = 1; cnt < marker_size; cnt++)
3230 if (line[cnt] != firstchar)
3231 return 0;
3232 /* line[1] through line[marker_size-1] are same as firstchar */
3233 if (len < marker_size + 1 || !isspace(line[marker_size]))
3234 return 0;
3235 return 1;
3236 }
3237
3238 static void checkdiff_consume_hunk(void *priv,
3239 long ob UNUSED, long on UNUSED,
3240 long nb, long nn UNUSED,
3241 const char *func UNUSED, long funclen UNUSED)
3242
3243 {
3244 struct checkdiff_t *data = priv;
3245 data->lineno = nb - 1;
3246 }
3247
3248 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3249 {
3250 struct checkdiff_t *data = priv;
3251 int marker_size = data->conflict_marker_size;
3252 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3253 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3254 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3255 char *err;
3256 const char *line_prefix;
3257
3258 assert(data->o);
3259 line_prefix = diff_line_prefix(data->o);
3260
3261 if (line[0] == '+') {
3262 unsigned bad;
3263 data->lineno++;
3264 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3265 data->status |= 1;
3266 fprintf(data->o->file,
3267 "%s%s:%d: leftover conflict marker\n",
3268 line_prefix, data->filename, data->lineno);
3269 }
3270 bad = ws_check(line + 1, len - 1, data->ws_rule);
3271 if (!bad)
3272 return 0;
3273 data->status |= bad;
3274 err = whitespace_error_string(bad);
3275 fprintf(data->o->file, "%s%s:%d: %s.\n",
3276 line_prefix, data->filename, data->lineno, err);
3277 free(err);
3278 emit_line(data->o, set, reset, line, 1);
3279 ws_check_emit(line + 1, len - 1, data->ws_rule,
3280 data->o->file, set, reset, ws);
3281 } else if (line[0] == ' ') {
3282 data->lineno++;
3283 }
3284 return 0;
3285 }
3286
3287 static unsigned char *deflate_it(char *data,
3288 unsigned long size,
3289 unsigned long *result_size)
3290 {
3291 int bound;
3292 unsigned char *deflated;
3293 git_zstream stream;
3294
3295 git_deflate_init(&stream, zlib_compression_level);
3296 bound = git_deflate_bound(&stream, size);
3297 deflated = xmalloc(bound);
3298 stream.next_out = deflated;
3299 stream.avail_out = bound;
3300
3301 stream.next_in = (unsigned char *)data;
3302 stream.avail_in = size;
3303 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3304 ; /* nothing */
3305 git_deflate_end(&stream);
3306 *result_size = stream.total_out;
3307 return deflated;
3308 }
3309
3310 static void emit_binary_diff_body(struct diff_options *o,
3311 mmfile_t *one, mmfile_t *two)
3312 {
3313 void *cp;
3314 void *delta;
3315 void *deflated;
3316 void *data;
3317 unsigned long orig_size;
3318 unsigned long delta_size;
3319 unsigned long deflate_size;
3320 unsigned long data_size;
3321
3322 /* We could do deflated delta, or we could do just deflated two,
3323 * whichever is smaller.
3324 */
3325 delta = NULL;
3326 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3327 if (one->size && two->size) {
3328 delta = diff_delta(one->ptr, one->size,
3329 two->ptr, two->size,
3330 &delta_size, deflate_size);
3331 if (delta) {
3332 void *to_free = delta;
3333 orig_size = delta_size;
3334 delta = deflate_it(delta, delta_size, &delta_size);
3335 free(to_free);
3336 }
3337 }
3338
3339 if (delta && delta_size < deflate_size) {
3340 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3341 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3342 s, strlen(s), 0);
3343 free(s);
3344 free(deflated);
3345 data = delta;
3346 data_size = delta_size;
3347 } else {
3348 char *s = xstrfmt("%lu", two->size);
3349 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3350 s, strlen(s), 0);
3351 free(s);
3352 free(delta);
3353 data = deflated;
3354 data_size = deflate_size;
3355 }
3356
3357 /* emit data encoded in base85 */
3358 cp = data;
3359 while (data_size) {
3360 int len;
3361 int bytes = (52 < data_size) ? 52 : data_size;
3362 char line[71];
3363 data_size -= bytes;
3364 if (bytes <= 26)
3365 line[0] = bytes + 'A' - 1;
3366 else
3367 line[0] = bytes - 26 + 'a' - 1;
3368 encode_85(line + 1, cp, bytes);
3369 cp = (char *) cp + bytes;
3370
3371 len = strlen(line);
3372 line[len++] = '\n';
3373 line[len] = '\0';
3374
3375 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3376 line, len, 0);
3377 }
3378 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3379 free(data);
3380 }
3381
3382 static void emit_binary_diff(struct diff_options *o,
3383 mmfile_t *one, mmfile_t *two)
3384 {
3385 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3386 emit_binary_diff_body(o, one, two);
3387 emit_binary_diff_body(o, two, one);
3388 }
3389
3390 int diff_filespec_is_binary(struct repository *r,
3391 struct diff_filespec *one)
3392 {
3393 struct diff_populate_filespec_options dpf_options = {
3394 .check_binary = 1,
3395 };
3396
3397 if (one->is_binary == -1) {
3398 diff_filespec_load_driver(one, r->index);
3399 if (one->driver->binary != -1)
3400 one->is_binary = one->driver->binary;
3401 else {
3402 if (!one->data && DIFF_FILE_VALID(one))
3403 diff_populate_filespec(r, one, &dpf_options);
3404 if (one->is_binary == -1 && one->data)
3405 one->is_binary = buffer_is_binary(one->data,
3406 one->size);
3407 if (one->is_binary == -1)
3408 one->is_binary = 0;
3409 }
3410 }
3411 return one->is_binary;
3412 }
3413
3414 static const struct userdiff_funcname *
3415 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3416 {
3417 diff_filespec_load_driver(one, o->repo->index);
3418 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3419 }
3420
3421 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3422 {
3423 if (!options->a_prefix)
3424 options->a_prefix = a;
3425 if (!options->b_prefix)
3426 options->b_prefix = b;
3427 }
3428
3429 void diff_set_noprefix(struct diff_options *options)
3430 {
3431 options->a_prefix = options->b_prefix = "";
3432 }
3433
3434 void diff_set_default_prefix(struct diff_options *options)
3435 {
3436 options->a_prefix = diff_src_prefix;
3437 options->b_prefix = diff_dst_prefix;
3438 }
3439
3440 struct userdiff_driver *get_textconv(struct repository *r,
3441 struct diff_filespec *one)
3442 {
3443 if (!DIFF_FILE_VALID(one))
3444 return NULL;
3445
3446 diff_filespec_load_driver(one, r->index);
3447 return userdiff_get_textconv(r, one->driver);
3448 }
3449
3450 static struct string_list *additional_headers(struct diff_options *o,
3451 const char *path)
3452 {
3453 if (!o->additional_path_headers)
3454 return NULL;
3455 return strmap_get(o->additional_path_headers, path);
3456 }
3457
3458 static void add_formatted_header(struct strbuf *msg,
3459 const char *header,
3460 const char *line_prefix,
3461 const char *meta,
3462 const char *reset)
3463 {
3464 const char *next, *newline;
3465
3466 for (next = header; *next; next = newline) {
3467 newline = strchrnul(next, '\n');
3468 strbuf_addf(msg, "%s%s%.*s%s\n", line_prefix, meta,
3469 (int)(newline - next), next, reset);
3470 if (*newline)
3471 newline++;
3472 }
3473 }
3474
3475 static void add_formatted_headers(struct strbuf *msg,
3476 struct string_list *more_headers,
3477 const char *line_prefix,
3478 const char *meta,
3479 const char *reset)
3480 {
3481 int i;
3482
3483 for (i = 0; i < more_headers->nr; i++)
3484 add_formatted_header(msg, more_headers->items[i].string,
3485 line_prefix, meta, reset);
3486 }
3487
3488 static int diff_filepair_is_phoney(struct diff_filespec *one,
3489 struct diff_filespec *two)
3490 {
3491 /*
3492 * This function specifically looks for pairs injected by
3493 * create_filepairs_for_header_only_notifications(). Such
3494 * pairs are "phoney" in that they do not represent any
3495 * content or even mode difference, but were inserted because
3496 * diff_queued_diff previously had no pair associated with
3497 * that path but we needed some pair to avoid losing the
3498 * "remerge CONFLICT" header associated with the path.
3499 */
3500 return !DIFF_FILE_VALID(one) && !DIFF_FILE_VALID(two);
3501 }
3502
3503 static int set_diff_algorithm(struct diff_options *opts,
3504 const char *alg)
3505 {
3506 long value = parse_algorithm_value(alg);
3507
3508 if (value < 0)
3509 return -1;
3510
3511 /* clear out previous settings */
3512 DIFF_XDL_CLR(opts, NEED_MINIMAL);
3513 opts->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3514 opts->xdl_opts |= value;
3515
3516 return 0;
3517 }
3518
3519 static void builtin_diff(const char *name_a,
3520 const char *name_b,
3521 struct diff_filespec *one,
3522 struct diff_filespec *two,
3523 const char *xfrm_msg,
3524 int must_show_header,
3525 struct diff_options *o,
3526 int complete_rewrite)
3527 {
3528 mmfile_t mf1, mf2;
3529 const char *lbl[2];
3530 char *a_one, *b_two;
3531 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3532 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3533 const char *a_prefix, *b_prefix;
3534 struct userdiff_driver *textconv_one = NULL;
3535 struct userdiff_driver *textconv_two = NULL;
3536 struct strbuf header = STRBUF_INIT;
3537 const char *line_prefix = diff_line_prefix(o);
3538
3539 diff_set_mnemonic_prefix(o, "a/", "b/");
3540 if (o->flags.reverse_diff) {
3541 a_prefix = o->b_prefix;
3542 b_prefix = o->a_prefix;
3543 } else {
3544 a_prefix = o->a_prefix;
3545 b_prefix = o->b_prefix;
3546 }
3547
3548 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3549 (!one->mode || S_ISGITLINK(one->mode)) &&
3550 (!two->mode || S_ISGITLINK(two->mode)) &&
3551 (!diff_filepair_is_phoney(one, two))) {
3552 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3553 &one->oid, &two->oid,
3554 two->dirty_submodule);
3555 return;
3556 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3557 (!one->mode || S_ISGITLINK(one->mode)) &&
3558 (!two->mode || S_ISGITLINK(two->mode)) &&
3559 (!diff_filepair_is_phoney(one, two))) {
3560 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3561 &one->oid, &two->oid,
3562 two->dirty_submodule);
3563 return;
3564 }
3565
3566 if (o->flags.allow_textconv) {
3567 textconv_one = get_textconv(o->repo, one);
3568 textconv_two = get_textconv(o->repo, two);
3569 }
3570
3571 /* Never use a non-valid filename anywhere if at all possible */
3572 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3573 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3574
3575 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3576 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3577 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3578 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3579 if (diff_filepair_is_phoney(one, two)) {
3580 /*
3581 * We should only reach this point for pairs generated from
3582 * create_filepairs_for_header_only_notifications(). For
3583 * these, we want to avoid the "/dev/null" special casing
3584 * above, because we do not want such pairs shown as either
3585 * "new file" or "deleted file" below.
3586 */
3587 lbl[0] = a_one;
3588 lbl[1] = b_two;
3589 }
3590 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3591 if (lbl[0][0] == '/') {
3592 /* /dev/null */
3593 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3594 if (xfrm_msg)
3595 strbuf_addstr(&header, xfrm_msg);
3596 o->found_changes = 1;
3597 must_show_header = 1;
3598 }
3599 else if (lbl[1][0] == '/') {
3600 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3601 if (xfrm_msg)
3602 strbuf_addstr(&header, xfrm_msg);
3603 o->found_changes = 1;
3604 must_show_header = 1;
3605 }
3606 else {
3607 if (one->mode != two->mode) {
3608 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3609 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3610 o->found_changes = 1;
3611 must_show_header = 1;
3612 }
3613 if (xfrm_msg)
3614 strbuf_addstr(&header, xfrm_msg);
3615
3616 /*
3617 * we do not run diff between different kind
3618 * of objects.
3619 */
3620 if ((one->mode ^ two->mode) & S_IFMT)
3621 goto free_ab_and_return;
3622 if (complete_rewrite &&
3623 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3624 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3625 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3626 header.buf, header.len, 0);
3627 strbuf_reset(&header);
3628 emit_rewrite_diff(name_a, name_b, one, two,
3629 textconv_one, textconv_two, o);
3630 o->found_changes = 1;
3631 goto free_ab_and_return;
3632 }
3633 }
3634
3635 if (o->irreversible_delete && lbl[1][0] == '/') {
3636 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3637 header.len, 0);
3638 strbuf_reset(&header);
3639 goto free_ab_and_return;
3640 } else if (!o->flags.text &&
3641 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3642 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3643 struct strbuf sb = STRBUF_INIT;
3644 if (!one->data && !two->data &&
3645 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3646 !o->flags.binary) {
3647 if (oideq(&one->oid, &two->oid)) {
3648 if (must_show_header)
3649 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3650 header.buf, header.len,
3651 0);
3652 goto free_ab_and_return;
3653 }
3654 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3655 header.buf, header.len, 0);
3656 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3657 diff_line_prefix(o), lbl[0], lbl[1]);
3658 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3659 sb.buf, sb.len, 0);
3660 strbuf_release(&sb);
3661 goto free_ab_and_return;
3662 }
3663 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3664 fill_mmfile(o->repo, &mf2, two) < 0)
3665 die("unable to read files to diff");
3666 /* Quite common confusing case */
3667 if (mf1.size == mf2.size &&
3668 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3669 if (must_show_header)
3670 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3671 header.buf, header.len, 0);
3672 goto free_ab_and_return;
3673 }
3674 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3675 strbuf_reset(&header);
3676 if (o->flags.binary)
3677 emit_binary_diff(o, &mf1, &mf2);
3678 else {
3679 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3680 diff_line_prefix(o), lbl[0], lbl[1]);
3681 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3682 sb.buf, sb.len, 0);
3683 strbuf_release(&sb);
3684 }
3685 o->found_changes = 1;
3686 } else {
3687 /* Crazy xdl interfaces.. */
3688 const char *diffopts;
3689 const char *v;
3690 xpparam_t xpp;
3691 xdemitconf_t xecfg;
3692 struct emit_callback ecbdata;
3693 const struct userdiff_funcname *pe;
3694
3695 if (must_show_header) {
3696 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3697 header.buf, header.len, 0);
3698 strbuf_reset(&header);
3699 }
3700
3701 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3702 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3703
3704 pe = diff_funcname_pattern(o, one);
3705 if (!pe)
3706 pe = diff_funcname_pattern(o, two);
3707
3708 memset(&xpp, 0, sizeof(xpp));
3709 memset(&xecfg, 0, sizeof(xecfg));
3710 memset(&ecbdata, 0, sizeof(ecbdata));
3711 if (o->flags.suppress_diff_headers)
3712 lbl[0] = NULL;
3713 ecbdata.label_path = lbl;
3714 ecbdata.color_diff = want_color(o->use_color);
3715 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3716 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3717 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3718 ecbdata.opt = o;
3719 if (header.len && !o->flags.suppress_diff_headers)
3720 ecbdata.header = &header;
3721 xpp.flags = o->xdl_opts;
3722 xpp.ignore_regex = o->ignore_regex;
3723 xpp.ignore_regex_nr = o->ignore_regex_nr;
3724 xpp.anchors = o->anchors;
3725 xpp.anchors_nr = o->anchors_nr;
3726 xecfg.ctxlen = o->context;
3727 xecfg.interhunkctxlen = o->interhunkcontext;
3728 xecfg.flags = XDL_EMIT_FUNCNAMES;
3729 if (o->flags.funccontext)
3730 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3731 if (pe)
3732 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3733
3734 diffopts = getenv("GIT_DIFF_OPTS");
3735 if (!diffopts)
3736 ;
3737 else if (skip_prefix(diffopts, "--unified=", &v))
3738 xecfg.ctxlen = strtoul(v, NULL, 10);
3739 else if (skip_prefix(diffopts, "-u", &v))
3740 xecfg.ctxlen = strtoul(v, NULL, 10);
3741
3742 if (o->word_diff)
3743 init_diff_words_data(&ecbdata, o, one, two);
3744 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3745 &ecbdata, &xpp, &xecfg))
3746 die("unable to generate diff for %s", one->path);
3747 if (o->word_diff)
3748 free_diff_words_data(&ecbdata);
3749 if (textconv_one)
3750 free(mf1.ptr);
3751 if (textconv_two)
3752 free(mf2.ptr);
3753 xdiff_clear_find_func(&xecfg);
3754 }
3755
3756 free_ab_and_return:
3757 strbuf_release(&header);
3758 diff_free_filespec_data(one);
3759 diff_free_filespec_data(two);
3760 free(a_one);
3761 free(b_two);
3762 return;
3763 }
3764
3765 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3766 {
3767 if (!is_renamed) {
3768 if (p->status == DIFF_STATUS_ADDED) {
3769 if (S_ISLNK(p->two->mode))
3770 return "new +l";
3771 else if ((p->two->mode & 0777) == 0755)
3772 return "new +x";
3773 else
3774 return "new";
3775 } else if (p->status == DIFF_STATUS_DELETED)
3776 return "gone";
3777 }
3778 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3779 return "mode -l";
3780 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3781 return "mode +l";
3782 else if ((p->one->mode & 0777) == 0644 &&
3783 (p->two->mode & 0777) == 0755)
3784 return "mode +x";
3785 else if ((p->one->mode & 0777) == 0755 &&
3786 (p->two->mode & 0777) == 0644)
3787 return "mode -x";
3788 return NULL;
3789 }
3790
3791 static void builtin_diffstat(const char *name_a, const char *name_b,
3792 struct diff_filespec *one,
3793 struct diff_filespec *two,
3794 struct diffstat_t *diffstat,
3795 struct diff_options *o,
3796 struct diff_filepair *p)
3797 {
3798 mmfile_t mf1, mf2;
3799 struct diffstat_file *data;
3800 int may_differ;
3801 int complete_rewrite = 0;
3802
3803 if (!DIFF_PAIR_UNMERGED(p)) {
3804 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3805 complete_rewrite = 1;
3806 }
3807
3808 data = diffstat_add(diffstat, name_a, name_b);
3809 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3810 if (o->flags.stat_with_summary)
3811 data->comments = get_compact_summary(p, data->is_renamed);
3812
3813 if (!one || !two) {
3814 data->is_unmerged = 1;
3815 return;
3816 }
3817
3818 /* saves some reads if true, not a guarantee of diff outcome */
3819 may_differ = !(one->oid_valid && two->oid_valid &&
3820 oideq(&one->oid, &two->oid));
3821
3822 if (diff_filespec_is_binary(o->repo, one) ||
3823 diff_filespec_is_binary(o->repo, two)) {
3824 data->is_binary = 1;
3825 if (!may_differ) {
3826 data->added = 0;
3827 data->deleted = 0;
3828 } else {
3829 data->added = diff_filespec_size(o->repo, two);
3830 data->deleted = diff_filespec_size(o->repo, one);
3831 }
3832 }
3833
3834 else if (complete_rewrite) {
3835 diff_populate_filespec(o->repo, one, NULL);
3836 diff_populate_filespec(o->repo, two, NULL);
3837 data->deleted = count_lines(one->data, one->size);
3838 data->added = count_lines(two->data, two->size);
3839 }
3840
3841 else if (may_differ) {
3842 /* Crazy xdl interfaces.. */
3843 xpparam_t xpp;
3844 xdemitconf_t xecfg;
3845
3846 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3847 fill_mmfile(o->repo, &mf2, two) < 0)
3848 die("unable to read files to diff");
3849
3850 memset(&xpp, 0, sizeof(xpp));
3851 memset(&xecfg, 0, sizeof(xecfg));
3852 xpp.flags = o->xdl_opts;
3853 xpp.ignore_regex = o->ignore_regex;
3854 xpp.ignore_regex_nr = o->ignore_regex_nr;
3855 xpp.anchors = o->anchors;
3856 xpp.anchors_nr = o->anchors_nr;
3857 xecfg.ctxlen = o->context;
3858 xecfg.interhunkctxlen = o->interhunkcontext;
3859 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3860 if (xdi_diff_outf(&mf1, &mf2, NULL,
3861 diffstat_consume, diffstat, &xpp, &xecfg))
3862 die("unable to generate diffstat for %s", one->path);
3863
3864 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3865 struct diffstat_file *file =
3866 diffstat->files[diffstat->nr - 1];
3867 /*
3868 * Omit diffstats of modified files where nothing changed.
3869 * Even if may_differ, this might be the case due to
3870 * ignoring whitespace changes, etc.
3871 *
3872 * But note that we special-case additions, deletions,
3873 * renames, and mode changes as adding an empty file,
3874 * for example is still of interest.
3875 */
3876 if ((p->status == DIFF_STATUS_MODIFIED)
3877 && !file->added
3878 && !file->deleted
3879 && one->mode == two->mode) {
3880 free_diffstat_file(file);
3881 diffstat->nr--;
3882 }
3883 }
3884 }
3885
3886 diff_free_filespec_data(one);
3887 diff_free_filespec_data(two);
3888 }
3889
3890 static void builtin_checkdiff(const char *name_a, const char *name_b,
3891 const char *attr_path,
3892 struct diff_filespec *one,
3893 struct diff_filespec *two,
3894 struct diff_options *o)
3895 {
3896 mmfile_t mf1, mf2;
3897 struct checkdiff_t data;
3898
3899 if (!two)
3900 return;
3901
3902 memset(&data, 0, sizeof(data));
3903 data.filename = name_b ? name_b : name_a;
3904 data.lineno = 0;
3905 data.o = o;
3906 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3907 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3908
3909 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3910 fill_mmfile(o->repo, &mf2, two) < 0)
3911 die("unable to read files to diff");
3912
3913 /*
3914 * All the other codepaths check both sides, but not checking
3915 * the "old" side here is deliberate. We are checking the newly
3916 * introduced changes, and as long as the "new" side is text, we
3917 * can and should check what it introduces.
3918 */
3919 if (diff_filespec_is_binary(o->repo, two))
3920 goto free_and_return;
3921 else {
3922 /* Crazy xdl interfaces.. */
3923 xpparam_t xpp;
3924 xdemitconf_t xecfg;
3925
3926 memset(&xpp, 0, sizeof(xpp));
3927 memset(&xecfg, 0, sizeof(xecfg));
3928 xecfg.ctxlen = 1; /* at least one context line */
3929 xpp.flags = 0;
3930 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3931 checkdiff_consume, &data,
3932 &xpp, &xecfg))
3933 die("unable to generate checkdiff for %s", one->path);
3934
3935 if (data.ws_rule & WS_BLANK_AT_EOF) {
3936 struct emit_callback ecbdata;
3937 int blank_at_eof;
3938
3939 ecbdata.ws_rule = data.ws_rule;
3940 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3941 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3942
3943 if (blank_at_eof) {
3944 static char *err;
3945 if (!err)
3946 err = whitespace_error_string(WS_BLANK_AT_EOF);
3947 fprintf(o->file, "%s:%d: %s.\n",
3948 data.filename, blank_at_eof, err);
3949 data.status = 1; /* report errors */
3950 }
3951 }
3952 }
3953 free_and_return:
3954 diff_free_filespec_data(one);
3955 diff_free_filespec_data(two);
3956 if (data.status)
3957 o->flags.check_failed = 1;
3958 }
3959
3960 struct diff_filespec *alloc_filespec(const char *path)
3961 {
3962 struct diff_filespec *spec;
3963
3964 FLEXPTR_ALLOC_STR(spec, path, path);
3965 spec->count = 1;
3966 spec->is_binary = -1;
3967 return spec;
3968 }
3969
3970 void free_filespec(struct diff_filespec *spec)
3971 {
3972 if (!--spec->count) {
3973 diff_free_filespec_data(spec);
3974 free(spec);
3975 }
3976 }
3977
3978 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3979 int oid_valid, unsigned short mode)
3980 {
3981 if (mode) {
3982 spec->mode = canon_mode(mode);
3983 oidcpy(&spec->oid, oid);
3984 spec->oid_valid = oid_valid;
3985 }
3986 }
3987
3988 /*
3989 * Given a name and sha1 pair, if the index tells us the file in
3990 * the work tree has that object contents, return true, so that
3991 * prepare_temp_file() does not have to inflate and extract.
3992 */
3993 static int reuse_worktree_file(struct index_state *istate,
3994 const char *name,
3995 const struct object_id *oid,
3996 int want_file)
3997 {
3998 const struct cache_entry *ce;
3999 struct stat st;
4000 int pos, len;
4001
4002 /*
4003 * We do not read the cache ourselves here, because the
4004 * benchmark with my previous version that always reads cache
4005 * shows that it makes things worse for diff-tree comparing
4006 * two linux-2.6 kernel trees in an already checked out work
4007 * tree. This is because most diff-tree comparisons deal with
4008 * only a small number of files, while reading the cache is
4009 * expensive for a large project, and its cost outweighs the
4010 * savings we get by not inflating the object to a temporary
4011 * file. Practically, this code only helps when we are used
4012 * by diff-cache --cached, which does read the cache before
4013 * calling us.
4014 */
4015 if (!istate->cache)
4016 return 0;
4017
4018 /* We want to avoid the working directory if our caller
4019 * doesn't need the data in a normal file, this system
4020 * is rather slow with its stat/open/mmap/close syscalls,
4021 * and the object is contained in a pack file. The pack
4022 * is probably already open and will be faster to obtain
4023 * the data through than the working directory. Loose
4024 * objects however would tend to be slower as they need
4025 * to be individually opened and inflated.
4026 */
4027 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
4028 return 0;
4029
4030 /*
4031 * Similarly, if we'd have to convert the file contents anyway, that
4032 * makes the optimization not worthwhile.
4033 */
4034 if (!want_file && would_convert_to_git(istate, name))
4035 return 0;
4036
4037 /*
4038 * If this path does not match our sparse-checkout definition,
4039 * then the file will not be in the working directory.
4040 */
4041 if (!path_in_sparse_checkout(name, istate))
4042 return 0;
4043
4044 len = strlen(name);
4045 pos = index_name_pos(istate, name, len);
4046 if (pos < 0)
4047 return 0;
4048 ce = istate->cache[pos];
4049
4050 /*
4051 * This is not the sha1 we are looking for, or
4052 * unreusable because it is not a regular file.
4053 */
4054 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
4055 return 0;
4056
4057 /*
4058 * If ce is marked as "assume unchanged", there is no
4059 * guarantee that work tree matches what we are looking for.
4060 */
4061 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
4062 return 0;
4063
4064 /*
4065 * If ce matches the file in the work tree, we can reuse it.
4066 */
4067 if (ce_uptodate(ce) ||
4068 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
4069 return 1;
4070
4071 return 0;
4072 }
4073
4074 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
4075 {
4076 struct strbuf buf = STRBUF_INIT;
4077 char *dirty = "";
4078
4079 /* Are we looking at the work tree? */
4080 if (s->dirty_submodule)
4081 dirty = "-dirty";
4082
4083 strbuf_addf(&buf, "Subproject commit %s%s\n",
4084 oid_to_hex(&s->oid), dirty);
4085 s->size = buf.len;
4086 if (size_only) {
4087 s->data = NULL;
4088 strbuf_release(&buf);
4089 } else {
4090 s->data = strbuf_detach(&buf, NULL);
4091 s->should_free = 1;
4092 }
4093 return 0;
4094 }
4095
4096 /*
4097 * While doing rename detection and pickaxe operation, we may need to
4098 * grab the data for the blob (or file) for our own in-core comparison.
4099 * diff_filespec has data and size fields for this purpose.
4100 */
4101 int diff_populate_filespec(struct repository *r,
4102 struct diff_filespec *s,
4103 const struct diff_populate_filespec_options *options)
4104 {
4105 int size_only = options ? options->check_size_only : 0;
4106 int check_binary = options ? options->check_binary : 0;
4107 int err = 0;
4108 int conv_flags = global_conv_flags_eol;
4109 /*
4110 * demote FAIL to WARN to allow inspecting the situation
4111 * instead of refusing.
4112 */
4113 if (conv_flags & CONV_EOL_RNDTRP_DIE)
4114 conv_flags = CONV_EOL_RNDTRP_WARN;
4115
4116 if (!DIFF_FILE_VALID(s))
4117 die("internal error: asking to populate invalid file.");
4118 if (S_ISDIR(s->mode))
4119 return -1;
4120
4121 if (s->data)
4122 return 0;
4123
4124 if (size_only && 0 < s->size)
4125 return 0;
4126
4127 if (S_ISGITLINK(s->mode))
4128 return diff_populate_gitlink(s, size_only);
4129
4130 if (!s->oid_valid ||
4131 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
4132 struct strbuf buf = STRBUF_INIT;
4133 struct stat st;
4134 int fd;
4135
4136 if (lstat(s->path, &st) < 0) {
4137 err_empty:
4138 err = -1;
4139 empty:
4140 s->data = (char *)"";
4141 s->size = 0;
4142 return err;
4143 }
4144 s->size = xsize_t(st.st_size);
4145 if (!s->size)
4146 goto empty;
4147 if (S_ISLNK(st.st_mode)) {
4148 struct strbuf sb = STRBUF_INIT;
4149
4150 if (strbuf_readlink(&sb, s->path, s->size))
4151 goto err_empty;
4152 s->size = sb.len;
4153 s->data = strbuf_detach(&sb, NULL);
4154 s->should_free = 1;
4155 return 0;
4156 }
4157
4158 /*
4159 * Even if the caller would be happy with getting
4160 * only the size, we cannot return early at this
4161 * point if the path requires us to run the content
4162 * conversion.
4163 */
4164 if (size_only && !would_convert_to_git(r->index, s->path))
4165 return 0;
4166
4167 /*
4168 * Note: this check uses xsize_t(st.st_size) that may
4169 * not be the true size of the blob after it goes
4170 * through convert_to_git(). This may not strictly be
4171 * correct, but the whole point of big_file_threshold
4172 * and is_binary check being that we want to avoid
4173 * opening the file and inspecting the contents, this
4174 * is probably fine.
4175 */
4176 if (check_binary &&
4177 s->size > big_file_threshold && s->is_binary == -1) {
4178 s->is_binary = 1;
4179 return 0;
4180 }
4181 fd = open(s->path, O_RDONLY);
4182 if (fd < 0)
4183 goto err_empty;
4184 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4185 close(fd);
4186 s->should_munmap = 1;
4187
4188 /*
4189 * Convert from working tree format to canonical git format
4190 */
4191 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4192 size_t size = 0;
4193 munmap(s->data, s->size);
4194 s->should_munmap = 0;
4195 s->data = strbuf_detach(&buf, &size);
4196 s->size = size;
4197 s->should_free = 1;
4198 }
4199 }
4200 else {
4201 struct object_info info = {
4202 .sizep = &s->size
4203 };
4204
4205 if (!(size_only || check_binary))
4206 /*
4207 * Set contentp, since there is no chance that merely
4208 * the size is sufficient.
4209 */
4210 info.contentp = &s->data;
4211
4212 if (options && options->missing_object_cb) {
4213 if (!oid_object_info_extended(r, &s->oid, &info,
4214 OBJECT_INFO_LOOKUP_REPLACE |
4215 OBJECT_INFO_SKIP_FETCH_OBJECT))
4216 goto object_read;
4217 options->missing_object_cb(options->missing_object_data);
4218 }
4219 if (oid_object_info_extended(r, &s->oid, &info,
4220 OBJECT_INFO_LOOKUP_REPLACE))
4221 die("unable to read %s", oid_to_hex(&s->oid));
4222
4223 object_read:
4224 if (size_only || check_binary) {
4225 if (size_only)
4226 return 0;
4227 if (s->size > big_file_threshold && s->is_binary == -1) {
4228 s->is_binary = 1;
4229 return 0;
4230 }
4231 }
4232 if (!info.contentp) {
4233 info.contentp = &s->data;
4234 if (oid_object_info_extended(r, &s->oid, &info,
4235 OBJECT_INFO_LOOKUP_REPLACE))
4236 die("unable to read %s", oid_to_hex(&s->oid));
4237 }
4238 s->should_free = 1;
4239 }
4240 return 0;
4241 }
4242
4243 void diff_free_filespec_blob(struct diff_filespec *s)
4244 {
4245 if (s->should_free)
4246 free(s->data);
4247 else if (s->should_munmap)
4248 munmap(s->data, s->size);
4249
4250 if (s->should_free || s->should_munmap) {
4251 s->should_free = s->should_munmap = 0;
4252 s->data = NULL;
4253 }
4254 }
4255
4256 void diff_free_filespec_data(struct diff_filespec *s)
4257 {
4258 if (!s)
4259 return;
4260
4261 diff_free_filespec_blob(s);
4262 FREE_AND_NULL(s->cnt_data);
4263 }
4264
4265 static void prep_temp_blob(struct index_state *istate,
4266 const char *path, struct diff_tempfile *temp,
4267 void *blob,
4268 unsigned long size,
4269 const struct object_id *oid,
4270 int mode)
4271 {
4272 struct strbuf buf = STRBUF_INIT;
4273 char *path_dup = xstrdup(path);
4274 const char *base = basename(path_dup);
4275 struct checkout_metadata meta;
4276
4277 init_checkout_metadata(&meta, NULL, NULL, oid);
4278
4279 temp->tempfile = mks_tempfile_dt("git-blob-XXXXXX", base);
4280 if (!temp->tempfile)
4281 die_errno("unable to create temp-file");
4282 if (convert_to_working_tree(istate, path,
4283 (const char *)blob, (size_t)size, &buf, &meta)) {
4284 blob = buf.buf;
4285 size = buf.len;
4286 }
4287 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4288 close_tempfile_gently(temp->tempfile))
4289 die_errno("unable to write temp-file");
4290 temp->name = get_tempfile_path(temp->tempfile);
4291 oid_to_hex_r(temp->hex, oid);
4292 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4293 strbuf_release(&buf);
4294 free(path_dup);
4295 }
4296
4297 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4298 struct diff_filespec *one)
4299 {
4300 struct diff_tempfile *temp = claim_diff_tempfile();
4301
4302 if (!DIFF_FILE_VALID(one)) {
4303 not_a_valid_file:
4304 /* A '-' entry produces this for file-2, and
4305 * a '+' entry produces this for file-1.
4306 */
4307 temp->name = "/dev/null";
4308 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4309 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4310 return temp;
4311 }
4312
4313 if (!S_ISGITLINK(one->mode) &&
4314 (!one->oid_valid ||
4315 reuse_worktree_file(r->index, one->path, &one->oid, 1))) {
4316 struct stat st;
4317 if (lstat(one->path, &st) < 0) {
4318 if (errno == ENOENT)
4319 goto not_a_valid_file;
4320 die_errno("stat(%s)", one->path);
4321 }
4322 if (S_ISLNK(st.st_mode)) {
4323 struct strbuf sb = STRBUF_INIT;
4324 if (strbuf_readlink(&sb, one->path, st.st_size) < 0)
4325 die_errno("readlink(%s)", one->path);
4326 prep_temp_blob(r->index, one->path, temp, sb.buf, sb.len,
4327 (one->oid_valid ?
4328 &one->oid : null_oid()),
4329 (one->oid_valid ?
4330 one->mode : S_IFLNK));
4331 strbuf_release(&sb);
4332 }
4333 else {
4334 /* we can borrow from the file in the work tree */
4335 temp->name = one->path;
4336 if (!one->oid_valid)
4337 oid_to_hex_r(temp->hex, null_oid());
4338 else
4339 oid_to_hex_r(temp->hex, &one->oid);
4340 /* Even though we may sometimes borrow the
4341 * contents from the work tree, we always want
4342 * one->mode. mode is trustworthy even when
4343 * !(one->oid_valid), as long as
4344 * DIFF_FILE_VALID(one).
4345 */
4346 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4347 }
4348 return temp;
4349 }
4350 else {
4351 if (diff_populate_filespec(r, one, NULL))
4352 die("cannot read data blob for %s", one->path);
4353 prep_temp_blob(r->index, one->path, temp,
4354 one->data, one->size,
4355 &one->oid, one->mode);
4356 }
4357 return temp;
4358 }
4359
4360 static void add_external_diff_name(struct repository *r,
4361 struct strvec *argv,
4362 struct diff_filespec *df)
4363 {
4364 struct diff_tempfile *temp = prepare_temp_file(r, df);
4365 strvec_push(argv, temp->name);
4366 strvec_push(argv, temp->hex);
4367 strvec_push(argv, temp->mode);
4368 }
4369
4370 /* An external diff command takes:
4371 *
4372 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4373 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4374 *
4375 */
4376 static void run_external_diff(const char *pgm,
4377 const char *name,
4378 const char *other,
4379 struct diff_filespec *one,
4380 struct diff_filespec *two,
4381 const char *xfrm_msg,
4382 struct diff_options *o)
4383 {
4384 struct child_process cmd = CHILD_PROCESS_INIT;
4385 struct diff_queue_struct *q = &diff_queued_diff;
4386
4387 strvec_push(&cmd.args, pgm);
4388 strvec_push(&cmd.args, name);
4389
4390 if (one && two) {
4391 add_external_diff_name(o->repo, &cmd.args, one);
4392 add_external_diff_name(o->repo, &cmd.args, two);
4393 if (other) {
4394 strvec_push(&cmd.args, other);
4395 if (xfrm_msg)
4396 strvec_push(&cmd.args, xfrm_msg);
4397 }
4398 }
4399
4400 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_COUNTER=%d",
4401 ++o->diff_path_counter);
4402 strvec_pushf(&cmd.env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4403
4404 diff_free_filespec_data(one);
4405 diff_free_filespec_data(two);
4406 cmd.use_shell = 1;
4407 if (run_command(&cmd))
4408 die(_("external diff died, stopping at %s"), name);
4409
4410 remove_tempfile();
4411 }
4412
4413 static int similarity_index(struct diff_filepair *p)
4414 {
4415 return p->score * 100 / MAX_SCORE;
4416 }
4417
4418 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4419 {
4420 if (startup_info->have_repository)
4421 return repo_find_unique_abbrev(the_repository, oid, abbrev);
4422 else {
4423 char *hex = oid_to_hex(oid);
4424 if (abbrev < 0)
4425 abbrev = FALLBACK_DEFAULT_ABBREV;
4426 if (abbrev > the_hash_algo->hexsz)
4427 BUG("oid abbreviation out of range: %d", abbrev);
4428 if (abbrev)
4429 hex[abbrev] = '\0';
4430 return hex;
4431 }
4432 }
4433
4434 static void fill_metainfo(struct strbuf *msg,
4435 const char *name,
4436 const char *other,
4437 struct diff_filespec *one,
4438 struct diff_filespec *two,
4439 struct diff_options *o,
4440 struct diff_filepair *p,
4441 int *must_show_header,
4442 int use_color)
4443 {
4444 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4445 const char *reset = diff_get_color(use_color, DIFF_RESET);
4446 const char *line_prefix = diff_line_prefix(o);
4447 struct string_list *more_headers = NULL;
4448
4449 *must_show_header = 1;
4450 strbuf_init(msg, PATH_MAX * 2 + 300);
4451 switch (p->status) {
4452 case DIFF_STATUS_COPIED:
4453 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4454 line_prefix, set, similarity_index(p));
4455 strbuf_addf(msg, "%s\n%s%scopy from ",
4456 reset, line_prefix, set);
4457 quote_c_style(name, msg, NULL, 0);
4458 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4459 quote_c_style(other, msg, NULL, 0);
4460 strbuf_addf(msg, "%s\n", reset);
4461 break;
4462 case DIFF_STATUS_RENAMED:
4463 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4464 line_prefix, set, similarity_index(p));
4465 strbuf_addf(msg, "%s\n%s%srename from ",
4466 reset, line_prefix, set);
4467 quote_c_style(name, msg, NULL, 0);
4468 strbuf_addf(msg, "%s\n%s%srename to ",
4469 reset, line_prefix, set);
4470 quote_c_style(other, msg, NULL, 0);
4471 strbuf_addf(msg, "%s\n", reset);
4472 break;
4473 case DIFF_STATUS_MODIFIED:
4474 if (p->score) {
4475 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4476 line_prefix,
4477 set, similarity_index(p), reset);
4478 break;
4479 }
4480 /* fallthru */
4481 default:
4482 *must_show_header = 0;
4483 }
4484 if ((more_headers = additional_headers(o, name))) {
4485 add_formatted_headers(msg, more_headers,
4486 line_prefix, set, reset);
4487 *must_show_header = 1;
4488 }
4489 if (one && two && !oideq(&one->oid, &two->oid)) {
4490 const unsigned hexsz = the_hash_algo->hexsz;
4491 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4492
4493 if (o->flags.full_index)
4494 abbrev = hexsz;
4495
4496 if (o->flags.binary) {
4497 mmfile_t mf;
4498 if ((!fill_mmfile(o->repo, &mf, one) &&
4499 diff_filespec_is_binary(o->repo, one)) ||
4500 (!fill_mmfile(o->repo, &mf, two) &&
4501 diff_filespec_is_binary(o->repo, two)))
4502 abbrev = hexsz;
4503 }
4504 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4505 diff_abbrev_oid(&one->oid, abbrev),
4506 diff_abbrev_oid(&two->oid, abbrev));
4507 if (one->mode == two->mode)
4508 strbuf_addf(msg, " %06o", one->mode);
4509 strbuf_addf(msg, "%s\n", reset);
4510 }
4511 }
4512
4513 static void run_diff_cmd(const char *pgm,
4514 const char *name,
4515 const char *other,
4516 const char *attr_path,
4517 struct diff_filespec *one,
4518 struct diff_filespec *two,
4519 struct strbuf *msg,
4520 struct diff_options *o,
4521 struct diff_filepair *p)
4522 {
4523 const char *xfrm_msg = NULL;
4524 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4525 int must_show_header = 0;
4526 struct userdiff_driver *drv = NULL;
4527
4528 if (o->flags.allow_external || !o->ignore_driver_algorithm)
4529 drv = userdiff_find_by_path(o->repo->index, attr_path);
4530
4531 if (o->flags.allow_external && drv && drv->external)
4532 pgm = drv->external;
4533
4534 if (msg) {
4535 /*
4536 * don't use colors when the header is intended for an
4537 * external diff driver
4538 */
4539 fill_metainfo(msg, name, other, one, two, o, p,
4540 &must_show_header,
4541 want_color(o->use_color) && !pgm);
4542 xfrm_msg = msg->len ? msg->buf : NULL;
4543 }
4544
4545 if (pgm) {
4546 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4547 return;
4548 }
4549 if (one && two) {
4550 if (!o->ignore_driver_algorithm && drv && drv->algorithm)
4551 set_diff_algorithm(o, drv->algorithm);
4552
4553 builtin_diff(name, other ? other : name,
4554 one, two, xfrm_msg, must_show_header,
4555 o, complete_rewrite);
4556 } else {
4557 fprintf(o->file, "* Unmerged path %s\n", name);
4558 }
4559 }
4560
4561 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4562 {
4563 if (DIFF_FILE_VALID(one)) {
4564 if (!one->oid_valid) {
4565 struct stat st;
4566 if (one->is_stdin) {
4567 oidclr(&one->oid);
4568 return;
4569 }
4570 if (lstat(one->path, &st) < 0)
4571 die_errno("stat '%s'", one->path);
4572 if (index_path(istate, &one->oid, one->path, &st, 0))
4573 die("cannot hash %s", one->path);
4574 }
4575 }
4576 else
4577 oidclr(&one->oid);
4578 }
4579
4580 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4581 {
4582 /* Strip the prefix but do not molest /dev/null and absolute paths */
4583 if (*namep && !is_absolute_path(*namep)) {
4584 *namep += prefix_length;
4585 if (**namep == '/')
4586 ++*namep;
4587 }
4588 if (*otherp && !is_absolute_path(*otherp)) {
4589 *otherp += prefix_length;
4590 if (**otherp == '/')
4591 ++*otherp;
4592 }
4593 }
4594
4595 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4596 {
4597 const char *pgm = external_diff();
4598 struct strbuf msg;
4599 struct diff_filespec *one = p->one;
4600 struct diff_filespec *two = p->two;
4601 const char *name;
4602 const char *other;
4603 const char *attr_path;
4604
4605 name = one->path;
4606 other = (strcmp(name, two->path) ? two->path : NULL);
4607 attr_path = name;
4608 if (o->prefix_length)
4609 strip_prefix(o->prefix_length, &name, &other);
4610
4611 if (!o->flags.allow_external)
4612 pgm = NULL;
4613
4614 if (DIFF_PAIR_UNMERGED(p)) {
4615 run_diff_cmd(pgm, name, NULL, attr_path,
4616 NULL, NULL, NULL, o, p);
4617 return;
4618 }
4619
4620 diff_fill_oid_info(one, o->repo->index);
4621 diff_fill_oid_info(two, o->repo->index);
4622
4623 if (!pgm &&
4624 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4625 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4626 /*
4627 * a filepair that changes between file and symlink
4628 * needs to be split into deletion and creation.
4629 */
4630 struct diff_filespec *null = alloc_filespec(two->path);
4631 run_diff_cmd(NULL, name, other, attr_path,
4632 one, null, &msg,
4633 o, p);
4634 free(null);
4635 strbuf_release(&msg);
4636
4637 null = alloc_filespec(one->path);
4638 run_diff_cmd(NULL, name, other, attr_path,
4639 null, two, &msg, o, p);
4640 free(null);
4641 }
4642 else
4643 run_diff_cmd(pgm, name, other, attr_path,
4644 one, two, &msg, o, p);
4645
4646 strbuf_release(&msg);
4647 }
4648
4649 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4650 struct diffstat_t *diffstat)
4651 {
4652 const char *name;
4653 const char *other;
4654
4655 if (!o->ignore_driver_algorithm) {
4656 struct userdiff_driver *drv = userdiff_find_by_path(o->repo->index,
4657 p->one->path);
4658
4659 if (drv && drv->algorithm)
4660 set_diff_algorithm(o, drv->algorithm);
4661 }
4662
4663 if (DIFF_PAIR_UNMERGED(p)) {
4664 /* unmerged */
4665 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4666 diffstat, o, p);
4667 return;
4668 }
4669
4670 name = p->one->path;
4671 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4672
4673 if (o->prefix_length)
4674 strip_prefix(o->prefix_length, &name, &other);
4675
4676 diff_fill_oid_info(p->one, o->repo->index);
4677 diff_fill_oid_info(p->two, o->repo->index);
4678
4679 builtin_diffstat(name, other, p->one, p->two,
4680 diffstat, o, p);
4681 }
4682
4683 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4684 {
4685 const char *name;
4686 const char *other;
4687 const char *attr_path;
4688
4689 if (DIFF_PAIR_UNMERGED(p)) {
4690 /* unmerged */
4691 return;
4692 }
4693
4694 name = p->one->path;
4695 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4696 attr_path = other ? other : name;
4697
4698 if (o->prefix_length)
4699 strip_prefix(o->prefix_length, &name, &other);
4700
4701 diff_fill_oid_info(p->one, o->repo->index);
4702 diff_fill_oid_info(p->two, o->repo->index);
4703
4704 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4705 }
4706
4707 void repo_diff_setup(struct repository *r, struct diff_options *options)
4708 {
4709 memcpy(options, &default_diff_options, sizeof(*options));
4710
4711 options->file = stdout;
4712 options->repo = r;
4713
4714 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4715 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4716 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4717 options->abbrev = DEFAULT_ABBREV;
4718 options->line_termination = '\n';
4719 options->break_opt = -1;
4720 options->rename_limit = -1;
4721 options->dirstat_permille = diff_dirstat_permille_default;
4722 options->context = diff_context_default;
4723 options->interhunkcontext = diff_interhunk_context_default;
4724 options->ws_error_highlight = ws_error_highlight_default;
4725 options->flags.rename_empty = 1;
4726 options->flags.relative_name = diff_relative;
4727 options->objfind = NULL;
4728
4729 /* pathchange left =NULL by default */
4730 options->change = diff_change;
4731 options->add_remove = diff_addremove;
4732 options->use_color = diff_use_color_default;
4733 options->detect_rename = diff_detect_rename_default;
4734 options->xdl_opts |= diff_algorithm;
4735 if (diff_indent_heuristic)
4736 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4737
4738 options->orderfile = diff_order_file_cfg;
4739
4740 if (!options->flags.ignore_submodule_set)
4741 options->flags.ignore_untracked_in_submodules = 1;
4742
4743 if (diff_no_prefix) {
4744 diff_set_noprefix(options);
4745 } else if (!diff_mnemonic_prefix) {
4746 diff_set_default_prefix(options);
4747 }
4748
4749 options->color_moved = diff_color_moved_default;
4750 options->color_moved_ws_handling = diff_color_moved_ws_default;
4751 }
4752
4753 static const char diff_status_letters[] = {
4754 DIFF_STATUS_ADDED,
4755 DIFF_STATUS_COPIED,
4756 DIFF_STATUS_DELETED,
4757 DIFF_STATUS_MODIFIED,
4758 DIFF_STATUS_RENAMED,
4759 DIFF_STATUS_TYPE_CHANGED,
4760 DIFF_STATUS_UNKNOWN,
4761 DIFF_STATUS_UNMERGED,
4762 DIFF_STATUS_FILTER_AON,
4763 DIFF_STATUS_FILTER_BROKEN,
4764 '\0',
4765 };
4766
4767 static unsigned int filter_bit['Z' + 1];
4768
4769 static void prepare_filter_bits(void)
4770 {
4771 int i;
4772
4773 if (!filter_bit[DIFF_STATUS_ADDED]) {
4774 for (i = 0; diff_status_letters[i]; i++)
4775 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4776 }
4777 }
4778
4779 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4780 {
4781 return opt->filter & filter_bit[(int) status];
4782 }
4783
4784 unsigned diff_filter_bit(char status)
4785 {
4786 prepare_filter_bits();
4787 return filter_bit[(int) status];
4788 }
4789
4790 int diff_check_follow_pathspec(struct pathspec *ps, int die_on_error)
4791 {
4792 unsigned forbidden_magic;
4793
4794 if (ps->nr != 1) {
4795 if (die_on_error)
4796 die(_("--follow requires exactly one pathspec"));
4797 return 0;
4798 }
4799
4800 forbidden_magic = ps->items[0].magic & ~(PATHSPEC_FROMTOP |
4801 PATHSPEC_LITERAL);
4802 if (forbidden_magic) {
4803 if (die_on_error) {
4804 struct strbuf sb = STRBUF_INIT;
4805 pathspec_magic_names(forbidden_magic, &sb);
4806 die(_("pathspec magic not supported by --follow: %s"),
4807 sb.buf);
4808 }
4809 return 0;
4810 }
4811
4812 return 1;
4813 }
4814
4815 void diff_setup_done(struct diff_options *options)
4816 {
4817 unsigned check_mask = DIFF_FORMAT_NAME |
4818 DIFF_FORMAT_NAME_STATUS |
4819 DIFF_FORMAT_CHECKDIFF |
4820 DIFF_FORMAT_NO_OUTPUT;
4821 /*
4822 * This must be signed because we're comparing against a potentially
4823 * negative value.
4824 */
4825 const int hexsz = the_hash_algo->hexsz;
4826
4827 if (options->set_default)
4828 options->set_default(options);
4829
4830 if (HAS_MULTI_BITS(options->output_format & check_mask))
4831 die(_("options '%s', '%s', '%s', and '%s' cannot be used together"),
4832 "--name-only", "--name-status", "--check", "-s");
4833
4834 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4835 die(_("options '%s', '%s', and '%s' cannot be used together"),
4836 "-G", "-S", "--find-object");
4837
4838 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4839 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s'"),
4840 "-G", "--pickaxe-regex", "--pickaxe-regex", "-S");
4841
4842 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4843 die(_("options '%s' and '%s' cannot be used together, use '%s' with '%s' and '%s'"),
4844 "--pickaxe-all", "--find-object", "--pickaxe-all", "-G", "-S");
4845
4846 /*
4847 * Most of the time we can say "there are changes"
4848 * only by checking if there are changed paths, but
4849 * --ignore-whitespace* options force us to look
4850 * inside contents.
4851 */
4852
4853 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4854 options->ignore_regex_nr)
4855 options->flags.diff_from_contents = 1;
4856 else
4857 options->flags.diff_from_contents = 0;
4858
4859 if (options->flags.find_copies_harder)
4860 options->detect_rename = DIFF_DETECT_COPY;
4861
4862 if (!options->flags.relative_name)
4863 options->prefix = NULL;
4864 if (options->prefix)
4865 options->prefix_length = strlen(options->prefix);
4866 else
4867 options->prefix_length = 0;
4868
4869 /*
4870 * --name-only, --name-status, --checkdiff, and -s
4871 * turn other output format off.
4872 */
4873 if (options->output_format & (DIFF_FORMAT_NAME |
4874 DIFF_FORMAT_NAME_STATUS |
4875 DIFF_FORMAT_CHECKDIFF |
4876 DIFF_FORMAT_NO_OUTPUT))
4877 options->output_format &= ~(DIFF_FORMAT_RAW |
4878 DIFF_FORMAT_NUMSTAT |
4879 DIFF_FORMAT_DIFFSTAT |
4880 DIFF_FORMAT_SHORTSTAT |
4881 DIFF_FORMAT_DIRSTAT |
4882 DIFF_FORMAT_SUMMARY |
4883 DIFF_FORMAT_PATCH);
4884
4885 /*
4886 * These cases always need recursive; we do not drop caller-supplied
4887 * recursive bits for other formats here.
4888 */
4889 if (options->output_format & (DIFF_FORMAT_PATCH |
4890 DIFF_FORMAT_NUMSTAT |
4891 DIFF_FORMAT_DIFFSTAT |
4892 DIFF_FORMAT_SHORTSTAT |
4893 DIFF_FORMAT_DIRSTAT |
4894 DIFF_FORMAT_SUMMARY |
4895 DIFF_FORMAT_CHECKDIFF))
4896 options->flags.recursive = 1;
4897 /*
4898 * Also pickaxe would not work very well if you do not say recursive
4899 */
4900 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4901 options->flags.recursive = 1;
4902 /*
4903 * When patches are generated, submodules diffed against the work tree
4904 * must be checked for dirtiness too so it can be shown in the output
4905 */
4906 if (options->output_format & DIFF_FORMAT_PATCH)
4907 options->flags.dirty_submodules = 1;
4908
4909 if (options->detect_rename && options->rename_limit < 0)
4910 options->rename_limit = diff_rename_limit_default;
4911 if (hexsz < options->abbrev)
4912 options->abbrev = hexsz; /* full */
4913
4914 /*
4915 * It does not make sense to show the first hit we happened
4916 * to have found. It does not make sense not to return with
4917 * exit code in such a case either.
4918 */
4919 if (options->flags.quick) {
4920 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4921 options->flags.exit_with_status = 1;
4922 }
4923
4924 options->diff_path_counter = 0;
4925
4926 if (options->flags.follow_renames)
4927 diff_check_follow_pathspec(&options->pathspec, 1);
4928
4929 if (!options->use_color || external_diff())
4930 options->color_moved = 0;
4931
4932 if (options->filter_not) {
4933 if (!options->filter)
4934 options->filter = ~filter_bit[DIFF_STATUS_FILTER_AON];
4935 options->filter &= ~options->filter_not;
4936 }
4937 }
4938
4939 int parse_long_opt(const char *opt, const char **argv,
4940 const char **optarg)
4941 {
4942 const char *arg = argv[0];
4943 if (!skip_prefix(arg, "--", &arg))
4944 return 0;
4945 if (!skip_prefix(arg, opt, &arg))
4946 return 0;
4947 if (*arg == '=') { /* stuck form: --option=value */
4948 *optarg = arg + 1;
4949 return 1;
4950 }
4951 if (*arg != '\0')
4952 return 0;
4953 /* separate form: --option value */
4954 if (!argv[1])
4955 die("Option '--%s' requires a value", opt);
4956 *optarg = argv[1];
4957 return 2;
4958 }
4959
4960 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4961 {
4962 struct diff_options *options = opt->value;
4963 int width = options->stat_width;
4964 int name_width = options->stat_name_width;
4965 int graph_width = options->stat_graph_width;
4966 int count = options->stat_count;
4967 char *end;
4968
4969 BUG_ON_OPT_NEG(unset);
4970
4971 if (!strcmp(opt->long_name, "stat")) {
4972 if (value) {
4973 width = strtoul(value, &end, 10);
4974 if (*end == ',')
4975 name_width = strtoul(end+1, &end, 10);
4976 if (*end == ',')
4977 count = strtoul(end+1, &end, 10);
4978 if (*end)
4979 return error(_("invalid --stat value: %s"), value);
4980 }
4981 } else if (!strcmp(opt->long_name, "stat-width")) {
4982 width = strtoul(value, &end, 10);
4983 if (*end)
4984 return error(_("%s expects a numerical value"),
4985 opt->long_name);
4986 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4987 name_width = strtoul(value, &end, 10);
4988 if (*end)
4989 return error(_("%s expects a numerical value"),
4990 opt->long_name);
4991 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4992 graph_width = strtoul(value, &end, 10);
4993 if (*end)
4994 return error(_("%s expects a numerical value"),
4995 opt->long_name);
4996 } else if (!strcmp(opt->long_name, "stat-count")) {
4997 count = strtoul(value, &end, 10);
4998 if (*end)
4999 return error(_("%s expects a numerical value"),
5000 opt->long_name);
5001 } else
5002 BUG("%s should not get here", opt->long_name);
5003
5004 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5005 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5006 options->stat_name_width = name_width;
5007 options->stat_graph_width = graph_width;
5008 options->stat_width = width;
5009 options->stat_count = count;
5010 return 0;
5011 }
5012
5013 static int parse_dirstat_opt(struct diff_options *options, const char *params)
5014 {
5015 struct strbuf errmsg = STRBUF_INIT;
5016 if (parse_dirstat_params(options, params, &errmsg))
5017 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
5018 errmsg.buf);
5019 strbuf_release(&errmsg);
5020 /*
5021 * The caller knows a dirstat-related option is given from the command
5022 * line; allow it to say "return this_function();"
5023 */
5024 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5025 options->output_format |= DIFF_FORMAT_DIRSTAT;
5026 return 1;
5027 }
5028
5029 static int diff_opt_diff_filter(const struct option *option,
5030 const char *optarg, int unset)
5031 {
5032 struct diff_options *opt = option->value;
5033 int i, optch;
5034
5035 BUG_ON_OPT_NEG(unset);
5036 prepare_filter_bits();
5037
5038 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
5039 unsigned int bit;
5040 int negate;
5041
5042 if ('a' <= optch && optch <= 'z') {
5043 negate = 1;
5044 optch = toupper(optch);
5045 } else {
5046 negate = 0;
5047 }
5048
5049 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
5050 if (!bit)
5051 return error(_("unknown change class '%c' in --diff-filter=%s"),
5052 optarg[i], optarg);
5053 if (negate)
5054 opt->filter_not |= bit;
5055 else
5056 opt->filter |= bit;
5057 }
5058 return 0;
5059 }
5060
5061 static void enable_patch_output(int *fmt)
5062 {
5063 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
5064 *fmt |= DIFF_FORMAT_PATCH;
5065 }
5066
5067 static int diff_opt_ws_error_highlight(const struct option *option,
5068 const char *arg, int unset)
5069 {
5070 struct diff_options *opt = option->value;
5071 int val = parse_ws_error_highlight(arg);
5072
5073 BUG_ON_OPT_NEG(unset);
5074 if (val < 0)
5075 return error(_("unknown value after ws-error-highlight=%.*s"),
5076 -1 - val, arg);
5077 opt->ws_error_highlight = val;
5078 return 0;
5079 }
5080
5081 static int diff_opt_find_object(const struct option *option,
5082 const char *arg, int unset)
5083 {
5084 struct diff_options *opt = option->value;
5085 struct object_id oid;
5086
5087 BUG_ON_OPT_NEG(unset);
5088 if (repo_get_oid(the_repository, arg, &oid))
5089 return error(_("unable to resolve '%s'"), arg);
5090
5091 if (!opt->objfind)
5092 CALLOC_ARRAY(opt->objfind, 1);
5093
5094 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
5095 opt->flags.recursive = 1;
5096 opt->flags.tree_in_recursive = 1;
5097 oidset_insert(opt->objfind, &oid);
5098 return 0;
5099 }
5100
5101 static int diff_opt_anchored(const struct option *opt,
5102 const char *arg, int unset)
5103 {
5104 struct diff_options *options = opt->value;
5105
5106 BUG_ON_OPT_NEG(unset);
5107 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5108 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5109 options->anchors_alloc);
5110 options->anchors[options->anchors_nr++] = xstrdup(arg);
5111 return 0;
5112 }
5113
5114 static int diff_opt_binary(const struct option *opt,
5115 const char *arg, int unset)
5116 {
5117 struct diff_options *options = opt->value;
5118
5119 BUG_ON_OPT_NEG(unset);
5120 BUG_ON_OPT_ARG(arg);
5121 enable_patch_output(&options->output_format);
5122 options->flags.binary = 1;
5123 return 0;
5124 }
5125
5126 static int diff_opt_break_rewrites(const struct option *opt,
5127 const char *arg, int unset)
5128 {
5129 int *break_opt = opt->value;
5130 int opt1, opt2;
5131
5132 BUG_ON_OPT_NEG(unset);
5133 if (!arg)
5134 arg = "";
5135 opt1 = parse_rename_score(&arg);
5136 if (*arg == 0)
5137 opt2 = 0;
5138 else if (*arg != '/')
5139 return error(_("%s expects <n>/<m> form"), opt->long_name);
5140 else {
5141 arg++;
5142 opt2 = parse_rename_score(&arg);
5143 }
5144 if (*arg != 0)
5145 return error(_("%s expects <n>/<m> form"), opt->long_name);
5146 *break_opt = opt1 | (opt2 << 16);
5147 return 0;
5148 }
5149
5150 static int diff_opt_char(const struct option *opt,
5151 const char *arg, int unset)
5152 {
5153 char *value = opt->value;
5154
5155 BUG_ON_OPT_NEG(unset);
5156 if (arg[1])
5157 return error(_("%s expects a character, got '%s'"),
5158 opt->long_name, arg);
5159 *value = arg[0];
5160 return 0;
5161 }
5162
5163 static int diff_opt_color_moved(const struct option *opt,
5164 const char *arg, int unset)
5165 {
5166 struct diff_options *options = opt->value;
5167
5168 if (unset) {
5169 options->color_moved = COLOR_MOVED_NO;
5170 } else if (!arg) {
5171 if (diff_color_moved_default)
5172 options->color_moved = diff_color_moved_default;
5173 if (options->color_moved == COLOR_MOVED_NO)
5174 options->color_moved = COLOR_MOVED_DEFAULT;
5175 } else {
5176 int cm = parse_color_moved(arg);
5177 if (cm < 0)
5178 return error(_("bad --color-moved argument: %s"), arg);
5179 options->color_moved = cm;
5180 }
5181 return 0;
5182 }
5183
5184 static int diff_opt_color_moved_ws(const struct option *opt,
5185 const char *arg, int unset)
5186 {
5187 struct diff_options *options = opt->value;
5188 unsigned cm;
5189
5190 if (unset) {
5191 options->color_moved_ws_handling = 0;
5192 return 0;
5193 }
5194
5195 cm = parse_color_moved_ws(arg);
5196 if (cm & COLOR_MOVED_WS_ERROR)
5197 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
5198 options->color_moved_ws_handling = cm;
5199 return 0;
5200 }
5201
5202 static int diff_opt_color_words(const struct option *opt,
5203 const char *arg, int unset)
5204 {
5205 struct diff_options *options = opt->value;
5206
5207 BUG_ON_OPT_NEG(unset);
5208 options->use_color = 1;
5209 options->word_diff = DIFF_WORDS_COLOR;
5210 options->word_regex = arg;
5211 return 0;
5212 }
5213
5214 static int diff_opt_compact_summary(const struct option *opt,
5215 const char *arg, int unset)
5216 {
5217 struct diff_options *options = opt->value;
5218
5219 BUG_ON_OPT_ARG(arg);
5220 if (unset) {
5221 options->flags.stat_with_summary = 0;
5222 } else {
5223 options->flags.stat_with_summary = 1;
5224 options->output_format &= ~DIFF_FORMAT_NO_OUTPUT;
5225 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5226 }
5227 return 0;
5228 }
5229
5230 static int diff_opt_diff_algorithm(const struct option *opt,
5231 const char *arg, int unset)
5232 {
5233 struct diff_options *options = opt->value;
5234
5235 BUG_ON_OPT_NEG(unset);
5236
5237 if (set_diff_algorithm(options, arg))
5238 return error(_("option diff-algorithm accepts \"myers\", "
5239 "\"minimal\", \"patience\" and \"histogram\""));
5240
5241 options->ignore_driver_algorithm = 1;
5242
5243 return 0;
5244 }
5245
5246 static int diff_opt_diff_algorithm_no_arg(const struct option *opt,
5247 const char *arg, int unset)
5248 {
5249 struct diff_options *options = opt->value;
5250
5251 BUG_ON_OPT_NEG(unset);
5252 BUG_ON_OPT_ARG(arg);
5253
5254 if (set_diff_algorithm(options, opt->long_name))
5255 BUG("available diff algorithms include \"myers\", "
5256 "\"minimal\", \"patience\" and \"histogram\"");
5257
5258 options->ignore_driver_algorithm = 1;
5259
5260 return 0;
5261 }
5262
5263 static int diff_opt_dirstat(const struct option *opt,
5264 const char *arg, int unset)
5265 {
5266 struct diff_options *options = opt->value;
5267
5268 BUG_ON_OPT_NEG(unset);
5269 if (!strcmp(opt->long_name, "cumulative")) {
5270 if (arg)
5271 BUG("how come --cumulative take a value?");
5272 arg = "cumulative";
5273 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5274 parse_dirstat_opt(options, "files");
5275 parse_dirstat_opt(options, arg ? arg : "");
5276 return 0;
5277 }
5278
5279 static int diff_opt_find_copies(const struct option *opt,
5280 const char *arg, int unset)
5281 {
5282 struct diff_options *options = opt->value;
5283
5284 BUG_ON_OPT_NEG(unset);
5285 if (!arg)
5286 arg = "";
5287 options->rename_score = parse_rename_score(&arg);
5288 if (*arg != 0)
5289 return error(_("invalid argument to %s"), opt->long_name);
5290
5291 if (options->detect_rename == DIFF_DETECT_COPY)
5292 options->flags.find_copies_harder = 1;
5293 else
5294 options->detect_rename = DIFF_DETECT_COPY;
5295
5296 return 0;
5297 }
5298
5299 static int diff_opt_find_renames(const struct option *opt,
5300 const char *arg, int unset)
5301 {
5302 struct diff_options *options = opt->value;
5303
5304 BUG_ON_OPT_NEG(unset);
5305 if (!arg)
5306 arg = "";
5307 options->rename_score = parse_rename_score(&arg);
5308 if (*arg != 0)
5309 return error(_("invalid argument to %s"), opt->long_name);
5310
5311 options->detect_rename = DIFF_DETECT_RENAME;
5312 return 0;
5313 }
5314
5315 static int diff_opt_follow(const struct option *opt,
5316 const char *arg, int unset)
5317 {
5318 struct diff_options *options = opt->value;
5319
5320 BUG_ON_OPT_ARG(arg);
5321 if (unset) {
5322 options->flags.follow_renames = 0;
5323 options->flags.default_follow_renames = 0;
5324 } else {
5325 options->flags.follow_renames = 1;
5326 }
5327 return 0;
5328 }
5329
5330 static int diff_opt_ignore_submodules(const struct option *opt,
5331 const char *arg, int unset)
5332 {
5333 struct diff_options *options = opt->value;
5334
5335 BUG_ON_OPT_NEG(unset);
5336 if (!arg)
5337 arg = "all";
5338 options->flags.override_submodule_config = 1;
5339 handle_ignore_submodules_arg(options, arg);
5340 return 0;
5341 }
5342
5343 static int diff_opt_line_prefix(const struct option *opt,
5344 const char *optarg, int unset)
5345 {
5346 struct diff_options *options = opt->value;
5347
5348 BUG_ON_OPT_NEG(unset);
5349 options->line_prefix = optarg;
5350 options->line_prefix_length = strlen(options->line_prefix);
5351 graph_setup_line_prefix(options);
5352 return 0;
5353 }
5354
5355 static int diff_opt_no_prefix(const struct option *opt,
5356 const char *optarg, int unset)
5357 {
5358 struct diff_options *options = opt->value;
5359
5360 BUG_ON_OPT_NEG(unset);
5361 BUG_ON_OPT_ARG(optarg);
5362 diff_set_noprefix(options);
5363 return 0;
5364 }
5365
5366 static int diff_opt_default_prefix(const struct option *opt,
5367 const char *optarg, int unset)
5368 {
5369 struct diff_options *options = opt->value;
5370
5371 BUG_ON_OPT_NEG(unset);
5372 BUG_ON_OPT_ARG(optarg);
5373 diff_src_prefix = "a/";
5374 diff_dst_prefix = "b/";
5375 diff_set_default_prefix(options);
5376 return 0;
5377 }
5378
5379 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5380 const struct option *opt,
5381 const char *arg, int unset)
5382 {
5383 struct diff_options *options = opt->value;
5384 char *path;
5385
5386 BUG_ON_OPT_NEG(unset);
5387 path = prefix_filename(ctx->prefix, arg);
5388 options->file = xfopen(path, "w");
5389 options->close_file = 1;
5390 if (options->use_color != GIT_COLOR_ALWAYS)
5391 options->use_color = GIT_COLOR_NEVER;
5392 free(path);
5393 return 0;
5394 }
5395
5396 static int diff_opt_patience(const struct option *opt,
5397 const char *arg, int unset)
5398 {
5399 struct diff_options *options = opt->value;
5400 int i;
5401
5402 BUG_ON_OPT_NEG(unset);
5403 BUG_ON_OPT_ARG(arg);
5404 /*
5405 * Both --patience and --anchored use PATIENCE_DIFF
5406 * internally, so remove any anchors previously
5407 * specified.
5408 */
5409 for (i = 0; i < options->anchors_nr; i++)
5410 free(options->anchors[i]);
5411 options->anchors_nr = 0;
5412 options->ignore_driver_algorithm = 1;
5413
5414 return set_diff_algorithm(options, "patience");
5415 }
5416
5417 static int diff_opt_ignore_regex(const struct option *opt,
5418 const char *arg, int unset)
5419 {
5420 struct diff_options *options = opt->value;
5421 regex_t *regex;
5422
5423 BUG_ON_OPT_NEG(unset);
5424 regex = xmalloc(sizeof(*regex));
5425 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5426 return error(_("invalid regex given to -I: '%s'"), arg);
5427 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5428 options->ignore_regex_alloc);
5429 options->ignore_regex[options->ignore_regex_nr++] = regex;
5430 return 0;
5431 }
5432
5433 static int diff_opt_pickaxe_regex(const struct option *opt,
5434 const char *arg, int unset)
5435 {
5436 struct diff_options *options = opt->value;
5437
5438 BUG_ON_OPT_NEG(unset);
5439 options->pickaxe = arg;
5440 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5441 return 0;
5442 }
5443
5444 static int diff_opt_pickaxe_string(const struct option *opt,
5445 const char *arg, int unset)
5446 {
5447 struct diff_options *options = opt->value;
5448
5449 BUG_ON_OPT_NEG(unset);
5450 options->pickaxe = arg;
5451 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5452 return 0;
5453 }
5454
5455 static int diff_opt_relative(const struct option *opt,
5456 const char *arg, int unset)
5457 {
5458 struct diff_options *options = opt->value;
5459
5460 options->flags.relative_name = !unset;
5461 if (arg)
5462 options->prefix = arg;
5463 return 0;
5464 }
5465
5466 static int diff_opt_submodule(const struct option *opt,
5467 const char *arg, int unset)
5468 {
5469 struct diff_options *options = opt->value;
5470
5471 BUG_ON_OPT_NEG(unset);
5472 if (!arg)
5473 arg = "log";
5474 if (parse_submodule_params(options, arg))
5475 return error(_("failed to parse --submodule option parameter: '%s'"),
5476 arg);
5477 return 0;
5478 }
5479
5480 static int diff_opt_textconv(const struct option *opt,
5481 const char *arg, int unset)
5482 {
5483 struct diff_options *options = opt->value;
5484
5485 BUG_ON_OPT_ARG(arg);
5486 if (unset) {
5487 options->flags.allow_textconv = 0;
5488 } else {
5489 options->flags.allow_textconv = 1;
5490 options->flags.textconv_set_via_cmdline = 1;
5491 }
5492 return 0;
5493 }
5494
5495 static int diff_opt_unified(const struct option *opt,
5496 const char *arg, int unset)
5497 {
5498 struct diff_options *options = opt->value;
5499 char *s;
5500
5501 BUG_ON_OPT_NEG(unset);
5502
5503 if (arg) {
5504 options->context = strtol(arg, &s, 10);
5505 if (*s)
5506 return error(_("%s expects a numerical value"), "--unified");
5507 }
5508 enable_patch_output(&options->output_format);
5509
5510 return 0;
5511 }
5512
5513 static int diff_opt_word_diff(const struct option *opt,
5514 const char *arg, int unset)
5515 {
5516 struct diff_options *options = opt->value;
5517
5518 BUG_ON_OPT_NEG(unset);
5519 if (arg) {
5520 if (!strcmp(arg, "plain"))
5521 options->word_diff = DIFF_WORDS_PLAIN;
5522 else if (!strcmp(arg, "color")) {
5523 options->use_color = 1;
5524 options->word_diff = DIFF_WORDS_COLOR;
5525 }
5526 else if (!strcmp(arg, "porcelain"))
5527 options->word_diff = DIFF_WORDS_PORCELAIN;
5528 else if (!strcmp(arg, "none"))
5529 options->word_diff = DIFF_WORDS_NONE;
5530 else
5531 return error(_("bad --word-diff argument: %s"), arg);
5532 } else {
5533 if (options->word_diff == DIFF_WORDS_NONE)
5534 options->word_diff = DIFF_WORDS_PLAIN;
5535 }
5536 return 0;
5537 }
5538
5539 static int diff_opt_word_diff_regex(const struct option *opt,
5540 const char *arg, int unset)
5541 {
5542 struct diff_options *options = opt->value;
5543
5544 BUG_ON_OPT_NEG(unset);
5545 if (options->word_diff == DIFF_WORDS_NONE)
5546 options->word_diff = DIFF_WORDS_PLAIN;
5547 options->word_regex = arg;
5548 return 0;
5549 }
5550
5551 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5552 {
5553 struct diff_options *options = opt->value;
5554
5555 BUG_ON_OPT_NEG(unset);
5556 if (!strcmp(opt->long_name, "skip-to"))
5557 options->skip_instead_of_rotate = 1;
5558 else
5559 options->skip_instead_of_rotate = 0;
5560 options->rotate_to = arg;
5561 return 0;
5562 }
5563
5564 /*
5565 * Consider adding new flags to __git_diff_common_options
5566 * in contrib/completion/git-completion.bash
5567 */
5568 struct option *add_diff_options(const struct option *opts,
5569 struct diff_options *options)
5570 {
5571 struct option parseopts[] = {
5572 OPT_GROUP(N_("Diff output format options")),
5573 OPT_BITOP('p', "patch", &options->output_format,
5574 N_("generate patch"),
5575 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5576 OPT_SET_INT('s', "no-patch", &options->output_format,
5577 N_("suppress diff output"), DIFF_FORMAT_NO_OUTPUT),
5578 OPT_BITOP('u', NULL, &options->output_format,
5579 N_("generate patch"),
5580 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5581 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5582 N_("generate diffs with <n> lines context"),
5583 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5584 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5585 N_("generate diffs with <n> lines context")),
5586 OPT_BITOP(0, "raw", &options->output_format,
5587 N_("generate the diff in raw format"),
5588 DIFF_FORMAT_RAW, DIFF_FORMAT_NO_OUTPUT),
5589 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5590 N_("synonym for '-p --raw'"),
5591 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5592 DIFF_FORMAT_NO_OUTPUT),
5593 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5594 N_("synonym for '-p --stat'"),
5595 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5596 DIFF_FORMAT_NO_OUTPUT),
5597 OPT_BITOP(0, "numstat", &options->output_format,
5598 N_("machine friendly --stat"),
5599 DIFF_FORMAT_NUMSTAT, DIFF_FORMAT_NO_OUTPUT),
5600 OPT_BITOP(0, "shortstat", &options->output_format,
5601 N_("output only the last line of --stat"),
5602 DIFF_FORMAT_SHORTSTAT, DIFF_FORMAT_NO_OUTPUT),
5603 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1>,<param2>..."),
5604 N_("output the distribution of relative amount of changes for each sub-directory"),
5605 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5606 diff_opt_dirstat),
5607 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5608 N_("synonym for --dirstat=cumulative"),
5609 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5610 diff_opt_dirstat),
5611 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1>,<param2>..."),
5612 N_("synonym for --dirstat=files,<param1>,<param2>..."),
5613 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5614 diff_opt_dirstat),
5615 OPT_BIT_F(0, "check", &options->output_format,
5616 N_("warn if changes introduce conflict markers or whitespace errors"),
5617 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5618 OPT_BITOP(0, "summary", &options->output_format,
5619 N_("condensed summary such as creations, renames and mode changes"),
5620 DIFF_FORMAT_SUMMARY, DIFF_FORMAT_NO_OUTPUT),
5621 OPT_BIT_F(0, "name-only", &options->output_format,
5622 N_("show only names of changed files"),
5623 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5624 OPT_BIT_F(0, "name-status", &options->output_format,
5625 N_("show only names and status of changed files"),
5626 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5627 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5628 N_("generate diffstat"),
5629 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5630 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5631 N_("generate diffstat with a given width"),
5632 PARSE_OPT_NONEG, diff_opt_stat),
5633 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5634 N_("generate diffstat with a given name width"),
5635 PARSE_OPT_NONEG, diff_opt_stat),
5636 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5637 N_("generate diffstat with a given graph width"),
5638 PARSE_OPT_NONEG, diff_opt_stat),
5639 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5640 N_("generate diffstat with limited lines"),
5641 PARSE_OPT_NONEG, diff_opt_stat),
5642 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5643 N_("generate compact summary in diffstat"),
5644 PARSE_OPT_NOARG, diff_opt_compact_summary),
5645 OPT_CALLBACK_F(0, "binary", options, NULL,
5646 N_("output a binary diff that can be applied"),
5647 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5648 OPT_BOOL(0, "full-index", &options->flags.full_index,
5649 N_("show full pre- and post-image object names on the \"index\" lines")),
5650 OPT_COLOR_FLAG(0, "color", &options->use_color,
5651 N_("show colored diff")),
5652 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5653 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5654 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5655 OPT_SET_INT('z', NULL, &options->line_termination,
5656 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5657 0),
5658 OPT__ABBREV(&options->abbrev),
5659 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5660 N_("show the given source prefix instead of \"a/\""),
5661 PARSE_OPT_NONEG),
5662 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5663 N_("show the given destination prefix instead of \"b/\""),
5664 PARSE_OPT_NONEG),
5665 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5666 N_("prepend an additional prefix to every line of output"),
5667 PARSE_OPT_NONEG, diff_opt_line_prefix),
5668 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5669 N_("do not show any source or destination prefix"),
5670 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5671 OPT_CALLBACK_F(0, "default-prefix", options, NULL,
5672 N_("use default prefixes a/ and b/"),
5673 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_default_prefix),
5674 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5675 N_("show context between diff hunks up to the specified number of lines"),
5676 PARSE_OPT_NONEG),
5677 OPT_CALLBACK_F(0, "output-indicator-new",
5678 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5679 N_("<char>"),
5680 N_("specify the character to indicate a new line instead of '+'"),
5681 PARSE_OPT_NONEG, diff_opt_char),
5682 OPT_CALLBACK_F(0, "output-indicator-old",
5683 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5684 N_("<char>"),
5685 N_("specify the character to indicate an old line instead of '-'"),
5686 PARSE_OPT_NONEG, diff_opt_char),
5687 OPT_CALLBACK_F(0, "output-indicator-context",
5688 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5689 N_("<char>"),
5690 N_("specify the character to indicate a context instead of ' '"),
5691 PARSE_OPT_NONEG, diff_opt_char),
5692
5693 OPT_GROUP(N_("Diff rename options")),
5694 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5695 N_("break complete rewrite changes into pairs of delete and create"),
5696 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5697 diff_opt_break_rewrites),
5698 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5699 N_("detect renames"),
5700 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5701 diff_opt_find_renames),
5702 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5703 N_("omit the preimage for deletes"),
5704 1, PARSE_OPT_NONEG),
5705 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5706 N_("detect copies"),
5707 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5708 diff_opt_find_copies),
5709 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5710 N_("use unmodified files as source to find copies")),
5711 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5712 N_("disable rename detection"),
5713 0, PARSE_OPT_NONEG),
5714 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5715 N_("use empty blobs as rename source")),
5716 OPT_CALLBACK_F(0, "follow", options, NULL,
5717 N_("continue listing the history of a file beyond renames"),
5718 PARSE_OPT_NOARG, diff_opt_follow),
5719 OPT_INTEGER('l', NULL, &options->rename_limit,
5720 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5721
5722 OPT_GROUP(N_("Diff algorithm options")),
5723 OPT_CALLBACK_F(0, "minimal", options, NULL,
5724 N_("produce the smallest possible diff"),
5725 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5726 diff_opt_diff_algorithm_no_arg),
5727 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5728 N_("ignore whitespace when comparing lines"),
5729 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5730 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5731 N_("ignore changes in amount of whitespace"),
5732 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5733 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5734 N_("ignore changes in whitespace at EOL"),
5735 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5736 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5737 N_("ignore carrier-return at the end of line"),
5738 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5739 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5740 N_("ignore changes whose lines are all blank"),
5741 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5742 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5743 N_("ignore changes whose all lines match <regex>"),
5744 0, diff_opt_ignore_regex),
5745 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5746 N_("heuristic to shift diff hunk boundaries for easy reading"),
5747 XDF_INDENT_HEURISTIC),
5748 OPT_CALLBACK_F(0, "patience", options, NULL,
5749 N_("generate diff using the \"patience diff\" algorithm"),
5750 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5751 diff_opt_patience),
5752 OPT_CALLBACK_F(0, "histogram", options, NULL,
5753 N_("generate diff using the \"histogram diff\" algorithm"),
5754 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5755 diff_opt_diff_algorithm_no_arg),
5756 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5757 N_("choose a diff algorithm"),
5758 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5759 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5760 N_("generate diff using the \"anchored diff\" algorithm"),
5761 PARSE_OPT_NONEG, diff_opt_anchored),
5762 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5763 N_("show word diff, using <mode> to delimit changed words"),
5764 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5765 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5766 N_("use <regex> to decide what a word is"),
5767 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5768 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5769 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5770 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5771 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5772 N_("moved lines of code are colored differently"),
5773 PARSE_OPT_OPTARG, diff_opt_color_moved),
5774 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5775 N_("how white spaces are ignored in --color-moved"),
5776 0, diff_opt_color_moved_ws),
5777
5778 OPT_GROUP(N_("Other diff options")),
5779 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5780 N_("when run from subdir, exclude changes outside and show relative paths"),
5781 PARSE_OPT_OPTARG,
5782 diff_opt_relative),
5783 OPT_BOOL('a', "text", &options->flags.text,
5784 N_("treat all files as text")),
5785 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5786 N_("swap two inputs, reverse the diff")),
5787 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5788 N_("exit with 1 if there were differences, 0 otherwise")),
5789 OPT_BOOL(0, "quiet", &options->flags.quick,
5790 N_("disable all output of the program")),
5791 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5792 N_("allow an external diff helper to be executed")),
5793 OPT_CALLBACK_F(0, "textconv", options, NULL,
5794 N_("run external text conversion filters when comparing binary files"),
5795 PARSE_OPT_NOARG, diff_opt_textconv),
5796 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5797 N_("ignore changes to submodules in the diff generation"),
5798 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5799 diff_opt_ignore_submodules),
5800 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5801 N_("specify how differences in submodules are shown"),
5802 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5803 diff_opt_submodule),
5804 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5805 N_("hide 'git add -N' entries from the index"),
5806 1, PARSE_OPT_NONEG),
5807 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5808 N_("treat 'git add -N' entries as real in the index"),
5809 0, PARSE_OPT_NONEG),
5810 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5811 N_("look for differences that change the number of occurrences of the specified string"),
5812 0, diff_opt_pickaxe_string),
5813 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5814 N_("look for differences that change the number of occurrences of the specified regex"),
5815 0, diff_opt_pickaxe_regex),
5816 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5817 N_("show all changes in the changeset with -S or -G"),
5818 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5819 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5820 N_("treat <string> in -S as extended POSIX regular expression"),
5821 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5822 OPT_FILENAME('O', NULL, &options->orderfile,
5823 N_("control the order in which files appear in the output")),
5824 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5825 N_("show the change in the specified path first"),
5826 PARSE_OPT_NONEG, diff_opt_rotate_to),
5827 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5828 N_("skip the output to the specified path"),
5829 PARSE_OPT_NONEG, diff_opt_rotate_to),
5830 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5831 N_("look for differences that change the number of occurrences of the specified object"),
5832 PARSE_OPT_NONEG, diff_opt_find_object),
5833 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5834 N_("select files by diff type"),
5835 PARSE_OPT_NONEG, diff_opt_diff_filter),
5836 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5837 N_("output to a specific file"),
5838 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5839
5840 OPT_END()
5841 };
5842
5843 return parse_options_concat(opts, parseopts);
5844 }
5845
5846 int diff_opt_parse(struct diff_options *options,
5847 const char **av, int ac, const char *prefix)
5848 {
5849 struct option no_options[] = { OPT_END() };
5850 struct option *parseopts = add_diff_options(no_options, options);
5851
5852 if (!prefix)
5853 prefix = "";
5854
5855 ac = parse_options(ac, av, prefix, parseopts, NULL,
5856 PARSE_OPT_KEEP_DASHDASH |
5857 PARSE_OPT_KEEP_UNKNOWN_OPT |
5858 PARSE_OPT_NO_INTERNAL_HELP |
5859 PARSE_OPT_ONE_SHOT |
5860 PARSE_OPT_STOP_AT_NON_OPTION);
5861 free(parseopts);
5862
5863 return ac;
5864 }
5865
5866 int parse_rename_score(const char **cp_p)
5867 {
5868 unsigned long num, scale;
5869 int ch, dot;
5870 const char *cp = *cp_p;
5871
5872 num = 0;
5873 scale = 1;
5874 dot = 0;
5875 for (;;) {
5876 ch = *cp;
5877 if ( !dot && ch == '.' ) {
5878 scale = 1;
5879 dot = 1;
5880 } else if ( ch == '%' ) {
5881 scale = dot ? scale*100 : 100;
5882 cp++; /* % is always at the end */
5883 break;
5884 } else if ( ch >= '0' && ch <= '9' ) {
5885 if ( scale < 100000 ) {
5886 scale *= 10;
5887 num = (num*10) + (ch-'0');
5888 }
5889 } else {
5890 break;
5891 }
5892 cp++;
5893 }
5894 *cp_p = cp;
5895
5896 /* user says num divided by scale and we say internally that
5897 * is MAX_SCORE * num / scale.
5898 */
5899 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5900 }
5901
5902 struct diff_queue_struct diff_queued_diff;
5903
5904 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5905 {
5906 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5907 queue->queue[queue->nr++] = dp;
5908 }
5909
5910 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5911 struct diff_filespec *one,
5912 struct diff_filespec *two)
5913 {
5914 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5915 dp->one = one;
5916 dp->two = two;
5917 if (queue)
5918 diff_q(queue, dp);
5919 return dp;
5920 }
5921
5922 void diff_free_filepair(struct diff_filepair *p)
5923 {
5924 free_filespec(p->one);
5925 free_filespec(p->two);
5926 free(p);
5927 }
5928
5929 void diff_free_queue(struct diff_queue_struct *q)
5930 {
5931 for (int i = 0; i < q->nr; i++)
5932 diff_free_filepair(q->queue[i]);
5933 free(q->queue);
5934 }
5935
5936 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5937 {
5938 int abblen;
5939 const char *abbrev;
5940
5941 /* Do we want all 40 hex characters? */
5942 if (len == the_hash_algo->hexsz)
5943 return oid_to_hex(oid);
5944
5945 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5946 abbrev = diff_abbrev_oid(oid, len);
5947
5948 if (!print_sha1_ellipsis())
5949 return abbrev;
5950
5951 abblen = strlen(abbrev);
5952
5953 /*
5954 * In well-behaved cases, where the abbreviated result is the
5955 * same as the requested length, append three dots after the
5956 * abbreviation (hence the whole logic is limited to the case
5957 * where abblen < 37); when the actual abbreviated result is a
5958 * bit longer than the requested length, we reduce the number
5959 * of dots so that they match the well-behaved ones. However,
5960 * if the actual abbreviation is longer than the requested
5961 * length by more than three, we give up on aligning, and add
5962 * three dots anyway, to indicate that the output is not the
5963 * full object name. Yes, this may be suboptimal, but this
5964 * appears only in "diff --raw --abbrev" output and it is not
5965 * worth the effort to change it now. Note that this would
5966 * likely to work fine when the automatic sizing of default
5967 * abbreviation length is used--we would be fed -1 in "len" in
5968 * that case, and will end up always appending three-dots, but
5969 * the automatic sizing is supposed to give abblen that ensures
5970 * uniqueness across all objects (statistically speaking).
5971 */
5972 if (abblen < the_hash_algo->hexsz - 3) {
5973 static char hex[GIT_MAX_HEXSZ + 1];
5974 if (len < abblen && abblen <= len + 2)
5975 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5976 else
5977 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5978 return hex;
5979 }
5980
5981 return oid_to_hex(oid);
5982 }
5983
5984 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5985 {
5986 int line_termination = opt->line_termination;
5987 int inter_name_termination = line_termination ? '\t' : '\0';
5988
5989 fprintf(opt->file, "%s", diff_line_prefix(opt));
5990 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5991 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5992 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5993 fprintf(opt->file, "%s ",
5994 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5995 }
5996 if (p->score) {
5997 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5998 inter_name_termination);
5999 } else {
6000 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
6001 }
6002
6003 if (p->status == DIFF_STATUS_COPIED ||
6004 p->status == DIFF_STATUS_RENAMED) {
6005 const char *name_a, *name_b;
6006 name_a = p->one->path;
6007 name_b = p->two->path;
6008 strip_prefix(opt->prefix_length, &name_a, &name_b);
6009 write_name_quoted(name_a, opt->file, inter_name_termination);
6010 write_name_quoted(name_b, opt->file, line_termination);
6011 } else {
6012 const char *name_a, *name_b;
6013 name_a = p->one->mode ? p->one->path : p->two->path;
6014 name_b = NULL;
6015 strip_prefix(opt->prefix_length, &name_a, &name_b);
6016 write_name_quoted(name_a, opt->file, line_termination);
6017 }
6018 }
6019
6020 int diff_unmodified_pair(struct diff_filepair *p)
6021 {
6022 /* This function is written stricter than necessary to support
6023 * the currently implemented transformers, but the idea is to
6024 * let transformers to produce diff_filepairs any way they want,
6025 * and filter and clean them up here before producing the output.
6026 */
6027 struct diff_filespec *one = p->one, *two = p->two;
6028
6029 if (DIFF_PAIR_UNMERGED(p))
6030 return 0; /* unmerged is interesting */
6031
6032 /* deletion, addition, mode or type change
6033 * and rename are all interesting.
6034 */
6035 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
6036 DIFF_PAIR_MODE_CHANGED(p) ||
6037 strcmp(one->path, two->path))
6038 return 0;
6039
6040 /* both are valid and point at the same path. that is, we are
6041 * dealing with a change.
6042 */
6043 if (one->oid_valid && two->oid_valid &&
6044 oideq(&one->oid, &two->oid) &&
6045 !one->dirty_submodule && !two->dirty_submodule)
6046 return 1; /* no change */
6047 if (!one->oid_valid && !two->oid_valid)
6048 return 1; /* both look at the same file on the filesystem. */
6049 return 0;
6050 }
6051
6052 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
6053 {
6054 int include_conflict_headers =
6055 (additional_headers(o, p->one->path) &&
6056 !o->pickaxe_opts &&
6057 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6058
6059 /*
6060 * Check if we can return early without showing a diff. Note that
6061 * diff_filepair only stores {oid, path, mode, is_valid}
6062 * information for each path, and thus diff_unmodified_pair() only
6063 * considers those bits of info. However, we do not want pairs
6064 * created by create_filepairs_for_header_only_notifications()
6065 * (which always look like unmodified pairs) to be ignored, so
6066 * return early if both p is unmodified AND we don't want to
6067 * include_conflict_headers.
6068 */
6069 if (diff_unmodified_pair(p) && !include_conflict_headers)
6070 return;
6071
6072 /* Actually, we can also return early to avoid showing tree diffs */
6073 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6074 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6075 return;
6076
6077 run_diff(p, o);
6078 }
6079
6080 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
6081 struct diffstat_t *diffstat)
6082 {
6083 if (diff_unmodified_pair(p))
6084 return;
6085
6086 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6087 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6088 return; /* no useful stat for tree diffs */
6089
6090 run_diffstat(p, o, diffstat);
6091 }
6092
6093 static void diff_flush_checkdiff(struct diff_filepair *p,
6094 struct diff_options *o)
6095 {
6096 if (diff_unmodified_pair(p))
6097 return;
6098
6099 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6100 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6101 return; /* nothing to check in tree diffs */
6102
6103 run_checkdiff(p, o);
6104 }
6105
6106 int diff_queue_is_empty(struct diff_options *o)
6107 {
6108 struct diff_queue_struct *q = &diff_queued_diff;
6109 int i;
6110 int include_conflict_headers =
6111 (o->additional_path_headers &&
6112 strmap_get_size(o->additional_path_headers) &&
6113 !o->pickaxe_opts &&
6114 (!o->filter || filter_bit_tst(DIFF_STATUS_UNMERGED, o)));
6115
6116 if (include_conflict_headers)
6117 return 0;
6118
6119 for (i = 0; i < q->nr; i++)
6120 if (!diff_unmodified_pair(q->queue[i]))
6121 return 0;
6122 return 1;
6123 }
6124
6125 #if DIFF_DEBUG
6126 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
6127 {
6128 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
6129 x, one ? one : "",
6130 s->path,
6131 DIFF_FILE_VALID(s) ? "valid" : "invalid",
6132 s->mode,
6133 s->oid_valid ? oid_to_hex(&s->oid) : "");
6134 fprintf(stderr, "queue[%d] %s size %lu\n",
6135 x, one ? one : "",
6136 s->size);
6137 }
6138
6139 void diff_debug_filepair(const struct diff_filepair *p, int i)
6140 {
6141 diff_debug_filespec(p->one, i, "one");
6142 diff_debug_filespec(p->two, i, "two");
6143 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
6144 p->score, p->status ? p->status : '?',
6145 p->one->rename_used, p->broken_pair);
6146 }
6147
6148 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
6149 {
6150 int i;
6151 if (msg)
6152 fprintf(stderr, "%s\n", msg);
6153 fprintf(stderr, "q->nr = %d\n", q->nr);
6154 for (i = 0; i < q->nr; i++) {
6155 struct diff_filepair *p = q->queue[i];
6156 diff_debug_filepair(p, i);
6157 }
6158 }
6159 #endif
6160
6161 static void diff_resolve_rename_copy(void)
6162 {
6163 int i;
6164 struct diff_filepair *p;
6165 struct diff_queue_struct *q = &diff_queued_diff;
6166
6167 diff_debug_queue("resolve-rename-copy", q);
6168
6169 for (i = 0; i < q->nr; i++) {
6170 p = q->queue[i];
6171 p->status = 0; /* undecided */
6172 if (DIFF_PAIR_UNMERGED(p))
6173 p->status = DIFF_STATUS_UNMERGED;
6174 else if (!DIFF_FILE_VALID(p->one))
6175 p->status = DIFF_STATUS_ADDED;
6176 else if (!DIFF_FILE_VALID(p->two))
6177 p->status = DIFF_STATUS_DELETED;
6178 else if (DIFF_PAIR_TYPE_CHANGED(p))
6179 p->status = DIFF_STATUS_TYPE_CHANGED;
6180
6181 /* from this point on, we are dealing with a pair
6182 * whose both sides are valid and of the same type, i.e.
6183 * either in-place edit or rename/copy edit.
6184 */
6185 else if (DIFF_PAIR_RENAME(p)) {
6186 /*
6187 * A rename might have re-connected a broken
6188 * pair up, causing the pathnames to be the
6189 * same again. If so, that's not a rename at
6190 * all, just a modification..
6191 *
6192 * Otherwise, see if this source was used for
6193 * multiple renames, in which case we decrement
6194 * the count, and call it a copy.
6195 */
6196 if (!strcmp(p->one->path, p->two->path))
6197 p->status = DIFF_STATUS_MODIFIED;
6198 else if (--p->one->rename_used > 0)
6199 p->status = DIFF_STATUS_COPIED;
6200 else
6201 p->status = DIFF_STATUS_RENAMED;
6202 }
6203 else if (!oideq(&p->one->oid, &p->two->oid) ||
6204 p->one->mode != p->two->mode ||
6205 p->one->dirty_submodule ||
6206 p->two->dirty_submodule ||
6207 is_null_oid(&p->one->oid))
6208 p->status = DIFF_STATUS_MODIFIED;
6209 else {
6210 /* This is a "no-change" entry and should not
6211 * happen anymore, but prepare for broken callers.
6212 */
6213 error("feeding unmodified %s to diffcore",
6214 p->one->path);
6215 p->status = DIFF_STATUS_UNKNOWN;
6216 }
6217 }
6218 diff_debug_queue("resolve-rename-copy done", q);
6219 }
6220
6221 static int check_pair_status(struct diff_filepair *p)
6222 {
6223 switch (p->status) {
6224 case DIFF_STATUS_UNKNOWN:
6225 return 0;
6226 case 0:
6227 die("internal error in diff-resolve-rename-copy");
6228 default:
6229 return 1;
6230 }
6231 }
6232
6233 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
6234 {
6235 int fmt = opt->output_format;
6236
6237 if (fmt & DIFF_FORMAT_CHECKDIFF)
6238 diff_flush_checkdiff(p, opt);
6239 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
6240 diff_flush_raw(p, opt);
6241 else if (fmt & DIFF_FORMAT_NAME) {
6242 const char *name_a, *name_b;
6243 name_a = p->two->path;
6244 name_b = NULL;
6245 strip_prefix(opt->prefix_length, &name_a, &name_b);
6246 fprintf(opt->file, "%s", diff_line_prefix(opt));
6247 write_name_quoted(name_a, opt->file, opt->line_termination);
6248 }
6249
6250 opt->found_changes = 1;
6251 }
6252
6253 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
6254 {
6255 struct strbuf sb = STRBUF_INIT;
6256 if (fs->mode)
6257 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
6258 else
6259 strbuf_addf(&sb, " %s ", newdelete);
6260
6261 quote_c_style(fs->path, &sb, NULL, 0);
6262 strbuf_addch(&sb, '\n');
6263 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6264 sb.buf, sb.len, 0);
6265 strbuf_release(&sb);
6266 }
6267
6268 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
6269 int show_name)
6270 {
6271 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
6272 struct strbuf sb = STRBUF_INIT;
6273 strbuf_addf(&sb, " mode change %06o => %06o",
6274 p->one->mode, p->two->mode);
6275 if (show_name) {
6276 strbuf_addch(&sb, ' ');
6277 quote_c_style(p->two->path, &sb, NULL, 0);
6278 }
6279 strbuf_addch(&sb, '\n');
6280 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6281 sb.buf, sb.len, 0);
6282 strbuf_release(&sb);
6283 }
6284 }
6285
6286 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6287 struct diff_filepair *p)
6288 {
6289 struct strbuf sb = STRBUF_INIT;
6290 struct strbuf names = STRBUF_INIT;
6291
6292 pprint_rename(&names, p->one->path, p->two->path);
6293 strbuf_addf(&sb, " %s %s (%d%%)\n",
6294 renamecopy, names.buf, similarity_index(p));
6295 strbuf_release(&names);
6296 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6297 sb.buf, sb.len, 0);
6298 show_mode_change(opt, p, 0);
6299 strbuf_release(&sb);
6300 }
6301
6302 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6303 {
6304 switch(p->status) {
6305 case DIFF_STATUS_DELETED:
6306 show_file_mode_name(opt, "delete", p->one);
6307 break;
6308 case DIFF_STATUS_ADDED:
6309 show_file_mode_name(opt, "create", p->two);
6310 break;
6311 case DIFF_STATUS_COPIED:
6312 show_rename_copy(opt, "copy", p);
6313 break;
6314 case DIFF_STATUS_RENAMED:
6315 show_rename_copy(opt, "rename", p);
6316 break;
6317 default:
6318 if (p->score) {
6319 struct strbuf sb = STRBUF_INIT;
6320 strbuf_addstr(&sb, " rewrite ");
6321 quote_c_style(p->two->path, &sb, NULL, 0);
6322 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6323 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6324 sb.buf, sb.len, 0);
6325 strbuf_release(&sb);
6326 }
6327 show_mode_change(opt, p, !p->score);
6328 break;
6329 }
6330 }
6331
6332 struct patch_id_t {
6333 git_hash_ctx *ctx;
6334 int patchlen;
6335 };
6336
6337 static int remove_space(char *line, int len)
6338 {
6339 int i;
6340 char *dst = line;
6341 unsigned char c;
6342
6343 for (i = 0; i < len; i++)
6344 if (!isspace((c = line[i])))
6345 *dst++ = c;
6346
6347 return dst - line;
6348 }
6349
6350 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6351 {
6352 unsigned char hash[GIT_MAX_RAWSZ];
6353 unsigned short carry = 0;
6354 int i;
6355
6356 the_hash_algo->final_fn(hash, ctx);
6357 the_hash_algo->init_fn(ctx);
6358 /* 20-byte sum, with carry */
6359 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6360 carry += result->hash[i] + hash[i];
6361 result->hash[i] = carry;
6362 carry >>= 8;
6363 }
6364 }
6365
6366 static int patch_id_consume(void *priv, char *line, unsigned long len)
6367 {
6368 struct patch_id_t *data = priv;
6369 int new_len;
6370
6371 if (len > 12 && starts_with(line, "\\ "))
6372 return 0;
6373 new_len = remove_space(line, len);
6374
6375 the_hash_algo->update_fn(data->ctx, line, new_len);
6376 data->patchlen += new_len;
6377 return 0;
6378 }
6379
6380 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6381 {
6382 the_hash_algo->update_fn(ctx, str, strlen(str));
6383 }
6384
6385 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6386 {
6387 /* large enough for 2^32 in octal */
6388 char buf[12];
6389 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6390 the_hash_algo->update_fn(ctx, buf, len);
6391 }
6392
6393 /* returns 0 upon success, and writes result into oid */
6394 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6395 {
6396 struct diff_queue_struct *q = &diff_queued_diff;
6397 int i;
6398 git_hash_ctx ctx;
6399 struct patch_id_t data;
6400
6401 the_hash_algo->init_fn(&ctx);
6402 memset(&data, 0, sizeof(struct patch_id_t));
6403 data.ctx = &ctx;
6404 oidclr(oid);
6405
6406 for (i = 0; i < q->nr; i++) {
6407 xpparam_t xpp;
6408 xdemitconf_t xecfg;
6409 mmfile_t mf1, mf2;
6410 struct diff_filepair *p = q->queue[i];
6411 int len1, len2;
6412
6413 memset(&xpp, 0, sizeof(xpp));
6414 memset(&xecfg, 0, sizeof(xecfg));
6415 if (p->status == 0)
6416 return error("internal diff status error");
6417 if (p->status == DIFF_STATUS_UNKNOWN)
6418 continue;
6419 if (diff_unmodified_pair(p))
6420 continue;
6421 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6422 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6423 continue;
6424 if (DIFF_PAIR_UNMERGED(p))
6425 continue;
6426
6427 diff_fill_oid_info(p->one, options->repo->index);
6428 diff_fill_oid_info(p->two, options->repo->index);
6429
6430 len1 = remove_space(p->one->path, strlen(p->one->path));
6431 len2 = remove_space(p->two->path, strlen(p->two->path));
6432 patch_id_add_string(&ctx, "diff--git");
6433 patch_id_add_string(&ctx, "a/");
6434 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6435 patch_id_add_string(&ctx, "b/");
6436 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6437
6438 if (p->one->mode == 0) {
6439 patch_id_add_string(&ctx, "newfilemode");
6440 patch_id_add_mode(&ctx, p->two->mode);
6441 } else if (p->two->mode == 0) {
6442 patch_id_add_string(&ctx, "deletedfilemode");
6443 patch_id_add_mode(&ctx, p->one->mode);
6444 } else if (p->one->mode != p->two->mode) {
6445 patch_id_add_string(&ctx, "oldmode");
6446 patch_id_add_mode(&ctx, p->one->mode);
6447 patch_id_add_string(&ctx, "newmode");
6448 patch_id_add_mode(&ctx, p->two->mode);
6449 }
6450
6451 if (diff_header_only) {
6452 /* don't do anything since we're only populating header info */
6453 } else if (diff_filespec_is_binary(options->repo, p->one) ||
6454 diff_filespec_is_binary(options->repo, p->two)) {
6455 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6456 the_hash_algo->hexsz);
6457 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6458 the_hash_algo->hexsz);
6459 } else {
6460 if (p->one->mode == 0) {
6461 patch_id_add_string(&ctx, "---/dev/null");
6462 patch_id_add_string(&ctx, "+++b/");
6463 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6464 } else if (p->two->mode == 0) {
6465 patch_id_add_string(&ctx, "---a/");
6466 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6467 patch_id_add_string(&ctx, "+++/dev/null");
6468 } else {
6469 patch_id_add_string(&ctx, "---a/");
6470 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6471 patch_id_add_string(&ctx, "+++b/");
6472 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6473 }
6474
6475 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6476 fill_mmfile(options->repo, &mf2, p->two) < 0)
6477 return error("unable to read files to diff");
6478 xpp.flags = 0;
6479 xecfg.ctxlen = 3;
6480 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6481 if (xdi_diff_outf(&mf1, &mf2, NULL,
6482 patch_id_consume, &data, &xpp, &xecfg))
6483 return error("unable to generate patch-id diff for %s",
6484 p->one->path);
6485 }
6486 flush_one_hunk(oid, &ctx);
6487 }
6488
6489 return 0;
6490 }
6491
6492 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
6493 {
6494 struct diff_queue_struct *q = &diff_queued_diff;
6495 int result = diff_get_patch_id(options, oid, diff_header_only);
6496
6497 diff_free_queue(q);
6498 DIFF_QUEUE_CLEAR(q);
6499
6500 return result;
6501 }
6502
6503 static int is_summary_empty(const struct diff_queue_struct *q)
6504 {
6505 int i;
6506
6507 for (i = 0; i < q->nr; i++) {
6508 const struct diff_filepair *p = q->queue[i];
6509
6510 switch (p->status) {
6511 case DIFF_STATUS_DELETED:
6512 case DIFF_STATUS_ADDED:
6513 case DIFF_STATUS_COPIED:
6514 case DIFF_STATUS_RENAMED:
6515 return 0;
6516 default:
6517 if (p->score)
6518 return 0;
6519 if (p->one->mode && p->two->mode &&
6520 p->one->mode != p->two->mode)
6521 return 0;
6522 break;
6523 }
6524 }
6525 return 1;
6526 }
6527
6528 static const char rename_limit_warning[] =
6529 N_("exhaustive rename detection was skipped due to too many files.");
6530
6531 static const char degrade_cc_to_c_warning[] =
6532 N_("only found copies from modified paths due to too many files.");
6533
6534 static const char rename_limit_advice[] =
6535 N_("you may want to set your %s variable to at least "
6536 "%d and retry the command.");
6537
6538 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6539 {
6540 fflush(stdout);
6541 if (degraded_cc)
6542 warning(_(degrade_cc_to_c_warning));
6543 else if (needed)
6544 warning(_(rename_limit_warning));
6545 else
6546 return;
6547 if (0 < needed)
6548 warning(_(rename_limit_advice), varname, needed);
6549 }
6550
6551 static void create_filepairs_for_header_only_notifications(struct diff_options *o)
6552 {
6553 struct strset present;
6554 struct diff_queue_struct *q = &diff_queued_diff;
6555 struct hashmap_iter iter;
6556 struct strmap_entry *e;
6557 int i;
6558
6559 strset_init_with_options(&present, /*pool*/ NULL, /*strdup*/ 0);
6560
6561 /*
6562 * Find out which paths exist in diff_queued_diff, preferring
6563 * one->path for any pair that has multiple paths.
6564 */
6565 for (i = 0; i < q->nr; i++) {
6566 struct diff_filepair *p = q->queue[i];
6567 char *path = p->one->path ? p->one->path : p->two->path;
6568
6569 if (strmap_contains(o->additional_path_headers, path))
6570 strset_add(&present, path);
6571 }
6572
6573 /*
6574 * Loop over paths in additional_path_headers; for each NOT already
6575 * in diff_queued_diff, create a synthetic filepair and insert that
6576 * into diff_queued_diff.
6577 */
6578 strmap_for_each_entry(o->additional_path_headers, &iter, e) {
6579 if (!strset_contains(&present, e->key)) {
6580 struct diff_filespec *one, *two;
6581 struct diff_filepair *p;
6582
6583 one = alloc_filespec(e->key);
6584 two = alloc_filespec(e->key);
6585 fill_filespec(one, null_oid(), 0, 0);
6586 fill_filespec(two, null_oid(), 0, 0);
6587 p = diff_queue(q, one, two);
6588 p->status = DIFF_STATUS_MODIFIED;
6589 }
6590 }
6591
6592 /* Re-sort the filepairs */
6593 diffcore_fix_diff_index();
6594
6595 /* Cleanup */
6596 strset_clear(&present);
6597 }
6598
6599 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6600 {
6601 int i;
6602 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6603 struct diff_queue_struct *q = &diff_queued_diff;
6604
6605 if (WSEH_NEW & WS_RULE_MASK)
6606 BUG("WS rules bit mask overlaps with diff symbol flags");
6607
6608 if (o->color_moved)
6609 o->emitted_symbols = &esm;
6610
6611 if (o->additional_path_headers)
6612 create_filepairs_for_header_only_notifications(o);
6613
6614 for (i = 0; i < q->nr; i++) {
6615 struct diff_filepair *p = q->queue[i];
6616 if (check_pair_status(p))
6617 diff_flush_patch(p, o);
6618 }
6619
6620 if (o->emitted_symbols) {
6621 if (o->color_moved) {
6622 struct mem_pool entry_pool;
6623 struct moved_entry_list *entry_list;
6624
6625 mem_pool_init(&entry_pool, 1024 * 1024);
6626 entry_list = add_lines_to_move_detection(o,
6627 &entry_pool);
6628 mark_color_as_moved(o, entry_list);
6629 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6630 dim_moved_lines(o);
6631
6632 mem_pool_discard(&entry_pool, 0);
6633 free(entry_list);
6634 }
6635
6636 for (i = 0; i < esm.nr; i++)
6637 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6638
6639 for (i = 0; i < esm.nr; i++)
6640 free((void *)esm.buf[i].line);
6641 esm.nr = 0;
6642
6643 o->emitted_symbols = NULL;
6644 }
6645 }
6646
6647 static void diff_free_file(struct diff_options *options)
6648 {
6649 if (options->close_file)
6650 fclose(options->file);
6651 }
6652
6653 static void diff_free_ignore_regex(struct diff_options *options)
6654 {
6655 int i;
6656
6657 for (i = 0; i < options->ignore_regex_nr; i++) {
6658 regfree(options->ignore_regex[i]);
6659 free(options->ignore_regex[i]);
6660 }
6661 free(options->ignore_regex);
6662 }
6663
6664 void diff_free(struct diff_options *options)
6665 {
6666 if (options->no_free)
6667 return;
6668
6669 diff_free_file(options);
6670 diff_free_ignore_regex(options);
6671 clear_pathspec(&options->pathspec);
6672 }
6673
6674 void diff_flush(struct diff_options *options)
6675 {
6676 struct diff_queue_struct *q = &diff_queued_diff;
6677 int i, output_format = options->output_format;
6678 int separator = 0;
6679 int dirstat_by_line = 0;
6680
6681 /*
6682 * Order: raw, stat, summary, patch
6683 * or: name/name-status/checkdiff (other bits clear)
6684 */
6685 if (!q->nr && !options->additional_path_headers)
6686 goto free_queue;
6687
6688 if (output_format & (DIFF_FORMAT_RAW |
6689 DIFF_FORMAT_NAME |
6690 DIFF_FORMAT_NAME_STATUS |
6691 DIFF_FORMAT_CHECKDIFF)) {
6692 for (i = 0; i < q->nr; i++) {
6693 struct diff_filepair *p = q->queue[i];
6694 if (check_pair_status(p))
6695 flush_one_pair(p, options);
6696 }
6697 separator++;
6698 }
6699
6700 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6701 dirstat_by_line = 1;
6702
6703 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6704 dirstat_by_line) {
6705 struct diffstat_t diffstat;
6706
6707 compute_diffstat(options, &diffstat, q);
6708 if (output_format & DIFF_FORMAT_NUMSTAT)
6709 show_numstat(&diffstat, options);
6710 if (output_format & DIFF_FORMAT_DIFFSTAT)
6711 show_stats(&diffstat, options);
6712 if (output_format & DIFF_FORMAT_SHORTSTAT)
6713 show_shortstats(&diffstat, options);
6714 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6715 show_dirstat_by_line(&diffstat, options);
6716 free_diffstat_info(&diffstat);
6717 separator++;
6718 }
6719 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6720 show_dirstat(options);
6721
6722 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6723 for (i = 0; i < q->nr; i++) {
6724 diff_summary(options, q->queue[i]);
6725 }
6726 separator++;
6727 }
6728
6729 if (output_format & DIFF_FORMAT_PATCH) {
6730 if (separator) {
6731 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6732 if (options->stat_sep)
6733 /* attach patch instead of inline */
6734 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6735 NULL, 0, 0);
6736 }
6737
6738 diff_flush_patch_all_file_pairs(options);
6739 }
6740
6741 if (output_format & DIFF_FORMAT_CALLBACK)
6742 options->format_callback(q, options, options->format_callback_data);
6743
6744 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6745 options->flags.exit_with_status &&
6746 options->flags.diff_from_contents) {
6747 /*
6748 * run diff_flush_patch for the exit status. setting
6749 * options->file to /dev/null should be safe, because we
6750 * aren't supposed to produce any output anyway.
6751 */
6752 diff_free_file(options);
6753 options->file = xfopen("/dev/null", "w");
6754 options->close_file = 1;
6755 options->color_moved = 0;
6756 for (i = 0; i < q->nr; i++) {
6757 struct diff_filepair *p = q->queue[i];
6758 if (check_pair_status(p))
6759 diff_flush_patch(p, options);
6760 if (options->found_changes)
6761 break;
6762 }
6763 }
6764
6765 free_queue:
6766 diff_free_queue(q);
6767 DIFF_QUEUE_CLEAR(q);
6768 diff_free(options);
6769
6770 /*
6771 * Report the content-level differences with HAS_CHANGES;
6772 * diff_addremove/diff_change does not set the bit when
6773 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6774 */
6775 if (options->flags.diff_from_contents) {
6776 if (options->found_changes)
6777 options->flags.has_changes = 1;
6778 else
6779 options->flags.has_changes = 0;
6780 }
6781 }
6782
6783 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6784 {
6785 return (((p->status == DIFF_STATUS_MODIFIED) &&
6786 ((p->score &&
6787 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6788 (!p->score &&
6789 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6790 ((p->status != DIFF_STATUS_MODIFIED) &&
6791 filter_bit_tst(p->status, options)));
6792 }
6793
6794 static void diffcore_apply_filter(struct diff_options *options)
6795 {
6796 int i;
6797 struct diff_queue_struct *q = &diff_queued_diff;
6798 struct diff_queue_struct outq;
6799
6800 DIFF_QUEUE_CLEAR(&outq);
6801
6802 if (!options->filter)
6803 return;
6804
6805 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6806 int found;
6807 for (i = found = 0; !found && i < q->nr; i++) {
6808 if (match_filter(options, q->queue[i]))
6809 found++;
6810 }
6811 if (found)
6812 return;
6813
6814 /* otherwise we will clear the whole queue
6815 * by copying the empty outq at the end of this
6816 * function, but first clear the current entries
6817 * in the queue.
6818 */
6819 for (i = 0; i < q->nr; i++)
6820 diff_free_filepair(q->queue[i]);
6821 }
6822 else {
6823 /* Only the matching ones */
6824 for (i = 0; i < q->nr; i++) {
6825 struct diff_filepair *p = q->queue[i];
6826 if (match_filter(options, p))
6827 diff_q(&outq, p);
6828 else
6829 diff_free_filepair(p);
6830 }
6831 }
6832 free(q->queue);
6833 *q = outq;
6834 }
6835
6836 /* Check whether two filespecs with the same mode and size are identical */
6837 static int diff_filespec_is_identical(struct repository *r,
6838 struct diff_filespec *one,
6839 struct diff_filespec *two)
6840 {
6841 if (S_ISGITLINK(one->mode))
6842 return 0;
6843 if (diff_populate_filespec(r, one, NULL))
6844 return 0;
6845 if (diff_populate_filespec(r, two, NULL))
6846 return 0;
6847 return !memcmp(one->data, two->data, one->size);
6848 }
6849
6850 static int diff_filespec_check_stat_unmatch(struct repository *r,
6851 struct diff_filepair *p)
6852 {
6853 struct diff_populate_filespec_options dpf_options = {
6854 .check_size_only = 1,
6855 .missing_object_cb = diff_queued_diff_prefetch,
6856 .missing_object_data = r,
6857 };
6858
6859 if (p->done_skip_stat_unmatch)
6860 return p->skip_stat_unmatch_result;
6861
6862 p->done_skip_stat_unmatch = 1;
6863 p->skip_stat_unmatch_result = 0;
6864 /*
6865 * 1. Entries that come from stat info dirtiness
6866 * always have both sides (iow, not create/delete),
6867 * one side of the object name is unknown, with
6868 * the same mode and size. Keep the ones that
6869 * do not match these criteria. They have real
6870 * differences.
6871 *
6872 * 2. At this point, the file is known to be modified,
6873 * with the same mode and size, and the object
6874 * name of one side is unknown. Need to inspect
6875 * the identical contents.
6876 */
6877 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6878 !DIFF_FILE_VALID(p->two) ||
6879 (p->one->oid_valid && p->two->oid_valid) ||
6880 (p->one->mode != p->two->mode) ||
6881 diff_populate_filespec(r, p->one, &dpf_options) ||
6882 diff_populate_filespec(r, p->two, &dpf_options) ||
6883 (p->one->size != p->two->size) ||
6884 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6885 p->skip_stat_unmatch_result = 1;
6886 return p->skip_stat_unmatch_result;
6887 }
6888
6889 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6890 {
6891 int i;
6892 struct diff_queue_struct *q = &diff_queued_diff;
6893 struct diff_queue_struct outq;
6894 DIFF_QUEUE_CLEAR(&outq);
6895
6896 for (i = 0; i < q->nr; i++) {
6897 struct diff_filepair *p = q->queue[i];
6898
6899 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6900 diff_q(&outq, p);
6901 else {
6902 /*
6903 * The caller can subtract 1 from skip_stat_unmatch
6904 * to determine how many paths were dirty only
6905 * due to stat info mismatch.
6906 */
6907 if (!diffopt->flags.no_index)
6908 diffopt->skip_stat_unmatch++;
6909 diff_free_filepair(p);
6910 }
6911 }
6912 free(q->queue);
6913 *q = outq;
6914 }
6915
6916 static int diffnamecmp(const void *a_, const void *b_)
6917 {
6918 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6919 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6920 const char *name_a, *name_b;
6921
6922 name_a = a->one ? a->one->path : a->two->path;
6923 name_b = b->one ? b->one->path : b->two->path;
6924 return strcmp(name_a, name_b);
6925 }
6926
6927 void diffcore_fix_diff_index(void)
6928 {
6929 struct diff_queue_struct *q = &diff_queued_diff;
6930 QSORT(q->queue, q->nr, diffnamecmp);
6931 }
6932
6933 void diff_add_if_missing(struct repository *r,
6934 struct oid_array *to_fetch,
6935 const struct diff_filespec *filespec)
6936 {
6937 if (filespec && filespec->oid_valid &&
6938 !S_ISGITLINK(filespec->mode) &&
6939 oid_object_info_extended(r, &filespec->oid, NULL,
6940 OBJECT_INFO_FOR_PREFETCH))
6941 oid_array_append(to_fetch, &filespec->oid);
6942 }
6943
6944 void diff_queued_diff_prefetch(void *repository)
6945 {
6946 struct repository *repo = repository;
6947 int i;
6948 struct diff_queue_struct *q = &diff_queued_diff;
6949 struct oid_array to_fetch = OID_ARRAY_INIT;
6950
6951 for (i = 0; i < q->nr; i++) {
6952 struct diff_filepair *p = q->queue[i];
6953 diff_add_if_missing(repo, &to_fetch, p->one);
6954 diff_add_if_missing(repo, &to_fetch, p->two);
6955 }
6956
6957 /*
6958 * NEEDSWORK: Consider deduplicating the OIDs sent.
6959 */
6960 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6961
6962 oid_array_clear(&to_fetch);
6963 }
6964
6965 void init_diffstat_widths(struct diff_options *options)
6966 {
6967 options->stat_width = -1; /* use full terminal width */
6968 options->stat_name_width = -1; /* respect diff.statNameWidth config */
6969 options->stat_graph_width = -1; /* respect diff.statGraphWidth config */
6970 }
6971
6972 void diffcore_std(struct diff_options *options)
6973 {
6974 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6975 DIFF_FORMAT_NUMSTAT |
6976 DIFF_FORMAT_PATCH |
6977 DIFF_FORMAT_SHORTSTAT |
6978 DIFF_FORMAT_DIRSTAT;
6979
6980 /*
6981 * Check if the user requested a blob-data-requiring diff output and/or
6982 * break-rewrite detection (which requires blob data). If yes, prefetch
6983 * the diff pairs.
6984 *
6985 * If no prefetching occurs, diffcore_rename() will prefetch if it
6986 * decides that it needs inexact rename detection.
6987 */
6988 if (options->repo == the_repository && repo_has_promisor_remote(the_repository) &&
6989 (options->output_format & output_formats_to_prefetch ||
6990 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6991 diff_queued_diff_prefetch(options->repo);
6992
6993 /* NOTE please keep the following in sync with diff_tree_combined() */
6994 if (options->skip_stat_unmatch)
6995 diffcore_skip_stat_unmatch(options);
6996 if (!options->found_follow) {
6997 /* See try_to_follow_renames() in tree-diff.c */
6998 if (options->break_opt != -1)
6999 diffcore_break(options->repo,
7000 options->break_opt);
7001 if (options->detect_rename)
7002 diffcore_rename(options);
7003 if (options->break_opt != -1)
7004 diffcore_merge_broken();
7005 }
7006 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
7007 diffcore_pickaxe(options);
7008 if (options->orderfile)
7009 diffcore_order(options->orderfile);
7010 if (options->rotate_to)
7011 diffcore_rotate(options);
7012 if (!options->found_follow)
7013 /* See try_to_follow_renames() in tree-diff.c */
7014 diff_resolve_rename_copy();
7015 diffcore_apply_filter(options);
7016
7017 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
7018 options->flags.has_changes = 1;
7019 else
7020 options->flags.has_changes = 0;
7021
7022 options->found_follow = 0;
7023 }
7024
7025 int diff_result_code(struct diff_options *opt)
7026 {
7027 int result = 0;
7028
7029 diff_warn_rename_limit("diff.renameLimit",
7030 opt->needed_rename_limit,
7031 opt->degraded_cc_to_c);
7032
7033 if (opt->flags.exit_with_status &&
7034 opt->flags.has_changes)
7035 result |= 01;
7036 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
7037 opt->flags.check_failed)
7038 result |= 02;
7039 return result;
7040 }
7041
7042 int diff_can_quit_early(struct diff_options *opt)
7043 {
7044 return (opt->flags.quick &&
7045 !opt->filter &&
7046 opt->flags.has_changes);
7047 }
7048
7049 /*
7050 * Shall changes to this submodule be ignored?
7051 *
7052 * Submodule changes can be configured to be ignored separately for each path,
7053 * but that configuration can be overridden from the command line.
7054 */
7055 static int is_submodule_ignored(const char *path, struct diff_options *options)
7056 {
7057 int ignored = 0;
7058 struct diff_flags orig_flags = options->flags;
7059 if (!options->flags.override_submodule_config)
7060 set_diffopt_flags_from_submodule_config(options, path);
7061 if (options->flags.ignore_submodules)
7062 ignored = 1;
7063 options->flags = orig_flags;
7064 return ignored;
7065 }
7066
7067 void compute_diffstat(struct diff_options *options,
7068 struct diffstat_t *diffstat,
7069 struct diff_queue_struct *q)
7070 {
7071 int i;
7072
7073 memset(diffstat, 0, sizeof(struct diffstat_t));
7074 for (i = 0; i < q->nr; i++) {
7075 struct diff_filepair *p = q->queue[i];
7076 if (check_pair_status(p))
7077 diff_flush_stat(p, options, diffstat);
7078 }
7079 options->found_changes = !!diffstat->nr;
7080 }
7081
7082 void diff_addremove(struct diff_options *options,
7083 int addremove, unsigned mode,
7084 const struct object_id *oid,
7085 int oid_valid,
7086 const char *concatpath, unsigned dirty_submodule)
7087 {
7088 struct diff_filespec *one, *two;
7089
7090 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
7091 return;
7092
7093 /* This may look odd, but it is a preparation for
7094 * feeding "there are unchanged files which should
7095 * not produce diffs, but when you are doing copy
7096 * detection you would need them, so here they are"
7097 * entries to the diff-core. They will be prefixed
7098 * with something like '=' or '*' (I haven't decided
7099 * which but should not make any difference).
7100 * Feeding the same new and old to diff_change()
7101 * also has the same effect.
7102 * Before the final output happens, they are pruned after
7103 * merged into rename/copy pairs as appropriate.
7104 */
7105 if (options->flags.reverse_diff)
7106 addremove = (addremove == '+' ? '-' :
7107 addremove == '-' ? '+' : addremove);
7108
7109 if (options->prefix &&
7110 strncmp(concatpath, options->prefix, options->prefix_length))
7111 return;
7112
7113 one = alloc_filespec(concatpath);
7114 two = alloc_filespec(concatpath);
7115
7116 if (addremove != '+')
7117 fill_filespec(one, oid, oid_valid, mode);
7118 if (addremove != '-') {
7119 fill_filespec(two, oid, oid_valid, mode);
7120 two->dirty_submodule = dirty_submodule;
7121 }
7122
7123 diff_queue(&diff_queued_diff, one, two);
7124 if (!options->flags.diff_from_contents)
7125 options->flags.has_changes = 1;
7126 }
7127
7128 void diff_change(struct diff_options *options,
7129 unsigned old_mode, unsigned new_mode,
7130 const struct object_id *old_oid,
7131 const struct object_id *new_oid,
7132 int old_oid_valid, int new_oid_valid,
7133 const char *concatpath,
7134 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
7135 {
7136 struct diff_filespec *one, *two;
7137 struct diff_filepair *p;
7138
7139 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
7140 is_submodule_ignored(concatpath, options))
7141 return;
7142
7143 if (options->flags.reverse_diff) {
7144 SWAP(old_mode, new_mode);
7145 SWAP(old_oid, new_oid);
7146 SWAP(old_oid_valid, new_oid_valid);
7147 SWAP(old_dirty_submodule, new_dirty_submodule);
7148 }
7149
7150 if (options->prefix &&
7151 strncmp(concatpath, options->prefix, options->prefix_length))
7152 return;
7153
7154 one = alloc_filespec(concatpath);
7155 two = alloc_filespec(concatpath);
7156 fill_filespec(one, old_oid, old_oid_valid, old_mode);
7157 fill_filespec(two, new_oid, new_oid_valid, new_mode);
7158 one->dirty_submodule = old_dirty_submodule;
7159 two->dirty_submodule = new_dirty_submodule;
7160 p = diff_queue(&diff_queued_diff, one, two);
7161
7162 if (options->flags.diff_from_contents)
7163 return;
7164
7165 if (options->flags.quick && options->skip_stat_unmatch &&
7166 !diff_filespec_check_stat_unmatch(options->repo, p)) {
7167 diff_free_filespec_data(p->one);
7168 diff_free_filespec_data(p->two);
7169 return;
7170 }
7171
7172 options->flags.has_changes = 1;
7173 }
7174
7175 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
7176 {
7177 struct diff_filepair *pair;
7178 struct diff_filespec *one, *two;
7179
7180 if (options->prefix &&
7181 strncmp(path, options->prefix, options->prefix_length))
7182 return NULL;
7183
7184 one = alloc_filespec(path);
7185 two = alloc_filespec(path);
7186 pair = diff_queue(&diff_queued_diff, one, two);
7187 pair->is_unmerged = 1;
7188 return pair;
7189 }
7190
7191 static char *run_textconv(struct repository *r,
7192 const char *pgm,
7193 struct diff_filespec *spec,
7194 size_t *outsize)
7195 {
7196 struct diff_tempfile *temp;
7197 struct child_process child = CHILD_PROCESS_INIT;
7198 struct strbuf buf = STRBUF_INIT;
7199 int err = 0;
7200
7201 temp = prepare_temp_file(r, spec);
7202 strvec_push(&child.args, pgm);
7203 strvec_push(&child.args, temp->name);
7204
7205 child.use_shell = 1;
7206 child.out = -1;
7207 if (start_command(&child)) {
7208 remove_tempfile();
7209 return NULL;
7210 }
7211
7212 if (strbuf_read(&buf, child.out, 0) < 0)
7213 err = error("error reading from textconv command '%s'", pgm);
7214 close(child.out);
7215
7216 if (finish_command(&child) || err) {
7217 strbuf_release(&buf);
7218 remove_tempfile();
7219 return NULL;
7220 }
7221 remove_tempfile();
7222
7223 return strbuf_detach(&buf, outsize);
7224 }
7225
7226 size_t fill_textconv(struct repository *r,
7227 struct userdiff_driver *driver,
7228 struct diff_filespec *df,
7229 char **outbuf)
7230 {
7231 size_t size;
7232
7233 if (!driver) {
7234 if (!DIFF_FILE_VALID(df)) {
7235 *outbuf = "";
7236 return 0;
7237 }
7238 if (diff_populate_filespec(r, df, NULL))
7239 die("unable to read files to diff");
7240 *outbuf = df->data;
7241 return df->size;
7242 }
7243
7244 if (!driver->textconv)
7245 BUG("fill_textconv called with non-textconv driver");
7246
7247 if (driver->textconv_cache && df->oid_valid) {
7248 *outbuf = notes_cache_get(driver->textconv_cache,
7249 &df->oid,
7250 &size);
7251 if (*outbuf)
7252 return size;
7253 }
7254
7255 *outbuf = run_textconv(r, driver->textconv, df, &size);
7256 if (!*outbuf)
7257 die("unable to read files to diff");
7258
7259 if (driver->textconv_cache && df->oid_valid) {
7260 /* ignore errors, as we might be in a readonly repository */
7261 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
7262 size);
7263 /*
7264 * we could save up changes and flush them all at the end,
7265 * but we would need an extra call after all diffing is done.
7266 * Since generating a cache entry is the slow path anyway,
7267 * this extra overhead probably isn't a big deal.
7268 */
7269 notes_cache_write(driver->textconv_cache);
7270 }
7271
7272 return size;
7273 }
7274
7275 int textconv_object(struct repository *r,
7276 const char *path,
7277 unsigned mode,
7278 const struct object_id *oid,
7279 int oid_valid,
7280 char **buf,
7281 unsigned long *buf_size)
7282 {
7283 struct diff_filespec *df;
7284 struct userdiff_driver *textconv;
7285
7286 df = alloc_filespec(path);
7287 fill_filespec(df, oid, oid_valid, mode);
7288 textconv = get_textconv(r, df);
7289 if (!textconv) {
7290 free_filespec(df);
7291 return 0;
7292 }
7293
7294 *buf_size = fill_textconv(r, textconv, df, buf);
7295 free_filespec(df);
7296 return 1;
7297 }
7298
7299 void setup_diff_pager(struct diff_options *opt)
7300 {
7301 /*
7302 * If the user asked for our exit code, then either they want --quiet
7303 * or --exit-code. We should definitely not bother with a pager in the
7304 * former case, as we will generate no output. Since we still properly
7305 * report our exit code even when a pager is run, we _could_ run a
7306 * pager with --exit-code. But since we have not done so historically,
7307 * and because it is easy to find people oneline advising "git diff
7308 * --exit-code" in hooks and other scripts, we do not do so.
7309 */
7310 if (!opt->flags.exit_with_status &&
7311 check_pager_config("diff") != 0)
7312 setup_pager();
7313 }
git for Debian