pathspec: create parse_long_magic function
[alt-git.git] / diff.c
blobec8728362dae5a3bf404f6782f9091d6a1c9a6c2
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "tempfile.h"
6 #include "quote.h"
7 #include "diff.h"
8 #include "diffcore.h"
9 #include "delta.h"
10 #include "xdiff-interface.h"
11 #include "color.h"
12 #include "attr.h"
13 #include "run-command.h"
14 #include "utf8.h"
15 #include "userdiff.h"
16 #include "submodule-config.h"
17 #include "submodule.h"
18 #include "ll-merge.h"
19 #include "string-list.h"
20 #include "argv-array.h"
21 #include "graph.h"
23 #ifdef NO_FAST_WORKING_DIRECTORY
24 #define FAST_WORKING_DIRECTORY 0
25 #else
26 #define FAST_WORKING_DIRECTORY 1
27 #endif
29 static int diff_detect_rename_default;
30 static int diff_indent_heuristic; /* experimental */
31 static int diff_compaction_heuristic; /* experimental */
32 static int diff_rename_limit_default = 400;
33 static int diff_suppress_blank_empty;
34 static int diff_use_color_default = -1;
35 static int diff_context_default = 3;
36 static const char *diff_word_regex_cfg;
37 static const char *external_diff_cmd_cfg;
38 static const char *diff_order_file_cfg;
39 int diff_auto_refresh_index = 1;
40 static int diff_mnemonic_prefix;
41 static int diff_no_prefix;
42 static int diff_stat_graph_width;
43 static int diff_dirstat_permille_default = 30;
44 static struct diff_options default_diff_options;
45 static long diff_algorithm;
46 static unsigned ws_error_highlight_default = WSEH_NEW;
48 static char diff_colors[][COLOR_MAXLEN] = {
49 GIT_COLOR_RESET,
50 GIT_COLOR_NORMAL, /* CONTEXT */
51 GIT_COLOR_BOLD, /* METAINFO */
52 GIT_COLOR_CYAN, /* FRAGINFO */
53 GIT_COLOR_RED, /* OLD */
54 GIT_COLOR_GREEN, /* NEW */
55 GIT_COLOR_YELLOW, /* COMMIT */
56 GIT_COLOR_BG_RED, /* WHITESPACE */
57 GIT_COLOR_NORMAL, /* FUNCINFO */
60 static NORETURN void die_want_option(const char *option_name)
62 die(_("option '%s' requires a value"), option_name);
65 static int parse_diff_color_slot(const char *var)
67 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
68 return DIFF_CONTEXT;
69 if (!strcasecmp(var, "meta"))
70 return DIFF_METAINFO;
71 if (!strcasecmp(var, "frag"))
72 return DIFF_FRAGINFO;
73 if (!strcasecmp(var, "old"))
74 return DIFF_FILE_OLD;
75 if (!strcasecmp(var, "new"))
76 return DIFF_FILE_NEW;
77 if (!strcasecmp(var, "commit"))
78 return DIFF_COMMIT;
79 if (!strcasecmp(var, "whitespace"))
80 return DIFF_WHITESPACE;
81 if (!strcasecmp(var, "func"))
82 return DIFF_FUNCINFO;
83 return -1;
86 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
87 struct strbuf *errmsg)
89 char *params_copy = xstrdup(params_string);
90 struct string_list params = STRING_LIST_INIT_NODUP;
91 int ret = 0;
92 int i;
94 if (*params_copy)
95 string_list_split_in_place(&params, params_copy, ',', -1);
96 for (i = 0; i < params.nr; i++) {
97 const char *p = params.items[i].string;
98 if (!strcmp(p, "changes")) {
99 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
100 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
101 } else if (!strcmp(p, "lines")) {
102 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
103 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
104 } else if (!strcmp(p, "files")) {
105 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
106 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
107 } else if (!strcmp(p, "noncumulative")) {
108 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
109 } else if (!strcmp(p, "cumulative")) {
110 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
111 } else if (isdigit(*p)) {
112 char *end;
113 int permille = strtoul(p, &end, 10) * 10;
114 if (*end == '.' && isdigit(*++end)) {
115 /* only use first digit */
116 permille += *end - '0';
117 /* .. and ignore any further digits */
118 while (isdigit(*++end))
119 ; /* nothing */
121 if (!*end)
122 options->dirstat_permille = permille;
123 else {
124 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
126 ret++;
128 } else {
129 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
130 ret++;
134 string_list_clear(&params, 0);
135 free(params_copy);
136 return ret;
139 static int parse_submodule_params(struct diff_options *options, const char *value)
141 if (!strcmp(value, "log"))
142 options->submodule_format = DIFF_SUBMODULE_LOG;
143 else if (!strcmp(value, "short"))
144 options->submodule_format = DIFF_SUBMODULE_SHORT;
145 else if (!strcmp(value, "diff"))
146 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
147 else
148 return -1;
149 return 0;
152 static int git_config_rename(const char *var, const char *value)
154 if (!value)
155 return DIFF_DETECT_RENAME;
156 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
157 return DIFF_DETECT_COPY;
158 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
161 long parse_algorithm_value(const char *value)
163 if (!value)
164 return -1;
165 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
166 return 0;
167 else if (!strcasecmp(value, "minimal"))
168 return XDF_NEED_MINIMAL;
169 else if (!strcasecmp(value, "patience"))
170 return XDF_PATIENCE_DIFF;
171 else if (!strcasecmp(value, "histogram"))
172 return XDF_HISTOGRAM_DIFF;
173 return -1;
176 static int parse_one_token(const char **arg, const char *token)
178 const char *rest;
179 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
180 *arg = rest;
181 return 1;
183 return 0;
186 static int parse_ws_error_highlight(const char *arg)
188 const char *orig_arg = arg;
189 unsigned val = 0;
191 while (*arg) {
192 if (parse_one_token(&arg, "none"))
193 val = 0;
194 else if (parse_one_token(&arg, "default"))
195 val = WSEH_NEW;
196 else if (parse_one_token(&arg, "all"))
197 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
198 else if (parse_one_token(&arg, "new"))
199 val |= WSEH_NEW;
200 else if (parse_one_token(&arg, "old"))
201 val |= WSEH_OLD;
202 else if (parse_one_token(&arg, "context"))
203 val |= WSEH_CONTEXT;
204 else {
205 return -1 - (int)(arg - orig_arg);
207 if (*arg)
208 arg++;
210 return val;
214 * These are to give UI layer defaults.
215 * The core-level commands such as git-diff-files should
216 * never be affected by the setting of diff.renames
217 * the user happens to have in the configuration file.
219 void init_diff_ui_defaults(void)
221 diff_detect_rename_default = 1;
224 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
226 if (!strcmp(var, "diff.indentheuristic")) {
227 diff_indent_heuristic = git_config_bool(var, value);
228 if (diff_indent_heuristic)
229 diff_compaction_heuristic = 0;
231 if (!strcmp(var, "diff.compactionheuristic")) {
232 diff_compaction_heuristic = git_config_bool(var, value);
233 if (diff_compaction_heuristic)
234 diff_indent_heuristic = 0;
236 return 0;
239 int git_diff_ui_config(const char *var, const char *value, void *cb)
241 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
242 diff_use_color_default = git_config_colorbool(var, value);
243 return 0;
245 if (!strcmp(var, "diff.context")) {
246 diff_context_default = git_config_int(var, value);
247 if (diff_context_default < 0)
248 return -1;
249 return 0;
251 if (!strcmp(var, "diff.renames")) {
252 diff_detect_rename_default = git_config_rename(var, value);
253 return 0;
255 if (!strcmp(var, "diff.autorefreshindex")) {
256 diff_auto_refresh_index = git_config_bool(var, value);
257 return 0;
259 if (!strcmp(var, "diff.mnemonicprefix")) {
260 diff_mnemonic_prefix = git_config_bool(var, value);
261 return 0;
263 if (!strcmp(var, "diff.noprefix")) {
264 diff_no_prefix = git_config_bool(var, value);
265 return 0;
267 if (!strcmp(var, "diff.statgraphwidth")) {
268 diff_stat_graph_width = git_config_int(var, value);
269 return 0;
271 if (!strcmp(var, "diff.external"))
272 return git_config_string(&external_diff_cmd_cfg, var, value);
273 if (!strcmp(var, "diff.wordregex"))
274 return git_config_string(&diff_word_regex_cfg, var, value);
275 if (!strcmp(var, "diff.orderfile"))
276 return git_config_pathname(&diff_order_file_cfg, var, value);
278 if (!strcmp(var, "diff.ignoresubmodules"))
279 handle_ignore_submodules_arg(&default_diff_options, value);
281 if (!strcmp(var, "diff.submodule")) {
282 if (parse_submodule_params(&default_diff_options, value))
283 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
284 value);
285 return 0;
288 if (!strcmp(var, "diff.algorithm")) {
289 diff_algorithm = parse_algorithm_value(value);
290 if (diff_algorithm < 0)
291 return -1;
292 return 0;
295 if (git_diff_heuristic_config(var, value, cb) < 0)
296 return -1;
298 if (!strcmp(var, "diff.wserrorhighlight")) {
299 int val = parse_ws_error_highlight(value);
300 if (val < 0)
301 return -1;
302 ws_error_highlight_default = val;
303 return 0;
306 if (git_color_config(var, value, cb) < 0)
307 return -1;
309 return git_diff_basic_config(var, value, cb);
312 int git_diff_basic_config(const char *var, const char *value, void *cb)
314 const char *name;
316 if (!strcmp(var, "diff.renamelimit")) {
317 diff_rename_limit_default = git_config_int(var, value);
318 return 0;
321 if (userdiff_config(var, value) < 0)
322 return -1;
324 if (skip_prefix(var, "diff.color.", &name) ||
325 skip_prefix(var, "color.diff.", &name)) {
326 int slot = parse_diff_color_slot(name);
327 if (slot < 0)
328 return 0;
329 if (!value)
330 return config_error_nonbool(var);
331 return color_parse(value, diff_colors[slot]);
334 /* like GNU diff's --suppress-blank-empty option */
335 if (!strcmp(var, "diff.suppressblankempty") ||
336 /* for backwards compatibility */
337 !strcmp(var, "diff.suppress-blank-empty")) {
338 diff_suppress_blank_empty = git_config_bool(var, value);
339 return 0;
342 if (!strcmp(var, "diff.dirstat")) {
343 struct strbuf errmsg = STRBUF_INIT;
344 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
345 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
346 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
347 errmsg.buf);
348 strbuf_release(&errmsg);
349 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
350 return 0;
353 if (starts_with(var, "submodule."))
354 return parse_submodule_config_option(var, value);
356 return git_default_config(var, value, cb);
359 static char *quote_two(const char *one, const char *two)
361 int need_one = quote_c_style(one, NULL, NULL, 1);
362 int need_two = quote_c_style(two, NULL, NULL, 1);
363 struct strbuf res = STRBUF_INIT;
365 if (need_one + need_two) {
366 strbuf_addch(&res, '"');
367 quote_c_style(one, &res, NULL, 1);
368 quote_c_style(two, &res, NULL, 1);
369 strbuf_addch(&res, '"');
370 } else {
371 strbuf_addstr(&res, one);
372 strbuf_addstr(&res, two);
374 return strbuf_detach(&res, NULL);
377 static const char *external_diff(void)
379 static const char *external_diff_cmd = NULL;
380 static int done_preparing = 0;
382 if (done_preparing)
383 return external_diff_cmd;
384 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
385 if (!external_diff_cmd)
386 external_diff_cmd = external_diff_cmd_cfg;
387 done_preparing = 1;
388 return external_diff_cmd;
392 * Keep track of files used for diffing. Sometimes such an entry
393 * refers to a temporary file, sometimes to an existing file, and
394 * sometimes to "/dev/null".
396 static struct diff_tempfile {
398 * filename external diff should read from, or NULL if this
399 * entry is currently not in use:
401 const char *name;
403 char hex[GIT_SHA1_HEXSZ + 1];
404 char mode[10];
407 * If this diff_tempfile instance refers to a temporary file,
408 * this tempfile object is used to manage its lifetime.
410 struct tempfile tempfile;
411 } diff_temp[2];
413 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
415 struct emit_callback {
416 int color_diff;
417 unsigned ws_rule;
418 int blank_at_eof_in_preimage;
419 int blank_at_eof_in_postimage;
420 int lno_in_preimage;
421 int lno_in_postimage;
422 sane_truncate_fn truncate;
423 const char **label_path;
424 struct diff_words_data *diff_words;
425 struct diff_options *opt;
426 struct strbuf *header;
429 static int count_lines(const char *data, int size)
431 int count, ch, completely_empty = 1, nl_just_seen = 0;
432 count = 0;
433 while (0 < size--) {
434 ch = *data++;
435 if (ch == '\n') {
436 count++;
437 nl_just_seen = 1;
438 completely_empty = 0;
440 else {
441 nl_just_seen = 0;
442 completely_empty = 0;
445 if (completely_empty)
446 return 0;
447 if (!nl_just_seen)
448 count++; /* no trailing newline */
449 return count;
452 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
454 if (!DIFF_FILE_VALID(one)) {
455 mf->ptr = (char *)""; /* does not matter */
456 mf->size = 0;
457 return 0;
459 else if (diff_populate_filespec(one, 0))
460 return -1;
462 mf->ptr = one->data;
463 mf->size = one->size;
464 return 0;
467 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
468 static unsigned long diff_filespec_size(struct diff_filespec *one)
470 if (!DIFF_FILE_VALID(one))
471 return 0;
472 diff_populate_filespec(one, CHECK_SIZE_ONLY);
473 return one->size;
476 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
478 char *ptr = mf->ptr;
479 long size = mf->size;
480 int cnt = 0;
482 if (!size)
483 return cnt;
484 ptr += size - 1; /* pointing at the very end */
485 if (*ptr != '\n')
486 ; /* incomplete line */
487 else
488 ptr--; /* skip the last LF */
489 while (mf->ptr < ptr) {
490 char *prev_eol;
491 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
492 if (*prev_eol == '\n')
493 break;
494 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
495 break;
496 cnt++;
497 ptr = prev_eol - 1;
499 return cnt;
502 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
503 struct emit_callback *ecbdata)
505 int l1, l2, at;
506 unsigned ws_rule = ecbdata->ws_rule;
507 l1 = count_trailing_blank(mf1, ws_rule);
508 l2 = count_trailing_blank(mf2, ws_rule);
509 if (l2 <= l1) {
510 ecbdata->blank_at_eof_in_preimage = 0;
511 ecbdata->blank_at_eof_in_postimage = 0;
512 return;
514 at = count_lines(mf1->ptr, mf1->size);
515 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
517 at = count_lines(mf2->ptr, mf2->size);
518 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
521 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
522 int first, const char *line, int len)
524 int has_trailing_newline, has_trailing_carriage_return;
525 int nofirst;
526 FILE *file = o->file;
528 fputs(diff_line_prefix(o), file);
530 if (len == 0) {
531 has_trailing_newline = (first == '\n');
532 has_trailing_carriage_return = (!has_trailing_newline &&
533 (first == '\r'));
534 nofirst = has_trailing_newline || has_trailing_carriage_return;
535 } else {
536 has_trailing_newline = (len > 0 && line[len-1] == '\n');
537 if (has_trailing_newline)
538 len--;
539 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
540 if (has_trailing_carriage_return)
541 len--;
542 nofirst = 0;
545 if (len || !nofirst) {
546 fputs(set, file);
547 if (!nofirst)
548 fputc(first, file);
549 fwrite(line, len, 1, file);
550 fputs(reset, file);
552 if (has_trailing_carriage_return)
553 fputc('\r', file);
554 if (has_trailing_newline)
555 fputc('\n', file);
558 static void emit_line(struct diff_options *o, const char *set, const char *reset,
559 const char *line, int len)
561 emit_line_0(o, set, reset, line[0], line+1, len-1);
564 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
566 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
567 ecbdata->blank_at_eof_in_preimage &&
568 ecbdata->blank_at_eof_in_postimage &&
569 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
570 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
571 return 0;
572 return ws_blank_line(line, len, ecbdata->ws_rule);
575 static void emit_line_checked(const char *reset,
576 struct emit_callback *ecbdata,
577 const char *line, int len,
578 enum color_diff color,
579 unsigned ws_error_highlight,
580 char sign)
582 const char *set = diff_get_color(ecbdata->color_diff, color);
583 const char *ws = NULL;
585 if (ecbdata->opt->ws_error_highlight & ws_error_highlight) {
586 ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
587 if (!*ws)
588 ws = NULL;
591 if (!ws)
592 emit_line_0(ecbdata->opt, set, reset, sign, line, len);
593 else if (sign == '+' && new_blank_line_at_eof(ecbdata, line, len))
594 /* Blank line at EOF - paint '+' as well */
595 emit_line_0(ecbdata->opt, ws, reset, sign, line, len);
596 else {
597 /* Emit just the prefix, then the rest. */
598 emit_line_0(ecbdata->opt, set, reset, sign, "", 0);
599 ws_check_emit(line, len, ecbdata->ws_rule,
600 ecbdata->opt->file, set, reset, ws);
604 static void emit_add_line(const char *reset,
605 struct emit_callback *ecbdata,
606 const char *line, int len)
608 emit_line_checked(reset, ecbdata, line, len,
609 DIFF_FILE_NEW, WSEH_NEW, '+');
612 static void emit_del_line(const char *reset,
613 struct emit_callback *ecbdata,
614 const char *line, int len)
616 emit_line_checked(reset, ecbdata, line, len,
617 DIFF_FILE_OLD, WSEH_OLD, '-');
620 static void emit_context_line(const char *reset,
621 struct emit_callback *ecbdata,
622 const char *line, int len)
624 emit_line_checked(reset, ecbdata, line, len,
625 DIFF_CONTEXT, WSEH_CONTEXT, ' ');
628 static void emit_hunk_header(struct emit_callback *ecbdata,
629 const char *line, int len)
631 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
632 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
633 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
634 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
635 static const char atat[2] = { '@', '@' };
636 const char *cp, *ep;
637 struct strbuf msgbuf = STRBUF_INIT;
638 int org_len = len;
639 int i = 1;
642 * As a hunk header must begin with "@@ -<old>, +<new> @@",
643 * it always is at least 10 bytes long.
645 if (len < 10 ||
646 memcmp(line, atat, 2) ||
647 !(ep = memmem(line + 2, len - 2, atat, 2))) {
648 emit_line(ecbdata->opt, context, reset, line, len);
649 return;
651 ep += 2; /* skip over @@ */
653 /* The hunk header in fraginfo color */
654 strbuf_addstr(&msgbuf, frag);
655 strbuf_add(&msgbuf, line, ep - line);
656 strbuf_addstr(&msgbuf, reset);
659 * trailing "\r\n"
661 for ( ; i < 3; i++)
662 if (line[len - i] == '\r' || line[len - i] == '\n')
663 len--;
665 /* blank before the func header */
666 for (cp = ep; ep - line < len; ep++)
667 if (*ep != ' ' && *ep != '\t')
668 break;
669 if (ep != cp) {
670 strbuf_addstr(&msgbuf, context);
671 strbuf_add(&msgbuf, cp, ep - cp);
672 strbuf_addstr(&msgbuf, reset);
675 if (ep < line + len) {
676 strbuf_addstr(&msgbuf, func);
677 strbuf_add(&msgbuf, ep, line + len - ep);
678 strbuf_addstr(&msgbuf, reset);
681 strbuf_add(&msgbuf, line + len, org_len - len);
682 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
683 strbuf_release(&msgbuf);
686 static struct diff_tempfile *claim_diff_tempfile(void) {
687 int i;
688 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
689 if (!diff_temp[i].name)
690 return diff_temp + i;
691 die("BUG: diff is failing to clean up its tempfiles");
694 static void remove_tempfile(void)
696 int i;
697 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
698 if (is_tempfile_active(&diff_temp[i].tempfile))
699 delete_tempfile(&diff_temp[i].tempfile);
700 diff_temp[i].name = NULL;
704 static void print_line_count(FILE *file, int count)
706 switch (count) {
707 case 0:
708 fprintf(file, "0,0");
709 break;
710 case 1:
711 fprintf(file, "1");
712 break;
713 default:
714 fprintf(file, "1,%d", count);
715 break;
719 static void emit_rewrite_lines(struct emit_callback *ecb,
720 int prefix, const char *data, int size)
722 const char *endp = NULL;
723 static const char *nneof = " No newline at end of file\n";
724 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
726 while (0 < size) {
727 int len;
729 endp = memchr(data, '\n', size);
730 len = endp ? (endp - data + 1) : size;
731 if (prefix != '+') {
732 ecb->lno_in_preimage++;
733 emit_del_line(reset, ecb, data, len);
734 } else {
735 ecb->lno_in_postimage++;
736 emit_add_line(reset, ecb, data, len);
738 size -= len;
739 data += len;
741 if (!endp) {
742 const char *context = diff_get_color(ecb->color_diff,
743 DIFF_CONTEXT);
744 putc('\n', ecb->opt->file);
745 emit_line_0(ecb->opt, context, reset, '\\',
746 nneof, strlen(nneof));
750 static void emit_rewrite_diff(const char *name_a,
751 const char *name_b,
752 struct diff_filespec *one,
753 struct diff_filespec *two,
754 struct userdiff_driver *textconv_one,
755 struct userdiff_driver *textconv_two,
756 struct diff_options *o)
758 int lc_a, lc_b;
759 const char *name_a_tab, *name_b_tab;
760 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
761 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
762 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
763 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
764 const char *a_prefix, *b_prefix;
765 char *data_one, *data_two;
766 size_t size_one, size_two;
767 struct emit_callback ecbdata;
768 const char *line_prefix = diff_line_prefix(o);
770 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
771 a_prefix = o->b_prefix;
772 b_prefix = o->a_prefix;
773 } else {
774 a_prefix = o->a_prefix;
775 b_prefix = o->b_prefix;
778 name_a += (*name_a == '/');
779 name_b += (*name_b == '/');
780 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
781 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
783 strbuf_reset(&a_name);
784 strbuf_reset(&b_name);
785 quote_two_c_style(&a_name, a_prefix, name_a, 0);
786 quote_two_c_style(&b_name, b_prefix, name_b, 0);
788 size_one = fill_textconv(textconv_one, one, &data_one);
789 size_two = fill_textconv(textconv_two, two, &data_two);
791 memset(&ecbdata, 0, sizeof(ecbdata));
792 ecbdata.color_diff = want_color(o->use_color);
793 ecbdata.ws_rule = whitespace_rule(name_b);
794 ecbdata.opt = o;
795 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
796 mmfile_t mf1, mf2;
797 mf1.ptr = (char *)data_one;
798 mf2.ptr = (char *)data_two;
799 mf1.size = size_one;
800 mf2.size = size_two;
801 check_blank_at_eof(&mf1, &mf2, &ecbdata);
803 ecbdata.lno_in_preimage = 1;
804 ecbdata.lno_in_postimage = 1;
806 lc_a = count_lines(data_one, size_one);
807 lc_b = count_lines(data_two, size_two);
808 fprintf(o->file,
809 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
810 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
811 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
812 line_prefix, fraginfo);
813 if (!o->irreversible_delete)
814 print_line_count(o->file, lc_a);
815 else
816 fprintf(o->file, "?,?");
817 fprintf(o->file, " +");
818 print_line_count(o->file, lc_b);
819 fprintf(o->file, " @@%s\n", reset);
820 if (lc_a && !o->irreversible_delete)
821 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
822 if (lc_b)
823 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
824 if (textconv_one)
825 free((char *)data_one);
826 if (textconv_two)
827 free((char *)data_two);
830 struct diff_words_buffer {
831 mmfile_t text;
832 long alloc;
833 struct diff_words_orig {
834 const char *begin, *end;
835 } *orig;
836 int orig_nr, orig_alloc;
839 static void diff_words_append(char *line, unsigned long len,
840 struct diff_words_buffer *buffer)
842 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
843 line++;
844 len--;
845 memcpy(buffer->text.ptr + buffer->text.size, line, len);
846 buffer->text.size += len;
847 buffer->text.ptr[buffer->text.size] = '\0';
850 struct diff_words_style_elem {
851 const char *prefix;
852 const char *suffix;
853 const char *color; /* NULL; filled in by the setup code if
854 * color is enabled */
857 struct diff_words_style {
858 enum diff_words_type type;
859 struct diff_words_style_elem new, old, ctx;
860 const char *newline;
863 static struct diff_words_style diff_words_styles[] = {
864 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
865 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
866 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
869 struct diff_words_data {
870 struct diff_words_buffer minus, plus;
871 const char *current_plus;
872 int last_minus;
873 struct diff_options *opt;
874 regex_t *word_regex;
875 enum diff_words_type type;
876 struct diff_words_style *style;
879 static int fn_out_diff_words_write_helper(FILE *fp,
880 struct diff_words_style_elem *st_el,
881 const char *newline,
882 size_t count, const char *buf,
883 const char *line_prefix)
885 int print = 0;
887 while (count) {
888 char *p = memchr(buf, '\n', count);
889 if (print)
890 fputs(line_prefix, fp);
891 if (p != buf) {
892 if (st_el->color && fputs(st_el->color, fp) < 0)
893 return -1;
894 if (fputs(st_el->prefix, fp) < 0 ||
895 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
896 fputs(st_el->suffix, fp) < 0)
897 return -1;
898 if (st_el->color && *st_el->color
899 && fputs(GIT_COLOR_RESET, fp) < 0)
900 return -1;
902 if (!p)
903 return 0;
904 if (fputs(newline, fp) < 0)
905 return -1;
906 count -= p + 1 - buf;
907 buf = p + 1;
908 print = 1;
910 return 0;
914 * '--color-words' algorithm can be described as:
916 * 1. collect a the minus/plus lines of a diff hunk, divided into
917 * minus-lines and plus-lines;
919 * 2. break both minus-lines and plus-lines into words and
920 * place them into two mmfile_t with one word for each line;
922 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
924 * And for the common parts of the both file, we output the plus side text.
925 * diff_words->current_plus is used to trace the current position of the plus file
926 * which printed. diff_words->last_minus is used to trace the last minus word
927 * printed.
929 * For '--graph' to work with '--color-words', we need to output the graph prefix
930 * on each line of color words output. Generally, there are two conditions on
931 * which we should output the prefix.
933 * 1. diff_words->last_minus == 0 &&
934 * diff_words->current_plus == diff_words->plus.text.ptr
936 * that is: the plus text must start as a new line, and if there is no minus
937 * word printed, a graph prefix must be printed.
939 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
940 * *(diff_words->current_plus - 1) == '\n'
942 * that is: a graph prefix must be printed following a '\n'
944 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
946 if ((diff_words->last_minus == 0 &&
947 diff_words->current_plus == diff_words->plus.text.ptr) ||
948 (diff_words->current_plus > diff_words->plus.text.ptr &&
949 *(diff_words->current_plus - 1) == '\n')) {
950 return 1;
951 } else {
952 return 0;
956 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
958 struct diff_words_data *diff_words = priv;
959 struct diff_words_style *style = diff_words->style;
960 int minus_first, minus_len, plus_first, plus_len;
961 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
962 struct diff_options *opt = diff_words->opt;
963 const char *line_prefix;
965 if (line[0] != '@' || parse_hunk_header(line, len,
966 &minus_first, &minus_len, &plus_first, &plus_len))
967 return;
969 assert(opt);
970 line_prefix = diff_line_prefix(opt);
972 /* POSIX requires that first be decremented by one if len == 0... */
973 if (minus_len) {
974 minus_begin = diff_words->minus.orig[minus_first].begin;
975 minus_end =
976 diff_words->minus.orig[minus_first + minus_len - 1].end;
977 } else
978 minus_begin = minus_end =
979 diff_words->minus.orig[minus_first].end;
981 if (plus_len) {
982 plus_begin = diff_words->plus.orig[plus_first].begin;
983 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
984 } else
985 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
987 if (color_words_output_graph_prefix(diff_words)) {
988 fputs(line_prefix, diff_words->opt->file);
990 if (diff_words->current_plus != plus_begin) {
991 fn_out_diff_words_write_helper(diff_words->opt->file,
992 &style->ctx, style->newline,
993 plus_begin - diff_words->current_plus,
994 diff_words->current_plus, line_prefix);
995 if (*(plus_begin - 1) == '\n')
996 fputs(line_prefix, diff_words->opt->file);
998 if (minus_begin != minus_end) {
999 fn_out_diff_words_write_helper(diff_words->opt->file,
1000 &style->old, style->newline,
1001 minus_end - minus_begin, minus_begin,
1002 line_prefix);
1004 if (plus_begin != plus_end) {
1005 fn_out_diff_words_write_helper(diff_words->opt->file,
1006 &style->new, style->newline,
1007 plus_end - plus_begin, plus_begin,
1008 line_prefix);
1011 diff_words->current_plus = plus_end;
1012 diff_words->last_minus = minus_first;
1015 /* This function starts looking at *begin, and returns 0 iff a word was found. */
1016 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1017 int *begin, int *end)
1019 if (word_regex && *begin < buffer->size) {
1020 regmatch_t match[1];
1021 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1022 buffer->size - *begin, 1, match, 0)) {
1023 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1024 '\n', match[0].rm_eo - match[0].rm_so);
1025 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1026 *begin += match[0].rm_so;
1027 return *begin >= *end;
1029 return -1;
1032 /* find the next word */
1033 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1034 (*begin)++;
1035 if (*begin >= buffer->size)
1036 return -1;
1038 /* find the end of the word */
1039 *end = *begin + 1;
1040 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1041 (*end)++;
1043 return 0;
1047 * This function splits the words in buffer->text, stores the list with
1048 * newline separator into out, and saves the offsets of the original words
1049 * in buffer->orig.
1051 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1052 regex_t *word_regex)
1054 int i, j;
1055 long alloc = 0;
1057 out->size = 0;
1058 out->ptr = NULL;
1060 /* fake an empty "0th" word */
1061 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1062 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1063 buffer->orig_nr = 1;
1065 for (i = 0; i < buffer->text.size; i++) {
1066 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1067 return;
1069 /* store original boundaries */
1070 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1071 buffer->orig_alloc);
1072 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1073 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1074 buffer->orig_nr++;
1076 /* store one word */
1077 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1078 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1079 out->ptr[out->size + j - i] = '\n';
1080 out->size += j - i + 1;
1082 i = j - 1;
1086 /* this executes the word diff on the accumulated buffers */
1087 static void diff_words_show(struct diff_words_data *diff_words)
1089 xpparam_t xpp;
1090 xdemitconf_t xecfg;
1091 mmfile_t minus, plus;
1092 struct diff_words_style *style = diff_words->style;
1094 struct diff_options *opt = diff_words->opt;
1095 const char *line_prefix;
1097 assert(opt);
1098 line_prefix = diff_line_prefix(opt);
1100 /* special case: only removal */
1101 if (!diff_words->plus.text.size) {
1102 fputs(line_prefix, diff_words->opt->file);
1103 fn_out_diff_words_write_helper(diff_words->opt->file,
1104 &style->old, style->newline,
1105 diff_words->minus.text.size,
1106 diff_words->minus.text.ptr, line_prefix);
1107 diff_words->minus.text.size = 0;
1108 return;
1111 diff_words->current_plus = diff_words->plus.text.ptr;
1112 diff_words->last_minus = 0;
1114 memset(&xpp, 0, sizeof(xpp));
1115 memset(&xecfg, 0, sizeof(xecfg));
1116 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1117 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1118 xpp.flags = 0;
1119 /* as only the hunk header will be parsed, we need a 0-context */
1120 xecfg.ctxlen = 0;
1121 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1122 &xpp, &xecfg))
1123 die("unable to generate word diff");
1124 free(minus.ptr);
1125 free(plus.ptr);
1126 if (diff_words->current_plus != diff_words->plus.text.ptr +
1127 diff_words->plus.text.size) {
1128 if (color_words_output_graph_prefix(diff_words))
1129 fputs(line_prefix, diff_words->opt->file);
1130 fn_out_diff_words_write_helper(diff_words->opt->file,
1131 &style->ctx, style->newline,
1132 diff_words->plus.text.ptr + diff_words->plus.text.size
1133 - diff_words->current_plus, diff_words->current_plus,
1134 line_prefix);
1136 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1139 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
1140 static void diff_words_flush(struct emit_callback *ecbdata)
1142 if (ecbdata->diff_words->minus.text.size ||
1143 ecbdata->diff_words->plus.text.size)
1144 diff_words_show(ecbdata->diff_words);
1147 static void diff_filespec_load_driver(struct diff_filespec *one)
1149 /* Use already-loaded driver */
1150 if (one->driver)
1151 return;
1153 if (S_ISREG(one->mode))
1154 one->driver = userdiff_find_by_path(one->path);
1156 /* Fallback to default settings */
1157 if (!one->driver)
1158 one->driver = userdiff_find_by_name("default");
1161 static const char *userdiff_word_regex(struct diff_filespec *one)
1163 diff_filespec_load_driver(one);
1164 return one->driver->word_regex;
1167 static void init_diff_words_data(struct emit_callback *ecbdata,
1168 struct diff_options *orig_opts,
1169 struct diff_filespec *one,
1170 struct diff_filespec *two)
1172 int i;
1173 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1174 memcpy(o, orig_opts, sizeof(struct diff_options));
1176 ecbdata->diff_words =
1177 xcalloc(1, sizeof(struct diff_words_data));
1178 ecbdata->diff_words->type = o->word_diff;
1179 ecbdata->diff_words->opt = o;
1180 if (!o->word_regex)
1181 o->word_regex = userdiff_word_regex(one);
1182 if (!o->word_regex)
1183 o->word_regex = userdiff_word_regex(two);
1184 if (!o->word_regex)
1185 o->word_regex = diff_word_regex_cfg;
1186 if (o->word_regex) {
1187 ecbdata->diff_words->word_regex = (regex_t *)
1188 xmalloc(sizeof(regex_t));
1189 if (regcomp(ecbdata->diff_words->word_regex,
1190 o->word_regex,
1191 REG_EXTENDED | REG_NEWLINE))
1192 die ("Invalid regular expression: %s",
1193 o->word_regex);
1195 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1196 if (o->word_diff == diff_words_styles[i].type) {
1197 ecbdata->diff_words->style =
1198 &diff_words_styles[i];
1199 break;
1202 if (want_color(o->use_color)) {
1203 struct diff_words_style *st = ecbdata->diff_words->style;
1204 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1205 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1206 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1210 static void free_diff_words_data(struct emit_callback *ecbdata)
1212 if (ecbdata->diff_words) {
1213 diff_words_flush(ecbdata);
1214 free (ecbdata->diff_words->opt);
1215 free (ecbdata->diff_words->minus.text.ptr);
1216 free (ecbdata->diff_words->minus.orig);
1217 free (ecbdata->diff_words->plus.text.ptr);
1218 free (ecbdata->diff_words->plus.orig);
1219 if (ecbdata->diff_words->word_regex) {
1220 regfree(ecbdata->diff_words->word_regex);
1221 free(ecbdata->diff_words->word_regex);
1223 free(ecbdata->diff_words);
1224 ecbdata->diff_words = NULL;
1228 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1230 if (want_color(diff_use_color))
1231 return diff_colors[ix];
1232 return "";
1235 const char *diff_line_prefix(struct diff_options *opt)
1237 struct strbuf *msgbuf;
1238 if (!opt->output_prefix)
1239 return "";
1241 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1242 return msgbuf->buf;
1245 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1247 const char *cp;
1248 unsigned long allot;
1249 size_t l = len;
1251 if (ecb->truncate)
1252 return ecb->truncate(line, len);
1253 cp = line;
1254 allot = l;
1255 while (0 < l) {
1256 (void) utf8_width(&cp, &l);
1257 if (!cp)
1258 break; /* truncated in the middle? */
1260 return allot - l;
1263 static void find_lno(const char *line, struct emit_callback *ecbdata)
1265 const char *p;
1266 ecbdata->lno_in_preimage = 0;
1267 ecbdata->lno_in_postimage = 0;
1268 p = strchr(line, '-');
1269 if (!p)
1270 return; /* cannot happen */
1271 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1272 p = strchr(p, '+');
1273 if (!p)
1274 return; /* cannot happen */
1275 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1278 static void fn_out_consume(void *priv, char *line, unsigned long len)
1280 struct emit_callback *ecbdata = priv;
1281 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1282 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1283 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1284 struct diff_options *o = ecbdata->opt;
1285 const char *line_prefix = diff_line_prefix(o);
1287 o->found_changes = 1;
1289 if (ecbdata->header) {
1290 fprintf(o->file, "%s", ecbdata->header->buf);
1291 strbuf_reset(ecbdata->header);
1292 ecbdata->header = NULL;
1295 if (ecbdata->label_path[0]) {
1296 const char *name_a_tab, *name_b_tab;
1298 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1299 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1301 fprintf(o->file, "%s%s--- %s%s%s\n",
1302 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1303 fprintf(o->file, "%s%s+++ %s%s%s\n",
1304 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1305 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1308 if (diff_suppress_blank_empty
1309 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1310 line[0] = '\n';
1311 len = 1;
1314 if (line[0] == '@') {
1315 if (ecbdata->diff_words)
1316 diff_words_flush(ecbdata);
1317 len = sane_truncate_line(ecbdata, line, len);
1318 find_lno(line, ecbdata);
1319 emit_hunk_header(ecbdata, line, len);
1320 if (line[len-1] != '\n')
1321 putc('\n', o->file);
1322 return;
1325 if (ecbdata->diff_words) {
1326 if (line[0] == '-') {
1327 diff_words_append(line, len,
1328 &ecbdata->diff_words->minus);
1329 return;
1330 } else if (line[0] == '+') {
1331 diff_words_append(line, len,
1332 &ecbdata->diff_words->plus);
1333 return;
1334 } else if (starts_with(line, "\\ ")) {
1336 * Eat the "no newline at eof" marker as if we
1337 * saw a "+" or "-" line with nothing on it,
1338 * and return without diff_words_flush() to
1339 * defer processing. If this is the end of
1340 * preimage, more "+" lines may come after it.
1342 return;
1344 diff_words_flush(ecbdata);
1345 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1346 emit_line(o, context, reset, line, len);
1347 fputs("~\n", o->file);
1348 } else {
1350 * Skip the prefix character, if any. With
1351 * diff_suppress_blank_empty, there may be
1352 * none.
1354 if (line[0] != '\n') {
1355 line++;
1356 len--;
1358 emit_line(o, context, reset, line, len);
1360 return;
1363 switch (line[0]) {
1364 case '+':
1365 ecbdata->lno_in_postimage++;
1366 emit_add_line(reset, ecbdata, line + 1, len - 1);
1367 break;
1368 case '-':
1369 ecbdata->lno_in_preimage++;
1370 emit_del_line(reset, ecbdata, line + 1, len - 1);
1371 break;
1372 case ' ':
1373 ecbdata->lno_in_postimage++;
1374 ecbdata->lno_in_preimage++;
1375 emit_context_line(reset, ecbdata, line + 1, len - 1);
1376 break;
1377 default:
1378 /* incomplete line at the end */
1379 ecbdata->lno_in_preimage++;
1380 emit_line(o, diff_get_color(ecbdata->color_diff, DIFF_CONTEXT),
1381 reset, line, len);
1382 break;
1386 static char *pprint_rename(const char *a, const char *b)
1388 const char *old = a;
1389 const char *new = b;
1390 struct strbuf name = STRBUF_INIT;
1391 int pfx_length, sfx_length;
1392 int pfx_adjust_for_slash;
1393 int len_a = strlen(a);
1394 int len_b = strlen(b);
1395 int a_midlen, b_midlen;
1396 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1397 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1399 if (qlen_a || qlen_b) {
1400 quote_c_style(a, &name, NULL, 0);
1401 strbuf_addstr(&name, " => ");
1402 quote_c_style(b, &name, NULL, 0);
1403 return strbuf_detach(&name, NULL);
1406 /* Find common prefix */
1407 pfx_length = 0;
1408 while (*old && *new && *old == *new) {
1409 if (*old == '/')
1410 pfx_length = old - a + 1;
1411 old++;
1412 new++;
1415 /* Find common suffix */
1416 old = a + len_a;
1417 new = b + len_b;
1418 sfx_length = 0;
1420 * If there is a common prefix, it must end in a slash. In
1421 * that case we let this loop run 1 into the prefix to see the
1422 * same slash.
1424 * If there is no common prefix, we cannot do this as it would
1425 * underrun the input strings.
1427 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
1428 while (a + pfx_length - pfx_adjust_for_slash <= old &&
1429 b + pfx_length - pfx_adjust_for_slash <= new &&
1430 *old == *new) {
1431 if (*old == '/')
1432 sfx_length = len_a - (old - a);
1433 old--;
1434 new--;
1438 * pfx{mid-a => mid-b}sfx
1439 * {pfx-a => pfx-b}sfx
1440 * pfx{sfx-a => sfx-b}
1441 * name-a => name-b
1443 a_midlen = len_a - pfx_length - sfx_length;
1444 b_midlen = len_b - pfx_length - sfx_length;
1445 if (a_midlen < 0)
1446 a_midlen = 0;
1447 if (b_midlen < 0)
1448 b_midlen = 0;
1450 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1451 if (pfx_length + sfx_length) {
1452 strbuf_add(&name, a, pfx_length);
1453 strbuf_addch(&name, '{');
1455 strbuf_add(&name, a + pfx_length, a_midlen);
1456 strbuf_addstr(&name, " => ");
1457 strbuf_add(&name, b + pfx_length, b_midlen);
1458 if (pfx_length + sfx_length) {
1459 strbuf_addch(&name, '}');
1460 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1462 return strbuf_detach(&name, NULL);
1465 struct diffstat_t {
1466 int nr;
1467 int alloc;
1468 struct diffstat_file {
1469 char *from_name;
1470 char *name;
1471 char *print_name;
1472 unsigned is_unmerged:1;
1473 unsigned is_binary:1;
1474 unsigned is_renamed:1;
1475 unsigned is_interesting:1;
1476 uintmax_t added, deleted;
1477 } **files;
1480 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1481 const char *name_a,
1482 const char *name_b)
1484 struct diffstat_file *x;
1485 x = xcalloc(1, sizeof(*x));
1486 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
1487 diffstat->files[diffstat->nr++] = x;
1488 if (name_b) {
1489 x->from_name = xstrdup(name_a);
1490 x->name = xstrdup(name_b);
1491 x->is_renamed = 1;
1493 else {
1494 x->from_name = NULL;
1495 x->name = xstrdup(name_a);
1497 return x;
1500 static void diffstat_consume(void *priv, char *line, unsigned long len)
1502 struct diffstat_t *diffstat = priv;
1503 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1505 if (line[0] == '+')
1506 x->added++;
1507 else if (line[0] == '-')
1508 x->deleted++;
1511 const char mime_boundary_leader[] = "------------";
1513 static int scale_linear(int it, int width, int max_change)
1515 if (!it)
1516 return 0;
1518 * make sure that at least one '-' or '+' is printed if
1519 * there is any change to this path. The easiest way is to
1520 * scale linearly as if the alloted width is one column shorter
1521 * than it is, and then add 1 to the result.
1523 return 1 + (it * (width - 1) / max_change);
1526 static void show_name(FILE *file,
1527 const char *prefix, const char *name, int len)
1529 fprintf(file, " %s%-*s |", prefix, len, name);
1532 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1534 if (cnt <= 0)
1535 return;
1536 fprintf(file, "%s", set);
1537 while (cnt--)
1538 putc(ch, file);
1539 fprintf(file, "%s", reset);
1542 static void fill_print_name(struct diffstat_file *file)
1544 char *pname;
1546 if (file->print_name)
1547 return;
1549 if (!file->is_renamed) {
1550 struct strbuf buf = STRBUF_INIT;
1551 if (quote_c_style(file->name, &buf, NULL, 0)) {
1552 pname = strbuf_detach(&buf, NULL);
1553 } else {
1554 pname = file->name;
1555 strbuf_release(&buf);
1557 } else {
1558 pname = pprint_rename(file->from_name, file->name);
1560 file->print_name = pname;
1563 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1565 struct strbuf sb = STRBUF_INIT;
1566 int ret;
1568 if (!files) {
1569 assert(insertions == 0 && deletions == 0);
1570 return fprintf(fp, "%s\n", " 0 files changed");
1573 strbuf_addf(&sb,
1574 (files == 1) ? " %d file changed" : " %d files changed",
1575 files);
1578 * For binary diff, the caller may want to print "x files
1579 * changed" with insertions == 0 && deletions == 0.
1581 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1582 * is probably less confusing (i.e skip over "2 files changed
1583 * but nothing about added/removed lines? Is this a bug in Git?").
1585 if (insertions || deletions == 0) {
1586 strbuf_addf(&sb,
1587 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
1588 insertions);
1591 if (deletions || insertions == 0) {
1592 strbuf_addf(&sb,
1593 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
1594 deletions);
1596 strbuf_addch(&sb, '\n');
1597 ret = fputs(sb.buf, fp);
1598 strbuf_release(&sb);
1599 return ret;
1602 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1604 int i, len, add, del, adds = 0, dels = 0;
1605 uintmax_t max_change = 0, max_len = 0;
1606 int total_files = data->nr, count;
1607 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1608 const char *reset, *add_c, *del_c;
1609 const char *line_prefix = "";
1610 int extra_shown = 0;
1612 if (data->nr == 0)
1613 return;
1615 line_prefix = diff_line_prefix(options);
1616 count = options->stat_count ? options->stat_count : data->nr;
1618 reset = diff_get_color_opt(options, DIFF_RESET);
1619 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1620 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1623 * Find the longest filename and max number of changes
1625 for (i = 0; (i < count) && (i < data->nr); i++) {
1626 struct diffstat_file *file = data->files[i];
1627 uintmax_t change = file->added + file->deleted;
1629 if (!file->is_interesting && (change == 0)) {
1630 count++; /* not shown == room for one more */
1631 continue;
1633 fill_print_name(file);
1634 len = strlen(file->print_name);
1635 if (max_len < len)
1636 max_len = len;
1638 if (file->is_unmerged) {
1639 /* "Unmerged" is 8 characters */
1640 bin_width = bin_width < 8 ? 8 : bin_width;
1641 continue;
1643 if (file->is_binary) {
1644 /* "Bin XXX -> YYY bytes" */
1645 int w = 14 + decimal_width(file->added)
1646 + decimal_width(file->deleted);
1647 bin_width = bin_width < w ? w : bin_width;
1648 /* Display change counts aligned with "Bin" */
1649 number_width = 3;
1650 continue;
1653 if (max_change < change)
1654 max_change = change;
1656 count = i; /* where we can stop scanning in data->files[] */
1659 * We have width = stat_width or term_columns() columns total.
1660 * We want a maximum of min(max_len, stat_name_width) for the name part.
1661 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1662 * We also need 1 for " " and 4 + decimal_width(max_change)
1663 * for " | NNNN " and one the empty column at the end, altogether
1664 * 6 + decimal_width(max_change).
1666 * If there's not enough space, we will use the smaller of
1667 * stat_name_width (if set) and 5/8*width for the filename,
1668 * and the rest for constant elements + graph part, but no more
1669 * than stat_graph_width for the graph part.
1670 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1671 * for the standard terminal size).
1673 * In other words: stat_width limits the maximum width, and
1674 * stat_name_width fixes the maximum width of the filename,
1675 * and is also used to divide available columns if there
1676 * aren't enough.
1678 * Binary files are displayed with "Bin XXX -> YYY bytes"
1679 * instead of the change count and graph. This part is treated
1680 * similarly to the graph part, except that it is not
1681 * "scaled". If total width is too small to accommodate the
1682 * guaranteed minimum width of the filename part and the
1683 * separators and this message, this message will "overflow"
1684 * making the line longer than the maximum width.
1687 if (options->stat_width == -1)
1688 width = term_columns() - strlen(line_prefix);
1689 else
1690 width = options->stat_width ? options->stat_width : 80;
1691 number_width = decimal_width(max_change) > number_width ?
1692 decimal_width(max_change) : number_width;
1694 if (options->stat_graph_width == -1)
1695 options->stat_graph_width = diff_stat_graph_width;
1698 * Guarantee 3/8*16==6 for the graph part
1699 * and 5/8*16==10 for the filename part
1701 if (width < 16 + 6 + number_width)
1702 width = 16 + 6 + number_width;
1705 * First assign sizes that are wanted, ignoring available width.
1706 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1707 * starting from "XXX" should fit in graph_width.
1709 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1710 if (options->stat_graph_width &&
1711 options->stat_graph_width < graph_width)
1712 graph_width = options->stat_graph_width;
1714 name_width = (options->stat_name_width > 0 &&
1715 options->stat_name_width < max_len) ?
1716 options->stat_name_width : max_len;
1719 * Adjust adjustable widths not to exceed maximum width
1721 if (name_width + number_width + 6 + graph_width > width) {
1722 if (graph_width > width * 3/8 - number_width - 6) {
1723 graph_width = width * 3/8 - number_width - 6;
1724 if (graph_width < 6)
1725 graph_width = 6;
1728 if (options->stat_graph_width &&
1729 graph_width > options->stat_graph_width)
1730 graph_width = options->stat_graph_width;
1731 if (name_width > width - number_width - 6 - graph_width)
1732 name_width = width - number_width - 6 - graph_width;
1733 else
1734 graph_width = width - number_width - 6 - name_width;
1738 * From here name_width is the width of the name area,
1739 * and graph_width is the width of the graph area.
1740 * max_change is used to scale graph properly.
1742 for (i = 0; i < count; i++) {
1743 const char *prefix = "";
1744 struct diffstat_file *file = data->files[i];
1745 char *name = file->print_name;
1746 uintmax_t added = file->added;
1747 uintmax_t deleted = file->deleted;
1748 int name_len;
1750 if (!file->is_interesting && (added + deleted == 0))
1751 continue;
1754 * "scale" the filename
1756 len = name_width;
1757 name_len = strlen(name);
1758 if (name_width < name_len) {
1759 char *slash;
1760 prefix = "...";
1761 len -= 3;
1762 name += name_len - len;
1763 slash = strchr(name, '/');
1764 if (slash)
1765 name = slash;
1768 if (file->is_binary) {
1769 fprintf(options->file, "%s", line_prefix);
1770 show_name(options->file, prefix, name, len);
1771 fprintf(options->file, " %*s", number_width, "Bin");
1772 if (!added && !deleted) {
1773 putc('\n', options->file);
1774 continue;
1776 fprintf(options->file, " %s%"PRIuMAX"%s",
1777 del_c, deleted, reset);
1778 fprintf(options->file, " -> ");
1779 fprintf(options->file, "%s%"PRIuMAX"%s",
1780 add_c, added, reset);
1781 fprintf(options->file, " bytes");
1782 fprintf(options->file, "\n");
1783 continue;
1785 else if (file->is_unmerged) {
1786 fprintf(options->file, "%s", line_prefix);
1787 show_name(options->file, prefix, name, len);
1788 fprintf(options->file, " Unmerged\n");
1789 continue;
1793 * scale the add/delete
1795 add = added;
1796 del = deleted;
1798 if (graph_width <= max_change) {
1799 int total = scale_linear(add + del, graph_width, max_change);
1800 if (total < 2 && add && del)
1801 /* width >= 2 due to the sanity check */
1802 total = 2;
1803 if (add < del) {
1804 add = scale_linear(add, graph_width, max_change);
1805 del = total - add;
1806 } else {
1807 del = scale_linear(del, graph_width, max_change);
1808 add = total - del;
1811 fprintf(options->file, "%s", line_prefix);
1812 show_name(options->file, prefix, name, len);
1813 fprintf(options->file, " %*"PRIuMAX"%s",
1814 number_width, added + deleted,
1815 added + deleted ? " " : "");
1816 show_graph(options->file, '+', add, add_c, reset);
1817 show_graph(options->file, '-', del, del_c, reset);
1818 fprintf(options->file, "\n");
1821 for (i = 0; i < data->nr; i++) {
1822 struct diffstat_file *file = data->files[i];
1823 uintmax_t added = file->added;
1824 uintmax_t deleted = file->deleted;
1826 if (file->is_unmerged ||
1827 (!file->is_interesting && (added + deleted == 0))) {
1828 total_files--;
1829 continue;
1832 if (!file->is_binary) {
1833 adds += added;
1834 dels += deleted;
1836 if (i < count)
1837 continue;
1838 if (!extra_shown)
1839 fprintf(options->file, "%s ...\n", line_prefix);
1840 extra_shown = 1;
1842 fprintf(options->file, "%s", line_prefix);
1843 print_stat_summary(options->file, total_files, adds, dels);
1846 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1848 int i, adds = 0, dels = 0, total_files = data->nr;
1850 if (data->nr == 0)
1851 return;
1853 for (i = 0; i < data->nr; i++) {
1854 int added = data->files[i]->added;
1855 int deleted= data->files[i]->deleted;
1857 if (data->files[i]->is_unmerged ||
1858 (!data->files[i]->is_interesting && (added + deleted == 0))) {
1859 total_files--;
1860 } else if (!data->files[i]->is_binary) { /* don't count bytes */
1861 adds += added;
1862 dels += deleted;
1865 fprintf(options->file, "%s", diff_line_prefix(options));
1866 print_stat_summary(options->file, total_files, adds, dels);
1869 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1871 int i;
1873 if (data->nr == 0)
1874 return;
1876 for (i = 0; i < data->nr; i++) {
1877 struct diffstat_file *file = data->files[i];
1879 fprintf(options->file, "%s", diff_line_prefix(options));
1881 if (file->is_binary)
1882 fprintf(options->file, "-\t-\t");
1883 else
1884 fprintf(options->file,
1885 "%"PRIuMAX"\t%"PRIuMAX"\t",
1886 file->added, file->deleted);
1887 if (options->line_termination) {
1888 fill_print_name(file);
1889 if (!file->is_renamed)
1890 write_name_quoted(file->name, options->file,
1891 options->line_termination);
1892 else {
1893 fputs(file->print_name, options->file);
1894 putc(options->line_termination, options->file);
1896 } else {
1897 if (file->is_renamed) {
1898 putc('\0', options->file);
1899 write_name_quoted(file->from_name, options->file, '\0');
1901 write_name_quoted(file->name, options->file, '\0');
1906 struct dirstat_file {
1907 const char *name;
1908 unsigned long changed;
1911 struct dirstat_dir {
1912 struct dirstat_file *files;
1913 int alloc, nr, permille, cumulative;
1916 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1917 unsigned long changed, const char *base, int baselen)
1919 unsigned long this_dir = 0;
1920 unsigned int sources = 0;
1921 const char *line_prefix = diff_line_prefix(opt);
1923 while (dir->nr) {
1924 struct dirstat_file *f = dir->files;
1925 int namelen = strlen(f->name);
1926 unsigned long this;
1927 char *slash;
1929 if (namelen < baselen)
1930 break;
1931 if (memcmp(f->name, base, baselen))
1932 break;
1933 slash = strchr(f->name + baselen, '/');
1934 if (slash) {
1935 int newbaselen = slash + 1 - f->name;
1936 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1937 sources++;
1938 } else {
1939 this = f->changed;
1940 dir->files++;
1941 dir->nr--;
1942 sources += 2;
1944 this_dir += this;
1948 * We don't report dirstat's for
1949 * - the top level
1950 * - or cases where everything came from a single directory
1951 * under this directory (sources == 1).
1953 if (baselen && sources != 1) {
1954 if (this_dir) {
1955 int permille = this_dir * 1000 / changed;
1956 if (permille >= dir->permille) {
1957 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1958 permille / 10, permille % 10, baselen, base);
1959 if (!dir->cumulative)
1960 return 0;
1964 return this_dir;
1967 static int dirstat_compare(const void *_a, const void *_b)
1969 const struct dirstat_file *a = _a;
1970 const struct dirstat_file *b = _b;
1971 return strcmp(a->name, b->name);
1974 static void show_dirstat(struct diff_options *options)
1976 int i;
1977 unsigned long changed;
1978 struct dirstat_dir dir;
1979 struct diff_queue_struct *q = &diff_queued_diff;
1981 dir.files = NULL;
1982 dir.alloc = 0;
1983 dir.nr = 0;
1984 dir.permille = options->dirstat_permille;
1985 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1987 changed = 0;
1988 for (i = 0; i < q->nr; i++) {
1989 struct diff_filepair *p = q->queue[i];
1990 const char *name;
1991 unsigned long copied, added, damage;
1992 int content_changed;
1994 name = p->two->path ? p->two->path : p->one->path;
1996 if (p->one->oid_valid && p->two->oid_valid)
1997 content_changed = oidcmp(&p->one->oid, &p->two->oid);
1998 else
1999 content_changed = 1;
2001 if (!content_changed) {
2003 * The SHA1 has not changed, so pre-/post-content is
2004 * identical. We can therefore skip looking at the
2005 * file contents altogether.
2007 damage = 0;
2008 goto found_damage;
2011 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2013 * In --dirstat-by-file mode, we don't really need to
2014 * look at the actual file contents at all.
2015 * The fact that the SHA1 changed is enough for us to
2016 * add this file to the list of results
2017 * (with each file contributing equal damage).
2019 damage = 1;
2020 goto found_damage;
2023 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2024 diff_populate_filespec(p->one, 0);
2025 diff_populate_filespec(p->two, 0);
2026 diffcore_count_changes(p->one, p->two, NULL, NULL,
2027 &copied, &added);
2028 diff_free_filespec_data(p->one);
2029 diff_free_filespec_data(p->two);
2030 } else if (DIFF_FILE_VALID(p->one)) {
2031 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2032 copied = added = 0;
2033 diff_free_filespec_data(p->one);
2034 } else if (DIFF_FILE_VALID(p->two)) {
2035 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2036 copied = 0;
2037 added = p->two->size;
2038 diff_free_filespec_data(p->two);
2039 } else
2040 continue;
2043 * Original minus copied is the removed material,
2044 * added is the new material. They are both damages
2045 * made to the preimage.
2046 * If the resulting damage is zero, we know that
2047 * diffcore_count_changes() considers the two entries to
2048 * be identical, but since content_changed is true, we
2049 * know that there must have been _some_ kind of change,
2050 * so we force all entries to have damage > 0.
2052 damage = (p->one->size - copied) + added;
2053 if (!damage)
2054 damage = 1;
2056 found_damage:
2057 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2058 dir.files[dir.nr].name = name;
2059 dir.files[dir.nr].changed = damage;
2060 changed += damage;
2061 dir.nr++;
2064 /* This can happen even with many files, if everything was renames */
2065 if (!changed)
2066 return;
2068 /* Show all directories with more than x% of the changes */
2069 QSORT(dir.files, dir.nr, dirstat_compare);
2070 gather_dirstat(options, &dir, changed, "", 0);
2073 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2075 int i;
2076 unsigned long changed;
2077 struct dirstat_dir dir;
2079 if (data->nr == 0)
2080 return;
2082 dir.files = NULL;
2083 dir.alloc = 0;
2084 dir.nr = 0;
2085 dir.permille = options->dirstat_permille;
2086 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2088 changed = 0;
2089 for (i = 0; i < data->nr; i++) {
2090 struct diffstat_file *file = data->files[i];
2091 unsigned long damage = file->added + file->deleted;
2092 if (file->is_binary)
2094 * binary files counts bytes, not lines. Must find some
2095 * way to normalize binary bytes vs. textual lines.
2096 * The following heuristic assumes that there are 64
2097 * bytes per "line".
2098 * This is stupid and ugly, but very cheap...
2100 damage = (damage + 63) / 64;
2101 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2102 dir.files[dir.nr].name = file->name;
2103 dir.files[dir.nr].changed = damage;
2104 changed += damage;
2105 dir.nr++;
2108 /* This can happen even with many files, if everything was renames */
2109 if (!changed)
2110 return;
2112 /* Show all directories with more than x% of the changes */
2113 QSORT(dir.files, dir.nr, dirstat_compare);
2114 gather_dirstat(options, &dir, changed, "", 0);
2117 static void free_diffstat_info(struct diffstat_t *diffstat)
2119 int i;
2120 for (i = 0; i < diffstat->nr; i++) {
2121 struct diffstat_file *f = diffstat->files[i];
2122 if (f->name != f->print_name)
2123 free(f->print_name);
2124 free(f->name);
2125 free(f->from_name);
2126 free(f);
2128 free(diffstat->files);
2131 struct checkdiff_t {
2132 const char *filename;
2133 int lineno;
2134 int conflict_marker_size;
2135 struct diff_options *o;
2136 unsigned ws_rule;
2137 unsigned status;
2140 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2142 char firstchar;
2143 int cnt;
2145 if (len < marker_size + 1)
2146 return 0;
2147 firstchar = line[0];
2148 switch (firstchar) {
2149 case '=': case '>': case '<': case '|':
2150 break;
2151 default:
2152 return 0;
2154 for (cnt = 1; cnt < marker_size; cnt++)
2155 if (line[cnt] != firstchar)
2156 return 0;
2157 /* line[1] thru line[marker_size-1] are same as firstchar */
2158 if (len < marker_size + 1 || !isspace(line[marker_size]))
2159 return 0;
2160 return 1;
2163 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2165 struct checkdiff_t *data = priv;
2166 int marker_size = data->conflict_marker_size;
2167 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2168 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2169 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2170 char *err;
2171 const char *line_prefix;
2173 assert(data->o);
2174 line_prefix = diff_line_prefix(data->o);
2176 if (line[0] == '+') {
2177 unsigned bad;
2178 data->lineno++;
2179 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2180 data->status |= 1;
2181 fprintf(data->o->file,
2182 "%s%s:%d: leftover conflict marker\n",
2183 line_prefix, data->filename, data->lineno);
2185 bad = ws_check(line + 1, len - 1, data->ws_rule);
2186 if (!bad)
2187 return;
2188 data->status |= bad;
2189 err = whitespace_error_string(bad);
2190 fprintf(data->o->file, "%s%s:%d: %s.\n",
2191 line_prefix, data->filename, data->lineno, err);
2192 free(err);
2193 emit_line(data->o, set, reset, line, 1);
2194 ws_check_emit(line + 1, len - 1, data->ws_rule,
2195 data->o->file, set, reset, ws);
2196 } else if (line[0] == ' ') {
2197 data->lineno++;
2198 } else if (line[0] == '@') {
2199 char *plus = strchr(line, '+');
2200 if (plus)
2201 data->lineno = strtol(plus, NULL, 10) - 1;
2202 else
2203 die("invalid diff");
2207 static unsigned char *deflate_it(char *data,
2208 unsigned long size,
2209 unsigned long *result_size)
2211 int bound;
2212 unsigned char *deflated;
2213 git_zstream stream;
2215 git_deflate_init(&stream, zlib_compression_level);
2216 bound = git_deflate_bound(&stream, size);
2217 deflated = xmalloc(bound);
2218 stream.next_out = deflated;
2219 stream.avail_out = bound;
2221 stream.next_in = (unsigned char *)data;
2222 stream.avail_in = size;
2223 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2224 ; /* nothing */
2225 git_deflate_end(&stream);
2226 *result_size = stream.total_out;
2227 return deflated;
2230 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two,
2231 const char *prefix)
2233 void *cp;
2234 void *delta;
2235 void *deflated;
2236 void *data;
2237 unsigned long orig_size;
2238 unsigned long delta_size;
2239 unsigned long deflate_size;
2240 unsigned long data_size;
2242 /* We could do deflated delta, or we could do just deflated two,
2243 * whichever is smaller.
2245 delta = NULL;
2246 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2247 if (one->size && two->size) {
2248 delta = diff_delta(one->ptr, one->size,
2249 two->ptr, two->size,
2250 &delta_size, deflate_size);
2251 if (delta) {
2252 void *to_free = delta;
2253 orig_size = delta_size;
2254 delta = deflate_it(delta, delta_size, &delta_size);
2255 free(to_free);
2259 if (delta && delta_size < deflate_size) {
2260 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2261 free(deflated);
2262 data = delta;
2263 data_size = delta_size;
2265 else {
2266 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2267 free(delta);
2268 data = deflated;
2269 data_size = deflate_size;
2272 /* emit data encoded in base85 */
2273 cp = data;
2274 while (data_size) {
2275 int bytes = (52 < data_size) ? 52 : data_size;
2276 char line[70];
2277 data_size -= bytes;
2278 if (bytes <= 26)
2279 line[0] = bytes + 'A' - 1;
2280 else
2281 line[0] = bytes - 26 + 'a' - 1;
2282 encode_85(line + 1, cp, bytes);
2283 cp = (char *) cp + bytes;
2284 fprintf(file, "%s", prefix);
2285 fputs(line, file);
2286 fputc('\n', file);
2288 fprintf(file, "%s\n", prefix);
2289 free(data);
2292 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two,
2293 const char *prefix)
2295 fprintf(file, "%sGIT binary patch\n", prefix);
2296 emit_binary_diff_body(file, one, two, prefix);
2297 emit_binary_diff_body(file, two, one, prefix);
2300 int diff_filespec_is_binary(struct diff_filespec *one)
2302 if (one->is_binary == -1) {
2303 diff_filespec_load_driver(one);
2304 if (one->driver->binary != -1)
2305 one->is_binary = one->driver->binary;
2306 else {
2307 if (!one->data && DIFF_FILE_VALID(one))
2308 diff_populate_filespec(one, CHECK_BINARY);
2309 if (one->is_binary == -1 && one->data)
2310 one->is_binary = buffer_is_binary(one->data,
2311 one->size);
2312 if (one->is_binary == -1)
2313 one->is_binary = 0;
2316 return one->is_binary;
2319 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2321 diff_filespec_load_driver(one);
2322 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2325 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2327 if (!options->a_prefix)
2328 options->a_prefix = a;
2329 if (!options->b_prefix)
2330 options->b_prefix = b;
2333 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2335 if (!DIFF_FILE_VALID(one))
2336 return NULL;
2338 diff_filespec_load_driver(one);
2339 return userdiff_get_textconv(one->driver);
2342 static void builtin_diff(const char *name_a,
2343 const char *name_b,
2344 struct diff_filespec *one,
2345 struct diff_filespec *two,
2346 const char *xfrm_msg,
2347 int must_show_header,
2348 struct diff_options *o,
2349 int complete_rewrite)
2351 mmfile_t mf1, mf2;
2352 const char *lbl[2];
2353 char *a_one, *b_two;
2354 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
2355 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2356 const char *a_prefix, *b_prefix;
2357 struct userdiff_driver *textconv_one = NULL;
2358 struct userdiff_driver *textconv_two = NULL;
2359 struct strbuf header = STRBUF_INIT;
2360 const char *line_prefix = diff_line_prefix(o);
2362 diff_set_mnemonic_prefix(o, "a/", "b/");
2363 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2364 a_prefix = o->b_prefix;
2365 b_prefix = o->a_prefix;
2366 } else {
2367 a_prefix = o->a_prefix;
2368 b_prefix = o->b_prefix;
2371 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
2372 (!one->mode || S_ISGITLINK(one->mode)) &&
2373 (!two->mode || S_ISGITLINK(two->mode))) {
2374 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2375 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2376 show_submodule_summary(o->file, one->path ? one->path : two->path,
2377 line_prefix,
2378 &one->oid, &two->oid,
2379 two->dirty_submodule,
2380 meta, del, add, reset);
2381 return;
2382 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
2383 (!one->mode || S_ISGITLINK(one->mode)) &&
2384 (!two->mode || S_ISGITLINK(two->mode))) {
2385 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2386 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2387 show_submodule_inline_diff(o->file, one->path ? one->path : two->path,
2388 line_prefix,
2389 &one->oid, &two->oid,
2390 two->dirty_submodule,
2391 meta, del, add, reset, o);
2392 return;
2395 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2396 textconv_one = get_textconv(one);
2397 textconv_two = get_textconv(two);
2400 /* Never use a non-valid filename anywhere if at all possible */
2401 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2402 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2404 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2405 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2406 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2407 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2408 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
2409 if (lbl[0][0] == '/') {
2410 /* /dev/null */
2411 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
2412 if (xfrm_msg)
2413 strbuf_addstr(&header, xfrm_msg);
2414 must_show_header = 1;
2416 else if (lbl[1][0] == '/') {
2417 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
2418 if (xfrm_msg)
2419 strbuf_addstr(&header, xfrm_msg);
2420 must_show_header = 1;
2422 else {
2423 if (one->mode != two->mode) {
2424 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
2425 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
2426 must_show_header = 1;
2428 if (xfrm_msg)
2429 strbuf_addstr(&header, xfrm_msg);
2432 * we do not run diff between different kind
2433 * of objects.
2435 if ((one->mode ^ two->mode) & S_IFMT)
2436 goto free_ab_and_return;
2437 if (complete_rewrite &&
2438 (textconv_one || !diff_filespec_is_binary(one)) &&
2439 (textconv_two || !diff_filespec_is_binary(two))) {
2440 fprintf(o->file, "%s", header.buf);
2441 strbuf_reset(&header);
2442 emit_rewrite_diff(name_a, name_b, one, two,
2443 textconv_one, textconv_two, o);
2444 o->found_changes = 1;
2445 goto free_ab_and_return;
2449 if (o->irreversible_delete && lbl[1][0] == '/') {
2450 fprintf(o->file, "%s", header.buf);
2451 strbuf_reset(&header);
2452 goto free_ab_and_return;
2453 } else if (!DIFF_OPT_TST(o, TEXT) &&
2454 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2455 (!textconv_two && diff_filespec_is_binary(two)) )) {
2456 if (!one->data && !two->data &&
2457 S_ISREG(one->mode) && S_ISREG(two->mode) &&
2458 !DIFF_OPT_TST(o, BINARY)) {
2459 if (!oidcmp(&one->oid, &two->oid)) {
2460 if (must_show_header)
2461 fprintf(o->file, "%s", header.buf);
2462 goto free_ab_and_return;
2464 fprintf(o->file, "%s", header.buf);
2465 fprintf(o->file, "%sBinary files %s and %s differ\n",
2466 line_prefix, lbl[0], lbl[1]);
2467 goto free_ab_and_return;
2469 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2470 die("unable to read files to diff");
2471 /* Quite common confusing case */
2472 if (mf1.size == mf2.size &&
2473 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2474 if (must_show_header)
2475 fprintf(o->file, "%s", header.buf);
2476 goto free_ab_and_return;
2478 fprintf(o->file, "%s", header.buf);
2479 strbuf_reset(&header);
2480 if (DIFF_OPT_TST(o, BINARY))
2481 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2482 else
2483 fprintf(o->file, "%sBinary files %s and %s differ\n",
2484 line_prefix, lbl[0], lbl[1]);
2485 o->found_changes = 1;
2486 } else {
2487 /* Crazy xdl interfaces.. */
2488 const char *diffopts = getenv("GIT_DIFF_OPTS");
2489 const char *v;
2490 xpparam_t xpp;
2491 xdemitconf_t xecfg;
2492 struct emit_callback ecbdata;
2493 const struct userdiff_funcname *pe;
2495 if (must_show_header) {
2496 fprintf(o->file, "%s", header.buf);
2497 strbuf_reset(&header);
2500 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2501 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2503 pe = diff_funcname_pattern(one);
2504 if (!pe)
2505 pe = diff_funcname_pattern(two);
2507 memset(&xpp, 0, sizeof(xpp));
2508 memset(&xecfg, 0, sizeof(xecfg));
2509 memset(&ecbdata, 0, sizeof(ecbdata));
2510 ecbdata.label_path = lbl;
2511 ecbdata.color_diff = want_color(o->use_color);
2512 ecbdata.ws_rule = whitespace_rule(name_b);
2513 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2514 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2515 ecbdata.opt = o;
2516 ecbdata.header = header.len ? &header : NULL;
2517 xpp.flags = o->xdl_opts;
2518 xecfg.ctxlen = o->context;
2519 xecfg.interhunkctxlen = o->interhunkcontext;
2520 xecfg.flags = XDL_EMIT_FUNCNAMES;
2521 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2522 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2523 if (pe)
2524 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2525 if (!diffopts)
2527 else if (skip_prefix(diffopts, "--unified=", &v))
2528 xecfg.ctxlen = strtoul(v, NULL, 10);
2529 else if (skip_prefix(diffopts, "-u", &v))
2530 xecfg.ctxlen = strtoul(v, NULL, 10);
2531 if (o->word_diff)
2532 init_diff_words_data(&ecbdata, o, one, two);
2533 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2534 &xpp, &xecfg))
2535 die("unable to generate diff for %s", one->path);
2536 if (o->word_diff)
2537 free_diff_words_data(&ecbdata);
2538 if (textconv_one)
2539 free(mf1.ptr);
2540 if (textconv_two)
2541 free(mf2.ptr);
2542 xdiff_clear_find_func(&xecfg);
2545 free_ab_and_return:
2546 strbuf_release(&header);
2547 diff_free_filespec_data(one);
2548 diff_free_filespec_data(two);
2549 free(a_one);
2550 free(b_two);
2551 return;
2554 static void builtin_diffstat(const char *name_a, const char *name_b,
2555 struct diff_filespec *one,
2556 struct diff_filespec *two,
2557 struct diffstat_t *diffstat,
2558 struct diff_options *o,
2559 struct diff_filepair *p)
2561 mmfile_t mf1, mf2;
2562 struct diffstat_file *data;
2563 int same_contents;
2564 int complete_rewrite = 0;
2566 if (!DIFF_PAIR_UNMERGED(p)) {
2567 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2568 complete_rewrite = 1;
2571 data = diffstat_add(diffstat, name_a, name_b);
2572 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
2574 if (!one || !two) {
2575 data->is_unmerged = 1;
2576 return;
2579 same_contents = !oidcmp(&one->oid, &two->oid);
2581 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2582 data->is_binary = 1;
2583 if (same_contents) {
2584 data->added = 0;
2585 data->deleted = 0;
2586 } else {
2587 data->added = diff_filespec_size(two);
2588 data->deleted = diff_filespec_size(one);
2592 else if (complete_rewrite) {
2593 diff_populate_filespec(one, 0);
2594 diff_populate_filespec(two, 0);
2595 data->deleted = count_lines(one->data, one->size);
2596 data->added = count_lines(two->data, two->size);
2599 else if (!same_contents) {
2600 /* Crazy xdl interfaces.. */
2601 xpparam_t xpp;
2602 xdemitconf_t xecfg;
2604 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2605 die("unable to read files to diff");
2607 memset(&xpp, 0, sizeof(xpp));
2608 memset(&xecfg, 0, sizeof(xecfg));
2609 xpp.flags = o->xdl_opts;
2610 xecfg.ctxlen = o->context;
2611 xecfg.interhunkctxlen = o->interhunkcontext;
2612 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2613 &xpp, &xecfg))
2614 die("unable to generate diffstat for %s", one->path);
2617 diff_free_filespec_data(one);
2618 diff_free_filespec_data(two);
2621 static void builtin_checkdiff(const char *name_a, const char *name_b,
2622 const char *attr_path,
2623 struct diff_filespec *one,
2624 struct diff_filespec *two,
2625 struct diff_options *o)
2627 mmfile_t mf1, mf2;
2628 struct checkdiff_t data;
2630 if (!two)
2631 return;
2633 memset(&data, 0, sizeof(data));
2634 data.filename = name_b ? name_b : name_a;
2635 data.lineno = 0;
2636 data.o = o;
2637 data.ws_rule = whitespace_rule(attr_path);
2638 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2640 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2641 die("unable to read files to diff");
2644 * All the other codepaths check both sides, but not checking
2645 * the "old" side here is deliberate. We are checking the newly
2646 * introduced changes, and as long as the "new" side is text, we
2647 * can and should check what it introduces.
2649 if (diff_filespec_is_binary(two))
2650 goto free_and_return;
2651 else {
2652 /* Crazy xdl interfaces.. */
2653 xpparam_t xpp;
2654 xdemitconf_t xecfg;
2656 memset(&xpp, 0, sizeof(xpp));
2657 memset(&xecfg, 0, sizeof(xecfg));
2658 xecfg.ctxlen = 1; /* at least one context line */
2659 xpp.flags = 0;
2660 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2661 &xpp, &xecfg))
2662 die("unable to generate checkdiff for %s", one->path);
2664 if (data.ws_rule & WS_BLANK_AT_EOF) {
2665 struct emit_callback ecbdata;
2666 int blank_at_eof;
2668 ecbdata.ws_rule = data.ws_rule;
2669 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2670 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2672 if (blank_at_eof) {
2673 static char *err;
2674 if (!err)
2675 err = whitespace_error_string(WS_BLANK_AT_EOF);
2676 fprintf(o->file, "%s:%d: %s.\n",
2677 data.filename, blank_at_eof, err);
2678 data.status = 1; /* report errors */
2682 free_and_return:
2683 diff_free_filespec_data(one);
2684 diff_free_filespec_data(two);
2685 if (data.status)
2686 DIFF_OPT_SET(o, CHECK_FAILED);
2689 struct diff_filespec *alloc_filespec(const char *path)
2691 struct diff_filespec *spec;
2693 FLEXPTR_ALLOC_STR(spec, path, path);
2694 spec->count = 1;
2695 spec->is_binary = -1;
2696 return spec;
2699 void free_filespec(struct diff_filespec *spec)
2701 if (!--spec->count) {
2702 diff_free_filespec_data(spec);
2703 free(spec);
2707 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2708 int sha1_valid, unsigned short mode)
2710 if (mode) {
2711 spec->mode = canon_mode(mode);
2712 hashcpy(spec->oid.hash, sha1);
2713 spec->oid_valid = sha1_valid;
2718 * Given a name and sha1 pair, if the index tells us the file in
2719 * the work tree has that object contents, return true, so that
2720 * prepare_temp_file() does not have to inflate and extract.
2722 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2724 const struct cache_entry *ce;
2725 struct stat st;
2726 int pos, len;
2729 * We do not read the cache ourselves here, because the
2730 * benchmark with my previous version that always reads cache
2731 * shows that it makes things worse for diff-tree comparing
2732 * two linux-2.6 kernel trees in an already checked out work
2733 * tree. This is because most diff-tree comparisons deal with
2734 * only a small number of files, while reading the cache is
2735 * expensive for a large project, and its cost outweighs the
2736 * savings we get by not inflating the object to a temporary
2737 * file. Practically, this code only helps when we are used
2738 * by diff-cache --cached, which does read the cache before
2739 * calling us.
2741 if (!active_cache)
2742 return 0;
2744 /* We want to avoid the working directory if our caller
2745 * doesn't need the data in a normal file, this system
2746 * is rather slow with its stat/open/mmap/close syscalls,
2747 * and the object is contained in a pack file. The pack
2748 * is probably already open and will be faster to obtain
2749 * the data through than the working directory. Loose
2750 * objects however would tend to be slower as they need
2751 * to be individually opened and inflated.
2753 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2754 return 0;
2757 * Similarly, if we'd have to convert the file contents anyway, that
2758 * makes the optimization not worthwhile.
2760 if (!want_file && would_convert_to_git(name))
2761 return 0;
2763 len = strlen(name);
2764 pos = cache_name_pos(name, len);
2765 if (pos < 0)
2766 return 0;
2767 ce = active_cache[pos];
2770 * This is not the sha1 we are looking for, or
2771 * unreusable because it is not a regular file.
2773 if (hashcmp(sha1, ce->oid.hash) || !S_ISREG(ce->ce_mode))
2774 return 0;
2777 * If ce is marked as "assume unchanged", there is no
2778 * guarantee that work tree matches what we are looking for.
2780 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2781 return 0;
2784 * If ce matches the file in the work tree, we can reuse it.
2786 if (ce_uptodate(ce) ||
2787 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2788 return 1;
2790 return 0;
2793 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2795 struct strbuf buf = STRBUF_INIT;
2796 char *dirty = "";
2798 /* Are we looking at the work tree? */
2799 if (s->dirty_submodule)
2800 dirty = "-dirty";
2802 strbuf_addf(&buf, "Subproject commit %s%s\n",
2803 oid_to_hex(&s->oid), dirty);
2804 s->size = buf.len;
2805 if (size_only) {
2806 s->data = NULL;
2807 strbuf_release(&buf);
2808 } else {
2809 s->data = strbuf_detach(&buf, NULL);
2810 s->should_free = 1;
2812 return 0;
2816 * While doing rename detection and pickaxe operation, we may need to
2817 * grab the data for the blob (or file) for our own in-core comparison.
2818 * diff_filespec has data and size fields for this purpose.
2820 int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
2822 int size_only = flags & CHECK_SIZE_ONLY;
2823 int err = 0;
2825 * demote FAIL to WARN to allow inspecting the situation
2826 * instead of refusing.
2828 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
2829 ? SAFE_CRLF_WARN
2830 : safe_crlf);
2832 if (!DIFF_FILE_VALID(s))
2833 die("internal error: asking to populate invalid file.");
2834 if (S_ISDIR(s->mode))
2835 return -1;
2837 if (s->data)
2838 return 0;
2840 if (size_only && 0 < s->size)
2841 return 0;
2843 if (S_ISGITLINK(s->mode))
2844 return diff_populate_gitlink(s, size_only);
2846 if (!s->oid_valid ||
2847 reuse_worktree_file(s->path, s->oid.hash, 0)) {
2848 struct strbuf buf = STRBUF_INIT;
2849 struct stat st;
2850 int fd;
2852 if (lstat(s->path, &st) < 0) {
2853 if (errno == ENOENT) {
2854 err_empty:
2855 err = -1;
2856 empty:
2857 s->data = (char *)"";
2858 s->size = 0;
2859 return err;
2862 s->size = xsize_t(st.st_size);
2863 if (!s->size)
2864 goto empty;
2865 if (S_ISLNK(st.st_mode)) {
2866 struct strbuf sb = STRBUF_INIT;
2868 if (strbuf_readlink(&sb, s->path, s->size))
2869 goto err_empty;
2870 s->size = sb.len;
2871 s->data = strbuf_detach(&sb, NULL);
2872 s->should_free = 1;
2873 return 0;
2875 if (size_only)
2876 return 0;
2877 if ((flags & CHECK_BINARY) &&
2878 s->size > big_file_threshold && s->is_binary == -1) {
2879 s->is_binary = 1;
2880 return 0;
2882 fd = open(s->path, O_RDONLY);
2883 if (fd < 0)
2884 goto err_empty;
2885 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2886 close(fd);
2887 s->should_munmap = 1;
2890 * Convert from working tree format to canonical git format
2892 if (convert_to_git(s->path, s->data, s->size, &buf, crlf_warn)) {
2893 size_t size = 0;
2894 munmap(s->data, s->size);
2895 s->should_munmap = 0;
2896 s->data = strbuf_detach(&buf, &size);
2897 s->size = size;
2898 s->should_free = 1;
2901 else {
2902 enum object_type type;
2903 if (size_only || (flags & CHECK_BINARY)) {
2904 type = sha1_object_info(s->oid.hash, &s->size);
2905 if (type < 0)
2906 die("unable to read %s",
2907 oid_to_hex(&s->oid));
2908 if (size_only)
2909 return 0;
2910 if (s->size > big_file_threshold && s->is_binary == -1) {
2911 s->is_binary = 1;
2912 return 0;
2915 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
2916 if (!s->data)
2917 die("unable to read %s", oid_to_hex(&s->oid));
2918 s->should_free = 1;
2920 return 0;
2923 void diff_free_filespec_blob(struct diff_filespec *s)
2925 if (s->should_free)
2926 free(s->data);
2927 else if (s->should_munmap)
2928 munmap(s->data, s->size);
2930 if (s->should_free || s->should_munmap) {
2931 s->should_free = s->should_munmap = 0;
2932 s->data = NULL;
2936 void diff_free_filespec_data(struct diff_filespec *s)
2938 diff_free_filespec_blob(s);
2939 free(s->cnt_data);
2940 s->cnt_data = NULL;
2943 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2944 void *blob,
2945 unsigned long size,
2946 const struct object_id *oid,
2947 int mode)
2949 int fd;
2950 struct strbuf buf = STRBUF_INIT;
2951 struct strbuf template = STRBUF_INIT;
2952 char *path_dup = xstrdup(path);
2953 const char *base = basename(path_dup);
2955 /* Generate "XXXXXX_basename.ext" */
2956 strbuf_addstr(&template, "XXXXXX_");
2957 strbuf_addstr(&template, base);
2959 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
2960 if (fd < 0)
2961 die_errno("unable to create temp-file");
2962 if (convert_to_working_tree(path,
2963 (const char *)blob, (size_t)size, &buf)) {
2964 blob = buf.buf;
2965 size = buf.len;
2967 if (write_in_full(fd, blob, size) != size)
2968 die_errno("unable to write temp-file");
2969 close_tempfile(&temp->tempfile);
2970 temp->name = get_tempfile_path(&temp->tempfile);
2971 oid_to_hex_r(temp->hex, oid);
2972 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
2973 strbuf_release(&buf);
2974 strbuf_release(&template);
2975 free(path_dup);
2978 static struct diff_tempfile *prepare_temp_file(const char *name,
2979 struct diff_filespec *one)
2981 struct diff_tempfile *temp = claim_diff_tempfile();
2983 if (!DIFF_FILE_VALID(one)) {
2984 not_a_valid_file:
2985 /* A '-' entry produces this for file-2, and
2986 * a '+' entry produces this for file-1.
2988 temp->name = "/dev/null";
2989 xsnprintf(temp->hex, sizeof(temp->hex), ".");
2990 xsnprintf(temp->mode, sizeof(temp->mode), ".");
2991 return temp;
2994 if (!S_ISGITLINK(one->mode) &&
2995 (!one->oid_valid ||
2996 reuse_worktree_file(name, one->oid.hash, 1))) {
2997 struct stat st;
2998 if (lstat(name, &st) < 0) {
2999 if (errno == ENOENT)
3000 goto not_a_valid_file;
3001 die_errno("stat(%s)", name);
3003 if (S_ISLNK(st.st_mode)) {
3004 struct strbuf sb = STRBUF_INIT;
3005 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3006 die_errno("readlink(%s)", name);
3007 prep_temp_blob(name, temp, sb.buf, sb.len,
3008 (one->oid_valid ?
3009 &one->oid : &null_oid),
3010 (one->oid_valid ?
3011 one->mode : S_IFLNK));
3012 strbuf_release(&sb);
3014 else {
3015 /* we can borrow from the file in the work tree */
3016 temp->name = name;
3017 if (!one->oid_valid)
3018 sha1_to_hex_r(temp->hex, null_sha1);
3019 else
3020 sha1_to_hex_r(temp->hex, one->oid.hash);
3021 /* Even though we may sometimes borrow the
3022 * contents from the work tree, we always want
3023 * one->mode. mode is trustworthy even when
3024 * !(one->sha1_valid), as long as
3025 * DIFF_FILE_VALID(one).
3027 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3029 return temp;
3031 else {
3032 if (diff_populate_filespec(one, 0))
3033 die("cannot read data blob for %s", one->path);
3034 prep_temp_blob(name, temp, one->data, one->size,
3035 &one->oid, one->mode);
3037 return temp;
3040 static void add_external_diff_name(struct argv_array *argv,
3041 const char *name,
3042 struct diff_filespec *df)
3044 struct diff_tempfile *temp = prepare_temp_file(name, df);
3045 argv_array_push(argv, temp->name);
3046 argv_array_push(argv, temp->hex);
3047 argv_array_push(argv, temp->mode);
3050 /* An external diff command takes:
3052 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3053 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3056 static void run_external_diff(const char *pgm,
3057 const char *name,
3058 const char *other,
3059 struct diff_filespec *one,
3060 struct diff_filespec *two,
3061 const char *xfrm_msg,
3062 int complete_rewrite,
3063 struct diff_options *o)
3065 struct argv_array argv = ARGV_ARRAY_INIT;
3066 struct argv_array env = ARGV_ARRAY_INIT;
3067 struct diff_queue_struct *q = &diff_queued_diff;
3069 argv_array_push(&argv, pgm);
3070 argv_array_push(&argv, name);
3072 if (one && two) {
3073 add_external_diff_name(&argv, name, one);
3074 if (!other)
3075 add_external_diff_name(&argv, name, two);
3076 else {
3077 add_external_diff_name(&argv, other, two);
3078 argv_array_push(&argv, other);
3079 argv_array_push(&argv, xfrm_msg);
3083 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3084 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3086 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3087 die(_("external diff died, stopping at %s"), name);
3089 remove_tempfile();
3090 argv_array_clear(&argv);
3091 argv_array_clear(&env);
3094 static int similarity_index(struct diff_filepair *p)
3096 return p->score * 100 / MAX_SCORE;
3099 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3101 if (startup_info->have_repository)
3102 return find_unique_abbrev(oid->hash, abbrev);
3103 else {
3104 char *hex = oid_to_hex(oid);
3105 if (abbrev < 0)
3106 abbrev = FALLBACK_DEFAULT_ABBREV;
3107 if (abbrev > GIT_SHA1_HEXSZ)
3108 die("BUG: oid abbreviation out of range: %d", abbrev);
3109 hex[abbrev] = '\0';
3110 return hex;
3114 static void fill_metainfo(struct strbuf *msg,
3115 const char *name,
3116 const char *other,
3117 struct diff_filespec *one,
3118 struct diff_filespec *two,
3119 struct diff_options *o,
3120 struct diff_filepair *p,
3121 int *must_show_header,
3122 int use_color)
3124 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3125 const char *reset = diff_get_color(use_color, DIFF_RESET);
3126 const char *line_prefix = diff_line_prefix(o);
3128 *must_show_header = 1;
3129 strbuf_init(msg, PATH_MAX * 2 + 300);
3130 switch (p->status) {
3131 case DIFF_STATUS_COPIED:
3132 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3133 line_prefix, set, similarity_index(p));
3134 strbuf_addf(msg, "%s\n%s%scopy from ",
3135 reset, line_prefix, set);
3136 quote_c_style(name, msg, NULL, 0);
3137 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3138 quote_c_style(other, msg, NULL, 0);
3139 strbuf_addf(msg, "%s\n", reset);
3140 break;
3141 case DIFF_STATUS_RENAMED:
3142 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3143 line_prefix, set, similarity_index(p));
3144 strbuf_addf(msg, "%s\n%s%srename from ",
3145 reset, line_prefix, set);
3146 quote_c_style(name, msg, NULL, 0);
3147 strbuf_addf(msg, "%s\n%s%srename to ",
3148 reset, line_prefix, set);
3149 quote_c_style(other, msg, NULL, 0);
3150 strbuf_addf(msg, "%s\n", reset);
3151 break;
3152 case DIFF_STATUS_MODIFIED:
3153 if (p->score) {
3154 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3155 line_prefix,
3156 set, similarity_index(p), reset);
3157 break;
3159 /* fallthru */
3160 default:
3161 *must_show_header = 0;
3163 if (one && two && oidcmp(&one->oid, &two->oid)) {
3164 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3166 if (DIFF_OPT_TST(o, BINARY)) {
3167 mmfile_t mf;
3168 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3169 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3170 abbrev = 40;
3172 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3173 diff_abbrev_oid(&one->oid, abbrev),
3174 diff_abbrev_oid(&two->oid, abbrev));
3175 if (one->mode == two->mode)
3176 strbuf_addf(msg, " %06o", one->mode);
3177 strbuf_addf(msg, "%s\n", reset);
3181 static void run_diff_cmd(const char *pgm,
3182 const char *name,
3183 const char *other,
3184 const char *attr_path,
3185 struct diff_filespec *one,
3186 struct diff_filespec *two,
3187 struct strbuf *msg,
3188 struct diff_options *o,
3189 struct diff_filepair *p)
3191 const char *xfrm_msg = NULL;
3192 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3193 int must_show_header = 0;
3196 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3197 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3198 if (drv && drv->external)
3199 pgm = drv->external;
3202 if (msg) {
3204 * don't use colors when the header is intended for an
3205 * external diff driver
3207 fill_metainfo(msg, name, other, one, two, o, p,
3208 &must_show_header,
3209 want_color(o->use_color) && !pgm);
3210 xfrm_msg = msg->len ? msg->buf : NULL;
3213 if (pgm) {
3214 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3215 complete_rewrite, o);
3216 return;
3218 if (one && two)
3219 builtin_diff(name, other ? other : name,
3220 one, two, xfrm_msg, must_show_header,
3221 o, complete_rewrite);
3222 else
3223 fprintf(o->file, "* Unmerged path %s\n", name);
3226 static void diff_fill_sha1_info(struct diff_filespec *one)
3228 if (DIFF_FILE_VALID(one)) {
3229 if (!one->oid_valid) {
3230 struct stat st;
3231 if (one->is_stdin) {
3232 oidclr(&one->oid);
3233 return;
3235 if (lstat(one->path, &st) < 0)
3236 die_errno("stat '%s'", one->path);
3237 if (index_path(one->oid.hash, one->path, &st, 0))
3238 die("cannot hash %s", one->path);
3241 else
3242 oidclr(&one->oid);
3245 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3247 /* Strip the prefix but do not molest /dev/null and absolute paths */
3248 if (*namep && **namep != '/') {
3249 *namep += prefix_length;
3250 if (**namep == '/')
3251 ++*namep;
3253 if (*otherp && **otherp != '/') {
3254 *otherp += prefix_length;
3255 if (**otherp == '/')
3256 ++*otherp;
3260 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3262 const char *pgm = external_diff();
3263 struct strbuf msg;
3264 struct diff_filespec *one = p->one;
3265 struct diff_filespec *two = p->two;
3266 const char *name;
3267 const char *other;
3268 const char *attr_path;
3270 name = p->one->path;
3271 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3272 attr_path = name;
3273 if (o->prefix_length)
3274 strip_prefix(o->prefix_length, &name, &other);
3276 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3277 pgm = NULL;
3279 if (DIFF_PAIR_UNMERGED(p)) {
3280 run_diff_cmd(pgm, name, NULL, attr_path,
3281 NULL, NULL, NULL, o, p);
3282 return;
3285 diff_fill_sha1_info(one);
3286 diff_fill_sha1_info(two);
3288 if (!pgm &&
3289 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3290 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3292 * a filepair that changes between file and symlink
3293 * needs to be split into deletion and creation.
3295 struct diff_filespec *null = alloc_filespec(two->path);
3296 run_diff_cmd(NULL, name, other, attr_path,
3297 one, null, &msg, o, p);
3298 free(null);
3299 strbuf_release(&msg);
3301 null = alloc_filespec(one->path);
3302 run_diff_cmd(NULL, name, other, attr_path,
3303 null, two, &msg, o, p);
3304 free(null);
3306 else
3307 run_diff_cmd(pgm, name, other, attr_path,
3308 one, two, &msg, o, p);
3310 strbuf_release(&msg);
3313 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3314 struct diffstat_t *diffstat)
3316 const char *name;
3317 const char *other;
3319 if (DIFF_PAIR_UNMERGED(p)) {
3320 /* unmerged */
3321 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
3322 return;
3325 name = p->one->path;
3326 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3328 if (o->prefix_length)
3329 strip_prefix(o->prefix_length, &name, &other);
3331 diff_fill_sha1_info(p->one);
3332 diff_fill_sha1_info(p->two);
3334 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
3337 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3339 const char *name;
3340 const char *other;
3341 const char *attr_path;
3343 if (DIFF_PAIR_UNMERGED(p)) {
3344 /* unmerged */
3345 return;
3348 name = p->one->path;
3349 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3350 attr_path = other ? other : name;
3352 if (o->prefix_length)
3353 strip_prefix(o->prefix_length, &name, &other);
3355 diff_fill_sha1_info(p->one);
3356 diff_fill_sha1_info(p->two);
3358 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3361 void diff_setup(struct diff_options *options)
3363 memcpy(options, &default_diff_options, sizeof(*options));
3365 options->file = stdout;
3367 options->line_termination = '\n';
3368 options->break_opt = -1;
3369 options->rename_limit = -1;
3370 options->dirstat_permille = diff_dirstat_permille_default;
3371 options->context = diff_context_default;
3372 options->ws_error_highlight = ws_error_highlight_default;
3373 DIFF_OPT_SET(options, RENAME_EMPTY);
3375 /* pathchange left =NULL by default */
3376 options->change = diff_change;
3377 options->add_remove = diff_addremove;
3378 options->use_color = diff_use_color_default;
3379 options->detect_rename = diff_detect_rename_default;
3380 options->xdl_opts |= diff_algorithm;
3381 if (diff_indent_heuristic)
3382 DIFF_XDL_SET(options, INDENT_HEURISTIC);
3383 else if (diff_compaction_heuristic)
3384 DIFF_XDL_SET(options, COMPACTION_HEURISTIC);
3386 options->orderfile = diff_order_file_cfg;
3388 if (diff_no_prefix) {
3389 options->a_prefix = options->b_prefix = "";
3390 } else if (!diff_mnemonic_prefix) {
3391 options->a_prefix = "a/";
3392 options->b_prefix = "b/";
3396 void diff_setup_done(struct diff_options *options)
3398 int count = 0;
3400 if (options->set_default)
3401 options->set_default(options);
3403 if (options->output_format & DIFF_FORMAT_NAME)
3404 count++;
3405 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3406 count++;
3407 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3408 count++;
3409 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3410 count++;
3411 if (count > 1)
3412 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
3415 * Most of the time we can say "there are changes"
3416 * only by checking if there are changed paths, but
3417 * --ignore-whitespace* options force us to look
3418 * inside contents.
3421 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3422 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3423 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3424 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3425 else
3426 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3428 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3429 options->detect_rename = DIFF_DETECT_COPY;
3431 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3432 options->prefix = NULL;
3433 if (options->prefix)
3434 options->prefix_length = strlen(options->prefix);
3435 else
3436 options->prefix_length = 0;
3438 if (options->output_format & (DIFF_FORMAT_NAME |
3439 DIFF_FORMAT_NAME_STATUS |
3440 DIFF_FORMAT_CHECKDIFF |
3441 DIFF_FORMAT_NO_OUTPUT))
3442 options->output_format &= ~(DIFF_FORMAT_RAW |
3443 DIFF_FORMAT_NUMSTAT |
3444 DIFF_FORMAT_DIFFSTAT |
3445 DIFF_FORMAT_SHORTSTAT |
3446 DIFF_FORMAT_DIRSTAT |
3447 DIFF_FORMAT_SUMMARY |
3448 DIFF_FORMAT_PATCH);
3451 * These cases always need recursive; we do not drop caller-supplied
3452 * recursive bits for other formats here.
3454 if (options->output_format & (DIFF_FORMAT_PATCH |
3455 DIFF_FORMAT_NUMSTAT |
3456 DIFF_FORMAT_DIFFSTAT |
3457 DIFF_FORMAT_SHORTSTAT |
3458 DIFF_FORMAT_DIRSTAT |
3459 DIFF_FORMAT_SUMMARY |
3460 DIFF_FORMAT_CHECKDIFF))
3461 DIFF_OPT_SET(options, RECURSIVE);
3463 * Also pickaxe would not work very well if you do not say recursive
3465 if (options->pickaxe)
3466 DIFF_OPT_SET(options, RECURSIVE);
3468 * When patches are generated, submodules diffed against the work tree
3469 * must be checked for dirtiness too so it can be shown in the output
3471 if (options->output_format & DIFF_FORMAT_PATCH)
3472 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3474 if (options->detect_rename && options->rename_limit < 0)
3475 options->rename_limit = diff_rename_limit_default;
3476 if (options->setup & DIFF_SETUP_USE_CACHE) {
3477 if (!active_cache)
3478 /* read-cache does not die even when it fails
3479 * so it is safe for us to do this here. Also
3480 * it does not smudge active_cache or active_nr
3481 * when it fails, so we do not have to worry about
3482 * cleaning it up ourselves either.
3484 read_cache();
3486 if (40 < options->abbrev)
3487 options->abbrev = 40; /* full */
3490 * It does not make sense to show the first hit we happened
3491 * to have found. It does not make sense not to return with
3492 * exit code in such a case either.
3494 if (DIFF_OPT_TST(options, QUICK)) {
3495 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3496 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3499 options->diff_path_counter = 0;
3501 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
3502 die(_("--follow requires exactly one pathspec"));
3505 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3507 char c, *eq;
3508 int len;
3510 if (*arg != '-')
3511 return 0;
3512 c = *++arg;
3513 if (!c)
3514 return 0;
3515 if (c == arg_short) {
3516 c = *++arg;
3517 if (!c)
3518 return 1;
3519 if (val && isdigit(c)) {
3520 char *end;
3521 int n = strtoul(arg, &end, 10);
3522 if (*end)
3523 return 0;
3524 *val = n;
3525 return 1;
3527 return 0;
3529 if (c != '-')
3530 return 0;
3531 arg++;
3532 eq = strchrnul(arg, '=');
3533 len = eq - arg;
3534 if (!len || strncmp(arg, arg_long, len))
3535 return 0;
3536 if (*eq) {
3537 int n;
3538 char *end;
3539 if (!isdigit(*++eq))
3540 return 0;
3541 n = strtoul(eq, &end, 10);
3542 if (*end)
3543 return 0;
3544 *val = n;
3546 return 1;
3549 static int diff_scoreopt_parse(const char *opt);
3551 static inline int short_opt(char opt, const char **argv,
3552 const char **optarg)
3554 const char *arg = argv[0];
3555 if (arg[0] != '-' || arg[1] != opt)
3556 return 0;
3557 if (arg[2] != '\0') {
3558 *optarg = arg + 2;
3559 return 1;
3561 if (!argv[1])
3562 die("Option '%c' requires a value", opt);
3563 *optarg = argv[1];
3564 return 2;
3567 int parse_long_opt(const char *opt, const char **argv,
3568 const char **optarg)
3570 const char *arg = argv[0];
3571 if (!skip_prefix(arg, "--", &arg))
3572 return 0;
3573 if (!skip_prefix(arg, opt, &arg))
3574 return 0;
3575 if (*arg == '=') { /* stuck form: --option=value */
3576 *optarg = arg + 1;
3577 return 1;
3579 if (*arg != '\0')
3580 return 0;
3581 /* separate form: --option value */
3582 if (!argv[1])
3583 die("Option '--%s' requires a value", opt);
3584 *optarg = argv[1];
3585 return 2;
3588 static int stat_opt(struct diff_options *options, const char **av)
3590 const char *arg = av[0];
3591 char *end;
3592 int width = options->stat_width;
3593 int name_width = options->stat_name_width;
3594 int graph_width = options->stat_graph_width;
3595 int count = options->stat_count;
3596 int argcount = 1;
3598 if (!skip_prefix(arg, "--stat", &arg))
3599 die("BUG: stat option does not begin with --stat: %s", arg);
3600 end = (char *)arg;
3602 switch (*arg) {
3603 case '-':
3604 if (skip_prefix(arg, "-width", &arg)) {
3605 if (*arg == '=')
3606 width = strtoul(arg + 1, &end, 10);
3607 else if (!*arg && !av[1])
3608 die_want_option("--stat-width");
3609 else if (!*arg) {
3610 width = strtoul(av[1], &end, 10);
3611 argcount = 2;
3613 } else if (skip_prefix(arg, "-name-width", &arg)) {
3614 if (*arg == '=')
3615 name_width = strtoul(arg + 1, &end, 10);
3616 else if (!*arg && !av[1])
3617 die_want_option("--stat-name-width");
3618 else if (!*arg) {
3619 name_width = strtoul(av[1], &end, 10);
3620 argcount = 2;
3622 } else if (skip_prefix(arg, "-graph-width", &arg)) {
3623 if (*arg == '=')
3624 graph_width = strtoul(arg + 1, &end, 10);
3625 else if (!*arg && !av[1])
3626 die_want_option("--stat-graph-width");
3627 else if (!*arg) {
3628 graph_width = strtoul(av[1], &end, 10);
3629 argcount = 2;
3631 } else if (skip_prefix(arg, "-count", &arg)) {
3632 if (*arg == '=')
3633 count = strtoul(arg + 1, &end, 10);
3634 else if (!*arg && !av[1])
3635 die_want_option("--stat-count");
3636 else if (!*arg) {
3637 count = strtoul(av[1], &end, 10);
3638 argcount = 2;
3641 break;
3642 case '=':
3643 width = strtoul(arg+1, &end, 10);
3644 if (*end == ',')
3645 name_width = strtoul(end+1, &end, 10);
3646 if (*end == ',')
3647 count = strtoul(end+1, &end, 10);
3650 /* Important! This checks all the error cases! */
3651 if (*end)
3652 return 0;
3653 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3654 options->stat_name_width = name_width;
3655 options->stat_graph_width = graph_width;
3656 options->stat_width = width;
3657 options->stat_count = count;
3658 return argcount;
3661 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3663 struct strbuf errmsg = STRBUF_INIT;
3664 if (parse_dirstat_params(options, params, &errmsg))
3665 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3666 errmsg.buf);
3667 strbuf_release(&errmsg);
3669 * The caller knows a dirstat-related option is given from the command
3670 * line; allow it to say "return this_function();"
3672 options->output_format |= DIFF_FORMAT_DIRSTAT;
3673 return 1;
3676 static int parse_submodule_opt(struct diff_options *options, const char *value)
3678 if (parse_submodule_params(options, value))
3679 die(_("Failed to parse --submodule option parameter: '%s'"),
3680 value);
3681 return 1;
3684 static const char diff_status_letters[] = {
3685 DIFF_STATUS_ADDED,
3686 DIFF_STATUS_COPIED,
3687 DIFF_STATUS_DELETED,
3688 DIFF_STATUS_MODIFIED,
3689 DIFF_STATUS_RENAMED,
3690 DIFF_STATUS_TYPE_CHANGED,
3691 DIFF_STATUS_UNKNOWN,
3692 DIFF_STATUS_UNMERGED,
3693 DIFF_STATUS_FILTER_AON,
3694 DIFF_STATUS_FILTER_BROKEN,
3695 '\0',
3698 static unsigned int filter_bit['Z' + 1];
3700 static void prepare_filter_bits(void)
3702 int i;
3704 if (!filter_bit[DIFF_STATUS_ADDED]) {
3705 for (i = 0; diff_status_letters[i]; i++)
3706 filter_bit[(int) diff_status_letters[i]] = (1 << i);
3710 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
3712 return opt->filter & filter_bit[(int) status];
3715 static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
3717 int i, optch;
3719 prepare_filter_bits();
3722 * If there is a negation e.g. 'd' in the input, and we haven't
3723 * initialized the filter field with another --diff-filter, start
3724 * from full set of bits, except for AON.
3726 if (!opt->filter) {
3727 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3728 if (optch < 'a' || 'z' < optch)
3729 continue;
3730 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
3731 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
3732 break;
3736 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
3737 unsigned int bit;
3738 int negate;
3740 if ('a' <= optch && optch <= 'z') {
3741 negate = 1;
3742 optch = toupper(optch);
3743 } else {
3744 negate = 0;
3747 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
3748 if (!bit)
3749 return optarg[i];
3750 if (negate)
3751 opt->filter &= ~bit;
3752 else
3753 opt->filter |= bit;
3755 return 0;
3758 static void enable_patch_output(int *fmt) {
3759 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
3760 *fmt |= DIFF_FORMAT_PATCH;
3763 static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
3765 int val = parse_ws_error_highlight(arg);
3767 if (val < 0) {
3768 error("unknown value after ws-error-highlight=%.*s",
3769 -1 - val, arg);
3770 return 0;
3772 opt->ws_error_highlight = val;
3773 return 1;
3776 int diff_opt_parse(struct diff_options *options,
3777 const char **av, int ac, const char *prefix)
3779 const char *arg = av[0];
3780 const char *optarg;
3781 int argcount;
3783 if (!prefix)
3784 prefix = "";
3786 /* Output format options */
3787 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
3788 || opt_arg(arg, 'U', "unified", &options->context))
3789 enable_patch_output(&options->output_format);
3790 else if (!strcmp(arg, "--raw"))
3791 options->output_format |= DIFF_FORMAT_RAW;
3792 else if (!strcmp(arg, "--patch-with-raw")) {
3793 enable_patch_output(&options->output_format);
3794 options->output_format |= DIFF_FORMAT_RAW;
3795 } else if (!strcmp(arg, "--numstat"))
3796 options->output_format |= DIFF_FORMAT_NUMSTAT;
3797 else if (!strcmp(arg, "--shortstat"))
3798 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3799 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3800 return parse_dirstat_opt(options, "");
3801 else if (skip_prefix(arg, "-X", &arg))
3802 return parse_dirstat_opt(options, arg);
3803 else if (skip_prefix(arg, "--dirstat=", &arg))
3804 return parse_dirstat_opt(options, arg);
3805 else if (!strcmp(arg, "--cumulative"))
3806 return parse_dirstat_opt(options, "cumulative");
3807 else if (!strcmp(arg, "--dirstat-by-file"))
3808 return parse_dirstat_opt(options, "files");
3809 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
3810 parse_dirstat_opt(options, "files");
3811 return parse_dirstat_opt(options, arg);
3813 else if (!strcmp(arg, "--check"))
3814 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3815 else if (!strcmp(arg, "--summary"))
3816 options->output_format |= DIFF_FORMAT_SUMMARY;
3817 else if (!strcmp(arg, "--patch-with-stat")) {
3818 enable_patch_output(&options->output_format);
3819 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3820 } else if (!strcmp(arg, "--name-only"))
3821 options->output_format |= DIFF_FORMAT_NAME;
3822 else if (!strcmp(arg, "--name-status"))
3823 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3824 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
3825 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3826 else if (starts_with(arg, "--stat"))
3827 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3828 return stat_opt(options, av);
3830 /* renames options */
3831 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
3832 !strcmp(arg, "--break-rewrites")) {
3833 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3834 return error("invalid argument to -B: %s", arg+2);
3836 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
3837 !strcmp(arg, "--find-renames")) {
3838 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3839 return error("invalid argument to -M: %s", arg+2);
3840 options->detect_rename = DIFF_DETECT_RENAME;
3842 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3843 options->irreversible_delete = 1;
3845 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
3846 !strcmp(arg, "--find-copies")) {
3847 if (options->detect_rename == DIFF_DETECT_COPY)
3848 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3849 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3850 return error("invalid argument to -C: %s", arg+2);
3851 options->detect_rename = DIFF_DETECT_COPY;
3853 else if (!strcmp(arg, "--no-renames"))
3854 options->detect_rename = 0;
3855 else if (!strcmp(arg, "--rename-empty"))
3856 DIFF_OPT_SET(options, RENAME_EMPTY);
3857 else if (!strcmp(arg, "--no-rename-empty"))
3858 DIFF_OPT_CLR(options, RENAME_EMPTY);
3859 else if (!strcmp(arg, "--relative"))
3860 DIFF_OPT_SET(options, RELATIVE_NAME);
3861 else if (skip_prefix(arg, "--relative=", &arg)) {
3862 DIFF_OPT_SET(options, RELATIVE_NAME);
3863 options->prefix = arg;
3866 /* xdiff options */
3867 else if (!strcmp(arg, "--minimal"))
3868 DIFF_XDL_SET(options, NEED_MINIMAL);
3869 else if (!strcmp(arg, "--no-minimal"))
3870 DIFF_XDL_CLR(options, NEED_MINIMAL);
3871 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3872 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3873 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3874 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3875 else if (!strcmp(arg, "--ignore-space-at-eol"))
3876 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3877 else if (!strcmp(arg, "--ignore-blank-lines"))
3878 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
3879 else if (!strcmp(arg, "--indent-heuristic")) {
3880 DIFF_XDL_SET(options, INDENT_HEURISTIC);
3881 DIFF_XDL_CLR(options, COMPACTION_HEURISTIC);
3882 } else if (!strcmp(arg, "--no-indent-heuristic"))
3883 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
3884 else if (!strcmp(arg, "--compaction-heuristic")) {
3885 DIFF_XDL_SET(options, COMPACTION_HEURISTIC);
3886 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
3887 } else if (!strcmp(arg, "--no-compaction-heuristic"))
3888 DIFF_XDL_CLR(options, COMPACTION_HEURISTIC);
3889 else if (!strcmp(arg, "--patience"))
3890 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3891 else if (!strcmp(arg, "--histogram"))
3892 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3893 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
3894 long value = parse_algorithm_value(optarg);
3895 if (value < 0)
3896 return error("option diff-algorithm accepts \"myers\", "
3897 "\"minimal\", \"patience\" and \"histogram\"");
3898 /* clear out previous settings */
3899 DIFF_XDL_CLR(options, NEED_MINIMAL);
3900 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
3901 options->xdl_opts |= value;
3902 return argcount;
3905 /* flags options */
3906 else if (!strcmp(arg, "--binary")) {
3907 enable_patch_output(&options->output_format);
3908 DIFF_OPT_SET(options, BINARY);
3910 else if (!strcmp(arg, "--full-index"))
3911 DIFF_OPT_SET(options, FULL_INDEX);
3912 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3913 DIFF_OPT_SET(options, TEXT);
3914 else if (!strcmp(arg, "-R"))
3915 DIFF_OPT_SET(options, REVERSE_DIFF);
3916 else if (!strcmp(arg, "--find-copies-harder"))
3917 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3918 else if (!strcmp(arg, "--follow"))
3919 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3920 else if (!strcmp(arg, "--no-follow")) {
3921 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
3922 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
3923 } else if (!strcmp(arg, "--color"))
3924 options->use_color = 1;
3925 else if (skip_prefix(arg, "--color=", &arg)) {
3926 int value = git_config_colorbool(NULL, arg);
3927 if (value < 0)
3928 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3929 options->use_color = value;
3931 else if (!strcmp(arg, "--no-color"))
3932 options->use_color = 0;
3933 else if (!strcmp(arg, "--color-words")) {
3934 options->use_color = 1;
3935 options->word_diff = DIFF_WORDS_COLOR;
3937 else if (skip_prefix(arg, "--color-words=", &arg)) {
3938 options->use_color = 1;
3939 options->word_diff = DIFF_WORDS_COLOR;
3940 options->word_regex = arg;
3942 else if (!strcmp(arg, "--word-diff")) {
3943 if (options->word_diff == DIFF_WORDS_NONE)
3944 options->word_diff = DIFF_WORDS_PLAIN;
3946 else if (skip_prefix(arg, "--word-diff=", &arg)) {
3947 if (!strcmp(arg, "plain"))
3948 options->word_diff = DIFF_WORDS_PLAIN;
3949 else if (!strcmp(arg, "color")) {
3950 options->use_color = 1;
3951 options->word_diff = DIFF_WORDS_COLOR;
3953 else if (!strcmp(arg, "porcelain"))
3954 options->word_diff = DIFF_WORDS_PORCELAIN;
3955 else if (!strcmp(arg, "none"))
3956 options->word_diff = DIFF_WORDS_NONE;
3957 else
3958 die("bad --word-diff argument: %s", arg);
3960 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3961 if (options->word_diff == DIFF_WORDS_NONE)
3962 options->word_diff = DIFF_WORDS_PLAIN;
3963 options->word_regex = optarg;
3964 return argcount;
3966 else if (!strcmp(arg, "--exit-code"))
3967 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3968 else if (!strcmp(arg, "--quiet"))
3969 DIFF_OPT_SET(options, QUICK);
3970 else if (!strcmp(arg, "--ext-diff"))
3971 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3972 else if (!strcmp(arg, "--no-ext-diff"))
3973 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3974 else if (!strcmp(arg, "--textconv"))
3975 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3976 else if (!strcmp(arg, "--no-textconv"))
3977 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3978 else if (!strcmp(arg, "--ignore-submodules")) {
3979 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3980 handle_ignore_submodules_arg(options, "all");
3981 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
3982 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3983 handle_ignore_submodules_arg(options, arg);
3984 } else if (!strcmp(arg, "--submodule"))
3985 options->submodule_format = DIFF_SUBMODULE_LOG;
3986 else if (skip_prefix(arg, "--submodule=", &arg))
3987 return parse_submodule_opt(options, arg);
3988 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
3989 return parse_ws_error_highlight_opt(options, arg);
3990 else if (!strcmp(arg, "--ita-invisible-in-index"))
3991 options->ita_invisible_in_index = 1;
3992 else if (!strcmp(arg, "--ita-visible-in-index"))
3993 options->ita_invisible_in_index = 0;
3995 /* misc options */
3996 else if (!strcmp(arg, "-z"))
3997 options->line_termination = 0;
3998 else if ((argcount = short_opt('l', av, &optarg))) {
3999 options->rename_limit = strtoul(optarg, NULL, 10);
4000 return argcount;
4002 else if ((argcount = short_opt('S', av, &optarg))) {
4003 options->pickaxe = optarg;
4004 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4005 return argcount;
4006 } else if ((argcount = short_opt('G', av, &optarg))) {
4007 options->pickaxe = optarg;
4008 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4009 return argcount;
4011 else if (!strcmp(arg, "--pickaxe-all"))
4012 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4013 else if (!strcmp(arg, "--pickaxe-regex"))
4014 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4015 else if ((argcount = short_opt('O', av, &optarg))) {
4016 const char *path = prefix_filename(prefix, strlen(prefix), optarg);
4017 options->orderfile = xstrdup(path);
4018 return argcount;
4020 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4021 int offending = parse_diff_filter_opt(optarg, options);
4022 if (offending)
4023 die("unknown change class '%c' in --diff-filter=%s",
4024 offending, optarg);
4025 return argcount;
4027 else if (!strcmp(arg, "--abbrev"))
4028 options->abbrev = DEFAULT_ABBREV;
4029 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4030 options->abbrev = strtoul(arg, NULL, 10);
4031 if (options->abbrev < MINIMUM_ABBREV)
4032 options->abbrev = MINIMUM_ABBREV;
4033 else if (40 < options->abbrev)
4034 options->abbrev = 40;
4036 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4037 options->a_prefix = optarg;
4038 return argcount;
4040 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4041 options->line_prefix = optarg;
4042 options->line_prefix_length = strlen(options->line_prefix);
4043 graph_setup_line_prefix(options);
4044 return argcount;
4046 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4047 options->b_prefix = optarg;
4048 return argcount;
4050 else if (!strcmp(arg, "--no-prefix"))
4051 options->a_prefix = options->b_prefix = "";
4052 else if (opt_arg(arg, '\0', "inter-hunk-context",
4053 &options->interhunkcontext))
4055 else if (!strcmp(arg, "-W"))
4056 DIFF_OPT_SET(options, FUNCCONTEXT);
4057 else if (!strcmp(arg, "--function-context"))
4058 DIFF_OPT_SET(options, FUNCCONTEXT);
4059 else if (!strcmp(arg, "--no-function-context"))
4060 DIFF_OPT_CLR(options, FUNCCONTEXT);
4061 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4062 const char *path = prefix_filename(prefix, strlen(prefix), optarg);
4063 options->file = fopen(path, "w");
4064 if (!options->file)
4065 die_errno("Could not open '%s'", path);
4066 options->close_file = 1;
4067 if (options->use_color != GIT_COLOR_ALWAYS)
4068 options->use_color = GIT_COLOR_NEVER;
4069 return argcount;
4070 } else
4071 return 0;
4072 return 1;
4075 int parse_rename_score(const char **cp_p)
4077 unsigned long num, scale;
4078 int ch, dot;
4079 const char *cp = *cp_p;
4081 num = 0;
4082 scale = 1;
4083 dot = 0;
4084 for (;;) {
4085 ch = *cp;
4086 if ( !dot && ch == '.' ) {
4087 scale = 1;
4088 dot = 1;
4089 } else if ( ch == '%' ) {
4090 scale = dot ? scale*100 : 100;
4091 cp++; /* % is always at the end */
4092 break;
4093 } else if ( ch >= '0' && ch <= '9' ) {
4094 if ( scale < 100000 ) {
4095 scale *= 10;
4096 num = (num*10) + (ch-'0');
4098 } else {
4099 break;
4101 cp++;
4103 *cp_p = cp;
4105 /* user says num divided by scale and we say internally that
4106 * is MAX_SCORE * num / scale.
4108 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4111 static int diff_scoreopt_parse(const char *opt)
4113 int opt1, opt2, cmd;
4115 if (*opt++ != '-')
4116 return -1;
4117 cmd = *opt++;
4118 if (cmd == '-') {
4119 /* convert the long-form arguments into short-form versions */
4120 if (skip_prefix(opt, "break-rewrites", &opt)) {
4121 if (*opt == 0 || *opt++ == '=')
4122 cmd = 'B';
4123 } else if (skip_prefix(opt, "find-copies", &opt)) {
4124 if (*opt == 0 || *opt++ == '=')
4125 cmd = 'C';
4126 } else if (skip_prefix(opt, "find-renames", &opt)) {
4127 if (*opt == 0 || *opt++ == '=')
4128 cmd = 'M';
4131 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4132 return -1; /* that is not a -M, -C, or -B option */
4134 opt1 = parse_rename_score(&opt);
4135 if (cmd != 'B')
4136 opt2 = 0;
4137 else {
4138 if (*opt == 0)
4139 opt2 = 0;
4140 else if (*opt != '/')
4141 return -1; /* we expect -B80/99 or -B80 */
4142 else {
4143 opt++;
4144 opt2 = parse_rename_score(&opt);
4147 if (*opt != 0)
4148 return -1;
4149 return opt1 | (opt2 << 16);
4152 struct diff_queue_struct diff_queued_diff;
4154 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4156 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4157 queue->queue[queue->nr++] = dp;
4160 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4161 struct diff_filespec *one,
4162 struct diff_filespec *two)
4164 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4165 dp->one = one;
4166 dp->two = two;
4167 if (queue)
4168 diff_q(queue, dp);
4169 return dp;
4172 void diff_free_filepair(struct diff_filepair *p)
4174 free_filespec(p->one);
4175 free_filespec(p->two);
4176 free(p);
4179 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4181 int abblen;
4182 const char *abbrev;
4184 if (len == GIT_SHA1_HEXSZ)
4185 return oid_to_hex(oid);
4187 abbrev = diff_abbrev_oid(oid, len);
4188 abblen = strlen(abbrev);
4191 * In well-behaved cases, where the abbbreviated result is the
4192 * same as the requested length, append three dots after the
4193 * abbreviation (hence the whole logic is limited to the case
4194 * where abblen < 37); when the actual abbreviated result is a
4195 * bit longer than the requested length, we reduce the number
4196 * of dots so that they match the well-behaved ones. However,
4197 * if the actual abbreviation is longer than the requested
4198 * length by more than three, we give up on aligning, and add
4199 * three dots anyway, to indicate that the output is not the
4200 * full object name. Yes, this may be suboptimal, but this
4201 * appears only in "diff --raw --abbrev" output and it is not
4202 * worth the effort to change it now. Note that this would
4203 * likely to work fine when the automatic sizing of default
4204 * abbreviation length is used--we would be fed -1 in "len" in
4205 * that case, and will end up always appending three-dots, but
4206 * the automatic sizing is supposed to give abblen that ensures
4207 * uniqueness across all objects (statistically speaking).
4209 if (abblen < GIT_SHA1_HEXSZ - 3) {
4210 static char hex[GIT_SHA1_HEXSZ + 1];
4211 if (len < abblen && abblen <= len + 2)
4212 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
4213 else
4214 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
4215 return hex;
4218 return oid_to_hex(oid);
4221 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4223 int line_termination = opt->line_termination;
4224 int inter_name_termination = line_termination ? '\t' : '\0';
4226 fprintf(opt->file, "%s", diff_line_prefix(opt));
4227 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4228 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4229 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
4230 fprintf(opt->file, "%s ",
4231 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
4233 if (p->score) {
4234 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4235 inter_name_termination);
4236 } else {
4237 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
4240 if (p->status == DIFF_STATUS_COPIED ||
4241 p->status == DIFF_STATUS_RENAMED) {
4242 const char *name_a, *name_b;
4243 name_a = p->one->path;
4244 name_b = p->two->path;
4245 strip_prefix(opt->prefix_length, &name_a, &name_b);
4246 write_name_quoted(name_a, opt->file, inter_name_termination);
4247 write_name_quoted(name_b, opt->file, line_termination);
4248 } else {
4249 const char *name_a, *name_b;
4250 name_a = p->one->mode ? p->one->path : p->two->path;
4251 name_b = NULL;
4252 strip_prefix(opt->prefix_length, &name_a, &name_b);
4253 write_name_quoted(name_a, opt->file, line_termination);
4257 int diff_unmodified_pair(struct diff_filepair *p)
4259 /* This function is written stricter than necessary to support
4260 * the currently implemented transformers, but the idea is to
4261 * let transformers to produce diff_filepairs any way they want,
4262 * and filter and clean them up here before producing the output.
4264 struct diff_filespec *one = p->one, *two = p->two;
4266 if (DIFF_PAIR_UNMERGED(p))
4267 return 0; /* unmerged is interesting */
4269 /* deletion, addition, mode or type change
4270 * and rename are all interesting.
4272 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
4273 DIFF_PAIR_MODE_CHANGED(p) ||
4274 strcmp(one->path, two->path))
4275 return 0;
4277 /* both are valid and point at the same path. that is, we are
4278 * dealing with a change.
4280 if (one->oid_valid && two->oid_valid &&
4281 !oidcmp(&one->oid, &two->oid) &&
4282 !one->dirty_submodule && !two->dirty_submodule)
4283 return 1; /* no change */
4284 if (!one->oid_valid && !two->oid_valid)
4285 return 1; /* both look at the same file on the filesystem. */
4286 return 0;
4289 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
4291 if (diff_unmodified_pair(p))
4292 return;
4294 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4295 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4296 return; /* no tree diffs in patch format */
4298 run_diff(p, o);
4301 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
4302 struct diffstat_t *diffstat)
4304 if (diff_unmodified_pair(p))
4305 return;
4307 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4308 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4309 return; /* no useful stat for tree diffs */
4311 run_diffstat(p, o, diffstat);
4314 static void diff_flush_checkdiff(struct diff_filepair *p,
4315 struct diff_options *o)
4317 if (diff_unmodified_pair(p))
4318 return;
4320 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4321 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4322 return; /* nothing to check in tree diffs */
4324 run_checkdiff(p, o);
4327 int diff_queue_is_empty(void)
4329 struct diff_queue_struct *q = &diff_queued_diff;
4330 int i;
4331 for (i = 0; i < q->nr; i++)
4332 if (!diff_unmodified_pair(q->queue[i]))
4333 return 0;
4334 return 1;
4337 #if DIFF_DEBUG
4338 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
4340 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
4341 x, one ? one : "",
4342 s->path,
4343 DIFF_FILE_VALID(s) ? "valid" : "invalid",
4344 s->mode,
4345 s->oid_valid ? oid_to_hex(&s->oid) : "");
4346 fprintf(stderr, "queue[%d] %s size %lu\n",
4347 x, one ? one : "",
4348 s->size);
4351 void diff_debug_filepair(const struct diff_filepair *p, int i)
4353 diff_debug_filespec(p->one, i, "one");
4354 diff_debug_filespec(p->two, i, "two");
4355 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
4356 p->score, p->status ? p->status : '?',
4357 p->one->rename_used, p->broken_pair);
4360 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4362 int i;
4363 if (msg)
4364 fprintf(stderr, "%s\n", msg);
4365 fprintf(stderr, "q->nr = %d\n", q->nr);
4366 for (i = 0; i < q->nr; i++) {
4367 struct diff_filepair *p = q->queue[i];
4368 diff_debug_filepair(p, i);
4371 #endif
4373 static void diff_resolve_rename_copy(void)
4375 int i;
4376 struct diff_filepair *p;
4377 struct diff_queue_struct *q = &diff_queued_diff;
4379 diff_debug_queue("resolve-rename-copy", q);
4381 for (i = 0; i < q->nr; i++) {
4382 p = q->queue[i];
4383 p->status = 0; /* undecided */
4384 if (DIFF_PAIR_UNMERGED(p))
4385 p->status = DIFF_STATUS_UNMERGED;
4386 else if (!DIFF_FILE_VALID(p->one))
4387 p->status = DIFF_STATUS_ADDED;
4388 else if (!DIFF_FILE_VALID(p->two))
4389 p->status = DIFF_STATUS_DELETED;
4390 else if (DIFF_PAIR_TYPE_CHANGED(p))
4391 p->status = DIFF_STATUS_TYPE_CHANGED;
4393 /* from this point on, we are dealing with a pair
4394 * whose both sides are valid and of the same type, i.e.
4395 * either in-place edit or rename/copy edit.
4397 else if (DIFF_PAIR_RENAME(p)) {
4399 * A rename might have re-connected a broken
4400 * pair up, causing the pathnames to be the
4401 * same again. If so, that's not a rename at
4402 * all, just a modification..
4404 * Otherwise, see if this source was used for
4405 * multiple renames, in which case we decrement
4406 * the count, and call it a copy.
4408 if (!strcmp(p->one->path, p->two->path))
4409 p->status = DIFF_STATUS_MODIFIED;
4410 else if (--p->one->rename_used > 0)
4411 p->status = DIFF_STATUS_COPIED;
4412 else
4413 p->status = DIFF_STATUS_RENAMED;
4415 else if (oidcmp(&p->one->oid, &p->two->oid) ||
4416 p->one->mode != p->two->mode ||
4417 p->one->dirty_submodule ||
4418 p->two->dirty_submodule ||
4419 is_null_oid(&p->one->oid))
4420 p->status = DIFF_STATUS_MODIFIED;
4421 else {
4422 /* This is a "no-change" entry and should not
4423 * happen anymore, but prepare for broken callers.
4425 error("feeding unmodified %s to diffcore",
4426 p->one->path);
4427 p->status = DIFF_STATUS_UNKNOWN;
4430 diff_debug_queue("resolve-rename-copy done", q);
4433 static int check_pair_status(struct diff_filepair *p)
4435 switch (p->status) {
4436 case DIFF_STATUS_UNKNOWN:
4437 return 0;
4438 case 0:
4439 die("internal error in diff-resolve-rename-copy");
4440 default:
4441 return 1;
4445 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4447 int fmt = opt->output_format;
4449 if (fmt & DIFF_FORMAT_CHECKDIFF)
4450 diff_flush_checkdiff(p, opt);
4451 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4452 diff_flush_raw(p, opt);
4453 else if (fmt & DIFF_FORMAT_NAME) {
4454 const char *name_a, *name_b;
4455 name_a = p->two->path;
4456 name_b = NULL;
4457 strip_prefix(opt->prefix_length, &name_a, &name_b);
4458 write_name_quoted(name_a, opt->file, opt->line_termination);
4462 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4464 if (fs->mode)
4465 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4466 else
4467 fprintf(file, " %s ", newdelete);
4468 write_name_quoted(fs->path, file, '\n');
4472 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4473 const char *line_prefix)
4475 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4476 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4477 p->two->mode, show_name ? ' ' : '\n');
4478 if (show_name) {
4479 write_name_quoted(p->two->path, file, '\n');
4484 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4485 const char *line_prefix)
4487 char *names = pprint_rename(p->one->path, p->two->path);
4489 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4490 free(names);
4491 show_mode_change(file, p, 0, line_prefix);
4494 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4496 FILE *file = opt->file;
4497 const char *line_prefix = diff_line_prefix(opt);
4499 switch(p->status) {
4500 case DIFF_STATUS_DELETED:
4501 fputs(line_prefix, file);
4502 show_file_mode_name(file, "delete", p->one);
4503 break;
4504 case DIFF_STATUS_ADDED:
4505 fputs(line_prefix, file);
4506 show_file_mode_name(file, "create", p->two);
4507 break;
4508 case DIFF_STATUS_COPIED:
4509 fputs(line_prefix, file);
4510 show_rename_copy(file, "copy", p, line_prefix);
4511 break;
4512 case DIFF_STATUS_RENAMED:
4513 fputs(line_prefix, file);
4514 show_rename_copy(file, "rename", p, line_prefix);
4515 break;
4516 default:
4517 if (p->score) {
4518 fprintf(file, "%s rewrite ", line_prefix);
4519 write_name_quoted(p->two->path, file, ' ');
4520 fprintf(file, "(%d%%)\n", similarity_index(p));
4522 show_mode_change(file, p, !p->score, line_prefix);
4523 break;
4527 struct patch_id_t {
4528 git_SHA_CTX *ctx;
4529 int patchlen;
4532 static int remove_space(char *line, int len)
4534 int i;
4535 char *dst = line;
4536 unsigned char c;
4538 for (i = 0; i < len; i++)
4539 if (!isspace((c = line[i])))
4540 *dst++ = c;
4542 return dst - line;
4545 static void patch_id_consume(void *priv, char *line, unsigned long len)
4547 struct patch_id_t *data = priv;
4548 int new_len;
4550 /* Ignore line numbers when computing the SHA1 of the patch */
4551 if (starts_with(line, "@@ -"))
4552 return;
4554 new_len = remove_space(line, len);
4556 git_SHA1_Update(data->ctx, line, new_len);
4557 data->patchlen += new_len;
4560 /* returns 0 upon success, and writes result into sha1 */
4561 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1, int diff_header_only)
4563 struct diff_queue_struct *q = &diff_queued_diff;
4564 int i;
4565 git_SHA_CTX ctx;
4566 struct patch_id_t data;
4567 char buffer[PATH_MAX * 4 + 20];
4569 git_SHA1_Init(&ctx);
4570 memset(&data, 0, sizeof(struct patch_id_t));
4571 data.ctx = &ctx;
4573 for (i = 0; i < q->nr; i++) {
4574 xpparam_t xpp;
4575 xdemitconf_t xecfg;
4576 mmfile_t mf1, mf2;
4577 struct diff_filepair *p = q->queue[i];
4578 int len1, len2;
4580 memset(&xpp, 0, sizeof(xpp));
4581 memset(&xecfg, 0, sizeof(xecfg));
4582 if (p->status == 0)
4583 return error("internal diff status error");
4584 if (p->status == DIFF_STATUS_UNKNOWN)
4585 continue;
4586 if (diff_unmodified_pair(p))
4587 continue;
4588 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4589 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4590 continue;
4591 if (DIFF_PAIR_UNMERGED(p))
4592 continue;
4594 diff_fill_sha1_info(p->one);
4595 diff_fill_sha1_info(p->two);
4597 len1 = remove_space(p->one->path, strlen(p->one->path));
4598 len2 = remove_space(p->two->path, strlen(p->two->path));
4599 if (p->one->mode == 0)
4600 len1 = snprintf(buffer, sizeof(buffer),
4601 "diff--gita/%.*sb/%.*s"
4602 "newfilemode%06o"
4603 "---/dev/null"
4604 "+++b/%.*s",
4605 len1, p->one->path,
4606 len2, p->two->path,
4607 p->two->mode,
4608 len2, p->two->path);
4609 else if (p->two->mode == 0)
4610 len1 = snprintf(buffer, sizeof(buffer),
4611 "diff--gita/%.*sb/%.*s"
4612 "deletedfilemode%06o"
4613 "---a/%.*s"
4614 "+++/dev/null",
4615 len1, p->one->path,
4616 len2, p->two->path,
4617 p->one->mode,
4618 len1, p->one->path);
4619 else
4620 len1 = snprintf(buffer, sizeof(buffer),
4621 "diff--gita/%.*sb/%.*s"
4622 "---a/%.*s"
4623 "+++b/%.*s",
4624 len1, p->one->path,
4625 len2, p->two->path,
4626 len1, p->one->path,
4627 len2, p->two->path);
4628 git_SHA1_Update(&ctx, buffer, len1);
4630 if (diff_header_only)
4631 continue;
4633 if (fill_mmfile(&mf1, p->one) < 0 ||
4634 fill_mmfile(&mf2, p->two) < 0)
4635 return error("unable to read files to diff");
4637 if (diff_filespec_is_binary(p->one) ||
4638 diff_filespec_is_binary(p->two)) {
4639 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
4640 40);
4641 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
4642 40);
4643 continue;
4646 xpp.flags = 0;
4647 xecfg.ctxlen = 3;
4648 xecfg.flags = 0;
4649 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4650 &xpp, &xecfg))
4651 return error("unable to generate patch-id diff for %s",
4652 p->one->path);
4655 git_SHA1_Final(sha1, &ctx);
4656 return 0;
4659 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1, int diff_header_only)
4661 struct diff_queue_struct *q = &diff_queued_diff;
4662 int i;
4663 int result = diff_get_patch_id(options, sha1, diff_header_only);
4665 for (i = 0; i < q->nr; i++)
4666 diff_free_filepair(q->queue[i]);
4668 free(q->queue);
4669 DIFF_QUEUE_CLEAR(q);
4671 return result;
4674 static int is_summary_empty(const struct diff_queue_struct *q)
4676 int i;
4678 for (i = 0; i < q->nr; i++) {
4679 const struct diff_filepair *p = q->queue[i];
4681 switch (p->status) {
4682 case DIFF_STATUS_DELETED:
4683 case DIFF_STATUS_ADDED:
4684 case DIFF_STATUS_COPIED:
4685 case DIFF_STATUS_RENAMED:
4686 return 0;
4687 default:
4688 if (p->score)
4689 return 0;
4690 if (p->one->mode && p->two->mode &&
4691 p->one->mode != p->two->mode)
4692 return 0;
4693 break;
4696 return 1;
4699 static const char rename_limit_warning[] =
4700 N_("inexact rename detection was skipped due to too many files.");
4702 static const char degrade_cc_to_c_warning[] =
4703 N_("only found copies from modified paths due to too many files.");
4705 static const char rename_limit_advice[] =
4706 N_("you may want to set your %s variable to at least "
4707 "%d and retry the command.");
4709 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4711 if (degraded_cc)
4712 warning(_(degrade_cc_to_c_warning));
4713 else if (needed)
4714 warning(_(rename_limit_warning));
4715 else
4716 return;
4717 if (0 < needed && needed < 32767)
4718 warning(_(rename_limit_advice), varname, needed);
4721 void diff_flush(struct diff_options *options)
4723 struct diff_queue_struct *q = &diff_queued_diff;
4724 int i, output_format = options->output_format;
4725 int separator = 0;
4726 int dirstat_by_line = 0;
4729 * Order: raw, stat, summary, patch
4730 * or: name/name-status/checkdiff (other bits clear)
4732 if (!q->nr)
4733 goto free_queue;
4735 if (output_format & (DIFF_FORMAT_RAW |
4736 DIFF_FORMAT_NAME |
4737 DIFF_FORMAT_NAME_STATUS |
4738 DIFF_FORMAT_CHECKDIFF)) {
4739 for (i = 0; i < q->nr; i++) {
4740 struct diff_filepair *p = q->queue[i];
4741 if (check_pair_status(p))
4742 flush_one_pair(p, options);
4744 separator++;
4747 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4748 dirstat_by_line = 1;
4750 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4751 dirstat_by_line) {
4752 struct diffstat_t diffstat;
4754 memset(&diffstat, 0, sizeof(struct diffstat_t));
4755 for (i = 0; i < q->nr; i++) {
4756 struct diff_filepair *p = q->queue[i];
4757 if (check_pair_status(p))
4758 diff_flush_stat(p, options, &diffstat);
4760 if (output_format & DIFF_FORMAT_NUMSTAT)
4761 show_numstat(&diffstat, options);
4762 if (output_format & DIFF_FORMAT_DIFFSTAT)
4763 show_stats(&diffstat, options);
4764 if (output_format & DIFF_FORMAT_SHORTSTAT)
4765 show_shortstats(&diffstat, options);
4766 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
4767 show_dirstat_by_line(&diffstat, options);
4768 free_diffstat_info(&diffstat);
4769 separator++;
4771 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4772 show_dirstat(options);
4774 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4775 for (i = 0; i < q->nr; i++) {
4776 diff_summary(options, q->queue[i]);
4778 separator++;
4781 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4782 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4783 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4785 * run diff_flush_patch for the exit status. setting
4786 * options->file to /dev/null should be safe, because we
4787 * aren't supposed to produce any output anyway.
4789 if (options->close_file)
4790 fclose(options->file);
4791 options->file = fopen("/dev/null", "w");
4792 if (!options->file)
4793 die_errno("Could not open /dev/null");
4794 options->close_file = 1;
4795 for (i = 0; i < q->nr; i++) {
4796 struct diff_filepair *p = q->queue[i];
4797 if (check_pair_status(p))
4798 diff_flush_patch(p, options);
4799 if (options->found_changes)
4800 break;
4804 if (output_format & DIFF_FORMAT_PATCH) {
4805 if (separator) {
4806 fprintf(options->file, "%s%c",
4807 diff_line_prefix(options),
4808 options->line_termination);
4809 if (options->stat_sep) {
4810 /* attach patch instead of inline */
4811 fputs(options->stat_sep, options->file);
4815 for (i = 0; i < q->nr; i++) {
4816 struct diff_filepair *p = q->queue[i];
4817 if (check_pair_status(p))
4818 diff_flush_patch(p, options);
4822 if (output_format & DIFF_FORMAT_CALLBACK)
4823 options->format_callback(q, options, options->format_callback_data);
4825 for (i = 0; i < q->nr; i++)
4826 diff_free_filepair(q->queue[i]);
4827 free_queue:
4828 free(q->queue);
4829 DIFF_QUEUE_CLEAR(q);
4830 if (options->close_file)
4831 fclose(options->file);
4834 * Report the content-level differences with HAS_CHANGES;
4835 * diff_addremove/diff_change does not set the bit when
4836 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4838 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4839 if (options->found_changes)
4840 DIFF_OPT_SET(options, HAS_CHANGES);
4841 else
4842 DIFF_OPT_CLR(options, HAS_CHANGES);
4846 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
4848 return (((p->status == DIFF_STATUS_MODIFIED) &&
4849 ((p->score &&
4850 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
4851 (!p->score &&
4852 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
4853 ((p->status != DIFF_STATUS_MODIFIED) &&
4854 filter_bit_tst(p->status, options)));
4857 static void diffcore_apply_filter(struct diff_options *options)
4859 int i;
4860 struct diff_queue_struct *q = &diff_queued_diff;
4861 struct diff_queue_struct outq;
4863 DIFF_QUEUE_CLEAR(&outq);
4865 if (!options->filter)
4866 return;
4868 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
4869 int found;
4870 for (i = found = 0; !found && i < q->nr; i++) {
4871 if (match_filter(options, q->queue[i]))
4872 found++;
4874 if (found)
4875 return;
4877 /* otherwise we will clear the whole queue
4878 * by copying the empty outq at the end of this
4879 * function, but first clear the current entries
4880 * in the queue.
4882 for (i = 0; i < q->nr; i++)
4883 diff_free_filepair(q->queue[i]);
4885 else {
4886 /* Only the matching ones */
4887 for (i = 0; i < q->nr; i++) {
4888 struct diff_filepair *p = q->queue[i];
4889 if (match_filter(options, p))
4890 diff_q(&outq, p);
4891 else
4892 diff_free_filepair(p);
4895 free(q->queue);
4896 *q = outq;
4899 /* Check whether two filespecs with the same mode and size are identical */
4900 static int diff_filespec_is_identical(struct diff_filespec *one,
4901 struct diff_filespec *two)
4903 if (S_ISGITLINK(one->mode))
4904 return 0;
4905 if (diff_populate_filespec(one, 0))
4906 return 0;
4907 if (diff_populate_filespec(two, 0))
4908 return 0;
4909 return !memcmp(one->data, two->data, one->size);
4912 static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
4914 if (p->done_skip_stat_unmatch)
4915 return p->skip_stat_unmatch_result;
4917 p->done_skip_stat_unmatch = 1;
4918 p->skip_stat_unmatch_result = 0;
4920 * 1. Entries that come from stat info dirtiness
4921 * always have both sides (iow, not create/delete),
4922 * one side of the object name is unknown, with
4923 * the same mode and size. Keep the ones that
4924 * do not match these criteria. They have real
4925 * differences.
4927 * 2. At this point, the file is known to be modified,
4928 * with the same mode and size, and the object
4929 * name of one side is unknown. Need to inspect
4930 * the identical contents.
4932 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4933 !DIFF_FILE_VALID(p->two) ||
4934 (p->one->oid_valid && p->two->oid_valid) ||
4935 (p->one->mode != p->two->mode) ||
4936 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
4937 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
4938 (p->one->size != p->two->size) ||
4939 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4940 p->skip_stat_unmatch_result = 1;
4941 return p->skip_stat_unmatch_result;
4944 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4946 int i;
4947 struct diff_queue_struct *q = &diff_queued_diff;
4948 struct diff_queue_struct outq;
4949 DIFF_QUEUE_CLEAR(&outq);
4951 for (i = 0; i < q->nr; i++) {
4952 struct diff_filepair *p = q->queue[i];
4954 if (diff_filespec_check_stat_unmatch(p))
4955 diff_q(&outq, p);
4956 else {
4958 * The caller can subtract 1 from skip_stat_unmatch
4959 * to determine how many paths were dirty only
4960 * due to stat info mismatch.
4962 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4963 diffopt->skip_stat_unmatch++;
4964 diff_free_filepair(p);
4967 free(q->queue);
4968 *q = outq;
4971 static int diffnamecmp(const void *a_, const void *b_)
4973 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4974 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4975 const char *name_a, *name_b;
4977 name_a = a->one ? a->one->path : a->two->path;
4978 name_b = b->one ? b->one->path : b->two->path;
4979 return strcmp(name_a, name_b);
4982 void diffcore_fix_diff_index(struct diff_options *options)
4984 struct diff_queue_struct *q = &diff_queued_diff;
4985 QSORT(q->queue, q->nr, diffnamecmp);
4988 void diffcore_std(struct diff_options *options)
4990 /* NOTE please keep the following in sync with diff_tree_combined() */
4991 if (options->skip_stat_unmatch)
4992 diffcore_skip_stat_unmatch(options);
4993 if (!options->found_follow) {
4994 /* See try_to_follow_renames() in tree-diff.c */
4995 if (options->break_opt != -1)
4996 diffcore_break(options->break_opt);
4997 if (options->detect_rename)
4998 diffcore_rename(options);
4999 if (options->break_opt != -1)
5000 diffcore_merge_broken();
5002 if (options->pickaxe)
5003 diffcore_pickaxe(options);
5004 if (options->orderfile)
5005 diffcore_order(options->orderfile);
5006 if (!options->found_follow)
5007 /* See try_to_follow_renames() in tree-diff.c */
5008 diff_resolve_rename_copy();
5009 diffcore_apply_filter(options);
5011 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5012 DIFF_OPT_SET(options, HAS_CHANGES);
5013 else
5014 DIFF_OPT_CLR(options, HAS_CHANGES);
5016 options->found_follow = 0;
5019 int diff_result_code(struct diff_options *opt, int status)
5021 int result = 0;
5023 diff_warn_rename_limit("diff.renameLimit",
5024 opt->needed_rename_limit,
5025 opt->degraded_cc_to_c);
5026 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5027 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5028 return status;
5029 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5030 DIFF_OPT_TST(opt, HAS_CHANGES))
5031 result |= 01;
5032 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5033 DIFF_OPT_TST(opt, CHECK_FAILED))
5034 result |= 02;
5035 return result;
5038 int diff_can_quit_early(struct diff_options *opt)
5040 return (DIFF_OPT_TST(opt, QUICK) &&
5041 !opt->filter &&
5042 DIFF_OPT_TST(opt, HAS_CHANGES));
5046 * Shall changes to this submodule be ignored?
5048 * Submodule changes can be configured to be ignored separately for each path,
5049 * but that configuration can be overridden from the command line.
5051 static int is_submodule_ignored(const char *path, struct diff_options *options)
5053 int ignored = 0;
5054 unsigned orig_flags = options->flags;
5055 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5056 set_diffopt_flags_from_submodule_config(options, path);
5057 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5058 ignored = 1;
5059 options->flags = orig_flags;
5060 return ignored;
5063 void diff_addremove(struct diff_options *options,
5064 int addremove, unsigned mode,
5065 const unsigned char *sha1,
5066 int sha1_valid,
5067 const char *concatpath, unsigned dirty_submodule)
5069 struct diff_filespec *one, *two;
5071 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5072 return;
5074 /* This may look odd, but it is a preparation for
5075 * feeding "there are unchanged files which should
5076 * not produce diffs, but when you are doing copy
5077 * detection you would need them, so here they are"
5078 * entries to the diff-core. They will be prefixed
5079 * with something like '=' or '*' (I haven't decided
5080 * which but should not make any difference).
5081 * Feeding the same new and old to diff_change()
5082 * also has the same effect.
5083 * Before the final output happens, they are pruned after
5084 * merged into rename/copy pairs as appropriate.
5086 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5087 addremove = (addremove == '+' ? '-' :
5088 addremove == '-' ? '+' : addremove);
5090 if (options->prefix &&
5091 strncmp(concatpath, options->prefix, options->prefix_length))
5092 return;
5094 one = alloc_filespec(concatpath);
5095 two = alloc_filespec(concatpath);
5097 if (addremove != '+')
5098 fill_filespec(one, sha1, sha1_valid, mode);
5099 if (addremove != '-') {
5100 fill_filespec(two, sha1, sha1_valid, mode);
5101 two->dirty_submodule = dirty_submodule;
5104 diff_queue(&diff_queued_diff, one, two);
5105 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5106 DIFF_OPT_SET(options, HAS_CHANGES);
5109 void diff_change(struct diff_options *options,
5110 unsigned old_mode, unsigned new_mode,
5111 const unsigned char *old_sha1,
5112 const unsigned char *new_sha1,
5113 int old_sha1_valid, int new_sha1_valid,
5114 const char *concatpath,
5115 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5117 struct diff_filespec *one, *two;
5118 struct diff_filepair *p;
5120 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5121 is_submodule_ignored(concatpath, options))
5122 return;
5124 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5125 unsigned tmp;
5126 const unsigned char *tmp_c;
5127 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
5128 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
5129 tmp = old_sha1_valid; old_sha1_valid = new_sha1_valid;
5130 new_sha1_valid = tmp;
5131 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
5132 new_dirty_submodule = tmp;
5135 if (options->prefix &&
5136 strncmp(concatpath, options->prefix, options->prefix_length))
5137 return;
5139 one = alloc_filespec(concatpath);
5140 two = alloc_filespec(concatpath);
5141 fill_filespec(one, old_sha1, old_sha1_valid, old_mode);
5142 fill_filespec(two, new_sha1, new_sha1_valid, new_mode);
5143 one->dirty_submodule = old_dirty_submodule;
5144 two->dirty_submodule = new_dirty_submodule;
5145 p = diff_queue(&diff_queued_diff, one, two);
5147 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5148 return;
5150 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5151 !diff_filespec_check_stat_unmatch(p))
5152 return;
5154 DIFF_OPT_SET(options, HAS_CHANGES);
5157 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5159 struct diff_filepair *pair;
5160 struct diff_filespec *one, *two;
5162 if (options->prefix &&
5163 strncmp(path, options->prefix, options->prefix_length))
5164 return NULL;
5166 one = alloc_filespec(path);
5167 two = alloc_filespec(path);
5168 pair = diff_queue(&diff_queued_diff, one, two);
5169 pair->is_unmerged = 1;
5170 return pair;
5173 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5174 size_t *outsize)
5176 struct diff_tempfile *temp;
5177 const char *argv[3];
5178 const char **arg = argv;
5179 struct child_process child = CHILD_PROCESS_INIT;
5180 struct strbuf buf = STRBUF_INIT;
5181 int err = 0;
5183 temp = prepare_temp_file(spec->path, spec);
5184 *arg++ = pgm;
5185 *arg++ = temp->name;
5186 *arg = NULL;
5188 child.use_shell = 1;
5189 child.argv = argv;
5190 child.out = -1;
5191 if (start_command(&child)) {
5192 remove_tempfile();
5193 return NULL;
5196 if (strbuf_read(&buf, child.out, 0) < 0)
5197 err = error("error reading from textconv command '%s'", pgm);
5198 close(child.out);
5200 if (finish_command(&child) || err) {
5201 strbuf_release(&buf);
5202 remove_tempfile();
5203 return NULL;
5205 remove_tempfile();
5207 return strbuf_detach(&buf, outsize);
5210 size_t fill_textconv(struct userdiff_driver *driver,
5211 struct diff_filespec *df,
5212 char **outbuf)
5214 size_t size;
5216 if (!driver) {
5217 if (!DIFF_FILE_VALID(df)) {
5218 *outbuf = "";
5219 return 0;
5221 if (diff_populate_filespec(df, 0))
5222 die("unable to read files to diff");
5223 *outbuf = df->data;
5224 return df->size;
5227 if (!driver->textconv)
5228 die("BUG: fill_textconv called with non-textconv driver");
5230 if (driver->textconv_cache && df->oid_valid) {
5231 *outbuf = notes_cache_get(driver->textconv_cache,
5232 df->oid.hash,
5233 &size);
5234 if (*outbuf)
5235 return size;
5238 *outbuf = run_textconv(driver->textconv, df, &size);
5239 if (!*outbuf)
5240 die("unable to read files to diff");
5242 if (driver->textconv_cache && df->oid_valid) {
5243 /* ignore errors, as we might be in a readonly repository */
5244 notes_cache_put(driver->textconv_cache, df->oid.hash, *outbuf,
5245 size);
5247 * we could save up changes and flush them all at the end,
5248 * but we would need an extra call after all diffing is done.
5249 * Since generating a cache entry is the slow path anyway,
5250 * this extra overhead probably isn't a big deal.
5252 notes_cache_write(driver->textconv_cache);
5255 return size;
5258 void setup_diff_pager(struct diff_options *opt)
5261 * If the user asked for our exit code, then either they want --quiet
5262 * or --exit-code. We should definitely not bother with a pager in the
5263 * former case, as we will generate no output. Since we still properly
5264 * report our exit code even when a pager is run, we _could_ run a
5265 * pager with --exit-code. But since we have not done so historically,
5266 * and because it is easy to find people oneline advising "git diff
5267 * --exit-code" in hooks and other scripts, we do not do so.
5269 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5270 check_pager_config("diff") != 0)
5271 setup_pager();