c5c7739ce3452a8de97d6b6baf882cea3cbb9ef5
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 "ll-merge.h"
22#include "string-list.h"
23#include "argv-array.h"
24#include "graph.h"
25#include "packfile.h"
26#include "help.h"
27#ifdef NO_FAST_WORKING_DIRECTORY
29#define FAST_WORKING_DIRECTORY 0
30#else
31#define FAST_WORKING_DIRECTORY 1
32#endif
33static int diff_detect_rename_default;
35static int diff_indent_heuristic = 1;
36static int diff_rename_limit_default = 400;
37static int diff_suppress_blank_empty;
38static int diff_use_color_default = -1;
39static int diff_color_moved_default;
40static int diff_color_moved_ws_default;
41static int diff_context_default = 3;
42static int diff_interhunk_context_default;
43static const char *diff_word_regex_cfg;
44static const char *external_diff_cmd_cfg;
45static const char *diff_order_file_cfg;
46int diff_auto_refresh_index = 1;
47static int diff_mnemonic_prefix;
48static int diff_no_prefix;
49static int diff_stat_graph_width;
50static int diff_dirstat_permille_default = 30;
51static struct diff_options default_diff_options;
52static long diff_algorithm;
53static unsigned ws_error_highlight_default = WSEH_NEW;
54static char diff_colors[][COLOR_MAXLEN] = {
56 GIT_COLOR_RESET,
57 GIT_COLOR_NORMAL, /* CONTEXT */
58 GIT_COLOR_BOLD, /* METAINFO */
59 GIT_COLOR_CYAN, /* FRAGINFO */
60 GIT_COLOR_RED, /* OLD */
61 GIT_COLOR_GREEN, /* NEW */
62 GIT_COLOR_YELLOW, /* COMMIT */
63 GIT_COLOR_BG_RED, /* WHITESPACE */
64 GIT_COLOR_NORMAL, /* FUNCINFO */
65 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
66 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
67 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
68 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
69 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
70 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
71 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
72 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
73 GIT_COLOR_FAINT, /* CONTEXT_DIM */
74 GIT_COLOR_FAINT_RED, /* OLD_DIM */
75 GIT_COLOR_FAINT_GREEN, /* NEW_DIM */
76 GIT_COLOR_BOLD, /* CONTEXT_BOLD */
77 GIT_COLOR_BOLD_RED, /* OLD_BOLD */
78 GIT_COLOR_BOLD_GREEN, /* NEW_BOLD */
79};
80static const char *color_diff_slots[] = {
82 [DIFF_CONTEXT] = "context",
83 [DIFF_METAINFO] = "meta",
84 [DIFF_FRAGINFO] = "frag",
85 [DIFF_FILE_OLD] = "old",
86 [DIFF_FILE_NEW] = "new",
87 [DIFF_COMMIT] = "commit",
88 [DIFF_WHITESPACE] = "whitespace",
89 [DIFF_FUNCINFO] = "func",
90 [DIFF_FILE_OLD_MOVED] = "oldMoved",
91 [DIFF_FILE_OLD_MOVED_ALT] = "oldMovedAlternative",
92 [DIFF_FILE_OLD_MOVED_DIM] = "oldMovedDimmed",
93 [DIFF_FILE_OLD_MOVED_ALT_DIM] = "oldMovedAlternativeDimmed",
94 [DIFF_FILE_NEW_MOVED] = "newMoved",
95 [DIFF_FILE_NEW_MOVED_ALT] = "newMovedAlternative",
96 [DIFF_FILE_NEW_MOVED_DIM] = "newMovedDimmed",
97 [DIFF_FILE_NEW_MOVED_ALT_DIM] = "newMovedAlternativeDimmed",
98 [DIFF_CONTEXT_DIM] = "contextDimmed",
99 [DIFF_FILE_OLD_DIM] = "oldDimmed",
100 [DIFF_FILE_NEW_DIM] = "newDimmed",
101 [DIFF_CONTEXT_BOLD] = "contextBold",
102 [DIFF_FILE_OLD_BOLD] = "oldBold",
103 [DIFF_FILE_NEW_BOLD] = "newBold",
104};
105static NORETURN void die_want_option(const char *option_name)
107{
108 die(_("option '%s' requires a value"), option_name);
109}
110define_list_config_array_extra(color_diff_slots, {"plain"});
112static int parse_diff_color_slot(const char *var)
114{
115 if (!strcasecmp(var, "plain"))
116 return DIFF_CONTEXT;
117 return LOOKUP_CONFIG(color_diff_slots, var);
118}
119static int parse_dirstat_params(struct diff_options *options, const char *params_string,
121 struct strbuf *errmsg)
122{
123 char *params_copy = xstrdup(params_string);
124 struct string_list params = STRING_LIST_INIT_NODUP;
125 int ret = 0;
126 int i;
127 if (*params_copy)
129 string_list_split_in_place(¶ms, 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 */
154 }
155 if (!*end)
156 options->dirstat_permille = permille;
157 else {
158 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
159 p);
160 ret++;
161 }
162 } else {
163 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
164 ret++;
165 }
166 }
168 string_list_clear(¶ms, 0);
169 free(params_copy);
170 return ret;
171}
172static int parse_submodule_params(struct diff_options *options, const char *value)
174{
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;
181 else
182 return -1;
183 return 0;
184}
185int git_config_rename(const char *var, const char *value)
187{
188 if (!value)
189 return DIFF_DETECT_RENAME;
190 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
191 return DIFF_DETECT_COPY;
192 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
193}
194long parse_algorithm_value(const char *value)
196{
197 if (!value)
198 return -1;
199 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
200 return 0;
201 else if (!strcasecmp(value, "minimal"))
202 return XDF_NEED_MINIMAL;
203 else if (!strcasecmp(value, "patience"))
204 return XDF_PATIENCE_DIFF;
205 else if (!strcasecmp(value, "histogram"))
206 return XDF_HISTOGRAM_DIFF;
207 return -1;
208}
209static int parse_one_token(const char **arg, const char *token)
211{
212 const char *rest;
213 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
214 *arg = rest;
215 return 1;
216 }
217 return 0;
218}
219static int parse_ws_error_highlight(const char *arg)
221{
222 const char *orig_arg = arg;
223 unsigned val = 0;
224 while (*arg) {
226 if (parse_one_token(&arg, "none"))
227 val = 0;
228 else if (parse_one_token(&arg, "default"))
229 val = WSEH_NEW;
230 else if (parse_one_token(&arg, "all"))
231 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
232 else if (parse_one_token(&arg, "new"))
233 val |= WSEH_NEW;
234 else if (parse_one_token(&arg, "old"))
235 val |= WSEH_OLD;
236 else if (parse_one_token(&arg, "context"))
237 val |= WSEH_CONTEXT;
238 else {
239 return -1 - (int)(arg - orig_arg);
240 }
241 if (*arg)
242 arg++;
243 }
244 return val;
245}
246/*
248 * These are to give UI layer defaults.
249 * The core-level commands such as git-diff-files should
250 * never be affected by the setting of diff.renames
251 * the user happens to have in the configuration file.
252 */
253void init_diff_ui_defaults(void)
254{
255 diff_detect_rename_default = DIFF_DETECT_RENAME;
256}
257int git_diff_heuristic_config(const char *var, const char *value, void *cb)
259{
260 if (!strcmp(var, "diff.indentheuristic"))
261 diff_indent_heuristic = git_config_bool(var, value);
262 return 0;
263}
264static int parse_color_moved(const char *arg)
266{
267 switch (git_parse_maybe_bool(arg)) {
268 case 0:
269 return COLOR_MOVED_NO;
270 case 1:
271 return COLOR_MOVED_DEFAULT;
272 default:
273 break;
274 }
275 if (!strcmp(arg, "no"))
277 return COLOR_MOVED_NO;
278 else if (!strcmp(arg, "plain"))
279 return COLOR_MOVED_PLAIN;
280 else if (!strcmp(arg, "blocks"))
281 return COLOR_MOVED_BLOCKS;
282 else if (!strcmp(arg, "zebra"))
283 return COLOR_MOVED_ZEBRA;
284 else if (!strcmp(arg, "default"))
285 return COLOR_MOVED_DEFAULT;
286 else if (!strcmp(arg, "dimmed_zebra"))
287 return COLOR_MOVED_ZEBRA_DIM;
288 else
289 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed_zebra', 'plain'"));
290}
291static int parse_color_moved_ws(const char *arg)
293{
294 int ret = 0;
295 struct string_list l = STRING_LIST_INIT_DUP;
296 struct string_list_item *i;
297 string_list_split(&l, arg, ',', -1);
299 for_each_string_list_item(i, &l) {
301 struct strbuf sb = STRBUF_INIT;
302 strbuf_addstr(&sb, i->string);
303 strbuf_trim(&sb);
304 if (!strcmp(sb.buf, "ignore-space-change"))
306 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
307 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
308 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
309 else if (!strcmp(sb.buf, "ignore-all-space"))
310 ret |= XDF_IGNORE_WHITESPACE;
311 else if (!strcmp(sb.buf, "allow-indentation-change"))
312 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
313 else
314 error(_("ignoring unknown color-moved-ws mode '%s'"), sb.buf);
315 strbuf_release(&sb);
317 }
318 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
320 (ret & XDF_WHITESPACE_FLAGS))
321 die(_("color-moved-ws: allow-indentation-change cannot be combined with other white space modes"));
322 string_list_clear(&l, 0);
324 return ret;
326}
327int git_diff_ui_config(const char *var, const char *value, void *cb)
329{
330 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
331 diff_use_color_default = git_config_colorbool(var, value);
332 return 0;
333 }
334 if (!strcmp(var, "diff.colormoved")) {
335 int cm = parse_color_moved(value);
336 if (cm < 0)
337 return -1;
338 diff_color_moved_default = cm;
339 return 0;
340 }
341 if (!strcmp(var, "diff.colormovedws")) {
342 int cm = parse_color_moved_ws(value);
343 if (cm < 0)
344 return -1;
345 diff_color_moved_ws_default = cm;
346 return 0;
347 }
348 if (!strcmp(var, "diff.context")) {
349 diff_context_default = git_config_int(var, value);
350 if (diff_context_default < 0)
351 return -1;
352 return 0;
353 }
354 if (!strcmp(var, "diff.interhunkcontext")) {
355 diff_interhunk_context_default = git_config_int(var, value);
356 if (diff_interhunk_context_default < 0)
357 return -1;
358 return 0;
359 }
360 if (!strcmp(var, "diff.renames")) {
361 diff_detect_rename_default = git_config_rename(var, value);
362 return 0;
363 }
364 if (!strcmp(var, "diff.autorefreshindex")) {
365 diff_auto_refresh_index = git_config_bool(var, value);
366 return 0;
367 }
368 if (!strcmp(var, "diff.mnemonicprefix")) {
369 diff_mnemonic_prefix = git_config_bool(var, value);
370 return 0;
371 }
372 if (!strcmp(var, "diff.noprefix")) {
373 diff_no_prefix = git_config_bool(var, value);
374 return 0;
375 }
376 if (!strcmp(var, "diff.statgraphwidth")) {
377 diff_stat_graph_width = git_config_int(var, value);
378 return 0;
379 }
380 if (!strcmp(var, "diff.external"))
381 return git_config_string(&external_diff_cmd_cfg, var, value);
382 if (!strcmp(var, "diff.wordregex"))
383 return git_config_string(&diff_word_regex_cfg, var, value);
384 if (!strcmp(var, "diff.orderfile"))
385 return git_config_pathname(&diff_order_file_cfg, var, value);
386 if (!strcmp(var, "diff.ignoresubmodules"))
388 handle_ignore_submodules_arg(&default_diff_options, value);
389 if (!strcmp(var, "diff.submodule")) {
391 if (parse_submodule_params(&default_diff_options, value))
392 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
393 value);
394 return 0;
395 }
396 if (!strcmp(var, "diff.algorithm")) {
398 diff_algorithm = parse_algorithm_value(value);
399 if (diff_algorithm < 0)
400 return -1;
401 return 0;
402 }
403 if (!strcmp(var, "diff.wserrorhighlight")) {
405 int val = parse_ws_error_highlight(value);
406 if (val < 0)
407 return -1;
408 ws_error_highlight_default = val;
409 return 0;
410 }
411 if (git_color_config(var, value, cb) < 0)
413 return -1;
414 return git_diff_basic_config(var, value, cb);
416}
417int git_diff_basic_config(const char *var, const char *value, void *cb)
419{
420 const char *name;
421 if (!strcmp(var, "diff.renamelimit")) {
423 diff_rename_limit_default = git_config_int(var, value);
424 return 0;
425 }
426 if (userdiff_config(var, value) < 0)
428 return -1;
429 if (skip_prefix(var, "diff.color.", &name) ||
431 skip_prefix(var, "color.diff.", &name)) {
432 int slot = parse_diff_color_slot(name);
433 if (slot < 0)
434 return 0;
435 if (!value)
436 return config_error_nonbool(var);
437 return color_parse(value, diff_colors[slot]);
438 }
439 /* like GNU diff's --suppress-blank-empty option */
441 if (!strcmp(var, "diff.suppressblankempty") ||
442 /* for backwards compatibility */
443 !strcmp(var, "diff.suppress-blank-empty")) {
444 diff_suppress_blank_empty = git_config_bool(var, value);
445 return 0;
446 }
447 if (!strcmp(var, "diff.dirstat")) {
449 struct strbuf errmsg = STRBUF_INIT;
450 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
451 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
452 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
453 errmsg.buf);
454 strbuf_release(&errmsg);
455 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
456 return 0;
457 }
458 if (git_diff_heuristic_config(var, value, cb) < 0)
460 return -1;
461 return git_default_config(var, value, cb);
463}
464static char *quote_two(const char *one, const char *two)
466{
467 int need_one = quote_c_style(one, NULL, NULL, 1);
468 int need_two = quote_c_style(two, NULL, NULL, 1);
469 struct strbuf res = STRBUF_INIT;
470 if (need_one + need_two) {
472 strbuf_addch(&res, '"');
473 quote_c_style(one, &res, NULL, 1);
474 quote_c_style(two, &res, NULL, 1);
475 strbuf_addch(&res, '"');
476 } else {
477 strbuf_addstr(&res, one);
478 strbuf_addstr(&res, two);
479 }
480 return strbuf_detach(&res, NULL);
481}
482static const char *external_diff(void)
484{
485 static const char *external_diff_cmd = NULL;
486 static int done_preparing = 0;
487 if (done_preparing)
489 return external_diff_cmd;
490 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
491 if (!external_diff_cmd)
492 external_diff_cmd = external_diff_cmd_cfg;
493 done_preparing = 1;
494 return external_diff_cmd;
495}
496/*
498 * Keep track of files used for diffing. Sometimes such an entry
499 * refers to a temporary file, sometimes to an existing file, and
500 * sometimes to "/dev/null".
501 */
502static struct diff_tempfile {
503 /*
504 * filename external diff should read from, or NULL if this
505 * entry is currently not in use:
506 */
507 const char *name;
508 char hex[GIT_MAX_HEXSZ + 1];
510 char mode[10];
511 /*
513 * If this diff_tempfile instance refers to a temporary file,
514 * this tempfile object is used to manage its lifetime.
515 */
516 struct tempfile *tempfile;
517} diff_temp[2];
518struct emit_callback {
520 int color_diff;
521 unsigned ws_rule;
522 int blank_at_eof_in_preimage;
523 int blank_at_eof_in_postimage;
524 int lno_in_preimage;
525 int lno_in_postimage;
526 const char **label_path;
527 struct diff_words_data *diff_words;
528 struct diff_options *opt;
529 struct strbuf *header;
530};
531static int count_lines(const char *data, int size)
533{
534 int count, ch, completely_empty = 1, nl_just_seen = 0;
535 count = 0;
536 while (0 < size--) {
537 ch = *data++;
538 if (ch == '\n') {
539 count++;
540 nl_just_seen = 1;
541 completely_empty = 0;
542 }
543 else {
544 nl_just_seen = 0;
545 completely_empty = 0;
546 }
547 }
548 if (completely_empty)
549 return 0;
550 if (!nl_just_seen)
551 count++; /* no trailing newline */
552 return count;
553}
554static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
556{
557 if (!DIFF_FILE_VALID(one)) {
558 mf->ptr = (char *)""; /* does not matter */
559 mf->size = 0;
560 return 0;
561 }
562 else if (diff_populate_filespec(one, 0))
563 return -1;
564 mf->ptr = one->data;
566 mf->size = one->size;
567 return 0;
568}
569/* like fill_mmfile, but only for size, so we can avoid retrieving blob */
571static unsigned long diff_filespec_size(struct diff_filespec *one)
572{
573 if (!DIFF_FILE_VALID(one))
574 return 0;
575 diff_populate_filespec(one, CHECK_SIZE_ONLY);
576 return one->size;
577}
578static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
580{
581 char *ptr = mf->ptr;
582 long size = mf->size;
583 int cnt = 0;
584 if (!size)
586 return cnt;
587 ptr += size - 1; /* pointing at the very end */
588 if (*ptr != '\n')
589 ; /* incomplete line */
590 else
591 ptr--; /* skip the last LF */
592 while (mf->ptr < ptr) {
593 char *prev_eol;
594 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
595 if (*prev_eol == '\n')
596 break;
597 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
598 break;
599 cnt++;
600 ptr = prev_eol - 1;
601 }
602 return cnt;
603}
604static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
606 struct emit_callback *ecbdata)
607{
608 int l1, l2, at;
609 unsigned ws_rule = ecbdata->ws_rule;
610 l1 = count_trailing_blank(mf1, ws_rule);
611 l2 = count_trailing_blank(mf2, ws_rule);
612 if (l2 <= l1) {
613 ecbdata->blank_at_eof_in_preimage = 0;
614 ecbdata->blank_at_eof_in_postimage = 0;
615 return;
616 }
617 at = count_lines(mf1->ptr, mf1->size);
618 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
619 at = count_lines(mf2->ptr, mf2->size);
621 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
622}
623static void emit_line_0(struct diff_options *o,
625 const char *set_sign, const char *set, unsigned reverse, const char *reset,
626 int first, const char *line, int len)
627{
628 int has_trailing_newline, has_trailing_carriage_return;
629 int needs_reset = 0; /* at the end of the line */
630 FILE *file = o->file;
631 fputs(diff_line_prefix(o), file);
633 has_trailing_newline = (len > 0 && line[len-1] == '\n');
635 if (has_trailing_newline)
636 len--;
637 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
639 if (has_trailing_carriage_return)
640 len--;
641 if (!len && !first)
643 goto end_of_line;
644 if (reverse && want_color(o->use_color)) {
646 fputs(GIT_COLOR_REVERSE, file);
647 needs_reset = 1;
648 }
649 if (set_sign) {
651 fputs(set_sign, file);
652 needs_reset = 1;
653 }
654 if (first)
656 fputc(first, file);
657 if (!len)
659 goto end_of_line;
660 if (set) {
662 if (set_sign && set != set_sign)
663 fputs(reset, file);
664 fputs(set, file);
665 needs_reset = 1;
666 }
667 fwrite(line, len, 1, file);
668 needs_reset = 1; /* 'line' may contain color codes. */
669end_of_line:
671 if (needs_reset)
672 fputs(reset, file);
673 if (has_trailing_carriage_return)
674 fputc('\r', file);
675 if (has_trailing_newline)
676 fputc('\n', file);
677}
678static void emit_line(struct diff_options *o, const char *set, const char *reset,
680 const char *line, int len)
681{
682 emit_line_0(o, set, NULL, 0, reset, 0, line, len);
683}
684enum diff_symbol {
686 DIFF_SYMBOL_BINARY_DIFF_HEADER,
687 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
688 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
689 DIFF_SYMBOL_BINARY_DIFF_BODY,
690 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
691 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
692 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
693 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
694 DIFF_SYMBOL_STATS_LINE,
695 DIFF_SYMBOL_WORD_DIFF,
696 DIFF_SYMBOL_STAT_SEP,
697 DIFF_SYMBOL_SUMMARY,
698 DIFF_SYMBOL_SUBMODULE_ADD,
699 DIFF_SYMBOL_SUBMODULE_DEL,
700 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
701 DIFF_SYMBOL_SUBMODULE_MODIFIED,
702 DIFF_SYMBOL_SUBMODULE_HEADER,
703 DIFF_SYMBOL_SUBMODULE_ERROR,
704 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
705 DIFF_SYMBOL_REWRITE_DIFF,
706 DIFF_SYMBOL_BINARY_FILES,
707 DIFF_SYMBOL_HEADER,
708 DIFF_SYMBOL_FILEPAIR_PLUS,
709 DIFF_SYMBOL_FILEPAIR_MINUS,
710 DIFF_SYMBOL_WORDS_PORCELAIN,
711 DIFF_SYMBOL_WORDS,
712 DIFF_SYMBOL_CONTEXT,
713 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
714 DIFF_SYMBOL_PLUS,
715 DIFF_SYMBOL_MINUS,
716 DIFF_SYMBOL_NO_LF_EOF,
717 DIFF_SYMBOL_CONTEXT_FRAGINFO,
718 DIFF_SYMBOL_CONTEXT_MARKER,
719 DIFF_SYMBOL_SEPARATOR
720};
721/*
722 * Flags for content lines:
723 * 0..12 are whitespace rules
724 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
725 * 16 is marking if the line is blank at EOF
726 */
727#define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
728#define DIFF_SYMBOL_MOVED_LINE (1<<17)
729#define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
730#define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
731#define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
732/*
734 * This struct is used when we need to buffer the output of the diff output.
735 *
736 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
737 * into the pre/post image file. This pointer could be a union with the
738 * line pointer. By storing an offset into the file instead of the literal line,
739 * we can decrease the memory footprint for the buffered output. At first we
740 * may want to only have indirection for the content lines, but we could also
741 * enhance the state for emitting prefabricated lines, e.g. the similarity
742 * score line or hunk/file headers would only need to store a number or path
743 * and then the output can be constructed later on depending on state.
744 */
745struct emitted_diff_symbol {
746 const char *line;
747 int len;
748 int flags;
749 enum diff_symbol s;
750};
751#define EMITTED_DIFF_SYMBOL_INIT {NULL}
752struct emitted_diff_symbols {
754 struct emitted_diff_symbol *buf;
755 int nr, alloc;
756};
757#define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
758static void append_emitted_diff_symbol(struct diff_options *o,
760 struct emitted_diff_symbol *e)
761{
762 struct emitted_diff_symbol *f;
763 ALLOC_GROW(o->emitted_symbols->buf,
765 o->emitted_symbols->nr + 1,
766 o->emitted_symbols->alloc);
767 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
768 memcpy(f, e, sizeof(struct emitted_diff_symbol));
770 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
771}
772struct moved_entry {
774 struct hashmap_entry ent;
775 const struct emitted_diff_symbol *es;
776 struct moved_entry *next_line;
777 struct ws_delta *wsd;
778};
779/**
781 * The struct ws_delta holds white space differences between moved lines, i.e.
782 * between '+' and '-' lines that have been detected to be a move.
783 * The string contains the difference in leading white spaces, before the
784 * rest of the line is compared using the white space config for move
785 * coloring. The current_longer indicates if the first string in the
786 * comparision is longer than the second.
787 */
788struct ws_delta {
789 char *string;
790 unsigned int current_longer : 1;
791};
792#define WS_DELTA_INIT { NULL, 0 }
793static int compute_ws_delta(const struct emitted_diff_symbol *a,
795 const struct emitted_diff_symbol *b,
796 struct ws_delta *out)
797{
798 const struct emitted_diff_symbol *longer = a->len > b->len ? a : b;
799 const struct emitted_diff_symbol *shorter = a->len > b->len ? b : a;
800 int d = longer->len - shorter->len;
801 out->string = xmemdupz(longer->line, d);
803 out->current_longer = (a == longer);
804 return !strncmp(longer->line + d, shorter->line, shorter->len);
806}
807static int cmp_in_block_with_wsd(const struct diff_options *o,
809 const struct moved_entry *cur,
810 const struct moved_entry *match,
811 struct moved_entry *pmb,
812 int n)
813{
814 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
815 int al = cur->es->len, cl = l->len;
816 const char *a = cur->es->line,
817 *b = match->es->line,
818 *c = l->line;
819 int wslen;
821 /*
823 * We need to check if 'cur' is equal to 'match'.
824 * As those are from the same (+/-) side, we do not need to adjust for
825 * indent changes. However these were found using fuzzy matching
826 * so we do have to check if they are equal.
827 */
828 if (strcmp(a, b))
829 return 1;
830 if (!pmb->wsd)
832 /*
833 * No white space delta was carried forward? This can happen
834 * when we exit early in this function and do not carry
835 * forward ws.
836 */
837 return 1;
838 /*
840 * The indent changes of the block are known and carried forward in
841 * pmb->wsd; however we need to check if the indent changes of the
842 * current line are still the same as before.
843 *
844 * To do so we need to compare 'l' to 'cur', adjusting the
845 * one of them for the white spaces, depending which was longer.
846 */
847 wslen = strlen(pmb->wsd->string);
849 if (pmb->wsd->current_longer) {
850 c += wslen;
851 cl -= wslen;
852 } else {
853 a += wslen;
854 al -= wslen;
855 }
856 if (strcmp(a, c))
858 return 1;
859 return 0;
861}
862static int moved_entry_cmp(const void *hashmap_cmp_fn_data,
864 const void *entry,
865 const void *entry_or_key,
866 const void *keydata)
867{
868 const struct diff_options *diffopt = hashmap_cmp_fn_data;
869 const struct moved_entry *a = entry;
870 const struct moved_entry *b = entry_or_key;
871 unsigned flags = diffopt->color_moved_ws_handling
872 & XDF_WHITESPACE_FLAGS;
873 if (diffopt->color_moved_ws_handling &
875 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
876 /*
877 * As there is not specific white space config given,
878 * we'd need to check for a new block, so ignore all
879 * white space. The setup of the white space
880 * configuration for the next block is done else where
881 */
882 flags |= XDF_IGNORE_WHITESPACE;
883 return !xdiff_compare_lines(a->es->line, a->es->len,
885 b->es->line, b->es->len,
886 flags);
887}
888static struct moved_entry *prepare_entry(struct diff_options *o,
890 int line_no)
891{
892 struct moved_entry *ret = xmalloc(sizeof(*ret));
893 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
894 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
895 ret->ent.hash = xdiff_hash_string(l->line, l->len, flags);
897 ret->es = l;
898 ret->next_line = NULL;
899 ret->wsd = NULL;
900 return ret;
902}
903static void add_lines_to_move_detection(struct diff_options *o,
905 struct hashmap *add_lines,
906 struct hashmap *del_lines)
907{
908 struct moved_entry *prev_line = NULL;
909 int n;
911 for (n = 0; n < o->emitted_symbols->nr; n++) {
912 struct hashmap *hm;
913 struct moved_entry *key;
914 switch (o->emitted_symbols->buf[n].s) {
916 case DIFF_SYMBOL_PLUS:
917 hm = add_lines;
918 break;
919 case DIFF_SYMBOL_MINUS:
920 hm = del_lines;
921 break;
922 default:
923 prev_line = NULL;
924 continue;
925 }
926 key = prepare_entry(o, n);
928 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
929 prev_line->next_line = key;
930 hashmap_add(hm, key);
932 prev_line = key;
933 }
934}
935static void pmb_advance_or_null(struct diff_options *o,
937 struct moved_entry *match,
938 struct hashmap *hm,
939 struct moved_entry **pmb,
940 int pmb_nr)
941{
942 int i;
943 for (i = 0; i < pmb_nr; i++) {
944 struct moved_entry *prev = pmb[i];
945 struct moved_entry *cur = (prev && prev->next_line) ?
946 prev->next_line : NULL;
947 if (cur && !hm->cmpfn(o, cur, match, NULL)) {
948 pmb[i] = cur;
949 } else {
950 pmb[i] = NULL;
951 }
952 }
953}
954static void pmb_advance_or_null_multi_match(struct diff_options *o,
956 struct moved_entry *match,
957 struct hashmap *hm,
958 struct moved_entry **pmb,
959 int pmb_nr, int n)
960{
961 int i;
962 char *got_match = xcalloc(1, pmb_nr);
963 for (; match; match = hashmap_get_next(hm, match)) {
965 for (i = 0; i < pmb_nr; i++) {
966 struct moved_entry *prev = pmb[i];
967 struct moved_entry *cur = (prev && prev->next_line) ?
968 prev->next_line : NULL;
969 if (!cur)
970 continue;
971 if (!cmp_in_block_with_wsd(o, cur, match, pmb[i], n))
972 got_match[i] |= 1;
973 }
974 }
975 for (i = 0; i < pmb_nr; i++) {
977 if (got_match[i]) {
978 /* Carry the white space delta forward */
979 pmb[i]->next_line->wsd = pmb[i]->wsd;
980 pmb[i] = pmb[i]->next_line;
981 } else
982 pmb[i] = NULL;
983 }
984}
985static int shrink_potential_moved_blocks(struct moved_entry **pmb,
987 int pmb_nr)
988{
989 int lp, rp;
990 /* Shrink the set of potential block to the remaining running */
992 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
993 while (lp < pmb_nr && pmb[lp])
994 lp++;
995 /* lp points at the first NULL now */
996 while (rp > -1 && !pmb[rp])
998 rp--;
999 /* rp points at the last non-NULL */
1000 if (lp < pmb_nr && rp > -1 && lp < rp) {
1002 pmb[lp] = pmb[rp];
1003 if (pmb[rp]->wsd) {
1004 free(pmb[rp]->wsd->string);
1005 FREE_AND_NULL(pmb[rp]->wsd);
1006 }
1007 pmb[rp] = NULL;
1008 rp--;
1009 lp++;
1010 }
1011 }
1012 /* Remember the number of running sets */
1014 return rp + 1;
1015}
1016/*
1018 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1019 *
1020 * Otherwise, if the last block has fewer alphanumeric characters than
1021 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1022 * that block.
1023 *
1024 * The last block consists of the (n - block_length)'th line up to but not
1025 * including the nth line.
1026 *
1027 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1028 * Think of a way to unify them.
1029 */
1030static void adjust_last_block(struct diff_options *o, int n, int block_length)
1031{
1032 int i, alnum_count = 0;
1033 if (o->color_moved == COLOR_MOVED_PLAIN)
1034 return;
1035 for (i = 1; i < block_length + 1; i++) {
1036 const char *c = o->emitted_symbols->buf[n - i].line;
1037 for (; *c; c++) {
1038 if (!isalnum(*c))
1039 continue;
1040 alnum_count++;
1041 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1042 return;
1043 }
1044 }
1045 for (i = 1; i < block_length + 1; i++)
1046 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
1047}
1048/* Find blocks of moved code, delegate actual coloring decision to helper */
1050static void mark_color_as_moved(struct diff_options *o,
1051 struct hashmap *add_lines,
1052 struct hashmap *del_lines)
1053{
1054 struct moved_entry **pmb = NULL; /* potentially moved blocks */
1055 int pmb_nr = 0, pmb_alloc = 0;
1056 int n, flipped_block = 1, block_length = 0;
1057 for (n = 0; n < o->emitted_symbols->nr; n++) {
1060 struct hashmap *hm = NULL;
1061 struct moved_entry *key;
1062 struct moved_entry *match = NULL;
1063 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1064 switch (l->s) {
1066 case DIFF_SYMBOL_PLUS:
1067 hm = del_lines;
1068 key = prepare_entry(o, n);
1069 match = hashmap_get(hm, key, NULL);
1070 free(key);
1071 break;
1072 case DIFF_SYMBOL_MINUS:
1073 hm = add_lines;
1074 key = prepare_entry(o, n);
1075 match = hashmap_get(hm, key, NULL);
1076 free(key);
1077 break;
1078 default:
1079 flipped_block = 1;
1080 }
1081 if (!match) {
1083 adjust_last_block(o, n, block_length);
1084 pmb_nr = 0;
1085 block_length = 0;
1086 continue;
1087 }
1088 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1090 if (o->color_moved == COLOR_MOVED_PLAIN)
1092 continue;
1093 if (o->color_moved_ws_handling &
1095 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1096 pmb_advance_or_null_multi_match(o, match, hm, pmb, pmb_nr, n);
1097 else
1098 pmb_advance_or_null(o, match, hm, pmb, pmb_nr);
1099 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
1101 if (pmb_nr == 0) {
1103 /*
1104 * The current line is the start of a new block.
1105 * Setup the set of potential blocks.
1106 */
1107 for (; match; match = hashmap_get_next(hm, match)) {
1108 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1109 if (o->color_moved_ws_handling &
1110 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) {
1111 struct ws_delta *wsd = xmalloc(sizeof(*match->wsd));
1112 if (compute_ws_delta(l, match->es, wsd)) {
1113 match->wsd = wsd;
1114 pmb[pmb_nr++] = match;
1115 } else
1116 free(wsd);
1117 } else {
1118 pmb[pmb_nr++] = match;
1119 }
1120 }
1121 flipped_block = (flipped_block + 1) % 2;
1123 adjust_last_block(o, n, block_length);
1125 block_length = 0;
1126 }
1127 block_length++;
1129 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1131 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1132 }
1133 adjust_last_block(o, n, block_length);
1134 free(pmb);
1136}
1137#define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1139 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1140static void dim_moved_lines(struct diff_options *o)
1141{
1142 int n;
1143 for (n = 0; n < o->emitted_symbols->nr; n++) {
1144 struct emitted_diff_symbol *prev = (n != 0) ?
1145 &o->emitted_symbols->buf[n - 1] : NULL;
1146 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1147 struct emitted_diff_symbol *next =
1148 (n < o->emitted_symbols->nr - 1) ?
1149 &o->emitted_symbols->buf[n + 1] : NULL;
1150 /* Not a plus or minus line? */
1152 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1153 continue;
1154 /* Not a moved line? */
1156 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1157 continue;
1158 /*
1160 * If prev or next are not a plus or minus line,
1161 * pretend they don't exist
1162 */
1163 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1164 prev->s != DIFF_SYMBOL_MINUS)
1165 prev = NULL;
1166 if (next && next->s != DIFF_SYMBOL_PLUS &&
1167 next->s != DIFF_SYMBOL_MINUS)
1168 next = NULL;
1169 /* Inside a block? */
1171 if ((prev &&
1172 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1173 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1174 (next &&
1175 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1176 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1177 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1178 continue;
1179 }
1180 /* Check if we are at an interesting bound: */
1182 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1183 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1184 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1185 continue;
1186 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1187 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1188 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1189 continue;
1190 /*
1192 * The boundary to prev and next are not interesting,
1193 * so this line is not interesting as a whole
1194 */
1195 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1196 }
1197}
1198static void emit_line_ws_markup(struct diff_options *o,
1200 const char *set_sign, const char *set,
1201 const char *reset,
1202 char sign, const char *line, int len,
1203 unsigned ws_rule, int blank_at_eof)
1204{
1205 const char *ws = NULL;
1206 if (o->ws_error_highlight & ws_rule) {
1208 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1209 if (!*ws)
1210 ws = NULL;
1211 }
1212 if (!ws && !set_sign)
1214 emit_line_0(o, set, NULL, 0, reset, sign, line, len);
1215 else if (!ws) {
1216 emit_line_0(o, set_sign, set, !!set_sign, reset, sign, line, len);
1217 } else if (blank_at_eof)
1218 /* Blank line at EOF - paint '+' as well */
1219 emit_line_0(o, ws, NULL, 0, reset, sign, line, len);
1220 else {
1221 /* Emit just the prefix, then the rest. */
1222 emit_line_0(o, set_sign ? set_sign : set, NULL, !!set_sign, reset,
1223 sign, "", 0);
1224 ws_check_emit(line, len, ws_rule,
1225 o->file, set, reset, ws);
1226 }
1227}
1228static void emit_diff_symbol_from_struct(struct diff_options *o,
1230 struct emitted_diff_symbol *eds)
1231{
1232 static const char *nneof = " No newline at end of file\n";
1233 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1234 struct strbuf sb = STRBUF_INIT;
1235 enum diff_symbol s = eds->s;
1237 const char *line = eds->line;
1238 int len = eds->len;
1239 unsigned flags = eds->flags;
1240 switch (s) {
1242 case DIFF_SYMBOL_NO_LF_EOF:
1243 context = diff_get_color_opt(o, DIFF_CONTEXT);
1244 reset = diff_get_color_opt(o, DIFF_RESET);
1245 putc('\n', o->file);
1246 emit_line_0(o, context, NULL, 0, reset, '\\',
1247 nneof, strlen(nneof));
1248 break;
1249 case DIFF_SYMBOL_SUBMODULE_HEADER:
1250 case DIFF_SYMBOL_SUBMODULE_ERROR:
1251 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1252 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1253 case DIFF_SYMBOL_SUMMARY:
1254 case DIFF_SYMBOL_STATS_LINE:
1255 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1256 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1257 emit_line(o, "", "", line, len);
1258 break;
1259 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1260 case DIFF_SYMBOL_CONTEXT_MARKER:
1261 context = diff_get_color_opt(o, DIFF_CONTEXT);
1262 reset = diff_get_color_opt(o, DIFF_RESET);
1263 emit_line(o, context, reset, line, len);
1264 break;
1265 case DIFF_SYMBOL_SEPARATOR:
1266 fprintf(o->file, "%s%c",
1267 diff_line_prefix(o),
1268 o->line_termination);
1269 break;
1270 case DIFF_SYMBOL_CONTEXT:
1271 set = diff_get_color_opt(o, DIFF_CONTEXT);
1272 reset = diff_get_color_opt(o, DIFF_RESET);
1273 set_sign = NULL;
1274 if (o->flags.dual_color_diffed_diffs) {
1275 char c = !len ? 0 : line[0];
1276 if (c == '+')
1278 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1279 else if (c == '@')
1280 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1281 else if (c == '-')
1282 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1283 }
1284 emit_line_ws_markup(o, set_sign, set, reset, ' ', line, len,
1285 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1286 break;
1287 case DIFF_SYMBOL_PLUS:
1288 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1289 DIFF_SYMBOL_MOVED_LINE_ALT |
1290 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1291 case DIFF_SYMBOL_MOVED_LINE |
1292 DIFF_SYMBOL_MOVED_LINE_ALT |
1293 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1294 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1295 break;
1296 case DIFF_SYMBOL_MOVED_LINE |
1297 DIFF_SYMBOL_MOVED_LINE_ALT:
1298 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1299 break;
1300 case DIFF_SYMBOL_MOVED_LINE |
1301 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1302 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1303 break;
1304 case DIFF_SYMBOL_MOVED_LINE:
1305 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1306 break;
1307 default:
1308 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1309 }
1310 reset = diff_get_color_opt(o, DIFF_RESET);
1311 if (!o->flags.dual_color_diffed_diffs)
1312 set_sign = NULL;
1313 else {
1314 char c = !len ? 0 : line[0];
1315 set_sign = set;
1317 if (c == '-')
1318 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1319 else if (c == '@')
1320 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1321 else if (c == '+')
1322 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1323 else
1324 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1325 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1326 }
1327 emit_line_ws_markup(o, set_sign, set, reset, '+', line, len,
1328 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1329 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1330 break;
1331 case DIFF_SYMBOL_MINUS:
1332 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1333 DIFF_SYMBOL_MOVED_LINE_ALT |
1334 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1335 case DIFF_SYMBOL_MOVED_LINE |
1336 DIFF_SYMBOL_MOVED_LINE_ALT |
1337 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1338 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1339 break;
1340 case DIFF_SYMBOL_MOVED_LINE |
1341 DIFF_SYMBOL_MOVED_LINE_ALT:
1342 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1343 break;
1344 case DIFF_SYMBOL_MOVED_LINE |
1345 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1346 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1347 break;
1348 case DIFF_SYMBOL_MOVED_LINE:
1349 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1350 break;
1351 default:
1352 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1353 }
1354 reset = diff_get_color_opt(o, DIFF_RESET);
1355 if (!o->flags.dual_color_diffed_diffs)
1356 set_sign = NULL;
1357 else {
1358 char c = !len ? 0 : line[0];
1359 set_sign = set;
1361 if (c == '+')
1362 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1363 else if (c == '@')
1364 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1365 else if (c == '-')
1366 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1367 else
1368 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1369 }
1370 emit_line_ws_markup(o, set_sign, set, reset, '-', line, len,
1371 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1372 break;
1373 case DIFF_SYMBOL_WORDS_PORCELAIN:
1374 context = diff_get_color_opt(o, DIFF_CONTEXT);
1375 reset = diff_get_color_opt(o, DIFF_RESET);
1376 emit_line(o, context, reset, line, len);
1377 fputs("~\n", o->file);
1378 break;
1379 case DIFF_SYMBOL_WORDS:
1380 context = diff_get_color_opt(o, DIFF_CONTEXT);
1381 reset = diff_get_color_opt(o, DIFF_RESET);
1382 /*
1383 * Skip the prefix character, if any. With
1384 * diff_suppress_blank_empty, there may be
1385 * none.
1386 */
1387 if (line[0] != '\n') {
1388 line++;
1389 len--;
1390 }
1391 emit_line(o, context, reset, line, len);
1392 break;
1393 case DIFF_SYMBOL_FILEPAIR_PLUS:
1394 meta = diff_get_color_opt(o, DIFF_METAINFO);
1395 reset = diff_get_color_opt(o, DIFF_RESET);
1396 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1397 line, reset,
1398 strchr(line, ' ') ? "\t" : "");
1399 break;
1400 case DIFF_SYMBOL_FILEPAIR_MINUS:
1401 meta = diff_get_color_opt(o, DIFF_METAINFO);
1402 reset = diff_get_color_opt(o, DIFF_RESET);
1403 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1404 line, reset,
1405 strchr(line, ' ') ? "\t" : "");
1406 break;
1407 case DIFF_SYMBOL_BINARY_FILES:
1408 case DIFF_SYMBOL_HEADER:
1409 fprintf(o->file, "%s", line);
1410 break;
1411 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1412 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1413 break;
1414 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1415 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1416 break;
1417 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1418 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1419 break;
1420 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1421 fputs(diff_line_prefix(o), o->file);
1422 fputc('\n', o->file);
1423 break;
1424 case DIFF_SYMBOL_REWRITE_DIFF:
1425 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1426 reset = diff_get_color_opt(o, DIFF_RESET);
1427 emit_line(o, fraginfo, reset, line, len);
1428 break;
1429 case DIFF_SYMBOL_SUBMODULE_ADD:
1430 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1431 reset = diff_get_color_opt(o, DIFF_RESET);
1432 emit_line(o, set, reset, line, len);
1433 break;
1434 case DIFF_SYMBOL_SUBMODULE_DEL:
1435 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1436 reset = diff_get_color_opt(o, DIFF_RESET);
1437 emit_line(o, set, reset, line, len);
1438 break;
1439 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1440 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1441 diff_line_prefix(o), line);
1442 break;
1443 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1444 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1445 diff_line_prefix(o), line);
1446 break;
1447 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1448 emit_line(o, "", "", " 0 files changed\n",
1449 strlen(" 0 files changed\n"));
1450 break;
1451 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1452 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1453 break;
1454 case DIFF_SYMBOL_WORD_DIFF:
1455 fprintf(o->file, "%.*s", len, line);
1456 break;
1457 case DIFF_SYMBOL_STAT_SEP:
1458 fputs(o->stat_sep, o->file);
1459 break;
1460 default:
1461 BUG("unknown diff symbol");
1462 }
1463 strbuf_release(&sb);
1464}
1465static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1467 const char *line, int len, unsigned flags)
1468{
1469 struct emitted_diff_symbol e = {line, len, flags, s};
1470 if (o->emitted_symbols)
1472 append_emitted_diff_symbol(o, &e);
1473 else
1474 emit_diff_symbol_from_struct(o, &e);
1475}
1476void diff_emit_submodule_del(struct diff_options *o, const char *line)
1478{
1479 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1480}
1481void diff_emit_submodule_add(struct diff_options *o, const char *line)
1483{
1484 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1485}
1486void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1488{
1489 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1490 path, strlen(path), 0);
1491}
1492void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1494{
1495 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1496 path, strlen(path), 0);
1497}
1498void diff_emit_submodule_header(struct diff_options *o, const char *header)
1500{
1501 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1502 header, strlen(header), 0);
1503}
1504void diff_emit_submodule_error(struct diff_options *o, const char *err)
1506{
1507 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1508}
1509void diff_emit_submodule_pipethrough(struct diff_options *o,
1511 const char *line, int len)
1512{
1513 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1514}
1515static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1517{
1518 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1519 ecbdata->blank_at_eof_in_preimage &&
1520 ecbdata->blank_at_eof_in_postimage &&
1521 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1522 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1523 return 0;
1524 return ws_blank_line(line, len, ecbdata->ws_rule);
1525}
1526static void emit_add_line(const char *reset,
1528 struct emit_callback *ecbdata,
1529 const char *line, int len)
1530{
1531 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1532 if (new_blank_line_at_eof(ecbdata, line, len))
1533 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1534 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1536}
1537static void emit_del_line(const char *reset,
1539 struct emit_callback *ecbdata,
1540 const char *line, int len)
1541{
1542 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1543 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1544}
1545static void emit_context_line(const char *reset,
1547 struct emit_callback *ecbdata,
1548 const char *line, int len)
1549{
1550 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1551 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1552}
1553static void emit_hunk_header(struct emit_callback *ecbdata,
1555 const char *line, int len)
1556{
1557 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1558 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1559 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1560 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1561 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1562 static const char atat[2] = { '@', '@' };
1563 const char *cp, *ep;
1564 struct strbuf msgbuf = STRBUF_INIT;
1565 int org_len = len;
1566 int i = 1;
1567 /*
1569 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1570 * it always is at least 10 bytes long.
1571 */
1572 if (len < 10 ||
1573 memcmp(line, atat, 2) ||
1574 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1575 emit_diff_symbol(ecbdata->opt,
1576 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1577 return;
1578 }
1579 ep += 2; /* skip over @@ */
1580 /* The hunk header in fraginfo color */
1582 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1583 strbuf_addstr(&msgbuf, reverse);
1584 strbuf_addstr(&msgbuf, frag);
1585 strbuf_add(&msgbuf, line, ep - line);
1586 strbuf_addstr(&msgbuf, reset);
1587 /*
1589 * trailing "\r\n"
1590 */
1591 for ( ; i < 3; i++)
1592 if (line[len - i] == '\r' || line[len - i] == '\n')
1593 len--;
1594 /* blank before the func header */
1596 for (cp = ep; ep - line < len; ep++)
1597 if (*ep != ' ' && *ep != '\t')
1598 break;
1599 if (ep != cp) {
1600 strbuf_addstr(&msgbuf, context);
1601 strbuf_add(&msgbuf, cp, ep - cp);
1602 strbuf_addstr(&msgbuf, reset);
1603 }
1604 if (ep < line + len) {
1606 strbuf_addstr(&msgbuf, func);
1607 strbuf_add(&msgbuf, ep, line + len - ep);
1608 strbuf_addstr(&msgbuf, reset);
1609 }
1610 strbuf_add(&msgbuf, line + len, org_len - len);
1612 strbuf_complete_line(&msgbuf);
1613 emit_diff_symbol(ecbdata->opt,
1614 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1615 strbuf_release(&msgbuf);
1616}
1617static struct diff_tempfile *claim_diff_tempfile(void) {
1619 int i;
1620 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1621 if (!diff_temp[i].name)
1622 return diff_temp + i;
1623 BUG("diff is failing to clean up its tempfiles");
1624}
1625static void remove_tempfile(void)
1627{
1628 int i;
1629 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1630 if (is_tempfile_active(diff_temp[i].tempfile))
1631 delete_tempfile(&diff_temp[i].tempfile);
1632 diff_temp[i].name = NULL;
1633 }
1634}
1635static void add_line_count(struct strbuf *out, int count)
1637{
1638 switch (count) {
1639 case 0:
1640 strbuf_addstr(out, "0,0");
1641 break;
1642 case 1:
1643 strbuf_addstr(out, "1");
1644 break;
1645 default:
1646 strbuf_addf(out, "1,%d", count);
1647 break;
1648 }
1649}
1650static void emit_rewrite_lines(struct emit_callback *ecb,
1652 int prefix, const char *data, int size)
1653{
1654 const char *endp = NULL;
1655 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1656 while (0 < size) {
1658 int len;
1659 endp = memchr(data, '\n', size);
1661 len = endp ? (endp - data + 1) : size;
1662 if (prefix != '+') {
1663 ecb->lno_in_preimage++;
1664 emit_del_line(reset, ecb, data, len);
1665 } else {
1666 ecb->lno_in_postimage++;
1667 emit_add_line(reset, ecb, data, len);
1668 }
1669 size -= len;
1670 data += len;
1671 }
1672 if (!endp)
1673 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1674}
1675static void emit_rewrite_diff(const char *name_a,
1677 const char *name_b,
1678 struct diff_filespec *one,
1679 struct diff_filespec *two,
1680 struct userdiff_driver *textconv_one,
1681 struct userdiff_driver *textconv_two,
1682 struct diff_options *o)
1683{
1684 int lc_a, lc_b;
1685 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1686 const char *a_prefix, *b_prefix;
1687 char *data_one, *data_two;
1688 size_t size_one, size_two;
1689 struct emit_callback ecbdata;
1690 struct strbuf out = STRBUF_INIT;
1691 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1693 a_prefix = o->b_prefix;
1694 b_prefix = o->a_prefix;
1695 } else {
1696 a_prefix = o->a_prefix;
1697 b_prefix = o->b_prefix;
1698 }
1699 name_a += (*name_a == '/');
1701 name_b += (*name_b == '/');
1702 strbuf_reset(&a_name);
1704 strbuf_reset(&b_name);
1705 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1706 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1707 size_one = fill_textconv(textconv_one, one, &data_one);
1709 size_two = fill_textconv(textconv_two, two, &data_two);
1710 memset(&ecbdata, 0, sizeof(ecbdata));
1712 ecbdata.color_diff = want_color(o->use_color);
1713 ecbdata.ws_rule = whitespace_rule(name_b);
1714 ecbdata.opt = o;
1715 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1716 mmfile_t mf1, mf2;
1717 mf1.ptr = (char *)data_one;
1718 mf2.ptr = (char *)data_two;
1719 mf1.size = size_one;
1720 mf2.size = size_two;
1721 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1722 }
1723 ecbdata.lno_in_preimage = 1;
1724 ecbdata.lno_in_postimage = 1;
1725 lc_a = count_lines(data_one, size_one);
1727 lc_b = count_lines(data_two, size_two);
1728 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1730 a_name.buf, a_name.len, 0);
1731 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1732 b_name.buf, b_name.len, 0);
1733 strbuf_addstr(&out, "@@ -");
1735 if (!o->irreversible_delete)
1736 add_line_count(&out, lc_a);
1737 else
1738 strbuf_addstr(&out, "?,?");
1739 strbuf_addstr(&out, " +");
1740 add_line_count(&out, lc_b);
1741 strbuf_addstr(&out, " @@\n");
1742 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1743 strbuf_release(&out);
1744 if (lc_a && !o->irreversible_delete)
1746 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1747 if (lc_b)
1748 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1749 if (textconv_one)
1750 free((char *)data_one);
1751 if (textconv_two)
1752 free((char *)data_two);
1753}
1754struct diff_words_buffer {
1756 mmfile_t text;
1757 unsigned long alloc;
1758 struct diff_words_orig {
1759 const char *begin, *end;
1760 } *orig;
1761 int orig_nr, orig_alloc;
1762};
1763static void diff_words_append(char *line, unsigned long len,
1765 struct diff_words_buffer *buffer)
1766{
1767 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1768 line++;
1769 len--;
1770 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1771 buffer->text.size += len;
1772 buffer->text.ptr[buffer->text.size] = '\0';
1773}
1774struct diff_words_style_elem {
1776 const char *prefix;
1777 const char *suffix;
1778 const char *color; /* NULL; filled in by the setup code if
1779 * color is enabled */
1780};
1781struct diff_words_style {
1783 enum diff_words_type type;
1784 struct diff_words_style_elem new_word, old_word, ctx;
1785 const char *newline;
1786};
1787static struct diff_words_style diff_words_styles[] = {
1789 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1790 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1791 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1792};
1793struct diff_words_data {
1795 struct diff_words_buffer minus, plus;
1796 const char *current_plus;
1797 int last_minus;
1798 struct diff_options *opt;
1799 regex_t *word_regex;
1800 enum diff_words_type type;
1801 struct diff_words_style *style;
1802};
1803static int fn_out_diff_words_write_helper(struct diff_options *o,
1805 struct diff_words_style_elem *st_el,
1806 const char *newline,
1807 size_t count, const char *buf)
1808{
1809 int print = 0;
1810 struct strbuf sb = STRBUF_INIT;
1811 while (count) {
1813 char *p = memchr(buf, '\n', count);
1814 if (print)
1815 strbuf_addstr(&sb, diff_line_prefix(o));
1816 if (p != buf) {
1818 const char *reset = st_el->color && *st_el->color ?
1819 GIT_COLOR_RESET : NULL;
1820 if (st_el->color && *st_el->color)
1821 strbuf_addstr(&sb, st_el->color);
1822 strbuf_addstr(&sb, st_el->prefix);
1823 strbuf_add(&sb, buf, p ? p - buf : count);
1824 strbuf_addstr(&sb, st_el->suffix);
1825 if (reset)
1826 strbuf_addstr(&sb, reset);
1827 }
1828 if (!p)
1829 goto out;
1830 strbuf_addstr(&sb, newline);
1832 count -= p + 1 - buf;
1833 buf = p + 1;
1834 print = 1;
1835 if (count) {
1836 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1837 sb.buf, sb.len, 0);
1838 strbuf_reset(&sb);
1839 }
1840 }
1841out:
1843 if (sb.len)
1844 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1845 sb.buf, sb.len, 0);
1846 strbuf_release(&sb);
1847 return 0;
1848}
1849/*
1851 * '--color-words' algorithm can be described as:
1852 *
1853 * 1. collect the minus/plus lines of a diff hunk, divided into
1854 * minus-lines and plus-lines;
1855 *
1856 * 2. break both minus-lines and plus-lines into words and
1857 * place them into two mmfile_t with one word for each line;
1858 *
1859 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1860 *
1861 * And for the common parts of the both file, we output the plus side text.
1862 * diff_words->current_plus is used to trace the current position of the plus file
1863 * which printed. diff_words->last_minus is used to trace the last minus word
1864 * printed.
1865 *
1866 * For '--graph' to work with '--color-words', we need to output the graph prefix
1867 * on each line of color words output. Generally, there are two conditions on
1868 * which we should output the prefix.
1869 *
1870 * 1. diff_words->last_minus == 0 &&
1871 * diff_words->current_plus == diff_words->plus.text.ptr
1872 *
1873 * that is: the plus text must start as a new line, and if there is no minus
1874 * word printed, a graph prefix must be printed.
1875 *
1876 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1877 * *(diff_words->current_plus - 1) == '\n'
1878 *
1879 * that is: a graph prefix must be printed following a '\n'
1880 */
1881static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1882{
1883 if ((diff_words->last_minus == 0 &&
1884 diff_words->current_plus == diff_words->plus.text.ptr) ||
1885 (diff_words->current_plus > diff_words->plus.text.ptr &&
1886 *(diff_words->current_plus - 1) == '\n')) {
1887 return 1;
1888 } else {
1889 return 0;
1890 }
1891}
1892static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1894{
1895 struct diff_words_data *diff_words = priv;
1896 struct diff_words_style *style = diff_words->style;
1897 int minus_first, minus_len, plus_first, plus_len;
1898 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1899 struct diff_options *opt = diff_words->opt;
1900 const char *line_prefix;
1901 if (line[0] != '@' || parse_hunk_header(line, len,
1903 &minus_first, &minus_len, &plus_first, &plus_len))
1904 return;
1905 assert(opt);
1907 line_prefix = diff_line_prefix(opt);
1908 /* POSIX requires that first be decremented by one if len == 0... */
1910 if (minus_len) {
1911 minus_begin = diff_words->minus.orig[minus_first].begin;
1912 minus_end =
1913 diff_words->minus.orig[minus_first + minus_len - 1].end;
1914 } else
1915 minus_begin = minus_end =
1916 diff_words->minus.orig[minus_first].end;
1917 if (plus_len) {
1919 plus_begin = diff_words->plus.orig[plus_first].begin;
1920 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1921 } else
1922 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1923 if (color_words_output_graph_prefix(diff_words)) {
1925 fputs(line_prefix, diff_words->opt->file);
1926 }
1927 if (diff_words->current_plus != plus_begin) {
1928 fn_out_diff_words_write_helper(diff_words->opt,
1929 &style->ctx, style->newline,
1930 plus_begin - diff_words->current_plus,
1931 diff_words->current_plus);
1932 }
1933 if (minus_begin != minus_end) {
1934 fn_out_diff_words_write_helper(diff_words->opt,
1935 &style->old_word, style->newline,
1936 minus_end - minus_begin, minus_begin);
1937 }
1938 if (plus_begin != plus_end) {
1939 fn_out_diff_words_write_helper(diff_words->opt,
1940 &style->new_word, style->newline,
1941 plus_end - plus_begin, plus_begin);
1942 }
1943 diff_words->current_plus = plus_end;
1945 diff_words->last_minus = minus_first;
1946}
1947/* This function starts looking at *begin, and returns 0 iff a word was found. */
1949static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1950 int *begin, int *end)
1951{
1952 if (word_regex && *begin < buffer->size) {
1953 regmatch_t match[1];
1954 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1955 buffer->size - *begin, 1, match, 0)) {
1956 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1957 '\n', match[0].rm_eo - match[0].rm_so);
1958 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1959 *begin += match[0].rm_so;
1960 return *begin >= *end;
1961 }
1962 return -1;
1963 }
1964 /* find the next word */
1966 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1967 (*begin)++;
1968 if (*begin >= buffer->size)
1969 return -1;
1970 /* find the end of the word */
1972 *end = *begin + 1;
1973 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1974 (*end)++;
1975 return 0;
1977}
1978/*
1980 * This function splits the words in buffer->text, stores the list with
1981 * newline separator into out, and saves the offsets of the original words
1982 * in buffer->orig.
1983 */
1984static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1985 regex_t *word_regex)
1986{
1987 int i, j;
1988 long alloc = 0;
1989 out->size = 0;
1991 out->ptr = NULL;
1992 /* fake an empty "0th" word */
1994 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1995 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1996 buffer->orig_nr = 1;
1997 for (i = 0; i < buffer->text.size; i++) {
1999 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
2000 return;
2001 /* store original boundaries */
2003 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
2004 buffer->orig_alloc);
2005 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
2006 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
2007 buffer->orig_nr++;
2008 /* store one word */
2010 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2011 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2012 out->ptr[out->size + j - i] = '\n';
2013 out->size += j - i + 1;
2014 i = j - 1;
2016 }
2017}
2018/* this executes the word diff on the accumulated buffers */
2020static void diff_words_show(struct diff_words_data *diff_words)
2021{
2022 xpparam_t xpp;
2023 xdemitconf_t xecfg;
2024 mmfile_t minus, plus;
2025 struct diff_words_style *style = diff_words->style;
2026 struct diff_options *opt = diff_words->opt;
2028 const char *line_prefix;
2029 assert(opt);
2031 line_prefix = diff_line_prefix(opt);
2032 /* special case: only removal */
2034 if (!diff_words->plus.text.size) {
2035 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2036 line_prefix, strlen(line_prefix), 0);
2037 fn_out_diff_words_write_helper(diff_words->opt,
2038 &style->old_word, style->newline,
2039 diff_words->minus.text.size,
2040 diff_words->minus.text.ptr);
2041 diff_words->minus.text.size = 0;
2042 return;
2043 }
2044 diff_words->current_plus = diff_words->plus.text.ptr;
2046 diff_words->last_minus = 0;
2047 memset(&xpp, 0, sizeof(xpp));
2049 memset(&xecfg, 0, sizeof(xecfg));
2050 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2051 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2052 xpp.flags = 0;
2053 /* as only the hunk header will be parsed, we need a 0-context */
2054 xecfg.ctxlen = 0;
2055 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
2056 &xpp, &xecfg))
2057 die("unable to generate word diff");
2058 free(minus.ptr);
2059 free(plus.ptr);
2060 if (diff_words->current_plus != diff_words->plus.text.ptr +
2061 diff_words->plus.text.size) {
2062 if (color_words_output_graph_prefix(diff_words))
2063 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2064 line_prefix, strlen(line_prefix), 0);
2065 fn_out_diff_words_write_helper(diff_words->opt,
2066 &style->ctx, style->newline,
2067 diff_words->plus.text.ptr + diff_words->plus.text.size
2068 - diff_words->current_plus, diff_words->current_plus);
2069 }
2070 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2071}
2072/* In "color-words" mode, show word-diff of words accumulated in the buffer */
2074static void diff_words_flush(struct emit_callback *ecbdata)
2075{
2076 struct diff_options *wo = ecbdata->diff_words->opt;
2077 if (ecbdata->diff_words->minus.text.size ||
2079 ecbdata->diff_words->plus.text.size)
2080 diff_words_show(ecbdata->diff_words);
2081 if (wo->emitted_symbols) {
2083 struct diff_options *o = ecbdata->opt;
2084 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2085 int i;
2086 /*
2088 * NEEDSWORK:
2089 * Instead of appending each, concat all words to a line?
2090 */
2091 for (i = 0; i < wol->nr; i++)
2092 append_emitted_diff_symbol(o, &wol->buf[i]);
2093 for (i = 0; i < wol->nr; i++)
2095 free((void *)wol->buf[i].line);
2096 wol->nr = 0;
2098 }
2099}
2100static void diff_filespec_load_driver(struct diff_filespec *one)
2102{
2103 /* Use already-loaded driver */
2104 if (one->driver)
2105 return;
2106 if (S_ISREG(one->mode))
2108 one->driver = userdiff_find_by_path(one->path);
2109 /* Fallback to default settings */
2111 if (!one->driver)
2112 one->driver = userdiff_find_by_name("default");
2113}
2114static const char *userdiff_word_regex(struct diff_filespec *one)
2116{
2117 diff_filespec_load_driver(one);
2118 return one->driver->word_regex;
2119}
2120static void init_diff_words_data(struct emit_callback *ecbdata,
2122 struct diff_options *orig_opts,
2123 struct diff_filespec *one,
2124 struct diff_filespec *two)
2125{
2126 int i;
2127 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2128 memcpy(o, orig_opts, sizeof(struct diff_options));
2129 ecbdata->diff_words =
2131 xcalloc(1, sizeof(struct diff_words_data));
2132 ecbdata->diff_words->type = o->word_diff;
2133 ecbdata->diff_words->opt = o;
2134 if (orig_opts->emitted_symbols)
2136 o->emitted_symbols =
2137 xcalloc(1, sizeof(struct emitted_diff_symbols));
2138 if (!o->word_regex)
2140 o->word_regex = userdiff_word_regex(one);
2141 if (!o->word_regex)
2142 o->word_regex = userdiff_word_regex(two);
2143 if (!o->word_regex)
2144 o->word_regex = diff_word_regex_cfg;
2145 if (o->word_regex) {
2146 ecbdata->diff_words->word_regex = (regex_t *)
2147 xmalloc(sizeof(regex_t));
2148 if (regcomp(ecbdata->diff_words->word_regex,
2149 o->word_regex,
2150 REG_EXTENDED | REG_NEWLINE))
2151 die ("Invalid regular expression: %s",
2152 o->word_regex);
2153 }
2154 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2155 if (o->word_diff == diff_words_styles[i].type) {
2156 ecbdata->diff_words->style =
2157 &diff_words_styles[i];
2158 break;
2159 }
2160 }
2161 if (want_color(o->use_color)) {
2162 struct diff_words_style *st = ecbdata->diff_words->style;
2163 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2164 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2165 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2166 }
2167}
2168static void free_diff_words_data(struct emit_callback *ecbdata)
2170{
2171 if (ecbdata->diff_words) {
2172 diff_words_flush(ecbdata);
2173 free (ecbdata->diff_words->opt->emitted_symbols);
2174 free (ecbdata->diff_words->opt);
2175 free (ecbdata->diff_words->minus.text.ptr);
2176 free (ecbdata->diff_words->minus.orig);
2177 free (ecbdata->diff_words->plus.text.ptr);
2178 free (ecbdata->diff_words->plus.orig);
2179 if (ecbdata->diff_words->word_regex) {
2180 regfree(ecbdata->diff_words->word_regex);
2181 free(ecbdata->diff_words->word_regex);
2182 }
2183 FREE_AND_NULL(ecbdata->diff_words);
2184 }
2185}
2186const char *diff_get_color(int diff_use_color, enum color_diff ix)
2188{
2189 if (want_color(diff_use_color))
2190 return diff_colors[ix];
2191 return "";
2192}
2193const char *diff_line_prefix(struct diff_options *opt)
2195{
2196 struct strbuf *msgbuf;
2197 if (!opt->output_prefix)
2198 return "";
2199 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2201 return msgbuf->buf;
2202}
2203static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
2205{
2206 const char *cp;
2207 unsigned long allot;
2208 size_t l = len;
2209 cp = line;
2211 allot = l;
2212 while (0 < l) {
2213 (void) utf8_width(&cp, &l);
2214 if (!cp)
2215 break; /* truncated in the middle? */
2216 }
2217 return allot - l;
2218}
2219static void find_lno(const char *line, struct emit_callback *ecbdata)
2221{
2222 const char *p;
2223 ecbdata->lno_in_preimage = 0;
2224 ecbdata->lno_in_postimage = 0;
2225 p = strchr(line, '-');
2226 if (!p)
2227 return; /* cannot happen */
2228 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2229 p = strchr(p, '+');
2230 if (!p)
2231 return; /* cannot happen */
2232 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2233}
2234static void fn_out_consume(void *priv, char *line, unsigned long len)
2236{
2237 struct emit_callback *ecbdata = priv;
2238 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2239 struct diff_options *o = ecbdata->opt;
2240 o->found_changes = 1;
2242 if (ecbdata->header) {
2244 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2245 ecbdata->header->buf, ecbdata->header->len, 0);
2246 strbuf_reset(ecbdata->header);
2247 ecbdata->header = NULL;
2248 }
2249 if (ecbdata->label_path[0]) {
2251 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2252 ecbdata->label_path[0],
2253 strlen(ecbdata->label_path[0]), 0);
2254 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2255 ecbdata->label_path[1],
2256 strlen(ecbdata->label_path[1]), 0);
2257 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2258 }
2259 if (diff_suppress_blank_empty
2261 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2262 line[0] = '\n';
2263 len = 1;
2264 }
2265 if (line[0] == '@') {
2267 if (ecbdata->diff_words)
2268 diff_words_flush(ecbdata);
2269 len = sane_truncate_line(ecbdata, line, len);
2270 find_lno(line, ecbdata);
2271 emit_hunk_header(ecbdata, line, len);
2272 return;
2273 }
2274 if (ecbdata->diff_words) {
2276 enum diff_symbol s =
2277 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2278 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2279 if (line[0] == '-') {
2280 diff_words_append(line, len,
2281 &ecbdata->diff_words->minus);
2282 return;
2283 } else if (line[0] == '+') {
2284 diff_words_append(line, len,
2285 &ecbdata->diff_words->plus);
2286 return;
2287 } else if (starts_with(line, "\\ ")) {
2288 /*
2289 * Eat the "no newline at eof" marker as if we
2290 * saw a "+" or "-" line with nothing on it,
2291 * and return without diff_words_flush() to
2292 * defer processing. If this is the end of
2293 * preimage, more "+" lines may come after it.
2294 */
2295 return;
2296 }
2297 diff_words_flush(ecbdata);
2298 emit_diff_symbol(o, s, line, len, 0);
2299 return;
2300 }
2301 switch (line[0]) {
2303 case '+':
2304 ecbdata->lno_in_postimage++;
2305 emit_add_line(reset, ecbdata, line + 1, len - 1);
2306 break;
2307 case '-':
2308 ecbdata->lno_in_preimage++;
2309 emit_del_line(reset, ecbdata, line + 1, len - 1);
2310 break;
2311 case ' ':
2312 ecbdata->lno_in_postimage++;
2313 ecbdata->lno_in_preimage++;
2314 emit_context_line(reset, ecbdata, line + 1, len - 1);
2315 break;
2316 default:
2317 /* incomplete line at the end */
2318 ecbdata->lno_in_preimage++;
2319 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2320 line, len, 0);
2321 break;
2322 }
2323}
2324static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2326{
2327 const char *old_name = a;
2328 const char *new_name = b;
2329 int pfx_length, sfx_length;
2330 int pfx_adjust_for_slash;
2331 int len_a = strlen(a);
2332 int len_b = strlen(b);
2333 int a_midlen, b_midlen;
2334 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2335 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2336 if (qlen_a || qlen_b) {
2338 quote_c_style(a, name, NULL, 0);
2339 strbuf_addstr(name, " => ");
2340 quote_c_style(b, name, NULL, 0);
2341 return;
2342 }
2343 /* Find common prefix */
2345 pfx_length = 0;
2346 while (*old_name && *new_name && *old_name == *new_name) {
2347 if (*old_name == '/')
2348 pfx_length = old_name - a + 1;
2349 old_name++;
2350 new_name++;
2351 }
2352 /* Find common suffix */
2354 old_name = a + len_a;
2355 new_name = b + len_b;
2356 sfx_length = 0;
2357 /*
2358 * If there is a common prefix, it must end in a slash. In
2359 * that case we let this loop run 1 into the prefix to see the
2360 * same slash.
2361 *
2362 * If there is no common prefix, we cannot do this as it would
2363 * underrun the input strings.
2364 */
2365 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2366 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2367 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2368 *old_name == *new_name) {
2369 if (*old_name == '/')
2370 sfx_length = len_a - (old_name - a);
2371 old_name--;
2372 new_name--;
2373 }
2374 /*
2376 * pfx{mid-a => mid-b}sfx
2377 * {pfx-a => pfx-b}sfx
2378 * pfx{sfx-a => sfx-b}
2379 * name-a => name-b
2380 */
2381 a_midlen = len_a - pfx_length - sfx_length;
2382 b_midlen = len_b - pfx_length - sfx_length;
2383 if (a_midlen < 0)
2384 a_midlen = 0;
2385 if (b_midlen < 0)
2386 b_midlen = 0;
2387 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2389 if (pfx_length + sfx_length) {
2390 strbuf_add(name, a, pfx_length);
2391 strbuf_addch(name, '{');
2392 }
2393 strbuf_add(name, a + pfx_length, a_midlen);
2394 strbuf_addstr(name, " => ");
2395 strbuf_add(name, b + pfx_length, b_midlen);
2396 if (pfx_length + sfx_length) {
2397 strbuf_addch(name, '}');
2398 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2399 }
2400}
2401struct diffstat_t {
2403 int nr;
2404 int alloc;
2405 struct diffstat_file {
2406 char *from_name;
2407 char *name;
2408 char *print_name;
2409 const char *comments;
2410 unsigned is_unmerged:1;
2411 unsigned is_binary:1;
2412 unsigned is_renamed:1;
2413 unsigned is_interesting:1;
2414 uintmax_t added, deleted;
2415 } **files;
2416};
2417static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2419 const char *name_a,
2420 const char *name_b)
2421{
2422 struct diffstat_file *x;
2423 x = xcalloc(1, sizeof(*x));
2424 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2425 diffstat->files[diffstat->nr++] = x;
2426 if (name_b) {
2427 x->from_name = xstrdup(name_a);
2428 x->name = xstrdup(name_b);
2429 x->is_renamed = 1;
2430 }
2431 else {
2432 x->from_name = NULL;
2433 x->name = xstrdup(name_a);
2434 }
2435 return x;
2436}
2437static void diffstat_consume(void *priv, char *line, unsigned long len)
2439{
2440 struct diffstat_t *diffstat = priv;
2441 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2442 if (line[0] == '+')
2444 x->added++;
2445 else if (line[0] == '-')
2446 x->deleted++;
2447}
2448const char mime_boundary_leader[] = "------------";
2450static int scale_linear(int it, int width, int max_change)
2452{
2453 if (!it)
2454 return 0;
2455 /*
2456 * make sure that at least one '-' or '+' is printed if
2457 * there is any change to this path. The easiest way is to
2458 * scale linearly as if the alloted width is one column shorter
2459 * than it is, and then add 1 to the result.
2460 */
2461 return 1 + (it * (width - 1) / max_change);
2462}
2463static void show_graph(struct strbuf *out, char ch, int cnt,
2465 const char *set, const char *reset)
2466{
2467 if (cnt <= 0)
2468 return;
2469 strbuf_addstr(out, set);
2470 strbuf_addchars(out, ch, cnt);
2471 strbuf_addstr(out, reset);
2472}
2473static void fill_print_name(struct diffstat_file *file)
2475{
2476 struct strbuf pname = STRBUF_INIT;
2477 if (file->print_name)
2479 return;
2480 if (file->is_renamed)
2482 pprint_rename(&pname, file->from_name, file->name);
2483 else
2484 quote_c_style(file->name, &pname, NULL, 0);
2485 if (file->comments)
2487 strbuf_addf(&pname, " (%s)", file->comments);
2488 file->print_name = strbuf_detach(&pname, NULL);
2490}
2491static void print_stat_summary_inserts_deletes(struct diff_options *options,
2493 int files, int insertions, int deletions)
2494{
2495 struct strbuf sb = STRBUF_INIT;
2496 if (!files) {
2498 assert(insertions == 0 && deletions == 0);
2499 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2500 NULL, 0, 0);
2501 return;
2502 }
2503 strbuf_addf(&sb,
2505 (files == 1) ? " %d file changed" : " %d files changed",
2506 files);
2507 /*
2509 * For binary diff, the caller may want to print "x files
2510 * changed" with insertions == 0 && deletions == 0.
2511 *
2512 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2513 * is probably less confusing (i.e skip over "2 files changed
2514 * but nothing about added/removed lines? Is this a bug in Git?").
2515 */
2516 if (insertions || deletions == 0) {
2517 strbuf_addf(&sb,
2518 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2519 insertions);
2520 }
2521 if (deletions || insertions == 0) {
2523 strbuf_addf(&sb,
2524 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2525 deletions);
2526 }
2527 strbuf_addch(&sb, '\n');
2528 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2529 sb.buf, sb.len, 0);
2530 strbuf_release(&sb);
2531}
2532void print_stat_summary(FILE *fp, int files,
2534 int insertions, int deletions)
2535{
2536 struct diff_options o;
2537 memset(&o, 0, sizeof(o));
2538 o.file = fp;
2539 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2541}
2542static void show_stats(struct diffstat_t *data, struct diff_options *options)
2544{
2545 int i, len, add, del, adds = 0, dels = 0;
2546 uintmax_t max_change = 0, max_len = 0;
2547 int total_files = data->nr, count;
2548 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2549 const char *reset, *add_c, *del_c;
2550 int extra_shown = 0;
2551 const char *line_prefix = diff_line_prefix(options);
2552 struct strbuf out = STRBUF_INIT;
2553 if (data->nr == 0)
2555 return;
2556 count = options->stat_count ? options->stat_count : data->nr;
2558 reset = diff_get_color_opt(options, DIFF_RESET);
2560 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2561 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2562 /*
2564 * Find the longest filename and max number of changes
2565 */
2566 for (i = 0; (i < count) && (i < data->nr); i++) {
2567 struct diffstat_file *file = data->files[i];
2568 uintmax_t change = file->added + file->deleted;
2569 if (!file->is_interesting && (change == 0)) {
2571 count++; /* not shown == room for one more */
2572 continue;
2573 }
2574 fill_print_name(file);
2575 len = strlen(file->print_name);
2576 if (max_len < len)
2577 max_len = len;
2578 if (file->is_unmerged) {
2580 /* "Unmerged" is 8 characters */
2581 bin_width = bin_width < 8 ? 8 : bin_width;
2582 continue;
2583 }
2584 if (file->is_binary) {
2585 /* "Bin XXX -> YYY bytes" */
2586 int w = 14 + decimal_width(file->added)
2587 + decimal_width(file->deleted);
2588 bin_width = bin_width < w ? w : bin_width;
2589 /* Display change counts aligned with "Bin" */
2590 number_width = 3;
2591 continue;
2592 }
2593 if (max_change < change)
2595 max_change = change;
2596 }
2597 count = i; /* where we can stop scanning in data->files[] */
2598 /*
2600 * We have width = stat_width or term_columns() columns total.
2601 * We want a maximum of min(max_len, stat_name_width) for the name part.
2602 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2603 * We also need 1 for " " and 4 + decimal_width(max_change)
2604 * for " | NNNN " and one the empty column at the end, altogether
2605 * 6 + decimal_width(max_change).
2606 *
2607 * If there's not enough space, we will use the smaller of
2608 * stat_name_width (if set) and 5/8*width for the filename,
2609 * and the rest for constant elements + graph part, but no more
2610 * than stat_graph_width for the graph part.
2611 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2612 * for the standard terminal size).
2613 *
2614 * In other words: stat_width limits the maximum width, and
2615 * stat_name_width fixes the maximum width of the filename,
2616 * and is also used to divide available columns if there
2617 * aren't enough.
2618 *
2619 * Binary files are displayed with "Bin XXX -> YYY bytes"
2620 * instead of the change count and graph. This part is treated
2621 * similarly to the graph part, except that it is not
2622 * "scaled". If total width is too small to accommodate the
2623 * guaranteed minimum width of the filename part and the
2624 * separators and this message, this message will "overflow"
2625 * making the line longer than the maximum width.
2626 */
2627 if (options->stat_width == -1)
2629 width = term_columns() - strlen(line_prefix);
2630 else
2631 width = options->stat_width ? options->stat_width : 80;
2632 number_width = decimal_width(max_change) > number_width ?
2633 decimal_width(max_change) : number_width;
2634 if (options->stat_graph_width == -1)
2636 options->stat_graph_width = diff_stat_graph_width;
2637 /*
2639 * Guarantee 3/8*16==6 for the graph part
2640 * and 5/8*16==10 for the filename part
2641 */
2642 if (width < 16 + 6 + number_width)
2643 width = 16 + 6 + number_width;
2644 /*
2646 * First assign sizes that are wanted, ignoring available width.
2647 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2648 * starting from "XXX" should fit in graph_width.
2649 */
2650 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2651 if (options->stat_graph_width &&
2652 options->stat_graph_width < graph_width)
2653 graph_width = options->stat_graph_width;
2654 name_width = (options->stat_name_width > 0 &&
2656 options->stat_name_width < max_len) ?
2657 options->stat_name_width : max_len;
2658 /*
2660 * Adjust adjustable widths not to exceed maximum width
2661 */
2662 if (name_width + number_width + 6 + graph_width > width) {
2663 if (graph_width > width * 3/8 - number_width - 6) {
2664 graph_width = width * 3/8 - number_width - 6;
2665 if (graph_width < 6)
2666 graph_width = 6;
2667 }
2668 if (options->stat_graph_width &&
2670 graph_width > options->stat_graph_width)
2671 graph_width = options->stat_graph_width;
2672 if (name_width > width - number_width - 6 - graph_width)
2673 name_width = width - number_width - 6 - graph_width;
2674 else
2675 graph_width = width - number_width - 6 - name_width;
2676 }
2677 /*
2679 * From here name_width is the width of the name area,
2680 * and graph_width is the width of the graph area.
2681 * max_change is used to scale graph properly.
2682 */
2683 for (i = 0; i < count; i++) {
2684 const char *prefix = "";
2685 struct diffstat_file *file = data->files[i];
2686 char *name = file->print_name;
2687 uintmax_t added = file->added;
2688 uintmax_t deleted = file->deleted;
2689 int name_len;
2690 if (!file->is_interesting && (added + deleted == 0))
2692 continue;
2693 /*
2695 * "scale" the filename
2696 */
2697 len = name_width;
2698 name_len = strlen(name);
2699 if (name_width < name_len) {
2700 char *slash;
2701 prefix = "...";
2702 len -= 3;
2703 name += name_len - len;
2704 slash = strchr(name, '/');
2705 if (slash)
2706 name = slash;
2707 }
2708 if (file->is_binary) {
2710 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2711 strbuf_addf(&out, " %*s", number_width, "Bin");
2712 if (!added && !deleted) {
2713 strbuf_addch(&out, '\n');
2714 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2715 out.buf, out.len, 0);
2716 strbuf_reset(&out);
2717 continue;
2718 }
2719 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2720 del_c, deleted, reset);
2721 strbuf_addstr(&out, " -> ");
2722 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2723 add_c, added, reset);
2724 strbuf_addstr(&out, " bytes\n");
2725 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2726 out.buf, out.len, 0);
2727 strbuf_reset(&out);
2728 continue;
2729 }
2730 else if (file->is_unmerged) {
2731 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2732 strbuf_addstr(&out, " Unmerged\n");
2733 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2734 out.buf, out.len, 0);
2735 strbuf_reset(&out);
2736 continue;
2737 }
2738 /*
2740 * scale the add/delete
2741 */
2742 add = added;
2743 del = deleted;
2744 if (graph_width <= max_change) {
2746 int total = scale_linear(add + del, graph_width, max_change);
2747 if (total < 2 && add && del)
2748 /* width >= 2 due to the sanity check */
2749 total = 2;
2750 if (add < del) {
2751 add = scale_linear(add, graph_width, max_change);
2752 del = total - add;
2753 } else {
2754 del = scale_linear(del, graph_width, max_change);
2755 add = total - del;
2756 }
2757 }
2758 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2759 strbuf_addf(&out, " %*"PRIuMAX"%s",
2760 number_width, added + deleted,
2761 added + deleted ? " " : "");
2762 show_graph(&out, '+', add, add_c, reset);
2763 show_graph(&out, '-', del, del_c, reset);
2764 strbuf_addch(&out, '\n');
2765 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2766 out.buf, out.len, 0);
2767 strbuf_reset(&out);
2768 }
2769 for (i = 0; i < data->nr; i++) {
2771 struct diffstat_file *file = data->files[i];
2772 uintmax_t added = file->added;
2773 uintmax_t deleted = file->deleted;
2774 if (file->is_unmerged ||
2776 (!file->is_interesting && (added + deleted == 0))) {
2777 total_files--;
2778 continue;
2779 }
2780 if (!file->is_binary) {
2782 adds += added;
2783 dels += deleted;
2784 }
2785 if (i < count)
2786 continue;
2787 if (!extra_shown)
2788 emit_diff_symbol(options,
2789 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2790 NULL, 0, 0);
2791 extra_shown = 1;
2792 }
2793 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2795 strbuf_release(&out);
2796}
2797static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2799{
2800 int i, adds = 0, dels = 0, total_files = data->nr;
2801 if (data->nr == 0)
2803 return;
2804 for (i = 0; i < data->nr; i++) {
2806 int added = data->files[i]->added;
2807 int deleted = data->files[i]->deleted;
2808 if (data->files[i]->is_unmerged ||
2810 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2811 total_files--;
2812 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2813 adds += added;
2814 dels += deleted;
2815 }
2816 }
2817 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2818}
2819static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2821{
2822 int i;
2823 if (data->nr == 0)
2825 return;
2826 for (i = 0; i < data->nr; i++) {
2828 struct diffstat_file *file = data->files[i];
2829 fprintf(options->file, "%s", diff_line_prefix(options));
2831 if (file->is_binary)
2833 fprintf(options->file, "-\t-\t");
2834 else
2835 fprintf(options->file,
2836 "%"PRIuMAX"\t%"PRIuMAX"\t",
2837 file->added, file->deleted);
2838 if (options->line_termination) {
2839 fill_print_name(file);
2840 if (!file->is_renamed)
2841 write_name_quoted(file->name, options->file,
2842 options->line_termination);
2843 else {
2844 fputs(file->print_name, options->file);
2845 putc(options->line_termination, options->file);
2846 }
2847 } else {
2848 if (file->is_renamed) {
2849 putc('\0', options->file);
2850 write_name_quoted(file->from_name, options->file, '\0');
2851 }
2852 write_name_quoted(file->name, options->file, '\0');
2853 }
2854 }
2855}
2856struct dirstat_file {
2858 const char *name;
2859 unsigned long changed;
2860};
2861struct dirstat_dir {
2863 struct dirstat_file *files;
2864 int alloc, nr, permille, cumulative;
2865};
2866static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2868 unsigned long changed, const char *base, int baselen)
2869{
2870 unsigned long sum_changes = 0;
2871 unsigned int sources = 0;
2872 const char *line_prefix = diff_line_prefix(opt);
2873 while (dir->nr) {
2875 struct dirstat_file *f = dir->files;
2876 int namelen = strlen(f->name);
2877 unsigned long changes;
2878 char *slash;
2879 if (namelen < baselen)
2881 break;
2882 if (memcmp(f->name, base, baselen))
2883 break;
2884 slash = strchr(f->name + baselen, '/');
2885 if (slash) {
2886 int newbaselen = slash + 1 - f->name;
2887 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2888 sources++;
2889 } else {
2890 changes = f->changed;
2891 dir->files++;
2892 dir->nr--;
2893 sources += 2;
2894 }
2895 sum_changes += changes;
2896 }
2897 /*
2899 * We don't report dirstat's for
2900 * - the top level
2901 * - or cases where everything came from a single directory
2902 * under this directory (sources == 1).
2903 */
2904 if (baselen && sources != 1) {
2905 if (sum_changes) {
2906 int permille = sum_changes * 1000 / changed;
2907 if (permille >= dir->permille) {
2908 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2909 permille / 10, permille % 10, baselen, base);
2910 if (!dir->cumulative)
2911 return 0;
2912 }
2913 }
2914 }
2915 return sum_changes;
2916}
2917static int dirstat_compare(const void *_a, const void *_b)
2919{
2920 const struct dirstat_file *a = _a;
2921 const struct dirstat_file *b = _b;
2922 return strcmp(a->name, b->name);
2923}
2924static void show_dirstat(struct diff_options *options)
2926{
2927 int i;
2928 unsigned long changed;
2929 struct dirstat_dir dir;
2930 struct diff_queue_struct *q = &diff_queued_diff;
2931 dir.files = NULL;
2933 dir.alloc = 0;
2934 dir.nr = 0;
2935 dir.permille = options->dirstat_permille;
2936 dir.cumulative = options->flags.dirstat_cumulative;
2937 changed = 0;
2939 for (i = 0; i < q->nr; i++) {
2940 struct diff_filepair *p = q->queue[i];
2941 const char *name;
2942 unsigned long copied, added, damage;
2943 int content_changed;
2944 name = p->two->path ? p->two->path : p->one->path;
2946 if (p->one->oid_valid && p->two->oid_valid)
2948 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2949 else
2950 content_changed = 1;
2951 if (!content_changed) {
2953 /*
2954 * The SHA1 has not changed, so pre-/post-content is
2955 * identical. We can therefore skip looking at the
2956 * file contents altogether.
2957 */
2958 damage = 0;
2959 goto found_damage;
2960 }
2961 if (options->flags.dirstat_by_file) {
2963 /*
2964 * In --dirstat-by-file mode, we don't really need to
2965 * look at the actual file contents at all.
2966 * The fact that the SHA1 changed is enough for us to
2967 * add this file to the list of results
2968 * (with each file contributing equal damage).
2969 */
2970 damage = 1;
2971 goto found_damage;
2972 }
2973 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2975 diff_populate_filespec(p->one, 0);
2976 diff_populate_filespec(p->two, 0);
2977 diffcore_count_changes(p->one, p->two, NULL, NULL,
2978 &copied, &added);
2979 diff_free_filespec_data(p->one);
2980 diff_free_filespec_data(p->two);
2981 } else if (DIFF_FILE_VALID(p->one)) {
2982 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2983 copied = added = 0;
2984 diff_free_filespec_data(p->one);
2985 } else if (DIFF_FILE_VALID(p->two)) {
2986 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2987 copied = 0;
2988 added = p->two->size;
2989 diff_free_filespec_data(p->two);
2990 } else
2991 continue;
2992 /*
2994 * Original minus copied is the removed material,
2995 * added is the new material. They are both damages
2996 * made to the preimage.
2997 * If the resulting damage is zero, we know that
2998 * diffcore_count_changes() considers the two entries to
2999 * be identical, but since content_changed is true, we
3000 * know that there must have been _some_ kind of change,
3001 * so we force all entries to have damage > 0.
3002 */
3003 damage = (p->one->size - copied) + added;
3004 if (!damage)
3005 damage = 1;
3006found_damage:
3008 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3009 dir.files[dir.nr].name = name;
3010 dir.files[dir.nr].changed = damage;
3011 changed += damage;
3012 dir.nr++;
3013 }
3014 /* This can happen even with many files, if everything was renames */
3016 if (!changed)
3017 return;
3018 /* Show all directories with more than x% of the changes */
3020 QSORT(dir.files, dir.nr, dirstat_compare);
3021 gather_dirstat(options, &dir, changed, "", 0);
3022}
3023static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3025{
3026 int i;
3027 unsigned long changed;
3028 struct dirstat_dir dir;
3029 if (data->nr == 0)
3031 return;
3032 dir.files = NULL;
3034 dir.alloc = 0;
3035 dir.nr = 0;
3036 dir.permille = options->dirstat_permille;
3037 dir.cumulative = options->flags.dirstat_cumulative;
3038 changed = 0;
3040 for (i = 0; i < data->nr; i++) {
3041 struct diffstat_file *file = data->files[i];
3042 unsigned long damage = file->added + file->deleted;
3043 if (file->is_binary)
3044 /*
3045 * binary files counts bytes, not lines. Must find some
3046 * way to normalize binary bytes vs. textual lines.
3047 * The following heuristic assumes that there are 64
3048 * bytes per "line".
3049 * This is stupid and ugly, but very cheap...
3050 */
3051 damage = DIV_ROUND_UP(damage, 64);
3052 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3053 dir.files[dir.nr].name = file->name;
3054 dir.files[dir.nr].changed = damage;
3055 changed += damage;
3056 dir.nr++;
3057 }
3058 /* This can happen even with many files, if everything was renames */
3060 if (!changed)
3061 return;
3062 /* Show all directories with more than x% of the changes */
3064 QSORT(dir.files, dir.nr, dirstat_compare);
3065 gather_dirstat(options, &dir, changed, "", 0);
3066}
3067static void free_diffstat_info(struct diffstat_t *diffstat)
3069{
3070 int i;
3071 for (i = 0; i < diffstat->nr; i++) {
3072 struct diffstat_file *f = diffstat->files[i];
3073 free(f->print_name);
3074 free(f->name);
3075 free(f->from_name);
3076 free(f);
3077 }
3078 free(diffstat->files);
3079}
3080struct checkdiff_t {
3082 const char *filename;
3083 int lineno;
3084 int conflict_marker_size;
3085 struct diff_options *o;
3086 unsigned ws_rule;
3087 unsigned status;
3088};
3089static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3091{
3092 char firstchar;
3093 int cnt;
3094 if (len < marker_size + 1)
3096 return 0;
3097 firstchar = line[0];
3098 switch (firstchar) {
3099 case '=': case '>': case '<': case '|':
3100 break;
3101 default:
3102 return 0;
3103 }
3104 for (cnt = 1; cnt < marker_size; cnt++)
3105 if (line[cnt] != firstchar)
3106 return 0;
3107 /* line[1] thru line[marker_size-1] are same as firstchar */
3108 if (len < marker_size + 1 || !isspace(line[marker_size]))
3109 return 0;
3110 return 1;
3111}
3112static void checkdiff_consume(void *priv, char *line, unsigned long len)
3114{
3115 struct checkdiff_t *data = priv;
3116 int marker_size = data->conflict_marker_size;
3117 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3118 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3119 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3120 char *err;
3121 const char *line_prefix;
3122 assert(data->o);
3124 line_prefix = diff_line_prefix(data->o);
3125 if (line[0] == '+') {
3127 unsigned bad;
3128 data->lineno++;
3129 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3130 data->status |= 1;
3131 fprintf(data->o->file,
3132 "%s%s:%d: leftover conflict marker\n",
3133 line_prefix, data->filename, data->lineno);
3134 }
3135 bad = ws_check(line + 1, len - 1, data->ws_rule);
3136 if (!bad)
3137 return;
3138 data->status |= bad;
3139 err = whitespace_error_string(bad);
3140 fprintf(data->o->file, "%s%s:%d: %s.\n",
3141 line_prefix, data->filename, data->lineno, err);
3142 free(err);
3143 emit_line(data->o, set, reset, line, 1);
3144 ws_check_emit(line + 1, len - 1, data->ws_rule,
3145 data->o->file, set, reset, ws);
3146 } else if (line[0] == ' ') {
3147 data->lineno++;
3148 } else if (line[0] == '@') {
3149 char *plus = strchr(line, '+');
3150 if (plus)
3151 data->lineno = strtol(plus, NULL, 10) - 1;
3152 else
3153 die("invalid diff");
3154 }
3155}
3156static unsigned char *deflate_it(char *data,
3158 unsigned long size,
3159 unsigned long *result_size)
3160{
3161 int bound;
3162 unsigned char *deflated;
3163 git_zstream stream;
3164 git_deflate_init(&stream, zlib_compression_level);
3166 bound = git_deflate_bound(&stream, size);
3167 deflated = xmalloc(bound);
3168 stream.next_out = deflated;
3169 stream.avail_out = bound;
3170 stream.next_in = (unsigned char *)data;
3172 stream.avail_in = size;
3173 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3174 ; /* nothing */
3175 git_deflate_end(&stream);
3176 *result_size = stream.total_out;
3177 return deflated;
3178}
3179static void emit_binary_diff_body(struct diff_options *o,
3181 mmfile_t *one, mmfile_t *two)
3182{
3183 void *cp;
3184 void *delta;
3185 void *deflated;
3186 void *data;
3187 unsigned long orig_size;
3188 unsigned long delta_size;
3189 unsigned long deflate_size;
3190 unsigned long data_size;
3191 /* We could do deflated delta, or we could do just deflated two,
3193 * whichever is smaller.
3194 */
3195 delta = NULL;
3196 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3197 if (one->size && two->size) {
3198 delta = diff_delta(one->ptr, one->size,
3199 two->ptr, two->size,
3200 &delta_size, deflate_size);
3201 if (delta) {
3202 void *to_free = delta;
3203 orig_size = delta_size;
3204 delta = deflate_it(delta, delta_size, &delta_size);
3205 free(to_free);
3206 }
3207 }
3208 if (delta && delta_size < deflate_size) {
3210 char *s = xstrfmt("%lu", orig_size);
3211 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3212 s, strlen(s), 0);
3213 free(s);
3214 free(deflated);
3215 data = delta;
3216 data_size = delta_size;
3217 } else {
3218 char *s = xstrfmt("%lu", two->size);
3219 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3220 s, strlen(s), 0);
3221 free(s);
3222 free(delta);
3223 data = deflated;
3224 data_size = deflate_size;
3225 }
3226 /* emit data encoded in base85 */
3228 cp = data;
3229 while (data_size) {
3230 int len;
3231 int bytes = (52 < data_size) ? 52 : data_size;
3232 char line[71];
3233 data_size -= bytes;
3234 if (bytes <= 26)
3235 line[0] = bytes + 'A' - 1;
3236 else
3237 line[0] = bytes - 26 + 'a' - 1;
3238 encode_85(line + 1, cp, bytes);
3239 cp = (char *) cp + bytes;
3240 len = strlen(line);
3242 line[len++] = '\n';
3243 line[len] = '\0';
3244 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3246 line, len, 0);
3247 }
3248 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3249 free(data);
3250}
3251static void emit_binary_diff(struct diff_options *o,
3253 mmfile_t *one, mmfile_t *two)
3254{
3255 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3256 emit_binary_diff_body(o, one, two);
3257 emit_binary_diff_body(o, two, one);
3258}
3259int diff_filespec_is_binary(struct diff_filespec *one)
3261{
3262 if (one->is_binary == -1) {
3263 diff_filespec_load_driver(one);
3264 if (one->driver->binary != -1)
3265 one->is_binary = one->driver->binary;
3266 else {
3267 if (!one->data && DIFF_FILE_VALID(one))
3268 diff_populate_filespec(one, CHECK_BINARY);
3269 if (one->is_binary == -1 && one->data)
3270 one->is_binary = buffer_is_binary(one->data,
3271 one->size);
3272 if (one->is_binary == -1)
3273 one->is_binary = 0;
3274 }
3275 }
3276 return one->is_binary;
3277}
3278static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3280{
3281 diff_filespec_load_driver(one);
3282 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3283}
3284void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3286{
3287 if (!options->a_prefix)
3288 options->a_prefix = a;
3289 if (!options->b_prefix)
3290 options->b_prefix = b;
3291}
3292struct userdiff_driver *get_textconv(struct diff_filespec *one)
3294{
3295 if (!DIFF_FILE_VALID(one))
3296 return NULL;
3297 diff_filespec_load_driver(one);
3299 return userdiff_get_textconv(one->driver);
3300}
3301static void builtin_diff(const char *name_a,
3303 const char *name_b,
3304 struct diff_filespec *one,
3305 struct diff_filespec *two,
3306 const char *xfrm_msg,
3307 int must_show_header,
3308 struct diff_options *o,
3309 int complete_rewrite)
3310{
3311 mmfile_t mf1, mf2;
3312 const char *lbl[2];
3313 char *a_one, *b_two;
3314 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3315 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3316 const char *a_prefix, *b_prefix;
3317 struct userdiff_driver *textconv_one = NULL;
3318 struct userdiff_driver *textconv_two = NULL;
3319 struct strbuf header = STRBUF_INIT;
3320 const char *line_prefix = diff_line_prefix(o);
3321 diff_set_mnemonic_prefix(o, "a/", "b/");
3323 if (o->flags.reverse_diff) {
3324 a_prefix = o->b_prefix;
3325 b_prefix = o->a_prefix;
3326 } else {
3327 a_prefix = o->a_prefix;
3328 b_prefix = o->b_prefix;
3329 }
3330 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3332 (!one->mode || S_ISGITLINK(one->mode)) &&
3333 (!two->mode || S_ISGITLINK(two->mode))) {
3334 show_submodule_summary(o, one->path ? one->path : two->path,
3335 &one->oid, &two->oid,
3336 two->dirty_submodule);
3337 return;
3338 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3339 (!one->mode || S_ISGITLINK(one->mode)) &&
3340 (!two->mode || S_ISGITLINK(two->mode))) {
3341 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3342 &one->oid, &two->oid,
3343 two->dirty_submodule);
3344 return;
3345 }
3346 if (o->flags.allow_textconv) {
3348 textconv_one = get_textconv(one);
3349 textconv_two = get_textconv(two);
3350 }
3351 /* Never use a non-valid filename anywhere if at all possible */
3353 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3354 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3355 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3357 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3358 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3359 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3360 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3361 if (lbl[0][0] == '/') {
3362 /* /dev/null */
3363 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3364 if (xfrm_msg)
3365 strbuf_addstr(&header, xfrm_msg);
3366 must_show_header = 1;
3367 }
3368 else if (lbl[1][0] == '/') {
3369 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3370 if (xfrm_msg)
3371 strbuf_addstr(&header, xfrm_msg);
3372 must_show_header = 1;
3373 }
3374 else {
3375 if (one->mode != two->mode) {
3376 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3377 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3378 must_show_header = 1;
3379 }
3380 if (xfrm_msg)
3381 strbuf_addstr(&header, xfrm_msg);
3382 /*
3384 * we do not run diff between different kind
3385 * of objects.
3386 */
3387 if ((one->mode ^ two->mode) & S_IFMT)
3388 goto free_ab_and_return;
3389 if (complete_rewrite &&
3390 (textconv_one || !diff_filespec_is_binary(one)) &&
3391 (textconv_two || !diff_filespec_is_binary(two))) {
3392 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3393 header.buf, header.len, 0);
3394 strbuf_reset(&header);
3395 emit_rewrite_diff(name_a, name_b, one, two,
3396 textconv_one, textconv_two, o);
3397 o->found_changes = 1;
3398 goto free_ab_and_return;
3399 }
3400 }
3401 if (o->irreversible_delete && lbl[1][0] == '/') {
3403 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3404 header.len, 0);
3405 strbuf_reset(&header);
3406 goto free_ab_and_return;
3407 } else if (!o->flags.text &&
3408 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3409 (!textconv_two && diff_filespec_is_binary(two)) )) {
3410 struct strbuf sb = STRBUF_INIT;
3411 if (!one->data && !two->data &&
3412 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3413 !o->flags.binary) {
3414 if (!oidcmp(&one->oid, &two->oid)) {
3415 if (must_show_header)
3416 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3417 header.buf, header.len,
3418 0);
3419 goto free_ab_and_return;
3420 }
3421 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3422 header.buf, header.len, 0);
3423 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3424 diff_line_prefix(o), lbl[0], lbl[1]);
3425 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3426 sb.buf, sb.len, 0);
3427 strbuf_release(&sb);
3428 goto free_ab_and_return;
3429 }
3430 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3431 die("unable to read files to diff");
3432 /* Quite common confusing case */
3433 if (mf1.size == mf2.size &&
3434 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3435 if (must_show_header)
3436 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3437 header.buf, header.len, 0);
3438 goto free_ab_and_return;
3439 }
3440 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3441 strbuf_reset(&header);
3442 if (o->flags.binary)
3443 emit_binary_diff(o, &mf1, &mf2);
3444 else {
3445 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3446 diff_line_prefix(o), lbl[0], lbl[1]);
3447 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3448 sb.buf, sb.len, 0);
3449 strbuf_release(&sb);
3450 }
3451 o->found_changes = 1;
3452 } else {
3453 /* Crazy xdl interfaces.. */
3454 const char *diffopts = getenv("GIT_DIFF_OPTS");
3455 const char *v;
3456 xpparam_t xpp;
3457 xdemitconf_t xecfg;
3458 struct emit_callback ecbdata;
3459 const struct userdiff_funcname *pe;
3460 if (must_show_header) {
3462 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3463 header.buf, header.len, 0);
3464 strbuf_reset(&header);
3465 }
3466 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3468 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3469 pe = diff_funcname_pattern(one);
3471 if (!pe)
3472 pe = diff_funcname_pattern(two);
3473 memset(&xpp, 0, sizeof(xpp));
3475 memset(&xecfg, 0, sizeof(xecfg));
3476 memset(&ecbdata, 0, sizeof(ecbdata));
3477 if (o->flags.suppress_diff_headers)
3478 lbl[0] = NULL;
3479 ecbdata.label_path = lbl;
3480 ecbdata.color_diff = want_color(o->use_color);
3481 ecbdata.ws_rule = whitespace_rule(name_b);
3482 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3483 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3484 ecbdata.opt = o;
3485 if (header.len && !o->flags.suppress_diff_headers)
3486 ecbdata.header = &header;
3487 xpp.flags = o->xdl_opts;
3488 xpp.anchors = o->anchors;
3489 xpp.anchors_nr = o->anchors_nr;
3490 xecfg.ctxlen = o->context;
3491 xecfg.interhunkctxlen = o->interhunkcontext;
3492 xecfg.flags = XDL_EMIT_FUNCNAMES;
3493 if (o->flags.funccontext)
3494 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3495 if (pe)
3496 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3497 if (!diffopts)
3498 ;
3499 else if (skip_prefix(diffopts, "--unified=", &v))
3500 xecfg.ctxlen = strtoul(v, NULL, 10);
3501 else if (skip_prefix(diffopts, "-u", &v))
3502 xecfg.ctxlen = strtoul(v, NULL, 10);
3503 if (o->word_diff)
3504 init_diff_words_data(&ecbdata, o, one, two);
3505 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
3506 &xpp, &xecfg))
3507 die("unable to generate diff for %s", one->path);
3508 if (o->word_diff)
3509 free_diff_words_data(&ecbdata);
3510 if (textconv_one)
3511 free(mf1.ptr);
3512 if (textconv_two)
3513 free(mf2.ptr);
3514 xdiff_clear_find_func(&xecfg);
3515 }
3516 free_ab_and_return:
3518 strbuf_release(&header);
3519 diff_free_filespec_data(one);
3520 diff_free_filespec_data(two);
3521 free(a_one);
3522 free(b_two);
3523 return;
3524}
3525static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3527{
3528 if (!is_renamed) {
3529 if (p->status == DIFF_STATUS_ADDED) {
3530 if (S_ISLNK(p->two->mode))
3531 return "new +l";
3532 else if ((p->two->mode & 0777) == 0755)
3533 return "new +x";
3534 else
3535 return "new";
3536 } else if (p->status == DIFF_STATUS_DELETED)
3537 return "gone";
3538 }
3539 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3540 return "mode -l";
3541 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3542 return "mode +l";
3543 else if ((p->one->mode & 0777) == 0644 &&
3544 (p->two->mode & 0777) == 0755)
3545 return "mode +x";
3546 else if ((p->one->mode & 0777) == 0755 &&
3547 (p->two->mode & 0777) == 0644)
3548 return "mode -x";
3549 return NULL;
3550}
3551static void builtin_diffstat(const char *name_a, const char *name_b,
3553 struct diff_filespec *one,
3554 struct diff_filespec *two,
3555 struct diffstat_t *diffstat,
3556 struct diff_options *o,
3557 struct diff_filepair *p)
3558{
3559 mmfile_t mf1, mf2;
3560 struct diffstat_file *data;
3561 int same_contents;
3562 int complete_rewrite = 0;
3563 if (!DIFF_PAIR_UNMERGED(p)) {
3565 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3566 complete_rewrite = 1;
3567 }
3568 data = diffstat_add(diffstat, name_a, name_b);
3570 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3571 if (o->flags.stat_with_summary)
3572 data->comments = get_compact_summary(p, data->is_renamed);
3573 if (!one || !two) {
3575 data->is_unmerged = 1;
3576 return;
3577 }
3578 same_contents = !oidcmp(&one->oid, &two->oid);
3580 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3582 data->is_binary = 1;
3583 if (same_contents) {
3584 data->added = 0;
3585 data->deleted = 0;
3586 } else {
3587 data->added = diff_filespec_size(two);
3588 data->deleted = diff_filespec_size(one);
3589 }
3590 }
3591 else if (complete_rewrite) {
3593 diff_populate_filespec(one, 0);
3594 diff_populate_filespec(two, 0);
3595 data->deleted = count_lines(one->data, one->size);
3596 data->added = count_lines(two->data, two->size);
3597 }
3598 else if (!same_contents) {
3600 /* Crazy xdl interfaces.. */
3601 xpparam_t xpp;
3602 xdemitconf_t xecfg;
3603 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3605 die("unable to read files to diff");
3606 memset(&xpp, 0, sizeof(xpp));
3608 memset(&xecfg, 0, sizeof(xecfg));
3609 xpp.flags = o->xdl_opts;
3610 xpp.anchors = o->anchors;
3611 xpp.anchors_nr = o->anchors_nr;
3612 xecfg.ctxlen = o->context;
3613 xecfg.interhunkctxlen = o->interhunkcontext;
3614 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
3615 &xpp, &xecfg))
3616 die("unable to generate diffstat for %s", one->path);
3617 }
3618 diff_free_filespec_data(one);
3620 diff_free_filespec_data(two);
3621}
3622static void builtin_checkdiff(const char *name_a, const char *name_b,
3624 const char *attr_path,
3625 struct diff_filespec *one,
3626 struct diff_filespec *two,
3627 struct diff_options *o)
3628{
3629 mmfile_t mf1, mf2;
3630 struct checkdiff_t data;
3631 if (!two)
3633 return;
3634 memset(&data, 0, sizeof(data));
3636 data.filename = name_b ? name_b : name_a;
3637 data.lineno = 0;
3638 data.o = o;
3639 data.ws_rule = whitespace_rule(attr_path);
3640 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3641 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3643 die("unable to read files to diff");
3644 /*
3646 * All the other codepaths check both sides, but not checking
3647 * the "old" side here is deliberate. We are checking the newly
3648 * introduced changes, and as long as the "new" side is text, we
3649 * can and should check what it introduces.
3650 */
3651 if (diff_filespec_is_binary(two))
3652 goto free_and_return;
3653 else {
3654 /* Crazy xdl interfaces.. */
3655 xpparam_t xpp;
3656 xdemitconf_t xecfg;
3657 memset(&xpp, 0, sizeof(xpp));
3659 memset(&xecfg, 0, sizeof(xecfg));
3660 xecfg.ctxlen = 1; /* at least one context line */
3661 xpp.flags = 0;
3662 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
3663 &xpp, &xecfg))
3664 die("unable to generate checkdiff for %s", one->path);
3665 if (data.ws_rule & WS_BLANK_AT_EOF) {
3667 struct emit_callback ecbdata;
3668 int blank_at_eof;
3669 ecbdata.ws_rule = data.ws_rule;
3671 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3672 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3673 if (blank_at_eof) {
3675 static char *err;
3676 if (!err)
3677 err = whitespace_error_string(WS_BLANK_AT_EOF);
3678 fprintf(o->file, "%s:%d: %s.\n",
3679 data.filename, blank_at_eof, err);
3680 data.status = 1; /* report errors */
3681 }
3682 }
3683 }
3684 free_and_return:
3685 diff_free_filespec_data(one);
3686 diff_free_filespec_data(two);
3687 if (data.status)
3688 o->flags.check_failed = 1;
3689}
3690struct diff_filespec *alloc_filespec(const char *path)
3692{
3693 struct diff_filespec *spec;
3694 FLEXPTR_ALLOC_STR(spec, path, path);
3696 spec->count = 1;
3697 spec->is_binary = -1;
3698 return spec;
3699}
3700void free_filespec(struct diff_filespec *spec)
3702{
3703 if (!--spec->count) {
3704 diff_free_filespec_data(spec);
3705 free(spec);
3706 }
3707}
3708void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3710 int oid_valid, unsigned short mode)
3711{
3712 if (mode) {
3713 spec->mode = canon_mode(mode);
3714 oidcpy(&spec->oid, oid);
3715 spec->oid_valid = oid_valid;
3716 }
3717}
3718/*
3720 * Given a name and sha1 pair, if the index tells us the file in
3721 * the work tree has that object contents, return true, so that
3722 * prepare_temp_file() does not have to inflate and extract.
3723 */
3724static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3725{
3726 const struct cache_entry *ce;
3727 struct stat st;
3728 int pos, len;
3729 /*
3731 * We do not read the cache ourselves here, because the
3732 * benchmark with my previous version that always reads cache
3733 * shows that it makes things worse for diff-tree comparing
3734 * two linux-2.6 kernel trees in an already checked out work
3735 * tree. This is because most diff-tree comparisons deal with
3736 * only a small number of files, while reading the cache is
3737 * expensive for a large project, and its cost outweighs the
3738 * savings we get by not inflating the object to a temporary
3739 * file. Practically, this code only helps when we are used
3740 * by diff-cache --cached, which does read the cache before
3741 * calling us.
3742 */
3743 if (!active_cache)
3744 return 0;
3745 /* We want to avoid the working directory if our caller
3747 * doesn't need the data in a normal file, this system
3748 * is rather slow with its stat/open/mmap/close syscalls,
3749 * and the object is contained in a pack file. The pack
3750 * is probably already open and will be faster to obtain
3751 * the data through than the working directory. Loose
3752 * objects however would tend to be slower as they need
3753 * to be individually opened and inflated.
3754 */
3755 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3756 return 0;
3757 /*
3759 * Similarly, if we'd have to convert the file contents anyway, that
3760 * makes the optimization not worthwhile.
3761 */
3762 if (!want_file && would_convert_to_git(&the_index, name))
3763 return 0;
3764 len = strlen(name);
3766 pos = cache_name_pos(name, len);
3767 if (pos < 0)
3768 return 0;
3769 ce = active_cache[pos];
3770 /*
3772 * This is not the sha1 we are looking for, or
3773 * unreusable because it is not a regular file.
3774 */
3775 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3776 return 0;
3777 /*
3779 * If ce is marked as "assume unchanged", there is no
3780 * guarantee that work tree matches what we are looking for.
3781 */
3782 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3783 return 0;
3784 /*
3786 * If ce matches the file in the work tree, we can reuse it.
3787 */
3788 if (ce_uptodate(ce) ||
3789 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3790 return 1;
3791 return 0;
3793}
3794static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3796{
3797 struct strbuf buf = STRBUF_INIT;
3798 char *dirty = "";
3799 /* Are we looking at the work tree? */
3801 if (s->dirty_submodule)
3802 dirty = "-dirty";
3803 strbuf_addf(&buf, "Subproject commit %s%s\n",
3805 oid_to_hex(&s->oid), dirty);
3806 s->size = buf.len;
3807 if (size_only) {
3808 s->data = NULL;
3809 strbuf_release(&buf);
3810 } else {
3811 s->data = strbuf_detach(&buf, NULL);
3812 s->should_free = 1;
3813 }
3814 return 0;
3815}
3816/*
3818 * While doing rename detection and pickaxe operation, we may need to
3819 * grab the data for the blob (or file) for our own in-core comparison.
3820 * diff_filespec has data and size fields for this purpose.
3821 */
3822int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3823{
3824 int size_only = flags & CHECK_SIZE_ONLY;
3825 int err = 0;
3826 int conv_flags = global_conv_flags_eol;
3827 /*
3828 * demote FAIL to WARN to allow inspecting the situation
3829 * instead of refusing.
3830 */
3831 if (conv_flags & CONV_EOL_RNDTRP_DIE)
3832 conv_flags = CONV_EOL_RNDTRP_WARN;
3833 if (!DIFF_FILE_VALID(s))
3835 die("internal error: asking to populate invalid file.");
3836 if (S_ISDIR(s->mode))
3837 return -1;
3838 if (s->data)
3840 return 0;
3841 if (size_only && 0 < s->size)
3843 return 0;
3844 if (S_ISGITLINK(s->mode))
3846 return diff_populate_gitlink(s, size_only);
3847 if (!s->oid_valid ||
3849 reuse_worktree_file(s->path, &s->oid, 0)) {
3850 struct strbuf buf = STRBUF_INIT;
3851 struct stat st;
3852 int fd;
3853 if (lstat(s->path, &st) < 0) {
3855 err_empty:
3856 err = -1;
3857 empty:
3858 s->data = (char *)"";
3859 s->size = 0;
3860 return err;
3861 }
3862 s->size = xsize_t(st.st_size);
3863 if (!s->size)
3864 goto empty;
3865 if (S_ISLNK(st.st_mode)) {
3866 struct strbuf sb = STRBUF_INIT;
3867 if (strbuf_readlink(&sb, s->path, s->size))
3869 goto err_empty;
3870 s->size = sb.len;
3871 s->data = strbuf_detach(&sb, NULL);
3872 s->should_free = 1;
3873 return 0;
3874 }
3875 /*
3877 * Even if the caller would be happy with getting
3878 * only the size, we cannot return early at this
3879 * point if the path requires us to run the content
3880 * conversion.
3881 */
3882 if (size_only && !would_convert_to_git(&the_index, s->path))
3883 return 0;
3884 /*
3886 * Note: this check uses xsize_t(st.st_size) that may
3887 * not be the true size of the blob after it goes
3888 * through convert_to_git(). This may not strictly be
3889 * correct, but the whole point of big_file_threshold
3890 * and is_binary check being that we want to avoid
3891 * opening the file and inspecting the contents, this
3892 * is probably fine.
3893 */
3894 if ((flags & CHECK_BINARY) &&
3895 s->size > big_file_threshold && s->is_binary == -1) {
3896 s->is_binary = 1;
3897 return 0;
3898 }
3899 fd = open(s->path, O_RDONLY);
3900 if (fd < 0)
3901 goto err_empty;
3902 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3903 close(fd);
3904 s->should_munmap = 1;
3905 /*
3907 * Convert from working tree format to canonical git format
3908 */
3909 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, conv_flags)) {
3910 size_t size = 0;
3911 munmap(s->data, s->size);
3912 s->should_munmap = 0;
3913 s->data = strbuf_detach(&buf, &size);
3914 s->size = size;
3915 s->should_free = 1;
3916 }
3917 }
3918 else {
3919 enum object_type type;
3920 if (size_only || (flags & CHECK_BINARY)) {
3921 type = oid_object_info(the_repository, &s->oid,
3922 &s->size);
3923 if (type < 0)
3924 die("unable to read %s",
3925 oid_to_hex(&s->oid));
3926 if (size_only)
3927 return 0;
3928 if (s->size > big_file_threshold && s->is_binary == -1) {
3929 s->is_binary = 1;
3930 return 0;
3931 }
3932 }
3933 s->data = read_object_file(&s->oid, &type, &s->size);
3934 if (!s->data)
3935 die("unable to read %s", oid_to_hex(&s->oid));
3936 s->should_free = 1;
3937 }
3938 return 0;
3939}
3940void diff_free_filespec_blob(struct diff_filespec *s)
3942{
3943 if (s->should_free)
3944 free(s->data);
3945 else if (s->should_munmap)
3946 munmap(s->data, s->size);
3947 if (s->should_free || s->should_munmap) {
3949 s->should_free = s->should_munmap = 0;
3950 s->data = NULL;
3951 }
3952}
3953void diff_free_filespec_data(struct diff_filespec *s)
3955{
3956 diff_free_filespec_blob(s);
3957 FREE_AND_NULL(s->cnt_data);
3958}
3959static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3961 void *blob,
3962 unsigned long size,
3963 const struct object_id *oid,
3964 int mode)
3965{
3966 struct strbuf buf = STRBUF_INIT;
3967 struct strbuf tempfile = STRBUF_INIT;
3968 char *path_dup = xstrdup(path);
3969 const char *base = basename(path_dup);
3970 /* Generate "XXXXXX_basename.ext" */
3972 strbuf_addstr(&tempfile, "XXXXXX_");
3973 strbuf_addstr(&tempfile, base);
3974 temp->tempfile = mks_tempfile_ts(tempfile.buf, strlen(base) + 1);
3976 if (!temp->tempfile)
3977 die_errno("unable to create temp-file");
3978 if (convert_to_working_tree(path,
3979 (const char *)blob, (size_t)size, &buf)) {
3980 blob = buf.buf;
3981 size = buf.len;
3982 }
3983 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
3984 close_tempfile_gently(temp->tempfile))
3985 die_errno("unable to write temp-file");
3986 temp->name = get_tempfile_path(temp->tempfile);
3987 oid_to_hex_r(temp->hex, oid);
3988 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3989 strbuf_release(&buf);
3990 strbuf_release(&tempfile);
3991 free(path_dup);
3992}
3993static struct diff_tempfile *prepare_temp_file(const char *name,
3995 struct diff_filespec *one)
3996{
3997 struct diff_tempfile *temp = claim_diff_tempfile();
3998 if (!DIFF_FILE_VALID(one)) {
4000 not_a_valid_file:
4001 /* A '-' entry produces this for file-2, and
4002 * a '+' entry produces this for file-1.
4003 */
4004 temp->name = "/dev/null";
4005 xsnprintf(temp->hex, sizeof(temp->hex), ".");
4006 xsnprintf(temp->mode, sizeof(temp->mode), ".");
4007 return temp;
4008 }
4009 if (!S_ISGITLINK(one->mode) &&
4011 (!one->oid_valid ||
4012 reuse_worktree_file(name, &one->oid, 1))) {
4013 struct stat st;
4014 if (lstat(name, &st) < 0) {
4015 if (errno == ENOENT)
4016 goto not_a_valid_file;
4017 die_errno("stat(%s)", name);
4018 }
4019 if (S_ISLNK(st.st_mode)) {
4020 struct strbuf sb = STRBUF_INIT;
4021 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4022 die_errno("readlink(%s)", name);
4023 prep_temp_blob(name, temp, sb.buf, sb.len,
4024 (one->oid_valid ?
4025 &one->oid : &null_oid),
4026 (one->oid_valid ?
4027 one->mode : S_IFLNK));
4028 strbuf_release(&sb);
4029 }
4030 else {
4031 /* we can borrow from the file in the work tree */
4032 temp->name = name;
4033 if (!one->oid_valid)
4034 oid_to_hex_r(temp->hex, &null_oid);
4035 else
4036 oid_to_hex_r(temp->hex, &one->oid);
4037 /* Even though we may sometimes borrow the
4038 * contents from the work tree, we always want
4039 * one->mode. mode is trustworthy even when
4040 * !(one->oid_valid), as long as
4041 * DIFF_FILE_VALID(one).
4042 */
4043 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4044 }
4045 return temp;
4046 }
4047 else {
4048 if (diff_populate_filespec(one, 0))
4049 die("cannot read data blob for %s", one->path);
4050 prep_temp_blob(name, temp, one->data, one->size,
4051 &one->oid, one->mode);
4052 }
4053 return temp;
4054}
4055static void add_external_diff_name(struct argv_array *argv,
4057 const char *name,
4058 struct diff_filespec *df)
4059{
4060 struct diff_tempfile *temp = prepare_temp_file(name, df);
4061 argv_array_push(argv, temp->name);
4062 argv_array_push(argv, temp->hex);
4063 argv_array_push(argv, temp->mode);
4064}
4065/* An external diff command takes:
4067 *
4068 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4069 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4070 *
4071 */
4072static void run_external_diff(const char *pgm,
4073 const char *name,
4074 const char *other,
4075 struct diff_filespec *one,
4076 struct diff_filespec *two,
4077 const char *xfrm_msg,
4078 int complete_rewrite,
4079 struct diff_options *o)
4080{
4081 struct argv_array argv = ARGV_ARRAY_INIT;
4082 struct argv_array env = ARGV_ARRAY_INIT;
4083 struct diff_queue_struct *q = &diff_queued_diff;
4084 argv_array_push(&argv, pgm);
4086 argv_array_push(&argv, name);
4087 if (one && two) {
4089 add_external_diff_name(&argv, name, one);
4090 if (!other)
4091 add_external_diff_name(&argv, name, two);
4092 else {
4093 add_external_diff_name(&argv, other, two);
4094 argv_array_push(&argv, other);
4095 argv_array_push(&argv, xfrm_msg);
4096 }
4097 }
4098 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
4100 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4101 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
4103 die(_("external diff died, stopping at %s"), name);
4104 remove_tempfile();
4106 argv_array_clear(&argv);
4107 argv_array_clear(&env);
4108}
4109static int similarity_index(struct diff_filepair *p)
4111{
4112 return p->score * 100 / MAX_SCORE;
4113}
4114static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4116{
4117 if (startup_info->have_repository)
4118 return find_unique_abbrev(oid, abbrev);
4119 else {
4120 char *hex = oid_to_hex(oid);
4121 if (abbrev < 0)
4122 abbrev = FALLBACK_DEFAULT_ABBREV;
4123 if (abbrev > the_hash_algo->hexsz)
4124 BUG("oid abbreviation out of range: %d", abbrev);
4125 if (abbrev)
4126 hex[abbrev] = '\0';
4127 return hex;
4128 }
4129}
4130static void fill_metainfo(struct strbuf *msg,
4132 const char *name,
4133 const char *other,
4134 struct diff_filespec *one,
4135 struct diff_filespec *two,
4136 struct diff_options *o,
4137 struct diff_filepair *p,
4138 int *must_show_header,
4139 int use_color)
4140{
4141 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4142 const char *reset = diff_get_color(use_color, DIFF_RESET);
4143 const char *line_prefix = diff_line_prefix(o);
4144 *must_show_header = 1;
4146 strbuf_init(msg, PATH_MAX * 2 + 300);
4147 switch (p->status) {
4148 case DIFF_STATUS_COPIED:
4149 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4150 line_prefix, set, similarity_index(p));
4151 strbuf_addf(msg, "%s\n%s%scopy from ",
4152 reset, line_prefix, set);
4153 quote_c_style(name, msg, NULL, 0);
4154 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4155 quote_c_style(other, msg, NULL, 0);
4156 strbuf_addf(msg, "%s\n", reset);
4157 break;
4158 case DIFF_STATUS_RENAMED:
4159 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4160 line_prefix, set, similarity_index(p));
4161 strbuf_addf(msg, "%s\n%s%srename from ",
4162 reset, line_prefix, set);
4163 quote_c_style(name, msg, NULL, 0);
4164 strbuf_addf(msg, "%s\n%s%srename to ",
4165 reset, line_prefix, set);
4166 quote_c_style(other, msg, NULL, 0);
4167 strbuf_addf(msg, "%s\n", reset);
4168 break;
4169 case DIFF_STATUS_MODIFIED:
4170 if (p->score) {
4171 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4172 line_prefix,
4173 set, similarity_index(p), reset);
4174 break;
4175 }
4176 /* fallthru */
4177 default:
4178 *must_show_header = 0;
4179 }
4180 if (one && two && oidcmp(&one->oid, &two->oid)) {
4181 const unsigned hexsz = the_hash_algo->hexsz;
4182 int abbrev = o->flags.full_index ? hexsz : DEFAULT_ABBREV;
4183 if (o->flags.binary) {
4185 mmfile_t mf;
4186 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
4187 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
4188 abbrev = hexsz;
4189 }
4190 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4191 diff_abbrev_oid(&one->oid, abbrev),
4192 diff_abbrev_oid(&two->oid, abbrev));
4193 if (one->mode == two->mode)
4194 strbuf_addf(msg, " %06o", one->mode);
4195 strbuf_addf(msg, "%s\n", reset);
4196 }
4197}
4198static void run_diff_cmd(const char *pgm,
4200 const char *name,
4201 const char *other,
4202 const char *attr_path,
4203 struct diff_filespec *one,
4204 struct diff_filespec *two,
4205 struct strbuf *msg,
4206 struct diff_options *o,
4207 struct diff_filepair *p)
4208{
4209 const char *xfrm_msg = NULL;
4210 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4211 int must_show_header = 0;
4212 if (o->flags.allow_external) {
4215 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
4216 if (drv && drv->external)
4217 pgm = drv->external;
4218 }
4219 if (msg) {
4221 /*
4222 * don't use colors when the header is intended for an
4223 * external diff driver
4224 */
4225 fill_metainfo(msg, name, other, one, two, o, p,
4226 &must_show_header,
4227 want_color(o->use_color) && !pgm);
4228 xfrm_msg = msg->len ? msg->buf : NULL;
4229 }
4230 if (pgm) {
4232 run_external_diff(pgm, name, other, one, two, xfrm_msg,
4233 complete_rewrite, o);
4234 return;
4235 }
4236 if (one && two)
4237 builtin_diff(name, other ? other : name,
4238 one, two, xfrm_msg, must_show_header,
4239 o, complete_rewrite);
4240 else
4241 fprintf(o->file, "* Unmerged path %s\n", name);
4242}
4243static void diff_fill_oid_info(struct diff_filespec *one)
4245{
4246 if (DIFF_FILE_VALID(one)) {
4247 if (!one->oid_valid) {
4248 struct stat st;
4249 if (one->is_stdin) {
4250 oidclr(&one->oid);
4251 return;
4252 }
4253 if (lstat(one->path, &st) < 0)
4254 die_errno("stat '%s'", one->path);
4255 if (index_path(&one->oid, one->path, &st, 0))
4256 die("cannot hash %s", one->path);
4257 }
4258 }
4259 else
4260 oidclr(&one->oid);
4261}
4262static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4264{
4265 /* Strip the prefix but do not molest /dev/null and absolute paths */
4266 if (*namep && **namep != '/') {
4267 *namep += prefix_length;
4268 if (**namep == '/')
4269 ++*namep;
4270 }
4271 if (*otherp && **otherp != '/') {
4272 *otherp += prefix_length;
4273 if (**otherp == '/')
4274 ++*otherp;
4275 }
4276}
4277static void run_diff(struct diff_filepair *p, struct diff_options *o)
4279{
4280 const char *pgm = external_diff();
4281 struct strbuf msg;
4282 struct diff_filespec *one = p->one;
4283 struct diff_filespec *two = p->two;
4284 const char *name;
4285 const char *other;
4286 const char *attr_path;
4287 name = one->path;
4289 other = (strcmp(name, two->path) ? two->path : NULL);
4290 attr_path = name;
4291 if (o->prefix_length)
4292 strip_prefix(o->prefix_length, &name, &other);
4293 if (!o->flags.allow_external)
4295 pgm = NULL;
4296 if (DIFF_PAIR_UNMERGED(p)) {
4298 run_diff_cmd(pgm, name, NULL, attr_path,
4299 NULL, NULL, NULL, o, p);
4300 return;
4301 }
4302 diff_fill_oid_info(one);
4304 diff_fill_oid_info(two);
4305 if (!pgm &&
4307 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4308 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4309 /*
4310 * a filepair that changes between file and symlink
4311 * needs to be split into deletion and creation.
4312 */
4313 struct diff_filespec *null = alloc_filespec(two->path);
4314 run_diff_cmd(NULL, name, other, attr_path,
4315 one, null, &msg, o, p);
4316 free(null);
4317 strbuf_release(&msg);
4318 null = alloc_filespec(one->path);
4320 run_diff_cmd(NULL, name, other, attr_path,
4321 null, two, &msg, o, p);
4322 free(null);
4323 }
4324 else
4325 run_diff_cmd(pgm, name, other, attr_path,
4326 one, two, &msg, o, p);
4327 strbuf_release(&msg);
4329}
4330static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4332 struct diffstat_t *diffstat)
4333{
4334 const char *name;
4335 const char *other;
4336 if (DIFF_PAIR_UNMERGED(p)) {
4338 /* unmerged */
4339 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4340 return;
4341 }
4342 name = p->one->path;
4344 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4345 if (o->prefix_length)
4347 strip_prefix(o->prefix_length, &name, &other);
4348 diff_fill_oid_info(p->one);
4350 diff_fill_oid_info(p->two);
4351 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4353}
4354static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4356{
4357 const char *name;
4358 const char *other;
4359 const char *attr_path;
4360 if (DIFF_PAIR_UNMERGED(p)) {
4362 /* unmerged */
4363 return;
4364 }
4365 name = p->one->path;
4367 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4368 attr_path = other ? other : name;
4369 if (o->prefix_length)
4371 strip_prefix(o->prefix_length, &name, &other);
4372 diff_fill_oid_info(p->one);
4374 diff_fill_oid_info(p->two);
4375 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4377}
4378void diff_setup(struct diff_options *options)
4380{
4381 memcpy(options, &default_diff_options, sizeof(*options));
4382 options->file = stdout;
4384 options->abbrev = DEFAULT_ABBREV;
4386 options->line_termination = '\n';
4387 options->break_opt = -1;
4388 options->rename_limit = -1;
4389 options->dirstat_permille = diff_dirstat_permille_default;
4390 options->context = diff_context_default;
4391 options->interhunkcontext = diff_interhunk_context_default;
4392 options->ws_error_highlight = ws_error_highlight_default;
4393 options->flags.rename_empty = 1;
4394 options->objfind = NULL;
4395 /* pathchange left =NULL by default */
4397 options->change = diff_change;
4398 options->add_remove = diff_addremove;
4399 options->use_color = diff_use_color_default;
4400 options->detect_rename = diff_detect_rename_default;
4401 options->xdl_opts |= diff_algorithm;
4402 if (diff_indent_heuristic)
4403 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4404 options->orderfile = diff_order_file_cfg;
4406 if (diff_no_prefix) {
4408 options->a_prefix = options->b_prefix = "";
4409 } else if (!diff_mnemonic_prefix) {
4410 options->a_prefix = "a/";
4411 options->b_prefix = "b/";
4412 }
4413 options->color_moved = diff_color_moved_default;
4415 options->color_moved_ws_handling = diff_color_moved_ws_default;
4416}
4417void diff_setup_done(struct diff_options *options)
4419{
4420 unsigned check_mask = DIFF_FORMAT_NAME |
4421 DIFF_FORMAT_NAME_STATUS |
4422 DIFF_FORMAT_CHECKDIFF |
4423 DIFF_FORMAT_NO_OUTPUT;
4424 /*
4425 * This must be signed because we're comparing against a potentially
4426 * negative value.
4427 */
4428 const int hexsz = the_hash_algo->hexsz;
4429 if (options->set_default)
4431 options->set_default(options);
4432 if (HAS_MULTI_BITS(options->output_format & check_mask))
4434 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4435 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4437 die(_("-G, -S and --find-object are mutually exclusive"));
4438 /*
4440 * Most of the time we can say "there are changes"
4441 * only by checking if there are changed paths, but
4442 * --ignore-whitespace* options force us to look
4443 * inside contents.
4444 */
4445 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS))
4447 options->flags.diff_from_contents = 1;
4448 else
4449 options->flags.diff_from_contents = 0;
4450 if (options->flags.find_copies_harder)
4452 options->detect_rename = DIFF_DETECT_COPY;
4453 if (!options->flags.relative_name)
4455 options->prefix = NULL;
4456 if (options->prefix)
4457 options->prefix_length = strlen(options->prefix);
4458 else
4459 options->prefix_length = 0;
4460 if (options->output_format & (DIFF_FORMAT_NAME |
4462 DIFF_FORMAT_NAME_STATUS |
4463 DIFF_FORMAT_CHECKDIFF |
4464 DIFF_FORMAT_NO_OUTPUT))
4465 options->output_format &= ~(DIFF_FORMAT_RAW |
4466 DIFF_FORMAT_NUMSTAT |
4467 DIFF_FORMAT_DIFFSTAT |
4468 DIFF_FORMAT_SHORTSTAT |
4469 DIFF_FORMAT_DIRSTAT |
4470 DIFF_FORMAT_SUMMARY |
4471 DIFF_FORMAT_PATCH);
4472 /*
4474 * These cases always need recursive; we do not drop caller-supplied
4475 * recursive bits for other formats here.
4476 */
4477 if (options->output_format & (DIFF_FORMAT_PATCH |
4478 DIFF_FORMAT_NUMSTAT |
4479 DIFF_FORMAT_DIFFSTAT |
4480 DIFF_FORMAT_SHORTSTAT |
4481 DIFF_FORMAT_DIRSTAT |
4482 DIFF_FORMAT_SUMMARY |
4483 DIFF_FORMAT_CHECKDIFF))
4484 options->flags.recursive = 1;
4485 /*
4486 * Also pickaxe would not work very well if you do not say recursive
4487 */
4488 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4489 options->flags.recursive = 1;
4490 /*
4491 * When patches are generated, submodules diffed against the work tree
4492 * must be checked for dirtiness too so it can be shown in the output
4493 */
4494 if (options->output_format & DIFF_FORMAT_PATCH)
4495 options->flags.dirty_submodules = 1;
4496 if (options->detect_rename && options->rename_limit < 0)
4498 options->rename_limit = diff_rename_limit_default;
4499 if (options->setup & DIFF_SETUP_USE_CACHE) {
4500 if (!active_cache)
4501 /* read-cache does not die even when it fails
4502 * so it is safe for us to do this here. Also
4503 * it does not smudge active_cache or active_nr
4504 * when it fails, so we do not have to worry about
4505 * cleaning it up ourselves either.
4506 */
4507 read_cache();
4508 }
4509 if (hexsz < options->abbrev)
4510 options->abbrev = hexsz; /* full */
4511 /*
4513 * It does not make sense to show the first hit we happened
4514 * to have found. It does not make sense not to return with
4515 * exit code in such a case either.
4516 */
4517 if (options->flags.quick) {
4518 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4519 options->flags.exit_with_status = 1;
4520 }
4521 options->diff_path_counter = 0;
4523 if (options->flags.follow_renames && options->pathspec.nr != 1)
4525 die(_("--follow requires exactly one pathspec"));
4526 if (!options->use_color || external_diff())
4528 options->color_moved = 0;
4529}
4530static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4532{
4533 char c, *eq;
4534 int len;
4535 if (*arg != '-')
4537 return 0;
4538 c = *++arg;
4539 if (!c)
4540 return 0;
4541 if (c == arg_short) {
4542 c = *++arg;
4543 if (!c)
4544 return 1;
4545 if (val && isdigit(c)) {
4546 char *end;
4547 int n = strtoul(arg, &end, 10);
4548 if (*end)
4549 return 0;
4550 *val = n;
4551 return 1;
4552 }
4553 return 0;
4554 }
4555 if (c != '-')
4556 return 0;
4557 arg++;
4558 eq = strchrnul(arg, '=');
4559 len = eq - arg;
4560 if (!len || strncmp(arg, arg_long, len))
4561 return 0;
4562 if (*eq) {
4563 int n;
4564 char *end;
4565 if (!isdigit(*++eq))
4566 return 0;
4567 n = strtoul(eq, &end, 10);
4568 if (*end)
4569 return 0;
4570 *val = n;
4571 }
4572 return 1;
4573}
4574static int diff_scoreopt_parse(const char *opt);
4576static inline int short_opt(char opt, const char **argv,
4578 const char **optarg)
4579{
4580 const char *arg = argv[0];
4581 if (arg[0] != '-' || arg[1] != opt)
4582 return 0;
4583 if (arg[2] != '\0') {
4584 *optarg = arg + 2;
4585 return 1;
4586 }
4587 if (!argv[1])
4588 die("Option '%c' requires a value", opt);
4589 *optarg = argv[1];
4590 return 2;
4591}
4592int parse_long_opt(const char *opt, const char **argv,
4594 const char **optarg)
4595{
4596 const char *arg = argv[0];
4597 if (!skip_prefix(arg, "--", &arg))
4598 return 0;
4599 if (!skip_prefix(arg, opt, &arg))
4600 return 0;
4601 if (*arg == '=') { /* stuck form: --option=value */
4602 *optarg = arg + 1;
4603 return 1;
4604 }
4605 if (*arg != '\0')
4606 return 0;
4607 /* separate form: --option value */
4608 if (!argv[1])
4609 die("Option '--%s' requires a value", opt);
4610 *optarg = argv[1];
4611 return 2;
4612}
4613static int stat_opt(struct diff_options *options, const char **av)
4615{
4616 const char *arg = av[0];
4617 char *end;
4618 int width = options->stat_width;
4619 int name_width = options->stat_name_width;
4620 int graph_width = options->stat_graph_width;
4621 int count = options->stat_count;
4622 int argcount = 1;
4623 if (!skip_prefix(arg, "--stat", &arg))
4625 BUG("stat option does not begin with --stat: %s", arg);
4626 end = (char *)arg;
4627 switch (*arg) {
4629 case '-':
4630 if (skip_prefix(arg, "-width", &arg)) {
4631 if (*arg == '=')
4632 width = strtoul(arg + 1, &end, 10);
4633 else if (!*arg && !av[1])
4634 die_want_option("--stat-width");
4635 else if (!*arg) {
4636 width = strtoul(av[1], &end, 10);
4637 argcount = 2;
4638 }
4639 } else if (skip_prefix(arg, "-name-width", &arg)) {
4640 if (*arg == '=')
4641 name_width = strtoul(arg + 1, &end, 10);
4642 else if (!*arg && !av[1])
4643 die_want_option("--stat-name-width");
4644 else if (!*arg) {
4645 name_width = strtoul(av[1], &end, 10);
4646 argcount = 2;
4647 }
4648 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4649 if (*arg == '=')
4650 graph_width = strtoul(arg + 1, &end, 10);
4651 else if (!*arg && !av[1])
4652 die_want_option("--stat-graph-width");
4653 else if (!*arg) {
4654 graph_width = strtoul(av[1], &end, 10);
4655 argcount = 2;
4656 }
4657 } else if (skip_prefix(arg, "-count", &arg)) {
4658 if (*arg == '=')
4659 count = strtoul(arg + 1, &end, 10);
4660 else if (!*arg && !av[1])
4661 die_want_option("--stat-count");
4662 else if (!*arg) {
4663 count = strtoul(av[1], &end, 10);
4664 argcount = 2;
4665 }
4666 }
4667 break;
4668 case '=':
4669 width = strtoul(arg+1, &end, 10);
4670 if (*end == ',')
4671 name_width = strtoul(end+1, &end, 10);
4672 if (*end == ',')
4673 count = strtoul(end+1, &end, 10);
4674 }
4675 /* Important! This checks all the error cases! */
4677 if (*end)
4678 return 0;
4679 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4680 options->stat_name_width = name_width;
4681 options->stat_graph_width = graph_width;
4682 options->stat_width = width;
4683 options->stat_count = count;
4684 return argcount;
4685}
4686static int parse_dirstat_opt(struct diff_options *options, const char *params)
4688{
4689 struct strbuf errmsg = STRBUF_INIT;
4690 if (parse_dirstat_params(options, params, &errmsg))
4691 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4692 errmsg.buf);
4693 strbuf_release(&errmsg);
4694 /*
4695 * The caller knows a dirstat-related option is given from the command
4696 * line; allow it to say "return this_function();"
4697 */
4698 options->output_format |= DIFF_FORMAT_DIRSTAT;
4699 return 1;
4700}
4701static int parse_submodule_opt(struct diff_options *options, const char *value)
4703{
4704 if (parse_submodule_params(options, value))
4705 die(_("Failed to parse --submodule option parameter: '%s'"),
4706 value);
4707 return 1;
4708}
4709static const char diff_status_letters[] = {
4711 DIFF_STATUS_ADDED,
4712 DIFF_STATUS_COPIED,
4713 DIFF_STATUS_DELETED,
4714 DIFF_STATUS_MODIFIED,
4715 DIFF_STATUS_RENAMED,
4716 DIFF_STATUS_TYPE_CHANGED,
4717 DIFF_STATUS_UNKNOWN,
4718 DIFF_STATUS_UNMERGED,
4719 DIFF_STATUS_FILTER_AON,
4720 DIFF_STATUS_FILTER_BROKEN,
4721 '\0',
4722};
4723static unsigned int filter_bit['Z' + 1];
4725static void prepare_filter_bits(void)
4727{
4728 int i;
4729 if (!filter_bit[DIFF_STATUS_ADDED]) {
4731 for (i = 0; diff_status_letters[i]; i++)
4732 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4733 }
4734}
4735static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4737{
4738 return opt->filter & filter_bit[(int) status];
4739}
4740static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4742{
4743 int i, optch;
4744 prepare_filter_bits();
4746 /*
4748 * If there is a negation e.g. 'd' in the input, and we haven't
4749 * initialized the filter field with another --diff-filter, start
4750 * from full set of bits, except for AON.
4751 */
4752 if (!opt->filter) {
4753 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4754 if (optch < 'a' || 'z' < optch)
4755 continue;
4756 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4757 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4758 break;
4759 }
4760 }
4761 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4763 unsigned int bit;
4764 int negate;
4765 if ('a' <= optch && optch <= 'z') {
4767 negate = 1;
4768 optch = toupper(optch);
4769 } else {
4770 negate = 0;
4771 }
4772 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4774 if (!bit)
4775 return optarg[i];
4776 if (negate)
4777 opt->filter &= ~bit;
4778 else
4779 opt->filter |= bit;
4780 }
4781 return 0;
4782}
4783static void enable_patch_output(int *fmt) {
4785 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4786 *fmt |= DIFF_FORMAT_PATCH;
4787}
4788static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4790{
4791 int val = parse_ws_error_highlight(arg);
4792 if (val < 0) {
4794 error("unknown value after ws-error-highlight=%.*s",
4795 -1 - val, arg);
4796 return 0;
4797 }
4798 opt->ws_error_highlight = val;
4799 return 1;
4800}
4801static int parse_objfind_opt(struct diff_options *opt, const char *arg)
4803{
4804 struct object_id oid;
4805 if (get_oid(arg, &oid))
4807 return error("unable to resolve '%s'", arg);
4808 if (!opt->objfind)
4810 opt->objfind = xcalloc(1, sizeof(*opt->objfind));
4811 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4813 opt->flags.recursive = 1;
4814 opt->flags.tree_in_recursive = 1;
4815 oidset_insert(opt->objfind, &oid);
4816 return 1;
4817}
4818int diff_opt_parse(struct diff_options *options,
4820 const char **av, int ac, const char *prefix)
4821{
4822 const char *arg = av[0];
4823 const char *optarg;
4824 int argcount;
4825 if (!prefix)
4827 prefix = "";
4828 /* Output format options */
4830 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4831 || opt_arg(arg, 'U', "unified", &options->context))
4832 enable_patch_output(&options->output_format);
4833 else if (!strcmp(arg, "--raw"))
4834 options->output_format |= DIFF_FORMAT_RAW;
4835 else if (!strcmp(arg, "--patch-with-raw")) {
4836 enable_patch_output(&options->output_format);
4837 options->output_format |= DIFF_FORMAT_RAW;
4838 } else if (!strcmp(arg, "--numstat"))
4839 options->output_format |= DIFF_FORMAT_NUMSTAT;
4840 else if (!strcmp(arg, "--shortstat"))
4841 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4842 else if (skip_prefix(arg, "-X", &arg) ||
4843 skip_to_optional_arg(arg, "--dirstat", &arg))
4844 return parse_dirstat_opt(options, arg);
4845 else if (!strcmp(arg, "--cumulative"))
4846 return parse_dirstat_opt(options, "cumulative");
4847 else if (skip_to_optional_arg(arg, "--dirstat-by-file", &arg)) {
4848 parse_dirstat_opt(options, "files");
4849 return parse_dirstat_opt(options, arg);
4850 }
4851 else if (!strcmp(arg, "--check"))
4852 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4853 else if (!strcmp(arg, "--summary"))
4854 options->output_format |= DIFF_FORMAT_SUMMARY;
4855 else if (!strcmp(arg, "--patch-with-stat")) {
4856 enable_patch_output(&options->output_format);
4857 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4858 } else if (!strcmp(arg, "--name-only"))
4859 options->output_format |= DIFF_FORMAT_NAME;
4860 else if (!strcmp(arg, "--name-status"))
4861 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4862 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4863 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4864 else if (starts_with(arg, "--stat"))
4865 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4866 return stat_opt(options, av);
4867 else if (!strcmp(arg, "--compact-summary")) {
4868 options->flags.stat_with_summary = 1;
4869 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4870 } else if (!strcmp(arg, "--no-compact-summary"))
4871 options->flags.stat_with_summary = 0;
4872 /* renames options */
4874 else if (starts_with(arg, "-B") ||
4875 skip_to_optional_arg(arg, "--break-rewrites", NULL)) {
4876 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4877 return error("invalid argument to -B: %s", arg+2);
4878 }
4879 else if (starts_with(arg, "-M") ||
4880 skip_to_optional_arg(arg, "--find-renames", NULL)) {
4881 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4882 return error("invalid argument to -M: %s", arg+2);
4883 options->detect_rename = DIFF_DETECT_RENAME;
4884 }
4885 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4886 options->irreversible_delete = 1;
4887 }
4888 else if (starts_with(arg, "-C") ||
4889 skip_to_optional_arg(arg, "--find-copies", NULL)) {
4890 if (options->detect_rename == DIFF_DETECT_COPY)
4891 options->flags.find_copies_harder = 1;
4892 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4893 return error("invalid argument to -C: %s", arg+2);
4894 options->detect_rename = DIFF_DETECT_COPY;
4895 }
4896 else if (!strcmp(arg, "--no-renames"))
4897 options->detect_rename = 0;
4898 else if (!strcmp(arg, "--rename-empty"))
4899 options->flags.rename_empty = 1;
4900 else if (!strcmp(arg, "--no-rename-empty"))
4901 options->flags.rename_empty = 0;
4902 else if (skip_to_optional_arg_default(arg, "--relative", &arg, NULL)) {
4903 options->flags.relative_name = 1;
4904 if (arg)
4905 options->prefix = arg;
4906 }
4907 /* xdiff options */
4909 else if (!strcmp(arg, "--minimal"))
4910 DIFF_XDL_SET(options, NEED_MINIMAL);
4911 else if (!strcmp(arg, "--no-minimal"))
4912 DIFF_XDL_CLR(options, NEED_MINIMAL);
4913 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4914 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4915 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4916 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4917 else if (!strcmp(arg, "--ignore-space-at-eol"))
4918 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4919 else if (!strcmp(arg, "--ignore-cr-at-eol"))
4920 DIFF_XDL_SET(options, IGNORE_CR_AT_EOL);
4921 else if (!strcmp(arg, "--ignore-blank-lines"))
4922 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4923 else if (!strcmp(arg, "--indent-heuristic"))
4924 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4925 else if (!strcmp(arg, "--no-indent-heuristic"))
4926 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4927 else if (!strcmp(arg, "--patience")) {
4928 int i;
4929 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4930 /*
4931 * Both --patience and --anchored use PATIENCE_DIFF
4932 * internally, so remove any anchors previously
4933 * specified.
4934 */
4935 for (i = 0; i < options->anchors_nr; i++)
4936 free(options->anchors[i]);
4937 options->anchors_nr = 0;
4938 } else if (!strcmp(arg, "--histogram"))
4939 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4940 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4941 long value = parse_algorithm_value(optarg);
4942 if (value < 0)
4943 return error("option diff-algorithm accepts \"myers\", "
4944 "\"minimal\", \"patience\" and \"histogram\"");
4945 /* clear out previous settings */
4946 DIFF_XDL_CLR(options, NEED_MINIMAL);
4947 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4948 options->xdl_opts |= value;
4949 return argcount;
4950 } else if (skip_prefix(arg, "--anchored=", &arg)) {
4951 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4952 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
4953 options->anchors_alloc);
4954 options->anchors[options->anchors_nr++] = xstrdup(arg);
4955 }
4956 /* flags options */
4958 else if (!strcmp(arg, "--binary")) {
4959 enable_patch_output(&options->output_format);
4960 options->flags.binary = 1;
4961 }
4962 else if (!strcmp(arg, "--full-index"))
4963 options->flags.full_index = 1;
4964 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4965 options->flags.text = 1;
4966 else if (!strcmp(arg, "-R"))
4967 options->flags.reverse_diff = 1;
4968 else if (!strcmp(arg, "--find-copies-harder"))
4969 options->flags.find_copies_harder = 1;
4970 else if (!strcmp(arg, "--follow"))
4971 options->flags.follow_renames = 1;
4972 else if (!strcmp(arg, "--no-follow")) {
4973 options->flags.follow_renames = 0;
4974 options->flags.default_follow_renames = 0;
4975 } else if (skip_to_optional_arg_default(arg, "--color", &arg, "always")) {
4976 int value = git_config_colorbool(NULL, arg);
4977 if (value < 0)
4978 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4979 options->use_color = value;
4980 }
4981 else if (!strcmp(arg, "--no-color"))
4982 options->use_color = 0;
4983 else if (!strcmp(arg, "--color-moved")) {
4984 if (diff_color_moved_default)
4985 options->color_moved = diff_color_moved_default;
4986 if (options->color_moved == COLOR_MOVED_NO)
4987 options->color_moved = COLOR_MOVED_DEFAULT;
4988 } else if (!strcmp(arg, "--no-color-moved"))
4989 options->color_moved = COLOR_MOVED_NO;
4990 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4991 int cm = parse_color_moved(arg);
4992 if (cm < 0)
4993 die("bad --color-moved argument: %s", arg);
4994 options->color_moved = cm;
4995 } else if (skip_prefix(arg, "--color-moved-ws=", &arg)) {
4996 options->color_moved_ws_handling = parse_color_moved_ws(arg);
4997 } else if (skip_to_optional_arg_default(arg, "--color-words", &options->word_regex, NULL)) {
4998 options->use_color = 1;
4999 options->word_diff = DIFF_WORDS_COLOR;
5000 }
5001 else if (!strcmp(arg, "--word-diff")) {
5002 if (options->word_diff == DIFF_WORDS_NONE)
5003 options->word_diff = DIFF_WORDS_PLAIN;
5004 }
5005 else if (skip_prefix(arg, "--word-diff=", &arg)) {
5006 if (!strcmp(arg, "plain"))
5007 options->word_diff = DIFF_WORDS_PLAIN;
5008 else if (!strcmp(arg, "color")) {
5009 options->use_color = 1;
5010 options->word_diff = DIFF_WORDS_COLOR;
5011 }
5012 else if (!strcmp(arg, "porcelain"))
5013 options->word_diff = DIFF_WORDS_PORCELAIN;
5014 else if (!strcmp(arg, "none"))
5015 options->word_diff = DIFF_WORDS_NONE;
5016 else
5017 die("bad --word-diff argument: %s", arg);
5018 }
5019 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
5020 if (options->word_diff == DIFF_WORDS_NONE)
5021 options->word_diff = DIFF_WORDS_PLAIN;
5022 options->word_regex = optarg;
5023 return argcount;
5024 }
5025 else if (!strcmp(arg, "--exit-code"))
5026 options->flags.exit_with_status = 1;
5027 else if (!strcmp(arg, "--quiet"))
5028 options->flags.quick = 1;
5029 else if (!strcmp(arg, "--ext-diff"))
5030 options->flags.allow_external = 1;
5031 else if (!strcmp(arg, "--no-ext-diff"))
5032 options->flags.allow_external = 0;
5033 else if (!strcmp(arg, "--textconv")) {
5034 options->flags.allow_textconv = 1;
5035 options->flags.textconv_set_via_cmdline = 1;
5036 } else if (!strcmp(arg, "--no-textconv"))
5037 options->flags.allow_textconv = 0;
5038 else if (skip_to_optional_arg_default(arg, "--ignore-submodules", &arg, "all")) {
5039 options->flags.override_submodule_config = 1;
5040 handle_ignore_submodules_arg(options, arg);
5041 } else if (skip_to_optional_arg_default(arg, "--submodule", &arg, "log"))
5042 return parse_submodule_opt(options, arg);
5043 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
5044 return parse_ws_error_highlight_opt(options, arg);
5045 else if (!strcmp(arg, "--ita-invisible-in-index"))
5046 options->ita_invisible_in_index = 1;
5047 else if (!strcmp(arg, "--ita-visible-in-index"))
5048 options->ita_invisible_in_index = 0;
5049 /* misc options */
5051 else if (!strcmp(arg, "-z"))
5052 options->line_termination = 0;
5053 else if ((argcount = short_opt('l', av, &optarg))) {
5054 options->rename_limit = strtoul(optarg, NULL, 10);
5055 return argcount;
5056 }
5057 else if ((argcount = short_opt('S', av, &optarg))) {
5058 options->pickaxe = optarg;
5059 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5060 return argcount;
5061 } else if ((argcount = short_opt('G', av, &optarg))) {
5062 options->pickaxe = optarg;
5063 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5064 return argcount;
5065 }
5066 else if (!strcmp(arg, "--pickaxe-all"))
5067 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
5068 else if (!strcmp(arg, "--pickaxe-regex"))
5069 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
5070 else if ((argcount = short_opt('O', av, &optarg))) {
5071 options->orderfile = prefix_filename(prefix, optarg);
5072 return argcount;
5073 } else if (skip_prefix(arg, "--find-object=", &arg))
5074 return parse_objfind_opt(options, arg);
5075 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
5076 int offending = parse_diff_filter_opt(optarg, options);
5077 if (offending)
5078 die("unknown change class '%c' in --diff-filter=%s",
5079 offending, optarg);
5080 return argcount;
5081 }
5082 else if (!strcmp(arg, "--no-abbrev"))
5083 options->abbrev = 0;
5084 else if (!strcmp(arg, "--abbrev"))
5085 options->abbrev = DEFAULT_ABBREV;
5086 else if (skip_prefix(arg, "--abbrev=", &arg)) {
5087 options->abbrev = strtoul(arg, NULL, 10);
5088 if (options->abbrev < MINIMUM_ABBREV)
5089 options->abbrev = MINIMUM_ABBREV;
5090 else if (the_hash_algo->hexsz < options->abbrev)
5091 options->abbrev = the_hash_algo->hexsz;
5092 }
5093 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
5094 options->a_prefix = optarg;
5095 return argcount;
5096 }
5097 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
5098 options->line_prefix = optarg;
5099 options->line_prefix_length = strlen(options->line_prefix);
5100 graph_setup_line_prefix(options);
5101 return argcount;
5102 }
5103 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
5104 options->b_prefix = optarg;
5105 return argcount;
5106 }
5107 else if (!strcmp(arg, "--no-prefix"))
5108 options->a_prefix = options->b_prefix = "";
5109 else if (opt_arg(arg, '\0', "inter-hunk-context",
5110 &options->interhunkcontext))
5111 ;
5112 else if (!strcmp(arg, "-W"))
5113 options->flags.funccontext = 1;
5114 else if (!strcmp(arg, "--function-context"))
5115 options->flags.funccontext = 1;
5116 else if (!strcmp(arg, "--no-function-context"))
5117 options->flags.funccontext = 0;
5118 else if ((argcount = parse_long_opt("output", av, &optarg))) {
5119 char *path = prefix_filename(prefix, optarg);
5120 options->file = xfopen(path, "w");
5121 options->close_file = 1;
5122 if (options->use_color != GIT_COLOR_ALWAYS)
5123 options->use_color = GIT_COLOR_NEVER;
5124 free(path);
5125 return argcount;
5126 } else
5127 return 0;
5128 return 1;
5129}
5130int parse_rename_score(const char **cp_p)
5132{
5133 unsigned long num, scale;
5134 int ch, dot;
5135 const char *cp = *cp_p;
5136 num = 0;
5138 scale = 1;
5139 dot = 0;
5140 for (;;) {
5141 ch = *cp;
5142 if ( !dot && ch == '.' ) {
5143 scale = 1;
5144 dot = 1;
5145 } else if ( ch == '%' ) {
5146 scale = dot ? scale*100 : 100;
5147 cp++; /* % is always at the end */
5148 break;
5149 } else if ( ch >= '0' && ch <= '9' ) {
5150 if ( scale < 100000 ) {
5151 scale *= 10;
5152 num = (num*10) + (ch-'0');
5153 }
5154 } else {
5155 break;
5156 }
5157 cp++;
5158 }
5159 *cp_p = cp;
5160 /* user says num divided by scale and we say internally that
5162 * is MAX_SCORE * num / scale.
5163 */
5164 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5165}
5166static int diff_scoreopt_parse(const char *opt)
5168{
5169 int opt1, opt2, cmd;
5170 if (*opt++ != '-')
5172 return -1;
5173 cmd = *opt++;
5174 if (cmd == '-') {
5175 /* convert the long-form arguments into short-form versions */
5176 if (skip_prefix(opt, "break-rewrites", &opt)) {
5177 if (*opt == 0 || *opt++ == '=')
5178 cmd = 'B';
5179 } else if (skip_prefix(opt, "find-copies", &opt)) {
5180 if (*opt == 0 || *opt++ == '=')
5181 cmd = 'C';
5182 } else if (skip_prefix(opt, "find-renames", &opt)) {
5183 if (*opt == 0 || *opt++ == '=')
5184 cmd = 'M';
5185 }
5186 }
5187 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
5188 return -1; /* that is not a -M, -C, or -B option */
5189 opt1 = parse_rename_score(&opt);
5191 if (cmd != 'B')
5192 opt2 = 0;
5193 else {
5194 if (*opt == 0)
5195 opt2 = 0;
5196 else if (*opt != '/')
5197 return -1; /* we expect -B80/99 or -B80 */
5198 else {
5199 opt++;
5200 opt2 = parse_rename_score(&opt);
5201 }
5202 }
5203 if (*opt != 0)
5204 return -1;
5205 return opt1 | (opt2 << 16);
5206}
5207struct diff_queue_struct diff_queued_diff;
5209void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5211{
5212 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5213 queue->queue[queue->nr++] = dp;
5214}
5215struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5217 struct diff_filespec *one,
5218 struct diff_filespec *two)
5219{
5220 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5221 dp->one = one;
5222 dp->two = two;
5223 if (queue)
5224 diff_q(queue, dp);
5225 return dp;
5226}
5227void diff_free_filepair(struct diff_filepair *p)
5229{
5230 free_filespec(p->one);
5231 free_filespec(p->two);
5232 free(p);
5233}
5234const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5236{
5237 int abblen;
5238 const char *abbrev;
5239 /* Do we want all 40 hex characters? */
5241 if (len == the_hash_algo->hexsz)
5242 return oid_to_hex(oid);
5243 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5245 abbrev = diff_abbrev_oid(oid, len);
5246 if (!print_sha1_ellipsis())
5248 return abbrev;
5249 abblen = strlen(abbrev);
5251 /*
5253 * In well-behaved cases, where the abbreviated result is the
5254 * same as the requested length, append three dots after the
5255 * abbreviation (hence the whole logic is limited to the case
5256 * where abblen < 37); when the actual abbreviated result is a
5257 * bit longer than the requested length, we reduce the number
5258 * of dots so that they match the well-behaved ones. However,
5259 * if the actual abbreviation is longer than the requested
5260 * length by more than three, we give up on aligning, and add
5261 * three dots anyway, to indicate that the output is not the
5262 * full object name. Yes, this may be suboptimal, but this
5263 * appears only in "diff --raw --abbrev" output and it is not
5264 * worth the effort to change it now. Note that this would
5265 * likely to work fine when the automatic sizing of default
5266 * abbreviation length is used--we would be fed -1 in "len" in
5267 * that case, and will end up always appending three-dots, but
5268 * the automatic sizing is supposed to give abblen that ensures
5269 * uniqueness across all objects (statistically speaking).
5270 */
5271 if (abblen < the_hash_algo->hexsz - 3) {
5272 static char hex[GIT_MAX_HEXSZ + 1];
5273 if (len < abblen && abblen <= len + 2)
5274 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5275 else
5276 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5277 return hex;
5278 }
5279 return oid_to_hex(oid);
5281}
5282static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5284{
5285 int line_termination = opt->line_termination;
5286 int inter_name_termination = line_termination ? '\t' : '\0';
5287 fprintf(opt->file, "%s", diff_line_prefix(opt));
5289 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5290 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5291 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5292 fprintf(opt->file, "%s ",
5293 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5294 }
5295 if (p->score) {
5296 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5297 inter_name_termination);
5298 } else {
5299 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5300 }
5301 if (p->status == DIFF_STATUS_COPIED ||
5303 p->status == DIFF_STATUS_RENAMED) {
5304 const char *name_a, *name_b;
5305 name_a = p->one->path;
5306 name_b = p->two->path;
5307 strip_prefix(opt->prefix_length, &name_a, &name_b);
5308 write_name_quoted(name_a, opt->file, inter_name_termination);
5309 write_name_quoted(name_b, opt->file, line_termination);
5310 } else {
5311 const char *name_a, *name_b;
5312 name_a = p->one->mode ? p->one->path : p->two->path;
5313 name_b = NULL;
5314 strip_prefix(opt->prefix_length, &name_a, &name_b);
5315 write_name_quoted(name_a, opt->file, line_termination);
5316 }
5317}
5318int diff_unmodified_pair(struct diff_filepair *p)
5320{
5321 /* This function is written stricter than necessary to support
5322 * the currently implemented transformers, but the idea is to
5323 * let transformers to produce diff_filepairs any way they want,
5324 * and filter and clean them up here before producing the output.
5325 */
5326 struct diff_filespec *one = p->one, *two = p->two;
5327 if (DIFF_PAIR_UNMERGED(p))
5329 return 0; /* unmerged is interesting */
5330 /* deletion, addition, mode or type change
5332 * and rename are all interesting.
5333 */
5334 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5335 DIFF_PAIR_MODE_CHANGED(p) ||
5336 strcmp(one->path, two->path))
5337 return 0;
5338 /* both are valid and point at the same path. that is, we are
5340 * dealing with a change.
5341 */
5342 if (one->oid_valid && two->oid_valid &&
5343 !oidcmp(&one->oid, &two->oid) &&
5344 !one->dirty_submodule && !two->dirty_submodule)
5345 return 1; /* no change */
5346 if (!one->oid_valid && !two->oid_valid)
5347 return 1; /* both look at the same file on the filesystem. */
5348 return 0;
5349}
5350static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5352{
5353 if (diff_unmodified_pair(p))
5354 return;
5355 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5357 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5358 return; /* no tree diffs in patch format */
5359 run_diff(p, o);
5361}
5362static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5364 struct diffstat_t *diffstat)
5365{
5366 if (diff_unmodified_pair(p))
5367 return;
5368 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5370 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5371 return; /* no useful stat for tree diffs */
5372 run_diffstat(p, o, diffstat);
5374}
5375static void diff_flush_checkdiff(struct diff_filepair *p,
5377 struct diff_options *o)
5378{
5379 if (diff_unmodified_pair(p))
5380 return;
5381 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5383 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5384 return; /* nothing to check in tree diffs */
5385 run_checkdiff(p, o);
5387}
5388int diff_queue_is_empty(void)
5390{
5391 struct diff_queue_struct *q = &diff_queued_diff;
5392 int i;
5393 for (i = 0; i < q->nr; i++)
5394 if (!diff_unmodified_pair(q->queue[i]))
5395 return 0;
5396 return 1;
5397}
5398#if DIFF_DEBUG
5400void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5401{
5402 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5403 x, one ? one : "",
5404 s->path,
5405 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5406 s->mode,
5407 s->oid_valid ? oid_to_hex(&s->oid) : "");
5408 fprintf(stderr, "queue[%d] %s size %lu\n",
5409 x, one ? one : "",
5410 s->size);
5411}
5412void diff_debug_filepair(const struct diff_filepair *p, int i)
5414{
5415 diff_debug_filespec(p->one, i, "one");
5416 diff_debug_filespec(p->two, i, "two");
5417 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5418 p->score, p->status ? p->status : '?',
5419 p->one->rename_used, p->broken_pair);
5420}
5421void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5423{
5424 int i;
5425 if (msg)
5426 fprintf(stderr, "%s\n", msg);
5427 fprintf(stderr, "q->nr = %d\n", q->nr);
5428 for (i = 0; i < q->nr; i++) {
5429 struct diff_filepair *p = q->queue[i];
5430 diff_debug_filepair(p, i);
5431 }
5432}
5433#endif
5434static void diff_resolve_rename_copy(void)
5436{
5437 int i;
5438 struct diff_filepair *p;
5439 struct diff_queue_struct *q = &diff_queued_diff;
5440 diff_debug_queue("resolve-rename-copy", q);
5442 for (i = 0; i < q->nr; i++) {
5444 p = q->queue[i];
5445 p->status = 0; /* undecided */
5446 if (DIFF_PAIR_UNMERGED(p))
5447 p->status = DIFF_STATUS_UNMERGED;
5448 else if (!DIFF_FILE_VALID(p->one))
5449 p->status = DIFF_STATUS_ADDED;
5450 else if (!DIFF_FILE_VALID(p->two))
5451 p->status = DIFF_STATUS_DELETED;
5452 else if (DIFF_PAIR_TYPE_CHANGED(p))
5453 p->status = DIFF_STATUS_TYPE_CHANGED;
5454 /* from this point on, we are dealing with a pair
5456 * whose both sides are valid and of the same type, i.e.
5457 * either in-place edit or rename/copy edit.
5458 */
5459 else if (DIFF_PAIR_RENAME(p)) {
5460 /*
5461 * A rename might have re-connected a broken
5462 * pair up, causing the pathnames to be the
5463 * same again. If so, that's not a rename at
5464 * all, just a modification..
5465 *
5466 * Otherwise, see if this source was used for
5467 * multiple renames, in which case we decrement
5468 * the count, and call it a copy.
5469 */
5470 if (!strcmp(p->one->path, p->two->path))
5471 p->status = DIFF_STATUS_MODIFIED;
5472 else if (--p->one->rename_used > 0)
5473 p->status = DIFF_STATUS_COPIED;
5474 else
5475 p->status = DIFF_STATUS_RENAMED;
5476 }
5477 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5478 p->one->mode != p->two->mode ||
5479 p->one->dirty_submodule ||
5480 p->two->dirty_submodule ||
5481 is_null_oid(&p->one->oid))
5482 p->status = DIFF_STATUS_MODIFIED;
5483 else {
5484 /* This is a "no-change" entry and should not
5485 * happen anymore, but prepare for broken callers.
5486 */
5487 error("feeding unmodified %s to diffcore",
5488 p->one->path);
5489 p->status = DIFF_STATUS_UNKNOWN;
5490 }
5491 }
5492 diff_debug_queue("resolve-rename-copy done", q);
5493}
5494static int check_pair_status(struct diff_filepair *p)
5496{
5497 switch (p->status) {
5498 case DIFF_STATUS_UNKNOWN:
5499 return 0;
5500 case 0:
5501 die("internal error in diff-resolve-rename-copy");
5502 default:
5503 return 1;
5504 }
5505}
5506static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5508{
5509 int fmt = opt->output_format;
5510 if (fmt & DIFF_FORMAT_CHECKDIFF)
5512 diff_flush_checkdiff(p, opt);
5513 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5514 diff_flush_raw(p, opt);
5515 else if (fmt & DIFF_FORMAT_NAME) {
5516 const char *name_a, *name_b;
5517 name_a = p->two->path;
5518 name_b = NULL;
5519 strip_prefix(opt->prefix_length, &name_a, &name_b);
5520 fprintf(opt->file, "%s", diff_line_prefix(opt));
5521 write_name_quoted(name_a, opt->file, opt->line_termination);
5522 }
5523}
5524static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5526{
5527 struct strbuf sb = STRBUF_INIT;
5528 if (fs->mode)
5529 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5530 else
5531 strbuf_addf(&sb, " %s ", newdelete);
5532 quote_c_style(fs->path, &sb, NULL, 0);
5534 strbuf_addch(&sb, '\n');
5535 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5536 sb.buf, sb.len, 0);
5537 strbuf_release(&sb);
5538}
5539static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5541 int show_name)
5542{
5543 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5544 struct strbuf sb = STRBUF_INIT;
5545 strbuf_addf(&sb, " mode change %06o => %06o",
5546 p->one->mode, p->two->mode);
5547 if (show_name) {
5548 strbuf_addch(&sb, ' ');
5549 quote_c_style(p->two->path, &sb, NULL, 0);
5550 }
5551 strbuf_addch(&sb, '\n');
5552 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5553 sb.buf, sb.len, 0);
5554 strbuf_release(&sb);
5555 }
5556}
5557static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5559 struct diff_filepair *p)
5560{
5561 struct strbuf sb = STRBUF_INIT;
5562 struct strbuf names = STRBUF_INIT;
5563 pprint_rename(&names, p->one->path, p->two->path);
5565 strbuf_addf(&sb, " %s %s (%d%%)\n",
5566 renamecopy, names.buf, similarity_index(p));
5567 strbuf_release(&names);
5568 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5569 sb.buf, sb.len, 0);
5570 show_mode_change(opt, p, 0);
5571 strbuf_release(&sb);
5572}
5573static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5575{
5576 switch(p->status) {
5577 case DIFF_STATUS_DELETED:
5578 show_file_mode_name(opt, "delete", p->one);
5579 break;
5580 case DIFF_STATUS_ADDED:
5581 show_file_mode_name(opt, "create", p->two);
5582 break;
5583 case DIFF_STATUS_COPIED:
5584 show_rename_copy(opt, "copy", p);
5585 break;
5586 case DIFF_STATUS_RENAMED:
5587 show_rename_copy(opt, "rename", p);
5588 break;
5589 default:
5590 if (p->score) {
5591 struct strbuf sb = STRBUF_INIT;
5592 strbuf_addstr(&sb, " rewrite ");
5593 quote_c_style(p->two->path, &sb, NULL, 0);
5594 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5595 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5596 sb.buf, sb.len, 0);
5597 strbuf_release(&sb);
5598 }
5599 show_mode_change(opt, p, !p->score);
5600 break;
5601 }
5602}
5603struct patch_id_t {
5605 git_SHA_CTX *ctx;
5606 int patchlen;
5607};
5608static int remove_space(char *line, int len)
5610{
5611 int i;
5612 char *dst = line;
5613 unsigned char c;
5614 for (i = 0; i < len; i++)
5616 if (!isspace((c = line[i])))
5617 *dst++ = c;
5618 return dst - line;
5620}
5621static void patch_id_consume(void *priv, char *line, unsigned long len)
5623{
5624 struct patch_id_t *data = priv;
5625 int new_len;
5626 /* Ignore line numbers when computing the SHA1 of the patch */
5628 if (starts_with(line, "@@ -"))
5629 return;
5630 new_len = remove_space(line, len);
5632 git_SHA1_Update(data->ctx, line, new_len);
5634 data->patchlen += new_len;
5635}
5636static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5638{
5639 git_SHA1_Update(ctx, str, strlen(str));
5640}
5641static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5643{
5644 /* large enough for 2^32 in octal */
5645 char buf[12];
5646 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5647 git_SHA1_Update(ctx, buf, len);
5648}
5649/* returns 0 upon success, and writes result into sha1 */
5651static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5652{
5653 struct diff_queue_struct *q = &diff_queued_diff;
5654 int i;
5655 git_SHA_CTX ctx;
5656 struct patch_id_t data;
5657 git_SHA1_Init(&ctx);
5659 memset(&data, 0, sizeof(struct patch_id_t));
5660 data.ctx = &ctx;
5661 for (i = 0; i < q->nr; i++) {
5663 xpparam_t xpp;
5664 xdemitconf_t xecfg;
5665 mmfile_t mf1, mf2;
5666 struct diff_filepair *p = q->queue[i];
5667 int len1, len2;
5668 memset(&xpp, 0, sizeof(xpp));
5670 memset(&xecfg, 0, sizeof(xecfg));
5671 if (p->status == 0)
5672 return error("internal diff status error");
5673 if (p->status == DIFF_STATUS_UNKNOWN)
5674 continue;
5675 if (diff_unmodified_pair(p))
5676 continue;
5677 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5678 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5679 continue;
5680 if (DIFF_PAIR_UNMERGED(p))
5681 continue;
5682 diff_fill_oid_info(p->one);
5684 diff_fill_oid_info(p->two);
5685 len1 = remove_space(p->one->path, strlen(p->one->path));
5687 len2 = remove_space(p->two->path, strlen(p->two->path));
5688 patch_id_add_string(&ctx, "diff--git");
5689 patch_id_add_string(&ctx, "a/");
5690 git_SHA1_Update(&ctx, p->one->path, len1);
5691 patch_id_add_string(&ctx, "b/");
5692 git_SHA1_Update(&ctx, p->two->path, len2);
5693 if (p->one->mode == 0) {
5695 patch_id_add_string(&ctx, "newfilemode");
5696 patch_id_add_mode(&ctx, p->two->mode);
5697 patch_id_add_string(&ctx, "---/dev/null");
5698 patch_id_add_string(&ctx, "+++b/");
5699 git_SHA1_Update(&ctx, p->two->path, len2);
5700 } else if (p->two->mode == 0) {
5701 patch_id_add_string(&ctx, "deletedfilemode");
5702 patch_id_add_mode(&ctx, p->one->mode);
5703 patch_id_add_string(&ctx, "---a/");
5704 git_SHA1_Update(&ctx, p->one->path, len1);
5705 patch_id_add_string(&ctx, "+++/dev/null");
5706 } else {
5707 patch_id_add_string(&ctx, "---a/");
5708 git_SHA1_Update(&ctx, p->one->path, len1);
5709 patch_id_add_string(&ctx, "+++b/");
5710 git_SHA1_Update(&ctx, p->two->path, len2);
5711 }
5712 if (diff_header_only)
5714 continue;
5715 if (fill_mmfile(&mf1, p->one) < 0 ||
5717 fill_mmfile(&mf2, p->two) < 0)
5718 return error("unable to read files to diff");
5719 if (diff_filespec_is_binary(p->one) ||
5721 diff_filespec_is_binary(p->two)) {
5722 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5723 GIT_SHA1_HEXSZ);
5724 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5725 GIT_SHA1_HEXSZ);
5726 continue;
5727 }
5728 xpp.flags = 0;
5730 xecfg.ctxlen = 3;
5731 xecfg.flags = 0;
5732 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
5733 &xpp, &xecfg))
5734 return error("unable to generate patch-id diff for %s",
5735 p->one->path);
5736 }
5737 git_SHA1_Final(oid->hash, &ctx);
5739 return 0;
5740}
5741int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5743{
5744 struct diff_queue_struct *q = &diff_queued_diff;
5745 int i;
5746 int result = diff_get_patch_id(options, oid, diff_header_only);
5747 for (i = 0; i < q->nr; i++)
5749 diff_free_filepair(q->queue[i]);
5750 free(q->queue);
5752 DIFF_QUEUE_CLEAR(q);
5753 return result;
5755}
5756static int is_summary_empty(const struct diff_queue_struct *q)
5758{
5759 int i;
5760 for (i = 0; i < q->nr; i++) {
5762 const struct diff_filepair *p = q->queue[i];
5763 switch (p->status) {
5765 case DIFF_STATUS_DELETED:
5766 case DIFF_STATUS_ADDED:
5767 case DIFF_STATUS_COPIED:
5768 case DIFF_STATUS_RENAMED:
5769 return 0;
5770 default:
5771 if (p->score)
5772 return 0;
5773 if (p->one->mode && p->two->mode &&
5774 p->one->mode != p->two->mode)
5775 return 0;
5776 break;
5777 }
5778 }
5779 return 1;
5780}
5781static const char rename_limit_warning[] =
5783N_("inexact rename detection was skipped due to too many files.");
5784static const char degrade_cc_to_c_warning[] =
5786N_("only found copies from modified paths due to too many files.");
5787static const char rename_limit_advice[] =
5789N_("you may want to set your %s variable to at least "
5790 "%d and retry the command.");
5791void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5793{
5794 fflush(stdout);
5795 if (degraded_cc)
5796 warning(_(degrade_cc_to_c_warning));
5797 else if (needed)
5798 warning(_(rename_limit_warning));
5799 else
5800 return;
5801 if (0 < needed)
5802 warning(_(rename_limit_advice), varname, needed);
5803}
5804static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5806{
5807 int i;
5808 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5809 struct diff_queue_struct *q = &diff_queued_diff;
5810 if (WSEH_NEW & WS_RULE_MASK)
5812 BUG("WS rules bit mask overlaps with diff symbol flags");
5813 if (o->color_moved)
5815 o->emitted_symbols = &esm;
5816 for (i = 0; i < q->nr; i++) {
5818 struct diff_filepair *p = q->queue[i];
5819 if (check_pair_status(p))
5820 diff_flush_patch(p, o);
5821 }
5822 if (o->emitted_symbols) {
5824 if (o->color_moved) {
5825 struct hashmap add_lines, del_lines;
5826 if (o->color_moved_ws_handling &
5828 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
5829 o->color_moved_ws_handling |= XDF_IGNORE_WHITESPACE;
5830 hashmap_init(&del_lines, moved_entry_cmp, o, 0);
5832 hashmap_init(&add_lines, moved_entry_cmp, o, 0);
5833 add_lines_to_move_detection(o, &add_lines, &del_lines);
5835 mark_color_as_moved(o, &add_lines, &del_lines);
5836 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5837 dim_moved_lines(o);
5838 hashmap_free(&add_lines, 0);
5840 hashmap_free(&del_lines, 0);
5841 }
5842 for (i = 0; i < esm.nr; i++)
5844 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5845 for (i = 0; i < esm.nr; i++)
5847 free((void *)esm.buf[i].line);
5848 }
5849 esm.nr = 0;
5850}
5851void diff_flush(struct diff_options *options)
5853{
5854 struct diff_queue_struct *q = &diff_queued_diff;
5855 int i, output_format = options->output_format;
5856 int separator = 0;
5857 int dirstat_by_line = 0;
5858 /*
5860 * Order: raw, stat, summary, patch
5861 * or: name/name-status/checkdiff (other bits clear)
5862 */
5863 if (!q->nr)
5864 goto free_queue;
5865 if (output_format & (DIFF_FORMAT_RAW |
5867 DIFF_FORMAT_NAME |
5868 DIFF_FORMAT_NAME_STATUS |
5869 DIFF_FORMAT_CHECKDIFF)) {
5870 for (i = 0; i < q->nr; i++) {
5871 struct diff_filepair *p = q->queue[i];
5872 if (check_pair_status(p))
5873 flush_one_pair(p, options);
5874 }
5875 separator++;
5876 }
5877 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
5879 dirstat_by_line = 1;
5880 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5882 dirstat_by_line) {
5883 struct diffstat_t diffstat;
5884 memset(&diffstat, 0, sizeof(struct diffstat_t));
5886 for (i = 0; i < q->nr; i++) {
5887 struct diff_filepair *p = q->queue[i];
5888 if (check_pair_status(p))
5889 diff_flush_stat(p, options, &diffstat);
5890 }
5891 if (output_format & DIFF_FORMAT_NUMSTAT)
5892 show_numstat(&diffstat, options);
5893 if (output_format & DIFF_FORMAT_DIFFSTAT)
5894 show_stats(&diffstat, options);
5895 if (output_format & DIFF_FORMAT_SHORTSTAT)
5896 show_shortstats(&diffstat, options);
5897 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5898 show_dirstat_by_line(&diffstat, options);
5899 free_diffstat_info(&diffstat);
5900 separator++;
5901 }
5902 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5903 show_dirstat(options);
5904 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5906 for (i = 0; i < q->nr; i++) {
5907 diff_summary(options, q->queue[i]);
5908 }
5909 separator++;
5910 }
5911 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5913 options->flags.exit_with_status &&
5914 options->flags.diff_from_contents) {
5915 /*
5916 * run diff_flush_patch for the exit status. setting
5917 * options->file to /dev/null should be safe, because we
5918 * aren't supposed to produce any output anyway.
5919 */
5920 if (options->close_file)
5921 fclose(options->file);
5922 options->file = xfopen("/dev/null", "w");
5923 options->close_file = 1;
5924 options->color_moved = 0;
5925 for (i = 0; i < q->nr; i++) {
5926 struct diff_filepair *p = q->queue[i];
5927 if (check_pair_status(p))
5928 diff_flush_patch(p, options);
5929 if (options->found_changes)
5930 break;
5931 }
5932 }
5933 if (output_format & DIFF_FORMAT_PATCH) {
5935 if (separator) {
5936 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5937 if (options->stat_sep)
5938 /* attach patch instead of inline */
5939 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5940 NULL, 0, 0);
5941 }
5942 diff_flush_patch_all_file_pairs(options);
5944 }
5945 if (output_format & DIFF_FORMAT_CALLBACK)
5947 options->format_callback(q, options, options->format_callback_data);
5948 for (i = 0; i < q->nr; i++)
5950 diff_free_filepair(q->queue[i]);
5951free_queue:
5952 free(q->queue);
5953 DIFF_QUEUE_CLEAR(q);
5954 if (options->close_file)
5955 fclose(options->file);
5956 /*
5958 * Report the content-level differences with HAS_CHANGES;
5959 * diff_addremove/diff_change does not set the bit when
5960 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5961 */
5962 if (options->flags.diff_from_contents) {
5963 if (options->found_changes)
5964 options->flags.has_changes = 1;
5965 else
5966 options->flags.has_changes = 0;
5967 }
5968}
5969static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5971{
5972 return (((p->status == DIFF_STATUS_MODIFIED) &&
5973 ((p->score &&
5974 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5975 (!p->score &&
5976 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5977 ((p->status != DIFF_STATUS_MODIFIED) &&
5978 filter_bit_tst(p->status, options)));
5979}
5980static void diffcore_apply_filter(struct diff_options *options)
5982{
5983 int i;
5984 struct diff_queue_struct *q = &diff_queued_diff;
5985 struct diff_queue_struct outq;
5986 DIFF_QUEUE_CLEAR(&outq);
5988 if (!options->filter)
5990 return;
5991 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5993 int found;
5994 for (i = found = 0; !found && i < q->nr; i++) {
5995 if (match_filter(options, q->queue[i]))
5996 found++;
5997 }
5998 if (found)
5999 return;
6000 /* otherwise we will clear the whole queue
6002 * by copying the empty outq at the end of this
6003 * function, but first clear the current entries
6004 * in the queue.
6005 */
6006 for (i = 0; i < q->nr; i++)
6007 diff_free_filepair(q->queue[i]);
6008 }
6009 else {
6010 /* Only the matching ones */
6011 for (i = 0; i < q->nr; i++) {
6012 struct diff_filepair *p = q->queue[i];
6013 if (match_filter(options, p))
6014 diff_q(&outq, p);
6015 else
6016 diff_free_filepair(p);
6017 }
6018 }
6019 free(q->queue);
6020 *q = outq;
6021}
6022/* Check whether two filespecs with the same mode and size are identical */
6024static int diff_filespec_is_identical(struct diff_filespec *one,
6025 struct diff_filespec *two)
6026{
6027 if (S_ISGITLINK(one->mode))
6028 return 0;
6029 if (diff_populate_filespec(one, 0))
6030 return 0;
6031 if (diff_populate_filespec(two, 0))
6032 return 0;
6033 return !memcmp(one->data, two->data, one->size);
6034}
6035static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
6037{
6038 if (p->done_skip_stat_unmatch)
6039 return p->skip_stat_unmatch_result;
6040 p->done_skip_stat_unmatch = 1;
6042 p->skip_stat_unmatch_result = 0;
6043 /*
6044 * 1. Entries that come from stat info dirtiness
6045 * always have both sides (iow, not create/delete),
6046 * one side of the object name is unknown, with
6047 * the same mode and size. Keep the ones that
6048 * do not match these criteria. They have real
6049 * differences.
6050 *
6051 * 2. At this point, the file is known to be modified,
6052 * with the same mode and size, and the object
6053 * name of one side is unknown. Need to inspect
6054 * the identical contents.
6055 */
6056 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6057 !DIFF_FILE_VALID(p->two) ||
6058 (p->one->oid_valid && p->two->oid_valid) ||
6059 (p->one->mode != p->two->mode) ||
6060 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
6061 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
6062 (p->one->size != p->two->size) ||
6063 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
6064 p->skip_stat_unmatch_result = 1;
6065 return p->skip_stat_unmatch_result;
6066}
6067static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6069{
6070 int i;
6071 struct diff_queue_struct *q = &diff_queued_diff;
6072 struct diff_queue_struct outq;
6073 DIFF_QUEUE_CLEAR(&outq);
6074 for (i = 0; i < q->nr; i++) {
6076 struct diff_filepair *p = q->queue[i];
6077 if (diff_filespec_check_stat_unmatch(p))
6079 diff_q(&outq, p);
6080 else {
6081 /*
6082 * The caller can subtract 1 from skip_stat_unmatch
6083 * to determine how many paths were dirty only
6084 * due to stat info mismatch.
6085 */
6086 if (!diffopt->flags.no_index)
6087 diffopt->skip_stat_unmatch++;
6088 diff_free_filepair(p);
6089 }
6090 }
6091 free(q->queue);
6092 *q = outq;
6093}
6094static int diffnamecmp(const void *a_, const void *b_)
6096{
6097 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6098 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6099 const char *name_a, *name_b;
6100 name_a = a->one ? a->one->path : a->two->path;
6102 name_b = b->one ? b->one->path : b->two->path;
6103 return strcmp(name_a, name_b);
6104}
6105void diffcore_fix_diff_index(struct diff_options *options)
6107{
6108 struct diff_queue_struct *q = &diff_queued_diff;
6109 QSORT(q->queue, q->nr, diffnamecmp);
6110}
6111void diffcore_std(struct diff_options *options)
6113{
6114 /* NOTE please keep the following in sync with diff_tree_combined() */
6115 if (options->skip_stat_unmatch)
6116 diffcore_skip_stat_unmatch(options);
6117 if (!options->found_follow) {
6118 /* See try_to_follow_renames() in tree-diff.c */
6119 if (options->break_opt != -1)
6120 diffcore_break(options->break_opt);
6121 if (options->detect_rename)
6122 diffcore_rename(options);
6123 if (options->break_opt != -1)
6124 diffcore_merge_broken();
6125 }
6126 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6127 diffcore_pickaxe(options);
6128 if (options->orderfile)
6129 diffcore_order(options->orderfile);
6130 if (!options->found_follow)
6131 /* See try_to_follow_renames() in tree-diff.c */
6132 diff_resolve_rename_copy();
6133 diffcore_apply_filter(options);
6134 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6136 options->flags.has_changes = 1;
6137 else
6138 options->flags.has_changes = 0;
6139 options->found_follow = 0;
6141}
6142int diff_result_code(struct diff_options *opt, int status)
6144{
6145 int result = 0;
6146 diff_warn_rename_limit("diff.renameLimit",
6148 opt->needed_rename_limit,
6149 opt->degraded_cc_to_c);
6150 if (!opt->flags.exit_with_status &&
6151 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6152 return status;
6153 if (opt->flags.exit_with_status &&
6154 opt->flags.has_changes)
6155 result |= 01;
6156 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6157 opt->flags.check_failed)
6158 result |= 02;
6159 return result;
6160}
6161int diff_can_quit_early(struct diff_options *opt)
6163{
6164 return (opt->flags.quick &&
6165 !opt->filter &&
6166 opt->flags.has_changes);
6167}
6168/*
6170 * Shall changes to this submodule be ignored?
6171 *
6172 * Submodule changes can be configured to be ignored separately for each path,
6173 * but that configuration can be overridden from the command line.
6174 */
6175static int is_submodule_ignored(const char *path, struct diff_options *options)
6176{
6177 int ignored = 0;
6178 struct diff_flags orig_flags = options->flags;
6179 if (!options->flags.override_submodule_config)
6180 set_diffopt_flags_from_submodule_config(options, path);
6181 if (options->flags.ignore_submodules)
6182 ignored = 1;
6183 options->flags = orig_flags;
6184 return ignored;
6185}
6186void diff_addremove(struct diff_options *options,
6188 int addremove, unsigned mode,
6189 const struct object_id *oid,
6190 int oid_valid,
6191 const char *concatpath, unsigned dirty_submodule)
6192{
6193 struct diff_filespec *one, *two;
6194 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6196 return;
6197 /* This may look odd, but it is a preparation for
6199 * feeding "there are unchanged files which should
6200 * not produce diffs, but when you are doing copy
6201 * detection you would need them, so here they are"
6202 * entries to the diff-core. They will be prefixed
6203 * with something like '=' or '*' (I haven't decided
6204 * which but should not make any difference).
6205 * Feeding the same new and old to diff_change()
6206 * also has the same effect.
6207 * Before the final output happens, they are pruned after
6208 * merged into rename/copy pairs as appropriate.
6209 */
6210 if (options->flags.reverse_diff)
6211 addremove = (addremove == '+' ? '-' :
6212 addremove == '-' ? '+' : addremove);
6213 if (options->prefix &&
6215 strncmp(concatpath, options->prefix, options->prefix_length))
6216 return;
6217 one = alloc_filespec(concatpath);
6219 two = alloc_filespec(concatpath);
6220 if (addremove != '+')
6222 fill_filespec(one, oid, oid_valid, mode);
6223 if (addremove != '-') {
6224 fill_filespec(two, oid, oid_valid, mode);
6225 two->dirty_submodule = dirty_submodule;
6226 }
6227 diff_queue(&diff_queued_diff, one, two);
6229 if (!options->flags.diff_from_contents)
6230 options->flags.has_changes = 1;
6231}
6232void diff_change(struct diff_options *options,
6234 unsigned old_mode, unsigned new_mode,
6235 const struct object_id *old_oid,
6236 const struct object_id *new_oid,
6237 int old_oid_valid, int new_oid_valid,
6238 const char *concatpath,
6239 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6240{
6241 struct diff_filespec *one, *two;
6242 struct diff_filepair *p;
6243 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6245 is_submodule_ignored(concatpath, options))
6246 return;
6247 if (options->flags.reverse_diff) {
6249 SWAP(old_mode, new_mode);
6250 SWAP(old_oid, new_oid);
6251 SWAP(old_oid_valid, new_oid_valid);
6252 SWAP(old_dirty_submodule, new_dirty_submodule);
6253 }
6254 if (options->prefix &&
6256 strncmp(concatpath, options->prefix, options->prefix_length))
6257 return;
6258 one = alloc_filespec(concatpath);
6260 two = alloc_filespec(concatpath);
6261 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6262 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6263 one->dirty_submodule = old_dirty_submodule;
6264 two->dirty_submodule = new_dirty_submodule;
6265 p = diff_queue(&diff_queued_diff, one, two);
6266 if (options->flags.diff_from_contents)
6268 return;
6269 if (options->flags.quick && options->skip_stat_unmatch &&
6271 !diff_filespec_check_stat_unmatch(p))
6272 return;
6273 options->flags.has_changes = 1;
6275}
6276struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6278{
6279 struct diff_filepair *pair;
6280 struct diff_filespec *one, *two;
6281 if (options->prefix &&
6283 strncmp(path, options->prefix, options->prefix_length))
6284 return NULL;
6285 one = alloc_filespec(path);
6287 two = alloc_filespec(path);
6288 pair = diff_queue(&diff_queued_diff, one, two);
6289 pair->is_unmerged = 1;
6290 return pair;
6291}
6292static char *run_textconv(const char *pgm, struct diff_filespec *spec,
6294 size_t *outsize)
6295{
6296 struct diff_tempfile *temp;
6297 const char *argv[3];
6298 const char **arg = argv;
6299 struct child_process child = CHILD_PROCESS_INIT;
6300 struct strbuf buf = STRBUF_INIT;
6301 int err = 0;
6302 temp = prepare_temp_file(spec->path, spec);
6304 *arg++ = pgm;
6305 *arg++ = temp->name;
6306 *arg = NULL;
6307 child.use_shell = 1;
6309 child.argv = argv;
6310 child.out = -1;
6311 if (start_command(&child)) {
6312 remove_tempfile();
6313 return NULL;
6314 }
6315 if (strbuf_read(&buf, child.out, 0) < 0)
6317 err = error("error reading from textconv command '%s'", pgm);
6318 close(child.out);
6319 if (finish_command(&child) || err) {
6321 strbuf_release(&buf);
6322 remove_tempfile();
6323 return NULL;
6324 }
6325 remove_tempfile();
6326 return strbuf_detach(&buf, outsize);
6328}
6329size_t fill_textconv(struct userdiff_driver *driver,
6331 struct diff_filespec *df,
6332 char **outbuf)
6333{
6334 size_t size;
6335 if (!driver) {
6337 if (!DIFF_FILE_VALID(df)) {
6338 *outbuf = "";
6339 return 0;
6340 }
6341 if (diff_populate_filespec(df, 0))
6342 die("unable to read files to diff");
6343 *outbuf = df->data;
6344 return df->size;
6345 }
6346 if (!driver->textconv)
6348 BUG("fill_textconv called with non-textconv driver");
6349 if (driver->textconv_cache && df->oid_valid) {
6351 *outbuf = notes_cache_get(driver->textconv_cache,
6352 &df->oid,
6353 &size);
6354 if (*outbuf)
6355 return size;
6356 }
6357 *outbuf = run_textconv(driver->textconv, df, &size);
6359 if (!*outbuf)
6360 die("unable to read files to diff");
6361 if (driver->textconv_cache && df->oid_valid) {
6363 /* ignore errors, as we might be in a readonly repository */
6364 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6365 size);
6366 /*
6367 * we could save up changes and flush them all at the end,
6368 * but we would need an extra call after all diffing is done.
6369 * Since generating a cache entry is the slow path anyway,
6370 * this extra overhead probably isn't a big deal.
6371 */
6372 notes_cache_write(driver->textconv_cache);
6373 }
6374 return size;
6376}
6377int textconv_object(const char *path,
6379 unsigned mode,
6380 const struct object_id *oid,
6381 int oid_valid,
6382 char **buf,
6383 unsigned long *buf_size)
6384{
6385 struct diff_filespec *df;
6386 struct userdiff_driver *textconv;
6387 df = alloc_filespec(path);
6389 fill_filespec(df, oid, oid_valid, mode);
6390 textconv = get_textconv(df);
6391 if (!textconv) {
6392 free_filespec(df);
6393 return 0;
6394 }
6395 *buf_size = fill_textconv(textconv, df, buf);
6397 free_filespec(df);
6398 return 1;
6399}
6400void setup_diff_pager(struct diff_options *opt)
6402{
6403 /*
6404 * If the user asked for our exit code, then either they want --quiet
6405 * or --exit-code. We should definitely not bother with a pager in the
6406 * former case, as we will generate no output. Since we still properly
6407 * report our exit code even when a pager is run, we _could_ run a
6408 * pager with --exit-code. But since we have not done so historically,
6409 * and because it is easy to find people oneline advising "git diff
6410 * --exit-code" in hooks and other scripts, we do not do so.
6411 */
6412 if (!opt->flags.exit_with_status &&
6413 check_pager_config("diff") != 0)
6414 setup_pager();
6415}