Merge branch 'pw/diff-color-moved-fix'
[git/debian.git] / diff.c
blob0b4f9b52c58eb156f2b076dbb0e1163ec6db6df3
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "config.h"
6 #include "tempfile.h"
7 #include "quote.h"
8 #include "diff.h"
9 #include "diffcore.h"
10 #include "delta.h"
11 #include "xdiff-interface.h"
12 #include "color.h"
13 #include "attr.h"
14 #include "run-command.h"
15 #include "utf8.h"
16 #include "object-store.h"
17 #include "userdiff.h"
18 #include "submodule-config.h"
19 #include "submodule.h"
20 #include "hashmap.h"
21 #include "mem-pool.h"
22 #include "ll-merge.h"
23 #include "string-list.h"
24 #include "strvec.h"
25 #include "graph.h"
26 #include "packfile.h"
27 #include "parse-options.h"
28 #include "help.h"
29 #include "promisor-remote.h"
30 #include "dir.h"
32 #ifdef NO_FAST_WORKING_DIRECTORY
33 #define FAST_WORKING_DIRECTORY 0
34 #else
35 #define FAST_WORKING_DIRECTORY 1
36 #endif
38 static int diff_detect_rename_default;
39 static int diff_indent_heuristic = 1;
40 static int diff_rename_limit_default = 1000;
41 static int diff_suppress_blank_empty;
42 static int diff_use_color_default = -1;
43 static int diff_color_moved_default;
44 static int diff_color_moved_ws_default;
45 static int diff_context_default = 3;
46 static int diff_interhunk_context_default;
47 static const char *diff_word_regex_cfg;
48 static const char *external_diff_cmd_cfg;
49 static const char *diff_order_file_cfg;
50 int diff_auto_refresh_index = 1;
51 static int diff_mnemonic_prefix;
52 static int diff_no_prefix;
53 static int diff_relative;
54 static int diff_stat_graph_width;
55 static int diff_dirstat_permille_default = 30;
56 static struct diff_options default_diff_options;
57 static long diff_algorithm;
58 static unsigned ws_error_highlight_default = WSEH_NEW;
60 static char diff_colors[][COLOR_MAXLEN] = {
61 GIT_COLOR_RESET,
62 GIT_COLOR_NORMAL, /* CONTEXT */
63 GIT_COLOR_BOLD, /* METAINFO */
64 GIT_COLOR_CYAN, /* FRAGINFO */
65 GIT_COLOR_RED, /* OLD */
66 GIT_COLOR_GREEN, /* NEW */
67 GIT_COLOR_YELLOW, /* COMMIT */
68 GIT_COLOR_BG_RED, /* WHITESPACE */
69 GIT_COLOR_NORMAL, /* FUNCINFO */
70 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
71 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
72 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
73 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
74 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
75 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
76 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
77 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
78 GIT_COLOR_FAINT, /* CONTEXT_DIM */
79 GIT_COLOR_FAINT_RED, /* OLD_DIM */
80 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
81 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
82 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
83 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
86 static const char *color_diff_slots[] = {
87 [DIFF_CONTEXT] = "context",
88 [DIFF_METAINFO] = "meta",
89 [DIFF_FRAGINFO] = "frag",
90 [DIFF_FILE_OLD] = "old",
91 [DIFF_FILE_NEW] = "new",
92 [DIFF_COMMIT] = "commit",
93 [DIFF_WHITESPACE] = "whitespace",
94 [DIFF_FUNCINFO] = "func",
95 [DIFF_FILE_OLD_MOVED] = "oldMoved",
96 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
97 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
98 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
99 [DIFF_FILE_NEW_MOVED] = "newMoved",
100 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
101 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
102 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
103 [DIFF_CONTEXT_DIM] = "contextDimmed",
104 [DIFF_FILE_OLD_DIM] = "oldDimmed",
105 [DIFF_FILE_NEW_DIM] = "newDimmed",
106 [DIFF_CONTEXT_BOLD] = "contextBold",
107 [DIFF_FILE_OLD_BOLD] = "oldBold",
108 [DIFF_FILE_NEW_BOLD] = "newBold",
111 define_list_config_array_extra(color_diff_slots, {"plain"});
113 static int parse_diff_color_slot(const char *var)
115 if (!strcasecmp(var, "plain"))
116 return DIFF_CONTEXT;
117 return LOOKUP_CONFIG(color_diff_slots, var);
120 static int parse_dirstat_params(struct diff_options *options, const char *params_string,
121 struct strbuf *errmsg)
123 char *params_copy = xstrdup(params_string);
124 struct string_list params = STRING_LIST_INIT_NODUP;
125 int ret = 0;
126 int i;
128 if (*params_copy)
129 string_list_split_in_place(&params, params_copy, ',', -1);
130 for (i = 0; i < params.nr; i++) {
131 const char *p = params.items[i].string;
132 if (!strcmp(p, "changes")) {
133 options->flags.dirstat_by_line = 0;
134 options->flags.dirstat_by_file = 0;
135 } else if (!strcmp(p, "lines")) {
136 options->flags.dirstat_by_line = 1;
137 options->flags.dirstat_by_file = 0;
138 } else if (!strcmp(p, "files")) {
139 options->flags.dirstat_by_line = 0;
140 options->flags.dirstat_by_file = 1;
141 } else if (!strcmp(p, "noncumulative")) {
142 options->flags.dirstat_cumulative = 0;
143 } else if (!strcmp(p, "cumulative")) {
144 options->flags.dirstat_cumulative = 1;
145 } else if (isdigit(*p)) {
146 char *end;
147 int permille = strtoul(p, &end, 10) * 10;
148 if (*end == '.' && isdigit(*++end)) {
149 /* only use first digit */
150 permille += *end - '0';
151 /* .. and ignore any further digits */
152 while (isdigit(*++end))
153 ; /* nothing */
155 if (!*end)
156 options->dirstat_permille = permille;
157 else {
158 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
160 ret++;
162 } else {
163 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
164 ret++;
168 string_list_clear(&params, 0);
169 free(params_copy);
170 return ret;
173 static int parse_submodule_params(struct diff_options *options, const char *value)
175 if (!strcmp(value, "log"))
176 options->submodule_format = DIFF_SUBMODULE_LOG;
177 else if (!strcmp(value, "short"))
178 options->submodule_format = DIFF_SUBMODULE_SHORT;
179 else if (!strcmp(value, "diff"))
180 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
182 * Please update $__git_diff_submodule_formats in
183 * git-completion.bash when you add new formats.
185 else
186 return -1;
187 return 0;
190 int git_config_rename(const char *var, const char *value)
192 if (!value)
193 return DIFF_DETECT_RENAME;
194 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
195 return DIFF_DETECT_COPY;
196 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
199 long parse_algorithm_value(const char *value)
201 if (!value)
202 return -1;
203 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
204 return 0;
205 else if (!strcasecmp(value, "minimal"))
206 return XDF_NEED_MINIMAL;
207 else if (!strcasecmp(value, "patience"))
208 return XDF_PATIENCE_DIFF;
209 else if (!strcasecmp(value, "histogram"))
210 return XDF_HISTOGRAM_DIFF;
212 * Please update $__git_diff_algorithms in git-completion.bash
213 * when you add new algorithms.
215 return -1;
218 static int parse_one_token(const char **arg, const char *token)
220 const char *rest;
221 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
222 *arg = rest;
223 return 1;
225 return 0;
228 static int parse_ws_error_highlight(const char *arg)
230 const char *orig_arg = arg;
231 unsigned val = 0;
233 while (*arg) {
234 if (parse_one_token(&arg, "none"))
235 val = 0;
236 else if (parse_one_token(&arg, "default"))
237 val = WSEH_NEW;
238 else if (parse_one_token(&arg, "all"))
239 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
240 else if (parse_one_token(&arg, "new"))
241 val |= WSEH_NEW;
242 else if (parse_one_token(&arg, "old"))
243 val |= WSEH_OLD;
244 else if (parse_one_token(&arg, "context"))
245 val |= WSEH_CONTEXT;
246 else {
247 return -1 - (int)(arg - orig_arg);
249 if (*arg)
250 arg++;
252 return val;
256 * These are to give UI layer defaults.
257 * The core-level commands such as git-diff-files should
258 * never be affected by the setting of diff.renames
259 * the user happens to have in the configuration file.
261 void init_diff_ui_defaults(void)
263 diff_detect_rename_default = DIFF_DETECT_RENAME;
266 int git_diff_heuristic_config(const char *var, const char *value, void *cb)
268 if (!strcmp(var, "diff.indentheuristic"))
269 diff_indent_heuristic = git_config_bool(var, value);
270 return 0;
273 static int parse_color_moved(const char *arg)
275 switch (git_parse_maybe_bool(arg)) {
276 case 0:
277 return COLOR_MOVED_NO;
278 case 1:
279 return COLOR_MOVED_DEFAULT;
280 default:
281 break;
284 if (!strcmp(arg, "no"))
285 return COLOR_MOVED_NO;
286 else if (!strcmp(arg, "plain"))
287 return COLOR_MOVED_PLAIN;
288 else if (!strcmp(arg, "blocks"))
289 return COLOR_MOVED_BLOCKS;
290 else if (!strcmp(arg, "zebra"))
291 return COLOR_MOVED_ZEBRA;
292 else if (!strcmp(arg, "default"))
293 return COLOR_MOVED_DEFAULT;
294 else if (!strcmp(arg, "dimmed-zebra"))
295 return COLOR_MOVED_ZEBRA_DIM;
296 else if (!strcmp(arg, "dimmed_zebra"))
297 return COLOR_MOVED_ZEBRA_DIM;
298 else
299 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
302 static unsigned parse_color_moved_ws(const char *arg)
304 int ret = 0;
305 struct string_list l = STRING_LIST_INIT_DUP;
306 struct string_list_item *i;
308 string_list_split(&l, arg, ',', -1);
310 for_each_string_list_item(i, &l) {
311 struct strbuf sb = STRBUF_INIT;
312 strbuf_addstr(&sb, i->string);
313 strbuf_trim(&sb);
315 if (!strcmp(sb.buf, "no"))
316 ret = 0;
317 else if (!strcmp(sb.buf, "ignore-space-change"))
318 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
319 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
320 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
321 else if (!strcmp(sb.buf, "ignore-all-space"))
322 ret |= XDF_IGNORE_WHITESPACE;
323 else if (!strcmp(sb.buf, "allow-indentation-change"))
324 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
325 else {
326 ret |= COLOR_MOVED_WS_ERROR;
327 error(_("unknown color-moved-ws mode '%s', possible values are 'ignore-space-change', 'ignore-space-at-eol', 'ignore-all-space', 'allow-indentation-change'"), sb.buf);
330 strbuf_release(&sb);
333 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
334 (ret & XDF_WHITESPACE_FLAGS)) {
335 error(_("color-moved-ws: allow-indentation-change cannot be combined with other whitespace modes"));
336 ret |= COLOR_MOVED_WS_ERROR;
339 string_list_clear(&l, 0);
341 return ret;
344 int git_diff_ui_config(const char *var, const char *value, void *cb)
346 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
347 diff_use_color_default = git_config_colorbool(var, value);
348 return 0;
350 if (!strcmp(var, "diff.colormoved")) {
351 int cm = parse_color_moved(value);
352 if (cm < 0)
353 return -1;
354 diff_color_moved_default = cm;
355 return 0;
357 if (!strcmp(var, "diff.colormovedws")) {
358 unsigned cm = parse_color_moved_ws(value);
359 if (cm & COLOR_MOVED_WS_ERROR)
360 return -1;
361 diff_color_moved_ws_default = cm;
362 return 0;
364 if (!strcmp(var, "diff.context")) {
365 diff_context_default = git_config_int(var, value);
366 if (diff_context_default < 0)
367 return -1;
368 return 0;
370 if (!strcmp(var, "diff.interhunkcontext")) {
371 diff_interhunk_context_default = git_config_int(var, value);
372 if (diff_interhunk_context_default < 0)
373 return -1;
374 return 0;
376 if (!strcmp(var, "diff.renames")) {
377 diff_detect_rename_default = git_config_rename(var, value);
378 return 0;
380 if (!strcmp(var, "diff.autorefreshindex")) {
381 diff_auto_refresh_index = git_config_bool(var, value);
382 return 0;
384 if (!strcmp(var, "diff.mnemonicprefix")) {
385 diff_mnemonic_prefix = git_config_bool(var, value);
386 return 0;
388 if (!strcmp(var, "diff.noprefix")) {
389 diff_no_prefix = git_config_bool(var, value);
390 return 0;
392 if (!strcmp(var, "diff.relative")) {
393 diff_relative = git_config_bool(var, value);
394 return 0;
396 if (!strcmp(var, "diff.statgraphwidth")) {
397 diff_stat_graph_width = git_config_int(var, value);
398 return 0;
400 if (!strcmp(var, "diff.external"))
401 return git_config_string(&external_diff_cmd_cfg, var, value);
402 if (!strcmp(var, "diff.wordregex"))
403 return git_config_string(&diff_word_regex_cfg, var, value);
404 if (!strcmp(var, "diff.orderfile"))
405 return git_config_pathname(&diff_order_file_cfg, var, value);
407 if (!strcmp(var, "diff.ignoresubmodules"))
408 handle_ignore_submodules_arg(&default_diff_options, value);
410 if (!strcmp(var, "diff.submodule")) {
411 if (parse_submodule_params(&default_diff_options, value))
412 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
413 value);
414 return 0;
417 if (!strcmp(var, "diff.algorithm")) {
418 diff_algorithm = parse_algorithm_value(value);
419 if (diff_algorithm < 0)
420 return -1;
421 return 0;
424 if (git_color_config(var, value, cb) < 0)
425 return -1;
427 return git_diff_basic_config(var, value, cb);
430 int git_diff_basic_config(const char *var, const char *value, void *cb)
432 const char *name;
434 if (!strcmp(var, "diff.renamelimit")) {
435 diff_rename_limit_default = git_config_int(var, value);
436 return 0;
439 if (userdiff_config(var, value) < 0)
440 return -1;
442 if (skip_prefix(var, "diff.color.", &name) ||
443 skip_prefix(var, "color.diff.", &name)) {
444 int slot = parse_diff_color_slot(name);
445 if (slot < 0)
446 return 0;
447 if (!value)
448 return config_error_nonbool(var);
449 return color_parse(value, diff_colors[slot]);
452 if (!strcmp(var, "diff.wserrorhighlight")) {
453 int val = parse_ws_error_highlight(value);
454 if (val < 0)
455 return -1;
456 ws_error_highlight_default = val;
457 return 0;
460 /* like GNU diff's --suppress-blank-empty option */
461 if (!strcmp(var, "diff.suppressblankempty") ||
462 /* for backwards compatibility */
463 !strcmp(var, "diff.suppress-blank-empty")) {
464 diff_suppress_blank_empty = git_config_bool(var, value);
465 return 0;
468 if (!strcmp(var, "diff.dirstat")) {
469 struct strbuf errmsg = STRBUF_INIT;
470 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
471 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
472 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
473 errmsg.buf);
474 strbuf_release(&errmsg);
475 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
476 return 0;
479 if (git_diff_heuristic_config(var, value, cb) < 0)
480 return -1;
482 return git_default_config(var, value, cb);
485 static char *quote_two(const char *one, const char *two)
487 int need_one = quote_c_style(one, NULL, NULL, CQUOTE_NODQ);
488 int need_two = quote_c_style(two, NULL, NULL, CQUOTE_NODQ);
489 struct strbuf res = STRBUF_INIT;
491 if (need_one + need_two) {
492 strbuf_addch(&res, '"');
493 quote_c_style(one, &res, NULL, CQUOTE_NODQ);
494 quote_c_style(two, &res, NULL, CQUOTE_NODQ);
495 strbuf_addch(&res, '"');
496 } else {
497 strbuf_addstr(&res, one);
498 strbuf_addstr(&res, two);
500 return strbuf_detach(&res, NULL);
503 static const char *external_diff(void)
505 static const char *external_diff_cmd = NULL;
506 static int done_preparing = 0;
508 if (done_preparing)
509 return external_diff_cmd;
510 external_diff_cmd = xstrdup_or_null(getenv("GIT_EXTERNAL_DIFF"));
511 if (!external_diff_cmd)
512 external_diff_cmd = external_diff_cmd_cfg;
513 done_preparing = 1;
514 return external_diff_cmd;
518 * Keep track of files used for diffing. Sometimes such an entry
519 * refers to a temporary file, sometimes to an existing file, and
520 * sometimes to "/dev/null".
522 static struct diff_tempfile {
524 * filename external diff should read from, or NULL if this
525 * entry is currently not in use:
527 const char *name;
529 char hex[GIT_MAX_HEXSZ + 1];
530 char mode[10];
533 * If this diff_tempfile instance refers to a temporary file,
534 * this tempfile object is used to manage its lifetime.
536 struct tempfile *tempfile;
537 } diff_temp[2];
539 struct emit_callback {
540 int color_diff;
541 unsigned ws_rule;
542 int blank_at_eof_in_preimage;
543 int blank_at_eof_in_postimage;
544 int lno_in_preimage;
545 int lno_in_postimage;
546 const char **label_path;
547 struct diff_words_data *diff_words;
548 struct diff_options *opt;
549 struct strbuf *header;
552 static int count_lines(const char *data, int size)
554 int count, ch, completely_empty = 1, nl_just_seen = 0;
555 count = 0;
556 while (0 < size--) {
557 ch = *data++;
558 if (ch == '\n') {
559 count++;
560 nl_just_seen = 1;
561 completely_empty = 0;
563 else {
564 nl_just_seen = 0;
565 completely_empty = 0;
568 if (completely_empty)
569 return 0;
570 if (!nl_just_seen)
571 count++; /* no trailing newline */
572 return count;
575 static int fill_mmfile(struct repository *r, mmfile_t *mf,
576 struct diff_filespec *one)
578 if (!DIFF_FILE_VALID(one)) {
579 mf->ptr = (char *)""; /* does not matter */
580 mf->size = 0;
581 return 0;
583 else if (diff_populate_filespec(r, one, NULL))
584 return -1;
586 mf->ptr = one->data;
587 mf->size = one->size;
588 return 0;
591 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
592 static unsigned long diff_filespec_size(struct repository *r,
593 struct diff_filespec *one)
595 struct diff_populate_filespec_options dpf_options = {
596 .check_size_only = 1,
599 if (!DIFF_FILE_VALID(one))
600 return 0;
601 diff_populate_filespec(r, one, &dpf_options);
602 return one->size;
605 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
607 char *ptr = mf->ptr;
608 long size = mf->size;
609 int cnt = 0;
611 if (!size)
612 return cnt;
613 ptr += size - 1; /* pointing at the very end */
614 if (*ptr != '\n')
615 ; /* incomplete line */
616 else
617 ptr--; /* skip the last LF */
618 while (mf->ptr < ptr) {
619 char *prev_eol;
620 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
621 if (*prev_eol == '\n')
622 break;
623 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
624 break;
625 cnt++;
626 ptr = prev_eol - 1;
628 return cnt;
631 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
632 struct emit_callback *ecbdata)
634 int l1, l2, at;
635 unsigned ws_rule = ecbdata->ws_rule;
636 l1 = count_trailing_blank(mf1, ws_rule);
637 l2 = count_trailing_blank(mf2, ws_rule);
638 if (l2 <= l1) {
639 ecbdata->blank_at_eof_in_preimage = 0;
640 ecbdata->blank_at_eof_in_postimage = 0;
641 return;
643 at = count_lines(mf1->ptr, mf1->size);
644 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
646 at = count_lines(mf2->ptr, mf2->size);
647 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
650 static void emit_line_0(struct diff_options *o,
651 const char *set_sign, const char *set, unsigned reverse, const char *reset,
652 int first, const char *line, int len)
654 int has_trailing_newline, has_trailing_carriage_return;
655 int needs_reset = 0; /* at the end of the line */
656 FILE *file = o->file;
658 fputs(diff_line_prefix(o), file);
660 has_trailing_newline = (len > 0 && line[len-1] == '\n');
661 if (has_trailing_newline)
662 len--;
664 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
665 if (has_trailing_carriage_return)
666 len--;
668 if (!len && !first)
669 goto end_of_line;
671 if (reverse && want_color(o->use_color)) {
672 fputs(GIT_COLOR_REVERSE, file);
673 needs_reset = 1;
676 if (set_sign) {
677 fputs(set_sign, file);
678 needs_reset = 1;
681 if (first)
682 fputc(first, file);
684 if (!len)
685 goto end_of_line;
687 if (set) {
688 if (set_sign && set != set_sign)
689 fputs(reset, file);
690 fputs(set, file);
691 needs_reset = 1;
693 fwrite(line, len, 1, file);
694 needs_reset = 1; /* 'line' may contain color codes. */
696 end_of_line:
697 if (needs_reset)
698 fputs(reset, file);
699 if (has_trailing_carriage_return)
700 fputc('\r', file);
701 if (has_trailing_newline)
702 fputc('\n', file);
705 static void emit_line(struct diff_options *o, const char *set, const char *reset,
706 const char *line, int len)
708 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
711 enum diff_symbol {
712 DIFF_SYMBOL_BINARY_DIFF_HEADER,
713 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
714 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
715 DIFF_SYMBOL_BINARY_DIFF_BODY,
716 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
717 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
718 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
719 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
720 DIFF_SYMBOL_STATS_LINE,
721 DIFF_SYMBOL_WORD_DIFF,
722 DIFF_SYMBOL_STAT_SEP,
723 DIFF_SYMBOL_SUMMARY,
724 DIFF_SYMBOL_SUBMODULE_ADD,
725 DIFF_SYMBOL_SUBMODULE_DEL,
726 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
727 DIFF_SYMBOL_SUBMODULE_MODIFIED,
728 DIFF_SYMBOL_SUBMODULE_HEADER,
729 DIFF_SYMBOL_SUBMODULE_ERROR,
730 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
731 DIFF_SYMBOL_REWRITE_DIFF,
732 DIFF_SYMBOL_BINARY_FILES,
733 DIFF_SYMBOL_HEADER,
734 DIFF_SYMBOL_FILEPAIR_PLUS,
735 DIFF_SYMBOL_FILEPAIR_MINUS,
736 DIFF_SYMBOL_WORDS_PORCELAIN,
737 DIFF_SYMBOL_WORDS,
738 DIFF_SYMBOL_CONTEXT,
739 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
740 DIFF_SYMBOL_PLUS,
741 DIFF_SYMBOL_MINUS,
742 DIFF_SYMBOL_NO_LF_EOF,
743 DIFF_SYMBOL_CONTEXT_FRAGINFO,
744 DIFF_SYMBOL_CONTEXT_MARKER,
745 DIFF_SYMBOL_SEPARATOR
748 * Flags for content lines:
749 * 0..12 are whitespace rules
750 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
751 * 16 is marking if the line is blank at EOF
753 #define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
754 #define DIFF_SYMBOL_MOVED_LINE (1<<17)
755 #define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
756 #define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
757 #define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
760 * This struct is used when we need to buffer the output of the diff output.
762 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
763 * into the pre/post image file. This pointer could be a union with the
764 * line pointer. By storing an offset into the file instead of the literal line,
765 * we can decrease the memory footprint for the buffered output. At first we
766 * may want to only have indirection for the content lines, but we could also
767 * enhance the state for emitting prefabricated lines, e.g. the similarity
768 * score line or hunk/file headers would only need to store a number or path
769 * and then the output can be constructed later on depending on state.
771 struct emitted_diff_symbol {
772 const char *line;
773 int len;
774 int flags;
775 int indent_off; /* Offset to first non-whitespace character */
776 int indent_width; /* The visual width of the indentation */
777 unsigned id;
778 enum diff_symbol s;
780 #define EMITTED_DIFF_SYMBOL_INIT { 0 }
782 struct emitted_diff_symbols {
783 struct emitted_diff_symbol *buf;
784 int nr, alloc;
786 #define EMITTED_DIFF_SYMBOLS_INIT { 0 }
788 static void append_emitted_diff_symbol(struct diff_options *o,
789 struct emitted_diff_symbol *e)
791 struct emitted_diff_symbol *f;
793 ALLOC_GROW(o->emitted_symbols->buf,
794 o->emitted_symbols->nr + 1,
795 o->emitted_symbols->alloc);
796 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
798 memcpy(f, e, sizeof(struct emitted_diff_symbol));
799 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
802 struct moved_entry {
803 const struct emitted_diff_symbol *es;
804 struct moved_entry *next_line;
805 struct moved_entry *next_match;
808 struct moved_block {
809 struct moved_entry *match;
810 int wsd; /* The whitespace delta of this block */
813 #define INDENT_BLANKLINE INT_MIN
815 static void fill_es_indent_data(struct emitted_diff_symbol *es)
817 unsigned int off = 0, i;
818 int width = 0, tab_width = es->flags & WS_TAB_WIDTH_MASK;
819 const char *s = es->line;
820 const int len = es->len;
822 /* skip any \v \f \r at start of indentation */
823 while (s[off] == '\f' || s[off] == '\v' ||
824 (s[off] == '\r' && off < len - 1))
825 off++;
827 /* calculate the visual width of indentation */
828 while(1) {
829 if (s[off] == ' ') {
830 width++;
831 off++;
832 } else if (s[off] == '\t') {
833 width += tab_width - (width % tab_width);
834 while (s[++off] == '\t')
835 width += tab_width;
836 } else {
837 break;
841 /* check if this line is blank */
842 for (i = off; i < len; i++)
843 if (!isspace(s[i]))
844 break;
846 if (i == len) {
847 es->indent_width = INDENT_BLANKLINE;
848 es->indent_off = len;
849 } else {
850 es->indent_off = off;
851 es->indent_width = width;
855 static int compute_ws_delta(const struct emitted_diff_symbol *a,
856 const struct emitted_diff_symbol *b)
858 int a_width = a->indent_width,
859 b_width = b->indent_width;
861 if (a_width == INDENT_BLANKLINE && b_width == INDENT_BLANKLINE)
862 return INDENT_BLANKLINE;
864 return a_width - b_width;
867 static int cmp_in_block_with_wsd(const struct moved_entry *cur,
868 const struct emitted_diff_symbol *l,
869 struct moved_block *pmb)
871 int a_width = cur->es->indent_width, b_width = l->indent_width;
872 int delta;
874 /* The text of each line must match */
875 if (cur->es->id != l->id)
876 return 1;
879 * If 'l' and 'cur' are both blank then we don't need to check the
880 * indent. We only need to check cur as we know the strings match.
881 * */
882 if (a_width == INDENT_BLANKLINE)
883 return 0;
886 * The indent changes of the block are known and stored in pmb->wsd;
887 * however we need to check if the indent changes of the current line
888 * match those of the current block.
890 delta = b_width - a_width;
893 * If the previous lines of this block were all blank then set its
894 * whitespace delta.
896 if (pmb->wsd == INDENT_BLANKLINE)
897 pmb->wsd = delta;
899 return delta != pmb->wsd;
902 struct interned_diff_symbol {
903 struct hashmap_entry ent;
904 struct emitted_diff_symbol *es;
907 static int interned_diff_symbol_cmp(const void *hashmap_cmp_fn_data,
908 const struct hashmap_entry *eptr,
909 const struct hashmap_entry *entry_or_key,
910 const void *keydata)
912 const struct diff_options *diffopt = hashmap_cmp_fn_data;
913 const struct emitted_diff_symbol *a, *b;
914 unsigned flags = diffopt->color_moved_ws_handling
915 & XDF_WHITESPACE_FLAGS;
917 a = container_of(eptr, const struct interned_diff_symbol, ent)->es;
918 b = container_of(entry_or_key, const struct interned_diff_symbol, ent)->es;
920 return !xdiff_compare_lines(a->line + a->indent_off,
921 a->len - a->indent_off,
922 b->line + b->indent_off,
923 b->len - b->indent_off, flags);
926 static void prepare_entry(struct diff_options *o, struct emitted_diff_symbol *l,
927 struct interned_diff_symbol *s)
929 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
930 unsigned int hash = xdiff_hash_string(l->line + l->indent_off,
931 l->len - l->indent_off, flags);
933 hashmap_entry_init(&s->ent, hash);
934 s->es = l;
937 struct moved_entry_list {
938 struct moved_entry *add, *del;
941 static struct moved_entry_list *add_lines_to_move_detection(struct diff_options *o,
942 struct mem_pool *entry_mem_pool)
944 struct moved_entry *prev_line = NULL;
945 struct mem_pool interned_pool;
946 struct hashmap interned_map;
947 struct moved_entry_list *entry_list = NULL;
948 size_t entry_list_alloc = 0;
949 unsigned id = 0;
950 int n;
952 hashmap_init(&interned_map, interned_diff_symbol_cmp, o, 8096);
953 mem_pool_init(&interned_pool, 1024 * 1024);
955 for (n = 0; n < o->emitted_symbols->nr; n++) {
956 struct interned_diff_symbol key;
957 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
958 struct interned_diff_symbol *s;
959 struct moved_entry *entry;
961 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS) {
962 prev_line = NULL;
963 continue;
966 if (o->color_moved_ws_handling &
967 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
968 fill_es_indent_data(l);
970 prepare_entry(o, l, &key);
971 s = hashmap_get_entry(&interned_map, &key, ent, &key.ent);
972 if (s) {
973 l->id = s->es->id;
974 } else {
975 l->id = id;
976 ALLOC_GROW_BY(entry_list, id, 1, entry_list_alloc);
977 hashmap_add(&interned_map,
978 memcpy(mem_pool_alloc(&interned_pool,
979 sizeof(key)),
980 &key, sizeof(key)));
982 entry = mem_pool_alloc(entry_mem_pool, sizeof(*entry));
983 entry->es = l;
984 entry->next_line = NULL;
985 if (prev_line && prev_line->es->s == l->s)
986 prev_line->next_line = entry;
987 prev_line = entry;
988 if (l->s == DIFF_SYMBOL_PLUS) {
989 entry->next_match = entry_list[l->id].add;
990 entry_list[l->id].add = entry;
991 } else {
992 entry->next_match = entry_list[l->id].del;
993 entry_list[l->id].del = entry;
997 hashmap_clear(&interned_map);
998 mem_pool_discard(&interned_pool, 0);
1000 return entry_list;
1003 static void pmb_advance_or_null(struct diff_options *o,
1004 struct emitted_diff_symbol *l,
1005 struct moved_block *pmb,
1006 int *pmb_nr)
1008 int i, j;
1010 for (i = 0, j = 0; i < *pmb_nr; i++) {
1011 int match;
1012 struct moved_entry *prev = pmb[i].match;
1013 struct moved_entry *cur = (prev && prev->next_line) ?
1014 prev->next_line : NULL;
1016 if (o->color_moved_ws_handling &
1017 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1018 match = cur &&
1019 !cmp_in_block_with_wsd(cur, l, &pmb[i]);
1020 else
1021 match = cur && cur->es->id == l->id;
1023 if (match) {
1024 pmb[j] = pmb[i];
1025 pmb[j++].match = cur;
1028 *pmb_nr = j;
1031 static void fill_potential_moved_blocks(struct diff_options *o,
1032 struct moved_entry *match,
1033 struct emitted_diff_symbol *l,
1034 struct moved_block **pmb_p,
1035 int *pmb_alloc_p, int *pmb_nr_p)
1038 struct moved_block *pmb = *pmb_p;
1039 int pmb_alloc = *pmb_alloc_p, pmb_nr = *pmb_nr_p;
1042 * The current line is the start of a new block.
1043 * Setup the set of potential blocks.
1045 for (; match; match = match->next_match) {
1046 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1047 if (o->color_moved_ws_handling &
1048 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1049 pmb[pmb_nr].wsd = compute_ws_delta(l, match->es);
1050 else
1051 pmb[pmb_nr].wsd = 0;
1052 pmb[pmb_nr++].match = match;
1055 *pmb_p = pmb;
1056 *pmb_alloc_p = pmb_alloc;
1057 *pmb_nr_p = pmb_nr;
1061 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1063 * Otherwise, if the last block has fewer alphanumeric characters than
1064 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1065 * that block.
1067 * The last block consists of the (n - block_length)'th line up to but not
1068 * including the nth line.
1070 * Returns 0 if the last block is empty or is unset by this function, non zero
1071 * otherwise.
1073 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1074 * Think of a way to unify them.
1076 #define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1077 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1078 static int adjust_last_block(struct diff_options *o, int n, int block_length)
1080 int i, alnum_count = 0;
1081 if (o->color_moved == COLOR_MOVED_PLAIN)
1082 return block_length;
1083 for (i = 1; i < block_length + 1; i++) {
1084 const char *c = o->emitted_symbols->buf[n - i].line;
1085 for (; *c; c++) {
1086 if (!isalnum(*c))
1087 continue;
1088 alnum_count++;
1089 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1090 return 1;
1093 for (i = 1; i < block_length + 1; i++)
1094 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK;
1095 return 0;
1098 /* Find blocks of moved code, delegate actual coloring decision to helper */
1099 static void mark_color_as_moved(struct diff_options *o,
1100 struct moved_entry_list *entry_list)
1102 struct moved_block *pmb = NULL; /* potentially moved blocks */
1103 int pmb_nr = 0, pmb_alloc = 0;
1104 int n, flipped_block = 0, block_length = 0;
1105 enum diff_symbol moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1108 for (n = 0; n < o->emitted_symbols->nr; n++) {
1109 struct moved_entry *match = NULL;
1110 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1112 switch (l->s) {
1113 case DIFF_SYMBOL_PLUS:
1114 match = entry_list[l->id].del;
1115 break;
1116 case DIFF_SYMBOL_MINUS:
1117 match = entry_list[l->id].add;
1118 break;
1119 default:
1120 flipped_block = 0;
1123 if (pmb_nr && (!match || l->s != moved_symbol)) {
1124 if (!adjust_last_block(o, n, block_length) &&
1125 block_length > 1) {
1127 * Rewind in case there is another match
1128 * starting at the second line of the block
1130 match = NULL;
1131 n -= block_length;
1133 pmb_nr = 0;
1134 block_length = 0;
1135 flipped_block = 0;
1137 if (!match) {
1138 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1139 continue;
1142 if (o->color_moved == COLOR_MOVED_PLAIN) {
1143 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1144 continue;
1147 pmb_advance_or_null(o, l, pmb, &pmb_nr);
1149 if (pmb_nr == 0) {
1150 int contiguous = adjust_last_block(o, n, block_length);
1152 if (!contiguous && block_length > 1)
1154 * Rewind in case there is another match
1155 * starting at the second line of the block
1157 n -= block_length;
1158 else
1159 fill_potential_moved_blocks(o, match, l,
1160 &pmb, &pmb_alloc,
1161 &pmb_nr);
1163 if (contiguous && pmb_nr && moved_symbol == l->s)
1164 flipped_block = (flipped_block + 1) % 2;
1165 else
1166 flipped_block = 0;
1168 if (pmb_nr)
1169 moved_symbol = l->s;
1170 else
1171 moved_symbol = DIFF_SYMBOL_BINARY_DIFF_HEADER;
1173 block_length = 0;
1176 if (pmb_nr) {
1177 block_length++;
1178 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1179 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1180 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1183 adjust_last_block(o, n, block_length);
1185 free(pmb);
1188 static void dim_moved_lines(struct diff_options *o)
1190 int n;
1191 for (n = 0; n < o->emitted_symbols->nr; n++) {
1192 struct emitted_diff_symbol *prev = (n != 0) ?
1193 &o->emitted_symbols->buf[n - 1] : NULL;
1194 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1195 struct emitted_diff_symbol *next =
1196 (n < o->emitted_symbols->nr - 1) ?
1197 &o->emitted_symbols->buf[n + 1] : NULL;
1199 /* Not a plus or minus line? */
1200 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1201 continue;
1203 /* Not a moved line? */
1204 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1205 continue;
1208 * If prev or next are not a plus or minus line,
1209 * pretend they don't exist
1211 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1212 prev->s != DIFF_SYMBOL_MINUS)
1213 prev = NULL;
1214 if (next && next->s != DIFF_SYMBOL_PLUS &&
1215 next->s != DIFF_SYMBOL_MINUS)
1216 next = NULL;
1218 /* Inside a block? */
1219 if ((prev &&
1220 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1221 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1222 (next &&
1223 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1224 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1225 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1226 continue;
1229 /* Check if we are at an interesting bound: */
1230 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1231 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1232 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1233 continue;
1234 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1235 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1236 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1237 continue;
1240 * The boundary to prev and next are not interesting,
1241 * so this line is not interesting as a whole
1243 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1247 static void emit_line_ws_markup(struct diff_options *o,
1248 const char *set_sign, const char *set,
1249 const char *reset,
1250 int sign_index, const char *line, int len,
1251 unsigned ws_rule, int blank_at_eof)
1253 const char *ws = NULL;
1254 int sign = o->output_indicators[sign_index];
1256 if (o->ws_error_highlight & ws_rule) {
1257 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1258 if (!*ws)
1259 ws = NULL;
1262 if (!ws && !set_sign)
1263 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1264 else if (!ws) {
1265 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1266 } else if (blank_at_eof)
1267 /* Blank line at EOF - paint '+' as well */
1268 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1269 else {
1270 /* Emit just the prefix, then the rest. */
1271 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1272 sign, "", 0);
1273 ws_check_emit(line, len, ws_rule,
1274 o->file, set, reset, ws);
1278 static void emit_diff_symbol_from_struct(struct diff_options *o,
1279 struct emitted_diff_symbol *eds)
1281 static const char *nneof = " No newline at end of file\n";
1282 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1283 struct strbuf sb = STRBUF_INIT;
1285 enum diff_symbol s = eds->s;
1286 const char *line = eds->line;
1287 int len = eds->len;
1288 unsigned flags = eds->flags;
1290 switch (s) {
1291 case DIFF_SYMBOL_NO_LF_EOF:
1292 context = diff_get_color_opt(o, DIFF_CONTEXT);
1293 reset = diff_get_color_opt(o, DIFF_RESET);
1294 putc('\n', o->file);
1295 emit_line_0(o, context, NULL, 0, reset, '\\',
1296 nneof, strlen(nneof));
1297 break;
1298 case DIFF_SYMBOL_SUBMODULE_HEADER:
1299 case DIFF_SYMBOL_SUBMODULE_ERROR:
1300 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1301 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1302 case DIFF_SYMBOL_SUMMARY:
1303 case DIFF_SYMBOL_STATS_LINE:
1304 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1305 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1306 emit_line(o, "", "", line, len);
1307 break;
1308 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1309 case DIFF_SYMBOL_CONTEXT_MARKER:
1310 context = diff_get_color_opt(o, DIFF_CONTEXT);
1311 reset = diff_get_color_opt(o, DIFF_RESET);
1312 emit_line(o, context, reset, line, len);
1313 break;
1314 case DIFF_SYMBOL_SEPARATOR:
1315 fprintf(o->file, "%s%c",
1316 diff_line_prefix(o),
1317 o->line_termination);
1318 break;
1319 case DIFF_SYMBOL_CONTEXT:
1320 set = diff_get_color_opt(o, DIFF_CONTEXT);
1321 reset = diff_get_color_opt(o, DIFF_RESET);
1322 set_sign = NULL;
1323 if (o->flags.dual_color_diffed_diffs) {
1324 char c = !len ? 0 : line[0];
1326 if (c == '+')
1327 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1328 else if (c == '@')
1329 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1330 else if (c == '-')
1331 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1333 emit_line_ws_markup(o, set_sign, set, reset,
1334 OUTPUT_INDICATOR_CONTEXT, line, len,
1335 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1336 break;
1337 case DIFF_SYMBOL_PLUS:
1338 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1339 DIFF_SYMBOL_MOVED_LINE_ALT |
1340 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1341 case DIFF_SYMBOL_MOVED_LINE |
1342 DIFF_SYMBOL_MOVED_LINE_ALT |
1343 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1344 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1345 break;
1346 case DIFF_SYMBOL_MOVED_LINE |
1347 DIFF_SYMBOL_MOVED_LINE_ALT:
1348 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1349 break;
1350 case DIFF_SYMBOL_MOVED_LINE |
1351 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1352 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1353 break;
1354 case DIFF_SYMBOL_MOVED_LINE:
1355 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1356 break;
1357 default:
1358 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1360 reset = diff_get_color_opt(o, DIFF_RESET);
1361 if (!o->flags.dual_color_diffed_diffs)
1362 set_sign = NULL;
1363 else {
1364 char c = !len ? 0 : line[0];
1366 set_sign = set;
1367 if (c == '-')
1368 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1369 else if (c == '@')
1370 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1371 else if (c == '+')
1372 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1373 else
1374 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1375 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1377 emit_line_ws_markup(o, set_sign, set, reset,
1378 OUTPUT_INDICATOR_NEW, line, len,
1379 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1380 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1381 break;
1382 case DIFF_SYMBOL_MINUS:
1383 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1384 DIFF_SYMBOL_MOVED_LINE_ALT |
1385 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1386 case DIFF_SYMBOL_MOVED_LINE |
1387 DIFF_SYMBOL_MOVED_LINE_ALT |
1388 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1389 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1390 break;
1391 case DIFF_SYMBOL_MOVED_LINE |
1392 DIFF_SYMBOL_MOVED_LINE_ALT:
1393 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1394 break;
1395 case DIFF_SYMBOL_MOVED_LINE |
1396 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1397 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1398 break;
1399 case DIFF_SYMBOL_MOVED_LINE:
1400 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1401 break;
1402 default:
1403 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1405 reset = diff_get_color_opt(o, DIFF_RESET);
1406 if (!o->flags.dual_color_diffed_diffs)
1407 set_sign = NULL;
1408 else {
1409 char c = !len ? 0 : line[0];
1411 set_sign = set;
1412 if (c == '+')
1413 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1414 else if (c == '@')
1415 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1416 else if (c == '-')
1417 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1418 else
1419 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1421 emit_line_ws_markup(o, set_sign, set, reset,
1422 OUTPUT_INDICATOR_OLD, line, len,
1423 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1424 break;
1425 case DIFF_SYMBOL_WORDS_PORCELAIN:
1426 context = diff_get_color_opt(o, DIFF_CONTEXT);
1427 reset = diff_get_color_opt(o, DIFF_RESET);
1428 emit_line(o, context, reset, line, len);
1429 fputs("~\n", o->file);
1430 break;
1431 case DIFF_SYMBOL_WORDS:
1432 context = diff_get_color_opt(o, DIFF_CONTEXT);
1433 reset = diff_get_color_opt(o, DIFF_RESET);
1435 * Skip the prefix character, if any. With
1436 * diff_suppress_blank_empty, there may be
1437 * none.
1439 if (line[0] != '\n') {
1440 line++;
1441 len--;
1443 emit_line(o, context, reset, line, len);
1444 break;
1445 case DIFF_SYMBOL_FILEPAIR_PLUS:
1446 meta = diff_get_color_opt(o, DIFF_METAINFO);
1447 reset = diff_get_color_opt(o, DIFF_RESET);
1448 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1449 line, reset,
1450 strchr(line, ' ') ? "\t" : "");
1451 break;
1452 case DIFF_SYMBOL_FILEPAIR_MINUS:
1453 meta = diff_get_color_opt(o, DIFF_METAINFO);
1454 reset = diff_get_color_opt(o, DIFF_RESET);
1455 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1456 line, reset,
1457 strchr(line, ' ') ? "\t" : "");
1458 break;
1459 case DIFF_SYMBOL_BINARY_FILES:
1460 case DIFF_SYMBOL_HEADER:
1461 fprintf(o->file, "%s", line);
1462 break;
1463 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1464 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1465 break;
1466 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1467 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1468 break;
1469 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1470 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1471 break;
1472 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1473 fputs(diff_line_prefix(o), o->file);
1474 fputc('\n', o->file);
1475 break;
1476 case DIFF_SYMBOL_REWRITE_DIFF:
1477 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1478 reset = diff_get_color_opt(o, DIFF_RESET);
1479 emit_line(o, fraginfo, reset, line, len);
1480 break;
1481 case DIFF_SYMBOL_SUBMODULE_ADD:
1482 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1483 reset = diff_get_color_opt(o, DIFF_RESET);
1484 emit_line(o, set, reset, line, len);
1485 break;
1486 case DIFF_SYMBOL_SUBMODULE_DEL:
1487 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1488 reset = diff_get_color_opt(o, DIFF_RESET);
1489 emit_line(o, set, reset, line, len);
1490 break;
1491 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1492 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1493 diff_line_prefix(o), line);
1494 break;
1495 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1496 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1497 diff_line_prefix(o), line);
1498 break;
1499 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1500 emit_line(o, "", "", " 0 files changed\n",
1501 strlen(" 0 files changed\n"));
1502 break;
1503 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1504 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1505 break;
1506 case DIFF_SYMBOL_WORD_DIFF:
1507 fprintf(o->file, "%.*s", len, line);
1508 break;
1509 case DIFF_SYMBOL_STAT_SEP:
1510 fputs(o->stat_sep, o->file);
1511 break;
1512 default:
1513 BUG("unknown diff symbol");
1515 strbuf_release(&sb);
1518 static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1519 const char *line, int len, unsigned flags)
1521 struct emitted_diff_symbol e = {
1522 .line = line, .len = len, .flags = flags, .s = s
1525 if (o->emitted_symbols)
1526 append_emitted_diff_symbol(o, &e);
1527 else
1528 emit_diff_symbol_from_struct(o, &e);
1531 void diff_emit_submodule_del(struct diff_options *o, const char *line)
1533 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1536 void diff_emit_submodule_add(struct diff_options *o, const char *line)
1538 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1541 void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1543 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1544 path, strlen(path), 0);
1547 void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1549 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1550 path, strlen(path), 0);
1553 void diff_emit_submodule_header(struct diff_options *o, const char *header)
1555 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1556 header, strlen(header), 0);
1559 void diff_emit_submodule_error(struct diff_options *o, const char *err)
1561 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1564 void diff_emit_submodule_pipethrough(struct diff_options *o,
1565 const char *line, int len)
1567 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1570 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1572 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1573 ecbdata->blank_at_eof_in_preimage &&
1574 ecbdata->blank_at_eof_in_postimage &&
1575 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1576 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1577 return 0;
1578 return ws_blank_line(line, len, ecbdata->ws_rule);
1581 static void emit_add_line(struct emit_callback *ecbdata,
1582 const char *line, int len)
1584 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1585 if (new_blank_line_at_eof(ecbdata, line, len))
1586 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1588 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1591 static void emit_del_line(struct emit_callback *ecbdata,
1592 const char *line, int len)
1594 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1595 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1598 static void emit_context_line(struct emit_callback *ecbdata,
1599 const char *line, int len)
1601 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1602 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1605 static void emit_hunk_header(struct emit_callback *ecbdata,
1606 const char *line, int len)
1608 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1609 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1610 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1611 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1612 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1613 static const char atat[2] = { '@', '@' };
1614 const char *cp, *ep;
1615 struct strbuf msgbuf = STRBUF_INIT;
1616 int org_len = len;
1617 int i = 1;
1620 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1621 * it always is at least 10 bytes long.
1623 if (len < 10 ||
1624 memcmp(line, atat, 2) ||
1625 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1626 emit_diff_symbol(ecbdata->opt,
1627 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1628 return;
1630 ep += 2; /* skip over @@ */
1632 /* The hunk header in fraginfo color */
1633 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1634 strbuf_addstr(&msgbuf, reverse);
1635 strbuf_addstr(&msgbuf, frag);
1636 if (ecbdata->opt->flags.suppress_hunk_header_line_count)
1637 strbuf_add(&msgbuf, atat, sizeof(atat));
1638 else
1639 strbuf_add(&msgbuf, line, ep - line);
1640 strbuf_addstr(&msgbuf, reset);
1643 * trailing "\r\n"
1645 for ( ; i < 3; i++)
1646 if (line[len - i] == '\r' || line[len - i] == '\n')
1647 len--;
1649 /* blank before the func header */
1650 for (cp = ep; ep - line < len; ep++)
1651 if (*ep != ' ' && *ep != '\t')
1652 break;
1653 if (ep != cp) {
1654 strbuf_addstr(&msgbuf, context);
1655 strbuf_add(&msgbuf, cp, ep - cp);
1656 strbuf_addstr(&msgbuf, reset);
1659 if (ep < line + len) {
1660 strbuf_addstr(&msgbuf, func);
1661 strbuf_add(&msgbuf, ep, line + len - ep);
1662 strbuf_addstr(&msgbuf, reset);
1665 strbuf_add(&msgbuf, line + len, org_len - len);
1666 strbuf_complete_line(&msgbuf);
1667 emit_diff_symbol(ecbdata->opt,
1668 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1669 strbuf_release(&msgbuf);
1672 static struct diff_tempfile *claim_diff_tempfile(void)
1674 int i;
1675 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1676 if (!diff_temp[i].name)
1677 return diff_temp + i;
1678 BUG("diff is failing to clean up its tempfiles");
1681 static void remove_tempfile(void)
1683 int i;
1684 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1685 if (is_tempfile_active(diff_temp[i].tempfile))
1686 delete_tempfile(&diff_temp[i].tempfile);
1687 diff_temp[i].name = NULL;
1691 static void add_line_count(struct strbuf *out, int count)
1693 switch (count) {
1694 case 0:
1695 strbuf_addstr(out, "0,0");
1696 break;
1697 case 1:
1698 strbuf_addstr(out, "1");
1699 break;
1700 default:
1701 strbuf_addf(out, "1,%d", count);
1702 break;
1706 static void emit_rewrite_lines(struct emit_callback *ecb,
1707 int prefix, const char *data, int size)
1709 const char *endp = NULL;
1711 while (0 < size) {
1712 int len;
1714 endp = memchr(data, '\n', size);
1715 len = endp ? (endp - data + 1) : size;
1716 if (prefix != '+') {
1717 ecb->lno_in_preimage++;
1718 emit_del_line(ecb, data, len);
1719 } else {
1720 ecb->lno_in_postimage++;
1721 emit_add_line(ecb, data, len);
1723 size -= len;
1724 data += len;
1726 if (!endp)
1727 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1730 static void emit_rewrite_diff(const char *name_a,
1731 const char *name_b,
1732 struct diff_filespec *one,
1733 struct diff_filespec *two,
1734 struct userdiff_driver *textconv_one,
1735 struct userdiff_driver *textconv_two,
1736 struct diff_options *o)
1738 int lc_a, lc_b;
1739 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1740 const char *a_prefix, *b_prefix;
1741 char *data_one, *data_two;
1742 size_t size_one, size_two;
1743 struct emit_callback ecbdata;
1744 struct strbuf out = STRBUF_INIT;
1746 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1747 a_prefix = o->b_prefix;
1748 b_prefix = o->a_prefix;
1749 } else {
1750 a_prefix = o->a_prefix;
1751 b_prefix = o->b_prefix;
1754 name_a += (*name_a == '/');
1755 name_b += (*name_b == '/');
1757 strbuf_reset(&a_name);
1758 strbuf_reset(&b_name);
1759 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1760 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1762 size_one = fill_textconv(o->repo, textconv_one, one, &data_one);
1763 size_two = fill_textconv(o->repo, textconv_two, two, &data_two);
1765 memset(&ecbdata, 0, sizeof(ecbdata));
1766 ecbdata.color_diff = want_color(o->use_color);
1767 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
1768 ecbdata.opt = o;
1769 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1770 mmfile_t mf1, mf2;
1771 mf1.ptr = (char *)data_one;
1772 mf2.ptr = (char *)data_two;
1773 mf1.size = size_one;
1774 mf2.size = size_two;
1775 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1777 ecbdata.lno_in_preimage = 1;
1778 ecbdata.lno_in_postimage = 1;
1780 lc_a = count_lines(data_one, size_one);
1781 lc_b = count_lines(data_two, size_two);
1783 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1784 a_name.buf, a_name.len, 0);
1785 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1786 b_name.buf, b_name.len, 0);
1788 strbuf_addstr(&out, "@@ -");
1789 if (!o->irreversible_delete)
1790 add_line_count(&out, lc_a);
1791 else
1792 strbuf_addstr(&out, "?,?");
1793 strbuf_addstr(&out, " +");
1794 add_line_count(&out, lc_b);
1795 strbuf_addstr(&out, " @@\n");
1796 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1797 strbuf_release(&out);
1799 if (lc_a && !o->irreversible_delete)
1800 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1801 if (lc_b)
1802 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1803 if (textconv_one)
1804 free((char *)data_one);
1805 if (textconv_two)
1806 free((char *)data_two);
1809 struct diff_words_buffer {
1810 mmfile_t text;
1811 unsigned long alloc;
1812 struct diff_words_orig {
1813 const char *begin, *end;
1814 } *orig;
1815 int orig_nr, orig_alloc;
1818 static void diff_words_append(char *line, unsigned long len,
1819 struct diff_words_buffer *buffer)
1821 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1822 line++;
1823 len--;
1824 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1825 buffer->text.size += len;
1826 buffer->text.ptr[buffer->text.size] = '\0';
1829 struct diff_words_style_elem {
1830 const char *prefix;
1831 const char *suffix;
1832 const char *color; /* NULL; filled in by the setup code if
1833 * color is enabled */
1836 struct diff_words_style {
1837 enum diff_words_type type;
1838 struct diff_words_style_elem new_word, old_word, ctx;
1839 const char *newline;
1842 static struct diff_words_style diff_words_styles[] = {
1843 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1844 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1845 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1848 struct diff_words_data {
1849 struct diff_words_buffer minus, plus;
1850 const char *current_plus;
1851 int last_minus;
1852 struct diff_options *opt;
1853 regex_t *word_regex;
1854 enum diff_words_type type;
1855 struct diff_words_style *style;
1858 static int fn_out_diff_words_write_helper(struct diff_options *o,
1859 struct diff_words_style_elem *st_el,
1860 const char *newline,
1861 size_t count, const char *buf)
1863 int print = 0;
1864 struct strbuf sb = STRBUF_INIT;
1866 while (count) {
1867 char *p = memchr(buf, '\n', count);
1868 if (print)
1869 strbuf_addstr(&sb, diff_line_prefix(o));
1871 if (p != buf) {
1872 const char *reset = st_el->color && *st_el->color ?
1873 GIT_COLOR_RESET : NULL;
1874 if (st_el->color && *st_el->color)
1875 strbuf_addstr(&sb, st_el->color);
1876 strbuf_addstr(&sb, st_el->prefix);
1877 strbuf_add(&sb, buf, p ? p - buf : count);
1878 strbuf_addstr(&sb, st_el->suffix);
1879 if (reset)
1880 strbuf_addstr(&sb, reset);
1882 if (!p)
1883 goto out;
1885 strbuf_addstr(&sb, newline);
1886 count -= p + 1 - buf;
1887 buf = p + 1;
1888 print = 1;
1889 if (count) {
1890 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1891 sb.buf, sb.len, 0);
1892 strbuf_reset(&sb);
1896 out:
1897 if (sb.len)
1898 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1899 sb.buf, sb.len, 0);
1900 strbuf_release(&sb);
1901 return 0;
1905 * '--color-words' algorithm can be described as:
1907 * 1. collect the minus/plus lines of a diff hunk, divided into
1908 * minus-lines and plus-lines;
1910 * 2. break both minus-lines and plus-lines into words and
1911 * place them into two mmfile_t with one word for each line;
1913 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1915 * And for the common parts of the both file, we output the plus side text.
1916 * diff_words->current_plus is used to trace the current position of the plus file
1917 * which printed. diff_words->last_minus is used to trace the last minus word
1918 * printed.
1920 * For '--graph' to work with '--color-words', we need to output the graph prefix
1921 * on each line of color words output. Generally, there are two conditions on
1922 * which we should output the prefix.
1924 * 1. diff_words->last_minus == 0 &&
1925 * diff_words->current_plus == diff_words->plus.text.ptr
1927 * that is: the plus text must start as a new line, and if there is no minus
1928 * word printed, a graph prefix must be printed.
1930 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1931 * *(diff_words->current_plus - 1) == '\n'
1933 * that is: a graph prefix must be printed following a '\n'
1935 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1937 if ((diff_words->last_minus == 0 &&
1938 diff_words->current_plus == diff_words->plus.text.ptr) ||
1939 (diff_words->current_plus > diff_words->plus.text.ptr &&
1940 *(diff_words->current_plus - 1) == '\n')) {
1941 return 1;
1942 } else {
1943 return 0;
1947 static void fn_out_diff_words_aux(void *priv,
1948 long minus_first, long minus_len,
1949 long plus_first, long plus_len,
1950 const char *func, long funclen)
1952 struct diff_words_data *diff_words = priv;
1953 struct diff_words_style *style = diff_words->style;
1954 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1955 struct diff_options *opt = diff_words->opt;
1956 const char *line_prefix;
1958 assert(opt);
1959 line_prefix = diff_line_prefix(opt);
1961 /* POSIX requires that first be decremented by one if len == 0... */
1962 if (minus_len) {
1963 minus_begin = diff_words->minus.orig[minus_first].begin;
1964 minus_end =
1965 diff_words->minus.orig[minus_first + minus_len - 1].end;
1966 } else
1967 minus_begin = minus_end =
1968 diff_words->minus.orig[minus_first].end;
1970 if (plus_len) {
1971 plus_begin = diff_words->plus.orig[plus_first].begin;
1972 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1973 } else
1974 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1976 if (color_words_output_graph_prefix(diff_words)) {
1977 fputs(line_prefix, diff_words->opt->file);
1979 if (diff_words->current_plus != plus_begin) {
1980 fn_out_diff_words_write_helper(diff_words->opt,
1981 &style->ctx, style->newline,
1982 plus_begin - diff_words->current_plus,
1983 diff_words->current_plus);
1985 if (minus_begin != minus_end) {
1986 fn_out_diff_words_write_helper(diff_words->opt,
1987 &style->old_word, style->newline,
1988 minus_end - minus_begin, minus_begin);
1990 if (plus_begin != plus_end) {
1991 fn_out_diff_words_write_helper(diff_words->opt,
1992 &style->new_word, style->newline,
1993 plus_end - plus_begin, plus_begin);
1996 diff_words->current_plus = plus_end;
1997 diff_words->last_minus = minus_first;
2000 /* This function starts looking at *begin, and returns 0 iff a word was found. */
2001 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
2002 int *begin, int *end)
2004 while (word_regex && *begin < buffer->size) {
2005 regmatch_t match[1];
2006 if (!regexec_buf(word_regex, buffer->ptr + *begin,
2007 buffer->size - *begin, 1, match, 0)) {
2008 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
2009 '\n', match[0].rm_eo - match[0].rm_so);
2010 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
2011 *begin += match[0].rm_so;
2012 if (*begin == *end)
2013 (*begin)++;
2014 else
2015 return *begin > *end;
2016 } else {
2017 return -1;
2021 /* find the next word */
2022 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
2023 (*begin)++;
2024 if (*begin >= buffer->size)
2025 return -1;
2027 /* find the end of the word */
2028 *end = *begin + 1;
2029 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
2030 (*end)++;
2032 return 0;
2036 * This function splits the words in buffer->text, stores the list with
2037 * newline separator into out, and saves the offsets of the original words
2038 * in buffer->orig.
2040 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
2041 regex_t *word_regex)
2043 int i, j;
2044 long alloc = 0;
2046 out->size = 0;
2047 out->ptr = NULL;
2049 /* fake an empty "0th" word */
2050 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
2051 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
2052 buffer->orig_nr = 1;
2054 for (i = 0; i < buffer->text.size; i++) {
2055 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2056 return;
2058 /* store original boundaries */
2059 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2060 buffer->orig_alloc);
2061 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2062 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2063 buffer->orig_nr++;
2065 /* store one word */
2066 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2067 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2068 out->ptr[out->size + j - i] = '\n';
2069 out->size += j - i + 1;
2071 i = j - 1;
2075 /* this executes the word diff on the accumulated buffers */
2076 static void diff_words_show(struct diff_words_data *diff_words)
2078 xpparam_t xpp;
2079 xdemitconf_t xecfg;
2080 mmfile_t minus, plus;
2081 struct diff_words_style *style = diff_words->style;
2083 struct diff_options *opt = diff_words->opt;
2084 const char *line_prefix;
2086 assert(opt);
2087 line_prefix = diff_line_prefix(opt);
2089 /* special case: only removal */
2090 if (!diff_words->plus.text.size) {
2091 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2092 line_prefix, strlen(line_prefix), 0);
2093 fn_out_diff_words_write_helper(diff_words->opt,
2094 &style->old_word, style->newline,
2095 diff_words->minus.text.size,
2096 diff_words->minus.text.ptr);
2097 diff_words->minus.text.size = 0;
2098 return;
2101 diff_words->current_plus = diff_words->plus.text.ptr;
2102 diff_words->last_minus = 0;
2104 memset(&xpp, 0, sizeof(xpp));
2105 memset(&xecfg, 0, sizeof(xecfg));
2106 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2107 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2108 xpp.flags = 0;
2109 /* as only the hunk header will be parsed, we need a 0-context */
2110 xecfg.ctxlen = 0;
2111 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2112 diff_words, &xpp, &xecfg))
2113 die("unable to generate word diff");
2114 free(minus.ptr);
2115 free(plus.ptr);
2116 if (diff_words->current_plus != diff_words->plus.text.ptr +
2117 diff_words->plus.text.size) {
2118 if (color_words_output_graph_prefix(diff_words))
2119 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2120 line_prefix, strlen(line_prefix), 0);
2121 fn_out_diff_words_write_helper(diff_words->opt,
2122 &style->ctx, style->newline,
2123 diff_words->plus.text.ptr + diff_words->plus.text.size
2124 - diff_words->current_plus, diff_words->current_plus);
2126 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2129 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
2130 static void diff_words_flush(struct emit_callback *ecbdata)
2132 struct diff_options *wo = ecbdata->diff_words->opt;
2134 if (ecbdata->diff_words->minus.text.size ||
2135 ecbdata->diff_words->plus.text.size)
2136 diff_words_show(ecbdata->diff_words);
2138 if (wo->emitted_symbols) {
2139 struct diff_options *o = ecbdata->opt;
2140 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2141 int i;
2144 * NEEDSWORK:
2145 * Instead of appending each, concat all words to a line?
2147 for (i = 0; i < wol->nr; i++)
2148 append_emitted_diff_symbol(o, &wol->buf[i]);
2150 for (i = 0; i < wol->nr; i++)
2151 free((void *)wol->buf[i].line);
2153 wol->nr = 0;
2157 static void diff_filespec_load_driver(struct diff_filespec *one,
2158 struct index_state *istate)
2160 /* Use already-loaded driver */
2161 if (one->driver)
2162 return;
2164 if (S_ISREG(one->mode))
2165 one->driver = userdiff_find_by_path(istate, one->path);
2167 /* Fallback to default settings */
2168 if (!one->driver)
2169 one->driver = userdiff_find_by_name("default");
2172 static const char *userdiff_word_regex(struct diff_filespec *one,
2173 struct index_state *istate)
2175 diff_filespec_load_driver(one, istate);
2176 return one->driver->word_regex;
2179 static void init_diff_words_data(struct emit_callback *ecbdata,
2180 struct diff_options *orig_opts,
2181 struct diff_filespec *one,
2182 struct diff_filespec *two)
2184 int i;
2185 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2186 memcpy(o, orig_opts, sizeof(struct diff_options));
2188 CALLOC_ARRAY(ecbdata->diff_words, 1);
2189 ecbdata->diff_words->type = o->word_diff;
2190 ecbdata->diff_words->opt = o;
2192 if (orig_opts->emitted_symbols)
2193 CALLOC_ARRAY(o->emitted_symbols, 1);
2195 if (!o->word_regex)
2196 o->word_regex = userdiff_word_regex(one, o->repo->index);
2197 if (!o->word_regex)
2198 o->word_regex = userdiff_word_regex(two, o->repo->index);
2199 if (!o->word_regex)
2200 o->word_regex = diff_word_regex_cfg;
2201 if (o->word_regex) {
2202 ecbdata->diff_words->word_regex = (regex_t *)
2203 xmalloc(sizeof(regex_t));
2204 if (regcomp(ecbdata->diff_words->word_regex,
2205 o->word_regex,
2206 REG_EXTENDED | REG_NEWLINE))
2207 die("invalid regular expression: %s",
2208 o->word_regex);
2210 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2211 if (o->word_diff == diff_words_styles[i].type) {
2212 ecbdata->diff_words->style =
2213 &diff_words_styles[i];
2214 break;
2217 if (want_color(o->use_color)) {
2218 struct diff_words_style *st = ecbdata->diff_words->style;
2219 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2220 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2221 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2225 static void free_diff_words_data(struct emit_callback *ecbdata)
2227 if (ecbdata->diff_words) {
2228 diff_words_flush(ecbdata);
2229 free (ecbdata->diff_words->opt->emitted_symbols);
2230 free (ecbdata->diff_words->opt);
2231 free (ecbdata->diff_words->minus.text.ptr);
2232 free (ecbdata->diff_words->minus.orig);
2233 free (ecbdata->diff_words->plus.text.ptr);
2234 free (ecbdata->diff_words->plus.orig);
2235 if (ecbdata->diff_words->word_regex) {
2236 regfree(ecbdata->diff_words->word_regex);
2237 free(ecbdata->diff_words->word_regex);
2239 FREE_AND_NULL(ecbdata->diff_words);
2243 const char *diff_get_color(int diff_use_color, enum color_diff ix)
2245 if (want_color(diff_use_color))
2246 return diff_colors[ix];
2247 return "";
2250 const char *diff_line_prefix(struct diff_options *opt)
2252 struct strbuf *msgbuf;
2253 if (!opt->output_prefix)
2254 return "";
2256 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2257 return msgbuf->buf;
2260 static unsigned long sane_truncate_line(char *line, unsigned long len)
2262 const char *cp;
2263 unsigned long allot;
2264 size_t l = len;
2266 cp = line;
2267 allot = l;
2268 while (0 < l) {
2269 (void) utf8_width(&cp, &l);
2270 if (!cp)
2271 break; /* truncated in the middle? */
2273 return allot - l;
2276 static void find_lno(const char *line, struct emit_callback *ecbdata)
2278 const char *p;
2279 ecbdata->lno_in_preimage = 0;
2280 ecbdata->lno_in_postimage = 0;
2281 p = strchr(line, '-');
2282 if (!p)
2283 return; /* cannot happen */
2284 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2285 p = strchr(p, '+');
2286 if (!p)
2287 return; /* cannot happen */
2288 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2291 static int fn_out_consume(void *priv, char *line, unsigned long len)
2293 struct emit_callback *ecbdata = priv;
2294 struct diff_options *o = ecbdata->opt;
2296 o->found_changes = 1;
2298 if (ecbdata->header) {
2299 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2300 ecbdata->header->buf, ecbdata->header->len, 0);
2301 strbuf_reset(ecbdata->header);
2302 ecbdata->header = NULL;
2305 if (ecbdata->label_path[0]) {
2306 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2307 ecbdata->label_path[0],
2308 strlen(ecbdata->label_path[0]), 0);
2309 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2310 ecbdata->label_path[1],
2311 strlen(ecbdata->label_path[1]), 0);
2312 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2315 if (diff_suppress_blank_empty
2316 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2317 line[0] = '\n';
2318 len = 1;
2321 if (line[0] == '@') {
2322 if (ecbdata->diff_words)
2323 diff_words_flush(ecbdata);
2324 len = sane_truncate_line(line, len);
2325 find_lno(line, ecbdata);
2326 emit_hunk_header(ecbdata, line, len);
2327 return 0;
2330 if (ecbdata->diff_words) {
2331 enum diff_symbol s =
2332 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2333 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2334 if (line[0] == '-') {
2335 diff_words_append(line, len,
2336 &ecbdata->diff_words->minus);
2337 return 0;
2338 } else if (line[0] == '+') {
2339 diff_words_append(line, len,
2340 &ecbdata->diff_words->plus);
2341 return 0;
2342 } else if (starts_with(line, "\\ ")) {
2344 * Eat the "no newline at eof" marker as if we
2345 * saw a "+" or "-" line with nothing on it,
2346 * and return without diff_words_flush() to
2347 * defer processing. If this is the end of
2348 * preimage, more "+" lines may come after it.
2350 return 0;
2352 diff_words_flush(ecbdata);
2353 emit_diff_symbol(o, s, line, len, 0);
2354 return 0;
2357 switch (line[0]) {
2358 case '+':
2359 ecbdata->lno_in_postimage++;
2360 emit_add_line(ecbdata, line + 1, len - 1);
2361 break;
2362 case '-':
2363 ecbdata->lno_in_preimage++;
2364 emit_del_line(ecbdata, line + 1, len - 1);
2365 break;
2366 case ' ':
2367 ecbdata->lno_in_postimage++;
2368 ecbdata->lno_in_preimage++;
2369 emit_context_line(ecbdata, line + 1, len - 1);
2370 break;
2371 default:
2372 /* incomplete line at the end */
2373 ecbdata->lno_in_preimage++;
2374 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2375 line, len, 0);
2376 break;
2378 return 0;
2381 static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2383 const char *old_name = a;
2384 const char *new_name = b;
2385 int pfx_length, sfx_length;
2386 int pfx_adjust_for_slash;
2387 int len_a = strlen(a);
2388 int len_b = strlen(b);
2389 int a_midlen, b_midlen;
2390 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2391 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2393 if (qlen_a || qlen_b) {
2394 quote_c_style(a, name, NULL, 0);
2395 strbuf_addstr(name, " => ");
2396 quote_c_style(b, name, NULL, 0);
2397 return;
2400 /* Find common prefix */
2401 pfx_length = 0;
2402 while (*old_name && *new_name && *old_name == *new_name) {
2403 if (*old_name == '/')
2404 pfx_length = old_name - a + 1;
2405 old_name++;
2406 new_name++;
2409 /* Find common suffix */
2410 old_name = a + len_a;
2411 new_name = b + len_b;
2412 sfx_length = 0;
2414 * If there is a common prefix, it must end in a slash. In
2415 * that case we let this loop run 1 into the prefix to see the
2416 * same slash.
2418 * If there is no common prefix, we cannot do this as it would
2419 * underrun the input strings.
2421 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2422 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2423 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2424 *old_name == *new_name) {
2425 if (*old_name == '/')
2426 sfx_length = len_a - (old_name - a);
2427 old_name--;
2428 new_name--;
2432 * pfx{mid-a => mid-b}sfx
2433 * {pfx-a => pfx-b}sfx
2434 * pfx{sfx-a => sfx-b}
2435 * name-a => name-b
2437 a_midlen = len_a - pfx_length - sfx_length;
2438 b_midlen = len_b - pfx_length - sfx_length;
2439 if (a_midlen < 0)
2440 a_midlen = 0;
2441 if (b_midlen < 0)
2442 b_midlen = 0;
2444 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2445 if (pfx_length + sfx_length) {
2446 strbuf_add(name, a, pfx_length);
2447 strbuf_addch(name, '{');
2449 strbuf_add(name, a + pfx_length, a_midlen);
2450 strbuf_addstr(name, " => ");
2451 strbuf_add(name, b + pfx_length, b_midlen);
2452 if (pfx_length + sfx_length) {
2453 strbuf_addch(name, '}');
2454 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2458 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2459 const char *name_a,
2460 const char *name_b)
2462 struct diffstat_file *x;
2463 CALLOC_ARRAY(x, 1);
2464 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2465 diffstat->files[diffstat->nr++] = x;
2466 if (name_b) {
2467 x->from_name = xstrdup(name_a);
2468 x->name = xstrdup(name_b);
2469 x->is_renamed = 1;
2471 else {
2472 x->from_name = NULL;
2473 x->name = xstrdup(name_a);
2475 return x;
2478 static int diffstat_consume(void *priv, char *line, unsigned long len)
2480 struct diffstat_t *diffstat = priv;
2481 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2483 if (line[0] == '+')
2484 x->added++;
2485 else if (line[0] == '-')
2486 x->deleted++;
2487 return 0;
2490 const char mime_boundary_leader[] = "------------";
2492 static int scale_linear(int it, int width, int max_change)
2494 if (!it)
2495 return 0;
2497 * make sure that at least one '-' or '+' is printed if
2498 * there is any change to this path. The easiest way is to
2499 * scale linearly as if the allotted width is one column shorter
2500 * than it is, and then add 1 to the result.
2502 return 1 + (it * (width - 1) / max_change);
2505 static void show_graph(struct strbuf *out, char ch, int cnt,
2506 const char *set, const char *reset)
2508 if (cnt <= 0)
2509 return;
2510 strbuf_addstr(out, set);
2511 strbuf_addchars(out, ch, cnt);
2512 strbuf_addstr(out, reset);
2515 static void fill_print_name(struct diffstat_file *file)
2517 struct strbuf pname = STRBUF_INIT;
2519 if (file->print_name)
2520 return;
2522 if (file->is_renamed)
2523 pprint_rename(&pname, file->from_name, file->name);
2524 else
2525 quote_c_style(file->name, &pname, NULL, 0);
2527 if (file->comments)
2528 strbuf_addf(&pname, " (%s)", file->comments);
2530 file->print_name = strbuf_detach(&pname, NULL);
2533 static void print_stat_summary_inserts_deletes(struct diff_options *options,
2534 int files, int insertions, int deletions)
2536 struct strbuf sb = STRBUF_INIT;
2538 if (!files) {
2539 assert(insertions == 0 && deletions == 0);
2540 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2541 NULL, 0, 0);
2542 return;
2545 strbuf_addf(&sb,
2546 (files == 1) ? " %d file changed" : " %d files changed",
2547 files);
2550 * For binary diff, the caller may want to print "x files
2551 * changed" with insertions == 0 && deletions == 0.
2553 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2554 * is probably less confusing (i.e skip over "2 files changed
2555 * but nothing about added/removed lines? Is this a bug in Git?").
2557 if (insertions || deletions == 0) {
2558 strbuf_addf(&sb,
2559 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2560 insertions);
2563 if (deletions || insertions == 0) {
2564 strbuf_addf(&sb,
2565 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2566 deletions);
2568 strbuf_addch(&sb, '\n');
2569 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2570 sb.buf, sb.len, 0);
2571 strbuf_release(&sb);
2574 void print_stat_summary(FILE *fp, int files,
2575 int insertions, int deletions)
2577 struct diff_options o;
2578 memset(&o, 0, sizeof(o));
2579 o.file = fp;
2581 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2584 static void show_stats(struct diffstat_t *data, struct diff_options *options)
2586 int i, len, add, del, adds = 0, dels = 0;
2587 uintmax_t max_change = 0, max_len = 0;
2588 int total_files = data->nr, count;
2589 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2590 const char *reset, *add_c, *del_c;
2591 int extra_shown = 0;
2592 const char *line_prefix = diff_line_prefix(options);
2593 struct strbuf out = STRBUF_INIT;
2595 if (data->nr == 0)
2596 return;
2598 count = options->stat_count ? options->stat_count : data->nr;
2600 reset = diff_get_color_opt(options, DIFF_RESET);
2601 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2602 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2605 * Find the longest filename and max number of changes
2607 for (i = 0; (i < count) && (i < data->nr); i++) {
2608 struct diffstat_file *file = data->files[i];
2609 uintmax_t change = file->added + file->deleted;
2611 if (!file->is_interesting && (change == 0)) {
2612 count++; /* not shown == room for one more */
2613 continue;
2615 fill_print_name(file);
2616 len = strlen(file->print_name);
2617 if (max_len < len)
2618 max_len = len;
2620 if (file->is_unmerged) {
2621 /* "Unmerged" is 8 characters */
2622 bin_width = bin_width < 8 ? 8 : bin_width;
2623 continue;
2625 if (file->is_binary) {
2626 /* "Bin XXX -> YYY bytes" */
2627 int w = 14 + decimal_width(file->added)
2628 + decimal_width(file->deleted);
2629 bin_width = bin_width < w ? w : bin_width;
2630 /* Display change counts aligned with "Bin" */
2631 number_width = 3;
2632 continue;
2635 if (max_change < change)
2636 max_change = change;
2638 count = i; /* where we can stop scanning in data->files[] */
2641 * We have width = stat_width or term_columns() columns total.
2642 * We want a maximum of min(max_len, stat_name_width) for the name part.
2643 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2644 * We also need 1 for " " and 4 + decimal_width(max_change)
2645 * for " | NNNN " and one the empty column at the end, altogether
2646 * 6 + decimal_width(max_change).
2648 * If there's not enough space, we will use the smaller of
2649 * stat_name_width (if set) and 5/8*width for the filename,
2650 * and the rest for constant elements + graph part, but no more
2651 * than stat_graph_width for the graph part.
2652 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2653 * for the standard terminal size).
2655 * In other words: stat_width limits the maximum width, and
2656 * stat_name_width fixes the maximum width of the filename,
2657 * and is also used to divide available columns if there
2658 * aren't enough.
2660 * Binary files are displayed with "Bin XXX -> YYY bytes"
2661 * instead of the change count and graph. This part is treated
2662 * similarly to the graph part, except that it is not
2663 * "scaled". If total width is too small to accommodate the
2664 * guaranteed minimum width of the filename part and the
2665 * separators and this message, this message will "overflow"
2666 * making the line longer than the maximum width.
2669 if (options->stat_width == -1)
2670 width = term_columns() - strlen(line_prefix);
2671 else
2672 width = options->stat_width ? options->stat_width : 80;
2673 number_width = decimal_width(max_change) > number_width ?
2674 decimal_width(max_change) : number_width;
2676 if (options->stat_graph_width == -1)
2677 options->stat_graph_width = diff_stat_graph_width;
2680 * Guarantee 3/8*16==6 for the graph part
2681 * and 5/8*16==10 for the filename part
2683 if (width < 16 + 6 + number_width)
2684 width = 16 + 6 + number_width;
2687 * First assign sizes that are wanted, ignoring available width.
2688 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2689 * starting from "XXX" should fit in graph_width.
2691 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2692 if (options->stat_graph_width &&
2693 options->stat_graph_width < graph_width)
2694 graph_width = options->stat_graph_width;
2696 name_width = (options->stat_name_width > 0 &&
2697 options->stat_name_width < max_len) ?
2698 options->stat_name_width : max_len;
2701 * Adjust adjustable widths not to exceed maximum width
2703 if (name_width + number_width + 6 + graph_width > width) {
2704 if (graph_width > width * 3/8 - number_width - 6) {
2705 graph_width = width * 3/8 - number_width - 6;
2706 if (graph_width < 6)
2707 graph_width = 6;
2710 if (options->stat_graph_width &&
2711 graph_width > options->stat_graph_width)
2712 graph_width = options->stat_graph_width;
2713 if (name_width > width - number_width - 6 - graph_width)
2714 name_width = width - number_width - 6 - graph_width;
2715 else
2716 graph_width = width - number_width - 6 - name_width;
2720 * From here name_width is the width of the name area,
2721 * and graph_width is the width of the graph area.
2722 * max_change is used to scale graph properly.
2724 for (i = 0; i < count; i++) {
2725 const char *prefix = "";
2726 struct diffstat_file *file = data->files[i];
2727 char *name = file->print_name;
2728 uintmax_t added = file->added;
2729 uintmax_t deleted = file->deleted;
2730 int name_len;
2732 if (!file->is_interesting && (added + deleted == 0))
2733 continue;
2736 * "scale" the filename
2738 len = name_width;
2739 name_len = strlen(name);
2740 if (name_width < name_len) {
2741 char *slash;
2742 prefix = "...";
2743 len -= 3;
2744 name += name_len - len;
2745 slash = strchr(name, '/');
2746 if (slash)
2747 name = slash;
2750 if (file->is_binary) {
2751 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2752 strbuf_addf(&out, " %*s", number_width, "Bin");
2753 if (!added && !deleted) {
2754 strbuf_addch(&out, '\n');
2755 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2756 out.buf, out.len, 0);
2757 strbuf_reset(&out);
2758 continue;
2760 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2761 del_c, deleted, reset);
2762 strbuf_addstr(&out, " -> ");
2763 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2764 add_c, added, reset);
2765 strbuf_addstr(&out, " bytes\n");
2766 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2767 out.buf, out.len, 0);
2768 strbuf_reset(&out);
2769 continue;
2771 else if (file->is_unmerged) {
2772 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2773 strbuf_addstr(&out, " Unmerged\n");
2774 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2775 out.buf, out.len, 0);
2776 strbuf_reset(&out);
2777 continue;
2781 * scale the add/delete
2783 add = added;
2784 del = deleted;
2786 if (graph_width <= max_change) {
2787 int total = scale_linear(add + del, graph_width, max_change);
2788 if (total < 2 && add && del)
2789 /* width >= 2 due to the sanity check */
2790 total = 2;
2791 if (add < del) {
2792 add = scale_linear(add, graph_width, max_change);
2793 del = total - add;
2794 } else {
2795 del = scale_linear(del, graph_width, max_change);
2796 add = total - del;
2799 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2800 strbuf_addf(&out, " %*"PRIuMAX"%s",
2801 number_width, added + deleted,
2802 added + deleted ? " " : "");
2803 show_graph(&out, '+', add, add_c, reset);
2804 show_graph(&out, '-', del, del_c, reset);
2805 strbuf_addch(&out, '\n');
2806 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2807 out.buf, out.len, 0);
2808 strbuf_reset(&out);
2811 for (i = 0; i < data->nr; i++) {
2812 struct diffstat_file *file = data->files[i];
2813 uintmax_t added = file->added;
2814 uintmax_t deleted = file->deleted;
2816 if (file->is_unmerged ||
2817 (!file->is_interesting && (added + deleted == 0))) {
2818 total_files--;
2819 continue;
2822 if (!file->is_binary) {
2823 adds += added;
2824 dels += deleted;
2826 if (i < count)
2827 continue;
2828 if (!extra_shown)
2829 emit_diff_symbol(options,
2830 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2831 NULL, 0, 0);
2832 extra_shown = 1;
2835 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2836 strbuf_release(&out);
2839 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2841 int i, adds = 0, dels = 0, total_files = data->nr;
2843 if (data->nr == 0)
2844 return;
2846 for (i = 0; i < data->nr; i++) {
2847 int added = data->files[i]->added;
2848 int deleted = data->files[i]->deleted;
2850 if (data->files[i]->is_unmerged ||
2851 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2852 total_files--;
2853 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2854 adds += added;
2855 dels += deleted;
2858 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2861 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2863 int i;
2865 if (data->nr == 0)
2866 return;
2868 for (i = 0; i < data->nr; i++) {
2869 struct diffstat_file *file = data->files[i];
2871 fprintf(options->file, "%s", diff_line_prefix(options));
2873 if (file->is_binary)
2874 fprintf(options->file, "-\t-\t");
2875 else
2876 fprintf(options->file,
2877 "%"PRIuMAX"\t%"PRIuMAX"\t",
2878 file->added, file->deleted);
2879 if (options->line_termination) {
2880 fill_print_name(file);
2881 if (!file->is_renamed)
2882 write_name_quoted(file->name, options->file,
2883 options->line_termination);
2884 else {
2885 fputs(file->print_name, options->file);
2886 putc(options->line_termination, options->file);
2888 } else {
2889 if (file->is_renamed) {
2890 putc('\0', options->file);
2891 write_name_quoted(file->from_name, options->file, '\0');
2893 write_name_quoted(file->name, options->file, '\0');
2898 struct dirstat_file {
2899 const char *name;
2900 unsigned long changed;
2903 struct dirstat_dir {
2904 struct dirstat_file *files;
2905 int alloc, nr, permille, cumulative;
2908 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2909 unsigned long changed, const char *base, int baselen)
2911 unsigned long sum_changes = 0;
2912 unsigned int sources = 0;
2913 const char *line_prefix = diff_line_prefix(opt);
2915 while (dir->nr) {
2916 struct dirstat_file *f = dir->files;
2917 int namelen = strlen(f->name);
2918 unsigned long changes;
2919 char *slash;
2921 if (namelen < baselen)
2922 break;
2923 if (memcmp(f->name, base, baselen))
2924 break;
2925 slash = strchr(f->name + baselen, '/');
2926 if (slash) {
2927 int newbaselen = slash + 1 - f->name;
2928 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2929 sources++;
2930 } else {
2931 changes = f->changed;
2932 dir->files++;
2933 dir->nr--;
2934 sources += 2;
2936 sum_changes += changes;
2940 * We don't report dirstat's for
2941 * - the top level
2942 * - or cases where everything came from a single directory
2943 * under this directory (sources == 1).
2945 if (baselen && sources != 1) {
2946 if (sum_changes) {
2947 int permille = sum_changes * 1000 / changed;
2948 if (permille >= dir->permille) {
2949 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2950 permille / 10, permille % 10, baselen, base);
2951 if (!dir->cumulative)
2952 return 0;
2956 return sum_changes;
2959 static int dirstat_compare(const void *_a, const void *_b)
2961 const struct dirstat_file *a = _a;
2962 const struct dirstat_file *b = _b;
2963 return strcmp(a->name, b->name);
2966 static void show_dirstat(struct diff_options *options)
2968 int i;
2969 unsigned long changed;
2970 struct dirstat_dir dir;
2971 struct diff_queue_struct *q = &diff_queued_diff;
2973 dir.files = NULL;
2974 dir.alloc = 0;
2975 dir.nr = 0;
2976 dir.permille = options->dirstat_permille;
2977 dir.cumulative = options->flags.dirstat_cumulative;
2979 changed = 0;
2980 for (i = 0; i < q->nr; i++) {
2981 struct diff_filepair *p = q->queue[i];
2982 const char *name;
2983 unsigned long copied, added, damage;
2984 struct diff_populate_filespec_options dpf_options = {
2985 .check_size_only = 1,
2988 name = p->two->path ? p->two->path : p->one->path;
2990 if (p->one->oid_valid && p->two->oid_valid &&
2991 oideq(&p->one->oid, &p->two->oid)) {
2993 * The SHA1 has not changed, so pre-/post-content is
2994 * identical. We can therefore skip looking at the
2995 * file contents altogether.
2997 damage = 0;
2998 goto found_damage;
3001 if (options->flags.dirstat_by_file) {
3003 * In --dirstat-by-file mode, we don't really need to
3004 * look at the actual file contents at all.
3005 * The fact that the SHA1 changed is enough for us to
3006 * add this file to the list of results
3007 * (with each file contributing equal damage).
3009 damage = 1;
3010 goto found_damage;
3013 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
3014 diff_populate_filespec(options->repo, p->one, NULL);
3015 diff_populate_filespec(options->repo, p->two, NULL);
3016 diffcore_count_changes(options->repo,
3017 p->one, p->two, NULL, NULL,
3018 &copied, &added);
3019 diff_free_filespec_data(p->one);
3020 diff_free_filespec_data(p->two);
3021 } else if (DIFF_FILE_VALID(p->one)) {
3022 diff_populate_filespec(options->repo, p->one, &dpf_options);
3023 copied = added = 0;
3024 diff_free_filespec_data(p->one);
3025 } else if (DIFF_FILE_VALID(p->two)) {
3026 diff_populate_filespec(options->repo, p->two, &dpf_options);
3027 copied = 0;
3028 added = p->two->size;
3029 diff_free_filespec_data(p->two);
3030 } else
3031 continue;
3034 * Original minus copied is the removed material,
3035 * added is the new material. They are both damages
3036 * made to the preimage.
3037 * If the resulting damage is zero, we know that
3038 * diffcore_count_changes() considers the two entries to
3039 * be identical, but since the oid changed, we
3040 * know that there must have been _some_ kind of change,
3041 * so we force all entries to have damage > 0.
3043 damage = (p->one->size - copied) + added;
3044 if (!damage)
3045 damage = 1;
3047 found_damage:
3048 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3049 dir.files[dir.nr].name = name;
3050 dir.files[dir.nr].changed = damage;
3051 changed += damage;
3052 dir.nr++;
3055 /* This can happen even with many files, if everything was renames */
3056 if (!changed)
3057 return;
3059 /* Show all directories with more than x% of the changes */
3060 QSORT(dir.files, dir.nr, dirstat_compare);
3061 gather_dirstat(options, &dir, changed, "", 0);
3064 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3066 int i;
3067 unsigned long changed;
3068 struct dirstat_dir dir;
3070 if (data->nr == 0)
3071 return;
3073 dir.files = NULL;
3074 dir.alloc = 0;
3075 dir.nr = 0;
3076 dir.permille = options->dirstat_permille;
3077 dir.cumulative = options->flags.dirstat_cumulative;
3079 changed = 0;
3080 for (i = 0; i < data->nr; i++) {
3081 struct diffstat_file *file = data->files[i];
3082 unsigned long damage = file->added + file->deleted;
3083 if (file->is_binary)
3085 * binary files counts bytes, not lines. Must find some
3086 * way to normalize binary bytes vs. textual lines.
3087 * The following heuristic assumes that there are 64
3088 * bytes per "line".
3089 * This is stupid and ugly, but very cheap...
3091 damage = DIV_ROUND_UP(damage, 64);
3092 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3093 dir.files[dir.nr].name = file->name;
3094 dir.files[dir.nr].changed = damage;
3095 changed += damage;
3096 dir.nr++;
3099 /* This can happen even with many files, if everything was renames */
3100 if (!changed)
3101 return;
3103 /* Show all directories with more than x% of the changes */
3104 QSORT(dir.files, dir.nr, dirstat_compare);
3105 gather_dirstat(options, &dir, changed, "", 0);
3108 static void free_diffstat_file(struct diffstat_file *f)
3110 free(f->print_name);
3111 free(f->name);
3112 free(f->from_name);
3113 free(f);
3116 void free_diffstat_info(struct diffstat_t *diffstat)
3118 int i;
3119 for (i = 0; i < diffstat->nr; i++)
3120 free_diffstat_file(diffstat->files[i]);
3121 free(diffstat->files);
3124 struct checkdiff_t {
3125 const char *filename;
3126 int lineno;
3127 int conflict_marker_size;
3128 struct diff_options *o;
3129 unsigned ws_rule;
3130 unsigned status;
3133 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3135 char firstchar;
3136 int cnt;
3138 if (len < marker_size + 1)
3139 return 0;
3140 firstchar = line[0];
3141 switch (firstchar) {
3142 case '=': case '>': case '<': case '|':
3143 break;
3144 default:
3145 return 0;
3147 for (cnt = 1; cnt < marker_size; cnt++)
3148 if (line[cnt] != firstchar)
3149 return 0;
3150 /* line[1] through line[marker_size-1] are same as firstchar */
3151 if (len < marker_size + 1 || !isspace(line[marker_size]))
3152 return 0;
3153 return 1;
3156 static void checkdiff_consume_hunk(void *priv,
3157 long ob, long on, long nb, long nn,
3158 const char *func, long funclen)
3161 struct checkdiff_t *data = priv;
3162 data->lineno = nb - 1;
3165 static int checkdiff_consume(void *priv, char *line, unsigned long len)
3167 struct checkdiff_t *data = priv;
3168 int marker_size = data->conflict_marker_size;
3169 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3170 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3171 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3172 char *err;
3173 const char *line_prefix;
3175 assert(data->o);
3176 line_prefix = diff_line_prefix(data->o);
3178 if (line[0] == '+') {
3179 unsigned bad;
3180 data->lineno++;
3181 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3182 data->status |= 1;
3183 fprintf(data->o->file,
3184 "%s%s:%d: leftover conflict marker\n",
3185 line_prefix, data->filename, data->lineno);
3187 bad = ws_check(line + 1, len - 1, data->ws_rule);
3188 if (!bad)
3189 return 0;
3190 data->status |= bad;
3191 err = whitespace_error_string(bad);
3192 fprintf(data->o->file, "%s%s:%d: %s.\n",
3193 line_prefix, data->filename, data->lineno, err);
3194 free(err);
3195 emit_line(data->o, set, reset, line, 1);
3196 ws_check_emit(line + 1, len - 1, data->ws_rule,
3197 data->o->file, set, reset, ws);
3198 } else if (line[0] == ' ') {
3199 data->lineno++;
3201 return 0;
3204 static unsigned char *deflate_it(char *data,
3205 unsigned long size,
3206 unsigned long *result_size)
3208 int bound;
3209 unsigned char *deflated;
3210 git_zstream stream;
3212 git_deflate_init(&stream, zlib_compression_level);
3213 bound = git_deflate_bound(&stream, size);
3214 deflated = xmalloc(bound);
3215 stream.next_out = deflated;
3216 stream.avail_out = bound;
3218 stream.next_in = (unsigned char *)data;
3219 stream.avail_in = size;
3220 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3221 ; /* nothing */
3222 git_deflate_end(&stream);
3223 *result_size = stream.total_out;
3224 return deflated;
3227 static void emit_binary_diff_body(struct diff_options *o,
3228 mmfile_t *one, mmfile_t *two)
3230 void *cp;
3231 void *delta;
3232 void *deflated;
3233 void *data;
3234 unsigned long orig_size;
3235 unsigned long delta_size;
3236 unsigned long deflate_size;
3237 unsigned long data_size;
3239 /* We could do deflated delta, or we could do just deflated two,
3240 * whichever is smaller.
3242 delta = NULL;
3243 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3244 if (one->size && two->size) {
3245 delta = diff_delta(one->ptr, one->size,
3246 two->ptr, two->size,
3247 &delta_size, deflate_size);
3248 if (delta) {
3249 void *to_free = delta;
3250 orig_size = delta_size;
3251 delta = deflate_it(delta, delta_size, &delta_size);
3252 free(to_free);
3256 if (delta && delta_size < deflate_size) {
3257 char *s = xstrfmt("%"PRIuMAX , (uintmax_t)orig_size);
3258 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3259 s, strlen(s), 0);
3260 free(s);
3261 free(deflated);
3262 data = delta;
3263 data_size = delta_size;
3264 } else {
3265 char *s = xstrfmt("%lu", two->size);
3266 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3267 s, strlen(s), 0);
3268 free(s);
3269 free(delta);
3270 data = deflated;
3271 data_size = deflate_size;
3274 /* emit data encoded in base85 */
3275 cp = data;
3276 while (data_size) {
3277 int len;
3278 int bytes = (52 < data_size) ? 52 : data_size;
3279 char line[71];
3280 data_size -= bytes;
3281 if (bytes <= 26)
3282 line[0] = bytes + 'A' - 1;
3283 else
3284 line[0] = bytes - 26 + 'a' - 1;
3285 encode_85(line + 1, cp, bytes);
3286 cp = (char *) cp + bytes;
3288 len = strlen(line);
3289 line[len++] = '\n';
3290 line[len] = '\0';
3292 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3293 line, len, 0);
3295 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3296 free(data);
3299 static void emit_binary_diff(struct diff_options *o,
3300 mmfile_t *one, mmfile_t *two)
3302 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3303 emit_binary_diff_body(o, one, two);
3304 emit_binary_diff_body(o, two, one);
3307 int diff_filespec_is_binary(struct repository *r,
3308 struct diff_filespec *one)
3310 struct diff_populate_filespec_options dpf_options = {
3311 .check_binary = 1,
3314 if (one->is_binary == -1) {
3315 diff_filespec_load_driver(one, r->index);
3316 if (one->driver->binary != -1)
3317 one->is_binary = one->driver->binary;
3318 else {
3319 if (!one->data && DIFF_FILE_VALID(one))
3320 diff_populate_filespec(r, one, &dpf_options);
3321 if (one->is_binary == -1 && one->data)
3322 one->is_binary = buffer_is_binary(one->data,
3323 one->size);
3324 if (one->is_binary == -1)
3325 one->is_binary = 0;
3328 return one->is_binary;
3331 static const struct userdiff_funcname *
3332 diff_funcname_pattern(struct diff_options *o, struct diff_filespec *one)
3334 diff_filespec_load_driver(one, o->repo->index);
3335 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3338 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3340 if (!options->a_prefix)
3341 options->a_prefix = a;
3342 if (!options->b_prefix)
3343 options->b_prefix = b;
3346 struct userdiff_driver *get_textconv(struct repository *r,
3347 struct diff_filespec *one)
3349 if (!DIFF_FILE_VALID(one))
3350 return NULL;
3352 diff_filespec_load_driver(one, r->index);
3353 return userdiff_get_textconv(r, one->driver);
3356 static void builtin_diff(const char *name_a,
3357 const char *name_b,
3358 struct diff_filespec *one,
3359 struct diff_filespec *two,
3360 const char *xfrm_msg,
3361 int must_show_header,
3362 struct diff_options *o,
3363 int complete_rewrite)
3365 mmfile_t mf1, mf2;
3366 const char *lbl[2];
3367 char *a_one, *b_two;
3368 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3369 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3370 const char *a_prefix, *b_prefix;
3371 struct userdiff_driver *textconv_one = NULL;
3372 struct userdiff_driver *textconv_two = NULL;
3373 struct strbuf header = STRBUF_INIT;
3374 const char *line_prefix = diff_line_prefix(o);
3376 diff_set_mnemonic_prefix(o, "a/", "b/");
3377 if (o->flags.reverse_diff) {
3378 a_prefix = o->b_prefix;
3379 b_prefix = o->a_prefix;
3380 } else {
3381 a_prefix = o->a_prefix;
3382 b_prefix = o->b_prefix;
3385 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3386 (!one->mode || S_ISGITLINK(one->mode)) &&
3387 (!two->mode || S_ISGITLINK(two->mode))) {
3388 show_submodule_diff_summary(o, one->path ? one->path : two->path,
3389 &one->oid, &two->oid,
3390 two->dirty_submodule);
3391 return;
3392 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3393 (!one->mode || S_ISGITLINK(one->mode)) &&
3394 (!two->mode || S_ISGITLINK(two->mode))) {
3395 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3396 &one->oid, &two->oid,
3397 two->dirty_submodule);
3398 return;
3401 if (o->flags.allow_textconv) {
3402 textconv_one = get_textconv(o->repo, one);
3403 textconv_two = get_textconv(o->repo, two);
3406 /* Never use a non-valid filename anywhere if at all possible */
3407 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3408 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3410 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3411 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3412 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3413 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3414 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3415 if (lbl[0][0] == '/') {
3416 /* /dev/null */
3417 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3418 if (xfrm_msg)
3419 strbuf_addstr(&header, xfrm_msg);
3420 must_show_header = 1;
3422 else if (lbl[1][0] == '/') {
3423 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3424 if (xfrm_msg)
3425 strbuf_addstr(&header, xfrm_msg);
3426 must_show_header = 1;
3428 else {
3429 if (one->mode != two->mode) {
3430 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3431 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3432 must_show_header = 1;
3434 if (xfrm_msg)
3435 strbuf_addstr(&header, xfrm_msg);
3438 * we do not run diff between different kind
3439 * of objects.
3441 if ((one->mode ^ two->mode) & S_IFMT)
3442 goto free_ab_and_return;
3443 if (complete_rewrite &&
3444 (textconv_one || !diff_filespec_is_binary(o->repo, one)) &&
3445 (textconv_two || !diff_filespec_is_binary(o->repo, two))) {
3446 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3447 header.buf, header.len, 0);
3448 strbuf_reset(&header);
3449 emit_rewrite_diff(name_a, name_b, one, two,
3450 textconv_one, textconv_two, o);
3451 o->found_changes = 1;
3452 goto free_ab_and_return;
3456 if (o->irreversible_delete && lbl[1][0] == '/') {
3457 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3458 header.len, 0);
3459 strbuf_reset(&header);
3460 goto free_ab_and_return;
3461 } else if (!o->flags.text &&
3462 ( (!textconv_one && diff_filespec_is_binary(o->repo, one)) ||
3463 (!textconv_two && diff_filespec_is_binary(o->repo, two)) )) {
3464 struct strbuf sb = STRBUF_INIT;
3465 if (!one->data && !two->data &&
3466 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3467 !o->flags.binary) {
3468 if (oideq(&one->oid, &two->oid)) {
3469 if (must_show_header)
3470 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3471 header.buf, header.len,
3473 goto free_ab_and_return;
3475 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3476 header.buf, header.len, 0);
3477 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3478 diff_line_prefix(o), lbl[0], lbl[1]);
3479 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3480 sb.buf, sb.len, 0);
3481 strbuf_release(&sb);
3482 goto free_ab_and_return;
3484 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3485 fill_mmfile(o->repo, &mf2, two) < 0)
3486 die("unable to read files to diff");
3487 /* Quite common confusing case */
3488 if (mf1.size == mf2.size &&
3489 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3490 if (must_show_header)
3491 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3492 header.buf, header.len, 0);
3493 goto free_ab_and_return;
3495 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3496 strbuf_reset(&header);
3497 if (o->flags.binary)
3498 emit_binary_diff(o, &mf1, &mf2);
3499 else {
3500 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3501 diff_line_prefix(o), lbl[0], lbl[1]);
3502 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3503 sb.buf, sb.len, 0);
3504 strbuf_release(&sb);
3506 o->found_changes = 1;
3507 } else {
3508 /* Crazy xdl interfaces.. */
3509 const char *diffopts;
3510 const char *v;
3511 xpparam_t xpp;
3512 xdemitconf_t xecfg;
3513 struct emit_callback ecbdata;
3514 const struct userdiff_funcname *pe;
3516 if (must_show_header) {
3517 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3518 header.buf, header.len, 0);
3519 strbuf_reset(&header);
3522 mf1.size = fill_textconv(o->repo, textconv_one, one, &mf1.ptr);
3523 mf2.size = fill_textconv(o->repo, textconv_two, two, &mf2.ptr);
3525 pe = diff_funcname_pattern(o, one);
3526 if (!pe)
3527 pe = diff_funcname_pattern(o, two);
3529 memset(&xpp, 0, sizeof(xpp));
3530 memset(&xecfg, 0, sizeof(xecfg));
3531 memset(&ecbdata, 0, sizeof(ecbdata));
3532 if (o->flags.suppress_diff_headers)
3533 lbl[0] = NULL;
3534 ecbdata.label_path = lbl;
3535 ecbdata.color_diff = want_color(o->use_color);
3536 ecbdata.ws_rule = whitespace_rule(o->repo->index, name_b);
3537 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3538 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3539 ecbdata.opt = o;
3540 if (header.len && !o->flags.suppress_diff_headers)
3541 ecbdata.header = &header;
3542 xpp.flags = o->xdl_opts;
3543 xpp.ignore_regex = o->ignore_regex;
3544 xpp.ignore_regex_nr = o->ignore_regex_nr;
3545 xpp.anchors = o->anchors;
3546 xpp.anchors_nr = o->anchors_nr;
3547 xecfg.ctxlen = o->context;
3548 xecfg.interhunkctxlen = o->interhunkcontext;
3549 xecfg.flags = XDL_EMIT_FUNCNAMES;
3550 if (o->flags.funccontext)
3551 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3552 if (pe)
3553 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3555 diffopts = getenv("GIT_DIFF_OPTS");
3556 if (!diffopts)
3558 else if (skip_prefix(diffopts, "--unified=", &v))
3559 xecfg.ctxlen = strtoul(v, NULL, 10);
3560 else if (skip_prefix(diffopts, "-u", &v))
3561 xecfg.ctxlen = strtoul(v, NULL, 10);
3563 if (o->word_diff)
3564 init_diff_words_data(&ecbdata, o, one, two);
3565 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3566 &ecbdata, &xpp, &xecfg))
3567 die("unable to generate diff for %s", one->path);
3568 if (o->word_diff)
3569 free_diff_words_data(&ecbdata);
3570 if (textconv_one)
3571 free(mf1.ptr);
3572 if (textconv_two)
3573 free(mf2.ptr);
3574 xdiff_clear_find_func(&xecfg);
3577 free_ab_and_return:
3578 strbuf_release(&header);
3579 diff_free_filespec_data(one);
3580 diff_free_filespec_data(two);
3581 free(a_one);
3582 free(b_two);
3583 return;
3586 static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3588 if (!is_renamed) {
3589 if (p->status == DIFF_STATUS_ADDED) {
3590 if (S_ISLNK(p->two->mode))
3591 return "new +l";
3592 else if ((p->two->mode & 0777) == 0755)
3593 return "new +x";
3594 else
3595 return "new";
3596 } else if (p->status == DIFF_STATUS_DELETED)
3597 return "gone";
3599 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3600 return "mode -l";
3601 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3602 return "mode +l";
3603 else if ((p->one->mode & 0777) == 0644 &&
3604 (p->two->mode & 0777) == 0755)
3605 return "mode +x";
3606 else if ((p->one->mode & 0777) == 0755 &&
3607 (p->two->mode & 0777) == 0644)
3608 return "mode -x";
3609 return NULL;
3612 static void builtin_diffstat(const char *name_a, const char *name_b,
3613 struct diff_filespec *one,
3614 struct diff_filespec *two,
3615 struct diffstat_t *diffstat,
3616 struct diff_options *o,
3617 struct diff_filepair *p)
3619 mmfile_t mf1, mf2;
3620 struct diffstat_file *data;
3621 int may_differ;
3622 int complete_rewrite = 0;
3624 if (!DIFF_PAIR_UNMERGED(p)) {
3625 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3626 complete_rewrite = 1;
3629 data = diffstat_add(diffstat, name_a, name_b);
3630 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3631 if (o->flags.stat_with_summary)
3632 data->comments = get_compact_summary(p, data->is_renamed);
3634 if (!one || !two) {
3635 data->is_unmerged = 1;
3636 return;
3639 /* saves some reads if true, not a guarantee of diff outcome */
3640 may_differ = !(one->oid_valid && two->oid_valid &&
3641 oideq(&one->oid, &two->oid));
3643 if (diff_filespec_is_binary(o->repo, one) ||
3644 diff_filespec_is_binary(o->repo, two)) {
3645 data->is_binary = 1;
3646 if (!may_differ) {
3647 data->added = 0;
3648 data->deleted = 0;
3649 } else {
3650 data->added = diff_filespec_size(o->repo, two);
3651 data->deleted = diff_filespec_size(o->repo, one);
3655 else if (complete_rewrite) {
3656 diff_populate_filespec(o->repo, one, NULL);
3657 diff_populate_filespec(o->repo, two, NULL);
3658 data->deleted = count_lines(one->data, one->size);
3659 data->added = count_lines(two->data, two->size);
3662 else if (may_differ) {
3663 /* Crazy xdl interfaces.. */
3664 xpparam_t xpp;
3665 xdemitconf_t xecfg;
3667 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3668 fill_mmfile(o->repo, &mf2, two) < 0)
3669 die("unable to read files to diff");
3671 memset(&xpp, 0, sizeof(xpp));
3672 memset(&xecfg, 0, sizeof(xecfg));
3673 xpp.flags = o->xdl_opts;
3674 xpp.ignore_regex = o->ignore_regex;
3675 xpp.ignore_regex_nr = o->ignore_regex_nr;
3676 xpp.anchors = o->anchors;
3677 xpp.anchors_nr = o->anchors_nr;
3678 xecfg.ctxlen = o->context;
3679 xecfg.interhunkctxlen = o->interhunkcontext;
3680 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
3681 if (xdi_diff_outf(&mf1, &mf2, NULL,
3682 diffstat_consume, diffstat, &xpp, &xecfg))
3683 die("unable to generate diffstat for %s", one->path);
3685 if (DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two)) {
3686 struct diffstat_file *file =
3687 diffstat->files[diffstat->nr - 1];
3689 * Omit diffstats of modified files where nothing changed.
3690 * Even if may_differ, this might be the case due to
3691 * ignoring whitespace changes, etc.
3693 * But note that we special-case additions, deletions,
3694 * renames, and mode changes as adding an empty file,
3695 * for example is still of interest.
3697 if ((p->status == DIFF_STATUS_MODIFIED)
3698 && !file->added
3699 && !file->deleted
3700 && one->mode == two->mode) {
3701 free_diffstat_file(file);
3702 diffstat->nr--;
3707 diff_free_filespec_data(one);
3708 diff_free_filespec_data(two);
3711 static void builtin_checkdiff(const char *name_a, const char *name_b,
3712 const char *attr_path,
3713 struct diff_filespec *one,
3714 struct diff_filespec *two,
3715 struct diff_options *o)
3717 mmfile_t mf1, mf2;
3718 struct checkdiff_t data;
3720 if (!two)
3721 return;
3723 memset(&data, 0, sizeof(data));
3724 data.filename = name_b ? name_b : name_a;
3725 data.lineno = 0;
3726 data.o = o;
3727 data.ws_rule = whitespace_rule(o->repo->index, attr_path);
3728 data.conflict_marker_size = ll_merge_marker_size(o->repo->index, attr_path);
3730 if (fill_mmfile(o->repo, &mf1, one) < 0 ||
3731 fill_mmfile(o->repo, &mf2, two) < 0)
3732 die("unable to read files to diff");
3735 * All the other codepaths check both sides, but not checking
3736 * the "old" side here is deliberate. We are checking the newly
3737 * introduced changes, and as long as the "new" side is text, we
3738 * can and should check what it introduces.
3740 if (diff_filespec_is_binary(o->repo, two))
3741 goto free_and_return;
3742 else {
3743 /* Crazy xdl interfaces.. */
3744 xpparam_t xpp;
3745 xdemitconf_t xecfg;
3747 memset(&xpp, 0, sizeof(xpp));
3748 memset(&xecfg, 0, sizeof(xecfg));
3749 xecfg.ctxlen = 1; /* at least one context line */
3750 xpp.flags = 0;
3751 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume_hunk,
3752 checkdiff_consume, &data,
3753 &xpp, &xecfg))
3754 die("unable to generate checkdiff for %s", one->path);
3756 if (data.ws_rule & WS_BLANK_AT_EOF) {
3757 struct emit_callback ecbdata;
3758 int blank_at_eof;
3760 ecbdata.ws_rule = data.ws_rule;
3761 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3762 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3764 if (blank_at_eof) {
3765 static char *err;
3766 if (!err)
3767 err = whitespace_error_string(WS_BLANK_AT_EOF);
3768 fprintf(o->file, "%s:%d: %s.\n",
3769 data.filename, blank_at_eof, err);
3770 data.status = 1; /* report errors */
3774 free_and_return:
3775 diff_free_filespec_data(one);
3776 diff_free_filespec_data(two);
3777 if (data.status)
3778 o->flags.check_failed = 1;
3781 struct diff_filespec *alloc_filespec(const char *path)
3783 struct diff_filespec *spec;
3785 FLEXPTR_ALLOC_STR(spec, path, path);
3786 spec->count = 1;
3787 spec->is_binary = -1;
3788 return spec;
3791 void free_filespec(struct diff_filespec *spec)
3793 if (!--spec->count) {
3794 diff_free_filespec_data(spec);
3795 free(spec);
3799 void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3800 int oid_valid, unsigned short mode)
3802 if (mode) {
3803 spec->mode = canon_mode(mode);
3804 oidcpy(&spec->oid, oid);
3805 spec->oid_valid = oid_valid;
3810 * Given a name and sha1 pair, if the index tells us the file in
3811 * the work tree has that object contents, return true, so that
3812 * prepare_temp_file() does not have to inflate and extract.
3814 static int reuse_worktree_file(struct index_state *istate,
3815 const char *name,
3816 const struct object_id *oid,
3817 int want_file)
3819 const struct cache_entry *ce;
3820 struct stat st;
3821 int pos, len;
3824 * We do not read the cache ourselves here, because the
3825 * benchmark with my previous version that always reads cache
3826 * shows that it makes things worse for diff-tree comparing
3827 * two linux-2.6 kernel trees in an already checked out work
3828 * tree. This is because most diff-tree comparisons deal with
3829 * only a small number of files, while reading the cache is
3830 * expensive for a large project, and its cost outweighs the
3831 * savings we get by not inflating the object to a temporary
3832 * file. Practically, this code only helps when we are used
3833 * by diff-cache --cached, which does read the cache before
3834 * calling us.
3836 if (!istate->cache)
3837 return 0;
3839 /* We want to avoid the working directory if our caller
3840 * doesn't need the data in a normal file, this system
3841 * is rather slow with its stat/open/mmap/close syscalls,
3842 * and the object is contained in a pack file. The pack
3843 * is probably already open and will be faster to obtain
3844 * the data through than the working directory. Loose
3845 * objects however would tend to be slower as they need
3846 * to be individually opened and inflated.
3848 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3849 return 0;
3852 * Similarly, if we'd have to convert the file contents anyway, that
3853 * makes the optimization not worthwhile.
3855 if (!want_file && would_convert_to_git(istate, name))
3856 return 0;
3859 * If this path does not match our sparse-checkout definition,
3860 * then the file will not be in the working directory.
3862 if (!path_in_sparse_checkout(name, istate))
3863 return 0;
3865 len = strlen(name);
3866 pos = index_name_pos(istate, name, len);
3867 if (pos < 0)
3868 return 0;
3869 ce = istate->cache[pos];
3872 * This is not the sha1 we are looking for, or
3873 * unreusable because it is not a regular file.
3875 if (!oideq(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3876 return 0;
3879 * If ce is marked as "assume unchanged", there is no
3880 * guarantee that work tree matches what we are looking for.
3882 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3883 return 0;
3886 * If ce matches the file in the work tree, we can reuse it.
3888 if (ce_uptodate(ce) ||
3889 (!lstat(name, &st) && !ie_match_stat(istate, ce, &st, 0)))
3890 return 1;
3892 return 0;
3895 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3897 struct strbuf buf = STRBUF_INIT;
3898 char *dirty = "";
3900 /* Are we looking at the work tree? */
3901 if (s->dirty_submodule)
3902 dirty = "-dirty";
3904 strbuf_addf(&buf, "Subproject commit %s%s\n",
3905 oid_to_hex(&s->oid), dirty);
3906 s->size = buf.len;
3907 if (size_only) {
3908 s->data = NULL;
3909 strbuf_release(&buf);
3910 } else {
3911 s->data = strbuf_detach(&buf, NULL);
3912 s->should_free = 1;
3914 return 0;
3918 * While doing rename detection and pickaxe operation, we may need to
3919 * grab the data for the blob (or file) for our own in-core comparison.
3920 * diff_filespec has data and size fields for this purpose.
3922 int diff_populate_filespec(struct repository *r,
3923 struct diff_filespec *s,
3924 const struct diff_populate_filespec_options *options)
3926 int size_only = options ? options->check_size_only : 0;
3927 int check_binary = options ? options->check_binary : 0;
3928 int err = 0;
3929 int conv_flags = global_conv_flags_eol;
3931 * demote FAIL to WARN to allow inspecting the situation
3932 * instead of refusing.
3934 if (conv_flags & CONV_EOL_RNDTRP_DIE)
3935 conv_flags = CONV_EOL_RNDTRP_WARN;
3937 if (!DIFF_FILE_VALID(s))
3938 die("internal error: asking to populate invalid file.");
3939 if (S_ISDIR(s->mode))
3940 return -1;
3942 if (s->data)
3943 return 0;
3945 if (size_only && 0 < s->size)
3946 return 0;
3948 if (S_ISGITLINK(s->mode))
3949 return diff_populate_gitlink(s, size_only);
3951 if (!s->oid_valid ||
3952 reuse_worktree_file(r->index, s->path, &s->oid, 0)) {
3953 struct strbuf buf = STRBUF_INIT;
3954 struct stat st;
3955 int fd;
3957 if (lstat(s->path, &st) < 0) {
3958 err_empty:
3959 err = -1;
3960 empty:
3961 s->data = (char *)"";
3962 s->size = 0;
3963 return err;
3965 s->size = xsize_t(st.st_size);
3966 if (!s->size)
3967 goto empty;
3968 if (S_ISLNK(st.st_mode)) {
3969 struct strbuf sb = STRBUF_INIT;
3971 if (strbuf_readlink(&sb, s->path, s->size))
3972 goto err_empty;
3973 s->size = sb.len;
3974 s->data = strbuf_detach(&sb, NULL);
3975 s->should_free = 1;
3976 return 0;
3980 * Even if the caller would be happy with getting
3981 * only the size, we cannot return early at this
3982 * point if the path requires us to run the content
3983 * conversion.
3985 if (size_only && !would_convert_to_git(r->index, s->path))
3986 return 0;
3989 * Note: this check uses xsize_t(st.st_size) that may
3990 * not be the true size of the blob after it goes
3991 * through convert_to_git(). This may not strictly be
3992 * correct, but the whole point of big_file_threshold
3993 * and is_binary check being that we want to avoid
3994 * opening the file and inspecting the contents, this
3995 * is probably fine.
3997 if (check_binary &&
3998 s->size > big_file_threshold && s->is_binary == -1) {
3999 s->is_binary = 1;
4000 return 0;
4002 fd = open(s->path, O_RDONLY);
4003 if (fd < 0)
4004 goto err_empty;
4005 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
4006 close(fd);
4007 s->should_munmap = 1;
4010 * Convert from working tree format to canonical git format
4012 if (convert_to_git(r->index, s->path, s->data, s->size, &buf, conv_flags)) {
4013 size_t size = 0;
4014 munmap(s->data, s->size);
4015 s->should_munmap = 0;
4016 s->data = strbuf_detach(&buf, &size);
4017 s->size = size;
4018 s->should_free = 1;
4021 else {
4022 struct object_info info = {
4023 .sizep = &s->size
4026 if (!(size_only || check_binary))
4028 * Set contentp, since there is no chance that merely
4029 * the size is sufficient.
4031 info.contentp = &s->data;
4033 if (options && options->missing_object_cb) {
4034 if (!oid_object_info_extended(r, &s->oid, &info,
4035 OBJECT_INFO_LOOKUP_REPLACE |
4036 OBJECT_INFO_SKIP_FETCH_OBJECT))
4037 goto object_read;
4038 options->missing_object_cb(options->missing_object_data);
4040 if (oid_object_info_extended(r, &s->oid, &info,
4041 OBJECT_INFO_LOOKUP_REPLACE))
4042 die("unable to read %s", oid_to_hex(&s->oid));
4044 object_read:
4045 if (size_only || check_binary) {
4046 if (size_only)
4047 return 0;
4048 if (s->size > big_file_threshold && s->is_binary == -1) {
4049 s->is_binary = 1;
4050 return 0;
4053 if (!info.contentp) {
4054 info.contentp = &s->data;
4055 if (oid_object_info_extended(r, &s->oid, &info,
4056 OBJECT_INFO_LOOKUP_REPLACE))
4057 die("unable to read %s", oid_to_hex(&s->oid));
4059 s->should_free = 1;
4061 return 0;
4064 void diff_free_filespec_blob(struct diff_filespec *s)
4066 if (s->should_free)
4067 free(s->data);
4068 else if (s->should_munmap)
4069 munmap(s->data, s->size);
4071 if (s->should_free || s->should_munmap) {
4072 s->should_free = s->should_munmap = 0;
4073 s->data = NULL;
4077 void diff_free_filespec_data(struct diff_filespec *s)
4079 if (!s)
4080 return;
4082 diff_free_filespec_blob(s);
4083 FREE_AND_NULL(s->cnt_data);
4086 static void prep_temp_blob(struct index_state *istate,
4087 const char *path, struct diff_tempfile *temp,
4088 void *blob,
4089 unsigned long size,
4090 const struct object_id *oid,
4091 int mode)
4093 struct strbuf buf = STRBUF_INIT;
4094 struct strbuf tempfile = STRBUF_INIT;
4095 char *path_dup = xstrdup(path);
4096 const char *base = basename(path_dup);
4097 struct checkout_metadata meta;
4099 init_checkout_metadata(&meta, NULL, NULL, oid);
4101 /* Generate "XXXXXX_basename.ext" */
4102 strbuf_addstr(&tempfile, "XXXXXX_");
4103 strbuf_addstr(&tempfile, base);
4105 temp->tempfile = mks_tempfile_ts(tempfile.buf, strlen(base) + 1);
4106 if (!temp->tempfile)
4107 die_errno("unable to create temp-file");
4108 if (convert_to_working_tree(istate, path,
4109 (const char *)blob, (size_t)size, &buf, &meta)) {
4110 blob = buf.buf;
4111 size = buf.len;
4113 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
4114 close_tempfile_gently(temp->tempfile))
4115 die_errno("unable to write temp-file");
4116 temp->name = get_tempfile_path(temp->tempfile);
4117 oid_to_hex_r(temp->hex, oid);
4118 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
4119 strbuf_release(&buf);
4120 strbuf_release(&tempfile);
4121 free(path_dup);
4124 static struct diff_tempfile *prepare_temp_file(struct repository *r,
4125 const char *name,
4126 struct diff_filespec *one)
4128 struct diff_tempfile *temp = claim_diff_tempfile();
4130 if (!DIFF_FILE_VALID(one)) {
4131 not_a_valid_file:
4132 /* A '-' entry produces this for file-2, and
4133 * a '+' entry produces this for file-1.
4135 temp->name = "/dev/null";
4136 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4137 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4138 return temp;
4141 if (!S_ISGITLINK(one->mode) &&
4142 (!one->oid_valid ||
4143 reuse_worktree_file(r->index, name, &one->oid, 1))) {
4144 struct stat st;
4145 if (lstat(name, &st) < 0) {
4146 if (errno == ENOENT)
4147 goto not_a_valid_file;
4148 die_errno("stat(%s)", name);
4150 if (S_ISLNK(st.st_mode)) {
4151 struct strbuf sb = STRBUF_INIT;
4152 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4153 die_errno("readlink(%s)", name);
4154 prep_temp_blob(r->index, name, temp, sb.buf, sb.len,
4155 (one->oid_valid ?
4156 &one->oid : null_oid()),
4157 (one->oid_valid ?
4158 one->mode : S_IFLNK));
4159 strbuf_release(&sb);
4161 else {
4162 /* we can borrow from the file in the work tree */
4163 temp->name = name;
4164 if (!one->oid_valid)
4165 oid_to_hex_r(temp->hex, null_oid());
4166 else
4167 oid_to_hex_r(temp->hex, &one->oid);
4168 /* Even though we may sometimes borrow the
4169 * contents from the work tree, we always want
4170 * one->mode. mode is trustworthy even when
4171 * !(one->oid_valid), as long as
4172 * DIFF_FILE_VALID(one).
4174 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4176 return temp;
4178 else {
4179 if (diff_populate_filespec(r, one, NULL))
4180 die("cannot read data blob for %s", one->path);
4181 prep_temp_blob(r->index, name, temp,
4182 one->data, one->size,
4183 &one->oid, one->mode);
4185 return temp;
4188 static void add_external_diff_name(struct repository *r,
4189 struct strvec *argv,
4190 const char *name,
4191 struct diff_filespec *df)
4193 struct diff_tempfile *temp = prepare_temp_file(r, name, df);
4194 strvec_push(argv, temp->name);
4195 strvec_push(argv, temp->hex);
4196 strvec_push(argv, temp->mode);
4199 /* An external diff command takes:
4201 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4202 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4205 static void run_external_diff(const char *pgm,
4206 const char *name,
4207 const char *other,
4208 struct diff_filespec *one,
4209 struct diff_filespec *two,
4210 const char *xfrm_msg,
4211 struct diff_options *o)
4213 struct strvec argv = STRVEC_INIT;
4214 struct strvec env = STRVEC_INIT;
4215 struct diff_queue_struct *q = &diff_queued_diff;
4217 strvec_push(&argv, pgm);
4218 strvec_push(&argv, name);
4220 if (one && two) {
4221 add_external_diff_name(o->repo, &argv, name, one);
4222 if (!other)
4223 add_external_diff_name(o->repo, &argv, name, two);
4224 else {
4225 add_external_diff_name(o->repo, &argv, other, two);
4226 strvec_push(&argv, other);
4227 strvec_push(&argv, xfrm_msg);
4231 strvec_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
4232 strvec_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4234 diff_free_filespec_data(one);
4235 diff_free_filespec_data(two);
4236 if (run_command_v_opt_cd_env(argv.v, RUN_USING_SHELL, NULL, env.v))
4237 die(_("external diff died, stopping at %s"), name);
4239 remove_tempfile();
4240 strvec_clear(&argv);
4241 strvec_clear(&env);
4244 static int similarity_index(struct diff_filepair *p)
4246 return p->score * 100 / MAX_SCORE;
4249 static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4251 if (startup_info->have_repository)
4252 return find_unique_abbrev(oid, abbrev);
4253 else {
4254 char *hex = oid_to_hex(oid);
4255 if (abbrev < 0)
4256 abbrev = FALLBACK_DEFAULT_ABBREV;
4257 if (abbrev > the_hash_algo->hexsz)
4258 BUG("oid abbreviation out of range: %d", abbrev);
4259 if (abbrev)
4260 hex[abbrev] = '\0';
4261 return hex;
4265 static void fill_metainfo(struct strbuf *msg,
4266 const char *name,
4267 const char *other,
4268 struct diff_filespec *one,
4269 struct diff_filespec *two,
4270 struct diff_options *o,
4271 struct diff_filepair *p,
4272 int *must_show_header,
4273 int use_color)
4275 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4276 const char *reset = diff_get_color(use_color, DIFF_RESET);
4277 const char *line_prefix = diff_line_prefix(o);
4279 *must_show_header = 1;
4280 strbuf_init(msg, PATH_MAX * 2 + 300);
4281 switch (p->status) {
4282 case DIFF_STATUS_COPIED:
4283 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4284 line_prefix, set, similarity_index(p));
4285 strbuf_addf(msg, "%s\n%s%scopy from ",
4286 reset, line_prefix, set);
4287 quote_c_style(name, msg, NULL, 0);
4288 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4289 quote_c_style(other, msg, NULL, 0);
4290 strbuf_addf(msg, "%s\n", reset);
4291 break;
4292 case DIFF_STATUS_RENAMED:
4293 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4294 line_prefix, set, similarity_index(p));
4295 strbuf_addf(msg, "%s\n%s%srename from ",
4296 reset, line_prefix, set);
4297 quote_c_style(name, msg, NULL, 0);
4298 strbuf_addf(msg, "%s\n%s%srename to ",
4299 reset, line_prefix, set);
4300 quote_c_style(other, msg, NULL, 0);
4301 strbuf_addf(msg, "%s\n", reset);
4302 break;
4303 case DIFF_STATUS_MODIFIED:
4304 if (p->score) {
4305 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4306 line_prefix,
4307 set, similarity_index(p), reset);
4308 break;
4310 /* fallthru */
4311 default:
4312 *must_show_header = 0;
4314 if (one && two && !oideq(&one->oid, &two->oid)) {
4315 const unsigned hexsz = the_hash_algo->hexsz;
4316 int abbrev = o->abbrev ? o->abbrev : DEFAULT_ABBREV;
4318 if (o->flags.full_index)
4319 abbrev = hexsz;
4321 if (o->flags.binary) {
4322 mmfile_t mf;
4323 if ((!fill_mmfile(o->repo, &mf, one) &&
4324 diff_filespec_is_binary(o->repo, one)) ||
4325 (!fill_mmfile(o->repo, &mf, two) &&
4326 diff_filespec_is_binary(o->repo, two)))
4327 abbrev = hexsz;
4329 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4330 diff_abbrev_oid(&one->oid, abbrev),
4331 diff_abbrev_oid(&two->oid, abbrev));
4332 if (one->mode == two->mode)
4333 strbuf_addf(msg, " %06o", one->mode);
4334 strbuf_addf(msg, "%s\n", reset);
4338 static void run_diff_cmd(const char *pgm,
4339 const char *name,
4340 const char *other,
4341 const char *attr_path,
4342 struct diff_filespec *one,
4343 struct diff_filespec *two,
4344 struct strbuf *msg,
4345 struct diff_options *o,
4346 struct diff_filepair *p)
4348 const char *xfrm_msg = NULL;
4349 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4350 int must_show_header = 0;
4353 if (o->flags.allow_external) {
4354 struct userdiff_driver *drv;
4356 drv = userdiff_find_by_path(o->repo->index, attr_path);
4357 if (drv && drv->external)
4358 pgm = drv->external;
4361 if (msg) {
4363 * don't use colors when the header is intended for an
4364 * external diff driver
4366 fill_metainfo(msg, name, other, one, two, o, p,
4367 &must_show_header,
4368 want_color(o->use_color) && !pgm);
4369 xfrm_msg = msg->len ? msg->buf : NULL;
4372 if (pgm) {
4373 run_external_diff(pgm, name, other, one, two, xfrm_msg, o);
4374 return;
4376 if (one && two)
4377 builtin_diff(name, other ? other : name,
4378 one, two, xfrm_msg, must_show_header,
4379 o, complete_rewrite);
4380 else
4381 fprintf(o->file, "* Unmerged path %s\n", name);
4384 static void diff_fill_oid_info(struct diff_filespec *one, struct index_state *istate)
4386 if (DIFF_FILE_VALID(one)) {
4387 if (!one->oid_valid) {
4388 struct stat st;
4389 if (one->is_stdin) {
4390 oidclr(&one->oid);
4391 return;
4393 if (lstat(one->path, &st) < 0)
4394 die_errno("stat '%s'", one->path);
4395 if (index_path(istate, &one->oid, one->path, &st, 0))
4396 die("cannot hash %s", one->path);
4399 else
4400 oidclr(&one->oid);
4403 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4405 /* Strip the prefix but do not molest /dev/null and absolute paths */
4406 if (*namep && !is_absolute_path(*namep)) {
4407 *namep += prefix_length;
4408 if (**namep == '/')
4409 ++*namep;
4411 if (*otherp && !is_absolute_path(*otherp)) {
4412 *otherp += prefix_length;
4413 if (**otherp == '/')
4414 ++*otherp;
4418 static void run_diff(struct diff_filepair *p, struct diff_options *o)
4420 const char *pgm = external_diff();
4421 struct strbuf msg;
4422 struct diff_filespec *one = p->one;
4423 struct diff_filespec *two = p->two;
4424 const char *name;
4425 const char *other;
4426 const char *attr_path;
4428 name = one->path;
4429 other = (strcmp(name, two->path) ? two->path : NULL);
4430 attr_path = name;
4431 if (o->prefix_length)
4432 strip_prefix(o->prefix_length, &name, &other);
4434 if (!o->flags.allow_external)
4435 pgm = NULL;
4437 if (DIFF_PAIR_UNMERGED(p)) {
4438 run_diff_cmd(pgm, name, NULL, attr_path,
4439 NULL, NULL, NULL, o, p);
4440 return;
4443 diff_fill_oid_info(one, o->repo->index);
4444 diff_fill_oid_info(two, o->repo->index);
4446 if (!pgm &&
4447 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4448 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4450 * a filepair that changes between file and symlink
4451 * needs to be split into deletion and creation.
4453 struct diff_filespec *null = alloc_filespec(two->path);
4454 run_diff_cmd(NULL, name, other, attr_path,
4455 one, null, &msg,
4456 o, p);
4457 free(null);
4458 strbuf_release(&msg);
4460 null = alloc_filespec(one->path);
4461 run_diff_cmd(NULL, name, other, attr_path,
4462 null, two, &msg, o, p);
4463 free(null);
4465 else
4466 run_diff_cmd(pgm, name, other, attr_path,
4467 one, two, &msg, o, p);
4469 strbuf_release(&msg);
4472 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4473 struct diffstat_t *diffstat)
4475 const char *name;
4476 const char *other;
4478 if (DIFF_PAIR_UNMERGED(p)) {
4479 /* unmerged */
4480 builtin_diffstat(p->one->path, NULL, NULL, NULL,
4481 diffstat, o, p);
4482 return;
4485 name = p->one->path;
4486 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4488 if (o->prefix_length)
4489 strip_prefix(o->prefix_length, &name, &other);
4491 diff_fill_oid_info(p->one, o->repo->index);
4492 diff_fill_oid_info(p->two, o->repo->index);
4494 builtin_diffstat(name, other, p->one, p->two,
4495 diffstat, o, p);
4498 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4500 const char *name;
4501 const char *other;
4502 const char *attr_path;
4504 if (DIFF_PAIR_UNMERGED(p)) {
4505 /* unmerged */
4506 return;
4509 name = p->one->path;
4510 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4511 attr_path = other ? other : name;
4513 if (o->prefix_length)
4514 strip_prefix(o->prefix_length, &name, &other);
4516 diff_fill_oid_info(p->one, o->repo->index);
4517 diff_fill_oid_info(p->two, o->repo->index);
4519 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4522 static void prep_parse_options(struct diff_options *options);
4524 void repo_diff_setup(struct repository *r, struct diff_options *options)
4526 memcpy(options, &default_diff_options, sizeof(*options));
4528 options->file = stdout;
4529 options->repo = r;
4531 options->output_indicators[OUTPUT_INDICATOR_NEW] = '+';
4532 options->output_indicators[OUTPUT_INDICATOR_OLD] = '-';
4533 options->output_indicators[OUTPUT_INDICATOR_CONTEXT] = ' ';
4534 options->abbrev = DEFAULT_ABBREV;
4535 options->line_termination = '\n';
4536 options->break_opt = -1;
4537 options->rename_limit = -1;
4538 options->dirstat_permille = diff_dirstat_permille_default;
4539 options->context = diff_context_default;
4540 options->interhunkcontext = diff_interhunk_context_default;
4541 options->ws_error_highlight = ws_error_highlight_default;
4542 options->flags.rename_empty = 1;
4543 options->flags.relative_name = diff_relative;
4544 options->objfind = NULL;
4546 /* pathchange left =NULL by default */
4547 options->change = diff_change;
4548 options->add_remove = diff_addremove;
4549 options->use_color = diff_use_color_default;
4550 options->detect_rename = diff_detect_rename_default;
4551 options->xdl_opts |= diff_algorithm;
4552 if (diff_indent_heuristic)
4553 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4555 options->orderfile = diff_order_file_cfg;
4557 if (!options->flags.ignore_submodule_set)
4558 options->flags.ignore_untracked_in_submodules = 1;
4560 if (diff_no_prefix) {
4561 options->a_prefix = options->b_prefix = "";
4562 } else if (!diff_mnemonic_prefix) {
4563 options->a_prefix = "a/";
4564 options->b_prefix = "b/";
4567 options->color_moved = diff_color_moved_default;
4568 options->color_moved_ws_handling = diff_color_moved_ws_default;
4570 prep_parse_options(options);
4573 void diff_setup_done(struct diff_options *options)
4575 unsigned check_mask = DIFF_FORMAT_NAME |
4576 DIFF_FORMAT_NAME_STATUS |
4577 DIFF_FORMAT_CHECKDIFF |
4578 DIFF_FORMAT_NO_OUTPUT;
4580 * This must be signed because we're comparing against a potentially
4581 * negative value.
4583 const int hexsz = the_hash_algo->hexsz;
4585 if (options->set_default)
4586 options->set_default(options);
4588 if (HAS_MULTI_BITS(options->output_format & check_mask))
4589 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4591 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4592 die(_("-G, -S and --find-object are mutually exclusive"));
4594 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_G_REGEX_MASK))
4595 die(_("-G and --pickaxe-regex are mutually exclusive, use --pickaxe-regex with -S"));
4597 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_ALL_OBJFIND_MASK))
4598 die(_("--pickaxe-all and --find-object are mutually exclusive, use --pickaxe-all with -G and -S"));
4601 * Most of the time we can say "there are changes"
4602 * only by checking if there are changed paths, but
4603 * --ignore-whitespace* options force us to look
4604 * inside contents.
4607 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS) ||
4608 options->ignore_regex_nr)
4609 options->flags.diff_from_contents = 1;
4610 else
4611 options->flags.diff_from_contents = 0;
4613 if (options->flags.find_copies_harder)
4614 options->detect_rename = DIFF_DETECT_COPY;
4616 if (!options->flags.relative_name)
4617 options->prefix = NULL;
4618 if (options->prefix)
4619 options->prefix_length = strlen(options->prefix);
4620 else
4621 options->prefix_length = 0;
4623 if (options->output_format & (DIFF_FORMAT_NAME |
4624 DIFF_FORMAT_NAME_STATUS |
4625 DIFF_FORMAT_CHECKDIFF |
4626 DIFF_FORMAT_NO_OUTPUT))
4627 options->output_format &= ~(DIFF_FORMAT_RAW |
4628 DIFF_FORMAT_NUMSTAT |
4629 DIFF_FORMAT_DIFFSTAT |
4630 DIFF_FORMAT_SHORTSTAT |
4631 DIFF_FORMAT_DIRSTAT |
4632 DIFF_FORMAT_SUMMARY |
4633 DIFF_FORMAT_PATCH);
4636 * These cases always need recursive; we do not drop caller-supplied
4637 * recursive bits for other formats here.
4639 if (options->output_format & (DIFF_FORMAT_PATCH |
4640 DIFF_FORMAT_NUMSTAT |
4641 DIFF_FORMAT_DIFFSTAT |
4642 DIFF_FORMAT_SHORTSTAT |
4643 DIFF_FORMAT_DIRSTAT |
4644 DIFF_FORMAT_SUMMARY |
4645 DIFF_FORMAT_CHECKDIFF))
4646 options->flags.recursive = 1;
4648 * Also pickaxe would not work very well if you do not say recursive
4650 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4651 options->flags.recursive = 1;
4653 * When patches are generated, submodules diffed against the work tree
4654 * must be checked for dirtiness too so it can be shown in the output
4656 if (options->output_format & DIFF_FORMAT_PATCH)
4657 options->flags.dirty_submodules = 1;
4659 if (options->detect_rename && options->rename_limit < 0)
4660 options->rename_limit = diff_rename_limit_default;
4661 if (hexsz < options->abbrev)
4662 options->abbrev = hexsz; /* full */
4665 * It does not make sense to show the first hit we happened
4666 * to have found. It does not make sense not to return with
4667 * exit code in such a case either.
4669 if (options->flags.quick) {
4670 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4671 options->flags.exit_with_status = 1;
4674 options->diff_path_counter = 0;
4676 if (options->flags.follow_renames && options->pathspec.nr != 1)
4677 die(_("--follow requires exactly one pathspec"));
4679 if (!options->use_color || external_diff())
4680 options->color_moved = 0;
4682 FREE_AND_NULL(options->parseopts);
4685 int parse_long_opt(const char *opt, const char **argv,
4686 const char **optarg)
4688 const char *arg = argv[0];
4689 if (!skip_prefix(arg, "--", &arg))
4690 return 0;
4691 if (!skip_prefix(arg, opt, &arg))
4692 return 0;
4693 if (*arg == '=') { /* stuck form: --option=value */
4694 *optarg = arg + 1;
4695 return 1;
4697 if (*arg != '\0')
4698 return 0;
4699 /* separate form: --option value */
4700 if (!argv[1])
4701 die("Option '--%s' requires a value", opt);
4702 *optarg = argv[1];
4703 return 2;
4706 static int diff_opt_stat(const struct option *opt, const char *value, int unset)
4708 struct diff_options *options = opt->value;
4709 int width = options->stat_width;
4710 int name_width = options->stat_name_width;
4711 int graph_width = options->stat_graph_width;
4712 int count = options->stat_count;
4713 char *end;
4715 BUG_ON_OPT_NEG(unset);
4717 if (!strcmp(opt->long_name, "stat")) {
4718 if (value) {
4719 width = strtoul(value, &end, 10);
4720 if (*end == ',')
4721 name_width = strtoul(end+1, &end, 10);
4722 if (*end == ',')
4723 count = strtoul(end+1, &end, 10);
4724 if (*end)
4725 return error(_("invalid --stat value: %s"), value);
4727 } else if (!strcmp(opt->long_name, "stat-width")) {
4728 width = strtoul(value, &end, 10);
4729 if (*end)
4730 return error(_("%s expects a numerical value"),
4731 opt->long_name);
4732 } else if (!strcmp(opt->long_name, "stat-name-width")) {
4733 name_width = strtoul(value, &end, 10);
4734 if (*end)
4735 return error(_("%s expects a numerical value"),
4736 opt->long_name);
4737 } else if (!strcmp(opt->long_name, "stat-graph-width")) {
4738 graph_width = strtoul(value, &end, 10);
4739 if (*end)
4740 return error(_("%s expects a numerical value"),
4741 opt->long_name);
4742 } else if (!strcmp(opt->long_name, "stat-count")) {
4743 count = strtoul(value, &end, 10);
4744 if (*end)
4745 return error(_("%s expects a numerical value"),
4746 opt->long_name);
4747 } else
4748 BUG("%s should not get here", opt->long_name);
4750 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4751 options->stat_name_width = name_width;
4752 options->stat_graph_width = graph_width;
4753 options->stat_width = width;
4754 options->stat_count = count;
4755 return 0;
4758 static int parse_dirstat_opt(struct diff_options *options, const char *params)
4760 struct strbuf errmsg = STRBUF_INIT;
4761 if (parse_dirstat_params(options, params, &errmsg))
4762 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4763 errmsg.buf);
4764 strbuf_release(&errmsg);
4766 * The caller knows a dirstat-related option is given from the command
4767 * line; allow it to say "return this_function();"
4769 options->output_format |= DIFF_FORMAT_DIRSTAT;
4770 return 1;
4773 static const char diff_status_letters[] = {
4774 DIFF_STATUS_ADDED,
4775 DIFF_STATUS_COPIED,
4776 DIFF_STATUS_DELETED,
4777 DIFF_STATUS_MODIFIED,
4778 DIFF_STATUS_RENAMED,
4779 DIFF_STATUS_TYPE_CHANGED,
4780 DIFF_STATUS_UNKNOWN,
4781 DIFF_STATUS_UNMERGED,
4782 DIFF_STATUS_FILTER_AON,
4783 DIFF_STATUS_FILTER_BROKEN,
4784 '\0',
4787 static unsigned int filter_bit['Z' + 1];
4789 static void prepare_filter_bits(void)
4791 int i;
4793 if (!filter_bit[DIFF_STATUS_ADDED]) {
4794 for (i = 0; diff_status_letters[i]; i++)
4795 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4799 static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4801 return opt->filter & filter_bit[(int) status];
4804 unsigned diff_filter_bit(char status)
4806 prepare_filter_bits();
4807 return filter_bit[(int) status];
4810 static int diff_opt_diff_filter(const struct option *option,
4811 const char *optarg, int unset)
4813 struct diff_options *opt = option->value;
4814 int i, optch;
4816 BUG_ON_OPT_NEG(unset);
4817 prepare_filter_bits();
4820 * If there is a negation e.g. 'd' in the input, and we haven't
4821 * initialized the filter field with another --diff-filter, start
4822 * from full set of bits, except for AON.
4824 if (!opt->filter) {
4825 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4826 if (optch < 'a' || 'z' < optch)
4827 continue;
4828 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4829 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4830 break;
4834 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4835 unsigned int bit;
4836 int negate;
4838 if ('a' <= optch && optch <= 'z') {
4839 negate = 1;
4840 optch = toupper(optch);
4841 } else {
4842 negate = 0;
4845 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4846 if (!bit)
4847 return error(_("unknown change class '%c' in --diff-filter=%s"),
4848 optarg[i], optarg);
4849 if (negate)
4850 opt->filter &= ~bit;
4851 else
4852 opt->filter |= bit;
4854 return 0;
4857 static void enable_patch_output(int *fmt)
4859 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4860 *fmt |= DIFF_FORMAT_PATCH;
4863 static int diff_opt_ws_error_highlight(const struct option *option,
4864 const char *arg, int unset)
4866 struct diff_options *opt = option->value;
4867 int val = parse_ws_error_highlight(arg);
4869 BUG_ON_OPT_NEG(unset);
4870 if (val < 0)
4871 return error(_("unknown value after ws-error-highlight=%.*s"),
4872 -1 - val, arg);
4873 opt->ws_error_highlight = val;
4874 return 0;
4877 static int diff_opt_find_object(const struct option *option,
4878 const char *arg, int unset)
4880 struct diff_options *opt = option->value;
4881 struct object_id oid;
4883 BUG_ON_OPT_NEG(unset);
4884 if (get_oid(arg, &oid))
4885 return error(_("unable to resolve '%s'"), arg);
4887 if (!opt->objfind)
4888 CALLOC_ARRAY(opt->objfind, 1);
4890 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4891 opt->flags.recursive = 1;
4892 opt->flags.tree_in_recursive = 1;
4893 oidset_insert(opt->objfind, &oid);
4894 return 0;
4897 static int diff_opt_anchored(const struct option *opt,
4898 const char *arg, int unset)
4900 struct diff_options *options = opt->value;
4902 BUG_ON_OPT_NEG(unset);
4903 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4904 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
4905 options->anchors_alloc);
4906 options->anchors[options->anchors_nr++] = xstrdup(arg);
4907 return 0;
4910 static int diff_opt_binary(const struct option *opt,
4911 const char *arg, int unset)
4913 struct diff_options *options = opt->value;
4915 BUG_ON_OPT_NEG(unset);
4916 BUG_ON_OPT_ARG(arg);
4917 enable_patch_output(&options->output_format);
4918 options->flags.binary = 1;
4919 return 0;
4922 static int diff_opt_break_rewrites(const struct option *opt,
4923 const char *arg, int unset)
4925 int *break_opt = opt->value;
4926 int opt1, opt2;
4928 BUG_ON_OPT_NEG(unset);
4929 if (!arg)
4930 arg = "";
4931 opt1 = parse_rename_score(&arg);
4932 if (*arg == 0)
4933 opt2 = 0;
4934 else if (*arg != '/')
4935 return error(_("%s expects <n>/<m> form"), opt->long_name);
4936 else {
4937 arg++;
4938 opt2 = parse_rename_score(&arg);
4940 if (*arg != 0)
4941 return error(_("%s expects <n>/<m> form"), opt->long_name);
4942 *break_opt = opt1 | (opt2 << 16);
4943 return 0;
4946 static int diff_opt_char(const struct option *opt,
4947 const char *arg, int unset)
4949 char *value = opt->value;
4951 BUG_ON_OPT_NEG(unset);
4952 if (arg[1])
4953 return error(_("%s expects a character, got '%s'"),
4954 opt->long_name, arg);
4955 *value = arg[0];
4956 return 0;
4959 static int diff_opt_color_moved(const struct option *opt,
4960 const char *arg, int unset)
4962 struct diff_options *options = opt->value;
4964 if (unset) {
4965 options->color_moved = COLOR_MOVED_NO;
4966 } else if (!arg) {
4967 if (diff_color_moved_default)
4968 options->color_moved = diff_color_moved_default;
4969 if (options->color_moved == COLOR_MOVED_NO)
4970 options->color_moved = COLOR_MOVED_DEFAULT;
4971 } else {
4972 int cm = parse_color_moved(arg);
4973 if (cm < 0)
4974 return error(_("bad --color-moved argument: %s"), arg);
4975 options->color_moved = cm;
4977 return 0;
4980 static int diff_opt_color_moved_ws(const struct option *opt,
4981 const char *arg, int unset)
4983 struct diff_options *options = opt->value;
4984 unsigned cm;
4986 if (unset) {
4987 options->color_moved_ws_handling = 0;
4988 return 0;
4991 cm = parse_color_moved_ws(arg);
4992 if (cm & COLOR_MOVED_WS_ERROR)
4993 return error(_("invalid mode '%s' in --color-moved-ws"), arg);
4994 options->color_moved_ws_handling = cm;
4995 return 0;
4998 static int diff_opt_color_words(const struct option *opt,
4999 const char *arg, int unset)
5001 struct diff_options *options = opt->value;
5003 BUG_ON_OPT_NEG(unset);
5004 options->use_color = 1;
5005 options->word_diff = DIFF_WORDS_COLOR;
5006 options->word_regex = arg;
5007 return 0;
5010 static int diff_opt_compact_summary(const struct option *opt,
5011 const char *arg, int unset)
5013 struct diff_options *options = opt->value;
5015 BUG_ON_OPT_ARG(arg);
5016 if (unset) {
5017 options->flags.stat_with_summary = 0;
5018 } else {
5019 options->flags.stat_with_summary = 1;
5020 options->output_format |= DIFF_FORMAT_DIFFSTAT;
5022 return 0;
5025 static int diff_opt_diff_algorithm(const struct option *opt,
5026 const char *arg, int unset)
5028 struct diff_options *options = opt->value;
5029 long value = parse_algorithm_value(arg);
5031 BUG_ON_OPT_NEG(unset);
5032 if (value < 0)
5033 return error(_("option diff-algorithm accepts \"myers\", "
5034 "\"minimal\", \"patience\" and \"histogram\""));
5036 /* clear out previous settings */
5037 DIFF_XDL_CLR(options, NEED_MINIMAL);
5038 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
5039 options->xdl_opts |= value;
5040 return 0;
5043 static int diff_opt_dirstat(const struct option *opt,
5044 const char *arg, int unset)
5046 struct diff_options *options = opt->value;
5048 BUG_ON_OPT_NEG(unset);
5049 if (!strcmp(opt->long_name, "cumulative")) {
5050 if (arg)
5051 BUG("how come --cumulative take a value?");
5052 arg = "cumulative";
5053 } else if (!strcmp(opt->long_name, "dirstat-by-file"))
5054 parse_dirstat_opt(options, "files");
5055 parse_dirstat_opt(options, arg ? arg : "");
5056 return 0;
5059 static int diff_opt_find_copies(const struct option *opt,
5060 const char *arg, int unset)
5062 struct diff_options *options = opt->value;
5064 BUG_ON_OPT_NEG(unset);
5065 if (!arg)
5066 arg = "";
5067 options->rename_score = parse_rename_score(&arg);
5068 if (*arg != 0)
5069 return error(_("invalid argument to %s"), opt->long_name);
5071 if (options->detect_rename == DIFF_DETECT_COPY)
5072 options->flags.find_copies_harder = 1;
5073 else
5074 options->detect_rename = DIFF_DETECT_COPY;
5076 return 0;
5079 static int diff_opt_find_renames(const struct option *opt,
5080 const char *arg, int unset)
5082 struct diff_options *options = opt->value;
5084 BUG_ON_OPT_NEG(unset);
5085 if (!arg)
5086 arg = "";
5087 options->rename_score = parse_rename_score(&arg);
5088 if (*arg != 0)
5089 return error(_("invalid argument to %s"), opt->long_name);
5091 options->detect_rename = DIFF_DETECT_RENAME;
5092 return 0;
5095 static int diff_opt_follow(const struct option *opt,
5096 const char *arg, int unset)
5098 struct diff_options *options = opt->value;
5100 BUG_ON_OPT_ARG(arg);
5101 if (unset) {
5102 options->flags.follow_renames = 0;
5103 options->flags.default_follow_renames = 0;
5104 } else {
5105 options->flags.follow_renames = 1;
5107 return 0;
5110 static int diff_opt_ignore_submodules(const struct option *opt,
5111 const char *arg, int unset)
5113 struct diff_options *options = opt->value;
5115 BUG_ON_OPT_NEG(unset);
5116 if (!arg)
5117 arg = "all";
5118 options->flags.override_submodule_config = 1;
5119 handle_ignore_submodules_arg(options, arg);
5120 return 0;
5123 static int diff_opt_line_prefix(const struct option *opt,
5124 const char *optarg, int unset)
5126 struct diff_options *options = opt->value;
5128 BUG_ON_OPT_NEG(unset);
5129 options->line_prefix = optarg;
5130 options->line_prefix_length = strlen(options->line_prefix);
5131 graph_setup_line_prefix(options);
5132 return 0;
5135 static int diff_opt_no_prefix(const struct option *opt,
5136 const char *optarg, int unset)
5138 struct diff_options *options = opt->value;
5140 BUG_ON_OPT_NEG(unset);
5141 BUG_ON_OPT_ARG(optarg);
5142 options->a_prefix = "";
5143 options->b_prefix = "";
5144 return 0;
5147 static enum parse_opt_result diff_opt_output(struct parse_opt_ctx_t *ctx,
5148 const struct option *opt,
5149 const char *arg, int unset)
5151 struct diff_options *options = opt->value;
5152 char *path;
5154 BUG_ON_OPT_NEG(unset);
5155 path = prefix_filename(ctx->prefix, arg);
5156 options->file = xfopen(path, "w");
5157 options->close_file = 1;
5158 if (options->use_color != GIT_COLOR_ALWAYS)
5159 options->use_color = GIT_COLOR_NEVER;
5160 free(path);
5161 return 0;
5164 static int diff_opt_patience(const struct option *opt,
5165 const char *arg, int unset)
5167 struct diff_options *options = opt->value;
5168 int i;
5170 BUG_ON_OPT_NEG(unset);
5171 BUG_ON_OPT_ARG(arg);
5172 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
5174 * Both --patience and --anchored use PATIENCE_DIFF
5175 * internally, so remove any anchors previously
5176 * specified.
5178 for (i = 0; i < options->anchors_nr; i++)
5179 free(options->anchors[i]);
5180 options->anchors_nr = 0;
5181 return 0;
5184 static int diff_opt_ignore_regex(const struct option *opt,
5185 const char *arg, int unset)
5187 struct diff_options *options = opt->value;
5188 regex_t *regex;
5190 BUG_ON_OPT_NEG(unset);
5191 regex = xmalloc(sizeof(*regex));
5192 if (regcomp(regex, arg, REG_EXTENDED | REG_NEWLINE))
5193 return error(_("invalid regex given to -I: '%s'"), arg);
5194 ALLOC_GROW(options->ignore_regex, options->ignore_regex_nr + 1,
5195 options->ignore_regex_alloc);
5196 options->ignore_regex[options->ignore_regex_nr++] = regex;
5197 return 0;
5200 static int diff_opt_pickaxe_regex(const struct option *opt,
5201 const char *arg, int unset)
5203 struct diff_options *options = opt->value;
5205 BUG_ON_OPT_NEG(unset);
5206 options->pickaxe = arg;
5207 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5208 return 0;
5211 static int diff_opt_pickaxe_string(const struct option *opt,
5212 const char *arg, int unset)
5214 struct diff_options *options = opt->value;
5216 BUG_ON_OPT_NEG(unset);
5217 options->pickaxe = arg;
5218 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5219 return 0;
5222 static int diff_opt_relative(const struct option *opt,
5223 const char *arg, int unset)
5225 struct diff_options *options = opt->value;
5227 options->flags.relative_name = !unset;
5228 if (arg)
5229 options->prefix = arg;
5230 return 0;
5233 static int diff_opt_submodule(const struct option *opt,
5234 const char *arg, int unset)
5236 struct diff_options *options = opt->value;
5238 BUG_ON_OPT_NEG(unset);
5239 if (!arg)
5240 arg = "log";
5241 if (parse_submodule_params(options, arg))
5242 return error(_("failed to parse --submodule option parameter: '%s'"),
5243 arg);
5244 return 0;
5247 static int diff_opt_textconv(const struct option *opt,
5248 const char *arg, int unset)
5250 struct diff_options *options = opt->value;
5252 BUG_ON_OPT_ARG(arg);
5253 if (unset) {
5254 options->flags.allow_textconv = 0;
5255 } else {
5256 options->flags.allow_textconv = 1;
5257 options->flags.textconv_set_via_cmdline = 1;
5259 return 0;
5262 static int diff_opt_unified(const struct option *opt,
5263 const char *arg, int unset)
5265 struct diff_options *options = opt->value;
5266 char *s;
5268 BUG_ON_OPT_NEG(unset);
5270 if (arg) {
5271 options->context = strtol(arg, &s, 10);
5272 if (*s)
5273 return error(_("%s expects a numerical value"), "--unified");
5275 enable_patch_output(&options->output_format);
5277 return 0;
5280 static int diff_opt_word_diff(const struct option *opt,
5281 const char *arg, int unset)
5283 struct diff_options *options = opt->value;
5285 BUG_ON_OPT_NEG(unset);
5286 if (arg) {
5287 if (!strcmp(arg, "plain"))
5288 options->word_diff = DIFF_WORDS_PLAIN;
5289 else if (!strcmp(arg, "color")) {
5290 options->use_color = 1;
5291 options->word_diff = DIFF_WORDS_COLOR;
5293 else if (!strcmp(arg, "porcelain"))
5294 options->word_diff = DIFF_WORDS_PORCELAIN;
5295 else if (!strcmp(arg, "none"))
5296 options->word_diff = DIFF_WORDS_NONE;
5297 else
5298 return error(_("bad --word-diff argument: %s"), arg);
5299 } else {
5300 if (options->word_diff == DIFF_WORDS_NONE)
5301 options->word_diff = DIFF_WORDS_PLAIN;
5303 return 0;
5306 static int diff_opt_word_diff_regex(const struct option *opt,
5307 const char *arg, int unset)
5309 struct diff_options *options = opt->value;
5311 BUG_ON_OPT_NEG(unset);
5312 if (options->word_diff == DIFF_WORDS_NONE)
5313 options->word_diff = DIFF_WORDS_PLAIN;
5314 options->word_regex = arg;
5315 return 0;
5318 static int diff_opt_rotate_to(const struct option *opt, const char *arg, int unset)
5320 struct diff_options *options = opt->value;
5322 BUG_ON_OPT_NEG(unset);
5323 if (!strcmp(opt->long_name, "skip-to"))
5324 options->skip_instead_of_rotate = 1;
5325 else
5326 options->skip_instead_of_rotate = 0;
5327 options->rotate_to = arg;
5328 return 0;
5331 static void prep_parse_options(struct diff_options *options)
5333 struct option parseopts[] = {
5334 OPT_GROUP(N_("Diff output format options")),
5335 OPT_BITOP('p', "patch", &options->output_format,
5336 N_("generate patch"),
5337 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5338 OPT_BIT_F('s', "no-patch", &options->output_format,
5339 N_("suppress diff output"),
5340 DIFF_FORMAT_NO_OUTPUT, PARSE_OPT_NONEG),
5341 OPT_BITOP('u', NULL, &options->output_format,
5342 N_("generate patch"),
5343 DIFF_FORMAT_PATCH, DIFF_FORMAT_NO_OUTPUT),
5344 OPT_CALLBACK_F('U', "unified", options, N_("<n>"),
5345 N_("generate diffs with <n> lines context"),
5346 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_unified),
5347 OPT_BOOL('W', "function-context", &options->flags.funccontext,
5348 N_("generate diffs with <n> lines context")),
5349 OPT_BIT_F(0, "raw", &options->output_format,
5350 N_("generate the diff in raw format"),
5351 DIFF_FORMAT_RAW, PARSE_OPT_NONEG),
5352 OPT_BITOP(0, "patch-with-raw", &options->output_format,
5353 N_("synonym for '-p --raw'"),
5354 DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW,
5355 DIFF_FORMAT_NO_OUTPUT),
5356 OPT_BITOP(0, "patch-with-stat", &options->output_format,
5357 N_("synonym for '-p --stat'"),
5358 DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT,
5359 DIFF_FORMAT_NO_OUTPUT),
5360 OPT_BIT_F(0, "numstat", &options->output_format,
5361 N_("machine friendly --stat"),
5362 DIFF_FORMAT_NUMSTAT, PARSE_OPT_NONEG),
5363 OPT_BIT_F(0, "shortstat", &options->output_format,
5364 N_("output only the last line of --stat"),
5365 DIFF_FORMAT_SHORTSTAT, PARSE_OPT_NONEG),
5366 OPT_CALLBACK_F('X', "dirstat", options, N_("<param1,param2>..."),
5367 N_("output the distribution of relative amount of changes for each sub-directory"),
5368 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5369 diff_opt_dirstat),
5370 OPT_CALLBACK_F(0, "cumulative", options, NULL,
5371 N_("synonym for --dirstat=cumulative"),
5372 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5373 diff_opt_dirstat),
5374 OPT_CALLBACK_F(0, "dirstat-by-file", options, N_("<param1,param2>..."),
5375 N_("synonym for --dirstat=files,param1,param2..."),
5376 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5377 diff_opt_dirstat),
5378 OPT_BIT_F(0, "check", &options->output_format,
5379 N_("warn if changes introduce conflict markers or whitespace errors"),
5380 DIFF_FORMAT_CHECKDIFF, PARSE_OPT_NONEG),
5381 OPT_BIT_F(0, "summary", &options->output_format,
5382 N_("condensed summary such as creations, renames and mode changes"),
5383 DIFF_FORMAT_SUMMARY, PARSE_OPT_NONEG),
5384 OPT_BIT_F(0, "name-only", &options->output_format,
5385 N_("show only names of changed files"),
5386 DIFF_FORMAT_NAME, PARSE_OPT_NONEG),
5387 OPT_BIT_F(0, "name-status", &options->output_format,
5388 N_("show only names and status of changed files"),
5389 DIFF_FORMAT_NAME_STATUS, PARSE_OPT_NONEG),
5390 OPT_CALLBACK_F(0, "stat", options, N_("<width>[,<name-width>[,<count>]]"),
5391 N_("generate diffstat"),
5392 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_stat),
5393 OPT_CALLBACK_F(0, "stat-width", options, N_("<width>"),
5394 N_("generate diffstat with a given width"),
5395 PARSE_OPT_NONEG, diff_opt_stat),
5396 OPT_CALLBACK_F(0, "stat-name-width", options, N_("<width>"),
5397 N_("generate diffstat with a given name width"),
5398 PARSE_OPT_NONEG, diff_opt_stat),
5399 OPT_CALLBACK_F(0, "stat-graph-width", options, N_("<width>"),
5400 N_("generate diffstat with a given graph width"),
5401 PARSE_OPT_NONEG, diff_opt_stat),
5402 OPT_CALLBACK_F(0, "stat-count", options, N_("<count>"),
5403 N_("generate diffstat with limited lines"),
5404 PARSE_OPT_NONEG, diff_opt_stat),
5405 OPT_CALLBACK_F(0, "compact-summary", options, NULL,
5406 N_("generate compact summary in diffstat"),
5407 PARSE_OPT_NOARG, diff_opt_compact_summary),
5408 OPT_CALLBACK_F(0, "binary", options, NULL,
5409 N_("output a binary diff that can be applied"),
5410 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_binary),
5411 OPT_BOOL(0, "full-index", &options->flags.full_index,
5412 N_("show full pre- and post-image object names on the \"index\" lines")),
5413 OPT_COLOR_FLAG(0, "color", &options->use_color,
5414 N_("show colored diff")),
5415 OPT_CALLBACK_F(0, "ws-error-highlight", options, N_("<kind>"),
5416 N_("highlight whitespace errors in the 'context', 'old' or 'new' lines in the diff"),
5417 PARSE_OPT_NONEG, diff_opt_ws_error_highlight),
5418 OPT_SET_INT('z', NULL, &options->line_termination,
5419 N_("do not munge pathnames and use NULs as output field terminators in --raw or --numstat"),
5421 OPT__ABBREV(&options->abbrev),
5422 OPT_STRING_F(0, "src-prefix", &options->a_prefix, N_("<prefix>"),
5423 N_("show the given source prefix instead of \"a/\""),
5424 PARSE_OPT_NONEG),
5425 OPT_STRING_F(0, "dst-prefix", &options->b_prefix, N_("<prefix>"),
5426 N_("show the given destination prefix instead of \"b/\""),
5427 PARSE_OPT_NONEG),
5428 OPT_CALLBACK_F(0, "line-prefix", options, N_("<prefix>"),
5429 N_("prepend an additional prefix to every line of output"),
5430 PARSE_OPT_NONEG, diff_opt_line_prefix),
5431 OPT_CALLBACK_F(0, "no-prefix", options, NULL,
5432 N_("do not show any source or destination prefix"),
5433 PARSE_OPT_NONEG | PARSE_OPT_NOARG, diff_opt_no_prefix),
5434 OPT_INTEGER_F(0, "inter-hunk-context", &options->interhunkcontext,
5435 N_("show context between diff hunks up to the specified number of lines"),
5436 PARSE_OPT_NONEG),
5437 OPT_CALLBACK_F(0, "output-indicator-new",
5438 &options->output_indicators[OUTPUT_INDICATOR_NEW],
5439 N_("<char>"),
5440 N_("specify the character to indicate a new line instead of '+'"),
5441 PARSE_OPT_NONEG, diff_opt_char),
5442 OPT_CALLBACK_F(0, "output-indicator-old",
5443 &options->output_indicators[OUTPUT_INDICATOR_OLD],
5444 N_("<char>"),
5445 N_("specify the character to indicate an old line instead of '-'"),
5446 PARSE_OPT_NONEG, diff_opt_char),
5447 OPT_CALLBACK_F(0, "output-indicator-context",
5448 &options->output_indicators[OUTPUT_INDICATOR_CONTEXT],
5449 N_("<char>"),
5450 N_("specify the character to indicate a context instead of ' '"),
5451 PARSE_OPT_NONEG, diff_opt_char),
5453 OPT_GROUP(N_("Diff rename options")),
5454 OPT_CALLBACK_F('B', "break-rewrites", &options->break_opt, N_("<n>[/<m>]"),
5455 N_("break complete rewrite changes into pairs of delete and create"),
5456 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5457 diff_opt_break_rewrites),
5458 OPT_CALLBACK_F('M', "find-renames", options, N_("<n>"),
5459 N_("detect renames"),
5460 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5461 diff_opt_find_renames),
5462 OPT_SET_INT_F('D', "irreversible-delete", &options->irreversible_delete,
5463 N_("omit the preimage for deletes"),
5464 1, PARSE_OPT_NONEG),
5465 OPT_CALLBACK_F('C', "find-copies", options, N_("<n>"),
5466 N_("detect copies"),
5467 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5468 diff_opt_find_copies),
5469 OPT_BOOL(0, "find-copies-harder", &options->flags.find_copies_harder,
5470 N_("use unmodified files as source to find copies")),
5471 OPT_SET_INT_F(0, "no-renames", &options->detect_rename,
5472 N_("disable rename detection"),
5473 0, PARSE_OPT_NONEG),
5474 OPT_BOOL(0, "rename-empty", &options->flags.rename_empty,
5475 N_("use empty blobs as rename source")),
5476 OPT_CALLBACK_F(0, "follow", options, NULL,
5477 N_("continue listing the history of a file beyond renames"),
5478 PARSE_OPT_NOARG, diff_opt_follow),
5479 OPT_INTEGER('l', NULL, &options->rename_limit,
5480 N_("prevent rename/copy detection if the number of rename/copy targets exceeds given limit")),
5482 OPT_GROUP(N_("Diff algorithm options")),
5483 OPT_BIT(0, "minimal", &options->xdl_opts,
5484 N_("produce the smallest possible diff"),
5485 XDF_NEED_MINIMAL),
5486 OPT_BIT_F('w', "ignore-all-space", &options->xdl_opts,
5487 N_("ignore whitespace when comparing lines"),
5488 XDF_IGNORE_WHITESPACE, PARSE_OPT_NONEG),
5489 OPT_BIT_F('b', "ignore-space-change", &options->xdl_opts,
5490 N_("ignore changes in amount of whitespace"),
5491 XDF_IGNORE_WHITESPACE_CHANGE, PARSE_OPT_NONEG),
5492 OPT_BIT_F(0, "ignore-space-at-eol", &options->xdl_opts,
5493 N_("ignore changes in whitespace at EOL"),
5494 XDF_IGNORE_WHITESPACE_AT_EOL, PARSE_OPT_NONEG),
5495 OPT_BIT_F(0, "ignore-cr-at-eol", &options->xdl_opts,
5496 N_("ignore carrier-return at the end of line"),
5497 XDF_IGNORE_CR_AT_EOL, PARSE_OPT_NONEG),
5498 OPT_BIT_F(0, "ignore-blank-lines", &options->xdl_opts,
5499 N_("ignore changes whose lines are all blank"),
5500 XDF_IGNORE_BLANK_LINES, PARSE_OPT_NONEG),
5501 OPT_CALLBACK_F('I', "ignore-matching-lines", options, N_("<regex>"),
5502 N_("ignore changes whose all lines match <regex>"),
5503 0, diff_opt_ignore_regex),
5504 OPT_BIT(0, "indent-heuristic", &options->xdl_opts,
5505 N_("heuristic to shift diff hunk boundaries for easy reading"),
5506 XDF_INDENT_HEURISTIC),
5507 OPT_CALLBACK_F(0, "patience", options, NULL,
5508 N_("generate diff using the \"patience diff\" algorithm"),
5509 PARSE_OPT_NONEG | PARSE_OPT_NOARG,
5510 diff_opt_patience),
5511 OPT_BITOP(0, "histogram", &options->xdl_opts,
5512 N_("generate diff using the \"histogram diff\" algorithm"),
5513 XDF_HISTOGRAM_DIFF, XDF_DIFF_ALGORITHM_MASK),
5514 OPT_CALLBACK_F(0, "diff-algorithm", options, N_("<algorithm>"),
5515 N_("choose a diff algorithm"),
5516 PARSE_OPT_NONEG, diff_opt_diff_algorithm),
5517 OPT_CALLBACK_F(0, "anchored", options, N_("<text>"),
5518 N_("generate diff using the \"anchored diff\" algorithm"),
5519 PARSE_OPT_NONEG, diff_opt_anchored),
5520 OPT_CALLBACK_F(0, "word-diff", options, N_("<mode>"),
5521 N_("show word diff, using <mode> to delimit changed words"),
5522 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_word_diff),
5523 OPT_CALLBACK_F(0, "word-diff-regex", options, N_("<regex>"),
5524 N_("use <regex> to decide what a word is"),
5525 PARSE_OPT_NONEG, diff_opt_word_diff_regex),
5526 OPT_CALLBACK_F(0, "color-words", options, N_("<regex>"),
5527 N_("equivalent to --word-diff=color --word-diff-regex=<regex>"),
5528 PARSE_OPT_NONEG | PARSE_OPT_OPTARG, diff_opt_color_words),
5529 OPT_CALLBACK_F(0, "color-moved", options, N_("<mode>"),
5530 N_("moved lines of code are colored differently"),
5531 PARSE_OPT_OPTARG, diff_opt_color_moved),
5532 OPT_CALLBACK_F(0, "color-moved-ws", options, N_("<mode>"),
5533 N_("how white spaces are ignored in --color-moved"),
5534 0, diff_opt_color_moved_ws),
5536 OPT_GROUP(N_("Other diff options")),
5537 OPT_CALLBACK_F(0, "relative", options, N_("<prefix>"),
5538 N_("when run from subdir, exclude changes outside and show relative paths"),
5539 PARSE_OPT_OPTARG,
5540 diff_opt_relative),
5541 OPT_BOOL('a', "text", &options->flags.text,
5542 N_("treat all files as text")),
5543 OPT_BOOL('R', NULL, &options->flags.reverse_diff,
5544 N_("swap two inputs, reverse the diff")),
5545 OPT_BOOL(0, "exit-code", &options->flags.exit_with_status,
5546 N_("exit with 1 if there were differences, 0 otherwise")),
5547 OPT_BOOL(0, "quiet", &options->flags.quick,
5548 N_("disable all output of the program")),
5549 OPT_BOOL(0, "ext-diff", &options->flags.allow_external,
5550 N_("allow an external diff helper to be executed")),
5551 OPT_CALLBACK_F(0, "textconv", options, NULL,
5552 N_("run external text conversion filters when comparing binary files"),
5553 PARSE_OPT_NOARG, diff_opt_textconv),
5554 OPT_CALLBACK_F(0, "ignore-submodules", options, N_("<when>"),
5555 N_("ignore changes to submodules in the diff generation"),
5556 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5557 diff_opt_ignore_submodules),
5558 OPT_CALLBACK_F(0, "submodule", options, N_("<format>"),
5559 N_("specify how differences in submodules are shown"),
5560 PARSE_OPT_NONEG | PARSE_OPT_OPTARG,
5561 diff_opt_submodule),
5562 OPT_SET_INT_F(0, "ita-invisible-in-index", &options->ita_invisible_in_index,
5563 N_("hide 'git add -N' entries from the index"),
5564 1, PARSE_OPT_NONEG),
5565 OPT_SET_INT_F(0, "ita-visible-in-index", &options->ita_invisible_in_index,
5566 N_("treat 'git add -N' entries as real in the index"),
5567 0, PARSE_OPT_NONEG),
5568 OPT_CALLBACK_F('S', NULL, options, N_("<string>"),
5569 N_("look for differences that change the number of occurrences of the specified string"),
5570 0, diff_opt_pickaxe_string),
5571 OPT_CALLBACK_F('G', NULL, options, N_("<regex>"),
5572 N_("look for differences that change the number of occurrences of the specified regex"),
5573 0, diff_opt_pickaxe_regex),
5574 OPT_BIT_F(0, "pickaxe-all", &options->pickaxe_opts,
5575 N_("show all changes in the changeset with -S or -G"),
5576 DIFF_PICKAXE_ALL, PARSE_OPT_NONEG),
5577 OPT_BIT_F(0, "pickaxe-regex", &options->pickaxe_opts,
5578 N_("treat <string> in -S as extended POSIX regular expression"),
5579 DIFF_PICKAXE_REGEX, PARSE_OPT_NONEG),
5580 OPT_FILENAME('O', NULL, &options->orderfile,
5581 N_("control the order in which files appear in the output")),
5582 OPT_CALLBACK_F(0, "rotate-to", options, N_("<path>"),
5583 N_("show the change in the specified path first"),
5584 PARSE_OPT_NONEG, diff_opt_rotate_to),
5585 OPT_CALLBACK_F(0, "skip-to", options, N_("<path>"),
5586 N_("skip the output to the specified path"),
5587 PARSE_OPT_NONEG, diff_opt_rotate_to),
5588 OPT_CALLBACK_F(0, "find-object", options, N_("<object-id>"),
5589 N_("look for differences that change the number of occurrences of the specified object"),
5590 PARSE_OPT_NONEG, diff_opt_find_object),
5591 OPT_CALLBACK_F(0, "diff-filter", options, N_("[(A|C|D|M|R|T|U|X|B)...[*]]"),
5592 N_("select files by diff type"),
5593 PARSE_OPT_NONEG, diff_opt_diff_filter),
5594 { OPTION_CALLBACK, 0, "output", options, N_("<file>"),
5595 N_("Output to a specific file"),
5596 PARSE_OPT_NONEG, NULL, 0, diff_opt_output },
5598 OPT_END()
5601 ALLOC_ARRAY(options->parseopts, ARRAY_SIZE(parseopts));
5602 memcpy(options->parseopts, parseopts, sizeof(parseopts));
5605 int diff_opt_parse(struct diff_options *options,
5606 const char **av, int ac, const char *prefix)
5608 if (!prefix)
5609 prefix = "";
5611 ac = parse_options(ac, av, prefix, options->parseopts, NULL,
5612 PARSE_OPT_KEEP_DASHDASH |
5613 PARSE_OPT_KEEP_UNKNOWN |
5614 PARSE_OPT_NO_INTERNAL_HELP |
5615 PARSE_OPT_ONE_SHOT |
5616 PARSE_OPT_STOP_AT_NON_OPTION);
5618 return ac;
5621 int parse_rename_score(const char **cp_p)
5623 unsigned long num, scale;
5624 int ch, dot;
5625 const char *cp = *cp_p;
5627 num = 0;
5628 scale = 1;
5629 dot = 0;
5630 for (;;) {
5631 ch = *cp;
5632 if ( !dot && ch == '.' ) {
5633 scale = 1;
5634 dot = 1;
5635 } else if ( ch == '%' ) {
5636 scale = dot ? scale*100 : 100;
5637 cp++; /* % is always at the end */
5638 break;
5639 } else if ( ch >= '0' && ch <= '9' ) {
5640 if ( scale < 100000 ) {
5641 scale *= 10;
5642 num = (num*10) + (ch-'0');
5644 } else {
5645 break;
5647 cp++;
5649 *cp_p = cp;
5651 /* user says num divided by scale and we say internally that
5652 * is MAX_SCORE * num / scale.
5654 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5657 struct diff_queue_struct diff_queued_diff;
5659 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5661 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5662 queue->queue[queue->nr++] = dp;
5665 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5666 struct diff_filespec *one,
5667 struct diff_filespec *two)
5669 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5670 dp->one = one;
5671 dp->two = two;
5672 if (queue)
5673 diff_q(queue, dp);
5674 return dp;
5677 void diff_free_filepair(struct diff_filepair *p)
5679 free_filespec(p->one);
5680 free_filespec(p->two);
5681 free(p);
5684 const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5686 int abblen;
5687 const char *abbrev;
5689 /* Do we want all 40 hex characters? */
5690 if (len == the_hash_algo->hexsz)
5691 return oid_to_hex(oid);
5693 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5694 abbrev = diff_abbrev_oid(oid, len);
5696 if (!print_sha1_ellipsis())
5697 return abbrev;
5699 abblen = strlen(abbrev);
5702 * In well-behaved cases, where the abbreviated result is the
5703 * same as the requested length, append three dots after the
5704 * abbreviation (hence the whole logic is limited to the case
5705 * where abblen < 37); when the actual abbreviated result is a
5706 * bit longer than the requested length, we reduce the number
5707 * of dots so that they match the well-behaved ones. However,
5708 * if the actual abbreviation is longer than the requested
5709 * length by more than three, we give up on aligning, and add
5710 * three dots anyway, to indicate that the output is not the
5711 * full object name. Yes, this may be suboptimal, but this
5712 * appears only in "diff --raw --abbrev" output and it is not
5713 * worth the effort to change it now. Note that this would
5714 * likely to work fine when the automatic sizing of default
5715 * abbreviation length is used--we would be fed -1 in "len" in
5716 * that case, and will end up always appending three-dots, but
5717 * the automatic sizing is supposed to give abblen that ensures
5718 * uniqueness across all objects (statistically speaking).
5720 if (abblen < the_hash_algo->hexsz - 3) {
5721 static char hex[GIT_MAX_HEXSZ + 1];
5722 if (len < abblen && abblen <= len + 2)
5723 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5724 else
5725 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5726 return hex;
5729 return oid_to_hex(oid);
5732 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5734 int line_termination = opt->line_termination;
5735 int inter_name_termination = line_termination ? '\t' : '\0';
5737 fprintf(opt->file, "%s", diff_line_prefix(opt));
5738 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5739 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5740 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5741 fprintf(opt->file, "%s ",
5742 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5744 if (p->score) {
5745 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5746 inter_name_termination);
5747 } else {
5748 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5751 if (p->status == DIFF_STATUS_COPIED ||
5752 p->status == DIFF_STATUS_RENAMED) {
5753 const char *name_a, *name_b;
5754 name_a = p->one->path;
5755 name_b = p->two->path;
5756 strip_prefix(opt->prefix_length, &name_a, &name_b);
5757 write_name_quoted(name_a, opt->file, inter_name_termination);
5758 write_name_quoted(name_b, opt->file, line_termination);
5759 } else {
5760 const char *name_a, *name_b;
5761 name_a = p->one->mode ? p->one->path : p->two->path;
5762 name_b = NULL;
5763 strip_prefix(opt->prefix_length, &name_a, &name_b);
5764 write_name_quoted(name_a, opt->file, line_termination);
5768 int diff_unmodified_pair(struct diff_filepair *p)
5770 /* This function is written stricter than necessary to support
5771 * the currently implemented transformers, but the idea is to
5772 * let transformers to produce diff_filepairs any way they want,
5773 * and filter and clean them up here before producing the output.
5775 struct diff_filespec *one = p->one, *two = p->two;
5777 if (DIFF_PAIR_UNMERGED(p))
5778 return 0; /* unmerged is interesting */
5780 /* deletion, addition, mode or type change
5781 * and rename are all interesting.
5783 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5784 DIFF_PAIR_MODE_CHANGED(p) ||
5785 strcmp(one->path, two->path))
5786 return 0;
5788 /* both are valid and point at the same path. that is, we are
5789 * dealing with a change.
5791 if (one->oid_valid && two->oid_valid &&
5792 oideq(&one->oid, &two->oid) &&
5793 !one->dirty_submodule && !two->dirty_submodule)
5794 return 1; /* no change */
5795 if (!one->oid_valid && !two->oid_valid)
5796 return 1; /* both look at the same file on the filesystem. */
5797 return 0;
5800 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5802 if (diff_unmodified_pair(p))
5803 return;
5805 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5806 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5807 return; /* no tree diffs in patch format */
5809 run_diff(p, o);
5812 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5813 struct diffstat_t *diffstat)
5815 if (diff_unmodified_pair(p))
5816 return;
5818 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5819 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5820 return; /* no useful stat for tree diffs */
5822 run_diffstat(p, o, diffstat);
5825 static void diff_flush_checkdiff(struct diff_filepair *p,
5826 struct diff_options *o)
5828 if (diff_unmodified_pair(p))
5829 return;
5831 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5832 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5833 return; /* nothing to check in tree diffs */
5835 run_checkdiff(p, o);
5838 int diff_queue_is_empty(void)
5840 struct diff_queue_struct *q = &diff_queued_diff;
5841 int i;
5842 for (i = 0; i < q->nr; i++)
5843 if (!diff_unmodified_pair(q->queue[i]))
5844 return 0;
5845 return 1;
5848 #if DIFF_DEBUG
5849 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5851 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5852 x, one ? one : "",
5853 s->path,
5854 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5855 s->mode,
5856 s->oid_valid ? oid_to_hex(&s->oid) : "");
5857 fprintf(stderr, "queue[%d] %s size %lu\n",
5858 x, one ? one : "",
5859 s->size);
5862 void diff_debug_filepair(const struct diff_filepair *p, int i)
5864 diff_debug_filespec(p->one, i, "one");
5865 diff_debug_filespec(p->two, i, "two");
5866 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5867 p->score, p->status ? p->status : '?',
5868 p->one->rename_used, p->broken_pair);
5871 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5873 int i;
5874 if (msg)
5875 fprintf(stderr, "%s\n", msg);
5876 fprintf(stderr, "q->nr = %d\n", q->nr);
5877 for (i = 0; i < q->nr; i++) {
5878 struct diff_filepair *p = q->queue[i];
5879 diff_debug_filepair(p, i);
5882 #endif
5884 static void diff_resolve_rename_copy(void)
5886 int i;
5887 struct diff_filepair *p;
5888 struct diff_queue_struct *q = &diff_queued_diff;
5890 diff_debug_queue("resolve-rename-copy", q);
5892 for (i = 0; i < q->nr; i++) {
5893 p = q->queue[i];
5894 p->status = 0; /* undecided */
5895 if (DIFF_PAIR_UNMERGED(p))
5896 p->status = DIFF_STATUS_UNMERGED;
5897 else if (!DIFF_FILE_VALID(p->one))
5898 p->status = DIFF_STATUS_ADDED;
5899 else if (!DIFF_FILE_VALID(p->two))
5900 p->status = DIFF_STATUS_DELETED;
5901 else if (DIFF_PAIR_TYPE_CHANGED(p))
5902 p->status = DIFF_STATUS_TYPE_CHANGED;
5904 /* from this point on, we are dealing with a pair
5905 * whose both sides are valid and of the same type, i.e.
5906 * either in-place edit or rename/copy edit.
5908 else if (DIFF_PAIR_RENAME(p)) {
5910 * A rename might have re-connected a broken
5911 * pair up, causing the pathnames to be the
5912 * same again. If so, that's not a rename at
5913 * all, just a modification..
5915 * Otherwise, see if this source was used for
5916 * multiple renames, in which case we decrement
5917 * the count, and call it a copy.
5919 if (!strcmp(p->one->path, p->two->path))
5920 p->status = DIFF_STATUS_MODIFIED;
5921 else if (--p->one->rename_used > 0)
5922 p->status = DIFF_STATUS_COPIED;
5923 else
5924 p->status = DIFF_STATUS_RENAMED;
5926 else if (!oideq(&p->one->oid, &p->two->oid) ||
5927 p->one->mode != p->two->mode ||
5928 p->one->dirty_submodule ||
5929 p->two->dirty_submodule ||
5930 is_null_oid(&p->one->oid))
5931 p->status = DIFF_STATUS_MODIFIED;
5932 else {
5933 /* This is a "no-change" entry and should not
5934 * happen anymore, but prepare for broken callers.
5936 error("feeding unmodified %s to diffcore",
5937 p->one->path);
5938 p->status = DIFF_STATUS_UNKNOWN;
5941 diff_debug_queue("resolve-rename-copy done", q);
5944 static int check_pair_status(struct diff_filepair *p)
5946 switch (p->status) {
5947 case DIFF_STATUS_UNKNOWN:
5948 return 0;
5949 case 0:
5950 die("internal error in diff-resolve-rename-copy");
5951 default:
5952 return 1;
5956 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5958 int fmt = opt->output_format;
5960 if (fmt & DIFF_FORMAT_CHECKDIFF)
5961 diff_flush_checkdiff(p, opt);
5962 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5963 diff_flush_raw(p, opt);
5964 else if (fmt & DIFF_FORMAT_NAME) {
5965 const char *name_a, *name_b;
5966 name_a = p->two->path;
5967 name_b = NULL;
5968 strip_prefix(opt->prefix_length, &name_a, &name_b);
5969 fprintf(opt->file, "%s", diff_line_prefix(opt));
5970 write_name_quoted(name_a, opt->file, opt->line_termination);
5974 static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5976 struct strbuf sb = STRBUF_INIT;
5977 if (fs->mode)
5978 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5979 else
5980 strbuf_addf(&sb, " %s ", newdelete);
5982 quote_c_style(fs->path, &sb, NULL, 0);
5983 strbuf_addch(&sb, '\n');
5984 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5985 sb.buf, sb.len, 0);
5986 strbuf_release(&sb);
5989 static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5990 int show_name)
5992 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5993 struct strbuf sb = STRBUF_INIT;
5994 strbuf_addf(&sb, " mode change %06o => %06o",
5995 p->one->mode, p->two->mode);
5996 if (show_name) {
5997 strbuf_addch(&sb, ' ');
5998 quote_c_style(p->two->path, &sb, NULL, 0);
6000 strbuf_addch(&sb, '\n');
6001 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6002 sb.buf, sb.len, 0);
6003 strbuf_release(&sb);
6007 static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
6008 struct diff_filepair *p)
6010 struct strbuf sb = STRBUF_INIT;
6011 struct strbuf names = STRBUF_INIT;
6013 pprint_rename(&names, p->one->path, p->two->path);
6014 strbuf_addf(&sb, " %s %s (%d%%)\n",
6015 renamecopy, names.buf, similarity_index(p));
6016 strbuf_release(&names);
6017 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6018 sb.buf, sb.len, 0);
6019 show_mode_change(opt, p, 0);
6020 strbuf_release(&sb);
6023 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
6025 switch(p->status) {
6026 case DIFF_STATUS_DELETED:
6027 show_file_mode_name(opt, "delete", p->one);
6028 break;
6029 case DIFF_STATUS_ADDED:
6030 show_file_mode_name(opt, "create", p->two);
6031 break;
6032 case DIFF_STATUS_COPIED:
6033 show_rename_copy(opt, "copy", p);
6034 break;
6035 case DIFF_STATUS_RENAMED:
6036 show_rename_copy(opt, "rename", p);
6037 break;
6038 default:
6039 if (p->score) {
6040 struct strbuf sb = STRBUF_INIT;
6041 strbuf_addstr(&sb, " rewrite ");
6042 quote_c_style(p->two->path, &sb, NULL, 0);
6043 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
6044 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
6045 sb.buf, sb.len, 0);
6046 strbuf_release(&sb);
6048 show_mode_change(opt, p, !p->score);
6049 break;
6053 struct patch_id_t {
6054 git_hash_ctx *ctx;
6055 int patchlen;
6058 static int remove_space(char *line, int len)
6060 int i;
6061 char *dst = line;
6062 unsigned char c;
6064 for (i = 0; i < len; i++)
6065 if (!isspace((c = line[i])))
6066 *dst++ = c;
6068 return dst - line;
6071 void flush_one_hunk(struct object_id *result, git_hash_ctx *ctx)
6073 unsigned char hash[GIT_MAX_RAWSZ];
6074 unsigned short carry = 0;
6075 int i;
6077 the_hash_algo->final_fn(hash, ctx);
6078 the_hash_algo->init_fn(ctx);
6079 /* 20-byte sum, with carry */
6080 for (i = 0; i < the_hash_algo->rawsz; ++i) {
6081 carry += result->hash[i] + hash[i];
6082 result->hash[i] = carry;
6083 carry >>= 8;
6087 static int patch_id_consume(void *priv, char *line, unsigned long len)
6089 struct patch_id_t *data = priv;
6090 int new_len;
6092 if (len > 12 && starts_with(line, "\\ "))
6093 return 0;
6094 new_len = remove_space(line, len);
6096 the_hash_algo->update_fn(data->ctx, line, new_len);
6097 data->patchlen += new_len;
6098 return 0;
6101 static void patch_id_add_string(git_hash_ctx *ctx, const char *str)
6103 the_hash_algo->update_fn(ctx, str, strlen(str));
6106 static void patch_id_add_mode(git_hash_ctx *ctx, unsigned mode)
6108 /* large enough for 2^32 in octal */
6109 char buf[12];
6110 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
6111 the_hash_algo->update_fn(ctx, buf, len);
6114 /* returns 0 upon success, and writes result into oid */
6115 static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only, int stable)
6117 struct diff_queue_struct *q = &diff_queued_diff;
6118 int i;
6119 git_hash_ctx ctx;
6120 struct patch_id_t data;
6122 the_hash_algo->init_fn(&ctx);
6123 memset(&data, 0, sizeof(struct patch_id_t));
6124 data.ctx = &ctx;
6125 oidclr(oid);
6127 for (i = 0; i < q->nr; i++) {
6128 xpparam_t xpp;
6129 xdemitconf_t xecfg;
6130 mmfile_t mf1, mf2;
6131 struct diff_filepair *p = q->queue[i];
6132 int len1, len2;
6134 memset(&xpp, 0, sizeof(xpp));
6135 memset(&xecfg, 0, sizeof(xecfg));
6136 if (p->status == 0)
6137 return error("internal diff status error");
6138 if (p->status == DIFF_STATUS_UNKNOWN)
6139 continue;
6140 if (diff_unmodified_pair(p))
6141 continue;
6142 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
6143 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
6144 continue;
6145 if (DIFF_PAIR_UNMERGED(p))
6146 continue;
6148 diff_fill_oid_info(p->one, options->repo->index);
6149 diff_fill_oid_info(p->two, options->repo->index);
6151 len1 = remove_space(p->one->path, strlen(p->one->path));
6152 len2 = remove_space(p->two->path, strlen(p->two->path));
6153 patch_id_add_string(&ctx, "diff--git");
6154 patch_id_add_string(&ctx, "a/");
6155 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6156 patch_id_add_string(&ctx, "b/");
6157 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6159 if (p->one->mode == 0) {
6160 patch_id_add_string(&ctx, "newfilemode");
6161 patch_id_add_mode(&ctx, p->two->mode);
6162 patch_id_add_string(&ctx, "---/dev/null");
6163 patch_id_add_string(&ctx, "+++b/");
6164 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6165 } else if (p->two->mode == 0) {
6166 patch_id_add_string(&ctx, "deletedfilemode");
6167 patch_id_add_mode(&ctx, p->one->mode);
6168 patch_id_add_string(&ctx, "---a/");
6169 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6170 patch_id_add_string(&ctx, "+++/dev/null");
6171 } else {
6172 patch_id_add_string(&ctx, "---a/");
6173 the_hash_algo->update_fn(&ctx, p->one->path, len1);
6174 patch_id_add_string(&ctx, "+++b/");
6175 the_hash_algo->update_fn(&ctx, p->two->path, len2);
6178 if (diff_header_only)
6179 continue;
6181 if (fill_mmfile(options->repo, &mf1, p->one) < 0 ||
6182 fill_mmfile(options->repo, &mf2, p->two) < 0)
6183 return error("unable to read files to diff");
6185 if (diff_filespec_is_binary(options->repo, p->one) ||
6186 diff_filespec_is_binary(options->repo, p->two)) {
6187 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->one->oid),
6188 the_hash_algo->hexsz);
6189 the_hash_algo->update_fn(&ctx, oid_to_hex(&p->two->oid),
6190 the_hash_algo->hexsz);
6191 continue;
6194 xpp.flags = 0;
6195 xecfg.ctxlen = 3;
6196 xecfg.flags = XDL_EMIT_NO_HUNK_HDR;
6197 if (xdi_diff_outf(&mf1, &mf2, NULL,
6198 patch_id_consume, &data, &xpp, &xecfg))
6199 return error("unable to generate patch-id diff for %s",
6200 p->one->path);
6202 if (stable)
6203 flush_one_hunk(oid, &ctx);
6206 if (!stable)
6207 the_hash_algo->final_oid_fn(oid, &ctx);
6209 return 0;
6212 int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only, int stable)
6214 struct diff_queue_struct *q = &diff_queued_diff;
6215 int i;
6216 int result = diff_get_patch_id(options, oid, diff_header_only, stable);
6218 for (i = 0; i < q->nr; i++)
6219 diff_free_filepair(q->queue[i]);
6221 free(q->queue);
6222 DIFF_QUEUE_CLEAR(q);
6224 return result;
6227 static int is_summary_empty(const struct diff_queue_struct *q)
6229 int i;
6231 for (i = 0; i < q->nr; i++) {
6232 const struct diff_filepair *p = q->queue[i];
6234 switch (p->status) {
6235 case DIFF_STATUS_DELETED:
6236 case DIFF_STATUS_ADDED:
6237 case DIFF_STATUS_COPIED:
6238 case DIFF_STATUS_RENAMED:
6239 return 0;
6240 default:
6241 if (p->score)
6242 return 0;
6243 if (p->one->mode && p->two->mode &&
6244 p->one->mode != p->two->mode)
6245 return 0;
6246 break;
6249 return 1;
6252 static const char rename_limit_warning[] =
6253 N_("exhaustive rename detection was skipped due to too many files.");
6255 static const char degrade_cc_to_c_warning[] =
6256 N_("only found copies from modified paths due to too many files.");
6258 static const char rename_limit_advice[] =
6259 N_("you may want to set your %s variable to at least "
6260 "%d and retry the command.");
6262 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
6264 fflush(stdout);
6265 if (degraded_cc)
6266 warning(_(degrade_cc_to_c_warning));
6267 else if (needed)
6268 warning(_(rename_limit_warning));
6269 else
6270 return;
6271 if (0 < needed)
6272 warning(_(rename_limit_advice), varname, needed);
6275 static void diff_flush_patch_all_file_pairs(struct diff_options *o)
6277 int i;
6278 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
6279 struct diff_queue_struct *q = &diff_queued_diff;
6281 if (WSEH_NEW & WS_RULE_MASK)
6282 BUG("WS rules bit mask overlaps with diff symbol flags");
6284 if (o->color_moved)
6285 o->emitted_symbols = &esm;
6287 for (i = 0; i < q->nr; i++) {
6288 struct diff_filepair *p = q->queue[i];
6289 if (check_pair_status(p))
6290 diff_flush_patch(p, o);
6293 if (o->emitted_symbols) {
6294 if (o->color_moved) {
6295 struct mem_pool entry_pool;
6296 struct moved_entry_list *entry_list;
6298 mem_pool_init(&entry_pool, 1024 * 1024);
6299 entry_list = add_lines_to_move_detection(o,
6300 &entry_pool);
6301 mark_color_as_moved(o, entry_list);
6302 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
6303 dim_moved_lines(o);
6305 mem_pool_discard(&entry_pool, 0);
6306 free(entry_list);
6309 for (i = 0; i < esm.nr; i++)
6310 emit_diff_symbol_from_struct(o, &esm.buf[i]);
6312 for (i = 0; i < esm.nr; i++)
6313 free((void *)esm.buf[i].line);
6314 esm.nr = 0;
6316 o->emitted_symbols = NULL;
6320 static void diff_free_file(struct diff_options *options)
6322 if (options->close_file)
6323 fclose(options->file);
6326 static void diff_free_ignore_regex(struct diff_options *options)
6328 int i;
6330 for (i = 0; i < options->ignore_regex_nr; i++) {
6331 regfree(options->ignore_regex[i]);
6332 free(options->ignore_regex[i]);
6334 free(options->ignore_regex);
6337 void diff_free(struct diff_options *options)
6339 if (options->no_free)
6340 return;
6342 diff_free_file(options);
6343 diff_free_ignore_regex(options);
6346 void diff_flush(struct diff_options *options)
6348 struct diff_queue_struct *q = &diff_queued_diff;
6349 int i, output_format = options->output_format;
6350 int separator = 0;
6351 int dirstat_by_line = 0;
6354 * Order: raw, stat, summary, patch
6355 * or: name/name-status/checkdiff (other bits clear)
6357 if (!q->nr)
6358 goto free_queue;
6360 if (output_format & (DIFF_FORMAT_RAW |
6361 DIFF_FORMAT_NAME |
6362 DIFF_FORMAT_NAME_STATUS |
6363 DIFF_FORMAT_CHECKDIFF)) {
6364 for (i = 0; i < q->nr; i++) {
6365 struct diff_filepair *p = q->queue[i];
6366 if (check_pair_status(p))
6367 flush_one_pair(p, options);
6369 separator++;
6372 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
6373 dirstat_by_line = 1;
6375 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
6376 dirstat_by_line) {
6377 struct diffstat_t diffstat;
6379 compute_diffstat(options, &diffstat, q);
6380 if (output_format & DIFF_FORMAT_NUMSTAT)
6381 show_numstat(&diffstat, options);
6382 if (output_format & DIFF_FORMAT_DIFFSTAT)
6383 show_stats(&diffstat, options);
6384 if (output_format & DIFF_FORMAT_SHORTSTAT)
6385 show_shortstats(&diffstat, options);
6386 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
6387 show_dirstat_by_line(&diffstat, options);
6388 free_diffstat_info(&diffstat);
6389 separator++;
6391 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
6392 show_dirstat(options);
6394 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
6395 for (i = 0; i < q->nr; i++) {
6396 diff_summary(options, q->queue[i]);
6398 separator++;
6401 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
6402 options->flags.exit_with_status &&
6403 options->flags.diff_from_contents) {
6405 * run diff_flush_patch for the exit status. setting
6406 * options->file to /dev/null should be safe, because we
6407 * aren't supposed to produce any output anyway.
6409 diff_free_file(options);
6410 options->file = xfopen("/dev/null", "w");
6411 options->close_file = 1;
6412 options->color_moved = 0;
6413 for (i = 0; i < q->nr; i++) {
6414 struct diff_filepair *p = q->queue[i];
6415 if (check_pair_status(p))
6416 diff_flush_patch(p, options);
6417 if (options->found_changes)
6418 break;
6422 if (output_format & DIFF_FORMAT_PATCH) {
6423 if (separator) {
6424 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
6425 if (options->stat_sep)
6426 /* attach patch instead of inline */
6427 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
6428 NULL, 0, 0);
6431 diff_flush_patch_all_file_pairs(options);
6434 if (output_format & DIFF_FORMAT_CALLBACK)
6435 options->format_callback(q, options, options->format_callback_data);
6437 for (i = 0; i < q->nr; i++)
6438 diff_free_filepair(q->queue[i]);
6439 free_queue:
6440 free(q->queue);
6441 DIFF_QUEUE_CLEAR(q);
6442 diff_free(options);
6445 * Report the content-level differences with HAS_CHANGES;
6446 * diff_addremove/diff_change does not set the bit when
6447 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
6449 if (options->flags.diff_from_contents) {
6450 if (options->found_changes)
6451 options->flags.has_changes = 1;
6452 else
6453 options->flags.has_changes = 0;
6457 static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
6459 return (((p->status == DIFF_STATUS_MODIFIED) &&
6460 ((p->score &&
6461 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
6462 (!p->score &&
6463 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
6464 ((p->status != DIFF_STATUS_MODIFIED) &&
6465 filter_bit_tst(p->status, options)));
6468 static void diffcore_apply_filter(struct diff_options *options)
6470 int i;
6471 struct diff_queue_struct *q = &diff_queued_diff;
6472 struct diff_queue_struct outq;
6474 DIFF_QUEUE_CLEAR(&outq);
6476 if (!options->filter)
6477 return;
6479 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
6480 int found;
6481 for (i = found = 0; !found && i < q->nr; i++) {
6482 if (match_filter(options, q->queue[i]))
6483 found++;
6485 if (found)
6486 return;
6488 /* otherwise we will clear the whole queue
6489 * by copying the empty outq at the end of this
6490 * function, but first clear the current entries
6491 * in the queue.
6493 for (i = 0; i < q->nr; i++)
6494 diff_free_filepair(q->queue[i]);
6496 else {
6497 /* Only the matching ones */
6498 for (i = 0; i < q->nr; i++) {
6499 struct diff_filepair *p = q->queue[i];
6500 if (match_filter(options, p))
6501 diff_q(&outq, p);
6502 else
6503 diff_free_filepair(p);
6506 free(q->queue);
6507 *q = outq;
6510 /* Check whether two filespecs with the same mode and size are identical */
6511 static int diff_filespec_is_identical(struct repository *r,
6512 struct diff_filespec *one,
6513 struct diff_filespec *two)
6515 if (S_ISGITLINK(one->mode))
6516 return 0;
6517 if (diff_populate_filespec(r, one, NULL))
6518 return 0;
6519 if (diff_populate_filespec(r, two, NULL))
6520 return 0;
6521 return !memcmp(one->data, two->data, one->size);
6524 static int diff_filespec_check_stat_unmatch(struct repository *r,
6525 struct diff_filepair *p)
6527 struct diff_populate_filespec_options dpf_options = {
6528 .check_size_only = 1,
6529 .missing_object_cb = diff_queued_diff_prefetch,
6530 .missing_object_data = r,
6533 if (p->done_skip_stat_unmatch)
6534 return p->skip_stat_unmatch_result;
6536 p->done_skip_stat_unmatch = 1;
6537 p->skip_stat_unmatch_result = 0;
6539 * 1. Entries that come from stat info dirtiness
6540 * always have both sides (iow, not create/delete),
6541 * one side of the object name is unknown, with
6542 * the same mode and size. Keep the ones that
6543 * do not match these criteria. They have real
6544 * differences.
6546 * 2. At this point, the file is known to be modified,
6547 * with the same mode and size, and the object
6548 * name of one side is unknown. Need to inspect
6549 * the identical contents.
6551 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6552 !DIFF_FILE_VALID(p->two) ||
6553 (p->one->oid_valid && p->two->oid_valid) ||
6554 (p->one->mode != p->two->mode) ||
6555 diff_populate_filespec(r, p->one, &dpf_options) ||
6556 diff_populate_filespec(r, p->two, &dpf_options) ||
6557 (p->one->size != p->two->size) ||
6558 !diff_filespec_is_identical(r, p->one, p->two)) /* (2) */
6559 p->skip_stat_unmatch_result = 1;
6560 return p->skip_stat_unmatch_result;
6563 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6565 int i;
6566 struct diff_queue_struct *q = &diff_queued_diff;
6567 struct diff_queue_struct outq;
6568 DIFF_QUEUE_CLEAR(&outq);
6570 for (i = 0; i < q->nr; i++) {
6571 struct diff_filepair *p = q->queue[i];
6573 if (diff_filespec_check_stat_unmatch(diffopt->repo, p))
6574 diff_q(&outq, p);
6575 else {
6577 * The caller can subtract 1 from skip_stat_unmatch
6578 * to determine how many paths were dirty only
6579 * due to stat info mismatch.
6581 if (!diffopt->flags.no_index)
6582 diffopt->skip_stat_unmatch++;
6583 diff_free_filepair(p);
6586 free(q->queue);
6587 *q = outq;
6590 static int diffnamecmp(const void *a_, const void *b_)
6592 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6593 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6594 const char *name_a, *name_b;
6596 name_a = a->one ? a->one->path : a->two->path;
6597 name_b = b->one ? b->one->path : b->two->path;
6598 return strcmp(name_a, name_b);
6601 void diffcore_fix_diff_index(void)
6603 struct diff_queue_struct *q = &diff_queued_diff;
6604 QSORT(q->queue, q->nr, diffnamecmp);
6607 void diff_add_if_missing(struct repository *r,
6608 struct oid_array *to_fetch,
6609 const struct diff_filespec *filespec)
6611 if (filespec && filespec->oid_valid &&
6612 !S_ISGITLINK(filespec->mode) &&
6613 oid_object_info_extended(r, &filespec->oid, NULL,
6614 OBJECT_INFO_FOR_PREFETCH))
6615 oid_array_append(to_fetch, &filespec->oid);
6618 void diff_queued_diff_prefetch(void *repository)
6620 struct repository *repo = repository;
6621 int i;
6622 struct diff_queue_struct *q = &diff_queued_diff;
6623 struct oid_array to_fetch = OID_ARRAY_INIT;
6625 for (i = 0; i < q->nr; i++) {
6626 struct diff_filepair *p = q->queue[i];
6627 diff_add_if_missing(repo, &to_fetch, p->one);
6628 diff_add_if_missing(repo, &to_fetch, p->two);
6632 * NEEDSWORK: Consider deduplicating the OIDs sent.
6634 promisor_remote_get_direct(repo, to_fetch.oid, to_fetch.nr);
6636 oid_array_clear(&to_fetch);
6639 void diffcore_std(struct diff_options *options)
6641 int output_formats_to_prefetch = DIFF_FORMAT_DIFFSTAT |
6642 DIFF_FORMAT_NUMSTAT |
6643 DIFF_FORMAT_PATCH |
6644 DIFF_FORMAT_SHORTSTAT |
6645 DIFF_FORMAT_DIRSTAT;
6648 * Check if the user requested a blob-data-requiring diff output and/or
6649 * break-rewrite detection (which requires blob data). If yes, prefetch
6650 * the diff pairs.
6652 * If no prefetching occurs, diffcore_rename() will prefetch if it
6653 * decides that it needs inexact rename detection.
6655 if (options->repo == the_repository && has_promisor_remote() &&
6656 (options->output_format & output_formats_to_prefetch ||
6657 options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
6658 diff_queued_diff_prefetch(options->repo);
6660 /* NOTE please keep the following in sync with diff_tree_combined() */
6661 if (options->skip_stat_unmatch)
6662 diffcore_skip_stat_unmatch(options);
6663 if (!options->found_follow) {
6664 /* See try_to_follow_renames() in tree-diff.c */
6665 if (options->break_opt != -1)
6666 diffcore_break(options->repo,
6667 options->break_opt);
6668 if (options->detect_rename)
6669 diffcore_rename(options);
6670 if (options->break_opt != -1)
6671 diffcore_merge_broken();
6673 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6674 diffcore_pickaxe(options);
6675 if (options->orderfile)
6676 diffcore_order(options->orderfile);
6677 if (options->rotate_to)
6678 diffcore_rotate(options);
6679 if (!options->found_follow)
6680 /* See try_to_follow_renames() in tree-diff.c */
6681 diff_resolve_rename_copy();
6682 diffcore_apply_filter(options);
6684 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6685 options->flags.has_changes = 1;
6686 else
6687 options->flags.has_changes = 0;
6689 options->found_follow = 0;
6692 int diff_result_code(struct diff_options *opt, int status)
6694 int result = 0;
6696 diff_warn_rename_limit("diff.renameLimit",
6697 opt->needed_rename_limit,
6698 opt->degraded_cc_to_c);
6699 if (!opt->flags.exit_with_status &&
6700 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6701 return status;
6702 if (opt->flags.exit_with_status &&
6703 opt->flags.has_changes)
6704 result |= 01;
6705 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6706 opt->flags.check_failed)
6707 result |= 02;
6708 return result;
6711 int diff_can_quit_early(struct diff_options *opt)
6713 return (opt->flags.quick &&
6714 !opt->filter &&
6715 opt->flags.has_changes);
6719 * Shall changes to this submodule be ignored?
6721 * Submodule changes can be configured to be ignored separately for each path,
6722 * but that configuration can be overridden from the command line.
6724 static int is_submodule_ignored(const char *path, struct diff_options *options)
6726 int ignored = 0;
6727 struct diff_flags orig_flags = options->flags;
6728 if (!options->flags.override_submodule_config)
6729 set_diffopt_flags_from_submodule_config(options, path);
6730 if (options->flags.ignore_submodules)
6731 ignored = 1;
6732 options->flags = orig_flags;
6733 return ignored;
6736 void compute_diffstat(struct diff_options *options,
6737 struct diffstat_t *diffstat,
6738 struct diff_queue_struct *q)
6740 int i;
6742 memset(diffstat, 0, sizeof(struct diffstat_t));
6743 for (i = 0; i < q->nr; i++) {
6744 struct diff_filepair *p = q->queue[i];
6745 if (check_pair_status(p))
6746 diff_flush_stat(p, options, diffstat);
6750 void diff_addremove(struct diff_options *options,
6751 int addremove, unsigned mode,
6752 const struct object_id *oid,
6753 int oid_valid,
6754 const char *concatpath, unsigned dirty_submodule)
6756 struct diff_filespec *one, *two;
6758 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6759 return;
6761 /* This may look odd, but it is a preparation for
6762 * feeding "there are unchanged files which should
6763 * not produce diffs, but when you are doing copy
6764 * detection you would need them, so here they are"
6765 * entries to the diff-core. They will be prefixed
6766 * with something like '=' or '*' (I haven't decided
6767 * which but should not make any difference).
6768 * Feeding the same new and old to diff_change()
6769 * also has the same effect.
6770 * Before the final output happens, they are pruned after
6771 * merged into rename/copy pairs as appropriate.
6773 if (options->flags.reverse_diff)
6774 addremove = (addremove == '+' ? '-' :
6775 addremove == '-' ? '+' : addremove);
6777 if (options->prefix &&
6778 strncmp(concatpath, options->prefix, options->prefix_length))
6779 return;
6781 one = alloc_filespec(concatpath);
6782 two = alloc_filespec(concatpath);
6784 if (addremove != '+')
6785 fill_filespec(one, oid, oid_valid, mode);
6786 if (addremove != '-') {
6787 fill_filespec(two, oid, oid_valid, mode);
6788 two->dirty_submodule = dirty_submodule;
6791 diff_queue(&diff_queued_diff, one, two);
6792 if (!options->flags.diff_from_contents)
6793 options->flags.has_changes = 1;
6796 void diff_change(struct diff_options *options,
6797 unsigned old_mode, unsigned new_mode,
6798 const struct object_id *old_oid,
6799 const struct object_id *new_oid,
6800 int old_oid_valid, int new_oid_valid,
6801 const char *concatpath,
6802 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6804 struct diff_filespec *one, *two;
6805 struct diff_filepair *p;
6807 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6808 is_submodule_ignored(concatpath, options))
6809 return;
6811 if (options->flags.reverse_diff) {
6812 SWAP(old_mode, new_mode);
6813 SWAP(old_oid, new_oid);
6814 SWAP(old_oid_valid, new_oid_valid);
6815 SWAP(old_dirty_submodule, new_dirty_submodule);
6818 if (options->prefix &&
6819 strncmp(concatpath, options->prefix, options->prefix_length))
6820 return;
6822 one = alloc_filespec(concatpath);
6823 two = alloc_filespec(concatpath);
6824 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6825 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6826 one->dirty_submodule = old_dirty_submodule;
6827 two->dirty_submodule = new_dirty_submodule;
6828 p = diff_queue(&diff_queued_diff, one, two);
6830 if (options->flags.diff_from_contents)
6831 return;
6833 if (options->flags.quick && options->skip_stat_unmatch &&
6834 !diff_filespec_check_stat_unmatch(options->repo, p)) {
6835 diff_free_filespec_data(p->one);
6836 diff_free_filespec_data(p->two);
6837 return;
6840 options->flags.has_changes = 1;
6843 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6845 struct diff_filepair *pair;
6846 struct diff_filespec *one, *two;
6848 if (options->prefix &&
6849 strncmp(path, options->prefix, options->prefix_length))
6850 return NULL;
6852 one = alloc_filespec(path);
6853 two = alloc_filespec(path);
6854 pair = diff_queue(&diff_queued_diff, one, two);
6855 pair->is_unmerged = 1;
6856 return pair;
6859 static char *run_textconv(struct repository *r,
6860 const char *pgm,
6861 struct diff_filespec *spec,
6862 size_t *outsize)
6864 struct diff_tempfile *temp;
6865 struct child_process child = CHILD_PROCESS_INIT;
6866 struct strbuf buf = STRBUF_INIT;
6867 int err = 0;
6869 temp = prepare_temp_file(r, spec->path, spec);
6870 strvec_push(&child.args, pgm);
6871 strvec_push(&child.args, temp->name);
6873 child.use_shell = 1;
6874 child.out = -1;
6875 if (start_command(&child)) {
6876 remove_tempfile();
6877 return NULL;
6880 if (strbuf_read(&buf, child.out, 0) < 0)
6881 err = error("error reading from textconv command '%s'", pgm);
6882 close(child.out);
6884 if (finish_command(&child) || err) {
6885 strbuf_release(&buf);
6886 remove_tempfile();
6887 return NULL;
6889 remove_tempfile();
6891 return strbuf_detach(&buf, outsize);
6894 size_t fill_textconv(struct repository *r,
6895 struct userdiff_driver *driver,
6896 struct diff_filespec *df,
6897 char **outbuf)
6899 size_t size;
6901 if (!driver) {
6902 if (!DIFF_FILE_VALID(df)) {
6903 *outbuf = "";
6904 return 0;
6906 if (diff_populate_filespec(r, df, NULL))
6907 die("unable to read files to diff");
6908 *outbuf = df->data;
6909 return df->size;
6912 if (!driver->textconv)
6913 BUG("fill_textconv called with non-textconv driver");
6915 if (driver->textconv_cache && df->oid_valid) {
6916 *outbuf = notes_cache_get(driver->textconv_cache,
6917 &df->oid,
6918 &size);
6919 if (*outbuf)
6920 return size;
6923 *outbuf = run_textconv(r, driver->textconv, df, &size);
6924 if (!*outbuf)
6925 die("unable to read files to diff");
6927 if (driver->textconv_cache && df->oid_valid) {
6928 /* ignore errors, as we might be in a readonly repository */
6929 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6930 size);
6932 * we could save up changes and flush them all at the end,
6933 * but we would need an extra call after all diffing is done.
6934 * Since generating a cache entry is the slow path anyway,
6935 * this extra overhead probably isn't a big deal.
6937 notes_cache_write(driver->textconv_cache);
6940 return size;
6943 int textconv_object(struct repository *r,
6944 const char *path,
6945 unsigned mode,
6946 const struct object_id *oid,
6947 int oid_valid,
6948 char **buf,
6949 unsigned long *buf_size)
6951 struct diff_filespec *df;
6952 struct userdiff_driver *textconv;
6954 df = alloc_filespec(path);
6955 fill_filespec(df, oid, oid_valid, mode);
6956 textconv = get_textconv(r, df);
6957 if (!textconv) {
6958 free_filespec(df);
6959 return 0;
6962 *buf_size = fill_textconv(r, textconv, df, buf);
6963 free_filespec(df);
6964 return 1;
6967 void setup_diff_pager(struct diff_options *opt)
6970 * If the user asked for our exit code, then either they want --quiet
6971 * or --exit-code. We should definitely not bother with a pager in the
6972 * former case, as we will generate no output. Since we still properly
6973 * report our exit code even when a pager is run, we _could_ run a
6974 * pager with --exit-code. But since we have not done so historically,
6975 * and because it is easy to find people oneline advising "git diff
6976 * --exit-code" in hooks and other scripts, we do not do so.
6978 if (!opt->flags.exit_with_status &&
6979 check_pager_config("diff") != 0)
6980 setup_pager();