?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Crawling towards -rc2
[git/debian.git] / diff.c
blobd76bb937c1e49bc5cd3ecaf9f498a90543def822
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 "hashmap.h"
20 #include "ll-merge.h"
21 #include "string-list.h"
22 #include "argv-array.h"
23 #include "graph.h"
24 #include "packfile.h"
25
26 #ifdef NO_FAST_WORKING_DIRECTORY
27 #define FAST_WORKING_DIRECTORY 0
28 #else
29 #define FAST_WORKING_DIRECTORY 1
30 #endif
31
32 static int diff_detect_rename_default;
33 static int diff_indent_heuristic = 1;
34 static int diff_rename_limit_default = 400;
35 static int diff_suppress_blank_empty;
36 static int diff_use_color_default = -1;
37 static int diff_color_moved_default;
38 static int diff_context_default = 3;
39 static int diff_interhunk_context_default;
40 static const char *diff_word_regex_cfg;
41 static const char *external_diff_cmd_cfg;
42 static const char *diff_order_file_cfg;
43 int diff_auto_refresh_index = 1;
44 static int diff_mnemonic_prefix;
45 static int diff_no_prefix;
46 static int diff_stat_graph_width;
47 static int diff_dirstat_permille_default = 30;
48 static struct diff_options default_diff_options;
49 static long diff_algorithm;
50 static unsigned ws_error_highlight_default = WSEH_NEW;
51
52 static char diff_colors[][COLOR_MAXLEN] = {
53 GIT_COLOR_RESET,
54 GIT_COLOR_NORMAL, /* CONTEXT */
55 GIT_COLOR_BOLD, /* METAINFO */
56 GIT_COLOR_CYAN, /* FRAGINFO */
57 GIT_COLOR_RED, /* OLD */
58 GIT_COLOR_GREEN, /* NEW */
59 GIT_COLOR_YELLOW, /* COMMIT */
60 GIT_COLOR_BG_RED, /* WHITESPACE */
61 GIT_COLOR_NORMAL, /* FUNCINFO */
62 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
63 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
64 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
65 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
66 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
67 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
68 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
69 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
70 };
71
72 static NORETURN void die_want_option(const char *option_name)
73 {
74 die(_("option '%s' requires a value"), option_name);
75 }
76
77 static int parse_diff_color_slot(const char *var)
78 {
79 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
80 return DIFF_CONTEXT;
81 if (!strcasecmp(var, "meta"))
82 return DIFF_METAINFO;
83 if (!strcasecmp(var, "frag"))
84 return DIFF_FRAGINFO;
85 if (!strcasecmp(var, "old"))
86 return DIFF_FILE_OLD;
87 if (!strcasecmp(var, "new"))
88 return DIFF_FILE_NEW;
89 if (!strcasecmp(var, "commit"))
90 return DIFF_COMMIT;
91 if (!strcasecmp(var, "whitespace"))
92 return DIFF_WHITESPACE;
93 if (!strcasecmp(var, "func"))
94 return DIFF_FUNCINFO;
95 if (!strcasecmp(var, "oldmoved"))
96 return DIFF_FILE_OLD_MOVED;
97 if (!strcasecmp(var, "oldmovedalternative"))
98 return DIFF_FILE_OLD_MOVED_ALT;
99 if (!strcasecmp(var, "oldmoveddimmed"))
100 return DIFF_FILE_OLD_MOVED_DIM;
101 if (!strcasecmp(var, "oldmovedalternativedimmed"))
102 return DIFF_FILE_OLD_MOVED_ALT_DIM;
103 if (!strcasecmp(var, "newmoved"))
104 return DIFF_FILE_NEW_MOVED;
105 if (!strcasecmp(var, "newmovedalternative"))
106 return DIFF_FILE_NEW_MOVED_ALT;
107 if (!strcasecmp(var, "newmoveddimmed"))
108 return DIFF_FILE_NEW_MOVED_DIM;
109 if (!strcasecmp(var, "newmovedalternativedimmed"))
110 return DIFF_FILE_NEW_MOVED_ALT_DIM;
111 return -1;
112 }
113
114 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
115 struct strbuf *errmsg)
116 {
117 char *params_copy = xstrdup(params_string);
118 struct string_list params = STRING_LIST_INIT_NODUP;
119 int ret = 0;
120 int i;
121
122 if (*params_copy)
123 string_list_split_in_place(&params, params_copy, ',', -1);
124 for (i = 0; i < params.nr; i++) {
125 const char *p = params.items[i].string;
126 if (!strcmp(p, "changes")) {
127 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
128 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
129 } else if (!strcmp(p, "lines")) {
130 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
131 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
132 } else if (!strcmp(p, "files")) {
133 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
134 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
135 } else if (!strcmp(p, "noncumulative")) {
136 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
137 } else if (!strcmp(p, "cumulative")) {
138 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
139 } else if (isdigit(*p)) {
140 char *end;
141 int permille = strtoul(p, &end, 10) * 10;
142 if (*end == '.' && isdigit(*++end)) {
143 /* only use first digit */
144 permille += *end - '0';
145 /* .. and ignore any further digits */
146 while (isdigit(*++end))
147 ; /* nothing */
148 }
149 if (!*end)
150 options->dirstat_permille = permille;
151 else {
152 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
153 p);
154 ret++;
155 }
156 } else {
157 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
158 ret++;
159 }
160
161 }
162 string_list_clear(&params, 0);
163 free(params_copy);
164 return ret;
165 }
166
167 static int parse_submodule_params(struct diff_options *options, const char *value)
168 {
169 if (!strcmp(value, "log"))
170 options->submodule_format = DIFF_SUBMODULE_LOG;
171 else if (!strcmp(value, "short"))
172 options->submodule_format = DIFF_SUBMODULE_SHORT;
173 else if (!strcmp(value, "diff"))
174 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
175 else
176 return -1;
177 return 0;
178 }
179
180 static int git_config_rename(const char *var, const char *value)
181 {
182 if (!value)
183 return DIFF_DETECT_RENAME;
184 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
185 return DIFF_DETECT_COPY;
186 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
187 }
188
189 long parse_algorithm_value(const char *value)
190 {
191 if (!value)
192 return -1;
193 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
194 return 0;
195 else if (!strcasecmp(value, "minimal"))
196 return XDF_NEED_MINIMAL;
197 else if (!strcasecmp(value, "patience"))
198 return XDF_PATIENCE_DIFF;
199 else if (!strcasecmp(value, "histogram"))
200 return XDF_HISTOGRAM_DIFF;
201 return -1;
202 }
203
204 static int parse_one_token(const char **arg, const char *token)
205 {
206 const char *rest;
207 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
208 *arg = rest;
209 return 1;
210 }
211 return 0;
212 }
213
214 static int parse_ws_error_highlight(const char *arg)
215 {
216 const char *orig_arg = arg;
217 unsigned val = 0;
218
219 while (*arg) {
220 if (parse_one_token(&arg, "none"))
221 val = 0;
222 else if (parse_one_token(&arg, "default"))
223 val = WSEH_NEW;
224 else if (parse_one_token(&arg, "all"))
225 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
226 else if (parse_one_token(&arg, "new"))
227 val |= WSEH_NEW;
228 else if (parse_one_token(&arg, "old"))
229 val |= WSEH_OLD;
230 else if (parse_one_token(&arg, "context"))
231 val |= WSEH_CONTEXT;
232 else {
233 return -1 - (int)(arg - orig_arg);
234 }
235 if (*arg)
236 arg++;
237 }
238 return val;
239 }
240
241 /*
242 * These are to give UI layer defaults.
243 * The core-level commands such as git-diff-files should
244 * never be affected by the setting of diff.renames
245 * the user happens to have in the configuration file.
246 */
247 void init_diff_ui_defaults(void)
248 {
249 diff_detect_rename_default = 1;
250 }
251
252 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
253 {
254 if (!strcmp(var, "diff.indentheuristic"))
255 diff_indent_heuristic = git_config_bool(var, value);
256 return 0;
257 }
258
259 static int parse_color_moved(const char *arg)
260 {
261 switch (git_parse_maybe_bool(arg)) {
262 case 0:
263 return COLOR_MOVED_NO;
264 case 1:
265 return COLOR_MOVED_DEFAULT;
266 default:
267 break;
268 }
269
270 if (!strcmp(arg, "no"))
271 return COLOR_MOVED_NO;
272 else if (!strcmp(arg, "plain"))
273 return COLOR_MOVED_PLAIN;
274 else if (!strcmp(arg, "zebra"))
275 return COLOR_MOVED_ZEBRA;
276 else if (!strcmp(arg, "default"))
277 return COLOR_MOVED_DEFAULT;
278 else if (!strcmp(arg, "dimmed_zebra"))
279 return COLOR_MOVED_ZEBRA_DIM;
280 else
281 return error(_("color moved setting must be one of 'no', 'default', 'zebra', 'dimmed_zebra', 'plain'"));
282 }
283
284 int git_diff_ui_config(const char *var, const char *value, void *cb)
285 {
286 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
287 diff_use_color_default = git_config_colorbool(var, value);
288 return 0;
289 }
290 if (!strcmp(var, "diff.colormoved")) {
291 int cm = parse_color_moved(value);
292 if (cm < 0)
293 return -1;
294 diff_color_moved_default = cm;
295 return 0;
296 }
297 if (!strcmp(var, "diff.context")) {
298 diff_context_default = git_config_int(var, value);
299 if (diff_context_default < 0)
300 return -1;
301 return 0;
302 }
303 if (!strcmp(var, "diff.interhunkcontext")) {
304 diff_interhunk_context_default = git_config_int(var, value);
305 if (diff_interhunk_context_default < 0)
306 return -1;
307 return 0;
308 }
309 if (!strcmp(var, "diff.renames")) {
310 diff_detect_rename_default = git_config_rename(var, value);
311 return 0;
312 }
313 if (!strcmp(var, "diff.autorefreshindex")) {
314 diff_auto_refresh_index = git_config_bool(var, value);
315 return 0;
316 }
317 if (!strcmp(var, "diff.mnemonicprefix")) {
318 diff_mnemonic_prefix = git_config_bool(var, value);
319 return 0;
320 }
321 if (!strcmp(var, "diff.noprefix")) {
322 diff_no_prefix = git_config_bool(var, value);
323 return 0;
324 }
325 if (!strcmp(var, "diff.statgraphwidth")) {
326 diff_stat_graph_width = git_config_int(var, value);
327 return 0;
328 }
329 if (!strcmp(var, "diff.external"))
330 return git_config_string(&external_diff_cmd_cfg, var, value);
331 if (!strcmp(var, "diff.wordregex"))
332 return git_config_string(&diff_word_regex_cfg, var, value);
333 if (!strcmp(var, "diff.orderfile"))
334 return git_config_pathname(&diff_order_file_cfg, var, value);
335
336 if (!strcmp(var, "diff.ignoresubmodules"))
337 handle_ignore_submodules_arg(&default_diff_options, value);
338
339 if (!strcmp(var, "diff.submodule")) {
340 if (parse_submodule_params(&default_diff_options, value))
341 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
342 value);
343 return 0;
344 }
345
346 if (!strcmp(var, "diff.algorithm")) {
347 diff_algorithm = parse_algorithm_value(value);
348 if (diff_algorithm < 0)
349 return -1;
350 return 0;
351 }
352
353 if (!strcmp(var, "diff.wserrorhighlight")) {
354 int val = parse_ws_error_highlight(value);
355 if (val < 0)
356 return -1;
357 ws_error_highlight_default = val;
358 return 0;
359 }
360
361 return git_diff_basic_config(var, value, cb);
362 }
363
364 int git_diff_basic_config(const char *var, const char *value, void *cb)
365 {
366 const char *name;
367
368 if (!strcmp(var, "diff.renamelimit")) {
369 diff_rename_limit_default = git_config_int(var, value);
370 return 0;
371 }
372
373 if (userdiff_config(var, value) < 0)
374 return -1;
375
376 if (skip_prefix(var, "diff.color.", &name) ||
377 skip_prefix(var, "color.diff.", &name)) {
378 int slot = parse_diff_color_slot(name);
379 if (slot < 0)
380 return 0;
381 if (!value)
382 return config_error_nonbool(var);
383 return color_parse(value, diff_colors[slot]);
384 }
385
386 /* like GNU diff's --suppress-blank-empty option */
387 if (!strcmp(var, "diff.suppressblankempty") ||
388 /* for backwards compatibility */
389 !strcmp(var, "diff.suppress-blank-empty")) {
390 diff_suppress_blank_empty = git_config_bool(var, value);
391 return 0;
392 }
393
394 if (!strcmp(var, "diff.dirstat")) {
395 struct strbuf errmsg = STRBUF_INIT;
396 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
397 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
398 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
399 errmsg.buf);
400 strbuf_release(&errmsg);
401 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
402 return 0;
403 }
404
405 if (git_diff_heuristic_config(var, value, cb) < 0)
406 return -1;
407
408 return git_default_config(var, value, cb);
409 }
410
411 static char *quote_two(const char *one, const char *two)
412 {
413 int need_one = quote_c_style(one, NULL, NULL, 1);
414 int need_two = quote_c_style(two, NULL, NULL, 1);
415 struct strbuf res = STRBUF_INIT;
416
417 if (need_one + need_two) {
418 strbuf_addch(&res, '"');
419 quote_c_style(one, &res, NULL, 1);
420 quote_c_style(two, &res, NULL, 1);
421 strbuf_addch(&res, '"');
422 } else {
423 strbuf_addstr(&res, one);
424 strbuf_addstr(&res, two);
425 }
426 return strbuf_detach(&res, NULL);
427 }
428
429 static const char *external_diff(void)
430 {
431 static const char *external_diff_cmd = NULL;
432 static int done_preparing = 0;
433
434 if (done_preparing)
435 return external_diff_cmd;
436 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
437 if (!external_diff_cmd)
438 external_diff_cmd = external_diff_cmd_cfg;
439 done_preparing = 1;
440 return external_diff_cmd;
441 }
442
443 /*
444 * Keep track of files used for diffing. Sometimes such an entry
445 * refers to a temporary file, sometimes to an existing file, and
446 * sometimes to "/dev/null".
447 */
448 static struct diff_tempfile {
449 /*
450 * filename external diff should read from, or NULL if this
451 * entry is currently not in use:
452 */
453 const char *name;
454
455 char hex[GIT_MAX_HEXSZ + 1];
456 char mode[10];
457
458 /*
459 * If this diff_tempfile instance refers to a temporary file,
460 * this tempfile object is used to manage its lifetime.
461 */
462 struct tempfile *tempfile;
463 } diff_temp[2];
464
465 struct emit_callback {
466 int color_diff;
467 unsigned ws_rule;
468 int blank_at_eof_in_preimage;
469 int blank_at_eof_in_postimage;
470 int lno_in_preimage;
471 int lno_in_postimage;
472 const char **label_path;
473 struct diff_words_data *diff_words;
474 struct diff_options *opt;
475 struct strbuf *header;
476 };
477
478 static int count_lines(const char *data, int size)
479 {
480 int count, ch, completely_empty = 1, nl_just_seen = 0;
481 count = 0;
482 while (0 < size--) {
483 ch = *data++;
484 if (ch == '\n') {
485 count++;
486 nl_just_seen = 1;
487 completely_empty = 0;
488 }
489 else {
490 nl_just_seen = 0;
491 completely_empty = 0;
492 }
493 }
494 if (completely_empty)
495 return 0;
496 if (!nl_just_seen)
497 count++; /* no trailing newline */
498 return count;
499 }
500
501 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
502 {
503 if (!DIFF_FILE_VALID(one)) {
504 mf->ptr = (char *)""; /* does not matter */
505 mf->size = 0;
506 return 0;
507 }
508 else if (diff_populate_filespec(one, 0))
509 return -1;
510
511 mf->ptr = one->data;
512 mf->size = one->size;
513 return 0;
514 }
515
516 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
517 static unsigned long diff_filespec_size(struct diff_filespec *one)
518 {
519 if (!DIFF_FILE_VALID(one))
520 return 0;
521 diff_populate_filespec(one, CHECK_SIZE_ONLY);
522 return one->size;
523 }
524
525 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
526 {
527 char *ptr = mf->ptr;
528 long size = mf->size;
529 int cnt = 0;
530
531 if (!size)
532 return cnt;
533 ptr += size - 1; /* pointing at the very end */
534 if (*ptr != '\n')
535 ; /* incomplete line */
536 else
537 ptr--; /* skip the last LF */
538 while (mf->ptr < ptr) {
539 char *prev_eol;
540 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
541 if (*prev_eol == '\n')
542 break;
543 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
544 break;
545 cnt++;
546 ptr = prev_eol - 1;
547 }
548 return cnt;
549 }
550
551 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
552 struct emit_callback *ecbdata)
553 {
554 int l1, l2, at;
555 unsigned ws_rule = ecbdata->ws_rule;
556 l1 = count_trailing_blank(mf1, ws_rule);
557 l2 = count_trailing_blank(mf2, ws_rule);
558 if (l2 <= l1) {
559 ecbdata->blank_at_eof_in_preimage = 0;
560 ecbdata->blank_at_eof_in_postimage = 0;
561 return;
562 }
563 at = count_lines(mf1->ptr, mf1->size);
564 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
565
566 at = count_lines(mf2->ptr, mf2->size);
567 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
568 }
569
570 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
571 int first, const char *line, int len)
572 {
573 int has_trailing_newline, has_trailing_carriage_return;
574 int nofirst;
575 FILE *file = o->file;
576
577 fputs(diff_line_prefix(o), file);
578
579 if (len == 0) {
580 has_trailing_newline = (first == '\n');
581 has_trailing_carriage_return = (!has_trailing_newline &&
582 (first == '\r'));
583 nofirst = has_trailing_newline || has_trailing_carriage_return;
584 } else {
585 has_trailing_newline = (len > 0 && line[len-1] == '\n');
586 if (has_trailing_newline)
587 len--;
588 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
589 if (has_trailing_carriage_return)
590 len--;
591 nofirst = 0;
592 }
593
594 if (len || !nofirst) {
595 fputs(set, file);
596 if (!nofirst)
597 fputc(first, file);
598 fwrite(line, len, 1, file);
599 fputs(reset, file);
600 }
601 if (has_trailing_carriage_return)
602 fputc('\r', file);
603 if (has_trailing_newline)
604 fputc('\n', file);
605 }
606
607 static void emit_line(struct diff_options *o, const char *set, const char *reset,
608 const char *line, int len)
609 {
610 emit_line_0(o, set, reset, line[0], line+1, len-1);
611 }
612
613 enum diff_symbol {
614 DIFF_SYMBOL_BINARY_DIFF_HEADER,
615 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
616 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
617 DIFF_SYMBOL_BINARY_DIFF_BODY,
618 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
619 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
620 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
621 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
622 DIFF_SYMBOL_STATS_LINE,
623 DIFF_SYMBOL_WORD_DIFF,
624 DIFF_SYMBOL_STAT_SEP,
625 DIFF_SYMBOL_SUMMARY,
626 DIFF_SYMBOL_SUBMODULE_ADD,
627 DIFF_SYMBOL_SUBMODULE_DEL,
628 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
629 DIFF_SYMBOL_SUBMODULE_MODIFIED,
630 DIFF_SYMBOL_SUBMODULE_HEADER,
631 DIFF_SYMBOL_SUBMODULE_ERROR,
632 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
633 DIFF_SYMBOL_REWRITE_DIFF,
634 DIFF_SYMBOL_BINARY_FILES,
635 DIFF_SYMBOL_HEADER,
636 DIFF_SYMBOL_FILEPAIR_PLUS,
637 DIFF_SYMBOL_FILEPAIR_MINUS,
638 DIFF_SYMBOL_WORDS_PORCELAIN,
639 DIFF_SYMBOL_WORDS,
640 DIFF_SYMBOL_CONTEXT,
641 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
642 DIFF_SYMBOL_PLUS,
643 DIFF_SYMBOL_MINUS,
644 DIFF_SYMBOL_NO_LF_EOF,
645 DIFF_SYMBOL_CONTEXT_FRAGINFO,
646 DIFF_SYMBOL_CONTEXT_MARKER,
647 DIFF_SYMBOL_SEPARATOR
648 };
649 /*
650 * Flags for content lines:
651 * 0..12 are whitespace rules
652 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
653 * 16 is marking if the line is blank at EOF
654 */
655 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
656 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
657 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
658 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
659 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
660
661 /*
662 * This struct is used when we need to buffer the output of the diff output.
663 *
664 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
665 * into the pre/post image file. This pointer could be a union with the
666 * line pointer. By storing an offset into the file instead of the literal line,
667 * we can decrease the memory footprint for the buffered output. At first we
668 * may want to only have indirection for the content lines, but we could also
669 * enhance the state for emitting prefabricated lines, e.g. the similarity
670 * score line or hunk/file headers would only need to store a number or path
671 * and then the output can be constructed later on depending on state.
672 */
673 struct emitted_diff_symbol {
674 const char *line;
675 int len;
676 int flags;
677 enum diff_symbol s;
678 };
679 #define EMITTED_DIFF_SYMBOL_INIT {NULL}
680
681 struct emitted_diff_symbols {
682 struct emitted_diff_symbol *buf;
683 int nr, alloc;
684 };
685 #define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
686
687 static void append_emitted_diff_symbol(struct diff_options *o,
688 struct emitted_diff_symbol *e)
689 {
690 struct emitted_diff_symbol *f;
691
692 ALLOC_GROW(o->emitted_symbols->buf,
693 o->emitted_symbols->nr + 1,
694 o->emitted_symbols->alloc);
695 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
696
697 memcpy(f, e, sizeof(struct emitted_diff_symbol));
698 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
699 }
700
701 struct moved_entry {
702 struct hashmap_entry ent;
703 const struct emitted_diff_symbol *es;
704 struct moved_entry *next_line;
705 };
706
707 static int next_byte(const char **cp, const char **endp,
708 const struct diff_options *diffopt)
709 {
710 int retval;
711
712 if (*cp > *endp)
713 return -1;
714
715 if (isspace(**cp)) {
716 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_CHANGE)) {
717 while (*cp < *endp && isspace(**cp))
718 (*cp)++;
719 /*
720 * After skipping a couple of whitespaces,
721 * we still have to account for one space.
722 */
723 return (int)' ';
724 }
725
726 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE)) {
727 while (*cp < *endp && isspace(**cp))
728 (*cp)++;
729 /* return the first non-ws character via the usual below */
730 }
731 }
732
733 retval = (unsigned char)(**cp);
734 (*cp)++;
735 return retval;
736 }
737
738 static int moved_entry_cmp(const struct diff_options *diffopt,
739 const struct moved_entry *a,
740 const struct moved_entry *b,
741 const void *keydata)
742 {
743 const char *ap = a->es->line, *ae = a->es->line + a->es->len;
744 const char *bp = b->es->line, *be = b->es->line + b->es->len;
745
746 if (!(diffopt->xdl_opts & XDF_WHITESPACE_FLAGS))
747 return a->es->len != b->es->len || memcmp(ap, bp, a->es->len);
748
749 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_AT_EOL)) {
750 while (ae > ap && isspace(*ae))
751 ae--;
752 while (be > bp && isspace(*be))
753 be--;
754 }
755
756 while (1) {
757 int ca, cb;
758 ca = next_byte(&ap, &ae, diffopt);
759 cb = next_byte(&bp, &be, diffopt);
760 if (ca != cb)
761 return 1;
762 if (ca < 0)
763 return 0;
764 }
765 }
766
767 static unsigned get_string_hash(struct emitted_diff_symbol *es, struct diff_options *o)
768 {
769 if (o->xdl_opts & XDF_WHITESPACE_FLAGS) {
770 static struct strbuf sb = STRBUF_INIT;
771 const char *ap = es->line, *ae = es->line + es->len;
772 int c;
773
774 strbuf_reset(&sb);
775 while (ae > ap && isspace(*ae))
776 ae--;
777 while ((c = next_byte(&ap, &ae, o)) > 0)
778 strbuf_addch(&sb, c);
779
780 return memhash(sb.buf, sb.len);
781 } else {
782 return memhash(es->line, es->len);
783 }
784 }
785
786 static struct moved_entry *prepare_entry(struct diff_options *o,
787 int line_no)
788 {
789 struct moved_entry *ret = xmalloc(sizeof(*ret));
790 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
791
792 ret->ent.hash = get_string_hash(l, o);
793 ret->es = l;
794 ret->next_line = NULL;
795
796 return ret;
797 }
798
799 static void add_lines_to_move_detection(struct diff_options *o,
800 struct hashmap *add_lines,
801 struct hashmap *del_lines)
802 {
803 struct moved_entry *prev_line = NULL;
804
805 int n;
806 for (n = 0; n < o->emitted_symbols->nr; n++) {
807 struct hashmap *hm;
808 struct moved_entry *key;
809
810 switch (o->emitted_symbols->buf[n].s) {
811 case DIFF_SYMBOL_PLUS:
812 hm = add_lines;
813 break;
814 case DIFF_SYMBOL_MINUS:
815 hm = del_lines;
816 break;
817 default:
818 prev_line = NULL;
819 continue;
820 }
821
822 key = prepare_entry(o, n);
823 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
824 prev_line->next_line = key;
825
826 hashmap_add(hm, key);
827 prev_line = key;
828 }
829 }
830
831 static int shrink_potential_moved_blocks(struct moved_entry **pmb,
832 int pmb_nr)
833 {
834 int lp, rp;
835
836 /* Shrink the set of potential block to the remaining running */
837 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
838 while (lp < pmb_nr && pmb[lp])
839 lp++;
840 /* lp points at the first NULL now */
841
842 while (rp > -1 && !pmb[rp])
843 rp--;
844 /* rp points at the last non-NULL */
845
846 if (lp < pmb_nr && rp > -1 && lp < rp) {
847 pmb[lp] = pmb[rp];
848 pmb[rp] = NULL;
849 rp--;
850 lp++;
851 }
852 }
853
854 /* Remember the number of running sets */
855 return rp + 1;
856 }
857
858 /*
859 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
860 *
861 * Otherwise, if the last block has fewer alphanumeric characters than
862 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
863 * that block.
864 *
865 * The last block consists of the (n - block_length)'th line up to but not
866 * including the nth line.
867 *
868 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
869 * Think of a way to unify them.
870 */
871 static void adjust_last_block(struct diff_options *o, int n, int block_length)
872 {
873 int i, alnum_count = 0;
874 if (o->color_moved == COLOR_MOVED_PLAIN)
875 return;
876 for (i = 1; i < block_length + 1; i++) {
877 const char *c = o->emitted_symbols->buf[n - i].line;
878 for (; *c; c++) {
879 if (!isalnum(*c))
880 continue;
881 alnum_count++;
882 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
883 return;
884 }
885 }
886 for (i = 1; i < block_length + 1; i++)
887 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
888 }
889
890 /* Find blocks of moved code, delegate actual coloring decision to helper */
891 static void mark_color_as_moved(struct diff_options *o,
892 struct hashmap *add_lines,
893 struct hashmap *del_lines)
894 {
895 struct moved_entry **pmb = NULL; /* potentially moved blocks */
896 int pmb_nr = 0, pmb_alloc = 0;
897 int n, flipped_block = 1, block_length = 0;
898
899
900 for (n = 0; n < o->emitted_symbols->nr; n++) {
901 struct hashmap *hm = NULL;
902 struct moved_entry *key;
903 struct moved_entry *match = NULL;
904 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
905 int i;
906
907 switch (l->s) {
908 case DIFF_SYMBOL_PLUS:
909 hm = del_lines;
910 key = prepare_entry(o, n);
911 match = hashmap_get(hm, key, o);
912 free(key);
913 break;
914 case DIFF_SYMBOL_MINUS:
915 hm = add_lines;
916 key = prepare_entry(o, n);
917 match = hashmap_get(hm, key, o);
918 free(key);
919 break;
920 default:
921 flipped_block = 1;
922 }
923
924 if (!match) {
925 adjust_last_block(o, n, block_length);
926 pmb_nr = 0;
927 block_length = 0;
928 continue;
929 }
930
931 l->flags |= DIFF_SYMBOL_MOVED_LINE;
932
933 if (o->color_moved == COLOR_MOVED_PLAIN)
934 continue;
935
936 /* Check any potential block runs, advance each or nullify */
937 for (i = 0; i < pmb_nr; i++) {
938 struct moved_entry *p = pmb[i];
939 struct moved_entry *pnext = (p && p->next_line) ?
940 p->next_line : NULL;
941 if (pnext && !hm->cmpfn(o, pnext, match, NULL)) {
942 pmb[i] = p->next_line;
943 } else {
944 pmb[i] = NULL;
945 }
946 }
947
948 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
949
950 if (pmb_nr == 0) {
951 /*
952 * The current line is the start of a new block.
953 * Setup the set of potential blocks.
954 */
955 for (; match; match = hashmap_get_next(hm, match)) {
956 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
957 pmb[pmb_nr++] = match;
958 }
959
960 flipped_block = (flipped_block + 1) % 2;
961
962 adjust_last_block(o, n, block_length);
963 block_length = 0;
964 }
965
966 block_length++;
967
968 if (flipped_block)
969 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
970 }
971 adjust_last_block(o, n, block_length);
972
973 free(pmb);
974 }
975
976 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
977 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
978 static void dim_moved_lines(struct diff_options *o)
979 {
980 int n;
981 for (n = 0; n < o->emitted_symbols->nr; n++) {
982 struct emitted_diff_symbol *prev = (n != 0) ?
983 &o->emitted_symbols->buf[n - 1] : NULL;
984 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
985 struct emitted_diff_symbol *next =
986 (n < o->emitted_symbols->nr - 1) ?
987 &o->emitted_symbols->buf[n + 1] : NULL;
988
989 /* Not a plus or minus line? */
990 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
991 continue;
992
993 /* Not a moved line? */
994 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
995 continue;
996
997 /*
998 * If prev or next are not a plus or minus line,
999 * pretend they don't exist
1000 */
1001 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1002 prev->s != DIFF_SYMBOL_MINUS)
1003 prev = NULL;
1004 if (next && next->s != DIFF_SYMBOL_PLUS &&
1005 next->s != DIFF_SYMBOL_MINUS)
1006 next = NULL;
1007
1008 /* Inside a block? */
1009 if ((prev &&
1010 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1011 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1012 (next &&
1013 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1014 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1015 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1016 continue;
1017 }
1018
1019 /* Check if we are at an interesting bound: */
1020 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1021 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1022 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1023 continue;
1024 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1025 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1026 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1027 continue;
1028
1029 /*
1030 * The boundary to prev and next are not interesting,
1031 * so this line is not interesting as a whole
1032 */
1033 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1034 }
1035 }
1036
1037 static void emit_line_ws_markup(struct diff_options *o,
1038 const char *set, const char *reset,
1039 const char *line, int len, char sign,
1040 unsigned ws_rule, int blank_at_eof)
1041 {
1042 const char *ws = NULL;
1043
1044 if (o->ws_error_highlight & ws_rule) {
1045 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1046 if (!*ws)
1047 ws = NULL;
1048 }
1049
1050 if (!ws)
1051 emit_line_0(o, set, reset, sign, line, len);
1052 else if (blank_at_eof)
1053 /* Blank line at EOF - paint '+' as well */
1054 emit_line_0(o, ws, reset, sign, line, len);
1055 else {
1056 /* Emit just the prefix, then the rest. */
1057 emit_line_0(o, set, reset, sign, "", 0);
1058 ws_check_emit(line, len, ws_rule,
1059 o->file, set, reset, ws);
1060 }
1061 }
1062
1063 static void emit_diff_symbol_from_struct(struct diff_options *o,
1064 struct emitted_diff_symbol *eds)
1065 {
1066 static const char *nneof = " No newline at end of file\n";
1067 const char *context, *reset, *set, *meta, *fraginfo;
1068 struct strbuf sb = STRBUF_INIT;
1069
1070 enum diff_symbol s = eds->s;
1071 const char *line = eds->line;
1072 int len = eds->len;
1073 unsigned flags = eds->flags;
1074
1075 switch (s) {
1076 case DIFF_SYMBOL_NO_LF_EOF:
1077 context = diff_get_color_opt(o, DIFF_CONTEXT);
1078 reset = diff_get_color_opt(o, DIFF_RESET);
1079 putc('\n', o->file);
1080 emit_line_0(o, context, reset, '\\',
1081 nneof, strlen(nneof));
1082 break;
1083 case DIFF_SYMBOL_SUBMODULE_HEADER:
1084 case DIFF_SYMBOL_SUBMODULE_ERROR:
1085 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1086 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1087 case DIFF_SYMBOL_SUMMARY:
1088 case DIFF_SYMBOL_STATS_LINE:
1089 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1090 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1091 emit_line(o, "", "", line, len);
1092 break;
1093 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1094 case DIFF_SYMBOL_CONTEXT_MARKER:
1095 context = diff_get_color_opt(o, DIFF_CONTEXT);
1096 reset = diff_get_color_opt(o, DIFF_RESET);
1097 emit_line(o, context, reset, line, len);
1098 break;
1099 case DIFF_SYMBOL_SEPARATOR:
1100 fprintf(o->file, "%s%c",
1101 diff_line_prefix(o),
1102 o->line_termination);
1103 break;
1104 case DIFF_SYMBOL_CONTEXT:
1105 set = diff_get_color_opt(o, DIFF_CONTEXT);
1106 reset = diff_get_color_opt(o, DIFF_RESET);
1107 emit_line_ws_markup(o, set, reset, line, len, ' ',
1108 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1109 break;
1110 case DIFF_SYMBOL_PLUS:
1111 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1112 DIFF_SYMBOL_MOVED_LINE_ALT |
1113 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1114 case DIFF_SYMBOL_MOVED_LINE |
1115 DIFF_SYMBOL_MOVED_LINE_ALT |
1116 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1117 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1118 break;
1119 case DIFF_SYMBOL_MOVED_LINE |
1120 DIFF_SYMBOL_MOVED_LINE_ALT:
1121 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1122 break;
1123 case DIFF_SYMBOL_MOVED_LINE |
1124 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1125 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1126 break;
1127 case DIFF_SYMBOL_MOVED_LINE:
1128 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1129 break;
1130 default:
1131 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1132 }
1133 reset = diff_get_color_opt(o, DIFF_RESET);
1134 emit_line_ws_markup(o, set, reset, line, len, '+',
1135 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1136 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1137 break;
1138 case DIFF_SYMBOL_MINUS:
1139 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1140 DIFF_SYMBOL_MOVED_LINE_ALT |
1141 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1142 case DIFF_SYMBOL_MOVED_LINE |
1143 DIFF_SYMBOL_MOVED_LINE_ALT |
1144 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1145 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1146 break;
1147 case DIFF_SYMBOL_MOVED_LINE |
1148 DIFF_SYMBOL_MOVED_LINE_ALT:
1149 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1150 break;
1151 case DIFF_SYMBOL_MOVED_LINE |
1152 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1153 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1154 break;
1155 case DIFF_SYMBOL_MOVED_LINE:
1156 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1157 break;
1158 default:
1159 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1160 }
1161 reset = diff_get_color_opt(o, DIFF_RESET);
1162 emit_line_ws_markup(o, set, reset, line, len, '-',
1163 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1164 break;
1165 case DIFF_SYMBOL_WORDS_PORCELAIN:
1166 context = diff_get_color_opt(o, DIFF_CONTEXT);
1167 reset = diff_get_color_opt(o, DIFF_RESET);
1168 emit_line(o, context, reset, line, len);
1169 fputs("~\n", o->file);
1170 break;
1171 case DIFF_SYMBOL_WORDS:
1172 context = diff_get_color_opt(o, DIFF_CONTEXT);
1173 reset = diff_get_color_opt(o, DIFF_RESET);
1174 /*
1175 * Skip the prefix character, if any. With
1176 * diff_suppress_blank_empty, there may be
1177 * none.
1178 */
1179 if (line[0] != '\n') {
1180 line++;
1181 len--;
1182 }
1183 emit_line(o, context, reset, line, len);
1184 break;
1185 case DIFF_SYMBOL_FILEPAIR_PLUS:
1186 meta = diff_get_color_opt(o, DIFF_METAINFO);
1187 reset = diff_get_color_opt(o, DIFF_RESET);
1188 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1189 line, reset,
1190 strchr(line, ' ') ? "\t" : "");
1191 break;
1192 case DIFF_SYMBOL_FILEPAIR_MINUS:
1193 meta = diff_get_color_opt(o, DIFF_METAINFO);
1194 reset = diff_get_color_opt(o, DIFF_RESET);
1195 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1196 line, reset,
1197 strchr(line, ' ') ? "\t" : "");
1198 break;
1199 case DIFF_SYMBOL_BINARY_FILES:
1200 case DIFF_SYMBOL_HEADER:
1201 fprintf(o->file, "%s", line);
1202 break;
1203 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1204 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1205 break;
1206 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1207 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1208 break;
1209 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1210 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1211 break;
1212 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1213 fputs(diff_line_prefix(o), o->file);
1214 fputc('\n', o->file);
1215 break;
1216 case DIFF_SYMBOL_REWRITE_DIFF:
1217 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1218 reset = diff_get_color_opt(o, DIFF_RESET);
1219 emit_line(o, fraginfo, reset, line, len);
1220 break;
1221 case DIFF_SYMBOL_SUBMODULE_ADD:
1222 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1223 reset = diff_get_color_opt(o, DIFF_RESET);
1224 emit_line(o, set, reset, line, len);
1225 break;
1226 case DIFF_SYMBOL_SUBMODULE_DEL:
1227 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1228 reset = diff_get_color_opt(o, DIFF_RESET);
1229 emit_line(o, set, reset, line, len);
1230 break;
1231 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1232 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1233 diff_line_prefix(o), line);
1234 break;
1235 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1236 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1237 diff_line_prefix(o), line);
1238 break;
1239 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1240 emit_line(o, "", "", " 0 files changed\n",
1241 strlen(" 0 files changed\n"));
1242 break;
1243 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1244 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1245 break;
1246 case DIFF_SYMBOL_WORD_DIFF:
1247 fprintf(o->file, "%.*s", len, line);
1248 break;
1249 case DIFF_SYMBOL_STAT_SEP:
1250 fputs(o->stat_sep, o->file);
1251 break;
1252 default:
1253 die("BUG: unknown diff symbol");
1254 }
1255 strbuf_release(&sb);
1256 }
1257
1258 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1259 const char *line, int len, unsigned flags)
1260 {
1261 struct emitted_diff_symbol e = {line, len, flags, s};
1262
1263 if (o->emitted_symbols)
1264 append_emitted_diff_symbol(o, &e);
1265 else
1266 emit_diff_symbol_from_struct(o, &e);
1267 }
1268
1269 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1270 {
1271 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1272 }
1273
1274 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1275 {
1276 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1277 }
1278
1279 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1280 {
1281 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1282 path, strlen(path), 0);
1283 }
1284
1285 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1286 {
1287 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1288 path, strlen(path), 0);
1289 }
1290
1291 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1292 {
1293 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1294 header, strlen(header), 0);
1295 }
1296
1297 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1298 {
1299 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1300 }
1301
1302 void diff_emit_submodule_pipethrough(struct diff_options *o,
1303 const char *line, int len)
1304 {
1305 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1306 }
1307
1308 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1309 {
1310 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1311 ecbdata->blank_at_eof_in_preimage &&
1312 ecbdata->blank_at_eof_in_postimage &&
1313 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1314 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1315 return 0;
1316 return ws_blank_line(line, len, ecbdata->ws_rule);
1317 }
1318
1319 static void emit_add_line(const char *reset,
1320 struct emit_callback *ecbdata,
1321 const char *line, int len)
1322 {
1323 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1324 if (new_blank_line_at_eof(ecbdata, line, len))
1325 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1326
1327 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1328 }
1329
1330 static void emit_del_line(const char *reset,
1331 struct emit_callback *ecbdata,
1332 const char *line, int len)
1333 {
1334 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1335 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1336 }
1337
1338 static void emit_context_line(const char *reset,
1339 struct emit_callback *ecbdata,
1340 const char *line, int len)
1341 {
1342 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1343 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1344 }
1345
1346 static void emit_hunk_header(struct emit_callback *ecbdata,
1347 const char *line, int len)
1348 {
1349 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1350 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1351 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1352 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1353 static const char atat[2] = { '@', '@' };
1354 const char *cp, *ep;
1355 struct strbuf msgbuf = STRBUF_INIT;
1356 int org_len = len;
1357 int i = 1;
1358
1359 /*
1360 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1361 * it always is at least 10 bytes long.
1362 */
1363 if (len < 10 ||
1364 memcmp(line, atat, 2) ||
1365 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1366 emit_diff_symbol(ecbdata->opt,
1367 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1368 return;
1369 }
1370 ep += 2; /* skip over @@ */
1371
1372 /* The hunk header in fraginfo color */
1373 strbuf_addstr(&msgbuf, frag);
1374 strbuf_add(&msgbuf, line, ep - line);
1375 strbuf_addstr(&msgbuf, reset);
1376
1377 /*
1378 * trailing "\r\n"
1379 */
1380 for ( ; i < 3; i++)
1381 if (line[len - i] == '\r' || line[len - i] == '\n')
1382 len--;
1383
1384 /* blank before the func header */
1385 for (cp = ep; ep - line < len; ep++)
1386 if (*ep != ' ' && *ep != '\t')
1387 break;
1388 if (ep != cp) {
1389 strbuf_addstr(&msgbuf, context);
1390 strbuf_add(&msgbuf, cp, ep - cp);
1391 strbuf_addstr(&msgbuf, reset);
1392 }
1393
1394 if (ep < line + len) {
1395 strbuf_addstr(&msgbuf, func);
1396 strbuf_add(&msgbuf, ep, line + len - ep);
1397 strbuf_addstr(&msgbuf, reset);
1398 }
1399
1400 strbuf_add(&msgbuf, line + len, org_len - len);
1401 strbuf_complete_line(&msgbuf);
1402 emit_diff_symbol(ecbdata->opt,
1403 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1404 strbuf_release(&msgbuf);
1405 }
1406
1407 static struct diff_tempfile *claim_diff_tempfile(void) {
1408 int i;
1409 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1410 if (!diff_temp[i].name)
1411 return diff_temp + i;
1412 die("BUG: diff is failing to clean up its tempfiles");
1413 }
1414
1415 static void remove_tempfile(void)
1416 {
1417 int i;
1418 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1419 if (is_tempfile_active(diff_temp[i].tempfile))
1420 delete_tempfile(&diff_temp[i].tempfile);
1421 diff_temp[i].name = NULL;
1422 }
1423 }
1424
1425 static void add_line_count(struct strbuf *out, int count)
1426 {
1427 switch (count) {
1428 case 0:
1429 strbuf_addstr(out, "0,0");
1430 break;
1431 case 1:
1432 strbuf_addstr(out, "1");
1433 break;
1434 default:
1435 strbuf_addf(out, "1,%d", count);
1436 break;
1437 }
1438 }
1439
1440 static void emit_rewrite_lines(struct emit_callback *ecb,
1441 int prefix, const char *data, int size)
1442 {
1443 const char *endp = NULL;
1444 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1445
1446 while (0 < size) {
1447 int len;
1448
1449 endp = memchr(data, '\n', size);
1450 len = endp ? (endp - data + 1) : size;
1451 if (prefix != '+') {
1452 ecb->lno_in_preimage++;
1453 emit_del_line(reset, ecb, data, len);
1454 } else {
1455 ecb->lno_in_postimage++;
1456 emit_add_line(reset, ecb, data, len);
1457 }
1458 size -= len;
1459 data += len;
1460 }
1461 if (!endp)
1462 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1463 }
1464
1465 static void emit_rewrite_diff(const char *name_a,
1466 const char *name_b,
1467 struct diff_filespec *one,
1468 struct diff_filespec *two,
1469 struct userdiff_driver *textconv_one,
1470 struct userdiff_driver *textconv_two,
1471 struct diff_options *o)
1472 {
1473 int lc_a, lc_b;
1474 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1475 const char *a_prefix, *b_prefix;
1476 char *data_one, *data_two;
1477 size_t size_one, size_two;
1478 struct emit_callback ecbdata;
1479 struct strbuf out = STRBUF_INIT;
1480
1481 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
1482 a_prefix = o->b_prefix;
1483 b_prefix = o->a_prefix;
1484 } else {
1485 a_prefix = o->a_prefix;
1486 b_prefix = o->b_prefix;
1487 }
1488
1489 name_a += (*name_a == '/');
1490 name_b += (*name_b == '/');
1491
1492 strbuf_reset(&a_name);
1493 strbuf_reset(&b_name);
1494 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1495 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1496
1497 size_one = fill_textconv(textconv_one, one, &data_one);
1498 size_two = fill_textconv(textconv_two, two, &data_two);
1499
1500 memset(&ecbdata, 0, sizeof(ecbdata));
1501 ecbdata.color_diff = want_color(o->use_color);
1502 ecbdata.ws_rule = whitespace_rule(name_b);
1503 ecbdata.opt = o;
1504 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1505 mmfile_t mf1, mf2;
1506 mf1.ptr = (char *)data_one;
1507 mf2.ptr = (char *)data_two;
1508 mf1.size = size_one;
1509 mf2.size = size_two;
1510 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1511 }
1512 ecbdata.lno_in_preimage = 1;
1513 ecbdata.lno_in_postimage = 1;
1514
1515 lc_a = count_lines(data_one, size_one);
1516 lc_b = count_lines(data_two, size_two);
1517
1518 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1519 a_name.buf, a_name.len, 0);
1520 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1521 b_name.buf, b_name.len, 0);
1522
1523 strbuf_addstr(&out, "@@ -");
1524 if (!o->irreversible_delete)
1525 add_line_count(&out, lc_a);
1526 else
1527 strbuf_addstr(&out, "?,?");
1528 strbuf_addstr(&out, " +");
1529 add_line_count(&out, lc_b);
1530 strbuf_addstr(&out, " @@\n");
1531 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1532 strbuf_release(&out);
1533
1534 if (lc_a && !o->irreversible_delete)
1535 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1536 if (lc_b)
1537 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1538 if (textconv_one)
1539 free((char *)data_one);
1540 if (textconv_two)
1541 free((char *)data_two);
1542 }
1543
1544 struct diff_words_buffer {
1545 mmfile_t text;
1546 unsigned long alloc;
1547 struct diff_words_orig {
1548 const char *begin, *end;
1549 } *orig;
1550 int orig_nr, orig_alloc;
1551 };
1552
1553 static void diff_words_append(char *line, unsigned long len,
1554 struct diff_words_buffer *buffer)
1555 {
1556 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1557 line++;
1558 len--;
1559 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1560 buffer->text.size += len;
1561 buffer->text.ptr[buffer->text.size] = '\0';
1562 }
1563
1564 struct diff_words_style_elem {
1565 const char *prefix;
1566 const char *suffix;
1567 const char *color; /* NULL; filled in by the setup code if
1568 * color is enabled */
1569 };
1570
1571 struct diff_words_style {
1572 enum diff_words_type type;
1573 struct diff_words_style_elem new, old, ctx;
1574 const char *newline;
1575 };
1576
1577 static struct diff_words_style diff_words_styles[] = {
1578 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1579 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1580 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1581 };
1582
1583 struct diff_words_data {
1584 struct diff_words_buffer minus, plus;
1585 const char *current_plus;
1586 int last_minus;
1587 struct diff_options *opt;
1588 regex_t *word_regex;
1589 enum diff_words_type type;
1590 struct diff_words_style *style;
1591 };
1592
1593 static int fn_out_diff_words_write_helper(struct diff_options *o,
1594 struct diff_words_style_elem *st_el,
1595 const char *newline,
1596 size_t count, const char *buf)
1597 {
1598 int print = 0;
1599 struct strbuf sb = STRBUF_INIT;
1600
1601 while (count) {
1602 char *p = memchr(buf, '\n', count);
1603 if (print)
1604 strbuf_addstr(&sb, diff_line_prefix(o));
1605
1606 if (p != buf) {
1607 const char *reset = st_el->color && *st_el->color ?
1608 GIT_COLOR_RESET : NULL;
1609 if (st_el->color && *st_el->color)
1610 strbuf_addstr(&sb, st_el->color);
1611 strbuf_addstr(&sb, st_el->prefix);
1612 strbuf_add(&sb, buf, p ? p - buf : count);
1613 strbuf_addstr(&sb, st_el->suffix);
1614 if (reset)
1615 strbuf_addstr(&sb, reset);
1616 }
1617 if (!p)
1618 goto out;
1619
1620 strbuf_addstr(&sb, newline);
1621 count -= p + 1 - buf;
1622 buf = p + 1;
1623 print = 1;
1624 if (count) {
1625 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1626 sb.buf, sb.len, 0);
1627 strbuf_reset(&sb);
1628 }
1629 }
1630
1631 out:
1632 if (sb.len)
1633 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1634 sb.buf, sb.len, 0);
1635 strbuf_release(&sb);
1636 return 0;
1637 }
1638
1639 /*
1640 * '--color-words' algorithm can be described as:
1641 *
1642 * 1. collect the minus/plus lines of a diff hunk, divided into
1643 * minus-lines and plus-lines;
1644 *
1645 * 2. break both minus-lines and plus-lines into words and
1646 * place them into two mmfile_t with one word for each line;
1647 *
1648 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1649 *
1650 * And for the common parts of the both file, we output the plus side text.
1651 * diff_words->current_plus is used to trace the current position of the plus file
1652 * which printed. diff_words->last_minus is used to trace the last minus word
1653 * printed.
1654 *
1655 * For '--graph' to work with '--color-words', we need to output the graph prefix
1656 * on each line of color words output. Generally, there are two conditions on
1657 * which we should output the prefix.
1658 *
1659 * 1. diff_words->last_minus == 0 &&
1660 * diff_words->current_plus == diff_words->plus.text.ptr
1661 *
1662 * that is: the plus text must start as a new line, and if there is no minus
1663 * word printed, a graph prefix must be printed.
1664 *
1665 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1666 * *(diff_words->current_plus - 1) == '\n'
1667 *
1668 * that is: a graph prefix must be printed following a '\n'
1669 */
1670 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1671 {
1672 if ((diff_words->last_minus == 0 &&
1673 diff_words->current_plus == diff_words->plus.text.ptr) ||
1674 (diff_words->current_plus > diff_words->plus.text.ptr &&
1675 *(diff_words->current_plus - 1) == '\n')) {
1676 return 1;
1677 } else {
1678 return 0;
1679 }
1680 }
1681
1682 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1683 {
1684 struct diff_words_data *diff_words = priv;
1685 struct diff_words_style *style = diff_words->style;
1686 int minus_first, minus_len, plus_first, plus_len;
1687 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1688 struct diff_options *opt = diff_words->opt;
1689 const char *line_prefix;
1690
1691 if (line[0] != '@' || parse_hunk_header(line, len,
1692 &minus_first, &minus_len, &plus_first, &plus_len))
1693 return;
1694
1695 assert(opt);
1696 line_prefix = diff_line_prefix(opt);
1697
1698 /* POSIX requires that first be decremented by one if len == 0... */
1699 if (minus_len) {
1700 minus_begin = diff_words->minus.orig[minus_first].begin;
1701 minus_end =
1702 diff_words->minus.orig[minus_first + minus_len - 1].end;
1703 } else
1704 minus_begin = minus_end =
1705 diff_words->minus.orig[minus_first].end;
1706
1707 if (plus_len) {
1708 plus_begin = diff_words->plus.orig[plus_first].begin;
1709 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1710 } else
1711 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1712
1713 if (color_words_output_graph_prefix(diff_words)) {
1714 fputs(line_prefix, diff_words->opt->file);
1715 }
1716 if (diff_words->current_plus != plus_begin) {
1717 fn_out_diff_words_write_helper(diff_words->opt,
1718 &style->ctx, style->newline,
1719 plus_begin - diff_words->current_plus,
1720 diff_words->current_plus);
1721 }
1722 if (minus_begin != minus_end) {
1723 fn_out_diff_words_write_helper(diff_words->opt,
1724 &style->old, style->newline,
1725 minus_end - minus_begin, minus_begin);
1726 }
1727 if (plus_begin != plus_end) {
1728 fn_out_diff_words_write_helper(diff_words->opt,
1729 &style->new, style->newline,
1730 plus_end - plus_begin, plus_begin);
1731 }
1732
1733 diff_words->current_plus = plus_end;
1734 diff_words->last_minus = minus_first;
1735 }
1736
1737 /* This function starts looking at *begin, and returns 0 iff a word was found. */
1738 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1739 int *begin, int *end)
1740 {
1741 if (word_regex && *begin < buffer->size) {
1742 regmatch_t match[1];
1743 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1744 buffer->size - *begin, 1, match, 0)) {
1745 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1746 '\n', match[0].rm_eo - match[0].rm_so);
1747 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1748 *begin += match[0].rm_so;
1749 return *begin >= *end;
1750 }
1751 return -1;
1752 }
1753
1754 /* find the next word */
1755 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1756 (*begin)++;
1757 if (*begin >= buffer->size)
1758 return -1;
1759
1760 /* find the end of the word */
1761 *end = *begin + 1;
1762 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1763 (*end)++;
1764
1765 return 0;
1766 }
1767
1768 /*
1769 * This function splits the words in buffer->text, stores the list with
1770 * newline separator into out, and saves the offsets of the original words
1771 * in buffer->orig.
1772 */
1773 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1774 regex_t *word_regex)
1775 {
1776 int i, j;
1777 long alloc = 0;
1778
1779 out->size = 0;
1780 out->ptr = NULL;
1781
1782 /* fake an empty "0th" word */
1783 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1784 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1785 buffer->orig_nr = 1;
1786
1787 for (i = 0; i < buffer->text.size; i++) {
1788 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1789 return;
1790
1791 /* store original boundaries */
1792 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1793 buffer->orig_alloc);
1794 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1795 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1796 buffer->orig_nr++;
1797
1798 /* store one word */
1799 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1800 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1801 out->ptr[out->size + j - i] = '\n';
1802 out->size += j - i + 1;
1803
1804 i = j - 1;
1805 }
1806 }
1807
1808 /* this executes the word diff on the accumulated buffers */
1809 static void diff_words_show(struct diff_words_data *diff_words)
1810 {
1811 xpparam_t xpp;
1812 xdemitconf_t xecfg;
1813 mmfile_t minus, plus;
1814 struct diff_words_style *style = diff_words->style;
1815
1816 struct diff_options *opt = diff_words->opt;
1817 const char *line_prefix;
1818
1819 assert(opt);
1820 line_prefix = diff_line_prefix(opt);
1821
1822 /* special case: only removal */
1823 if (!diff_words->plus.text.size) {
1824 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1825 line_prefix, strlen(line_prefix), 0);
1826 fn_out_diff_words_write_helper(diff_words->opt,
1827 &style->old, style->newline,
1828 diff_words->minus.text.size,
1829 diff_words->minus.text.ptr);
1830 diff_words->minus.text.size = 0;
1831 return;
1832 }
1833
1834 diff_words->current_plus = diff_words->plus.text.ptr;
1835 diff_words->last_minus = 0;
1836
1837 memset(&xpp, 0, sizeof(xpp));
1838 memset(&xecfg, 0, sizeof(xecfg));
1839 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1840 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1841 xpp.flags = 0;
1842 /* as only the hunk header will be parsed, we need a 0-context */
1843 xecfg.ctxlen = 0;
1844 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1845 &xpp, &xecfg))
1846 die("unable to generate word diff");
1847 free(minus.ptr);
1848 free(plus.ptr);
1849 if (diff_words->current_plus != diff_words->plus.text.ptr +
1850 diff_words->plus.text.size) {
1851 if (color_words_output_graph_prefix(diff_words))
1852 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1853 line_prefix, strlen(line_prefix), 0);
1854 fn_out_diff_words_write_helper(diff_words->opt,
1855 &style->ctx, style->newline,
1856 diff_words->plus.text.ptr + diff_words->plus.text.size
1857 - diff_words->current_plus, diff_words->current_plus);
1858 }
1859 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1860 }
1861
1862 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1863 static void diff_words_flush(struct emit_callback *ecbdata)
1864 {
1865 struct diff_options *wo = ecbdata->diff_words->opt;
1866
1867 if (ecbdata->diff_words->minus.text.size ||
1868 ecbdata->diff_words->plus.text.size)
1869 diff_words_show(ecbdata->diff_words);
1870
1871 if (wo->emitted_symbols) {
1872 struct diff_options *o = ecbdata->opt;
1873 struct emitted_diff_symbols *wol = wo->emitted_symbols;
1874 int i;
1875
1876 /*
1877 * NEEDSWORK:
1878 * Instead of appending each, concat all words to a line?
1879 */
1880 for (i = 0; i < wol->nr; i++)
1881 append_emitted_diff_symbol(o, &wol->buf[i]);
1882
1883 for (i = 0; i < wol->nr; i++)
1884 free((void *)wol->buf[i].line);
1885
1886 wol->nr = 0;
1887 }
1888 }
1889
1890 static void diff_filespec_load_driver(struct diff_filespec *one)
1891 {
1892 /* Use already-loaded driver */
1893 if (one->driver)
1894 return;
1895
1896 if (S_ISREG(one->mode))
1897 one->driver = userdiff_find_by_path(one->path);
1898
1899 /* Fallback to default settings */
1900 if (!one->driver)
1901 one->driver = userdiff_find_by_name("default");
1902 }
1903
1904 static const char *userdiff_word_regex(struct diff_filespec *one)
1905 {
1906 diff_filespec_load_driver(one);
1907 return one->driver->word_regex;
1908 }
1909
1910 static void init_diff_words_data(struct emit_callback *ecbdata,
1911 struct diff_options *orig_opts,
1912 struct diff_filespec *one,
1913 struct diff_filespec *two)
1914 {
1915 int i;
1916 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1917 memcpy(o, orig_opts, sizeof(struct diff_options));
1918
1919 ecbdata->diff_words =
1920 xcalloc(1, sizeof(struct diff_words_data));
1921 ecbdata->diff_words->type = o->word_diff;
1922 ecbdata->diff_words->opt = o;
1923
1924 if (orig_opts->emitted_symbols)
1925 o->emitted_symbols =
1926 xcalloc(1, sizeof(struct emitted_diff_symbols));
1927
1928 if (!o->word_regex)
1929 o->word_regex = userdiff_word_regex(one);
1930 if (!o->word_regex)
1931 o->word_regex = userdiff_word_regex(two);
1932 if (!o->word_regex)
1933 o->word_regex = diff_word_regex_cfg;
1934 if (o->word_regex) {
1935 ecbdata->diff_words->word_regex = (regex_t *)
1936 xmalloc(sizeof(regex_t));
1937 if (regcomp(ecbdata->diff_words->word_regex,
1938 o->word_regex,
1939 REG_EXTENDED | REG_NEWLINE))
1940 die ("Invalid regular expression: %s",
1941 o->word_regex);
1942 }
1943 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1944 if (o->word_diff == diff_words_styles[i].type) {
1945 ecbdata->diff_words->style =
1946 &diff_words_styles[i];
1947 break;
1948 }
1949 }
1950 if (want_color(o->use_color)) {
1951 struct diff_words_style *st = ecbdata->diff_words->style;
1952 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1953 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1954 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1955 }
1956 }
1957
1958 static void free_diff_words_data(struct emit_callback *ecbdata)
1959 {
1960 if (ecbdata->diff_words) {
1961 diff_words_flush(ecbdata);
1962 free (ecbdata->diff_words->opt->emitted_symbols);
1963 free (ecbdata->diff_words->opt);
1964 free (ecbdata->diff_words->minus.text.ptr);
1965 free (ecbdata->diff_words->minus.orig);
1966 free (ecbdata->diff_words->plus.text.ptr);
1967 free (ecbdata->diff_words->plus.orig);
1968 if (ecbdata->diff_words->word_regex) {
1969 regfree(ecbdata->diff_words->word_regex);
1970 free(ecbdata->diff_words->word_regex);
1971 }
1972 FREE_AND_NULL(ecbdata->diff_words);
1973 }
1974 }
1975
1976 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1977 {
1978 if (want_color(diff_use_color))
1979 return diff_colors[ix];
1980 return "";
1981 }
1982
1983 const char *diff_line_prefix(struct diff_options *opt)
1984 {
1985 struct strbuf *msgbuf;
1986 if (!opt->output_prefix)
1987 return "";
1988
1989 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1990 return msgbuf->buf;
1991 }
1992
1993 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1994 {
1995 const char *cp;
1996 unsigned long allot;
1997 size_t l = len;
1998
1999 cp = line;
2000 allot = l;
2001 while (0 < l) {
2002 (void) utf8_width(&cp, &l);
2003 if (!cp)
2004 break; /* truncated in the middle? */
2005 }
2006 return allot - l;
2007 }
2008
2009 static void find_lno(const char *line, struct emit_callback *ecbdata)
2010 {
2011 const char *p;
2012 ecbdata->lno_in_preimage = 0;
2013 ecbdata->lno_in_postimage = 0;
2014 p = strchr(line, '-');
2015 if (!p)
2016 return; /* cannot happen */
2017 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2018 p = strchr(p, '+');
2019 if (!p)
2020 return; /* cannot happen */
2021 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2022 }
2023
2024 static void fn_out_consume(void *priv, char *line, unsigned long len)
2025 {
2026 struct emit_callback *ecbdata = priv;
2027 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2028 struct diff_options *o = ecbdata->opt;
2029
2030 o->found_changes = 1;
2031
2032 if (ecbdata->header) {
2033 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2034 ecbdata->header->buf, ecbdata->header->len, 0);
2035 strbuf_reset(ecbdata->header);
2036 ecbdata->header = NULL;
2037 }
2038
2039 if (ecbdata->label_path[0]) {
2040 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2041 ecbdata->label_path[0],
2042 strlen(ecbdata->label_path[0]), 0);
2043 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2044 ecbdata->label_path[1],
2045 strlen(ecbdata->label_path[1]), 0);
2046 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2047 }
2048
2049 if (diff_suppress_blank_empty
2050 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2051 line[0] = '\n';
2052 len = 1;
2053 }
2054
2055 if (line[0] == '@') {
2056 if (ecbdata->diff_words)
2057 diff_words_flush(ecbdata);
2058 len = sane_truncate_line(ecbdata, line, len);
2059 find_lno(line, ecbdata);
2060 emit_hunk_header(ecbdata, line, len);
2061 return;
2062 }
2063
2064 if (ecbdata->diff_words) {
2065 enum diff_symbol s =
2066 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2067 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2068 if (line[0] == '-') {
2069 diff_words_append(line, len,
2070 &ecbdata->diff_words->minus);
2071 return;
2072 } else if (line[0] == '+') {
2073 diff_words_append(line, len,
2074 &ecbdata->diff_words->plus);
2075 return;
2076 } else if (starts_with(line, "\\ ")) {
2077 /*
2078 * Eat the "no newline at eof" marker as if we
2079 * saw a "+" or "-" line with nothing on it,
2080 * and return without diff_words_flush() to
2081 * defer processing. If this is the end of
2082 * preimage, more "+" lines may come after it.
2083 */
2084 return;
2085 }
2086 diff_words_flush(ecbdata);
2087 emit_diff_symbol(o, s, line, len, 0);
2088 return;
2089 }
2090
2091 switch (line[0]) {
2092 case '+':
2093 ecbdata->lno_in_postimage++;
2094 emit_add_line(reset, ecbdata, line + 1, len - 1);
2095 break;
2096 case '-':
2097 ecbdata->lno_in_preimage++;
2098 emit_del_line(reset, ecbdata, line + 1, len - 1);
2099 break;
2100 case ' ':
2101 ecbdata->lno_in_postimage++;
2102 ecbdata->lno_in_preimage++;
2103 emit_context_line(reset, ecbdata, line + 1, len - 1);
2104 break;
2105 default:
2106 /* incomplete line at the end */
2107 ecbdata->lno_in_preimage++;
2108 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2109 line, len, 0);
2110 break;
2111 }
2112 }
2113
2114 static char *pprint_rename(const char *a, const char *b)
2115 {
2116 const char *old = a;
2117 const char *new = b;
2118 struct strbuf name = STRBUF_INIT;
2119 int pfx_length, sfx_length;
2120 int pfx_adjust_for_slash;
2121 int len_a = strlen(a);
2122 int len_b = strlen(b);
2123 int a_midlen, b_midlen;
2124 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2125 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2126
2127 if (qlen_a || qlen_b) {
2128 quote_c_style(a, &name, NULL, 0);
2129 strbuf_addstr(&name, " => ");
2130 quote_c_style(b, &name, NULL, 0);
2131 return strbuf_detach(&name, NULL);
2132 }
2133
2134 /* Find common prefix */
2135 pfx_length = 0;
2136 while (*old && *new && *old == *new) {
2137 if (*old == '/')
2138 pfx_length = old - a + 1;
2139 old++;
2140 new++;
2141 }
2142
2143 /* Find common suffix */
2144 old = a + len_a;
2145 new = b + len_b;
2146 sfx_length = 0;
2147 /*
2148 * If there is a common prefix, it must end in a slash. In
2149 * that case we let this loop run 1 into the prefix to see the
2150 * same slash.
2151 *
2152 * If there is no common prefix, we cannot do this as it would
2153 * underrun the input strings.
2154 */
2155 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2156 while (a + pfx_length - pfx_adjust_for_slash <= old &&
2157 b + pfx_length - pfx_adjust_for_slash <= new &&
2158 *old == *new) {
2159 if (*old == '/')
2160 sfx_length = len_a - (old - a);
2161 old--;
2162 new--;
2163 }
2164
2165 /*
2166 * pfx{mid-a => mid-b}sfx
2167 * {pfx-a => pfx-b}sfx
2168 * pfx{sfx-a => sfx-b}
2169 * name-a => name-b
2170 */
2171 a_midlen = len_a - pfx_length - sfx_length;
2172 b_midlen = len_b - pfx_length - sfx_length;
2173 if (a_midlen < 0)
2174 a_midlen = 0;
2175 if (b_midlen < 0)
2176 b_midlen = 0;
2177
2178 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2179 if (pfx_length + sfx_length) {
2180 strbuf_add(&name, a, pfx_length);
2181 strbuf_addch(&name, '{');
2182 }
2183 strbuf_add(&name, a + pfx_length, a_midlen);
2184 strbuf_addstr(&name, " => ");
2185 strbuf_add(&name, b + pfx_length, b_midlen);
2186 if (pfx_length + sfx_length) {
2187 strbuf_addch(&name, '}');
2188 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
2189 }
2190 return strbuf_detach(&name, NULL);
2191 }
2192
2193 struct diffstat_t {
2194 int nr;
2195 int alloc;
2196 struct diffstat_file {
2197 char *from_name;
2198 char *name;
2199 char *print_name;
2200 unsigned is_unmerged:1;
2201 unsigned is_binary:1;
2202 unsigned is_renamed:1;
2203 unsigned is_interesting:1;
2204 uintmax_t added, deleted;
2205 } **files;
2206 };
2207
2208 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2209 const char *name_a,
2210 const char *name_b)
2211 {
2212 struct diffstat_file *x;
2213 x = xcalloc(1, sizeof(*x));
2214 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2215 diffstat->files[diffstat->nr++] = x;
2216 if (name_b) {
2217 x->from_name = xstrdup(name_a);
2218 x->name = xstrdup(name_b);
2219 x->is_renamed = 1;
2220 }
2221 else {
2222 x->from_name = NULL;
2223 x->name = xstrdup(name_a);
2224 }
2225 return x;
2226 }
2227
2228 static void diffstat_consume(void *priv, char *line, unsigned long len)
2229 {
2230 struct diffstat_t *diffstat = priv;
2231 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2232
2233 if (line[0] == '+')
2234 x->added++;
2235 else if (line[0] == '-')
2236 x->deleted++;
2237 }
2238
2239 const char mime_boundary_leader[] = "------------";
2240
2241 static int scale_linear(int it, int width, int max_change)
2242 {
2243 if (!it)
2244 return 0;
2245 /*
2246 * make sure that at least one '-' or '+' is printed if
2247 * there is any change to this path. The easiest way is to
2248 * scale linearly as if the alloted width is one column shorter
2249 * than it is, and then add 1 to the result.
2250 */
2251 return 1 + (it * (width - 1) / max_change);
2252 }
2253
2254 static void show_graph(struct strbuf *out, char ch, int cnt,
2255 const char *set, const char *reset)
2256 {
2257 if (cnt <= 0)
2258 return;
2259 strbuf_addstr(out, set);
2260 strbuf_addchars(out, ch, cnt);
2261 strbuf_addstr(out, reset);
2262 }
2263
2264 static void fill_print_name(struct diffstat_file *file)
2265 {
2266 char *pname;
2267
2268 if (file->print_name)
2269 return;
2270
2271 if (!file->is_renamed) {
2272 struct strbuf buf = STRBUF_INIT;
2273 if (quote_c_style(file->name, &buf, NULL, 0)) {
2274 pname = strbuf_detach(&buf, NULL);
2275 } else {
2276 pname = file->name;
2277 strbuf_release(&buf);
2278 }
2279 } else {
2280 pname = pprint_rename(file->from_name, file->name);
2281 }
2282 file->print_name = pname;
2283 }
2284
2285 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2286 int files, int insertions, int deletions)
2287 {
2288 struct strbuf sb = STRBUF_INIT;
2289
2290 if (!files) {
2291 assert(insertions == 0 && deletions == 0);
2292 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2293 NULL, 0, 0);
2294 return;
2295 }
2296
2297 strbuf_addf(&sb,
2298 (files == 1) ? " %d file changed" : " %d files changed",
2299 files);
2300
2301 /*
2302 * For binary diff, the caller may want to print "x files
2303 * changed" with insertions == 0 && deletions == 0.
2304 *
2305 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2306 * is probably less confusing (i.e skip over "2 files changed
2307 * but nothing about added/removed lines? Is this a bug in Git?").
2308 */
2309 if (insertions || deletions == 0) {
2310 strbuf_addf(&sb,
2311 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2312 insertions);
2313 }
2314
2315 if (deletions || insertions == 0) {
2316 strbuf_addf(&sb,
2317 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2318 deletions);
2319 }
2320 strbuf_addch(&sb, '\n');
2321 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2322 sb.buf, sb.len, 0);
2323 strbuf_release(&sb);
2324 }
2325
2326 void print_stat_summary(FILE *fp, int files,
2327 int insertions, int deletions)
2328 {
2329 struct diff_options o;
2330 memset(&o, 0, sizeof(o));
2331 o.file = fp;
2332
2333 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2334 }
2335
2336 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2337 {
2338 int i, len, add, del, adds = 0, dels = 0;
2339 uintmax_t max_change = 0, max_len = 0;
2340 int total_files = data->nr, count;
2341 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2342 const char *reset, *add_c, *del_c;
2343 int extra_shown = 0;
2344 const char *line_prefix = diff_line_prefix(options);
2345 struct strbuf out = STRBUF_INIT;
2346
2347 if (data->nr == 0)
2348 return;
2349
2350 count = options->stat_count ? options->stat_count : data->nr;
2351
2352 reset = diff_get_color_opt(options, DIFF_RESET);
2353 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2354 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2355
2356 /*
2357 * Find the longest filename and max number of changes
2358 */
2359 for (i = 0; (i < count) && (i < data->nr); i++) {
2360 struct diffstat_file *file = data->files[i];
2361 uintmax_t change = file->added + file->deleted;
2362
2363 if (!file->is_interesting && (change == 0)) {
2364 count++; /* not shown == room for one more */
2365 continue;
2366 }
2367 fill_print_name(file);
2368 len = strlen(file->print_name);
2369 if (max_len < len)
2370 max_len = len;
2371
2372 if (file->is_unmerged) {
2373 /* "Unmerged" is 8 characters */
2374 bin_width = bin_width < 8 ? 8 : bin_width;
2375 continue;
2376 }
2377 if (file->is_binary) {
2378 /* "Bin XXX -> YYY bytes" */
2379 int w = 14 + decimal_width(file->added)
2380 + decimal_width(file->deleted);
2381 bin_width = bin_width < w ? w : bin_width;
2382 /* Display change counts aligned with "Bin" */
2383 number_width = 3;
2384 continue;
2385 }
2386
2387 if (max_change < change)
2388 max_change = change;
2389 }
2390 count = i; /* where we can stop scanning in data->files[] */
2391
2392 /*
2393 * We have width = stat_width or term_columns() columns total.
2394 * We want a maximum of min(max_len, stat_name_width) for the name part.
2395 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2396 * We also need 1 for " " and 4 + decimal_width(max_change)
2397 * for " | NNNN " and one the empty column at the end, altogether
2398 * 6 + decimal_width(max_change).
2399 *
2400 * If there's not enough space, we will use the smaller of
2401 * stat_name_width (if set) and 5/8*width for the filename,
2402 * and the rest for constant elements + graph part, but no more
2403 * than stat_graph_width for the graph part.
2404 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2405 * for the standard terminal size).
2406 *
2407 * In other words: stat_width limits the maximum width, and
2408 * stat_name_width fixes the maximum width of the filename,
2409 * and is also used to divide available columns if there
2410 * aren't enough.
2411 *
2412 * Binary files are displayed with "Bin XXX -> YYY bytes"
2413 * instead of the change count and graph. This part is treated
2414 * similarly to the graph part, except that it is not
2415 * "scaled". If total width is too small to accommodate the
2416 * guaranteed minimum width of the filename part and the
2417 * separators and this message, this message will "overflow"
2418 * making the line longer than the maximum width.
2419 */
2420
2421 if (options->stat_width == -1)
2422 width = term_columns() - strlen(line_prefix);
2423 else
2424 width = options->stat_width ? options->stat_width : 80;
2425 number_width = decimal_width(max_change) > number_width ?
2426 decimal_width(max_change) : number_width;
2427
2428 if (options->stat_graph_width == -1)
2429 options->stat_graph_width = diff_stat_graph_width;
2430
2431 /*
2432 * Guarantee 3/8*16==6 for the graph part
2433 * and 5/8*16==10 for the filename part
2434 */
2435 if (width < 16 + 6 + number_width)
2436 width = 16 + 6 + number_width;
2437
2438 /*
2439 * First assign sizes that are wanted, ignoring available width.
2440 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2441 * starting from "XXX" should fit in graph_width.
2442 */
2443 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2444 if (options->stat_graph_width &&
2445 options->stat_graph_width < graph_width)
2446 graph_width = options->stat_graph_width;
2447
2448 name_width = (options->stat_name_width > 0 &&
2449 options->stat_name_width < max_len) ?
2450 options->stat_name_width : max_len;
2451
2452 /*
2453 * Adjust adjustable widths not to exceed maximum width
2454 */
2455 if (name_width + number_width + 6 + graph_width > width) {
2456 if (graph_width > width * 3/8 - number_width - 6) {
2457 graph_width = width * 3/8 - number_width - 6;
2458 if (graph_width < 6)
2459 graph_width = 6;
2460 }
2461
2462 if (options->stat_graph_width &&
2463 graph_width > options->stat_graph_width)
2464 graph_width = options->stat_graph_width;
2465 if (name_width > width - number_width - 6 - graph_width)
2466 name_width = width - number_width - 6 - graph_width;
2467 else
2468 graph_width = width - number_width - 6 - name_width;
2469 }
2470
2471 /*
2472 * From here name_width is the width of the name area,
2473 * and graph_width is the width of the graph area.
2474 * max_change is used to scale graph properly.
2475 */
2476 for (i = 0; i < count; i++) {
2477 const char *prefix = "";
2478 struct diffstat_file *file = data->files[i];
2479 char *name = file->print_name;
2480 uintmax_t added = file->added;
2481 uintmax_t deleted = file->deleted;
2482 int name_len;
2483
2484 if (!file->is_interesting && (added + deleted == 0))
2485 continue;
2486
2487 /*
2488 * "scale" the filename
2489 */
2490 len = name_width;
2491 name_len = strlen(name);
2492 if (name_width < name_len) {
2493 char *slash;
2494 prefix = "...";
2495 len -= 3;
2496 name += name_len - len;
2497 slash = strchr(name, '/');
2498 if (slash)
2499 name = slash;
2500 }
2501
2502 if (file->is_binary) {
2503 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2504 strbuf_addf(&out, " %*s", number_width, "Bin");
2505 if (!added && !deleted) {
2506 strbuf_addch(&out, '\n');
2507 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2508 out.buf, out.len, 0);
2509 strbuf_reset(&out);
2510 continue;
2511 }
2512 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2513 del_c, deleted, reset);
2514 strbuf_addstr(&out, " -> ");
2515 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2516 add_c, added, reset);
2517 strbuf_addstr(&out, " bytes\n");
2518 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2519 out.buf, out.len, 0);
2520 strbuf_reset(&out);
2521 continue;
2522 }
2523 else if (file->is_unmerged) {
2524 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2525 strbuf_addstr(&out, " Unmerged\n");
2526 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2527 out.buf, out.len, 0);
2528 strbuf_reset(&out);
2529 continue;
2530 }
2531
2532 /*
2533 * scale the add/delete
2534 */
2535 add = added;
2536 del = deleted;
2537
2538 if (graph_width <= max_change) {
2539 int total = scale_linear(add + del, graph_width, max_change);
2540 if (total < 2 && add && del)
2541 /* width >= 2 due to the sanity check */
2542 total = 2;
2543 if (add < del) {
2544 add = scale_linear(add, graph_width, max_change);
2545 del = total - add;
2546 } else {
2547 del = scale_linear(del, graph_width, max_change);
2548 add = total - del;
2549 }
2550 }
2551 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2552 strbuf_addf(&out, " %*"PRIuMAX"%s",
2553 number_width, added + deleted,
2554 added + deleted ? " " : "");
2555 show_graph(&out, '+', add, add_c, reset);
2556 show_graph(&out, '-', del, del_c, reset);
2557 strbuf_addch(&out, '\n');
2558 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2559 out.buf, out.len, 0);
2560 strbuf_reset(&out);
2561 }
2562
2563 for (i = 0; i < data->nr; i++) {
2564 struct diffstat_file *file = data->files[i];
2565 uintmax_t added = file->added;
2566 uintmax_t deleted = file->deleted;
2567
2568 if (file->is_unmerged ||
2569 (!file->is_interesting && (added + deleted == 0))) {
2570 total_files--;
2571 continue;
2572 }
2573
2574 if (!file->is_binary) {
2575 adds += added;
2576 dels += deleted;
2577 }
2578 if (i < count)
2579 continue;
2580 if (!extra_shown)
2581 emit_diff_symbol(options,
2582 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2583 NULL, 0, 0);
2584 extra_shown = 1;
2585 }
2586
2587 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2588 strbuf_release(&out);
2589 }
2590
2591 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2592 {
2593 int i, adds = 0, dels = 0, total_files = data->nr;
2594
2595 if (data->nr == 0)
2596 return;
2597
2598 for (i = 0; i < data->nr; i++) {
2599 int added = data->files[i]->added;
2600 int deleted = data->files[i]->deleted;
2601
2602 if (data->files[i]->is_unmerged ||
2603 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2604 total_files--;
2605 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2606 adds += added;
2607 dels += deleted;
2608 }
2609 }
2610 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2611 }
2612
2613 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2614 {
2615 int i;
2616
2617 if (data->nr == 0)
2618 return;
2619
2620 for (i = 0; i < data->nr; i++) {
2621 struct diffstat_file *file = data->files[i];
2622
2623 fprintf(options->file, "%s", diff_line_prefix(options));
2624
2625 if (file->is_binary)
2626 fprintf(options->file, "-\t-\t");
2627 else
2628 fprintf(options->file,
2629 "%"PRIuMAX"\t%"PRIuMAX"\t",
2630 file->added, file->deleted);
2631 if (options->line_termination) {
2632 fill_print_name(file);
2633 if (!file->is_renamed)
2634 write_name_quoted(file->name, options->file,
2635 options->line_termination);
2636 else {
2637 fputs(file->print_name, options->file);
2638 putc(options->line_termination, options->file);
2639 }
2640 } else {
2641 if (file->is_renamed) {
2642 putc('\0', options->file);
2643 write_name_quoted(file->from_name, options->file, '\0');
2644 }
2645 write_name_quoted(file->name, options->file, '\0');
2646 }
2647 }
2648 }
2649
2650 struct dirstat_file {
2651 const char *name;
2652 unsigned long changed;
2653 };
2654
2655 struct dirstat_dir {
2656 struct dirstat_file *files;
2657 int alloc, nr, permille, cumulative;
2658 };
2659
2660 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2661 unsigned long changed, const char *base, int baselen)
2662 {
2663 unsigned long this_dir = 0;
2664 unsigned int sources = 0;
2665 const char *line_prefix = diff_line_prefix(opt);
2666
2667 while (dir->nr) {
2668 struct dirstat_file *f = dir->files;
2669 int namelen = strlen(f->name);
2670 unsigned long this;
2671 char *slash;
2672
2673 if (namelen < baselen)
2674 break;
2675 if (memcmp(f->name, base, baselen))
2676 break;
2677 slash = strchr(f->name + baselen, '/');
2678 if (slash) {
2679 int newbaselen = slash + 1 - f->name;
2680 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2681 sources++;
2682 } else {
2683 this = f->changed;
2684 dir->files++;
2685 dir->nr--;
2686 sources += 2;
2687 }
2688 this_dir += this;
2689 }
2690
2691 /*
2692 * We don't report dirstat's for
2693 * - the top level
2694 * - or cases where everything came from a single directory
2695 * under this directory (sources == 1).
2696 */
2697 if (baselen && sources != 1) {
2698 if (this_dir) {
2699 int permille = this_dir * 1000 / changed;
2700 if (permille >= dir->permille) {
2701 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2702 permille / 10, permille % 10, baselen, base);
2703 if (!dir->cumulative)
2704 return 0;
2705 }
2706 }
2707 }
2708 return this_dir;
2709 }
2710
2711 static int dirstat_compare(const void *_a, const void *_b)
2712 {
2713 const struct dirstat_file *a = _a;
2714 const struct dirstat_file *b = _b;
2715 return strcmp(a->name, b->name);
2716 }
2717
2718 static void show_dirstat(struct diff_options *options)
2719 {
2720 int i;
2721 unsigned long changed;
2722 struct dirstat_dir dir;
2723 struct diff_queue_struct *q = &diff_queued_diff;
2724
2725 dir.files = NULL;
2726 dir.alloc = 0;
2727 dir.nr = 0;
2728 dir.permille = options->dirstat_permille;
2729 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2730
2731 changed = 0;
2732 for (i = 0; i < q->nr; i++) {
2733 struct diff_filepair *p = q->queue[i];
2734 const char *name;
2735 unsigned long copied, added, damage;
2736 int content_changed;
2737
2738 name = p->two->path ? p->two->path : p->one->path;
2739
2740 if (p->one->oid_valid && p->two->oid_valid)
2741 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2742 else
2743 content_changed = 1;
2744
2745 if (!content_changed) {
2746 /*
2747 * The SHA1 has not changed, so pre-/post-content is
2748 * identical. We can therefore skip looking at the
2749 * file contents altogether.
2750 */
2751 damage = 0;
2752 goto found_damage;
2753 }
2754
2755 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2756 /*
2757 * In --dirstat-by-file mode, we don't really need to
2758 * look at the actual file contents at all.
2759 * The fact that the SHA1 changed is enough for us to
2760 * add this file to the list of results
2761 * (with each file contributing equal damage).
2762 */
2763 damage = 1;
2764 goto found_damage;
2765 }
2766
2767 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2768 diff_populate_filespec(p->one, 0);
2769 diff_populate_filespec(p->two, 0);
2770 diffcore_count_changes(p->one, p->two, NULL, NULL,
2771 &copied, &added);
2772 diff_free_filespec_data(p->one);
2773 diff_free_filespec_data(p->two);
2774 } else if (DIFF_FILE_VALID(p->one)) {
2775 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2776 copied = added = 0;
2777 diff_free_filespec_data(p->one);
2778 } else if (DIFF_FILE_VALID(p->two)) {
2779 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2780 copied = 0;
2781 added = p->two->size;
2782 diff_free_filespec_data(p->two);
2783 } else
2784 continue;
2785
2786 /*
2787 * Original minus copied is the removed material,
2788 * added is the new material. They are both damages
2789 * made to the preimage.
2790 * If the resulting damage is zero, we know that
2791 * diffcore_count_changes() considers the two entries to
2792 * be identical, but since content_changed is true, we
2793 * know that there must have been _some_ kind of change,
2794 * so we force all entries to have damage > 0.
2795 */
2796 damage = (p->one->size - copied) + added;
2797 if (!damage)
2798 damage = 1;
2799
2800 found_damage:
2801 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2802 dir.files[dir.nr].name = name;
2803 dir.files[dir.nr].changed = damage;
2804 changed += damage;
2805 dir.nr++;
2806 }
2807
2808 /* This can happen even with many files, if everything was renames */
2809 if (!changed)
2810 return;
2811
2812 /* Show all directories with more than x% of the changes */
2813 QSORT(dir.files, dir.nr, dirstat_compare);
2814 gather_dirstat(options, &dir, changed, "", 0);
2815 }
2816
2817 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2818 {
2819 int i;
2820 unsigned long changed;
2821 struct dirstat_dir dir;
2822
2823 if (data->nr == 0)
2824 return;
2825
2826 dir.files = NULL;
2827 dir.alloc = 0;
2828 dir.nr = 0;
2829 dir.permille = options->dirstat_permille;
2830 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2831
2832 changed = 0;
2833 for (i = 0; i < data->nr; i++) {
2834 struct diffstat_file *file = data->files[i];
2835 unsigned long damage = file->added + file->deleted;
2836 if (file->is_binary)
2837 /*
2838 * binary files counts bytes, not lines. Must find some
2839 * way to normalize binary bytes vs. textual lines.
2840 * The following heuristic assumes that there are 64
2841 * bytes per "line".
2842 * This is stupid and ugly, but very cheap...
2843 */
2844 damage = DIV_ROUND_UP(damage, 64);
2845 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2846 dir.files[dir.nr].name = file->name;
2847 dir.files[dir.nr].changed = damage;
2848 changed += damage;
2849 dir.nr++;
2850 }
2851
2852 /* This can happen even with many files, if everything was renames */
2853 if (!changed)
2854 return;
2855
2856 /* Show all directories with more than x% of the changes */
2857 QSORT(dir.files, dir.nr, dirstat_compare);
2858 gather_dirstat(options, &dir, changed, "", 0);
2859 }
2860
2861 static void free_diffstat_info(struct diffstat_t *diffstat)
2862 {
2863 int i;
2864 for (i = 0; i < diffstat->nr; i++) {
2865 struct diffstat_file *f = diffstat->files[i];
2866 if (f->name != f->print_name)
2867 free(f->print_name);
2868 free(f->name);
2869 free(f->from_name);
2870 free(f);
2871 }
2872 free(diffstat->files);
2873 }
2874
2875 struct checkdiff_t {
2876 const char *filename;
2877 int lineno;
2878 int conflict_marker_size;
2879 struct diff_options *o;
2880 unsigned ws_rule;
2881 unsigned status;
2882 };
2883
2884 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2885 {
2886 char firstchar;
2887 int cnt;
2888
2889 if (len < marker_size + 1)
2890 return 0;
2891 firstchar = line[0];
2892 switch (firstchar) {
2893 case '=': case '>': case '<': case '|':
2894 break;
2895 default:
2896 return 0;
2897 }
2898 for (cnt = 1; cnt < marker_size; cnt++)
2899 if (line[cnt] != firstchar)
2900 return 0;
2901 /* line[1] thru line[marker_size-1] are same as firstchar */
2902 if (len < marker_size + 1 || !isspace(line[marker_size]))
2903 return 0;
2904 return 1;
2905 }
2906
2907 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2908 {
2909 struct checkdiff_t *data = priv;
2910 int marker_size = data->conflict_marker_size;
2911 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2912 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2913 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2914 char *err;
2915 const char *line_prefix;
2916
2917 assert(data->o);
2918 line_prefix = diff_line_prefix(data->o);
2919
2920 if (line[0] == '+') {
2921 unsigned bad;
2922 data->lineno++;
2923 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2924 data->status |= 1;
2925 fprintf(data->o->file,
2926 "%s%s:%d: leftover conflict marker\n",
2927 line_prefix, data->filename, data->lineno);
2928 }
2929 bad = ws_check(line + 1, len - 1, data->ws_rule);
2930 if (!bad)
2931 return;
2932 data->status |= bad;
2933 err = whitespace_error_string(bad);
2934 fprintf(data->o->file, "%s%s:%d: %s.\n",
2935 line_prefix, data->filename, data->lineno, err);
2936 free(err);
2937 emit_line(data->o, set, reset, line, 1);
2938 ws_check_emit(line + 1, len - 1, data->ws_rule,
2939 data->o->file, set, reset, ws);
2940 } else if (line[0] == ' ') {
2941 data->lineno++;
2942 } else if (line[0] == '@') {
2943 char *plus = strchr(line, '+');
2944 if (plus)
2945 data->lineno = strtol(plus, NULL, 10) - 1;
2946 else
2947 die("invalid diff");
2948 }
2949 }
2950
2951 static unsigned char *deflate_it(char *data,
2952 unsigned long size,
2953 unsigned long *result_size)
2954 {
2955 int bound;
2956 unsigned char *deflated;
2957 git_zstream stream;
2958
2959 git_deflate_init(&stream, zlib_compression_level);
2960 bound = git_deflate_bound(&stream, size);
2961 deflated = xmalloc(bound);
2962 stream.next_out = deflated;
2963 stream.avail_out = bound;
2964
2965 stream.next_in = (unsigned char *)data;
2966 stream.avail_in = size;
2967 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2968 ; /* nothing */
2969 git_deflate_end(&stream);
2970 *result_size = stream.total_out;
2971 return deflated;
2972 }
2973
2974 static void emit_binary_diff_body(struct diff_options *o,
2975 mmfile_t *one, mmfile_t *two)
2976 {
2977 void *cp;
2978 void *delta;
2979 void *deflated;
2980 void *data;
2981 unsigned long orig_size;
2982 unsigned long delta_size;
2983 unsigned long deflate_size;
2984 unsigned long data_size;
2985
2986 /* We could do deflated delta, or we could do just deflated two,
2987 * whichever is smaller.
2988 */
2989 delta = NULL;
2990 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2991 if (one->size && two->size) {
2992 delta = diff_delta(one->ptr, one->size,
2993 two->ptr, two->size,
2994 &delta_size, deflate_size);
2995 if (delta) {
2996 void *to_free = delta;
2997 orig_size = delta_size;
2998 delta = deflate_it(delta, delta_size, &delta_size);
2999 free(to_free);
3000 }
3001 }
3002
3003 if (delta && delta_size < deflate_size) {
3004 char *s = xstrfmt("%lu", orig_size);
3005 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3006 s, strlen(s), 0);
3007 free(s);
3008 free(deflated);
3009 data = delta;
3010 data_size = delta_size;
3011 } else {
3012 char *s = xstrfmt("%lu", two->size);
3013 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3014 s, strlen(s), 0);
3015 free(s);
3016 free(delta);
3017 data = deflated;
3018 data_size = deflate_size;
3019 }
3020
3021 /* emit data encoded in base85 */
3022 cp = data;
3023 while (data_size) {
3024 int len;
3025 int bytes = (52 < data_size) ? 52 : data_size;
3026 char line[71];
3027 data_size -= bytes;
3028 if (bytes <= 26)
3029 line[0] = bytes + 'A' - 1;
3030 else
3031 line[0] = bytes - 26 + 'a' - 1;
3032 encode_85(line + 1, cp, bytes);
3033 cp = (char *) cp + bytes;
3034
3035 len = strlen(line);
3036 line[len++] = '\n';
3037 line[len] = '\0';
3038
3039 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3040 line, len, 0);
3041 }
3042 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3043 free(data);
3044 }
3045
3046 static void emit_binary_diff(struct diff_options *o,
3047 mmfile_t *one, mmfile_t *two)
3048 {
3049 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3050 emit_binary_diff_body(o, one, two);
3051 emit_binary_diff_body(o, two, one);
3052 }
3053
3054 int diff_filespec_is_binary(struct diff_filespec *one)
3055 {
3056 if (one->is_binary == -1) {
3057 diff_filespec_load_driver(one);
3058 if (one->driver->binary != -1)
3059 one->is_binary = one->driver->binary;
3060 else {
3061 if (!one->data && DIFF_FILE_VALID(one))
3062 diff_populate_filespec(one, CHECK_BINARY);
3063 if (one->is_binary == -1 && one->data)
3064 one->is_binary = buffer_is_binary(one->data,
3065 one->size);
3066 if (one->is_binary == -1)
3067 one->is_binary = 0;
3068 }
3069 }
3070 return one->is_binary;
3071 }
3072
3073 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3074 {
3075 diff_filespec_load_driver(one);
3076 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3077 }
3078
3079 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3080 {
3081 if (!options->a_prefix)
3082 options->a_prefix = a;
3083 if (!options->b_prefix)
3084 options->b_prefix = b;
3085 }
3086
3087 struct userdiff_driver *get_textconv(struct diff_filespec *one)
3088 {
3089 if (!DIFF_FILE_VALID(one))
3090 return NULL;
3091
3092 diff_filespec_load_driver(one);
3093 return userdiff_get_textconv(one->driver);
3094 }
3095
3096 static void builtin_diff(const char *name_a,
3097 const char *name_b,
3098 struct diff_filespec *one,
3099 struct diff_filespec *two,
3100 const char *xfrm_msg,
3101 int must_show_header,
3102 struct diff_options *o,
3103 int complete_rewrite)
3104 {
3105 mmfile_t mf1, mf2;
3106 const char *lbl[2];
3107 char *a_one, *b_two;
3108 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3109 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3110 const char *a_prefix, *b_prefix;
3111 struct userdiff_driver *textconv_one = NULL;
3112 struct userdiff_driver *textconv_two = NULL;
3113 struct strbuf header = STRBUF_INIT;
3114 const char *line_prefix = diff_line_prefix(o);
3115
3116 diff_set_mnemonic_prefix(o, "a/", "b/");
3117 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
3118 a_prefix = o->b_prefix;
3119 b_prefix = o->a_prefix;
3120 } else {
3121 a_prefix = o->a_prefix;
3122 b_prefix = o->b_prefix;
3123 }
3124
3125 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3126 (!one->mode || S_ISGITLINK(one->mode)) &&
3127 (!two->mode || S_ISGITLINK(two->mode))) {
3128 show_submodule_summary(o, one->path ? one->path : two->path,
3129 &one->oid, &two->oid,
3130 two->dirty_submodule);
3131 return;
3132 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3133 (!one->mode || S_ISGITLINK(one->mode)) &&
3134 (!two->mode || S_ISGITLINK(two->mode))) {
3135 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3136 &one->oid, &two->oid,
3137 two->dirty_submodule);
3138 return;
3139 }
3140
3141 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
3142 textconv_one = get_textconv(one);
3143 textconv_two = get_textconv(two);
3144 }
3145
3146 /* Never use a non-valid filename anywhere if at all possible */
3147 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3148 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3149
3150 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3151 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3152 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3153 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3154 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3155 if (lbl[0][0] == '/') {
3156 /* /dev/null */
3157 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3158 if (xfrm_msg)
3159 strbuf_addstr(&header, xfrm_msg);
3160 must_show_header = 1;
3161 }
3162 else if (lbl[1][0] == '/') {
3163 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3164 if (xfrm_msg)
3165 strbuf_addstr(&header, xfrm_msg);
3166 must_show_header = 1;
3167 }
3168 else {
3169 if (one->mode != two->mode) {
3170 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3171 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3172 must_show_header = 1;
3173 }
3174 if (xfrm_msg)
3175 strbuf_addstr(&header, xfrm_msg);
3176
3177 /*
3178 * we do not run diff between different kind
3179 * of objects.
3180 */
3181 if ((one->mode ^ two->mode) & S_IFMT)
3182 goto free_ab_and_return;
3183 if (complete_rewrite &&
3184 (textconv_one || !diff_filespec_is_binary(one)) &&
3185 (textconv_two || !diff_filespec_is_binary(two))) {
3186 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3187 header.buf, header.len, 0);
3188 strbuf_reset(&header);
3189 emit_rewrite_diff(name_a, name_b, one, two,
3190 textconv_one, textconv_two, o);
3191 o->found_changes = 1;
3192 goto free_ab_and_return;
3193 }
3194 }
3195
3196 if (o->irreversible_delete && lbl[1][0] == '/') {
3197 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3198 header.len, 0);
3199 strbuf_reset(&header);
3200 goto free_ab_and_return;
3201 } else if (!DIFF_OPT_TST(o, TEXT) &&
3202 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3203 (!textconv_two && diff_filespec_is_binary(two)) )) {
3204 struct strbuf sb = STRBUF_INIT;
3205 if (!one->data && !two->data &&
3206 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3207 !DIFF_OPT_TST(o, BINARY)) {
3208 if (!oidcmp(&one->oid, &two->oid)) {
3209 if (must_show_header)
3210 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3211 header.buf, header.len,
3212 0);
3213 goto free_ab_and_return;
3214 }
3215 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3216 header.buf, header.len, 0);
3217 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3218 diff_line_prefix(o), lbl[0], lbl[1]);
3219 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3220 sb.buf, sb.len, 0);
3221 strbuf_release(&sb);
3222 goto free_ab_and_return;
3223 }
3224 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3225 die("unable to read files to diff");
3226 /* Quite common confusing case */
3227 if (mf1.size == mf2.size &&
3228 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3229 if (must_show_header)
3230 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3231 header.buf, header.len, 0);
3232 goto free_ab_and_return;
3233 }
3234 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3235 strbuf_reset(&header);
3236 if (DIFF_OPT_TST(o, BINARY))
3237 emit_binary_diff(o, &mf1, &mf2);
3238 else {
3239 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3240 diff_line_prefix(o), lbl[0], lbl[1]);
3241 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3242 sb.buf, sb.len, 0);
3243 strbuf_release(&sb);
3244 }
3245 o->found_changes = 1;
3246 } else {
3247 /* Crazy xdl interfaces.. */
3248 const char *diffopts = getenv("GIT_DIFF_OPTS");
3249 const char *v;
3250 xpparam_t xpp;
3251 xdemitconf_t xecfg;
3252 struct emit_callback ecbdata;
3253 const struct userdiff_funcname *pe;
3254
3255 if (must_show_header) {
3256 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3257 header.buf, header.len, 0);
3258 strbuf_reset(&header);
3259 }
3260
3261 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3262 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3263
3264 pe = diff_funcname_pattern(one);
3265 if (!pe)
3266 pe = diff_funcname_pattern(two);
3267
3268 memset(&xpp, 0, sizeof(xpp));
3269 memset(&xecfg, 0, sizeof(xecfg));
3270 memset(&ecbdata, 0, sizeof(ecbdata));
3271 ecbdata.label_path = lbl;
3272 ecbdata.color_diff = want_color(o->use_color);
3273 ecbdata.ws_rule = whitespace_rule(name_b);
3274 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3275 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3276 ecbdata.opt = o;
3277 ecbdata.header = header.len ? &header : NULL;
3278 xpp.flags = o->xdl_opts;
3279 xecfg.ctxlen = o->context;
3280 xecfg.interhunkctxlen = o->interhunkcontext;
3281 xecfg.flags = XDL_EMIT_FUNCNAMES;
3282 if (DIFF_OPT_TST(o, FUNCCONTEXT))
3283 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3284 if (pe)
3285 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3286 if (!diffopts)
3287 ;
3288 else if (skip_prefix(diffopts, "--unified=", &v))
3289 xecfg.ctxlen = strtoul(v, NULL, 10);
3290 else if (skip_prefix(diffopts, "-u", &v))
3291 xecfg.ctxlen = strtoul(v, NULL, 10);
3292 if (o->word_diff)
3293 init_diff_words_data(&ecbdata, o, one, two);
3294 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
3295 &xpp, &xecfg))
3296 die("unable to generate diff for %s", one->path);
3297 if (o->word_diff)
3298 free_diff_words_data(&ecbdata);
3299 if (textconv_one)
3300 free(mf1.ptr);
3301 if (textconv_two)
3302 free(mf2.ptr);
3303 xdiff_clear_find_func(&xecfg);
3304 }
3305
3306 free_ab_and_return:
3307 strbuf_release(&header);
3308 diff_free_filespec_data(one);
3309 diff_free_filespec_data(two);
3310 free(a_one);
3311 free(b_two);
3312 return;
3313 }
3314
3315 static void builtin_diffstat(const char *name_a, const char *name_b,
3316 struct diff_filespec *one,
3317 struct diff_filespec *two,
3318 struct diffstat_t *diffstat,
3319 struct diff_options *o,
3320 struct diff_filepair *p)
3321 {
3322 mmfile_t mf1, mf2;
3323 struct diffstat_file *data;
3324 int same_contents;
3325 int complete_rewrite = 0;
3326
3327 if (!DIFF_PAIR_UNMERGED(p)) {
3328 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3329 complete_rewrite = 1;
3330 }
3331
3332 data = diffstat_add(diffstat, name_a, name_b);
3333 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3334
3335 if (!one || !two) {
3336 data->is_unmerged = 1;
3337 return;
3338 }
3339
3340 same_contents = !oidcmp(&one->oid, &two->oid);
3341
3342 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3343 data->is_binary = 1;
3344 if (same_contents) {
3345 data->added = 0;
3346 data->deleted = 0;
3347 } else {
3348 data->added = diff_filespec_size(two);
3349 data->deleted = diff_filespec_size(one);
3350 }
3351 }
3352
3353 else if (complete_rewrite) {
3354 diff_populate_filespec(one, 0);
3355 diff_populate_filespec(two, 0);
3356 data->deleted = count_lines(one->data, one->size);
3357 data->added = count_lines(two->data, two->size);
3358 }
3359
3360 else if (!same_contents) {
3361 /* Crazy xdl interfaces.. */
3362 xpparam_t xpp;
3363 xdemitconf_t xecfg;
3364
3365 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3366 die("unable to read files to diff");
3367
3368 memset(&xpp, 0, sizeof(xpp));
3369 memset(&xecfg, 0, sizeof(xecfg));
3370 xpp.flags = o->xdl_opts;
3371 xecfg.ctxlen = o->context;
3372 xecfg.interhunkctxlen = o->interhunkcontext;
3373 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
3374 &xpp, &xecfg))
3375 die("unable to generate diffstat for %s", one->path);
3376 }
3377
3378 diff_free_filespec_data(one);
3379 diff_free_filespec_data(two);
3380 }
3381
3382 static void builtin_checkdiff(const char *name_a, const char *name_b,
3383 const char *attr_path,
3384 struct diff_filespec *one,
3385 struct diff_filespec *two,
3386 struct diff_options *o)
3387 {
3388 mmfile_t mf1, mf2;
3389 struct checkdiff_t data;
3390
3391 if (!two)
3392 return;
3393
3394 memset(&data, 0, sizeof(data));
3395 data.filename = name_b ? name_b : name_a;
3396 data.lineno = 0;
3397 data.o = o;
3398 data.ws_rule = whitespace_rule(attr_path);
3399 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3400
3401 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3402 die("unable to read files to diff");
3403
3404 /*
3405 * All the other codepaths check both sides, but not checking
3406 * the "old" side here is deliberate. We are checking the newly
3407 * introduced changes, and as long as the "new" side is text, we
3408 * can and should check what it introduces.
3409 */
3410 if (diff_filespec_is_binary(two))
3411 goto free_and_return;
3412 else {
3413 /* Crazy xdl interfaces.. */
3414 xpparam_t xpp;
3415 xdemitconf_t xecfg;
3416
3417 memset(&xpp, 0, sizeof(xpp));
3418 memset(&xecfg, 0, sizeof(xecfg));
3419 xecfg.ctxlen = 1; /* at least one context line */
3420 xpp.flags = 0;
3421 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
3422 &xpp, &xecfg))
3423 die("unable to generate checkdiff for %s", one->path);
3424
3425 if (data.ws_rule & WS_BLANK_AT_EOF) {
3426 struct emit_callback ecbdata;
3427 int blank_at_eof;
3428
3429 ecbdata.ws_rule = data.ws_rule;
3430 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3431 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3432
3433 if (blank_at_eof) {
3434 static char *err;
3435 if (!err)
3436 err = whitespace_error_string(WS_BLANK_AT_EOF);
3437 fprintf(o->file, "%s:%d: %s.\n",
3438 data.filename, blank_at_eof, err);
3439 data.status = 1; /* report errors */
3440 }
3441 }
3442 }
3443 free_and_return:
3444 diff_free_filespec_data(one);
3445 diff_free_filespec_data(two);
3446 if (data.status)
3447 DIFF_OPT_SET(o, CHECK_FAILED);
3448 }
3449
3450 struct diff_filespec *alloc_filespec(const char *path)
3451 {
3452 struct diff_filespec *spec;
3453
3454 FLEXPTR_ALLOC_STR(spec, path, path);
3455 spec->count = 1;
3456 spec->is_binary = -1;
3457 return spec;
3458 }
3459
3460 void free_filespec(struct diff_filespec *spec)
3461 {
3462 if (!--spec->count) {
3463 diff_free_filespec_data(spec);
3464 free(spec);
3465 }
3466 }
3467
3468 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3469 int oid_valid, unsigned short mode)
3470 {
3471 if (mode) {
3472 spec->mode = canon_mode(mode);
3473 oidcpy(&spec->oid, oid);
3474 spec->oid_valid = oid_valid;
3475 }
3476 }
3477
3478 /*
3479 * Given a name and sha1 pair, if the index tells us the file in
3480 * the work tree has that object contents, return true, so that
3481 * prepare_temp_file() does not have to inflate and extract.
3482 */
3483 static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3484 {
3485 const struct cache_entry *ce;
3486 struct stat st;
3487 int pos, len;
3488
3489 /*
3490 * We do not read the cache ourselves here, because the
3491 * benchmark with my previous version that always reads cache
3492 * shows that it makes things worse for diff-tree comparing
3493 * two linux-2.6 kernel trees in an already checked out work
3494 * tree. This is because most diff-tree comparisons deal with
3495 * only a small number of files, while reading the cache is
3496 * expensive for a large project, and its cost outweighs the
3497 * savings we get by not inflating the object to a temporary
3498 * file. Practically, this code only helps when we are used
3499 * by diff-cache --cached, which does read the cache before
3500 * calling us.
3501 */
3502 if (!active_cache)
3503 return 0;
3504
3505 /* We want to avoid the working directory if our caller
3506 * doesn't need the data in a normal file, this system
3507 * is rather slow with its stat/open/mmap/close syscalls,
3508 * and the object is contained in a pack file. The pack
3509 * is probably already open and will be faster to obtain
3510 * the data through than the working directory. Loose
3511 * objects however would tend to be slower as they need
3512 * to be individually opened and inflated.
3513 */
3514 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(oid->hash))
3515 return 0;
3516
3517 /*
3518 * Similarly, if we'd have to convert the file contents anyway, that
3519 * makes the optimization not worthwhile.
3520 */
3521 if (!want_file && would_convert_to_git(&the_index, name))
3522 return 0;
3523
3524 len = strlen(name);
3525 pos = cache_name_pos(name, len);
3526 if (pos < 0)
3527 return 0;
3528 ce = active_cache[pos];
3529
3530 /*
3531 * This is not the sha1 we are looking for, or
3532 * unreusable because it is not a regular file.
3533 */
3534 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3535 return 0;
3536
3537 /*
3538 * If ce is marked as "assume unchanged", there is no
3539 * guarantee that work tree matches what we are looking for.
3540 */
3541 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3542 return 0;
3543
3544 /*
3545 * If ce matches the file in the work tree, we can reuse it.
3546 */
3547 if (ce_uptodate(ce) ||
3548 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3549 return 1;
3550
3551 return 0;
3552 }
3553
3554 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3555 {
3556 struct strbuf buf = STRBUF_INIT;
3557 char *dirty = "";
3558
3559 /* Are we looking at the work tree? */
3560 if (s->dirty_submodule)
3561 dirty = "-dirty";
3562
3563 strbuf_addf(&buf, "Subproject commit %s%s\n",
3564 oid_to_hex(&s->oid), dirty);
3565 s->size = buf.len;
3566 if (size_only) {
3567 s->data = NULL;
3568 strbuf_release(&buf);
3569 } else {
3570 s->data = strbuf_detach(&buf, NULL);
3571 s->should_free = 1;
3572 }
3573 return 0;
3574 }
3575
3576 /*
3577 * While doing rename detection and pickaxe operation, we may need to
3578 * grab the data for the blob (or file) for our own in-core comparison.
3579 * diff_filespec has data and size fields for this purpose.
3580 */
3581 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3582 {
3583 int size_only = flags & CHECK_SIZE_ONLY;
3584 int err = 0;
3585 /*
3586 * demote FAIL to WARN to allow inspecting the situation
3587 * instead of refusing.
3588 */
3589 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
3590 ? SAFE_CRLF_WARN
3591 : safe_crlf);
3592
3593 if (!DIFF_FILE_VALID(s))
3594 die("internal error: asking to populate invalid file.");
3595 if (S_ISDIR(s->mode))
3596 return -1;
3597
3598 if (s->data)
3599 return 0;
3600
3601 if (size_only && 0 < s->size)
3602 return 0;
3603
3604 if (S_ISGITLINK(s->mode))
3605 return diff_populate_gitlink(s, size_only);
3606
3607 if (!s->oid_valid ||
3608 reuse_worktree_file(s->path, &s->oid, 0)) {
3609 struct strbuf buf = STRBUF_INIT;
3610 struct stat st;
3611 int fd;
3612
3613 if (lstat(s->path, &st) < 0) {
3614 if (errno == ENOENT) {
3615 err_empty:
3616 err = -1;
3617 empty:
3618 s->data = (char *)"";
3619 s->size = 0;
3620 return err;
3621 }
3622 }
3623 s->size = xsize_t(st.st_size);
3624 if (!s->size)
3625 goto empty;
3626 if (S_ISLNK(st.st_mode)) {
3627 struct strbuf sb = STRBUF_INIT;
3628
3629 if (strbuf_readlink(&sb, s->path, s->size))
3630 goto err_empty;
3631 s->size = sb.len;
3632 s->data = strbuf_detach(&sb, NULL);
3633 s->should_free = 1;
3634 return 0;
3635 }
3636
3637 /*
3638 * Even if the caller would be happy with getting
3639 * only the size, we cannot return early at this
3640 * point if the path requires us to run the content
3641 * conversion.
3642 */
3643 if (size_only && !would_convert_to_git(&the_index, s->path))
3644 return 0;
3645
3646 /*
3647 * Note: this check uses xsize_t(st.st_size) that may
3648 * not be the true size of the blob after it goes
3649 * through convert_to_git(). This may not strictly be
3650 * correct, but the whole point of big_file_threshold
3651 * and is_binary check being that we want to avoid
3652 * opening the file and inspecting the contents, this
3653 * is probably fine.
3654 */
3655 if ((flags & CHECK_BINARY) &&
3656 s->size > big_file_threshold && s->is_binary == -1) {
3657 s->is_binary = 1;
3658 return 0;
3659 }
3660 fd = open(s->path, O_RDONLY);
3661 if (fd < 0)
3662 goto err_empty;
3663 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3664 close(fd);
3665 s->should_munmap = 1;
3666
3667 /*
3668 * Convert from working tree format to canonical git format
3669 */
3670 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, crlf_warn)) {
3671 size_t size = 0;
3672 munmap(s->data, s->size);
3673 s->should_munmap = 0;
3674 s->data = strbuf_detach(&buf, &size);
3675 s->size = size;
3676 s->should_free = 1;
3677 }
3678 }
3679 else {
3680 enum object_type type;
3681 if (size_only || (flags & CHECK_BINARY)) {
3682 type = sha1_object_info(s->oid.hash, &s->size);
3683 if (type < 0)
3684 die("unable to read %s",
3685 oid_to_hex(&s->oid));
3686 if (size_only)
3687 return 0;
3688 if (s->size > big_file_threshold && s->is_binary == -1) {
3689 s->is_binary = 1;
3690 return 0;
3691 }
3692 }
3693 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
3694 if (!s->data)
3695 die("unable to read %s", oid_to_hex(&s->oid));
3696 s->should_free = 1;
3697 }
3698 return 0;
3699 }
3700
3701 void diff_free_filespec_blob(struct diff_filespec *s)
3702 {
3703 if (s->should_free)
3704 free(s->data);
3705 else if (s->should_munmap)
3706 munmap(s->data, s->size);
3707
3708 if (s->should_free || s->should_munmap) {
3709 s->should_free = s->should_munmap = 0;
3710 s->data = NULL;
3711 }
3712 }
3713
3714 void diff_free_filespec_data(struct diff_filespec *s)
3715 {
3716 diff_free_filespec_blob(s);
3717 FREE_AND_NULL(s->cnt_data);
3718 }
3719
3720 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3721 void *blob,
3722 unsigned long size,
3723 const struct object_id *oid,
3724 int mode)
3725 {
3726 struct strbuf buf = STRBUF_INIT;
3727 struct strbuf template = STRBUF_INIT;
3728 char *path_dup = xstrdup(path);
3729 const char *base = basename(path_dup);
3730
3731 /* Generate "XXXXXX_basename.ext" */
3732 strbuf_addstr(&template, "XXXXXX_");
3733 strbuf_addstr(&template, base);
3734
3735 temp->tempfile = mks_tempfile_ts(template.buf, strlen(base) + 1);
3736 if (!temp->tempfile)
3737 die_errno("unable to create temp-file");
3738 if (convert_to_working_tree(path,
3739 (const char *)blob, (size_t)size, &buf)) {
3740 blob = buf.buf;
3741 size = buf.len;
3742 }
3743 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
3744 close_tempfile_gently(temp->tempfile))
3745 die_errno("unable to write temp-file");
3746 temp->name = get_tempfile_path(temp->tempfile);
3747 oid_to_hex_r(temp->hex, oid);
3748 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3749 strbuf_release(&buf);
3750 strbuf_release(&template);
3751 free(path_dup);
3752 }
3753
3754 static struct diff_tempfile *prepare_temp_file(const char *name,
3755 struct diff_filespec *one)
3756 {
3757 struct diff_tempfile *temp = claim_diff_tempfile();
3758
3759 if (!DIFF_FILE_VALID(one)) {
3760 not_a_valid_file:
3761 /* A '-' entry produces this for file-2, and
3762 * a '+' entry produces this for file-1.
3763 */
3764 temp->name = "/dev/null";
3765 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3766 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3767 return temp;
3768 }
3769
3770 if (!S_ISGITLINK(one->mode) &&
3771 (!one->oid_valid ||
3772 reuse_worktree_file(name, &one->oid, 1))) {
3773 struct stat st;
3774 if (lstat(name, &st) < 0) {
3775 if (errno == ENOENT)
3776 goto not_a_valid_file;
3777 die_errno("stat(%s)", name);
3778 }
3779 if (S_ISLNK(st.st_mode)) {
3780 struct strbuf sb = STRBUF_INIT;
3781 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3782 die_errno("readlink(%s)", name);
3783 prep_temp_blob(name, temp, sb.buf, sb.len,
3784 (one->oid_valid ?
3785 &one->oid : &null_oid),
3786 (one->oid_valid ?
3787 one->mode : S_IFLNK));
3788 strbuf_release(&sb);
3789 }
3790 else {
3791 /* we can borrow from the file in the work tree */
3792 temp->name = name;
3793 if (!one->oid_valid)
3794 oid_to_hex_r(temp->hex, &null_oid);
3795 else
3796 oid_to_hex_r(temp->hex, &one->oid);
3797 /* Even though we may sometimes borrow the
3798 * contents from the work tree, we always want
3799 * one->mode. mode is trustworthy even when
3800 * !(one->oid_valid), as long as
3801 * DIFF_FILE_VALID(one).
3802 */
3803 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3804 }
3805 return temp;
3806 }
3807 else {
3808 if (diff_populate_filespec(one, 0))
3809 die("cannot read data blob for %s", one->path);
3810 prep_temp_blob(name, temp, one->data, one->size,
3811 &one->oid, one->mode);
3812 }
3813 return temp;
3814 }
3815
3816 static void add_external_diff_name(struct argv_array *argv,
3817 const char *name,
3818 struct diff_filespec *df)
3819 {
3820 struct diff_tempfile *temp = prepare_temp_file(name, df);
3821 argv_array_push(argv, temp->name);
3822 argv_array_push(argv, temp->hex);
3823 argv_array_push(argv, temp->mode);
3824 }
3825
3826 /* An external diff command takes:
3827 *
3828 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3829 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3830 *
3831 */
3832 static void run_external_diff(const char *pgm,
3833 const char *name,
3834 const char *other,
3835 struct diff_filespec *one,
3836 struct diff_filespec *two,
3837 const char *xfrm_msg,
3838 int complete_rewrite,
3839 struct diff_options *o)
3840 {
3841 struct argv_array argv = ARGV_ARRAY_INIT;
3842 struct argv_array env = ARGV_ARRAY_INIT;
3843 struct diff_queue_struct *q = &diff_queued_diff;
3844
3845 argv_array_push(&argv, pgm);
3846 argv_array_push(&argv, name);
3847
3848 if (one && two) {
3849 add_external_diff_name(&argv, name, one);
3850 if (!other)
3851 add_external_diff_name(&argv, name, two);
3852 else {
3853 add_external_diff_name(&argv, other, two);
3854 argv_array_push(&argv, other);
3855 argv_array_push(&argv, xfrm_msg);
3856 }
3857 }
3858
3859 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3860 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3861
3862 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3863 die(_("external diff died, stopping at %s"), name);
3864
3865 remove_tempfile();
3866 argv_array_clear(&argv);
3867 argv_array_clear(&env);
3868 }
3869
3870 static int similarity_index(struct diff_filepair *p)
3871 {
3872 return p->score * 100 / MAX_SCORE;
3873 }
3874
3875 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3876 {
3877 if (startup_info->have_repository)
3878 return find_unique_abbrev(oid->hash, abbrev);
3879 else {
3880 char *hex = oid_to_hex(oid);
3881 if (abbrev < 0)
3882 abbrev = FALLBACK_DEFAULT_ABBREV;
3883 if (abbrev > GIT_SHA1_HEXSZ)
3884 die("BUG: oid abbreviation out of range: %d", abbrev);
3885 if (abbrev)
3886 hex[abbrev] = '\0';
3887 return hex;
3888 }
3889 }
3890
3891 static void fill_metainfo(struct strbuf *msg,
3892 const char *name,
3893 const char *other,
3894 struct diff_filespec *one,
3895 struct diff_filespec *two,
3896 struct diff_options *o,
3897 struct diff_filepair *p,
3898 int *must_show_header,
3899 int use_color)
3900 {
3901 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3902 const char *reset = diff_get_color(use_color, DIFF_RESET);
3903 const char *line_prefix = diff_line_prefix(o);
3904
3905 *must_show_header = 1;
3906 strbuf_init(msg, PATH_MAX * 2 + 300);
3907 switch (p->status) {
3908 case DIFF_STATUS_COPIED:
3909 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3910 line_prefix, set, similarity_index(p));
3911 strbuf_addf(msg, "%s\n%s%scopy from ",
3912 reset, line_prefix, set);
3913 quote_c_style(name, msg, NULL, 0);
3914 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3915 quote_c_style(other, msg, NULL, 0);
3916 strbuf_addf(msg, "%s\n", reset);
3917 break;
3918 case DIFF_STATUS_RENAMED:
3919 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3920 line_prefix, set, similarity_index(p));
3921 strbuf_addf(msg, "%s\n%s%srename from ",
3922 reset, line_prefix, set);
3923 quote_c_style(name, msg, NULL, 0);
3924 strbuf_addf(msg, "%s\n%s%srename to ",
3925 reset, line_prefix, set);
3926 quote_c_style(other, msg, NULL, 0);
3927 strbuf_addf(msg, "%s\n", reset);
3928 break;
3929 case DIFF_STATUS_MODIFIED:
3930 if (p->score) {
3931 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3932 line_prefix,
3933 set, similarity_index(p), reset);
3934 break;
3935 }
3936 /* fallthru */
3937 default:
3938 *must_show_header = 0;
3939 }
3940 if (one && two && oidcmp(&one->oid, &two->oid)) {
3941 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3942
3943 if (DIFF_OPT_TST(o, BINARY)) {
3944 mmfile_t mf;
3945 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3946 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3947 abbrev = 40;
3948 }
3949 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3950 diff_abbrev_oid(&one->oid, abbrev),
3951 diff_abbrev_oid(&two->oid, abbrev));
3952 if (one->mode == two->mode)
3953 strbuf_addf(msg, " %06o", one->mode);
3954 strbuf_addf(msg, "%s\n", reset);
3955 }
3956 }
3957
3958 static void run_diff_cmd(const char *pgm,
3959 const char *name,
3960 const char *other,
3961 const char *attr_path,
3962 struct diff_filespec *one,
3963 struct diff_filespec *two,
3964 struct strbuf *msg,
3965 struct diff_options *o,
3966 struct diff_filepair *p)
3967 {
3968 const char *xfrm_msg = NULL;
3969 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3970 int must_show_header = 0;
3971
3972
3973 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3974 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3975 if (drv && drv->external)
3976 pgm = drv->external;
3977 }
3978
3979 if (msg) {
3980 /*
3981 * don't use colors when the header is intended for an
3982 * external diff driver
3983 */
3984 fill_metainfo(msg, name, other, one, two, o, p,
3985 &must_show_header,
3986 want_color(o->use_color) && !pgm);
3987 xfrm_msg = msg->len ? msg->buf : NULL;
3988 }
3989
3990 if (pgm) {
3991 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3992 complete_rewrite, o);
3993 return;
3994 }
3995 if (one && two)
3996 builtin_diff(name, other ? other : name,
3997 one, two, xfrm_msg, must_show_header,
3998 o, complete_rewrite);
3999 else
4000 fprintf(o->file, "* Unmerged path %s\n", name);
4001 }
4002
4003 static void diff_fill_oid_info(struct diff_filespec *one)
4004 {
4005 if (DIFF_FILE_VALID(one)) {
4006 if (!one->oid_valid) {
4007 struct stat st;
4008 if (one->is_stdin) {
4009 oidclr(&one->oid);
4010 return;
4011 }
4012 if (lstat(one->path, &st) < 0)
4013 die_errno("stat '%s'", one->path);
4014 if (index_path(&one->oid, one->path, &st, 0))
4015 die("cannot hash %s", one->path);
4016 }
4017 }
4018 else
4019 oidclr(&one->oid);
4020 }
4021
4022 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4023 {
4024 /* Strip the prefix but do not molest /dev/null and absolute paths */
4025 if (*namep && **namep != '/') {
4026 *namep += prefix_length;
4027 if (**namep == '/')
4028 ++*namep;
4029 }
4030 if (*otherp && **otherp != '/') {
4031 *otherp += prefix_length;
4032 if (**otherp == '/')
4033 ++*otherp;
4034 }
4035 }
4036
4037 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4038 {
4039 const char *pgm = external_diff();
4040 struct strbuf msg;
4041 struct diff_filespec *one = p->one;
4042 struct diff_filespec *two = p->two;
4043 const char *name;
4044 const char *other;
4045 const char *attr_path;
4046
4047 name = one->path;
4048 other = (strcmp(name, two->path) ? two->path : NULL);
4049 attr_path = name;
4050 if (o->prefix_length)
4051 strip_prefix(o->prefix_length, &name, &other);
4052
4053 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
4054 pgm = NULL;
4055
4056 if (DIFF_PAIR_UNMERGED(p)) {
4057 run_diff_cmd(pgm, name, NULL, attr_path,
4058 NULL, NULL, NULL, o, p);
4059 return;
4060 }
4061
4062 diff_fill_oid_info(one);
4063 diff_fill_oid_info(two);
4064
4065 if (!pgm &&
4066 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4067 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4068 /*
4069 * a filepair that changes between file and symlink
4070 * needs to be split into deletion and creation.
4071 */
4072 struct diff_filespec *null = alloc_filespec(two->path);
4073 run_diff_cmd(NULL, name, other, attr_path,
4074 one, null, &msg, o, p);
4075 free(null);
4076 strbuf_release(&msg);
4077
4078 null = alloc_filespec(one->path);
4079 run_diff_cmd(NULL, name, other, attr_path,
4080 null, two, &msg, o, p);
4081 free(null);
4082 }
4083 else
4084 run_diff_cmd(pgm, name, other, attr_path,
4085 one, two, &msg, o, p);
4086
4087 strbuf_release(&msg);
4088 }
4089
4090 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4091 struct diffstat_t *diffstat)
4092 {
4093 const char *name;
4094 const char *other;
4095
4096 if (DIFF_PAIR_UNMERGED(p)) {
4097 /* unmerged */
4098 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4099 return;
4100 }
4101
4102 name = p->one->path;
4103 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4104
4105 if (o->prefix_length)
4106 strip_prefix(o->prefix_length, &name, &other);
4107
4108 diff_fill_oid_info(p->one);
4109 diff_fill_oid_info(p->two);
4110
4111 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4112 }
4113
4114 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4115 {
4116 const char *name;
4117 const char *other;
4118 const char *attr_path;
4119
4120 if (DIFF_PAIR_UNMERGED(p)) {
4121 /* unmerged */
4122 return;
4123 }
4124
4125 name = p->one->path;
4126 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4127 attr_path = other ? other : name;
4128
4129 if (o->prefix_length)
4130 strip_prefix(o->prefix_length, &name, &other);
4131
4132 diff_fill_oid_info(p->one);
4133 diff_fill_oid_info(p->two);
4134
4135 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4136 }
4137
4138 void diff_setup(struct diff_options *options)
4139 {
4140 memcpy(options, &default_diff_options, sizeof(*options));
4141
4142 options->file = stdout;
4143
4144 options->abbrev = DEFAULT_ABBREV;
4145 options->line_termination = '\n';
4146 options->break_opt = -1;
4147 options->rename_limit = -1;
4148 options->dirstat_permille = diff_dirstat_permille_default;
4149 options->context = diff_context_default;
4150 options->interhunkcontext = diff_interhunk_context_default;
4151 options->ws_error_highlight = ws_error_highlight_default;
4152 DIFF_OPT_SET(options, RENAME_EMPTY);
4153
4154 /* pathchange left =NULL by default */
4155 options->change = diff_change;
4156 options->add_remove = diff_addremove;
4157 options->use_color = diff_use_color_default;
4158 options->detect_rename = diff_detect_rename_default;
4159 options->xdl_opts |= diff_algorithm;
4160 if (diff_indent_heuristic)
4161 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4162
4163 options->orderfile = diff_order_file_cfg;
4164
4165 if (diff_no_prefix) {
4166 options->a_prefix = options->b_prefix = "";
4167 } else if (!diff_mnemonic_prefix) {
4168 options->a_prefix = "a/";
4169 options->b_prefix = "b/";
4170 }
4171
4172 options->color_moved = diff_color_moved_default;
4173 }
4174
4175 void diff_setup_done(struct diff_options *options)
4176 {
4177 int count = 0;
4178
4179 if (options->set_default)
4180 options->set_default(options);
4181
4182 if (options->output_format & DIFF_FORMAT_NAME)
4183 count++;
4184 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
4185 count++;
4186 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
4187 count++;
4188 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
4189 count++;
4190 if (count > 1)
4191 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4192
4193 /*
4194 * Most of the time we can say "there are changes"
4195 * only by checking if there are changed paths, but
4196 * --ignore-whitespace* options force us to look
4197 * inside contents.
4198 */
4199
4200 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
4201 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
4202 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
4203 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
4204 else
4205 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
4206
4207 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
4208 options->detect_rename = DIFF_DETECT_COPY;
4209
4210 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
4211 options->prefix = NULL;
4212 if (options->prefix)
4213 options->prefix_length = strlen(options->prefix);
4214 else
4215 options->prefix_length = 0;
4216
4217 if (options->output_format & (DIFF_FORMAT_NAME |
4218 DIFF_FORMAT_NAME_STATUS |
4219 DIFF_FORMAT_CHECKDIFF |
4220 DIFF_FORMAT_NO_OUTPUT))
4221 options->output_format &= ~(DIFF_FORMAT_RAW |
4222 DIFF_FORMAT_NUMSTAT |
4223 DIFF_FORMAT_DIFFSTAT |
4224 DIFF_FORMAT_SHORTSTAT |
4225 DIFF_FORMAT_DIRSTAT |
4226 DIFF_FORMAT_SUMMARY |
4227 DIFF_FORMAT_PATCH);
4228
4229 /*
4230 * These cases always need recursive; we do not drop caller-supplied
4231 * recursive bits for other formats here.
4232 */
4233 if (options->output_format & (DIFF_FORMAT_PATCH |
4234 DIFF_FORMAT_NUMSTAT |
4235 DIFF_FORMAT_DIFFSTAT |
4236 DIFF_FORMAT_SHORTSTAT |
4237 DIFF_FORMAT_DIRSTAT |
4238 DIFF_FORMAT_SUMMARY |
4239 DIFF_FORMAT_CHECKDIFF))
4240 DIFF_OPT_SET(options, RECURSIVE);
4241 /*
4242 * Also pickaxe would not work very well if you do not say recursive
4243 */
4244 if (options->pickaxe)
4245 DIFF_OPT_SET(options, RECURSIVE);
4246 /*
4247 * When patches are generated, submodules diffed against the work tree
4248 * must be checked for dirtiness too so it can be shown in the output
4249 */
4250 if (options->output_format & DIFF_FORMAT_PATCH)
4251 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
4252
4253 if (options->detect_rename && options->rename_limit < 0)
4254 options->rename_limit = diff_rename_limit_default;
4255 if (options->setup & DIFF_SETUP_USE_CACHE) {
4256 if (!active_cache)
4257 /* read-cache does not die even when it fails
4258 * so it is safe for us to do this here. Also
4259 * it does not smudge active_cache or active_nr
4260 * when it fails, so we do not have to worry about
4261 * cleaning it up ourselves either.
4262 */
4263 read_cache();
4264 }
4265 if (40 < options->abbrev)
4266 options->abbrev = 40; /* full */
4267
4268 /*
4269 * It does not make sense to show the first hit we happened
4270 * to have found. It does not make sense not to return with
4271 * exit code in such a case either.
4272 */
4273 if (DIFF_OPT_TST(options, QUICK)) {
4274 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4275 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4276 }
4277
4278 options->diff_path_counter = 0;
4279
4280 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
4281 die(_("--follow requires exactly one pathspec"));
4282
4283 if (!options->use_color || external_diff())
4284 options->color_moved = 0;
4285 }
4286
4287 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4288 {
4289 char c, *eq;
4290 int len;
4291
4292 if (*arg != '-')
4293 return 0;
4294 c = *++arg;
4295 if (!c)
4296 return 0;
4297 if (c == arg_short) {
4298 c = *++arg;
4299 if (!c)
4300 return 1;
4301 if (val && isdigit(c)) {
4302 char *end;
4303 int n = strtoul(arg, &end, 10);
4304 if (*end)
4305 return 0;
4306 *val = n;
4307 return 1;
4308 }
4309 return 0;
4310 }
4311 if (c != '-')
4312 return 0;
4313 arg++;
4314 eq = strchrnul(arg, '=');
4315 len = eq - arg;
4316 if (!len || strncmp(arg, arg_long, len))
4317 return 0;
4318 if (*eq) {
4319 int n;
4320 char *end;
4321 if (!isdigit(*++eq))
4322 return 0;
4323 n = strtoul(eq, &end, 10);
4324 if (*end)
4325 return 0;
4326 *val = n;
4327 }
4328 return 1;
4329 }
4330
4331 static int diff_scoreopt_parse(const char *opt);
4332
4333 static inline int short_opt(char opt, const char **argv,
4334 const char **optarg)
4335 {
4336 const char *arg = argv[0];
4337 if (arg[0] != '-' || arg[1] != opt)
4338 return 0;
4339 if (arg[2] != '\0') {
4340 *optarg = arg + 2;
4341 return 1;
4342 }
4343 if (!argv[1])
4344 die("Option '%c' requires a value", opt);
4345 *optarg = argv[1];
4346 return 2;
4347 }
4348
4349 int parse_long_opt(const char *opt, const char **argv,
4350 const char **optarg)
4351 {
4352 const char *arg = argv[0];
4353 if (!skip_prefix(arg, "--", &arg))
4354 return 0;
4355 if (!skip_prefix(arg, opt, &arg))
4356 return 0;
4357 if (*arg == '=') { /* stuck form: --option=value */
4358 *optarg = arg + 1;
4359 return 1;
4360 }
4361 if (*arg != '\0')
4362 return 0;
4363 /* separate form: --option value */
4364 if (!argv[1])
4365 die("Option '--%s' requires a value", opt);
4366 *optarg = argv[1];
4367 return 2;
4368 }
4369
4370 static int stat_opt(struct diff_options *options, const char **av)
4371 {
4372 const char *arg = av[0];
4373 char *end;
4374 int width = options->stat_width;
4375 int name_width = options->stat_name_width;
4376 int graph_width = options->stat_graph_width;
4377 int count = options->stat_count;
4378 int argcount = 1;
4379
4380 if (!skip_prefix(arg, "--stat", &arg))
4381 die("BUG: stat option does not begin with --stat: %s", arg);
4382 end = (char *)arg;
4383
4384 switch (*arg) {
4385 case '-':
4386 if (skip_prefix(arg, "-width", &arg)) {
4387 if (*arg == '=')
4388 width = strtoul(arg + 1, &end, 10);
4389 else if (!*arg && !av[1])
4390 die_want_option("--stat-width");
4391 else if (!*arg) {
4392 width = strtoul(av[1], &end, 10);
4393 argcount = 2;
4394 }
4395 } else if (skip_prefix(arg, "-name-width", &arg)) {
4396 if (*arg == '=')
4397 name_width = strtoul(arg + 1, &end, 10);
4398 else if (!*arg && !av[1])
4399 die_want_option("--stat-name-width");
4400 else if (!*arg) {
4401 name_width = strtoul(av[1], &end, 10);
4402 argcount = 2;
4403 }
4404 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4405 if (*arg == '=')
4406 graph_width = strtoul(arg + 1, &end, 10);
4407 else if (!*arg && !av[1])
4408 die_want_option("--stat-graph-width");
4409 else if (!*arg) {
4410 graph_width = strtoul(av[1], &end, 10);
4411 argcount = 2;
4412 }
4413 } else if (skip_prefix(arg, "-count", &arg)) {
4414 if (*arg == '=')
4415 count = strtoul(arg + 1, &end, 10);
4416 else if (!*arg && !av[1])
4417 die_want_option("--stat-count");
4418 else if (!*arg) {
4419 count = strtoul(av[1], &end, 10);
4420 argcount = 2;
4421 }
4422 }
4423 break;
4424 case '=':
4425 width = strtoul(arg+1, &end, 10);
4426 if (*end == ',')
4427 name_width = strtoul(end+1, &end, 10);
4428 if (*end == ',')
4429 count = strtoul(end+1, &end, 10);
4430 }
4431
4432 /* Important! This checks all the error cases! */
4433 if (*end)
4434 return 0;
4435 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4436 options->stat_name_width = name_width;
4437 options->stat_graph_width = graph_width;
4438 options->stat_width = width;
4439 options->stat_count = count;
4440 return argcount;
4441 }
4442
4443 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4444 {
4445 struct strbuf errmsg = STRBUF_INIT;
4446 if (parse_dirstat_params(options, params, &errmsg))
4447 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4448 errmsg.buf);
4449 strbuf_release(&errmsg);
4450 /*
4451 * The caller knows a dirstat-related option is given from the command
4452 * line; allow it to say "return this_function();"
4453 */
4454 options->output_format |= DIFF_FORMAT_DIRSTAT;
4455 return 1;
4456 }
4457
4458 static int parse_submodule_opt(struct diff_options *options, const char *value)
4459 {
4460 if (parse_submodule_params(options, value))
4461 die(_("Failed to parse --submodule option parameter: '%s'"),
4462 value);
4463 return 1;
4464 }
4465
4466 static const char diff_status_letters[] = {
4467 DIFF_STATUS_ADDED,
4468 DIFF_STATUS_COPIED,
4469 DIFF_STATUS_DELETED,
4470 DIFF_STATUS_MODIFIED,
4471 DIFF_STATUS_RENAMED,
4472 DIFF_STATUS_TYPE_CHANGED,
4473 DIFF_STATUS_UNKNOWN,
4474 DIFF_STATUS_UNMERGED,
4475 DIFF_STATUS_FILTER_AON,
4476 DIFF_STATUS_FILTER_BROKEN,
4477 '\0',
4478 };
4479
4480 static unsigned int filter_bit['Z' + 1];
4481
4482 static void prepare_filter_bits(void)
4483 {
4484 int i;
4485
4486 if (!filter_bit[DIFF_STATUS_ADDED]) {
4487 for (i = 0; diff_status_letters[i]; i++)
4488 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4489 }
4490 }
4491
4492 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4493 {
4494 return opt->filter & filter_bit[(int) status];
4495 }
4496
4497 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4498 {
4499 int i, optch;
4500
4501 prepare_filter_bits();
4502
4503 /*
4504 * If there is a negation e.g. 'd' in the input, and we haven't
4505 * initialized the filter field with another --diff-filter, start
4506 * from full set of bits, except for AON.
4507 */
4508 if (!opt->filter) {
4509 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4510 if (optch < 'a' || 'z' < optch)
4511 continue;
4512 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4513 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4514 break;
4515 }
4516 }
4517
4518 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4519 unsigned int bit;
4520 int negate;
4521
4522 if ('a' <= optch && optch <= 'z') {
4523 negate = 1;
4524 optch = toupper(optch);
4525 } else {
4526 negate = 0;
4527 }
4528
4529 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4530 if (!bit)
4531 return optarg[i];
4532 if (negate)
4533 opt->filter &= ~bit;
4534 else
4535 opt->filter |= bit;
4536 }
4537 return 0;
4538 }
4539
4540 static void enable_patch_output(int *fmt) {
4541 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4542 *fmt |= DIFF_FORMAT_PATCH;
4543 }
4544
4545 static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4546 {
4547 int val = parse_ws_error_highlight(arg);
4548
4549 if (val < 0) {
4550 error("unknown value after ws-error-highlight=%.*s",
4551 -1 - val, arg);
4552 return 0;
4553 }
4554 opt->ws_error_highlight = val;
4555 return 1;
4556 }
4557
4558 int diff_opt_parse(struct diff_options *options,
4559 const char **av, int ac, const char *prefix)
4560 {
4561 const char *arg = av[0];
4562 const char *optarg;
4563 int argcount;
4564
4565 if (!prefix)
4566 prefix = "";
4567
4568 /* Output format options */
4569 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4570 || opt_arg(arg, 'U', "unified", &options->context))
4571 enable_patch_output(&options->output_format);
4572 else if (!strcmp(arg, "--raw"))
4573 options->output_format |= DIFF_FORMAT_RAW;
4574 else if (!strcmp(arg, "--patch-with-raw")) {
4575 enable_patch_output(&options->output_format);
4576 options->output_format |= DIFF_FORMAT_RAW;
4577 } else if (!strcmp(arg, "--numstat"))
4578 options->output_format |= DIFF_FORMAT_NUMSTAT;
4579 else if (!strcmp(arg, "--shortstat"))
4580 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4581 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
4582 return parse_dirstat_opt(options, "");
4583 else if (skip_prefix(arg, "-X", &arg))
4584 return parse_dirstat_opt(options, arg);
4585 else if (skip_prefix(arg, "--dirstat=", &arg))
4586 return parse_dirstat_opt(options, arg);
4587 else if (!strcmp(arg, "--cumulative"))
4588 return parse_dirstat_opt(options, "cumulative");
4589 else if (!strcmp(arg, "--dirstat-by-file"))
4590 return parse_dirstat_opt(options, "files");
4591 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
4592 parse_dirstat_opt(options, "files");
4593 return parse_dirstat_opt(options, arg);
4594 }
4595 else if (!strcmp(arg, "--check"))
4596 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4597 else if (!strcmp(arg, "--summary"))
4598 options->output_format |= DIFF_FORMAT_SUMMARY;
4599 else if (!strcmp(arg, "--patch-with-stat")) {
4600 enable_patch_output(&options->output_format);
4601 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4602 } else if (!strcmp(arg, "--name-only"))
4603 options->output_format |= DIFF_FORMAT_NAME;
4604 else if (!strcmp(arg, "--name-status"))
4605 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4606 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4607 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4608 else if (starts_with(arg, "--stat"))
4609 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4610 return stat_opt(options, av);
4611
4612 /* renames options */
4613 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
4614 !strcmp(arg, "--break-rewrites")) {
4615 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4616 return error("invalid argument to -B: %s", arg+2);
4617 }
4618 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
4619 !strcmp(arg, "--find-renames")) {
4620 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4621 return error("invalid argument to -M: %s", arg+2);
4622 options->detect_rename = DIFF_DETECT_RENAME;
4623 }
4624 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4625 options->irreversible_delete = 1;
4626 }
4627 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
4628 !strcmp(arg, "--find-copies")) {
4629 if (options->detect_rename == DIFF_DETECT_COPY)
4630 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4631 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4632 return error("invalid argument to -C: %s", arg+2);
4633 options->detect_rename = DIFF_DETECT_COPY;
4634 }
4635 else if (!strcmp(arg, "--no-renames"))
4636 options->detect_rename = 0;
4637 else if (!strcmp(arg, "--rename-empty"))
4638 DIFF_OPT_SET(options, RENAME_EMPTY);
4639 else if (!strcmp(arg, "--no-rename-empty"))
4640 DIFF_OPT_CLR(options, RENAME_EMPTY);
4641 else if (!strcmp(arg, "--relative"))
4642 DIFF_OPT_SET(options, RELATIVE_NAME);
4643 else if (skip_prefix(arg, "--relative=", &arg)) {
4644 DIFF_OPT_SET(options, RELATIVE_NAME);
4645 options->prefix = arg;
4646 }
4647
4648 /* xdiff options */
4649 else if (!strcmp(arg, "--minimal"))
4650 DIFF_XDL_SET(options, NEED_MINIMAL);
4651 else if (!strcmp(arg, "--no-minimal"))
4652 DIFF_XDL_CLR(options, NEED_MINIMAL);
4653 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4654 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4655 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4656 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4657 else if (!strcmp(arg, "--ignore-space-at-eol"))
4658 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4659 else if (!strcmp(arg, "--ignore-blank-lines"))
4660 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4661 else if (!strcmp(arg, "--indent-heuristic"))
4662 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4663 else if (!strcmp(arg, "--no-indent-heuristic"))
4664 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4665 else if (!strcmp(arg, "--patience"))
4666 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4667 else if (!strcmp(arg, "--histogram"))
4668 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4669 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4670 long value = parse_algorithm_value(optarg);
4671 if (value < 0)
4672 return error("option diff-algorithm accepts \"myers\", "
4673 "\"minimal\", \"patience\" and \"histogram\"");
4674 /* clear out previous settings */
4675 DIFF_XDL_CLR(options, NEED_MINIMAL);
4676 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4677 options->xdl_opts |= value;
4678 return argcount;
4679 }
4680
4681 /* flags options */
4682 else if (!strcmp(arg, "--binary")) {
4683 enable_patch_output(&options->output_format);
4684 DIFF_OPT_SET(options, BINARY);
4685 }
4686 else if (!strcmp(arg, "--full-index"))
4687 DIFF_OPT_SET(options, FULL_INDEX);
4688 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4689 DIFF_OPT_SET(options, TEXT);
4690 else if (!strcmp(arg, "-R"))
4691 DIFF_OPT_SET(options, REVERSE_DIFF);
4692 else if (!strcmp(arg, "--find-copies-harder"))
4693 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4694 else if (!strcmp(arg, "--follow"))
4695 DIFF_OPT_SET(options, FOLLOW_RENAMES);
4696 else if (!strcmp(arg, "--no-follow")) {
4697 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
4698 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
4699 } else if (!strcmp(arg, "--color"))
4700 options->use_color = 1;
4701 else if (skip_prefix(arg, "--color=", &arg)) {
4702 int value = git_config_colorbool(NULL, arg);
4703 if (value < 0)
4704 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4705 options->use_color = value;
4706 }
4707 else if (!strcmp(arg, "--no-color"))
4708 options->use_color = 0;
4709 else if (!strcmp(arg, "--color-moved")) {
4710 if (diff_color_moved_default)
4711 options->color_moved = diff_color_moved_default;
4712 if (options->color_moved == COLOR_MOVED_NO)
4713 options->color_moved = COLOR_MOVED_DEFAULT;
4714 } else if (!strcmp(arg, "--no-color-moved"))
4715 options->color_moved = COLOR_MOVED_NO;
4716 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4717 int cm = parse_color_moved(arg);
4718 if (cm < 0)
4719 die("bad --color-moved argument: %s", arg);
4720 options->color_moved = cm;
4721 } else if (!strcmp(arg, "--color-words")) {
4722 options->use_color = 1;
4723 options->word_diff = DIFF_WORDS_COLOR;
4724 }
4725 else if (skip_prefix(arg, "--color-words=", &arg)) {
4726 options->use_color = 1;
4727 options->word_diff = DIFF_WORDS_COLOR;
4728 options->word_regex = arg;
4729 }
4730 else if (!strcmp(arg, "--word-diff")) {
4731 if (options->word_diff == DIFF_WORDS_NONE)
4732 options->word_diff = DIFF_WORDS_PLAIN;
4733 }
4734 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4735 if (!strcmp(arg, "plain"))
4736 options->word_diff = DIFF_WORDS_PLAIN;
4737 else if (!strcmp(arg, "color")) {
4738 options->use_color = 1;
4739 options->word_diff = DIFF_WORDS_COLOR;
4740 }
4741 else if (!strcmp(arg, "porcelain"))
4742 options->word_diff = DIFF_WORDS_PORCELAIN;
4743 else if (!strcmp(arg, "none"))
4744 options->word_diff = DIFF_WORDS_NONE;
4745 else
4746 die("bad --word-diff argument: %s", arg);
4747 }
4748 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
4749 if (options->word_diff == DIFF_WORDS_NONE)
4750 options->word_diff = DIFF_WORDS_PLAIN;
4751 options->word_regex = optarg;
4752 return argcount;
4753 }
4754 else if (!strcmp(arg, "--exit-code"))
4755 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4756 else if (!strcmp(arg, "--quiet"))
4757 DIFF_OPT_SET(options, QUICK);
4758 else if (!strcmp(arg, "--ext-diff"))
4759 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
4760 else if (!strcmp(arg, "--no-ext-diff"))
4761 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
4762 else if (!strcmp(arg, "--textconv"))
4763 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
4764 else if (!strcmp(arg, "--no-textconv"))
4765 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
4766 else if (!strcmp(arg, "--ignore-submodules")) {
4767 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4768 handle_ignore_submodules_arg(options, "all");
4769 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
4770 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4771 handle_ignore_submodules_arg(options, arg);
4772 } else if (!strcmp(arg, "--submodule"))
4773 options->submodule_format = DIFF_SUBMODULE_LOG;
4774 else if (skip_prefix(arg, "--submodule=", &arg))
4775 return parse_submodule_opt(options, arg);
4776 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
4777 return parse_ws_error_highlight_opt(options, arg);
4778 else if (!strcmp(arg, "--ita-invisible-in-index"))
4779 options->ita_invisible_in_index = 1;
4780 else if (!strcmp(arg, "--ita-visible-in-index"))
4781 options->ita_invisible_in_index = 0;
4782
4783 /* misc options */
4784 else if (!strcmp(arg, "-z"))
4785 options->line_termination = 0;
4786 else if ((argcount = short_opt('l', av, &optarg))) {
4787 options->rename_limit = strtoul(optarg, NULL, 10);
4788 return argcount;
4789 }
4790 else if ((argcount = short_opt('S', av, &optarg))) {
4791 options->pickaxe = optarg;
4792 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4793 return argcount;
4794 } else if ((argcount = short_opt('G', av, &optarg))) {
4795 options->pickaxe = optarg;
4796 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4797 return argcount;
4798 }
4799 else if (!strcmp(arg, "--pickaxe-all"))
4800 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4801 else if (!strcmp(arg, "--pickaxe-regex"))
4802 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4803 else if ((argcount = short_opt('O', av, &optarg))) {
4804 options->orderfile = prefix_filename(prefix, optarg);
4805 return argcount;
4806 }
4807 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4808 int offending = parse_diff_filter_opt(optarg, options);
4809 if (offending)
4810 die("unknown change class '%c' in --diff-filter=%s",
4811 offending, optarg);
4812 return argcount;
4813 }
4814 else if (!strcmp(arg, "--no-abbrev"))
4815 options->abbrev = 0;
4816 else if (!strcmp(arg, "--abbrev"))
4817 options->abbrev = DEFAULT_ABBREV;
4818 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4819 options->abbrev = strtoul(arg, NULL, 10);
4820 if (options->abbrev < MINIMUM_ABBREV)
4821 options->abbrev = MINIMUM_ABBREV;
4822 else if (40 < options->abbrev)
4823 options->abbrev = 40;
4824 }
4825 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4826 options->a_prefix = optarg;
4827 return argcount;
4828 }
4829 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4830 options->line_prefix = optarg;
4831 options->line_prefix_length = strlen(options->line_prefix);
4832 graph_setup_line_prefix(options);
4833 return argcount;
4834 }
4835 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4836 options->b_prefix = optarg;
4837 return argcount;
4838 }
4839 else if (!strcmp(arg, "--no-prefix"))
4840 options->a_prefix = options->b_prefix = "";
4841 else if (opt_arg(arg, '\0', "inter-hunk-context",
4842 &options->interhunkcontext))
4843 ;
4844 else if (!strcmp(arg, "-W"))
4845 DIFF_OPT_SET(options, FUNCCONTEXT);
4846 else if (!strcmp(arg, "--function-context"))
4847 DIFF_OPT_SET(options, FUNCCONTEXT);
4848 else if (!strcmp(arg, "--no-function-context"))
4849 DIFF_OPT_CLR(options, FUNCCONTEXT);
4850 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4851 char *path = prefix_filename(prefix, optarg);
4852 options->file = xfopen(path, "w");
4853 options->close_file = 1;
4854 if (options->use_color != GIT_COLOR_ALWAYS)
4855 options->use_color = GIT_COLOR_NEVER;
4856 free(path);
4857 return argcount;
4858 } else
4859 return 0;
4860 return 1;
4861 }
4862
4863 int parse_rename_score(const char **cp_p)
4864 {
4865 unsigned long num, scale;
4866 int ch, dot;
4867 const char *cp = *cp_p;
4868
4869 num = 0;
4870 scale = 1;
4871 dot = 0;
4872 for (;;) {
4873 ch = *cp;
4874 if ( !dot && ch == '.' ) {
4875 scale = 1;
4876 dot = 1;
4877 } else if ( ch == '%' ) {
4878 scale = dot ? scale*100 : 100;
4879 cp++; /* % is always at the end */
4880 break;
4881 } else if ( ch >= '0' && ch <= '9' ) {
4882 if ( scale < 100000 ) {
4883 scale *= 10;
4884 num = (num*10) + (ch-'0');
4885 }
4886 } else {
4887 break;
4888 }
4889 cp++;
4890 }
4891 *cp_p = cp;
4892
4893 /* user says num divided by scale and we say internally that
4894 * is MAX_SCORE * num / scale.
4895 */
4896 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4897 }
4898
4899 static int diff_scoreopt_parse(const char *opt)
4900 {
4901 int opt1, opt2, cmd;
4902
4903 if (*opt++ != '-')
4904 return -1;
4905 cmd = *opt++;
4906 if (cmd == '-') {
4907 /* convert the long-form arguments into short-form versions */
4908 if (skip_prefix(opt, "break-rewrites", &opt)) {
4909 if (*opt == 0 || *opt++ == '=')
4910 cmd = 'B';
4911 } else if (skip_prefix(opt, "find-copies", &opt)) {
4912 if (*opt == 0 || *opt++ == '=')
4913 cmd = 'C';
4914 } else if (skip_prefix(opt, "find-renames", &opt)) {
4915 if (*opt == 0 || *opt++ == '=')
4916 cmd = 'M';
4917 }
4918 }
4919 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4920 return -1; /* that is not a -M, -C, or -B option */
4921
4922 opt1 = parse_rename_score(&opt);
4923 if (cmd != 'B')
4924 opt2 = 0;
4925 else {
4926 if (*opt == 0)
4927 opt2 = 0;
4928 else if (*opt != '/')
4929 return -1; /* we expect -B80/99 or -B80 */
4930 else {
4931 opt++;
4932 opt2 = parse_rename_score(&opt);
4933 }
4934 }
4935 if (*opt != 0)
4936 return -1;
4937 return opt1 | (opt2 << 16);
4938 }
4939
4940 struct diff_queue_struct diff_queued_diff;
4941
4942 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4943 {
4944 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4945 queue->queue[queue->nr++] = dp;
4946 }
4947
4948 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4949 struct diff_filespec *one,
4950 struct diff_filespec *two)
4951 {
4952 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4953 dp->one = one;
4954 dp->two = two;
4955 if (queue)
4956 diff_q(queue, dp);
4957 return dp;
4958 }
4959
4960 void diff_free_filepair(struct diff_filepair *p)
4961 {
4962 free_filespec(p->one);
4963 free_filespec(p->two);
4964 free(p);
4965 }
4966
4967 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4968 {
4969 int abblen;
4970 const char *abbrev;
4971
4972 if (len == GIT_SHA1_HEXSZ)
4973 return oid_to_hex(oid);
4974
4975 abbrev = diff_abbrev_oid(oid, len);
4976 abblen = strlen(abbrev);
4977
4978 /*
4979 * In well-behaved cases, where the abbbreviated result is the
4980 * same as the requested length, append three dots after the
4981 * abbreviation (hence the whole logic is limited to the case
4982 * where abblen < 37); when the actual abbreviated result is a
4983 * bit longer than the requested length, we reduce the number
4984 * of dots so that they match the well-behaved ones. However,
4985 * if the actual abbreviation is longer than the requested
4986 * length by more than three, we give up on aligning, and add
4987 * three dots anyway, to indicate that the output is not the
4988 * full object name. Yes, this may be suboptimal, but this
4989 * appears only in "diff --raw --abbrev" output and it is not
4990 * worth the effort to change it now. Note that this would
4991 * likely to work fine when the automatic sizing of default
4992 * abbreviation length is used--we would be fed -1 in "len" in
4993 * that case, and will end up always appending three-dots, but
4994 * the automatic sizing is supposed to give abblen that ensures
4995 * uniqueness across all objects (statistically speaking).
4996 */
4997 if (abblen < GIT_SHA1_HEXSZ - 3) {
4998 static char hex[GIT_MAX_HEXSZ + 1];
4999 if (len < abblen && abblen <= len + 2)
5000 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5001 else
5002 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5003 return hex;
5004 }
5005
5006 return oid_to_hex(oid);
5007 }
5008
5009 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5010 {
5011 int line_termination = opt->line_termination;
5012 int inter_name_termination = line_termination ? '\t' : '\0';
5013
5014 fprintf(opt->file, "%s", diff_line_prefix(opt));
5015 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5016 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5017 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5018 fprintf(opt->file, "%s ",
5019 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5020 }
5021 if (p->score) {
5022 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5023 inter_name_termination);
5024 } else {
5025 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5026 }
5027
5028 if (p->status == DIFF_STATUS_COPIED ||
5029 p->status == DIFF_STATUS_RENAMED) {
5030 const char *name_a, *name_b;
5031 name_a = p->one->path;
5032 name_b = p->two->path;
5033 strip_prefix(opt->prefix_length, &name_a, &name_b);
5034 write_name_quoted(name_a, opt->file, inter_name_termination);
5035 write_name_quoted(name_b, opt->file, line_termination);
5036 } else {
5037 const char *name_a, *name_b;
5038 name_a = p->one->mode ? p->one->path : p->two->path;
5039 name_b = NULL;
5040 strip_prefix(opt->prefix_length, &name_a, &name_b);
5041 write_name_quoted(name_a, opt->file, line_termination);
5042 }
5043 }
5044
5045 int diff_unmodified_pair(struct diff_filepair *p)
5046 {
5047 /* This function is written stricter than necessary to support
5048 * the currently implemented transformers, but the idea is to
5049 * let transformers to produce diff_filepairs any way they want,
5050 * and filter and clean them up here before producing the output.
5051 */
5052 struct diff_filespec *one = p->one, *two = p->two;
5053
5054 if (DIFF_PAIR_UNMERGED(p))
5055 return 0; /* unmerged is interesting */
5056
5057 /* deletion, addition, mode or type change
5058 * and rename are all interesting.
5059 */
5060 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5061 DIFF_PAIR_MODE_CHANGED(p) ||
5062 strcmp(one->path, two->path))
5063 return 0;
5064
5065 /* both are valid and point at the same path. that is, we are
5066 * dealing with a change.
5067 */
5068 if (one->oid_valid && two->oid_valid &&
5069 !oidcmp(&one->oid, &two->oid) &&
5070 !one->dirty_submodule && !two->dirty_submodule)
5071 return 1; /* no change */
5072 if (!one->oid_valid && !two->oid_valid)
5073 return 1; /* both look at the same file on the filesystem. */
5074 return 0;
5075 }
5076
5077 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5078 {
5079 if (diff_unmodified_pair(p))
5080 return;
5081
5082 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5083 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5084 return; /* no tree diffs in patch format */
5085
5086 run_diff(p, o);
5087 }
5088
5089 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5090 struct diffstat_t *diffstat)
5091 {
5092 if (diff_unmodified_pair(p))
5093 return;
5094
5095 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5096 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5097 return; /* no useful stat for tree diffs */
5098
5099 run_diffstat(p, o, diffstat);
5100 }
5101
5102 static void diff_flush_checkdiff(struct diff_filepair *p,
5103 struct diff_options *o)
5104 {
5105 if (diff_unmodified_pair(p))
5106 return;
5107
5108 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5109 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5110 return; /* nothing to check in tree diffs */
5111
5112 run_checkdiff(p, o);
5113 }
5114
5115 int diff_queue_is_empty(void)
5116 {
5117 struct diff_queue_struct *q = &diff_queued_diff;
5118 int i;
5119 for (i = 0; i < q->nr; i++)
5120 if (!diff_unmodified_pair(q->queue[i]))
5121 return 0;
5122 return 1;
5123 }
5124
5125 #if DIFF_DEBUG
5126 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5127 {
5128 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5129 x, one ? one : "",
5130 s->path,
5131 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5132 s->mode,
5133 s->oid_valid ? oid_to_hex(&s->oid) : "");
5134 fprintf(stderr, "queue[%d] %s size %lu\n",
5135 x, one ? one : "",
5136 s->size);
5137 }
5138
5139 void diff_debug_filepair(const struct diff_filepair *p, int i)
5140 {
5141 diff_debug_filespec(p->one, i, "one");
5142 diff_debug_filespec(p->two, i, "two");
5143 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5144 p->score, p->status ? p->status : '?',
5145 p->one->rename_used, p->broken_pair);
5146 }
5147
5148 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5149 {
5150 int i;
5151 if (msg)
5152 fprintf(stderr, "%s\n", msg);
5153 fprintf(stderr, "q->nr = %d\n", q->nr);
5154 for (i = 0; i < q->nr; i++) {
5155 struct diff_filepair *p = q->queue[i];
5156 diff_debug_filepair(p, i);
5157 }
5158 }
5159 #endif
5160
5161 static void diff_resolve_rename_copy(void)
5162 {
5163 int i;
5164 struct diff_filepair *p;
5165 struct diff_queue_struct *q = &diff_queued_diff;
5166
5167 diff_debug_queue("resolve-rename-copy", q);
5168
5169 for (i = 0; i < q->nr; i++) {
5170 p = q->queue[i];
5171 p->status = 0; /* undecided */
5172 if (DIFF_PAIR_UNMERGED(p))
5173 p->status = DIFF_STATUS_UNMERGED;
5174 else if (!DIFF_FILE_VALID(p->one))
5175 p->status = DIFF_STATUS_ADDED;
5176 else if (!DIFF_FILE_VALID(p->two))
5177 p->status = DIFF_STATUS_DELETED;
5178 else if (DIFF_PAIR_TYPE_CHANGED(p))
5179 p->status = DIFF_STATUS_TYPE_CHANGED;
5180
5181 /* from this point on, we are dealing with a pair
5182 * whose both sides are valid and of the same type, i.e.
5183 * either in-place edit or rename/copy edit.
5184 */
5185 else if (DIFF_PAIR_RENAME(p)) {
5186 /*
5187 * A rename might have re-connected a broken
5188 * pair up, causing the pathnames to be the
5189 * same again. If so, that's not a rename at
5190 * all, just a modification..
5191 *
5192 * Otherwise, see if this source was used for
5193 * multiple renames, in which case we decrement
5194 * the count, and call it a copy.
5195 */
5196 if (!strcmp(p->one->path, p->two->path))
5197 p->status = DIFF_STATUS_MODIFIED;
5198 else if (--p->one->rename_used > 0)
5199 p->status = DIFF_STATUS_COPIED;
5200 else
5201 p->status = DIFF_STATUS_RENAMED;
5202 }
5203 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5204 p->one->mode != p->two->mode ||
5205 p->one->dirty_submodule ||
5206 p->two->dirty_submodule ||
5207 is_null_oid(&p->one->oid))
5208 p->status = DIFF_STATUS_MODIFIED;
5209 else {
5210 /* This is a "no-change" entry and should not
5211 * happen anymore, but prepare for broken callers.
5212 */
5213 error("feeding unmodified %s to diffcore",
5214 p->one->path);
5215 p->status = DIFF_STATUS_UNKNOWN;
5216 }
5217 }
5218 diff_debug_queue("resolve-rename-copy done", q);
5219 }
5220
5221 static int check_pair_status(struct diff_filepair *p)
5222 {
5223 switch (p->status) {
5224 case DIFF_STATUS_UNKNOWN:
5225 return 0;
5226 case 0:
5227 die("internal error in diff-resolve-rename-copy");
5228 default:
5229 return 1;
5230 }
5231 }
5232
5233 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5234 {
5235 int fmt = opt->output_format;
5236
5237 if (fmt & DIFF_FORMAT_CHECKDIFF)
5238 diff_flush_checkdiff(p, opt);
5239 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5240 diff_flush_raw(p, opt);
5241 else if (fmt & DIFF_FORMAT_NAME) {
5242 const char *name_a, *name_b;
5243 name_a = p->two->path;
5244 name_b = NULL;
5245 strip_prefix(opt->prefix_length, &name_a, &name_b);
5246 fprintf(opt->file, "%s", diff_line_prefix(opt));
5247 write_name_quoted(name_a, opt->file, opt->line_termination);
5248 }
5249 }
5250
5251 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5252 {
5253 struct strbuf sb = STRBUF_INIT;
5254 if (fs->mode)
5255 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5256 else
5257 strbuf_addf(&sb, " %s ", newdelete);
5258
5259 quote_c_style(fs->path, &sb, NULL, 0);
5260 strbuf_addch(&sb, '\n');
5261 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5262 sb.buf, sb.len, 0);
5263 strbuf_release(&sb);
5264 }
5265
5266 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5267 int show_name)
5268 {
5269 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5270 struct strbuf sb = STRBUF_INIT;
5271 strbuf_addf(&sb, " mode change %06o => %06o",
5272 p->one->mode, p->two->mode);
5273 if (show_name) {
5274 strbuf_addch(&sb, ' ');
5275 quote_c_style(p->two->path, &sb, NULL, 0);
5276 }
5277 strbuf_addch(&sb, '\n');
5278 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5279 sb.buf, sb.len, 0);
5280 strbuf_release(&sb);
5281 }
5282 }
5283
5284 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5285 struct diff_filepair *p)
5286 {
5287 struct strbuf sb = STRBUF_INIT;
5288 char *names = pprint_rename(p->one->path, p->two->path);
5289 strbuf_addf(&sb, " %s %s (%d%%)\n",
5290 renamecopy, names, similarity_index(p));
5291 free(names);
5292 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5293 sb.buf, sb.len, 0);
5294 show_mode_change(opt, p, 0);
5295 strbuf_release(&sb);
5296 }
5297
5298 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5299 {
5300 switch(p->status) {
5301 case DIFF_STATUS_DELETED:
5302 show_file_mode_name(opt, "delete", p->one);
5303 break;
5304 case DIFF_STATUS_ADDED:
5305 show_file_mode_name(opt, "create", p->two);
5306 break;
5307 case DIFF_STATUS_COPIED:
5308 show_rename_copy(opt, "copy", p);
5309 break;
5310 case DIFF_STATUS_RENAMED:
5311 show_rename_copy(opt, "rename", p);
5312 break;
5313 default:
5314 if (p->score) {
5315 struct strbuf sb = STRBUF_INIT;
5316 strbuf_addstr(&sb, " rewrite ");
5317 quote_c_style(p->two->path, &sb, NULL, 0);
5318 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5319 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5320 sb.buf, sb.len, 0);
5321 strbuf_release(&sb);
5322 }
5323 show_mode_change(opt, p, !p->score);
5324 break;
5325 }
5326 }
5327
5328 struct patch_id_t {
5329 git_SHA_CTX *ctx;
5330 int patchlen;
5331 };
5332
5333 static int remove_space(char *line, int len)
5334 {
5335 int i;
5336 char *dst = line;
5337 unsigned char c;
5338
5339 for (i = 0; i < len; i++)
5340 if (!isspace((c = line[i])))
5341 *dst++ = c;
5342
5343 return dst - line;
5344 }
5345
5346 static void patch_id_consume(void *priv, char *line, unsigned long len)
5347 {
5348 struct patch_id_t *data = priv;
5349 int new_len;
5350
5351 /* Ignore line numbers when computing the SHA1 of the patch */
5352 if (starts_with(line, "@@ -"))
5353 return;
5354
5355 new_len = remove_space(line, len);
5356
5357 git_SHA1_Update(data->ctx, line, new_len);
5358 data->patchlen += new_len;
5359 }
5360
5361 static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5362 {
5363 git_SHA1_Update(ctx, str, strlen(str));
5364 }
5365
5366 static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5367 {
5368 /* large enough for 2^32 in octal */
5369 char buf[12];
5370 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5371 git_SHA1_Update(ctx, buf, len);
5372 }
5373
5374 /* returns 0 upon success, and writes result into sha1 */
5375 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5376 {
5377 struct diff_queue_struct *q = &diff_queued_diff;
5378 int i;
5379 git_SHA_CTX ctx;
5380 struct patch_id_t data;
5381
5382 git_SHA1_Init(&ctx);
5383 memset(&data, 0, sizeof(struct patch_id_t));
5384 data.ctx = &ctx;
5385
5386 for (i = 0; i < q->nr; i++) {
5387 xpparam_t xpp;
5388 xdemitconf_t xecfg;
5389 mmfile_t mf1, mf2;
5390 struct diff_filepair *p = q->queue[i];
5391 int len1, len2;
5392
5393 memset(&xpp, 0, sizeof(xpp));
5394 memset(&xecfg, 0, sizeof(xecfg));
5395 if (p->status == 0)
5396 return error("internal diff status error");
5397 if (p->status == DIFF_STATUS_UNKNOWN)
5398 continue;
5399 if (diff_unmodified_pair(p))
5400 continue;
5401 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5402 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5403 continue;
5404 if (DIFF_PAIR_UNMERGED(p))
5405 continue;
5406
5407 diff_fill_oid_info(p->one);
5408 diff_fill_oid_info(p->two);
5409
5410 len1 = remove_space(p->one->path, strlen(p->one->path));
5411 len2 = remove_space(p->two->path, strlen(p->two->path));
5412 patch_id_add_string(&ctx, "diff--git");
5413 patch_id_add_string(&ctx, "a/");
5414 git_SHA1_Update(&ctx, p->one->path, len1);
5415 patch_id_add_string(&ctx, "b/");
5416 git_SHA1_Update(&ctx, p->two->path, len2);
5417
5418 if (p->one->mode == 0) {
5419 patch_id_add_string(&ctx, "newfilemode");
5420 patch_id_add_mode(&ctx, p->two->mode);
5421 patch_id_add_string(&ctx, "---/dev/null");
5422 patch_id_add_string(&ctx, "+++b/");
5423 git_SHA1_Update(&ctx, p->two->path, len2);
5424 } else if (p->two->mode == 0) {
5425 patch_id_add_string(&ctx, "deletedfilemode");
5426 patch_id_add_mode(&ctx, p->one->mode);
5427 patch_id_add_string(&ctx, "---a/");
5428 git_SHA1_Update(&ctx, p->one->path, len1);
5429 patch_id_add_string(&ctx, "+++/dev/null");
5430 } else {
5431 patch_id_add_string(&ctx, "---a/");
5432 git_SHA1_Update(&ctx, p->one->path, len1);
5433 patch_id_add_string(&ctx, "+++b/");
5434 git_SHA1_Update(&ctx, p->two->path, len2);
5435 }
5436
5437 if (diff_header_only)
5438 continue;
5439
5440 if (fill_mmfile(&mf1, p->one) < 0 ||
5441 fill_mmfile(&mf2, p->two) < 0)
5442 return error("unable to read files to diff");
5443
5444 if (diff_filespec_is_binary(p->one) ||
5445 diff_filespec_is_binary(p->two)) {
5446 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5447 GIT_SHA1_HEXSZ);
5448 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5449 GIT_SHA1_HEXSZ);
5450 continue;
5451 }
5452
5453 xpp.flags = 0;
5454 xecfg.ctxlen = 3;
5455 xecfg.flags = 0;
5456 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
5457 &xpp, &xecfg))
5458 return error("unable to generate patch-id diff for %s",
5459 p->one->path);
5460 }
5461
5462 git_SHA1_Final(oid->hash, &ctx);
5463 return 0;
5464 }
5465
5466 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5467 {
5468 struct diff_queue_struct *q = &diff_queued_diff;
5469 int i;
5470 int result = diff_get_patch_id(options, oid, diff_header_only);
5471
5472 for (i = 0; i < q->nr; i++)
5473 diff_free_filepair(q->queue[i]);
5474
5475 free(q->queue);
5476 DIFF_QUEUE_CLEAR(q);
5477
5478 return result;
5479 }
5480
5481 static int is_summary_empty(const struct diff_queue_struct *q)
5482 {
5483 int i;
5484
5485 for (i = 0; i < q->nr; i++) {
5486 const struct diff_filepair *p = q->queue[i];
5487
5488 switch (p->status) {
5489 case DIFF_STATUS_DELETED:
5490 case DIFF_STATUS_ADDED:
5491 case DIFF_STATUS_COPIED:
5492 case DIFF_STATUS_RENAMED:
5493 return 0;
5494 default:
5495 if (p->score)
5496 return 0;
5497 if (p->one->mode && p->two->mode &&
5498 p->one->mode != p->two->mode)
5499 return 0;
5500 break;
5501 }
5502 }
5503 return 1;
5504 }
5505
5506 static const char rename_limit_warning[] =
5507 N_("inexact rename detection was skipped due to too many files.");
5508
5509 static const char degrade_cc_to_c_warning[] =
5510 N_("only found copies from modified paths due to too many files.");
5511
5512 static const char rename_limit_advice[] =
5513 N_("you may want to set your %s variable to at least "
5514 "%d and retry the command.");
5515
5516 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5517 {
5518 if (degraded_cc)
5519 warning(_(degrade_cc_to_c_warning));
5520 else if (needed)
5521 warning(_(rename_limit_warning));
5522 else
5523 return;
5524 if (0 < needed && needed < 32767)
5525 warning(_(rename_limit_advice), varname, needed);
5526 }
5527
5528 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5529 {
5530 int i;
5531 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5532 struct diff_queue_struct *q = &diff_queued_diff;
5533
5534 if (WSEH_NEW & WS_RULE_MASK)
5535 die("BUG: WS rules bit mask overlaps with diff symbol flags");
5536
5537 if (o->color_moved)
5538 o->emitted_symbols = &esm;
5539
5540 for (i = 0; i < q->nr; i++) {
5541 struct diff_filepair *p = q->queue[i];
5542 if (check_pair_status(p))
5543 diff_flush_patch(p, o);
5544 }
5545
5546 if (o->emitted_symbols) {
5547 if (o->color_moved) {
5548 struct hashmap add_lines, del_lines;
5549
5550 hashmap_init(&del_lines,
5551 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5552 hashmap_init(&add_lines,
5553 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5554
5555 add_lines_to_move_detection(o, &add_lines, &del_lines);
5556 mark_color_as_moved(o, &add_lines, &del_lines);
5557 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5558 dim_moved_lines(o);
5559
5560 hashmap_free(&add_lines, 0);
5561 hashmap_free(&del_lines, 0);
5562 }
5563
5564 for (i = 0; i < esm.nr; i++)
5565 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5566
5567 for (i = 0; i < esm.nr; i++)
5568 free((void *)esm.buf[i].line);
5569 }
5570 esm.nr = 0;
5571 }
5572
5573 void diff_flush(struct diff_options *options)
5574 {
5575 struct diff_queue_struct *q = &diff_queued_diff;
5576 int i, output_format = options->output_format;
5577 int separator = 0;
5578 int dirstat_by_line = 0;
5579
5580 /*
5581 * Order: raw, stat, summary, patch
5582 * or: name/name-status/checkdiff (other bits clear)
5583 */
5584 if (!q->nr)
5585 goto free_queue;
5586
5587 if (output_format & (DIFF_FORMAT_RAW |
5588 DIFF_FORMAT_NAME |
5589 DIFF_FORMAT_NAME_STATUS |
5590 DIFF_FORMAT_CHECKDIFF)) {
5591 for (i = 0; i < q->nr; i++) {
5592 struct diff_filepair *p = q->queue[i];
5593 if (check_pair_status(p))
5594 flush_one_pair(p, options);
5595 }
5596 separator++;
5597 }
5598
5599 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
5600 dirstat_by_line = 1;
5601
5602 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5603 dirstat_by_line) {
5604 struct diffstat_t diffstat;
5605
5606 memset(&diffstat, 0, sizeof(struct diffstat_t));
5607 for (i = 0; i < q->nr; i++) {
5608 struct diff_filepair *p = q->queue[i];
5609 if (check_pair_status(p))
5610 diff_flush_stat(p, options, &diffstat);
5611 }
5612 if (output_format & DIFF_FORMAT_NUMSTAT)
5613 show_numstat(&diffstat, options);
5614 if (output_format & DIFF_FORMAT_DIFFSTAT)
5615 show_stats(&diffstat, options);
5616 if (output_format & DIFF_FORMAT_SHORTSTAT)
5617 show_shortstats(&diffstat, options);
5618 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5619 show_dirstat_by_line(&diffstat, options);
5620 free_diffstat_info(&diffstat);
5621 separator++;
5622 }
5623 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5624 show_dirstat(options);
5625
5626 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5627 for (i = 0; i < q->nr; i++) {
5628 diff_summary(options, q->queue[i]);
5629 }
5630 separator++;
5631 }
5632
5633 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5634 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
5635 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5636 /*
5637 * run diff_flush_patch for the exit status. setting
5638 * options->file to /dev/null should be safe, because we
5639 * aren't supposed to produce any output anyway.
5640 */
5641 if (options->close_file)
5642 fclose(options->file);
5643 options->file = xfopen("/dev/null", "w");
5644 options->close_file = 1;
5645 options->color_moved = 0;
5646 for (i = 0; i < q->nr; i++) {
5647 struct diff_filepair *p = q->queue[i];
5648 if (check_pair_status(p))
5649 diff_flush_patch(p, options);
5650 if (options->found_changes)
5651 break;
5652 }
5653 }
5654
5655 if (output_format & DIFF_FORMAT_PATCH) {
5656 if (separator) {
5657 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5658 if (options->stat_sep)
5659 /* attach patch instead of inline */
5660 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5661 NULL, 0, 0);
5662 }
5663
5664 diff_flush_patch_all_file_pairs(options);
5665 }
5666
5667 if (output_format & DIFF_FORMAT_CALLBACK)
5668 options->format_callback(q, options, options->format_callback_data);
5669
5670 for (i = 0; i < q->nr; i++)
5671 diff_free_filepair(q->queue[i]);
5672 free_queue:
5673 free(q->queue);
5674 DIFF_QUEUE_CLEAR(q);
5675 if (options->close_file)
5676 fclose(options->file);
5677
5678 /*
5679 * Report the content-level differences with HAS_CHANGES;
5680 * diff_addremove/diff_change does not set the bit when
5681 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5682 */
5683 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5684 if (options->found_changes)
5685 DIFF_OPT_SET(options, HAS_CHANGES);
5686 else
5687 DIFF_OPT_CLR(options, HAS_CHANGES);
5688 }
5689 }
5690
5691 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5692 {
5693 return (((p->status == DIFF_STATUS_MODIFIED) &&
5694 ((p->score &&
5695 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5696 (!p->score &&
5697 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5698 ((p->status != DIFF_STATUS_MODIFIED) &&
5699 filter_bit_tst(p->status, options)));
5700 }
5701
5702 static void diffcore_apply_filter(struct diff_options *options)
5703 {
5704 int i;
5705 struct diff_queue_struct *q = &diff_queued_diff;
5706 struct diff_queue_struct outq;
5707
5708 DIFF_QUEUE_CLEAR(&outq);
5709
5710 if (!options->filter)
5711 return;
5712
5713 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5714 int found;
5715 for (i = found = 0; !found && i < q->nr; i++) {
5716 if (match_filter(options, q->queue[i]))
5717 found++;
5718 }
5719 if (found)
5720 return;
5721
5722 /* otherwise we will clear the whole queue
5723 * by copying the empty outq at the end of this
5724 * function, but first clear the current entries
5725 * in the queue.
5726 */
5727 for (i = 0; i < q->nr; i++)
5728 diff_free_filepair(q->queue[i]);
5729 }
5730 else {
5731 /* Only the matching ones */
5732 for (i = 0; i < q->nr; i++) {
5733 struct diff_filepair *p = q->queue[i];
5734 if (match_filter(options, p))
5735 diff_q(&outq, p);
5736 else
5737 diff_free_filepair(p);
5738 }
5739 }
5740 free(q->queue);
5741 *q = outq;
5742 }
5743
5744 /* Check whether two filespecs with the same mode and size are identical */
5745 static int diff_filespec_is_identical(struct diff_filespec *one,
5746 struct diff_filespec *two)
5747 {
5748 if (S_ISGITLINK(one->mode))
5749 return 0;
5750 if (diff_populate_filespec(one, 0))
5751 return 0;
5752 if (diff_populate_filespec(two, 0))
5753 return 0;
5754 return !memcmp(one->data, two->data, one->size);
5755 }
5756
5757 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
5758 {
5759 if (p->done_skip_stat_unmatch)
5760 return p->skip_stat_unmatch_result;
5761
5762 p->done_skip_stat_unmatch = 1;
5763 p->skip_stat_unmatch_result = 0;
5764 /*
5765 * 1. Entries that come from stat info dirtiness
5766 * always have both sides (iow, not create/delete),
5767 * one side of the object name is unknown, with
5768 * the same mode and size. Keep the ones that
5769 * do not match these criteria. They have real
5770 * differences.
5771 *
5772 * 2. At this point, the file is known to be modified,
5773 * with the same mode and size, and the object
5774 * name of one side is unknown. Need to inspect
5775 * the identical contents.
5776 */
5777 if (!DIFF_FILE_VALID(p->one) || /* (1) */
5778 !DIFF_FILE_VALID(p->two) ||
5779 (p->one->oid_valid && p->two->oid_valid) ||
5780 (p->one->mode != p->two->mode) ||
5781 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
5782 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
5783 (p->one->size != p->two->size) ||
5784 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
5785 p->skip_stat_unmatch_result = 1;
5786 return p->skip_stat_unmatch_result;
5787 }
5788
5789 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
5790 {
5791 int i;
5792 struct diff_queue_struct *q = &diff_queued_diff;
5793 struct diff_queue_struct outq;
5794 DIFF_QUEUE_CLEAR(&outq);
5795
5796 for (i = 0; i < q->nr; i++) {
5797 struct diff_filepair *p = q->queue[i];
5798
5799 if (diff_filespec_check_stat_unmatch(p))
5800 diff_q(&outq, p);
5801 else {
5802 /*
5803 * The caller can subtract 1 from skip_stat_unmatch
5804 * to determine how many paths were dirty only
5805 * due to stat info mismatch.
5806 */
5807 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
5808 diffopt->skip_stat_unmatch++;
5809 diff_free_filepair(p);
5810 }
5811 }
5812 free(q->queue);
5813 *q = outq;
5814 }
5815
5816 static int diffnamecmp(const void *a_, const void *b_)
5817 {
5818 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
5819 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
5820 const char *name_a, *name_b;
5821
5822 name_a = a->one ? a->one->path : a->two->path;
5823 name_b = b->one ? b->one->path : b->two->path;
5824 return strcmp(name_a, name_b);
5825 }
5826
5827 void diffcore_fix_diff_index(struct diff_options *options)
5828 {
5829 struct diff_queue_struct *q = &diff_queued_diff;
5830 QSORT(q->queue, q->nr, diffnamecmp);
5831 }
5832
5833 void diffcore_std(struct diff_options *options)
5834 {
5835 /* NOTE please keep the following in sync with diff_tree_combined() */
5836 if (options->skip_stat_unmatch)
5837 diffcore_skip_stat_unmatch(options);
5838 if (!options->found_follow) {
5839 /* See try_to_follow_renames() in tree-diff.c */
5840 if (options->break_opt != -1)
5841 diffcore_break(options->break_opt);
5842 if (options->detect_rename)
5843 diffcore_rename(options);
5844 if (options->break_opt != -1)
5845 diffcore_merge_broken();
5846 }
5847 if (options->pickaxe)
5848 diffcore_pickaxe(options);
5849 if (options->orderfile)
5850 diffcore_order(options->orderfile);
5851 if (!options->found_follow)
5852 /* See try_to_follow_renames() in tree-diff.c */
5853 diff_resolve_rename_copy();
5854 diffcore_apply_filter(options);
5855
5856 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5857 DIFF_OPT_SET(options, HAS_CHANGES);
5858 else
5859 DIFF_OPT_CLR(options, HAS_CHANGES);
5860
5861 options->found_follow = 0;
5862 }
5863
5864 int diff_result_code(struct diff_options *opt, int status)
5865 {
5866 int result = 0;
5867
5868 diff_warn_rename_limit("diff.renameLimit",
5869 opt->needed_rename_limit,
5870 opt->degraded_cc_to_c);
5871 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5872 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5873 return status;
5874 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5875 DIFF_OPT_TST(opt, HAS_CHANGES))
5876 result |= 01;
5877 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5878 DIFF_OPT_TST(opt, CHECK_FAILED))
5879 result |= 02;
5880 return result;
5881 }
5882
5883 int diff_can_quit_early(struct diff_options *opt)
5884 {
5885 return (DIFF_OPT_TST(opt, QUICK) &&
5886 !opt->filter &&
5887 DIFF_OPT_TST(opt, HAS_CHANGES));
5888 }
5889
5890 /*
5891 * Shall changes to this submodule be ignored?
5892 *
5893 * Submodule changes can be configured to be ignored separately for each path,
5894 * but that configuration can be overridden from the command line.
5895 */
5896 static int is_submodule_ignored(const char *path, struct diff_options *options)
5897 {
5898 int ignored = 0;
5899 unsigned orig_flags = options->flags;
5900 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5901 set_diffopt_flags_from_submodule_config(options, path);
5902 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5903 ignored = 1;
5904 options->flags = orig_flags;
5905 return ignored;
5906 }
5907
5908 void diff_addremove(struct diff_options *options,
5909 int addremove, unsigned mode,
5910 const struct object_id *oid,
5911 int oid_valid,
5912 const char *concatpath, unsigned dirty_submodule)
5913 {
5914 struct diff_filespec *one, *two;
5915
5916 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5917 return;
5918
5919 /* This may look odd, but it is a preparation for
5920 * feeding "there are unchanged files which should
5921 * not produce diffs, but when you are doing copy
5922 * detection you would need them, so here they are"
5923 * entries to the diff-core. They will be prefixed
5924 * with something like '=' or '*' (I haven't decided
5925 * which but should not make any difference).
5926 * Feeding the same new and old to diff_change()
5927 * also has the same effect.
5928 * Before the final output happens, they are pruned after
5929 * merged into rename/copy pairs as appropriate.
5930 */
5931 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5932 addremove = (addremove == '+' ? '-' :
5933 addremove == '-' ? '+' : addremove);
5934
5935 if (options->prefix &&
5936 strncmp(concatpath, options->prefix, options->prefix_length))
5937 return;
5938
5939 one = alloc_filespec(concatpath);
5940 two = alloc_filespec(concatpath);
5941
5942 if (addremove != '+')
5943 fill_filespec(one, oid, oid_valid, mode);
5944 if (addremove != '-') {
5945 fill_filespec(two, oid, oid_valid, mode);
5946 two->dirty_submodule = dirty_submodule;
5947 }
5948
5949 diff_queue(&diff_queued_diff, one, two);
5950 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5951 DIFF_OPT_SET(options, HAS_CHANGES);
5952 }
5953
5954 void diff_change(struct diff_options *options,
5955 unsigned old_mode, unsigned new_mode,
5956 const struct object_id *old_oid,
5957 const struct object_id *new_oid,
5958 int old_oid_valid, int new_oid_valid,
5959 const char *concatpath,
5960 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5961 {
5962 struct diff_filespec *one, *two;
5963 struct diff_filepair *p;
5964
5965 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5966 is_submodule_ignored(concatpath, options))
5967 return;
5968
5969 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5970 SWAP(old_mode, new_mode);
5971 SWAP(old_oid, new_oid);
5972 SWAP(old_oid_valid, new_oid_valid);
5973 SWAP(old_dirty_submodule, new_dirty_submodule);
5974 }
5975
5976 if (options->prefix &&
5977 strncmp(concatpath, options->prefix, options->prefix_length))
5978 return;
5979
5980 one = alloc_filespec(concatpath);
5981 two = alloc_filespec(concatpath);
5982 fill_filespec(one, old_oid, old_oid_valid, old_mode);
5983 fill_filespec(two, new_oid, new_oid_valid, new_mode);
5984 one->dirty_submodule = old_dirty_submodule;
5985 two->dirty_submodule = new_dirty_submodule;
5986 p = diff_queue(&diff_queued_diff, one, two);
5987
5988 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5989 return;
5990
5991 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5992 !diff_filespec_check_stat_unmatch(p))
5993 return;
5994
5995 DIFF_OPT_SET(options, HAS_CHANGES);
5996 }
5997
5998 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5999 {
6000 struct diff_filepair *pair;
6001 struct diff_filespec *one, *two;
6002
6003 if (options->prefix &&
6004 strncmp(path, options->prefix, options->prefix_length))
6005 return NULL;
6006
6007 one = alloc_filespec(path);
6008 two = alloc_filespec(path);
6009 pair = diff_queue(&diff_queued_diff, one, two);
6010 pair->is_unmerged = 1;
6011 return pair;
6012 }
6013
6014 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
6015 size_t *outsize)
6016 {
6017 struct diff_tempfile *temp;
6018 const char *argv[3];
6019 const char **arg = argv;
6020 struct child_process child = CHILD_PROCESS_INIT;
6021 struct strbuf buf = STRBUF_INIT;
6022 int err = 0;
6023
6024 temp = prepare_temp_file(spec->path, spec);
6025 *arg++ = pgm;
6026 *arg++ = temp->name;
6027 *arg = NULL;
6028
6029 child.use_shell = 1;
6030 child.argv = argv;
6031 child.out = -1;
6032 if (start_command(&child)) {
6033 remove_tempfile();
6034 return NULL;
6035 }
6036
6037 if (strbuf_read(&buf, child.out, 0) < 0)
6038 err = error("error reading from textconv command '%s'", pgm);
6039 close(child.out);
6040
6041 if (finish_command(&child) || err) {
6042 strbuf_release(&buf);
6043 remove_tempfile();
6044 return NULL;
6045 }
6046 remove_tempfile();
6047
6048 return strbuf_detach(&buf, outsize);
6049 }
6050
6051 size_t fill_textconv(struct userdiff_driver *driver,
6052 struct diff_filespec *df,
6053 char **outbuf)
6054 {
6055 size_t size;
6056
6057 if (!driver) {
6058 if (!DIFF_FILE_VALID(df)) {
6059 *outbuf = "";
6060 return 0;
6061 }
6062 if (diff_populate_filespec(df, 0))
6063 die("unable to read files to diff");
6064 *outbuf = df->data;
6065 return df->size;
6066 }
6067
6068 if (!driver->textconv)
6069 die("BUG: fill_textconv called with non-textconv driver");
6070
6071 if (driver->textconv_cache && df->oid_valid) {
6072 *outbuf = notes_cache_get(driver->textconv_cache,
6073 &df->oid,
6074 &size);
6075 if (*outbuf)
6076 return size;
6077 }
6078
6079 *outbuf = run_textconv(driver->textconv, df, &size);
6080 if (!*outbuf)
6081 die("unable to read files to diff");
6082
6083 if (driver->textconv_cache && df->oid_valid) {
6084 /* ignore errors, as we might be in a readonly repository */
6085 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6086 size);
6087 /*
6088 * we could save up changes and flush them all at the end,
6089 * but we would need an extra call after all diffing is done.
6090 * Since generating a cache entry is the slow path anyway,
6091 * this extra overhead probably isn't a big deal.
6092 */
6093 notes_cache_write(driver->textconv_cache);
6094 }
6095
6096 return size;
6097 }
6098
6099 int textconv_object(const char *path,
6100 unsigned mode,
6101 const struct object_id *oid,
6102 int oid_valid,
6103 char **buf,
6104 unsigned long *buf_size)
6105 {
6106 struct diff_filespec *df;
6107 struct userdiff_driver *textconv;
6108
6109 df = alloc_filespec(path);
6110 fill_filespec(df, oid, oid_valid, mode);
6111 textconv = get_textconv(df);
6112 if (!textconv) {
6113 free_filespec(df);
6114 return 0;
6115 }
6116
6117 *buf_size = fill_textconv(textconv, df, buf);
6118 free_filespec(df);
6119 return 1;
6120 }
6121
6122 void setup_diff_pager(struct diff_options *opt)
6123 {
6124 /*
6125 * If the user asked for our exit code, then either they want --quiet
6126 * or --exit-code. We should definitely not bother with a pager in the
6127 * former case, as we will generate no output. Since we still properly
6128 * report our exit code even when a pager is run, we _could_ run a
6129 * pager with --exit-code. But since we have not done so historically,
6130 * and because it is easy to find people oneline advising "git diff
6131 * --exit-code" in hooks and other scripts, we do not do so.
6132 */
6133 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
6134 check_pager_config("diff") != 0)
6135 setup_pager();
6136 }
git for Debian