?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff.c: emit_diff_symbol learns about DIFF_SYMBOL_STAT_SEP
[git/debian.git] / diff.c
blob5a9c55736d6f7ab37b708eb47af5015302467922
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 "userdiff.h"
17 #include "submodule-config.h"
18 #include "submodule.h"
19 #include "ll-merge.h"
20 #include "string-list.h"
21 #include "argv-array.h"
22 #include "graph.h"
23
24 #ifdef NO_FAST_WORKING_DIRECTORY
25 #define FAST_WORKING_DIRECTORY 0
26 #else
27 #define FAST_WORKING_DIRECTORY 1
28 #endif
29
30 static int diff_detect_rename_default;
31 static int diff_indent_heuristic = 1;
32 static int diff_rename_limit_default = 400;
33 static int diff_suppress_blank_empty;
34 static int diff_use_color_default = -1;
35 static int diff_context_default = 3;
36 static int diff_interhunk_context_default;
37 static const char *diff_word_regex_cfg;
38 static const char *external_diff_cmd_cfg;
39 static const char *diff_order_file_cfg;
40 int diff_auto_refresh_index = 1;
41 static int diff_mnemonic_prefix;
42 static int diff_no_prefix;
43 static int diff_stat_graph_width;
44 static int diff_dirstat_permille_default = 30;
45 static struct diff_options default_diff_options;
46 static long diff_algorithm;
47 static unsigned ws_error_highlight_default = WSEH_NEW;
48
49 static char diff_colors[][COLOR_MAXLEN] = {
50 GIT_COLOR_RESET,
51 GIT_COLOR_NORMAL, /* CONTEXT */
52 GIT_COLOR_BOLD, /* METAINFO */
53 GIT_COLOR_CYAN, /* FRAGINFO */
54 GIT_COLOR_RED, /* OLD */
55 GIT_COLOR_GREEN, /* NEW */
56 GIT_COLOR_YELLOW, /* COMMIT */
57 GIT_COLOR_BG_RED, /* WHITESPACE */
58 GIT_COLOR_NORMAL, /* FUNCINFO */
59 };
60
61 static NORETURN void die_want_option(const char *option_name)
62 {
63 die(_("option '%s' requires a value"), option_name);
64 }
65
66 static int parse_diff_color_slot(const char *var)
67 {
68 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
69 return DIFF_CONTEXT;
70 if (!strcasecmp(var, "meta"))
71 return DIFF_METAINFO;
72 if (!strcasecmp(var, "frag"))
73 return DIFF_FRAGINFO;
74 if (!strcasecmp(var, "old"))
75 return DIFF_FILE_OLD;
76 if (!strcasecmp(var, "new"))
77 return DIFF_FILE_NEW;
78 if (!strcasecmp(var, "commit"))
79 return DIFF_COMMIT;
80 if (!strcasecmp(var, "whitespace"))
81 return DIFF_WHITESPACE;
82 if (!strcasecmp(var, "func"))
83 return DIFF_FUNCINFO;
84 return -1;
85 }
86
87 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
88 struct strbuf *errmsg)
89 {
90 char *params_copy = xstrdup(params_string);
91 struct string_list params = STRING_LIST_INIT_NODUP;
92 int ret = 0;
93 int i;
94
95 if (*params_copy)
96 string_list_split_in_place(&params, params_copy, ',', -1);
97 for (i = 0; i < params.nr; i++) {
98 const char *p = params.items[i].string;
99 if (!strcmp(p, "changes")) {
100 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
101 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
102 } else if (!strcmp(p, "lines")) {
103 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
104 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
105 } else if (!strcmp(p, "files")) {
106 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
107 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
108 } else if (!strcmp(p, "noncumulative")) {
109 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
110 } else if (!strcmp(p, "cumulative")) {
111 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
112 } else if (isdigit(*p)) {
113 char *end;
114 int permille = strtoul(p, &end, 10) * 10;
115 if (*end == '.' && isdigit(*++end)) {
116 /* only use first digit */
117 permille += *end - '0';
118 /* .. and ignore any further digits */
119 while (isdigit(*++end))
120 ; /* nothing */
121 }
122 if (!*end)
123 options->dirstat_permille = permille;
124 else {
125 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
126 p);
127 ret++;
128 }
129 } else {
130 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
131 ret++;
132 }
133
134 }
135 string_list_clear(&params, 0);
136 free(params_copy);
137 return ret;
138 }
139
140 static int parse_submodule_params(struct diff_options *options, const char *value)
141 {
142 if (!strcmp(value, "log"))
143 options->submodule_format = DIFF_SUBMODULE_LOG;
144 else if (!strcmp(value, "short"))
145 options->submodule_format = DIFF_SUBMODULE_SHORT;
146 else if (!strcmp(value, "diff"))
147 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
148 else
149 return -1;
150 return 0;
151 }
152
153 static int git_config_rename(const char *var, const char *value)
154 {
155 if (!value)
156 return DIFF_DETECT_RENAME;
157 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
158 return DIFF_DETECT_COPY;
159 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
160 }
161
162 long parse_algorithm_value(const char *value)
163 {
164 if (!value)
165 return -1;
166 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
167 return 0;
168 else if (!strcasecmp(value, "minimal"))
169 return XDF_NEED_MINIMAL;
170 else if (!strcasecmp(value, "patience"))
171 return XDF_PATIENCE_DIFF;
172 else if (!strcasecmp(value, "histogram"))
173 return XDF_HISTOGRAM_DIFF;
174 return -1;
175 }
176
177 static int parse_one_token(const char **arg, const char *token)
178 {
179 const char *rest;
180 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
181 *arg = rest;
182 return 1;
183 }
184 return 0;
185 }
186
187 static int parse_ws_error_highlight(const char *arg)
188 {
189 const char *orig_arg = arg;
190 unsigned val = 0;
191
192 while (*arg) {
193 if (parse_one_token(&arg, "none"))
194 val = 0;
195 else if (parse_one_token(&arg, "default"))
196 val = WSEH_NEW;
197 else if (parse_one_token(&arg, "all"))
198 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
199 else if (parse_one_token(&arg, "new"))
200 val |= WSEH_NEW;
201 else if (parse_one_token(&arg, "old"))
202 val |= WSEH_OLD;
203 else if (parse_one_token(&arg, "context"))
204 val |= WSEH_CONTEXT;
205 else {
206 return -1 - (int)(arg - orig_arg);
207 }
208 if (*arg)
209 arg++;
210 }
211 return val;
212 }
213
214 /*
215 * These are to give UI layer defaults.
216 * The core-level commands such as git-diff-files should
217 * never be affected by the setting of diff.renames
218 * the user happens to have in the configuration file.
219 */
220 void init_diff_ui_defaults(void)
221 {
222 diff_detect_rename_default = 1;
223 }
224
225 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
226 {
227 if (!strcmp(var, "diff.indentheuristic"))
228 diff_indent_heuristic = git_config_bool(var, value);
229 return 0;
230 }
231
232 int git_diff_ui_config(const char *var, const char *value, void *cb)
233 {
234 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
235 diff_use_color_default = git_config_colorbool(var, value);
236 return 0;
237 }
238 if (!strcmp(var, "diff.context")) {
239 diff_context_default = git_config_int(var, value);
240 if (diff_context_default < 0)
241 return -1;
242 return 0;
243 }
244 if (!strcmp(var, "diff.interhunkcontext")) {
245 diff_interhunk_context_default = git_config_int(var, value);
246 if (diff_interhunk_context_default < 0)
247 return -1;
248 return 0;
249 }
250 if (!strcmp(var, "diff.renames")) {
251 diff_detect_rename_default = git_config_rename(var, value);
252 return 0;
253 }
254 if (!strcmp(var, "diff.autorefreshindex")) {
255 diff_auto_refresh_index = git_config_bool(var, value);
256 return 0;
257 }
258 if (!strcmp(var, "diff.mnemonicprefix")) {
259 diff_mnemonic_prefix = git_config_bool(var, value);
260 return 0;
261 }
262 if (!strcmp(var, "diff.noprefix")) {
263 diff_no_prefix = git_config_bool(var, value);
264 return 0;
265 }
266 if (!strcmp(var, "diff.statgraphwidth")) {
267 diff_stat_graph_width = git_config_int(var, value);
268 return 0;
269 }
270 if (!strcmp(var, "diff.external"))
271 return git_config_string(&external_diff_cmd_cfg, var, value);
272 if (!strcmp(var, "diff.wordregex"))
273 return git_config_string(&diff_word_regex_cfg, var, value);
274 if (!strcmp(var, "diff.orderfile"))
275 return git_config_pathname(&diff_order_file_cfg, var, value);
276
277 if (!strcmp(var, "diff.ignoresubmodules"))
278 handle_ignore_submodules_arg(&default_diff_options, value);
279
280 if (!strcmp(var, "diff.submodule")) {
281 if (parse_submodule_params(&default_diff_options, value))
282 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
283 value);
284 return 0;
285 }
286
287 if (!strcmp(var, "diff.algorithm")) {
288 diff_algorithm = parse_algorithm_value(value);
289 if (diff_algorithm < 0)
290 return -1;
291 return 0;
292 }
293
294 if (!strcmp(var, "diff.wserrorhighlight")) {
295 int val = parse_ws_error_highlight(value);
296 if (val < 0)
297 return -1;
298 ws_error_highlight_default = val;
299 return 0;
300 }
301
302 if (git_color_config(var, value, cb) < 0)
303 return -1;
304
305 return git_diff_basic_config(var, value, cb);
306 }
307
308 int git_diff_basic_config(const char *var, const char *value, void *cb)
309 {
310 const char *name;
311
312 if (!strcmp(var, "diff.renamelimit")) {
313 diff_rename_limit_default = git_config_int(var, value);
314 return 0;
315 }
316
317 if (userdiff_config(var, value) < 0)
318 return -1;
319
320 if (skip_prefix(var, "diff.color.", &name) ||
321 skip_prefix(var, "color.diff.", &name)) {
322 int slot = parse_diff_color_slot(name);
323 if (slot < 0)
324 return 0;
325 if (!value)
326 return config_error_nonbool(var);
327 return color_parse(value, diff_colors[slot]);
328 }
329
330 /* like GNU diff's --suppress-blank-empty option */
331 if (!strcmp(var, "diff.suppressblankempty") ||
332 /* for backwards compatibility */
333 !strcmp(var, "diff.suppress-blank-empty")) {
334 diff_suppress_blank_empty = git_config_bool(var, value);
335 return 0;
336 }
337
338 if (!strcmp(var, "diff.dirstat")) {
339 struct strbuf errmsg = STRBUF_INIT;
340 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
341 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
342 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
343 errmsg.buf);
344 strbuf_release(&errmsg);
345 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
346 return 0;
347 }
348
349 if (starts_with(var, "submodule."))
350 return parse_submodule_config_option(var, value);
351
352 if (git_diff_heuristic_config(var, value, cb) < 0)
353 return -1;
354
355 return git_default_config(var, value, cb);
356 }
357
358 static char *quote_two(const char *one, const char *two)
359 {
360 int need_one = quote_c_style(one, NULL, NULL, 1);
361 int need_two = quote_c_style(two, NULL, NULL, 1);
362 struct strbuf res = STRBUF_INIT;
363
364 if (need_one + need_two) {
365 strbuf_addch(&res, '"');
366 quote_c_style(one, &res, NULL, 1);
367 quote_c_style(two, &res, NULL, 1);
368 strbuf_addch(&res, '"');
369 } else {
370 strbuf_addstr(&res, one);
371 strbuf_addstr(&res, two);
372 }
373 return strbuf_detach(&res, NULL);
374 }
375
376 static const char *external_diff(void)
377 {
378 static const char *external_diff_cmd = NULL;
379 static int done_preparing = 0;
380
381 if (done_preparing)
382 return external_diff_cmd;
383 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
384 if (!external_diff_cmd)
385 external_diff_cmd = external_diff_cmd_cfg;
386 done_preparing = 1;
387 return external_diff_cmd;
388 }
389
390 /*
391 * Keep track of files used for diffing. Sometimes such an entry
392 * refers to a temporary file, sometimes to an existing file, and
393 * sometimes to "/dev/null".
394 */
395 static struct diff_tempfile {
396 /*
397 * filename external diff should read from, or NULL if this
398 * entry is currently not in use:
399 */
400 const char *name;
401
402 char hex[GIT_MAX_HEXSZ + 1];
403 char mode[10];
404
405 /*
406 * If this diff_tempfile instance refers to a temporary file,
407 * this tempfile object is used to manage its lifetime.
408 */
409 struct tempfile tempfile;
410 } diff_temp[2];
411
412 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
413
414 struct emit_callback {
415 int color_diff;
416 unsigned ws_rule;
417 int blank_at_eof_in_preimage;
418 int blank_at_eof_in_postimage;
419 int lno_in_preimage;
420 int lno_in_postimage;
421 sane_truncate_fn truncate;
422 const char **label_path;
423 struct diff_words_data *diff_words;
424 struct diff_options *opt;
425 struct strbuf *header;
426 };
427
428 static int count_lines(const char *data, int size)
429 {
430 int count, ch, completely_empty = 1, nl_just_seen = 0;
431 count = 0;
432 while (0 < size--) {
433 ch = *data++;
434 if (ch == '\n') {
435 count++;
436 nl_just_seen = 1;
437 completely_empty = 0;
438 }
439 else {
440 nl_just_seen = 0;
441 completely_empty = 0;
442 }
443 }
444 if (completely_empty)
445 return 0;
446 if (!nl_just_seen)
447 count++; /* no trailing newline */
448 return count;
449 }
450
451 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
452 {
453 if (!DIFF_FILE_VALID(one)) {
454 mf->ptr = (char *)""; /* does not matter */
455 mf->size = 0;
456 return 0;
457 }
458 else if (diff_populate_filespec(one, 0))
459 return -1;
460
461 mf->ptr = one->data;
462 mf->size = one->size;
463 return 0;
464 }
465
466 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
467 static unsigned long diff_filespec_size(struct diff_filespec *one)
468 {
469 if (!DIFF_FILE_VALID(one))
470 return 0;
471 diff_populate_filespec(one, CHECK_SIZE_ONLY);
472 return one->size;
473 }
474
475 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
476 {
477 char *ptr = mf->ptr;
478 long size = mf->size;
479 int cnt = 0;
480
481 if (!size)
482 return cnt;
483 ptr += size - 1; /* pointing at the very end */
484 if (*ptr != '\n')
485 ; /* incomplete line */
486 else
487 ptr--; /* skip the last LF */
488 while (mf->ptr < ptr) {
489 char *prev_eol;
490 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
491 if (*prev_eol == '\n')
492 break;
493 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
494 break;
495 cnt++;
496 ptr = prev_eol - 1;
497 }
498 return cnt;
499 }
500
501 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
502 struct emit_callback *ecbdata)
503 {
504 int l1, l2, at;
505 unsigned ws_rule = ecbdata->ws_rule;
506 l1 = count_trailing_blank(mf1, ws_rule);
507 l2 = count_trailing_blank(mf2, ws_rule);
508 if (l2 <= l1) {
509 ecbdata->blank_at_eof_in_preimage = 0;
510 ecbdata->blank_at_eof_in_postimage = 0;
511 return;
512 }
513 at = count_lines(mf1->ptr, mf1->size);
514 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
515
516 at = count_lines(mf2->ptr, mf2->size);
517 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
518 }
519
520 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
521 int first, const char *line, int len)
522 {
523 int has_trailing_newline, has_trailing_carriage_return;
524 int nofirst;
525 FILE *file = o->file;
526
527 fputs(diff_line_prefix(o), file);
528
529 if (len == 0) {
530 has_trailing_newline = (first == '\n');
531 has_trailing_carriage_return = (!has_trailing_newline &&
532 (first == '\r'));
533 nofirst = has_trailing_newline || has_trailing_carriage_return;
534 } else {
535 has_trailing_newline = (len > 0 && line[len-1] == '\n');
536 if (has_trailing_newline)
537 len--;
538 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
539 if (has_trailing_carriage_return)
540 len--;
541 nofirst = 0;
542 }
543
544 if (len || !nofirst) {
545 fputs(set, file);
546 if (!nofirst)
547 fputc(first, file);
548 fwrite(line, len, 1, file);
549 fputs(reset, file);
550 }
551 if (has_trailing_carriage_return)
552 fputc('\r', file);
553 if (has_trailing_newline)
554 fputc('\n', file);
555 }
556
557 static void emit_line(struct diff_options *o, const char *set, const char *reset,
558 const char *line, int len)
559 {
560 emit_line_0(o, set, reset, line[0], line+1, len-1);
561 }
562
563 enum diff_symbol {
564 DIFF_SYMBOL_BINARY_DIFF_HEADER,
565 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
566 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
567 DIFF_SYMBOL_BINARY_DIFF_BODY,
568 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
569 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
570 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
571 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
572 DIFF_SYMBOL_STATS_LINE,
573 DIFF_SYMBOL_WORD_DIFF,
574 DIFF_SYMBOL_STAT_SEP,
575 DIFF_SYMBOL_SUBMODULE_ADD,
576 DIFF_SYMBOL_SUBMODULE_DEL,
577 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
578 DIFF_SYMBOL_SUBMODULE_MODIFIED,
579 DIFF_SYMBOL_SUBMODULE_HEADER,
580 DIFF_SYMBOL_SUBMODULE_ERROR,
581 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
582 DIFF_SYMBOL_REWRITE_DIFF,
583 DIFF_SYMBOL_BINARY_FILES,
584 DIFF_SYMBOL_HEADER,
585 DIFF_SYMBOL_FILEPAIR_PLUS,
586 DIFF_SYMBOL_FILEPAIR_MINUS,
587 DIFF_SYMBOL_WORDS_PORCELAIN,
588 DIFF_SYMBOL_WORDS,
589 DIFF_SYMBOL_CONTEXT,
590 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
591 DIFF_SYMBOL_PLUS,
592 DIFF_SYMBOL_MINUS,
593 DIFF_SYMBOL_NO_LF_EOF,
594 DIFF_SYMBOL_CONTEXT_FRAGINFO,
595 DIFF_SYMBOL_CONTEXT_MARKER,
596 DIFF_SYMBOL_SEPARATOR
597 };
598 /*
599 * Flags for content lines:
600 * 0..12 are whitespace rules
601 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
602 * 16 is marking if the line is blank at EOF
603 */
604 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
605 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
606
607 static void emit_line_ws_markup(struct diff_options *o,
608 const char *set, const char *reset,
609 const char *line, int len, char sign,
610 unsigned ws_rule, int blank_at_eof)
611 {
612 const char *ws = NULL;
613
614 if (o->ws_error_highlight & ws_rule) {
615 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
616 if (!*ws)
617 ws = NULL;
618 }
619
620 if (!ws)
621 emit_line_0(o, set, reset, sign, line, len);
622 else if (blank_at_eof)
623 /* Blank line at EOF - paint '+' as well */
624 emit_line_0(o, ws, reset, sign, line, len);
625 else {
626 /* Emit just the prefix, then the rest. */
627 emit_line_0(o, set, reset, sign, "", 0);
628 ws_check_emit(line, len, ws_rule,
629 o->file, set, reset, ws);
630 }
631 }
632
633 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
634 const char *line, int len, unsigned flags)
635 {
636 static const char *nneof = " No newline at end of file\n";
637 const char *context, *reset, *set, *meta, *fraginfo;
638 struct strbuf sb = STRBUF_INIT;
639 switch (s) {
640 case DIFF_SYMBOL_NO_LF_EOF:
641 context = diff_get_color_opt(o, DIFF_CONTEXT);
642 reset = diff_get_color_opt(o, DIFF_RESET);
643 putc('\n', o->file);
644 emit_line_0(o, context, reset, '\\',
645 nneof, strlen(nneof));
646 break;
647 case DIFF_SYMBOL_SUBMODULE_HEADER:
648 case DIFF_SYMBOL_SUBMODULE_ERROR:
649 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
650 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
651 case DIFF_SYMBOL_STATS_LINE:
652 case DIFF_SYMBOL_BINARY_DIFF_BODY:
653 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
654 emit_line(o, "", "", line, len);
655 break;
656 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
657 case DIFF_SYMBOL_CONTEXT_MARKER:
658 context = diff_get_color_opt(o, DIFF_CONTEXT);
659 reset = diff_get_color_opt(o, DIFF_RESET);
660 emit_line(o, context, reset, line, len);
661 break;
662 case DIFF_SYMBOL_SEPARATOR:
663 fprintf(o->file, "%s%c",
664 diff_line_prefix(o),
665 o->line_termination);
666 break;
667 case DIFF_SYMBOL_CONTEXT:
668 set = diff_get_color_opt(o, DIFF_CONTEXT);
669 reset = diff_get_color_opt(o, DIFF_RESET);
670 emit_line_ws_markup(o, set, reset, line, len, ' ',
671 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
672 break;
673 case DIFF_SYMBOL_PLUS:
674 set = diff_get_color_opt(o, DIFF_FILE_NEW);
675 reset = diff_get_color_opt(o, DIFF_RESET);
676 emit_line_ws_markup(o, set, reset, line, len, '+',
677 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
678 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
679 break;
680 case DIFF_SYMBOL_MINUS:
681 set = diff_get_color_opt(o, DIFF_FILE_OLD);
682 reset = diff_get_color_opt(o, DIFF_RESET);
683 emit_line_ws_markup(o, set, reset, line, len, '-',
684 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
685 break;
686 case DIFF_SYMBOL_WORDS_PORCELAIN:
687 context = diff_get_color_opt(o, DIFF_CONTEXT);
688 reset = diff_get_color_opt(o, DIFF_RESET);
689 emit_line(o, context, reset, line, len);
690 fputs("~\n", o->file);
691 break;
692 case DIFF_SYMBOL_WORDS:
693 context = diff_get_color_opt(o, DIFF_CONTEXT);
694 reset = diff_get_color_opt(o, DIFF_RESET);
695 /*
696 * Skip the prefix character, if any. With
697 * diff_suppress_blank_empty, there may be
698 * none.
699 */
700 if (line[0] != '\n') {
701 line++;
702 len--;
703 }
704 emit_line(o, context, reset, line, len);
705 break;
706 case DIFF_SYMBOL_FILEPAIR_PLUS:
707 meta = diff_get_color_opt(o, DIFF_METAINFO);
708 reset = diff_get_color_opt(o, DIFF_RESET);
709 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
710 line, reset,
711 strchr(line, ' ') ? "\t" : "");
712 break;
713 case DIFF_SYMBOL_FILEPAIR_MINUS:
714 meta = diff_get_color_opt(o, DIFF_METAINFO);
715 reset = diff_get_color_opt(o, DIFF_RESET);
716 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
717 line, reset,
718 strchr(line, ' ') ? "\t" : "");
719 break;
720 case DIFF_SYMBOL_BINARY_FILES:
721 case DIFF_SYMBOL_HEADER:
722 fprintf(o->file, "%s", line);
723 break;
724 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
725 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
726 break;
727 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
728 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
729 break;
730 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
731 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
732 break;
733 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
734 fputs(diff_line_prefix(o), o->file);
735 fputc('\n', o->file);
736 break;
737 case DIFF_SYMBOL_REWRITE_DIFF:
738 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
739 reset = diff_get_color_opt(o, DIFF_RESET);
740 emit_line(o, fraginfo, reset, line, len);
741 break;
742 case DIFF_SYMBOL_SUBMODULE_ADD:
743 set = diff_get_color_opt(o, DIFF_FILE_NEW);
744 reset = diff_get_color_opt(o, DIFF_RESET);
745 emit_line(o, set, reset, line, len);
746 break;
747 case DIFF_SYMBOL_SUBMODULE_DEL:
748 set = diff_get_color_opt(o, DIFF_FILE_OLD);
749 reset = diff_get_color_opt(o, DIFF_RESET);
750 emit_line(o, set, reset, line, len);
751 break;
752 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
753 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
754 diff_line_prefix(o), line);
755 break;
756 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
757 fprintf(o->file, "%sSubmodule %s contains modified content\n",
758 diff_line_prefix(o), line);
759 break;
760 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
761 emit_line(o, "", "", " 0 files changed\n",
762 strlen(" 0 files changed\n"));
763 break;
764 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
765 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
766 break;
767 case DIFF_SYMBOL_WORD_DIFF:
768 fprintf(o->file, "%.*s", len, line);
769 break;
770 case DIFF_SYMBOL_STAT_SEP:
771 fputs(o->stat_sep, o->file);
772 break;
773 default:
774 die("BUG: unknown diff symbol");
775 }
776 strbuf_release(&sb);
777 }
778
779 void diff_emit_submodule_del(struct diff_options *o, const char *line)
780 {
781 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
782 }
783
784 void diff_emit_submodule_add(struct diff_options *o, const char *line)
785 {
786 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
787 }
788
789 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
790 {
791 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
792 path, strlen(path), 0);
793 }
794
795 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
796 {
797 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
798 path, strlen(path), 0);
799 }
800
801 void diff_emit_submodule_header(struct diff_options *o, const char *header)
802 {
803 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
804 header, strlen(header), 0);
805 }
806
807 void diff_emit_submodule_error(struct diff_options *o, const char *err)
808 {
809 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
810 }
811
812 void diff_emit_submodule_pipethrough(struct diff_options *o,
813 const char *line, int len)
814 {
815 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
816 }
817
818 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
819 {
820 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
821 ecbdata->blank_at_eof_in_preimage &&
822 ecbdata->blank_at_eof_in_postimage &&
823 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
824 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
825 return 0;
826 return ws_blank_line(line, len, ecbdata->ws_rule);
827 }
828
829 static void emit_add_line(const char *reset,
830 struct emit_callback *ecbdata,
831 const char *line, int len)
832 {
833 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
834 if (new_blank_line_at_eof(ecbdata, line, len))
835 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
836
837 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
838 }
839
840 static void emit_del_line(const char *reset,
841 struct emit_callback *ecbdata,
842 const char *line, int len)
843 {
844 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
845 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
846 }
847
848 static void emit_context_line(const char *reset,
849 struct emit_callback *ecbdata,
850 const char *line, int len)
851 {
852 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
853 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
854 }
855
856 static void emit_hunk_header(struct emit_callback *ecbdata,
857 const char *line, int len)
858 {
859 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
860 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
861 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
862 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
863 static const char atat[2] = { '@', '@' };
864 const char *cp, *ep;
865 struct strbuf msgbuf = STRBUF_INIT;
866 int org_len = len;
867 int i = 1;
868
869 /*
870 * As a hunk header must begin with "@@ -<old>, +<new> @@",
871 * it always is at least 10 bytes long.
872 */
873 if (len < 10 ||
874 memcmp(line, atat, 2) ||
875 !(ep = memmem(line + 2, len - 2, atat, 2))) {
876 emit_diff_symbol(ecbdata->opt,
877 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
878 return;
879 }
880 ep += 2; /* skip over @@ */
881
882 /* The hunk header in fraginfo color */
883 strbuf_addstr(&msgbuf, frag);
884 strbuf_add(&msgbuf, line, ep - line);
885 strbuf_addstr(&msgbuf, reset);
886
887 /*
888 * trailing "\r\n"
889 */
890 for ( ; i < 3; i++)
891 if (line[len - i] == '\r' || line[len - i] == '\n')
892 len--;
893
894 /* blank before the func header */
895 for (cp = ep; ep - line < len; ep++)
896 if (*ep != ' ' && *ep != '\t')
897 break;
898 if (ep != cp) {
899 strbuf_addstr(&msgbuf, context);
900 strbuf_add(&msgbuf, cp, ep - cp);
901 strbuf_addstr(&msgbuf, reset);
902 }
903
904 if (ep < line + len) {
905 strbuf_addstr(&msgbuf, func);
906 strbuf_add(&msgbuf, ep, line + len - ep);
907 strbuf_addstr(&msgbuf, reset);
908 }
909
910 strbuf_add(&msgbuf, line + len, org_len - len);
911 strbuf_complete_line(&msgbuf);
912 emit_diff_symbol(ecbdata->opt,
913 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
914 strbuf_release(&msgbuf);
915 }
916
917 static struct diff_tempfile *claim_diff_tempfile(void) {
918 int i;
919 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
920 if (!diff_temp[i].name)
921 return diff_temp + i;
922 die("BUG: diff is failing to clean up its tempfiles");
923 }
924
925 static void remove_tempfile(void)
926 {
927 int i;
928 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
929 if (is_tempfile_active(&diff_temp[i].tempfile))
930 delete_tempfile(&diff_temp[i].tempfile);
931 diff_temp[i].name = NULL;
932 }
933 }
934
935 static void add_line_count(struct strbuf *out, int count)
936 {
937 switch (count) {
938 case 0:
939 strbuf_addstr(out, "0,0");
940 break;
941 case 1:
942 strbuf_addstr(out, "1");
943 break;
944 default:
945 strbuf_addf(out, "1,%d", count);
946 break;
947 }
948 }
949
950 static void emit_rewrite_lines(struct emit_callback *ecb,
951 int prefix, const char *data, int size)
952 {
953 const char *endp = NULL;
954 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
955
956 while (0 < size) {
957 int len;
958
959 endp = memchr(data, '\n', size);
960 len = endp ? (endp - data + 1) : size;
961 if (prefix != '+') {
962 ecb->lno_in_preimage++;
963 emit_del_line(reset, ecb, data, len);
964 } else {
965 ecb->lno_in_postimage++;
966 emit_add_line(reset, ecb, data, len);
967 }
968 size -= len;
969 data += len;
970 }
971 if (!endp)
972 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
973 }
974
975 static void emit_rewrite_diff(const char *name_a,
976 const char *name_b,
977 struct diff_filespec *one,
978 struct diff_filespec *two,
979 struct userdiff_driver *textconv_one,
980 struct userdiff_driver *textconv_two,
981 struct diff_options *o)
982 {
983 int lc_a, lc_b;
984 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
985 const char *a_prefix, *b_prefix;
986 char *data_one, *data_two;
987 size_t size_one, size_two;
988 struct emit_callback ecbdata;
989 struct strbuf out = STRBUF_INIT;
990
991 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
992 a_prefix = o->b_prefix;
993 b_prefix = o->a_prefix;
994 } else {
995 a_prefix = o->a_prefix;
996 b_prefix = o->b_prefix;
997 }
998
999 name_a += (*name_a == '/');
1000 name_b += (*name_b == '/');
1001
1002 strbuf_reset(&a_name);
1003 strbuf_reset(&b_name);
1004 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1005 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1006
1007 size_one = fill_textconv(textconv_one, one, &data_one);
1008 size_two = fill_textconv(textconv_two, two, &data_two);
1009
1010 memset(&ecbdata, 0, sizeof(ecbdata));
1011 ecbdata.color_diff = want_color(o->use_color);
1012 ecbdata.ws_rule = whitespace_rule(name_b);
1013 ecbdata.opt = o;
1014 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1015 mmfile_t mf1, mf2;
1016 mf1.ptr = (char *)data_one;
1017 mf2.ptr = (char *)data_two;
1018 mf1.size = size_one;
1019 mf2.size = size_two;
1020 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1021 }
1022 ecbdata.lno_in_preimage = 1;
1023 ecbdata.lno_in_postimage = 1;
1024
1025 lc_a = count_lines(data_one, size_one);
1026 lc_b = count_lines(data_two, size_two);
1027
1028 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1029 a_name.buf, a_name.len, 0);
1030 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1031 b_name.buf, b_name.len, 0);
1032
1033 strbuf_addstr(&out, "@@ -");
1034 if (!o->irreversible_delete)
1035 add_line_count(&out, lc_a);
1036 else
1037 strbuf_addstr(&out, "?,?");
1038 strbuf_addstr(&out, " +");
1039 add_line_count(&out, lc_b);
1040 strbuf_addstr(&out, " @@\n");
1041 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1042 strbuf_release(&out);
1043
1044 if (lc_a && !o->irreversible_delete)
1045 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1046 if (lc_b)
1047 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1048 if (textconv_one)
1049 free((char *)data_one);
1050 if (textconv_two)
1051 free((char *)data_two);
1052 }
1053
1054 struct diff_words_buffer {
1055 mmfile_t text;
1056 long alloc;
1057 struct diff_words_orig {
1058 const char *begin, *end;
1059 } *orig;
1060 int orig_nr, orig_alloc;
1061 };
1062
1063 static void diff_words_append(char *line, unsigned long len,
1064 struct diff_words_buffer *buffer)
1065 {
1066 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1067 line++;
1068 len--;
1069 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1070 buffer->text.size += len;
1071 buffer->text.ptr[buffer->text.size] = '\0';
1072 }
1073
1074 struct diff_words_style_elem {
1075 const char *prefix;
1076 const char *suffix;
1077 const char *color; /* NULL; filled in by the setup code if
1078 * color is enabled */
1079 };
1080
1081 struct diff_words_style {
1082 enum diff_words_type type;
1083 struct diff_words_style_elem new, old, ctx;
1084 const char *newline;
1085 };
1086
1087 static struct diff_words_style diff_words_styles[] = {
1088 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1089 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1090 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1091 };
1092
1093 struct diff_words_data {
1094 struct diff_words_buffer minus, plus;
1095 const char *current_plus;
1096 int last_minus;
1097 struct diff_options *opt;
1098 regex_t *word_regex;
1099 enum diff_words_type type;
1100 struct diff_words_style *style;
1101 };
1102
1103 static int fn_out_diff_words_write_helper(struct diff_options *o,
1104 struct diff_words_style_elem *st_el,
1105 const char *newline,
1106 size_t count, const char *buf)
1107 {
1108 int print = 0;
1109 struct strbuf sb = STRBUF_INIT;
1110
1111 while (count) {
1112 char *p = memchr(buf, '\n', count);
1113 if (print)
1114 strbuf_addstr(&sb, diff_line_prefix(o));
1115
1116 if (p != buf) {
1117 const char *reset = st_el->color && *st_el->color ?
1118 GIT_COLOR_RESET : NULL;
1119 if (st_el->color && *st_el->color)
1120 strbuf_addstr(&sb, st_el->color);
1121 strbuf_addstr(&sb, st_el->prefix);
1122 strbuf_add(&sb, buf, p ? p - buf : count);
1123 strbuf_addstr(&sb, st_el->suffix);
1124 if (reset)
1125 strbuf_addstr(&sb, reset);
1126 }
1127 if (!p)
1128 goto out;
1129
1130 strbuf_addstr(&sb, newline);
1131 count -= p + 1 - buf;
1132 buf = p + 1;
1133 print = 1;
1134 if (count) {
1135 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1136 sb.buf, sb.len, 0);
1137 strbuf_reset(&sb);
1138 }
1139 }
1140
1141 out:
1142 if (sb.len)
1143 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1144 sb.buf, sb.len, 0);
1145 strbuf_release(&sb);
1146 return 0;
1147 }
1148
1149 /*
1150 * '--color-words' algorithm can be described as:
1151 *
1152 * 1. collect the minus/plus lines of a diff hunk, divided into
1153 * minus-lines and plus-lines;
1154 *
1155 * 2. break both minus-lines and plus-lines into words and
1156 * place them into two mmfile_t with one word for each line;
1157 *
1158 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1159 *
1160 * And for the common parts of the both file, we output the plus side text.
1161 * diff_words->current_plus is used to trace the current position of the plus file
1162 * which printed. diff_words->last_minus is used to trace the last minus word
1163 * printed.
1164 *
1165 * For '--graph' to work with '--color-words', we need to output the graph prefix
1166 * on each line of color words output. Generally, there are two conditions on
1167 * which we should output the prefix.
1168 *
1169 * 1. diff_words->last_minus == 0 &&
1170 * diff_words->current_plus == diff_words->plus.text.ptr
1171 *
1172 * that is: the plus text must start as a new line, and if there is no minus
1173 * word printed, a graph prefix must be printed.
1174 *
1175 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1176 * *(diff_words->current_plus - 1) == '\n'
1177 *
1178 * that is: a graph prefix must be printed following a '\n'
1179 */
1180 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1181 {
1182 if ((diff_words->last_minus == 0 &&
1183 diff_words->current_plus == diff_words->plus.text.ptr) ||
1184 (diff_words->current_plus > diff_words->plus.text.ptr &&
1185 *(diff_words->current_plus - 1) == '\n')) {
1186 return 1;
1187 } else {
1188 return 0;
1189 }
1190 }
1191
1192 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1193 {
1194 struct diff_words_data *diff_words = priv;
1195 struct diff_words_style *style = diff_words->style;
1196 int minus_first, minus_len, plus_first, plus_len;
1197 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1198 struct diff_options *opt = diff_words->opt;
1199 const char *line_prefix;
1200
1201 if (line[0] != '@' || parse_hunk_header(line, len,
1202 &minus_first, &minus_len, &plus_first, &plus_len))
1203 return;
1204
1205 assert(opt);
1206 line_prefix = diff_line_prefix(opt);
1207
1208 /* POSIX requires that first be decremented by one if len == 0... */
1209 if (minus_len) {
1210 minus_begin = diff_words->minus.orig[minus_first].begin;
1211 minus_end =
1212 diff_words->minus.orig[minus_first + minus_len - 1].end;
1213 } else
1214 minus_begin = minus_end =
1215 diff_words->minus.orig[minus_first].end;
1216
1217 if (plus_len) {
1218 plus_begin = diff_words->plus.orig[plus_first].begin;
1219 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1220 } else
1221 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1222
1223 if (color_words_output_graph_prefix(diff_words)) {
1224 fputs(line_prefix, diff_words->opt->file);
1225 }
1226 if (diff_words->current_plus != plus_begin) {
1227 fn_out_diff_words_write_helper(diff_words->opt,
1228 &style->ctx, style->newline,
1229 plus_begin - diff_words->current_plus,
1230 diff_words->current_plus);
1231 }
1232 if (minus_begin != minus_end) {
1233 fn_out_diff_words_write_helper(diff_words->opt,
1234 &style->old, style->newline,
1235 minus_end - minus_begin, minus_begin);
1236 }
1237 if (plus_begin != plus_end) {
1238 fn_out_diff_words_write_helper(diff_words->opt,
1239 &style->new, style->newline,
1240 plus_end - plus_begin, plus_begin);
1241 }
1242
1243 diff_words->current_plus = plus_end;
1244 diff_words->last_minus = minus_first;
1245 }
1246
1247 /* This function starts looking at *begin, and returns 0 iff a word was found. */
1248 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1249 int *begin, int *end)
1250 {
1251 if (word_regex && *begin < buffer->size) {
1252 regmatch_t match[1];
1253 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1254 buffer->size - *begin, 1, match, 0)) {
1255 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1256 '\n', match[0].rm_eo - match[0].rm_so);
1257 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1258 *begin += match[0].rm_so;
1259 return *begin >= *end;
1260 }
1261 return -1;
1262 }
1263
1264 /* find the next word */
1265 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1266 (*begin)++;
1267 if (*begin >= buffer->size)
1268 return -1;
1269
1270 /* find the end of the word */
1271 *end = *begin + 1;
1272 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1273 (*end)++;
1274
1275 return 0;
1276 }
1277
1278 /*
1279 * This function splits the words in buffer->text, stores the list with
1280 * newline separator into out, and saves the offsets of the original words
1281 * in buffer->orig.
1282 */
1283 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1284 regex_t *word_regex)
1285 {
1286 int i, j;
1287 long alloc = 0;
1288
1289 out->size = 0;
1290 out->ptr = NULL;
1291
1292 /* fake an empty "0th" word */
1293 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1294 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1295 buffer->orig_nr = 1;
1296
1297 for (i = 0; i < buffer->text.size; i++) {
1298 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1299 return;
1300
1301 /* store original boundaries */
1302 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1303 buffer->orig_alloc);
1304 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1305 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1306 buffer->orig_nr++;
1307
1308 /* store one word */
1309 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1310 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1311 out->ptr[out->size + j - i] = '\n';
1312 out->size += j - i + 1;
1313
1314 i = j - 1;
1315 }
1316 }
1317
1318 /* this executes the word diff on the accumulated buffers */
1319 static void diff_words_show(struct diff_words_data *diff_words)
1320 {
1321 xpparam_t xpp;
1322 xdemitconf_t xecfg;
1323 mmfile_t minus, plus;
1324 struct diff_words_style *style = diff_words->style;
1325
1326 struct diff_options *opt = diff_words->opt;
1327 const char *line_prefix;
1328
1329 assert(opt);
1330 line_prefix = diff_line_prefix(opt);
1331
1332 /* special case: only removal */
1333 if (!diff_words->plus.text.size) {
1334 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1335 line_prefix, strlen(line_prefix), 0);
1336 fn_out_diff_words_write_helper(diff_words->opt,
1337 &style->old, style->newline,
1338 diff_words->minus.text.size,
1339 diff_words->minus.text.ptr);
1340 diff_words->minus.text.size = 0;
1341 return;
1342 }
1343
1344 diff_words->current_plus = diff_words->plus.text.ptr;
1345 diff_words->last_minus = 0;
1346
1347 memset(&xpp, 0, sizeof(xpp));
1348 memset(&xecfg, 0, sizeof(xecfg));
1349 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1350 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1351 xpp.flags = 0;
1352 /* as only the hunk header will be parsed, we need a 0-context */
1353 xecfg.ctxlen = 0;
1354 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1355 &xpp, &xecfg))
1356 die("unable to generate word diff");
1357 free(minus.ptr);
1358 free(plus.ptr);
1359 if (diff_words->current_plus != diff_words->plus.text.ptr +
1360 diff_words->plus.text.size) {
1361 if (color_words_output_graph_prefix(diff_words))
1362 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1363 line_prefix, strlen(line_prefix), 0);
1364 fn_out_diff_words_write_helper(diff_words->opt,
1365 &style->ctx, style->newline,
1366 diff_words->plus.text.ptr + diff_words->plus.text.size
1367 - diff_words->current_plus, diff_words->current_plus);
1368 }
1369 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1370 }
1371
1372 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1373 static void diff_words_flush(struct emit_callback *ecbdata)
1374 {
1375 if (ecbdata->diff_words->minus.text.size ||
1376 ecbdata->diff_words->plus.text.size)
1377 diff_words_show(ecbdata->diff_words);
1378 }
1379
1380 static void diff_filespec_load_driver(struct diff_filespec *one)
1381 {
1382 /* Use already-loaded driver */
1383 if (one->driver)
1384 return;
1385
1386 if (S_ISREG(one->mode))
1387 one->driver = userdiff_find_by_path(one->path);
1388
1389 /* Fallback to default settings */
1390 if (!one->driver)
1391 one->driver = userdiff_find_by_name("default");
1392 }
1393
1394 static const char *userdiff_word_regex(struct diff_filespec *one)
1395 {
1396 diff_filespec_load_driver(one);
1397 return one->driver->word_regex;
1398 }
1399
1400 static void init_diff_words_data(struct emit_callback *ecbdata,
1401 struct diff_options *orig_opts,
1402 struct diff_filespec *one,
1403 struct diff_filespec *two)
1404 {
1405 int i;
1406 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1407 memcpy(o, orig_opts, sizeof(struct diff_options));
1408
1409 ecbdata->diff_words =
1410 xcalloc(1, sizeof(struct diff_words_data));
1411 ecbdata->diff_words->type = o->word_diff;
1412 ecbdata->diff_words->opt = o;
1413 if (!o->word_regex)
1414 o->word_regex = userdiff_word_regex(one);
1415 if (!o->word_regex)
1416 o->word_regex = userdiff_word_regex(two);
1417 if (!o->word_regex)
1418 o->word_regex = diff_word_regex_cfg;
1419 if (o->word_regex) {
1420 ecbdata->diff_words->word_regex = (regex_t *)
1421 xmalloc(sizeof(regex_t));
1422 if (regcomp(ecbdata->diff_words->word_regex,
1423 o->word_regex,
1424 REG_EXTENDED | REG_NEWLINE))
1425 die ("Invalid regular expression: %s",
1426 o->word_regex);
1427 }
1428 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1429 if (o->word_diff == diff_words_styles[i].type) {
1430 ecbdata->diff_words->style =
1431 &diff_words_styles[i];
1432 break;
1433 }
1434 }
1435 if (want_color(o->use_color)) {
1436 struct diff_words_style *st = ecbdata->diff_words->style;
1437 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1438 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1439 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1440 }
1441 }
1442
1443 static void free_diff_words_data(struct emit_callback *ecbdata)
1444 {
1445 if (ecbdata->diff_words) {
1446 diff_words_flush(ecbdata);
1447 free (ecbdata->diff_words->opt);
1448 free (ecbdata->diff_words->minus.text.ptr);
1449 free (ecbdata->diff_words->minus.orig);
1450 free (ecbdata->diff_words->plus.text.ptr);
1451 free (ecbdata->diff_words->plus.orig);
1452 if (ecbdata->diff_words->word_regex) {
1453 regfree(ecbdata->diff_words->word_regex);
1454 free(ecbdata->diff_words->word_regex);
1455 }
1456 FREE_AND_NULL(ecbdata->diff_words);
1457 }
1458 }
1459
1460 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1461 {
1462 if (want_color(diff_use_color))
1463 return diff_colors[ix];
1464 return "";
1465 }
1466
1467 const char *diff_line_prefix(struct diff_options *opt)
1468 {
1469 struct strbuf *msgbuf;
1470 if (!opt->output_prefix)
1471 return "";
1472
1473 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1474 return msgbuf->buf;
1475 }
1476
1477 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1478 {
1479 const char *cp;
1480 unsigned long allot;
1481 size_t l = len;
1482
1483 if (ecb->truncate)
1484 return ecb->truncate(line, len);
1485 cp = line;
1486 allot = l;
1487 while (0 < l) {
1488 (void) utf8_width(&cp, &l);
1489 if (!cp)
1490 break; /* truncated in the middle? */
1491 }
1492 return allot - l;
1493 }
1494
1495 static void find_lno(const char *line, struct emit_callback *ecbdata)
1496 {
1497 const char *p;
1498 ecbdata->lno_in_preimage = 0;
1499 ecbdata->lno_in_postimage = 0;
1500 p = strchr(line, '-');
1501 if (!p)
1502 return; /* cannot happen */
1503 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1504 p = strchr(p, '+');
1505 if (!p)
1506 return; /* cannot happen */
1507 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1508 }
1509
1510 static void fn_out_consume(void *priv, char *line, unsigned long len)
1511 {
1512 struct emit_callback *ecbdata = priv;
1513 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1514 struct diff_options *o = ecbdata->opt;
1515
1516 o->found_changes = 1;
1517
1518 if (ecbdata->header) {
1519 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
1520 ecbdata->header->buf, ecbdata->header->len, 0);
1521 strbuf_reset(ecbdata->header);
1522 ecbdata->header = NULL;
1523 }
1524
1525 if (ecbdata->label_path[0]) {
1526 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1527 ecbdata->label_path[0],
1528 strlen(ecbdata->label_path[0]), 0);
1529 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1530 ecbdata->label_path[1],
1531 strlen(ecbdata->label_path[1]), 0);
1532 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1533 }
1534
1535 if (diff_suppress_blank_empty
1536 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1537 line[0] = '\n';
1538 len = 1;
1539 }
1540
1541 if (line[0] == '@') {
1542 if (ecbdata->diff_words)
1543 diff_words_flush(ecbdata);
1544 len = sane_truncate_line(ecbdata, line, len);
1545 find_lno(line, ecbdata);
1546 emit_hunk_header(ecbdata, line, len);
1547 return;
1548 }
1549
1550 if (ecbdata->diff_words) {
1551 enum diff_symbol s =
1552 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
1553 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
1554 if (line[0] == '-') {
1555 diff_words_append(line, len,
1556 &ecbdata->diff_words->minus);
1557 return;
1558 } else if (line[0] == '+') {
1559 diff_words_append(line, len,
1560 &ecbdata->diff_words->plus);
1561 return;
1562 } else if (starts_with(line, "\\ ")) {
1563 /*
1564 * Eat the "no newline at eof" marker as if we
1565 * saw a "+" or "-" line with nothing on it,
1566 * and return without diff_words_flush() to
1567 * defer processing. If this is the end of
1568 * preimage, more "+" lines may come after it.
1569 */
1570 return;
1571 }
1572 diff_words_flush(ecbdata);
1573 emit_diff_symbol(o, s, line, len, 0);
1574 return;
1575 }
1576
1577 switch (line[0]) {
1578 case '+':
1579 ecbdata->lno_in_postimage++;
1580 emit_add_line(reset, ecbdata, line + 1, len - 1);
1581 break;
1582 case '-':
1583 ecbdata->lno_in_preimage++;
1584 emit_del_line(reset, ecbdata, line + 1, len - 1);
1585 break;
1586 case ' ':
1587 ecbdata->lno_in_postimage++;
1588 ecbdata->lno_in_preimage++;
1589 emit_context_line(reset, ecbdata, line + 1, len - 1);
1590 break;
1591 default:
1592 /* incomplete line at the end */
1593 ecbdata->lno_in_preimage++;
1594 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
1595 line, len, 0);
1596 break;
1597 }
1598 }
1599
1600 static char *pprint_rename(const char *a, const char *b)
1601 {
1602 const char *old = a;
1603 const char *new = b;
1604 struct strbuf name = STRBUF_INIT;
1605 int pfx_length, sfx_length;
1606 int pfx_adjust_for_slash;
1607 int len_a = strlen(a);
1608 int len_b = strlen(b);
1609 int a_midlen, b_midlen;
1610 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1611 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1612
1613 if (qlen_a || qlen_b) {
1614 quote_c_style(a, &name, NULL, 0);
1615 strbuf_addstr(&name, " => ");
1616 quote_c_style(b, &name, NULL, 0);
1617 return strbuf_detach(&name, NULL);
1618 }
1619
1620 /* Find common prefix */
1621 pfx_length = 0;
1622 while (*old && *new && *old == *new) {
1623 if (*old == '/')
1624 pfx_length = old - a + 1;
1625 old++;
1626 new++;
1627 }
1628
1629 /* Find common suffix */
1630 old = a + len_a;
1631 new = b + len_b;
1632 sfx_length = 0;
1633 /*
1634 * If there is a common prefix, it must end in a slash. In
1635 * that case we let this loop run 1 into the prefix to see the
1636 * same slash.
1637 *
1638 * If there is no common prefix, we cannot do this as it would
1639 * underrun the input strings.
1640 */
1641 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1642 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1643 b + pfx_length - pfx_adjust_for_slash <= new &&
1644 *old == *new) {
1645 if (*old == '/')
1646 sfx_length = len_a - (old - a);
1647 old--;
1648 new--;
1649 }
1650
1651 /*
1652 * pfx{mid-a => mid-b}sfx
1653 * {pfx-a => pfx-b}sfx
1654 * pfx{sfx-a => sfx-b}
1655 * name-a => name-b
1656 */
1657 a_midlen = len_a - pfx_length - sfx_length;
1658 b_midlen = len_b - pfx_length - sfx_length;
1659 if (a_midlen < 0)
1660 a_midlen = 0;
1661 if (b_midlen < 0)
1662 b_midlen = 0;
1663
1664 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1665 if (pfx_length + sfx_length) {
1666 strbuf_add(&name, a, pfx_length);
1667 strbuf_addch(&name, '{');
1668 }
1669 strbuf_add(&name, a + pfx_length, a_midlen);
1670 strbuf_addstr(&name, " => ");
1671 strbuf_add(&name, b + pfx_length, b_midlen);
1672 if (pfx_length + sfx_length) {
1673 strbuf_addch(&name, '}');
1674 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1675 }
1676 return strbuf_detach(&name, NULL);
1677 }
1678
1679 struct diffstat_t {
1680 int nr;
1681 int alloc;
1682 struct diffstat_file {
1683 char *from_name;
1684 char *name;
1685 char *print_name;
1686 unsigned is_unmerged:1;
1687 unsigned is_binary:1;
1688 unsigned is_renamed:1;
1689 unsigned is_interesting:1;
1690 uintmax_t added, deleted;
1691 } **files;
1692 };
1693
1694 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1695 const char *name_a,
1696 const char *name_b)
1697 {
1698 struct diffstat_file *x;
1699 x = xcalloc(1, sizeof(*x));
1700 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1701 diffstat->files[diffstat->nr++] = x;
1702 if (name_b) {
1703 x->from_name = xstrdup(name_a);
1704 x->name = xstrdup(name_b);
1705 x->is_renamed = 1;
1706 }
1707 else {
1708 x->from_name = NULL;
1709 x->name = xstrdup(name_a);
1710 }
1711 return x;
1712 }
1713
1714 static void diffstat_consume(void *priv, char *line, unsigned long len)
1715 {
1716 struct diffstat_t *diffstat = priv;
1717 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1718
1719 if (line[0] == '+')
1720 x->added++;
1721 else if (line[0] == '-')
1722 x->deleted++;
1723 }
1724
1725 const char mime_boundary_leader[] = "------------";
1726
1727 static int scale_linear(int it, int width, int max_change)
1728 {
1729 if (!it)
1730 return 0;
1731 /*
1732 * make sure that at least one '-' or '+' is printed if
1733 * there is any change to this path. The easiest way is to
1734 * scale linearly as if the alloted width is one column shorter
1735 * than it is, and then add 1 to the result.
1736 */
1737 return 1 + (it * (width - 1) / max_change);
1738 }
1739
1740 static void show_graph(struct strbuf *out, char ch, int cnt,
1741 const char *set, const char *reset)
1742 {
1743 if (cnt <= 0)
1744 return;
1745 strbuf_addstr(out, set);
1746 strbuf_addchars(out, ch, cnt);
1747 strbuf_addstr(out, reset);
1748 }
1749
1750 static void fill_print_name(struct diffstat_file *file)
1751 {
1752 char *pname;
1753
1754 if (file->print_name)
1755 return;
1756
1757 if (!file->is_renamed) {
1758 struct strbuf buf = STRBUF_INIT;
1759 if (quote_c_style(file->name, &buf, NULL, 0)) {
1760 pname = strbuf_detach(&buf, NULL);
1761 } else {
1762 pname = file->name;
1763 strbuf_release(&buf);
1764 }
1765 } else {
1766 pname = pprint_rename(file->from_name, file->name);
1767 }
1768 file->print_name = pname;
1769 }
1770
1771 static void print_stat_summary_inserts_deletes(struct diff_options *options,
1772 int files, int insertions, int deletions)
1773 {
1774 struct strbuf sb = STRBUF_INIT;
1775
1776 if (!files) {
1777 assert(insertions == 0 && deletions == 0);
1778 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
1779 NULL, 0, 0);
1780 return;
1781 }
1782
1783 strbuf_addf(&sb,
1784 (files == 1) ? " %d file changed" : " %d files changed",
1785 files);
1786
1787 /*
1788 * For binary diff, the caller may want to print "x files
1789 * changed" with insertions == 0 && deletions == 0.
1790 *
1791 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1792 * is probably less confusing (i.e skip over "2 files changed
1793 * but nothing about added/removed lines? Is this a bug in Git?").
1794 */
1795 if (insertions || deletions == 0) {
1796 strbuf_addf(&sb,
1797 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1798 insertions);
1799 }
1800
1801 if (deletions || insertions == 0) {
1802 strbuf_addf(&sb,
1803 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1804 deletions);
1805 }
1806 strbuf_addch(&sb, '\n');
1807 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
1808 sb.buf, sb.len, 0);
1809 strbuf_release(&sb);
1810 }
1811
1812 void print_stat_summary(FILE *fp, int files,
1813 int insertions, int deletions)
1814 {
1815 struct diff_options o;
1816 memset(&o, 0, sizeof(o));
1817 o.file = fp;
1818
1819 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
1820 }
1821
1822 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1823 {
1824 int i, len, add, del, adds = 0, dels = 0;
1825 uintmax_t max_change = 0, max_len = 0;
1826 int total_files = data->nr, count;
1827 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1828 const char *reset, *add_c, *del_c;
1829 int extra_shown = 0;
1830 const char *line_prefix = diff_line_prefix(options);
1831 struct strbuf out = STRBUF_INIT;
1832
1833 if (data->nr == 0)
1834 return;
1835
1836 count = options->stat_count ? options->stat_count : data->nr;
1837
1838 reset = diff_get_color_opt(options, DIFF_RESET);
1839 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1840 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1841
1842 /*
1843 * Find the longest filename and max number of changes
1844 */
1845 for (i = 0; (i < count) && (i < data->nr); i++) {
1846 struct diffstat_file *file = data->files[i];
1847 uintmax_t change = file->added + file->deleted;
1848
1849 if (!file->is_interesting && (change == 0)) {
1850 count++; /* not shown == room for one more */
1851 continue;
1852 }
1853 fill_print_name(file);
1854 len = strlen(file->print_name);
1855 if (max_len < len)
1856 max_len = len;
1857
1858 if (file->is_unmerged) {
1859 /* "Unmerged" is 8 characters */
1860 bin_width = bin_width < 8 ? 8 : bin_width;
1861 continue;
1862 }
1863 if (file->is_binary) {
1864 /* "Bin XXX -> YYY bytes" */
1865 int w = 14 + decimal_width(file->added)
1866 + decimal_width(file->deleted);
1867 bin_width = bin_width < w ? w : bin_width;
1868 /* Display change counts aligned with "Bin" */
1869 number_width = 3;
1870 continue;
1871 }
1872
1873 if (max_change < change)
1874 max_change = change;
1875 }
1876 count = i; /* where we can stop scanning in data->files[] */
1877
1878 /*
1879 * We have width = stat_width or term_columns() columns total.
1880 * We want a maximum of min(max_len, stat_name_width) for the name part.
1881 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1882 * We also need 1 for " " and 4 + decimal_width(max_change)
1883 * for " | NNNN " and one the empty column at the end, altogether
1884 * 6 + decimal_width(max_change).
1885 *
1886 * If there's not enough space, we will use the smaller of
1887 * stat_name_width (if set) and 5/8*width for the filename,
1888 * and the rest for constant elements + graph part, but no more
1889 * than stat_graph_width for the graph part.
1890 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1891 * for the standard terminal size).
1892 *
1893 * In other words: stat_width limits the maximum width, and
1894 * stat_name_width fixes the maximum width of the filename,
1895 * and is also used to divide available columns if there
1896 * aren't enough.
1897 *
1898 * Binary files are displayed with "Bin XXX -> YYY bytes"
1899 * instead of the change count and graph. This part is treated
1900 * similarly to the graph part, except that it is not
1901 * "scaled". If total width is too small to accommodate the
1902 * guaranteed minimum width of the filename part and the
1903 * separators and this message, this message will "overflow"
1904 * making the line longer than the maximum width.
1905 */
1906
1907 if (options->stat_width == -1)
1908 width = term_columns() - strlen(line_prefix);
1909 else
1910 width = options->stat_width ? options->stat_width : 80;
1911 number_width = decimal_width(max_change) > number_width ?
1912 decimal_width(max_change) : number_width;
1913
1914 if (options->stat_graph_width == -1)
1915 options->stat_graph_width = diff_stat_graph_width;
1916
1917 /*
1918 * Guarantee 3/8*16==6 for the graph part
1919 * and 5/8*16==10 for the filename part
1920 */
1921 if (width < 16 + 6 + number_width)
1922 width = 16 + 6 + number_width;
1923
1924 /*
1925 * First assign sizes that are wanted, ignoring available width.
1926 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1927 * starting from "XXX" should fit in graph_width.
1928 */
1929 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1930 if (options->stat_graph_width &&
1931 options->stat_graph_width < graph_width)
1932 graph_width = options->stat_graph_width;
1933
1934 name_width = (options->stat_name_width > 0 &&
1935 options->stat_name_width < max_len) ?
1936 options->stat_name_width : max_len;
1937
1938 /*
1939 * Adjust adjustable widths not to exceed maximum width
1940 */
1941 if (name_width + number_width + 6 + graph_width > width) {
1942 if (graph_width > width * 3/8 - number_width - 6) {
1943 graph_width = width * 3/8 - number_width - 6;
1944 if (graph_width < 6)
1945 graph_width = 6;
1946 }
1947
1948 if (options->stat_graph_width &&
1949 graph_width > options->stat_graph_width)
1950 graph_width = options->stat_graph_width;
1951 if (name_width > width - number_width - 6 - graph_width)
1952 name_width = width - number_width - 6 - graph_width;
1953 else
1954 graph_width = width - number_width - 6 - name_width;
1955 }
1956
1957 /*
1958 * From here name_width is the width of the name area,
1959 * and graph_width is the width of the graph area.
1960 * max_change is used to scale graph properly.
1961 */
1962 for (i = 0; i < count; i++) {
1963 const char *prefix = "";
1964 struct diffstat_file *file = data->files[i];
1965 char *name = file->print_name;
1966 uintmax_t added = file->added;
1967 uintmax_t deleted = file->deleted;
1968 int name_len;
1969
1970 if (!file->is_interesting && (added + deleted == 0))
1971 continue;
1972
1973 /*
1974 * "scale" the filename
1975 */
1976 len = name_width;
1977 name_len = strlen(name);
1978 if (name_width < name_len) {
1979 char *slash;
1980 prefix = "...";
1981 len -= 3;
1982 name += name_len - len;
1983 slash = strchr(name, '/');
1984 if (slash)
1985 name = slash;
1986 }
1987
1988 if (file->is_binary) {
1989 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
1990 strbuf_addf(&out, " %*s", number_width, "Bin");
1991 if (!added && !deleted) {
1992 strbuf_addch(&out, '\n');
1993 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
1994 out.buf, out.len, 0);
1995 strbuf_reset(&out);
1996 continue;
1997 }
1998 strbuf_addf(&out, " %s%"PRIuMAX"%s",
1999 del_c, deleted, reset);
2000 strbuf_addstr(&out, " -> ");
2001 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2002 add_c, added, reset);
2003 strbuf_addstr(&out, " bytes\n");
2004 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2005 out.buf, out.len, 0);
2006 strbuf_reset(&out);
2007 continue;
2008 }
2009 else if (file->is_unmerged) {
2010 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2011 strbuf_addstr(&out, " Unmerged\n");
2012 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2013 out.buf, out.len, 0);
2014 strbuf_reset(&out);
2015 continue;
2016 }
2017
2018 /*
2019 * scale the add/delete
2020 */
2021 add = added;
2022 del = deleted;
2023
2024 if (graph_width <= max_change) {
2025 int total = scale_linear(add + del, graph_width, max_change);
2026 if (total < 2 && add && del)
2027 /* width >= 2 due to the sanity check */
2028 total = 2;
2029 if (add < del) {
2030 add = scale_linear(add, graph_width, max_change);
2031 del = total - add;
2032 } else {
2033 del = scale_linear(del, graph_width, max_change);
2034 add = total - del;
2035 }
2036 }
2037 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2038 strbuf_addf(&out, " %*"PRIuMAX"%s",
2039 number_width, added + deleted,
2040 added + deleted ? " " : "");
2041 show_graph(&out, '+', add, add_c, reset);
2042 show_graph(&out, '-', del, del_c, reset);
2043 strbuf_addch(&out, '\n');
2044 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2045 out.buf, out.len, 0);
2046 strbuf_reset(&out);
2047 }
2048
2049 for (i = 0; i < data->nr; i++) {
2050 struct diffstat_file *file = data->files[i];
2051 uintmax_t added = file->added;
2052 uintmax_t deleted = file->deleted;
2053
2054 if (file->is_unmerged ||
2055 (!file->is_interesting && (added + deleted == 0))) {
2056 total_files--;
2057 continue;
2058 }
2059
2060 if (!file->is_binary) {
2061 adds += added;
2062 dels += deleted;
2063 }
2064 if (i < count)
2065 continue;
2066 if (!extra_shown)
2067 emit_diff_symbol(options,
2068 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2069 NULL, 0, 0);
2070 extra_shown = 1;
2071 }
2072
2073 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2074 }
2075
2076 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2077 {
2078 int i, adds = 0, dels = 0, total_files = data->nr;
2079
2080 if (data->nr == 0)
2081 return;
2082
2083 for (i = 0; i < data->nr; i++) {
2084 int added = data->files[i]->added;
2085 int deleted = data->files[i]->deleted;
2086
2087 if (data->files[i]->is_unmerged ||
2088 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2089 total_files--;
2090 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2091 adds += added;
2092 dels += deleted;
2093 }
2094 }
2095 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2096 }
2097
2098 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2099 {
2100 int i;
2101
2102 if (data->nr == 0)
2103 return;
2104
2105 for (i = 0; i < data->nr; i++) {
2106 struct diffstat_file *file = data->files[i];
2107
2108 fprintf(options->file, "%s", diff_line_prefix(options));
2109
2110 if (file->is_binary)
2111 fprintf(options->file, "-\t-\t");
2112 else
2113 fprintf(options->file,
2114 "%"PRIuMAX"\t%"PRIuMAX"\t",
2115 file->added, file->deleted);
2116 if (options->line_termination) {
2117 fill_print_name(file);
2118 if (!file->is_renamed)
2119 write_name_quoted(file->name, options->file,
2120 options->line_termination);
2121 else {
2122 fputs(file->print_name, options->file);
2123 putc(options->line_termination, options->file);
2124 }
2125 } else {
2126 if (file->is_renamed) {
2127 putc('\0', options->file);
2128 write_name_quoted(file->from_name, options->file, '\0');
2129 }
2130 write_name_quoted(file->name, options->file, '\0');
2131 }
2132 }
2133 }
2134
2135 struct dirstat_file {
2136 const char *name;
2137 unsigned long changed;
2138 };
2139
2140 struct dirstat_dir {
2141 struct dirstat_file *files;
2142 int alloc, nr, permille, cumulative;
2143 };
2144
2145 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2146 unsigned long changed, const char *base, int baselen)
2147 {
2148 unsigned long this_dir = 0;
2149 unsigned int sources = 0;
2150 const char *line_prefix = diff_line_prefix(opt);
2151
2152 while (dir->nr) {
2153 struct dirstat_file *f = dir->files;
2154 int namelen = strlen(f->name);
2155 unsigned long this;
2156 char *slash;
2157
2158 if (namelen < baselen)
2159 break;
2160 if (memcmp(f->name, base, baselen))
2161 break;
2162 slash = strchr(f->name + baselen, '/');
2163 if (slash) {
2164 int newbaselen = slash + 1 - f->name;
2165 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2166 sources++;
2167 } else {
2168 this = f->changed;
2169 dir->files++;
2170 dir->nr--;
2171 sources += 2;
2172 }
2173 this_dir += this;
2174 }
2175
2176 /*
2177 * We don't report dirstat's for
2178 * - the top level
2179 * - or cases where everything came from a single directory
2180 * under this directory (sources == 1).
2181 */
2182 if (baselen && sources != 1) {
2183 if (this_dir) {
2184 int permille = this_dir * 1000 / changed;
2185 if (permille >= dir->permille) {
2186 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2187 permille / 10, permille % 10, baselen, base);
2188 if (!dir->cumulative)
2189 return 0;
2190 }
2191 }
2192 }
2193 return this_dir;
2194 }
2195
2196 static int dirstat_compare(const void *_a, const void *_b)
2197 {
2198 const struct dirstat_file *a = _a;
2199 const struct dirstat_file *b = _b;
2200 return strcmp(a->name, b->name);
2201 }
2202
2203 static void show_dirstat(struct diff_options *options)
2204 {
2205 int i;
2206 unsigned long changed;
2207 struct dirstat_dir dir;
2208 struct diff_queue_struct *q = &diff_queued_diff;
2209
2210 dir.files = NULL;
2211 dir.alloc = 0;
2212 dir.nr = 0;
2213 dir.permille = options->dirstat_permille;
2214 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2215
2216 changed = 0;
2217 for (i = 0; i < q->nr; i++) {
2218 struct diff_filepair *p = q->queue[i];
2219 const char *name;
2220 unsigned long copied, added, damage;
2221 int content_changed;
2222
2223 name = p->two->path ? p->two->path : p->one->path;
2224
2225 if (p->one->oid_valid && p->two->oid_valid)
2226 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2227 else
2228 content_changed = 1;
2229
2230 if (!content_changed) {
2231 /*
2232 * The SHA1 has not changed, so pre-/post-content is
2233 * identical. We can therefore skip looking at the
2234 * file contents altogether.
2235 */
2236 damage = 0;
2237 goto found_damage;
2238 }
2239
2240 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2241 /*
2242 * In --dirstat-by-file mode, we don't really need to
2243 * look at the actual file contents at all.
2244 * The fact that the SHA1 changed is enough for us to
2245 * add this file to the list of results
2246 * (with each file contributing equal damage).
2247 */
2248 damage = 1;
2249 goto found_damage;
2250 }
2251
2252 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2253 diff_populate_filespec(p->one, 0);
2254 diff_populate_filespec(p->two, 0);
2255 diffcore_count_changes(p->one, p->two, NULL, NULL,
2256 &copied, &added);
2257 diff_free_filespec_data(p->one);
2258 diff_free_filespec_data(p->two);
2259 } else if (DIFF_FILE_VALID(p->one)) {
2260 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2261 copied = added = 0;
2262 diff_free_filespec_data(p->one);
2263 } else if (DIFF_FILE_VALID(p->two)) {
2264 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2265 copied = 0;
2266 added = p->two->size;
2267 diff_free_filespec_data(p->two);
2268 } else
2269 continue;
2270
2271 /*
2272 * Original minus copied is the removed material,
2273 * added is the new material. They are both damages
2274 * made to the preimage.
2275 * If the resulting damage is zero, we know that
2276 * diffcore_count_changes() considers the two entries to
2277 * be identical, but since content_changed is true, we
2278 * know that there must have been _some_ kind of change,
2279 * so we force all entries to have damage > 0.
2280 */
2281 damage = (p->one->size - copied) + added;
2282 if (!damage)
2283 damage = 1;
2284
2285 found_damage:
2286 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2287 dir.files[dir.nr].name = name;
2288 dir.files[dir.nr].changed = damage;
2289 changed += damage;
2290 dir.nr++;
2291 }
2292
2293 /* This can happen even with many files, if everything was renames */
2294 if (!changed)
2295 return;
2296
2297 /* Show all directories with more than x% of the changes */
2298 QSORT(dir.files, dir.nr, dirstat_compare);
2299 gather_dirstat(options, &dir, changed, "", 0);
2300 }
2301
2302 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2303 {
2304 int i;
2305 unsigned long changed;
2306 struct dirstat_dir dir;
2307
2308 if (data->nr == 0)
2309 return;
2310
2311 dir.files = NULL;
2312 dir.alloc = 0;
2313 dir.nr = 0;
2314 dir.permille = options->dirstat_permille;
2315 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2316
2317 changed = 0;
2318 for (i = 0; i < data->nr; i++) {
2319 struct diffstat_file *file = data->files[i];
2320 unsigned long damage = file->added + file->deleted;
2321 if (file->is_binary)
2322 /*
2323 * binary files counts bytes, not lines. Must find some
2324 * way to normalize binary bytes vs. textual lines.
2325 * The following heuristic assumes that there are 64
2326 * bytes per "line".
2327 * This is stupid and ugly, but very cheap...
2328 */
2329 damage = (damage + 63) / 64;
2330 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2331 dir.files[dir.nr].name = file->name;
2332 dir.files[dir.nr].changed = damage;
2333 changed += damage;
2334 dir.nr++;
2335 }
2336
2337 /* This can happen even with many files, if everything was renames */
2338 if (!changed)
2339 return;
2340
2341 /* Show all directories with more than x% of the changes */
2342 QSORT(dir.files, dir.nr, dirstat_compare);
2343 gather_dirstat(options, &dir, changed, "", 0);
2344 }
2345
2346 static void free_diffstat_info(struct diffstat_t *diffstat)
2347 {
2348 int i;
2349 for (i = 0; i < diffstat->nr; i++) {
2350 struct diffstat_file *f = diffstat->files[i];
2351 if (f->name != f->print_name)
2352 free(f->print_name);
2353 free(f->name);
2354 free(f->from_name);
2355 free(f);
2356 }
2357 free(diffstat->files);
2358 }
2359
2360 struct checkdiff_t {
2361 const char *filename;
2362 int lineno;
2363 int conflict_marker_size;
2364 struct diff_options *o;
2365 unsigned ws_rule;
2366 unsigned status;
2367 };
2368
2369 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2370 {
2371 char firstchar;
2372 int cnt;
2373
2374 if (len < marker_size + 1)
2375 return 0;
2376 firstchar = line[0];
2377 switch (firstchar) {
2378 case '=': case '>': case '<': case '|':
2379 break;
2380 default:
2381 return 0;
2382 }
2383 for (cnt = 1; cnt < marker_size; cnt++)
2384 if (line[cnt] != firstchar)
2385 return 0;
2386 /* line[1] thru line[marker_size-1] are same as firstchar */
2387 if (len < marker_size + 1 || !isspace(line[marker_size]))
2388 return 0;
2389 return 1;
2390 }
2391
2392 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2393 {
2394 struct checkdiff_t *data = priv;
2395 int marker_size = data->conflict_marker_size;
2396 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2397 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2398 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2399 char *err;
2400 const char *line_prefix;
2401
2402 assert(data->o);
2403 line_prefix = diff_line_prefix(data->o);
2404
2405 if (line[0] == '+') {
2406 unsigned bad;
2407 data->lineno++;
2408 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2409 data->status |= 1;
2410 fprintf(data->o->file,
2411 "%s%s:%d: leftover conflict marker\n",
2412 line_prefix, data->filename, data->lineno);
2413 }
2414 bad = ws_check(line + 1, len - 1, data->ws_rule);
2415 if (!bad)
2416 return;
2417 data->status |= bad;
2418 err = whitespace_error_string(bad);
2419 fprintf(data->o->file, "%s%s:%d: %s.\n",
2420 line_prefix, data->filename, data->lineno, err);
2421 free(err);
2422 emit_line(data->o, set, reset, line, 1);
2423 ws_check_emit(line + 1, len - 1, data->ws_rule,
2424 data->o->file, set, reset, ws);
2425 } else if (line[0] == ' ') {
2426 data->lineno++;
2427 } else if (line[0] == '@') {
2428 char *plus = strchr(line, '+');
2429 if (plus)
2430 data->lineno = strtol(plus, NULL, 10) - 1;
2431 else
2432 die("invalid diff");
2433 }
2434 }
2435
2436 static unsigned char *deflate_it(char *data,
2437 unsigned long size,
2438 unsigned long *result_size)
2439 {
2440 int bound;
2441 unsigned char *deflated;
2442 git_zstream stream;
2443
2444 git_deflate_init(&stream, zlib_compression_level);
2445 bound = git_deflate_bound(&stream, size);
2446 deflated = xmalloc(bound);
2447 stream.next_out = deflated;
2448 stream.avail_out = bound;
2449
2450 stream.next_in = (unsigned char *)data;
2451 stream.avail_in = size;
2452 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2453 ; /* nothing */
2454 git_deflate_end(&stream);
2455 *result_size = stream.total_out;
2456 return deflated;
2457 }
2458
2459 static void emit_binary_diff_body(struct diff_options *o,
2460 mmfile_t *one, mmfile_t *two)
2461 {
2462 void *cp;
2463 void *delta;
2464 void *deflated;
2465 void *data;
2466 unsigned long orig_size;
2467 unsigned long delta_size;
2468 unsigned long deflate_size;
2469 unsigned long data_size;
2470
2471 /* We could do deflated delta, or we could do just deflated two,
2472 * whichever is smaller.
2473 */
2474 delta = NULL;
2475 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2476 if (one->size && two->size) {
2477 delta = diff_delta(one->ptr, one->size,
2478 two->ptr, two->size,
2479 &delta_size, deflate_size);
2480 if (delta) {
2481 void *to_free = delta;
2482 orig_size = delta_size;
2483 delta = deflate_it(delta, delta_size, &delta_size);
2484 free(to_free);
2485 }
2486 }
2487
2488 if (delta && delta_size < deflate_size) {
2489 char *s = xstrfmt("%lu", orig_size);
2490 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
2491 s, strlen(s), 0);
2492 free(s);
2493 free(deflated);
2494 data = delta;
2495 data_size = delta_size;
2496 } else {
2497 char *s = xstrfmt("%lu", two->size);
2498 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
2499 s, strlen(s), 0);
2500 free(s);
2501 free(delta);
2502 data = deflated;
2503 data_size = deflate_size;
2504 }
2505
2506 /* emit data encoded in base85 */
2507 cp = data;
2508 while (data_size) {
2509 int len;
2510 int bytes = (52 < data_size) ? 52 : data_size;
2511 char line[71];
2512 data_size -= bytes;
2513 if (bytes <= 26)
2514 line[0] = bytes + 'A' - 1;
2515 else
2516 line[0] = bytes - 26 + 'a' - 1;
2517 encode_85(line + 1, cp, bytes);
2518 cp = (char *) cp + bytes;
2519
2520 len = strlen(line);
2521 line[len++] = '\n';
2522 line[len] = '\0';
2523
2524 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
2525 line, len, 0);
2526 }
2527 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
2528 free(data);
2529 }
2530
2531 static void emit_binary_diff(struct diff_options *o,
2532 mmfile_t *one, mmfile_t *two)
2533 {
2534 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
2535 emit_binary_diff_body(o, one, two);
2536 emit_binary_diff_body(o, two, one);
2537 }
2538
2539 int diff_filespec_is_binary(struct diff_filespec *one)
2540 {
2541 if (one->is_binary == -1) {
2542 diff_filespec_load_driver(one);
2543 if (one->driver->binary != -1)
2544 one->is_binary = one->driver->binary;
2545 else {
2546 if (!one->data && DIFF_FILE_VALID(one))
2547 diff_populate_filespec(one, CHECK_BINARY);
2548 if (one->is_binary == -1 && one->data)
2549 one->is_binary = buffer_is_binary(one->data,
2550 one->size);
2551 if (one->is_binary == -1)
2552 one->is_binary = 0;
2553 }
2554 }
2555 return one->is_binary;
2556 }
2557
2558 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2559 {
2560 diff_filespec_load_driver(one);
2561 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2562 }
2563
2564 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2565 {
2566 if (!options->a_prefix)
2567 options->a_prefix = a;
2568 if (!options->b_prefix)
2569 options->b_prefix = b;
2570 }
2571
2572 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2573 {
2574 if (!DIFF_FILE_VALID(one))
2575 return NULL;
2576
2577 diff_filespec_load_driver(one);
2578 return userdiff_get_textconv(one->driver);
2579 }
2580
2581 static void builtin_diff(const char *name_a,
2582 const char *name_b,
2583 struct diff_filespec *one,
2584 struct diff_filespec *two,
2585 const char *xfrm_msg,
2586 int must_show_header,
2587 struct diff_options *o,
2588 int complete_rewrite)
2589 {
2590 mmfile_t mf1, mf2;
2591 const char *lbl[2];
2592 char *a_one, *b_two;
2593 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2594 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2595 const char *a_prefix, *b_prefix;
2596 struct userdiff_driver *textconv_one = NULL;
2597 struct userdiff_driver *textconv_two = NULL;
2598 struct strbuf header = STRBUF_INIT;
2599 const char *line_prefix = diff_line_prefix(o);
2600
2601 diff_set_mnemonic_prefix(o, "a/", "b/");
2602 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2603 a_prefix = o->b_prefix;
2604 b_prefix = o->a_prefix;
2605 } else {
2606 a_prefix = o->a_prefix;
2607 b_prefix = o->b_prefix;
2608 }
2609
2610 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
2611 (!one->mode || S_ISGITLINK(one->mode)) &&
2612 (!two->mode || S_ISGITLINK(two->mode))) {
2613 show_submodule_summary(o, one->path ? one->path : two->path,
2614 &one->oid, &two->oid,
2615 two->dirty_submodule);
2616 return;
2617 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
2618 (!one->mode || S_ISGITLINK(one->mode)) &&
2619 (!two->mode || S_ISGITLINK(two->mode))) {
2620 show_submodule_inline_diff(o, one->path ? one->path : two->path,
2621 &one->oid, &two->oid,
2622 two->dirty_submodule);
2623 return;
2624 }
2625
2626 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2627 textconv_one = get_textconv(one);
2628 textconv_two = get_textconv(two);
2629 }
2630
2631 /* Never use a non-valid filename anywhere if at all possible */
2632 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2633 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2634
2635 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2636 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2637 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2638 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2639 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2640 if (lbl[0][0] == '/') {
2641 /* /dev/null */
2642 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2643 if (xfrm_msg)
2644 strbuf_addstr(&header, xfrm_msg);
2645 must_show_header = 1;
2646 }
2647 else if (lbl[1][0] == '/') {
2648 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2649 if (xfrm_msg)
2650 strbuf_addstr(&header, xfrm_msg);
2651 must_show_header = 1;
2652 }
2653 else {
2654 if (one->mode != two->mode) {
2655 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2656 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2657 must_show_header = 1;
2658 }
2659 if (xfrm_msg)
2660 strbuf_addstr(&header, xfrm_msg);
2661
2662 /*
2663 * we do not run diff between different kind
2664 * of objects.
2665 */
2666 if ((one->mode ^ two->mode) & S_IFMT)
2667 goto free_ab_and_return;
2668 if (complete_rewrite &&
2669 (textconv_one || !diff_filespec_is_binary(one)) &&
2670 (textconv_two || !diff_filespec_is_binary(two))) {
2671 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2672 header.buf, header.len, 0);
2673 strbuf_reset(&header);
2674 emit_rewrite_diff(name_a, name_b, one, two,
2675 textconv_one, textconv_two, o);
2676 o->found_changes = 1;
2677 goto free_ab_and_return;
2678 }
2679 }
2680
2681 if (o->irreversible_delete && lbl[1][0] == '/') {
2682 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
2683 header.len, 0);
2684 strbuf_reset(&header);
2685 goto free_ab_and_return;
2686 } else if (!DIFF_OPT_TST(o, TEXT) &&
2687 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2688 (!textconv_two && diff_filespec_is_binary(two)) )) {
2689 struct strbuf sb = STRBUF_INIT;
2690 if (!one->data && !two->data &&
2691 S_ISREG(one->mode) && S_ISREG(two->mode) &&
2692 !DIFF_OPT_TST(o, BINARY)) {
2693 if (!oidcmp(&one->oid, &two->oid)) {
2694 if (must_show_header)
2695 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2696 header.buf, header.len,
2697 0);
2698 goto free_ab_and_return;
2699 }
2700 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2701 header.buf, header.len, 0);
2702 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
2703 diff_line_prefix(o), lbl[0], lbl[1]);
2704 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
2705 sb.buf, sb.len, 0);
2706 strbuf_release(&sb);
2707 goto free_ab_and_return;
2708 }
2709 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2710 die("unable to read files to diff");
2711 /* Quite common confusing case */
2712 if (mf1.size == mf2.size &&
2713 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2714 if (must_show_header)
2715 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2716 header.buf, header.len, 0);
2717 goto free_ab_and_return;
2718 }
2719 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
2720 strbuf_reset(&header);
2721 if (DIFF_OPT_TST(o, BINARY))
2722 emit_binary_diff(o, &mf1, &mf2);
2723 else {
2724 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
2725 diff_line_prefix(o), lbl[0], lbl[1]);
2726 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
2727 sb.buf, sb.len, 0);
2728 strbuf_release(&sb);
2729 }
2730 o->found_changes = 1;
2731 } else {
2732 /* Crazy xdl interfaces.. */
2733 const char *diffopts = getenv("GIT_DIFF_OPTS");
2734 const char *v;
2735 xpparam_t xpp;
2736 xdemitconf_t xecfg;
2737 struct emit_callback ecbdata;
2738 const struct userdiff_funcname *pe;
2739
2740 if (must_show_header) {
2741 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2742 header.buf, header.len, 0);
2743 strbuf_reset(&header);
2744 }
2745
2746 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2747 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2748
2749 pe = diff_funcname_pattern(one);
2750 if (!pe)
2751 pe = diff_funcname_pattern(two);
2752
2753 memset(&xpp, 0, sizeof(xpp));
2754 memset(&xecfg, 0, sizeof(xecfg));
2755 memset(&ecbdata, 0, sizeof(ecbdata));
2756 ecbdata.label_path = lbl;
2757 ecbdata.color_diff = want_color(o->use_color);
2758 ecbdata.ws_rule = whitespace_rule(name_b);
2759 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2760 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2761 ecbdata.opt = o;
2762 ecbdata.header = header.len ? &header : NULL;
2763 xpp.flags = o->xdl_opts;
2764 xecfg.ctxlen = o->context;
2765 xecfg.interhunkctxlen = o->interhunkcontext;
2766 xecfg.flags = XDL_EMIT_FUNCNAMES;
2767 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2768 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2769 if (pe)
2770 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2771 if (!diffopts)
2772 ;
2773 else if (skip_prefix(diffopts, "--unified=", &v))
2774 xecfg.ctxlen = strtoul(v, NULL, 10);
2775 else if (skip_prefix(diffopts, "-u", &v))
2776 xecfg.ctxlen = strtoul(v, NULL, 10);
2777 if (o->word_diff)
2778 init_diff_words_data(&ecbdata, o, one, two);
2779 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2780 &xpp, &xecfg))
2781 die("unable to generate diff for %s", one->path);
2782 if (o->word_diff)
2783 free_diff_words_data(&ecbdata);
2784 if (textconv_one)
2785 free(mf1.ptr);
2786 if (textconv_two)
2787 free(mf2.ptr);
2788 xdiff_clear_find_func(&xecfg);
2789 }
2790
2791 free_ab_and_return:
2792 strbuf_release(&header);
2793 diff_free_filespec_data(one);
2794 diff_free_filespec_data(two);
2795 free(a_one);
2796 free(b_two);
2797 return;
2798 }
2799
2800 static void builtin_diffstat(const char *name_a, const char *name_b,
2801 struct diff_filespec *one,
2802 struct diff_filespec *two,
2803 struct diffstat_t *diffstat,
2804 struct diff_options *o,
2805 struct diff_filepair *p)
2806 {
2807 mmfile_t mf1, mf2;
2808 struct diffstat_file *data;
2809 int same_contents;
2810 int complete_rewrite = 0;
2811
2812 if (!DIFF_PAIR_UNMERGED(p)) {
2813 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2814 complete_rewrite = 1;
2815 }
2816
2817 data = diffstat_add(diffstat, name_a, name_b);
2818 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2819
2820 if (!one || !two) {
2821 data->is_unmerged = 1;
2822 return;
2823 }
2824
2825 same_contents = !oidcmp(&one->oid, &two->oid);
2826
2827 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2828 data->is_binary = 1;
2829 if (same_contents) {
2830 data->added = 0;
2831 data->deleted = 0;
2832 } else {
2833 data->added = diff_filespec_size(two);
2834 data->deleted = diff_filespec_size(one);
2835 }
2836 }
2837
2838 else if (complete_rewrite) {
2839 diff_populate_filespec(one, 0);
2840 diff_populate_filespec(two, 0);
2841 data->deleted = count_lines(one->data, one->size);
2842 data->added = count_lines(two->data, two->size);
2843 }
2844
2845 else if (!same_contents) {
2846 /* Crazy xdl interfaces.. */
2847 xpparam_t xpp;
2848 xdemitconf_t xecfg;
2849
2850 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2851 die("unable to read files to diff");
2852
2853 memset(&xpp, 0, sizeof(xpp));
2854 memset(&xecfg, 0, sizeof(xecfg));
2855 xpp.flags = o->xdl_opts;
2856 xecfg.ctxlen = o->context;
2857 xecfg.interhunkctxlen = o->interhunkcontext;
2858 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2859 &xpp, &xecfg))
2860 die("unable to generate diffstat for %s", one->path);
2861 }
2862
2863 diff_free_filespec_data(one);
2864 diff_free_filespec_data(two);
2865 }
2866
2867 static void builtin_checkdiff(const char *name_a, const char *name_b,
2868 const char *attr_path,
2869 struct diff_filespec *one,
2870 struct diff_filespec *two,
2871 struct diff_options *o)
2872 {
2873 mmfile_t mf1, mf2;
2874 struct checkdiff_t data;
2875
2876 if (!two)
2877 return;
2878
2879 memset(&data, 0, sizeof(data));
2880 data.filename = name_b ? name_b : name_a;
2881 data.lineno = 0;
2882 data.o = o;
2883 data.ws_rule = whitespace_rule(attr_path);
2884 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2885
2886 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2887 die("unable to read files to diff");
2888
2889 /*
2890 * All the other codepaths check both sides, but not checking
2891 * the "old" side here is deliberate. We are checking the newly
2892 * introduced changes, and as long as the "new" side is text, we
2893 * can and should check what it introduces.
2894 */
2895 if (diff_filespec_is_binary(two))
2896 goto free_and_return;
2897 else {
2898 /* Crazy xdl interfaces.. */
2899 xpparam_t xpp;
2900 xdemitconf_t xecfg;
2901
2902 memset(&xpp, 0, sizeof(xpp));
2903 memset(&xecfg, 0, sizeof(xecfg));
2904 xecfg.ctxlen = 1; /* at least one context line */
2905 xpp.flags = 0;
2906 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2907 &xpp, &xecfg))
2908 die("unable to generate checkdiff for %s", one->path);
2909
2910 if (data.ws_rule & WS_BLANK_AT_EOF) {
2911 struct emit_callback ecbdata;
2912 int blank_at_eof;
2913
2914 ecbdata.ws_rule = data.ws_rule;
2915 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2916 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2917
2918 if (blank_at_eof) {
2919 static char *err;
2920 if (!err)
2921 err = whitespace_error_string(WS_BLANK_AT_EOF);
2922 fprintf(o->file, "%s:%d: %s.\n",
2923 data.filename, blank_at_eof, err);
2924 data.status = 1; /* report errors */
2925 }
2926 }
2927 }
2928 free_and_return:
2929 diff_free_filespec_data(one);
2930 diff_free_filespec_data(two);
2931 if (data.status)
2932 DIFF_OPT_SET(o, CHECK_FAILED);
2933 }
2934
2935 struct diff_filespec *alloc_filespec(const char *path)
2936 {
2937 struct diff_filespec *spec;
2938
2939 FLEXPTR_ALLOC_STR(spec, path, path);
2940 spec->count = 1;
2941 spec->is_binary = -1;
2942 return spec;
2943 }
2944
2945 void free_filespec(struct diff_filespec *spec)
2946 {
2947 if (!--spec->count) {
2948 diff_free_filespec_data(spec);
2949 free(spec);
2950 }
2951 }
2952
2953 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
2954 int oid_valid, unsigned short mode)
2955 {
2956 if (mode) {
2957 spec->mode = canon_mode(mode);
2958 oidcpy(&spec->oid, oid);
2959 spec->oid_valid = oid_valid;
2960 }
2961 }
2962
2963 /*
2964 * Given a name and sha1 pair, if the index tells us the file in
2965 * the work tree has that object contents, return true, so that
2966 * prepare_temp_file() does not have to inflate and extract.
2967 */
2968 static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
2969 {
2970 const struct cache_entry *ce;
2971 struct stat st;
2972 int pos, len;
2973
2974 /*
2975 * We do not read the cache ourselves here, because the
2976 * benchmark with my previous version that always reads cache
2977 * shows that it makes things worse for diff-tree comparing
2978 * two linux-2.6 kernel trees in an already checked out work
2979 * tree. This is because most diff-tree comparisons deal with
2980 * only a small number of files, while reading the cache is
2981 * expensive for a large project, and its cost outweighs the
2982 * savings we get by not inflating the object to a temporary
2983 * file. Practically, this code only helps when we are used
2984 * by diff-cache --cached, which does read the cache before
2985 * calling us.
2986 */
2987 if (!active_cache)
2988 return 0;
2989
2990 /* We want to avoid the working directory if our caller
2991 * doesn't need the data in a normal file, this system
2992 * is rather slow with its stat/open/mmap/close syscalls,
2993 * and the object is contained in a pack file. The pack
2994 * is probably already open and will be faster to obtain
2995 * the data through than the working directory. Loose
2996 * objects however would tend to be slower as they need
2997 * to be individually opened and inflated.
2998 */
2999 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(oid->hash))
3000 return 0;
3001
3002 /*
3003 * Similarly, if we'd have to convert the file contents anyway, that
3004 * makes the optimization not worthwhile.
3005 */
3006 if (!want_file && would_convert_to_git(&the_index, name))
3007 return 0;
3008
3009 len = strlen(name);
3010 pos = cache_name_pos(name, len);
3011 if (pos < 0)
3012 return 0;
3013 ce = active_cache[pos];
3014
3015 /*
3016 * This is not the sha1 we are looking for, or
3017 * unreusable because it is not a regular file.
3018 */
3019 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3020 return 0;
3021
3022 /*
3023 * If ce is marked as "assume unchanged", there is no
3024 * guarantee that work tree matches what we are looking for.
3025 */
3026 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3027 return 0;
3028
3029 /*
3030 * If ce matches the file in the work tree, we can reuse it.
3031 */
3032 if (ce_uptodate(ce) ||
3033 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3034 return 1;
3035
3036 return 0;
3037 }
3038
3039 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3040 {
3041 struct strbuf buf = STRBUF_INIT;
3042 char *dirty = "";
3043
3044 /* Are we looking at the work tree? */
3045 if (s->dirty_submodule)
3046 dirty = "-dirty";
3047
3048 strbuf_addf(&buf, "Subproject commit %s%s\n",
3049 oid_to_hex(&s->oid), dirty);
3050 s->size = buf.len;
3051 if (size_only) {
3052 s->data = NULL;
3053 strbuf_release(&buf);
3054 } else {
3055 s->data = strbuf_detach(&buf, NULL);
3056 s->should_free = 1;
3057 }
3058 return 0;
3059 }
3060
3061 /*
3062 * While doing rename detection and pickaxe operation, we may need to
3063 * grab the data for the blob (or file) for our own in-core comparison.
3064 * diff_filespec has data and size fields for this purpose.
3065 */
3066 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3067 {
3068 int size_only = flags & CHECK_SIZE_ONLY;
3069 int err = 0;
3070 /*
3071 * demote FAIL to WARN to allow inspecting the situation
3072 * instead of refusing.
3073 */
3074 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
3075 ? SAFE_CRLF_WARN
3076 : safe_crlf);
3077
3078 if (!DIFF_FILE_VALID(s))
3079 die("internal error: asking to populate invalid file.");
3080 if (S_ISDIR(s->mode))
3081 return -1;
3082
3083 if (s->data)
3084 return 0;
3085
3086 if (size_only && 0 < s->size)
3087 return 0;
3088
3089 if (S_ISGITLINK(s->mode))
3090 return diff_populate_gitlink(s, size_only);
3091
3092 if (!s->oid_valid ||
3093 reuse_worktree_file(s->path, &s->oid, 0)) {
3094 struct strbuf buf = STRBUF_INIT;
3095 struct stat st;
3096 int fd;
3097
3098 if (lstat(s->path, &st) < 0) {
3099 if (errno == ENOENT) {
3100 err_empty:
3101 err = -1;
3102 empty:
3103 s->data = (char *)"";
3104 s->size = 0;
3105 return err;
3106 }
3107 }
3108 s->size = xsize_t(st.st_size);
3109 if (!s->size)
3110 goto empty;
3111 if (S_ISLNK(st.st_mode)) {
3112 struct strbuf sb = STRBUF_INIT;
3113
3114 if (strbuf_readlink(&sb, s->path, s->size))
3115 goto err_empty;
3116 s->size = sb.len;
3117 s->data = strbuf_detach(&sb, NULL);
3118 s->should_free = 1;
3119 return 0;
3120 }
3121
3122 /*
3123 * Even if the caller would be happy with getting
3124 * only the size, we cannot return early at this
3125 * point if the path requires us to run the content
3126 * conversion.
3127 */
3128 if (size_only && !would_convert_to_git(&the_index, s->path))
3129 return 0;
3130
3131 /*
3132 * Note: this check uses xsize_t(st.st_size) that may
3133 * not be the true size of the blob after it goes
3134 * through convert_to_git(). This may not strictly be
3135 * correct, but the whole point of big_file_threshold
3136 * and is_binary check being that we want to avoid
3137 * opening the file and inspecting the contents, this
3138 * is probably fine.
3139 */
3140 if ((flags & CHECK_BINARY) &&
3141 s->size > big_file_threshold && s->is_binary == -1) {
3142 s->is_binary = 1;
3143 return 0;
3144 }
3145 fd = open(s->path, O_RDONLY);
3146 if (fd < 0)
3147 goto err_empty;
3148 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3149 close(fd);
3150 s->should_munmap = 1;
3151
3152 /*
3153 * Convert from working tree format to canonical git format
3154 */
3155 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, crlf_warn)) {
3156 size_t size = 0;
3157 munmap(s->data, s->size);
3158 s->should_munmap = 0;
3159 s->data = strbuf_detach(&buf, &size);
3160 s->size = size;
3161 s->should_free = 1;
3162 }
3163 }
3164 else {
3165 enum object_type type;
3166 if (size_only || (flags & CHECK_BINARY)) {
3167 type = sha1_object_info(s->oid.hash, &s->size);
3168 if (type < 0)
3169 die("unable to read %s",
3170 oid_to_hex(&s->oid));
3171 if (size_only)
3172 return 0;
3173 if (s->size > big_file_threshold && s->is_binary == -1) {
3174 s->is_binary = 1;
3175 return 0;
3176 }
3177 }
3178 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
3179 if (!s->data)
3180 die("unable to read %s", oid_to_hex(&s->oid));
3181 s->should_free = 1;
3182 }
3183 return 0;
3184 }
3185
3186 void diff_free_filespec_blob(struct diff_filespec *s)
3187 {
3188 if (s->should_free)
3189 free(s->data);
3190 else if (s->should_munmap)
3191 munmap(s->data, s->size);
3192
3193 if (s->should_free || s->should_munmap) {
3194 s->should_free = s->should_munmap = 0;
3195 s->data = NULL;
3196 }
3197 }
3198
3199 void diff_free_filespec_data(struct diff_filespec *s)
3200 {
3201 diff_free_filespec_blob(s);
3202 FREE_AND_NULL(s->cnt_data);
3203 }
3204
3205 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3206 void *blob,
3207 unsigned long size,
3208 const struct object_id *oid,
3209 int mode)
3210 {
3211 int fd;
3212 struct strbuf buf = STRBUF_INIT;
3213 struct strbuf template = STRBUF_INIT;
3214 char *path_dup = xstrdup(path);
3215 const char *base = basename(path_dup);
3216
3217 /* Generate "XXXXXX_basename.ext" */
3218 strbuf_addstr(&template, "XXXXXX_");
3219 strbuf_addstr(&template, base);
3220
3221 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
3222 if (fd < 0)
3223 die_errno("unable to create temp-file");
3224 if (convert_to_working_tree(path,
3225 (const char *)blob, (size_t)size, &buf)) {
3226 blob = buf.buf;
3227 size = buf.len;
3228 }
3229 if (write_in_full(fd, blob, size) != size)
3230 die_errno("unable to write temp-file");
3231 close_tempfile(&temp->tempfile);
3232 temp->name = get_tempfile_path(&temp->tempfile);
3233 oid_to_hex_r(temp->hex, oid);
3234 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3235 strbuf_release(&buf);
3236 strbuf_release(&template);
3237 free(path_dup);
3238 }
3239
3240 static struct diff_tempfile *prepare_temp_file(const char *name,
3241 struct diff_filespec *one)
3242 {
3243 struct diff_tempfile *temp = claim_diff_tempfile();
3244
3245 if (!DIFF_FILE_VALID(one)) {
3246 not_a_valid_file:
3247 /* A '-' entry produces this for file-2, and
3248 * a '+' entry produces this for file-1.
3249 */
3250 temp->name = "/dev/null";
3251 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3252 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3253 return temp;
3254 }
3255
3256 if (!S_ISGITLINK(one->mode) &&
3257 (!one->oid_valid ||
3258 reuse_worktree_file(name, &one->oid, 1))) {
3259 struct stat st;
3260 if (lstat(name, &st) < 0) {
3261 if (errno == ENOENT)
3262 goto not_a_valid_file;
3263 die_errno("stat(%s)", name);
3264 }
3265 if (S_ISLNK(st.st_mode)) {
3266 struct strbuf sb = STRBUF_INIT;
3267 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3268 die_errno("readlink(%s)", name);
3269 prep_temp_blob(name, temp, sb.buf, sb.len,
3270 (one->oid_valid ?
3271 &one->oid : &null_oid),
3272 (one->oid_valid ?
3273 one->mode : S_IFLNK));
3274 strbuf_release(&sb);
3275 }
3276 else {
3277 /* we can borrow from the file in the work tree */
3278 temp->name = name;
3279 if (!one->oid_valid)
3280 oid_to_hex_r(temp->hex, &null_oid);
3281 else
3282 oid_to_hex_r(temp->hex, &one->oid);
3283 /* Even though we may sometimes borrow the
3284 * contents from the work tree, we always want
3285 * one->mode. mode is trustworthy even when
3286 * !(one->oid_valid), as long as
3287 * DIFF_FILE_VALID(one).
3288 */
3289 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3290 }
3291 return temp;
3292 }
3293 else {
3294 if (diff_populate_filespec(one, 0))
3295 die("cannot read data blob for %s", one->path);
3296 prep_temp_blob(name, temp, one->data, one->size,
3297 &one->oid, one->mode);
3298 }
3299 return temp;
3300 }
3301
3302 static void add_external_diff_name(struct argv_array *argv,
3303 const char *name,
3304 struct diff_filespec *df)
3305 {
3306 struct diff_tempfile *temp = prepare_temp_file(name, df);
3307 argv_array_push(argv, temp->name);
3308 argv_array_push(argv, temp->hex);
3309 argv_array_push(argv, temp->mode);
3310 }
3311
3312 /* An external diff command takes:
3313 *
3314 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3315 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3316 *
3317 */
3318 static void run_external_diff(const char *pgm,
3319 const char *name,
3320 const char *other,
3321 struct diff_filespec *one,
3322 struct diff_filespec *two,
3323 const char *xfrm_msg,
3324 int complete_rewrite,
3325 struct diff_options *o)
3326 {
3327 struct argv_array argv = ARGV_ARRAY_INIT;
3328 struct argv_array env = ARGV_ARRAY_INIT;
3329 struct diff_queue_struct *q = &diff_queued_diff;
3330
3331 argv_array_push(&argv, pgm);
3332 argv_array_push(&argv, name);
3333
3334 if (one && two) {
3335 add_external_diff_name(&argv, name, one);
3336 if (!other)
3337 add_external_diff_name(&argv, name, two);
3338 else {
3339 add_external_diff_name(&argv, other, two);
3340 argv_array_push(&argv, other);
3341 argv_array_push(&argv, xfrm_msg);
3342 }
3343 }
3344
3345 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3346 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3347
3348 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3349 die(_("external diff died, stopping at %s"), name);
3350
3351 remove_tempfile();
3352 argv_array_clear(&argv);
3353 argv_array_clear(&env);
3354 }
3355
3356 static int similarity_index(struct diff_filepair *p)
3357 {
3358 return p->score * 100 / MAX_SCORE;
3359 }
3360
3361 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3362 {
3363 if (startup_info->have_repository)
3364 return find_unique_abbrev(oid->hash, abbrev);
3365 else {
3366 char *hex = oid_to_hex(oid);
3367 if (abbrev < 0)
3368 abbrev = FALLBACK_DEFAULT_ABBREV;
3369 if (abbrev > GIT_SHA1_HEXSZ)
3370 die("BUG: oid abbreviation out of range: %d", abbrev);
3371 if (abbrev)
3372 hex[abbrev] = '\0';
3373 return hex;
3374 }
3375 }
3376
3377 static void fill_metainfo(struct strbuf *msg,
3378 const char *name,
3379 const char *other,
3380 struct diff_filespec *one,
3381 struct diff_filespec *two,
3382 struct diff_options *o,
3383 struct diff_filepair *p,
3384 int *must_show_header,
3385 int use_color)
3386 {
3387 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3388 const char *reset = diff_get_color(use_color, DIFF_RESET);
3389 const char *line_prefix = diff_line_prefix(o);
3390
3391 *must_show_header = 1;
3392 strbuf_init(msg, PATH_MAX * 2 + 300);
3393 switch (p->status) {
3394 case DIFF_STATUS_COPIED:
3395 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3396 line_prefix, set, similarity_index(p));
3397 strbuf_addf(msg, "%s\n%s%scopy from ",
3398 reset, line_prefix, set);
3399 quote_c_style(name, msg, NULL, 0);
3400 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3401 quote_c_style(other, msg, NULL, 0);
3402 strbuf_addf(msg, "%s\n", reset);
3403 break;
3404 case DIFF_STATUS_RENAMED:
3405 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3406 line_prefix, set, similarity_index(p));
3407 strbuf_addf(msg, "%s\n%s%srename from ",
3408 reset, line_prefix, set);
3409 quote_c_style(name, msg, NULL, 0);
3410 strbuf_addf(msg, "%s\n%s%srename to ",
3411 reset, line_prefix, set);
3412 quote_c_style(other, msg, NULL, 0);
3413 strbuf_addf(msg, "%s\n", reset);
3414 break;
3415 case DIFF_STATUS_MODIFIED:
3416 if (p->score) {
3417 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3418 line_prefix,
3419 set, similarity_index(p), reset);
3420 break;
3421 }
3422 /* fallthru */
3423 default:
3424 *must_show_header = 0;
3425 }
3426 if (one && two && oidcmp(&one->oid, &two->oid)) {
3427 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3428
3429 if (DIFF_OPT_TST(o, BINARY)) {
3430 mmfile_t mf;
3431 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3432 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3433 abbrev = 40;
3434 }
3435 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3436 diff_abbrev_oid(&one->oid, abbrev),
3437 diff_abbrev_oid(&two->oid, abbrev));
3438 if (one->mode == two->mode)
3439 strbuf_addf(msg, " %06o", one->mode);
3440 strbuf_addf(msg, "%s\n", reset);
3441 }
3442 }
3443
3444 static void run_diff_cmd(const char *pgm,
3445 const char *name,
3446 const char *other,
3447 const char *attr_path,
3448 struct diff_filespec *one,
3449 struct diff_filespec *two,
3450 struct strbuf *msg,
3451 struct diff_options *o,
3452 struct diff_filepair *p)
3453 {
3454 const char *xfrm_msg = NULL;
3455 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3456 int must_show_header = 0;
3457
3458
3459 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3460 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3461 if (drv && drv->external)
3462 pgm = drv->external;
3463 }
3464
3465 if (msg) {
3466 /*
3467 * don't use colors when the header is intended for an
3468 * external diff driver
3469 */
3470 fill_metainfo(msg, name, other, one, two, o, p,
3471 &must_show_header,
3472 want_color(o->use_color) && !pgm);
3473 xfrm_msg = msg->len ? msg->buf : NULL;
3474 }
3475
3476 if (pgm) {
3477 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3478 complete_rewrite, o);
3479 return;
3480 }
3481 if (one && two)
3482 builtin_diff(name, other ? other : name,
3483 one, two, xfrm_msg, must_show_header,
3484 o, complete_rewrite);
3485 else
3486 fprintf(o->file, "* Unmerged path %s\n", name);
3487 }
3488
3489 static void diff_fill_oid_info(struct diff_filespec *one)
3490 {
3491 if (DIFF_FILE_VALID(one)) {
3492 if (!one->oid_valid) {
3493 struct stat st;
3494 if (one->is_stdin) {
3495 oidclr(&one->oid);
3496 return;
3497 }
3498 if (lstat(one->path, &st) < 0)
3499 die_errno("stat '%s'", one->path);
3500 if (index_path(one->oid.hash, one->path, &st, 0))
3501 die("cannot hash %s", one->path);
3502 }
3503 }
3504 else
3505 oidclr(&one->oid);
3506 }
3507
3508 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3509 {
3510 /* Strip the prefix but do not molest /dev/null and absolute paths */
3511 if (*namep && **namep != '/') {
3512 *namep += prefix_length;
3513 if (**namep == '/')
3514 ++*namep;
3515 }
3516 if (*otherp && **otherp != '/') {
3517 *otherp += prefix_length;
3518 if (**otherp == '/')
3519 ++*otherp;
3520 }
3521 }
3522
3523 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3524 {
3525 const char *pgm = external_diff();
3526 struct strbuf msg;
3527 struct diff_filespec *one = p->one;
3528 struct diff_filespec *two = p->two;
3529 const char *name;
3530 const char *other;
3531 const char *attr_path;
3532
3533 name = one->path;
3534 other = (strcmp(name, two->path) ? two->path : NULL);
3535 attr_path = name;
3536 if (o->prefix_length)
3537 strip_prefix(o->prefix_length, &name, &other);
3538
3539 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3540 pgm = NULL;
3541
3542 if (DIFF_PAIR_UNMERGED(p)) {
3543 run_diff_cmd(pgm, name, NULL, attr_path,
3544 NULL, NULL, NULL, o, p);
3545 return;
3546 }
3547
3548 diff_fill_oid_info(one);
3549 diff_fill_oid_info(two);
3550
3551 if (!pgm &&
3552 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3553 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3554 /*
3555 * a filepair that changes between file and symlink
3556 * needs to be split into deletion and creation.
3557 */
3558 struct diff_filespec *null = alloc_filespec(two->path);
3559 run_diff_cmd(NULL, name, other, attr_path,
3560 one, null, &msg, o, p);
3561 free(null);
3562 strbuf_release(&msg);
3563
3564 null = alloc_filespec(one->path);
3565 run_diff_cmd(NULL, name, other, attr_path,
3566 null, two, &msg, o, p);
3567 free(null);
3568 }
3569 else
3570 run_diff_cmd(pgm, name, other, attr_path,
3571 one, two, &msg, o, p);
3572
3573 strbuf_release(&msg);
3574 }
3575
3576 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3577 struct diffstat_t *diffstat)
3578 {
3579 const char *name;
3580 const char *other;
3581
3582 if (DIFF_PAIR_UNMERGED(p)) {
3583 /* unmerged */
3584 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3585 return;
3586 }
3587
3588 name = p->one->path;
3589 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3590
3591 if (o->prefix_length)
3592 strip_prefix(o->prefix_length, &name, &other);
3593
3594 diff_fill_oid_info(p->one);
3595 diff_fill_oid_info(p->two);
3596
3597 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3598 }
3599
3600 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3601 {
3602 const char *name;
3603 const char *other;
3604 const char *attr_path;
3605
3606 if (DIFF_PAIR_UNMERGED(p)) {
3607 /* unmerged */
3608 return;
3609 }
3610
3611 name = p->one->path;
3612 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3613 attr_path = other ? other : name;
3614
3615 if (o->prefix_length)
3616 strip_prefix(o->prefix_length, &name, &other);
3617
3618 diff_fill_oid_info(p->one);
3619 diff_fill_oid_info(p->two);
3620
3621 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3622 }
3623
3624 void diff_setup(struct diff_options *options)
3625 {
3626 memcpy(options, &default_diff_options, sizeof(*options));
3627
3628 options->file = stdout;
3629
3630 options->abbrev = DEFAULT_ABBREV;
3631 options->line_termination = '\n';
3632 options->break_opt = -1;
3633 options->rename_limit = -1;
3634 options->dirstat_permille = diff_dirstat_permille_default;
3635 options->context = diff_context_default;
3636 options->interhunkcontext = diff_interhunk_context_default;
3637 options->ws_error_highlight = ws_error_highlight_default;
3638 DIFF_OPT_SET(options, RENAME_EMPTY);
3639
3640 /* pathchange left =NULL by default */
3641 options->change = diff_change;
3642 options->add_remove = diff_addremove;
3643 options->use_color = diff_use_color_default;
3644 options->detect_rename = diff_detect_rename_default;
3645 options->xdl_opts |= diff_algorithm;
3646 if (diff_indent_heuristic)
3647 DIFF_XDL_SET(options, INDENT_HEURISTIC);
3648
3649 options->orderfile = diff_order_file_cfg;
3650
3651 if (diff_no_prefix) {
3652 options->a_prefix = options->b_prefix = "";
3653 } else if (!diff_mnemonic_prefix) {
3654 options->a_prefix = "a/";
3655 options->b_prefix = "b/";
3656 }
3657 }
3658
3659 void diff_setup_done(struct diff_options *options)
3660 {
3661 int count = 0;
3662
3663 if (options->set_default)
3664 options->set_default(options);
3665
3666 if (options->output_format & DIFF_FORMAT_NAME)
3667 count++;
3668 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3669 count++;
3670 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3671 count++;
3672 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3673 count++;
3674 if (count > 1)
3675 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
3676
3677 /*
3678 * Most of the time we can say "there are changes"
3679 * only by checking if there are changed paths, but
3680 * --ignore-whitespace* options force us to look
3681 * inside contents.
3682 */
3683
3684 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3685 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3686 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3687 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3688 else
3689 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3690
3691 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3692 options->detect_rename = DIFF_DETECT_COPY;
3693
3694 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3695 options->prefix = NULL;
3696 if (options->prefix)
3697 options->prefix_length = strlen(options->prefix);
3698 else
3699 options->prefix_length = 0;
3700
3701 if (options->output_format & (DIFF_FORMAT_NAME |
3702 DIFF_FORMAT_NAME_STATUS |
3703 DIFF_FORMAT_CHECKDIFF |
3704 DIFF_FORMAT_NO_OUTPUT))
3705 options->output_format &= ~(DIFF_FORMAT_RAW |
3706 DIFF_FORMAT_NUMSTAT |
3707 DIFF_FORMAT_DIFFSTAT |
3708 DIFF_FORMAT_SHORTSTAT |
3709 DIFF_FORMAT_DIRSTAT |
3710 DIFF_FORMAT_SUMMARY |
3711 DIFF_FORMAT_PATCH);
3712
3713 /*
3714 * These cases always need recursive; we do not drop caller-supplied
3715 * recursive bits for other formats here.
3716 */
3717 if (options->output_format & (DIFF_FORMAT_PATCH |
3718 DIFF_FORMAT_NUMSTAT |
3719 DIFF_FORMAT_DIFFSTAT |
3720 DIFF_FORMAT_SHORTSTAT |
3721 DIFF_FORMAT_DIRSTAT |
3722 DIFF_FORMAT_SUMMARY |
3723 DIFF_FORMAT_CHECKDIFF))
3724 DIFF_OPT_SET(options, RECURSIVE);
3725 /*
3726 * Also pickaxe would not work very well if you do not say recursive
3727 */
3728 if (options->pickaxe)
3729 DIFF_OPT_SET(options, RECURSIVE);
3730 /*
3731 * When patches are generated, submodules diffed against the work tree
3732 * must be checked for dirtiness too so it can be shown in the output
3733 */
3734 if (options->output_format & DIFF_FORMAT_PATCH)
3735 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3736
3737 if (options->detect_rename && options->rename_limit < 0)
3738 options->rename_limit = diff_rename_limit_default;
3739 if (options->setup & DIFF_SETUP_USE_CACHE) {
3740 if (!active_cache)
3741 /* read-cache does not die even when it fails
3742 * so it is safe for us to do this here. Also
3743 * it does not smudge active_cache or active_nr
3744 * when it fails, so we do not have to worry about
3745 * cleaning it up ourselves either.
3746 */
3747 read_cache();
3748 }
3749 if (40 < options->abbrev)
3750 options->abbrev = 40; /* full */
3751
3752 /*
3753 * It does not make sense to show the first hit we happened
3754 * to have found. It does not make sense not to return with
3755 * exit code in such a case either.
3756 */
3757 if (DIFF_OPT_TST(options, QUICK)) {
3758 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3759 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3760 }
3761
3762 options->diff_path_counter = 0;
3763
3764 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3765 die(_("--follow requires exactly one pathspec"));
3766 }
3767
3768 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3769 {
3770 char c, *eq;
3771 int len;
3772
3773 if (*arg != '-')
3774 return 0;
3775 c = *++arg;
3776 if (!c)
3777 return 0;
3778 if (c == arg_short) {
3779 c = *++arg;
3780 if (!c)
3781 return 1;
3782 if (val && isdigit(c)) {
3783 char *end;
3784 int n = strtoul(arg, &end, 10);
3785 if (*end)
3786 return 0;
3787 *val = n;
3788 return 1;
3789 }
3790 return 0;
3791 }
3792 if (c != '-')
3793 return 0;
3794 arg++;
3795 eq = strchrnul(arg, '=');
3796 len = eq - arg;
3797 if (!len || strncmp(arg, arg_long, len))
3798 return 0;
3799 if (*eq) {
3800 int n;
3801 char *end;
3802 if (!isdigit(*++eq))
3803 return 0;
3804 n = strtoul(eq, &end, 10);
3805 if (*end)
3806 return 0;
3807 *val = n;
3808 }
3809 return 1;
3810 }
3811
3812 static int diff_scoreopt_parse(const char *opt);
3813
3814 static inline int short_opt(char opt, const char **argv,
3815 const char **optarg)
3816 {
3817 const char *arg = argv[0];
3818 if (arg[0] != '-' || arg[1] != opt)
3819 return 0;
3820 if (arg[2] != '\0') {
3821 *optarg = arg + 2;
3822 return 1;
3823 }
3824 if (!argv[1])
3825 die("Option '%c' requires a value", opt);
3826 *optarg = argv[1];
3827 return 2;
3828 }
3829
3830 int parse_long_opt(const char *opt, const char **argv,
3831 const char **optarg)
3832 {
3833 const char *arg = argv[0];
3834 if (!skip_prefix(arg, "--", &arg))
3835 return 0;
3836 if (!skip_prefix(arg, opt, &arg))
3837 return 0;
3838 if (*arg == '=') { /* stuck form: --option=value */
3839 *optarg = arg + 1;
3840 return 1;
3841 }
3842 if (*arg != '\0')
3843 return 0;
3844 /* separate form: --option value */
3845 if (!argv[1])
3846 die("Option '--%s' requires a value", opt);
3847 *optarg = argv[1];
3848 return 2;
3849 }
3850
3851 static int stat_opt(struct diff_options *options, const char **av)
3852 {
3853 const char *arg = av[0];
3854 char *end;
3855 int width = options->stat_width;
3856 int name_width = options->stat_name_width;
3857 int graph_width = options->stat_graph_width;
3858 int count = options->stat_count;
3859 int argcount = 1;
3860
3861 if (!skip_prefix(arg, "--stat", &arg))
3862 die("BUG: stat option does not begin with --stat: %s", arg);
3863 end = (char *)arg;
3864
3865 switch (*arg) {
3866 case '-':
3867 if (skip_prefix(arg, "-width", &arg)) {
3868 if (*arg == '=')
3869 width = strtoul(arg + 1, &end, 10);
3870 else if (!*arg && !av[1])
3871 die_want_option("--stat-width");
3872 else if (!*arg) {
3873 width = strtoul(av[1], &end, 10);
3874 argcount = 2;
3875 }
3876 } else if (skip_prefix(arg, "-name-width", &arg)) {
3877 if (*arg == '=')
3878 name_width = strtoul(arg + 1, &end, 10);
3879 else if (!*arg && !av[1])
3880 die_want_option("--stat-name-width");
3881 else if (!*arg) {
3882 name_width = strtoul(av[1], &end, 10);
3883 argcount = 2;
3884 }
3885 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3886 if (*arg == '=')
3887 graph_width = strtoul(arg + 1, &end, 10);
3888 else if (!*arg && !av[1])
3889 die_want_option("--stat-graph-width");
3890 else if (!*arg) {
3891 graph_width = strtoul(av[1], &end, 10);
3892 argcount = 2;
3893 }
3894 } else if (skip_prefix(arg, "-count", &arg)) {
3895 if (*arg == '=')
3896 count = strtoul(arg + 1, &end, 10);
3897 else if (!*arg && !av[1])
3898 die_want_option("--stat-count");
3899 else if (!*arg) {
3900 count = strtoul(av[1], &end, 10);
3901 argcount = 2;
3902 }
3903 }
3904 break;
3905 case '=':
3906 width = strtoul(arg+1, &end, 10);
3907 if (*end == ',')
3908 name_width = strtoul(end+1, &end, 10);
3909 if (*end == ',')
3910 count = strtoul(end+1, &end, 10);
3911 }
3912
3913 /* Important! This checks all the error cases! */
3914 if (*end)
3915 return 0;
3916 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3917 options->stat_name_width = name_width;
3918 options->stat_graph_width = graph_width;
3919 options->stat_width = width;
3920 options->stat_count = count;
3921 return argcount;
3922 }
3923
3924 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3925 {
3926 struct strbuf errmsg = STRBUF_INIT;
3927 if (parse_dirstat_params(options, params, &errmsg))
3928 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3929 errmsg.buf);
3930 strbuf_release(&errmsg);
3931 /*
3932 * The caller knows a dirstat-related option is given from the command
3933 * line; allow it to say "return this_function();"
3934 */
3935 options->output_format |= DIFF_FORMAT_DIRSTAT;
3936 return 1;
3937 }
3938
3939 static int parse_submodule_opt(struct diff_options *options, const char *value)
3940 {
3941 if (parse_submodule_params(options, value))
3942 die(_("Failed to parse --submodule option parameter: '%s'"),
3943 value);
3944 return 1;
3945 }
3946
3947 static const char diff_status_letters[] = {
3948 DIFF_STATUS_ADDED,
3949 DIFF_STATUS_COPIED,
3950 DIFF_STATUS_DELETED,
3951 DIFF_STATUS_MODIFIED,
3952 DIFF_STATUS_RENAMED,
3953 DIFF_STATUS_TYPE_CHANGED,
3954 DIFF_STATUS_UNKNOWN,
3955 DIFF_STATUS_UNMERGED,
3956 DIFF_STATUS_FILTER_AON,
3957 DIFF_STATUS_FILTER_BROKEN,
3958 '\0',
3959 };
3960
3961 static unsigned int filter_bit['Z' + 1];
3962
3963 static void prepare_filter_bits(void)
3964 {
3965 int i;
3966
3967 if (!filter_bit[DIFF_STATUS_ADDED]) {
3968 for (i = 0; diff_status_letters[i]; i++)
3969 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3970 }
3971 }
3972
3973 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3974 {
3975 return opt->filter & filter_bit[(int) status];
3976 }
3977
3978 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3979 {
3980 int i, optch;
3981
3982 prepare_filter_bits();
3983
3984 /*
3985 * If there is a negation e.g. 'd' in the input, and we haven't
3986 * initialized the filter field with another --diff-filter, start
3987 * from full set of bits, except for AON.
3988 */
3989 if (!opt->filter) {
3990 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3991 if (optch < 'a' || 'z' < optch)
3992 continue;
3993 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3994 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3995 break;
3996 }
3997 }
3998
3999 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4000 unsigned int bit;
4001 int negate;
4002
4003 if ('a' <= optch && optch <= 'z') {
4004 negate = 1;
4005 optch = toupper(optch);
4006 } else {
4007 negate = 0;
4008 }
4009
4010 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4011 if (!bit)
4012 return optarg[i];
4013 if (negate)
4014 opt->filter &= ~bit;
4015 else
4016 opt->filter |= bit;
4017 }
4018 return 0;
4019 }
4020
4021 static void enable_patch_output(int *fmt) {
4022 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4023 *fmt |= DIFF_FORMAT_PATCH;
4024 }
4025
4026 static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4027 {
4028 int val = parse_ws_error_highlight(arg);
4029
4030 if (val < 0) {
4031 error("unknown value after ws-error-highlight=%.*s",
4032 -1 - val, arg);
4033 return 0;
4034 }
4035 opt->ws_error_highlight = val;
4036 return 1;
4037 }
4038
4039 int diff_opt_parse(struct diff_options *options,
4040 const char **av, int ac, const char *prefix)
4041 {
4042 const char *arg = av[0];
4043 const char *optarg;
4044 int argcount;
4045
4046 if (!prefix)
4047 prefix = "";
4048
4049 /* Output format options */
4050 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4051 || opt_arg(arg, 'U', "unified", &options->context))
4052 enable_patch_output(&options->output_format);
4053 else if (!strcmp(arg, "--raw"))
4054 options->output_format |= DIFF_FORMAT_RAW;
4055 else if (!strcmp(arg, "--patch-with-raw")) {
4056 enable_patch_output(&options->output_format);
4057 options->output_format |= DIFF_FORMAT_RAW;
4058 } else if (!strcmp(arg, "--numstat"))
4059 options->output_format |= DIFF_FORMAT_NUMSTAT;
4060 else if (!strcmp(arg, "--shortstat"))
4061 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4062 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
4063 return parse_dirstat_opt(options, "");
4064 else if (skip_prefix(arg, "-X", &arg))
4065 return parse_dirstat_opt(options, arg);
4066 else if (skip_prefix(arg, "--dirstat=", &arg))
4067 return parse_dirstat_opt(options, arg);
4068 else if (!strcmp(arg, "--cumulative"))
4069 return parse_dirstat_opt(options, "cumulative");
4070 else if (!strcmp(arg, "--dirstat-by-file"))
4071 return parse_dirstat_opt(options, "files");
4072 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
4073 parse_dirstat_opt(options, "files");
4074 return parse_dirstat_opt(options, arg);
4075 }
4076 else if (!strcmp(arg, "--check"))
4077 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4078 else if (!strcmp(arg, "--summary"))
4079 options->output_format |= DIFF_FORMAT_SUMMARY;
4080 else if (!strcmp(arg, "--patch-with-stat")) {
4081 enable_patch_output(&options->output_format);
4082 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4083 } else if (!strcmp(arg, "--name-only"))
4084 options->output_format |= DIFF_FORMAT_NAME;
4085 else if (!strcmp(arg, "--name-status"))
4086 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4087 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4088 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4089 else if (starts_with(arg, "--stat"))
4090 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4091 return stat_opt(options, av);
4092
4093 /* renames options */
4094 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
4095 !strcmp(arg, "--break-rewrites")) {
4096 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4097 return error("invalid argument to -B: %s", arg+2);
4098 }
4099 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
4100 !strcmp(arg, "--find-renames")) {
4101 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4102 return error("invalid argument to -M: %s", arg+2);
4103 options->detect_rename = DIFF_DETECT_RENAME;
4104 }
4105 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4106 options->irreversible_delete = 1;
4107 }
4108 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
4109 !strcmp(arg, "--find-copies")) {
4110 if (options->detect_rename == DIFF_DETECT_COPY)
4111 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4112 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4113 return error("invalid argument to -C: %s", arg+2);
4114 options->detect_rename = DIFF_DETECT_COPY;
4115 }
4116 else if (!strcmp(arg, "--no-renames"))
4117 options->detect_rename = 0;
4118 else if (!strcmp(arg, "--rename-empty"))
4119 DIFF_OPT_SET(options, RENAME_EMPTY);
4120 else if (!strcmp(arg, "--no-rename-empty"))
4121 DIFF_OPT_CLR(options, RENAME_EMPTY);
4122 else if (!strcmp(arg, "--relative"))
4123 DIFF_OPT_SET(options, RELATIVE_NAME);
4124 else if (skip_prefix(arg, "--relative=", &arg)) {
4125 DIFF_OPT_SET(options, RELATIVE_NAME);
4126 options->prefix = arg;
4127 }
4128
4129 /* xdiff options */
4130 else if (!strcmp(arg, "--minimal"))
4131 DIFF_XDL_SET(options, NEED_MINIMAL);
4132 else if (!strcmp(arg, "--no-minimal"))
4133 DIFF_XDL_CLR(options, NEED_MINIMAL);
4134 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4135 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4136 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4137 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4138 else if (!strcmp(arg, "--ignore-space-at-eol"))
4139 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4140 else if (!strcmp(arg, "--ignore-blank-lines"))
4141 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4142 else if (!strcmp(arg, "--indent-heuristic"))
4143 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4144 else if (!strcmp(arg, "--no-indent-heuristic"))
4145 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4146 else if (!strcmp(arg, "--patience"))
4147 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4148 else if (!strcmp(arg, "--histogram"))
4149 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4150 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4151 long value = parse_algorithm_value(optarg);
4152 if (value < 0)
4153 return error("option diff-algorithm accepts \"myers\", "
4154 "\"minimal\", \"patience\" and \"histogram\"");
4155 /* clear out previous settings */
4156 DIFF_XDL_CLR(options, NEED_MINIMAL);
4157 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4158 options->xdl_opts |= value;
4159 return argcount;
4160 }
4161
4162 /* flags options */
4163 else if (!strcmp(arg, "--binary")) {
4164 enable_patch_output(&options->output_format);
4165 DIFF_OPT_SET(options, BINARY);
4166 }
4167 else if (!strcmp(arg, "--full-index"))
4168 DIFF_OPT_SET(options, FULL_INDEX);
4169 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4170 DIFF_OPT_SET(options, TEXT);
4171 else if (!strcmp(arg, "-R"))
4172 DIFF_OPT_SET(options, REVERSE_DIFF);
4173 else if (!strcmp(arg, "--find-copies-harder"))
4174 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4175 else if (!strcmp(arg, "--follow"))
4176 DIFF_OPT_SET(options, FOLLOW_RENAMES);
4177 else if (!strcmp(arg, "--no-follow")) {
4178 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
4179 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
4180 } else if (!strcmp(arg, "--color"))
4181 options->use_color = 1;
4182 else if (skip_prefix(arg, "--color=", &arg)) {
4183 int value = git_config_colorbool(NULL, arg);
4184 if (value < 0)
4185 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4186 options->use_color = value;
4187 }
4188 else if (!strcmp(arg, "--no-color"))
4189 options->use_color = 0;
4190 else if (!strcmp(arg, "--color-words")) {
4191 options->use_color = 1;
4192 options->word_diff = DIFF_WORDS_COLOR;
4193 }
4194 else if (skip_prefix(arg, "--color-words=", &arg)) {
4195 options->use_color = 1;
4196 options->word_diff = DIFF_WORDS_COLOR;
4197 options->word_regex = arg;
4198 }
4199 else if (!strcmp(arg, "--word-diff")) {
4200 if (options->word_diff == DIFF_WORDS_NONE)
4201 options->word_diff = DIFF_WORDS_PLAIN;
4202 }
4203 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4204 if (!strcmp(arg, "plain"))
4205 options->word_diff = DIFF_WORDS_PLAIN;
4206 else if (!strcmp(arg, "color")) {
4207 options->use_color = 1;
4208 options->word_diff = DIFF_WORDS_COLOR;
4209 }
4210 else if (!strcmp(arg, "porcelain"))
4211 options->word_diff = DIFF_WORDS_PORCELAIN;
4212 else if (!strcmp(arg, "none"))
4213 options->word_diff = DIFF_WORDS_NONE;
4214 else
4215 die("bad --word-diff argument: %s", arg);
4216 }
4217 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
4218 if (options->word_diff == DIFF_WORDS_NONE)
4219 options->word_diff = DIFF_WORDS_PLAIN;
4220 options->word_regex = optarg;
4221 return argcount;
4222 }
4223 else if (!strcmp(arg, "--exit-code"))
4224 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4225 else if (!strcmp(arg, "--quiet"))
4226 DIFF_OPT_SET(options, QUICK);
4227 else if (!strcmp(arg, "--ext-diff"))
4228 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
4229 else if (!strcmp(arg, "--no-ext-diff"))
4230 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
4231 else if (!strcmp(arg, "--textconv"))
4232 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
4233 else if (!strcmp(arg, "--no-textconv"))
4234 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
4235 else if (!strcmp(arg, "--ignore-submodules")) {
4236 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4237 handle_ignore_submodules_arg(options, "all");
4238 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
4239 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4240 handle_ignore_submodules_arg(options, arg);
4241 } else if (!strcmp(arg, "--submodule"))
4242 options->submodule_format = DIFF_SUBMODULE_LOG;
4243 else if (skip_prefix(arg, "--submodule=", &arg))
4244 return parse_submodule_opt(options, arg);
4245 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
4246 return parse_ws_error_highlight_opt(options, arg);
4247 else if (!strcmp(arg, "--ita-invisible-in-index"))
4248 options->ita_invisible_in_index = 1;
4249 else if (!strcmp(arg, "--ita-visible-in-index"))
4250 options->ita_invisible_in_index = 0;
4251
4252 /* misc options */
4253 else if (!strcmp(arg, "-z"))
4254 options->line_termination = 0;
4255 else if ((argcount = short_opt('l', av, &optarg))) {
4256 options->rename_limit = strtoul(optarg, NULL, 10);
4257 return argcount;
4258 }
4259 else if ((argcount = short_opt('S', av, &optarg))) {
4260 options->pickaxe = optarg;
4261 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4262 return argcount;
4263 } else if ((argcount = short_opt('G', av, &optarg))) {
4264 options->pickaxe = optarg;
4265 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4266 return argcount;
4267 }
4268 else if (!strcmp(arg, "--pickaxe-all"))
4269 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4270 else if (!strcmp(arg, "--pickaxe-regex"))
4271 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4272 else if ((argcount = short_opt('O', av, &optarg))) {
4273 options->orderfile = prefix_filename(prefix, optarg);
4274 return argcount;
4275 }
4276 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4277 int offending = parse_diff_filter_opt(optarg, options);
4278 if (offending)
4279 die("unknown change class '%c' in --diff-filter=%s",
4280 offending, optarg);
4281 return argcount;
4282 }
4283 else if (!strcmp(arg, "--no-abbrev"))
4284 options->abbrev = 0;
4285 else if (!strcmp(arg, "--abbrev"))
4286 options->abbrev = DEFAULT_ABBREV;
4287 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4288 options->abbrev = strtoul(arg, NULL, 10);
4289 if (options->abbrev < MINIMUM_ABBREV)
4290 options->abbrev = MINIMUM_ABBREV;
4291 else if (40 < options->abbrev)
4292 options->abbrev = 40;
4293 }
4294 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4295 options->a_prefix = optarg;
4296 return argcount;
4297 }
4298 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4299 options->line_prefix = optarg;
4300 options->line_prefix_length = strlen(options->line_prefix);
4301 graph_setup_line_prefix(options);
4302 return argcount;
4303 }
4304 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4305 options->b_prefix = optarg;
4306 return argcount;
4307 }
4308 else if (!strcmp(arg, "--no-prefix"))
4309 options->a_prefix = options->b_prefix = "";
4310 else if (opt_arg(arg, '\0', "inter-hunk-context",
4311 &options->interhunkcontext))
4312 ;
4313 else if (!strcmp(arg, "-W"))
4314 DIFF_OPT_SET(options, FUNCCONTEXT);
4315 else if (!strcmp(arg, "--function-context"))
4316 DIFF_OPT_SET(options, FUNCCONTEXT);
4317 else if (!strcmp(arg, "--no-function-context"))
4318 DIFF_OPT_CLR(options, FUNCCONTEXT);
4319 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4320 char *path = prefix_filename(prefix, optarg);
4321 options->file = xfopen(path, "w");
4322 options->close_file = 1;
4323 if (options->use_color != GIT_COLOR_ALWAYS)
4324 options->use_color = GIT_COLOR_NEVER;
4325 free(path);
4326 return argcount;
4327 } else
4328 return 0;
4329 return 1;
4330 }
4331
4332 int parse_rename_score(const char **cp_p)
4333 {
4334 unsigned long num, scale;
4335 int ch, dot;
4336 const char *cp = *cp_p;
4337
4338 num = 0;
4339 scale = 1;
4340 dot = 0;
4341 for (;;) {
4342 ch = *cp;
4343 if ( !dot && ch == '.' ) {
4344 scale = 1;
4345 dot = 1;
4346 } else if ( ch == '%' ) {
4347 scale = dot ? scale*100 : 100;
4348 cp++; /* % is always at the end */
4349 break;
4350 } else if ( ch >= '0' && ch <= '9' ) {
4351 if ( scale < 100000 ) {
4352 scale *= 10;
4353 num = (num*10) + (ch-'0');
4354 }
4355 } else {
4356 break;
4357 }
4358 cp++;
4359 }
4360 *cp_p = cp;
4361
4362 /* user says num divided by scale and we say internally that
4363 * is MAX_SCORE * num / scale.
4364 */
4365 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4366 }
4367
4368 static int diff_scoreopt_parse(const char *opt)
4369 {
4370 int opt1, opt2, cmd;
4371
4372 if (*opt++ != '-')
4373 return -1;
4374 cmd = *opt++;
4375 if (cmd == '-') {
4376 /* convert the long-form arguments into short-form versions */
4377 if (skip_prefix(opt, "break-rewrites", &opt)) {
4378 if (*opt == 0 || *opt++ == '=')
4379 cmd = 'B';
4380 } else if (skip_prefix(opt, "find-copies", &opt)) {
4381 if (*opt == 0 || *opt++ == '=')
4382 cmd = 'C';
4383 } else if (skip_prefix(opt, "find-renames", &opt)) {
4384 if (*opt == 0 || *opt++ == '=')
4385 cmd = 'M';
4386 }
4387 }
4388 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4389 return -1; /* that is not a -M, -C, or -B option */
4390
4391 opt1 = parse_rename_score(&opt);
4392 if (cmd != 'B')
4393 opt2 = 0;
4394 else {
4395 if (*opt == 0)
4396 opt2 = 0;
4397 else if (*opt != '/')
4398 return -1; /* we expect -B80/99 or -B80 */
4399 else {
4400 opt++;
4401 opt2 = parse_rename_score(&opt);
4402 }
4403 }
4404 if (*opt != 0)
4405 return -1;
4406 return opt1 | (opt2 << 16);
4407 }
4408
4409 struct diff_queue_struct diff_queued_diff;
4410
4411 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4412 {
4413 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4414 queue->queue[queue->nr++] = dp;
4415 }
4416
4417 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4418 struct diff_filespec *one,
4419 struct diff_filespec *two)
4420 {
4421 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4422 dp->one = one;
4423 dp->two = two;
4424 if (queue)
4425 diff_q(queue, dp);
4426 return dp;
4427 }
4428
4429 void diff_free_filepair(struct diff_filepair *p)
4430 {
4431 free_filespec(p->one);
4432 free_filespec(p->two);
4433 free(p);
4434 }
4435
4436 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4437 {
4438 int abblen;
4439 const char *abbrev;
4440
4441 if (len == GIT_SHA1_HEXSZ)
4442 return oid_to_hex(oid);
4443
4444 abbrev = diff_abbrev_oid(oid, len);
4445 abblen = strlen(abbrev);
4446
4447 /*
4448 * In well-behaved cases, where the abbbreviated result is the
4449 * same as the requested length, append three dots after the
4450 * abbreviation (hence the whole logic is limited to the case
4451 * where abblen < 37); when the actual abbreviated result is a
4452 * bit longer than the requested length, we reduce the number
4453 * of dots so that they match the well-behaved ones. However,
4454 * if the actual abbreviation is longer than the requested
4455 * length by more than three, we give up on aligning, and add
4456 * three dots anyway, to indicate that the output is not the
4457 * full object name. Yes, this may be suboptimal, but this
4458 * appears only in "diff --raw --abbrev" output and it is not
4459 * worth the effort to change it now. Note that this would
4460 * likely to work fine when the automatic sizing of default
4461 * abbreviation length is used--we would be fed -1 in "len" in
4462 * that case, and will end up always appending three-dots, but
4463 * the automatic sizing is supposed to give abblen that ensures
4464 * uniqueness across all objects (statistically speaking).
4465 */
4466 if (abblen < GIT_SHA1_HEXSZ - 3) {
4467 static char hex[GIT_MAX_HEXSZ + 1];
4468 if (len < abblen && abblen <= len + 2)
4469 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
4470 else
4471 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
4472 return hex;
4473 }
4474
4475 return oid_to_hex(oid);
4476 }
4477
4478 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4479 {
4480 int line_termination = opt->line_termination;
4481 int inter_name_termination = line_termination ? '\t' : '\0';
4482
4483 fprintf(opt->file, "%s", diff_line_prefix(opt));
4484 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4485 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4486 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
4487 fprintf(opt->file, "%s ",
4488 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
4489 }
4490 if (p->score) {
4491 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4492 inter_name_termination);
4493 } else {
4494 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4495 }
4496
4497 if (p->status == DIFF_STATUS_COPIED ||
4498 p->status == DIFF_STATUS_RENAMED) {
4499 const char *name_a, *name_b;
4500 name_a = p->one->path;
4501 name_b = p->two->path;
4502 strip_prefix(opt->prefix_length, &name_a, &name_b);
4503 write_name_quoted(name_a, opt->file, inter_name_termination);
4504 write_name_quoted(name_b, opt->file, line_termination);
4505 } else {
4506 const char *name_a, *name_b;
4507 name_a = p->one->mode ? p->one->path : p->two->path;
4508 name_b = NULL;
4509 strip_prefix(opt->prefix_length, &name_a, &name_b);
4510 write_name_quoted(name_a, opt->file, line_termination);
4511 }
4512 }
4513
4514 int diff_unmodified_pair(struct diff_filepair *p)
4515 {
4516 /* This function is written stricter than necessary to support
4517 * the currently implemented transformers, but the idea is to
4518 * let transformers to produce diff_filepairs any way they want,
4519 * and filter and clean them up here before producing the output.
4520 */
4521 struct diff_filespec *one = p->one, *two = p->two;
4522
4523 if (DIFF_PAIR_UNMERGED(p))
4524 return 0; /* unmerged is interesting */
4525
4526 /* deletion, addition, mode or type change
4527 * and rename are all interesting.
4528 */
4529 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4530 DIFF_PAIR_MODE_CHANGED(p) ||
4531 strcmp(one->path, two->path))
4532 return 0;
4533
4534 /* both are valid and point at the same path. that is, we are
4535 * dealing with a change.
4536 */
4537 if (one->oid_valid && two->oid_valid &&
4538 !oidcmp(&one->oid, &two->oid) &&
4539 !one->dirty_submodule && !two->dirty_submodule)
4540 return 1; /* no change */
4541 if (!one->oid_valid && !two->oid_valid)
4542 return 1; /* both look at the same file on the filesystem. */
4543 return 0;
4544 }
4545
4546 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4547 {
4548 if (diff_unmodified_pair(p))
4549 return;
4550
4551 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4552 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4553 return; /* no tree diffs in patch format */
4554
4555 run_diff(p, o);
4556 }
4557
4558 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4559 struct diffstat_t *diffstat)
4560 {
4561 if (diff_unmodified_pair(p))
4562 return;
4563
4564 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4565 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4566 return; /* no useful stat for tree diffs */
4567
4568 run_diffstat(p, o, diffstat);
4569 }
4570
4571 static void diff_flush_checkdiff(struct diff_filepair *p,
4572 struct diff_options *o)
4573 {
4574 if (diff_unmodified_pair(p))
4575 return;
4576
4577 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4578 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4579 return; /* nothing to check in tree diffs */
4580
4581 run_checkdiff(p, o);
4582 }
4583
4584 int diff_queue_is_empty(void)
4585 {
4586 struct diff_queue_struct *q = &diff_queued_diff;
4587 int i;
4588 for (i = 0; i < q->nr; i++)
4589 if (!diff_unmodified_pair(q->queue[i]))
4590 return 0;
4591 return 1;
4592 }
4593
4594 #if DIFF_DEBUG
4595 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4596 {
4597 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4598 x, one ? one : "",
4599 s->path,
4600 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4601 s->mode,
4602 s->oid_valid ? oid_to_hex(&s->oid) : "");
4603 fprintf(stderr, "queue[%d] %s size %lu\n",
4604 x, one ? one : "",
4605 s->size);
4606 }
4607
4608 void diff_debug_filepair(const struct diff_filepair *p, int i)
4609 {
4610 diff_debug_filespec(p->one, i, "one");
4611 diff_debug_filespec(p->two, i, "two");
4612 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4613 p->score, p->status ? p->status : '?',
4614 p->one->rename_used, p->broken_pair);
4615 }
4616
4617 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4618 {
4619 int i;
4620 if (msg)
4621 fprintf(stderr, "%s\n", msg);
4622 fprintf(stderr, "q->nr = %d\n", q->nr);
4623 for (i = 0; i < q->nr; i++) {
4624 struct diff_filepair *p = q->queue[i];
4625 diff_debug_filepair(p, i);
4626 }
4627 }
4628 #endif
4629
4630 static void diff_resolve_rename_copy(void)
4631 {
4632 int i;
4633 struct diff_filepair *p;
4634 struct diff_queue_struct *q = &diff_queued_diff;
4635
4636 diff_debug_queue("resolve-rename-copy", q);
4637
4638 for (i = 0; i < q->nr; i++) {
4639 p = q->queue[i];
4640 p->status = 0; /* undecided */
4641 if (DIFF_PAIR_UNMERGED(p))
4642 p->status = DIFF_STATUS_UNMERGED;
4643 else if (!DIFF_FILE_VALID(p->one))
4644 p->status = DIFF_STATUS_ADDED;
4645 else if (!DIFF_FILE_VALID(p->two))
4646 p->status = DIFF_STATUS_DELETED;
4647 else if (DIFF_PAIR_TYPE_CHANGED(p))
4648 p->status = DIFF_STATUS_TYPE_CHANGED;
4649
4650 /* from this point on, we are dealing with a pair
4651 * whose both sides are valid and of the same type, i.e.
4652 * either in-place edit or rename/copy edit.
4653 */
4654 else if (DIFF_PAIR_RENAME(p)) {
4655 /*
4656 * A rename might have re-connected a broken
4657 * pair up, causing the pathnames to be the
4658 * same again. If so, that's not a rename at
4659 * all, just a modification..
4660 *
4661 * Otherwise, see if this source was used for
4662 * multiple renames, in which case we decrement
4663 * the count, and call it a copy.
4664 */
4665 if (!strcmp(p->one->path, p->two->path))
4666 p->status = DIFF_STATUS_MODIFIED;
4667 else if (--p->one->rename_used > 0)
4668 p->status = DIFF_STATUS_COPIED;
4669 else
4670 p->status = DIFF_STATUS_RENAMED;
4671 }
4672 else if (oidcmp(&p->one->oid, &p->two->oid) ||
4673 p->one->mode != p->two->mode ||
4674 p->one->dirty_submodule ||
4675 p->two->dirty_submodule ||
4676 is_null_oid(&p->one->oid))
4677 p->status = DIFF_STATUS_MODIFIED;
4678 else {
4679 /* This is a "no-change" entry and should not
4680 * happen anymore, but prepare for broken callers.
4681 */
4682 error("feeding unmodified %s to diffcore",
4683 p->one->path);
4684 p->status = DIFF_STATUS_UNKNOWN;
4685 }
4686 }
4687 diff_debug_queue("resolve-rename-copy done", q);
4688 }
4689
4690 static int check_pair_status(struct diff_filepair *p)
4691 {
4692 switch (p->status) {
4693 case DIFF_STATUS_UNKNOWN:
4694 return 0;
4695 case 0:
4696 die("internal error in diff-resolve-rename-copy");
4697 default:
4698 return 1;
4699 }
4700 }
4701
4702 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4703 {
4704 int fmt = opt->output_format;
4705
4706 if (fmt & DIFF_FORMAT_CHECKDIFF)
4707 diff_flush_checkdiff(p, opt);
4708 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4709 diff_flush_raw(p, opt);
4710 else if (fmt & DIFF_FORMAT_NAME) {
4711 const char *name_a, *name_b;
4712 name_a = p->two->path;
4713 name_b = NULL;
4714 strip_prefix(opt->prefix_length, &name_a, &name_b);
4715 fprintf(opt->file, "%s", diff_line_prefix(opt));
4716 write_name_quoted(name_a, opt->file, opt->line_termination);
4717 }
4718 }
4719
4720 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4721 {
4722 if (fs->mode)
4723 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4724 else
4725 fprintf(file, " %s ", newdelete);
4726 write_name_quoted(fs->path, file, '\n');
4727 }
4728
4729
4730 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4731 const char *line_prefix)
4732 {
4733 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4734 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4735 p->two->mode, show_name ? ' ' : '\n');
4736 if (show_name) {
4737 write_name_quoted(p->two->path, file, '\n');
4738 }
4739 }
4740 }
4741
4742 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4743 const char *line_prefix)
4744 {
4745 char *names = pprint_rename(p->one->path, p->two->path);
4746
4747 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4748 free(names);
4749 show_mode_change(file, p, 0, line_prefix);
4750 }
4751
4752 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4753 {
4754 FILE *file = opt->file;
4755 const char *line_prefix = diff_line_prefix(opt);
4756
4757 switch(p->status) {
4758 case DIFF_STATUS_DELETED:
4759 fputs(line_prefix, file);
4760 show_file_mode_name(file, "delete", p->one);
4761 break;
4762 case DIFF_STATUS_ADDED:
4763 fputs(line_prefix, file);
4764 show_file_mode_name(file, "create", p->two);
4765 break;
4766 case DIFF_STATUS_COPIED:
4767 fputs(line_prefix, file);
4768 show_rename_copy(file, "copy", p, line_prefix);
4769 break;
4770 case DIFF_STATUS_RENAMED:
4771 fputs(line_prefix, file);
4772 show_rename_copy(file, "rename", p, line_prefix);
4773 break;
4774 default:
4775 if (p->score) {
4776 fprintf(file, "%s rewrite ", line_prefix);
4777 write_name_quoted(p->two->path, file, ' ');
4778 fprintf(file, "(%d%%)\n", similarity_index(p));
4779 }
4780 show_mode_change(file, p, !p->score, line_prefix);
4781 break;
4782 }
4783 }
4784
4785 struct patch_id_t {
4786 git_SHA_CTX *ctx;
4787 int patchlen;
4788 };
4789
4790 static int remove_space(char *line, int len)
4791 {
4792 int i;
4793 char *dst = line;
4794 unsigned char c;
4795
4796 for (i = 0; i < len; i++)
4797 if (!isspace((c = line[i])))
4798 *dst++ = c;
4799
4800 return dst - line;
4801 }
4802
4803 static void patch_id_consume(void *priv, char *line, unsigned long len)
4804 {
4805 struct patch_id_t *data = priv;
4806 int new_len;
4807
4808 /* Ignore line numbers when computing the SHA1 of the patch */
4809 if (starts_with(line, "@@ -"))
4810 return;
4811
4812 new_len = remove_space(line, len);
4813
4814 git_SHA1_Update(data->ctx, line, new_len);
4815 data->patchlen += new_len;
4816 }
4817
4818 static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
4819 {
4820 git_SHA1_Update(ctx, str, strlen(str));
4821 }
4822
4823 static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
4824 {
4825 /* large enough for 2^32 in octal */
4826 char buf[12];
4827 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
4828 git_SHA1_Update(ctx, buf, len);
4829 }
4830
4831 /* returns 0 upon success, and writes result into sha1 */
4832 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
4833 {
4834 struct diff_queue_struct *q = &diff_queued_diff;
4835 int i;
4836 git_SHA_CTX ctx;
4837 struct patch_id_t data;
4838
4839 git_SHA1_Init(&ctx);
4840 memset(&data, 0, sizeof(struct patch_id_t));
4841 data.ctx = &ctx;
4842
4843 for (i = 0; i < q->nr; i++) {
4844 xpparam_t xpp;
4845 xdemitconf_t xecfg;
4846 mmfile_t mf1, mf2;
4847 struct diff_filepair *p = q->queue[i];
4848 int len1, len2;
4849
4850 memset(&xpp, 0, sizeof(xpp));
4851 memset(&xecfg, 0, sizeof(xecfg));
4852 if (p->status == 0)
4853 return error("internal diff status error");
4854 if (p->status == DIFF_STATUS_UNKNOWN)
4855 continue;
4856 if (diff_unmodified_pair(p))
4857 continue;
4858 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4859 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4860 continue;
4861 if (DIFF_PAIR_UNMERGED(p))
4862 continue;
4863
4864 diff_fill_oid_info(p->one);
4865 diff_fill_oid_info(p->two);
4866
4867 len1 = remove_space(p->one->path, strlen(p->one->path));
4868 len2 = remove_space(p->two->path, strlen(p->two->path));
4869 patch_id_add_string(&ctx, "diff--git");
4870 patch_id_add_string(&ctx, "a/");
4871 git_SHA1_Update(&ctx, p->one->path, len1);
4872 patch_id_add_string(&ctx, "b/");
4873 git_SHA1_Update(&ctx, p->two->path, len2);
4874
4875 if (p->one->mode == 0) {
4876 patch_id_add_string(&ctx, "newfilemode");
4877 patch_id_add_mode(&ctx, p->two->mode);
4878 patch_id_add_string(&ctx, "---/dev/null");
4879 patch_id_add_string(&ctx, "+++b/");
4880 git_SHA1_Update(&ctx, p->two->path, len2);
4881 } else if (p->two->mode == 0) {
4882 patch_id_add_string(&ctx, "deletedfilemode");
4883 patch_id_add_mode(&ctx, p->one->mode);
4884 patch_id_add_string(&ctx, "---a/");
4885 git_SHA1_Update(&ctx, p->one->path, len1);
4886 patch_id_add_string(&ctx, "+++/dev/null");
4887 } else {
4888 patch_id_add_string(&ctx, "---a/");
4889 git_SHA1_Update(&ctx, p->one->path, len1);
4890 patch_id_add_string(&ctx, "+++b/");
4891 git_SHA1_Update(&ctx, p->two->path, len2);
4892 }
4893
4894 if (diff_header_only)
4895 continue;
4896
4897 if (fill_mmfile(&mf1, p->one) < 0 ||
4898 fill_mmfile(&mf2, p->two) < 0)
4899 return error("unable to read files to diff");
4900
4901 if (diff_filespec_is_binary(p->one) ||
4902 diff_filespec_is_binary(p->two)) {
4903 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
4904 GIT_SHA1_HEXSZ);
4905 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
4906 GIT_SHA1_HEXSZ);
4907 continue;
4908 }
4909
4910 xpp.flags = 0;
4911 xecfg.ctxlen = 3;
4912 xecfg.flags = 0;
4913 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4914 &xpp, &xecfg))
4915 return error("unable to generate patch-id diff for %s",
4916 p->one->path);
4917 }
4918
4919 git_SHA1_Final(oid->hash, &ctx);
4920 return 0;
4921 }
4922
4923 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
4924 {
4925 struct diff_queue_struct *q = &diff_queued_diff;
4926 int i;
4927 int result = diff_get_patch_id(options, oid, diff_header_only);
4928
4929 for (i = 0; i < q->nr; i++)
4930 diff_free_filepair(q->queue[i]);
4931
4932 free(q->queue);
4933 DIFF_QUEUE_CLEAR(q);
4934
4935 return result;
4936 }
4937
4938 static int is_summary_empty(const struct diff_queue_struct *q)
4939 {
4940 int i;
4941
4942 for (i = 0; i < q->nr; i++) {
4943 const struct diff_filepair *p = q->queue[i];
4944
4945 switch (p->status) {
4946 case DIFF_STATUS_DELETED:
4947 case DIFF_STATUS_ADDED:
4948 case DIFF_STATUS_COPIED:
4949 case DIFF_STATUS_RENAMED:
4950 return 0;
4951 default:
4952 if (p->score)
4953 return 0;
4954 if (p->one->mode && p->two->mode &&
4955 p->one->mode != p->two->mode)
4956 return 0;
4957 break;
4958 }
4959 }
4960 return 1;
4961 }
4962
4963 static const char rename_limit_warning[] =
4964 N_("inexact rename detection was skipped due to too many files.");
4965
4966 static const char degrade_cc_to_c_warning[] =
4967 N_("only found copies from modified paths due to too many files.");
4968
4969 static const char rename_limit_advice[] =
4970 N_("you may want to set your %s variable to at least "
4971 "%d and retry the command.");
4972
4973 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4974 {
4975 if (degraded_cc)
4976 warning(_(degrade_cc_to_c_warning));
4977 else if (needed)
4978 warning(_(rename_limit_warning));
4979 else
4980 return;
4981 if (0 < needed && needed < 32767)
4982 warning(_(rename_limit_advice), varname, needed);
4983 }
4984
4985 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
4986 {
4987 int i;
4988 struct diff_queue_struct *q = &diff_queued_diff;
4989
4990 if (WSEH_NEW & WS_RULE_MASK)
4991 die("BUG: WS rules bit mask overlaps with diff symbol flags");
4992
4993 for (i = 0; i < q->nr; i++) {
4994 struct diff_filepair *p = q->queue[i];
4995 if (check_pair_status(p))
4996 diff_flush_patch(p, o);
4997 }
4998 }
4999
5000 void diff_flush(struct diff_options *options)
5001 {
5002 struct diff_queue_struct *q = &diff_queued_diff;
5003 int i, output_format = options->output_format;
5004 int separator = 0;
5005 int dirstat_by_line = 0;
5006
5007 /*
5008 * Order: raw, stat, summary, patch
5009 * or: name/name-status/checkdiff (other bits clear)
5010 */
5011 if (!q->nr)
5012 goto free_queue;
5013
5014 if (output_format & (DIFF_FORMAT_RAW |
5015 DIFF_FORMAT_NAME |
5016 DIFF_FORMAT_NAME_STATUS |
5017 DIFF_FORMAT_CHECKDIFF)) {
5018 for (i = 0; i < q->nr; i++) {
5019 struct diff_filepair *p = q->queue[i];
5020 if (check_pair_status(p))
5021 flush_one_pair(p, options);
5022 }
5023 separator++;
5024 }
5025
5026 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
5027 dirstat_by_line = 1;
5028
5029 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5030 dirstat_by_line) {
5031 struct diffstat_t diffstat;
5032
5033 memset(&diffstat, 0, sizeof(struct diffstat_t));
5034 for (i = 0; i < q->nr; i++) {
5035 struct diff_filepair *p = q->queue[i];
5036 if (check_pair_status(p))
5037 diff_flush_stat(p, options, &diffstat);
5038 }
5039 if (output_format & DIFF_FORMAT_NUMSTAT)
5040 show_numstat(&diffstat, options);
5041 if (output_format & DIFF_FORMAT_DIFFSTAT)
5042 show_stats(&diffstat, options);
5043 if (output_format & DIFF_FORMAT_SHORTSTAT)
5044 show_shortstats(&diffstat, options);
5045 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5046 show_dirstat_by_line(&diffstat, options);
5047 free_diffstat_info(&diffstat);
5048 separator++;
5049 }
5050 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5051 show_dirstat(options);
5052
5053 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5054 for (i = 0; i < q->nr; i++) {
5055 diff_summary(options, q->queue[i]);
5056 }
5057 separator++;
5058 }
5059
5060 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5061 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
5062 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5063 /*
5064 * run diff_flush_patch for the exit status. setting
5065 * options->file to /dev/null should be safe, because we
5066 * aren't supposed to produce any output anyway.
5067 */
5068 if (options->close_file)
5069 fclose(options->file);
5070 options->file = xfopen("/dev/null", "w");
5071 options->close_file = 1;
5072 for (i = 0; i < q->nr; i++) {
5073 struct diff_filepair *p = q->queue[i];
5074 if (check_pair_status(p))
5075 diff_flush_patch(p, options);
5076 if (options->found_changes)
5077 break;
5078 }
5079 }
5080
5081 if (output_format & DIFF_FORMAT_PATCH) {
5082 if (separator) {
5083 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5084 if (options->stat_sep)
5085 /* attach patch instead of inline */
5086 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5087 NULL, 0, 0);
5088 }
5089
5090 diff_flush_patch_all_file_pairs(options);
5091 }
5092
5093 if (output_format & DIFF_FORMAT_CALLBACK)
5094 options->format_callback(q, options, options->format_callback_data);
5095
5096 for (i = 0; i < q->nr; i++)
5097 diff_free_filepair(q->queue[i]);
5098 free_queue:
5099 free(q->queue);
5100 DIFF_QUEUE_CLEAR(q);
5101 if (options->close_file)
5102 fclose(options->file);
5103
5104 /*
5105 * Report the content-level differences with HAS_CHANGES;
5106 * diff_addremove/diff_change does not set the bit when
5107 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5108 */
5109 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5110 if (options->found_changes)
5111 DIFF_OPT_SET(options, HAS_CHANGES);
5112 else
5113 DIFF_OPT_CLR(options, HAS_CHANGES);
5114 }
5115 }
5116
5117 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5118 {
5119 return (((p->status == DIFF_STATUS_MODIFIED) &&
5120 ((p->score &&
5121 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5122 (!p->score &&
5123 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5124 ((p->status != DIFF_STATUS_MODIFIED) &&
5125 filter_bit_tst(p->status, options)));
5126 }
5127
5128 static void diffcore_apply_filter(struct diff_options *options)
5129 {
5130 int i;
5131 struct diff_queue_struct *q = &diff_queued_diff;
5132 struct diff_queue_struct outq;
5133
5134 DIFF_QUEUE_CLEAR(&outq);
5135
5136 if (!options->filter)
5137 return;
5138
5139 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5140 int found;
5141 for (i = found = 0; !found && i < q->nr; i++) {
5142 if (match_filter(options, q->queue[i]))
5143 found++;
5144 }
5145 if (found)
5146 return;
5147
5148 /* otherwise we will clear the whole queue
5149 * by copying the empty outq at the end of this
5150 * function, but first clear the current entries
5151 * in the queue.
5152 */
5153 for (i = 0; i < q->nr; i++)
5154 diff_free_filepair(q->queue[i]);
5155 }
5156 else {
5157 /* Only the matching ones */
5158 for (i = 0; i < q->nr; i++) {
5159 struct diff_filepair *p = q->queue[i];
5160 if (match_filter(options, p))
5161 diff_q(&outq, p);
5162 else
5163 diff_free_filepair(p);
5164 }
5165 }
5166 free(q->queue);
5167 *q = outq;
5168 }
5169
5170 /* Check whether two filespecs with the same mode and size are identical */
5171 static int diff_filespec_is_identical(struct diff_filespec *one,
5172 struct diff_filespec *two)
5173 {
5174 if (S_ISGITLINK(one->mode))
5175 return 0;
5176 if (diff_populate_filespec(one, 0))
5177 return 0;
5178 if (diff_populate_filespec(two, 0))
5179 return 0;
5180 return !memcmp(one->data, two->data, one->size);
5181 }
5182
5183 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
5184 {
5185 if (p->done_skip_stat_unmatch)
5186 return p->skip_stat_unmatch_result;
5187
5188 p->done_skip_stat_unmatch = 1;
5189 p->skip_stat_unmatch_result = 0;
5190 /*
5191 * 1. Entries that come from stat info dirtiness
5192 * always have both sides (iow, not create/delete),
5193 * one side of the object name is unknown, with
5194 * the same mode and size. Keep the ones that
5195 * do not match these criteria. They have real
5196 * differences.
5197 *
5198 * 2. At this point, the file is known to be modified,
5199 * with the same mode and size, and the object
5200 * name of one side is unknown. Need to inspect
5201 * the identical contents.
5202 */
5203 if (!DIFF_FILE_VALID(p->one) || /* (1) */
5204 !DIFF_FILE_VALID(p->two) ||
5205 (p->one->oid_valid && p->two->oid_valid) ||
5206 (p->one->mode != p->two->mode) ||
5207 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
5208 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
5209 (p->one->size != p->two->size) ||
5210 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
5211 p->skip_stat_unmatch_result = 1;
5212 return p->skip_stat_unmatch_result;
5213 }
5214
5215 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
5216 {
5217 int i;
5218 struct diff_queue_struct *q = &diff_queued_diff;
5219 struct diff_queue_struct outq;
5220 DIFF_QUEUE_CLEAR(&outq);
5221
5222 for (i = 0; i < q->nr; i++) {
5223 struct diff_filepair *p = q->queue[i];
5224
5225 if (diff_filespec_check_stat_unmatch(p))
5226 diff_q(&outq, p);
5227 else {
5228 /*
5229 * The caller can subtract 1 from skip_stat_unmatch
5230 * to determine how many paths were dirty only
5231 * due to stat info mismatch.
5232 */
5233 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
5234 diffopt->skip_stat_unmatch++;
5235 diff_free_filepair(p);
5236 }
5237 }
5238 free(q->queue);
5239 *q = outq;
5240 }
5241
5242 static int diffnamecmp(const void *a_, const void *b_)
5243 {
5244 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
5245 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
5246 const char *name_a, *name_b;
5247
5248 name_a = a->one ? a->one->path : a->two->path;
5249 name_b = b->one ? b->one->path : b->two->path;
5250 return strcmp(name_a, name_b);
5251 }
5252
5253 void diffcore_fix_diff_index(struct diff_options *options)
5254 {
5255 struct diff_queue_struct *q = &diff_queued_diff;
5256 QSORT(q->queue, q->nr, diffnamecmp);
5257 }
5258
5259 void diffcore_std(struct diff_options *options)
5260 {
5261 /* NOTE please keep the following in sync with diff_tree_combined() */
5262 if (options->skip_stat_unmatch)
5263 diffcore_skip_stat_unmatch(options);
5264 if (!options->found_follow) {
5265 /* See try_to_follow_renames() in tree-diff.c */
5266 if (options->break_opt != -1)
5267 diffcore_break(options->break_opt);
5268 if (options->detect_rename)
5269 diffcore_rename(options);
5270 if (options->break_opt != -1)
5271 diffcore_merge_broken();
5272 }
5273 if (options->pickaxe)
5274 diffcore_pickaxe(options);
5275 if (options->orderfile)
5276 diffcore_order(options->orderfile);
5277 if (!options->found_follow)
5278 /* See try_to_follow_renames() in tree-diff.c */
5279 diff_resolve_rename_copy();
5280 diffcore_apply_filter(options);
5281
5282 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5283 DIFF_OPT_SET(options, HAS_CHANGES);
5284 else
5285 DIFF_OPT_CLR(options, HAS_CHANGES);
5286
5287 options->found_follow = 0;
5288 }
5289
5290 int diff_result_code(struct diff_options *opt, int status)
5291 {
5292 int result = 0;
5293
5294 diff_warn_rename_limit("diff.renameLimit",
5295 opt->needed_rename_limit,
5296 opt->degraded_cc_to_c);
5297 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5298 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5299 return status;
5300 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5301 DIFF_OPT_TST(opt, HAS_CHANGES))
5302 result |= 01;
5303 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5304 DIFF_OPT_TST(opt, CHECK_FAILED))
5305 result |= 02;
5306 return result;
5307 }
5308
5309 int diff_can_quit_early(struct diff_options *opt)
5310 {
5311 return (DIFF_OPT_TST(opt, QUICK) &&
5312 !opt->filter &&
5313 DIFF_OPT_TST(opt, HAS_CHANGES));
5314 }
5315
5316 /*
5317 * Shall changes to this submodule be ignored?
5318 *
5319 * Submodule changes can be configured to be ignored separately for each path,
5320 * but that configuration can be overridden from the command line.
5321 */
5322 static int is_submodule_ignored(const char *path, struct diff_options *options)
5323 {
5324 int ignored = 0;
5325 unsigned orig_flags = options->flags;
5326 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5327 set_diffopt_flags_from_submodule_config(options, path);
5328 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5329 ignored = 1;
5330 options->flags = orig_flags;
5331 return ignored;
5332 }
5333
5334 void diff_addremove(struct diff_options *options,
5335 int addremove, unsigned mode,
5336 const struct object_id *oid,
5337 int oid_valid,
5338 const char *concatpath, unsigned dirty_submodule)
5339 {
5340 struct diff_filespec *one, *two;
5341
5342 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5343 return;
5344
5345 /* This may look odd, but it is a preparation for
5346 * feeding "there are unchanged files which should
5347 * not produce diffs, but when you are doing copy
5348 * detection you would need them, so here they are"
5349 * entries to the diff-core. They will be prefixed
5350 * with something like '=' or '*' (I haven't decided
5351 * which but should not make any difference).
5352 * Feeding the same new and old to diff_change()
5353 * also has the same effect.
5354 * Before the final output happens, they are pruned after
5355 * merged into rename/copy pairs as appropriate.
5356 */
5357 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5358 addremove = (addremove == '+' ? '-' :
5359 addremove == '-' ? '+' : addremove);
5360
5361 if (options->prefix &&
5362 strncmp(concatpath, options->prefix, options->prefix_length))
5363 return;
5364
5365 one = alloc_filespec(concatpath);
5366 two = alloc_filespec(concatpath);
5367
5368 if (addremove != '+')
5369 fill_filespec(one, oid, oid_valid, mode);
5370 if (addremove != '-') {
5371 fill_filespec(two, oid, oid_valid, mode);
5372 two->dirty_submodule = dirty_submodule;
5373 }
5374
5375 diff_queue(&diff_queued_diff, one, two);
5376 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5377 DIFF_OPT_SET(options, HAS_CHANGES);
5378 }
5379
5380 void diff_change(struct diff_options *options,
5381 unsigned old_mode, unsigned new_mode,
5382 const struct object_id *old_oid,
5383 const struct object_id *new_oid,
5384 int old_oid_valid, int new_oid_valid,
5385 const char *concatpath,
5386 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5387 {
5388 struct diff_filespec *one, *two;
5389 struct diff_filepair *p;
5390
5391 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5392 is_submodule_ignored(concatpath, options))
5393 return;
5394
5395 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5396 SWAP(old_mode, new_mode);
5397 SWAP(old_oid, new_oid);
5398 SWAP(old_oid_valid, new_oid_valid);
5399 SWAP(old_dirty_submodule, new_dirty_submodule);
5400 }
5401
5402 if (options->prefix &&
5403 strncmp(concatpath, options->prefix, options->prefix_length))
5404 return;
5405
5406 one = alloc_filespec(concatpath);
5407 two = alloc_filespec(concatpath);
5408 fill_filespec(one, old_oid, old_oid_valid, old_mode);
5409 fill_filespec(two, new_oid, new_oid_valid, new_mode);
5410 one->dirty_submodule = old_dirty_submodule;
5411 two->dirty_submodule = new_dirty_submodule;
5412 p = diff_queue(&diff_queued_diff, one, two);
5413
5414 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5415 return;
5416
5417 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5418 !diff_filespec_check_stat_unmatch(p))
5419 return;
5420
5421 DIFF_OPT_SET(options, HAS_CHANGES);
5422 }
5423
5424 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5425 {
5426 struct diff_filepair *pair;
5427 struct diff_filespec *one, *two;
5428
5429 if (options->prefix &&
5430 strncmp(path, options->prefix, options->prefix_length))
5431 return NULL;
5432
5433 one = alloc_filespec(path);
5434 two = alloc_filespec(path);
5435 pair = diff_queue(&diff_queued_diff, one, two);
5436 pair->is_unmerged = 1;
5437 return pair;
5438 }
5439
5440 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5441 size_t *outsize)
5442 {
5443 struct diff_tempfile *temp;
5444 const char *argv[3];
5445 const char **arg = argv;
5446 struct child_process child = CHILD_PROCESS_INIT;
5447 struct strbuf buf = STRBUF_INIT;
5448 int err = 0;
5449
5450 temp = prepare_temp_file(spec->path, spec);
5451 *arg++ = pgm;
5452 *arg++ = temp->name;
5453 *arg = NULL;
5454
5455 child.use_shell = 1;
5456 child.argv = argv;
5457 child.out = -1;
5458 if (start_command(&child)) {
5459 remove_tempfile();
5460 return NULL;
5461 }
5462
5463 if (strbuf_read(&buf, child.out, 0) < 0)
5464 err = error("error reading from textconv command '%s'", pgm);
5465 close(child.out);
5466
5467 if (finish_command(&child) || err) {
5468 strbuf_release(&buf);
5469 remove_tempfile();
5470 return NULL;
5471 }
5472 remove_tempfile();
5473
5474 return strbuf_detach(&buf, outsize);
5475 }
5476
5477 size_t fill_textconv(struct userdiff_driver *driver,
5478 struct diff_filespec *df,
5479 char **outbuf)
5480 {
5481 size_t size;
5482
5483 if (!driver) {
5484 if (!DIFF_FILE_VALID(df)) {
5485 *outbuf = "";
5486 return 0;
5487 }
5488 if (diff_populate_filespec(df, 0))
5489 die("unable to read files to diff");
5490 *outbuf = df->data;
5491 return df->size;
5492 }
5493
5494 if (!driver->textconv)
5495 die("BUG: fill_textconv called with non-textconv driver");
5496
5497 if (driver->textconv_cache && df->oid_valid) {
5498 *outbuf = notes_cache_get(driver->textconv_cache,
5499 &df->oid,
5500 &size);
5501 if (*outbuf)
5502 return size;
5503 }
5504
5505 *outbuf = run_textconv(driver->textconv, df, &size);
5506 if (!*outbuf)
5507 die("unable to read files to diff");
5508
5509 if (driver->textconv_cache && df->oid_valid) {
5510 /* ignore errors, as we might be in a readonly repository */
5511 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
5512 size);
5513 /*
5514 * we could save up changes and flush them all at the end,
5515 * but we would need an extra call after all diffing is done.
5516 * Since generating a cache entry is the slow path anyway,
5517 * this extra overhead probably isn't a big deal.
5518 */
5519 notes_cache_write(driver->textconv_cache);
5520 }
5521
5522 return size;
5523 }
5524
5525 int textconv_object(const char *path,
5526 unsigned mode,
5527 const struct object_id *oid,
5528 int oid_valid,
5529 char **buf,
5530 unsigned long *buf_size)
5531 {
5532 struct diff_filespec *df;
5533 struct userdiff_driver *textconv;
5534
5535 df = alloc_filespec(path);
5536 fill_filespec(df, oid, oid_valid, mode);
5537 textconv = get_textconv(df);
5538 if (!textconv) {
5539 free_filespec(df);
5540 return 0;
5541 }
5542
5543 *buf_size = fill_textconv(textconv, df, buf);
5544 free_filespec(df);
5545 return 1;
5546 }
5547
5548 void setup_diff_pager(struct diff_options *opt)
5549 {
5550 /*
5551 * If the user asked for our exit code, then either they want --quiet
5552 * or --exit-code. We should definitely not bother with a pager in the
5553 * former case, as we will generate no output. Since we still properly
5554 * report our exit code even when a pager is run, we _could_ run a
5555 * pager with --exit-code. But since we have not done so historically,
5556 * and because it is easy to find people oneline advising "git diff
5557 * --exit-code" in hooks and other scripts, we do not do so.
5558 */
5559 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5560 check_pager_config("diff") != 0)
5561 setup_pager();
5562 }
git for Debian