?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff: define block by number of alphanumeric chars
[git.git] / diff.c
blobc50fcc7ea5c8bc2e28a976332bfe1d3e48338ad5
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
25 #ifdef NO_FAST_WORKING_DIRECTORY
26 #define FAST_WORKING_DIRECTORY 0
27 #else
28 #define FAST_WORKING_DIRECTORY 1
29 #endif
30
31 static int diff_detect_rename_default;
32 static int diff_indent_heuristic = 1;
33 static int diff_rename_limit_default = 400;
34 static int diff_suppress_blank_empty;
35 static int diff_use_color_default = -1;
36 static int diff_color_moved_default;
37 static int diff_context_default = 3;
38 static int diff_interhunk_context_default;
39 static const char *diff_word_regex_cfg;
40 static const char *external_diff_cmd_cfg;
41 static const char *diff_order_file_cfg;
42 int diff_auto_refresh_index = 1;
43 static int diff_mnemonic_prefix;
44 static int diff_no_prefix;
45 static int diff_stat_graph_width;
46 static int diff_dirstat_permille_default = 30;
47 static struct diff_options default_diff_options;
48 static long diff_algorithm;
49 static unsigned ws_error_highlight_default = WSEH_NEW;
50
51 static char diff_colors[][COLOR_MAXLEN] = {
52 GIT_COLOR_RESET,
53 GIT_COLOR_NORMAL, /* CONTEXT */
54 GIT_COLOR_BOLD, /* METAINFO */
55 GIT_COLOR_CYAN, /* FRAGINFO */
56 GIT_COLOR_RED, /* OLD */
57 GIT_COLOR_GREEN, /* NEW */
58 GIT_COLOR_YELLOW, /* COMMIT */
59 GIT_COLOR_BG_RED, /* WHITESPACE */
60 GIT_COLOR_NORMAL, /* FUNCINFO */
61 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
62 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
63 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
64 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
65 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
66 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
67 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
68 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
69 };
70
71 static NORETURN void die_want_option(const char *option_name)
72 {
73 die(_("option '%s' requires a value"), option_name);
74 }
75
76 static int parse_diff_color_slot(const char *var)
77 {
78 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
79 return DIFF_CONTEXT;
80 if (!strcasecmp(var, "meta"))
81 return DIFF_METAINFO;
82 if (!strcasecmp(var, "frag"))
83 return DIFF_FRAGINFO;
84 if (!strcasecmp(var, "old"))
85 return DIFF_FILE_OLD;
86 if (!strcasecmp(var, "new"))
87 return DIFF_FILE_NEW;
88 if (!strcasecmp(var, "commit"))
89 return DIFF_COMMIT;
90 if (!strcasecmp(var, "whitespace"))
91 return DIFF_WHITESPACE;
92 if (!strcasecmp(var, "func"))
93 return DIFF_FUNCINFO;
94 if (!strcasecmp(var, "oldmoved"))
95 return DIFF_FILE_OLD_MOVED;
96 if (!strcasecmp(var, "oldmovedalternative"))
97 return DIFF_FILE_OLD_MOVED_ALT;
98 if (!strcasecmp(var, "oldmoveddimmed"))
99 return DIFF_FILE_OLD_MOVED_DIM;
100 if (!strcasecmp(var, "oldmovedalternativedimmed"))
101 return DIFF_FILE_OLD_MOVED_ALT_DIM;
102 if (!strcasecmp(var, "newmoved"))
103 return DIFF_FILE_NEW_MOVED;
104 if (!strcasecmp(var, "newmovedalternative"))
105 return DIFF_FILE_NEW_MOVED_ALT;
106 if (!strcasecmp(var, "newmoveddimmed"))
107 return DIFF_FILE_NEW_MOVED_DIM;
108 if (!strcasecmp(var, "newmovedalternativedimmed"))
109 return DIFF_FILE_NEW_MOVED_ALT_DIM;
110 return -1;
111 }
112
113 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
114 struct strbuf *errmsg)
115 {
116 char *params_copy = xstrdup(params_string);
117 struct string_list params = STRING_LIST_INIT_NODUP;
118 int ret = 0;
119 int i;
120
121 if (*params_copy)
122 string_list_split_in_place(&params, params_copy, ',', -1);
123 for (i = 0; i < params.nr; i++) {
124 const char *p = params.items[i].string;
125 if (!strcmp(p, "changes")) {
126 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
127 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
128 } else if (!strcmp(p, "lines")) {
129 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
130 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
131 } else if (!strcmp(p, "files")) {
132 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
133 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
134 } else if (!strcmp(p, "noncumulative")) {
135 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
136 } else if (!strcmp(p, "cumulative")) {
137 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
138 } else if (isdigit(*p)) {
139 char *end;
140 int permille = strtoul(p, &end, 10) * 10;
141 if (*end == '.' && isdigit(*++end)) {
142 /* only use first digit */
143 permille += *end - '0';
144 /* .. and ignore any further digits */
145 while (isdigit(*++end))
146 ; /* nothing */
147 }
148 if (!*end)
149 options->dirstat_permille = permille;
150 else {
151 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
152 p);
153 ret++;
154 }
155 } else {
156 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
157 ret++;
158 }
159
160 }
161 string_list_clear(&params, 0);
162 free(params_copy);
163 return ret;
164 }
165
166 static int parse_submodule_params(struct diff_options *options, const char *value)
167 {
168 if (!strcmp(value, "log"))
169 options->submodule_format = DIFF_SUBMODULE_LOG;
170 else if (!strcmp(value, "short"))
171 options->submodule_format = DIFF_SUBMODULE_SHORT;
172 else if (!strcmp(value, "diff"))
173 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
174 else
175 return -1;
176 return 0;
177 }
178
179 static int git_config_rename(const char *var, const char *value)
180 {
181 if (!value)
182 return DIFF_DETECT_RENAME;
183 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
184 return DIFF_DETECT_COPY;
185 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
186 }
187
188 long parse_algorithm_value(const char *value)
189 {
190 if (!value)
191 return -1;
192 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
193 return 0;
194 else if (!strcasecmp(value, "minimal"))
195 return XDF_NEED_MINIMAL;
196 else if (!strcasecmp(value, "patience"))
197 return XDF_PATIENCE_DIFF;
198 else if (!strcasecmp(value, "histogram"))
199 return XDF_HISTOGRAM_DIFF;
200 return -1;
201 }
202
203 static int parse_one_token(const char **arg, const char *token)
204 {
205 const char *rest;
206 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
207 *arg = rest;
208 return 1;
209 }
210 return 0;
211 }
212
213 static int parse_ws_error_highlight(const char *arg)
214 {
215 const char *orig_arg = arg;
216 unsigned val = 0;
217
218 while (*arg) {
219 if (parse_one_token(&arg, "none"))
220 val = 0;
221 else if (parse_one_token(&arg, "default"))
222 val = WSEH_NEW;
223 else if (parse_one_token(&arg, "all"))
224 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
225 else if (parse_one_token(&arg, "new"))
226 val |= WSEH_NEW;
227 else if (parse_one_token(&arg, "old"))
228 val |= WSEH_OLD;
229 else if (parse_one_token(&arg, "context"))
230 val |= WSEH_CONTEXT;
231 else {
232 return -1 - (int)(arg - orig_arg);
233 }
234 if (*arg)
235 arg++;
236 }
237 return val;
238 }
239
240 /*
241 * These are to give UI layer defaults.
242 * The core-level commands such as git-diff-files should
243 * never be affected by the setting of diff.renames
244 * the user happens to have in the configuration file.
245 */
246 void init_diff_ui_defaults(void)
247 {
248 diff_detect_rename_default = 1;
249 }
250
251 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
252 {
253 if (!strcmp(var, "diff.indentheuristic"))
254 diff_indent_heuristic = git_config_bool(var, value);
255 return 0;
256 }
257
258 static int parse_color_moved(const char *arg)
259 {
260 switch (git_parse_maybe_bool(arg)) {
261 case 0:
262 return COLOR_MOVED_NO;
263 case 1:
264 return COLOR_MOVED_DEFAULT;
265 default:
266 break;
267 }
268
269 if (!strcmp(arg, "no"))
270 return COLOR_MOVED_NO;
271 else if (!strcmp(arg, "plain"))
272 return COLOR_MOVED_PLAIN;
273 else if (!strcmp(arg, "zebra"))
274 return COLOR_MOVED_ZEBRA;
275 else if (!strcmp(arg, "default"))
276 return COLOR_MOVED_DEFAULT;
277 else if (!strcmp(arg, "dimmed_zebra"))
278 return COLOR_MOVED_ZEBRA_DIM;
279 else
280 return error(_("color moved setting must be one of 'no', 'default', 'zebra', 'dimmed_zebra', 'plain'"));
281 }
282
283 int git_diff_ui_config(const char *var, const char *value, void *cb)
284 {
285 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
286 diff_use_color_default = git_config_colorbool(var, value);
287 return 0;
288 }
289 if (!strcmp(var, "diff.colormoved")) {
290 int cm = parse_color_moved(value);
291 if (cm < 0)
292 return -1;
293 diff_color_moved_default = cm;
294 return 0;
295 }
296 if (!strcmp(var, "diff.context")) {
297 diff_context_default = git_config_int(var, value);
298 if (diff_context_default < 0)
299 return -1;
300 return 0;
301 }
302 if (!strcmp(var, "diff.interhunkcontext")) {
303 diff_interhunk_context_default = git_config_int(var, value);
304 if (diff_interhunk_context_default < 0)
305 return -1;
306 return 0;
307 }
308 if (!strcmp(var, "diff.renames")) {
309 diff_detect_rename_default = git_config_rename(var, value);
310 return 0;
311 }
312 if (!strcmp(var, "diff.autorefreshindex")) {
313 diff_auto_refresh_index = git_config_bool(var, value);
314 return 0;
315 }
316 if (!strcmp(var, "diff.mnemonicprefix")) {
317 diff_mnemonic_prefix = git_config_bool(var, value);
318 return 0;
319 }
320 if (!strcmp(var, "diff.noprefix")) {
321 diff_no_prefix = git_config_bool(var, value);
322 return 0;
323 }
324 if (!strcmp(var, "diff.statgraphwidth")) {
325 diff_stat_graph_width = git_config_int(var, value);
326 return 0;
327 }
328 if (!strcmp(var, "diff.external"))
329 return git_config_string(&external_diff_cmd_cfg, var, value);
330 if (!strcmp(var, "diff.wordregex"))
331 return git_config_string(&diff_word_regex_cfg, var, value);
332 if (!strcmp(var, "diff.orderfile"))
333 return git_config_pathname(&diff_order_file_cfg, var, value);
334
335 if (!strcmp(var, "diff.ignoresubmodules"))
336 handle_ignore_submodules_arg(&default_diff_options, value);
337
338 if (!strcmp(var, "diff.submodule")) {
339 if (parse_submodule_params(&default_diff_options, value))
340 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
341 value);
342 return 0;
343 }
344
345 if (!strcmp(var, "diff.algorithm")) {
346 diff_algorithm = parse_algorithm_value(value);
347 if (diff_algorithm < 0)
348 return -1;
349 return 0;
350 }
351
352 if (!strcmp(var, "diff.wserrorhighlight")) {
353 int val = parse_ws_error_highlight(value);
354 if (val < 0)
355 return -1;
356 ws_error_highlight_default = val;
357 return 0;
358 }
359
360 if (git_color_config(var, value, cb) < 0)
361 return -1;
362
363 return git_diff_basic_config(var, value, cb);
364 }
365
366 int git_diff_basic_config(const char *var, const char *value, void *cb)
367 {
368 const char *name;
369
370 if (!strcmp(var, "diff.renamelimit")) {
371 diff_rename_limit_default = git_config_int(var, value);
372 return 0;
373 }
374
375 if (userdiff_config(var, value) < 0)
376 return -1;
377
378 if (skip_prefix(var, "diff.color.", &name) ||
379 skip_prefix(var, "color.diff.", &name)) {
380 int slot = parse_diff_color_slot(name);
381 if (slot < 0)
382 return 0;
383 if (!value)
384 return config_error_nonbool(var);
385 return color_parse(value, diff_colors[slot]);
386 }
387
388 /* like GNU diff's --suppress-blank-empty option */
389 if (!strcmp(var, "diff.suppressblankempty") ||
390 /* for backwards compatibility */
391 !strcmp(var, "diff.suppress-blank-empty")) {
392 diff_suppress_blank_empty = git_config_bool(var, value);
393 return 0;
394 }
395
396 if (!strcmp(var, "diff.dirstat")) {
397 struct strbuf errmsg = STRBUF_INIT;
398 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
399 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
400 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
401 errmsg.buf);
402 strbuf_release(&errmsg);
403 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
404 return 0;
405 }
406
407 if (starts_with(var, "submodule."))
408 return parse_submodule_config_option(var, value);
409
410 if (git_diff_heuristic_config(var, value, cb) < 0)
411 return -1;
412
413 return git_default_config(var, value, cb);
414 }
415
416 static char *quote_two(const char *one, const char *two)
417 {
418 int need_one = quote_c_style(one, NULL, NULL, 1);
419 int need_two = quote_c_style(two, NULL, NULL, 1);
420 struct strbuf res = STRBUF_INIT;
421
422 if (need_one + need_two) {
423 strbuf_addch(&res, '"');
424 quote_c_style(one, &res, NULL, 1);
425 quote_c_style(two, &res, NULL, 1);
426 strbuf_addch(&res, '"');
427 } else {
428 strbuf_addstr(&res, one);
429 strbuf_addstr(&res, two);
430 }
431 return strbuf_detach(&res, NULL);
432 }
433
434 static const char *external_diff(void)
435 {
436 static const char *external_diff_cmd = NULL;
437 static int done_preparing = 0;
438
439 if (done_preparing)
440 return external_diff_cmd;
441 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
442 if (!external_diff_cmd)
443 external_diff_cmd = external_diff_cmd_cfg;
444 done_preparing = 1;
445 return external_diff_cmd;
446 }
447
448 /*
449 * Keep track of files used for diffing. Sometimes such an entry
450 * refers to a temporary file, sometimes to an existing file, and
451 * sometimes to "/dev/null".
452 */
453 static struct diff_tempfile {
454 /*
455 * filename external diff should read from, or NULL if this
456 * entry is currently not in use:
457 */
458 const char *name;
459
460 char hex[GIT_MAX_HEXSZ + 1];
461 char mode[10];
462
463 /*
464 * If this diff_tempfile instance refers to a temporary file,
465 * this tempfile object is used to manage its lifetime.
466 */
467 struct tempfile tempfile;
468 } diff_temp[2];
469
470 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
471
472 struct emit_callback {
473 int color_diff;
474 unsigned ws_rule;
475 int blank_at_eof_in_preimage;
476 int blank_at_eof_in_postimage;
477 int lno_in_preimage;
478 int lno_in_postimage;
479 sane_truncate_fn truncate;
480 const char **label_path;
481 struct diff_words_data *diff_words;
482 struct diff_options *opt;
483 struct strbuf *header;
484 };
485
486 static int count_lines(const char *data, int size)
487 {
488 int count, ch, completely_empty = 1, nl_just_seen = 0;
489 count = 0;
490 while (0 < size--) {
491 ch = *data++;
492 if (ch == '\n') {
493 count++;
494 nl_just_seen = 1;
495 completely_empty = 0;
496 }
497 else {
498 nl_just_seen = 0;
499 completely_empty = 0;
500 }
501 }
502 if (completely_empty)
503 return 0;
504 if (!nl_just_seen)
505 count++; /* no trailing newline */
506 return count;
507 }
508
509 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
510 {
511 if (!DIFF_FILE_VALID(one)) {
512 mf->ptr = (char *)""; /* does not matter */
513 mf->size = 0;
514 return 0;
515 }
516 else if (diff_populate_filespec(one, 0))
517 return -1;
518
519 mf->ptr = one->data;
520 mf->size = one->size;
521 return 0;
522 }
523
524 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
525 static unsigned long diff_filespec_size(struct diff_filespec *one)
526 {
527 if (!DIFF_FILE_VALID(one))
528 return 0;
529 diff_populate_filespec(one, CHECK_SIZE_ONLY);
530 return one->size;
531 }
532
533 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
534 {
535 char *ptr = mf->ptr;
536 long size = mf->size;
537 int cnt = 0;
538
539 if (!size)
540 return cnt;
541 ptr += size - 1; /* pointing at the very end */
542 if (*ptr != '\n')
543 ; /* incomplete line */
544 else
545 ptr--; /* skip the last LF */
546 while (mf->ptr < ptr) {
547 char *prev_eol;
548 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
549 if (*prev_eol == '\n')
550 break;
551 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
552 break;
553 cnt++;
554 ptr = prev_eol - 1;
555 }
556 return cnt;
557 }
558
559 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
560 struct emit_callback *ecbdata)
561 {
562 int l1, l2, at;
563 unsigned ws_rule = ecbdata->ws_rule;
564 l1 = count_trailing_blank(mf1, ws_rule);
565 l2 = count_trailing_blank(mf2, ws_rule);
566 if (l2 <= l1) {
567 ecbdata->blank_at_eof_in_preimage = 0;
568 ecbdata->blank_at_eof_in_postimage = 0;
569 return;
570 }
571 at = count_lines(mf1->ptr, mf1->size);
572 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
573
574 at = count_lines(mf2->ptr, mf2->size);
575 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
576 }
577
578 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
579 int first, const char *line, int len)
580 {
581 int has_trailing_newline, has_trailing_carriage_return;
582 int nofirst;
583 FILE *file = o->file;
584
585 fputs(diff_line_prefix(o), file);
586
587 if (len == 0) {
588 has_trailing_newline = (first == '\n');
589 has_trailing_carriage_return = (!has_trailing_newline &&
590 (first == '\r'));
591 nofirst = has_trailing_newline || has_trailing_carriage_return;
592 } else {
593 has_trailing_newline = (len > 0 && line[len-1] == '\n');
594 if (has_trailing_newline)
595 len--;
596 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
597 if (has_trailing_carriage_return)
598 len--;
599 nofirst = 0;
600 }
601
602 if (len || !nofirst) {
603 fputs(set, file);
604 if (!nofirst)
605 fputc(first, file);
606 fwrite(line, len, 1, file);
607 fputs(reset, file);
608 }
609 if (has_trailing_carriage_return)
610 fputc('\r', file);
611 if (has_trailing_newline)
612 fputc('\n', file);
613 }
614
615 static void emit_line(struct diff_options *o, const char *set, const char *reset,
616 const char *line, int len)
617 {
618 emit_line_0(o, set, reset, line[0], line+1, len-1);
619 }
620
621 enum diff_symbol {
622 DIFF_SYMBOL_BINARY_DIFF_HEADER,
623 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
624 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
625 DIFF_SYMBOL_BINARY_DIFF_BODY,
626 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
627 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
628 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
629 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
630 DIFF_SYMBOL_STATS_LINE,
631 DIFF_SYMBOL_WORD_DIFF,
632 DIFF_SYMBOL_STAT_SEP,
633 DIFF_SYMBOL_SUMMARY,
634 DIFF_SYMBOL_SUBMODULE_ADD,
635 DIFF_SYMBOL_SUBMODULE_DEL,
636 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
637 DIFF_SYMBOL_SUBMODULE_MODIFIED,
638 DIFF_SYMBOL_SUBMODULE_HEADER,
639 DIFF_SYMBOL_SUBMODULE_ERROR,
640 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
641 DIFF_SYMBOL_REWRITE_DIFF,
642 DIFF_SYMBOL_BINARY_FILES,
643 DIFF_SYMBOL_HEADER,
644 DIFF_SYMBOL_FILEPAIR_PLUS,
645 DIFF_SYMBOL_FILEPAIR_MINUS,
646 DIFF_SYMBOL_WORDS_PORCELAIN,
647 DIFF_SYMBOL_WORDS,
648 DIFF_SYMBOL_CONTEXT,
649 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
650 DIFF_SYMBOL_PLUS,
651 DIFF_SYMBOL_MINUS,
652 DIFF_SYMBOL_NO_LF_EOF,
653 DIFF_SYMBOL_CONTEXT_FRAGINFO,
654 DIFF_SYMBOL_CONTEXT_MARKER,
655 DIFF_SYMBOL_SEPARATOR
656 };
657 /*
658 * Flags for content lines:
659 * 0..12 are whitespace rules
660 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
661 * 16 is marking if the line is blank at EOF
662 */
663 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
664 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
665 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
666 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
667 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
668
669 /*
670 * This struct is used when we need to buffer the output of the diff output.
671 *
672 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
673 * into the pre/post image file. This pointer could be a union with the
674 * line pointer. By storing an offset into the file instead of the literal line,
675 * we can decrease the memory footprint for the buffered output. At first we
676 * may want to only have indirection for the content lines, but we could also
677 * enhance the state for emitting prefabricated lines, e.g. the similarity
678 * score line or hunk/file headers would only need to store a number or path
679 * and then the output can be constructed later on depending on state.
680 */
681 struct emitted_diff_symbol {
682 const char *line;
683 int len;
684 int flags;
685 enum diff_symbol s;
686 };
687 #define EMITTED_DIFF_SYMBOL_INIT {NULL}
688
689 struct emitted_diff_symbols {
690 struct emitted_diff_symbol *buf;
691 int nr, alloc;
692 };
693 #define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
694
695 static void append_emitted_diff_symbol(struct diff_options *o,
696 struct emitted_diff_symbol *e)
697 {
698 struct emitted_diff_symbol *f;
699
700 ALLOC_GROW(o->emitted_symbols->buf,
701 o->emitted_symbols->nr + 1,
702 o->emitted_symbols->alloc);
703 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
704
705 memcpy(f, e, sizeof(struct emitted_diff_symbol));
706 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
707 }
708
709 struct moved_entry {
710 struct hashmap_entry ent;
711 const struct emitted_diff_symbol *es;
712 struct moved_entry *next_line;
713 };
714
715 static int next_byte(const char **cp, const char **endp,
716 const struct diff_options *diffopt)
717 {
718 int retval;
719
720 if (*cp > *endp)
721 return -1;
722
723 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_CHANGE)) {
724 while (*cp < *endp && isspace(**cp))
725 (*cp)++;
726 /*
727 * After skipping a couple of whitespaces, we still have to
728 * account for one space.
729 */
730 return (int)' ';
731 }
732
733 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE)) {
734 while (*cp < *endp && isspace(**cp))
735 (*cp)++;
736 /* return the first non-ws character via the usual below */
737 }
738
739 retval = (unsigned char)(**cp);
740 (*cp)++;
741 return retval;
742 }
743
744 static int moved_entry_cmp(const struct diff_options *diffopt,
745 const struct moved_entry *a,
746 const struct moved_entry *b,
747 const void *keydata)
748 {
749 const char *ap = a->es->line, *ae = a->es->line + a->es->len;
750 const char *bp = b->es->line, *be = b->es->line + b->es->len;
751
752 if (!(diffopt->xdl_opts & XDF_WHITESPACE_FLAGS))
753 return a->es->len != b->es->len || memcmp(ap, bp, a->es->len);
754
755 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_AT_EOL)) {
756 while (ae > ap && isspace(*ae))
757 ae--;
758 while (be > bp && isspace(*be))
759 be--;
760 }
761
762 while (1) {
763 int ca, cb;
764 ca = next_byte(&ap, &ae, diffopt);
765 cb = next_byte(&bp, &be, diffopt);
766 if (ca != cb)
767 return 1;
768 if (ca < 0)
769 return 0;
770 }
771 }
772
773 static unsigned get_string_hash(struct emitted_diff_symbol *es, struct diff_options *o)
774 {
775 if (o->xdl_opts & XDF_WHITESPACE_FLAGS) {
776 static struct strbuf sb = STRBUF_INIT;
777 const char *ap = es->line, *ae = es->line + es->len;
778 int c;
779
780 strbuf_reset(&sb);
781 while (ae > ap && isspace(*ae))
782 ae--;
783 while ((c = next_byte(&ap, &ae, o)) > 0)
784 strbuf_addch(&sb, c);
785
786 return memhash(sb.buf, sb.len);
787 } else {
788 return memhash(es->line, es->len);
789 }
790 }
791
792 static struct moved_entry *prepare_entry(struct diff_options *o,
793 int line_no)
794 {
795 struct moved_entry *ret = xmalloc(sizeof(*ret));
796 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
797
798 ret->ent.hash = get_string_hash(l, o);
799 ret->es = l;
800 ret->next_line = NULL;
801
802 return ret;
803 }
804
805 static void add_lines_to_move_detection(struct diff_options *o,
806 struct hashmap *add_lines,
807 struct hashmap *del_lines)
808 {
809 struct moved_entry *prev_line = NULL;
810
811 int n;
812 for (n = 0; n < o->emitted_symbols->nr; n++) {
813 struct hashmap *hm;
814 struct moved_entry *key;
815
816 switch (o->emitted_symbols->buf[n].s) {
817 case DIFF_SYMBOL_PLUS:
818 hm = add_lines;
819 break;
820 case DIFF_SYMBOL_MINUS:
821 hm = del_lines;
822 break;
823 default:
824 prev_line = NULL;
825 continue;
826 }
827
828 key = prepare_entry(o, n);
829 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
830 prev_line->next_line = key;
831
832 hashmap_add(hm, key);
833 prev_line = key;
834 }
835 }
836
837 static int shrink_potential_moved_blocks(struct moved_entry **pmb,
838 int pmb_nr)
839 {
840 int lp, rp;
841
842 /* Shrink the set of potential block to the remaining running */
843 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
844 while (lp < pmb_nr && pmb[lp])
845 lp++;
846 /* lp points at the first NULL now */
847
848 while (rp > -1 && !pmb[rp])
849 rp--;
850 /* rp points at the last non-NULL */
851
852 if (lp < pmb_nr && rp > -1 && lp < rp) {
853 pmb[lp] = pmb[rp];
854 pmb[rp] = NULL;
855 rp--;
856 lp++;
857 }
858 }
859
860 /* Remember the number of running sets */
861 return rp + 1;
862 }
863
864 /*
865 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
866 *
867 * Otherwise, if the last block has fewer alphanumeric characters than
868 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
869 * that block.
870 *
871 * The last block consists of the (n - block_length)'th line up to but not
872 * including the nth line.
873 *
874 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
875 * Think of a way to unify them.
876 */
877 static void adjust_last_block(struct diff_options *o, int n, int block_length)
878 {
879 int i, alnum_count = 0;
880 if (o->color_moved == COLOR_MOVED_PLAIN)
881 return;
882 for (i = 1; i < block_length + 1; i++) {
883 const char *c = o->emitted_symbols->buf[n - i].line;
884 for (; *c; c++) {
885 if (!isalnum(*c))
886 continue;
887 alnum_count++;
888 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
889 return;
890 }
891 }
892 for (i = 1; i < block_length + 1; i++)
893 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
894 }
895
896 /* Find blocks of moved code, delegate actual coloring decision to helper */
897 static void mark_color_as_moved(struct diff_options *o,
898 struct hashmap *add_lines,
899 struct hashmap *del_lines)
900 {
901 struct moved_entry **pmb = NULL; /* potentially moved blocks */
902 int pmb_nr = 0, pmb_alloc = 0;
903 int n, flipped_block = 1, block_length = 0;
904
905
906 for (n = 0; n < o->emitted_symbols->nr; n++) {
907 struct hashmap *hm = NULL;
908 struct moved_entry *key;
909 struct moved_entry *match = NULL;
910 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
911 int i;
912
913 switch (l->s) {
914 case DIFF_SYMBOL_PLUS:
915 hm = del_lines;
916 key = prepare_entry(o, n);
917 match = hashmap_get(hm, key, o);
918 free(key);
919 break;
920 case DIFF_SYMBOL_MINUS:
921 hm = add_lines;
922 key = prepare_entry(o, n);
923 match = hashmap_get(hm, key, o);
924 free(key);
925 break;
926 default:
927 flipped_block = 1;
928 }
929
930 if (!match) {
931 adjust_last_block(o, n, block_length);
932 pmb_nr = 0;
933 block_length = 0;
934 continue;
935 }
936
937 l->flags |= DIFF_SYMBOL_MOVED_LINE;
938
939 if (o->color_moved == COLOR_MOVED_PLAIN)
940 continue;
941
942 /* Check any potential block runs, advance each or nullify */
943 for (i = 0; i < pmb_nr; i++) {
944 struct moved_entry *p = pmb[i];
945 struct moved_entry *pnext = (p && p->next_line) ?
946 p->next_line : NULL;
947 if (pnext && !hm->cmpfn(o, pnext, match, NULL)) {
948 pmb[i] = p->next_line;
949 } else {
950 pmb[i] = NULL;
951 }
952 }
953
954 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
955
956 if (pmb_nr == 0) {
957 /*
958 * The current line is the start of a new block.
959 * Setup the set of potential blocks.
960 */
961 for (; match; match = hashmap_get_next(hm, match)) {
962 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
963 pmb[pmb_nr++] = match;
964 }
965
966 flipped_block = (flipped_block + 1) % 2;
967
968 adjust_last_block(o, n, block_length);
969 block_length = 0;
970 }
971
972 block_length++;
973
974 if (flipped_block)
975 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
976 }
977 adjust_last_block(o, n, block_length);
978
979 free(pmb);
980 }
981
982 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
983 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
984 static void dim_moved_lines(struct diff_options *o)
985 {
986 int n;
987 for (n = 0; n < o->emitted_symbols->nr; n++) {
988 struct emitted_diff_symbol *prev = (n != 0) ?
989 &o->emitted_symbols->buf[n - 1] : NULL;
990 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
991 struct emitted_diff_symbol *next =
992 (n < o->emitted_symbols->nr - 1) ?
993 &o->emitted_symbols->buf[n + 1] : NULL;
994
995 /* Not a plus or minus line? */
996 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
997 continue;
998
999 /* Not a moved line? */
1000 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1001 continue;
1002
1003 /*
1004 * If prev or next are not a plus or minus line,
1005 * pretend they don't exist
1006 */
1007 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1008 prev->s != DIFF_SYMBOL_MINUS)
1009 prev = NULL;
1010 if (next && next->s != DIFF_SYMBOL_PLUS &&
1011 next->s != DIFF_SYMBOL_MINUS)
1012 next = NULL;
1013
1014 /* Inside a block? */
1015 if ((prev &&
1016 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1017 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1018 (next &&
1019 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1020 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1021 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1022 continue;
1023 }
1024
1025 /* Check if we are at an interesting bound: */
1026 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1027 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1028 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1029 continue;
1030 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1031 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1032 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1033 continue;
1034
1035 /*
1036 * The boundary to prev and next are not interesting,
1037 * so this line is not interesting as a whole
1038 */
1039 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1040 }
1041 }
1042
1043 static void emit_line_ws_markup(struct diff_options *o,
1044 const char *set, const char *reset,
1045 const char *line, int len, char sign,
1046 unsigned ws_rule, int blank_at_eof)
1047 {
1048 const char *ws = NULL;
1049
1050 if (o->ws_error_highlight & ws_rule) {
1051 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1052 if (!*ws)
1053 ws = NULL;
1054 }
1055
1056 if (!ws)
1057 emit_line_0(o, set, reset, sign, line, len);
1058 else if (blank_at_eof)
1059 /* Blank line at EOF - paint '+' as well */
1060 emit_line_0(o, ws, reset, sign, line, len);
1061 else {
1062 /* Emit just the prefix, then the rest. */
1063 emit_line_0(o, set, reset, sign, "", 0);
1064 ws_check_emit(line, len, ws_rule,
1065 o->file, set, reset, ws);
1066 }
1067 }
1068
1069 static void emit_diff_symbol_from_struct(struct diff_options *o,
1070 struct emitted_diff_symbol *eds)
1071 {
1072 static const char *nneof = " No newline at end of file\n";
1073 const char *context, *reset, *set, *meta, *fraginfo;
1074 struct strbuf sb = STRBUF_INIT;
1075
1076 enum diff_symbol s = eds->s;
1077 const char *line = eds->line;
1078 int len = eds->len;
1079 unsigned flags = eds->flags;
1080
1081 switch (s) {
1082 case DIFF_SYMBOL_NO_LF_EOF:
1083 context = diff_get_color_opt(o, DIFF_CONTEXT);
1084 reset = diff_get_color_opt(o, DIFF_RESET);
1085 putc('\n', o->file);
1086 emit_line_0(o, context, reset, '\\',
1087 nneof, strlen(nneof));
1088 break;
1089 case DIFF_SYMBOL_SUBMODULE_HEADER:
1090 case DIFF_SYMBOL_SUBMODULE_ERROR:
1091 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1092 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1093 case DIFF_SYMBOL_SUMMARY:
1094 case DIFF_SYMBOL_STATS_LINE:
1095 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1096 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1097 emit_line(o, "", "", line, len);
1098 break;
1099 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1100 case DIFF_SYMBOL_CONTEXT_MARKER:
1101 context = diff_get_color_opt(o, DIFF_CONTEXT);
1102 reset = diff_get_color_opt(o, DIFF_RESET);
1103 emit_line(o, context, reset, line, len);
1104 break;
1105 case DIFF_SYMBOL_SEPARATOR:
1106 fprintf(o->file, "%s%c",
1107 diff_line_prefix(o),
1108 o->line_termination);
1109 break;
1110 case DIFF_SYMBOL_CONTEXT:
1111 set = diff_get_color_opt(o, DIFF_CONTEXT);
1112 reset = diff_get_color_opt(o, DIFF_RESET);
1113 emit_line_ws_markup(o, set, reset, line, len, ' ',
1114 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1115 break;
1116 case DIFF_SYMBOL_PLUS:
1117 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1118 DIFF_SYMBOL_MOVED_LINE_ALT |
1119 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1120 case DIFF_SYMBOL_MOVED_LINE |
1121 DIFF_SYMBOL_MOVED_LINE_ALT |
1122 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1123 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1124 break;
1125 case DIFF_SYMBOL_MOVED_LINE |
1126 DIFF_SYMBOL_MOVED_LINE_ALT:
1127 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1128 break;
1129 case DIFF_SYMBOL_MOVED_LINE |
1130 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1131 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1132 break;
1133 case DIFF_SYMBOL_MOVED_LINE:
1134 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1135 break;
1136 default:
1137 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1138 }
1139 reset = diff_get_color_opt(o, DIFF_RESET);
1140 emit_line_ws_markup(o, set, reset, line, len, '+',
1141 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1142 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1143 break;
1144 case DIFF_SYMBOL_MINUS:
1145 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1146 DIFF_SYMBOL_MOVED_LINE_ALT |
1147 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1148 case DIFF_SYMBOL_MOVED_LINE |
1149 DIFF_SYMBOL_MOVED_LINE_ALT |
1150 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1151 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1152 break;
1153 case DIFF_SYMBOL_MOVED_LINE |
1154 DIFF_SYMBOL_MOVED_LINE_ALT:
1155 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1156 break;
1157 case DIFF_SYMBOL_MOVED_LINE |
1158 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1159 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1160 break;
1161 case DIFF_SYMBOL_MOVED_LINE:
1162 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1163 break;
1164 default:
1165 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1166 }
1167 reset = diff_get_color_opt(o, DIFF_RESET);
1168 emit_line_ws_markup(o, set, reset, line, len, '-',
1169 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1170 break;
1171 case DIFF_SYMBOL_WORDS_PORCELAIN:
1172 context = diff_get_color_opt(o, DIFF_CONTEXT);
1173 reset = diff_get_color_opt(o, DIFF_RESET);
1174 emit_line(o, context, reset, line, len);
1175 fputs("~\n", o->file);
1176 break;
1177 case DIFF_SYMBOL_WORDS:
1178 context = diff_get_color_opt(o, DIFF_CONTEXT);
1179 reset = diff_get_color_opt(o, DIFF_RESET);
1180 /*
1181 * Skip the prefix character, if any. With
1182 * diff_suppress_blank_empty, there may be
1183 * none.
1184 */
1185 if (line[0] != '\n') {
1186 line++;
1187 len--;
1188 }
1189 emit_line(o, context, reset, line, len);
1190 break;
1191 case DIFF_SYMBOL_FILEPAIR_PLUS:
1192 meta = diff_get_color_opt(o, DIFF_METAINFO);
1193 reset = diff_get_color_opt(o, DIFF_RESET);
1194 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1195 line, reset,
1196 strchr(line, ' ') ? "\t" : "");
1197 break;
1198 case DIFF_SYMBOL_FILEPAIR_MINUS:
1199 meta = diff_get_color_opt(o, DIFF_METAINFO);
1200 reset = diff_get_color_opt(o, DIFF_RESET);
1201 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1202 line, reset,
1203 strchr(line, ' ') ? "\t" : "");
1204 break;
1205 case DIFF_SYMBOL_BINARY_FILES:
1206 case DIFF_SYMBOL_HEADER:
1207 fprintf(o->file, "%s", line);
1208 break;
1209 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1210 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1211 break;
1212 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1213 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1214 break;
1215 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1216 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1217 break;
1218 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1219 fputs(diff_line_prefix(o), o->file);
1220 fputc('\n', o->file);
1221 break;
1222 case DIFF_SYMBOL_REWRITE_DIFF:
1223 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1224 reset = diff_get_color_opt(o, DIFF_RESET);
1225 emit_line(o, fraginfo, reset, line, len);
1226 break;
1227 case DIFF_SYMBOL_SUBMODULE_ADD:
1228 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1229 reset = diff_get_color_opt(o, DIFF_RESET);
1230 emit_line(o, set, reset, line, len);
1231 break;
1232 case DIFF_SYMBOL_SUBMODULE_DEL:
1233 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1234 reset = diff_get_color_opt(o, DIFF_RESET);
1235 emit_line(o, set, reset, line, len);
1236 break;
1237 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1238 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1239 diff_line_prefix(o), line);
1240 break;
1241 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1242 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1243 diff_line_prefix(o), line);
1244 break;
1245 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1246 emit_line(o, "", "", " 0 files changed\n",
1247 strlen(" 0 files changed\n"));
1248 break;
1249 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1250 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1251 break;
1252 case DIFF_SYMBOL_WORD_DIFF:
1253 fprintf(o->file, "%.*s", len, line);
1254 break;
1255 case DIFF_SYMBOL_STAT_SEP:
1256 fputs(o->stat_sep, o->file);
1257 break;
1258 default:
1259 die("BUG: unknown diff symbol");
1260 }
1261 strbuf_release(&sb);
1262 }
1263
1264 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1265 const char *line, int len, unsigned flags)
1266 {
1267 struct emitted_diff_symbol e = {line, len, flags, s};
1268
1269 if (o->emitted_symbols)
1270 append_emitted_diff_symbol(o, &e);
1271 else
1272 emit_diff_symbol_from_struct(o, &e);
1273 }
1274
1275 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1276 {
1277 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1278 }
1279
1280 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1281 {
1282 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1283 }
1284
1285 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1286 {
1287 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1288 path, strlen(path), 0);
1289 }
1290
1291 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1292 {
1293 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1294 path, strlen(path), 0);
1295 }
1296
1297 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1298 {
1299 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1300 header, strlen(header), 0);
1301 }
1302
1303 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1304 {
1305 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1306 }
1307
1308 void diff_emit_submodule_pipethrough(struct diff_options *o,
1309 const char *line, int len)
1310 {
1311 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1312 }
1313
1314 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1315 {
1316 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1317 ecbdata->blank_at_eof_in_preimage &&
1318 ecbdata->blank_at_eof_in_postimage &&
1319 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1320 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1321 return 0;
1322 return ws_blank_line(line, len, ecbdata->ws_rule);
1323 }
1324
1325 static void emit_add_line(const char *reset,
1326 struct emit_callback *ecbdata,
1327 const char *line, int len)
1328 {
1329 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1330 if (new_blank_line_at_eof(ecbdata, line, len))
1331 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1332
1333 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1334 }
1335
1336 static void emit_del_line(const char *reset,
1337 struct emit_callback *ecbdata,
1338 const char *line, int len)
1339 {
1340 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1341 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1342 }
1343
1344 static void emit_context_line(const char *reset,
1345 struct emit_callback *ecbdata,
1346 const char *line, int len)
1347 {
1348 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1349 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1350 }
1351
1352 static void emit_hunk_header(struct emit_callback *ecbdata,
1353 const char *line, int len)
1354 {
1355 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1356 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1357 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1358 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1359 static const char atat[2] = { '@', '@' };
1360 const char *cp, *ep;
1361 struct strbuf msgbuf = STRBUF_INIT;
1362 int org_len = len;
1363 int i = 1;
1364
1365 /*
1366 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1367 * it always is at least 10 bytes long.
1368 */
1369 if (len < 10 ||
1370 memcmp(line, atat, 2) ||
1371 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1372 emit_diff_symbol(ecbdata->opt,
1373 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1374 return;
1375 }
1376 ep += 2; /* skip over @@ */
1377
1378 /* The hunk header in fraginfo color */
1379 strbuf_addstr(&msgbuf, frag);
1380 strbuf_add(&msgbuf, line, ep - line);
1381 strbuf_addstr(&msgbuf, reset);
1382
1383 /*
1384 * trailing "\r\n"
1385 */
1386 for ( ; i < 3; i++)
1387 if (line[len - i] == '\r' || line[len - i] == '\n')
1388 len--;
1389
1390 /* blank before the func header */
1391 for (cp = ep; ep - line < len; ep++)
1392 if (*ep != ' ' && *ep != '\t')
1393 break;
1394 if (ep != cp) {
1395 strbuf_addstr(&msgbuf, context);
1396 strbuf_add(&msgbuf, cp, ep - cp);
1397 strbuf_addstr(&msgbuf, reset);
1398 }
1399
1400 if (ep < line + len) {
1401 strbuf_addstr(&msgbuf, func);
1402 strbuf_add(&msgbuf, ep, line + len - ep);
1403 strbuf_addstr(&msgbuf, reset);
1404 }
1405
1406 strbuf_add(&msgbuf, line + len, org_len - len);
1407 strbuf_complete_line(&msgbuf);
1408 emit_diff_symbol(ecbdata->opt,
1409 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1410 strbuf_release(&msgbuf);
1411 }
1412
1413 static struct diff_tempfile *claim_diff_tempfile(void) {
1414 int i;
1415 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1416 if (!diff_temp[i].name)
1417 return diff_temp + i;
1418 die("BUG: diff is failing to clean up its tempfiles");
1419 }
1420
1421 static void remove_tempfile(void)
1422 {
1423 int i;
1424 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1425 if (is_tempfile_active(&diff_temp[i].tempfile))
1426 delete_tempfile(&diff_temp[i].tempfile);
1427 diff_temp[i].name = NULL;
1428 }
1429 }
1430
1431 static void add_line_count(struct strbuf *out, int count)
1432 {
1433 switch (count) {
1434 case 0:
1435 strbuf_addstr(out, "0,0");
1436 break;
1437 case 1:
1438 strbuf_addstr(out, "1");
1439 break;
1440 default:
1441 strbuf_addf(out, "1,%d", count);
1442 break;
1443 }
1444 }
1445
1446 static void emit_rewrite_lines(struct emit_callback *ecb,
1447 int prefix, const char *data, int size)
1448 {
1449 const char *endp = NULL;
1450 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1451
1452 while (0 < size) {
1453 int len;
1454
1455 endp = memchr(data, '\n', size);
1456 len = endp ? (endp - data + 1) : size;
1457 if (prefix != '+') {
1458 ecb->lno_in_preimage++;
1459 emit_del_line(reset, ecb, data, len);
1460 } else {
1461 ecb->lno_in_postimage++;
1462 emit_add_line(reset, ecb, data, len);
1463 }
1464 size -= len;
1465 data += len;
1466 }
1467 if (!endp)
1468 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1469 }
1470
1471 static void emit_rewrite_diff(const char *name_a,
1472 const char *name_b,
1473 struct diff_filespec *one,
1474 struct diff_filespec *two,
1475 struct userdiff_driver *textconv_one,
1476 struct userdiff_driver *textconv_two,
1477 struct diff_options *o)
1478 {
1479 int lc_a, lc_b;
1480 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1481 const char *a_prefix, *b_prefix;
1482 char *data_one, *data_two;
1483 size_t size_one, size_two;
1484 struct emit_callback ecbdata;
1485 struct strbuf out = STRBUF_INIT;
1486
1487 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
1488 a_prefix = o->b_prefix;
1489 b_prefix = o->a_prefix;
1490 } else {
1491 a_prefix = o->a_prefix;
1492 b_prefix = o->b_prefix;
1493 }
1494
1495 name_a += (*name_a == '/');
1496 name_b += (*name_b == '/');
1497
1498 strbuf_reset(&a_name);
1499 strbuf_reset(&b_name);
1500 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1501 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1502
1503 size_one = fill_textconv(textconv_one, one, &data_one);
1504 size_two = fill_textconv(textconv_two, two, &data_two);
1505
1506 memset(&ecbdata, 0, sizeof(ecbdata));
1507 ecbdata.color_diff = want_color(o->use_color);
1508 ecbdata.ws_rule = whitespace_rule(name_b);
1509 ecbdata.opt = o;
1510 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1511 mmfile_t mf1, mf2;
1512 mf1.ptr = (char *)data_one;
1513 mf2.ptr = (char *)data_two;
1514 mf1.size = size_one;
1515 mf2.size = size_two;
1516 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1517 }
1518 ecbdata.lno_in_preimage = 1;
1519 ecbdata.lno_in_postimage = 1;
1520
1521 lc_a = count_lines(data_one, size_one);
1522 lc_b = count_lines(data_two, size_two);
1523
1524 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1525 a_name.buf, a_name.len, 0);
1526 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1527 b_name.buf, b_name.len, 0);
1528
1529 strbuf_addstr(&out, "@@ -");
1530 if (!o->irreversible_delete)
1531 add_line_count(&out, lc_a);
1532 else
1533 strbuf_addstr(&out, "?,?");
1534 strbuf_addstr(&out, " +");
1535 add_line_count(&out, lc_b);
1536 strbuf_addstr(&out, " @@\n");
1537 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1538 strbuf_release(&out);
1539
1540 if (lc_a && !o->irreversible_delete)
1541 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1542 if (lc_b)
1543 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1544 if (textconv_one)
1545 free((char *)data_one);
1546 if (textconv_two)
1547 free((char *)data_two);
1548 }
1549
1550 struct diff_words_buffer {
1551 mmfile_t text;
1552 long alloc;
1553 struct diff_words_orig {
1554 const char *begin, *end;
1555 } *orig;
1556 int orig_nr, orig_alloc;
1557 };
1558
1559 static void diff_words_append(char *line, unsigned long len,
1560 struct diff_words_buffer *buffer)
1561 {
1562 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1563 line++;
1564 len--;
1565 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1566 buffer->text.size += len;
1567 buffer->text.ptr[buffer->text.size] = '\0';
1568 }
1569
1570 struct diff_words_style_elem {
1571 const char *prefix;
1572 const char *suffix;
1573 const char *color; /* NULL; filled in by the setup code if
1574 * color is enabled */
1575 };
1576
1577 struct diff_words_style {
1578 enum diff_words_type type;
1579 struct diff_words_style_elem new, old, ctx;
1580 const char *newline;
1581 };
1582
1583 static struct diff_words_style diff_words_styles[] = {
1584 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1585 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1586 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1587 };
1588
1589 struct diff_words_data {
1590 struct diff_words_buffer minus, plus;
1591 const char *current_plus;
1592 int last_minus;
1593 struct diff_options *opt;
1594 regex_t *word_regex;
1595 enum diff_words_type type;
1596 struct diff_words_style *style;
1597 };
1598
1599 static int fn_out_diff_words_write_helper(struct diff_options *o,
1600 struct diff_words_style_elem *st_el,
1601 const char *newline,
1602 size_t count, const char *buf)
1603 {
1604 int print = 0;
1605 struct strbuf sb = STRBUF_INIT;
1606
1607 while (count) {
1608 char *p = memchr(buf, '\n', count);
1609 if (print)
1610 strbuf_addstr(&sb, diff_line_prefix(o));
1611
1612 if (p != buf) {
1613 const char *reset = st_el->color && *st_el->color ?
1614 GIT_COLOR_RESET : NULL;
1615 if (st_el->color && *st_el->color)
1616 strbuf_addstr(&sb, st_el->color);
1617 strbuf_addstr(&sb, st_el->prefix);
1618 strbuf_add(&sb, buf, p ? p - buf : count);
1619 strbuf_addstr(&sb, st_el->suffix);
1620 if (reset)
1621 strbuf_addstr(&sb, reset);
1622 }
1623 if (!p)
1624 goto out;
1625
1626 strbuf_addstr(&sb, newline);
1627 count -= p + 1 - buf;
1628 buf = p + 1;
1629 print = 1;
1630 if (count) {
1631 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1632 sb.buf, sb.len, 0);
1633 strbuf_reset(&sb);
1634 }
1635 }
1636
1637 out:
1638 if (sb.len)
1639 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1640 sb.buf, sb.len, 0);
1641 strbuf_release(&sb);
1642 return 0;
1643 }
1644
1645 /*
1646 * '--color-words' algorithm can be described as:
1647 *
1648 * 1. collect the minus/plus lines of a diff hunk, divided into
1649 * minus-lines and plus-lines;
1650 *
1651 * 2. break both minus-lines and plus-lines into words and
1652 * place them into two mmfile_t with one word for each line;
1653 *
1654 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1655 *
1656 * And for the common parts of the both file, we output the plus side text.
1657 * diff_words->current_plus is used to trace the current position of the plus file
1658 * which printed. diff_words->last_minus is used to trace the last minus word
1659 * printed.
1660 *
1661 * For '--graph' to work with '--color-words', we need to output the graph prefix
1662 * on each line of color words output. Generally, there are two conditions on
1663 * which we should output the prefix.
1664 *
1665 * 1. diff_words->last_minus == 0 &&
1666 * diff_words->current_plus == diff_words->plus.text.ptr
1667 *
1668 * that is: the plus text must start as a new line, and if there is no minus
1669 * word printed, a graph prefix must be printed.
1670 *
1671 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1672 * *(diff_words->current_plus - 1) == '\n'
1673 *
1674 * that is: a graph prefix must be printed following a '\n'
1675 */
1676 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1677 {
1678 if ((diff_words->last_minus == 0 &&
1679 diff_words->current_plus == diff_words->plus.text.ptr) ||
1680 (diff_words->current_plus > diff_words->plus.text.ptr &&
1681 *(diff_words->current_plus - 1) == '\n')) {
1682 return 1;
1683 } else {
1684 return 0;
1685 }
1686 }
1687
1688 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1689 {
1690 struct diff_words_data *diff_words = priv;
1691 struct diff_words_style *style = diff_words->style;
1692 int minus_first, minus_len, plus_first, plus_len;
1693 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1694 struct diff_options *opt = diff_words->opt;
1695 const char *line_prefix;
1696
1697 if (line[0] != '@' || parse_hunk_header(line, len,
1698 &minus_first, &minus_len, &plus_first, &plus_len))
1699 return;
1700
1701 assert(opt);
1702 line_prefix = diff_line_prefix(opt);
1703
1704 /* POSIX requires that first be decremented by one if len == 0... */
1705 if (minus_len) {
1706 minus_begin = diff_words->minus.orig[minus_first].begin;
1707 minus_end =
1708 diff_words->minus.orig[minus_first + minus_len - 1].end;
1709 } else
1710 minus_begin = minus_end =
1711 diff_words->minus.orig[minus_first].end;
1712
1713 if (plus_len) {
1714 plus_begin = diff_words->plus.orig[plus_first].begin;
1715 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1716 } else
1717 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1718
1719 if (color_words_output_graph_prefix(diff_words)) {
1720 fputs(line_prefix, diff_words->opt->file);
1721 }
1722 if (diff_words->current_plus != plus_begin) {
1723 fn_out_diff_words_write_helper(diff_words->opt,
1724 &style->ctx, style->newline,
1725 plus_begin - diff_words->current_plus,
1726 diff_words->current_plus);
1727 }
1728 if (minus_begin != minus_end) {
1729 fn_out_diff_words_write_helper(diff_words->opt,
1730 &style->old, style->newline,
1731 minus_end - minus_begin, minus_begin);
1732 }
1733 if (plus_begin != plus_end) {
1734 fn_out_diff_words_write_helper(diff_words->opt,
1735 &style->new, style->newline,
1736 plus_end - plus_begin, plus_begin);
1737 }
1738
1739 diff_words->current_plus = plus_end;
1740 diff_words->last_minus = minus_first;
1741 }
1742
1743 /* This function starts looking at *begin, and returns 0 iff a word was found. */
1744 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1745 int *begin, int *end)
1746 {
1747 if (word_regex && *begin < buffer->size) {
1748 regmatch_t match[1];
1749 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1750 buffer->size - *begin, 1, match, 0)) {
1751 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1752 '\n', match[0].rm_eo - match[0].rm_so);
1753 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1754 *begin += match[0].rm_so;
1755 return *begin >= *end;
1756 }
1757 return -1;
1758 }
1759
1760 /* find the next word */
1761 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1762 (*begin)++;
1763 if (*begin >= buffer->size)
1764 return -1;
1765
1766 /* find the end of the word */
1767 *end = *begin + 1;
1768 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1769 (*end)++;
1770
1771 return 0;
1772 }
1773
1774 /*
1775 * This function splits the words in buffer->text, stores the list with
1776 * newline separator into out, and saves the offsets of the original words
1777 * in buffer->orig.
1778 */
1779 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1780 regex_t *word_regex)
1781 {
1782 int i, j;
1783 long alloc = 0;
1784
1785 out->size = 0;
1786 out->ptr = NULL;
1787
1788 /* fake an empty "0th" word */
1789 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1790 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1791 buffer->orig_nr = 1;
1792
1793 for (i = 0; i < buffer->text.size; i++) {
1794 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1795 return;
1796
1797 /* store original boundaries */
1798 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1799 buffer->orig_alloc);
1800 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1801 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1802 buffer->orig_nr++;
1803
1804 /* store one word */
1805 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1806 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1807 out->ptr[out->size + j - i] = '\n';
1808 out->size += j - i + 1;
1809
1810 i = j - 1;
1811 }
1812 }
1813
1814 /* this executes the word diff on the accumulated buffers */
1815 static void diff_words_show(struct diff_words_data *diff_words)
1816 {
1817 xpparam_t xpp;
1818 xdemitconf_t xecfg;
1819 mmfile_t minus, plus;
1820 struct diff_words_style *style = diff_words->style;
1821
1822 struct diff_options *opt = diff_words->opt;
1823 const char *line_prefix;
1824
1825 assert(opt);
1826 line_prefix = diff_line_prefix(opt);
1827
1828 /* special case: only removal */
1829 if (!diff_words->plus.text.size) {
1830 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1831 line_prefix, strlen(line_prefix), 0);
1832 fn_out_diff_words_write_helper(diff_words->opt,
1833 &style->old, style->newline,
1834 diff_words->minus.text.size,
1835 diff_words->minus.text.ptr);
1836 diff_words->minus.text.size = 0;
1837 return;
1838 }
1839
1840 diff_words->current_plus = diff_words->plus.text.ptr;
1841 diff_words->last_minus = 0;
1842
1843 memset(&xpp, 0, sizeof(xpp));
1844 memset(&xecfg, 0, sizeof(xecfg));
1845 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1846 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1847 xpp.flags = 0;
1848 /* as only the hunk header will be parsed, we need a 0-context */
1849 xecfg.ctxlen = 0;
1850 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1851 &xpp, &xecfg))
1852 die("unable to generate word diff");
1853 free(minus.ptr);
1854 free(plus.ptr);
1855 if (diff_words->current_plus != diff_words->plus.text.ptr +
1856 diff_words->plus.text.size) {
1857 if (color_words_output_graph_prefix(diff_words))
1858 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1859 line_prefix, strlen(line_prefix), 0);
1860 fn_out_diff_words_write_helper(diff_words->opt,
1861 &style->ctx, style->newline,
1862 diff_words->plus.text.ptr + diff_words->plus.text.size
1863 - diff_words->current_plus, diff_words->current_plus);
1864 }
1865 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1866 }
1867
1868 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1869 static void diff_words_flush(struct emit_callback *ecbdata)
1870 {
1871 struct diff_options *wo = ecbdata->diff_words->opt;
1872
1873 if (ecbdata->diff_words->minus.text.size ||
1874 ecbdata->diff_words->plus.text.size)
1875 diff_words_show(ecbdata->diff_words);
1876
1877 if (wo->emitted_symbols) {
1878 struct diff_options *o = ecbdata->opt;
1879 struct emitted_diff_symbols *wol = wo->emitted_symbols;
1880 int i;
1881
1882 /*
1883 * NEEDSWORK:
1884 * Instead of appending each, concat all words to a line?
1885 */
1886 for (i = 0; i < wol->nr; i++)
1887 append_emitted_diff_symbol(o, &wol->buf[i]);
1888
1889 for (i = 0; i < wol->nr; i++)
1890 free((void *)wol->buf[i].line);
1891
1892 wol->nr = 0;
1893 }
1894 }
1895
1896 static void diff_filespec_load_driver(struct diff_filespec *one)
1897 {
1898 /* Use already-loaded driver */
1899 if (one->driver)
1900 return;
1901
1902 if (S_ISREG(one->mode))
1903 one->driver = userdiff_find_by_path(one->path);
1904
1905 /* Fallback to default settings */
1906 if (!one->driver)
1907 one->driver = userdiff_find_by_name("default");
1908 }
1909
1910 static const char *userdiff_word_regex(struct diff_filespec *one)
1911 {
1912 diff_filespec_load_driver(one);
1913 return one->driver->word_regex;
1914 }
1915
1916 static void init_diff_words_data(struct emit_callback *ecbdata,
1917 struct diff_options *orig_opts,
1918 struct diff_filespec *one,
1919 struct diff_filespec *two)
1920 {
1921 int i;
1922 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1923 memcpy(o, orig_opts, sizeof(struct diff_options));
1924
1925 ecbdata->diff_words =
1926 xcalloc(1, sizeof(struct diff_words_data));
1927 ecbdata->diff_words->type = o->word_diff;
1928 ecbdata->diff_words->opt = o;
1929
1930 if (orig_opts->emitted_symbols)
1931 o->emitted_symbols =
1932 xcalloc(1, sizeof(struct emitted_diff_symbols));
1933
1934 if (!o->word_regex)
1935 o->word_regex = userdiff_word_regex(one);
1936 if (!o->word_regex)
1937 o->word_regex = userdiff_word_regex(two);
1938 if (!o->word_regex)
1939 o->word_regex = diff_word_regex_cfg;
1940 if (o->word_regex) {
1941 ecbdata->diff_words->word_regex = (regex_t *)
1942 xmalloc(sizeof(regex_t));
1943 if (regcomp(ecbdata->diff_words->word_regex,
1944 o->word_regex,
1945 REG_EXTENDED | REG_NEWLINE))
1946 die ("Invalid regular expression: %s",
1947 o->word_regex);
1948 }
1949 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1950 if (o->word_diff == diff_words_styles[i].type) {
1951 ecbdata->diff_words->style =
1952 &diff_words_styles[i];
1953 break;
1954 }
1955 }
1956 if (want_color(o->use_color)) {
1957 struct diff_words_style *st = ecbdata->diff_words->style;
1958 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1959 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1960 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1961 }
1962 }
1963
1964 static void free_diff_words_data(struct emit_callback *ecbdata)
1965 {
1966 if (ecbdata->diff_words) {
1967 diff_words_flush(ecbdata);
1968 free (ecbdata->diff_words->opt->emitted_symbols);
1969 free (ecbdata->diff_words->opt);
1970 free (ecbdata->diff_words->minus.text.ptr);
1971 free (ecbdata->diff_words->minus.orig);
1972 free (ecbdata->diff_words->plus.text.ptr);
1973 free (ecbdata->diff_words->plus.orig);
1974 if (ecbdata->diff_words->word_regex) {
1975 regfree(ecbdata->diff_words->word_regex);
1976 free(ecbdata->diff_words->word_regex);
1977 }
1978 FREE_AND_NULL(ecbdata->diff_words);
1979 }
1980 }
1981
1982 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1983 {
1984 if (want_color(diff_use_color))
1985 return diff_colors[ix];
1986 return "";
1987 }
1988
1989 const char *diff_line_prefix(struct diff_options *opt)
1990 {
1991 struct strbuf *msgbuf;
1992 if (!opt->output_prefix)
1993 return "";
1994
1995 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1996 return msgbuf->buf;
1997 }
1998
1999 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
2000 {
2001 const char *cp;
2002 unsigned long allot;
2003 size_t l = len;
2004
2005 if (ecb->truncate)
2006 return ecb->truncate(line, len);
2007 cp = line;
2008 allot = l;
2009 while (0 < l) {
2010 (void) utf8_width(&cp, &l);
2011 if (!cp)
2012 break; /* truncated in the middle? */
2013 }
2014 return allot - l;
2015 }
2016
2017 static void find_lno(const char *line, struct emit_callback *ecbdata)
2018 {
2019 const char *p;
2020 ecbdata->lno_in_preimage = 0;
2021 ecbdata->lno_in_postimage = 0;
2022 p = strchr(line, '-');
2023 if (!p)
2024 return; /* cannot happen */
2025 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2026 p = strchr(p, '+');
2027 if (!p)
2028 return; /* cannot happen */
2029 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2030 }
2031
2032 static void fn_out_consume(void *priv, char *line, unsigned long len)
2033 {
2034 struct emit_callback *ecbdata = priv;
2035 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2036 struct diff_options *o = ecbdata->opt;
2037
2038 o->found_changes = 1;
2039
2040 if (ecbdata->header) {
2041 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2042 ecbdata->header->buf, ecbdata->header->len, 0);
2043 strbuf_reset(ecbdata->header);
2044 ecbdata->header = NULL;
2045 }
2046
2047 if (ecbdata->label_path[0]) {
2048 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2049 ecbdata->label_path[0],
2050 strlen(ecbdata->label_path[0]), 0);
2051 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2052 ecbdata->label_path[1],
2053 strlen(ecbdata->label_path[1]), 0);
2054 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2055 }
2056
2057 if (diff_suppress_blank_empty
2058 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2059 line[0] = '\n';
2060 len = 1;
2061 }
2062
2063 if (line[0] == '@') {
2064 if (ecbdata->diff_words)
2065 diff_words_flush(ecbdata);
2066 len = sane_truncate_line(ecbdata, line, len);
2067 find_lno(line, ecbdata);
2068 emit_hunk_header(ecbdata, line, len);
2069 return;
2070 }
2071
2072 if (ecbdata->diff_words) {
2073 enum diff_symbol s =
2074 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2075 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2076 if (line[0] == '-') {
2077 diff_words_append(line, len,
2078 &ecbdata->diff_words->minus);
2079 return;
2080 } else if (line[0] == '+') {
2081 diff_words_append(line, len,
2082 &ecbdata->diff_words->plus);
2083 return;
2084 } else if (starts_with(line, "\\ ")) {
2085 /*
2086 * Eat the "no newline at eof" marker as if we
2087 * saw a "+" or "-" line with nothing on it,
2088 * and return without diff_words_flush() to
2089 * defer processing. If this is the end of
2090 * preimage, more "+" lines may come after it.
2091 */
2092 return;
2093 }
2094 diff_words_flush(ecbdata);
2095 emit_diff_symbol(o, s, line, len, 0);
2096 return;
2097 }
2098
2099 switch (line[0]) {
2100 case '+':
2101 ecbdata->lno_in_postimage++;
2102 emit_add_line(reset, ecbdata, line + 1, len - 1);
2103 break;
2104 case '-':
2105 ecbdata->lno_in_preimage++;
2106 emit_del_line(reset, ecbdata, line + 1, len - 1);
2107 break;
2108 case ' ':
2109 ecbdata->lno_in_postimage++;
2110 ecbdata->lno_in_preimage++;
2111 emit_context_line(reset, ecbdata, line + 1, len - 1);
2112 break;
2113 default:
2114 /* incomplete line at the end */
2115 ecbdata->lno_in_preimage++;
2116 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2117 line, len, 0);
2118 break;
2119 }
2120 }
2121
2122 static char *pprint_rename(const char *a, const char *b)
2123 {
2124 const char *old = a;
2125 const char *new = b;
2126 struct strbuf name = STRBUF_INIT;
2127 int pfx_length, sfx_length;
2128 int pfx_adjust_for_slash;
2129 int len_a = strlen(a);
2130 int len_b = strlen(b);
2131 int a_midlen, b_midlen;
2132 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2133 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2134
2135 if (qlen_a || qlen_b) {
2136 quote_c_style(a, &name, NULL, 0);
2137 strbuf_addstr(&name, " => ");
2138 quote_c_style(b, &name, NULL, 0);
2139 return strbuf_detach(&name, NULL);
2140 }
2141
2142 /* Find common prefix */
2143 pfx_length = 0;
2144 while (*old && *new && *old == *new) {
2145 if (*old == '/')
2146 pfx_length = old - a + 1;
2147 old++;
2148 new++;
2149 }
2150
2151 /* Find common suffix */
2152 old = a + len_a;
2153 new = b + len_b;
2154 sfx_length = 0;
2155 /*
2156 * If there is a common prefix, it must end in a slash. In
2157 * that case we let this loop run 1 into the prefix to see the
2158 * same slash.
2159 *
2160 * If there is no common prefix, we cannot do this as it would
2161 * underrun the input strings.
2162 */
2163 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2164 while (a + pfx_length - pfx_adjust_for_slash <= old &&
2165 b + pfx_length - pfx_adjust_for_slash <= new &&
2166 *old == *new) {
2167 if (*old == '/')
2168 sfx_length = len_a - (old - a);
2169 old--;
2170 new--;
2171 }
2172
2173 /*
2174 * pfx{mid-a => mid-b}sfx
2175 * {pfx-a => pfx-b}sfx
2176 * pfx{sfx-a => sfx-b}
2177 * name-a => name-b
2178 */
2179 a_midlen = len_a - pfx_length - sfx_length;
2180 b_midlen = len_b - pfx_length - sfx_length;
2181 if (a_midlen < 0)
2182 a_midlen = 0;
2183 if (b_midlen < 0)
2184 b_midlen = 0;
2185
2186 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2187 if (pfx_length + sfx_length) {
2188 strbuf_add(&name, a, pfx_length);
2189 strbuf_addch(&name, '{');
2190 }
2191 strbuf_add(&name, a + pfx_length, a_midlen);
2192 strbuf_addstr(&name, " => ");
2193 strbuf_add(&name, b + pfx_length, b_midlen);
2194 if (pfx_length + sfx_length) {
2195 strbuf_addch(&name, '}');
2196 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
2197 }
2198 return strbuf_detach(&name, NULL);
2199 }
2200
2201 struct diffstat_t {
2202 int nr;
2203 int alloc;
2204 struct diffstat_file {
2205 char *from_name;
2206 char *name;
2207 char *print_name;
2208 unsigned is_unmerged:1;
2209 unsigned is_binary:1;
2210 unsigned is_renamed:1;
2211 unsigned is_interesting:1;
2212 uintmax_t added, deleted;
2213 } **files;
2214 };
2215
2216 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2217 const char *name_a,
2218 const char *name_b)
2219 {
2220 struct diffstat_file *x;
2221 x = xcalloc(1, sizeof(*x));
2222 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2223 diffstat->files[diffstat->nr++] = x;
2224 if (name_b) {
2225 x->from_name = xstrdup(name_a);
2226 x->name = xstrdup(name_b);
2227 x->is_renamed = 1;
2228 }
2229 else {
2230 x->from_name = NULL;
2231 x->name = xstrdup(name_a);
2232 }
2233 return x;
2234 }
2235
2236 static void diffstat_consume(void *priv, char *line, unsigned long len)
2237 {
2238 struct diffstat_t *diffstat = priv;
2239 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2240
2241 if (line[0] == '+')
2242 x->added++;
2243 else if (line[0] == '-')
2244 x->deleted++;
2245 }
2246
2247 const char mime_boundary_leader[] = "------------";
2248
2249 static int scale_linear(int it, int width, int max_change)
2250 {
2251 if (!it)
2252 return 0;
2253 /*
2254 * make sure that at least one '-' or '+' is printed if
2255 * there is any change to this path. The easiest way is to
2256 * scale linearly as if the alloted width is one column shorter
2257 * than it is, and then add 1 to the result.
2258 */
2259 return 1 + (it * (width - 1) / max_change);
2260 }
2261
2262 static void show_graph(struct strbuf *out, char ch, int cnt,
2263 const char *set, const char *reset)
2264 {
2265 if (cnt <= 0)
2266 return;
2267 strbuf_addstr(out, set);
2268 strbuf_addchars(out, ch, cnt);
2269 strbuf_addstr(out, reset);
2270 }
2271
2272 static void fill_print_name(struct diffstat_file *file)
2273 {
2274 char *pname;
2275
2276 if (file->print_name)
2277 return;
2278
2279 if (!file->is_renamed) {
2280 struct strbuf buf = STRBUF_INIT;
2281 if (quote_c_style(file->name, &buf, NULL, 0)) {
2282 pname = strbuf_detach(&buf, NULL);
2283 } else {
2284 pname = file->name;
2285 strbuf_release(&buf);
2286 }
2287 } else {
2288 pname = pprint_rename(file->from_name, file->name);
2289 }
2290 file->print_name = pname;
2291 }
2292
2293 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2294 int files, int insertions, int deletions)
2295 {
2296 struct strbuf sb = STRBUF_INIT;
2297
2298 if (!files) {
2299 assert(insertions == 0 && deletions == 0);
2300 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2301 NULL, 0, 0);
2302 return;
2303 }
2304
2305 strbuf_addf(&sb,
2306 (files == 1) ? " %d file changed" : " %d files changed",
2307 files);
2308
2309 /*
2310 * For binary diff, the caller may want to print "x files
2311 * changed" with insertions == 0 && deletions == 0.
2312 *
2313 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2314 * is probably less confusing (i.e skip over "2 files changed
2315 * but nothing about added/removed lines? Is this a bug in Git?").
2316 */
2317 if (insertions || deletions == 0) {
2318 strbuf_addf(&sb,
2319 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2320 insertions);
2321 }
2322
2323 if (deletions || insertions == 0) {
2324 strbuf_addf(&sb,
2325 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2326 deletions);
2327 }
2328 strbuf_addch(&sb, '\n');
2329 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2330 sb.buf, sb.len, 0);
2331 strbuf_release(&sb);
2332 }
2333
2334 void print_stat_summary(FILE *fp, int files,
2335 int insertions, int deletions)
2336 {
2337 struct diff_options o;
2338 memset(&o, 0, sizeof(o));
2339 o.file = fp;
2340
2341 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2342 }
2343
2344 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2345 {
2346 int i, len, add, del, adds = 0, dels = 0;
2347 uintmax_t max_change = 0, max_len = 0;
2348 int total_files = data->nr, count;
2349 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2350 const char *reset, *add_c, *del_c;
2351 int extra_shown = 0;
2352 const char *line_prefix = diff_line_prefix(options);
2353 struct strbuf out = STRBUF_INIT;
2354
2355 if (data->nr == 0)
2356 return;
2357
2358 count = options->stat_count ? options->stat_count : data->nr;
2359
2360 reset = diff_get_color_opt(options, DIFF_RESET);
2361 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2362 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2363
2364 /*
2365 * Find the longest filename and max number of changes
2366 */
2367 for (i = 0; (i < count) && (i < data->nr); i++) {
2368 struct diffstat_file *file = data->files[i];
2369 uintmax_t change = file->added + file->deleted;
2370
2371 if (!file->is_interesting && (change == 0)) {
2372 count++; /* not shown == room for one more */
2373 continue;
2374 }
2375 fill_print_name(file);
2376 len = strlen(file->print_name);
2377 if (max_len < len)
2378 max_len = len;
2379
2380 if (file->is_unmerged) {
2381 /* "Unmerged" is 8 characters */
2382 bin_width = bin_width < 8 ? 8 : bin_width;
2383 continue;
2384 }
2385 if (file->is_binary) {
2386 /* "Bin XXX -> YYY bytes" */
2387 int w = 14 + decimal_width(file->added)
2388 + decimal_width(file->deleted);
2389 bin_width = bin_width < w ? w : bin_width;
2390 /* Display change counts aligned with "Bin" */
2391 number_width = 3;
2392 continue;
2393 }
2394
2395 if (max_change < change)
2396 max_change = change;
2397 }
2398 count = i; /* where we can stop scanning in data->files[] */
2399
2400 /*
2401 * We have width = stat_width or term_columns() columns total.
2402 * We want a maximum of min(max_len, stat_name_width) for the name part.
2403 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2404 * We also need 1 for " " and 4 + decimal_width(max_change)
2405 * for " | NNNN " and one the empty column at the end, altogether
2406 * 6 + decimal_width(max_change).
2407 *
2408 * If there's not enough space, we will use the smaller of
2409 * stat_name_width (if set) and 5/8*width for the filename,
2410 * and the rest for constant elements + graph part, but no more
2411 * than stat_graph_width for the graph part.
2412 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2413 * for the standard terminal size).
2414 *
2415 * In other words: stat_width limits the maximum width, and
2416 * stat_name_width fixes the maximum width of the filename,
2417 * and is also used to divide available columns if there
2418 * aren't enough.
2419 *
2420 * Binary files are displayed with "Bin XXX -> YYY bytes"
2421 * instead of the change count and graph. This part is treated
2422 * similarly to the graph part, except that it is not
2423 * "scaled". If total width is too small to accommodate the
2424 * guaranteed minimum width of the filename part and the
2425 * separators and this message, this message will "overflow"
2426 * making the line longer than the maximum width.
2427 */
2428
2429 if (options->stat_width == -1)
2430 width = term_columns() - strlen(line_prefix);
2431 else
2432 width = options->stat_width ? options->stat_width : 80;
2433 number_width = decimal_width(max_change) > number_width ?
2434 decimal_width(max_change) : number_width;
2435
2436 if (options->stat_graph_width == -1)
2437 options->stat_graph_width = diff_stat_graph_width;
2438
2439 /*
2440 * Guarantee 3/8*16==6 for the graph part
2441 * and 5/8*16==10 for the filename part
2442 */
2443 if (width < 16 + 6 + number_width)
2444 width = 16 + 6 + number_width;
2445
2446 /*
2447 * First assign sizes that are wanted, ignoring available width.
2448 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2449 * starting from "XXX" should fit in graph_width.
2450 */
2451 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2452 if (options->stat_graph_width &&
2453 options->stat_graph_width < graph_width)
2454 graph_width = options->stat_graph_width;
2455
2456 name_width = (options->stat_name_width > 0 &&
2457 options->stat_name_width < max_len) ?
2458 options->stat_name_width : max_len;
2459
2460 /*
2461 * Adjust adjustable widths not to exceed maximum width
2462 */
2463 if (name_width + number_width + 6 + graph_width > width) {
2464 if (graph_width > width * 3/8 - number_width - 6) {
2465 graph_width = width * 3/8 - number_width - 6;
2466 if (graph_width < 6)
2467 graph_width = 6;
2468 }
2469
2470 if (options->stat_graph_width &&
2471 graph_width > options->stat_graph_width)
2472 graph_width = options->stat_graph_width;
2473 if (name_width > width - number_width - 6 - graph_width)
2474 name_width = width - number_width - 6 - graph_width;
2475 else
2476 graph_width = width - number_width - 6 - name_width;
2477 }
2478
2479 /*
2480 * From here name_width is the width of the name area,
2481 * and graph_width is the width of the graph area.
2482 * max_change is used to scale graph properly.
2483 */
2484 for (i = 0; i < count; i++) {
2485 const char *prefix = "";
2486 struct diffstat_file *file = data->files[i];
2487 char *name = file->print_name;
2488 uintmax_t added = file->added;
2489 uintmax_t deleted = file->deleted;
2490 int name_len;
2491
2492 if (!file->is_interesting && (added + deleted == 0))
2493 continue;
2494
2495 /*
2496 * "scale" the filename
2497 */
2498 len = name_width;
2499 name_len = strlen(name);
2500 if (name_width < name_len) {
2501 char *slash;
2502 prefix = "...";
2503 len -= 3;
2504 name += name_len - len;
2505 slash = strchr(name, '/');
2506 if (slash)
2507 name = slash;
2508 }
2509
2510 if (file->is_binary) {
2511 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2512 strbuf_addf(&out, " %*s", number_width, "Bin");
2513 if (!added && !deleted) {
2514 strbuf_addch(&out, '\n');
2515 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2516 out.buf, out.len, 0);
2517 strbuf_reset(&out);
2518 continue;
2519 }
2520 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2521 del_c, deleted, reset);
2522 strbuf_addstr(&out, " -> ");
2523 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2524 add_c, added, reset);
2525 strbuf_addstr(&out, " bytes\n");
2526 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2527 out.buf, out.len, 0);
2528 strbuf_reset(&out);
2529 continue;
2530 }
2531 else if (file->is_unmerged) {
2532 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2533 strbuf_addstr(&out, " Unmerged\n");
2534 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2535 out.buf, out.len, 0);
2536 strbuf_reset(&out);
2537 continue;
2538 }
2539
2540 /*
2541 * scale the add/delete
2542 */
2543 add = added;
2544 del = deleted;
2545
2546 if (graph_width <= max_change) {
2547 int total = scale_linear(add + del, graph_width, max_change);
2548 if (total < 2 && add && del)
2549 /* width >= 2 due to the sanity check */
2550 total = 2;
2551 if (add < del) {
2552 add = scale_linear(add, graph_width, max_change);
2553 del = total - add;
2554 } else {
2555 del = scale_linear(del, graph_width, max_change);
2556 add = total - del;
2557 }
2558 }
2559 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2560 strbuf_addf(&out, " %*"PRIuMAX"%s",
2561 number_width, added + deleted,
2562 added + deleted ? " " : "");
2563 show_graph(&out, '+', add, add_c, reset);
2564 show_graph(&out, '-', del, del_c, reset);
2565 strbuf_addch(&out, '\n');
2566 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2567 out.buf, out.len, 0);
2568 strbuf_reset(&out);
2569 }
2570
2571 for (i = 0; i < data->nr; i++) {
2572 struct diffstat_file *file = data->files[i];
2573 uintmax_t added = file->added;
2574 uintmax_t deleted = file->deleted;
2575
2576 if (file->is_unmerged ||
2577 (!file->is_interesting && (added + deleted == 0))) {
2578 total_files--;
2579 continue;
2580 }
2581
2582 if (!file->is_binary) {
2583 adds += added;
2584 dels += deleted;
2585 }
2586 if (i < count)
2587 continue;
2588 if (!extra_shown)
2589 emit_diff_symbol(options,
2590 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2591 NULL, 0, 0);
2592 extra_shown = 1;
2593 }
2594
2595 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2596 }
2597
2598 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2599 {
2600 int i, adds = 0, dels = 0, total_files = data->nr;
2601
2602 if (data->nr == 0)
2603 return;
2604
2605 for (i = 0; i < data->nr; i++) {
2606 int added = data->files[i]->added;
2607 int deleted = data->files[i]->deleted;
2608
2609 if (data->files[i]->is_unmerged ||
2610 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2611 total_files--;
2612 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2613 adds += added;
2614 dels += deleted;
2615 }
2616 }
2617 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2618 }
2619
2620 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2621 {
2622 int i;
2623
2624 if (data->nr == 0)
2625 return;
2626
2627 for (i = 0; i < data->nr; i++) {
2628 struct diffstat_file *file = data->files[i];
2629
2630 fprintf(options->file, "%s", diff_line_prefix(options));
2631
2632 if (file->is_binary)
2633 fprintf(options->file, "-\t-\t");
2634 else
2635 fprintf(options->file,
2636 "%"PRIuMAX"\t%"PRIuMAX"\t",
2637 file->added, file->deleted);
2638 if (options->line_termination) {
2639 fill_print_name(file);
2640 if (!file->is_renamed)
2641 write_name_quoted(file->name, options->file,
2642 options->line_termination);
2643 else {
2644 fputs(file->print_name, options->file);
2645 putc(options->line_termination, options->file);
2646 }
2647 } else {
2648 if (file->is_renamed) {
2649 putc('\0', options->file);
2650 write_name_quoted(file->from_name, options->file, '\0');
2651 }
2652 write_name_quoted(file->name, options->file, '\0');
2653 }
2654 }
2655 }
2656
2657 struct dirstat_file {
2658 const char *name;
2659 unsigned long changed;
2660 };
2661
2662 struct dirstat_dir {
2663 struct dirstat_file *files;
2664 int alloc, nr, permille, cumulative;
2665 };
2666
2667 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2668 unsigned long changed, const char *base, int baselen)
2669 {
2670 unsigned long this_dir = 0;
2671 unsigned int sources = 0;
2672 const char *line_prefix = diff_line_prefix(opt);
2673
2674 while (dir->nr) {
2675 struct dirstat_file *f = dir->files;
2676 int namelen = strlen(f->name);
2677 unsigned long this;
2678 char *slash;
2679
2680 if (namelen < baselen)
2681 break;
2682 if (memcmp(f->name, base, baselen))
2683 break;
2684 slash = strchr(f->name + baselen, '/');
2685 if (slash) {
2686 int newbaselen = slash + 1 - f->name;
2687 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2688 sources++;
2689 } else {
2690 this = f->changed;
2691 dir->files++;
2692 dir->nr--;
2693 sources += 2;
2694 }
2695 this_dir += this;
2696 }
2697
2698 /*
2699 * We don't report dirstat's for
2700 * - the top level
2701 * - or cases where everything came from a single directory
2702 * under this directory (sources == 1).
2703 */
2704 if (baselen && sources != 1) {
2705 if (this_dir) {
2706 int permille = this_dir * 1000 / changed;
2707 if (permille >= dir->permille) {
2708 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2709 permille / 10, permille % 10, baselen, base);
2710 if (!dir->cumulative)
2711 return 0;
2712 }
2713 }
2714 }
2715 return this_dir;
2716 }
2717
2718 static int dirstat_compare(const void *_a, const void *_b)
2719 {
2720 const struct dirstat_file *a = _a;
2721 const struct dirstat_file *b = _b;
2722 return strcmp(a->name, b->name);
2723 }
2724
2725 static void show_dirstat(struct diff_options *options)
2726 {
2727 int i;
2728 unsigned long changed;
2729 struct dirstat_dir dir;
2730 struct diff_queue_struct *q = &diff_queued_diff;
2731
2732 dir.files = NULL;
2733 dir.alloc = 0;
2734 dir.nr = 0;
2735 dir.permille = options->dirstat_permille;
2736 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2737
2738 changed = 0;
2739 for (i = 0; i < q->nr; i++) {
2740 struct diff_filepair *p = q->queue[i];
2741 const char *name;
2742 unsigned long copied, added, damage;
2743 int content_changed;
2744
2745 name = p->two->path ? p->two->path : p->one->path;
2746
2747 if (p->one->oid_valid && p->two->oid_valid)
2748 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2749 else
2750 content_changed = 1;
2751
2752 if (!content_changed) {
2753 /*
2754 * The SHA1 has not changed, so pre-/post-content is
2755 * identical. We can therefore skip looking at the
2756 * file contents altogether.
2757 */
2758 damage = 0;
2759 goto found_damage;
2760 }
2761
2762 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2763 /*
2764 * In --dirstat-by-file mode, we don't really need to
2765 * look at the actual file contents at all.
2766 * The fact that the SHA1 changed is enough for us to
2767 * add this file to the list of results
2768 * (with each file contributing equal damage).
2769 */
2770 damage = 1;
2771 goto found_damage;
2772 }
2773
2774 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2775 diff_populate_filespec(p->one, 0);
2776 diff_populate_filespec(p->two, 0);
2777 diffcore_count_changes(p->one, p->two, NULL, NULL,
2778 &copied, &added);
2779 diff_free_filespec_data(p->one);
2780 diff_free_filespec_data(p->two);
2781 } else if (DIFF_FILE_VALID(p->one)) {
2782 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2783 copied = added = 0;
2784 diff_free_filespec_data(p->one);
2785 } else if (DIFF_FILE_VALID(p->two)) {
2786 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2787 copied = 0;
2788 added = p->two->size;
2789 diff_free_filespec_data(p->two);
2790 } else
2791 continue;
2792
2793 /*
2794 * Original minus copied is the removed material,
2795 * added is the new material. They are both damages
2796 * made to the preimage.
2797 * If the resulting damage is zero, we know that
2798 * diffcore_count_changes() considers the two entries to
2799 * be identical, but since content_changed is true, we
2800 * know that there must have been _some_ kind of change,
2801 * so we force all entries to have damage > 0.
2802 */
2803 damage = (p->one->size - copied) + added;
2804 if (!damage)
2805 damage = 1;
2806
2807 found_damage:
2808 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2809 dir.files[dir.nr].name = name;
2810 dir.files[dir.nr].changed = damage;
2811 changed += damage;
2812 dir.nr++;
2813 }
2814
2815 /* This can happen even with many files, if everything was renames */
2816 if (!changed)
2817 return;
2818
2819 /* Show all directories with more than x% of the changes */
2820 QSORT(dir.files, dir.nr, dirstat_compare);
2821 gather_dirstat(options, &dir, changed, "", 0);
2822 }
2823
2824 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2825 {
2826 int i;
2827 unsigned long changed;
2828 struct dirstat_dir dir;
2829
2830 if (data->nr == 0)
2831 return;
2832
2833 dir.files = NULL;
2834 dir.alloc = 0;
2835 dir.nr = 0;
2836 dir.permille = options->dirstat_permille;
2837 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2838
2839 changed = 0;
2840 for (i = 0; i < data->nr; i++) {
2841 struct diffstat_file *file = data->files[i];
2842 unsigned long damage = file->added + file->deleted;
2843 if (file->is_binary)
2844 /*
2845 * binary files counts bytes, not lines. Must find some
2846 * way to normalize binary bytes vs. textual lines.
2847 * The following heuristic assumes that there are 64
2848 * bytes per "line".
2849 * This is stupid and ugly, but very cheap...
2850 */
2851 damage = (damage + 63) / 64;
2852 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2853 dir.files[dir.nr].name = file->name;
2854 dir.files[dir.nr].changed = damage;
2855 changed += damage;
2856 dir.nr++;
2857 }
2858
2859 /* This can happen even with many files, if everything was renames */
2860 if (!changed)
2861 return;
2862
2863 /* Show all directories with more than x% of the changes */
2864 QSORT(dir.files, dir.nr, dirstat_compare);
2865 gather_dirstat(options, &dir, changed, "", 0);
2866 }
2867
2868 static void free_diffstat_info(struct diffstat_t *diffstat)
2869 {
2870 int i;
2871 for (i = 0; i < diffstat->nr; i++) {
2872 struct diffstat_file *f = diffstat->files[i];
2873 if (f->name != f->print_name)
2874 free(f->print_name);
2875 free(f->name);
2876 free(f->from_name);
2877 free(f);
2878 }
2879 free(diffstat->files);
2880 }
2881
2882 struct checkdiff_t {
2883 const char *filename;
2884 int lineno;
2885 int conflict_marker_size;
2886 struct diff_options *o;
2887 unsigned ws_rule;
2888 unsigned status;
2889 };
2890
2891 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2892 {
2893 char firstchar;
2894 int cnt;
2895
2896 if (len < marker_size + 1)
2897 return 0;
2898 firstchar = line[0];
2899 switch (firstchar) {
2900 case '=': case '>': case '<': case '|':
2901 break;
2902 default:
2903 return 0;
2904 }
2905 for (cnt = 1; cnt < marker_size; cnt++)
2906 if (line[cnt] != firstchar)
2907 return 0;
2908 /* line[1] thru line[marker_size-1] are same as firstchar */
2909 if (len < marker_size + 1 || !isspace(line[marker_size]))
2910 return 0;
2911 return 1;
2912 }
2913
2914 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2915 {
2916 struct checkdiff_t *data = priv;
2917 int marker_size = data->conflict_marker_size;
2918 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2919 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2920 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2921 char *err;
2922 const char *line_prefix;
2923
2924 assert(data->o);
2925 line_prefix = diff_line_prefix(data->o);
2926
2927 if (line[0] == '+') {
2928 unsigned bad;
2929 data->lineno++;
2930 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2931 data->status |= 1;
2932 fprintf(data->o->file,
2933 "%s%s:%d: leftover conflict marker\n",
2934 line_prefix, data->filename, data->lineno);
2935 }
2936 bad = ws_check(line + 1, len - 1, data->ws_rule);
2937 if (!bad)
2938 return;
2939 data->status |= bad;
2940 err = whitespace_error_string(bad);
2941 fprintf(data->o->file, "%s%s:%d: %s.\n",
2942 line_prefix, data->filename, data->lineno, err);
2943 free(err);
2944 emit_line(data->o, set, reset, line, 1);
2945 ws_check_emit(line + 1, len - 1, data->ws_rule,
2946 data->o->file, set, reset, ws);
2947 } else if (line[0] == ' ') {
2948 data->lineno++;
2949 } else if (line[0] == '@') {
2950 char *plus = strchr(line, '+');
2951 if (plus)
2952 data->lineno = strtol(plus, NULL, 10) - 1;
2953 else
2954 die("invalid diff");
2955 }
2956 }
2957
2958 static unsigned char *deflate_it(char *data,
2959 unsigned long size,
2960 unsigned long *result_size)
2961 {
2962 int bound;
2963 unsigned char *deflated;
2964 git_zstream stream;
2965
2966 git_deflate_init(&stream, zlib_compression_level);
2967 bound = git_deflate_bound(&stream, size);
2968 deflated = xmalloc(bound);
2969 stream.next_out = deflated;
2970 stream.avail_out = bound;
2971
2972 stream.next_in = (unsigned char *)data;
2973 stream.avail_in = size;
2974 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2975 ; /* nothing */
2976 git_deflate_end(&stream);
2977 *result_size = stream.total_out;
2978 return deflated;
2979 }
2980
2981 static void emit_binary_diff_body(struct diff_options *o,
2982 mmfile_t *one, mmfile_t *two)
2983 {
2984 void *cp;
2985 void *delta;
2986 void *deflated;
2987 void *data;
2988 unsigned long orig_size;
2989 unsigned long delta_size;
2990 unsigned long deflate_size;
2991 unsigned long data_size;
2992
2993 /* We could do deflated delta, or we could do just deflated two,
2994 * whichever is smaller.
2995 */
2996 delta = NULL;
2997 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2998 if (one->size && two->size) {
2999 delta = diff_delta(one->ptr, one->size,
3000 two->ptr, two->size,
3001 &delta_size, deflate_size);
3002 if (delta) {
3003 void *to_free = delta;
3004 orig_size = delta_size;
3005 delta = deflate_it(delta, delta_size, &delta_size);
3006 free(to_free);
3007 }
3008 }
3009
3010 if (delta && delta_size < deflate_size) {
3011 char *s = xstrfmt("%lu", orig_size);
3012 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3013 s, strlen(s), 0);
3014 free(s);
3015 free(deflated);
3016 data = delta;
3017 data_size = delta_size;
3018 } else {
3019 char *s = xstrfmt("%lu", two->size);
3020 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3021 s, strlen(s), 0);
3022 free(s);
3023 free(delta);
3024 data = deflated;
3025 data_size = deflate_size;
3026 }
3027
3028 /* emit data encoded in base85 */
3029 cp = data;
3030 while (data_size) {
3031 int len;
3032 int bytes = (52 < data_size) ? 52 : data_size;
3033 char line[71];
3034 data_size -= bytes;
3035 if (bytes <= 26)
3036 line[0] = bytes + 'A' - 1;
3037 else
3038 line[0] = bytes - 26 + 'a' - 1;
3039 encode_85(line + 1, cp, bytes);
3040 cp = (char *) cp + bytes;
3041
3042 len = strlen(line);
3043 line[len++] = '\n';
3044 line[len] = '\0';
3045
3046 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3047 line, len, 0);
3048 }
3049 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3050 free(data);
3051 }
3052
3053 static void emit_binary_diff(struct diff_options *o,
3054 mmfile_t *one, mmfile_t *two)
3055 {
3056 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3057 emit_binary_diff_body(o, one, two);
3058 emit_binary_diff_body(o, two, one);
3059 }
3060
3061 int diff_filespec_is_binary(struct diff_filespec *one)
3062 {
3063 if (one->is_binary == -1) {
3064 diff_filespec_load_driver(one);
3065 if (one->driver->binary != -1)
3066 one->is_binary = one->driver->binary;
3067 else {
3068 if (!one->data && DIFF_FILE_VALID(one))
3069 diff_populate_filespec(one, CHECK_BINARY);
3070 if (one->is_binary == -1 && one->data)
3071 one->is_binary = buffer_is_binary(one->data,
3072 one->size);
3073 if (one->is_binary == -1)
3074 one->is_binary = 0;
3075 }
3076 }
3077 return one->is_binary;
3078 }
3079
3080 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3081 {
3082 diff_filespec_load_driver(one);
3083 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3084 }
3085
3086 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3087 {
3088 if (!options->a_prefix)
3089 options->a_prefix = a;
3090 if (!options->b_prefix)
3091 options->b_prefix = b;
3092 }
3093
3094 struct userdiff_driver *get_textconv(struct diff_filespec *one)
3095 {
3096 if (!DIFF_FILE_VALID(one))
3097 return NULL;
3098
3099 diff_filespec_load_driver(one);
3100 return userdiff_get_textconv(one->driver);
3101 }
3102
3103 static void builtin_diff(const char *name_a,
3104 const char *name_b,
3105 struct diff_filespec *one,
3106 struct diff_filespec *two,
3107 const char *xfrm_msg,
3108 int must_show_header,
3109 struct diff_options *o,
3110 int complete_rewrite)
3111 {
3112 mmfile_t mf1, mf2;
3113 const char *lbl[2];
3114 char *a_one, *b_two;
3115 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3116 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3117 const char *a_prefix, *b_prefix;
3118 struct userdiff_driver *textconv_one = NULL;
3119 struct userdiff_driver *textconv_two = NULL;
3120 struct strbuf header = STRBUF_INIT;
3121 const char *line_prefix = diff_line_prefix(o);
3122
3123 diff_set_mnemonic_prefix(o, "a/", "b/");
3124 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
3125 a_prefix = o->b_prefix;
3126 b_prefix = o->a_prefix;
3127 } else {
3128 a_prefix = o->a_prefix;
3129 b_prefix = o->b_prefix;
3130 }
3131
3132 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3133 (!one->mode || S_ISGITLINK(one->mode)) &&
3134 (!two->mode || S_ISGITLINK(two->mode))) {
3135 show_submodule_summary(o, one->path ? one->path : two->path,
3136 &one->oid, &two->oid,
3137 two->dirty_submodule);
3138 return;
3139 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3140 (!one->mode || S_ISGITLINK(one->mode)) &&
3141 (!two->mode || S_ISGITLINK(two->mode))) {
3142 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3143 &one->oid, &two->oid,
3144 two->dirty_submodule);
3145 return;
3146 }
3147
3148 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
3149 textconv_one = get_textconv(one);
3150 textconv_two = get_textconv(two);
3151 }
3152
3153 /* Never use a non-valid filename anywhere if at all possible */
3154 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3155 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3156
3157 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3158 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3159 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3160 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3161 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3162 if (lbl[0][0] == '/') {
3163 /* /dev/null */
3164 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3165 if (xfrm_msg)
3166 strbuf_addstr(&header, xfrm_msg);
3167 must_show_header = 1;
3168 }
3169 else if (lbl[1][0] == '/') {
3170 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3171 if (xfrm_msg)
3172 strbuf_addstr(&header, xfrm_msg);
3173 must_show_header = 1;
3174 }
3175 else {
3176 if (one->mode != two->mode) {
3177 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3178 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3179 must_show_header = 1;
3180 }
3181 if (xfrm_msg)
3182 strbuf_addstr(&header, xfrm_msg);
3183
3184 /*
3185 * we do not run diff between different kind
3186 * of objects.
3187 */
3188 if ((one->mode ^ two->mode) & S_IFMT)
3189 goto free_ab_and_return;
3190 if (complete_rewrite &&
3191 (textconv_one || !diff_filespec_is_binary(one)) &&
3192 (textconv_two || !diff_filespec_is_binary(two))) {
3193 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3194 header.buf, header.len, 0);
3195 strbuf_reset(&header);
3196 emit_rewrite_diff(name_a, name_b, one, two,
3197 textconv_one, textconv_two, o);
3198 o->found_changes = 1;
3199 goto free_ab_and_return;
3200 }
3201 }
3202
3203 if (o->irreversible_delete && lbl[1][0] == '/') {
3204 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3205 header.len, 0);
3206 strbuf_reset(&header);
3207 goto free_ab_and_return;
3208 } else if (!DIFF_OPT_TST(o, TEXT) &&
3209 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3210 (!textconv_two && diff_filespec_is_binary(two)) )) {
3211 struct strbuf sb = STRBUF_INIT;
3212 if (!one->data && !two->data &&
3213 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3214 !DIFF_OPT_TST(o, BINARY)) {
3215 if (!oidcmp(&one->oid, &two->oid)) {
3216 if (must_show_header)
3217 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3218 header.buf, header.len,
3219 0);
3220 goto free_ab_and_return;
3221 }
3222 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3223 header.buf, header.len, 0);
3224 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3225 diff_line_prefix(o), lbl[0], lbl[1]);
3226 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3227 sb.buf, sb.len, 0);
3228 strbuf_release(&sb);
3229 goto free_ab_and_return;
3230 }
3231 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3232 die("unable to read files to diff");
3233 /* Quite common confusing case */
3234 if (mf1.size == mf2.size &&
3235 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3236 if (must_show_header)
3237 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3238 header.buf, header.len, 0);
3239 goto free_ab_and_return;
3240 }
3241 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3242 strbuf_reset(&header);
3243 if (DIFF_OPT_TST(o, BINARY))
3244 emit_binary_diff(o, &mf1, &mf2);
3245 else {
3246 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3247 diff_line_prefix(o), lbl[0], lbl[1]);
3248 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3249 sb.buf, sb.len, 0);
3250 strbuf_release(&sb);
3251 }
3252 o->found_changes = 1;
3253 } else {
3254 /* Crazy xdl interfaces.. */
3255 const char *diffopts = getenv("GIT_DIFF_OPTS");
3256 const char *v;
3257 xpparam_t xpp;
3258 xdemitconf_t xecfg;
3259 struct emit_callback ecbdata;
3260 const struct userdiff_funcname *pe;
3261
3262 if (must_show_header) {
3263 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3264 header.buf, header.len, 0);
3265 strbuf_reset(&header);
3266 }
3267
3268 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3269 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3270
3271 pe = diff_funcname_pattern(one);
3272 if (!pe)
3273 pe = diff_funcname_pattern(two);
3274
3275 memset(&xpp, 0, sizeof(xpp));
3276 memset(&xecfg, 0, sizeof(xecfg));
3277 memset(&ecbdata, 0, sizeof(ecbdata));
3278 ecbdata.label_path = lbl;
3279 ecbdata.color_diff = want_color(o->use_color);
3280 ecbdata.ws_rule = whitespace_rule(name_b);
3281 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3282 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3283 ecbdata.opt = o;
3284 ecbdata.header = header.len ? &header : NULL;
3285 xpp.flags = o->xdl_opts;
3286 xecfg.ctxlen = o->context;
3287 xecfg.interhunkctxlen = o->interhunkcontext;
3288 xecfg.flags = XDL_EMIT_FUNCNAMES;
3289 if (DIFF_OPT_TST(o, FUNCCONTEXT))
3290 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3291 if (pe)
3292 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3293 if (!diffopts)
3294 ;
3295 else if (skip_prefix(diffopts, "--unified=", &v))
3296 xecfg.ctxlen = strtoul(v, NULL, 10);
3297 else if (skip_prefix(diffopts, "-u", &v))
3298 xecfg.ctxlen = strtoul(v, NULL, 10);
3299 if (o->word_diff)
3300 init_diff_words_data(&ecbdata, o, one, two);
3301 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
3302 &xpp, &xecfg))
3303 die("unable to generate diff for %s", one->path);
3304 if (o->word_diff)
3305 free_diff_words_data(&ecbdata);
3306 if (textconv_one)
3307 free(mf1.ptr);
3308 if (textconv_two)
3309 free(mf2.ptr);
3310 xdiff_clear_find_func(&xecfg);
3311 }
3312
3313 free_ab_and_return:
3314 strbuf_release(&header);
3315 diff_free_filespec_data(one);
3316 diff_free_filespec_data(two);
3317 free(a_one);
3318 free(b_two);
3319 return;
3320 }
3321
3322 static void builtin_diffstat(const char *name_a, const char *name_b,
3323 struct diff_filespec *one,
3324 struct diff_filespec *two,
3325 struct diffstat_t *diffstat,
3326 struct diff_options *o,
3327 struct diff_filepair *p)
3328 {
3329 mmfile_t mf1, mf2;
3330 struct diffstat_file *data;
3331 int same_contents;
3332 int complete_rewrite = 0;
3333
3334 if (!DIFF_PAIR_UNMERGED(p)) {
3335 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3336 complete_rewrite = 1;
3337 }
3338
3339 data = diffstat_add(diffstat, name_a, name_b);
3340 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3341
3342 if (!one || !two) {
3343 data->is_unmerged = 1;
3344 return;
3345 }
3346
3347 same_contents = !oidcmp(&one->oid, &two->oid);
3348
3349 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3350 data->is_binary = 1;
3351 if (same_contents) {
3352 data->added = 0;
3353 data->deleted = 0;
3354 } else {
3355 data->added = diff_filespec_size(two);
3356 data->deleted = diff_filespec_size(one);
3357 }
3358 }
3359
3360 else if (complete_rewrite) {
3361 diff_populate_filespec(one, 0);
3362 diff_populate_filespec(two, 0);
3363 data->deleted = count_lines(one->data, one->size);
3364 data->added = count_lines(two->data, two->size);
3365 }
3366
3367 else if (!same_contents) {
3368 /* Crazy xdl interfaces.. */
3369 xpparam_t xpp;
3370 xdemitconf_t xecfg;
3371
3372 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3373 die("unable to read files to diff");
3374
3375 memset(&xpp, 0, sizeof(xpp));
3376 memset(&xecfg, 0, sizeof(xecfg));
3377 xpp.flags = o->xdl_opts;
3378 xecfg.ctxlen = o->context;
3379 xecfg.interhunkctxlen = o->interhunkcontext;
3380 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
3381 &xpp, &xecfg))
3382 die("unable to generate diffstat for %s", one->path);
3383 }
3384
3385 diff_free_filespec_data(one);
3386 diff_free_filespec_data(two);
3387 }
3388
3389 static void builtin_checkdiff(const char *name_a, const char *name_b,
3390 const char *attr_path,
3391 struct diff_filespec *one,
3392 struct diff_filespec *two,
3393 struct diff_options *o)
3394 {
3395 mmfile_t mf1, mf2;
3396 struct checkdiff_t data;
3397
3398 if (!two)
3399 return;
3400
3401 memset(&data, 0, sizeof(data));
3402 data.filename = name_b ? name_b : name_a;
3403 data.lineno = 0;
3404 data.o = o;
3405 data.ws_rule = whitespace_rule(attr_path);
3406 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3407
3408 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3409 die("unable to read files to diff");
3410
3411 /*
3412 * All the other codepaths check both sides, but not checking
3413 * the "old" side here is deliberate. We are checking the newly
3414 * introduced changes, and as long as the "new" side is text, we
3415 * can and should check what it introduces.
3416 */
3417 if (diff_filespec_is_binary(two))
3418 goto free_and_return;
3419 else {
3420 /* Crazy xdl interfaces.. */
3421 xpparam_t xpp;
3422 xdemitconf_t xecfg;
3423
3424 memset(&xpp, 0, sizeof(xpp));
3425 memset(&xecfg, 0, sizeof(xecfg));
3426 xecfg.ctxlen = 1; /* at least one context line */
3427 xpp.flags = 0;
3428 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
3429 &xpp, &xecfg))
3430 die("unable to generate checkdiff for %s", one->path);
3431
3432 if (data.ws_rule & WS_BLANK_AT_EOF) {
3433 struct emit_callback ecbdata;
3434 int blank_at_eof;
3435
3436 ecbdata.ws_rule = data.ws_rule;
3437 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3438 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3439
3440 if (blank_at_eof) {
3441 static char *err;
3442 if (!err)
3443 err = whitespace_error_string(WS_BLANK_AT_EOF);
3444 fprintf(o->file, "%s:%d: %s.\n",
3445 data.filename, blank_at_eof, err);
3446 data.status = 1; /* report errors */
3447 }
3448 }
3449 }
3450 free_and_return:
3451 diff_free_filespec_data(one);
3452 diff_free_filespec_data(two);
3453 if (data.status)
3454 DIFF_OPT_SET(o, CHECK_FAILED);
3455 }
3456
3457 struct diff_filespec *alloc_filespec(const char *path)
3458 {
3459 struct diff_filespec *spec;
3460
3461 FLEXPTR_ALLOC_STR(spec, path, path);
3462 spec->count = 1;
3463 spec->is_binary = -1;
3464 return spec;
3465 }
3466
3467 void free_filespec(struct diff_filespec *spec)
3468 {
3469 if (!--spec->count) {
3470 diff_free_filespec_data(spec);
3471 free(spec);
3472 }
3473 }
3474
3475 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3476 int oid_valid, unsigned short mode)
3477 {
3478 if (mode) {
3479 spec->mode = canon_mode(mode);
3480 oidcpy(&spec->oid, oid);
3481 spec->oid_valid = oid_valid;
3482 }
3483 }
3484
3485 /*
3486 * Given a name and sha1 pair, if the index tells us the file in
3487 * the work tree has that object contents, return true, so that
3488 * prepare_temp_file() does not have to inflate and extract.
3489 */
3490 static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3491 {
3492 const struct cache_entry *ce;
3493 struct stat st;
3494 int pos, len;
3495
3496 /*
3497 * We do not read the cache ourselves here, because the
3498 * benchmark with my previous version that always reads cache
3499 * shows that it makes things worse for diff-tree comparing
3500 * two linux-2.6 kernel trees in an already checked out work
3501 * tree. This is because most diff-tree comparisons deal with
3502 * only a small number of files, while reading the cache is
3503 * expensive for a large project, and its cost outweighs the
3504 * savings we get by not inflating the object to a temporary
3505 * file. Practically, this code only helps when we are used
3506 * by diff-cache --cached, which does read the cache before
3507 * calling us.
3508 */
3509 if (!active_cache)
3510 return 0;
3511
3512 /* We want to avoid the working directory if our caller
3513 * doesn't need the data in a normal file, this system
3514 * is rather slow with its stat/open/mmap/close syscalls,
3515 * and the object is contained in a pack file. The pack
3516 * is probably already open and will be faster to obtain
3517 * the data through than the working directory. Loose
3518 * objects however would tend to be slower as they need
3519 * to be individually opened and inflated.
3520 */
3521 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(oid->hash))
3522 return 0;
3523
3524 /*
3525 * Similarly, if we'd have to convert the file contents anyway, that
3526 * makes the optimization not worthwhile.
3527 */
3528 if (!want_file && would_convert_to_git(&the_index, name))
3529 return 0;
3530
3531 len = strlen(name);
3532 pos = cache_name_pos(name, len);
3533 if (pos < 0)
3534 return 0;
3535 ce = active_cache[pos];
3536
3537 /*
3538 * This is not the sha1 we are looking for, or
3539 * unreusable because it is not a regular file.
3540 */
3541 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3542 return 0;
3543
3544 /*
3545 * If ce is marked as "assume unchanged", there is no
3546 * guarantee that work tree matches what we are looking for.
3547 */
3548 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3549 return 0;
3550
3551 /*
3552 * If ce matches the file in the work tree, we can reuse it.
3553 */
3554 if (ce_uptodate(ce) ||
3555 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3556 return 1;
3557
3558 return 0;
3559 }
3560
3561 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3562 {
3563 struct strbuf buf = STRBUF_INIT;
3564 char *dirty = "";
3565
3566 /* Are we looking at the work tree? */
3567 if (s->dirty_submodule)
3568 dirty = "-dirty";
3569
3570 strbuf_addf(&buf, "Subproject commit %s%s\n",
3571 oid_to_hex(&s->oid), dirty);
3572 s->size = buf.len;
3573 if (size_only) {
3574 s->data = NULL;
3575 strbuf_release(&buf);
3576 } else {
3577 s->data = strbuf_detach(&buf, NULL);
3578 s->should_free = 1;
3579 }
3580 return 0;
3581 }
3582
3583 /*
3584 * While doing rename detection and pickaxe operation, we may need to
3585 * grab the data for the blob (or file) for our own in-core comparison.
3586 * diff_filespec has data and size fields for this purpose.
3587 */
3588 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3589 {
3590 int size_only = flags & CHECK_SIZE_ONLY;
3591 int err = 0;
3592 /*
3593 * demote FAIL to WARN to allow inspecting the situation
3594 * instead of refusing.
3595 */
3596 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
3597 ? SAFE_CRLF_WARN
3598 : safe_crlf);
3599
3600 if (!DIFF_FILE_VALID(s))
3601 die("internal error: asking to populate invalid file.");
3602 if (S_ISDIR(s->mode))
3603 return -1;
3604
3605 if (s->data)
3606 return 0;
3607
3608 if (size_only && 0 < s->size)
3609 return 0;
3610
3611 if (S_ISGITLINK(s->mode))
3612 return diff_populate_gitlink(s, size_only);
3613
3614 if (!s->oid_valid ||
3615 reuse_worktree_file(s->path, &s->oid, 0)) {
3616 struct strbuf buf = STRBUF_INIT;
3617 struct stat st;
3618 int fd;
3619
3620 if (lstat(s->path, &st) < 0) {
3621 if (errno == ENOENT) {
3622 err_empty:
3623 err = -1;
3624 empty:
3625 s->data = (char *)"";
3626 s->size = 0;
3627 return err;
3628 }
3629 }
3630 s->size = xsize_t(st.st_size);
3631 if (!s->size)
3632 goto empty;
3633 if (S_ISLNK(st.st_mode)) {
3634 struct strbuf sb = STRBUF_INIT;
3635
3636 if (strbuf_readlink(&sb, s->path, s->size))
3637 goto err_empty;
3638 s->size = sb.len;
3639 s->data = strbuf_detach(&sb, NULL);
3640 s->should_free = 1;
3641 return 0;
3642 }
3643
3644 /*
3645 * Even if the caller would be happy with getting
3646 * only the size, we cannot return early at this
3647 * point if the path requires us to run the content
3648 * conversion.
3649 */
3650 if (size_only && !would_convert_to_git(&the_index, s->path))
3651 return 0;
3652
3653 /*
3654 * Note: this check uses xsize_t(st.st_size) that may
3655 * not be the true size of the blob after it goes
3656 * through convert_to_git(). This may not strictly be
3657 * correct, but the whole point of big_file_threshold
3658 * and is_binary check being that we want to avoid
3659 * opening the file and inspecting the contents, this
3660 * is probably fine.
3661 */
3662 if ((flags & CHECK_BINARY) &&
3663 s->size > big_file_threshold && s->is_binary == -1) {
3664 s->is_binary = 1;
3665 return 0;
3666 }
3667 fd = open(s->path, O_RDONLY);
3668 if (fd < 0)
3669 goto err_empty;
3670 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3671 close(fd);
3672 s->should_munmap = 1;
3673
3674 /*
3675 * Convert from working tree format to canonical git format
3676 */
3677 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, crlf_warn)) {
3678 size_t size = 0;
3679 munmap(s->data, s->size);
3680 s->should_munmap = 0;
3681 s->data = strbuf_detach(&buf, &size);
3682 s->size = size;
3683 s->should_free = 1;
3684 }
3685 }
3686 else {
3687 enum object_type type;
3688 if (size_only || (flags & CHECK_BINARY)) {
3689 type = sha1_object_info(s->oid.hash, &s->size);
3690 if (type < 0)
3691 die("unable to read %s",
3692 oid_to_hex(&s->oid));
3693 if (size_only)
3694 return 0;
3695 if (s->size > big_file_threshold && s->is_binary == -1) {
3696 s->is_binary = 1;
3697 return 0;
3698 }
3699 }
3700 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
3701 if (!s->data)
3702 die("unable to read %s", oid_to_hex(&s->oid));
3703 s->should_free = 1;
3704 }
3705 return 0;
3706 }
3707
3708 void diff_free_filespec_blob(struct diff_filespec *s)
3709 {
3710 if (s->should_free)
3711 free(s->data);
3712 else if (s->should_munmap)
3713 munmap(s->data, s->size);
3714
3715 if (s->should_free || s->should_munmap) {
3716 s->should_free = s->should_munmap = 0;
3717 s->data = NULL;
3718 }
3719 }
3720
3721 void diff_free_filespec_data(struct diff_filespec *s)
3722 {
3723 diff_free_filespec_blob(s);
3724 FREE_AND_NULL(s->cnt_data);
3725 }
3726
3727 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3728 void *blob,
3729 unsigned long size,
3730 const struct object_id *oid,
3731 int mode)
3732 {
3733 int fd;
3734 struct strbuf buf = STRBUF_INIT;
3735 struct strbuf template = STRBUF_INIT;
3736 char *path_dup = xstrdup(path);
3737 const char *base = basename(path_dup);
3738
3739 /* Generate "XXXXXX_basename.ext" */
3740 strbuf_addstr(&template, "XXXXXX_");
3741 strbuf_addstr(&template, base);
3742
3743 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
3744 if (fd < 0)
3745 die_errno("unable to create temp-file");
3746 if (convert_to_working_tree(path,
3747 (const char *)blob, (size_t)size, &buf)) {
3748 blob = buf.buf;
3749 size = buf.len;
3750 }
3751 if (write_in_full(fd, blob, size) != size)
3752 die_errno("unable to write temp-file");
3753 close_tempfile(&temp->tempfile);
3754 temp->name = get_tempfile_path(&temp->tempfile);
3755 oid_to_hex_r(temp->hex, oid);
3756 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3757 strbuf_release(&buf);
3758 strbuf_release(&template);
3759 free(path_dup);
3760 }
3761
3762 static struct diff_tempfile *prepare_temp_file(const char *name,
3763 struct diff_filespec *one)
3764 {
3765 struct diff_tempfile *temp = claim_diff_tempfile();
3766
3767 if (!DIFF_FILE_VALID(one)) {
3768 not_a_valid_file:
3769 /* A '-' entry produces this for file-2, and
3770 * a '+' entry produces this for file-1.
3771 */
3772 temp->name = "/dev/null";
3773 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3774 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3775 return temp;
3776 }
3777
3778 if (!S_ISGITLINK(one->mode) &&
3779 (!one->oid_valid ||
3780 reuse_worktree_file(name, &one->oid, 1))) {
3781 struct stat st;
3782 if (lstat(name, &st) < 0) {
3783 if (errno == ENOENT)
3784 goto not_a_valid_file;
3785 die_errno("stat(%s)", name);
3786 }
3787 if (S_ISLNK(st.st_mode)) {
3788 struct strbuf sb = STRBUF_INIT;
3789 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3790 die_errno("readlink(%s)", name);
3791 prep_temp_blob(name, temp, sb.buf, sb.len,
3792 (one->oid_valid ?
3793 &one->oid : &null_oid),
3794 (one->oid_valid ?
3795 one->mode : S_IFLNK));
3796 strbuf_release(&sb);
3797 }
3798 else {
3799 /* we can borrow from the file in the work tree */
3800 temp->name = name;
3801 if (!one->oid_valid)
3802 oid_to_hex_r(temp->hex, &null_oid);
3803 else
3804 oid_to_hex_r(temp->hex, &one->oid);
3805 /* Even though we may sometimes borrow the
3806 * contents from the work tree, we always want
3807 * one->mode. mode is trustworthy even when
3808 * !(one->oid_valid), as long as
3809 * DIFF_FILE_VALID(one).
3810 */
3811 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3812 }
3813 return temp;
3814 }
3815 else {
3816 if (diff_populate_filespec(one, 0))
3817 die("cannot read data blob for %s", one->path);
3818 prep_temp_blob(name, temp, one->data, one->size,
3819 &one->oid, one->mode);
3820 }
3821 return temp;
3822 }
3823
3824 static void add_external_diff_name(struct argv_array *argv,
3825 const char *name,
3826 struct diff_filespec *df)
3827 {
3828 struct diff_tempfile *temp = prepare_temp_file(name, df);
3829 argv_array_push(argv, temp->name);
3830 argv_array_push(argv, temp->hex);
3831 argv_array_push(argv, temp->mode);
3832 }
3833
3834 /* An external diff command takes:
3835 *
3836 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3837 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3838 *
3839 */
3840 static void run_external_diff(const char *pgm,
3841 const char *name,
3842 const char *other,
3843 struct diff_filespec *one,
3844 struct diff_filespec *two,
3845 const char *xfrm_msg,
3846 int complete_rewrite,
3847 struct diff_options *o)
3848 {
3849 struct argv_array argv = ARGV_ARRAY_INIT;
3850 struct argv_array env = ARGV_ARRAY_INIT;
3851 struct diff_queue_struct *q = &diff_queued_diff;
3852
3853 argv_array_push(&argv, pgm);
3854 argv_array_push(&argv, name);
3855
3856 if (one && two) {
3857 add_external_diff_name(&argv, name, one);
3858 if (!other)
3859 add_external_diff_name(&argv, name, two);
3860 else {
3861 add_external_diff_name(&argv, other, two);
3862 argv_array_push(&argv, other);
3863 argv_array_push(&argv, xfrm_msg);
3864 }
3865 }
3866
3867 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3868 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3869
3870 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3871 die(_("external diff died, stopping at %s"), name);
3872
3873 remove_tempfile();
3874 argv_array_clear(&argv);
3875 argv_array_clear(&env);
3876 }
3877
3878 static int similarity_index(struct diff_filepair *p)
3879 {
3880 return p->score * 100 / MAX_SCORE;
3881 }
3882
3883 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3884 {
3885 if (startup_info->have_repository)
3886 return find_unique_abbrev(oid->hash, abbrev);
3887 else {
3888 char *hex = oid_to_hex(oid);
3889 if (abbrev < 0)
3890 abbrev = FALLBACK_DEFAULT_ABBREV;
3891 if (abbrev > GIT_SHA1_HEXSZ)
3892 die("BUG: oid abbreviation out of range: %d", abbrev);
3893 if (abbrev)
3894 hex[abbrev] = '\0';
3895 return hex;
3896 }
3897 }
3898
3899 static void fill_metainfo(struct strbuf *msg,
3900 const char *name,
3901 const char *other,
3902 struct diff_filespec *one,
3903 struct diff_filespec *two,
3904 struct diff_options *o,
3905 struct diff_filepair *p,
3906 int *must_show_header,
3907 int use_color)
3908 {
3909 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3910 const char *reset = diff_get_color(use_color, DIFF_RESET);
3911 const char *line_prefix = diff_line_prefix(o);
3912
3913 *must_show_header = 1;
3914 strbuf_init(msg, PATH_MAX * 2 + 300);
3915 switch (p->status) {
3916 case DIFF_STATUS_COPIED:
3917 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3918 line_prefix, set, similarity_index(p));
3919 strbuf_addf(msg, "%s\n%s%scopy from ",
3920 reset, line_prefix, set);
3921 quote_c_style(name, msg, NULL, 0);
3922 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3923 quote_c_style(other, msg, NULL, 0);
3924 strbuf_addf(msg, "%s\n", reset);
3925 break;
3926 case DIFF_STATUS_RENAMED:
3927 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3928 line_prefix, set, similarity_index(p));
3929 strbuf_addf(msg, "%s\n%s%srename from ",
3930 reset, line_prefix, set);
3931 quote_c_style(name, msg, NULL, 0);
3932 strbuf_addf(msg, "%s\n%s%srename to ",
3933 reset, line_prefix, set);
3934 quote_c_style(other, msg, NULL, 0);
3935 strbuf_addf(msg, "%s\n", reset);
3936 break;
3937 case DIFF_STATUS_MODIFIED:
3938 if (p->score) {
3939 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3940 line_prefix,
3941 set, similarity_index(p), reset);
3942 break;
3943 }
3944 /* fallthru */
3945 default:
3946 *must_show_header = 0;
3947 }
3948 if (one && two && oidcmp(&one->oid, &two->oid)) {
3949 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3950
3951 if (DIFF_OPT_TST(o, BINARY)) {
3952 mmfile_t mf;
3953 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3954 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3955 abbrev = 40;
3956 }
3957 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3958 diff_abbrev_oid(&one->oid, abbrev),
3959 diff_abbrev_oid(&two->oid, abbrev));
3960 if (one->mode == two->mode)
3961 strbuf_addf(msg, " %06o", one->mode);
3962 strbuf_addf(msg, "%s\n", reset);
3963 }
3964 }
3965
3966 static void run_diff_cmd(const char *pgm,
3967 const char *name,
3968 const char *other,
3969 const char *attr_path,
3970 struct diff_filespec *one,
3971 struct diff_filespec *two,
3972 struct strbuf *msg,
3973 struct diff_options *o,
3974 struct diff_filepair *p)
3975 {
3976 const char *xfrm_msg = NULL;
3977 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3978 int must_show_header = 0;
3979
3980
3981 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3982 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3983 if (drv && drv->external)
3984 pgm = drv->external;
3985 }
3986
3987 if (msg) {
3988 /*
3989 * don't use colors when the header is intended for an
3990 * external diff driver
3991 */
3992 fill_metainfo(msg, name, other, one, two, o, p,
3993 &must_show_header,
3994 want_color(o->use_color) && !pgm);
3995 xfrm_msg = msg->len ? msg->buf : NULL;
3996 }
3997
3998 if (pgm) {
3999 run_external_diff(pgm, name, other, one, two, xfrm_msg,
4000 complete_rewrite, o);
4001 return;
4002 }
4003 if (one && two)
4004 builtin_diff(name, other ? other : name,
4005 one, two, xfrm_msg, must_show_header,
4006 o, complete_rewrite);
4007 else
4008 fprintf(o->file, "* Unmerged path %s\n", name);
4009 }
4010
4011 static void diff_fill_oid_info(struct diff_filespec *one)
4012 {
4013 if (DIFF_FILE_VALID(one)) {
4014 if (!one->oid_valid) {
4015 struct stat st;
4016 if (one->is_stdin) {
4017 oidclr(&one->oid);
4018 return;
4019 }
4020 if (lstat(one->path, &st) < 0)
4021 die_errno("stat '%s'", one->path);
4022 if (index_path(one->oid.hash, one->path, &st, 0))
4023 die("cannot hash %s", one->path);
4024 }
4025 }
4026 else
4027 oidclr(&one->oid);
4028 }
4029
4030 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4031 {
4032 /* Strip the prefix but do not molest /dev/null and absolute paths */
4033 if (*namep && **namep != '/') {
4034 *namep += prefix_length;
4035 if (**namep == '/')
4036 ++*namep;
4037 }
4038 if (*otherp && **otherp != '/') {
4039 *otherp += prefix_length;
4040 if (**otherp == '/')
4041 ++*otherp;
4042 }
4043 }
4044
4045 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4046 {
4047 const char *pgm = external_diff();
4048 struct strbuf msg;
4049 struct diff_filespec *one = p->one;
4050 struct diff_filespec *two = p->two;
4051 const char *name;
4052 const char *other;
4053 const char *attr_path;
4054
4055 name = one->path;
4056 other = (strcmp(name, two->path) ? two->path : NULL);
4057 attr_path = name;
4058 if (o->prefix_length)
4059 strip_prefix(o->prefix_length, &name, &other);
4060
4061 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
4062 pgm = NULL;
4063
4064 if (DIFF_PAIR_UNMERGED(p)) {
4065 run_diff_cmd(pgm, name, NULL, attr_path,
4066 NULL, NULL, NULL, o, p);
4067 return;
4068 }
4069
4070 diff_fill_oid_info(one);
4071 diff_fill_oid_info(two);
4072
4073 if (!pgm &&
4074 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4075 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4076 /*
4077 * a filepair that changes between file and symlink
4078 * needs to be split into deletion and creation.
4079 */
4080 struct diff_filespec *null = alloc_filespec(two->path);
4081 run_diff_cmd(NULL, name, other, attr_path,
4082 one, null, &msg, o, p);
4083 free(null);
4084 strbuf_release(&msg);
4085
4086 null = alloc_filespec(one->path);
4087 run_diff_cmd(NULL, name, other, attr_path,
4088 null, two, &msg, o, p);
4089 free(null);
4090 }
4091 else
4092 run_diff_cmd(pgm, name, other, attr_path,
4093 one, two, &msg, o, p);
4094
4095 strbuf_release(&msg);
4096 }
4097
4098 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4099 struct diffstat_t *diffstat)
4100 {
4101 const char *name;
4102 const char *other;
4103
4104 if (DIFF_PAIR_UNMERGED(p)) {
4105 /* unmerged */
4106 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4107 return;
4108 }
4109
4110 name = p->one->path;
4111 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4112
4113 if (o->prefix_length)
4114 strip_prefix(o->prefix_length, &name, &other);
4115
4116 diff_fill_oid_info(p->one);
4117 diff_fill_oid_info(p->two);
4118
4119 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4120 }
4121
4122 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4123 {
4124 const char *name;
4125 const char *other;
4126 const char *attr_path;
4127
4128 if (DIFF_PAIR_UNMERGED(p)) {
4129 /* unmerged */
4130 return;
4131 }
4132
4133 name = p->one->path;
4134 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4135 attr_path = other ? other : name;
4136
4137 if (o->prefix_length)
4138 strip_prefix(o->prefix_length, &name, &other);
4139
4140 diff_fill_oid_info(p->one);
4141 diff_fill_oid_info(p->two);
4142
4143 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4144 }
4145
4146 void diff_setup(struct diff_options *options)
4147 {
4148 memcpy(options, &default_diff_options, sizeof(*options));
4149
4150 options->file = stdout;
4151
4152 options->abbrev = DEFAULT_ABBREV;
4153 options->line_termination = '\n';
4154 options->break_opt = -1;
4155 options->rename_limit = -1;
4156 options->dirstat_permille = diff_dirstat_permille_default;
4157 options->context = diff_context_default;
4158 options->interhunkcontext = diff_interhunk_context_default;
4159 options->ws_error_highlight = ws_error_highlight_default;
4160 DIFF_OPT_SET(options, RENAME_EMPTY);
4161
4162 /* pathchange left =NULL by default */
4163 options->change = diff_change;
4164 options->add_remove = diff_addremove;
4165 options->use_color = diff_use_color_default;
4166 options->detect_rename = diff_detect_rename_default;
4167 options->xdl_opts |= diff_algorithm;
4168 if (diff_indent_heuristic)
4169 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4170
4171 options->orderfile = diff_order_file_cfg;
4172
4173 if (diff_no_prefix) {
4174 options->a_prefix = options->b_prefix = "";
4175 } else if (!diff_mnemonic_prefix) {
4176 options->a_prefix = "a/";
4177 options->b_prefix = "b/";
4178 }
4179
4180 options->color_moved = diff_color_moved_default;
4181 }
4182
4183 void diff_setup_done(struct diff_options *options)
4184 {
4185 int count = 0;
4186
4187 if (options->set_default)
4188 options->set_default(options);
4189
4190 if (options->output_format & DIFF_FORMAT_NAME)
4191 count++;
4192 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
4193 count++;
4194 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
4195 count++;
4196 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
4197 count++;
4198 if (count > 1)
4199 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4200
4201 /*
4202 * Most of the time we can say "there are changes"
4203 * only by checking if there are changed paths, but
4204 * --ignore-whitespace* options force us to look
4205 * inside contents.
4206 */
4207
4208 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
4209 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
4210 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
4211 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
4212 else
4213 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
4214
4215 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
4216 options->detect_rename = DIFF_DETECT_COPY;
4217
4218 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
4219 options->prefix = NULL;
4220 if (options->prefix)
4221 options->prefix_length = strlen(options->prefix);
4222 else
4223 options->prefix_length = 0;
4224
4225 if (options->output_format & (DIFF_FORMAT_NAME |
4226 DIFF_FORMAT_NAME_STATUS |
4227 DIFF_FORMAT_CHECKDIFF |
4228 DIFF_FORMAT_NO_OUTPUT))
4229 options->output_format &= ~(DIFF_FORMAT_RAW |
4230 DIFF_FORMAT_NUMSTAT |
4231 DIFF_FORMAT_DIFFSTAT |
4232 DIFF_FORMAT_SHORTSTAT |
4233 DIFF_FORMAT_DIRSTAT |
4234 DIFF_FORMAT_SUMMARY |
4235 DIFF_FORMAT_PATCH);
4236
4237 /*
4238 * These cases always need recursive; we do not drop caller-supplied
4239 * recursive bits for other formats here.
4240 */
4241 if (options->output_format & (DIFF_FORMAT_PATCH |
4242 DIFF_FORMAT_NUMSTAT |
4243 DIFF_FORMAT_DIFFSTAT |
4244 DIFF_FORMAT_SHORTSTAT |
4245 DIFF_FORMAT_DIRSTAT |
4246 DIFF_FORMAT_SUMMARY |
4247 DIFF_FORMAT_CHECKDIFF))
4248 DIFF_OPT_SET(options, RECURSIVE);
4249 /*
4250 * Also pickaxe would not work very well if you do not say recursive
4251 */
4252 if (options->pickaxe)
4253 DIFF_OPT_SET(options, RECURSIVE);
4254 /*
4255 * When patches are generated, submodules diffed against the work tree
4256 * must be checked for dirtiness too so it can be shown in the output
4257 */
4258 if (options->output_format & DIFF_FORMAT_PATCH)
4259 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
4260
4261 if (options->detect_rename && options->rename_limit < 0)
4262 options->rename_limit = diff_rename_limit_default;
4263 if (options->setup & DIFF_SETUP_USE_CACHE) {
4264 if (!active_cache)
4265 /* read-cache does not die even when it fails
4266 * so it is safe for us to do this here. Also
4267 * it does not smudge active_cache or active_nr
4268 * when it fails, so we do not have to worry about
4269 * cleaning it up ourselves either.
4270 */
4271 read_cache();
4272 }
4273 if (40 < options->abbrev)
4274 options->abbrev = 40; /* full */
4275
4276 /*
4277 * It does not make sense to show the first hit we happened
4278 * to have found. It does not make sense not to return with
4279 * exit code in such a case either.
4280 */
4281 if (DIFF_OPT_TST(options, QUICK)) {
4282 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4283 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4284 }
4285
4286 options->diff_path_counter = 0;
4287
4288 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
4289 die(_("--follow requires exactly one pathspec"));
4290
4291 if (!options->use_color || external_diff())
4292 options->color_moved = 0;
4293 }
4294
4295 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4296 {
4297 char c, *eq;
4298 int len;
4299
4300 if (*arg != '-')
4301 return 0;
4302 c = *++arg;
4303 if (!c)
4304 return 0;
4305 if (c == arg_short) {
4306 c = *++arg;
4307 if (!c)
4308 return 1;
4309 if (val && isdigit(c)) {
4310 char *end;
4311 int n = strtoul(arg, &end, 10);
4312 if (*end)
4313 return 0;
4314 *val = n;
4315 return 1;
4316 }
4317 return 0;
4318 }
4319 if (c != '-')
4320 return 0;
4321 arg++;
4322 eq = strchrnul(arg, '=');
4323 len = eq - arg;
4324 if (!len || strncmp(arg, arg_long, len))
4325 return 0;
4326 if (*eq) {
4327 int n;
4328 char *end;
4329 if (!isdigit(*++eq))
4330 return 0;
4331 n = strtoul(eq, &end, 10);
4332 if (*end)
4333 return 0;
4334 *val = n;
4335 }
4336 return 1;
4337 }
4338
4339 static int diff_scoreopt_parse(const char *opt);
4340
4341 static inline int short_opt(char opt, const char **argv,
4342 const char **optarg)
4343 {
4344 const char *arg = argv[0];
4345 if (arg[0] != '-' || arg[1] != opt)
4346 return 0;
4347 if (arg[2] != '\0') {
4348 *optarg = arg + 2;
4349 return 1;
4350 }
4351 if (!argv[1])
4352 die("Option '%c' requires a value", opt);
4353 *optarg = argv[1];
4354 return 2;
4355 }
4356
4357 int parse_long_opt(const char *opt, const char **argv,
4358 const char **optarg)
4359 {
4360 const char *arg = argv[0];
4361 if (!skip_prefix(arg, "--", &arg))
4362 return 0;
4363 if (!skip_prefix(arg, opt, &arg))
4364 return 0;
4365 if (*arg == '=') { /* stuck form: --option=value */
4366 *optarg = arg + 1;
4367 return 1;
4368 }
4369 if (*arg != '\0')
4370 return 0;
4371 /* separate form: --option value */
4372 if (!argv[1])
4373 die("Option '--%s' requires a value", opt);
4374 *optarg = argv[1];
4375 return 2;
4376 }
4377
4378 static int stat_opt(struct diff_options *options, const char **av)
4379 {
4380 const char *arg = av[0];
4381 char *end;
4382 int width = options->stat_width;
4383 int name_width = options->stat_name_width;
4384 int graph_width = options->stat_graph_width;
4385 int count = options->stat_count;
4386 int argcount = 1;
4387
4388 if (!skip_prefix(arg, "--stat", &arg))
4389 die("BUG: stat option does not begin with --stat: %s", arg);
4390 end = (char *)arg;
4391
4392 switch (*arg) {
4393 case '-':
4394 if (skip_prefix(arg, "-width", &arg)) {
4395 if (*arg == '=')
4396 width = strtoul(arg + 1, &end, 10);
4397 else if (!*arg && !av[1])
4398 die_want_option("--stat-width");
4399 else if (!*arg) {
4400 width = strtoul(av[1], &end, 10);
4401 argcount = 2;
4402 }
4403 } else if (skip_prefix(arg, "-name-width", &arg)) {
4404 if (*arg == '=')
4405 name_width = strtoul(arg + 1, &end, 10);
4406 else if (!*arg && !av[1])
4407 die_want_option("--stat-name-width");
4408 else if (!*arg) {
4409 name_width = strtoul(av[1], &end, 10);
4410 argcount = 2;
4411 }
4412 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4413 if (*arg == '=')
4414 graph_width = strtoul(arg + 1, &end, 10);
4415 else if (!*arg && !av[1])
4416 die_want_option("--stat-graph-width");
4417 else if (!*arg) {
4418 graph_width = strtoul(av[1], &end, 10);
4419 argcount = 2;
4420 }
4421 } else if (skip_prefix(arg, "-count", &arg)) {
4422 if (*arg == '=')
4423 count = strtoul(arg + 1, &end, 10);
4424 else if (!*arg && !av[1])
4425 die_want_option("--stat-count");
4426 else if (!*arg) {
4427 count = strtoul(av[1], &end, 10);
4428 argcount = 2;
4429 }
4430 }
4431 break;
4432 case '=':
4433 width = strtoul(arg+1, &end, 10);
4434 if (*end == ',')
4435 name_width = strtoul(end+1, &end, 10);
4436 if (*end == ',')
4437 count = strtoul(end+1, &end, 10);
4438 }
4439
4440 /* Important! This checks all the error cases! */
4441 if (*end)
4442 return 0;
4443 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4444 options->stat_name_width = name_width;
4445 options->stat_graph_width = graph_width;
4446 options->stat_width = width;
4447 options->stat_count = count;
4448 return argcount;
4449 }
4450
4451 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4452 {
4453 struct strbuf errmsg = STRBUF_INIT;
4454 if (parse_dirstat_params(options, params, &errmsg))
4455 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4456 errmsg.buf);
4457 strbuf_release(&errmsg);
4458 /*
4459 * The caller knows a dirstat-related option is given from the command
4460 * line; allow it to say "return this_function();"
4461 */
4462 options->output_format |= DIFF_FORMAT_DIRSTAT;
4463 return 1;
4464 }
4465
4466 static int parse_submodule_opt(struct diff_options *options, const char *value)
4467 {
4468 if (parse_submodule_params(options, value))
4469 die(_("Failed to parse --submodule option parameter: '%s'"),
4470 value);
4471 return 1;
4472 }
4473
4474 static const char diff_status_letters[] = {
4475 DIFF_STATUS_ADDED,
4476 DIFF_STATUS_COPIED,
4477 DIFF_STATUS_DELETED,
4478 DIFF_STATUS_MODIFIED,
4479 DIFF_STATUS_RENAMED,
4480 DIFF_STATUS_TYPE_CHANGED,
4481 DIFF_STATUS_UNKNOWN,
4482 DIFF_STATUS_UNMERGED,
4483 DIFF_STATUS_FILTER_AON,
4484 DIFF_STATUS_FILTER_BROKEN,
4485 '\0',
4486 };
4487
4488 static unsigned int filter_bit['Z' + 1];
4489
4490 static void prepare_filter_bits(void)
4491 {
4492 int i;
4493
4494 if (!filter_bit[DIFF_STATUS_ADDED]) {
4495 for (i = 0; diff_status_letters[i]; i++)
4496 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4497 }
4498 }
4499
4500 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4501 {
4502 return opt->filter & filter_bit[(int) status];
4503 }
4504
4505 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4506 {
4507 int i, optch;
4508
4509 prepare_filter_bits();
4510
4511 /*
4512 * If there is a negation e.g. 'd' in the input, and we haven't
4513 * initialized the filter field with another --diff-filter, start
4514 * from full set of bits, except for AON.
4515 */
4516 if (!opt->filter) {
4517 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4518 if (optch < 'a' || 'z' < optch)
4519 continue;
4520 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4521 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4522 break;
4523 }
4524 }
4525
4526 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4527 unsigned int bit;
4528 int negate;
4529
4530 if ('a' <= optch && optch <= 'z') {
4531 negate = 1;
4532 optch = toupper(optch);
4533 } else {
4534 negate = 0;
4535 }
4536
4537 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4538 if (!bit)
4539 return optarg[i];
4540 if (negate)
4541 opt->filter &= ~bit;
4542 else
4543 opt->filter |= bit;
4544 }
4545 return 0;
4546 }
4547
4548 static void enable_patch_output(int *fmt) {
4549 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4550 *fmt |= DIFF_FORMAT_PATCH;
4551 }
4552
4553 static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4554 {
4555 int val = parse_ws_error_highlight(arg);
4556
4557 if (val < 0) {
4558 error("unknown value after ws-error-highlight=%.*s",
4559 -1 - val, arg);
4560 return 0;
4561 }
4562 opt->ws_error_highlight = val;
4563 return 1;
4564 }
4565
4566 int diff_opt_parse(struct diff_options *options,
4567 const char **av, int ac, const char *prefix)
4568 {
4569 const char *arg = av[0];
4570 const char *optarg;
4571 int argcount;
4572
4573 if (!prefix)
4574 prefix = "";
4575
4576 /* Output format options */
4577 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4578 || opt_arg(arg, 'U', "unified", &options->context))
4579 enable_patch_output(&options->output_format);
4580 else if (!strcmp(arg, "--raw"))
4581 options->output_format |= DIFF_FORMAT_RAW;
4582 else if (!strcmp(arg, "--patch-with-raw")) {
4583 enable_patch_output(&options->output_format);
4584 options->output_format |= DIFF_FORMAT_RAW;
4585 } else if (!strcmp(arg, "--numstat"))
4586 options->output_format |= DIFF_FORMAT_NUMSTAT;
4587 else if (!strcmp(arg, "--shortstat"))
4588 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4589 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
4590 return parse_dirstat_opt(options, "");
4591 else if (skip_prefix(arg, "-X", &arg))
4592 return parse_dirstat_opt(options, arg);
4593 else if (skip_prefix(arg, "--dirstat=", &arg))
4594 return parse_dirstat_opt(options, arg);
4595 else if (!strcmp(arg, "--cumulative"))
4596 return parse_dirstat_opt(options, "cumulative");
4597 else if (!strcmp(arg, "--dirstat-by-file"))
4598 return parse_dirstat_opt(options, "files");
4599 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
4600 parse_dirstat_opt(options, "files");
4601 return parse_dirstat_opt(options, arg);
4602 }
4603 else if (!strcmp(arg, "--check"))
4604 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4605 else if (!strcmp(arg, "--summary"))
4606 options->output_format |= DIFF_FORMAT_SUMMARY;
4607 else if (!strcmp(arg, "--patch-with-stat")) {
4608 enable_patch_output(&options->output_format);
4609 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4610 } else if (!strcmp(arg, "--name-only"))
4611 options->output_format |= DIFF_FORMAT_NAME;
4612 else if (!strcmp(arg, "--name-status"))
4613 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4614 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4615 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4616 else if (starts_with(arg, "--stat"))
4617 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4618 return stat_opt(options, av);
4619
4620 /* renames options */
4621 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
4622 !strcmp(arg, "--break-rewrites")) {
4623 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4624 return error("invalid argument to -B: %s", arg+2);
4625 }
4626 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
4627 !strcmp(arg, "--find-renames")) {
4628 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4629 return error("invalid argument to -M: %s", arg+2);
4630 options->detect_rename = DIFF_DETECT_RENAME;
4631 }
4632 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4633 options->irreversible_delete = 1;
4634 }
4635 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
4636 !strcmp(arg, "--find-copies")) {
4637 if (options->detect_rename == DIFF_DETECT_COPY)
4638 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4639 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4640 return error("invalid argument to -C: %s", arg+2);
4641 options->detect_rename = DIFF_DETECT_COPY;
4642 }
4643 else if (!strcmp(arg, "--no-renames"))
4644 options->detect_rename = 0;
4645 else if (!strcmp(arg, "--rename-empty"))
4646 DIFF_OPT_SET(options, RENAME_EMPTY);
4647 else if (!strcmp(arg, "--no-rename-empty"))
4648 DIFF_OPT_CLR(options, RENAME_EMPTY);
4649 else if (!strcmp(arg, "--relative"))
4650 DIFF_OPT_SET(options, RELATIVE_NAME);
4651 else if (skip_prefix(arg, "--relative=", &arg)) {
4652 DIFF_OPT_SET(options, RELATIVE_NAME);
4653 options->prefix = arg;
4654 }
4655
4656 /* xdiff options */
4657 else if (!strcmp(arg, "--minimal"))
4658 DIFF_XDL_SET(options, NEED_MINIMAL);
4659 else if (!strcmp(arg, "--no-minimal"))
4660 DIFF_XDL_CLR(options, NEED_MINIMAL);
4661 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4662 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4663 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4664 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4665 else if (!strcmp(arg, "--ignore-space-at-eol"))
4666 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4667 else if (!strcmp(arg, "--ignore-blank-lines"))
4668 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4669 else if (!strcmp(arg, "--indent-heuristic"))
4670 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4671 else if (!strcmp(arg, "--no-indent-heuristic"))
4672 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4673 else if (!strcmp(arg, "--patience"))
4674 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4675 else if (!strcmp(arg, "--histogram"))
4676 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4677 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4678 long value = parse_algorithm_value(optarg);
4679 if (value < 0)
4680 return error("option diff-algorithm accepts \"myers\", "
4681 "\"minimal\", \"patience\" and \"histogram\"");
4682 /* clear out previous settings */
4683 DIFF_XDL_CLR(options, NEED_MINIMAL);
4684 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4685 options->xdl_opts |= value;
4686 return argcount;
4687 }
4688
4689 /* flags options */
4690 else if (!strcmp(arg, "--binary")) {
4691 enable_patch_output(&options->output_format);
4692 DIFF_OPT_SET(options, BINARY);
4693 }
4694 else if (!strcmp(arg, "--full-index"))
4695 DIFF_OPT_SET(options, FULL_INDEX);
4696 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4697 DIFF_OPT_SET(options, TEXT);
4698 else if (!strcmp(arg, "-R"))
4699 DIFF_OPT_SET(options, REVERSE_DIFF);
4700 else if (!strcmp(arg, "--find-copies-harder"))
4701 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4702 else if (!strcmp(arg, "--follow"))
4703 DIFF_OPT_SET(options, FOLLOW_RENAMES);
4704 else if (!strcmp(arg, "--no-follow")) {
4705 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
4706 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
4707 } else if (!strcmp(arg, "--color"))
4708 options->use_color = 1;
4709 else if (skip_prefix(arg, "--color=", &arg)) {
4710 int value = git_config_colorbool(NULL, arg);
4711 if (value < 0)
4712 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4713 options->use_color = value;
4714 }
4715 else if (!strcmp(arg, "--no-color"))
4716 options->use_color = 0;
4717 else if (!strcmp(arg, "--color-moved")) {
4718 if (diff_color_moved_default)
4719 options->color_moved = diff_color_moved_default;
4720 if (options->color_moved == COLOR_MOVED_NO)
4721 options->color_moved = COLOR_MOVED_DEFAULT;
4722 } else if (!strcmp(arg, "--no-color-moved"))
4723 options->color_moved = COLOR_MOVED_NO;
4724 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4725 int cm = parse_color_moved(arg);
4726 if (cm < 0)
4727 die("bad --color-moved argument: %s", arg);
4728 options->color_moved = cm;
4729 } else if (!strcmp(arg, "--color-words")) {
4730 options->use_color = 1;
4731 options->word_diff = DIFF_WORDS_COLOR;
4732 }
4733 else if (skip_prefix(arg, "--color-words=", &arg)) {
4734 options->use_color = 1;
4735 options->word_diff = DIFF_WORDS_COLOR;
4736 options->word_regex = arg;
4737 }
4738 else if (!strcmp(arg, "--word-diff")) {
4739 if (options->word_diff == DIFF_WORDS_NONE)
4740 options->word_diff = DIFF_WORDS_PLAIN;
4741 }
4742 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4743 if (!strcmp(arg, "plain"))
4744 options->word_diff = DIFF_WORDS_PLAIN;
4745 else if (!strcmp(arg, "color")) {
4746 options->use_color = 1;
4747 options->word_diff = DIFF_WORDS_COLOR;
4748 }
4749 else if (!strcmp(arg, "porcelain"))
4750 options->word_diff = DIFF_WORDS_PORCELAIN;
4751 else if (!strcmp(arg, "none"))
4752 options->word_diff = DIFF_WORDS_NONE;
4753 else
4754 die("bad --word-diff argument: %s", arg);
4755 }
4756 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
4757 if (options->word_diff == DIFF_WORDS_NONE)
4758 options->word_diff = DIFF_WORDS_PLAIN;
4759 options->word_regex = optarg;
4760 return argcount;
4761 }
4762 else if (!strcmp(arg, "--exit-code"))
4763 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4764 else if (!strcmp(arg, "--quiet"))
4765 DIFF_OPT_SET(options, QUICK);
4766 else if (!strcmp(arg, "--ext-diff"))
4767 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
4768 else if (!strcmp(arg, "--no-ext-diff"))
4769 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
4770 else if (!strcmp(arg, "--textconv"))
4771 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
4772 else if (!strcmp(arg, "--no-textconv"))
4773 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
4774 else if (!strcmp(arg, "--ignore-submodules")) {
4775 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4776 handle_ignore_submodules_arg(options, "all");
4777 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
4778 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4779 handle_ignore_submodules_arg(options, arg);
4780 } else if (!strcmp(arg, "--submodule"))
4781 options->submodule_format = DIFF_SUBMODULE_LOG;
4782 else if (skip_prefix(arg, "--submodule=", &arg))
4783 return parse_submodule_opt(options, arg);
4784 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
4785 return parse_ws_error_highlight_opt(options, arg);
4786 else if (!strcmp(arg, "--ita-invisible-in-index"))
4787 options->ita_invisible_in_index = 1;
4788 else if (!strcmp(arg, "--ita-visible-in-index"))
4789 options->ita_invisible_in_index = 0;
4790
4791 /* misc options */
4792 else if (!strcmp(arg, "-z"))
4793 options->line_termination = 0;
4794 else if ((argcount = short_opt('l', av, &optarg))) {
4795 options->rename_limit = strtoul(optarg, NULL, 10);
4796 return argcount;
4797 }
4798 else if ((argcount = short_opt('S', av, &optarg))) {
4799 options->pickaxe = optarg;
4800 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4801 return argcount;
4802 } else if ((argcount = short_opt('G', av, &optarg))) {
4803 options->pickaxe = optarg;
4804 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4805 return argcount;
4806 }
4807 else if (!strcmp(arg, "--pickaxe-all"))
4808 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4809 else if (!strcmp(arg, "--pickaxe-regex"))
4810 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4811 else if ((argcount = short_opt('O', av, &optarg))) {
4812 options->orderfile = prefix_filename(prefix, optarg);
4813 return argcount;
4814 }
4815 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4816 int offending = parse_diff_filter_opt(optarg, options);
4817 if (offending)
4818 die("unknown change class '%c' in --diff-filter=%s",
4819 offending, optarg);
4820 return argcount;
4821 }
4822 else if (!strcmp(arg, "--no-abbrev"))
4823 options->abbrev = 0;
4824 else if (!strcmp(arg, "--abbrev"))
4825 options->abbrev = DEFAULT_ABBREV;
4826 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4827 options->abbrev = strtoul(arg, NULL, 10);
4828 if (options->abbrev < MINIMUM_ABBREV)
4829 options->abbrev = MINIMUM_ABBREV;
4830 else if (40 < options->abbrev)
4831 options->abbrev = 40;
4832 }
4833 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4834 options->a_prefix = optarg;
4835 return argcount;
4836 }
4837 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4838 options->line_prefix = optarg;
4839 options->line_prefix_length = strlen(options->line_prefix);
4840 graph_setup_line_prefix(options);
4841 return argcount;
4842 }
4843 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4844 options->b_prefix = optarg;
4845 return argcount;
4846 }
4847 else if (!strcmp(arg, "--no-prefix"))
4848 options->a_prefix = options->b_prefix = "";
4849 else if (opt_arg(arg, '\0', "inter-hunk-context",
4850 &options->interhunkcontext))
4851 ;
4852 else if (!strcmp(arg, "-W"))
4853 DIFF_OPT_SET(options, FUNCCONTEXT);
4854 else if (!strcmp(arg, "--function-context"))
4855 DIFF_OPT_SET(options, FUNCCONTEXT);
4856 else if (!strcmp(arg, "--no-function-context"))
4857 DIFF_OPT_CLR(options, FUNCCONTEXT);
4858 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4859 char *path = prefix_filename(prefix, optarg);
4860 options->file = xfopen(path, "w");
4861 options->close_file = 1;
4862 if (options->use_color != GIT_COLOR_ALWAYS)
4863 options->use_color = GIT_COLOR_NEVER;
4864 free(path);
4865 return argcount;
4866 } else
4867 return 0;
4868 return 1;
4869 }
4870
4871 int parse_rename_score(const char **cp_p)
4872 {
4873 unsigned long num, scale;
4874 int ch, dot;
4875 const char *cp = *cp_p;
4876
4877 num = 0;
4878 scale = 1;
4879 dot = 0;
4880 for (;;) {
4881 ch = *cp;
4882 if ( !dot && ch == '.' ) {
4883 scale = 1;
4884 dot = 1;
4885 } else if ( ch == '%' ) {
4886 scale = dot ? scale*100 : 100;
4887 cp++; /* % is always at the end */
4888 break;
4889 } else if ( ch >= '0' && ch <= '9' ) {
4890 if ( scale < 100000 ) {
4891 scale *= 10;
4892 num = (num*10) + (ch-'0');
4893 }
4894 } else {
4895 break;
4896 }
4897 cp++;
4898 }
4899 *cp_p = cp;
4900
4901 /* user says num divided by scale and we say internally that
4902 * is MAX_SCORE * num / scale.
4903 */
4904 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4905 }
4906
4907 static int diff_scoreopt_parse(const char *opt)
4908 {
4909 int opt1, opt2, cmd;
4910
4911 if (*opt++ != '-')
4912 return -1;
4913 cmd = *opt++;
4914 if (cmd == '-') {
4915 /* convert the long-form arguments into short-form versions */
4916 if (skip_prefix(opt, "break-rewrites", &opt)) {
4917 if (*opt == 0 || *opt++ == '=')
4918 cmd = 'B';
4919 } else if (skip_prefix(opt, "find-copies", &opt)) {
4920 if (*opt == 0 || *opt++ == '=')
4921 cmd = 'C';
4922 } else if (skip_prefix(opt, "find-renames", &opt)) {
4923 if (*opt == 0 || *opt++ == '=')
4924 cmd = 'M';
4925 }
4926 }
4927 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4928 return -1; /* that is not a -M, -C, or -B option */
4929
4930 opt1 = parse_rename_score(&opt);
4931 if (cmd != 'B')
4932 opt2 = 0;
4933 else {
4934 if (*opt == 0)
4935 opt2 = 0;
4936 else if (*opt != '/')
4937 return -1; /* we expect -B80/99 or -B80 */
4938 else {
4939 opt++;
4940 opt2 = parse_rename_score(&opt);
4941 }
4942 }
4943 if (*opt != 0)
4944 return -1;
4945 return opt1 | (opt2 << 16);
4946 }
4947
4948 struct diff_queue_struct diff_queued_diff;
4949
4950 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4951 {
4952 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4953 queue->queue[queue->nr++] = dp;
4954 }
4955
4956 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4957 struct diff_filespec *one,
4958 struct diff_filespec *two)
4959 {
4960 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4961 dp->one = one;
4962 dp->two = two;
4963 if (queue)
4964 diff_q(queue, dp);
4965 return dp;
4966 }
4967
4968 void diff_free_filepair(struct diff_filepair *p)
4969 {
4970 free_filespec(p->one);
4971 free_filespec(p->two);
4972 free(p);
4973 }
4974
4975 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4976 {
4977 int abblen;
4978 const char *abbrev;
4979
4980 if (len == GIT_SHA1_HEXSZ)
4981 return oid_to_hex(oid);
4982
4983 abbrev = diff_abbrev_oid(oid, len);
4984 abblen = strlen(abbrev);
4985
4986 /*
4987 * In well-behaved cases, where the abbbreviated result is the
4988 * same as the requested length, append three dots after the
4989 * abbreviation (hence the whole logic is limited to the case
4990 * where abblen < 37); when the actual abbreviated result is a
4991 * bit longer than the requested length, we reduce the number
4992 * of dots so that they match the well-behaved ones. However,
4993 * if the actual abbreviation is longer than the requested
4994 * length by more than three, we give up on aligning, and add
4995 * three dots anyway, to indicate that the output is not the
4996 * full object name. Yes, this may be suboptimal, but this
4997 * appears only in "diff --raw --abbrev" output and it is not
4998 * worth the effort to change it now. Note that this would
4999 * likely to work fine when the automatic sizing of default
5000 * abbreviation length is used--we would be fed -1 in "len" in
5001 * that case, and will end up always appending three-dots, but
5002 * the automatic sizing is supposed to give abblen that ensures
5003 * uniqueness across all objects (statistically speaking).
5004 */
5005 if (abblen < GIT_SHA1_HEXSZ - 3) {
5006 static char hex[GIT_MAX_HEXSZ + 1];
5007 if (len < abblen && abblen <= len + 2)
5008 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5009 else
5010 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5011 return hex;
5012 }
5013
5014 return oid_to_hex(oid);
5015 }
5016
5017 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5018 {
5019 int line_termination = opt->line_termination;
5020 int inter_name_termination = line_termination ? '\t' : '\0';
5021
5022 fprintf(opt->file, "%s", diff_line_prefix(opt));
5023 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5024 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5025 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5026 fprintf(opt->file, "%s ",
5027 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5028 }
5029 if (p->score) {
5030 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5031 inter_name_termination);
5032 } else {
5033 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5034 }
5035
5036 if (p->status == DIFF_STATUS_COPIED ||
5037 p->status == DIFF_STATUS_RENAMED) {
5038 const char *name_a, *name_b;
5039 name_a = p->one->path;
5040 name_b = p->two->path;
5041 strip_prefix(opt->prefix_length, &name_a, &name_b);
5042 write_name_quoted(name_a, opt->file, inter_name_termination);
5043 write_name_quoted(name_b, opt->file, line_termination);
5044 } else {
5045 const char *name_a, *name_b;
5046 name_a = p->one->mode ? p->one->path : p->two->path;
5047 name_b = NULL;
5048 strip_prefix(opt->prefix_length, &name_a, &name_b);
5049 write_name_quoted(name_a, opt->file, line_termination);
5050 }
5051 }
5052
5053 int diff_unmodified_pair(struct diff_filepair *p)
5054 {
5055 /* This function is written stricter than necessary to support
5056 * the currently implemented transformers, but the idea is to
5057 * let transformers to produce diff_filepairs any way they want,
5058 * and filter and clean them up here before producing the output.
5059 */
5060 struct diff_filespec *one = p->one, *two = p->two;
5061
5062 if (DIFF_PAIR_UNMERGED(p))
5063 return 0; /* unmerged is interesting */
5064
5065 /* deletion, addition, mode or type change
5066 * and rename are all interesting.
5067 */
5068 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5069 DIFF_PAIR_MODE_CHANGED(p) ||
5070 strcmp(one->path, two->path))
5071 return 0;
5072
5073 /* both are valid and point at the same path. that is, we are
5074 * dealing with a change.
5075 */
5076 if (one->oid_valid && two->oid_valid &&
5077 !oidcmp(&one->oid, &two->oid) &&
5078 !one->dirty_submodule && !two->dirty_submodule)
5079 return 1; /* no change */
5080 if (!one->oid_valid && !two->oid_valid)
5081 return 1; /* both look at the same file on the filesystem. */
5082 return 0;
5083 }
5084
5085 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5086 {
5087 if (diff_unmodified_pair(p))
5088 return;
5089
5090 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5091 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5092 return; /* no tree diffs in patch format */
5093
5094 run_diff(p, o);
5095 }
5096
5097 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5098 struct diffstat_t *diffstat)
5099 {
5100 if (diff_unmodified_pair(p))
5101 return;
5102
5103 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5104 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5105 return; /* no useful stat for tree diffs */
5106
5107 run_diffstat(p, o, diffstat);
5108 }
5109
5110 static void diff_flush_checkdiff(struct diff_filepair *p,
5111 struct diff_options *o)
5112 {
5113 if (diff_unmodified_pair(p))
5114 return;
5115
5116 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5117 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5118 return; /* nothing to check in tree diffs */
5119
5120 run_checkdiff(p, o);
5121 }
5122
5123 int diff_queue_is_empty(void)
5124 {
5125 struct diff_queue_struct *q = &diff_queued_diff;
5126 int i;
5127 for (i = 0; i < q->nr; i++)
5128 if (!diff_unmodified_pair(q->queue[i]))
5129 return 0;
5130 return 1;
5131 }
5132
5133 #if DIFF_DEBUG
5134 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5135 {
5136 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5137 x, one ? one : "",
5138 s->path,
5139 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5140 s->mode,
5141 s->oid_valid ? oid_to_hex(&s->oid) : "");
5142 fprintf(stderr, "queue[%d] %s size %lu\n",
5143 x, one ? one : "",
5144 s->size);
5145 }
5146
5147 void diff_debug_filepair(const struct diff_filepair *p, int i)
5148 {
5149 diff_debug_filespec(p->one, i, "one");
5150 diff_debug_filespec(p->two, i, "two");
5151 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5152 p->score, p->status ? p->status : '?',
5153 p->one->rename_used, p->broken_pair);
5154 }
5155
5156 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5157 {
5158 int i;
5159 if (msg)
5160 fprintf(stderr, "%s\n", msg);
5161 fprintf(stderr, "q->nr = %d\n", q->nr);
5162 for (i = 0; i < q->nr; i++) {
5163 struct diff_filepair *p = q->queue[i];
5164 diff_debug_filepair(p, i);
5165 }
5166 }
5167 #endif
5168
5169 static void diff_resolve_rename_copy(void)
5170 {
5171 int i;
5172 struct diff_filepair *p;
5173 struct diff_queue_struct *q = &diff_queued_diff;
5174
5175 diff_debug_queue("resolve-rename-copy", q);
5176
5177 for (i = 0; i < q->nr; i++) {
5178 p = q->queue[i];
5179 p->status = 0; /* undecided */
5180 if (DIFF_PAIR_UNMERGED(p))
5181 p->status = DIFF_STATUS_UNMERGED;
5182 else if (!DIFF_FILE_VALID(p->one))
5183 p->status = DIFF_STATUS_ADDED;
5184 else if (!DIFF_FILE_VALID(p->two))
5185 p->status = DIFF_STATUS_DELETED;
5186 else if (DIFF_PAIR_TYPE_CHANGED(p))
5187 p->status = DIFF_STATUS_TYPE_CHANGED;
5188
5189 /* from this point on, we are dealing with a pair
5190 * whose both sides are valid and of the same type, i.e.
5191 * either in-place edit or rename/copy edit.
5192 */
5193 else if (DIFF_PAIR_RENAME(p)) {
5194 /*
5195 * A rename might have re-connected a broken
5196 * pair up, causing the pathnames to be the
5197 * same again. If so, that's not a rename at
5198 * all, just a modification..
5199 *
5200 * Otherwise, see if this source was used for
5201 * multiple renames, in which case we decrement
5202 * the count, and call it a copy.
5203 */
5204 if (!strcmp(p->one->path, p->two->path))
5205 p->status = DIFF_STATUS_MODIFIED;
5206 else if (--p->one->rename_used > 0)
5207 p->status = DIFF_STATUS_COPIED;
5208 else
5209 p->status = DIFF_STATUS_RENAMED;
5210 }
5211 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5212 p->one->mode != p->two->mode ||
5213 p->one->dirty_submodule ||
5214 p->two->dirty_submodule ||
5215 is_null_oid(&p->one->oid))
5216 p->status = DIFF_STATUS_MODIFIED;
5217 else {
5218 /* This is a "no-change" entry and should not
5219 * happen anymore, but prepare for broken callers.
5220 */
5221 error("feeding unmodified %s to diffcore",
5222 p->one->path);
5223 p->status = DIFF_STATUS_UNKNOWN;
5224 }
5225 }
5226 diff_debug_queue("resolve-rename-copy done", q);
5227 }
5228
5229 static int check_pair_status(struct diff_filepair *p)
5230 {
5231 switch (p->status) {
5232 case DIFF_STATUS_UNKNOWN:
5233 return 0;
5234 case 0:
5235 die("internal error in diff-resolve-rename-copy");
5236 default:
5237 return 1;
5238 }
5239 }
5240
5241 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5242 {
5243 int fmt = opt->output_format;
5244
5245 if (fmt & DIFF_FORMAT_CHECKDIFF)
5246 diff_flush_checkdiff(p, opt);
5247 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5248 diff_flush_raw(p, opt);
5249 else if (fmt & DIFF_FORMAT_NAME) {
5250 const char *name_a, *name_b;
5251 name_a = p->two->path;
5252 name_b = NULL;
5253 strip_prefix(opt->prefix_length, &name_a, &name_b);
5254 fprintf(opt->file, "%s", diff_line_prefix(opt));
5255 write_name_quoted(name_a, opt->file, opt->line_termination);
5256 }
5257 }
5258
5259 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5260 {
5261 struct strbuf sb = STRBUF_INIT;
5262 if (fs->mode)
5263 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5264 else
5265 strbuf_addf(&sb, " %s ", newdelete);
5266
5267 quote_c_style(fs->path, &sb, NULL, 0);
5268 strbuf_addch(&sb, '\n');
5269 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5270 sb.buf, sb.len, 0);
5271 strbuf_release(&sb);
5272 }
5273
5274 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5275 int show_name)
5276 {
5277 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5278 struct strbuf sb = STRBUF_INIT;
5279 strbuf_addf(&sb, " mode change %06o => %06o",
5280 p->one->mode, p->two->mode);
5281 if (show_name) {
5282 strbuf_addch(&sb, ' ');
5283 quote_c_style(p->two->path, &sb, NULL, 0);
5284 }
5285 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5286 sb.buf, sb.len, 0);
5287 strbuf_release(&sb);
5288 }
5289 }
5290
5291 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5292 struct diff_filepair *p)
5293 {
5294 struct strbuf sb = STRBUF_INIT;
5295 char *names = pprint_rename(p->one->path, p->two->path);
5296 strbuf_addf(&sb, " %s %s (%d%%)\n",
5297 renamecopy, names, similarity_index(p));
5298 free(names);
5299 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5300 sb.buf, sb.len, 0);
5301 show_mode_change(opt, p, 0);
5302 }
5303
5304 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5305 {
5306 switch(p->status) {
5307 case DIFF_STATUS_DELETED:
5308 show_file_mode_name(opt, "delete", p->one);
5309 break;
5310 case DIFF_STATUS_ADDED:
5311 show_file_mode_name(opt, "create", p->two);
5312 break;
5313 case DIFF_STATUS_COPIED:
5314 show_rename_copy(opt, "copy", p);
5315 break;
5316 case DIFF_STATUS_RENAMED:
5317 show_rename_copy(opt, "rename", p);
5318 break;
5319 default:
5320 if (p->score) {
5321 struct strbuf sb = STRBUF_INIT;
5322 strbuf_addstr(&sb, " rewrite ");
5323 quote_c_style(p->two->path, &sb, NULL, 0);
5324 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5325 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5326 sb.buf, sb.len, 0);
5327 }
5328 show_mode_change(opt, p, !p->score);
5329 break;
5330 }
5331 }
5332
5333 struct patch_id_t {
5334 git_SHA_CTX *ctx;
5335 int patchlen;
5336 };
5337
5338 static int remove_space(char *line, int len)
5339 {
5340 int i;
5341 char *dst = line;
5342 unsigned char c;
5343
5344 for (i = 0; i < len; i++)
5345 if (!isspace((c = line[i])))
5346 *dst++ = c;
5347
5348 return dst - line;
5349 }
5350
5351 static void patch_id_consume(void *priv, char *line, unsigned long len)
5352 {
5353 struct patch_id_t *data = priv;
5354 int new_len;
5355
5356 /* Ignore line numbers when computing the SHA1 of the patch */
5357 if (starts_with(line, "@@ -"))
5358 return;
5359
5360 new_len = remove_space(line, len);
5361
5362 git_SHA1_Update(data->ctx, line, new_len);
5363 data->patchlen += new_len;
5364 }
5365
5366 static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5367 {
5368 git_SHA1_Update(ctx, str, strlen(str));
5369 }
5370
5371 static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5372 {
5373 /* large enough for 2^32 in octal */
5374 char buf[12];
5375 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5376 git_SHA1_Update(ctx, buf, len);
5377 }
5378
5379 /* returns 0 upon success, and writes result into sha1 */
5380 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5381 {
5382 struct diff_queue_struct *q = &diff_queued_diff;
5383 int i;
5384 git_SHA_CTX ctx;
5385 struct patch_id_t data;
5386
5387 git_SHA1_Init(&ctx);
5388 memset(&data, 0, sizeof(struct patch_id_t));
5389 data.ctx = &ctx;
5390
5391 for (i = 0; i < q->nr; i++) {
5392 xpparam_t xpp;
5393 xdemitconf_t xecfg;
5394 mmfile_t mf1, mf2;
5395 struct diff_filepair *p = q->queue[i];
5396 int len1, len2;
5397
5398 memset(&xpp, 0, sizeof(xpp));
5399 memset(&xecfg, 0, sizeof(xecfg));
5400 if (p->status == 0)
5401 return error("internal diff status error");
5402 if (p->status == DIFF_STATUS_UNKNOWN)
5403 continue;
5404 if (diff_unmodified_pair(p))
5405 continue;
5406 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5407 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5408 continue;
5409 if (DIFF_PAIR_UNMERGED(p))
5410 continue;
5411
5412 diff_fill_oid_info(p->one);
5413 diff_fill_oid_info(p->two);
5414
5415 len1 = remove_space(p->one->path, strlen(p->one->path));
5416 len2 = remove_space(p->two->path, strlen(p->two->path));
5417 patch_id_add_string(&ctx, "diff--git");
5418 patch_id_add_string(&ctx, "a/");
5419 git_SHA1_Update(&ctx, p->one->path, len1);
5420 patch_id_add_string(&ctx, "b/");
5421 git_SHA1_Update(&ctx, p->two->path, len2);
5422
5423 if (p->one->mode == 0) {
5424 patch_id_add_string(&ctx, "newfilemode");
5425 patch_id_add_mode(&ctx, p->two->mode);
5426 patch_id_add_string(&ctx, "---/dev/null");
5427 patch_id_add_string(&ctx, "+++b/");
5428 git_SHA1_Update(&ctx, p->two->path, len2);
5429 } else if (p->two->mode == 0) {
5430 patch_id_add_string(&ctx, "deletedfilemode");
5431 patch_id_add_mode(&ctx, p->one->mode);
5432 patch_id_add_string(&ctx, "---a/");
5433 git_SHA1_Update(&ctx, p->one->path, len1);
5434 patch_id_add_string(&ctx, "+++/dev/null");
5435 } else {
5436 patch_id_add_string(&ctx, "---a/");
5437 git_SHA1_Update(&ctx, p->one->path, len1);
5438 patch_id_add_string(&ctx, "+++b/");
5439 git_SHA1_Update(&ctx, p->two->path, len2);
5440 }
5441
5442 if (diff_header_only)
5443 continue;
5444
5445 if (fill_mmfile(&mf1, p->one) < 0 ||
5446 fill_mmfile(&mf2, p->two) < 0)
5447 return error("unable to read files to diff");
5448
5449 if (diff_filespec_is_binary(p->one) ||
5450 diff_filespec_is_binary(p->two)) {
5451 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5452 GIT_SHA1_HEXSZ);
5453 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5454 GIT_SHA1_HEXSZ);
5455 continue;
5456 }
5457
5458 xpp.flags = 0;
5459 xecfg.ctxlen = 3;
5460 xecfg.flags = 0;
5461 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
5462 &xpp, &xecfg))
5463 return error("unable to generate patch-id diff for %s",
5464 p->one->path);
5465 }
5466
5467 git_SHA1_Final(oid->hash, &ctx);
5468 return 0;
5469 }
5470
5471 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5472 {
5473 struct diff_queue_struct *q = &diff_queued_diff;
5474 int i;
5475 int result = diff_get_patch_id(options, oid, diff_header_only);
5476
5477 for (i = 0; i < q->nr; i++)
5478 diff_free_filepair(q->queue[i]);
5479
5480 free(q->queue);
5481 DIFF_QUEUE_CLEAR(q);
5482
5483 return result;
5484 }
5485
5486 static int is_summary_empty(const struct diff_queue_struct *q)
5487 {
5488 int i;
5489
5490 for (i = 0; i < q->nr; i++) {
5491 const struct diff_filepair *p = q->queue[i];
5492
5493 switch (p->status) {
5494 case DIFF_STATUS_DELETED:
5495 case DIFF_STATUS_ADDED:
5496 case DIFF_STATUS_COPIED:
5497 case DIFF_STATUS_RENAMED:
5498 return 0;
5499 default:
5500 if (p->score)
5501 return 0;
5502 if (p->one->mode && p->two->mode &&
5503 p->one->mode != p->two->mode)
5504 return 0;
5505 break;
5506 }
5507 }
5508 return 1;
5509 }
5510
5511 static const char rename_limit_warning[] =
5512 N_("inexact rename detection was skipped due to too many files.");
5513
5514 static const char degrade_cc_to_c_warning[] =
5515 N_("only found copies from modified paths due to too many files.");
5516
5517 static const char rename_limit_advice[] =
5518 N_("you may want to set your %s variable to at least "
5519 "%d and retry the command.");
5520
5521 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5522 {
5523 if (degraded_cc)
5524 warning(_(degrade_cc_to_c_warning));
5525 else if (needed)
5526 warning(_(rename_limit_warning));
5527 else
5528 return;
5529 if (0 < needed && needed < 32767)
5530 warning(_(rename_limit_advice), varname, needed);
5531 }
5532
5533 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5534 {
5535 int i;
5536 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5537 struct diff_queue_struct *q = &diff_queued_diff;
5538
5539 if (WSEH_NEW & WS_RULE_MASK)
5540 die("BUG: WS rules bit mask overlaps with diff symbol flags");
5541
5542 if (o->color_moved)
5543 o->emitted_symbols = &esm;
5544
5545 for (i = 0; i < q->nr; i++) {
5546 struct diff_filepair *p = q->queue[i];
5547 if (check_pair_status(p))
5548 diff_flush_patch(p, o);
5549 }
5550
5551 if (o->emitted_symbols) {
5552 if (o->color_moved) {
5553 struct hashmap add_lines, del_lines;
5554
5555 hashmap_init(&del_lines,
5556 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5557 hashmap_init(&add_lines,
5558 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5559
5560 add_lines_to_move_detection(o, &add_lines, &del_lines);
5561 mark_color_as_moved(o, &add_lines, &del_lines);
5562 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5563 dim_moved_lines(o);
5564
5565 hashmap_free(&add_lines, 0);
5566 hashmap_free(&del_lines, 0);
5567 }
5568
5569 for (i = 0; i < esm.nr; i++)
5570 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5571
5572 for (i = 0; i < esm.nr; i++)
5573 free((void *)esm.buf[i].line);
5574 }
5575 esm.nr = 0;
5576 }
5577
5578 void diff_flush(struct diff_options *options)
5579 {
5580 struct diff_queue_struct *q = &diff_queued_diff;
5581 int i, output_format = options->output_format;
5582 int separator = 0;
5583 int dirstat_by_line = 0;
5584
5585 /*
5586 * Order: raw, stat, summary, patch
5587 * or: name/name-status/checkdiff (other bits clear)
5588 */
5589 if (!q->nr)
5590 goto free_queue;
5591
5592 if (output_format & (DIFF_FORMAT_RAW |
5593 DIFF_FORMAT_NAME |
5594 DIFF_FORMAT_NAME_STATUS |
5595 DIFF_FORMAT_CHECKDIFF)) {
5596 for (i = 0; i < q->nr; i++) {
5597 struct diff_filepair *p = q->queue[i];
5598 if (check_pair_status(p))
5599 flush_one_pair(p, options);
5600 }
5601 separator++;
5602 }
5603
5604 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
5605 dirstat_by_line = 1;
5606
5607 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5608 dirstat_by_line) {
5609 struct diffstat_t diffstat;
5610
5611 memset(&diffstat, 0, sizeof(struct diffstat_t));
5612 for (i = 0; i < q->nr; i++) {
5613 struct diff_filepair *p = q->queue[i];
5614 if (check_pair_status(p))
5615 diff_flush_stat(p, options, &diffstat);
5616 }
5617 if (output_format & DIFF_FORMAT_NUMSTAT)
5618 show_numstat(&diffstat, options);
5619 if (output_format & DIFF_FORMAT_DIFFSTAT)
5620 show_stats(&diffstat, options);
5621 if (output_format & DIFF_FORMAT_SHORTSTAT)
5622 show_shortstats(&diffstat, options);
5623 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5624 show_dirstat_by_line(&diffstat, options);
5625 free_diffstat_info(&diffstat);
5626 separator++;
5627 }
5628 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5629 show_dirstat(options);
5630
5631 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5632 for (i = 0; i < q->nr; i++) {
5633 diff_summary(options, q->queue[i]);
5634 }
5635 separator++;
5636 }
5637
5638 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5639 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
5640 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5641 /*
5642 * run diff_flush_patch for the exit status. setting
5643 * options->file to /dev/null should be safe, because we
5644 * aren't supposed to produce any output anyway.
5645 */
5646 if (options->close_file)
5647 fclose(options->file);
5648 options->file = xfopen("/dev/null", "w");
5649 options->close_file = 1;
5650 options->color_moved = 0;
5651 for (i = 0; i < q->nr; i++) {
5652 struct diff_filepair *p = q->queue[i];
5653 if (check_pair_status(p))
5654 diff_flush_patch(p, options);
5655 if (options->found_changes)
5656 break;
5657 }
5658 }
5659
5660 if (output_format & DIFF_FORMAT_PATCH) {
5661 if (separator) {
5662 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5663 if (options->stat_sep)
5664 /* attach patch instead of inline */
5665 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5666 NULL, 0, 0);
5667 }
5668
5669 diff_flush_patch_all_file_pairs(options);
5670 }
5671
5672 if (output_format & DIFF_FORMAT_CALLBACK)
5673 options->format_callback(q, options, options->format_callback_data);
5674
5675 for (i = 0; i < q->nr; i++)
5676 diff_free_filepair(q->queue[i]);
5677 free_queue:
5678 free(q->queue);
5679 DIFF_QUEUE_CLEAR(q);
5680 if (options->close_file)
5681 fclose(options->file);
5682
5683 /*
5684 * Report the content-level differences with HAS_CHANGES;
5685 * diff_addremove/diff_change does not set the bit when
5686 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5687 */
5688 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5689 if (options->found_changes)
5690 DIFF_OPT_SET(options, HAS_CHANGES);
5691 else
5692 DIFF_OPT_CLR(options, HAS_CHANGES);
5693 }
5694 }
5695
5696 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5697 {
5698 return (((p->status == DIFF_STATUS_MODIFIED) &&
5699 ((p->score &&
5700 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5701 (!p->score &&
5702 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5703 ((p->status != DIFF_STATUS_MODIFIED) &&
5704 filter_bit_tst(p->status, options)));
5705 }
5706
5707 static void diffcore_apply_filter(struct diff_options *options)
5708 {
5709 int i;
5710 struct diff_queue_struct *q = &diff_queued_diff;
5711 struct diff_queue_struct outq;
5712
5713 DIFF_QUEUE_CLEAR(&outq);
5714
5715 if (!options->filter)
5716 return;
5717
5718 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5719 int found;
5720 for (i = found = 0; !found && i < q->nr; i++) {
5721 if (match_filter(options, q->queue[i]))
5722 found++;
5723 }
5724 if (found)
5725 return;
5726
5727 /* otherwise we will clear the whole queue
5728 * by copying the empty outq at the end of this
5729 * function, but first clear the current entries
5730 * in the queue.
5731 */
5732 for (i = 0; i < q->nr; i++)
5733 diff_free_filepair(q->queue[i]);
5734 }
5735 else {
5736 /* Only the matching ones */
5737 for (i = 0; i < q->nr; i++) {
5738 struct diff_filepair *p = q->queue[i];
5739 if (match_filter(options, p))
5740 diff_q(&outq, p);
5741 else
5742 diff_free_filepair(p);
5743 }
5744 }
5745 free(q->queue);
5746 *q = outq;
5747 }
5748
5749 /* Check whether two filespecs with the same mode and size are identical */
5750 static int diff_filespec_is_identical(struct diff_filespec *one,
5751 struct diff_filespec *two)
5752 {
5753 if (S_ISGITLINK(one->mode))
5754 return 0;
5755 if (diff_populate_filespec(one, 0))
5756 return 0;
5757 if (diff_populate_filespec(two, 0))
5758 return 0;
5759 return !memcmp(one->data, two->data, one->size);
5760 }
5761
5762 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
5763 {
5764 if (p->done_skip_stat_unmatch)
5765 return p->skip_stat_unmatch_result;
5766
5767 p->done_skip_stat_unmatch = 1;
5768 p->skip_stat_unmatch_result = 0;
5769 /*
5770 * 1. Entries that come from stat info dirtiness
5771 * always have both sides (iow, not create/delete),
5772 * one side of the object name is unknown, with
5773 * the same mode and size. Keep the ones that
5774 * do not match these criteria. They have real
5775 * differences.
5776 *
5777 * 2. At this point, the file is known to be modified,
5778 * with the same mode and size, and the object
5779 * name of one side is unknown. Need to inspect
5780 * the identical contents.
5781 */
5782 if (!DIFF_FILE_VALID(p->one) || /* (1) */
5783 !DIFF_FILE_VALID(p->two) ||
5784 (p->one->oid_valid && p->two->oid_valid) ||
5785 (p->one->mode != p->two->mode) ||
5786 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
5787 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
5788 (p->one->size != p->two->size) ||
5789 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
5790 p->skip_stat_unmatch_result = 1;
5791 return p->skip_stat_unmatch_result;
5792 }
5793
5794 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
5795 {
5796 int i;
5797 struct diff_queue_struct *q = &diff_queued_diff;
5798 struct diff_queue_struct outq;
5799 DIFF_QUEUE_CLEAR(&outq);
5800
5801 for (i = 0; i < q->nr; i++) {
5802 struct diff_filepair *p = q->queue[i];
5803
5804 if (diff_filespec_check_stat_unmatch(p))
5805 diff_q(&outq, p);
5806 else {
5807 /*
5808 * The caller can subtract 1 from skip_stat_unmatch
5809 * to determine how many paths were dirty only
5810 * due to stat info mismatch.
5811 */
5812 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
5813 diffopt->skip_stat_unmatch++;
5814 diff_free_filepair(p);
5815 }
5816 }
5817 free(q->queue);
5818 *q = outq;
5819 }
5820
5821 static int diffnamecmp(const void *a_, const void *b_)
5822 {
5823 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
5824 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
5825 const char *name_a, *name_b;
5826
5827 name_a = a->one ? a->one->path : a->two->path;
5828 name_b = b->one ? b->one->path : b->two->path;
5829 return strcmp(name_a, name_b);
5830 }
5831
5832 void diffcore_fix_diff_index(struct diff_options *options)
5833 {
5834 struct diff_queue_struct *q = &diff_queued_diff;
5835 QSORT(q->queue, q->nr, diffnamecmp);
5836 }
5837
5838 void diffcore_std(struct diff_options *options)
5839 {
5840 /* NOTE please keep the following in sync with diff_tree_combined() */
5841 if (options->skip_stat_unmatch)
5842 diffcore_skip_stat_unmatch(options);
5843 if (!options->found_follow) {
5844 /* See try_to_follow_renames() in tree-diff.c */
5845 if (options->break_opt != -1)
5846 diffcore_break(options->break_opt);
5847 if (options->detect_rename)
5848 diffcore_rename(options);
5849 if (options->break_opt != -1)
5850 diffcore_merge_broken();
5851 }
5852 if (options->pickaxe)
5853 diffcore_pickaxe(options);
5854 if (options->orderfile)
5855 diffcore_order(options->orderfile);
5856 if (!options->found_follow)
5857 /* See try_to_follow_renames() in tree-diff.c */
5858 diff_resolve_rename_copy();
5859 diffcore_apply_filter(options);
5860
5861 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5862 DIFF_OPT_SET(options, HAS_CHANGES);
5863 else
5864 DIFF_OPT_CLR(options, HAS_CHANGES);
5865
5866 options->found_follow = 0;
5867 }
5868
5869 int diff_result_code(struct diff_options *opt, int status)
5870 {
5871 int result = 0;
5872
5873 diff_warn_rename_limit("diff.renameLimit",
5874 opt->needed_rename_limit,
5875 opt->degraded_cc_to_c);
5876 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5877 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5878 return status;
5879 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5880 DIFF_OPT_TST(opt, HAS_CHANGES))
5881 result |= 01;
5882 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5883 DIFF_OPT_TST(opt, CHECK_FAILED))
5884 result |= 02;
5885 return result;
5886 }
5887
5888 int diff_can_quit_early(struct diff_options *opt)
5889 {
5890 return (DIFF_OPT_TST(opt, QUICK) &&
5891 !opt->filter &&
5892 DIFF_OPT_TST(opt, HAS_CHANGES));
5893 }
5894
5895 /*
5896 * Shall changes to this submodule be ignored?
5897 *
5898 * Submodule changes can be configured to be ignored separately for each path,
5899 * but that configuration can be overridden from the command line.
5900 */
5901 static int is_submodule_ignored(const char *path, struct diff_options *options)
5902 {
5903 int ignored = 0;
5904 unsigned orig_flags = options->flags;
5905 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5906 set_diffopt_flags_from_submodule_config(options, path);
5907 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5908 ignored = 1;
5909 options->flags = orig_flags;
5910 return ignored;
5911 }
5912
5913 void diff_addremove(struct diff_options *options,
5914 int addremove, unsigned mode,
5915 const struct object_id *oid,
5916 int oid_valid,
5917 const char *concatpath, unsigned dirty_submodule)
5918 {
5919 struct diff_filespec *one, *two;
5920
5921 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5922 return;
5923
5924 /* This may look odd, but it is a preparation for
5925 * feeding "there are unchanged files which should
5926 * not produce diffs, but when you are doing copy
5927 * detection you would need them, so here they are"
5928 * entries to the diff-core. They will be prefixed
5929 * with something like '=' or '*' (I haven't decided
5930 * which but should not make any difference).
5931 * Feeding the same new and old to diff_change()
5932 * also has the same effect.
5933 * Before the final output happens, they are pruned after
5934 * merged into rename/copy pairs as appropriate.
5935 */
5936 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5937 addremove = (addremove == '+' ? '-' :
5938 addremove == '-' ? '+' : addremove);
5939
5940 if (options->prefix &&
5941 strncmp(concatpath, options->prefix, options->prefix_length))
5942 return;
5943
5944 one = alloc_filespec(concatpath);
5945 two = alloc_filespec(concatpath);
5946
5947 if (addremove != '+')
5948 fill_filespec(one, oid, oid_valid, mode);
5949 if (addremove != '-') {
5950 fill_filespec(two, oid, oid_valid, mode);
5951 two->dirty_submodule = dirty_submodule;
5952 }
5953
5954 diff_queue(&diff_queued_diff, one, two);
5955 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5956 DIFF_OPT_SET(options, HAS_CHANGES);
5957 }
5958
5959 void diff_change(struct diff_options *options,
5960 unsigned old_mode, unsigned new_mode,
5961 const struct object_id *old_oid,
5962 const struct object_id *new_oid,
5963 int old_oid_valid, int new_oid_valid,
5964 const char *concatpath,
5965 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5966 {
5967 struct diff_filespec *one, *two;
5968 struct diff_filepair *p;
5969
5970 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5971 is_submodule_ignored(concatpath, options))
5972 return;
5973
5974 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5975 SWAP(old_mode, new_mode);
5976 SWAP(old_oid, new_oid);
5977 SWAP(old_oid_valid, new_oid_valid);
5978 SWAP(old_dirty_submodule, new_dirty_submodule);
5979 }
5980
5981 if (options->prefix &&
5982 strncmp(concatpath, options->prefix, options->prefix_length))
5983 return;
5984
5985 one = alloc_filespec(concatpath);
5986 two = alloc_filespec(concatpath);
5987 fill_filespec(one, old_oid, old_oid_valid, old_mode);
5988 fill_filespec(two, new_oid, new_oid_valid, new_mode);
5989 one->dirty_submodule = old_dirty_submodule;
5990 two->dirty_submodule = new_dirty_submodule;
5991 p = diff_queue(&diff_queued_diff, one, two);
5992
5993 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5994 return;
5995
5996 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5997 !diff_filespec_check_stat_unmatch(p))
5998 return;
5999
6000 DIFF_OPT_SET(options, HAS_CHANGES);
6001 }
6002
6003 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6004 {
6005 struct diff_filepair *pair;
6006 struct diff_filespec *one, *two;
6007
6008 if (options->prefix &&
6009 strncmp(path, options->prefix, options->prefix_length))
6010 return NULL;
6011
6012 one = alloc_filespec(path);
6013 two = alloc_filespec(path);
6014 pair = diff_queue(&diff_queued_diff, one, two);
6015 pair->is_unmerged = 1;
6016 return pair;
6017 }
6018
6019 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
6020 size_t *outsize)
6021 {
6022 struct diff_tempfile *temp;
6023 const char *argv[3];
6024 const char **arg = argv;
6025 struct child_process child = CHILD_PROCESS_INIT;
6026 struct strbuf buf = STRBUF_INIT;
6027 int err = 0;
6028
6029 temp = prepare_temp_file(spec->path, spec);
6030 *arg++ = pgm;
6031 *arg++ = temp->name;
6032 *arg = NULL;
6033
6034 child.use_shell = 1;
6035 child.argv = argv;
6036 child.out = -1;
6037 if (start_command(&child)) {
6038 remove_tempfile();
6039 return NULL;
6040 }
6041
6042 if (strbuf_read(&buf, child.out, 0) < 0)
6043 err = error("error reading from textconv command '%s'", pgm);
6044 close(child.out);
6045
6046 if (finish_command(&child) || err) {
6047 strbuf_release(&buf);
6048 remove_tempfile();
6049 return NULL;
6050 }
6051 remove_tempfile();
6052
6053 return strbuf_detach(&buf, outsize);
6054 }
6055
6056 size_t fill_textconv(struct userdiff_driver *driver,
6057 struct diff_filespec *df,
6058 char **outbuf)
6059 {
6060 size_t size;
6061
6062 if (!driver) {
6063 if (!DIFF_FILE_VALID(df)) {
6064 *outbuf = "";
6065 return 0;
6066 }
6067 if (diff_populate_filespec(df, 0))
6068 die("unable to read files to diff");
6069 *outbuf = df->data;
6070 return df->size;
6071 }
6072
6073 if (!driver->textconv)
6074 die("BUG: fill_textconv called with non-textconv driver");
6075
6076 if (driver->textconv_cache && df->oid_valid) {
6077 *outbuf = notes_cache_get(driver->textconv_cache,
6078 &df->oid,
6079 &size);
6080 if (*outbuf)
6081 return size;
6082 }
6083
6084 *outbuf = run_textconv(driver->textconv, df, &size);
6085 if (!*outbuf)
6086 die("unable to read files to diff");
6087
6088 if (driver->textconv_cache && df->oid_valid) {
6089 /* ignore errors, as we might be in a readonly repository */
6090 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6091 size);
6092 /*
6093 * we could save up changes and flush them all at the end,
6094 * but we would need an extra call after all diffing is done.
6095 * Since generating a cache entry is the slow path anyway,
6096 * this extra overhead probably isn't a big deal.
6097 */
6098 notes_cache_write(driver->textconv_cache);
6099 }
6100
6101 return size;
6102 }
6103
6104 int textconv_object(const char *path,
6105 unsigned mode,
6106 const struct object_id *oid,
6107 int oid_valid,
6108 char **buf,
6109 unsigned long *buf_size)
6110 {
6111 struct diff_filespec *df;
6112 struct userdiff_driver *textconv;
6113
6114 df = alloc_filespec(path);
6115 fill_filespec(df, oid, oid_valid, mode);
6116 textconv = get_textconv(df);
6117 if (!textconv) {
6118 free_filespec(df);
6119 return 0;
6120 }
6121
6122 *buf_size = fill_textconv(textconv, df, buf);
6123 free_filespec(df);
6124 return 1;
6125 }
6126
6127 void setup_diff_pager(struct diff_options *opt)
6128 {
6129 /*
6130 * If the user asked for our exit code, then either they want --quiet
6131 * or --exit-code. We should definitely not bother with a pager in the
6132 * former case, as we will generate no output. Since we still properly
6133 * report our exit code even when a pager is run, we _could_ run a
6134 * pager with --exit-code. But since we have not done so historically,
6135 * and because it is easy to find people oneline advising "git diff
6136 * --exit-code" in hooks and other scripts, we do not do so.
6137 */
6138 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
6139 check_pager_config("diff") != 0)
6140 setup_pager();
6141 }
The Best VCS Around