?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'rs/sha1-file-close-mapped-file-on-error'
[git.git] / diff.c
blob1b5f27636061feb0ecf273f257c5e080c5ff7481
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "config.h"
6 #include "tempfile.h"
7 #include "quote.h"
8 #include "diff.h"
9 #include "diffcore.h"
10 #include "delta.h"
11 #include "xdiff-interface.h"
12 #include "color.h"
13 #include "attr.h"
14 #include "run-command.h"
15 #include "utf8.h"
16 #include "object-store.h"
17 #include "userdiff.h"
18 #include "submodule-config.h"
19 #include "submodule.h"
20 #include "hashmap.h"
21 #include "ll-merge.h"
22 #include "string-list.h"
23 #include "argv-array.h"
24 #include "graph.h"
25 #include "packfile.h"
26 #include "help.h"
27
28 #ifdef NO_FAST_WORKING_DIRECTORY
29 #define FAST_WORKING_DIRECTORY 0
30 #else
31 #define FAST_WORKING_DIRECTORY 1
32 #endif
33
34 static int diff_detect_rename_default;
35 static int diff_indent_heuristic = 1;
36 static int diff_rename_limit_default = 400;
37 static int diff_suppress_blank_empty;
38 static int diff_use_color_default = -1;
39 static int diff_color_moved_default;
40 static int diff_color_moved_ws_default;
41 static int diff_context_default = 3;
42 static int diff_interhunk_context_default;
43 static const char *diff_word_regex_cfg;
44 static const char *external_diff_cmd_cfg;
45 static const char *diff_order_file_cfg;
46 int diff_auto_refresh_index = 1;
47 static int diff_mnemonic_prefix;
48 static int diff_no_prefix;
49 static int diff_stat_graph_width;
50 static int diff_dirstat_permille_default = 30;
51 static struct diff_options default_diff_options;
52 static long diff_algorithm;
53 static unsigned ws_error_highlight_default = WSEH_NEW;
54
55 static char diff_colors[][COLOR_MAXLEN] = {
56 GIT_COLOR_RESET,
57 GIT_COLOR_NORMAL, /* CONTEXT */
58 GIT_COLOR_BOLD, /* METAINFO */
59 GIT_COLOR_CYAN, /* FRAGINFO */
60 GIT_COLOR_RED, /* OLD */
61 GIT_COLOR_GREEN, /* NEW */
62 GIT_COLOR_YELLOW, /* COMMIT */
63 GIT_COLOR_BG_RED, /* WHITESPACE */
64 GIT_COLOR_NORMAL, /* FUNCINFO */
65 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
66 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
67 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
68 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
69 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
70 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
71 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
72 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
73 GIT_COLOR_FAINT, /* CONTEXT_DIM */
74 GIT_COLOR_FAINT_RED, /* OLD_DIM */
75 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
76 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
77 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
78 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
79 };
80
81 static const char *color_diff_slots[] = {
82 [DIFF_CONTEXT] = "context",
83 [DIFF_METAINFO] = "meta",
84 [DIFF_FRAGINFO] = "frag",
85 [DIFF_FILE_OLD] = "old",
86 [DIFF_FILE_NEW] = "new",
87 [DIFF_COMMIT] = "commit",
88 [DIFF_WHITESPACE] = "whitespace",
89 [DIFF_FUNCINFO] = "func",
90 [DIFF_FILE_OLD_MOVED] = "oldMoved",
91 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
92 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
93 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
94 [DIFF_FILE_NEW_MOVED] = "newMoved",
95 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
96 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
97 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
98 [DIFF_CONTEXT_DIM] = "contextDimmed",
99 [DIFF_FILE_OLD_DIM] = "oldDimmed",
100 [DIFF_FILE_NEW_DIM] = "newDimmed",
101 [DIFF_CONTEXT_BOLD] = "contextBold",
102 [DIFF_FILE_OLD_BOLD] = "oldBold",
103 [DIFF_FILE_NEW_BOLD] = "newBold",
104 };
105
106 static NORETURN void die_want_option(const char *option_name)
107 {
108 die(_("option '%s' requires a value"), option_name);
109 }
110
111 define_list_config_array_extra(color_diff_slots, {"plain"});
112
113 static int parse_diff_color_slot(const char *var)
114 {
115 if (!strcasecmp(var, "plain"))
116 return DIFF_CONTEXT;
117 return LOOKUP_CONFIG(color_diff_slots, var);
118 }
119
120 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
121 struct strbuf *errmsg)
122 {
123 char *params_copy = xstrdup(params_string);
124 struct string_list params = STRING_LIST_INIT_NODUP;
125 int ret = 0;
126 int i;
127
128 if (*params_copy)
129 string_list_split_in_place(&params, params_copy, ',', -1);
130 for (i = 0; i < params.nr; i++) {
131 const char *p = params.items[i].string;
132 if (!strcmp(p, "changes")) {
133 options->flags.dirstat_by_line = 0;
134 options->flags.dirstat_by_file = 0;
135 } else if (!strcmp(p, "lines")) {
136 options->flags.dirstat_by_line = 1;
137 options->flags.dirstat_by_file = 0;
138 } else if (!strcmp(p, "files")) {
139 options->flags.dirstat_by_line = 0;
140 options->flags.dirstat_by_file = 1;
141 } else if (!strcmp(p, "noncumulative")) {
142 options->flags.dirstat_cumulative = 0;
143 } else if (!strcmp(p, "cumulative")) {
144 options->flags.dirstat_cumulative = 1;
145 } else if (isdigit(*p)) {
146 char *end;
147 int permille = strtoul(p, &end, 10) * 10;
148 if (*end == '.' && isdigit(*++end)) {
149 /* only use first digit */
150 permille += *end - '0';
151 /* .. and ignore any further digits */
152 while (isdigit(*++end))
153 ; /* nothing */
154 }
155 if (!*end)
156 options->dirstat_permille = permille;
157 else {
158 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
159 p);
160 ret++;
161 }
162 } else {
163 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
164 ret++;
165 }
166
167 }
168 string_list_clear(&params, 0);
169 free(params_copy);
170 return ret;
171 }
172
173 static int parse_submodule_params(struct diff_options *options, const char *value)
174 {
175 if (!strcmp(value, "log"))
176 options->submodule_format = DIFF_SUBMODULE_LOG;
177 else if (!strcmp(value, "short"))
178 options->submodule_format = DIFF_SUBMODULE_SHORT;
179 else if (!strcmp(value, "diff"))
180 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
181 else
182 return -1;
183 return 0;
184 }
185
186 int git_config_rename(const char *var, const char *value)
187 {
188 if (!value)
189 return DIFF_DETECT_RENAME;
190 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
191 return DIFF_DETECT_COPY;
192 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
193 }
194
195 long parse_algorithm_value(const char *value)
196 {
197 if (!value)
198 return -1;
199 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
200 return 0;
201 else if (!strcasecmp(value, "minimal"))
202 return XDF_NEED_MINIMAL;
203 else if (!strcasecmp(value, "patience"))
204 return XDF_PATIENCE_DIFF;
205 else if (!strcasecmp(value, "histogram"))
206 return XDF_HISTOGRAM_DIFF;
207 return -1;
208 }
209
210 static int parse_one_token(const char **arg, const char *token)
211 {
212 const char *rest;
213 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
214 *arg = rest;
215 return 1;
216 }
217 return 0;
218 }
219
220 static int parse_ws_error_highlight(const char *arg)
221 {
222 const char *orig_arg = arg;
223 unsigned val = 0;
224
225 while (*arg) {
226 if (parse_one_token(&arg, "none"))
227 val = 0;
228 else if (parse_one_token(&arg, "default"))
229 val = WSEH_NEW;
230 else if (parse_one_token(&arg, "all"))
231 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
232 else if (parse_one_token(&arg, "new"))
233 val |= WSEH_NEW;
234 else if (parse_one_token(&arg, "old"))
235 val |= WSEH_OLD;
236 else if (parse_one_token(&arg, "context"))
237 val |= WSEH_CONTEXT;
238 else {
239 return -1 - (int)(arg - orig_arg);
240 }
241 if (*arg)
242 arg++;
243 }
244 return val;
245 }
246
247 /*
248 * These are to give UI layer defaults.
249 * The core-level commands such as git-diff-files should
250 * never be affected by the setting of diff.renames
251 * the user happens to have in the configuration file.
252 */
253 void init_diff_ui_defaults(void)
254 {
255 diff_detect_rename_default = DIFF_DETECT_RENAME;
256 }
257
258 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
259 {
260 if (!strcmp(var, "diff.indentheuristic"))
261 diff_indent_heuristic = git_config_bool(var, value);
262 return 0;
263 }
264
265 static int parse_color_moved(const char *arg)
266 {
267 switch (git_parse_maybe_bool(arg)) {
268 case 0:
269 return COLOR_MOVED_NO;
270 case 1:
271 return COLOR_MOVED_DEFAULT;
272 default:
273 break;
274 }
275
276 if (!strcmp(arg, "no"))
277 return COLOR_MOVED_NO;
278 else if (!strcmp(arg, "plain"))
279 return COLOR_MOVED_PLAIN;
280 else if (!strcmp(arg, "blocks"))
281 return COLOR_MOVED_BLOCKS;
282 else if (!strcmp(arg, "zebra"))
283 return COLOR_MOVED_ZEBRA;
284 else if (!strcmp(arg, "default"))
285 return COLOR_MOVED_DEFAULT;
286 else if (!strcmp(arg, "dimmed-zebra"))
287 return COLOR_MOVED_ZEBRA_DIM;
288 else if (!strcmp(arg, "dimmed_zebra"))
289 return COLOR_MOVED_ZEBRA_DIM;
290 else
291 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
292 }
293
294 static unsigned parse_color_moved_ws(const char *arg)
295 {
296 int ret = 0;
297 struct string_list l = STRING_LIST_INIT_DUP;
298 struct string_list_item *i;
299
300 string_list_split(&l, arg, ',', -1);
301
302 for_each_string_list_item(i, &l) {
303 struct strbuf sb = STRBUF_INIT;
304 strbuf_addstr(&sb, i->string);
305 strbuf_trim(&sb);
306
307 if (!strcmp(sb.buf, "ignore-space-change"))
308 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
309 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
310 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
311 else if (!strcmp(sb.buf, "ignore-all-space"))
312 ret |= XDF_IGNORE_WHITESPACE;
313 else if (!strcmp(sb.buf, "allow-indentation-change"))
314 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
315 else {
316 ret |= COLOR_MOVED_WS_ERROR;
317 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);
318 }
319
320 strbuf_release(&sb);
321 }
322
323 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
324 (ret & XDF_WHITESPACE_FLAGS)) {
325 error(_("color-moved-ws: allow-indentation-change cannot be combined with other white space modes"));
326 ret |= COLOR_MOVED_WS_ERROR;
327 }
328
329 string_list_clear(&l, 0);
330
331 return ret;
332 }
333
334 int git_diff_ui_config(const char *var, const char *value, void *cb)
335 {
336 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
337 diff_use_color_default = git_config_colorbool(var, value);
338 return 0;
339 }
340 if (!strcmp(var, "diff.colormoved")) {
341 int cm = parse_color_moved(value);
342 if (cm < 0)
343 return -1;
344 diff_color_moved_default = cm;
345 return 0;
346 }
347 if (!strcmp(var, "diff.colormovedws")) {
348 unsigned cm = parse_color_moved_ws(value);
349 if (cm & COLOR_MOVED_WS_ERROR)
350 return -1;
351 diff_color_moved_ws_default = cm;
352 return 0;
353 }
354 if (!strcmp(var, "diff.context")) {
355 diff_context_default = git_config_int(var, value);
356 if (diff_context_default < 0)
357 return -1;
358 return 0;
359 }
360 if (!strcmp(var, "diff.interhunkcontext")) {
361 diff_interhunk_context_default = git_config_int(var, value);
362 if (diff_interhunk_context_default < 0)
363 return -1;
364 return 0;
365 }
366 if (!strcmp(var, "diff.renames")) {
367 diff_detect_rename_default = git_config_rename(var, value);
368 return 0;
369 }
370 if (!strcmp(var, "diff.autorefreshindex")) {
371 diff_auto_refresh_index = git_config_bool(var, value);
372 return 0;
373 }
374 if (!strcmp(var, "diff.mnemonicprefix")) {
375 diff_mnemonic_prefix = git_config_bool(var, value);
376 return 0;
377 }
378 if (!strcmp(var, "diff.noprefix")) {
379 diff_no_prefix = git_config_bool(var, value);
380 return 0;
381 }
382 if (!strcmp(var, "diff.statgraphwidth")) {
383 diff_stat_graph_width = git_config_int(var, value);
384 return 0;
385 }
386 if (!strcmp(var, "diff.external"))
387 return git_config_string(&external_diff_cmd_cfg, var, value);
388 if (!strcmp(var, "diff.wordregex"))
389 return git_config_string(&diff_word_regex_cfg, var, value);
390 if (!strcmp(var, "diff.orderfile"))
391 return git_config_pathname(&diff_order_file_cfg, var, value);
392
393 if (!strcmp(var, "diff.ignoresubmodules"))
394 handle_ignore_submodules_arg(&default_diff_options, value);
395
396 if (!strcmp(var, "diff.submodule")) {
397 if (parse_submodule_params(&default_diff_options, value))
398 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
399 value);
400 return 0;
401 }
402
403 if (!strcmp(var, "diff.algorithm")) {
404 diff_algorithm = parse_algorithm_value(value);
405 if (diff_algorithm < 0)
406 return -1;
407 return 0;
408 }
409
410 if (!strcmp(var, "diff.wserrorhighlight")) {
411 int val = parse_ws_error_highlight(value);
412 if (val < 0)
413 return -1;
414 ws_error_highlight_default = val;
415 return 0;
416 }
417
418 if (git_color_config(var, value, cb) < 0)
419 return -1;
420
421 return git_diff_basic_config(var, value, cb);
422 }
423
424 int git_diff_basic_config(const char *var, const char *value, void *cb)
425 {
426 const char *name;
427
428 if (!strcmp(var, "diff.renamelimit")) {
429 diff_rename_limit_default = git_config_int(var, value);
430 return 0;
431 }
432
433 if (userdiff_config(var, value) < 0)
434 return -1;
435
436 if (skip_prefix(var, "diff.color.", &name) ||
437 skip_prefix(var, "color.diff.", &name)) {
438 int slot = parse_diff_color_slot(name);
439 if (slot < 0)
440 return 0;
441 if (!value)
442 return config_error_nonbool(var);
443 return color_parse(value, diff_colors[slot]);
444 }
445
446 /* like GNU diff's --suppress-blank-empty option */
447 if (!strcmp(var, "diff.suppressblankempty") ||
448 /* for backwards compatibility */
449 !strcmp(var, "diff.suppress-blank-empty")) {
450 diff_suppress_blank_empty = git_config_bool(var, value);
451 return 0;
452 }
453
454 if (!strcmp(var, "diff.dirstat")) {
455 struct strbuf errmsg = STRBUF_INIT;
456 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
457 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
458 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
459 errmsg.buf);
460 strbuf_release(&errmsg);
461 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
462 return 0;
463 }
464
465 if (git_diff_heuristic_config(var, value, cb) < 0)
466 return -1;
467
468 return git_default_config(var, value, cb);
469 }
470
471 static char *quote_two(const char *one, const char *two)
472 {
473 int need_one = quote_c_style(one, NULL, NULL, 1);
474 int need_two = quote_c_style(two, NULL, NULL, 1);
475 struct strbuf res = STRBUF_INIT;
476
477 if (need_one + need_two) {
478 strbuf_addch(&res, '"');
479 quote_c_style(one, &res, NULL, 1);
480 quote_c_style(two, &res, NULL, 1);
481 strbuf_addch(&res, '"');
482 } else {
483 strbuf_addstr(&res, one);
484 strbuf_addstr(&res, two);
485 }
486 return strbuf_detach(&res, NULL);
487 }
488
489 static const char *external_diff(void)
490 {
491 static const char *external_diff_cmd = NULL;
492 static int done_preparing = 0;
493
494 if (done_preparing)
495 return external_diff_cmd;
496 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
497 if (!external_diff_cmd)
498 external_diff_cmd = external_diff_cmd_cfg;
499 done_preparing = 1;
500 return external_diff_cmd;
501 }
502
503 /*
504 * Keep track of files used for diffing. Sometimes such an entry
505 * refers to a temporary file, sometimes to an existing file, and
506 * sometimes to "/dev/null".
507 */
508 static struct diff_tempfile {
509 /*
510 * filename external diff should read from, or NULL if this
511 * entry is currently not in use:
512 */
513 const char *name;
514
515 char hex[GIT_MAX_HEXSZ + 1];
516 char mode[10];
517
518 /*
519 * If this diff_tempfile instance refers to a temporary file,
520 * this tempfile object is used to manage its lifetime.
521 */
522 struct tempfile *tempfile;
523 } diff_temp[2];
524
525 struct emit_callback {
526 int color_diff;
527 unsigned ws_rule;
528 int blank_at_eof_in_preimage;
529 int blank_at_eof_in_postimage;
530 int lno_in_preimage;
531 int lno_in_postimage;
532 const char **label_path;
533 struct diff_words_data *diff_words;
534 struct diff_options *opt;
535 struct strbuf *header;
536 };
537
538 static int count_lines(const char *data, int size)
539 {
540 int count, ch, completely_empty = 1, nl_just_seen = 0;
541 count = 0;
542 while (0 < size--) {
543 ch = *data++;
544 if (ch == '\n') {
545 count++;
546 nl_just_seen = 1;
547 completely_empty = 0;
548 }
549 else {
550 nl_just_seen = 0;
551 completely_empty = 0;
552 }
553 }
554 if (completely_empty)
555 return 0;
556 if (!nl_just_seen)
557 count++; /* no trailing newline */
558 return count;
559 }
560
561 static int fill_mmfile(struct repository *r, mmfile_t *mf,
562 struct diff_filespec *one)
563 {
564 if (!DIFF_FILE_VALID(one)) {
565 mf->ptr = (char *)""; /* does not matter */
566 mf->size = 0;
567 return 0;
568 }
569 else if (diff_populate_filespec(r, one, 0))
570 return -1;
571
572 mf->ptr = one->data;
573 mf->size = one->size;
574 return 0;
575 }
576
577 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
578 static unsigned long diff_filespec_size(struct repository *r,
579 struct diff_filespec *one)
580 {
581 if (!DIFF_FILE_VALID(one))
582 return 0;
583 diff_populate_filespec(r, one, CHECK_SIZE_ONLY);
584 return one->size;
585 }
586
587 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
588 {
589 char *ptr = mf->ptr;
590 long size = mf->size;
591 int cnt = 0;
592
593 if (!size)
594 return cnt;
595 ptr += size - 1; /* pointing at the very end */
596 if (*ptr != '\n')
597 ; /* incomplete line */
598 else
599 ptr--; /* skip the last LF */
600 while (mf->ptr < ptr) {
601 char *prev_eol;
602 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
603 if (*prev_eol == '\n')
604 break;
605 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
606 break;
607 cnt++;
608 ptr = prev_eol - 1;
609 }
610 return cnt;
611 }
612
613 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
614 struct emit_callback *ecbdata)
615 {
616 int l1, l2, at;
617 unsigned ws_rule = ecbdata->ws_rule;
618 l1 = count_trailing_blank(mf1, ws_rule);
619 l2 = count_trailing_blank(mf2, ws_rule);
620 if (l2 <= l1) {
621 ecbdata->blank_at_eof_in_preimage = 0;
622 ecbdata->blank_at_eof_in_postimage = 0;
623 return;
624 }
625 at = count_lines(mf1->ptr, mf1->size);
626 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
627
628 at = count_lines(mf2->ptr, mf2->size);
629 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
630 }
631
632 static void emit_line_0(struct diff_options *o,
633 const char *set_sign, const char *set, unsigned reverse, const char *reset,
634 int first, const char *line, int len)
635 {
636 int has_trailing_newline, has_trailing_carriage_return;
637 int needs_reset = 0; /* at the end of the line */
638 FILE *file = o->file;
639
640 fputs(diff_line_prefix(o), file);
641
642 has_trailing_newline = (len > 0 && line[len-1] == '\n');
643 if (has_trailing_newline)
644 len--;
645
646 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
647 if (has_trailing_carriage_return)
648 len--;
649
650 if (!len && !first)
651 goto end_of_line;
652
653 if (reverse && want_color(o->use_color)) {
654 fputs(GIT_COLOR_REVERSE, file);
655 needs_reset = 1;
656 }
657
658 if (set_sign) {
659 fputs(set_sign, file);
660 needs_reset = 1;
661 }
662
663 if (first)
664 fputc(first, file);
665
666 if (!len)
667 goto end_of_line;
668
669 if (set) {
670 if (set_sign && set != set_sign)
671 fputs(reset, file);
672 fputs(set, file);
673 needs_reset = 1;
674 }
675 fwrite(line, len, 1, file);
676 needs_reset = 1; /* 'line' may contain color codes. */
677
678 end_of_line:
679 if (needs_reset)
680 fputs(reset, file);
681 if (has_trailing_carriage_return)
682 fputc('\r', file);
683 if (has_trailing_newline)
684 fputc('\n', file);
685 }
686
687 static void emit_line(struct diff_options *o, const char *set, const char *reset,
688 const char *line, int len)
689 {
690 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
691 }
692
693 enum diff_symbol {
694 DIFF_SYMBOL_BINARY_DIFF_HEADER,
695 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
696 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
697 DIFF_SYMBOL_BINARY_DIFF_BODY,
698 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
699 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
700 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
701 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
702 DIFF_SYMBOL_STATS_LINE,
703 DIFF_SYMBOL_WORD_DIFF,
704 DIFF_SYMBOL_STAT_SEP,
705 DIFF_SYMBOL_SUMMARY,
706 DIFF_SYMBOL_SUBMODULE_ADD,
707 DIFF_SYMBOL_SUBMODULE_DEL,
708 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
709 DIFF_SYMBOL_SUBMODULE_MODIFIED,
710 DIFF_SYMBOL_SUBMODULE_HEADER,
711 DIFF_SYMBOL_SUBMODULE_ERROR,
712 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
713 DIFF_SYMBOL_REWRITE_DIFF,
714 DIFF_SYMBOL_BINARY_FILES,
715 DIFF_SYMBOL_HEADER,
716 DIFF_SYMBOL_FILEPAIR_PLUS,
717 DIFF_SYMBOL_FILEPAIR_MINUS,
718 DIFF_SYMBOL_WORDS_PORCELAIN,
719 DIFF_SYMBOL_WORDS,
720 DIFF_SYMBOL_CONTEXT,
721 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
722 DIFF_SYMBOL_PLUS,
723 DIFF_SYMBOL_MINUS,
724 DIFF_SYMBOL_NO_LF_EOF,
725 DIFF_SYMBOL_CONTEXT_FRAGINFO,
726 DIFF_SYMBOL_CONTEXT_MARKER,
727 DIFF_SYMBOL_SEPARATOR
728 };
729 /*
730 * Flags for content lines:
731 * 0..12 are whitespace rules
732 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
733 * 16 is marking if the line is blank at EOF
734 */
735 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
736 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
737 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
738 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
739 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
740
741 /*
742 * This struct is used when we need to buffer the output of the diff output.
743 *
744 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
745 * into the pre/post image file. This pointer could be a union with the
746 * line pointer. By storing an offset into the file instead of the literal line,
747 * we can decrease the memory footprint for the buffered output. At first we
748 * may want to only have indirection for the content lines, but we could also
749 * enhance the state for emitting prefabricated lines, e.g. the similarity
750 * score line or hunk/file headers would only need to store a number or path
751 * and then the output can be constructed later on depending on state.
752 */
753 struct emitted_diff_symbol {
754 const char *line;
755 int len;
756 int flags;
757 enum diff_symbol s;
758 };
759 #define EMITTED_DIFF_SYMBOL_INIT {NULL}
760
761 struct emitted_diff_symbols {
762 struct emitted_diff_symbol *buf;
763 int nr, alloc;
764 };
765 #define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
766
767 static void append_emitted_diff_symbol(struct diff_options *o,
768 struct emitted_diff_symbol *e)
769 {
770 struct emitted_diff_symbol *f;
771
772 ALLOC_GROW(o->emitted_symbols->buf,
773 o->emitted_symbols->nr + 1,
774 o->emitted_symbols->alloc);
775 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
776
777 memcpy(f, e, sizeof(struct emitted_diff_symbol));
778 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
779 }
780
781 struct moved_entry {
782 struct hashmap_entry ent;
783 const struct emitted_diff_symbol *es;
784 struct moved_entry *next_line;
785 };
786
787 /**
788 * The struct ws_delta holds white space differences between moved lines, i.e.
789 * between '+' and '-' lines that have been detected to be a move.
790 * The string contains the difference in leading white spaces, before the
791 * rest of the line is compared using the white space config for move
792 * coloring. The current_longer indicates if the first string in the
793 * comparision is longer than the second.
794 */
795 struct ws_delta {
796 char *string;
797 unsigned int current_longer : 1;
798 };
799 #define WS_DELTA_INIT { NULL, 0 }
800
801 struct moved_block {
802 struct moved_entry *match;
803 struct ws_delta wsd;
804 };
805
806 static void moved_block_clear(struct moved_block *b)
807 {
808 FREE_AND_NULL(b->wsd.string);
809 b->match = NULL;
810 }
811
812 static int compute_ws_delta(const struct emitted_diff_symbol *a,
813 const struct emitted_diff_symbol *b,
814 struct ws_delta *out)
815 {
816 const struct emitted_diff_symbol *longer = a->len > b->len ? a : b;
817 const struct emitted_diff_symbol *shorter = a->len > b->len ? b : a;
818 int d = longer->len - shorter->len;
819
820 if (strncmp(longer->line + d, shorter->line, shorter->len))
821 return 0;
822
823 out->string = xmemdupz(longer->line, d);
824 out->current_longer = (a == longer);
825
826 return 1;
827 }
828
829 static int cmp_in_block_with_wsd(const struct diff_options *o,
830 const struct moved_entry *cur,
831 const struct moved_entry *match,
832 struct moved_block *pmb,
833 int n)
834 {
835 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
836 int al = cur->es->len, cl = l->len;
837 const char *a = cur->es->line,
838 *b = match->es->line,
839 *c = l->line;
840
841 int wslen;
842
843 /*
844 * We need to check if 'cur' is equal to 'match'.
845 * As those are from the same (+/-) side, we do not need to adjust for
846 * indent changes. However these were found using fuzzy matching
847 * so we do have to check if they are equal.
848 */
849 if (strcmp(a, b))
850 return 1;
851
852 if (!pmb->wsd.string)
853 /*
854 * The white space delta is not active? This can happen
855 * when we exit early in this function.
856 */
857 return 1;
858
859 /*
860 * The indent changes of the block are known and stored in
861 * pmb->wsd; however we need to check if the indent changes of the
862 * current line are still the same as before.
863 *
864 * To do so we need to compare 'l' to 'cur', adjusting the
865 * one of them for the white spaces, depending which was longer.
866 */
867
868 wslen = strlen(pmb->wsd.string);
869 if (pmb->wsd.current_longer) {
870 c += wslen;
871 cl -= wslen;
872 } else {
873 a += wslen;
874 al -= wslen;
875 }
876
877 if (al != cl || memcmp(a, c, al))
878 return 1;
879
880 return 0;
881 }
882
883 static int moved_entry_cmp(const void *hashmap_cmp_fn_data,
884 const void *entry,
885 const void *entry_or_key,
886 const void *keydata)
887 {
888 const struct diff_options *diffopt = hashmap_cmp_fn_data;
889 const struct moved_entry *a = entry;
890 const struct moved_entry *b = entry_or_key;
891 unsigned flags = diffopt->color_moved_ws_handling
892 & XDF_WHITESPACE_FLAGS;
893
894 if (diffopt->color_moved_ws_handling &
895 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
896 /*
897 * As there is not specific white space config given,
898 * we'd need to check for a new block, so ignore all
899 * white space. The setup of the white space
900 * configuration for the next block is done else where
901 */
902 flags |= XDF_IGNORE_WHITESPACE;
903
904 return !xdiff_compare_lines(a->es->line, a->es->len,
905 b->es->line, b->es->len,
906 flags);
907 }
908
909 static struct moved_entry *prepare_entry(struct diff_options *o,
910 int line_no)
911 {
912 struct moved_entry *ret = xmalloc(sizeof(*ret));
913 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
914 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
915
916 ret->ent.hash = xdiff_hash_string(l->line, l->len, flags);
917 ret->es = l;
918 ret->next_line = NULL;
919
920 return ret;
921 }
922
923 static void add_lines_to_move_detection(struct diff_options *o,
924 struct hashmap *add_lines,
925 struct hashmap *del_lines)
926 {
927 struct moved_entry *prev_line = NULL;
928
929 int n;
930 for (n = 0; n < o->emitted_symbols->nr; n++) {
931 struct hashmap *hm;
932 struct moved_entry *key;
933
934 switch (o->emitted_symbols->buf[n].s) {
935 case DIFF_SYMBOL_PLUS:
936 hm = add_lines;
937 break;
938 case DIFF_SYMBOL_MINUS:
939 hm = del_lines;
940 break;
941 default:
942 prev_line = NULL;
943 continue;
944 }
945
946 key = prepare_entry(o, n);
947 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
948 prev_line->next_line = key;
949
950 hashmap_add(hm, key);
951 prev_line = key;
952 }
953 }
954
955 static void pmb_advance_or_null(struct diff_options *o,
956 struct moved_entry *match,
957 struct hashmap *hm,
958 struct moved_block *pmb,
959 int pmb_nr)
960 {
961 int i;
962 for (i = 0; i < pmb_nr; i++) {
963 struct moved_entry *prev = pmb[i].match;
964 struct moved_entry *cur = (prev && prev->next_line) ?
965 prev->next_line : NULL;
966 if (cur && !hm->cmpfn(o, cur, match, NULL)) {
967 pmb[i].match = cur;
968 } else {
969 pmb[i].match = NULL;
970 }
971 }
972 }
973
974 static void pmb_advance_or_null_multi_match(struct diff_options *o,
975 struct moved_entry *match,
976 struct hashmap *hm,
977 struct moved_block *pmb,
978 int pmb_nr, int n)
979 {
980 int i;
981 char *got_match = xcalloc(1, pmb_nr);
982
983 for (; match; match = hashmap_get_next(hm, match)) {
984 for (i = 0; i < pmb_nr; i++) {
985 struct moved_entry *prev = pmb[i].match;
986 struct moved_entry *cur = (prev && prev->next_line) ?
987 prev->next_line : NULL;
988 if (!cur)
989 continue;
990 if (!cmp_in_block_with_wsd(o, cur, match, &pmb[i], n))
991 got_match[i] |= 1;
992 }
993 }
994
995 for (i = 0; i < pmb_nr; i++) {
996 if (got_match[i]) {
997 /* Advance to the next line */
998 pmb[i].match = pmb[i].match->next_line;
999 } else {
1000 moved_block_clear(&pmb[i]);
1001 }
1002 }
1003
1004 free(got_match);
1005 }
1006
1007 static int shrink_potential_moved_blocks(struct moved_block *pmb,
1008 int pmb_nr)
1009 {
1010 int lp, rp;
1011
1012 /* Shrink the set of potential block to the remaining running */
1013 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
1014 while (lp < pmb_nr && pmb[lp].match)
1015 lp++;
1016 /* lp points at the first NULL now */
1017
1018 while (rp > -1 && !pmb[rp].match)
1019 rp--;
1020 /* rp points at the last non-NULL */
1021
1022 if (lp < pmb_nr && rp > -1 && lp < rp) {
1023 pmb[lp] = pmb[rp];
1024 pmb[rp].match = NULL;
1025 pmb[rp].wsd.string = NULL;
1026 rp--;
1027 lp++;
1028 }
1029 }
1030
1031 /* Remember the number of running sets */
1032 return rp + 1;
1033 }
1034
1035 /*
1036 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1037 *
1038 * Otherwise, if the last block has fewer alphanumeric characters than
1039 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1040 * that block.
1041 *
1042 * The last block consists of the (n - block_length)'th line up to but not
1043 * including the nth line.
1044 *
1045 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1046 * Think of a way to unify them.
1047 */
1048 static void adjust_last_block(struct diff_options *o, int n, int block_length)
1049 {
1050 int i, alnum_count = 0;
1051 if (o->color_moved == COLOR_MOVED_PLAIN)
1052 return;
1053 for (i = 1; i < block_length + 1; i++) {
1054 const char *c = o->emitted_symbols->buf[n - i].line;
1055 for (; *c; c++) {
1056 if (!isalnum(*c))
1057 continue;
1058 alnum_count++;
1059 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1060 return;
1061 }
1062 }
1063 for (i = 1; i < block_length + 1; i++)
1064 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
1065 }
1066
1067 /* Find blocks of moved code, delegate actual coloring decision to helper */
1068 static void mark_color_as_moved(struct diff_options *o,
1069 struct hashmap *add_lines,
1070 struct hashmap *del_lines)
1071 {
1072 struct moved_block *pmb = NULL; /* potentially moved blocks */
1073 int pmb_nr = 0, pmb_alloc = 0;
1074 int n, flipped_block = 1, block_length = 0;
1075
1076
1077 for (n = 0; n < o->emitted_symbols->nr; n++) {
1078 struct hashmap *hm = NULL;
1079 struct moved_entry *key;
1080 struct moved_entry *match = NULL;
1081 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1082
1083 switch (l->s) {
1084 case DIFF_SYMBOL_PLUS:
1085 hm = del_lines;
1086 key = prepare_entry(o, n);
1087 match = hashmap_get(hm, key, NULL);
1088 free(key);
1089 break;
1090 case DIFF_SYMBOL_MINUS:
1091 hm = add_lines;
1092 key = prepare_entry(o, n);
1093 match = hashmap_get(hm, key, NULL);
1094 free(key);
1095 break;
1096 default:
1097 flipped_block = 1;
1098 }
1099
1100 if (!match) {
1101 int i;
1102
1103 adjust_last_block(o, n, block_length);
1104 for(i = 0; i < pmb_nr; i++)
1105 moved_block_clear(&pmb[i]);
1106 pmb_nr = 0;
1107 block_length = 0;
1108 continue;
1109 }
1110
1111 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1112
1113 if (o->color_moved == COLOR_MOVED_PLAIN)
1114 continue;
1115
1116 if (o->color_moved_ws_handling &
1117 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1118 pmb_advance_or_null_multi_match(o, match, hm, pmb, pmb_nr, n);
1119 else
1120 pmb_advance_or_null(o, match, hm, pmb, pmb_nr);
1121
1122 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
1123
1124 if (pmb_nr == 0) {
1125 /*
1126 * The current line is the start of a new block.
1127 * Setup the set of potential blocks.
1128 */
1129 for (; match; match = hashmap_get_next(hm, match)) {
1130 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1131 if (o->color_moved_ws_handling &
1132 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) {
1133 if (compute_ws_delta(l, match->es,
1134 &pmb[pmb_nr].wsd))
1135 pmb[pmb_nr++].match = match;
1136 } else {
1137 pmb[pmb_nr].wsd.string = NULL;
1138 pmb[pmb_nr++].match = match;
1139 }
1140 }
1141
1142 flipped_block = (flipped_block + 1) % 2;
1143
1144 adjust_last_block(o, n, block_length);
1145 block_length = 0;
1146 }
1147
1148 block_length++;
1149
1150 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1151 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1152 }
1153 adjust_last_block(o, n, block_length);
1154
1155 for(n = 0; n < pmb_nr; n++)
1156 moved_block_clear(&pmb[n]);
1157 free(pmb);
1158 }
1159
1160 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1161 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1162 static void dim_moved_lines(struct diff_options *o)
1163 {
1164 int n;
1165 for (n = 0; n < o->emitted_symbols->nr; n++) {
1166 struct emitted_diff_symbol *prev = (n != 0) ?
1167 &o->emitted_symbols->buf[n - 1] : NULL;
1168 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1169 struct emitted_diff_symbol *next =
1170 (n < o->emitted_symbols->nr - 1) ?
1171 &o->emitted_symbols->buf[n + 1] : NULL;
1172
1173 /* Not a plus or minus line? */
1174 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1175 continue;
1176
1177 /* Not a moved line? */
1178 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1179 continue;
1180
1181 /*
1182 * If prev or next are not a plus or minus line,
1183 * pretend they don't exist
1184 */
1185 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1186 prev->s != DIFF_SYMBOL_MINUS)
1187 prev = NULL;
1188 if (next && next->s != DIFF_SYMBOL_PLUS &&
1189 next->s != DIFF_SYMBOL_MINUS)
1190 next = NULL;
1191
1192 /* Inside a block? */
1193 if ((prev &&
1194 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1195 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1196 (next &&
1197 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1198 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1199 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1200 continue;
1201 }
1202
1203 /* Check if we are at an interesting bound: */
1204 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1205 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1206 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1207 continue;
1208 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1209 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1210 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1211 continue;
1212
1213 /*
1214 * The boundary to prev and next are not interesting,
1215 * so this line is not interesting as a whole
1216 */
1217 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1218 }
1219 }
1220
1221 static void emit_line_ws_markup(struct diff_options *o,
1222 const char *set_sign, const char *set,
1223 const char *reset,
1224 int sign_index, const char *line, int len,
1225 unsigned ws_rule, int blank_at_eof)
1226 {
1227 const char *ws = NULL;
1228 int sign = o->output_indicators[sign_index];
1229
1230 if (o->ws_error_highlight & ws_rule) {
1231 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1232 if (!*ws)
1233 ws = NULL;
1234 }
1235
1236 if (!ws && !set_sign)
1237 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1238 else if (!ws) {
1239 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1240 } else if (blank_at_eof)
1241 /* Blank line at EOF - paint '+' as well */
1242 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1243 else {
1244 /* Emit just the prefix, then the rest. */
1245 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1246 sign, "", 0);
1247 ws_check_emit(line, len, ws_rule,
1248 o->file, set, reset, ws);
1249 }
1250 }
1251
1252 static void emit_diff_symbol_from_struct(struct diff_options *o,
1253 struct emitted_diff_symbol *eds)
1254 {
1255 static const char *nneof = " No newline at end of file\n";
1256 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1257 struct strbuf sb = STRBUF_INIT;
1258
1259 enum diff_symbol s = eds->s;
1260 const char *line = eds->line;
1261 int len = eds->len;
1262 unsigned flags = eds->flags;
1263
1264 switch (s) {
1265 case DIFF_SYMBOL_NO_LF_EOF:
1266 context = diff_get_color_opt(o, DIFF_CONTEXT);
1267 reset = diff_get_color_opt(o, DIFF_RESET);
1268 putc('\n', o->file);
1269 emit_line_0(o, context, NULL, 0, reset, '\\',
1270 nneof, strlen(nneof));
1271 break;
1272 case DIFF_SYMBOL_SUBMODULE_HEADER:
1273 case DIFF_SYMBOL_SUBMODULE_ERROR:
1274 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1275 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1276 case DIFF_SYMBOL_SUMMARY:
1277 case DIFF_SYMBOL_STATS_LINE:
1278 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1279 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1280 emit_line(o, "", "", line, len);
1281 break;
1282 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1283 case DIFF_SYMBOL_CONTEXT_MARKER:
1284 context = diff_get_color_opt(o, DIFF_CONTEXT);
1285 reset = diff_get_color_opt(o, DIFF_RESET);
1286 emit_line(o, context, reset, line, len);
1287 break;
1288 case DIFF_SYMBOL_SEPARATOR:
1289 fprintf(o->file, "%s%c",
1290 diff_line_prefix(o),
1291 o->line_termination);
1292 break;
1293 case DIFF_SYMBOL_CONTEXT:
1294 set = diff_get_color_opt(o, DIFF_CONTEXT);
1295 reset = diff_get_color_opt(o, DIFF_RESET);
1296 set_sign = NULL;
1297 if (o->flags.dual_color_diffed_diffs) {
1298 char c = !len ? 0 : line[0];
1299
1300 if (c == '+')
1301 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1302 else if (c == '@')
1303 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1304 else if (c == '-')
1305 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1306 }
1307 emit_line_ws_markup(o, set_sign, set, reset,
1308 OUTPUT_INDICATOR_CONTEXT, line, len,
1309 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1310 break;
1311 case DIFF_SYMBOL_PLUS:
1312 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1313 DIFF_SYMBOL_MOVED_LINE_ALT |
1314 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1315 case DIFF_SYMBOL_MOVED_LINE |
1316 DIFF_SYMBOL_MOVED_LINE_ALT |
1317 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1318 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1319 break;
1320 case DIFF_SYMBOL_MOVED_LINE |
1321 DIFF_SYMBOL_MOVED_LINE_ALT:
1322 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1323 break;
1324 case DIFF_SYMBOL_MOVED_LINE |
1325 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1326 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1327 break;
1328 case DIFF_SYMBOL_MOVED_LINE:
1329 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1330 break;
1331 default:
1332 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1333 }
1334 reset = diff_get_color_opt(o, DIFF_RESET);
1335 if (!o->flags.dual_color_diffed_diffs)
1336 set_sign = NULL;
1337 else {
1338 char c = !len ? 0 : line[0];
1339
1340 set_sign = set;
1341 if (c == '-')
1342 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1343 else if (c == '@')
1344 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1345 else if (c == '+')
1346 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1347 else
1348 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1349 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1350 }
1351 emit_line_ws_markup(o, set_sign, set, reset,
1352 OUTPUT_INDICATOR_NEW, line, len,
1353 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1354 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1355 break;
1356 case DIFF_SYMBOL_MINUS:
1357 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1358 DIFF_SYMBOL_MOVED_LINE_ALT |
1359 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1360 case DIFF_SYMBOL_MOVED_LINE |
1361 DIFF_SYMBOL_MOVED_LINE_ALT |
1362 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1363 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1364 break;
1365 case DIFF_SYMBOL_MOVED_LINE |
1366 DIFF_SYMBOL_MOVED_LINE_ALT:
1367 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1368 break;
1369 case DIFF_SYMBOL_MOVED_LINE |
1370 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1371 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1372 break;
1373 case DIFF_SYMBOL_MOVED_LINE:
1374 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1375 break;
1376 default:
1377 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1378 }
1379 reset = diff_get_color_opt(o, DIFF_RESET);
1380 if (!o->flags.dual_color_diffed_diffs)
1381 set_sign = NULL;
1382 else {
1383 char c = !len ? 0 : line[0];
1384
1385 set_sign = set;
1386 if (c == '+')
1387 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1388 else if (c == '@')
1389 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1390 else if (c == '-')
1391 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1392 else
1393 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1394 }
1395 emit_line_ws_markup(o, set_sign, set, reset,
1396 OUTPUT_INDICATOR_OLD, line, len,
1397 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1398 break;
1399 case DIFF_SYMBOL_WORDS_PORCELAIN:
1400 context = diff_get_color_opt(o, DIFF_CONTEXT);
1401 reset = diff_get_color_opt(o, DIFF_RESET);
1402 emit_line(o, context, reset, line, len);
1403 fputs("~\n", o->file);
1404 break;
1405 case DIFF_SYMBOL_WORDS:
1406 context = diff_get_color_opt(o, DIFF_CONTEXT);
1407 reset = diff_get_color_opt(o, DIFF_RESET);
1408 /*
1409 * Skip the prefix character, if any. With
1410 * diff_suppress_blank_empty, there may be
1411 * none.
1412 */
1413 if (line[0] != '\n') {
1414 line++;
1415 len--;
1416 }
1417 emit_line(o, context, reset, line, len);
1418 break;
1419 case DIFF_SYMBOL_FILEPAIR_PLUS:
1420 meta = diff_get_color_opt(o, DIFF_METAINFO);
1421 reset = diff_get_color_opt(o, DIFF_RESET);
1422 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1423 line, reset,
1424 strchr(line, ' ') ? "\t" : "");
1425 break;
1426 case DIFF_SYMBOL_FILEPAIR_MINUS:
1427 meta = diff_get_color_opt(o, DIFF_METAINFO);
1428 reset = diff_get_color_opt(o, DIFF_RESET);
1429 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1430 line, reset,
1431 strchr(line, ' ') ? "\t" : "");
1432 break;
1433 case DIFF_SYMBOL_BINARY_FILES:
1434 case DIFF_SYMBOL_HEADER:
1435 fprintf(o->file, "%s", line);
1436 break;
1437 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1438 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1439 break;
1440 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1441 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1442 break;
1443 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1444 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1445 break;
1446 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1447 fputs(diff_line_prefix(o), o->file);
1448 fputc('\n', o->file);
1449 break;
1450 case DIFF_SYMBOL_REWRITE_DIFF:
1451 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1452 reset = diff_get_color_opt(o, DIFF_RESET);
1453 emit_line(o, fraginfo, reset, line, len);
1454 break;
1455 case DIFF_SYMBOL_SUBMODULE_ADD:
1456 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1457 reset = diff_get_color_opt(o, DIFF_RESET);
1458 emit_line(o, set, reset, line, len);
1459 break;
1460 case DIFF_SYMBOL_SUBMODULE_DEL:
1461 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1462 reset = diff_get_color_opt(o, DIFF_RESET);
1463 emit_line(o, set, reset, line, len);
1464 break;
1465 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1466 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1467 diff_line_prefix(o), line);
1468 break;
1469 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1470 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1471 diff_line_prefix(o), line);
1472 break;
1473 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1474 emit_line(o, "", "", " 0 files changed\n",
1475 strlen(" 0 files changed\n"));
1476 break;
1477 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1478 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1479 break;
1480 case DIFF_SYMBOL_WORD_DIFF:
1481 fprintf(o->file, "%.*s", len, line);
1482 break;
1483 case DIFF_SYMBOL_STAT_SEP:
1484 fputs(o->stat_sep, o->file);
1485 break;
1486 default:
1487 BUG("unknown diff symbol");
1488 }
1489 strbuf_release(&sb);
1490 }
1491
1492 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1493 const char *line, int len, unsigned flags)
1494 {
1495 struct emitted_diff_symbol e = {line, len, flags, s};
1496
1497 if (o->emitted_symbols)
1498 append_emitted_diff_symbol(o, &e);
1499 else
1500 emit_diff_symbol_from_struct(o, &e);
1501 }
1502
1503 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1504 {
1505 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1506 }
1507
1508 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1509 {
1510 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1511 }
1512
1513 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1514 {
1515 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1516 path, strlen(path), 0);
1517 }
1518
1519 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1520 {
1521 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1522 path, strlen(path), 0);
1523 }
1524
1525 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1526 {
1527 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1528 header, strlen(header), 0);
1529 }
1530
1531 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1532 {
1533 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1534 }
1535
1536 void diff_emit_submodule_pipethrough(struct diff_options *o,
1537 const char *line, int len)
1538 {
1539 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1540 }
1541
1542 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1543 {
1544 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1545 ecbdata->blank_at_eof_in_preimage &&
1546 ecbdata->blank_at_eof_in_postimage &&
1547 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1548 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1549 return 0;
1550 return ws_blank_line(line, len, ecbdata->ws_rule);
1551 }
1552
1553 static void emit_add_line(const char *reset,
1554 struct emit_callback *ecbdata,
1555 const char *line, int len)
1556 {
1557 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1558 if (new_blank_line_at_eof(ecbdata, line, len))
1559 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1560
1561 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1562 }
1563
1564 static void emit_del_line(const char *reset,
1565 struct emit_callback *ecbdata,
1566 const char *line, int len)
1567 {
1568 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1569 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1570 }
1571
1572 static void emit_context_line(const char *reset,
1573 struct emit_callback *ecbdata,
1574 const char *line, int len)
1575 {
1576 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1577 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1578 }
1579
1580 static void emit_hunk_header(struct emit_callback *ecbdata,
1581 const char *line, int len)
1582 {
1583 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1584 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1585 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1586 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1587 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1588 static const char atat[2] = { '@', '@' };
1589 const char *cp, *ep;
1590 struct strbuf msgbuf = STRBUF_INIT;
1591 int org_len = len;
1592 int i = 1;
1593
1594 /*
1595 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1596 * it always is at least 10 bytes long.
1597 */
1598 if (len < 10 ||
1599 memcmp(line, atat, 2) ||
1600 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1601 emit_diff_symbol(ecbdata->opt,
1602 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1603 return;
1604 }
1605 ep += 2; /* skip over @@ */
1606
1607 /* The hunk header in fraginfo color */
1608 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1609 strbuf_addstr(&msgbuf, reverse);
1610 strbuf_addstr(&msgbuf, frag);
1611 strbuf_add(&msgbuf, line, ep - line);
1612 strbuf_addstr(&msgbuf, reset);
1613
1614 /*
1615 * trailing "\r\n"
1616 */
1617 for ( ; i < 3; i++)
1618 if (line[len - i] == '\r' || line[len - i] == '\n')
1619 len--;
1620
1621 /* blank before the func header */
1622 for (cp = ep; ep - line < len; ep++)
1623 if (*ep != ' ' && *ep != '\t')
1624 break;
1625 if (ep != cp) {
1626 strbuf_addstr(&msgbuf, context);
1627 strbuf_add(&msgbuf, cp, ep - cp);
1628 strbuf_addstr(&msgbuf, reset);
1629 }
1630
1631 if (ep < line + len) {
1632 strbuf_addstr(&msgbuf, func);
1633 strbuf_add(&msgbuf, ep, line + len - ep);
1634 strbuf_addstr(&msgbuf, reset);
1635 }
1636
1637 strbuf_add(&msgbuf, line + len, org_len - len);
1638 strbuf_complete_line(&msgbuf);
1639 emit_diff_symbol(ecbdata->opt,
1640 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1641 strbuf_release(&msgbuf);
1642 }
1643
1644 static struct diff_tempfile *claim_diff_tempfile(void)
1645 {
1646 int i;
1647 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1648 if (!diff_temp[i].name)
1649 return diff_temp + i;
1650 BUG("diff is failing to clean up its tempfiles");
1651 }
1652
1653 static void remove_tempfile(void)
1654 {
1655 int i;
1656 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1657 if (is_tempfile_active(diff_temp[i].tempfile))
1658 delete_tempfile(&diff_temp[i].tempfile);
1659 diff_temp[i].name = NULL;
1660 }
1661 }
1662
1663 static void add_line_count(struct strbuf *out, int count)
1664 {
1665 switch (count) {
1666 case 0:
1667 strbuf_addstr(out, "0,0");
1668 break;
1669 case 1:
1670 strbuf_addstr(out, "1");
1671 break;
1672 default:
1673 strbuf_addf(out, "1,%d", count);
1674 break;
1675 }
1676 }
1677
1678 static void emit_rewrite_lines(struct emit_callback *ecb,
1679 int prefix, const char *data, int size)
1680 {
1681 const char *endp = NULL;
1682 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1683
1684 while (0 < size) {
1685 int len;
1686
1687 endp = memchr(data, '\n', size);
1688 len = endp ? (endp - data + 1) : size;
1689 if (prefix != '+') {
1690 ecb->lno_in_preimage++;
1691 emit_del_line(reset, ecb, data, len);
1692 } else {
1693 ecb->lno_in_postimage++;
1694 emit_add_line(reset, ecb, data, len);
1695 }
1696 size -= len;
1697 data += len;
1698 }
1699 if (!endp)
1700 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1701 }
1702
1703 static void emit_rewrite_diff(const char *name_a,
1704 const char *name_b,
1705 struct diff_filespec *one,
1706 struct diff_filespec *two,
1707 struct userdiff_driver *textconv_one,
1708 struct userdiff_driver *textconv_two,
1709 struct diff_options *o)
1710 {
1711 int lc_a, lc_b;
1712 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1713 const char *a_prefix, *b_prefix;
1714 char *data_one, *data_two;
1715 size_t size_one, size_two;
1716 struct emit_callback ecbdata;
1717 struct strbuf out = STRBUF_INIT;
1718
1719 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1720 a_prefix = o->b_prefix;
1721 b_prefix = o->a_prefix;
1722 } else {
1723 a_prefix = o->a_prefix;
1724 b_prefix = o->b_prefix;
1725 }
1726
1727 name_a += (*name_a == '/');
1728 name_b += (*name_b == '/');
1729
1730 strbuf_reset(&a_name);
1731 strbuf_reset(&b_name);
1732 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1733 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1734
1735 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1736 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1737
1738 memset(&ecbdata, 0, sizeof(ecbdata));
1739 ecbdata.color_diff = want_color(o->use_color);
1740 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1741 ecbdata.opt = o;
1742 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1743 mmfile_t mf1, mf2;
1744 mf1.ptr = (char *)data_one;
1745 mf2.ptr = (char *)data_two;
1746 mf1.size = size_one;
1747 mf2.size = size_two;
1748 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1749 }
1750 ecbdata.lno_in_preimage = 1;
1751 ecbdata.lno_in_postimage = 1;
1752
1753 lc_a = count_lines(data_one, size_one);
1754 lc_b = count_lines(data_two, size_two);
1755
1756 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1757 a_name.buf, a_name.len, 0);
1758 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1759 b_name.buf, b_name.len, 0);
1760
1761 strbuf_addstr(&out, "@@ -");
1762 if (!o->irreversible_delete)
1763 add_line_count(&out, lc_a);
1764 else
1765 strbuf_addstr(&out, "?,?");
1766 strbuf_addstr(&out, " +");
1767 add_line_count(&out, lc_b);
1768 strbuf_addstr(&out, " @@\n");
1769 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1770 strbuf_release(&out);
1771
1772 if (lc_a && !o->irreversible_delete)
1773 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1774 if (lc_b)
1775 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1776 if (textconv_one)
1777 free((char *)data_one);
1778 if (textconv_two)
1779 free((char *)data_two);
1780 }
1781
1782 struct diff_words_buffer {
1783 mmfile_t text;
1784 unsigned long alloc;
1785 struct diff_words_orig {
1786 const char *begin, *end;
1787 } *orig;
1788 int orig_nr, orig_alloc;
1789 };
1790
1791 static void diff_words_append(char *line, unsigned long len,
1792 struct diff_words_buffer *buffer)
1793 {
1794 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1795 line++;
1796 len--;
1797 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1798 buffer->text.size += len;
1799 buffer->text.ptr[buffer->text.size] = '\0';
1800 }
1801
1802 struct diff_words_style_elem {
1803 const char *prefix;
1804 const char *suffix;
1805 const char *color; /* NULL; filled in by the setup code if
1806 * color is enabled */
1807 };
1808
1809 struct diff_words_style {
1810 enum diff_words_type type;
1811 struct diff_words_style_elem new_word, old_word, ctx;
1812 const char *newline;
1813 };
1814
1815 static struct diff_words_style diff_words_styles[] = {
1816 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1817 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1818 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1819 };
1820
1821 struct diff_words_data {
1822 struct diff_words_buffer minus, plus;
1823 const char *current_plus;
1824 int last_minus;
1825 struct diff_options *opt;
1826 regex_t *word_regex;
1827 enum diff_words_type type;
1828 struct diff_words_style *style;
1829 };
1830
1831 static int fn_out_diff_words_write_helper(struct diff_options *o,
1832 struct diff_words_style_elem *st_el,
1833 const char *newline,
1834 size_t count, const char *buf)
1835 {
1836 int print = 0;
1837 struct strbuf sb = STRBUF_INIT;
1838
1839 while (count) {
1840 char *p = memchr(buf, '\n', count);
1841 if (print)
1842 strbuf_addstr(&sb, diff_line_prefix(o));
1843
1844 if (p != buf) {
1845 const char *reset = st_el->color && *st_el->color ?
1846 GIT_COLOR_RESET : NULL;
1847 if (st_el->color && *st_el->color)
1848 strbuf_addstr(&sb, st_el->color);
1849 strbuf_addstr(&sb, st_el->prefix);
1850 strbuf_add(&sb, buf, p ? p - buf : count);
1851 strbuf_addstr(&sb, st_el->suffix);
1852 if (reset)
1853 strbuf_addstr(&sb, reset);
1854 }
1855 if (!p)
1856 goto out;
1857
1858 strbuf_addstr(&sb, newline);
1859 count -= p + 1 - buf;
1860 buf = p + 1;
1861 print = 1;
1862 if (count) {
1863 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1864 sb.buf, sb.len, 0);
1865 strbuf_reset(&sb);
1866 }
1867 }
1868
1869 out:
1870 if (sb.len)
1871 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1872 sb.buf, sb.len, 0);
1873 strbuf_release(&sb);
1874 return 0;
1875 }
1876
1877 /*
1878 * '--color-words' algorithm can be described as:
1879 *
1880 * 1. collect the minus/plus lines of a diff hunk, divided into
1881 * minus-lines and plus-lines;
1882 *
1883 * 2. break both minus-lines and plus-lines into words and
1884 * place them into two mmfile_t with one word for each line;
1885 *
1886 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1887 *
1888 * And for the common parts of the both file, we output the plus side text.
1889 * diff_words->current_plus is used to trace the current position of the plus file
1890 * which printed. diff_words->last_minus is used to trace the last minus word
1891 * printed.
1892 *
1893 * For '--graph' to work with '--color-words', we need to output the graph prefix
1894 * on each line of color words output. Generally, there are two conditions on
1895 * which we should output the prefix.
1896 *
1897 * 1. diff_words->last_minus == 0 &&
1898 * diff_words->current_plus == diff_words->plus.text.ptr
1899 *
1900 * that is: the plus text must start as a new line, and if there is no minus
1901 * word printed, a graph prefix must be printed.
1902 *
1903 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1904 * *(diff_words->current_plus - 1) == '\n'
1905 *
1906 * that is: a graph prefix must be printed following a '\n'
1907 */
1908 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1909 {
1910 if ((diff_words->last_minus == 0 &&
1911 diff_words->current_plus == diff_words->plus.text.ptr) ||
1912 (diff_words->current_plus > diff_words->plus.text.ptr &&
1913 *(diff_words->current_plus - 1) == '\n')) {
1914 return 1;
1915 } else {
1916 return 0;
1917 }
1918 }
1919
1920 static void fn_out_diff_words_aux(void *priv,
1921 long minus_first, long minus_len,
1922 long plus_first, long plus_len,
1923 const char *func, long funclen)
1924 {
1925 struct diff_words_data *diff_words = priv;
1926 struct diff_words_style *style = diff_words->style;
1927 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1928 struct diff_options *opt = diff_words->opt;
1929 const char *line_prefix;
1930
1931 assert(opt);
1932 line_prefix = diff_line_prefix(opt);
1933
1934 /* POSIX requires that first be decremented by one if len == 0... */
1935 if (minus_len) {
1936 minus_begin = diff_words->minus.orig[minus_first].begin;
1937 minus_end =
1938 diff_words->minus.orig[minus_first + minus_len - 1].end;
1939 } else
1940 minus_begin = minus_end =
1941 diff_words->minus.orig[minus_first].end;
1942
1943 if (plus_len) {
1944 plus_begin = diff_words->plus.orig[plus_first].begin;
1945 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1946 } else
1947 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1948
1949 if (color_words_output_graph_prefix(diff_words)) {
1950 fputs(line_prefix, diff_words->opt->file);
1951 }
1952 if (diff_words->current_plus != plus_begin) {
1953 fn_out_diff_words_write_helper(diff_words->opt,
1954 &style->ctx, style->newline,
1955 plus_begin - diff_words->current_plus,
1956 diff_words->current_plus);
1957 }
1958 if (minus_begin != minus_end) {
1959 fn_out_diff_words_write_helper(diff_words->opt,
1960 &style->old_word, style->newline,
1961 minus_end - minus_begin, minus_begin);
1962 }
1963 if (plus_begin != plus_end) {
1964 fn_out_diff_words_write_helper(diff_words->opt,
1965 &style->new_word, style->newline,
1966 plus_end - plus_begin, plus_begin);
1967 }
1968
1969 diff_words->current_plus = plus_end;
1970 diff_words->last_minus = minus_first;
1971 }
1972
1973 /* This function starts looking at *begin, and returns 0 iff a word was found. */
1974 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1975 int *begin, int *end)
1976 {
1977 if (word_regex && *begin < buffer->size) {
1978 regmatch_t match[1];
1979 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1980 buffer->size - *begin, 1, match, 0)) {
1981 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1982 '\n', match[0].rm_eo - match[0].rm_so);
1983 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1984 *begin += match[0].rm_so;
1985 return *begin >= *end;
1986 }
1987 return -1;
1988 }
1989
1990 /* find the next word */
1991 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1992 (*begin)++;
1993 if (*begin >= buffer->size)
1994 return -1;
1995
1996 /* find the end of the word */
1997 *end = *begin + 1;
1998 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1999 (*end)++;
2000
2001 return 0;
2002 }
2003
2004 /*
2005 * This function splits the words in buffer->text, stores the list with
2006 * newline separator into out, and saves the offsets of the original words
2007 * in buffer->orig.
2008 */
2009 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2010 regex_t *word_regex)
2011 {
2012 int i, j;
2013 long alloc = 0;
2014
2015 out->size = 0;
2016 out->ptr = NULL;
2017
2018 /* fake an empty "0th" word */
2019 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2020 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2021 buffer->orig_nr = 1;
2022
2023 for (i = 0; i < buffer->text.size; i++) {
2024 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2025 return;
2026
2027 /* store original boundaries */
2028 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2029 buffer->orig_alloc);
2030 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2031 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2032 buffer->orig_nr++;
2033
2034 /* store one word */
2035 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2036 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2037 out->ptr[out->size + j - i] = '\n';
2038 out->size += j - i + 1;
2039
2040 i = j - 1;
2041 }
2042 }
2043
2044 /* this executes the word diff on the accumulated buffers */
2045 static void diff_words_show(struct diff_words_data *diff_words)
2046 {
2047 xpparam_t xpp;
2048 xdemitconf_t xecfg;
2049 mmfile_t minus, plus;
2050 struct diff_words_style *style = diff_words->style;
2051
2052 struct diff_options *opt = diff_words->opt;
2053 const char *line_prefix;
2054
2055 assert(opt);
2056 line_prefix = diff_line_prefix(opt);
2057
2058 /* special case: only removal */
2059 if (!diff_words->plus.text.size) {
2060 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2061 line_prefix, strlen(line_prefix), 0);
2062 fn_out_diff_words_write_helper(diff_words->opt,
2063 &style->old_word, style->newline,
2064 diff_words->minus.text.size,
2065 diff_words->minus.text.ptr);
2066 diff_words->minus.text.size = 0;
2067 return;
2068 }
2069
2070 diff_words->current_plus = diff_words->plus.text.ptr;
2071 diff_words->last_minus = 0;
2072
2073 memset(&xpp, 0, sizeof(xpp));
2074 memset(&xecfg, 0, sizeof(xecfg));
2075 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2076 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2077 xpp.flags = 0;
2078 /* as only the hunk header will be parsed, we need a 0-context */
2079 xecfg.ctxlen = 0;
2080 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2081 diff_words, &xpp, &xecfg))
2082 die("unable to generate word diff");
2083 free(minus.ptr);
2084 free(plus.ptr);
2085 if (diff_words->current_plus != diff_words->plus.text.ptr +
2086 diff_words->plus.text.size) {
2087 if (color_words_output_graph_prefix(diff_words))
2088 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2089 line_prefix, strlen(line_prefix), 0);
2090 fn_out_diff_words_write_helper(diff_words->opt,
2091 &style->ctx, style->newline,
2092 diff_words->plus.text.ptr + diff_words->plus.text.size
2093 - diff_words->current_plus, diff_words->current_plus);
2094 }
2095 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2096 }
2097
2098 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2099 static void diff_words_flush(struct emit_callback *ecbdata)
2100 {
2101 struct diff_options *wo = ecbdata->diff_words->opt;
2102
2103 if (ecbdata->diff_words->minus.text.size ||
2104 ecbdata->diff_words->plus.text.size)
2105 diff_words_show(ecbdata->diff_words);
2106
2107 if (wo->emitted_symbols) {
2108 struct diff_options *o = ecbdata->opt;
2109 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2110 int i;
2111
2112 /*
2113 * NEEDSWORK:
2114 * Instead of appending each, concat all words to a line?
2115 */
2116 for (i = 0; i < wol->nr; i++)
2117 append_emitted_diff_symbol(o, &wol->buf[i]);
2118
2119 for (i = 0; i < wol->nr; i++)
2120 free((void *)wol->buf[i].line);
2121
2122 wol->nr = 0;
2123 }
2124 }
2125
2126 static void diff_filespec_load_driver(struct diff_filespec *one,
2127 struct index_state *istate)
2128 {
2129 /* Use already-loaded driver */
2130 if (one->driver)
2131 return;
2132
2133 if (S_ISREG(one->mode))
2134 one->driver = userdiff_find_by_path(istate, one->path);
2135
2136 /* Fallback to default settings */
2137 if (!one->driver)
2138 one->driver = userdiff_find_by_name("default");
2139 }
2140
2141 static const char *userdiff_word_regex(struct diff_filespec *one,
2142 struct index_state *istate)
2143 {
2144 diff_filespec_load_driver(one, istate);
2145 return one->driver->word_regex;
2146 }
2147
2148 static void init_diff_words_data(struct emit_callback *ecbdata,
2149 struct diff_options *orig_opts,
2150 struct diff_filespec *one,
2151 struct diff_filespec *two)
2152 {
2153 int i;
2154 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2155 memcpy(o, orig_opts, sizeof(struct diff_options));
2156
2157 ecbdata->diff_words =
2158 xcalloc(1, sizeof(struct diff_words_data));
2159 ecbdata->diff_words->type = o->word_diff;
2160 ecbdata->diff_words->opt = o;
2161
2162 if (orig_opts->emitted_symbols)
2163 o->emitted_symbols =
2164 xcalloc(1, sizeof(struct emitted_diff_symbols));
2165
2166 if (!o->word_regex)
2167 o->word_regex = userdiff_word_regex(one, o->repo->index);
2168 if (!o->word_regex)
2169 o->word_regex = userdiff_word_regex(two, o->repo->index);
2170 if (!o->word_regex)
2171 o->word_regex = diff_word_regex_cfg;
2172 if (o->word_regex) {
2173 ecbdata->diff_words->word_regex = (regex_t *)
2174 xmalloc(sizeof(regex_t));
2175 if (regcomp(ecbdata->diff_words->word_regex,
2176 o->word_regex,
2177 REG_EXTENDED | REG_NEWLINE))
2178 die("invalid regular expression: %s",
2179 o->word_regex);
2180 }
2181 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2182 if (o->word_diff == diff_words_styles[i].type) {
2183 ecbdata->diff_words->style =
2184 &diff_words_styles[i];
2185 break;
2186 }
2187 }
2188 if (want_color(o->use_color)) {
2189 struct diff_words_style *st = ecbdata->diff_words->style;
2190 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2191 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2192 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2193 }
2194 }
2195
2196 static void free_diff_words_data(struct emit_callback *ecbdata)
2197 {
2198 if (ecbdata->diff_words) {
2199 diff_words_flush(ecbdata);
2200 free (ecbdata->diff_words->opt->emitted_symbols);
2201 free (ecbdata->diff_words->opt);
2202 free (ecbdata->diff_words->minus.text.ptr);
2203 free (ecbdata->diff_words->minus.orig);
2204 free (ecbdata->diff_words->plus.text.ptr);
2205 free (ecbdata->diff_words->plus.orig);
2206 if (ecbdata->diff_words->word_regex) {
2207 regfree(ecbdata->diff_words->word_regex);
2208 free(ecbdata->diff_words->word_regex);
2209 }
2210 FREE_AND_NULL(ecbdata->diff_words);
2211 }
2212 }
2213
2214 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2215 {
2216 if (want_color(diff_use_color))
2217 return diff_colors[ix];
2218 return "";
2219 }
2220
2221 const char *diff_line_prefix(struct diff_options *opt)
2222 {
2223 struct strbuf *msgbuf;
2224 if (!opt->output_prefix)
2225 return "";
2226
2227 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2228 return msgbuf->buf;
2229 }
2230
2231 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
2232 {
2233 const char *cp;
2234 unsigned long allot;
2235 size_t l = len;
2236
2237 cp = line;
2238 allot = l;
2239 while (0 < l) {
2240 (void) utf8_width(&cp, &l);
2241 if (!cp)
2242 break; /* truncated in the middle? */
2243 }
2244 return allot - l;
2245 }
2246
2247 static void find_lno(const char *line, struct emit_callback *ecbdata)
2248 {
2249 const char *p;
2250 ecbdata->lno_in_preimage = 0;
2251 ecbdata->lno_in_postimage = 0;
2252 p = strchr(line, '-');
2253 if (!p)
2254 return; /* cannot happen */
2255 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2256 p = strchr(p, '+');
2257 if (!p)
2258 return; /* cannot happen */
2259 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2260 }
2261
2262 static void fn_out_consume(void *priv, char *line, unsigned long len)
2263 {
2264 struct emit_callback *ecbdata = priv;
2265 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2266 struct diff_options *o = ecbdata->opt;
2267
2268 o->found_changes = 1;
2269
2270 if (ecbdata->header) {
2271 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2272 ecbdata->header->buf, ecbdata->header->len, 0);
2273 strbuf_reset(ecbdata->header);
2274 ecbdata->header = NULL;
2275 }
2276
2277 if (ecbdata->label_path[0]) {
2278 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2279 ecbdata->label_path[0],
2280 strlen(ecbdata->label_path[0]), 0);
2281 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2282 ecbdata->label_path[1],
2283 strlen(ecbdata->label_path[1]), 0);
2284 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2285 }
2286
2287 if (diff_suppress_blank_empty
2288 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2289 line[0] = '\n';
2290 len = 1;
2291 }
2292
2293 if (line[0] == '@') {
2294 if (ecbdata->diff_words)
2295 diff_words_flush(ecbdata);
2296 len = sane_truncate_line(ecbdata, line, len);
2297 find_lno(line, ecbdata);
2298 emit_hunk_header(ecbdata, line, len);
2299 return;
2300 }
2301
2302 if (ecbdata->diff_words) {
2303 enum diff_symbol s =
2304 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2305 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2306 if (line[0] == '-') {
2307 diff_words_append(line, len,
2308 &ecbdata->diff_words->minus);
2309 return;
2310 } else if (line[0] == '+') {
2311 diff_words_append(line, len,
2312 &ecbdata->diff_words->plus);
2313 return;
2314 } else if (starts_with(line, "\\ ")) {
2315 /*
2316 * Eat the "no newline at eof" marker as if we
2317 * saw a "+" or "-" line with nothing on it,
2318 * and return without diff_words_flush() to
2319 * defer processing. If this is the end of
2320 * preimage, more "+" lines may come after it.
2321 */
2322 return;
2323 }
2324 diff_words_flush(ecbdata);
2325 emit_diff_symbol(o, s, line, len, 0);
2326 return;
2327 }
2328
2329 switch (line[0]) {
2330 case '+':
2331 ecbdata->lno_in_postimage++;
2332 emit_add_line(reset, ecbdata, line + 1, len - 1);
2333 break;
2334 case '-':
2335 ecbdata->lno_in_preimage++;
2336 emit_del_line(reset, ecbdata, line + 1, len - 1);
2337 break;
2338 case ' ':
2339 ecbdata->lno_in_postimage++;
2340 ecbdata->lno_in_preimage++;
2341 emit_context_line(reset, ecbdata, line + 1, len - 1);
2342 break;
2343 default:
2344 /* incomplete line at the end */
2345 ecbdata->lno_in_preimage++;
2346 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2347 line, len, 0);
2348 break;
2349 }
2350 }
2351
2352 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2353 {
2354 const char *old_name = a;
2355 const char *new_name = b;
2356 int pfx_length, sfx_length;
2357 int pfx_adjust_for_slash;
2358 int len_a = strlen(a);
2359 int len_b = strlen(b);
2360 int a_midlen, b_midlen;
2361 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2362 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2363
2364 if (qlen_a || qlen_b) {
2365 quote_c_style(a, name, NULL, 0);
2366 strbuf_addstr(name, " => ");
2367 quote_c_style(b, name, NULL, 0);
2368 return;
2369 }
2370
2371 /* Find common prefix */
2372 pfx_length = 0;
2373 while (*old_name && *new_name && *old_name == *new_name) {
2374 if (*old_name == '/')
2375 pfx_length = old_name - a + 1;
2376 old_name++;
2377 new_name++;
2378 }
2379
2380 /* Find common suffix */
2381 old_name = a + len_a;
2382 new_name = b + len_b;
2383 sfx_length = 0;
2384 /*
2385 * If there is a common prefix, it must end in a slash. In
2386 * that case we let this loop run 1 into the prefix to see the
2387 * same slash.
2388 *
2389 * If there is no common prefix, we cannot do this as it would
2390 * underrun the input strings.
2391 */
2392 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2393 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2394 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2395 *old_name == *new_name) {
2396 if (*old_name == '/')
2397 sfx_length = len_a - (old_name - a);
2398 old_name--;
2399 new_name--;
2400 }
2401
2402 /*
2403 * pfx{mid-a => mid-b}sfx
2404 * {pfx-a => pfx-b}sfx
2405 * pfx{sfx-a => sfx-b}
2406 * name-a => name-b
2407 */
2408 a_midlen = len_a - pfx_length - sfx_length;
2409 b_midlen = len_b - pfx_length - sfx_length;
2410 if (a_midlen < 0)
2411 a_midlen = 0;
2412 if (b_midlen < 0)
2413 b_midlen = 0;
2414
2415 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2416 if (pfx_length + sfx_length) {
2417 strbuf_add(name, a, pfx_length);
2418 strbuf_addch(name, '{');
2419 }
2420 strbuf_add(name, a + pfx_length, a_midlen);
2421 strbuf_addstr(name, " => ");
2422 strbuf_add(name, b + pfx_length, b_midlen);
2423 if (pfx_length + sfx_length) {
2424 strbuf_addch(name, '}');
2425 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2426 }
2427 }
2428
2429 struct diffstat_t {
2430 int nr;
2431 int alloc;
2432 struct diffstat_file {
2433 char *from_name;
2434 char *name;
2435 char *print_name;
2436 const char *comments;
2437 unsigned is_unmerged:1;
2438 unsigned is_binary:1;
2439 unsigned is_renamed:1;
2440 unsigned is_interesting:1;
2441 uintmax_t added, deleted;
2442 } **files;
2443 };
2444
2445 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2446 const char *name_a,
2447 const char *name_b)
2448 {
2449 struct diffstat_file *x;
2450 x = xcalloc(1, sizeof(*x));
2451 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2452 diffstat->files[diffstat->nr++] = x;
2453 if (name_b) {
2454 x->from_name = xstrdup(name_a);
2455 x->name = xstrdup(name_b);
2456 x->is_renamed = 1;
2457 }
2458 else {
2459 x->from_name = NULL;
2460 x->name = xstrdup(name_a);
2461 }
2462 return x;
2463 }
2464
2465 static void diffstat_consume(void *priv, char *line, unsigned long len)
2466 {
2467 struct diffstat_t *diffstat = priv;
2468 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2469
2470 if (line[0] == '+')
2471 x->added++;
2472 else if (line[0] == '-')
2473 x->deleted++;
2474 }
2475
2476 const char mime_boundary_leader[] = "------------";
2477
2478 static int scale_linear(int it, int width, int max_change)
2479 {
2480 if (!it)
2481 return 0;
2482 /*
2483 * make sure that at least one '-' or '+' is printed if
2484 * there is any change to this path. The easiest way is to
2485 * scale linearly as if the alloted width is one column shorter
2486 * than it is, and then add 1 to the result.
2487 */
2488 return 1 + (it * (width - 1) / max_change);
2489 }
2490
2491 static void show_graph(struct strbuf *out, char ch, int cnt,
2492 const char *set, const char *reset)
2493 {
2494 if (cnt <= 0)
2495 return;
2496 strbuf_addstr(out, set);
2497 strbuf_addchars(out, ch, cnt);
2498 strbuf_addstr(out, reset);
2499 }
2500
2501 static void fill_print_name(struct diffstat_file *file)
2502 {
2503 struct strbuf pname = STRBUF_INIT;
2504
2505 if (file->print_name)
2506 return;
2507
2508 if (file->is_renamed)
2509 pprint_rename(&pname, file->from_name, file->name);
2510 else
2511 quote_c_style(file->name, &pname, NULL, 0);
2512
2513 if (file->comments)
2514 strbuf_addf(&pname, " (%s)", file->comments);
2515
2516 file->print_name = strbuf_detach(&pname, NULL);
2517 }
2518
2519 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2520 int files, int insertions, int deletions)
2521 {
2522 struct strbuf sb = STRBUF_INIT;
2523
2524 if (!files) {
2525 assert(insertions == 0 && deletions == 0);
2526 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2527 NULL, 0, 0);
2528 return;
2529 }
2530
2531 strbuf_addf(&sb,
2532 (files == 1) ? " %d file changed" : " %d files changed",
2533 files);
2534
2535 /*
2536 * For binary diff, the caller may want to print "x files
2537 * changed" with insertions == 0 && deletions == 0.
2538 *
2539 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2540 * is probably less confusing (i.e skip over "2 files changed
2541 * but nothing about added/removed lines? Is this a bug in Git?").
2542 */
2543 if (insertions || deletions == 0) {
2544 strbuf_addf(&sb,
2545 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2546 insertions);
2547 }
2548
2549 if (deletions || insertions == 0) {
2550 strbuf_addf(&sb,
2551 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2552 deletions);
2553 }
2554 strbuf_addch(&sb, '\n');
2555 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2556 sb.buf, sb.len, 0);
2557 strbuf_release(&sb);
2558 }
2559
2560 void print_stat_summary(FILE *fp, int files,
2561 int insertions, int deletions)
2562 {
2563 struct diff_options o;
2564 memset(&o, 0, sizeof(o));
2565 o.file = fp;
2566
2567 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2568 }
2569
2570 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2571 {
2572 int i, len, add, del, adds = 0, dels = 0;
2573 uintmax_t max_change = 0, max_len = 0;
2574 int total_files = data->nr, count;
2575 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2576 const char *reset, *add_c, *del_c;
2577 int extra_shown = 0;
2578 const char *line_prefix = diff_line_prefix(options);
2579 struct strbuf out = STRBUF_INIT;
2580
2581 if (data->nr == 0)
2582 return;
2583
2584 count = options->stat_count ? options->stat_count : data->nr;
2585
2586 reset = diff_get_color_opt(options, DIFF_RESET);
2587 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2588 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2589
2590 /*
2591 * Find the longest filename and max number of changes
2592 */
2593 for (i = 0; (i < count) && (i < data->nr); i++) {
2594 struct diffstat_file *file = data->files[i];
2595 uintmax_t change = file->added + file->deleted;
2596
2597 if (!file->is_interesting && (change == 0)) {
2598 count++; /* not shown == room for one more */
2599 continue;
2600 }
2601 fill_print_name(file);
2602 len = strlen(file->print_name);
2603 if (max_len < len)
2604 max_len = len;
2605
2606 if (file->is_unmerged) {
2607 /* "Unmerged" is 8 characters */
2608 bin_width = bin_width < 8 ? 8 : bin_width;
2609 continue;
2610 }
2611 if (file->is_binary) {
2612 /* "Bin XXX -> YYY bytes" */
2613 int w = 14 + decimal_width(file->added)
2614 + decimal_width(file->deleted);
2615 bin_width = bin_width < w ? w : bin_width;
2616 /* Display change counts aligned with "Bin" */
2617 number_width = 3;
2618 continue;
2619 }
2620
2621 if (max_change < change)
2622 max_change = change;
2623 }
2624 count = i; /* where we can stop scanning in data->files[] */
2625
2626 /*
2627 * We have width = stat_width or term_columns() columns total.
2628 * We want a maximum of min(max_len, stat_name_width) for the name part.
2629 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2630 * We also need 1 for " " and 4 + decimal_width(max_change)
2631 * for " | NNNN " and one the empty column at the end, altogether
2632 * 6 + decimal_width(max_change).
2633 *
2634 * If there's not enough space, we will use the smaller of
2635 * stat_name_width (if set) and 5/8*width for the filename,
2636 * and the rest for constant elements + graph part, but no more
2637 * than stat_graph_width for the graph part.
2638 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2639 * for the standard terminal size).
2640 *
2641 * In other words: stat_width limits the maximum width, and
2642 * stat_name_width fixes the maximum width of the filename,
2643 * and is also used to divide available columns if there
2644 * aren't enough.
2645 *
2646 * Binary files are displayed with "Bin XXX -> YYY bytes"
2647 * instead of the change count and graph. This part is treated
2648 * similarly to the graph part, except that it is not
2649 * "scaled". If total width is too small to accommodate the
2650 * guaranteed minimum width of the filename part and the
2651 * separators and this message, this message will "overflow"
2652 * making the line longer than the maximum width.
2653 */
2654
2655 if (options->stat_width == -1)
2656 width = term_columns() - strlen(line_prefix);
2657 else
2658 width = options->stat_width ? options->stat_width : 80;
2659 number_width = decimal_width(max_change) > number_width ?
2660 decimal_width(max_change) : number_width;
2661
2662 if (options->stat_graph_width == -1)
2663 options->stat_graph_width = diff_stat_graph_width;
2664
2665 /*
2666 * Guarantee 3/8*16==6 for the graph part
2667 * and 5/8*16==10 for the filename part
2668 */
2669 if (width < 16 + 6 + number_width)
2670 width = 16 + 6 + number_width;
2671
2672 /*
2673 * First assign sizes that are wanted, ignoring available width.
2674 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2675 * starting from "XXX" should fit in graph_width.
2676 */
2677 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2678 if (options->stat_graph_width &&
2679 options->stat_graph_width < graph_width)
2680 graph_width = options->stat_graph_width;
2681
2682 name_width = (options->stat_name_width > 0 &&
2683 options->stat_name_width < max_len) ?
2684 options->stat_name_width : max_len;
2685
2686 /*
2687 * Adjust adjustable widths not to exceed maximum width
2688 */
2689 if (name_width + number_width + 6 + graph_width > width) {
2690 if (graph_width > width * 3/8 - number_width - 6) {
2691 graph_width = width * 3/8 - number_width - 6;
2692 if (graph_width < 6)
2693 graph_width = 6;
2694 }
2695
2696 if (options->stat_graph_width &&
2697 graph_width > options->stat_graph_width)
2698 graph_width = options->stat_graph_width;
2699 if (name_width > width - number_width - 6 - graph_width)
2700 name_width = width - number_width - 6 - graph_width;
2701 else
2702 graph_width = width - number_width - 6 - name_width;
2703 }
2704
2705 /*
2706 * From here name_width is the width of the name area,
2707 * and graph_width is the width of the graph area.
2708 * max_change is used to scale graph properly.
2709 */
2710 for (i = 0; i < count; i++) {
2711 const char *prefix = "";
2712 struct diffstat_file *file = data->files[i];
2713 char *name = file->print_name;
2714 uintmax_t added = file->added;
2715 uintmax_t deleted = file->deleted;
2716 int name_len;
2717
2718 if (!file->is_interesting && (added + deleted == 0))
2719 continue;
2720
2721 /*
2722 * "scale" the filename
2723 */
2724 len = name_width;
2725 name_len = strlen(name);
2726 if (name_width < name_len) {
2727 char *slash;
2728 prefix = "...";
2729 len -= 3;
2730 name += name_len - len;
2731 slash = strchr(name, '/');
2732 if (slash)
2733 name = slash;
2734 }
2735
2736 if (file->is_binary) {
2737 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2738 strbuf_addf(&out, " %*s", number_width, "Bin");
2739 if (!added && !deleted) {
2740 strbuf_addch(&out, '\n');
2741 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2742 out.buf, out.len, 0);
2743 strbuf_reset(&out);
2744 continue;
2745 }
2746 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2747 del_c, deleted, reset);
2748 strbuf_addstr(&out, " -> ");
2749 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2750 add_c, added, reset);
2751 strbuf_addstr(&out, " bytes\n");
2752 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2753 out.buf, out.len, 0);
2754 strbuf_reset(&out);
2755 continue;
2756 }
2757 else if (file->is_unmerged) {
2758 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2759 strbuf_addstr(&out, " Unmerged\n");
2760 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2761 out.buf, out.len, 0);
2762 strbuf_reset(&out);
2763 continue;
2764 }
2765
2766 /*
2767 * scale the add/delete
2768 */
2769 add = added;
2770 del = deleted;
2771
2772 if (graph_width <= max_change) {
2773 int total = scale_linear(add + del, graph_width, max_change);
2774 if (total < 2 && add && del)
2775 /* width >= 2 due to the sanity check */
2776 total = 2;
2777 if (add < del) {
2778 add = scale_linear(add, graph_width, max_change);
2779 del = total - add;
2780 } else {
2781 del = scale_linear(del, graph_width, max_change);
2782 add = total - del;
2783 }
2784 }
2785 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2786 strbuf_addf(&out, " %*"PRIuMAX"%s",
2787 number_width, added + deleted,
2788 added + deleted ? " " : "");
2789 show_graph(&out, '+', add, add_c, reset);
2790 show_graph(&out, '-', del, del_c, reset);
2791 strbuf_addch(&out, '\n');
2792 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2793 out.buf, out.len, 0);
2794 strbuf_reset(&out);
2795 }
2796
2797 for (i = 0; i < data->nr; i++) {
2798 struct diffstat_file *file = data->files[i];
2799 uintmax_t added = file->added;
2800 uintmax_t deleted = file->deleted;
2801
2802 if (file->is_unmerged ||
2803 (!file->is_interesting && (added + deleted == 0))) {
2804 total_files--;
2805 continue;
2806 }
2807
2808 if (!file->is_binary) {
2809 adds += added;
2810 dels += deleted;
2811 }
2812 if (i < count)
2813 continue;
2814 if (!extra_shown)
2815 emit_diff_symbol(options,
2816 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2817 NULL, 0, 0);
2818 extra_shown = 1;
2819 }
2820
2821 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2822 strbuf_release(&out);
2823 }
2824
2825 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2826 {
2827 int i, adds = 0, dels = 0, total_files = data->nr;
2828
2829 if (data->nr == 0)
2830 return;
2831
2832 for (i = 0; i < data->nr; i++) {
2833 int added = data->files[i]->added;
2834 int deleted = data->files[i]->deleted;
2835
2836 if (data->files[i]->is_unmerged ||
2837 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2838 total_files--;
2839 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2840 adds += added;
2841 dels += deleted;
2842 }
2843 }
2844 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2845 }
2846
2847 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2848 {
2849 int i;
2850
2851 if (data->nr == 0)
2852 return;
2853
2854 for (i = 0; i < data->nr; i++) {
2855 struct diffstat_file *file = data->files[i];
2856
2857 fprintf(options->file, "%s", diff_line_prefix(options));
2858
2859 if (file->is_binary)
2860 fprintf(options->file, "-\t-\t");
2861 else
2862 fprintf(options->file,
2863 "%"PRIuMAX"\t%"PRIuMAX"\t",
2864 file->added, file->deleted);
2865 if (options->line_termination) {
2866 fill_print_name(file);
2867 if (!file->is_renamed)
2868 write_name_quoted(file->name, options->file,
2869 options->line_termination);
2870 else {
2871 fputs(file->print_name, options->file);
2872 putc(options->line_termination, options->file);
2873 }
2874 } else {
2875 if (file->is_renamed) {
2876 putc('\0', options->file);
2877 write_name_quoted(file->from_name, options->file, '\0');
2878 }
2879 write_name_quoted(file->name, options->file, '\0');
2880 }
2881 }
2882 }
2883
2884 struct dirstat_file {
2885 const char *name;
2886 unsigned long changed;
2887 };
2888
2889 struct dirstat_dir {
2890 struct dirstat_file *files;
2891 int alloc, nr, permille, cumulative;
2892 };
2893
2894 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2895 unsigned long changed, const char *base, int baselen)
2896 {
2897 unsigned long sum_changes = 0;
2898 unsigned int sources = 0;
2899 const char *line_prefix = diff_line_prefix(opt);
2900
2901 while (dir->nr) {
2902 struct dirstat_file *f = dir->files;
2903 int namelen = strlen(f->name);
2904 unsigned long changes;
2905 char *slash;
2906
2907 if (namelen < baselen)
2908 break;
2909 if (memcmp(f->name, base, baselen))
2910 break;
2911 slash = strchr(f->name + baselen, '/');
2912 if (slash) {
2913 int newbaselen = slash + 1 - f->name;
2914 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2915 sources++;
2916 } else {
2917 changes = f->changed;
2918 dir->files++;
2919 dir->nr--;
2920 sources += 2;
2921 }
2922 sum_changes += changes;
2923 }
2924
2925 /*
2926 * We don't report dirstat's for
2927 * - the top level
2928 * - or cases where everything came from a single directory
2929 * under this directory (sources == 1).
2930 */
2931 if (baselen && sources != 1) {
2932 if (sum_changes) {
2933 int permille = sum_changes * 1000 / changed;
2934 if (permille >= dir->permille) {
2935 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2936 permille / 10, permille % 10, baselen, base);
2937 if (!dir->cumulative)
2938 return 0;
2939 }
2940 }
2941 }
2942 return sum_changes;
2943 }
2944
2945 static int dirstat_compare(const void *_a, const void *_b)
2946 {
2947 const struct dirstat_file *a = _a;
2948 const struct dirstat_file *b = _b;
2949 return strcmp(a->name, b->name);
2950 }
2951
2952 static void show_dirstat(struct diff_options *options)
2953 {
2954 int i;
2955 unsigned long changed;
2956 struct dirstat_dir dir;
2957 struct diff_queue_struct *q = &diff_queued_diff;
2958
2959 dir.files = NULL;
2960 dir.alloc = 0;
2961 dir.nr = 0;
2962 dir.permille = options->dirstat_permille;
2963 dir.cumulative = options->flags.dirstat_cumulative;
2964
2965 changed = 0;
2966 for (i = 0; i < q->nr; i++) {
2967 struct diff_filepair *p = q->queue[i];
2968 const char *name;
2969 unsigned long copied, added, damage;
2970
2971 name = p->two->path ? p->two->path : p->one->path;
2972
2973 if (p->one->oid_valid && p->two->oid_valid &&
2974 oideq(&p->one->oid, &p->two->oid)) {
2975 /*
2976 * The SHA1 has not changed, so pre-/post-content is
2977 * identical. We can therefore skip looking at the
2978 * file contents altogether.
2979 */
2980 damage = 0;
2981 goto found_damage;
2982 }
2983
2984 if (options->flags.dirstat_by_file) {
2985 /*
2986 * In --dirstat-by-file mode, we don't really need to
2987 * look at the actual file contents at all.
2988 * The fact that the SHA1 changed is enough for us to
2989 * add this file to the list of results
2990 * (with each file contributing equal damage).
2991 */
2992 damage = 1;
2993 goto found_damage;
2994 }
2995
2996 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2997 diff_populate_filespec(options->repo, p->one, 0);
2998 diff_populate_filespec(options->repo, p->two, 0);
2999 diffcore_count_changes(options->repo,
3000 p->one, p->two, NULL, NULL,
3001 &copied, &added);
3002 diff_free_filespec_data(p->one);
3003 diff_free_filespec_data(p->two);
3004 } else if (DIFF_FILE_VALID(p->one)) {
3005 diff_populate_filespec(options->repo, p->one, CHECK_SIZE_ONLY);
3006 copied = added = 0;
3007 diff_free_filespec_data(p->one);
3008 } else if (DIFF_FILE_VALID(p->two)) {
3009 diff_populate_filespec(options->repo, p->two, CHECK_SIZE_ONLY);
3010 copied = 0;
3011 added = p->two->size;
3012 diff_free_filespec_data(p->two);
3013 } else
3014 continue;
3015
3016 /*
3017 * Original minus copied is the removed material,
3018 * added is the new material. They are both damages
3019 * made to the preimage.
3020 * If the resulting damage is zero, we know that
3021 * diffcore_count_changes() considers the two entries to
3022 * be identical, but since the oid changed, we
3023 * know that there must have been _some_ kind of change,
3024 * so we force all entries to have damage > 0.
3025 */
3026 damage = (p->one->size - copied) + added;
3027 if (!damage)
3028 damage = 1;
3029
3030 found_damage:
3031 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3032 dir.files[dir.nr].name = name;
3033 dir.files[dir.nr].changed = damage;
3034 changed += damage;
3035 dir.nr++;
3036 }
3037
3038 /* This can happen even with many files, if everything was renames */
3039 if (!changed)
3040 return;
3041
3042 /* Show all directories with more than x% of the changes */
3043 QSORT(dir.files, dir.nr, dirstat_compare);
3044 gather_dirstat(options, &dir, changed, "", 0);
3045 }
3046
3047 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3048 {
3049 int i;
3050 unsigned long changed;
3051 struct dirstat_dir dir;
3052
3053 if (data->nr == 0)
3054 return;
3055
3056 dir.files = NULL;
3057 dir.alloc = 0;
3058 dir.nr = 0;
3059 dir.permille = options->dirstat_permille;
3060 dir.cumulative = options->flags.dirstat_cumulative;
3061
3062 changed = 0;
3063 for (i = 0; i < data->nr; i++) {
3064 struct diffstat_file *file = data->files[i];
3065 unsigned long damage = file->added + file->deleted;
3066 if (file->is_binary)
3067 /*
3068 * binary files counts bytes, not lines. Must find some
3069 * way to normalize binary bytes vs. textual lines.
3070 * The following heuristic assumes that there are 64
3071 * bytes per "line".
3072 * This is stupid and ugly, but very cheap...
3073 */
3074 damage = DIV_ROUND_UP(damage, 64);
3075 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3076 dir.files[dir.nr].name = file->name;
3077 dir.files[dir.nr].changed = damage;
3078 changed += damage;
3079 dir.nr++;
3080 }
3081
3082 /* This can happen even with many files, if everything was renames */
3083 if (!changed)
3084 return;
3085
3086 /* Show all directories with more than x% of the changes */
3087 QSORT(dir.files, dir.nr, dirstat_compare);
3088 gather_dirstat(options, &dir, changed, "", 0);
3089 }
3090
3091 static void free_diffstat_info(struct diffstat_t *diffstat)
3092 {
3093 int i;
3094 for (i = 0; i < diffstat->nr; i++) {
3095 struct diffstat_file *f = diffstat->files[i];
3096 free(f->print_name);
3097 free(f->name);
3098 free(f->from_name);
3099 free(f);
3100 }
3101 free(diffstat->files);
3102 }
3103
3104 struct checkdiff_t {
3105 const char *filename;
3106 int lineno;
3107 int conflict_marker_size;
3108 struct diff_options *o;
3109 unsigned ws_rule;
3110 unsigned status;
3111 };
3112
3113 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3114 {
3115 char firstchar;
3116 int cnt;
3117
3118 if (len < marker_size + 1)
3119 return 0;
3120 firstchar = line[0];
3121 switch (firstchar) {
3122 case '=': case '>': case '<': case '|':
3123 break;
3124 default:
3125 return 0;
3126 }
3127 for (cnt = 1; cnt < marker_size; cnt++)
3128 if (line[cnt] != firstchar)
3129 return 0;
3130 /* line[1] thru line[marker_size-1] are same as firstchar */
3131 if (len < marker_size + 1 || !isspace(line[marker_size]))
3132 return 0;
3133 return 1;
3134 }
3135
3136 static void checkdiff_consume_hunk(void *priv,
3137 long ob, long on, long nb, long nn,
3138 const char *func, long funclen)
3139
3140 {
3141 struct checkdiff_t *data = priv;
3142 data->lineno = nb - 1;
3143 }
3144
3145 static void checkdiff_consume(void *priv, char *line, unsigned long len)
3146 {
3147 struct checkdiff_t *data = priv;
3148 int marker_size = data->conflict_marker_size;
3149 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3150 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3151 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3152 char *err;
3153 const char *line_prefix;
3154
3155 assert(data->o);
3156 line_prefix = diff_line_prefix(data->o);
3157
3158 if (line[0] == '+') {
3159 unsigned bad;
3160 data->lineno++;
3161 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3162 data->status |= 1;
3163 fprintf(data->o->file,
3164 "%s%s:%d: leftover conflict marker\n",
3165 line_prefix, data->filename, data->lineno);
3166 }
3167 bad = ws_check(line + 1, len - 1, data->ws_rule);
3168 if (!bad)
3169 return;
3170 data->status |= bad;
3171 err = whitespace_error_string(bad);
3172 fprintf(data->o->file, "%s%s:%d: %s.\n",
3173 line_prefix, data->filename, data->lineno, err);
3174 free(err);
3175 emit_line(data->o, set, reset, line, 1);
3176 ws_check_emit(line + 1, len - 1, data->ws_rule,
3177 data->o->file, set, reset, ws);
3178 } else if (line[0] == ' ') {
3179 data->lineno++;
3180 }
3181 }
3182
3183 static unsigned char *deflate_it(char *data,
3184 unsigned long size,
3185 unsigned long *result_size)
3186 {
3187 int bound;
3188 unsigned char *deflated;
3189 git_zstream stream;
3190
3191 git_deflate_init(&stream, zlib_compression_level);
3192 bound = git_deflate_bound(&stream, size);
3193 deflated = xmalloc(bound);
3194 stream.next_out = deflated;
3195 stream.avail_out = bound;
3196
3197 stream.next_in = (unsigned char *)data;
3198 stream.avail_in = size;
3199 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3200 ; /* nothing */
3201 git_deflate_end(&stream);
3202 *result_size = stream.total_out;
3203 return deflated;
3204 }
3205
3206 static void emit_binary_diff_body(struct diff_options *o,
3207 mmfile_t *one, mmfile_t *two)
3208 {
3209 void *cp;
3210 void *delta;
3211 void *deflated;
3212 void *data;
3213 unsigned long orig_size;
3214 unsigned long delta_size;
3215 unsigned long deflate_size;
3216 unsigned long data_size;
3217
3218 /* We could do deflated delta, or we could do just deflated two,
3219 * whichever is smaller.
3220 */
3221 delta = NULL;
3222 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3223 if (one->size && two->size) {
3224 delta = diff_delta(one->ptr, one->size,
3225 two->ptr, two->size,
3226 &delta_size, deflate_size);
3227 if (delta) {
3228 void *to_free = delta;
3229 orig_size = delta_size;
3230 delta = deflate_it(delta, delta_size, &delta_size);
3231 free(to_free);
3232 }
3233 }
3234
3235 if (delta && delta_size < deflate_size) {
3236 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3237 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3238 s, strlen(s), 0);
3239 free(s);
3240 free(deflated);
3241 data = delta;
3242 data_size = delta_size;
3243 } else {
3244 char *s = xstrfmt("%lu", two->size);
3245 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3246 s, strlen(s), 0);
3247 free(s);
3248 free(delta);
3249 data = deflated;
3250 data_size = deflate_size;
3251 }
3252
3253 /* emit data encoded in base85 */
3254 cp = data;
3255 while (data_size) {
3256 int len;
3257 int bytes = (52 < data_size) ? 52 : data_size;
3258 char line[71];
3259 data_size -= bytes;
3260 if (bytes <= 26)
3261 line[0] = bytes + 'A' - 1;
3262 else
3263 line[0] = bytes - 26 + 'a' - 1;
3264 encode_85(line + 1, cp, bytes);
3265 cp = (char *) cp + bytes;
3266
3267 len = strlen(line);
3268 line[len++] = '\n';
3269 line[len] = '\0';
3270
3271 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3272 line, len, 0);
3273 }
3274 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3275 free(data);
3276 }
3277
3278 static void emit_binary_diff(struct diff_options *o,
3279 mmfile_t *one, mmfile_t *two)
3280 {
3281 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3282 emit_binary_diff_body(o, one, two);
3283 emit_binary_diff_body(o, two, one);
3284 }
3285
3286 int diff_filespec_is_binary(struct repository *r,
3287 struct diff_filespec *one)
3288 {
3289 if (one->is_binary == -1) {
3290 diff_filespec_load_driver(one, r->index);
3291 if (one->driver->binary != -1)
3292 one->is_binary = one->driver->binary;
3293 else {
3294 if (!one->data && DIFF_FILE_VALID(one))
3295 diff_populate_filespec(r, one, CHECK_BINARY);
3296 if (one->is_binary == -1 && one->data)
3297 one->is_binary = buffer_is_binary(one->data,
3298 one->size);
3299 if (one->is_binary == -1)
3300 one->is_binary = 0;
3301 }
3302 }
3303 return one->is_binary;
3304 }
3305
3306 static const struct userdiff_funcname *
3307 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3308 {
3309 diff_filespec_load_driver(one, o->repo->index);
3310 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3311 }
3312
3313 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3314 {
3315 if (!options->a_prefix)
3316 options->a_prefix = a;
3317 if (!options->b_prefix)
3318 options->b_prefix = b;
3319 }
3320
3321 struct userdiff_driver *get_textconv(struct repository *r,
3322 struct diff_filespec *one)
3323 {
3324 if (!DIFF_FILE_VALID(one))
3325 return NULL;
3326
3327 diff_filespec_load_driver(one, r->index);
3328 return userdiff_get_textconv(r, one->driver);
3329 }
3330
3331 static void builtin_diff(const char *name_a,
3332 const char *name_b,
3333 struct diff_filespec *one,
3334 struct diff_filespec *two,
3335 const char *xfrm_msg,
3336 int must_show_header,
3337 struct diff_options *o,
3338 int complete_rewrite)
3339 {
3340 mmfile_t mf1, mf2;
3341 const char *lbl[2];
3342 char *a_one, *b_two;
3343 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3344 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3345 const char *a_prefix, *b_prefix;
3346 struct userdiff_driver *textconv_one = NULL;
3347 struct userdiff_driver *textconv_two = NULL;
3348 struct strbuf header = STRBUF_INIT;
3349 const char *line_prefix = diff_line_prefix(o);
3350
3351 diff_set_mnemonic_prefix(o, "a/", "b/");
3352 if (o->flags.reverse_diff) {
3353 a_prefix = o->b_prefix;
3354 b_prefix = o->a_prefix;
3355 } else {
3356 a_prefix = o->a_prefix;
3357 b_prefix = o->b_prefix;
3358 }
3359
3360 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3361 (!one->mode || S_ISGITLINK(one->mode)) &&
3362 (!two->mode || S_ISGITLINK(two->mode))) {
3363 show_submodule_summary(o, one->path ? one->path : two->path,
3364 &one->oid, &two->oid,
3365 two->dirty_submodule);
3366 return;
3367 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3368 (!one->mode || S_ISGITLINK(one->mode)) &&
3369 (!two->mode || S_ISGITLINK(two->mode))) {
3370 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3371 &one->oid, &two->oid,
3372 two->dirty_submodule);
3373 return;
3374 }
3375
3376 if (o->flags.allow_textconv) {
3377 textconv_one = get_textconv(o->repo, one);
3378 textconv_two = get_textconv(o->repo, two);
3379 }
3380
3381 /* Never use a non-valid filename anywhere if at all possible */
3382 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3383 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3384
3385 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3386 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3387 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3388 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3389 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3390 if (lbl[0][0] == '/') {
3391 /* /dev/null */
3392 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3393 if (xfrm_msg)
3394 strbuf_addstr(&header, xfrm_msg);
3395 must_show_header = 1;
3396 }
3397 else if (lbl[1][0] == '/') {
3398 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3399 if (xfrm_msg)
3400 strbuf_addstr(&header, xfrm_msg);
3401 must_show_header = 1;
3402 }
3403 else {
3404 if (one->mode != two->mode) {
3405 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3406 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3407 must_show_header = 1;
3408 }
3409 if (xfrm_msg)
3410 strbuf_addstr(&header, xfrm_msg);
3411
3412 /*
3413 * we do not run diff between different kind
3414 * of objects.
3415 */
3416 if ((one->mode ^ two->mode) & S_IFMT)
3417 goto free_ab_and_return;
3418 if (complete_rewrite &&
3419 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3420 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3421 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3422 header.buf, header.len, 0);
3423 strbuf_reset(&header);
3424 emit_rewrite_diff(name_a, name_b, one, two,
3425 textconv_one, textconv_two, o);
3426 o->found_changes = 1;
3427 goto free_ab_and_return;
3428 }
3429 }
3430
3431 if (o->irreversible_delete && lbl[1][0] == '/') {
3432 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3433 header.len, 0);
3434 strbuf_reset(&header);
3435 goto free_ab_and_return;
3436 } else if (!o->flags.text &&
3437 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3438 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3439 struct strbuf sb = STRBUF_INIT;
3440 if (!one->data && !two->data &&
3441 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3442 !o->flags.binary) {
3443 if (oideq(&one->oid, &two->oid)) {
3444 if (must_show_header)
3445 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3446 header.buf, header.len,
3447 0);
3448 goto free_ab_and_return;
3449 }
3450 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3451 header.buf, header.len, 0);
3452 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3453 diff_line_prefix(o), lbl[0], lbl[1]);
3454 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3455 sb.buf, sb.len, 0);
3456 strbuf_release(&sb);
3457 goto free_ab_and_return;
3458 }
3459 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3460 fill_mmfile(o->repo, &mf2, two) < 0)
3461 die("unable to read files to diff");
3462 /* Quite common confusing case */
3463 if (mf1.size == mf2.size &&
3464 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3465 if (must_show_header)
3466 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3467 header.buf, header.len, 0);
3468 goto free_ab_and_return;
3469 }
3470 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3471 strbuf_reset(&header);
3472 if (o->flags.binary)
3473 emit_binary_diff(o, &mf1, &mf2);
3474 else {
3475 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3476 diff_line_prefix(o), lbl[0], lbl[1]);
3477 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3478 sb.buf, sb.len, 0);
3479 strbuf_release(&sb);
3480 }
3481 o->found_changes = 1;
3482 } else {
3483 /* Crazy xdl interfaces.. */
3484 const char *diffopts = getenv("GIT_DIFF_OPTS");
3485 const char *v;
3486 xpparam_t xpp;
3487 xdemitconf_t xecfg;
3488 struct emit_callback ecbdata;
3489 const struct userdiff_funcname *pe;
3490
3491 if (must_show_header) {
3492 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3493 header.buf, header.len, 0);
3494 strbuf_reset(&header);
3495 }
3496
3497 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3498 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3499
3500 pe = diff_funcname_pattern(o, one);
3501 if (!pe)
3502 pe = diff_funcname_pattern(o, two);
3503
3504 memset(&xpp, 0, sizeof(xpp));
3505 memset(&xecfg, 0, sizeof(xecfg));
3506 memset(&ecbdata, 0, sizeof(ecbdata));
3507 if (o->flags.suppress_diff_headers)
3508 lbl[0] = NULL;
3509 ecbdata.label_path = lbl;
3510 ecbdata.color_diff = want_color(o->use_color);
3511 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3512 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3513 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3514 ecbdata.opt = o;
3515 if (header.len && !o->flags.suppress_diff_headers)
3516 ecbdata.header = &header;
3517 xpp.flags = o->xdl_opts;
3518 xpp.anchors = o->anchors;
3519 xpp.anchors_nr = o->anchors_nr;
3520 xecfg.ctxlen = o->context;
3521 xecfg.interhunkctxlen = o->interhunkcontext;
3522 xecfg.flags = XDL_EMIT_FUNCNAMES;
3523 if (o->flags.funccontext)
3524 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3525 if (pe)
3526 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3527 if (!diffopts)
3528 ;
3529 else if (skip_prefix(diffopts, "--unified=", &v))
3530 xecfg.ctxlen = strtoul(v, NULL, 10);
3531 else if (skip_prefix(diffopts, "-u", &v))
3532 xecfg.ctxlen = strtoul(v, NULL, 10);
3533 if (o->word_diff)
3534 init_diff_words_data(&ecbdata, o, one, two);
3535 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3536 &ecbdata, &xpp, &xecfg))
3537 die("unable to generate diff for %s", one->path);
3538 if (o->word_diff)
3539 free_diff_words_data(&ecbdata);
3540 if (textconv_one)
3541 free(mf1.ptr);
3542 if (textconv_two)
3543 free(mf2.ptr);
3544 xdiff_clear_find_func(&xecfg);
3545 }
3546
3547 free_ab_and_return:
3548 strbuf_release(&header);
3549 diff_free_filespec_data(one);
3550 diff_free_filespec_data(two);
3551 free(a_one);
3552 free(b_two);
3553 return;
3554 }
3555
3556 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3557 {
3558 if (!is_renamed) {
3559 if (p->status == DIFF_STATUS_ADDED) {
3560 if (S_ISLNK(p->two->mode))
3561 return "new +l";
3562 else if ((p->two->mode & 0777) == 0755)
3563 return "new +x";
3564 else
3565 return "new";
3566 } else if (p->status == DIFF_STATUS_DELETED)
3567 return "gone";
3568 }
3569 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3570 return "mode -l";
3571 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3572 return "mode +l";
3573 else if ((p->one->mode & 0777) == 0644 &&
3574 (p->two->mode & 0777) == 0755)
3575 return "mode +x";
3576 else if ((p->one->mode & 0777) == 0755 &&
3577 (p->two->mode & 0777) == 0644)
3578 return "mode -x";
3579 return NULL;
3580 }
3581
3582 static void builtin_diffstat(const char *name_a, const char *name_b,
3583 struct diff_filespec *one,
3584 struct diff_filespec *two,
3585 struct diffstat_t *diffstat,
3586 struct diff_options *o,
3587 struct diff_filepair *p)
3588 {
3589 mmfile_t mf1, mf2;
3590 struct diffstat_file *data;
3591 int same_contents;
3592 int complete_rewrite = 0;
3593
3594 if (!DIFF_PAIR_UNMERGED(p)) {
3595 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3596 complete_rewrite = 1;
3597 }
3598
3599 data = diffstat_add(diffstat, name_a, name_b);
3600 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3601 if (o->flags.stat_with_summary)
3602 data->comments = get_compact_summary(p, data->is_renamed);
3603
3604 if (!one || !two) {
3605 data->is_unmerged = 1;
3606 return;
3607 }
3608
3609 same_contents = oideq(&one->oid, &two->oid);
3610
3611 if (diff_filespec_is_binary(o->repo, one) ||
3612 diff_filespec_is_binary(o->repo, two)) {
3613 data->is_binary = 1;
3614 if (same_contents) {
3615 data->added = 0;
3616 data->deleted = 0;
3617 } else {
3618 data->added = diff_filespec_size(o->repo, two);
3619 data->deleted = diff_filespec_size(o->repo, one);
3620 }
3621 }
3622
3623 else if (complete_rewrite) {
3624 diff_populate_filespec(o->repo, one, 0);
3625 diff_populate_filespec(o->repo, two, 0);
3626 data->deleted = count_lines(one->data, one->size);
3627 data->added = count_lines(two->data, two->size);
3628 }
3629
3630 else if (!same_contents) {
3631 /* Crazy xdl interfaces.. */
3632 xpparam_t xpp;
3633 xdemitconf_t xecfg;
3634
3635 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3636 fill_mmfile(o->repo, &mf2, two) < 0)
3637 die("unable to read files to diff");
3638
3639 memset(&xpp, 0, sizeof(xpp));
3640 memset(&xecfg, 0, sizeof(xecfg));
3641 xpp.flags = o->xdl_opts;
3642 xpp.anchors = o->anchors;
3643 xpp.anchors_nr = o->anchors_nr;
3644 xecfg.ctxlen = o->context;
3645 xecfg.interhunkctxlen = o->interhunkcontext;
3646 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
3647 diffstat_consume, diffstat, &xpp, &xecfg))
3648 die("unable to generate diffstat for %s", one->path);
3649 }
3650
3651 diff_free_filespec_data(one);
3652 diff_free_filespec_data(two);
3653 }
3654
3655 static void builtin_checkdiff(const char *name_a, const char *name_b,
3656 const char *attr_path,
3657 struct diff_filespec *one,
3658 struct diff_filespec *two,
3659 struct diff_options *o)
3660 {
3661 mmfile_t mf1, mf2;
3662 struct checkdiff_t data;
3663
3664 if (!two)
3665 return;
3666
3667 memset(&data, 0, sizeof(data));
3668 data.filename = name_b ? name_b : name_a;
3669 data.lineno = 0;
3670 data.o = o;
3671 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3672 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3673
3674 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3675 fill_mmfile(o->repo, &mf2, two) < 0)
3676 die("unable to read files to diff");
3677
3678 /*
3679 * All the other codepaths check both sides, but not checking
3680 * the "old" side here is deliberate. We are checking the newly
3681 * introduced changes, and as long as the "new" side is text, we
3682 * can and should check what it introduces.
3683 */
3684 if (diff_filespec_is_binary(o->repo, two))
3685 goto free_and_return;
3686 else {
3687 /* Crazy xdl interfaces.. */
3688 xpparam_t xpp;
3689 xdemitconf_t xecfg;
3690
3691 memset(&xpp, 0, sizeof(xpp));
3692 memset(&xecfg, 0, sizeof(xecfg));
3693 xecfg.ctxlen = 1; /* at least one context line */
3694 xpp.flags = 0;
3695 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3696 checkdiff_consume, &data,
3697 &xpp, &xecfg))
3698 die("unable to generate checkdiff for %s", one->path);
3699
3700 if (data.ws_rule & WS_BLANK_AT_EOF) {
3701 struct emit_callback ecbdata;
3702 int blank_at_eof;
3703
3704 ecbdata.ws_rule = data.ws_rule;
3705 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3706 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3707
3708 if (blank_at_eof) {
3709 static char *err;
3710 if (!err)
3711 err = whitespace_error_string(WS_BLANK_AT_EOF);
3712 fprintf(o->file, "%s:%d: %s.\n",
3713 data.filename, blank_at_eof, err);
3714 data.status = 1; /* report errors */
3715 }
3716 }
3717 }
3718 free_and_return:
3719 diff_free_filespec_data(one);
3720 diff_free_filespec_data(two);
3721 if (data.status)
3722 o->flags.check_failed = 1;
3723 }
3724
3725 struct diff_filespec *alloc_filespec(const char *path)
3726 {
3727 struct diff_filespec *spec;
3728
3729 FLEXPTR_ALLOC_STR(spec, path, path);
3730 spec->count = 1;
3731 spec->is_binary = -1;
3732 return spec;
3733 }
3734
3735 void free_filespec(struct diff_filespec *spec)
3736 {
3737 if (!--spec->count) {
3738 diff_free_filespec_data(spec);
3739 free(spec);
3740 }
3741 }
3742
3743 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3744 int oid_valid, unsigned short mode)
3745 {
3746 if (mode) {
3747 spec->mode = canon_mode(mode);
3748 oidcpy(&spec->oid, oid);
3749 spec->oid_valid = oid_valid;
3750 }
3751 }
3752
3753 /*
3754 * Given a name and sha1 pair, if the index tells us the file in
3755 * the work tree has that object contents, return true, so that
3756 * prepare_temp_file() does not have to inflate and extract.
3757 */
3758 static int reuse_worktree_file(struct index_state *istate,
3759 const char *name,
3760 const struct object_id *oid,
3761 int want_file)
3762 {
3763 const struct cache_entry *ce;
3764 struct stat st;
3765 int pos, len;
3766
3767 /*
3768 * We do not read the cache ourselves here, because the
3769 * benchmark with my previous version that always reads cache
3770 * shows that it makes things worse for diff-tree comparing
3771 * two linux-2.6 kernel trees in an already checked out work
3772 * tree. This is because most diff-tree comparisons deal with
3773 * only a small number of files, while reading the cache is
3774 * expensive for a large project, and its cost outweighs the
3775 * savings we get by not inflating the object to a temporary
3776 * file. Practically, this code only helps when we are used
3777 * by diff-cache --cached, which does read the cache before
3778 * calling us.
3779 */
3780 if (!istate->cache)
3781 return 0;
3782
3783 /* We want to avoid the working directory if our caller
3784 * doesn't need the data in a normal file, this system
3785 * is rather slow with its stat/open/mmap/close syscalls,
3786 * and the object is contained in a pack file. The pack
3787 * is probably already open and will be faster to obtain
3788 * the data through than the working directory. Loose
3789 * objects however would tend to be slower as they need
3790 * to be individually opened and inflated.
3791 */
3792 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3793 return 0;
3794
3795 /*
3796 * Similarly, if we'd have to convert the file contents anyway, that
3797 * makes the optimization not worthwhile.
3798 */
3799 if (!want_file && would_convert_to_git(istate, name))
3800 return 0;
3801
3802 len = strlen(name);
3803 pos = index_name_pos(istate, name, len);
3804 if (pos < 0)
3805 return 0;
3806 ce = istate->cache[pos];
3807
3808 /*
3809 * This is not the sha1 we are looking for, or
3810 * unreusable because it is not a regular file.
3811 */
3812 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3813 return 0;
3814
3815 /*
3816 * If ce is marked as "assume unchanged", there is no
3817 * guarantee that work tree matches what we are looking for.
3818 */
3819 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3820 return 0;
3821
3822 /*
3823 * If ce matches the file in the work tree, we can reuse it.
3824 */
3825 if (ce_uptodate(ce) ||
3826 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
3827 return 1;
3828
3829 return 0;
3830 }
3831
3832 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3833 {
3834 struct strbuf buf = STRBUF_INIT;
3835 char *dirty = "";
3836
3837 /* Are we looking at the work tree? */
3838 if (s->dirty_submodule)
3839 dirty = "-dirty";
3840
3841 strbuf_addf(&buf, "Subproject commit %s%s\n",
3842 oid_to_hex(&s->oid), dirty);
3843 s->size = buf.len;
3844 if (size_only) {
3845 s->data = NULL;
3846 strbuf_release(&buf);
3847 } else {
3848 s->data = strbuf_detach(&buf, NULL);
3849 s->should_free = 1;
3850 }
3851 return 0;
3852 }
3853
3854 /*
3855 * While doing rename detection and pickaxe operation, we may need to
3856 * grab the data for the blob (or file) for our own in-core comparison.
3857 * diff_filespec has data and size fields for this purpose.
3858 */
3859 int diff_populate_filespec(struct repository *r,
3860 struct diff_filespec *s,
3861 unsigned int flags)
3862 {
3863 int size_only = flags & CHECK_SIZE_ONLY;
3864 int err = 0;
3865 int conv_flags = global_conv_flags_eol;
3866 /*
3867 * demote FAIL to WARN to allow inspecting the situation
3868 * instead of refusing.
3869 */
3870 if (conv_flags & CONV_EOL_RNDTRP_DIE)
3871 conv_flags = CONV_EOL_RNDTRP_WARN;
3872
3873 if (!DIFF_FILE_VALID(s))
3874 die("internal error: asking to populate invalid file.");
3875 if (S_ISDIR(s->mode))
3876 return -1;
3877
3878 if (s->data)
3879 return 0;
3880
3881 if (size_only && 0 < s->size)
3882 return 0;
3883
3884 if (S_ISGITLINK(s->mode))
3885 return diff_populate_gitlink(s, size_only);
3886
3887 if (!s->oid_valid ||
3888 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
3889 struct strbuf buf = STRBUF_INIT;
3890 struct stat st;
3891 int fd;
3892
3893 if (lstat(s->path, &st) < 0) {
3894 err_empty:
3895 err = -1;
3896 empty:
3897 s->data = (char *)"";
3898 s->size = 0;
3899 return err;
3900 }
3901 s->size = xsize_t(st.st_size);
3902 if (!s->size)
3903 goto empty;
3904 if (S_ISLNK(st.st_mode)) {
3905 struct strbuf sb = STRBUF_INIT;
3906
3907 if (strbuf_readlink(&sb, s->path, s->size))
3908 goto err_empty;
3909 s->size = sb.len;
3910 s->data = strbuf_detach(&sb, NULL);
3911 s->should_free = 1;
3912 return 0;
3913 }
3914
3915 /*
3916 * Even if the caller would be happy with getting
3917 * only the size, we cannot return early at this
3918 * point if the path requires us to run the content
3919 * conversion.
3920 */
3921 if (size_only && !would_convert_to_git(r->index, s->path))
3922 return 0;
3923
3924 /*
3925 * Note: this check uses xsize_t(st.st_size) that may
3926 * not be the true size of the blob after it goes
3927 * through convert_to_git(). This may not strictly be
3928 * correct, but the whole point of big_file_threshold
3929 * and is_binary check being that we want to avoid
3930 * opening the file and inspecting the contents, this
3931 * is probably fine.
3932 */
3933 if ((flags & CHECK_BINARY) &&
3934 s->size > big_file_threshold && s->is_binary == -1) {
3935 s->is_binary = 1;
3936 return 0;
3937 }
3938 fd = open(s->path, O_RDONLY);
3939 if (fd < 0)
3940 goto err_empty;
3941 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3942 close(fd);
3943 s->should_munmap = 1;
3944
3945 /*
3946 * Convert from working tree format to canonical git format
3947 */
3948 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
3949 size_t size = 0;
3950 munmap(s->data, s->size);
3951 s->should_munmap = 0;
3952 s->data = strbuf_detach(&buf, &size);
3953 s->size = size;
3954 s->should_free = 1;
3955 }
3956 }
3957 else {
3958 enum object_type type;
3959 if (size_only || (flags & CHECK_BINARY)) {
3960 type = oid_object_info(r, &s->oid, &s->size);
3961 if (type < 0)
3962 die("unable to read %s",
3963 oid_to_hex(&s->oid));
3964 if (size_only)
3965 return 0;
3966 if (s->size > big_file_threshold && s->is_binary == -1) {
3967 s->is_binary = 1;
3968 return 0;
3969 }
3970 }
3971 s->data = read_object_file(&s->oid, &type, &s->size);
3972 if (!s->data)
3973 die("unable to read %s", oid_to_hex(&s->oid));
3974 s->should_free = 1;
3975 }
3976 return 0;
3977 }
3978
3979 void diff_free_filespec_blob(struct diff_filespec *s)
3980 {
3981 if (s->should_free)
3982 free(s->data);
3983 else if (s->should_munmap)
3984 munmap(s->data, s->size);
3985
3986 if (s->should_free || s->should_munmap) {
3987 s->should_free = s->should_munmap = 0;
3988 s->data = NULL;
3989 }
3990 }
3991
3992 void diff_free_filespec_data(struct diff_filespec *s)
3993 {
3994 diff_free_filespec_blob(s);
3995 FREE_AND_NULL(s->cnt_data);
3996 }
3997
3998 static void prep_temp_blob(struct index_state *istate,
3999 const char *path, struct diff_tempfile *temp,
4000 void *blob,
4001 unsigned long size,
4002 const struct object_id *oid,
4003 int mode)
4004 {
4005 struct strbuf buf = STRBUF_INIT;
4006 struct strbuf tempfile = STRBUF_INIT;
4007 char *path_dup = xstrdup(path);
4008 const char *base = basename(path_dup);
4009
4010 /* Generate "XXXXXX_basename.ext" */
4011 strbuf_addstr(&tempfile, "XXXXXX_");
4012 strbuf_addstr(&tempfile, base);
4013
4014 temp->tempfile = mks_tempfile_ts(tempfile.buf, strlen(base) + 1);
4015 if (!temp->tempfile)
4016 die_errno("unable to create temp-file");
4017 if (convert_to_working_tree(istate, path,
4018 (const char *)blob, (size_t)size, &buf)) {
4019 blob = buf.buf;
4020 size = buf.len;
4021 }
4022 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4023 close_tempfile_gently(temp->tempfile))
4024 die_errno("unable to write temp-file");
4025 temp->name = get_tempfile_path(temp->tempfile);
4026 oid_to_hex_r(temp->hex, oid);
4027 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4028 strbuf_release(&buf);
4029 strbuf_release(&tempfile);
4030 free(path_dup);
4031 }
4032
4033 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4034 const char *name,
4035 struct diff_filespec *one)
4036 {
4037 struct diff_tempfile *temp = claim_diff_tempfile();
4038
4039 if (!DIFF_FILE_VALID(one)) {
4040 not_a_valid_file:
4041 /* A '-' entry produces this for file-2, and
4042 * a '+' entry produces this for file-1.
4043 */
4044 temp->name = "/dev/null";
4045 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4046 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4047 return temp;
4048 }
4049
4050 if (!S_ISGITLINK(one->mode) &&
4051 (!one->oid_valid ||
4052 reuse_worktree_file(r->index, name, &one->oid, 1))) {
4053 struct stat st;
4054 if (lstat(name, &st) < 0) {
4055 if (errno == ENOENT)
4056 goto not_a_valid_file;
4057 die_errno("stat(%s)", name);
4058 }
4059 if (S_ISLNK(st.st_mode)) {
4060 struct strbuf sb = STRBUF_INIT;
4061 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4062 die_errno("readlink(%s)", name);
4063 prep_temp_blob(r->index, name, temp, sb.buf, sb.len,
4064 (one->oid_valid ?
4065 &one->oid : &null_oid),
4066 (one->oid_valid ?
4067 one->mode : S_IFLNK));
4068 strbuf_release(&sb);
4069 }
4070 else {
4071 /* we can borrow from the file in the work tree */
4072 temp->name = name;
4073 if (!one->oid_valid)
4074 oid_to_hex_r(temp->hex, &null_oid);
4075 else
4076 oid_to_hex_r(temp->hex, &one->oid);
4077 /* Even though we may sometimes borrow the
4078 * contents from the work tree, we always want
4079 * one->mode. mode is trustworthy even when
4080 * !(one->oid_valid), as long as
4081 * DIFF_FILE_VALID(one).
4082 */
4083 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4084 }
4085 return temp;
4086 }
4087 else {
4088 if (diff_populate_filespec(r, one, 0))
4089 die("cannot read data blob for %s", one->path);
4090 prep_temp_blob(r->index, name, temp,
4091 one->data, one->size,
4092 &one->oid, one->mode);
4093 }
4094 return temp;
4095 }
4096
4097 static void add_external_diff_name(struct repository *r,
4098 struct argv_array *argv,
4099 const char *name,
4100 struct diff_filespec *df)
4101 {
4102 struct diff_tempfile *temp = prepare_temp_file(r, name, df);
4103 argv_array_push(argv, temp->name);
4104 argv_array_push(argv, temp->hex);
4105 argv_array_push(argv, temp->mode);
4106 }
4107
4108 /* An external diff command takes:
4109 *
4110 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4111 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4112 *
4113 */
4114 static void run_external_diff(const char *pgm,
4115 const char *name,
4116 const char *other,
4117 struct diff_filespec *one,
4118 struct diff_filespec *two,
4119 const char *xfrm_msg,
4120 int complete_rewrite,
4121 struct diff_options *o)
4122 {
4123 struct argv_array argv = ARGV_ARRAY_INIT;
4124 struct argv_array env = ARGV_ARRAY_INIT;
4125 struct diff_queue_struct *q = &diff_queued_diff;
4126
4127 argv_array_push(&argv, pgm);
4128 argv_array_push(&argv, name);
4129
4130 if (one && two) {
4131 add_external_diff_name(o->repo, &argv, name, one);
4132 if (!other)
4133 add_external_diff_name(o->repo, &argv, name, two);
4134 else {
4135 add_external_diff_name(o->repo, &argv, other, two);
4136 argv_array_push(&argv, other);
4137 argv_array_push(&argv, xfrm_msg);
4138 }
4139 }
4140
4141 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
4142 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4143
4144 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
4145 die(_("external diff died, stopping at %s"), name);
4146
4147 remove_tempfile();
4148 argv_array_clear(&argv);
4149 argv_array_clear(&env);
4150 }
4151
4152 static int similarity_index(struct diff_filepair *p)
4153 {
4154 return p->score * 100 / MAX_SCORE;
4155 }
4156
4157 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4158 {
4159 if (startup_info->have_repository)
4160 return find_unique_abbrev(oid, abbrev);
4161 else {
4162 char *hex = oid_to_hex(oid);
4163 if (abbrev < 0)
4164 abbrev = FALLBACK_DEFAULT_ABBREV;
4165 if (abbrev > the_hash_algo->hexsz)
4166 BUG("oid abbreviation out of range: %d", abbrev);
4167 if (abbrev)
4168 hex[abbrev] = '\0';
4169 return hex;
4170 }
4171 }
4172
4173 static void fill_metainfo(struct strbuf *msg,
4174 const char *name,
4175 const char *other,
4176 struct diff_filespec *one,
4177 struct diff_filespec *two,
4178 struct diff_options *o,
4179 struct diff_filepair *p,
4180 int *must_show_header,
4181 int use_color)
4182 {
4183 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4184 const char *reset = diff_get_color(use_color, DIFF_RESET);
4185 const char *line_prefix = diff_line_prefix(o);
4186
4187 *must_show_header = 1;
4188 strbuf_init(msg, PATH_MAX * 2 + 300);
4189 switch (p->status) {
4190 case DIFF_STATUS_COPIED:
4191 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4192 line_prefix, set, similarity_index(p));
4193 strbuf_addf(msg, "%s\n%s%scopy from ",
4194 reset, line_prefix, set);
4195 quote_c_style(name, msg, NULL, 0);
4196 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4197 quote_c_style(other, msg, NULL, 0);
4198 strbuf_addf(msg, "%s\n", reset);
4199 break;
4200 case DIFF_STATUS_RENAMED:
4201 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4202 line_prefix, set, similarity_index(p));
4203 strbuf_addf(msg, "%s\n%s%srename from ",
4204 reset, line_prefix, set);
4205 quote_c_style(name, msg, NULL, 0);
4206 strbuf_addf(msg, "%s\n%s%srename to ",
4207 reset, line_prefix, set);
4208 quote_c_style(other, msg, NULL, 0);
4209 strbuf_addf(msg, "%s\n", reset);
4210 break;
4211 case DIFF_STATUS_MODIFIED:
4212 if (p->score) {
4213 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4214 line_prefix,
4215 set, similarity_index(p), reset);
4216 break;
4217 }
4218 /* fallthru */
4219 default:
4220 *must_show_header = 0;
4221 }
4222 if (one && two && !oideq(&one->oid, &two->oid)) {
4223 const unsigned hexsz = the_hash_algo->hexsz;
4224 int abbrev = o->flags.full_index ? hexsz : DEFAULT_ABBREV;
4225
4226 if (o->flags.binary) {
4227 mmfile_t mf;
4228 if ((!fill_mmfile(o->repo, &mf, one) &&
4229 diff_filespec_is_binary(o->repo, one)) ||
4230 (!fill_mmfile(o->repo, &mf, two) &&
4231 diff_filespec_is_binary(o->repo, two)))
4232 abbrev = hexsz;
4233 }
4234 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4235 diff_abbrev_oid(&one->oid, abbrev),
4236 diff_abbrev_oid(&two->oid, abbrev));
4237 if (one->mode == two->mode)
4238 strbuf_addf(msg, " %06o", one->mode);
4239 strbuf_addf(msg, "%s\n", reset);
4240 }
4241 }
4242
4243 static void run_diff_cmd(const char *pgm,
4244 const char *name,
4245 const char *other,
4246 const char *attr_path,
4247 struct diff_filespec *one,
4248 struct diff_filespec *two,
4249 struct strbuf *msg,
4250 struct diff_options *o,
4251 struct diff_filepair *p)
4252 {
4253 const char *xfrm_msg = NULL;
4254 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4255 int must_show_header = 0;
4256
4257
4258 if (o->flags.allow_external) {
4259 struct userdiff_driver *drv;
4260
4261 drv = userdiff_find_by_path(o->repo->index, attr_path);
4262 if (drv && drv->external)
4263 pgm = drv->external;
4264 }
4265
4266 if (msg) {
4267 /*
4268 * don't use colors when the header is intended for an
4269 * external diff driver
4270 */
4271 fill_metainfo(msg, name, other, one, two, o, p,
4272 &must_show_header,
4273 want_color(o->use_color) && !pgm);
4274 xfrm_msg = msg->len ? msg->buf : NULL;
4275 }
4276
4277 if (pgm) {
4278 run_external_diff(pgm, name, other, one, two, xfrm_msg,
4279 complete_rewrite, o);
4280 return;
4281 }
4282 if (one && two)
4283 builtin_diff(name, other ? other : name,
4284 one, two, xfrm_msg, must_show_header,
4285 o, complete_rewrite);
4286 else
4287 fprintf(o->file, "* Unmerged path %s\n", name);
4288 }
4289
4290 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4291 {
4292 if (DIFF_FILE_VALID(one)) {
4293 if (!one->oid_valid) {
4294 struct stat st;
4295 if (one->is_stdin) {
4296 oidclr(&one->oid);
4297 return;
4298 }
4299 if (lstat(one->path, &st) < 0)
4300 die_errno("stat '%s'", one->path);
4301 if (index_path(istate, &one->oid, one->path, &st, 0))
4302 die("cannot hash %s", one->path);
4303 }
4304 }
4305 else
4306 oidclr(&one->oid);
4307 }
4308
4309 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4310 {
4311 /* Strip the prefix but do not molest /dev/null and absolute paths */
4312 if (*namep && !is_absolute_path(*namep)) {
4313 *namep += prefix_length;
4314 if (**namep == '/')
4315 ++*namep;
4316 }
4317 if (*otherp && !is_absolute_path(*otherp)) {
4318 *otherp += prefix_length;
4319 if (**otherp == '/')
4320 ++*otherp;
4321 }
4322 }
4323
4324 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4325 {
4326 const char *pgm = external_diff();
4327 struct strbuf msg;
4328 struct diff_filespec *one = p->one;
4329 struct diff_filespec *two = p->two;
4330 const char *name;
4331 const char *other;
4332 const char *attr_path;
4333
4334 name = one->path;
4335 other = (strcmp(name, two->path) ? two->path : NULL);
4336 attr_path = name;
4337 if (o->prefix_length)
4338 strip_prefix(o->prefix_length, &name, &other);
4339
4340 if (!o->flags.allow_external)
4341 pgm = NULL;
4342
4343 if (DIFF_PAIR_UNMERGED(p)) {
4344 run_diff_cmd(pgm, name, NULL, attr_path,
4345 NULL, NULL, NULL, o, p);
4346 return;
4347 }
4348
4349 diff_fill_oid_info(one, o->repo->index);
4350 diff_fill_oid_info(two, o->repo->index);
4351
4352 if (!pgm &&
4353 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4354 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4355 /*
4356 * a filepair that changes between file and symlink
4357 * needs to be split into deletion and creation.
4358 */
4359 struct diff_filespec *null = alloc_filespec(two->path);
4360 run_diff_cmd(NULL, name, other, attr_path,
4361 one, null, &msg,
4362 o, p);
4363 free(null);
4364 strbuf_release(&msg);
4365
4366 null = alloc_filespec(one->path);
4367 run_diff_cmd(NULL, name, other, attr_path,
4368 null, two, &msg, o, p);
4369 free(null);
4370 }
4371 else
4372 run_diff_cmd(pgm, name, other, attr_path,
4373 one, two, &msg, o, p);
4374
4375 strbuf_release(&msg);
4376 }
4377
4378 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4379 struct diffstat_t *diffstat)
4380 {
4381 const char *name;
4382 const char *other;
4383
4384 if (DIFF_PAIR_UNMERGED(p)) {
4385 /* unmerged */
4386 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4387 diffstat, o, p);
4388 return;
4389 }
4390
4391 name = p->one->path;
4392 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4393
4394 if (o->prefix_length)
4395 strip_prefix(o->prefix_length, &name, &other);
4396
4397 diff_fill_oid_info(p->one, o->repo->index);
4398 diff_fill_oid_info(p->two, o->repo->index);
4399
4400 builtin_diffstat(name, other, p->one, p->two,
4401 diffstat, o, p);
4402 }
4403
4404 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4405 {
4406 const char *name;
4407 const char *other;
4408 const char *attr_path;
4409
4410 if (DIFF_PAIR_UNMERGED(p)) {
4411 /* unmerged */
4412 return;
4413 }
4414
4415 name = p->one->path;
4416 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4417 attr_path = other ? other : name;
4418
4419 if (o->prefix_length)
4420 strip_prefix(o->prefix_length, &name, &other);
4421
4422 diff_fill_oid_info(p->one, o->repo->index);
4423 diff_fill_oid_info(p->two, o->repo->index);
4424
4425 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4426 }
4427
4428 void repo_diff_setup(struct repository *r, struct diff_options *options)
4429 {
4430 memcpy(options, &default_diff_options, sizeof(*options));
4431
4432 options->file = stdout;
4433 options->repo = r;
4434
4435 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4436 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4437 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4438 options->abbrev = DEFAULT_ABBREV;
4439 options->line_termination = '\n';
4440 options->break_opt = -1;
4441 options->rename_limit = -1;
4442 options->dirstat_permille = diff_dirstat_permille_default;
4443 options->context = diff_context_default;
4444 options->interhunkcontext = diff_interhunk_context_default;
4445 options->ws_error_highlight = ws_error_highlight_default;
4446 options->flags.rename_empty = 1;
4447 options->objfind = NULL;
4448
4449 /* pathchange left =NULL by default */
4450 options->change = diff_change;
4451 options->add_remove = diff_addremove;
4452 options->use_color = diff_use_color_default;
4453 options->detect_rename = diff_detect_rename_default;
4454 options->xdl_opts |= diff_algorithm;
4455 if (diff_indent_heuristic)
4456 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4457
4458 options->orderfile = diff_order_file_cfg;
4459
4460 if (diff_no_prefix) {
4461 options->a_prefix = options->b_prefix = "";
4462 } else if (!diff_mnemonic_prefix) {
4463 options->a_prefix = "a/";
4464 options->b_prefix = "b/";
4465 }
4466
4467 options->color_moved = diff_color_moved_default;
4468 options->color_moved_ws_handling = diff_color_moved_ws_default;
4469 }
4470
4471 void diff_setup_done(struct diff_options *options)
4472 {
4473 unsigned check_mask = DIFF_FORMAT_NAME |
4474 DIFF_FORMAT_NAME_STATUS |
4475 DIFF_FORMAT_CHECKDIFF |
4476 DIFF_FORMAT_NO_OUTPUT;
4477 /*
4478 * This must be signed because we're comparing against a potentially
4479 * negative value.
4480 */
4481 const int hexsz = the_hash_algo->hexsz;
4482
4483 if (options->set_default)
4484 options->set_default(options);
4485
4486 if (HAS_MULTI_BITS(options->output_format & check_mask))
4487 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4488
4489 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4490 die(_("-G, -S and --find-object are mutually exclusive"));
4491
4492 /*
4493 * Most of the time we can say "there are changes"
4494 * only by checking if there are changed paths, but
4495 * --ignore-whitespace* options force us to look
4496 * inside contents.
4497 */
4498
4499 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS))
4500 options->flags.diff_from_contents = 1;
4501 else
4502 options->flags.diff_from_contents = 0;
4503
4504 if (options->flags.find_copies_harder)
4505 options->detect_rename = DIFF_DETECT_COPY;
4506
4507 if (!options->flags.relative_name)
4508 options->prefix = NULL;
4509 if (options->prefix)
4510 options->prefix_length = strlen(options->prefix);
4511 else
4512 options->prefix_length = 0;
4513
4514 if (options->output_format & (DIFF_FORMAT_NAME |
4515 DIFF_FORMAT_NAME_STATUS |
4516 DIFF_FORMAT_CHECKDIFF |
4517 DIFF_FORMAT_NO_OUTPUT))
4518 options->output_format &= ~(DIFF_FORMAT_RAW |
4519 DIFF_FORMAT_NUMSTAT |
4520 DIFF_FORMAT_DIFFSTAT |
4521 DIFF_FORMAT_SHORTSTAT |
4522 DIFF_FORMAT_DIRSTAT |
4523 DIFF_FORMAT_SUMMARY |
4524 DIFF_FORMAT_PATCH);
4525
4526 /*
4527 * These cases always need recursive; we do not drop caller-supplied
4528 * recursive bits for other formats here.
4529 */
4530 if (options->output_format & (DIFF_FORMAT_PATCH |
4531 DIFF_FORMAT_NUMSTAT |
4532 DIFF_FORMAT_DIFFSTAT |
4533 DIFF_FORMAT_SHORTSTAT |
4534 DIFF_FORMAT_DIRSTAT |
4535 DIFF_FORMAT_SUMMARY |
4536 DIFF_FORMAT_CHECKDIFF))
4537 options->flags.recursive = 1;
4538 /*
4539 * Also pickaxe would not work very well if you do not say recursive
4540 */
4541 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4542 options->flags.recursive = 1;
4543 /*
4544 * When patches are generated, submodules diffed against the work tree
4545 * must be checked for dirtiness too so it can be shown in the output
4546 */
4547 if (options->output_format & DIFF_FORMAT_PATCH)
4548 options->flags.dirty_submodules = 1;
4549
4550 if (options->detect_rename && options->rename_limit < 0)
4551 options->rename_limit = diff_rename_limit_default;
4552 if (hexsz < options->abbrev)
4553 options->abbrev = hexsz; /* full */
4554
4555 /*
4556 * It does not make sense to show the first hit we happened
4557 * to have found. It does not make sense not to return with
4558 * exit code in such a case either.
4559 */
4560 if (options->flags.quick) {
4561 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4562 options->flags.exit_with_status = 1;
4563 }
4564
4565 options->diff_path_counter = 0;
4566
4567 if (options->flags.follow_renames && options->pathspec.nr != 1)
4568 die(_("--follow requires exactly one pathspec"));
4569
4570 if (!options->use_color || external_diff())
4571 options->color_moved = 0;
4572 }
4573
4574 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4575 {
4576 char c, *eq;
4577 int len;
4578
4579 if (*arg != '-')
4580 return 0;
4581 c = *++arg;
4582 if (!c)
4583 return 0;
4584 if (c == arg_short) {
4585 c = *++arg;
4586 if (!c)
4587 return 1;
4588 if (val && isdigit(c)) {
4589 char *end;
4590 int n = strtoul(arg, &end, 10);
4591 if (*end)
4592 return 0;
4593 *val = n;
4594 return 1;
4595 }
4596 return 0;
4597 }
4598 if (c != '-')
4599 return 0;
4600 arg++;
4601 eq = strchrnul(arg, '=');
4602 len = eq - arg;
4603 if (!len || strncmp(arg, arg_long, len))
4604 return 0;
4605 if (*eq) {
4606 int n;
4607 char *end;
4608 if (!isdigit(*++eq))
4609 return 0;
4610 n = strtoul(eq, &end, 10);
4611 if (*end)
4612 return 0;
4613 *val = n;
4614 }
4615 return 1;
4616 }
4617
4618 static int diff_scoreopt_parse(const char *opt);
4619
4620 static inline int short_opt(char opt, const char **argv,
4621 const char **optarg)
4622 {
4623 const char *arg = argv[0];
4624 if (arg[0] != '-' || arg[1] != opt)
4625 return 0;
4626 if (arg[2] != '\0') {
4627 *optarg = arg + 2;
4628 return 1;
4629 }
4630 if (!argv[1])
4631 die("Option '%c' requires a value", opt);
4632 *optarg = argv[1];
4633 return 2;
4634 }
4635
4636 int parse_long_opt(const char *opt, const char **argv,
4637 const char **optarg)
4638 {
4639 const char *arg = argv[0];
4640 if (!skip_prefix(arg, "--", &arg))
4641 return 0;
4642 if (!skip_prefix(arg, opt, &arg))
4643 return 0;
4644 if (*arg == '=') { /* stuck form: --option=value */
4645 *optarg = arg + 1;
4646 return 1;
4647 }
4648 if (*arg != '\0')
4649 return 0;
4650 /* separate form: --option value */
4651 if (!argv[1])
4652 die("Option '--%s' requires a value", opt);
4653 *optarg = argv[1];
4654 return 2;
4655 }
4656
4657 static int stat_opt(struct diff_options *options, const char **av)
4658 {
4659 const char *arg = av[0];
4660 char *end;
4661 int width = options->stat_width;
4662 int name_width = options->stat_name_width;
4663 int graph_width = options->stat_graph_width;
4664 int count = options->stat_count;
4665 int argcount = 1;
4666
4667 if (!skip_prefix(arg, "--stat", &arg))
4668 BUG("stat option does not begin with --stat: %s", arg);
4669 end = (char *)arg;
4670
4671 switch (*arg) {
4672 case '-':
4673 if (skip_prefix(arg, "-width", &arg)) {
4674 if (*arg == '=')
4675 width = strtoul(arg + 1, &end, 10);
4676 else if (!*arg && !av[1])
4677 die_want_option("--stat-width");
4678 else if (!*arg) {
4679 width = strtoul(av[1], &end, 10);
4680 argcount = 2;
4681 }
4682 } else if (skip_prefix(arg, "-name-width", &arg)) {
4683 if (*arg == '=')
4684 name_width = strtoul(arg + 1, &end, 10);
4685 else if (!*arg && !av[1])
4686 die_want_option("--stat-name-width");
4687 else if (!*arg) {
4688 name_width = strtoul(av[1], &end, 10);
4689 argcount = 2;
4690 }
4691 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4692 if (*arg == '=')
4693 graph_width = strtoul(arg + 1, &end, 10);
4694 else if (!*arg && !av[1])
4695 die_want_option("--stat-graph-width");
4696 else if (!*arg) {
4697 graph_width = strtoul(av[1], &end, 10);
4698 argcount = 2;
4699 }
4700 } else if (skip_prefix(arg, "-count", &arg)) {
4701 if (*arg == '=')
4702 count = strtoul(arg + 1, &end, 10);
4703 else if (!*arg && !av[1])
4704 die_want_option("--stat-count");
4705 else if (!*arg) {
4706 count = strtoul(av[1], &end, 10);
4707 argcount = 2;
4708 }
4709 }
4710 break;
4711 case '=':
4712 width = strtoul(arg+1, &end, 10);
4713 if (*end == ',')
4714 name_width = strtoul(end+1, &end, 10);
4715 if (*end == ',')
4716 count = strtoul(end+1, &end, 10);
4717 }
4718
4719 /* Important! This checks all the error cases! */
4720 if (*end)
4721 return 0;
4722 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4723 options->stat_name_width = name_width;
4724 options->stat_graph_width = graph_width;
4725 options->stat_width = width;
4726 options->stat_count = count;
4727 return argcount;
4728 }
4729
4730 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4731 {
4732 struct strbuf errmsg = STRBUF_INIT;
4733 if (parse_dirstat_params(options, params, &errmsg))
4734 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4735 errmsg.buf);
4736 strbuf_release(&errmsg);
4737 /*
4738 * The caller knows a dirstat-related option is given from the command
4739 * line; allow it to say "return this_function();"
4740 */
4741 options->output_format |= DIFF_FORMAT_DIRSTAT;
4742 return 1;
4743 }
4744
4745 static int parse_submodule_opt(struct diff_options *options, const char *value)
4746 {
4747 if (parse_submodule_params(options, value))
4748 die(_("Failed to parse --submodule option parameter: '%s'"),
4749 value);
4750 return 1;
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 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4785 {
4786 int i, optch;
4787
4788 prepare_filter_bits();
4789
4790 /*
4791 * If there is a negation e.g. 'd' in the input, and we haven't
4792 * initialized the filter field with another --diff-filter, start
4793 * from full set of bits, except for AON.
4794 */
4795 if (!opt->filter) {
4796 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4797 if (optch < 'a' || 'z' < optch)
4798 continue;
4799 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4800 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4801 break;
4802 }
4803 }
4804
4805 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4806 unsigned int bit;
4807 int negate;
4808
4809 if ('a' <= optch && optch <= 'z') {
4810 negate = 1;
4811 optch = toupper(optch);
4812 } else {
4813 negate = 0;
4814 }
4815
4816 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4817 if (!bit)
4818 return optarg[i];
4819 if (negate)
4820 opt->filter &= ~bit;
4821 else
4822 opt->filter |= bit;
4823 }
4824 return 0;
4825 }
4826
4827 static void enable_patch_output(int *fmt)
4828 {
4829 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4830 *fmt |= DIFF_FORMAT_PATCH;
4831 }
4832
4833 static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4834 {
4835 int val = parse_ws_error_highlight(arg);
4836
4837 if (val < 0) {
4838 error("unknown value after ws-error-highlight=%.*s",
4839 -1 - val, arg);
4840 return 0;
4841 }
4842 opt->ws_error_highlight = val;
4843 return 1;
4844 }
4845
4846 static int parse_objfind_opt(struct diff_options *opt, const char *arg)
4847 {
4848 struct object_id oid;
4849
4850 if (get_oid(arg, &oid))
4851 return error("unable to resolve '%s'", arg);
4852
4853 if (!opt->objfind)
4854 opt->objfind = xcalloc(1, sizeof(*opt->objfind));
4855
4856 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4857 opt->flags.recursive = 1;
4858 opt->flags.tree_in_recursive = 1;
4859 oidset_insert(opt->objfind, &oid);
4860 return 1;
4861 }
4862
4863 int diff_opt_parse(struct diff_options *options,
4864 const char **av, int ac, const char *prefix)
4865 {
4866 const char *arg = av[0];
4867 const char *optarg;
4868 int argcount;
4869
4870 if (!prefix)
4871 prefix = "";
4872
4873 /* Output format options */
4874 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4875 || opt_arg(arg, 'U', "unified", &options->context))
4876 enable_patch_output(&options->output_format);
4877 else if (!strcmp(arg, "--raw"))
4878 options->output_format |= DIFF_FORMAT_RAW;
4879 else if (!strcmp(arg, "--patch-with-raw")) {
4880 enable_patch_output(&options->output_format);
4881 options->output_format |= DIFF_FORMAT_RAW;
4882 } else if (!strcmp(arg, "--numstat"))
4883 options->output_format |= DIFF_FORMAT_NUMSTAT;
4884 else if (!strcmp(arg, "--shortstat"))
4885 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4886 else if (skip_prefix(arg, "-X", &arg) ||
4887 skip_to_optional_arg(arg, "--dirstat", &arg))
4888 return parse_dirstat_opt(options, arg);
4889 else if (!strcmp(arg, "--cumulative"))
4890 return parse_dirstat_opt(options, "cumulative");
4891 else if (skip_to_optional_arg(arg, "--dirstat-by-file", &arg)) {
4892 parse_dirstat_opt(options, "files");
4893 return parse_dirstat_opt(options, arg);
4894 }
4895 else if (!strcmp(arg, "--check"))
4896 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4897 else if (!strcmp(arg, "--summary"))
4898 options->output_format |= DIFF_FORMAT_SUMMARY;
4899 else if (!strcmp(arg, "--patch-with-stat")) {
4900 enable_patch_output(&options->output_format);
4901 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4902 } else if (!strcmp(arg, "--name-only"))
4903 options->output_format |= DIFF_FORMAT_NAME;
4904 else if (!strcmp(arg, "--name-status"))
4905 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4906 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4907 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4908 else if (starts_with(arg, "--stat"))
4909 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4910 return stat_opt(options, av);
4911 else if (!strcmp(arg, "--compact-summary")) {
4912 options->flags.stat_with_summary = 1;
4913 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4914 } else if (!strcmp(arg, "--no-compact-summary"))
4915 options->flags.stat_with_summary = 0;
4916 else if (skip_prefix(arg, "--output-indicator-new=", &arg))
4917 options->output_indicators[OUTPUT_INDICATOR_NEW] = arg[0];
4918 else if (skip_prefix(arg, "--output-indicator-old=", &arg))
4919 options->output_indicators[OUTPUT_INDICATOR_OLD] = arg[0];
4920 else if (skip_prefix(arg, "--output-indicator-context=", &arg))
4921 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = arg[0];
4922
4923 /* renames options */
4924 else if (starts_with(arg, "-B") ||
4925 skip_to_optional_arg(arg, "--break-rewrites", NULL)) {
4926 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4927 return error("invalid argument to -B: %s", arg+2);
4928 }
4929 else if (starts_with(arg, "-M") ||
4930 skip_to_optional_arg(arg, "--find-renames", NULL)) {
4931 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4932 return error("invalid argument to -M: %s", arg+2);
4933 options->detect_rename = DIFF_DETECT_RENAME;
4934 }
4935 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4936 options->irreversible_delete = 1;
4937 }
4938 else if (starts_with(arg, "-C") ||
4939 skip_to_optional_arg(arg, "--find-copies", NULL)) {
4940 if (options->detect_rename == DIFF_DETECT_COPY)
4941 options->flags.find_copies_harder = 1;
4942 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4943 return error("invalid argument to -C: %s", arg+2);
4944 options->detect_rename = DIFF_DETECT_COPY;
4945 }
4946 else if (!strcmp(arg, "--no-renames"))
4947 options->detect_rename = 0;
4948 else if (!strcmp(arg, "--rename-empty"))
4949 options->flags.rename_empty = 1;
4950 else if (!strcmp(arg, "--no-rename-empty"))
4951 options->flags.rename_empty = 0;
4952 else if (skip_to_optional_arg_default(arg, "--relative", &arg, NULL)) {
4953 options->flags.relative_name = 1;
4954 if (arg)
4955 options->prefix = arg;
4956 }
4957
4958 /* xdiff options */
4959 else if (!strcmp(arg, "--minimal"))
4960 DIFF_XDL_SET(options, NEED_MINIMAL);
4961 else if (!strcmp(arg, "--no-minimal"))
4962 DIFF_XDL_CLR(options, NEED_MINIMAL);
4963 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4964 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4965 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4966 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4967 else if (!strcmp(arg, "--ignore-space-at-eol"))
4968 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4969 else if (!strcmp(arg, "--ignore-cr-at-eol"))
4970 DIFF_XDL_SET(options, IGNORE_CR_AT_EOL);
4971 else if (!strcmp(arg, "--ignore-blank-lines"))
4972 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4973 else if (!strcmp(arg, "--indent-heuristic"))
4974 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4975 else if (!strcmp(arg, "--no-indent-heuristic"))
4976 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4977 else if (!strcmp(arg, "--patience")) {
4978 int i;
4979 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4980 /*
4981 * Both --patience and --anchored use PATIENCE_DIFF
4982 * internally, so remove any anchors previously
4983 * specified.
4984 */
4985 for (i = 0; i < options->anchors_nr; i++)
4986 free(options->anchors[i]);
4987 options->anchors_nr = 0;
4988 } else if (!strcmp(arg, "--histogram"))
4989 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4990 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4991 long value = parse_algorithm_value(optarg);
4992 if (value < 0)
4993 return error("option diff-algorithm accepts \"myers\", "
4994 "\"minimal\", \"patience\" and \"histogram\"");
4995 /* clear out previous settings */
4996 DIFF_XDL_CLR(options, NEED_MINIMAL);
4997 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4998 options->xdl_opts |= value;
4999 return argcount;
5000 } else if (skip_prefix(arg, "--anchored=", &arg)) {
5001 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5002 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
5003 options->anchors_alloc);
5004 options->anchors[options->anchors_nr++] = xstrdup(arg);
5005 }
5006
5007 /* flags options */
5008 else if (!strcmp(arg, "--binary")) {
5009 enable_patch_output(&options->output_format);
5010 options->flags.binary = 1;
5011 }
5012 else if (!strcmp(arg, "--full-index"))
5013 options->flags.full_index = 1;
5014 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
5015 options->flags.text = 1;
5016 else if (!strcmp(arg, "-R"))
5017 options->flags.reverse_diff = 1;
5018 else if (!strcmp(arg, "--find-copies-harder"))
5019 options->flags.find_copies_harder = 1;
5020 else if (!strcmp(arg, "--follow"))
5021 options->flags.follow_renames = 1;
5022 else if (!strcmp(arg, "--no-follow")) {
5023 options->flags.follow_renames = 0;
5024 options->flags.default_follow_renames = 0;
5025 } else if (skip_to_optional_arg_default(arg, "--color", &arg, "always")) {
5026 int value = git_config_colorbool(NULL, arg);
5027 if (value < 0)
5028 return error("option `color' expects \"always\", \"auto\", or \"never\"");
5029 options->use_color = value;
5030 }
5031 else if (!strcmp(arg, "--no-color"))
5032 options->use_color = 0;
5033 else if (!strcmp(arg, "--color-moved")) {
5034 if (diff_color_moved_default)
5035 options->color_moved = diff_color_moved_default;
5036 if (options->color_moved == COLOR_MOVED_NO)
5037 options->color_moved = COLOR_MOVED_DEFAULT;
5038 } else if (!strcmp(arg, "--no-color-moved"))
5039 options->color_moved = COLOR_MOVED_NO;
5040 else if (skip_prefix(arg, "--color-moved=", &arg)) {
5041 int cm = parse_color_moved(arg);
5042 if (cm < 0)
5043 return error("bad --color-moved argument: %s", arg);
5044 options->color_moved = cm;
5045 } else if (skip_prefix(arg, "--color-moved-ws=", &arg)) {
5046 unsigned cm = parse_color_moved_ws(arg);
5047 if (cm & COLOR_MOVED_WS_ERROR)
5048 return -1;
5049 options->color_moved_ws_handling = cm;
5050 } else if (skip_to_optional_arg_default(arg, "--color-words", &options->word_regex, NULL)) {
5051 options->use_color = 1;
5052 options->word_diff = DIFF_WORDS_COLOR;
5053 }
5054 else if (!strcmp(arg, "--word-diff")) {
5055 if (options->word_diff == DIFF_WORDS_NONE)
5056 options->word_diff = DIFF_WORDS_PLAIN;
5057 }
5058 else if (skip_prefix(arg, "--word-diff=", &arg)) {
5059 if (!strcmp(arg, "plain"))
5060 options->word_diff = DIFF_WORDS_PLAIN;
5061 else if (!strcmp(arg, "color")) {
5062 options->use_color = 1;
5063 options->word_diff = DIFF_WORDS_COLOR;
5064 }
5065 else if (!strcmp(arg, "porcelain"))
5066 options->word_diff = DIFF_WORDS_PORCELAIN;
5067 else if (!strcmp(arg, "none"))
5068 options->word_diff = DIFF_WORDS_NONE;
5069 else
5070 die("bad --word-diff argument: %s", arg);
5071 }
5072 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
5073 if (options->word_diff == DIFF_WORDS_NONE)
5074 options->word_diff = DIFF_WORDS_PLAIN;
5075 options->word_regex = optarg;
5076 return argcount;
5077 }
5078 else if (!strcmp(arg, "--exit-code"))
5079 options->flags.exit_with_status = 1;
5080 else if (!strcmp(arg, "--quiet"))
5081 options->flags.quick = 1;
5082 else if (!strcmp(arg, "--ext-diff"))
5083 options->flags.allow_external = 1;
5084 else if (!strcmp(arg, "--no-ext-diff"))
5085 options->flags.allow_external = 0;
5086 else if (!strcmp(arg, "--textconv")) {
5087 options->flags.allow_textconv = 1;
5088 options->flags.textconv_set_via_cmdline = 1;
5089 } else if (!strcmp(arg, "--no-textconv"))
5090 options->flags.allow_textconv = 0;
5091 else if (skip_to_optional_arg_default(arg, "--ignore-submodules", &arg, "all")) {
5092 options->flags.override_submodule_config = 1;
5093 handle_ignore_submodules_arg(options, arg);
5094 } else if (skip_to_optional_arg_default(arg, "--submodule", &arg, "log"))
5095 return parse_submodule_opt(options, arg);
5096 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
5097 return parse_ws_error_highlight_opt(options, arg);
5098 else if (!strcmp(arg, "--ita-invisible-in-index"))
5099 options->ita_invisible_in_index = 1;
5100 else if (!strcmp(arg, "--ita-visible-in-index"))
5101 options->ita_invisible_in_index = 0;
5102
5103 /* misc options */
5104 else if (!strcmp(arg, "-z"))
5105 options->line_termination = 0;
5106 else if ((argcount = short_opt('l', av, &optarg))) {
5107 options->rename_limit = strtoul(optarg, NULL, 10);
5108 return argcount;
5109 }
5110 else if ((argcount = short_opt('S', av, &optarg))) {
5111 options->pickaxe = optarg;
5112 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5113 return argcount;
5114 } else if ((argcount = short_opt('G', av, &optarg))) {
5115 options->pickaxe = optarg;
5116 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5117 return argcount;
5118 }
5119 else if (!strcmp(arg, "--pickaxe-all"))
5120 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
5121 else if (!strcmp(arg, "--pickaxe-regex"))
5122 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
5123 else if ((argcount = short_opt('O', av, &optarg))) {
5124 options->orderfile = prefix_filename(prefix, optarg);
5125 return argcount;
5126 } else if (skip_prefix(arg, "--find-object=", &arg))
5127 return parse_objfind_opt(options, arg);
5128 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
5129 int offending = parse_diff_filter_opt(optarg, options);
5130 if (offending)
5131 die("unknown change class '%c' in --diff-filter=%s",
5132 offending, optarg);
5133 return argcount;
5134 }
5135 else if (!strcmp(arg, "--no-abbrev"))
5136 options->abbrev = 0;
5137 else if (!strcmp(arg, "--abbrev"))
5138 options->abbrev = DEFAULT_ABBREV;
5139 else if (skip_prefix(arg, "--abbrev=", &arg)) {
5140 options->abbrev = strtoul(arg, NULL, 10);
5141 if (options->abbrev < MINIMUM_ABBREV)
5142 options->abbrev = MINIMUM_ABBREV;
5143 else if (the_hash_algo->hexsz < options->abbrev)
5144 options->abbrev = the_hash_algo->hexsz;
5145 }
5146 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
5147 options->a_prefix = optarg;
5148 return argcount;
5149 }
5150 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
5151 options->line_prefix = optarg;
5152 options->line_prefix_length = strlen(options->line_prefix);
5153 graph_setup_line_prefix(options);
5154 return argcount;
5155 }
5156 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
5157 options->b_prefix = optarg;
5158 return argcount;
5159 }
5160 else if (!strcmp(arg, "--no-prefix"))
5161 options->a_prefix = options->b_prefix = "";
5162 else if (opt_arg(arg, '\0', "inter-hunk-context",
5163 &options->interhunkcontext))
5164 ;
5165 else if (!strcmp(arg, "-W"))
5166 options->flags.funccontext = 1;
5167 else if (!strcmp(arg, "--function-context"))
5168 options->flags.funccontext = 1;
5169 else if (!strcmp(arg, "--no-function-context"))
5170 options->flags.funccontext = 0;
5171 else if ((argcount = parse_long_opt("output", av, &optarg))) {
5172 char *path = prefix_filename(prefix, optarg);
5173 options->file = xfopen(path, "w");
5174 options->close_file = 1;
5175 if (options->use_color != GIT_COLOR_ALWAYS)
5176 options->use_color = GIT_COLOR_NEVER;
5177 free(path);
5178 return argcount;
5179 } else
5180 return 0;
5181 return 1;
5182 }
5183
5184 int parse_rename_score(const char **cp_p)
5185 {
5186 unsigned long num, scale;
5187 int ch, dot;
5188 const char *cp = *cp_p;
5189
5190 num = 0;
5191 scale = 1;
5192 dot = 0;
5193 for (;;) {
5194 ch = *cp;
5195 if ( !dot && ch == '.' ) {
5196 scale = 1;
5197 dot = 1;
5198 } else if ( ch == '%' ) {
5199 scale = dot ? scale*100 : 100;
5200 cp++; /* % is always at the end */
5201 break;
5202 } else if ( ch >= '0' && ch <= '9' ) {
5203 if ( scale < 100000 ) {
5204 scale *= 10;
5205 num = (num*10) + (ch-'0');
5206 }
5207 } else {
5208 break;
5209 }
5210 cp++;
5211 }
5212 *cp_p = cp;
5213
5214 /* user says num divided by scale and we say internally that
5215 * is MAX_SCORE * num / scale.
5216 */
5217 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5218 }
5219
5220 static int diff_scoreopt_parse(const char *opt)
5221 {
5222 int opt1, opt2, cmd;
5223
5224 if (*opt++ != '-')
5225 return -1;
5226 cmd = *opt++;
5227 if (cmd == '-') {
5228 /* convert the long-form arguments into short-form versions */
5229 if (skip_prefix(opt, "break-rewrites", &opt)) {
5230 if (*opt == 0 || *opt++ == '=')
5231 cmd = 'B';
5232 } else if (skip_prefix(opt, "find-copies", &opt)) {
5233 if (*opt == 0 || *opt++ == '=')
5234 cmd = 'C';
5235 } else if (skip_prefix(opt, "find-renames", &opt)) {
5236 if (*opt == 0 || *opt++ == '=')
5237 cmd = 'M';
5238 }
5239 }
5240 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
5241 return -1; /* that is not a -M, -C, or -B option */
5242
5243 opt1 = parse_rename_score(&opt);
5244 if (cmd != 'B')
5245 opt2 = 0;
5246 else {
5247 if (*opt == 0)
5248 opt2 = 0;
5249 else if (*opt != '/')
5250 return -1; /* we expect -B80/99 or -B80 */
5251 else {
5252 opt++;
5253 opt2 = parse_rename_score(&opt);
5254 }
5255 }
5256 if (*opt != 0)
5257 return -1;
5258 return opt1 | (opt2 << 16);
5259 }
5260
5261 struct diff_queue_struct diff_queued_diff;
5262
5263 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5264 {
5265 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5266 queue->queue[queue->nr++] = dp;
5267 }
5268
5269 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5270 struct diff_filespec *one,
5271 struct diff_filespec *two)
5272 {
5273 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5274 dp->one = one;
5275 dp->two = two;
5276 if (queue)
5277 diff_q(queue, dp);
5278 return dp;
5279 }
5280
5281 void diff_free_filepair(struct diff_filepair *p)
5282 {
5283 free_filespec(p->one);
5284 free_filespec(p->two);
5285 free(p);
5286 }
5287
5288 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5289 {
5290 int abblen;
5291 const char *abbrev;
5292
5293 /* Do we want all 40 hex characters? */
5294 if (len == the_hash_algo->hexsz)
5295 return oid_to_hex(oid);
5296
5297 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5298 abbrev = diff_abbrev_oid(oid, len);
5299
5300 if (!print_sha1_ellipsis())
5301 return abbrev;
5302
5303 abblen = strlen(abbrev);
5304
5305 /*
5306 * In well-behaved cases, where the abbreviated result is the
5307 * same as the requested length, append three dots after the
5308 * abbreviation (hence the whole logic is limited to the case
5309 * where abblen < 37); when the actual abbreviated result is a
5310 * bit longer than the requested length, we reduce the number
5311 * of dots so that they match the well-behaved ones. However,
5312 * if the actual abbreviation is longer than the requested
5313 * length by more than three, we give up on aligning, and add
5314 * three dots anyway, to indicate that the output is not the
5315 * full object name. Yes, this may be suboptimal, but this
5316 * appears only in "diff --raw --abbrev" output and it is not
5317 * worth the effort to change it now. Note that this would
5318 * likely to work fine when the automatic sizing of default
5319 * abbreviation length is used--we would be fed -1 in "len" in
5320 * that case, and will end up always appending three-dots, but
5321 * the automatic sizing is supposed to give abblen that ensures
5322 * uniqueness across all objects (statistically speaking).
5323 */
5324 if (abblen < the_hash_algo->hexsz - 3) {
5325 static char hex[GIT_MAX_HEXSZ + 1];
5326 if (len < abblen && abblen <= len + 2)
5327 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5328 else
5329 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5330 return hex;
5331 }
5332
5333 return oid_to_hex(oid);
5334 }
5335
5336 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5337 {
5338 int line_termination = opt->line_termination;
5339 int inter_name_termination = line_termination ? '\t' : '\0';
5340
5341 fprintf(opt->file, "%s", diff_line_prefix(opt));
5342 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5343 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5344 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5345 fprintf(opt->file, "%s ",
5346 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5347 }
5348 if (p->score) {
5349 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5350 inter_name_termination);
5351 } else {
5352 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5353 }
5354
5355 if (p->status == DIFF_STATUS_COPIED ||
5356 p->status == DIFF_STATUS_RENAMED) {
5357 const char *name_a, *name_b;
5358 name_a = p->one->path;
5359 name_b = p->two->path;
5360 strip_prefix(opt->prefix_length, &name_a, &name_b);
5361 write_name_quoted(name_a, opt->file, inter_name_termination);
5362 write_name_quoted(name_b, opt->file, line_termination);
5363 } else {
5364 const char *name_a, *name_b;
5365 name_a = p->one->mode ? p->one->path : p->two->path;
5366 name_b = NULL;
5367 strip_prefix(opt->prefix_length, &name_a, &name_b);
5368 write_name_quoted(name_a, opt->file, line_termination);
5369 }
5370 }
5371
5372 int diff_unmodified_pair(struct diff_filepair *p)
5373 {
5374 /* This function is written stricter than necessary to support
5375 * the currently implemented transformers, but the idea is to
5376 * let transformers to produce diff_filepairs any way they want,
5377 * and filter and clean them up here before producing the output.
5378 */
5379 struct diff_filespec *one = p->one, *two = p->two;
5380
5381 if (DIFF_PAIR_UNMERGED(p))
5382 return 0; /* unmerged is interesting */
5383
5384 /* deletion, addition, mode or type change
5385 * and rename are all interesting.
5386 */
5387 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5388 DIFF_PAIR_MODE_CHANGED(p) ||
5389 strcmp(one->path, two->path))
5390 return 0;
5391
5392 /* both are valid and point at the same path. that is, we are
5393 * dealing with a change.
5394 */
5395 if (one->oid_valid && two->oid_valid &&
5396 oideq(&one->oid, &two->oid) &&
5397 !one->dirty_submodule && !two->dirty_submodule)
5398 return 1; /* no change */
5399 if (!one->oid_valid && !two->oid_valid)
5400 return 1; /* both look at the same file on the filesystem. */
5401 return 0;
5402 }
5403
5404 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5405 {
5406 if (diff_unmodified_pair(p))
5407 return;
5408
5409 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5410 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5411 return; /* no tree diffs in patch format */
5412
5413 run_diff(p, o);
5414 }
5415
5416 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5417 struct diffstat_t *diffstat)
5418 {
5419 if (diff_unmodified_pair(p))
5420 return;
5421
5422 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5423 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5424 return; /* no useful stat for tree diffs */
5425
5426 run_diffstat(p, o, diffstat);
5427 }
5428
5429 static void diff_flush_checkdiff(struct diff_filepair *p,
5430 struct diff_options *o)
5431 {
5432 if (diff_unmodified_pair(p))
5433 return;
5434
5435 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5436 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5437 return; /* nothing to check in tree diffs */
5438
5439 run_checkdiff(p, o);
5440 }
5441
5442 int diff_queue_is_empty(void)
5443 {
5444 struct diff_queue_struct *q = &diff_queued_diff;
5445 int i;
5446 for (i = 0; i < q->nr; i++)
5447 if (!diff_unmodified_pair(q->queue[i]))
5448 return 0;
5449 return 1;
5450 }
5451
5452 #if DIFF_DEBUG
5453 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5454 {
5455 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5456 x, one ? one : "",
5457 s->path,
5458 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5459 s->mode,
5460 s->oid_valid ? oid_to_hex(&s->oid) : "");
5461 fprintf(stderr, "queue[%d] %s size %lu\n",
5462 x, one ? one : "",
5463 s->size);
5464 }
5465
5466 void diff_debug_filepair(const struct diff_filepair *p, int i)
5467 {
5468 diff_debug_filespec(p->one, i, "one");
5469 diff_debug_filespec(p->two, i, "two");
5470 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5471 p->score, p->status ? p->status : '?',
5472 p->one->rename_used, p->broken_pair);
5473 }
5474
5475 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5476 {
5477 int i;
5478 if (msg)
5479 fprintf(stderr, "%s\n", msg);
5480 fprintf(stderr, "q->nr = %d\n", q->nr);
5481 for (i = 0; i < q->nr; i++) {
5482 struct diff_filepair *p = q->queue[i];
5483 diff_debug_filepair(p, i);
5484 }
5485 }
5486 #endif
5487
5488 static void diff_resolve_rename_copy(void)
5489 {
5490 int i;
5491 struct diff_filepair *p;
5492 struct diff_queue_struct *q = &diff_queued_diff;
5493
5494 diff_debug_queue("resolve-rename-copy", q);
5495
5496 for (i = 0; i < q->nr; i++) {
5497 p = q->queue[i];
5498 p->status = 0; /* undecided */
5499 if (DIFF_PAIR_UNMERGED(p))
5500 p->status = DIFF_STATUS_UNMERGED;
5501 else if (!DIFF_FILE_VALID(p->one))
5502 p->status = DIFF_STATUS_ADDED;
5503 else if (!DIFF_FILE_VALID(p->two))
5504 p->status = DIFF_STATUS_DELETED;
5505 else if (DIFF_PAIR_TYPE_CHANGED(p))
5506 p->status = DIFF_STATUS_TYPE_CHANGED;
5507
5508 /* from this point on, we are dealing with a pair
5509 * whose both sides are valid and of the same type, i.e.
5510 * either in-place edit or rename/copy edit.
5511 */
5512 else if (DIFF_PAIR_RENAME(p)) {
5513 /*
5514 * A rename might have re-connected a broken
5515 * pair up, causing the pathnames to be the
5516 * same again. If so, that's not a rename at
5517 * all, just a modification..
5518 *
5519 * Otherwise, see if this source was used for
5520 * multiple renames, in which case we decrement
5521 * the count, and call it a copy.
5522 */
5523 if (!strcmp(p->one->path, p->two->path))
5524 p->status = DIFF_STATUS_MODIFIED;
5525 else if (--p->one->rename_used > 0)
5526 p->status = DIFF_STATUS_COPIED;
5527 else
5528 p->status = DIFF_STATUS_RENAMED;
5529 }
5530 else if (!oideq(&p->one->oid, &p->two->oid) ||
5531 p->one->mode != p->two->mode ||
5532 p->one->dirty_submodule ||
5533 p->two->dirty_submodule ||
5534 is_null_oid(&p->one->oid))
5535 p->status = DIFF_STATUS_MODIFIED;
5536 else {
5537 /* This is a "no-change" entry and should not
5538 * happen anymore, but prepare for broken callers.
5539 */
5540 error("feeding unmodified %s to diffcore",
5541 p->one->path);
5542 p->status = DIFF_STATUS_UNKNOWN;
5543 }
5544 }
5545 diff_debug_queue("resolve-rename-copy done", q);
5546 }
5547
5548 static int check_pair_status(struct diff_filepair *p)
5549 {
5550 switch (p->status) {
5551 case DIFF_STATUS_UNKNOWN:
5552 return 0;
5553 case 0:
5554 die("internal error in diff-resolve-rename-copy");
5555 default:
5556 return 1;
5557 }
5558 }
5559
5560 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5561 {
5562 int fmt = opt->output_format;
5563
5564 if (fmt & DIFF_FORMAT_CHECKDIFF)
5565 diff_flush_checkdiff(p, opt);
5566 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5567 diff_flush_raw(p, opt);
5568 else if (fmt & DIFF_FORMAT_NAME) {
5569 const char *name_a, *name_b;
5570 name_a = p->two->path;
5571 name_b = NULL;
5572 strip_prefix(opt->prefix_length, &name_a, &name_b);
5573 fprintf(opt->file, "%s", diff_line_prefix(opt));
5574 write_name_quoted(name_a, opt->file, opt->line_termination);
5575 }
5576 }
5577
5578 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5579 {
5580 struct strbuf sb = STRBUF_INIT;
5581 if (fs->mode)
5582 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5583 else
5584 strbuf_addf(&sb, " %s ", newdelete);
5585
5586 quote_c_style(fs->path, &sb, NULL, 0);
5587 strbuf_addch(&sb, '\n');
5588 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5589 sb.buf, sb.len, 0);
5590 strbuf_release(&sb);
5591 }
5592
5593 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5594 int show_name)
5595 {
5596 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5597 struct strbuf sb = STRBUF_INIT;
5598 strbuf_addf(&sb, " mode change %06o => %06o",
5599 p->one->mode, p->two->mode);
5600 if (show_name) {
5601 strbuf_addch(&sb, ' ');
5602 quote_c_style(p->two->path, &sb, NULL, 0);
5603 }
5604 strbuf_addch(&sb, '\n');
5605 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5606 sb.buf, sb.len, 0);
5607 strbuf_release(&sb);
5608 }
5609 }
5610
5611 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5612 struct diff_filepair *p)
5613 {
5614 struct strbuf sb = STRBUF_INIT;
5615 struct strbuf names = STRBUF_INIT;
5616
5617 pprint_rename(&names, p->one->path, p->two->path);
5618 strbuf_addf(&sb, " %s %s (%d%%)\n",
5619 renamecopy, names.buf, similarity_index(p));
5620 strbuf_release(&names);
5621 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5622 sb.buf, sb.len, 0);
5623 show_mode_change(opt, p, 0);
5624 strbuf_release(&sb);
5625 }
5626
5627 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5628 {
5629 switch(p->status) {
5630 case DIFF_STATUS_DELETED:
5631 show_file_mode_name(opt, "delete", p->one);
5632 break;
5633 case DIFF_STATUS_ADDED:
5634 show_file_mode_name(opt, "create", p->two);
5635 break;
5636 case DIFF_STATUS_COPIED:
5637 show_rename_copy(opt, "copy", p);
5638 break;
5639 case DIFF_STATUS_RENAMED:
5640 show_rename_copy(opt, "rename", p);
5641 break;
5642 default:
5643 if (p->score) {
5644 struct strbuf sb = STRBUF_INIT;
5645 strbuf_addstr(&sb, " rewrite ");
5646 quote_c_style(p->two->path, &sb, NULL, 0);
5647 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5648 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5649 sb.buf, sb.len, 0);
5650 strbuf_release(&sb);
5651 }
5652 show_mode_change(opt, p, !p->score);
5653 break;
5654 }
5655 }
5656
5657 struct patch_id_t {
5658 git_SHA_CTX *ctx;
5659 int patchlen;
5660 };
5661
5662 static int remove_space(char *line, int len)
5663 {
5664 int i;
5665 char *dst = line;
5666 unsigned char c;
5667
5668 for (i = 0; i < len; i++)
5669 if (!isspace((c = line[i])))
5670 *dst++ = c;
5671
5672 return dst - line;
5673 }
5674
5675 static void patch_id_consume(void *priv, char *line, unsigned long len)
5676 {
5677 struct patch_id_t *data = priv;
5678 int new_len;
5679
5680 new_len = remove_space(line, len);
5681
5682 git_SHA1_Update(data->ctx, line, new_len);
5683 data->patchlen += new_len;
5684 }
5685
5686 static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5687 {
5688 git_SHA1_Update(ctx, str, strlen(str));
5689 }
5690
5691 static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5692 {
5693 /* large enough for 2^32 in octal */
5694 char buf[12];
5695 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5696 git_SHA1_Update(ctx, buf, len);
5697 }
5698
5699 /* returns 0 upon success, and writes result into sha1 */
5700 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5701 {
5702 struct diff_queue_struct *q = &diff_queued_diff;
5703 int i;
5704 git_SHA_CTX ctx;
5705 struct patch_id_t data;
5706
5707 git_SHA1_Init(&ctx);
5708 memset(&data, 0, sizeof(struct patch_id_t));
5709 data.ctx = &ctx;
5710
5711 for (i = 0; i < q->nr; i++) {
5712 xpparam_t xpp;
5713 xdemitconf_t xecfg;
5714 mmfile_t mf1, mf2;
5715 struct diff_filepair *p = q->queue[i];
5716 int len1, len2;
5717
5718 memset(&xpp, 0, sizeof(xpp));
5719 memset(&xecfg, 0, sizeof(xecfg));
5720 if (p->status == 0)
5721 return error("internal diff status error");
5722 if (p->status == DIFF_STATUS_UNKNOWN)
5723 continue;
5724 if (diff_unmodified_pair(p))
5725 continue;
5726 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5727 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5728 continue;
5729 if (DIFF_PAIR_UNMERGED(p))
5730 continue;
5731
5732 diff_fill_oid_info(p->one, options->repo->index);
5733 diff_fill_oid_info(p->two, options->repo->index);
5734
5735 len1 = remove_space(p->one->path, strlen(p->one->path));
5736 len2 = remove_space(p->two->path, strlen(p->two->path));
5737 patch_id_add_string(&ctx, "diff--git");
5738 patch_id_add_string(&ctx, "a/");
5739 git_SHA1_Update(&ctx, p->one->path, len1);
5740 patch_id_add_string(&ctx, "b/");
5741 git_SHA1_Update(&ctx, p->two->path, len2);
5742
5743 if (p->one->mode == 0) {
5744 patch_id_add_string(&ctx, "newfilemode");
5745 patch_id_add_mode(&ctx, p->two->mode);
5746 patch_id_add_string(&ctx, "---/dev/null");
5747 patch_id_add_string(&ctx, "+++b/");
5748 git_SHA1_Update(&ctx, p->two->path, len2);
5749 } else if (p->two->mode == 0) {
5750 patch_id_add_string(&ctx, "deletedfilemode");
5751 patch_id_add_mode(&ctx, p->one->mode);
5752 patch_id_add_string(&ctx, "---a/");
5753 git_SHA1_Update(&ctx, p->one->path, len1);
5754 patch_id_add_string(&ctx, "+++/dev/null");
5755 } else {
5756 patch_id_add_string(&ctx, "---a/");
5757 git_SHA1_Update(&ctx, p->one->path, len1);
5758 patch_id_add_string(&ctx, "+++b/");
5759 git_SHA1_Update(&ctx, p->two->path, len2);
5760 }
5761
5762 if (diff_header_only)
5763 continue;
5764
5765 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
5766 fill_mmfile(options->repo, &mf2, p->two) < 0)
5767 return error("unable to read files to diff");
5768
5769 if (diff_filespec_is_binary(options->repo, p->one) ||
5770 diff_filespec_is_binary(options->repo, p->two)) {
5771 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5772 GIT_SHA1_HEXSZ);
5773 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5774 GIT_SHA1_HEXSZ);
5775 continue;
5776 }
5777
5778 xpp.flags = 0;
5779 xecfg.ctxlen = 3;
5780 xecfg.flags = 0;
5781 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
5782 patch_id_consume, &data, &xpp, &xecfg))
5783 return error("unable to generate patch-id diff for %s",
5784 p->one->path);
5785 }
5786
5787 git_SHA1_Final(oid->hash, &ctx);
5788 return 0;
5789 }
5790
5791 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5792 {
5793 struct diff_queue_struct *q = &diff_queued_diff;
5794 int i;
5795 int result = diff_get_patch_id(options, oid, diff_header_only);
5796
5797 for (i = 0; i < q->nr; i++)
5798 diff_free_filepair(q->queue[i]);
5799
5800 free(q->queue);
5801 DIFF_QUEUE_CLEAR(q);
5802
5803 return result;
5804 }
5805
5806 static int is_summary_empty(const struct diff_queue_struct *q)
5807 {
5808 int i;
5809
5810 for (i = 0; i < q->nr; i++) {
5811 const struct diff_filepair *p = q->queue[i];
5812
5813 switch (p->status) {
5814 case DIFF_STATUS_DELETED:
5815 case DIFF_STATUS_ADDED:
5816 case DIFF_STATUS_COPIED:
5817 case DIFF_STATUS_RENAMED:
5818 return 0;
5819 default:
5820 if (p->score)
5821 return 0;
5822 if (p->one->mode && p->two->mode &&
5823 p->one->mode != p->two->mode)
5824 return 0;
5825 break;
5826 }
5827 }
5828 return 1;
5829 }
5830
5831 static const char rename_limit_warning[] =
5832 N_("inexact rename detection was skipped due to too many files.");
5833
5834 static const char degrade_cc_to_c_warning[] =
5835 N_("only found copies from modified paths due to too many files.");
5836
5837 static const char rename_limit_advice[] =
5838 N_("you may want to set your %s variable to at least "
5839 "%d and retry the command.");
5840
5841 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5842 {
5843 fflush(stdout);
5844 if (degraded_cc)
5845 warning(_(degrade_cc_to_c_warning));
5846 else if (needed)
5847 warning(_(rename_limit_warning));
5848 else
5849 return;
5850 if (0 < needed)
5851 warning(_(rename_limit_advice), varname, needed);
5852 }
5853
5854 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5855 {
5856 int i;
5857 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5858 struct diff_queue_struct *q = &diff_queued_diff;
5859
5860 if (WSEH_NEW & WS_RULE_MASK)
5861 BUG("WS rules bit mask overlaps with diff symbol flags");
5862
5863 if (o->color_moved)
5864 o->emitted_symbols = &esm;
5865
5866 for (i = 0; i < q->nr; i++) {
5867 struct diff_filepair *p = q->queue[i];
5868 if (check_pair_status(p))
5869 diff_flush_patch(p, o);
5870 }
5871
5872 if (o->emitted_symbols) {
5873 if (o->color_moved) {
5874 struct hashmap add_lines, del_lines;
5875
5876 if (o->color_moved_ws_handling &
5877 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
5878 o->color_moved_ws_handling |= XDF_IGNORE_WHITESPACE;
5879
5880 hashmap_init(&del_lines, moved_entry_cmp, o, 0);
5881 hashmap_init(&add_lines, moved_entry_cmp, o, 0);
5882
5883 add_lines_to_move_detection(o, &add_lines, &del_lines);
5884 mark_color_as_moved(o, &add_lines, &del_lines);
5885 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5886 dim_moved_lines(o);
5887
5888 hashmap_free(&add_lines, 1);
5889 hashmap_free(&del_lines, 1);
5890 }
5891
5892 for (i = 0; i < esm.nr; i++)
5893 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5894
5895 for (i = 0; i < esm.nr; i++)
5896 free((void *)esm.buf[i].line);
5897 }
5898 esm.nr = 0;
5899 }
5900
5901 void diff_flush(struct diff_options *options)
5902 {
5903 struct diff_queue_struct *q = &diff_queued_diff;
5904 int i, output_format = options->output_format;
5905 int separator = 0;
5906 int dirstat_by_line = 0;
5907
5908 /*
5909 * Order: raw, stat, summary, patch
5910 * or: name/name-status/checkdiff (other bits clear)
5911 */
5912 if (!q->nr)
5913 goto free_queue;
5914
5915 if (output_format & (DIFF_FORMAT_RAW |
5916 DIFF_FORMAT_NAME |
5917 DIFF_FORMAT_NAME_STATUS |
5918 DIFF_FORMAT_CHECKDIFF)) {
5919 for (i = 0; i < q->nr; i++) {
5920 struct diff_filepair *p = q->queue[i];
5921 if (check_pair_status(p))
5922 flush_one_pair(p, options);
5923 }
5924 separator++;
5925 }
5926
5927 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
5928 dirstat_by_line = 1;
5929
5930 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5931 dirstat_by_line) {
5932 struct diffstat_t diffstat;
5933
5934 memset(&diffstat, 0, sizeof(struct diffstat_t));
5935 for (i = 0; i < q->nr; i++) {
5936 struct diff_filepair *p = q->queue[i];
5937 if (check_pair_status(p))
5938 diff_flush_stat(p, options, &diffstat);
5939 }
5940 if (output_format & DIFF_FORMAT_NUMSTAT)
5941 show_numstat(&diffstat, options);
5942 if (output_format & DIFF_FORMAT_DIFFSTAT)
5943 show_stats(&diffstat, options);
5944 if (output_format & DIFF_FORMAT_SHORTSTAT)
5945 show_shortstats(&diffstat, options);
5946 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5947 show_dirstat_by_line(&diffstat, options);
5948 free_diffstat_info(&diffstat);
5949 separator++;
5950 }
5951 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5952 show_dirstat(options);
5953
5954 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5955 for (i = 0; i < q->nr; i++) {
5956 diff_summary(options, q->queue[i]);
5957 }
5958 separator++;
5959 }
5960
5961 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5962 options->flags.exit_with_status &&
5963 options->flags.diff_from_contents) {
5964 /*
5965 * run diff_flush_patch for the exit status. setting
5966 * options->file to /dev/null should be safe, because we
5967 * aren't supposed to produce any output anyway.
5968 */
5969 if (options->close_file)
5970 fclose(options->file);
5971 options->file = xfopen("/dev/null", "w");
5972 options->close_file = 1;
5973 options->color_moved = 0;
5974 for (i = 0; i < q->nr; i++) {
5975 struct diff_filepair *p = q->queue[i];
5976 if (check_pair_status(p))
5977 diff_flush_patch(p, options);
5978 if (options->found_changes)
5979 break;
5980 }
5981 }
5982
5983 if (output_format & DIFF_FORMAT_PATCH) {
5984 if (separator) {
5985 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5986 if (options->stat_sep)
5987 /* attach patch instead of inline */
5988 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5989 NULL, 0, 0);
5990 }
5991
5992 diff_flush_patch_all_file_pairs(options);
5993 }
5994
5995 if (output_format & DIFF_FORMAT_CALLBACK)
5996 options->format_callback(q, options, options->format_callback_data);
5997
5998 for (i = 0; i < q->nr; i++)
5999 diff_free_filepair(q->queue[i]);
6000 free_queue:
6001 free(q->queue);
6002 DIFF_QUEUE_CLEAR(q);
6003 if (options->close_file)
6004 fclose(options->file);
6005
6006 /*
6007 * Report the content-level differences with HAS_CHANGES;
6008 * diff_addremove/diff_change does not set the bit when
6009 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6010 */
6011 if (options->flags.diff_from_contents) {
6012 if (options->found_changes)
6013 options->flags.has_changes = 1;
6014 else
6015 options->flags.has_changes = 0;
6016 }
6017 }
6018
6019 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6020 {
6021 return (((p->status == DIFF_STATUS_MODIFIED) &&
6022 ((p->score &&
6023 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6024 (!p->score &&
6025 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6026 ((p->status != DIFF_STATUS_MODIFIED) &&
6027 filter_bit_tst(p->status, options)));
6028 }
6029
6030 static void diffcore_apply_filter(struct diff_options *options)
6031 {
6032 int i;
6033 struct diff_queue_struct *q = &diff_queued_diff;
6034 struct diff_queue_struct outq;
6035
6036 DIFF_QUEUE_CLEAR(&outq);
6037
6038 if (!options->filter)
6039 return;
6040
6041 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6042 int found;
6043 for (i = found = 0; !found && i < q->nr; i++) {
6044 if (match_filter(options, q->queue[i]))
6045 found++;
6046 }
6047 if (found)
6048 return;
6049
6050 /* otherwise we will clear the whole queue
6051 * by copying the empty outq at the end of this
6052 * function, but first clear the current entries
6053 * in the queue.
6054 */
6055 for (i = 0; i < q->nr; i++)
6056 diff_free_filepair(q->queue[i]);
6057 }
6058 else {
6059 /* Only the matching ones */
6060 for (i = 0; i < q->nr; i++) {
6061 struct diff_filepair *p = q->queue[i];
6062 if (match_filter(options, p))
6063 diff_q(&outq, p);
6064 else
6065 diff_free_filepair(p);
6066 }
6067 }
6068 free(q->queue);
6069 *q = outq;
6070 }
6071
6072 /* Check whether two filespecs with the same mode and size are identical */
6073 static int diff_filespec_is_identical(struct repository *r,
6074 struct diff_filespec *one,
6075 struct diff_filespec *two)
6076 {
6077 if (S_ISGITLINK(one->mode))
6078 return 0;
6079 if (diff_populate_filespec(r, one, 0))
6080 return 0;
6081 if (diff_populate_filespec(r, two, 0))
6082 return 0;
6083 return !memcmp(one->data, two->data, one->size);
6084 }
6085
6086 static int diff_filespec_check_stat_unmatch(struct repository *r,
6087 struct diff_filepair *p)
6088 {
6089 if (p->done_skip_stat_unmatch)
6090 return p->skip_stat_unmatch_result;
6091
6092 p->done_skip_stat_unmatch = 1;
6093 p->skip_stat_unmatch_result = 0;
6094 /*
6095 * 1. Entries that come from stat info dirtiness
6096 * always have both sides (iow, not create/delete),
6097 * one side of the object name is unknown, with
6098 * the same mode and size. Keep the ones that
6099 * do not match these criteria. They have real
6100 * differences.
6101 *
6102 * 2. At this point, the file is known to be modified,
6103 * with the same mode and size, and the object
6104 * name of one side is unknown. Need to inspect
6105 * the identical contents.
6106 */
6107 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6108 !DIFF_FILE_VALID(p->two) ||
6109 (p->one->oid_valid && p->two->oid_valid) ||
6110 (p->one->mode != p->two->mode) ||
6111 diff_populate_filespec(r, p->one, CHECK_SIZE_ONLY) ||
6112 diff_populate_filespec(r, p->two, CHECK_SIZE_ONLY) ||
6113 (p->one->size != p->two->size) ||
6114 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6115 p->skip_stat_unmatch_result = 1;
6116 return p->skip_stat_unmatch_result;
6117 }
6118
6119 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6120 {
6121 int i;
6122 struct diff_queue_struct *q = &diff_queued_diff;
6123 struct diff_queue_struct outq;
6124 DIFF_QUEUE_CLEAR(&outq);
6125
6126 for (i = 0; i < q->nr; i++) {
6127 struct diff_filepair *p = q->queue[i];
6128
6129 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6130 diff_q(&outq, p);
6131 else {
6132 /*
6133 * The caller can subtract 1 from skip_stat_unmatch
6134 * to determine how many paths were dirty only
6135 * due to stat info mismatch.
6136 */
6137 if (!diffopt->flags.no_index)
6138 diffopt->skip_stat_unmatch++;
6139 diff_free_filepair(p);
6140 }
6141 }
6142 free(q->queue);
6143 *q = outq;
6144 }
6145
6146 static int diffnamecmp(const void *a_, const void *b_)
6147 {
6148 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6149 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6150 const char *name_a, *name_b;
6151
6152 name_a = a->one ? a->one->path : a->two->path;
6153 name_b = b->one ? b->one->path : b->two->path;
6154 return strcmp(name_a, name_b);
6155 }
6156
6157 void diffcore_fix_diff_index(struct diff_options *options)
6158 {
6159 struct diff_queue_struct *q = &diff_queued_diff;
6160 QSORT(q->queue, q->nr, diffnamecmp);
6161 }
6162
6163 void diffcore_std(struct diff_options *options)
6164 {
6165 /* NOTE please keep the following in sync with diff_tree_combined() */
6166 if (options->skip_stat_unmatch)
6167 diffcore_skip_stat_unmatch(options);
6168 if (!options->found_follow) {
6169 /* See try_to_follow_renames() in tree-diff.c */
6170 if (options->break_opt != -1)
6171 diffcore_break(options->repo,
6172 options->break_opt);
6173 if (options->detect_rename)
6174 diffcore_rename(options);
6175 if (options->break_opt != -1)
6176 diffcore_merge_broken();
6177 }
6178 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6179 diffcore_pickaxe(options);
6180 if (options->orderfile)
6181 diffcore_order(options->orderfile);
6182 if (!options->found_follow)
6183 /* See try_to_follow_renames() in tree-diff.c */
6184 diff_resolve_rename_copy();
6185 diffcore_apply_filter(options);
6186
6187 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6188 options->flags.has_changes = 1;
6189 else
6190 options->flags.has_changes = 0;
6191
6192 options->found_follow = 0;
6193 }
6194
6195 int diff_result_code(struct diff_options *opt, int status)
6196 {
6197 int result = 0;
6198
6199 diff_warn_rename_limit("diff.renameLimit",
6200 opt->needed_rename_limit,
6201 opt->degraded_cc_to_c);
6202 if (!opt->flags.exit_with_status &&
6203 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6204 return status;
6205 if (opt->flags.exit_with_status &&
6206 opt->flags.has_changes)
6207 result |= 01;
6208 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6209 opt->flags.check_failed)
6210 result |= 02;
6211 return result;
6212 }
6213
6214 int diff_can_quit_early(struct diff_options *opt)
6215 {
6216 return (opt->flags.quick &&
6217 !opt->filter &&
6218 opt->flags.has_changes);
6219 }
6220
6221 /*
6222 * Shall changes to this submodule be ignored?
6223 *
6224 * Submodule changes can be configured to be ignored separately for each path,
6225 * but that configuration can be overridden from the command line.
6226 */
6227 static int is_submodule_ignored(const char *path, struct diff_options *options)
6228 {
6229 int ignored = 0;
6230 struct diff_flags orig_flags = options->flags;
6231 if (!options->flags.override_submodule_config)
6232 set_diffopt_flags_from_submodule_config(options, path);
6233 if (options->flags.ignore_submodules)
6234 ignored = 1;
6235 options->flags = orig_flags;
6236 return ignored;
6237 }
6238
6239 void diff_addremove(struct diff_options *options,
6240 int addremove, unsigned mode,
6241 const struct object_id *oid,
6242 int oid_valid,
6243 const char *concatpath, unsigned dirty_submodule)
6244 {
6245 struct diff_filespec *one, *two;
6246
6247 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6248 return;
6249
6250 /* This may look odd, but it is a preparation for
6251 * feeding "there are unchanged files which should
6252 * not produce diffs, but when you are doing copy
6253 * detection you would need them, so here they are"
6254 * entries to the diff-core. They will be prefixed
6255 * with something like '=' or '*' (I haven't decided
6256 * which but should not make any difference).
6257 * Feeding the same new and old to diff_change()
6258 * also has the same effect.
6259 * Before the final output happens, they are pruned after
6260 * merged into rename/copy pairs as appropriate.
6261 */
6262 if (options->flags.reverse_diff)
6263 addremove = (addremove == '+' ? '-' :
6264 addremove == '-' ? '+' : addremove);
6265
6266 if (options->prefix &&
6267 strncmp(concatpath, options->prefix, options->prefix_length))
6268 return;
6269
6270 one = alloc_filespec(concatpath);
6271 two = alloc_filespec(concatpath);
6272
6273 if (addremove != '+')
6274 fill_filespec(one, oid, oid_valid, mode);
6275 if (addremove != '-') {
6276 fill_filespec(two, oid, oid_valid, mode);
6277 two->dirty_submodule = dirty_submodule;
6278 }
6279
6280 diff_queue(&diff_queued_diff, one, two);
6281 if (!options->flags.diff_from_contents)
6282 options->flags.has_changes = 1;
6283 }
6284
6285 void diff_change(struct diff_options *options,
6286 unsigned old_mode, unsigned new_mode,
6287 const struct object_id *old_oid,
6288 const struct object_id *new_oid,
6289 int old_oid_valid, int new_oid_valid,
6290 const char *concatpath,
6291 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6292 {
6293 struct diff_filespec *one, *two;
6294 struct diff_filepair *p;
6295
6296 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6297 is_submodule_ignored(concatpath, options))
6298 return;
6299
6300 if (options->flags.reverse_diff) {
6301 SWAP(old_mode, new_mode);
6302 SWAP(old_oid, new_oid);
6303 SWAP(old_oid_valid, new_oid_valid);
6304 SWAP(old_dirty_submodule, new_dirty_submodule);
6305 }
6306
6307 if (options->prefix &&
6308 strncmp(concatpath, options->prefix, options->prefix_length))
6309 return;
6310
6311 one = alloc_filespec(concatpath);
6312 two = alloc_filespec(concatpath);
6313 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6314 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6315 one->dirty_submodule = old_dirty_submodule;
6316 two->dirty_submodule = new_dirty_submodule;
6317 p = diff_queue(&diff_queued_diff, one, two);
6318
6319 if (options->flags.diff_from_contents)
6320 return;
6321
6322 if (options->flags.quick && options->skip_stat_unmatch &&
6323 !diff_filespec_check_stat_unmatch(options->repo, p))
6324 return;
6325
6326 options->flags.has_changes = 1;
6327 }
6328
6329 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6330 {
6331 struct diff_filepair *pair;
6332 struct diff_filespec *one, *two;
6333
6334 if (options->prefix &&
6335 strncmp(path, options->prefix, options->prefix_length))
6336 return NULL;
6337
6338 one = alloc_filespec(path);
6339 two = alloc_filespec(path);
6340 pair = diff_queue(&diff_queued_diff, one, two);
6341 pair->is_unmerged = 1;
6342 return pair;
6343 }
6344
6345 static char *run_textconv(struct repository *r,
6346 const char *pgm,
6347 struct diff_filespec *spec,
6348 size_t *outsize)
6349 {
6350 struct diff_tempfile *temp;
6351 const char *argv[3];
6352 const char **arg = argv;
6353 struct child_process child = CHILD_PROCESS_INIT;
6354 struct strbuf buf = STRBUF_INIT;
6355 int err = 0;
6356
6357 temp = prepare_temp_file(r, spec->path, spec);
6358 *arg++ = pgm;
6359 *arg++ = temp->name;
6360 *arg = NULL;
6361
6362 child.use_shell = 1;
6363 child.argv = argv;
6364 child.out = -1;
6365 if (start_command(&child)) {
6366 remove_tempfile();
6367 return NULL;
6368 }
6369
6370 if (strbuf_read(&buf, child.out, 0) < 0)
6371 err = error("error reading from textconv command '%s'", pgm);
6372 close(child.out);
6373
6374 if (finish_command(&child) || err) {
6375 strbuf_release(&buf);
6376 remove_tempfile();
6377 return NULL;
6378 }
6379 remove_tempfile();
6380
6381 return strbuf_detach(&buf, outsize);
6382 }
6383
6384 size_t fill_textconv(struct repository *r,
6385 struct userdiff_driver *driver,
6386 struct diff_filespec *df,
6387 char **outbuf)
6388 {
6389 size_t size;
6390
6391 if (!driver) {
6392 if (!DIFF_FILE_VALID(df)) {
6393 *outbuf = "";
6394 return 0;
6395 }
6396 if (diff_populate_filespec(r, df, 0))
6397 die("unable to read files to diff");
6398 *outbuf = df->data;
6399 return df->size;
6400 }
6401
6402 if (!driver->textconv)
6403 BUG("fill_textconv called with non-textconv driver");
6404
6405 if (driver->textconv_cache && df->oid_valid) {
6406 *outbuf = notes_cache_get(driver->textconv_cache,
6407 &df->oid,
6408 &size);
6409 if (*outbuf)
6410 return size;
6411 }
6412
6413 *outbuf = run_textconv(r, driver->textconv, df, &size);
6414 if (!*outbuf)
6415 die("unable to read files to diff");
6416
6417 if (driver->textconv_cache && df->oid_valid) {
6418 /* ignore errors, as we might be in a readonly repository */
6419 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6420 size);
6421 /*
6422 * we could save up changes and flush them all at the end,
6423 * but we would need an extra call after all diffing is done.
6424 * Since generating a cache entry is the slow path anyway,
6425 * this extra overhead probably isn't a big deal.
6426 */
6427 notes_cache_write(driver->textconv_cache);
6428 }
6429
6430 return size;
6431 }
6432
6433 int textconv_object(struct repository *r,
6434 const char *path,
6435 unsigned mode,
6436 const struct object_id *oid,
6437 int oid_valid,
6438 char **buf,
6439 unsigned long *buf_size)
6440 {
6441 struct diff_filespec *df;
6442 struct userdiff_driver *textconv;
6443
6444 df = alloc_filespec(path);
6445 fill_filespec(df, oid, oid_valid, mode);
6446 textconv = get_textconv(r, df);
6447 if (!textconv) {
6448 free_filespec(df);
6449 return 0;
6450 }
6451
6452 *buf_size = fill_textconv(r, textconv, df, buf);
6453 free_filespec(df);
6454 return 1;
6455 }
6456
6457 void setup_diff_pager(struct diff_options *opt)
6458 {
6459 /*
6460 * If the user asked for our exit code, then either they want --quiet
6461 * or --exit-code. We should definitely not bother with a pager in the
6462 * former case, as we will generate no output. Since we still properly
6463 * report our exit code even when a pager is run, we _could_ run a
6464 * pager with --exit-code. But since we have not done so historically,
6465 * and because it is easy to find people oneline advising "git diff
6466 * --exit-code" in hooks and other scripts, we do not do so.
6467 */
6468 if (!opt->flags.exit_with_status &&
6469 check_pager_config("diff") != 0)
6470 setup_pager();
6471 }
The Best VCS Around