be312bc20d65b53e2b09409246ee81bceaa68108
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 if (!strcmp(arg, "dimmed_zebra"))
289 return COLOR_MOVED_ZEBRA_DIM;
290 else
291 return error(_("color moved setting must be one of 'no', 'default', 'blocks', 'zebra', 'dimmed-zebra', 'plain'"));
292}
293static int parse_color_moved_ws(const char *arg)
295{
296 int ret = 0;
297 struct string_list l = STRING_LIST_INIT_DUP;
298 struct string_list_item *i;
299 string_list_split(&l, arg, ',', -1);
301 for_each_string_list_item(i, &l) {
303 struct strbuf sb = STRBUF_INIT;
304 strbuf_addstr(&sb, i->string);
305 strbuf_trim(&sb);
306 if (!strcmp(sb.buf, "ignore-space-change"))
308 ret |= XDF_IGNORE_WHITESPACE_CHANGE;
309 else if (!strcmp(sb.buf, "ignore-space-at-eol"))
310 ret |= XDF_IGNORE_WHITESPACE_AT_EOL;
311 else if (!strcmp(sb.buf, "ignore-all-space"))
312 ret |= XDF_IGNORE_WHITESPACE;
313 else if (!strcmp(sb.buf, "allow-indentation-change"))
314 ret |= COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE;
315 else
316 error(_("ignoring unknown color-moved-ws mode '%s'"), sb.buf);
317 strbuf_release(&sb);
319 }
320 if ((ret & COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) &&
322 (ret & XDF_WHITESPACE_FLAGS))
323 die(_("color-moved-ws: allow-indentation-change cannot be combined with other white space modes"));
324 string_list_clear(&l, 0);
326 return ret;
328}
329int git_diff_ui_config(const char *var, const char *value, void *cb)
331{
332 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
333 diff_use_color_default = git_config_colorbool(var, value);
334 return 0;
335 }
336 if (!strcmp(var, "diff.colormoved")) {
337 int cm = parse_color_moved(value);
338 if (cm < 0)
339 return -1;
340 diff_color_moved_default = cm;
341 return 0;
342 }
343 if (!strcmp(var, "diff.colormovedws")) {
344 int cm = parse_color_moved_ws(value);
345 if (cm < 0)
346 return -1;
347 diff_color_moved_ws_default = cm;
348 return 0;
349 }
350 if (!strcmp(var, "diff.context")) {
351 diff_context_default = git_config_int(var, value);
352 if (diff_context_default < 0)
353 return -1;
354 return 0;
355 }
356 if (!strcmp(var, "diff.interhunkcontext")) {
357 diff_interhunk_context_default = git_config_int(var, value);
358 if (diff_interhunk_context_default < 0)
359 return -1;
360 return 0;
361 }
362 if (!strcmp(var, "diff.renames")) {
363 diff_detect_rename_default = git_config_rename(var, value);
364 return 0;
365 }
366 if (!strcmp(var, "diff.autorefreshindex")) {
367 diff_auto_refresh_index = git_config_bool(var, value);
368 return 0;
369 }
370 if (!strcmp(var, "diff.mnemonicprefix")) {
371 diff_mnemonic_prefix = git_config_bool(var, value);
372 return 0;
373 }
374 if (!strcmp(var, "diff.noprefix")) {
375 diff_no_prefix = git_config_bool(var, value);
376 return 0;
377 }
378 if (!strcmp(var, "diff.statgraphwidth")) {
379 diff_stat_graph_width = git_config_int(var, value);
380 return 0;
381 }
382 if (!strcmp(var, "diff.external"))
383 return git_config_string(&external_diff_cmd_cfg, var, value);
384 if (!strcmp(var, "diff.wordregex"))
385 return git_config_string(&diff_word_regex_cfg, var, value);
386 if (!strcmp(var, "diff.orderfile"))
387 return git_config_pathname(&diff_order_file_cfg, var, value);
388 if (!strcmp(var, "diff.ignoresubmodules"))
390 handle_ignore_submodules_arg(&default_diff_options, value);
391 if (!strcmp(var, "diff.submodule")) {
393 if (parse_submodule_params(&default_diff_options, value))
394 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
395 value);
396 return 0;
397 }
398 if (!strcmp(var, "diff.algorithm")) {
400 diff_algorithm = parse_algorithm_value(value);
401 if (diff_algorithm < 0)
402 return -1;
403 return 0;
404 }
405 if (!strcmp(var, "diff.wserrorhighlight")) {
407 int val = parse_ws_error_highlight(value);
408 if (val < 0)
409 return -1;
410 ws_error_highlight_default = val;
411 return 0;
412 }
413 if (git_color_config(var, value, cb) < 0)
415 return -1;
416 return git_diff_basic_config(var, value, cb);
418}
419int git_diff_basic_config(const char *var, const char *value, void *cb)
421{
422 const char *name;
423 if (!strcmp(var, "diff.renamelimit")) {
425 diff_rename_limit_default = git_config_int(var, value);
426 return 0;
427 }
428 if (userdiff_config(var, value) < 0)
430 return -1;
431 if (skip_prefix(var, "diff.color.", &name) ||
433 skip_prefix(var, "color.diff.", &name)) {
434 int slot = parse_diff_color_slot(name);
435 if (slot < 0)
436 return 0;
437 if (!value)
438 return config_error_nonbool(var);
439 return color_parse(value, diff_colors[slot]);
440 }
441 /* like GNU diff's --suppress-blank-empty option */
443 if (!strcmp(var, "diff.suppressblankempty") ||
444 /* for backwards compatibility */
445 !strcmp(var, "diff.suppress-blank-empty")) {
446 diff_suppress_blank_empty = git_config_bool(var, value);
447 return 0;
448 }
449 if (!strcmp(var, "diff.dirstat")) {
451 struct strbuf errmsg = STRBUF_INIT;
452 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
453 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
454 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
455 errmsg.buf);
456 strbuf_release(&errmsg);
457 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
458 return 0;
459 }
460 if (git_diff_heuristic_config(var, value, cb) < 0)
462 return -1;
463 return git_default_config(var, value, cb);
465}
466static char *quote_two(const char *one, const char *two)
468{
469 int need_one = quote_c_style(one, NULL, NULL, 1);
470 int need_two = quote_c_style(two, NULL, NULL, 1);
471 struct strbuf res = STRBUF_INIT;
472 if (need_one + need_two) {
474 strbuf_addch(&res, '"');
475 quote_c_style(one, &res, NULL, 1);
476 quote_c_style(two, &res, NULL, 1);
477 strbuf_addch(&res, '"');
478 } else {
479 strbuf_addstr(&res, one);
480 strbuf_addstr(&res, two);
481 }
482 return strbuf_detach(&res, NULL);
483}
484static const char *external_diff(void)
486{
487 static const char *external_diff_cmd = NULL;
488 static int done_preparing = 0;
489 if (done_preparing)
491 return external_diff_cmd;
492 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
493 if (!external_diff_cmd)
494 external_diff_cmd = external_diff_cmd_cfg;
495 done_preparing = 1;
496 return external_diff_cmd;
497}
498/*
500 * Keep track of files used for diffing. Sometimes such an entry
501 * refers to a temporary file, sometimes to an existing file, and
502 * sometimes to "/dev/null".
503 */
504static struct diff_tempfile {
505 /*
506 * filename external diff should read from, or NULL if this
507 * entry is currently not in use:
508 */
509 const char *name;
510 char hex[GIT_MAX_HEXSZ + 1];
512 char mode[10];
513 /*
515 * If this diff_tempfile instance refers to a temporary file,
516 * this tempfile object is used to manage its lifetime.
517 */
518 struct tempfile *tempfile;
519} diff_temp[2];
520struct emit_callback {
522 int color_diff;
523 unsigned ws_rule;
524 int blank_at_eof_in_preimage;
525 int blank_at_eof_in_postimage;
526 int lno_in_preimage;
527 int lno_in_postimage;
528 const char **label_path;
529 struct diff_words_data *diff_words;
530 struct diff_options *opt;
531 struct strbuf *header;
532};
533static int count_lines(const char *data, int size)
535{
536 int count, ch, completely_empty = 1, nl_just_seen = 0;
537 count = 0;
538 while (0 < size--) {
539 ch = *data++;
540 if (ch == '\n') {
541 count++;
542 nl_just_seen = 1;
543 completely_empty = 0;
544 }
545 else {
546 nl_just_seen = 0;
547 completely_empty = 0;
548 }
549 }
550 if (completely_empty)
551 return 0;
552 if (!nl_just_seen)
553 count++; /* no trailing newline */
554 return count;
555}
556static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
558{
559 if (!DIFF_FILE_VALID(one)) {
560 mf->ptr = (char *)""; /* does not matter */
561 mf->size = 0;
562 return 0;
563 }
564 else if (diff_populate_filespec(one, 0))
565 return -1;
566 mf->ptr = one->data;
568 mf->size = one->size;
569 return 0;
570}
571/* like fill_mmfile, but only for size, so we can avoid retrieving blob */
573static unsigned long diff_filespec_size(struct diff_filespec *one)
574{
575 if (!DIFF_FILE_VALID(one))
576 return 0;
577 diff_populate_filespec(one, CHECK_SIZE_ONLY);
578 return one->size;
579}
580static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
582{
583 char *ptr = mf->ptr;
584 long size = mf->size;
585 int cnt = 0;
586 if (!size)
588 return cnt;
589 ptr += size - 1; /* pointing at the very end */
590 if (*ptr != '\n')
591 ; /* incomplete line */
592 else
593 ptr--; /* skip the last LF */
594 while (mf->ptr < ptr) {
595 char *prev_eol;
596 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
597 if (*prev_eol == '\n')
598 break;
599 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
600 break;
601 cnt++;
602 ptr = prev_eol - 1;
603 }
604 return cnt;
605}
606static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
608 struct emit_callback *ecbdata)
609{
610 int l1, l2, at;
611 unsigned ws_rule = ecbdata->ws_rule;
612 l1 = count_trailing_blank(mf1, ws_rule);
613 l2 = count_trailing_blank(mf2, ws_rule);
614 if (l2 <= l1) {
615 ecbdata->blank_at_eof_in_preimage = 0;
616 ecbdata->blank_at_eof_in_postimage = 0;
617 return;
618 }
619 at = count_lines(mf1->ptr, mf1->size);
620 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
621 at = count_lines(mf2->ptr, mf2->size);
623 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
624}
625static void emit_line_0(struct diff_options *o,
627 const char *set, unsigned reverse, const char *reset,
628 int first, const char *line, int len)
629{
630 int has_trailing_newline, has_trailing_carriage_return;
631 int nofirst;
632 FILE *file = o->file;
633 if (first)
635 fputs(diff_line_prefix(o), file);
636 else if (!len)
637 return;
638 if (len == 0) {
640 has_trailing_newline = (first == '\n');
641 has_trailing_carriage_return = (!has_trailing_newline &&
642 (first == '\r'));
643 nofirst = has_trailing_newline || has_trailing_carriage_return;
644 } else {
645 has_trailing_newline = (len > 0 && line[len-1] == '\n');
646 if (has_trailing_newline)
647 len--;
648 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
649 if (has_trailing_carriage_return)
650 len--;
651 nofirst = 0;
652 }
653 if (len || !nofirst) {
655 if (reverse && want_color(o->use_color))
656 fputs(GIT_COLOR_REVERSE, file);
657 fputs(set, file);
658 if (first && !nofirst)
659 fputc(first, file);
660 fwrite(line, len, 1, file);
661 fputs(reset, file);
662 }
663 if (has_trailing_carriage_return)
664 fputc('\r', file);
665 if (has_trailing_newline)
666 fputc('\n', file);
667}
668static void emit_line(struct diff_options *o, const char *set, const char *reset,
670 const char *line, int len)
671{
672 emit_line_0(o, set, 0, reset, line[0], line+1, len-1);
673}
674enum diff_symbol {
676 DIFF_SYMBOL_BINARY_DIFF_HEADER,
677 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
678 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
679 DIFF_SYMBOL_BINARY_DIFF_BODY,
680 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
681 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
682 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
683 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
684 DIFF_SYMBOL_STATS_LINE,
685 DIFF_SYMBOL_WORD_DIFF,
686 DIFF_SYMBOL_STAT_SEP,
687 DIFF_SYMBOL_SUMMARY,
688 DIFF_SYMBOL_SUBMODULE_ADD,
689 DIFF_SYMBOL_SUBMODULE_DEL,
690 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
691 DIFF_SYMBOL_SUBMODULE_MODIFIED,
692 DIFF_SYMBOL_SUBMODULE_HEADER,
693 DIFF_SYMBOL_SUBMODULE_ERROR,
694 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
695 DIFF_SYMBOL_REWRITE_DIFF,
696 DIFF_SYMBOL_BINARY_FILES,
697 DIFF_SYMBOL_HEADER,
698 DIFF_SYMBOL_FILEPAIR_PLUS,
699 DIFF_SYMBOL_FILEPAIR_MINUS,
700 DIFF_SYMBOL_WORDS_PORCELAIN,
701 DIFF_SYMBOL_WORDS,
702 DIFF_SYMBOL_CONTEXT,
703 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
704 DIFF_SYMBOL_PLUS,
705 DIFF_SYMBOL_MINUS,
706 DIFF_SYMBOL_NO_LF_EOF,
707 DIFF_SYMBOL_CONTEXT_FRAGINFO,
708 DIFF_SYMBOL_CONTEXT_MARKER,
709 DIFF_SYMBOL_SEPARATOR
710};
711/*
712 * Flags for content lines:
713 * 0..12 are whitespace rules
714 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
715 * 16 is marking if the line is blank at EOF
716 */
717#define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
718#define DIFF_SYMBOL_MOVED_LINE (1<<17)
719#define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
720#define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
721#define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
722/*
724 * This struct is used when we need to buffer the output of the diff output.
725 *
726 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
727 * into the pre/post image file. This pointer could be a union with the
728 * line pointer. By storing an offset into the file instead of the literal line,
729 * we can decrease the memory footprint for the buffered output. At first we
730 * may want to only have indirection for the content lines, but we could also
731 * enhance the state for emitting prefabricated lines, e.g. the similarity
732 * score line or hunk/file headers would only need to store a number or path
733 * and then the output can be constructed later on depending on state.
734 */
735struct emitted_diff_symbol {
736 const char *line;
737 int len;
738 int flags;
739 enum diff_symbol s;
740};
741#define EMITTED_DIFF_SYMBOL_INIT {NULL}
742struct emitted_diff_symbols {
744 struct emitted_diff_symbol *buf;
745 int nr, alloc;
746};
747#define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
748static void append_emitted_diff_symbol(struct diff_options *o,
750 struct emitted_diff_symbol *e)
751{
752 struct emitted_diff_symbol *f;
753 ALLOC_GROW(o->emitted_symbols->buf,
755 o->emitted_symbols->nr + 1,
756 o->emitted_symbols->alloc);
757 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
758 memcpy(f, e, sizeof(struct emitted_diff_symbol));
760 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
761}
762struct moved_entry {
764 struct hashmap_entry ent;
765 const struct emitted_diff_symbol *es;
766 struct moved_entry *next_line;
767 struct ws_delta *wsd;
768};
769/**
771 * The struct ws_delta holds white space differences between moved lines, i.e.
772 * between '+' and '-' lines that have been detected to be a move.
773 * The string contains the difference in leading white spaces, before the
774 * rest of the line is compared using the white space config for move
775 * coloring. The current_longer indicates if the first string in the
776 * comparision is longer than the second.
777 */
778struct ws_delta {
779 char *string;
780 unsigned int current_longer : 1;
781};
782#define WS_DELTA_INIT { NULL, 0 }
783static int compute_ws_delta(const struct emitted_diff_symbol *a,
785 const struct emitted_diff_symbol *b,
786 struct ws_delta *out)
787{
788 const struct emitted_diff_symbol *longer = a->len > b->len ? a : b;
789 const struct emitted_diff_symbol *shorter = a->len > b->len ? b : a;
790 int d = longer->len - shorter->len;
791 out->string = xmemdupz(longer->line, d);
793 out->current_longer = (a == longer);
794 return !strncmp(longer->line + d, shorter->line, shorter->len);
796}
797static int cmp_in_block_with_wsd(const struct diff_options *o,
799 const struct moved_entry *cur,
800 const struct moved_entry *match,
801 struct moved_entry *pmb,
802 int n)
803{
804 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
805 int al = cur->es->len, cl = l->len;
806 const char *a = cur->es->line,
807 *b = match->es->line,
808 *c = l->line;
809 int wslen;
811 /*
813 * We need to check if 'cur' is equal to 'match'.
814 * As those are from the same (+/-) side, we do not need to adjust for
815 * indent changes. However these were found using fuzzy matching
816 * so we do have to check if they are equal.
817 */
818 if (strcmp(a, b))
819 return 1;
820 if (!pmb->wsd)
822 /*
823 * No white space delta was carried forward? This can happen
824 * when we exit early in this function and do not carry
825 * forward ws.
826 */
827 return 1;
828 /*
830 * The indent changes of the block are known and carried forward in
831 * pmb->wsd; however we need to check if the indent changes of the
832 * current line are still the same as before.
833 *
834 * To do so we need to compare 'l' to 'cur', adjusting the
835 * one of them for the white spaces, depending which was longer.
836 */
837 wslen = strlen(pmb->wsd->string);
839 if (pmb->wsd->current_longer) {
840 c += wslen;
841 cl -= wslen;
842 } else {
843 a += wslen;
844 al -= wslen;
845 }
846 if (strcmp(a, c))
848 return 1;
849 return 0;
851}
852static int moved_entry_cmp(const void *hashmap_cmp_fn_data,
854 const void *entry,
855 const void *entry_or_key,
856 const void *keydata)
857{
858 const struct diff_options *diffopt = hashmap_cmp_fn_data;
859 const struct moved_entry *a = entry;
860 const struct moved_entry *b = entry_or_key;
861 unsigned flags = diffopt->color_moved_ws_handling
862 & XDF_WHITESPACE_FLAGS;
863 if (diffopt->color_moved_ws_handling &
865 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
866 /*
867 * As there is not specific white space config given,
868 * we'd need to check for a new block, so ignore all
869 * white space. The setup of the white space
870 * configuration for the next block is done else where
871 */
872 flags |= XDF_IGNORE_WHITESPACE;
873 return !xdiff_compare_lines(a->es->line, a->es->len,
875 b->es->line, b->es->len,
876 flags);
877}
878static struct moved_entry *prepare_entry(struct diff_options *o,
880 int line_no)
881{
882 struct moved_entry *ret = xmalloc(sizeof(*ret));
883 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
884 unsigned flags = o->color_moved_ws_handling & XDF_WHITESPACE_FLAGS;
885 ret->ent.hash = xdiff_hash_string(l->line, l->len, flags);
887 ret->es = l;
888 ret->next_line = NULL;
889 ret->wsd = NULL;
890 return ret;
892}
893static void add_lines_to_move_detection(struct diff_options *o,
895 struct hashmap *add_lines,
896 struct hashmap *del_lines)
897{
898 struct moved_entry *prev_line = NULL;
899 int n;
901 for (n = 0; n < o->emitted_symbols->nr; n++) {
902 struct hashmap *hm;
903 struct moved_entry *key;
904 switch (o->emitted_symbols->buf[n].s) {
906 case DIFF_SYMBOL_PLUS:
907 hm = add_lines;
908 break;
909 case DIFF_SYMBOL_MINUS:
910 hm = del_lines;
911 break;
912 default:
913 prev_line = NULL;
914 continue;
915 }
916 key = prepare_entry(o, n);
918 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
919 prev_line->next_line = key;
920 hashmap_add(hm, key);
922 prev_line = key;
923 }
924}
925static void pmb_advance_or_null(struct diff_options *o,
927 struct moved_entry *match,
928 struct hashmap *hm,
929 struct moved_entry **pmb,
930 int pmb_nr)
931{
932 int i;
933 for (i = 0; i < pmb_nr; i++) {
934 struct moved_entry *prev = pmb[i];
935 struct moved_entry *cur = (prev && prev->next_line) ?
936 prev->next_line : NULL;
937 if (cur && !hm->cmpfn(o, cur, match, NULL)) {
938 pmb[i] = cur;
939 } else {
940 pmb[i] = NULL;
941 }
942 }
943}
944static void pmb_advance_or_null_multi_match(struct diff_options *o,
946 struct moved_entry *match,
947 struct hashmap *hm,
948 struct moved_entry **pmb,
949 int pmb_nr, int n)
950{
951 int i;
952 char *got_match = xcalloc(1, pmb_nr);
953 for (; match; match = hashmap_get_next(hm, match)) {
955 for (i = 0; i < pmb_nr; i++) {
956 struct moved_entry *prev = pmb[i];
957 struct moved_entry *cur = (prev && prev->next_line) ?
958 prev->next_line : NULL;
959 if (!cur)
960 continue;
961 if (!cmp_in_block_with_wsd(o, cur, match, pmb[i], n))
962 got_match[i] |= 1;
963 }
964 }
965 for (i = 0; i < pmb_nr; i++) {
967 if (got_match[i]) {
968 /* Carry the white space delta forward */
969 pmb[i]->next_line->wsd = pmb[i]->wsd;
970 pmb[i] = pmb[i]->next_line;
971 } else
972 pmb[i] = NULL;
973 }
974}
975static int shrink_potential_moved_blocks(struct moved_entry **pmb,
977 int pmb_nr)
978{
979 int lp, rp;
980 /* Shrink the set of potential block to the remaining running */
982 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
983 while (lp < pmb_nr && pmb[lp])
984 lp++;
985 /* lp points at the first NULL now */
986 while (rp > -1 && !pmb[rp])
988 rp--;
989 /* rp points at the last non-NULL */
990 if (lp < pmb_nr && rp > -1 && lp < rp) {
992 pmb[lp] = pmb[rp];
993 if (pmb[rp]->wsd) {
994 free(pmb[rp]->wsd->string);
995 FREE_AND_NULL(pmb[rp]->wsd);
996 }
997 pmb[rp] = NULL;
998 rp--;
999 lp++;
1000 }
1001 }
1002 /* Remember the number of running sets */
1004 return rp + 1;
1005}
1006/*
1008 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
1009 *
1010 * Otherwise, if the last block has fewer alphanumeric characters than
1011 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
1012 * that block.
1013 *
1014 * The last block consists of the (n - block_length)'th line up to but not
1015 * including the nth line.
1016 *
1017 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
1018 * Think of a way to unify them.
1019 */
1020static void adjust_last_block(struct diff_options *o, int n, int block_length)
1021{
1022 int i, alnum_count = 0;
1023 if (o->color_moved == COLOR_MOVED_PLAIN)
1024 return;
1025 for (i = 1; i < block_length + 1; i++) {
1026 const char *c = o->emitted_symbols->buf[n - i].line;
1027 for (; *c; c++) {
1028 if (!isalnum(*c))
1029 continue;
1030 alnum_count++;
1031 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
1032 return;
1033 }
1034 }
1035 for (i = 1; i < block_length + 1; i++)
1036 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
1037}
1038/* Find blocks of moved code, delegate actual coloring decision to helper */
1040static void mark_color_as_moved(struct diff_options *o,
1041 struct hashmap *add_lines,
1042 struct hashmap *del_lines)
1043{
1044 struct moved_entry **pmb = NULL; /* potentially moved blocks */
1045 int pmb_nr = 0, pmb_alloc = 0;
1046 int n, flipped_block = 1, block_length = 0;
1047 for (n = 0; n < o->emitted_symbols->nr; n++) {
1050 struct hashmap *hm = NULL;
1051 struct moved_entry *key;
1052 struct moved_entry *match = NULL;
1053 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1054 switch (l->s) {
1056 case DIFF_SYMBOL_PLUS:
1057 hm = del_lines;
1058 key = prepare_entry(o, n);
1059 match = hashmap_get(hm, key, NULL);
1060 free(key);
1061 break;
1062 case DIFF_SYMBOL_MINUS:
1063 hm = add_lines;
1064 key = prepare_entry(o, n);
1065 match = hashmap_get(hm, key, NULL);
1066 free(key);
1067 break;
1068 default:
1069 flipped_block = 1;
1070 }
1071 if (!match) {
1073 adjust_last_block(o, n, block_length);
1074 pmb_nr = 0;
1075 block_length = 0;
1076 continue;
1077 }
1078 l->flags |= DIFF_SYMBOL_MOVED_LINE;
1080 if (o->color_moved == COLOR_MOVED_PLAIN)
1082 continue;
1083 if (o->color_moved_ws_handling &
1085 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
1086 pmb_advance_or_null_multi_match(o, match, hm, pmb, pmb_nr, n);
1087 else
1088 pmb_advance_or_null(o, match, hm, pmb, pmb_nr);
1089 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
1091 if (pmb_nr == 0) {
1093 /*
1094 * The current line is the start of a new block.
1095 * Setup the set of potential blocks.
1096 */
1097 for (; match; match = hashmap_get_next(hm, match)) {
1098 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
1099 if (o->color_moved_ws_handling &
1100 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE) {
1101 struct ws_delta *wsd = xmalloc(sizeof(*match->wsd));
1102 if (compute_ws_delta(l, match->es, wsd)) {
1103 match->wsd = wsd;
1104 pmb[pmb_nr++] = match;
1105 } else
1106 free(wsd);
1107 } else {
1108 pmb[pmb_nr++] = match;
1109 }
1110 }
1111 flipped_block = (flipped_block + 1) % 2;
1113 adjust_last_block(o, n, block_length);
1115 block_length = 0;
1116 }
1117 block_length++;
1119 if (flipped_block && o->color_moved != COLOR_MOVED_BLOCKS)
1121 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
1122 }
1123 adjust_last_block(o, n, block_length);
1124 free(pmb);
1126}
1127#define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
1129 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
1130static void dim_moved_lines(struct diff_options *o)
1131{
1132 int n;
1133 for (n = 0; n < o->emitted_symbols->nr; n++) {
1134 struct emitted_diff_symbol *prev = (n != 0) ?
1135 &o->emitted_symbols->buf[n - 1] : NULL;
1136 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
1137 struct emitted_diff_symbol *next =
1138 (n < o->emitted_symbols->nr - 1) ?
1139 &o->emitted_symbols->buf[n + 1] : NULL;
1140 /* Not a plus or minus line? */
1142 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
1143 continue;
1144 /* Not a moved line? */
1146 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
1147 continue;
1148 /*
1150 * If prev or next are not a plus or minus line,
1151 * pretend they don't exist
1152 */
1153 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1154 prev->s != DIFF_SYMBOL_MINUS)
1155 prev = NULL;
1156 if (next && next->s != DIFF_SYMBOL_PLUS &&
1157 next->s != DIFF_SYMBOL_MINUS)
1158 next = NULL;
1159 /* Inside a block? */
1161 if ((prev &&
1162 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1163 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1164 (next &&
1165 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1166 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1167 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1168 continue;
1169 }
1170 /* Check if we are at an interesting bound: */
1172 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1173 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1174 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1175 continue;
1176 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1177 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1178 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1179 continue;
1180 /*
1182 * The boundary to prev and next are not interesting,
1183 * so this line is not interesting as a whole
1184 */
1185 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1186 }
1187}
1188static void emit_line_ws_markup(struct diff_options *o,
1190 const char *set, const char *reset,
1191 const char *line, int len,
1192 const char *set_sign, char sign,
1193 unsigned ws_rule, int blank_at_eof)
1194{
1195 const char *ws = NULL;
1196 if (o->ws_error_highlight & ws_rule) {
1198 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1199 if (!*ws)
1200 ws = NULL;
1201 }
1202 if (!ws && !set_sign)
1204 emit_line_0(o, set, 0, reset, sign, line, len);
1205 else if (!ws) {
1206 /* Emit just the prefix, then the rest. */
1207 emit_line_0(o, set_sign ? set_sign : set, !!set_sign, reset,
1208 sign, "", 0);
1209 emit_line_0(o, set, 0, reset, 0, line, len);
1210 } else if (blank_at_eof)
1211 /* Blank line at EOF - paint '+' as well */
1212 emit_line_0(o, ws, 0, reset, sign, line, len);
1213 else {
1214 /* Emit just the prefix, then the rest. */
1215 emit_line_0(o, set_sign ? set_sign : set, !!set_sign, reset,
1216 sign, "", 0);
1217 ws_check_emit(line, len, ws_rule,
1218 o->file, set, reset, ws);
1219 }
1220}
1221static void emit_diff_symbol_from_struct(struct diff_options *o,
1223 struct emitted_diff_symbol *eds)
1224{
1225 static const char *nneof = " No newline at end of file\n";
1226 const char *context, *reset, *set, *set_sign, *meta, *fraginfo;
1227 struct strbuf sb = STRBUF_INIT;
1228 enum diff_symbol s = eds->s;
1230 const char *line = eds->line;
1231 int len = eds->len;
1232 unsigned flags = eds->flags;
1233 switch (s) {
1235 case DIFF_SYMBOL_NO_LF_EOF:
1236 context = diff_get_color_opt(o, DIFF_CONTEXT);
1237 reset = diff_get_color_opt(o, DIFF_RESET);
1238 putc('\n', o->file);
1239 emit_line_0(o, context, 0, reset, '\\',
1240 nneof, strlen(nneof));
1241 break;
1242 case DIFF_SYMBOL_SUBMODULE_HEADER:
1243 case DIFF_SYMBOL_SUBMODULE_ERROR:
1244 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1245 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1246 case DIFF_SYMBOL_SUMMARY:
1247 case DIFF_SYMBOL_STATS_LINE:
1248 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1249 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1250 emit_line(o, "", "", line, len);
1251 break;
1252 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1253 case DIFF_SYMBOL_CONTEXT_MARKER:
1254 context = diff_get_color_opt(o, DIFF_CONTEXT);
1255 reset = diff_get_color_opt(o, DIFF_RESET);
1256 emit_line(o, context, reset, line, len);
1257 break;
1258 case DIFF_SYMBOL_SEPARATOR:
1259 fprintf(o->file, "%s%c",
1260 diff_line_prefix(o),
1261 o->line_termination);
1262 break;
1263 case DIFF_SYMBOL_CONTEXT:
1264 set = diff_get_color_opt(o, DIFF_CONTEXT);
1265 reset = diff_get_color_opt(o, DIFF_RESET);
1266 set_sign = NULL;
1267 if (o->flags.dual_color_diffed_diffs) {
1268 char c = !len ? 0 : line[0];
1269 if (c == '+')
1271 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1272 else if (c == '@')
1273 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1274 else if (c == '-')
1275 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1276 }
1277 emit_line_ws_markup(o, set, reset, line, len, set_sign, ' ',
1278 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1279 break;
1280 case DIFF_SYMBOL_PLUS:
1281 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1282 DIFF_SYMBOL_MOVED_LINE_ALT |
1283 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1284 case DIFF_SYMBOL_MOVED_LINE |
1285 DIFF_SYMBOL_MOVED_LINE_ALT |
1286 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1287 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1288 break;
1289 case DIFF_SYMBOL_MOVED_LINE |
1290 DIFF_SYMBOL_MOVED_LINE_ALT:
1291 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1292 break;
1293 case DIFF_SYMBOL_MOVED_LINE |
1294 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1295 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1296 break;
1297 case DIFF_SYMBOL_MOVED_LINE:
1298 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1299 break;
1300 default:
1301 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1302 }
1303 reset = diff_get_color_opt(o, DIFF_RESET);
1304 if (!o->flags.dual_color_diffed_diffs)
1305 set_sign = NULL;
1306 else {
1307 char c = !len ? 0 : line[0];
1308 set_sign = set;
1310 if (c == '-')
1311 set = diff_get_color_opt(o, DIFF_FILE_OLD_BOLD);
1312 else if (c == '@')
1313 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1314 else if (c == '+')
1315 set = diff_get_color_opt(o, DIFF_FILE_NEW_BOLD);
1316 else
1317 set = diff_get_color_opt(o, DIFF_CONTEXT_BOLD);
1318 flags &= ~DIFF_SYMBOL_CONTENT_WS_MASK;
1319 }
1320 emit_line_ws_markup(o, set, reset, line, len, set_sign, '+',
1321 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1322 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1323 break;
1324 case DIFF_SYMBOL_MINUS:
1325 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1326 DIFF_SYMBOL_MOVED_LINE_ALT |
1327 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1328 case DIFF_SYMBOL_MOVED_LINE |
1329 DIFF_SYMBOL_MOVED_LINE_ALT |
1330 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1331 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1332 break;
1333 case DIFF_SYMBOL_MOVED_LINE |
1334 DIFF_SYMBOL_MOVED_LINE_ALT:
1335 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1336 break;
1337 case DIFF_SYMBOL_MOVED_LINE |
1338 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1339 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1340 break;
1341 case DIFF_SYMBOL_MOVED_LINE:
1342 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1343 break;
1344 default:
1345 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1346 }
1347 reset = diff_get_color_opt(o, DIFF_RESET);
1348 if (!o->flags.dual_color_diffed_diffs)
1349 set_sign = NULL;
1350 else {
1351 char c = !len ? 0 : line[0];
1352 set_sign = set;
1354 if (c == '+')
1355 set = diff_get_color_opt(o, DIFF_FILE_NEW_DIM);
1356 else if (c == '@')
1357 set = diff_get_color_opt(o, DIFF_FRAGINFO);
1358 else if (c == '-')
1359 set = diff_get_color_opt(o, DIFF_FILE_OLD_DIM);
1360 else
1361 set = diff_get_color_opt(o, DIFF_CONTEXT_DIM);
1362 }
1363 emit_line_ws_markup(o, set, reset, line, len, set_sign, '-',
1364 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1365 break;
1366 case DIFF_SYMBOL_WORDS_PORCELAIN:
1367 context = diff_get_color_opt(o, DIFF_CONTEXT);
1368 reset = diff_get_color_opt(o, DIFF_RESET);
1369 emit_line(o, context, reset, line, len);
1370 fputs("~\n", o->file);
1371 break;
1372 case DIFF_SYMBOL_WORDS:
1373 context = diff_get_color_opt(o, DIFF_CONTEXT);
1374 reset = diff_get_color_opt(o, DIFF_RESET);
1375 /*
1376 * Skip the prefix character, if any. With
1377 * diff_suppress_blank_empty, there may be
1378 * none.
1379 */
1380 if (line[0] != '\n') {
1381 line++;
1382 len--;
1383 }
1384 emit_line(o, context, reset, line, len);
1385 break;
1386 case DIFF_SYMBOL_FILEPAIR_PLUS:
1387 meta = diff_get_color_opt(o, DIFF_METAINFO);
1388 reset = diff_get_color_opt(o, DIFF_RESET);
1389 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1390 line, reset,
1391 strchr(line, ' ') ? "\t" : "");
1392 break;
1393 case DIFF_SYMBOL_FILEPAIR_MINUS:
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_BINARY_FILES:
1401 case DIFF_SYMBOL_HEADER:
1402 fprintf(o->file, "%s", line);
1403 break;
1404 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1405 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1406 break;
1407 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1408 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1409 break;
1410 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1411 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1412 break;
1413 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1414 fputs(diff_line_prefix(o), o->file);
1415 fputc('\n', o->file);
1416 break;
1417 case DIFF_SYMBOL_REWRITE_DIFF:
1418 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1419 reset = diff_get_color_opt(o, DIFF_RESET);
1420 emit_line(o, fraginfo, reset, line, len);
1421 break;
1422 case DIFF_SYMBOL_SUBMODULE_ADD:
1423 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1424 reset = diff_get_color_opt(o, DIFF_RESET);
1425 emit_line(o, set, reset, line, len);
1426 break;
1427 case DIFF_SYMBOL_SUBMODULE_DEL:
1428 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1429 reset = diff_get_color_opt(o, DIFF_RESET);
1430 emit_line(o, set, reset, line, len);
1431 break;
1432 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1433 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1434 diff_line_prefix(o), line);
1435 break;
1436 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1437 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1438 diff_line_prefix(o), line);
1439 break;
1440 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1441 emit_line(o, "", "", " 0 files changed\n",
1442 strlen(" 0 files changed\n"));
1443 break;
1444 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1445 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1446 break;
1447 case DIFF_SYMBOL_WORD_DIFF:
1448 fprintf(o->file, "%.*s", len, line);
1449 break;
1450 case DIFF_SYMBOL_STAT_SEP:
1451 fputs(o->stat_sep, o->file);
1452 break;
1453 default:
1454 BUG("unknown diff symbol");
1455 }
1456 strbuf_release(&sb);
1457}
1458static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1460 const char *line, int len, unsigned flags)
1461{
1462 struct emitted_diff_symbol e = {line, len, flags, s};
1463 if (o->emitted_symbols)
1465 append_emitted_diff_symbol(o, &e);
1466 else
1467 emit_diff_symbol_from_struct(o, &e);
1468}
1469void diff_emit_submodule_del(struct diff_options *o, const char *line)
1471{
1472 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1473}
1474void diff_emit_submodule_add(struct diff_options *o, const char *line)
1476{
1477 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1478}
1479void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1481{
1482 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1483 path, strlen(path), 0);
1484}
1485void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1487{
1488 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1489 path, strlen(path), 0);
1490}
1491void diff_emit_submodule_header(struct diff_options *o, const char *header)
1493{
1494 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1495 header, strlen(header), 0);
1496}
1497void diff_emit_submodule_error(struct diff_options *o, const char *err)
1499{
1500 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1501}
1502void diff_emit_submodule_pipethrough(struct diff_options *o,
1504 const char *line, int len)
1505{
1506 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1507}
1508static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1510{
1511 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1512 ecbdata->blank_at_eof_in_preimage &&
1513 ecbdata->blank_at_eof_in_postimage &&
1514 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1515 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1516 return 0;
1517 return ws_blank_line(line, len, ecbdata->ws_rule);
1518}
1519static void emit_add_line(const char *reset,
1521 struct emit_callback *ecbdata,
1522 const char *line, int len)
1523{
1524 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1525 if (new_blank_line_at_eof(ecbdata, line, len))
1526 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1527 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1529}
1530static void emit_del_line(const char *reset,
1532 struct emit_callback *ecbdata,
1533 const char *line, int len)
1534{
1535 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1536 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1537}
1538static void emit_context_line(const char *reset,
1540 struct emit_callback *ecbdata,
1541 const char *line, int len)
1542{
1543 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1544 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1545}
1546static void emit_hunk_header(struct emit_callback *ecbdata,
1548 const char *line, int len)
1549{
1550 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1551 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1552 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1553 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1554 const char *reverse = ecbdata->color_diff ? GIT_COLOR_REVERSE : "";
1555 static const char atat[2] = { '@', '@' };
1556 const char *cp, *ep;
1557 struct strbuf msgbuf = STRBUF_INIT;
1558 int org_len = len;
1559 int i = 1;
1560 /*
1562 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1563 * it always is at least 10 bytes long.
1564 */
1565 if (len < 10 ||
1566 memcmp(line, atat, 2) ||
1567 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1568 emit_diff_symbol(ecbdata->opt,
1569 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1570 return;
1571 }
1572 ep += 2; /* skip over @@ */
1573 /* The hunk header in fraginfo color */
1575 if (ecbdata->opt->flags.dual_color_diffed_diffs)
1576 strbuf_addstr(&msgbuf, reverse);
1577 strbuf_addstr(&msgbuf, frag);
1578 strbuf_add(&msgbuf, line, ep - line);
1579 strbuf_addstr(&msgbuf, reset);
1580 /*
1582 * trailing "\r\n"
1583 */
1584 for ( ; i < 3; i++)
1585 if (line[len - i] == '\r' || line[len - i] == '\n')
1586 len--;
1587 /* blank before the func header */
1589 for (cp = ep; ep - line < len; ep++)
1590 if (*ep != ' ' && *ep != '\t')
1591 break;
1592 if (ep != cp) {
1593 strbuf_addstr(&msgbuf, context);
1594 strbuf_add(&msgbuf, cp, ep - cp);
1595 strbuf_addstr(&msgbuf, reset);
1596 }
1597 if (ep < line + len) {
1599 strbuf_addstr(&msgbuf, func);
1600 strbuf_add(&msgbuf, ep, line + len - ep);
1601 strbuf_addstr(&msgbuf, reset);
1602 }
1603 strbuf_add(&msgbuf, line + len, org_len - len);
1605 strbuf_complete_line(&msgbuf);
1606 emit_diff_symbol(ecbdata->opt,
1607 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1608 strbuf_release(&msgbuf);
1609}
1610static struct diff_tempfile *claim_diff_tempfile(void) {
1612 int i;
1613 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1614 if (!diff_temp[i].name)
1615 return diff_temp + i;
1616 BUG("diff is failing to clean up its tempfiles");
1617}
1618static void remove_tempfile(void)
1620{
1621 int i;
1622 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1623 if (is_tempfile_active(diff_temp[i].tempfile))
1624 delete_tempfile(&diff_temp[i].tempfile);
1625 diff_temp[i].name = NULL;
1626 }
1627}
1628static void add_line_count(struct strbuf *out, int count)
1630{
1631 switch (count) {
1632 case 0:
1633 strbuf_addstr(out, "0,0");
1634 break;
1635 case 1:
1636 strbuf_addstr(out, "1");
1637 break;
1638 default:
1639 strbuf_addf(out, "1,%d", count);
1640 break;
1641 }
1642}
1643static void emit_rewrite_lines(struct emit_callback *ecb,
1645 int prefix, const char *data, int size)
1646{
1647 const char *endp = NULL;
1648 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1649 while (0 < size) {
1651 int len;
1652 endp = memchr(data, '\n', size);
1654 len = endp ? (endp - data + 1) : size;
1655 if (prefix != '+') {
1656 ecb->lno_in_preimage++;
1657 emit_del_line(reset, ecb, data, len);
1658 } else {
1659 ecb->lno_in_postimage++;
1660 emit_add_line(reset, ecb, data, len);
1661 }
1662 size -= len;
1663 data += len;
1664 }
1665 if (!endp)
1666 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1667}
1668static void emit_rewrite_diff(const char *name_a,
1670 const char *name_b,
1671 struct diff_filespec *one,
1672 struct diff_filespec *two,
1673 struct userdiff_driver *textconv_one,
1674 struct userdiff_driver *textconv_two,
1675 struct diff_options *o)
1676{
1677 int lc_a, lc_b;
1678 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1679 const char *a_prefix, *b_prefix;
1680 char *data_one, *data_two;
1681 size_t size_one, size_two;
1682 struct emit_callback ecbdata;
1683 struct strbuf out = STRBUF_INIT;
1684 if (diff_mnemonic_prefix && o->flags.reverse_diff) {
1686 a_prefix = o->b_prefix;
1687 b_prefix = o->a_prefix;
1688 } else {
1689 a_prefix = o->a_prefix;
1690 b_prefix = o->b_prefix;
1691 }
1692 name_a += (*name_a == '/');
1694 name_b += (*name_b == '/');
1695 strbuf_reset(&a_name);
1697 strbuf_reset(&b_name);
1698 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1699 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1700 size_one = fill_textconv(textconv_one, one, &data_one);
1702 size_two = fill_textconv(textconv_two, two, &data_two);
1703 memset(&ecbdata, 0, sizeof(ecbdata));
1705 ecbdata.color_diff = want_color(o->use_color);
1706 ecbdata.ws_rule = whitespace_rule(name_b);
1707 ecbdata.opt = o;
1708 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1709 mmfile_t mf1, mf2;
1710 mf1.ptr = (char *)data_one;
1711 mf2.ptr = (char *)data_two;
1712 mf1.size = size_one;
1713 mf2.size = size_two;
1714 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1715 }
1716 ecbdata.lno_in_preimage = 1;
1717 ecbdata.lno_in_postimage = 1;
1718 lc_a = count_lines(data_one, size_one);
1720 lc_b = count_lines(data_two, size_two);
1721 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1723 a_name.buf, a_name.len, 0);
1724 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1725 b_name.buf, b_name.len, 0);
1726 strbuf_addstr(&out, "@@ -");
1728 if (!o->irreversible_delete)
1729 add_line_count(&out, lc_a);
1730 else
1731 strbuf_addstr(&out, "?,?");
1732 strbuf_addstr(&out, " +");
1733 add_line_count(&out, lc_b);
1734 strbuf_addstr(&out, " @@\n");
1735 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1736 strbuf_release(&out);
1737 if (lc_a && !o->irreversible_delete)
1739 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1740 if (lc_b)
1741 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1742 if (textconv_one)
1743 free((char *)data_one);
1744 if (textconv_two)
1745 free((char *)data_two);
1746}
1747struct diff_words_buffer {
1749 mmfile_t text;
1750 unsigned long alloc;
1751 struct diff_words_orig {
1752 const char *begin, *end;
1753 } *orig;
1754 int orig_nr, orig_alloc;
1755};
1756static void diff_words_append(char *line, unsigned long len,
1758 struct diff_words_buffer *buffer)
1759{
1760 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1761 line++;
1762 len--;
1763 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1764 buffer->text.size += len;
1765 buffer->text.ptr[buffer->text.size] = '\0';
1766}
1767struct diff_words_style_elem {
1769 const char *prefix;
1770 const char *suffix;
1771 const char *color; /* NULL; filled in by the setup code if
1772 * color is enabled */
1773};
1774struct diff_words_style {
1776 enum diff_words_type type;
1777 struct diff_words_style_elem new_word, old_word, ctx;
1778 const char *newline;
1779};
1780static struct diff_words_style diff_words_styles[] = {
1782 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1783 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1784 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1785};
1786struct diff_words_data {
1788 struct diff_words_buffer minus, plus;
1789 const char *current_plus;
1790 int last_minus;
1791 struct diff_options *opt;
1792 regex_t *word_regex;
1793 enum diff_words_type type;
1794 struct diff_words_style *style;
1795};
1796static int fn_out_diff_words_write_helper(struct diff_options *o,
1798 struct diff_words_style_elem *st_el,
1799 const char *newline,
1800 size_t count, const char *buf)
1801{
1802 int print = 0;
1803 struct strbuf sb = STRBUF_INIT;
1804 while (count) {
1806 char *p = memchr(buf, '\n', count);
1807 if (print)
1808 strbuf_addstr(&sb, diff_line_prefix(o));
1809 if (p != buf) {
1811 const char *reset = st_el->color && *st_el->color ?
1812 GIT_COLOR_RESET : NULL;
1813 if (st_el->color && *st_el->color)
1814 strbuf_addstr(&sb, st_el->color);
1815 strbuf_addstr(&sb, st_el->prefix);
1816 strbuf_add(&sb, buf, p ? p - buf : count);
1817 strbuf_addstr(&sb, st_el->suffix);
1818 if (reset)
1819 strbuf_addstr(&sb, reset);
1820 }
1821 if (!p)
1822 goto out;
1823 strbuf_addstr(&sb, newline);
1825 count -= p + 1 - buf;
1826 buf = p + 1;
1827 print = 1;
1828 if (count) {
1829 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1830 sb.buf, sb.len, 0);
1831 strbuf_reset(&sb);
1832 }
1833 }
1834out:
1836 if (sb.len)
1837 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1838 sb.buf, sb.len, 0);
1839 strbuf_release(&sb);
1840 return 0;
1841}
1842/*
1844 * '--color-words' algorithm can be described as:
1845 *
1846 * 1. collect the minus/plus lines of a diff hunk, divided into
1847 * minus-lines and plus-lines;
1848 *
1849 * 2. break both minus-lines and plus-lines into words and
1850 * place them into two mmfile_t with one word for each line;
1851 *
1852 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1853 *
1854 * And for the common parts of the both file, we output the plus side text.
1855 * diff_words->current_plus is used to trace the current position of the plus file
1856 * which printed. diff_words->last_minus is used to trace the last minus word
1857 * printed.
1858 *
1859 * For '--graph' to work with '--color-words', we need to output the graph prefix
1860 * on each line of color words output. Generally, there are two conditions on
1861 * which we should output the prefix.
1862 *
1863 * 1. diff_words->last_minus == 0 &&
1864 * diff_words->current_plus == diff_words->plus.text.ptr
1865 *
1866 * that is: the plus text must start as a new line, and if there is no minus
1867 * word printed, a graph prefix must be printed.
1868 *
1869 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1870 * *(diff_words->current_plus - 1) == '\n'
1871 *
1872 * that is: a graph prefix must be printed following a '\n'
1873 */
1874static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1875{
1876 if ((diff_words->last_minus == 0 &&
1877 diff_words->current_plus == diff_words->plus.text.ptr) ||
1878 (diff_words->current_plus > diff_words->plus.text.ptr &&
1879 *(diff_words->current_plus - 1) == '\n')) {
1880 return 1;
1881 } else {
1882 return 0;
1883 }
1884}
1885static void fn_out_diff_words_aux(void *priv,
1887 long minus_first, long minus_len,
1888 long plus_first, long plus_len,
1889 const char *func, long funclen)
1890{
1891 struct diff_words_data *diff_words = priv;
1892 struct diff_words_style *style = diff_words->style;
1893 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1894 struct diff_options *opt = diff_words->opt;
1895 const char *line_prefix;
1896 assert(opt);
1898 line_prefix = diff_line_prefix(opt);
1899 /* POSIX requires that first be decremented by one if len == 0... */
1901 if (minus_len) {
1902 minus_begin = diff_words->minus.orig[minus_first].begin;
1903 minus_end =
1904 diff_words->minus.orig[minus_first + minus_len - 1].end;
1905 } else
1906 minus_begin = minus_end =
1907 diff_words->minus.orig[minus_first].end;
1908 if (plus_len) {
1910 plus_begin = diff_words->plus.orig[plus_first].begin;
1911 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1912 } else
1913 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1914 if (color_words_output_graph_prefix(diff_words)) {
1916 fputs(line_prefix, diff_words->opt->file);
1917 }
1918 if (diff_words->current_plus != plus_begin) {
1919 fn_out_diff_words_write_helper(diff_words->opt,
1920 &style->ctx, style->newline,
1921 plus_begin - diff_words->current_plus,
1922 diff_words->current_plus);
1923 }
1924 if (minus_begin != minus_end) {
1925 fn_out_diff_words_write_helper(diff_words->opt,
1926 &style->old_word, style->newline,
1927 minus_end - minus_begin, minus_begin);
1928 }
1929 if (plus_begin != plus_end) {
1930 fn_out_diff_words_write_helper(diff_words->opt,
1931 &style->new_word, style->newline,
1932 plus_end - plus_begin, plus_begin);
1933 }
1934 diff_words->current_plus = plus_end;
1936 diff_words->last_minus = minus_first;
1937}
1938/* This function starts looking at *begin, and returns 0 iff a word was found. */
1940static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1941 int *begin, int *end)
1942{
1943 if (word_regex && *begin < buffer->size) {
1944 regmatch_t match[1];
1945 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1946 buffer->size - *begin, 1, match, 0)) {
1947 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1948 '\n', match[0].rm_eo - match[0].rm_so);
1949 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1950 *begin += match[0].rm_so;
1951 return *begin >= *end;
1952 }
1953 return -1;
1954 }
1955 /* find the next word */
1957 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1958 (*begin)++;
1959 if (*begin >= buffer->size)
1960 return -1;
1961 /* find the end of the word */
1963 *end = *begin + 1;
1964 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1965 (*end)++;
1966 return 0;
1968}
1969/*
1971 * This function splits the words in buffer->text, stores the list with
1972 * newline separator into out, and saves the offsets of the original words
1973 * in buffer->orig.
1974 */
1975static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1976 regex_t *word_regex)
1977{
1978 int i, j;
1979 long alloc = 0;
1980 out->size = 0;
1982 out->ptr = NULL;
1983 /* fake an empty "0th" word */
1985 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1986 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1987 buffer->orig_nr = 1;
1988 for (i = 0; i < buffer->text.size; i++) {
1990 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1991 return;
1992 /* store original boundaries */
1994 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1995 buffer->orig_alloc);
1996 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1997 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1998 buffer->orig_nr++;
1999 /* store one word */
2001 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2002 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
2003 out->ptr[out->size + j - i] = '\n';
2004 out->size += j - i + 1;
2005 i = j - 1;
2007 }
2008}
2009/* this executes the word diff on the accumulated buffers */
2011static void diff_words_show(struct diff_words_data *diff_words)
2012{
2013 xpparam_t xpp;
2014 xdemitconf_t xecfg;
2015 mmfile_t minus, plus;
2016 struct diff_words_style *style = diff_words->style;
2017 struct diff_options *opt = diff_words->opt;
2019 const char *line_prefix;
2020 assert(opt);
2022 line_prefix = diff_line_prefix(opt);
2023 /* special case: only removal */
2025 if (!diff_words->plus.text.size) {
2026 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2027 line_prefix, strlen(line_prefix), 0);
2028 fn_out_diff_words_write_helper(diff_words->opt,
2029 &style->old_word, style->newline,
2030 diff_words->minus.text.size,
2031 diff_words->minus.text.ptr);
2032 diff_words->minus.text.size = 0;
2033 return;
2034 }
2035 diff_words->current_plus = diff_words->plus.text.ptr;
2037 diff_words->last_minus = 0;
2038 memset(&xpp, 0, sizeof(xpp));
2040 memset(&xecfg, 0, sizeof(xecfg));
2041 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
2042 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
2043 xpp.flags = 0;
2044 /* as only the hunk header will be parsed, we need a 0-context */
2045 xecfg.ctxlen = 0;
2046 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, NULL,
2047 diff_words, &xpp, &xecfg))
2048 die("unable to generate word diff");
2049 free(minus.ptr);
2050 free(plus.ptr);
2051 if (diff_words->current_plus != diff_words->plus.text.ptr +
2052 diff_words->plus.text.size) {
2053 if (color_words_output_graph_prefix(diff_words))
2054 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
2055 line_prefix, strlen(line_prefix), 0);
2056 fn_out_diff_words_write_helper(diff_words->opt,
2057 &style->ctx, style->newline,
2058 diff_words->plus.text.ptr + diff_words->plus.text.size
2059 - diff_words->current_plus, diff_words->current_plus);
2060 }
2061 diff_words->minus.text.size = diff_words->plus.text.size = 0;
2062}
2063/* In "color-words" mode, show word-diff of words accumulated in the buffer */
2065static void diff_words_flush(struct emit_callback *ecbdata)
2066{
2067 struct diff_options *wo = ecbdata->diff_words->opt;
2068 if (ecbdata->diff_words->minus.text.size ||
2070 ecbdata->diff_words->plus.text.size)
2071 diff_words_show(ecbdata->diff_words);
2072 if (wo->emitted_symbols) {
2074 struct diff_options *o = ecbdata->opt;
2075 struct emitted_diff_symbols *wol = wo->emitted_symbols;
2076 int i;
2077 /*
2079 * NEEDSWORK:
2080 * Instead of appending each, concat all words to a line?
2081 */
2082 for (i = 0; i < wol->nr; i++)
2083 append_emitted_diff_symbol(o, &wol->buf[i]);
2084 for (i = 0; i < wol->nr; i++)
2086 free((void *)wol->buf[i].line);
2087 wol->nr = 0;
2089 }
2090}
2091static void diff_filespec_load_driver(struct diff_filespec *one)
2093{
2094 /* Use already-loaded driver */
2095 if (one->driver)
2096 return;
2097 if (S_ISREG(one->mode))
2099 one->driver = userdiff_find_by_path(one->path);
2100 /* Fallback to default settings */
2102 if (!one->driver)
2103 one->driver = userdiff_find_by_name("default");
2104}
2105static const char *userdiff_word_regex(struct diff_filespec *one)
2107{
2108 diff_filespec_load_driver(one);
2109 return one->driver->word_regex;
2110}
2111static void init_diff_words_data(struct emit_callback *ecbdata,
2113 struct diff_options *orig_opts,
2114 struct diff_filespec *one,
2115 struct diff_filespec *two)
2116{
2117 int i;
2118 struct diff_options *o = xmalloc(sizeof(struct diff_options));
2119 memcpy(o, orig_opts, sizeof(struct diff_options));
2120 ecbdata->diff_words =
2122 xcalloc(1, sizeof(struct diff_words_data));
2123 ecbdata->diff_words->type = o->word_diff;
2124 ecbdata->diff_words->opt = o;
2125 if (orig_opts->emitted_symbols)
2127 o->emitted_symbols =
2128 xcalloc(1, sizeof(struct emitted_diff_symbols));
2129 if (!o->word_regex)
2131 o->word_regex = userdiff_word_regex(one);
2132 if (!o->word_regex)
2133 o->word_regex = userdiff_word_regex(two);
2134 if (!o->word_regex)
2135 o->word_regex = diff_word_regex_cfg;
2136 if (o->word_regex) {
2137 ecbdata->diff_words->word_regex = (regex_t *)
2138 xmalloc(sizeof(regex_t));
2139 if (regcomp(ecbdata->diff_words->word_regex,
2140 o->word_regex,
2141 REG_EXTENDED | REG_NEWLINE))
2142 die("invalid regular expression: %s",
2143 o->word_regex);
2144 }
2145 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2146 if (o->word_diff == diff_words_styles[i].type) {
2147 ecbdata->diff_words->style =
2148 &diff_words_styles[i];
2149 break;
2150 }
2151 }
2152 if (want_color(o->use_color)) {
2153 struct diff_words_style *st = ecbdata->diff_words->style;
2154 st->old_word.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2155 st->new_word.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2156 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
2157 }
2158}
2159static void free_diff_words_data(struct emit_callback *ecbdata)
2161{
2162 if (ecbdata->diff_words) {
2163 diff_words_flush(ecbdata);
2164 free (ecbdata->diff_words->opt->emitted_symbols);
2165 free (ecbdata->diff_words->opt);
2166 free (ecbdata->diff_words->minus.text.ptr);
2167 free (ecbdata->diff_words->minus.orig);
2168 free (ecbdata->diff_words->plus.text.ptr);
2169 free (ecbdata->diff_words->plus.orig);
2170 if (ecbdata->diff_words->word_regex) {
2171 regfree(ecbdata->diff_words->word_regex);
2172 free(ecbdata->diff_words->word_regex);
2173 }
2174 FREE_AND_NULL(ecbdata->diff_words);
2175 }
2176}
2177const char *diff_get_color(int diff_use_color, enum color_diff ix)
2179{
2180 if (want_color(diff_use_color))
2181 return diff_colors[ix];
2182 return "";
2183}
2184const char *diff_line_prefix(struct diff_options *opt)
2186{
2187 struct strbuf *msgbuf;
2188 if (!opt->output_prefix)
2189 return "";
2190 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
2192 return msgbuf->buf;
2193}
2194static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
2196{
2197 const char *cp;
2198 unsigned long allot;
2199 size_t l = len;
2200 cp = line;
2202 allot = l;
2203 while (0 < l) {
2204 (void) utf8_width(&cp, &l);
2205 if (!cp)
2206 break; /* truncated in the middle? */
2207 }
2208 return allot - l;
2209}
2210static void find_lno(const char *line, struct emit_callback *ecbdata)
2212{
2213 const char *p;
2214 ecbdata->lno_in_preimage = 0;
2215 ecbdata->lno_in_postimage = 0;
2216 p = strchr(line, '-');
2217 if (!p)
2218 return; /* cannot happen */
2219 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2220 p = strchr(p, '+');
2221 if (!p)
2222 return; /* cannot happen */
2223 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2224}
2225static void fn_out_consume(void *priv, char *line, unsigned long len)
2227{
2228 struct emit_callback *ecbdata = priv;
2229 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2230 struct diff_options *o = ecbdata->opt;
2231 o->found_changes = 1;
2233 if (ecbdata->header) {
2235 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2236 ecbdata->header->buf, ecbdata->header->len, 0);
2237 strbuf_reset(ecbdata->header);
2238 ecbdata->header = NULL;
2239 }
2240 if (ecbdata->label_path[0]) {
2242 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2243 ecbdata->label_path[0],
2244 strlen(ecbdata->label_path[0]), 0);
2245 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2246 ecbdata->label_path[1],
2247 strlen(ecbdata->label_path[1]), 0);
2248 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2249 }
2250 if (diff_suppress_blank_empty
2252 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2253 line[0] = '\n';
2254 len = 1;
2255 }
2256 if (line[0] == '@') {
2258 if (ecbdata->diff_words)
2259 diff_words_flush(ecbdata);
2260 len = sane_truncate_line(ecbdata, line, len);
2261 find_lno(line, ecbdata);
2262 emit_hunk_header(ecbdata, line, len);
2263 return;
2264 }
2265 if (ecbdata->diff_words) {
2267 enum diff_symbol s =
2268 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2269 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2270 if (line[0] == '-') {
2271 diff_words_append(line, len,
2272 &ecbdata->diff_words->minus);
2273 return;
2274 } else if (line[0] == '+') {
2275 diff_words_append(line, len,
2276 &ecbdata->diff_words->plus);
2277 return;
2278 } else if (starts_with(line, "\\ ")) {
2279 /*
2280 * Eat the "no newline at eof" marker as if we
2281 * saw a "+" or "-" line with nothing on it,
2282 * and return without diff_words_flush() to
2283 * defer processing. If this is the end of
2284 * preimage, more "+" lines may come after it.
2285 */
2286 return;
2287 }
2288 diff_words_flush(ecbdata);
2289 emit_diff_symbol(o, s, line, len, 0);
2290 return;
2291 }
2292 switch (line[0]) {
2294 case '+':
2295 ecbdata->lno_in_postimage++;
2296 emit_add_line(reset, ecbdata, line + 1, len - 1);
2297 break;
2298 case '-':
2299 ecbdata->lno_in_preimage++;
2300 emit_del_line(reset, ecbdata, line + 1, len - 1);
2301 break;
2302 case ' ':
2303 ecbdata->lno_in_postimage++;
2304 ecbdata->lno_in_preimage++;
2305 emit_context_line(reset, ecbdata, line + 1, len - 1);
2306 break;
2307 default:
2308 /* incomplete line at the end */
2309 ecbdata->lno_in_preimage++;
2310 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2311 line, len, 0);
2312 break;
2313 }
2314}
2315static void pprint_rename(struct strbuf *name, const char *a, const char *b)
2317{
2318 const char *old_name = a;
2319 const char *new_name = b;
2320 int pfx_length, sfx_length;
2321 int pfx_adjust_for_slash;
2322 int len_a = strlen(a);
2323 int len_b = strlen(b);
2324 int a_midlen, b_midlen;
2325 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2326 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2327 if (qlen_a || qlen_b) {
2329 quote_c_style(a, name, NULL, 0);
2330 strbuf_addstr(name, " => ");
2331 quote_c_style(b, name, NULL, 0);
2332 return;
2333 }
2334 /* Find common prefix */
2336 pfx_length = 0;
2337 while (*old_name && *new_name && *old_name == *new_name) {
2338 if (*old_name == '/')
2339 pfx_length = old_name - a + 1;
2340 old_name++;
2341 new_name++;
2342 }
2343 /* Find common suffix */
2345 old_name = a + len_a;
2346 new_name = b + len_b;
2347 sfx_length = 0;
2348 /*
2349 * If there is a common prefix, it must end in a slash. In
2350 * that case we let this loop run 1 into the prefix to see the
2351 * same slash.
2352 *
2353 * If there is no common prefix, we cannot do this as it would
2354 * underrun the input strings.
2355 */
2356 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2357 while (a + pfx_length - pfx_adjust_for_slash <= old_name &&
2358 b + pfx_length - pfx_adjust_for_slash <= new_name &&
2359 *old_name == *new_name) {
2360 if (*old_name == '/')
2361 sfx_length = len_a - (old_name - a);
2362 old_name--;
2363 new_name--;
2364 }
2365 /*
2367 * pfx{mid-a => mid-b}sfx
2368 * {pfx-a => pfx-b}sfx
2369 * pfx{sfx-a => sfx-b}
2370 * name-a => name-b
2371 */
2372 a_midlen = len_a - pfx_length - sfx_length;
2373 b_midlen = len_b - pfx_length - sfx_length;
2374 if (a_midlen < 0)
2375 a_midlen = 0;
2376 if (b_midlen < 0)
2377 b_midlen = 0;
2378 strbuf_grow(name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2380 if (pfx_length + sfx_length) {
2381 strbuf_add(name, a, pfx_length);
2382 strbuf_addch(name, '{');
2383 }
2384 strbuf_add(name, a + pfx_length, a_midlen);
2385 strbuf_addstr(name, " => ");
2386 strbuf_add(name, b + pfx_length, b_midlen);
2387 if (pfx_length + sfx_length) {
2388 strbuf_addch(name, '}');
2389 strbuf_add(name, a + len_a - sfx_length, sfx_length);
2390 }
2391}
2392struct diffstat_t {
2394 int nr;
2395 int alloc;
2396 struct diffstat_file {
2397 char *from_name;
2398 char *name;
2399 char *print_name;
2400 const char *comments;
2401 unsigned is_unmerged:1;
2402 unsigned is_binary:1;
2403 unsigned is_renamed:1;
2404 unsigned is_interesting:1;
2405 uintmax_t added, deleted;
2406 } **files;
2407};
2408static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2410 const char *name_a,
2411 const char *name_b)
2412{
2413 struct diffstat_file *x;
2414 x = xcalloc(1, sizeof(*x));
2415 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2416 diffstat->files[diffstat->nr++] = x;
2417 if (name_b) {
2418 x->from_name = xstrdup(name_a);
2419 x->name = xstrdup(name_b);
2420 x->is_renamed = 1;
2421 }
2422 else {
2423 x->from_name = NULL;
2424 x->name = xstrdup(name_a);
2425 }
2426 return x;
2427}
2428static void diffstat_consume(void *priv, char *line, unsigned long len)
2430{
2431 struct diffstat_t *diffstat = priv;
2432 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2433 if (line[0] == '+')
2435 x->added++;
2436 else if (line[0] == '-')
2437 x->deleted++;
2438}
2439const char mime_boundary_leader[] = "------------";
2441static int scale_linear(int it, int width, int max_change)
2443{
2444 if (!it)
2445 return 0;
2446 /*
2447 * make sure that at least one '-' or '+' is printed if
2448 * there is any change to this path. The easiest way is to
2449 * scale linearly as if the alloted width is one column shorter
2450 * than it is, and then add 1 to the result.
2451 */
2452 return 1 + (it * (width - 1) / max_change);
2453}
2454static void show_graph(struct strbuf *out, char ch, int cnt,
2456 const char *set, const char *reset)
2457{
2458 if (cnt <= 0)
2459 return;
2460 strbuf_addstr(out, set);
2461 strbuf_addchars(out, ch, cnt);
2462 strbuf_addstr(out, reset);
2463}
2464static void fill_print_name(struct diffstat_file *file)
2466{
2467 struct strbuf pname = STRBUF_INIT;
2468 if (file->print_name)
2470 return;
2471 if (file->is_renamed)
2473 pprint_rename(&pname, file->from_name, file->name);
2474 else
2475 quote_c_style(file->name, &pname, NULL, 0);
2476 if (file->comments)
2478 strbuf_addf(&pname, " (%s)", file->comments);
2479 file->print_name = strbuf_detach(&pname, NULL);
2481}
2482static void print_stat_summary_inserts_deletes(struct diff_options *options,
2484 int files, int insertions, int deletions)
2485{
2486 struct strbuf sb = STRBUF_INIT;
2487 if (!files) {
2489 assert(insertions == 0 && deletions == 0);
2490 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2491 NULL, 0, 0);
2492 return;
2493 }
2494 strbuf_addf(&sb,
2496 (files == 1) ? " %d file changed" : " %d files changed",
2497 files);
2498 /*
2500 * For binary diff, the caller may want to print "x files
2501 * changed" with insertions == 0 && deletions == 0.
2502 *
2503 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2504 * is probably less confusing (i.e skip over "2 files changed
2505 * but nothing about added/removed lines? Is this a bug in Git?").
2506 */
2507 if (insertions || deletions == 0) {
2508 strbuf_addf(&sb,
2509 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2510 insertions);
2511 }
2512 if (deletions || insertions == 0) {
2514 strbuf_addf(&sb,
2515 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2516 deletions);
2517 }
2518 strbuf_addch(&sb, '\n');
2519 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2520 sb.buf, sb.len, 0);
2521 strbuf_release(&sb);
2522}
2523void print_stat_summary(FILE *fp, int files,
2525 int insertions, int deletions)
2526{
2527 struct diff_options o;
2528 memset(&o, 0, sizeof(o));
2529 o.file = fp;
2530 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2532}
2533static void show_stats(struct diffstat_t *data, struct diff_options *options)
2535{
2536 int i, len, add, del, adds = 0, dels = 0;
2537 uintmax_t max_change = 0, max_len = 0;
2538 int total_files = data->nr, count;
2539 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2540 const char *reset, *add_c, *del_c;
2541 int extra_shown = 0;
2542 const char *line_prefix = diff_line_prefix(options);
2543 struct strbuf out = STRBUF_INIT;
2544 if (data->nr == 0)
2546 return;
2547 count = options->stat_count ? options->stat_count : data->nr;
2549 reset = diff_get_color_opt(options, DIFF_RESET);
2551 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2552 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2553 /*
2555 * Find the longest filename and max number of changes
2556 */
2557 for (i = 0; (i < count) && (i < data->nr); i++) {
2558 struct diffstat_file *file = data->files[i];
2559 uintmax_t change = file->added + file->deleted;
2560 if (!file->is_interesting && (change == 0)) {
2562 count++; /* not shown == room for one more */
2563 continue;
2564 }
2565 fill_print_name(file);
2566 len = strlen(file->print_name);
2567 if (max_len < len)
2568 max_len = len;
2569 if (file->is_unmerged) {
2571 /* "Unmerged" is 8 characters */
2572 bin_width = bin_width < 8 ? 8 : bin_width;
2573 continue;
2574 }
2575 if (file->is_binary) {
2576 /* "Bin XXX -> YYY bytes" */
2577 int w = 14 + decimal_width(file->added)
2578 + decimal_width(file->deleted);
2579 bin_width = bin_width < w ? w : bin_width;
2580 /* Display change counts aligned with "Bin" */
2581 number_width = 3;
2582 continue;
2583 }
2584 if (max_change < change)
2586 max_change = change;
2587 }
2588 count = i; /* where we can stop scanning in data->files[] */
2589 /*
2591 * We have width = stat_width or term_columns() columns total.
2592 * We want a maximum of min(max_len, stat_name_width) for the name part.
2593 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2594 * We also need 1 for " " and 4 + decimal_width(max_change)
2595 * for " | NNNN " and one the empty column at the end, altogether
2596 * 6 + decimal_width(max_change).
2597 *
2598 * If there's not enough space, we will use the smaller of
2599 * stat_name_width (if set) and 5/8*width for the filename,
2600 * and the rest for constant elements + graph part, but no more
2601 * than stat_graph_width for the graph part.
2602 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2603 * for the standard terminal size).
2604 *
2605 * In other words: stat_width limits the maximum width, and
2606 * stat_name_width fixes the maximum width of the filename,
2607 * and is also used to divide available columns if there
2608 * aren't enough.
2609 *
2610 * Binary files are displayed with "Bin XXX -> YYY bytes"
2611 * instead of the change count and graph. This part is treated
2612 * similarly to the graph part, except that it is not
2613 * "scaled". If total width is too small to accommodate the
2614 * guaranteed minimum width of the filename part and the
2615 * separators and this message, this message will "overflow"
2616 * making the line longer than the maximum width.
2617 */
2618 if (options->stat_width == -1)
2620 width = term_columns() - strlen(line_prefix);
2621 else
2622 width = options->stat_width ? options->stat_width : 80;
2623 number_width = decimal_width(max_change) > number_width ?
2624 decimal_width(max_change) : number_width;
2625 if (options->stat_graph_width == -1)
2627 options->stat_graph_width = diff_stat_graph_width;
2628 /*
2630 * Guarantee 3/8*16==6 for the graph part
2631 * and 5/8*16==10 for the filename part
2632 */
2633 if (width < 16 + 6 + number_width)
2634 width = 16 + 6 + number_width;
2635 /*
2637 * First assign sizes that are wanted, ignoring available width.
2638 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2639 * starting from "XXX" should fit in graph_width.
2640 */
2641 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2642 if (options->stat_graph_width &&
2643 options->stat_graph_width < graph_width)
2644 graph_width = options->stat_graph_width;
2645 name_width = (options->stat_name_width > 0 &&
2647 options->stat_name_width < max_len) ?
2648 options->stat_name_width : max_len;
2649 /*
2651 * Adjust adjustable widths not to exceed maximum width
2652 */
2653 if (name_width + number_width + 6 + graph_width > width) {
2654 if (graph_width > width * 3/8 - number_width - 6) {
2655 graph_width = width * 3/8 - number_width - 6;
2656 if (graph_width < 6)
2657 graph_width = 6;
2658 }
2659 if (options->stat_graph_width &&
2661 graph_width > options->stat_graph_width)
2662 graph_width = options->stat_graph_width;
2663 if (name_width > width - number_width - 6 - graph_width)
2664 name_width = width - number_width - 6 - graph_width;
2665 else
2666 graph_width = width - number_width - 6 - name_width;
2667 }
2668 /*
2670 * From here name_width is the width of the name area,
2671 * and graph_width is the width of the graph area.
2672 * max_change is used to scale graph properly.
2673 */
2674 for (i = 0; i < count; i++) {
2675 const char *prefix = "";
2676 struct diffstat_file *file = data->files[i];
2677 char *name = file->print_name;
2678 uintmax_t added = file->added;
2679 uintmax_t deleted = file->deleted;
2680 int name_len;
2681 if (!file->is_interesting && (added + deleted == 0))
2683 continue;
2684 /*
2686 * "scale" the filename
2687 */
2688 len = name_width;
2689 name_len = strlen(name);
2690 if (name_width < name_len) {
2691 char *slash;
2692 prefix = "...";
2693 len -= 3;
2694 name += name_len - len;
2695 slash = strchr(name, '/');
2696 if (slash)
2697 name = slash;
2698 }
2699 if (file->is_binary) {
2701 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2702 strbuf_addf(&out, " %*s", number_width, "Bin");
2703 if (!added && !deleted) {
2704 strbuf_addch(&out, '\n');
2705 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2706 out.buf, out.len, 0);
2707 strbuf_reset(&out);
2708 continue;
2709 }
2710 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2711 del_c, deleted, reset);
2712 strbuf_addstr(&out, " -> ");
2713 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2714 add_c, added, reset);
2715 strbuf_addstr(&out, " bytes\n");
2716 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2717 out.buf, out.len, 0);
2718 strbuf_reset(&out);
2719 continue;
2720 }
2721 else if (file->is_unmerged) {
2722 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2723 strbuf_addstr(&out, " Unmerged\n");
2724 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2725 out.buf, out.len, 0);
2726 strbuf_reset(&out);
2727 continue;
2728 }
2729 /*
2731 * scale the add/delete
2732 */
2733 add = added;
2734 del = deleted;
2735 if (graph_width <= max_change) {
2737 int total = scale_linear(add + del, graph_width, max_change);
2738 if (total < 2 && add && del)
2739 /* width >= 2 due to the sanity check */
2740 total = 2;
2741 if (add < del) {
2742 add = scale_linear(add, graph_width, max_change);
2743 del = total - add;
2744 } else {
2745 del = scale_linear(del, graph_width, max_change);
2746 add = total - del;
2747 }
2748 }
2749 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2750 strbuf_addf(&out, " %*"PRIuMAX"%s",
2751 number_width, added + deleted,
2752 added + deleted ? " " : "");
2753 show_graph(&out, '+', add, add_c, reset);
2754 show_graph(&out, '-', del, del_c, reset);
2755 strbuf_addch(&out, '\n');
2756 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2757 out.buf, out.len, 0);
2758 strbuf_reset(&out);
2759 }
2760 for (i = 0; i < data->nr; i++) {
2762 struct diffstat_file *file = data->files[i];
2763 uintmax_t added = file->added;
2764 uintmax_t deleted = file->deleted;
2765 if (file->is_unmerged ||
2767 (!file->is_interesting && (added + deleted == 0))) {
2768 total_files--;
2769 continue;
2770 }
2771 if (!file->is_binary) {
2773 adds += added;
2774 dels += deleted;
2775 }
2776 if (i < count)
2777 continue;
2778 if (!extra_shown)
2779 emit_diff_symbol(options,
2780 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2781 NULL, 0, 0);
2782 extra_shown = 1;
2783 }
2784 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2786 strbuf_release(&out);
2787}
2788static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2790{
2791 int i, adds = 0, dels = 0, total_files = data->nr;
2792 if (data->nr == 0)
2794 return;
2795 for (i = 0; i < data->nr; i++) {
2797 int added = data->files[i]->added;
2798 int deleted = data->files[i]->deleted;
2799 if (data->files[i]->is_unmerged ||
2801 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2802 total_files--;
2803 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2804 adds += added;
2805 dels += deleted;
2806 }
2807 }
2808 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2809}
2810static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2812{
2813 int i;
2814 if (data->nr == 0)
2816 return;
2817 for (i = 0; i < data->nr; i++) {
2819 struct diffstat_file *file = data->files[i];
2820 fprintf(options->file, "%s", diff_line_prefix(options));
2822 if (file->is_binary)
2824 fprintf(options->file, "-\t-\t");
2825 else
2826 fprintf(options->file,
2827 "%"PRIuMAX"\t%"PRIuMAX"\t",
2828 file->added, file->deleted);
2829 if (options->line_termination) {
2830 fill_print_name(file);
2831 if (!file->is_renamed)
2832 write_name_quoted(file->name, options->file,
2833 options->line_termination);
2834 else {
2835 fputs(file->print_name, options->file);
2836 putc(options->line_termination, options->file);
2837 }
2838 } else {
2839 if (file->is_renamed) {
2840 putc('\0', options->file);
2841 write_name_quoted(file->from_name, options->file, '\0');
2842 }
2843 write_name_quoted(file->name, options->file, '\0');
2844 }
2845 }
2846}
2847struct dirstat_file {
2849 const char *name;
2850 unsigned long changed;
2851};
2852struct dirstat_dir {
2854 struct dirstat_file *files;
2855 int alloc, nr, permille, cumulative;
2856};
2857static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2859 unsigned long changed, const char *base, int baselen)
2860{
2861 unsigned long sum_changes = 0;
2862 unsigned int sources = 0;
2863 const char *line_prefix = diff_line_prefix(opt);
2864 while (dir->nr) {
2866 struct dirstat_file *f = dir->files;
2867 int namelen = strlen(f->name);
2868 unsigned long changes;
2869 char *slash;
2870 if (namelen < baselen)
2872 break;
2873 if (memcmp(f->name, base, baselen))
2874 break;
2875 slash = strchr(f->name + baselen, '/');
2876 if (slash) {
2877 int newbaselen = slash + 1 - f->name;
2878 changes = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2879 sources++;
2880 } else {
2881 changes = f->changed;
2882 dir->files++;
2883 dir->nr--;
2884 sources += 2;
2885 }
2886 sum_changes += changes;
2887 }
2888 /*
2890 * We don't report dirstat's for
2891 * - the top level
2892 * - or cases where everything came from a single directory
2893 * under this directory (sources == 1).
2894 */
2895 if (baselen && sources != 1) {
2896 if (sum_changes) {
2897 int permille = sum_changes * 1000 / changed;
2898 if (permille >= dir->permille) {
2899 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2900 permille / 10, permille % 10, baselen, base);
2901 if (!dir->cumulative)
2902 return 0;
2903 }
2904 }
2905 }
2906 return sum_changes;
2907}
2908static int dirstat_compare(const void *_a, const void *_b)
2910{
2911 const struct dirstat_file *a = _a;
2912 const struct dirstat_file *b = _b;
2913 return strcmp(a->name, b->name);
2914}
2915static void show_dirstat(struct diff_options *options)
2917{
2918 int i;
2919 unsigned long changed;
2920 struct dirstat_dir dir;
2921 struct diff_queue_struct *q = &diff_queued_diff;
2922 dir.files = NULL;
2924 dir.alloc = 0;
2925 dir.nr = 0;
2926 dir.permille = options->dirstat_permille;
2927 dir.cumulative = options->flags.dirstat_cumulative;
2928 changed = 0;
2930 for (i = 0; i < q->nr; i++) {
2931 struct diff_filepair *p = q->queue[i];
2932 const char *name;
2933 unsigned long copied, added, damage;
2934 int content_changed;
2935 name = p->two->path ? p->two->path : p->one->path;
2937 if (p->one->oid_valid && p->two->oid_valid)
2939 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2940 else
2941 content_changed = 1;
2942 if (!content_changed) {
2944 /*
2945 * The SHA1 has not changed, so pre-/post-content is
2946 * identical. We can therefore skip looking at the
2947 * file contents altogether.
2948 */
2949 damage = 0;
2950 goto found_damage;
2951 }
2952 if (options->flags.dirstat_by_file) {
2954 /*
2955 * In --dirstat-by-file mode, we don't really need to
2956 * look at the actual file contents at all.
2957 * The fact that the SHA1 changed is enough for us to
2958 * add this file to the list of results
2959 * (with each file contributing equal damage).
2960 */
2961 damage = 1;
2962 goto found_damage;
2963 }
2964 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2966 diff_populate_filespec(p->one, 0);
2967 diff_populate_filespec(p->two, 0);
2968 diffcore_count_changes(p->one, p->two, NULL, NULL,
2969 &copied, &added);
2970 diff_free_filespec_data(p->one);
2971 diff_free_filespec_data(p->two);
2972 } else if (DIFF_FILE_VALID(p->one)) {
2973 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2974 copied = added = 0;
2975 diff_free_filespec_data(p->one);
2976 } else if (DIFF_FILE_VALID(p->two)) {
2977 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2978 copied = 0;
2979 added = p->two->size;
2980 diff_free_filespec_data(p->two);
2981 } else
2982 continue;
2983 /*
2985 * Original minus copied is the removed material,
2986 * added is the new material. They are both damages
2987 * made to the preimage.
2988 * If the resulting damage is zero, we know that
2989 * diffcore_count_changes() considers the two entries to
2990 * be identical, but since content_changed is true, we
2991 * know that there must have been _some_ kind of change,
2992 * so we force all entries to have damage > 0.
2993 */
2994 damage = (p->one->size - copied) + added;
2995 if (!damage)
2996 damage = 1;
2997found_damage:
2999 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3000 dir.files[dir.nr].name = name;
3001 dir.files[dir.nr].changed = damage;
3002 changed += damage;
3003 dir.nr++;
3004 }
3005 /* This can happen even with many files, if everything was renames */
3007 if (!changed)
3008 return;
3009 /* Show all directories with more than x% of the changes */
3011 QSORT(dir.files, dir.nr, dirstat_compare);
3012 gather_dirstat(options, &dir, changed, "", 0);
3013}
3014static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
3016{
3017 int i;
3018 unsigned long changed;
3019 struct dirstat_dir dir;
3020 if (data->nr == 0)
3022 return;
3023 dir.files = NULL;
3025 dir.alloc = 0;
3026 dir.nr = 0;
3027 dir.permille = options->dirstat_permille;
3028 dir.cumulative = options->flags.dirstat_cumulative;
3029 changed = 0;
3031 for (i = 0; i < data->nr; i++) {
3032 struct diffstat_file *file = data->files[i];
3033 unsigned long damage = file->added + file->deleted;
3034 if (file->is_binary)
3035 /*
3036 * binary files counts bytes, not lines. Must find some
3037 * way to normalize binary bytes vs. textual lines.
3038 * The following heuristic assumes that there are 64
3039 * bytes per "line".
3040 * This is stupid and ugly, but very cheap...
3041 */
3042 damage = DIV_ROUND_UP(damage, 64);
3043 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
3044 dir.files[dir.nr].name = file->name;
3045 dir.files[dir.nr].changed = damage;
3046 changed += damage;
3047 dir.nr++;
3048 }
3049 /* This can happen even with many files, if everything was renames */
3051 if (!changed)
3052 return;
3053 /* Show all directories with more than x% of the changes */
3055 QSORT(dir.files, dir.nr, dirstat_compare);
3056 gather_dirstat(options, &dir, changed, "", 0);
3057}
3058static void free_diffstat_info(struct diffstat_t *diffstat)
3060{
3061 int i;
3062 for (i = 0; i < diffstat->nr; i++) {
3063 struct diffstat_file *f = diffstat->files[i];
3064 free(f->print_name);
3065 free(f->name);
3066 free(f->from_name);
3067 free(f);
3068 }
3069 free(diffstat->files);
3070}
3071struct checkdiff_t {
3073 const char *filename;
3074 int lineno;
3075 int conflict_marker_size;
3076 struct diff_options *o;
3077 unsigned ws_rule;
3078 unsigned status;
3079};
3080static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
3082{
3083 char firstchar;
3084 int cnt;
3085 if (len < marker_size + 1)
3087 return 0;
3088 firstchar = line[0];
3089 switch (firstchar) {
3090 case '=': case '>': case '<': case '|':
3091 break;
3092 default:
3093 return 0;
3094 }
3095 for (cnt = 1; cnt < marker_size; cnt++)
3096 if (line[cnt] != firstchar)
3097 return 0;
3098 /* line[1] thru line[marker_size-1] are same as firstchar */
3099 if (len < marker_size + 1 || !isspace(line[marker_size]))
3100 return 0;
3101 return 1;
3102}
3103static void checkdiff_consume(void *priv, char *line, unsigned long len)
3105{
3106 struct checkdiff_t *data = priv;
3107 int marker_size = data->conflict_marker_size;
3108 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
3109 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
3110 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
3111 char *err;
3112 const char *line_prefix;
3113 assert(data->o);
3115 line_prefix = diff_line_prefix(data->o);
3116 if (line[0] == '+') {
3118 unsigned bad;
3119 data->lineno++;
3120 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
3121 data->status |= 1;
3122 fprintf(data->o->file,
3123 "%s%s:%d: leftover conflict marker\n",
3124 line_prefix, data->filename, data->lineno);
3125 }
3126 bad = ws_check(line + 1, len - 1, data->ws_rule);
3127 if (!bad)
3128 return;
3129 data->status |= bad;
3130 err = whitespace_error_string(bad);
3131 fprintf(data->o->file, "%s%s:%d: %s.\n",
3132 line_prefix, data->filename, data->lineno, err);
3133 free(err);
3134 emit_line(data->o, set, reset, line, 1);
3135 ws_check_emit(line + 1, len - 1, data->ws_rule,
3136 data->o->file, set, reset, ws);
3137 } else if (line[0] == ' ') {
3138 data->lineno++;
3139 } else if (line[0] == '@') {
3140 char *plus = strchr(line, '+');
3141 if (plus)
3142 data->lineno = strtol(plus, NULL, 10) - 1;
3143 else
3144 die("invalid diff");
3145 }
3146}
3147static unsigned char *deflate_it(char *data,
3149 unsigned long size,
3150 unsigned long *result_size)
3151{
3152 int bound;
3153 unsigned char *deflated;
3154 git_zstream stream;
3155 git_deflate_init(&stream, zlib_compression_level);
3157 bound = git_deflate_bound(&stream, size);
3158 deflated = xmalloc(bound);
3159 stream.next_out = deflated;
3160 stream.avail_out = bound;
3161 stream.next_in = (unsigned char *)data;
3163 stream.avail_in = size;
3164 while (git_deflate(&stream, Z_FINISH) == Z_OK)
3165 ; /* nothing */
3166 git_deflate_end(&stream);
3167 *result_size = stream.total_out;
3168 return deflated;
3169}
3170static void emit_binary_diff_body(struct diff_options *o,
3172 mmfile_t *one, mmfile_t *two)
3173{
3174 void *cp;
3175 void *delta;
3176 void *deflated;
3177 void *data;
3178 unsigned long orig_size;
3179 unsigned long delta_size;
3180 unsigned long deflate_size;
3181 unsigned long data_size;
3182 /* We could do deflated delta, or we could do just deflated two,
3184 * whichever is smaller.
3185 */
3186 delta = NULL;
3187 deflated = deflate_it(two->ptr, two->size, &deflate_size);
3188 if (one->size && two->size) {
3189 delta = diff_delta(one->ptr, one->size,
3190 two->ptr, two->size,
3191 &delta_size, deflate_size);
3192 if (delta) {
3193 void *to_free = delta;
3194 orig_size = delta_size;
3195 delta = deflate_it(delta, delta_size, &delta_size);
3196 free(to_free);
3197 }
3198 }
3199 if (delta && delta_size < deflate_size) {
3201 char *s = xstrfmt("%lu", orig_size);
3202 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3203 s, strlen(s), 0);
3204 free(s);
3205 free(deflated);
3206 data = delta;
3207 data_size = delta_size;
3208 } else {
3209 char *s = xstrfmt("%lu", two->size);
3210 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3211 s, strlen(s), 0);
3212 free(s);
3213 free(delta);
3214 data = deflated;
3215 data_size = deflate_size;
3216 }
3217 /* emit data encoded in base85 */
3219 cp = data;
3220 while (data_size) {
3221 int len;
3222 int bytes = (52 < data_size) ? 52 : data_size;
3223 char line[71];
3224 data_size -= bytes;
3225 if (bytes <= 26)
3226 line[0] = bytes + 'A' - 1;
3227 else
3228 line[0] = bytes - 26 + 'a' - 1;
3229 encode_85(line + 1, cp, bytes);
3230 cp = (char *) cp + bytes;
3231 len = strlen(line);
3233 line[len++] = '\n';
3234 line[len] = '\0';
3235 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3237 line, len, 0);
3238 }
3239 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3240 free(data);
3241}
3242static void emit_binary_diff(struct diff_options *o,
3244 mmfile_t *one, mmfile_t *two)
3245{
3246 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3247 emit_binary_diff_body(o, one, two);
3248 emit_binary_diff_body(o, two, one);
3249}
3250int diff_filespec_is_binary(struct diff_filespec *one)
3252{
3253 if (one->is_binary == -1) {
3254 diff_filespec_load_driver(one);
3255 if (one->driver->binary != -1)
3256 one->is_binary = one->driver->binary;
3257 else {
3258 if (!one->data && DIFF_FILE_VALID(one))
3259 diff_populate_filespec(one, CHECK_BINARY);
3260 if (one->is_binary == -1 && one->data)
3261 one->is_binary = buffer_is_binary(one->data,
3262 one->size);
3263 if (one->is_binary == -1)
3264 one->is_binary = 0;
3265 }
3266 }
3267 return one->is_binary;
3268}
3269static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3271{
3272 diff_filespec_load_driver(one);
3273 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3274}
3275void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3277{
3278 if (!options->a_prefix)
3279 options->a_prefix = a;
3280 if (!options->b_prefix)
3281 options->b_prefix = b;
3282}
3283struct userdiff_driver *get_textconv(struct diff_filespec *one)
3285{
3286 if (!DIFF_FILE_VALID(one))
3287 return NULL;
3288 diff_filespec_load_driver(one);
3290 return userdiff_get_textconv(one->driver);
3291}
3292static void builtin_diff(const char *name_a,
3294 const char *name_b,
3295 struct diff_filespec *one,
3296 struct diff_filespec *two,
3297 const char *xfrm_msg,
3298 int must_show_header,
3299 struct diff_options *o,
3300 int complete_rewrite)
3301{
3302 mmfile_t mf1, mf2;
3303 const char *lbl[2];
3304 char *a_one, *b_two;
3305 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3306 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3307 const char *a_prefix, *b_prefix;
3308 struct userdiff_driver *textconv_one = NULL;
3309 struct userdiff_driver *textconv_two = NULL;
3310 struct strbuf header = STRBUF_INIT;
3311 const char *line_prefix = diff_line_prefix(o);
3312 diff_set_mnemonic_prefix(o, "a/", "b/");
3314 if (o->flags.reverse_diff) {
3315 a_prefix = o->b_prefix;
3316 b_prefix = o->a_prefix;
3317 } else {
3318 a_prefix = o->a_prefix;
3319 b_prefix = o->b_prefix;
3320 }
3321 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3323 (!one->mode || S_ISGITLINK(one->mode)) &&
3324 (!two->mode || S_ISGITLINK(two->mode))) {
3325 show_submodule_summary(o, one->path ? one->path : two->path,
3326 &one->oid, &two->oid,
3327 two->dirty_submodule);
3328 return;
3329 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3330 (!one->mode || S_ISGITLINK(one->mode)) &&
3331 (!two->mode || S_ISGITLINK(two->mode))) {
3332 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3333 &one->oid, &two->oid,
3334 two->dirty_submodule);
3335 return;
3336 }
3337 if (o->flags.allow_textconv) {
3339 textconv_one = get_textconv(one);
3340 textconv_two = get_textconv(two);
3341 }
3342 /* Never use a non-valid filename anywhere if at all possible */
3344 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3345 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3346 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3348 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3349 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3350 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3351 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3352 if (lbl[0][0] == '/') {
3353 /* /dev/null */
3354 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3355 if (xfrm_msg)
3356 strbuf_addstr(&header, xfrm_msg);
3357 must_show_header = 1;
3358 }
3359 else if (lbl[1][0] == '/') {
3360 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3361 if (xfrm_msg)
3362 strbuf_addstr(&header, xfrm_msg);
3363 must_show_header = 1;
3364 }
3365 else {
3366 if (one->mode != two->mode) {
3367 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3368 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3369 must_show_header = 1;
3370 }
3371 if (xfrm_msg)
3372 strbuf_addstr(&header, xfrm_msg);
3373 /*
3375 * we do not run diff between different kind
3376 * of objects.
3377 */
3378 if ((one->mode ^ two->mode) & S_IFMT)
3379 goto free_ab_and_return;
3380 if (complete_rewrite &&
3381 (textconv_one || !diff_filespec_is_binary(one)) &&
3382 (textconv_two || !diff_filespec_is_binary(two))) {
3383 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3384 header.buf, header.len, 0);
3385 strbuf_reset(&header);
3386 emit_rewrite_diff(name_a, name_b, one, two,
3387 textconv_one, textconv_two, o);
3388 o->found_changes = 1;
3389 goto free_ab_and_return;
3390 }
3391 }
3392 if (o->irreversible_delete && lbl[1][0] == '/') {
3394 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3395 header.len, 0);
3396 strbuf_reset(&header);
3397 goto free_ab_and_return;
3398 } else if (!o->flags.text &&
3399 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3400 (!textconv_two && diff_filespec_is_binary(two)) )) {
3401 struct strbuf sb = STRBUF_INIT;
3402 if (!one->data && !two->data &&
3403 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3404 !o->flags.binary) {
3405 if (!oidcmp(&one->oid, &two->oid)) {
3406 if (must_show_header)
3407 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3408 header.buf, header.len,
3409 0);
3410 goto free_ab_and_return;
3411 }
3412 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3413 header.buf, header.len, 0);
3414 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3415 diff_line_prefix(o), lbl[0], lbl[1]);
3416 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3417 sb.buf, sb.len, 0);
3418 strbuf_release(&sb);
3419 goto free_ab_and_return;
3420 }
3421 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3422 die("unable to read files to diff");
3423 /* Quite common confusing case */
3424 if (mf1.size == mf2.size &&
3425 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3426 if (must_show_header)
3427 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3428 header.buf, header.len, 0);
3429 goto free_ab_and_return;
3430 }
3431 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3432 strbuf_reset(&header);
3433 if (o->flags.binary)
3434 emit_binary_diff(o, &mf1, &mf2);
3435 else {
3436 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3437 diff_line_prefix(o), lbl[0], lbl[1]);
3438 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3439 sb.buf, sb.len, 0);
3440 strbuf_release(&sb);
3441 }
3442 o->found_changes = 1;
3443 } else {
3444 /* Crazy xdl interfaces.. */
3445 const char *diffopts = getenv("GIT_DIFF_OPTS");
3446 const char *v;
3447 xpparam_t xpp;
3448 xdemitconf_t xecfg;
3449 struct emit_callback ecbdata;
3450 const struct userdiff_funcname *pe;
3451 if (must_show_header) {
3453 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3454 header.buf, header.len, 0);
3455 strbuf_reset(&header);
3456 }
3457 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3459 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3460 pe = diff_funcname_pattern(one);
3462 if (!pe)
3463 pe = diff_funcname_pattern(two);
3464 memset(&xpp, 0, sizeof(xpp));
3466 memset(&xecfg, 0, sizeof(xecfg));
3467 memset(&ecbdata, 0, sizeof(ecbdata));
3468 if (o->flags.suppress_diff_headers)
3469 lbl[0] = NULL;
3470 ecbdata.label_path = lbl;
3471 ecbdata.color_diff = want_color(o->use_color);
3472 ecbdata.ws_rule = whitespace_rule(name_b);
3473 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3474 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3475 ecbdata.opt = o;
3476 if (header.len && !o->flags.suppress_diff_headers)
3477 ecbdata.header = &header;
3478 xpp.flags = o->xdl_opts;
3479 xpp.anchors = o->anchors;
3480 xpp.anchors_nr = o->anchors_nr;
3481 xecfg.ctxlen = o->context;
3482 xecfg.interhunkctxlen = o->interhunkcontext;
3483 xecfg.flags = XDL_EMIT_FUNCNAMES;
3484 if (o->flags.funccontext)
3485 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3486 if (pe)
3487 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3488 if (!diffopts)
3489 ;
3490 else if (skip_prefix(diffopts, "--unified=", &v))
3491 xecfg.ctxlen = strtoul(v, NULL, 10);
3492 else if (skip_prefix(diffopts, "-u", &v))
3493 xecfg.ctxlen = strtoul(v, NULL, 10);
3494 if (o->word_diff)
3495 init_diff_words_data(&ecbdata, o, one, two);
3496 if (xdi_diff_outf(&mf1, &mf2, NULL, fn_out_consume,
3497 &ecbdata, &xpp, &xecfg))
3498 die("unable to generate diff for %s", one->path);
3499 if (o->word_diff)
3500 free_diff_words_data(&ecbdata);
3501 if (textconv_one)
3502 free(mf1.ptr);
3503 if (textconv_two)
3504 free(mf2.ptr);
3505 xdiff_clear_find_func(&xecfg);
3506 }
3507 free_ab_and_return:
3509 strbuf_release(&header);
3510 diff_free_filespec_data(one);
3511 diff_free_filespec_data(two);
3512 free(a_one);
3513 free(b_two);
3514 return;
3515}
3516static char *get_compact_summary(const struct diff_filepair *p, int is_renamed)
3518{
3519 if (!is_renamed) {
3520 if (p->status == DIFF_STATUS_ADDED) {
3521 if (S_ISLNK(p->two->mode))
3522 return "new +l";
3523 else if ((p->two->mode & 0777) == 0755)
3524 return "new +x";
3525 else
3526 return "new";
3527 } else if (p->status == DIFF_STATUS_DELETED)
3528 return "gone";
3529 }
3530 if (S_ISLNK(p->one->mode) && !S_ISLNK(p->two->mode))
3531 return "mode -l";
3532 else if (!S_ISLNK(p->one->mode) && S_ISLNK(p->two->mode))
3533 return "mode +l";
3534 else if ((p->one->mode & 0777) == 0644 &&
3535 (p->two->mode & 0777) == 0755)
3536 return "mode +x";
3537 else if ((p->one->mode & 0777) == 0755 &&
3538 (p->two->mode & 0777) == 0644)
3539 return "mode -x";
3540 return NULL;
3541}
3542static void builtin_diffstat(const char *name_a, const char *name_b,
3544 struct diff_filespec *one,
3545 struct diff_filespec *two,
3546 struct diffstat_t *diffstat,
3547 struct diff_options *o,
3548 struct diff_filepair *p)
3549{
3550 mmfile_t mf1, mf2;
3551 struct diffstat_file *data;
3552 int same_contents;
3553 int complete_rewrite = 0;
3554 if (!DIFF_PAIR_UNMERGED(p)) {
3556 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3557 complete_rewrite = 1;
3558 }
3559 data = diffstat_add(diffstat, name_a, name_b);
3561 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3562 if (o->flags.stat_with_summary)
3563 data->comments = get_compact_summary(p, data->is_renamed);
3564 if (!one || !two) {
3566 data->is_unmerged = 1;
3567 return;
3568 }
3569 same_contents = !oidcmp(&one->oid, &two->oid);
3571 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3573 data->is_binary = 1;
3574 if (same_contents) {
3575 data->added = 0;
3576 data->deleted = 0;
3577 } else {
3578 data->added = diff_filespec_size(two);
3579 data->deleted = diff_filespec_size(one);
3580 }
3581 }
3582 else if (complete_rewrite) {
3584 diff_populate_filespec(one, 0);
3585 diff_populate_filespec(two, 0);
3586 data->deleted = count_lines(one->data, one->size);
3587 data->added = count_lines(two->data, two->size);
3588 }
3589 else if (!same_contents) {
3591 /* Crazy xdl interfaces.. */
3592 xpparam_t xpp;
3593 xdemitconf_t xecfg;
3594 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3596 die("unable to read files to diff");
3597 memset(&xpp, 0, sizeof(xpp));
3599 memset(&xecfg, 0, sizeof(xecfg));
3600 xpp.flags = o->xdl_opts;
3601 xpp.anchors = o->anchors;
3602 xpp.anchors_nr = o->anchors_nr;
3603 xecfg.ctxlen = o->context;
3604 xecfg.interhunkctxlen = o->interhunkcontext;
3605 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
3606 diffstat_consume, diffstat, &xpp, &xecfg))
3607 die("unable to generate diffstat for %s", one->path);
3608 }
3609 diff_free_filespec_data(one);
3611 diff_free_filespec_data(two);
3612}
3613static void builtin_checkdiff(const char *name_a, const char *name_b,
3615 const char *attr_path,
3616 struct diff_filespec *one,
3617 struct diff_filespec *two,
3618 struct diff_options *o)
3619{
3620 mmfile_t mf1, mf2;
3621 struct checkdiff_t data;
3622 if (!two)
3624 return;
3625 memset(&data, 0, sizeof(data));
3627 data.filename = name_b ? name_b : name_a;
3628 data.lineno = 0;
3629 data.o = o;
3630 data.ws_rule = whitespace_rule(attr_path);
3631 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3632 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3634 die("unable to read files to diff");
3635 /*
3637 * All the other codepaths check both sides, but not checking
3638 * the "old" side here is deliberate. We are checking the newly
3639 * introduced changes, and as long as the "new" side is text, we
3640 * can and should check what it introduces.
3641 */
3642 if (diff_filespec_is_binary(two))
3643 goto free_and_return;
3644 else {
3645 /* Crazy xdl interfaces.. */
3646 xpparam_t xpp;
3647 xdemitconf_t xecfg;
3648 memset(&xpp, 0, sizeof(xpp));
3650 memset(&xecfg, 0, sizeof(xecfg));
3651 xecfg.ctxlen = 1; /* at least one context line */
3652 xpp.flags = 0;
3653 if (xdi_diff_outf(&mf1, &mf2, NULL, checkdiff_consume,
3654 &data, &xpp, &xecfg))
3655 die("unable to generate checkdiff for %s", one->path);
3656 if (data.ws_rule & WS_BLANK_AT_EOF) {
3658 struct emit_callback ecbdata;
3659 int blank_at_eof;
3660 ecbdata.ws_rule = data.ws_rule;
3662 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3663 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3664 if (blank_at_eof) {
3666 static char *err;
3667 if (!err)
3668 err = whitespace_error_string(WS_BLANK_AT_EOF);
3669 fprintf(o->file, "%s:%d: %s.\n",
3670 data.filename, blank_at_eof, err);
3671 data.status = 1; /* report errors */
3672 }
3673 }
3674 }
3675 free_and_return:
3676 diff_free_filespec_data(one);
3677 diff_free_filespec_data(two);
3678 if (data.status)
3679 o->flags.check_failed = 1;
3680}
3681struct diff_filespec *alloc_filespec(const char *path)
3683{
3684 struct diff_filespec *spec;
3685 FLEXPTR_ALLOC_STR(spec, path, path);
3687 spec->count = 1;
3688 spec->is_binary = -1;
3689 return spec;
3690}
3691void free_filespec(struct diff_filespec *spec)
3693{
3694 if (!--spec->count) {
3695 diff_free_filespec_data(spec);
3696 free(spec);
3697 }
3698}
3699void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3701 int oid_valid, unsigned short mode)
3702{
3703 if (mode) {
3704 spec->mode = canon_mode(mode);
3705 oidcpy(&spec->oid, oid);
3706 spec->oid_valid = oid_valid;
3707 }
3708}
3709/*
3711 * Given a name and sha1 pair, if the index tells us the file in
3712 * the work tree has that object contents, return true, so that
3713 * prepare_temp_file() does not have to inflate and extract.
3714 */
3715static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3716{
3717 const struct cache_entry *ce;
3718 struct stat st;
3719 int pos, len;
3720 /*
3722 * We do not read the cache ourselves here, because the
3723 * benchmark with my previous version that always reads cache
3724 * shows that it makes things worse for diff-tree comparing
3725 * two linux-2.6 kernel trees in an already checked out work
3726 * tree. This is because most diff-tree comparisons deal with
3727 * only a small number of files, while reading the cache is
3728 * expensive for a large project, and its cost outweighs the
3729 * savings we get by not inflating the object to a temporary
3730 * file. Practically, this code only helps when we are used
3731 * by diff-cache --cached, which does read the cache before
3732 * calling us.
3733 */
3734 if (!active_cache)
3735 return 0;
3736 /* We want to avoid the working directory if our caller
3738 * doesn't need the data in a normal file, this system
3739 * is rather slow with its stat/open/mmap/close syscalls,
3740 * and the object is contained in a pack file. The pack
3741 * is probably already open and will be faster to obtain
3742 * the data through than the working directory. Loose
3743 * objects however would tend to be slower as they need
3744 * to be individually opened and inflated.
3745 */
3746 if (!FAST_WORKING_DIRECTORY && !want_file && has_object_pack(oid))
3747 return 0;
3748 /*
3750 * Similarly, if we'd have to convert the file contents anyway, that
3751 * makes the optimization not worthwhile.
3752 */
3753 if (!want_file && would_convert_to_git(&the_index, name))
3754 return 0;
3755 len = strlen(name);
3757 pos = cache_name_pos(name, len);
3758 if (pos < 0)
3759 return 0;
3760 ce = active_cache[pos];
3761 /*
3763 * This is not the sha1 we are looking for, or
3764 * unreusable because it is not a regular file.
3765 */
3766 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3767 return 0;
3768 /*
3770 * If ce is marked as "assume unchanged", there is no
3771 * guarantee that work tree matches what we are looking for.
3772 */
3773 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3774 return 0;
3775 /*
3777 * If ce matches the file in the work tree, we can reuse it.
3778 */
3779 if (ce_uptodate(ce) ||
3780 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3781 return 1;
3782 return 0;
3784}
3785static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3787{
3788 struct strbuf buf = STRBUF_INIT;
3789 char *dirty = "";
3790 /* Are we looking at the work tree? */
3792 if (s->dirty_submodule)
3793 dirty = "-dirty";
3794 strbuf_addf(&buf, "Subproject commit %s%s\n",
3796 oid_to_hex(&s->oid), dirty);
3797 s->size = buf.len;
3798 if (size_only) {
3799 s->data = NULL;
3800 strbuf_release(&buf);
3801 } else {
3802 s->data = strbuf_detach(&buf, NULL);
3803 s->should_free = 1;
3804 }
3805 return 0;
3806}
3807/*
3809 * While doing rename detection and pickaxe operation, we may need to
3810 * grab the data for the blob (or file) for our own in-core comparison.
3811 * diff_filespec has data and size fields for this purpose.
3812 */
3813int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3814{
3815 int size_only = flags & CHECK_SIZE_ONLY;
3816 int err = 0;
3817 int conv_flags = global_conv_flags_eol;
3818 /*
3819 * demote FAIL to WARN to allow inspecting the situation
3820 * instead of refusing.
3821 */
3822 if (conv_flags & CONV_EOL_RNDTRP_DIE)
3823 conv_flags = CONV_EOL_RNDTRP_WARN;
3824 if (!DIFF_FILE_VALID(s))
3826 die("internal error: asking to populate invalid file.");
3827 if (S_ISDIR(s->mode))
3828 return -1;
3829 if (s->data)
3831 return 0;
3832 if (size_only && 0 < s->size)
3834 return 0;
3835 if (S_ISGITLINK(s->mode))
3837 return diff_populate_gitlink(s, size_only);
3838 if (!s->oid_valid ||
3840 reuse_worktree_file(s->path, &s->oid, 0)) {
3841 struct strbuf buf = STRBUF_INIT;
3842 struct stat st;
3843 int fd;
3844 if (lstat(s->path, &st) < 0) {
3846 err_empty:
3847 err = -1;
3848 empty:
3849 s->data = (char *)"";
3850 s->size = 0;
3851 return err;
3852 }
3853 s->size = xsize_t(st.st_size);
3854 if (!s->size)
3855 goto empty;
3856 if (S_ISLNK(st.st_mode)) {
3857 struct strbuf sb = STRBUF_INIT;
3858 if (strbuf_readlink(&sb, s->path, s->size))
3860 goto err_empty;
3861 s->size = sb.len;
3862 s->data = strbuf_detach(&sb, NULL);
3863 s->should_free = 1;
3864 return 0;
3865 }
3866 /*
3868 * Even if the caller would be happy with getting
3869 * only the size, we cannot return early at this
3870 * point if the path requires us to run the content
3871 * conversion.
3872 */
3873 if (size_only && !would_convert_to_git(&the_index, s->path))
3874 return 0;
3875 /*
3877 * Note: this check uses xsize_t(st.st_size) that may
3878 * not be the true size of the blob after it goes
3879 * through convert_to_git(). This may not strictly be
3880 * correct, but the whole point of big_file_threshold
3881 * and is_binary check being that we want to avoid
3882 * opening the file and inspecting the contents, this
3883 * is probably fine.
3884 */
3885 if ((flags & CHECK_BINARY) &&
3886 s->size > big_file_threshold && s->is_binary == -1) {
3887 s->is_binary = 1;
3888 return 0;
3889 }
3890 fd = open(s->path, O_RDONLY);
3891 if (fd < 0)
3892 goto err_empty;
3893 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3894 close(fd);
3895 s->should_munmap = 1;
3896 /*
3898 * Convert from working tree format to canonical git format
3899 */
3900 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, conv_flags)) {
3901 size_t size = 0;
3902 munmap(s->data, s->size);
3903 s->should_munmap = 0;
3904 s->data = strbuf_detach(&buf, &size);
3905 s->size = size;
3906 s->should_free = 1;
3907 }
3908 }
3909 else {
3910 enum object_type type;
3911 if (size_only || (flags & CHECK_BINARY)) {
3912 type = oid_object_info(the_repository, &s->oid,
3913 &s->size);
3914 if (type < 0)
3915 die("unable to read %s",
3916 oid_to_hex(&s->oid));
3917 if (size_only)
3918 return 0;
3919 if (s->size > big_file_threshold && s->is_binary == -1) {
3920 s->is_binary = 1;
3921 return 0;
3922 }
3923 }
3924 s->data = read_object_file(&s->oid, &type, &s->size);
3925 if (!s->data)
3926 die("unable to read %s", oid_to_hex(&s->oid));
3927 s->should_free = 1;
3928 }
3929 return 0;
3930}
3931void diff_free_filespec_blob(struct diff_filespec *s)
3933{
3934 if (s->should_free)
3935 free(s->data);
3936 else if (s->should_munmap)
3937 munmap(s->data, s->size);
3938 if (s->should_free || s->should_munmap) {
3940 s->should_free = s->should_munmap = 0;
3941 s->data = NULL;
3942 }
3943}
3944void diff_free_filespec_data(struct diff_filespec *s)
3946{
3947 diff_free_filespec_blob(s);
3948 FREE_AND_NULL(s->cnt_data);
3949}
3950static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3952 void *blob,
3953 unsigned long size,
3954 const struct object_id *oid,
3955 int mode)
3956{
3957 struct strbuf buf = STRBUF_INIT;
3958 struct strbuf tempfile = STRBUF_INIT;
3959 char *path_dup = xstrdup(path);
3960 const char *base = basename(path_dup);
3961 /* Generate "XXXXXX_basename.ext" */
3963 strbuf_addstr(&tempfile, "XXXXXX_");
3964 strbuf_addstr(&tempfile, base);
3965 temp->tempfile = mks_tempfile_ts(tempfile.buf, strlen(base) + 1);
3967 if (!temp->tempfile)
3968 die_errno("unable to create temp-file");
3969 if (convert_to_working_tree(&the_index, path,
3970 (const char *)blob, (size_t)size, &buf)) {
3971 blob = buf.buf;
3972 size = buf.len;
3973 }
3974 if (write_in_full(temp->tempfile->fd, blob, size) < 0 ||
3975 close_tempfile_gently(temp->tempfile))
3976 die_errno("unable to write temp-file");
3977 temp->name = get_tempfile_path(temp->tempfile);
3978 oid_to_hex_r(temp->hex, oid);
3979 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3980 strbuf_release(&buf);
3981 strbuf_release(&tempfile);
3982 free(path_dup);
3983}
3984static struct diff_tempfile *prepare_temp_file(const char *name,
3986 struct diff_filespec *one)
3987{
3988 struct diff_tempfile *temp = claim_diff_tempfile();
3989 if (!DIFF_FILE_VALID(one)) {
3991 not_a_valid_file:
3992 /* A '-' entry produces this for file-2, and
3993 * a '+' entry produces this for file-1.
3994 */
3995 temp->name = "/dev/null";
3996 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3997 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3998 return temp;
3999 }
4000 if (!S_ISGITLINK(one->mode) &&
4002 (!one->oid_valid ||
4003 reuse_worktree_file(name, &one->oid, 1))) {
4004 struct stat st;
4005 if (lstat(name, &st) < 0) {
4006 if (errno == ENOENT)
4007 goto not_a_valid_file;
4008 die_errno("stat(%s)", name);
4009 }
4010 if (S_ISLNK(st.st_mode)) {
4011 struct strbuf sb = STRBUF_INIT;
4012 if (strbuf_readlink(&sb, name, st.st_size) < 0)
4013 die_errno("readlink(%s)", name);
4014 prep_temp_blob(name, temp, sb.buf, sb.len,
4015 (one->oid_valid ?
4016 &one->oid : &null_oid),
4017 (one->oid_valid ?
4018 one->mode : S_IFLNK));
4019 strbuf_release(&sb);
4020 }
4021 else {
4022 /* we can borrow from the file in the work tree */
4023 temp->name = name;
4024 if (!one->oid_valid)
4025 oid_to_hex_r(temp->hex, &null_oid);
4026 else
4027 oid_to_hex_r(temp->hex, &one->oid);
4028 /* Even though we may sometimes borrow the
4029 * contents from the work tree, we always want
4030 * one->mode. mode is trustworthy even when
4031 * !(one->oid_valid), as long as
4032 * DIFF_FILE_VALID(one).
4033 */
4034 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
4035 }
4036 return temp;
4037 }
4038 else {
4039 if (diff_populate_filespec(one, 0))
4040 die("cannot read data blob for %s", one->path);
4041 prep_temp_blob(name, temp, one->data, one->size,
4042 &one->oid, one->mode);
4043 }
4044 return temp;
4045}
4046static void add_external_diff_name(struct argv_array *argv,
4048 const char *name,
4049 struct diff_filespec *df)
4050{
4051 struct diff_tempfile *temp = prepare_temp_file(name, df);
4052 argv_array_push(argv, temp->name);
4053 argv_array_push(argv, temp->hex);
4054 argv_array_push(argv, temp->mode);
4055}
4056/* An external diff command takes:
4058 *
4059 * diff-cmd name infile1 infile1-sha1 infile1-mode \
4060 * infile2 infile2-sha1 infile2-mode [ rename-to ]
4061 *
4062 */
4063static void run_external_diff(const char *pgm,
4064 const char *name,
4065 const char *other,
4066 struct diff_filespec *one,
4067 struct diff_filespec *two,
4068 const char *xfrm_msg,
4069 int complete_rewrite,
4070 struct diff_options *o)
4071{
4072 struct argv_array argv = ARGV_ARRAY_INIT;
4073 struct argv_array env = ARGV_ARRAY_INIT;
4074 struct diff_queue_struct *q = &diff_queued_diff;
4075 argv_array_push(&argv, pgm);
4077 argv_array_push(&argv, name);
4078 if (one && two) {
4080 add_external_diff_name(&argv, name, one);
4081 if (!other)
4082 add_external_diff_name(&argv, name, two);
4083 else {
4084 add_external_diff_name(&argv, other, two);
4085 argv_array_push(&argv, other);
4086 argv_array_push(&argv, xfrm_msg);
4087 }
4088 }
4089 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
4091 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
4092 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
4094 die(_("external diff died, stopping at %s"), name);
4095 remove_tempfile();
4097 argv_array_clear(&argv);
4098 argv_array_clear(&env);
4099}
4100static int similarity_index(struct diff_filepair *p)
4102{
4103 return p->score * 100 / MAX_SCORE;
4104}
4105static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
4107{
4108 if (startup_info->have_repository)
4109 return find_unique_abbrev(oid, abbrev);
4110 else {
4111 char *hex = oid_to_hex(oid);
4112 if (abbrev < 0)
4113 abbrev = FALLBACK_DEFAULT_ABBREV;
4114 if (abbrev > the_hash_algo->hexsz)
4115 BUG("oid abbreviation out of range: %d", abbrev);
4116 if (abbrev)
4117 hex[abbrev] = '\0';
4118 return hex;
4119 }
4120}
4121static void fill_metainfo(struct strbuf *msg,
4123 const char *name,
4124 const char *other,
4125 struct diff_filespec *one,
4126 struct diff_filespec *two,
4127 struct diff_options *o,
4128 struct diff_filepair *p,
4129 int *must_show_header,
4130 int use_color)
4131{
4132 const char *set = diff_get_color(use_color, DIFF_METAINFO);
4133 const char *reset = diff_get_color(use_color, DIFF_RESET);
4134 const char *line_prefix = diff_line_prefix(o);
4135 *must_show_header = 1;
4137 strbuf_init(msg, PATH_MAX * 2 + 300);
4138 switch (p->status) {
4139 case DIFF_STATUS_COPIED:
4140 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4141 line_prefix, set, similarity_index(p));
4142 strbuf_addf(msg, "%s\n%s%scopy from ",
4143 reset, line_prefix, set);
4144 quote_c_style(name, msg, NULL, 0);
4145 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
4146 quote_c_style(other, msg, NULL, 0);
4147 strbuf_addf(msg, "%s\n", reset);
4148 break;
4149 case DIFF_STATUS_RENAMED:
4150 strbuf_addf(msg, "%s%ssimilarity index %d%%",
4151 line_prefix, set, similarity_index(p));
4152 strbuf_addf(msg, "%s\n%s%srename from ",
4153 reset, line_prefix, set);
4154 quote_c_style(name, msg, NULL, 0);
4155 strbuf_addf(msg, "%s\n%s%srename to ",
4156 reset, line_prefix, set);
4157 quote_c_style(other, msg, NULL, 0);
4158 strbuf_addf(msg, "%s\n", reset);
4159 break;
4160 case DIFF_STATUS_MODIFIED:
4161 if (p->score) {
4162 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
4163 line_prefix,
4164 set, similarity_index(p), reset);
4165 break;
4166 }
4167 /* fallthru */
4168 default:
4169 *must_show_header = 0;
4170 }
4171 if (one && two && oidcmp(&one->oid, &two->oid)) {
4172 const unsigned hexsz = the_hash_algo->hexsz;
4173 int abbrev = o->flags.full_index ? hexsz : DEFAULT_ABBREV;
4174 if (o->flags.binary) {
4176 mmfile_t mf;
4177 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
4178 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
4179 abbrev = hexsz;
4180 }
4181 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
4182 diff_abbrev_oid(&one->oid, abbrev),
4183 diff_abbrev_oid(&two->oid, abbrev));
4184 if (one->mode == two->mode)
4185 strbuf_addf(msg, " %06o", one->mode);
4186 strbuf_addf(msg, "%s\n", reset);
4187 }
4188}
4189static void run_diff_cmd(const char *pgm,
4191 const char *name,
4192 const char *other,
4193 const char *attr_path,
4194 struct diff_filespec *one,
4195 struct diff_filespec *two,
4196 struct strbuf *msg,
4197 struct diff_options *o,
4198 struct diff_filepair *p)
4199{
4200 const char *xfrm_msg = NULL;
4201 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
4202 int must_show_header = 0;
4203 if (o->flags.allow_external) {
4206 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
4207 if (drv && drv->external)
4208 pgm = drv->external;
4209 }
4210 if (msg) {
4212 /*
4213 * don't use colors when the header is intended for an
4214 * external diff driver
4215 */
4216 fill_metainfo(msg, name, other, one, two, o, p,
4217 &must_show_header,
4218 want_color(o->use_color) && !pgm);
4219 xfrm_msg = msg->len ? msg->buf : NULL;
4220 }
4221 if (pgm) {
4223 run_external_diff(pgm, name, other, one, two, xfrm_msg,
4224 complete_rewrite, o);
4225 return;
4226 }
4227 if (one && two)
4228 builtin_diff(name, other ? other : name,
4229 one, two, xfrm_msg, must_show_header,
4230 o, complete_rewrite);
4231 else
4232 fprintf(o->file, "* Unmerged path %s\n", name);
4233}
4234static void diff_fill_oid_info(struct diff_filespec *one)
4236{
4237 if (DIFF_FILE_VALID(one)) {
4238 if (!one->oid_valid) {
4239 struct stat st;
4240 if (one->is_stdin) {
4241 oidclr(&one->oid);
4242 return;
4243 }
4244 if (lstat(one->path, &st) < 0)
4245 die_errno("stat '%s'", one->path);
4246 if (index_path(&one->oid, one->path, &st, 0))
4247 die("cannot hash %s", one->path);
4248 }
4249 }
4250 else
4251 oidclr(&one->oid);
4252}
4253static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4255{
4256 /* Strip the prefix but do not molest /dev/null and absolute paths */
4257 if (*namep && **namep != '/') {
4258 *namep += prefix_length;
4259 if (**namep == '/')
4260 ++*namep;
4261 }
4262 if (*otherp && **otherp != '/') {
4263 *otherp += prefix_length;
4264 if (**otherp == '/')
4265 ++*otherp;
4266 }
4267}
4268static void run_diff(struct diff_filepair *p, struct diff_options *o)
4270{
4271 const char *pgm = external_diff();
4272 struct strbuf msg;
4273 struct diff_filespec *one = p->one;
4274 struct diff_filespec *two = p->two;
4275 const char *name;
4276 const char *other;
4277 const char *attr_path;
4278 name = one->path;
4280 other = (strcmp(name, two->path) ? two->path : NULL);
4281 attr_path = name;
4282 if (o->prefix_length)
4283 strip_prefix(o->prefix_length, &name, &other);
4284 if (!o->flags.allow_external)
4286 pgm = NULL;
4287 if (DIFF_PAIR_UNMERGED(p)) {
4289 run_diff_cmd(pgm, name, NULL, attr_path,
4290 NULL, NULL, NULL, o, p);
4291 return;
4292 }
4293 diff_fill_oid_info(one);
4295 diff_fill_oid_info(two);
4296 if (!pgm &&
4298 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4299 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4300 /*
4301 * a filepair that changes between file and symlink
4302 * needs to be split into deletion and creation.
4303 */
4304 struct diff_filespec *null = alloc_filespec(two->path);
4305 run_diff_cmd(NULL, name, other, attr_path,
4306 one, null, &msg, o, p);
4307 free(null);
4308 strbuf_release(&msg);
4309 null = alloc_filespec(one->path);
4311 run_diff_cmd(NULL, name, other, attr_path,
4312 null, two, &msg, o, p);
4313 free(null);
4314 }
4315 else
4316 run_diff_cmd(pgm, name, other, attr_path,
4317 one, two, &msg, o, p);
4318 strbuf_release(&msg);
4320}
4321static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4323 struct diffstat_t *diffstat)
4324{
4325 const char *name;
4326 const char *other;
4327 if (DIFF_PAIR_UNMERGED(p)) {
4329 /* unmerged */
4330 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4331 return;
4332 }
4333 name = p->one->path;
4335 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4336 if (o->prefix_length)
4338 strip_prefix(o->prefix_length, &name, &other);
4339 diff_fill_oid_info(p->one);
4341 diff_fill_oid_info(p->two);
4342 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4344}
4345static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4347{
4348 const char *name;
4349 const char *other;
4350 const char *attr_path;
4351 if (DIFF_PAIR_UNMERGED(p)) {
4353 /* unmerged */
4354 return;
4355 }
4356 name = p->one->path;
4358 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4359 attr_path = other ? other : name;
4360 if (o->prefix_length)
4362 strip_prefix(o->prefix_length, &name, &other);
4363 diff_fill_oid_info(p->one);
4365 diff_fill_oid_info(p->two);
4366 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4368}
4369void diff_setup(struct diff_options *options)
4371{
4372 memcpy(options, &default_diff_options, sizeof(*options));
4373 options->file = stdout;
4375 options->abbrev = DEFAULT_ABBREV;
4377 options->line_termination = '\n';
4378 options->break_opt = -1;
4379 options->rename_limit = -1;
4380 options->dirstat_permille = diff_dirstat_permille_default;
4381 options->context = diff_context_default;
4382 options->interhunkcontext = diff_interhunk_context_default;
4383 options->ws_error_highlight = ws_error_highlight_default;
4384 options->flags.rename_empty = 1;
4385 options->objfind = NULL;
4386 /* pathchange left =NULL by default */
4388 options->change = diff_change;
4389 options->add_remove = diff_addremove;
4390 options->use_color = diff_use_color_default;
4391 options->detect_rename = diff_detect_rename_default;
4392 options->xdl_opts |= diff_algorithm;
4393 if (diff_indent_heuristic)
4394 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4395 options->orderfile = diff_order_file_cfg;
4397 if (diff_no_prefix) {
4399 options->a_prefix = options->b_prefix = "";
4400 } else if (!diff_mnemonic_prefix) {
4401 options->a_prefix = "a/";
4402 options->b_prefix = "b/";
4403 }
4404 options->color_moved = diff_color_moved_default;
4406 options->color_moved_ws_handling = diff_color_moved_ws_default;
4407}
4408void diff_setup_done(struct diff_options *options)
4410{
4411 unsigned check_mask = DIFF_FORMAT_NAME |
4412 DIFF_FORMAT_NAME_STATUS |
4413 DIFF_FORMAT_CHECKDIFF |
4414 DIFF_FORMAT_NO_OUTPUT;
4415 /*
4416 * This must be signed because we're comparing against a potentially
4417 * negative value.
4418 */
4419 const int hexsz = the_hash_algo->hexsz;
4420 if (options->set_default)
4422 options->set_default(options);
4423 if (HAS_MULTI_BITS(options->output_format & check_mask))
4425 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4426 if (HAS_MULTI_BITS(options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK))
4428 die(_("-G, -S and --find-object are mutually exclusive"));
4429 /*
4431 * Most of the time we can say "there are changes"
4432 * only by checking if there are changed paths, but
4433 * --ignore-whitespace* options force us to look
4434 * inside contents.
4435 */
4436 if ((options->xdl_opts & XDF_WHITESPACE_FLAGS))
4438 options->flags.diff_from_contents = 1;
4439 else
4440 options->flags.diff_from_contents = 0;
4441 if (options->flags.find_copies_harder)
4443 options->detect_rename = DIFF_DETECT_COPY;
4444 if (!options->flags.relative_name)
4446 options->prefix = NULL;
4447 if (options->prefix)
4448 options->prefix_length = strlen(options->prefix);
4449 else
4450 options->prefix_length = 0;
4451 if (options->output_format & (DIFF_FORMAT_NAME |
4453 DIFF_FORMAT_NAME_STATUS |
4454 DIFF_FORMAT_CHECKDIFF |
4455 DIFF_FORMAT_NO_OUTPUT))
4456 options->output_format &= ~(DIFF_FORMAT_RAW |
4457 DIFF_FORMAT_NUMSTAT |
4458 DIFF_FORMAT_DIFFSTAT |
4459 DIFF_FORMAT_SHORTSTAT |
4460 DIFF_FORMAT_DIRSTAT |
4461 DIFF_FORMAT_SUMMARY |
4462 DIFF_FORMAT_PATCH);
4463 /*
4465 * These cases always need recursive; we do not drop caller-supplied
4466 * recursive bits for other formats here.
4467 */
4468 if (options->output_format & (DIFF_FORMAT_PATCH |
4469 DIFF_FORMAT_NUMSTAT |
4470 DIFF_FORMAT_DIFFSTAT |
4471 DIFF_FORMAT_SHORTSTAT |
4472 DIFF_FORMAT_DIRSTAT |
4473 DIFF_FORMAT_SUMMARY |
4474 DIFF_FORMAT_CHECKDIFF))
4475 options->flags.recursive = 1;
4476 /*
4477 * Also pickaxe would not work very well if you do not say recursive
4478 */
4479 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
4480 options->flags.recursive = 1;
4481 /*
4482 * When patches are generated, submodules diffed against the work tree
4483 * must be checked for dirtiness too so it can be shown in the output
4484 */
4485 if (options->output_format & DIFF_FORMAT_PATCH)
4486 options->flags.dirty_submodules = 1;
4487 if (options->detect_rename && options->rename_limit < 0)
4489 options->rename_limit = diff_rename_limit_default;
4490 if (hexsz < options->abbrev)
4491 options->abbrev = hexsz; /* full */
4492 /*
4494 * It does not make sense to show the first hit we happened
4495 * to have found. It does not make sense not to return with
4496 * exit code in such a case either.
4497 */
4498 if (options->flags.quick) {
4499 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4500 options->flags.exit_with_status = 1;
4501 }
4502 options->diff_path_counter = 0;
4504 if (options->flags.follow_renames && options->pathspec.nr != 1)
4506 die(_("--follow requires exactly one pathspec"));
4507 if (!options->use_color || external_diff())
4509 options->color_moved = 0;
4510}
4511static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4513{
4514 char c, *eq;
4515 int len;
4516 if (*arg != '-')
4518 return 0;
4519 c = *++arg;
4520 if (!c)
4521 return 0;
4522 if (c == arg_short) {
4523 c = *++arg;
4524 if (!c)
4525 return 1;
4526 if (val && isdigit(c)) {
4527 char *end;
4528 int n = strtoul(arg, &end, 10);
4529 if (*end)
4530 return 0;
4531 *val = n;
4532 return 1;
4533 }
4534 return 0;
4535 }
4536 if (c != '-')
4537 return 0;
4538 arg++;
4539 eq = strchrnul(arg, '=');
4540 len = eq - arg;
4541 if (!len || strncmp(arg, arg_long, len))
4542 return 0;
4543 if (*eq) {
4544 int n;
4545 char *end;
4546 if (!isdigit(*++eq))
4547 return 0;
4548 n = strtoul(eq, &end, 10);
4549 if (*end)
4550 return 0;
4551 *val = n;
4552 }
4553 return 1;
4554}
4555static int diff_scoreopt_parse(const char *opt);
4557static inline int short_opt(char opt, const char **argv,
4559 const char **optarg)
4560{
4561 const char *arg = argv[0];
4562 if (arg[0] != '-' || arg[1] != opt)
4563 return 0;
4564 if (arg[2] != '\0') {
4565 *optarg = arg + 2;
4566 return 1;
4567 }
4568 if (!argv[1])
4569 die("Option '%c' requires a value", opt);
4570 *optarg = argv[1];
4571 return 2;
4572}
4573int parse_long_opt(const char *opt, const char **argv,
4575 const char **optarg)
4576{
4577 const char *arg = argv[0];
4578 if (!skip_prefix(arg, "--", &arg))
4579 return 0;
4580 if (!skip_prefix(arg, opt, &arg))
4581 return 0;
4582 if (*arg == '=') { /* stuck form: --option=value */
4583 *optarg = arg + 1;
4584 return 1;
4585 }
4586 if (*arg != '\0')
4587 return 0;
4588 /* separate form: --option value */
4589 if (!argv[1])
4590 die("Option '--%s' requires a value", opt);
4591 *optarg = argv[1];
4592 return 2;
4593}
4594static int stat_opt(struct diff_options *options, const char **av)
4596{
4597 const char *arg = av[0];
4598 char *end;
4599 int width = options->stat_width;
4600 int name_width = options->stat_name_width;
4601 int graph_width = options->stat_graph_width;
4602 int count = options->stat_count;
4603 int argcount = 1;
4604 if (!skip_prefix(arg, "--stat", &arg))
4606 BUG("stat option does not begin with --stat: %s", arg);
4607 end = (char *)arg;
4608 switch (*arg) {
4610 case '-':
4611 if (skip_prefix(arg, "-width", &arg)) {
4612 if (*arg == '=')
4613 width = strtoul(arg + 1, &end, 10);
4614 else if (!*arg && !av[1])
4615 die_want_option("--stat-width");
4616 else if (!*arg) {
4617 width = strtoul(av[1], &end, 10);
4618 argcount = 2;
4619 }
4620 } else if (skip_prefix(arg, "-name-width", &arg)) {
4621 if (*arg == '=')
4622 name_width = strtoul(arg + 1, &end, 10);
4623 else if (!*arg && !av[1])
4624 die_want_option("--stat-name-width");
4625 else if (!*arg) {
4626 name_width = strtoul(av[1], &end, 10);
4627 argcount = 2;
4628 }
4629 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4630 if (*arg == '=')
4631 graph_width = strtoul(arg + 1, &end, 10);
4632 else if (!*arg && !av[1])
4633 die_want_option("--stat-graph-width");
4634 else if (!*arg) {
4635 graph_width = strtoul(av[1], &end, 10);
4636 argcount = 2;
4637 }
4638 } else if (skip_prefix(arg, "-count", &arg)) {
4639 if (*arg == '=')
4640 count = strtoul(arg + 1, &end, 10);
4641 else if (!*arg && !av[1])
4642 die_want_option("--stat-count");
4643 else if (!*arg) {
4644 count = strtoul(av[1], &end, 10);
4645 argcount = 2;
4646 }
4647 }
4648 break;
4649 case '=':
4650 width = strtoul(arg+1, &end, 10);
4651 if (*end == ',')
4652 name_width = strtoul(end+1, &end, 10);
4653 if (*end == ',')
4654 count = strtoul(end+1, &end, 10);
4655 }
4656 /* Important! This checks all the error cases! */
4658 if (*end)
4659 return 0;
4660 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4661 options->stat_name_width = name_width;
4662 options->stat_graph_width = graph_width;
4663 options->stat_width = width;
4664 options->stat_count = count;
4665 return argcount;
4666}
4667static int parse_dirstat_opt(struct diff_options *options, const char *params)
4669{
4670 struct strbuf errmsg = STRBUF_INIT;
4671 if (parse_dirstat_params(options, params, &errmsg))
4672 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4673 errmsg.buf);
4674 strbuf_release(&errmsg);
4675 /*
4676 * The caller knows a dirstat-related option is given from the command
4677 * line; allow it to say "return this_function();"
4678 */
4679 options->output_format |= DIFF_FORMAT_DIRSTAT;
4680 return 1;
4681}
4682static int parse_submodule_opt(struct diff_options *options, const char *value)
4684{
4685 if (parse_submodule_params(options, value))
4686 die(_("Failed to parse --submodule option parameter: '%s'"),
4687 value);
4688 return 1;
4689}
4690static const char diff_status_letters[] = {
4692 DIFF_STATUS_ADDED,
4693 DIFF_STATUS_COPIED,
4694 DIFF_STATUS_DELETED,
4695 DIFF_STATUS_MODIFIED,
4696 DIFF_STATUS_RENAMED,
4697 DIFF_STATUS_TYPE_CHANGED,
4698 DIFF_STATUS_UNKNOWN,
4699 DIFF_STATUS_UNMERGED,
4700 DIFF_STATUS_FILTER_AON,
4701 DIFF_STATUS_FILTER_BROKEN,
4702 '\0',
4703};
4704static unsigned int filter_bit['Z' + 1];
4706static void prepare_filter_bits(void)
4708{
4709 int i;
4710 if (!filter_bit[DIFF_STATUS_ADDED]) {
4712 for (i = 0; diff_status_letters[i]; i++)
4713 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4714 }
4715}
4716static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4718{
4719 return opt->filter & filter_bit[(int) status];
4720}
4721static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4723{
4724 int i, optch;
4725 prepare_filter_bits();
4727 /*
4729 * If there is a negation e.g. 'd' in the input, and we haven't
4730 * initialized the filter field with another --diff-filter, start
4731 * from full set of bits, except for AON.
4732 */
4733 if (!opt->filter) {
4734 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4735 if (optch < 'a' || 'z' < optch)
4736 continue;
4737 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4738 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4739 break;
4740 }
4741 }
4742 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4744 unsigned int bit;
4745 int negate;
4746 if ('a' <= optch && optch <= 'z') {
4748 negate = 1;
4749 optch = toupper(optch);
4750 } else {
4751 negate = 0;
4752 }
4753 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4755 if (!bit)
4756 return optarg[i];
4757 if (negate)
4758 opt->filter &= ~bit;
4759 else
4760 opt->filter |= bit;
4761 }
4762 return 0;
4763}
4764static void enable_patch_output(int *fmt) {
4766 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4767 *fmt |= DIFF_FORMAT_PATCH;
4768}
4769static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4771{
4772 int val = parse_ws_error_highlight(arg);
4773 if (val < 0) {
4775 error("unknown value after ws-error-highlight=%.*s",
4776 -1 - val, arg);
4777 return 0;
4778 }
4779 opt->ws_error_highlight = val;
4780 return 1;
4781}
4782static int parse_objfind_opt(struct diff_options *opt, const char *arg)
4784{
4785 struct object_id oid;
4786 if (get_oid(arg, &oid))
4788 return error("unable to resolve '%s'", arg);
4789 if (!opt->objfind)
4791 opt->objfind = xcalloc(1, sizeof(*opt->objfind));
4792 opt->pickaxe_opts |= DIFF_PICKAXE_KIND_OBJFIND;
4794 opt->flags.recursive = 1;
4795 opt->flags.tree_in_recursive = 1;
4796 oidset_insert(opt->objfind, &oid);
4797 return 1;
4798}
4799int diff_opt_parse(struct diff_options *options,
4801 const char **av, int ac, const char *prefix)
4802{
4803 const char *arg = av[0];
4804 const char *optarg;
4805 int argcount;
4806 if (!prefix)
4808 prefix = "";
4809 /* Output format options */
4811 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4812 || opt_arg(arg, 'U', "unified", &options->context))
4813 enable_patch_output(&options->output_format);
4814 else if (!strcmp(arg, "--raw"))
4815 options->output_format |= DIFF_FORMAT_RAW;
4816 else if (!strcmp(arg, "--patch-with-raw")) {
4817 enable_patch_output(&options->output_format);
4818 options->output_format |= DIFF_FORMAT_RAW;
4819 } else if (!strcmp(arg, "--numstat"))
4820 options->output_format |= DIFF_FORMAT_NUMSTAT;
4821 else if (!strcmp(arg, "--shortstat"))
4822 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4823 else if (skip_prefix(arg, "-X", &arg) ||
4824 skip_to_optional_arg(arg, "--dirstat", &arg))
4825 return parse_dirstat_opt(options, arg);
4826 else if (!strcmp(arg, "--cumulative"))
4827 return parse_dirstat_opt(options, "cumulative");
4828 else if (skip_to_optional_arg(arg, "--dirstat-by-file", &arg)) {
4829 parse_dirstat_opt(options, "files");
4830 return parse_dirstat_opt(options, arg);
4831 }
4832 else if (!strcmp(arg, "--check"))
4833 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4834 else if (!strcmp(arg, "--summary"))
4835 options->output_format |= DIFF_FORMAT_SUMMARY;
4836 else if (!strcmp(arg, "--patch-with-stat")) {
4837 enable_patch_output(&options->output_format);
4838 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4839 } else if (!strcmp(arg, "--name-only"))
4840 options->output_format |= DIFF_FORMAT_NAME;
4841 else if (!strcmp(arg, "--name-status"))
4842 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4843 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4844 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4845 else if (starts_with(arg, "--stat"))
4846 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4847 return stat_opt(options, av);
4848 else if (!strcmp(arg, "--compact-summary")) {
4849 options->flags.stat_with_summary = 1;
4850 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4851 } else if (!strcmp(arg, "--no-compact-summary"))
4852 options->flags.stat_with_summary = 0;
4853 /* renames options */
4855 else if (starts_with(arg, "-B") ||
4856 skip_to_optional_arg(arg, "--break-rewrites", NULL)) {
4857 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4858 return error("invalid argument to -B: %s", arg+2);
4859 }
4860 else if (starts_with(arg, "-M") ||
4861 skip_to_optional_arg(arg, "--find-renames", NULL)) {
4862 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4863 return error("invalid argument to -M: %s", arg+2);
4864 options->detect_rename = DIFF_DETECT_RENAME;
4865 }
4866 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4867 options->irreversible_delete = 1;
4868 }
4869 else if (starts_with(arg, "-C") ||
4870 skip_to_optional_arg(arg, "--find-copies", NULL)) {
4871 if (options->detect_rename == DIFF_DETECT_COPY)
4872 options->flags.find_copies_harder = 1;
4873 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4874 return error("invalid argument to -C: %s", arg+2);
4875 options->detect_rename = DIFF_DETECT_COPY;
4876 }
4877 else if (!strcmp(arg, "--no-renames"))
4878 options->detect_rename = 0;
4879 else if (!strcmp(arg, "--rename-empty"))
4880 options->flags.rename_empty = 1;
4881 else if (!strcmp(arg, "--no-rename-empty"))
4882 options->flags.rename_empty = 0;
4883 else if (skip_to_optional_arg_default(arg, "--relative", &arg, NULL)) {
4884 options->flags.relative_name = 1;
4885 if (arg)
4886 options->prefix = arg;
4887 }
4888 /* xdiff options */
4890 else if (!strcmp(arg, "--minimal"))
4891 DIFF_XDL_SET(options, NEED_MINIMAL);
4892 else if (!strcmp(arg, "--no-minimal"))
4893 DIFF_XDL_CLR(options, NEED_MINIMAL);
4894 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4895 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4896 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4897 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4898 else if (!strcmp(arg, "--ignore-space-at-eol"))
4899 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4900 else if (!strcmp(arg, "--ignore-cr-at-eol"))
4901 DIFF_XDL_SET(options, IGNORE_CR_AT_EOL);
4902 else if (!strcmp(arg, "--ignore-blank-lines"))
4903 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4904 else if (!strcmp(arg, "--indent-heuristic"))
4905 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4906 else if (!strcmp(arg, "--no-indent-heuristic"))
4907 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4908 else if (!strcmp(arg, "--patience")) {
4909 int i;
4910 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4911 /*
4912 * Both --patience and --anchored use PATIENCE_DIFF
4913 * internally, so remove any anchors previously
4914 * specified.
4915 */
4916 for (i = 0; i < options->anchors_nr; i++)
4917 free(options->anchors[i]);
4918 options->anchors_nr = 0;
4919 } else if (!strcmp(arg, "--histogram"))
4920 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4921 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4922 long value = parse_algorithm_value(optarg);
4923 if (value < 0)
4924 return error("option diff-algorithm accepts \"myers\", "
4925 "\"minimal\", \"patience\" and \"histogram\"");
4926 /* clear out previous settings */
4927 DIFF_XDL_CLR(options, NEED_MINIMAL);
4928 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4929 options->xdl_opts |= value;
4930 return argcount;
4931 } else if (skip_prefix(arg, "--anchored=", &arg)) {
4932 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4933 ALLOC_GROW(options->anchors, options->anchors_nr + 1,
4934 options->anchors_alloc);
4935 options->anchors[options->anchors_nr++] = xstrdup(arg);
4936 }
4937 /* flags options */
4939 else if (!strcmp(arg, "--binary")) {
4940 enable_patch_output(&options->output_format);
4941 options->flags.binary = 1;
4942 }
4943 else if (!strcmp(arg, "--full-index"))
4944 options->flags.full_index = 1;
4945 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4946 options->flags.text = 1;
4947 else if (!strcmp(arg, "-R"))
4948 options->flags.reverse_diff = 1;
4949 else if (!strcmp(arg, "--find-copies-harder"))
4950 options->flags.find_copies_harder = 1;
4951 else if (!strcmp(arg, "--follow"))
4952 options->flags.follow_renames = 1;
4953 else if (!strcmp(arg, "--no-follow")) {
4954 options->flags.follow_renames = 0;
4955 options->flags.default_follow_renames = 0;
4956 } else if (skip_to_optional_arg_default(arg, "--color", &arg, "always")) {
4957 int value = git_config_colorbool(NULL, arg);
4958 if (value < 0)
4959 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4960 options->use_color = value;
4961 }
4962 else if (!strcmp(arg, "--no-color"))
4963 options->use_color = 0;
4964 else if (!strcmp(arg, "--color-moved")) {
4965 if (diff_color_moved_default)
4966 options->color_moved = diff_color_moved_default;
4967 if (options->color_moved == COLOR_MOVED_NO)
4968 options->color_moved = COLOR_MOVED_DEFAULT;
4969 } else if (!strcmp(arg, "--no-color-moved"))
4970 options->color_moved = COLOR_MOVED_NO;
4971 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4972 int cm = parse_color_moved(arg);
4973 if (cm < 0)
4974 die("bad --color-moved argument: %s", arg);
4975 options->color_moved = cm;
4976 } else if (skip_prefix(arg, "--color-moved-ws=", &arg)) {
4977 options->color_moved_ws_handling = parse_color_moved_ws(arg);
4978 } else if (skip_to_optional_arg_default(arg, "--color-words", &options->word_regex, NULL)) {
4979 options->use_color = 1;
4980 options->word_diff = DIFF_WORDS_COLOR;
4981 }
4982 else if (!strcmp(arg, "--word-diff")) {
4983 if (options->word_diff == DIFF_WORDS_NONE)
4984 options->word_diff = DIFF_WORDS_PLAIN;
4985 }
4986 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4987 if (!strcmp(arg, "plain"))
4988 options->word_diff = DIFF_WORDS_PLAIN;
4989 else if (!strcmp(arg, "color")) {
4990 options->use_color = 1;
4991 options->word_diff = DIFF_WORDS_COLOR;
4992 }
4993 else if (!strcmp(arg, "porcelain"))
4994 options->word_diff = DIFF_WORDS_PORCELAIN;
4995 else if (!strcmp(arg, "none"))
4996 options->word_diff = DIFF_WORDS_NONE;
4997 else
4998 die("bad --word-diff argument: %s", arg);
4999 }
5000 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
5001 if (options->word_diff == DIFF_WORDS_NONE)
5002 options->word_diff = DIFF_WORDS_PLAIN;
5003 options->word_regex = optarg;
5004 return argcount;
5005 }
5006 else if (!strcmp(arg, "--exit-code"))
5007 options->flags.exit_with_status = 1;
5008 else if (!strcmp(arg, "--quiet"))
5009 options->flags.quick = 1;
5010 else if (!strcmp(arg, "--ext-diff"))
5011 options->flags.allow_external = 1;
5012 else if (!strcmp(arg, "--no-ext-diff"))
5013 options->flags.allow_external = 0;
5014 else if (!strcmp(arg, "--textconv")) {
5015 options->flags.allow_textconv = 1;
5016 options->flags.textconv_set_via_cmdline = 1;
5017 } else if (!strcmp(arg, "--no-textconv"))
5018 options->flags.allow_textconv = 0;
5019 else if (skip_to_optional_arg_default(arg, "--ignore-submodules", &arg, "all")) {
5020 options->flags.override_submodule_config = 1;
5021 handle_ignore_submodules_arg(options, arg);
5022 } else if (skip_to_optional_arg_default(arg, "--submodule", &arg, "log"))
5023 return parse_submodule_opt(options, arg);
5024 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
5025 return parse_ws_error_highlight_opt(options, arg);
5026 else if (!strcmp(arg, "--ita-invisible-in-index"))
5027 options->ita_invisible_in_index = 1;
5028 else if (!strcmp(arg, "--ita-visible-in-index"))
5029 options->ita_invisible_in_index = 0;
5030 /* misc options */
5032 else if (!strcmp(arg, "-z"))
5033 options->line_termination = 0;
5034 else if ((argcount = short_opt('l', av, &optarg))) {
5035 options->rename_limit = strtoul(optarg, NULL, 10);
5036 return argcount;
5037 }
5038 else if ((argcount = short_opt('S', av, &optarg))) {
5039 options->pickaxe = optarg;
5040 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
5041 return argcount;
5042 } else if ((argcount = short_opt('G', av, &optarg))) {
5043 options->pickaxe = optarg;
5044 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
5045 return argcount;
5046 }
5047 else if (!strcmp(arg, "--pickaxe-all"))
5048 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
5049 else if (!strcmp(arg, "--pickaxe-regex"))
5050 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
5051 else if ((argcount = short_opt('O', av, &optarg))) {
5052 options->orderfile = prefix_filename(prefix, optarg);
5053 return argcount;
5054 } else if (skip_prefix(arg, "--find-object=", &arg))
5055 return parse_objfind_opt(options, arg);
5056 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
5057 int offending = parse_diff_filter_opt(optarg, options);
5058 if (offending)
5059 die("unknown change class '%c' in --diff-filter=%s",
5060 offending, optarg);
5061 return argcount;
5062 }
5063 else if (!strcmp(arg, "--no-abbrev"))
5064 options->abbrev = 0;
5065 else if (!strcmp(arg, "--abbrev"))
5066 options->abbrev = DEFAULT_ABBREV;
5067 else if (skip_prefix(arg, "--abbrev=", &arg)) {
5068 options->abbrev = strtoul(arg, NULL, 10);
5069 if (options->abbrev < MINIMUM_ABBREV)
5070 options->abbrev = MINIMUM_ABBREV;
5071 else if (the_hash_algo->hexsz < options->abbrev)
5072 options->abbrev = the_hash_algo->hexsz;
5073 }
5074 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
5075 options->a_prefix = optarg;
5076 return argcount;
5077 }
5078 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
5079 options->line_prefix = optarg;
5080 options->line_prefix_length = strlen(options->line_prefix);
5081 graph_setup_line_prefix(options);
5082 return argcount;
5083 }
5084 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
5085 options->b_prefix = optarg;
5086 return argcount;
5087 }
5088 else if (!strcmp(arg, "--no-prefix"))
5089 options->a_prefix = options->b_prefix = "";
5090 else if (opt_arg(arg, '\0', "inter-hunk-context",
5091 &options->interhunkcontext))
5092 ;
5093 else if (!strcmp(arg, "-W"))
5094 options->flags.funccontext = 1;
5095 else if (!strcmp(arg, "--function-context"))
5096 options->flags.funccontext = 1;
5097 else if (!strcmp(arg, "--no-function-context"))
5098 options->flags.funccontext = 0;
5099 else if ((argcount = parse_long_opt("output", av, &optarg))) {
5100 char *path = prefix_filename(prefix, optarg);
5101 options->file = xfopen(path, "w");
5102 options->close_file = 1;
5103 if (options->use_color != GIT_COLOR_ALWAYS)
5104 options->use_color = GIT_COLOR_NEVER;
5105 free(path);
5106 return argcount;
5107 } else
5108 return 0;
5109 return 1;
5110}
5111int parse_rename_score(const char **cp_p)
5113{
5114 unsigned long num, scale;
5115 int ch, dot;
5116 const char *cp = *cp_p;
5117 num = 0;
5119 scale = 1;
5120 dot = 0;
5121 for (;;) {
5122 ch = *cp;
5123 if ( !dot && ch == '.' ) {
5124 scale = 1;
5125 dot = 1;
5126 } else if ( ch == '%' ) {
5127 scale = dot ? scale*100 : 100;
5128 cp++; /* % is always at the end */
5129 break;
5130 } else if ( ch >= '0' && ch <= '9' ) {
5131 if ( scale < 100000 ) {
5132 scale *= 10;
5133 num = (num*10) + (ch-'0');
5134 }
5135 } else {
5136 break;
5137 }
5138 cp++;
5139 }
5140 *cp_p = cp;
5141 /* user says num divided by scale and we say internally that
5143 * is MAX_SCORE * num / scale.
5144 */
5145 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
5146}
5147static int diff_scoreopt_parse(const char *opt)
5149{
5150 int opt1, opt2, cmd;
5151 if (*opt++ != '-')
5153 return -1;
5154 cmd = *opt++;
5155 if (cmd == '-') {
5156 /* convert the long-form arguments into short-form versions */
5157 if (skip_prefix(opt, "break-rewrites", &opt)) {
5158 if (*opt == 0 || *opt++ == '=')
5159 cmd = 'B';
5160 } else if (skip_prefix(opt, "find-copies", &opt)) {
5161 if (*opt == 0 || *opt++ == '=')
5162 cmd = 'C';
5163 } else if (skip_prefix(opt, "find-renames", &opt)) {
5164 if (*opt == 0 || *opt++ == '=')
5165 cmd = 'M';
5166 }
5167 }
5168 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
5169 return -1; /* that is not a -M, -C, or -B option */
5170 opt1 = parse_rename_score(&opt);
5172 if (cmd != 'B')
5173 opt2 = 0;
5174 else {
5175 if (*opt == 0)
5176 opt2 = 0;
5177 else if (*opt != '/')
5178 return -1; /* we expect -B80/99 or -B80 */
5179 else {
5180 opt++;
5181 opt2 = parse_rename_score(&opt);
5182 }
5183 }
5184 if (*opt != 0)
5185 return -1;
5186 return opt1 | (opt2 << 16);
5187}
5188struct diff_queue_struct diff_queued_diff;
5190void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
5192{
5193 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
5194 queue->queue[queue->nr++] = dp;
5195}
5196struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
5198 struct diff_filespec *one,
5199 struct diff_filespec *two)
5200{
5201 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
5202 dp->one = one;
5203 dp->two = two;
5204 if (queue)
5205 diff_q(queue, dp);
5206 return dp;
5207}
5208void diff_free_filepair(struct diff_filepair *p)
5210{
5211 free_filespec(p->one);
5212 free_filespec(p->two);
5213 free(p);
5214}
5215const char *diff_aligned_abbrev(const struct object_id *oid, int len)
5217{
5218 int abblen;
5219 const char *abbrev;
5220 /* Do we want all 40 hex characters? */
5222 if (len == the_hash_algo->hexsz)
5223 return oid_to_hex(oid);
5224 /* An abbreviated value is fine, possibly followed by an ellipsis. */
5226 abbrev = diff_abbrev_oid(oid, len);
5227 if (!print_sha1_ellipsis())
5229 return abbrev;
5230 abblen = strlen(abbrev);
5232 /*
5234 * In well-behaved cases, where the abbreviated result is the
5235 * same as the requested length, append three dots after the
5236 * abbreviation (hence the whole logic is limited to the case
5237 * where abblen < 37); when the actual abbreviated result is a
5238 * bit longer than the requested length, we reduce the number
5239 * of dots so that they match the well-behaved ones. However,
5240 * if the actual abbreviation is longer than the requested
5241 * length by more than three, we give up on aligning, and add
5242 * three dots anyway, to indicate that the output is not the
5243 * full object name. Yes, this may be suboptimal, but this
5244 * appears only in "diff --raw --abbrev" output and it is not
5245 * worth the effort to change it now. Note that this would
5246 * likely to work fine when the automatic sizing of default
5247 * abbreviation length is used--we would be fed -1 in "len" in
5248 * that case, and will end up always appending three-dots, but
5249 * the automatic sizing is supposed to give abblen that ensures
5250 * uniqueness across all objects (statistically speaking).
5251 */
5252 if (abblen < the_hash_algo->hexsz - 3) {
5253 static char hex[GIT_MAX_HEXSZ + 1];
5254 if (len < abblen && abblen <= len + 2)
5255 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5256 else
5257 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5258 return hex;
5259 }
5260 return oid_to_hex(oid);
5262}
5263static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5265{
5266 int line_termination = opt->line_termination;
5267 int inter_name_termination = line_termination ? '\t' : '\0';
5268 fprintf(opt->file, "%s", diff_line_prefix(opt));
5270 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5271 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5272 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5273 fprintf(opt->file, "%s ",
5274 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5275 }
5276 if (p->score) {
5277 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5278 inter_name_termination);
5279 } else {
5280 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5281 }
5282 if (p->status == DIFF_STATUS_COPIED ||
5284 p->status == DIFF_STATUS_RENAMED) {
5285 const char *name_a, *name_b;
5286 name_a = p->one->path;
5287 name_b = p->two->path;
5288 strip_prefix(opt->prefix_length, &name_a, &name_b);
5289 write_name_quoted(name_a, opt->file, inter_name_termination);
5290 write_name_quoted(name_b, opt->file, line_termination);
5291 } else {
5292 const char *name_a, *name_b;
5293 name_a = p->one->mode ? p->one->path : p->two->path;
5294 name_b = NULL;
5295 strip_prefix(opt->prefix_length, &name_a, &name_b);
5296 write_name_quoted(name_a, opt->file, line_termination);
5297 }
5298}
5299int diff_unmodified_pair(struct diff_filepair *p)
5301{
5302 /* This function is written stricter than necessary to support
5303 * the currently implemented transformers, but the idea is to
5304 * let transformers to produce diff_filepairs any way they want,
5305 * and filter and clean them up here before producing the output.
5306 */
5307 struct diff_filespec *one = p->one, *two = p->two;
5308 if (DIFF_PAIR_UNMERGED(p))
5310 return 0; /* unmerged is interesting */
5311 /* deletion, addition, mode or type change
5313 * and rename are all interesting.
5314 */
5315 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5316 DIFF_PAIR_MODE_CHANGED(p) ||
5317 strcmp(one->path, two->path))
5318 return 0;
5319 /* both are valid and point at the same path. that is, we are
5321 * dealing with a change.
5322 */
5323 if (one->oid_valid && two->oid_valid &&
5324 !oidcmp(&one->oid, &two->oid) &&
5325 !one->dirty_submodule && !two->dirty_submodule)
5326 return 1; /* no change */
5327 if (!one->oid_valid && !two->oid_valid)
5328 return 1; /* both look at the same file on the filesystem. */
5329 return 0;
5330}
5331static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5333{
5334 if (diff_unmodified_pair(p))
5335 return;
5336 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5338 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5339 return; /* no tree diffs in patch format */
5340 run_diff(p, o);
5342}
5343static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5345 struct diffstat_t *diffstat)
5346{
5347 if (diff_unmodified_pair(p))
5348 return;
5349 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5351 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5352 return; /* no useful stat for tree diffs */
5353 run_diffstat(p, o, diffstat);
5355}
5356static void diff_flush_checkdiff(struct diff_filepair *p,
5358 struct diff_options *o)
5359{
5360 if (diff_unmodified_pair(p))
5361 return;
5362 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5364 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5365 return; /* nothing to check in tree diffs */
5366 run_checkdiff(p, o);
5368}
5369int diff_queue_is_empty(void)
5371{
5372 struct diff_queue_struct *q = &diff_queued_diff;
5373 int i;
5374 for (i = 0; i < q->nr; i++)
5375 if (!diff_unmodified_pair(q->queue[i]))
5376 return 0;
5377 return 1;
5378}
5379#if DIFF_DEBUG
5381void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5382{
5383 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5384 x, one ? one : "",
5385 s->path,
5386 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5387 s->mode,
5388 s->oid_valid ? oid_to_hex(&s->oid) : "");
5389 fprintf(stderr, "queue[%d] %s size %lu\n",
5390 x, one ? one : "",
5391 s->size);
5392}
5393void diff_debug_filepair(const struct diff_filepair *p, int i)
5395{
5396 diff_debug_filespec(p->one, i, "one");
5397 diff_debug_filespec(p->two, i, "two");
5398 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5399 p->score, p->status ? p->status : '?',
5400 p->one->rename_used, p->broken_pair);
5401}
5402void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5404{
5405 int i;
5406 if (msg)
5407 fprintf(stderr, "%s\n", msg);
5408 fprintf(stderr, "q->nr = %d\n", q->nr);
5409 for (i = 0; i < q->nr; i++) {
5410 struct diff_filepair *p = q->queue[i];
5411 diff_debug_filepair(p, i);
5412 }
5413}
5414#endif
5415static void diff_resolve_rename_copy(void)
5417{
5418 int i;
5419 struct diff_filepair *p;
5420 struct diff_queue_struct *q = &diff_queued_diff;
5421 diff_debug_queue("resolve-rename-copy", q);
5423 for (i = 0; i < q->nr; i++) {
5425 p = q->queue[i];
5426 p->status = 0; /* undecided */
5427 if (DIFF_PAIR_UNMERGED(p))
5428 p->status = DIFF_STATUS_UNMERGED;
5429 else if (!DIFF_FILE_VALID(p->one))
5430 p->status = DIFF_STATUS_ADDED;
5431 else if (!DIFF_FILE_VALID(p->two))
5432 p->status = DIFF_STATUS_DELETED;
5433 else if (DIFF_PAIR_TYPE_CHANGED(p))
5434 p->status = DIFF_STATUS_TYPE_CHANGED;
5435 /* from this point on, we are dealing with a pair
5437 * whose both sides are valid and of the same type, i.e.
5438 * either in-place edit or rename/copy edit.
5439 */
5440 else if (DIFF_PAIR_RENAME(p)) {
5441 /*
5442 * A rename might have re-connected a broken
5443 * pair up, causing the pathnames to be the
5444 * same again. If so, that's not a rename at
5445 * all, just a modification..
5446 *
5447 * Otherwise, see if this source was used for
5448 * multiple renames, in which case we decrement
5449 * the count, and call it a copy.
5450 */
5451 if (!strcmp(p->one->path, p->two->path))
5452 p->status = DIFF_STATUS_MODIFIED;
5453 else if (--p->one->rename_used > 0)
5454 p->status = DIFF_STATUS_COPIED;
5455 else
5456 p->status = DIFF_STATUS_RENAMED;
5457 }
5458 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5459 p->one->mode != p->two->mode ||
5460 p->one->dirty_submodule ||
5461 p->two->dirty_submodule ||
5462 is_null_oid(&p->one->oid))
5463 p->status = DIFF_STATUS_MODIFIED;
5464 else {
5465 /* This is a "no-change" entry and should not
5466 * happen anymore, but prepare for broken callers.
5467 */
5468 error("feeding unmodified %s to diffcore",
5469 p->one->path);
5470 p->status = DIFF_STATUS_UNKNOWN;
5471 }
5472 }
5473 diff_debug_queue("resolve-rename-copy done", q);
5474}
5475static int check_pair_status(struct diff_filepair *p)
5477{
5478 switch (p->status) {
5479 case DIFF_STATUS_UNKNOWN:
5480 return 0;
5481 case 0:
5482 die("internal error in diff-resolve-rename-copy");
5483 default:
5484 return 1;
5485 }
5486}
5487static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5489{
5490 int fmt = opt->output_format;
5491 if (fmt & DIFF_FORMAT_CHECKDIFF)
5493 diff_flush_checkdiff(p, opt);
5494 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5495 diff_flush_raw(p, opt);
5496 else if (fmt & DIFF_FORMAT_NAME) {
5497 const char *name_a, *name_b;
5498 name_a = p->two->path;
5499 name_b = NULL;
5500 strip_prefix(opt->prefix_length, &name_a, &name_b);
5501 fprintf(opt->file, "%s", diff_line_prefix(opt));
5502 write_name_quoted(name_a, opt->file, opt->line_termination);
5503 }
5504}
5505static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5507{
5508 struct strbuf sb = STRBUF_INIT;
5509 if (fs->mode)
5510 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5511 else
5512 strbuf_addf(&sb, " %s ", newdelete);
5513 quote_c_style(fs->path, &sb, NULL, 0);
5515 strbuf_addch(&sb, '\n');
5516 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5517 sb.buf, sb.len, 0);
5518 strbuf_release(&sb);
5519}
5520static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5522 int show_name)
5523{
5524 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5525 struct strbuf sb = STRBUF_INIT;
5526 strbuf_addf(&sb, " mode change %06o => %06o",
5527 p->one->mode, p->two->mode);
5528 if (show_name) {
5529 strbuf_addch(&sb, ' ');
5530 quote_c_style(p->two->path, &sb, NULL, 0);
5531 }
5532 strbuf_addch(&sb, '\n');
5533 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5534 sb.buf, sb.len, 0);
5535 strbuf_release(&sb);
5536 }
5537}
5538static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5540 struct diff_filepair *p)
5541{
5542 struct strbuf sb = STRBUF_INIT;
5543 struct strbuf names = STRBUF_INIT;
5544 pprint_rename(&names, p->one->path, p->two->path);
5546 strbuf_addf(&sb, " %s %s (%d%%)\n",
5547 renamecopy, names.buf, similarity_index(p));
5548 strbuf_release(&names);
5549 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5550 sb.buf, sb.len, 0);
5551 show_mode_change(opt, p, 0);
5552 strbuf_release(&sb);
5553}
5554static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5556{
5557 switch(p->status) {
5558 case DIFF_STATUS_DELETED:
5559 show_file_mode_name(opt, "delete", p->one);
5560 break;
5561 case DIFF_STATUS_ADDED:
5562 show_file_mode_name(opt, "create", p->two);
5563 break;
5564 case DIFF_STATUS_COPIED:
5565 show_rename_copy(opt, "copy", p);
5566 break;
5567 case DIFF_STATUS_RENAMED:
5568 show_rename_copy(opt, "rename", p);
5569 break;
5570 default:
5571 if (p->score) {
5572 struct strbuf sb = STRBUF_INIT;
5573 strbuf_addstr(&sb, " rewrite ");
5574 quote_c_style(p->two->path, &sb, NULL, 0);
5575 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5576 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5577 sb.buf, sb.len, 0);
5578 strbuf_release(&sb);
5579 }
5580 show_mode_change(opt, p, !p->score);
5581 break;
5582 }
5583}
5584struct patch_id_t {
5586 git_SHA_CTX *ctx;
5587 int patchlen;
5588};
5589static int remove_space(char *line, int len)
5591{
5592 int i;
5593 char *dst = line;
5594 unsigned char c;
5595 for (i = 0; i < len; i++)
5597 if (!isspace((c = line[i])))
5598 *dst++ = c;
5599 return dst - line;
5601}
5602static void patch_id_consume(void *priv, char *line, unsigned long len)
5604{
5605 struct patch_id_t *data = priv;
5606 int new_len;
5607 new_len = remove_space(line, len);
5609 git_SHA1_Update(data->ctx, line, new_len);
5611 data->patchlen += new_len;
5612}
5613static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5615{
5616 git_SHA1_Update(ctx, str, strlen(str));
5617}
5618static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5620{
5621 /* large enough for 2^32 in octal */
5622 char buf[12];
5623 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5624 git_SHA1_Update(ctx, buf, len);
5625}
5626/* returns 0 upon success, and writes result into sha1 */
5628static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5629{
5630 struct diff_queue_struct *q = &diff_queued_diff;
5631 int i;
5632 git_SHA_CTX ctx;
5633 struct patch_id_t data;
5634 git_SHA1_Init(&ctx);
5636 memset(&data, 0, sizeof(struct patch_id_t));
5637 data.ctx = &ctx;
5638 for (i = 0; i < q->nr; i++) {
5640 xpparam_t xpp;
5641 xdemitconf_t xecfg;
5642 mmfile_t mf1, mf2;
5643 struct diff_filepair *p = q->queue[i];
5644 int len1, len2;
5645 memset(&xpp, 0, sizeof(xpp));
5647 memset(&xecfg, 0, sizeof(xecfg));
5648 if (p->status == 0)
5649 return error("internal diff status error");
5650 if (p->status == DIFF_STATUS_UNKNOWN)
5651 continue;
5652 if (diff_unmodified_pair(p))
5653 continue;
5654 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5655 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5656 continue;
5657 if (DIFF_PAIR_UNMERGED(p))
5658 continue;
5659 diff_fill_oid_info(p->one);
5661 diff_fill_oid_info(p->two);
5662 len1 = remove_space(p->one->path, strlen(p->one->path));
5664 len2 = remove_space(p->two->path, strlen(p->two->path));
5665 patch_id_add_string(&ctx, "diff--git");
5666 patch_id_add_string(&ctx, "a/");
5667 git_SHA1_Update(&ctx, p->one->path, len1);
5668 patch_id_add_string(&ctx, "b/");
5669 git_SHA1_Update(&ctx, p->two->path, len2);
5670 if (p->one->mode == 0) {
5672 patch_id_add_string(&ctx, "newfilemode");
5673 patch_id_add_mode(&ctx, p->two->mode);
5674 patch_id_add_string(&ctx, "---/dev/null");
5675 patch_id_add_string(&ctx, "+++b/");
5676 git_SHA1_Update(&ctx, p->two->path, len2);
5677 } else if (p->two->mode == 0) {
5678 patch_id_add_string(&ctx, "deletedfilemode");
5679 patch_id_add_mode(&ctx, p->one->mode);
5680 patch_id_add_string(&ctx, "---a/");
5681 git_SHA1_Update(&ctx, p->one->path, len1);
5682 patch_id_add_string(&ctx, "+++/dev/null");
5683 } else {
5684 patch_id_add_string(&ctx, "---a/");
5685 git_SHA1_Update(&ctx, p->one->path, len1);
5686 patch_id_add_string(&ctx, "+++b/");
5687 git_SHA1_Update(&ctx, p->two->path, len2);
5688 }
5689 if (diff_header_only)
5691 continue;
5692 if (fill_mmfile(&mf1, p->one) < 0 ||
5694 fill_mmfile(&mf2, p->two) < 0)
5695 return error("unable to read files to diff");
5696 if (diff_filespec_is_binary(p->one) ||
5698 diff_filespec_is_binary(p->two)) {
5699 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5700 GIT_SHA1_HEXSZ);
5701 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5702 GIT_SHA1_HEXSZ);
5703 continue;
5704 }
5705 xpp.flags = 0;
5707 xecfg.ctxlen = 3;
5708 xecfg.flags = 0;
5709 if (xdi_diff_outf(&mf1, &mf2, discard_hunk_line,
5710 patch_id_consume, &data, &xpp, &xecfg))
5711 return error("unable to generate patch-id diff for %s",
5712 p->one->path);
5713 }
5714 git_SHA1_Final(oid->hash, &ctx);
5716 return 0;
5717}
5718int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5720{
5721 struct diff_queue_struct *q = &diff_queued_diff;
5722 int i;
5723 int result = diff_get_patch_id(options, oid, diff_header_only);
5724 for (i = 0; i < q->nr; i++)
5726 diff_free_filepair(q->queue[i]);
5727 free(q->queue);
5729 DIFF_QUEUE_CLEAR(q);
5730 return result;
5732}
5733static int is_summary_empty(const struct diff_queue_struct *q)
5735{
5736 int i;
5737 for (i = 0; i < q->nr; i++) {
5739 const struct diff_filepair *p = q->queue[i];
5740 switch (p->status) {
5742 case DIFF_STATUS_DELETED:
5743 case DIFF_STATUS_ADDED:
5744 case DIFF_STATUS_COPIED:
5745 case DIFF_STATUS_RENAMED:
5746 return 0;
5747 default:
5748 if (p->score)
5749 return 0;
5750 if (p->one->mode && p->two->mode &&
5751 p->one->mode != p->two->mode)
5752 return 0;
5753 break;
5754 }
5755 }
5756 return 1;
5757}
5758static const char rename_limit_warning[] =
5760N_("inexact rename detection was skipped due to too many files.");
5761static const char degrade_cc_to_c_warning[] =
5763N_("only found copies from modified paths due to too many files.");
5764static const char rename_limit_advice[] =
5766N_("you may want to set your %s variable to at least "
5767 "%d and retry the command.");
5768void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5770{
5771 fflush(stdout);
5772 if (degraded_cc)
5773 warning(_(degrade_cc_to_c_warning));
5774 else if (needed)
5775 warning(_(rename_limit_warning));
5776 else
5777 return;
5778 if (0 < needed)
5779 warning(_(rename_limit_advice), varname, needed);
5780}
5781static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5783{
5784 int i;
5785 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5786 struct diff_queue_struct *q = &diff_queued_diff;
5787 if (WSEH_NEW & WS_RULE_MASK)
5789 BUG("WS rules bit mask overlaps with diff symbol flags");
5790 if (o->color_moved)
5792 o->emitted_symbols = &esm;
5793 for (i = 0; i < q->nr; i++) {
5795 struct diff_filepair *p = q->queue[i];
5796 if (check_pair_status(p))
5797 diff_flush_patch(p, o);
5798 }
5799 if (o->emitted_symbols) {
5801 if (o->color_moved) {
5802 struct hashmap add_lines, del_lines;
5803 if (o->color_moved_ws_handling &
5805 COLOR_MOVED_WS_ALLOW_INDENTATION_CHANGE)
5806 o->color_moved_ws_handling |= XDF_IGNORE_WHITESPACE;
5807 hashmap_init(&del_lines, moved_entry_cmp, o, 0);
5809 hashmap_init(&add_lines, moved_entry_cmp, o, 0);
5810 add_lines_to_move_detection(o, &add_lines, &del_lines);
5812 mark_color_as_moved(o, &add_lines, &del_lines);
5813 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5814 dim_moved_lines(o);
5815 hashmap_free(&add_lines, 0);
5817 hashmap_free(&del_lines, 0);
5818 }
5819 for (i = 0; i < esm.nr; i++)
5821 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5822 for (i = 0; i < esm.nr; i++)
5824 free((void *)esm.buf[i].line);
5825 }
5826 esm.nr = 0;
5827}
5828void diff_flush(struct diff_options *options)
5830{
5831 struct diff_queue_struct *q = &diff_queued_diff;
5832 int i, output_format = options->output_format;
5833 int separator = 0;
5834 int dirstat_by_line = 0;
5835 /*
5837 * Order: raw, stat, summary, patch
5838 * or: name/name-status/checkdiff (other bits clear)
5839 */
5840 if (!q->nr)
5841 goto free_queue;
5842 if (output_format & (DIFF_FORMAT_RAW |
5844 DIFF_FORMAT_NAME |
5845 DIFF_FORMAT_NAME_STATUS |
5846 DIFF_FORMAT_CHECKDIFF)) {
5847 for (i = 0; i < q->nr; i++) {
5848 struct diff_filepair *p = q->queue[i];
5849 if (check_pair_status(p))
5850 flush_one_pair(p, options);
5851 }
5852 separator++;
5853 }
5854 if (output_format & DIFF_FORMAT_DIRSTAT && options->flags.dirstat_by_line)
5856 dirstat_by_line = 1;
5857 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5859 dirstat_by_line) {
5860 struct diffstat_t diffstat;
5861 memset(&diffstat, 0, sizeof(struct diffstat_t));
5863 for (i = 0; i < q->nr; i++) {
5864 struct diff_filepair *p = q->queue[i];
5865 if (check_pair_status(p))
5866 diff_flush_stat(p, options, &diffstat);
5867 }
5868 if (output_format & DIFF_FORMAT_NUMSTAT)
5869 show_numstat(&diffstat, options);
5870 if (output_format & DIFF_FORMAT_DIFFSTAT)
5871 show_stats(&diffstat, options);
5872 if (output_format & DIFF_FORMAT_SHORTSTAT)
5873 show_shortstats(&diffstat, options);
5874 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5875 show_dirstat_by_line(&diffstat, options);
5876 free_diffstat_info(&diffstat);
5877 separator++;
5878 }
5879 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5880 show_dirstat(options);
5881 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5883 for (i = 0; i < q->nr; i++) {
5884 diff_summary(options, q->queue[i]);
5885 }
5886 separator++;
5887 }
5888 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5890 options->flags.exit_with_status &&
5891 options->flags.diff_from_contents) {
5892 /*
5893 * run diff_flush_patch for the exit status. setting
5894 * options->file to /dev/null should be safe, because we
5895 * aren't supposed to produce any output anyway.
5896 */
5897 if (options->close_file)
5898 fclose(options->file);
5899 options->file = xfopen("/dev/null", "w");
5900 options->close_file = 1;
5901 options->color_moved = 0;
5902 for (i = 0; i < q->nr; i++) {
5903 struct diff_filepair *p = q->queue[i];
5904 if (check_pair_status(p))
5905 diff_flush_patch(p, options);
5906 if (options->found_changes)
5907 break;
5908 }
5909 }
5910 if (output_format & DIFF_FORMAT_PATCH) {
5912 if (separator) {
5913 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5914 if (options->stat_sep)
5915 /* attach patch instead of inline */
5916 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5917 NULL, 0, 0);
5918 }
5919 diff_flush_patch_all_file_pairs(options);
5921 }
5922 if (output_format & DIFF_FORMAT_CALLBACK)
5924 options->format_callback(q, options, options->format_callback_data);
5925 for (i = 0; i < q->nr; i++)
5927 diff_free_filepair(q->queue[i]);
5928free_queue:
5929 free(q->queue);
5930 DIFF_QUEUE_CLEAR(q);
5931 if (options->close_file)
5932 fclose(options->file);
5933 /*
5935 * Report the content-level differences with HAS_CHANGES;
5936 * diff_addremove/diff_change does not set the bit when
5937 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5938 */
5939 if (options->flags.diff_from_contents) {
5940 if (options->found_changes)
5941 options->flags.has_changes = 1;
5942 else
5943 options->flags.has_changes = 0;
5944 }
5945}
5946static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5948{
5949 return (((p->status == DIFF_STATUS_MODIFIED) &&
5950 ((p->score &&
5951 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5952 (!p->score &&
5953 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5954 ((p->status != DIFF_STATUS_MODIFIED) &&
5955 filter_bit_tst(p->status, options)));
5956}
5957static void diffcore_apply_filter(struct diff_options *options)
5959{
5960 int i;
5961 struct diff_queue_struct *q = &diff_queued_diff;
5962 struct diff_queue_struct outq;
5963 DIFF_QUEUE_CLEAR(&outq);
5965 if (!options->filter)
5967 return;
5968 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5970 int found;
5971 for (i = found = 0; !found && i < q->nr; i++) {
5972 if (match_filter(options, q->queue[i]))
5973 found++;
5974 }
5975 if (found)
5976 return;
5977 /* otherwise we will clear the whole queue
5979 * by copying the empty outq at the end of this
5980 * function, but first clear the current entries
5981 * in the queue.
5982 */
5983 for (i = 0; i < q->nr; i++)
5984 diff_free_filepair(q->queue[i]);
5985 }
5986 else {
5987 /* Only the matching ones */
5988 for (i = 0; i < q->nr; i++) {
5989 struct diff_filepair *p = q->queue[i];
5990 if (match_filter(options, p))
5991 diff_q(&outq, p);
5992 else
5993 diff_free_filepair(p);
5994 }
5995 }
5996 free(q->queue);
5997 *q = outq;
5998}
5999/* Check whether two filespecs with the same mode and size are identical */
6001static int diff_filespec_is_identical(struct diff_filespec *one,
6002 struct diff_filespec *two)
6003{
6004 if (S_ISGITLINK(one->mode))
6005 return 0;
6006 if (diff_populate_filespec(one, 0))
6007 return 0;
6008 if (diff_populate_filespec(two, 0))
6009 return 0;
6010 return !memcmp(one->data, two->data, one->size);
6011}
6012static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
6014{
6015 if (p->done_skip_stat_unmatch)
6016 return p->skip_stat_unmatch_result;
6017 p->done_skip_stat_unmatch = 1;
6019 p->skip_stat_unmatch_result = 0;
6020 /*
6021 * 1. Entries that come from stat info dirtiness
6022 * always have both sides (iow, not create/delete),
6023 * one side of the object name is unknown, with
6024 * the same mode and size. Keep the ones that
6025 * do not match these criteria. They have real
6026 * differences.
6027 *
6028 * 2. At this point, the file is known to be modified,
6029 * with the same mode and size, and the object
6030 * name of one side is unknown. Need to inspect
6031 * the identical contents.
6032 */
6033 if (!DIFF_FILE_VALID(p->one) || /* (1) */
6034 !DIFF_FILE_VALID(p->two) ||
6035 (p->one->oid_valid && p->two->oid_valid) ||
6036 (p->one->mode != p->two->mode) ||
6037 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
6038 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
6039 (p->one->size != p->two->size) ||
6040 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
6041 p->skip_stat_unmatch_result = 1;
6042 return p->skip_stat_unmatch_result;
6043}
6044static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
6046{
6047 int i;
6048 struct diff_queue_struct *q = &diff_queued_diff;
6049 struct diff_queue_struct outq;
6050 DIFF_QUEUE_CLEAR(&outq);
6051 for (i = 0; i < q->nr; i++) {
6053 struct diff_filepair *p = q->queue[i];
6054 if (diff_filespec_check_stat_unmatch(p))
6056 diff_q(&outq, p);
6057 else {
6058 /*
6059 * The caller can subtract 1 from skip_stat_unmatch
6060 * to determine how many paths were dirty only
6061 * due to stat info mismatch.
6062 */
6063 if (!diffopt->flags.no_index)
6064 diffopt->skip_stat_unmatch++;
6065 diff_free_filepair(p);
6066 }
6067 }
6068 free(q->queue);
6069 *q = outq;
6070}
6071static int diffnamecmp(const void *a_, const void *b_)
6073{
6074 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
6075 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
6076 const char *name_a, *name_b;
6077 name_a = a->one ? a->one->path : a->two->path;
6079 name_b = b->one ? b->one->path : b->two->path;
6080 return strcmp(name_a, name_b);
6081}
6082void diffcore_fix_diff_index(struct diff_options *options)
6084{
6085 struct diff_queue_struct *q = &diff_queued_diff;
6086 QSORT(q->queue, q->nr, diffnamecmp);
6087}
6088void diffcore_std(struct diff_options *options)
6090{
6091 /* NOTE please keep the following in sync with diff_tree_combined() */
6092 if (options->skip_stat_unmatch)
6093 diffcore_skip_stat_unmatch(options);
6094 if (!options->found_follow) {
6095 /* See try_to_follow_renames() in tree-diff.c */
6096 if (options->break_opt != -1)
6097 diffcore_break(options->break_opt);
6098 if (options->detect_rename)
6099 diffcore_rename(options);
6100 if (options->break_opt != -1)
6101 diffcore_merge_broken();
6102 }
6103 if (options->pickaxe_opts & DIFF_PICKAXE_KINDS_MASK)
6104 diffcore_pickaxe(options);
6105 if (options->orderfile)
6106 diffcore_order(options->orderfile);
6107 if (!options->found_follow)
6108 /* See try_to_follow_renames() in tree-diff.c */
6109 diff_resolve_rename_copy();
6110 diffcore_apply_filter(options);
6111 if (diff_queued_diff.nr && !options->flags.diff_from_contents)
6113 options->flags.has_changes = 1;
6114 else
6115 options->flags.has_changes = 0;
6116 options->found_follow = 0;
6118}
6119int diff_result_code(struct diff_options *opt, int status)
6121{
6122 int result = 0;
6123 diff_warn_rename_limit("diff.renameLimit",
6125 opt->needed_rename_limit,
6126 opt->degraded_cc_to_c);
6127 if (!opt->flags.exit_with_status &&
6128 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
6129 return status;
6130 if (opt->flags.exit_with_status &&
6131 opt->flags.has_changes)
6132 result |= 01;
6133 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
6134 opt->flags.check_failed)
6135 result |= 02;
6136 return result;
6137}
6138int diff_can_quit_early(struct diff_options *opt)
6140{
6141 return (opt->flags.quick &&
6142 !opt->filter &&
6143 opt->flags.has_changes);
6144}
6145/*
6147 * Shall changes to this submodule be ignored?
6148 *
6149 * Submodule changes can be configured to be ignored separately for each path,
6150 * but that configuration can be overridden from the command line.
6151 */
6152static int is_submodule_ignored(const char *path, struct diff_options *options)
6153{
6154 int ignored = 0;
6155 struct diff_flags orig_flags = options->flags;
6156 if (!options->flags.override_submodule_config)
6157 set_diffopt_flags_from_submodule_config(options, path);
6158 if (options->flags.ignore_submodules)
6159 ignored = 1;
6160 options->flags = orig_flags;
6161 return ignored;
6162}
6163void diff_addremove(struct diff_options *options,
6165 int addremove, unsigned mode,
6166 const struct object_id *oid,
6167 int oid_valid,
6168 const char *concatpath, unsigned dirty_submodule)
6169{
6170 struct diff_filespec *one, *two;
6171 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
6173 return;
6174 /* This may look odd, but it is a preparation for
6176 * feeding "there are unchanged files which should
6177 * not produce diffs, but when you are doing copy
6178 * detection you would need them, so here they are"
6179 * entries to the diff-core. They will be prefixed
6180 * with something like '=' or '*' (I haven't decided
6181 * which but should not make any difference).
6182 * Feeding the same new and old to diff_change()
6183 * also has the same effect.
6184 * Before the final output happens, they are pruned after
6185 * merged into rename/copy pairs as appropriate.
6186 */
6187 if (options->flags.reverse_diff)
6188 addremove = (addremove == '+' ? '-' :
6189 addremove == '-' ? '+' : addremove);
6190 if (options->prefix &&
6192 strncmp(concatpath, options->prefix, options->prefix_length))
6193 return;
6194 one = alloc_filespec(concatpath);
6196 two = alloc_filespec(concatpath);
6197 if (addremove != '+')
6199 fill_filespec(one, oid, oid_valid, mode);
6200 if (addremove != '-') {
6201 fill_filespec(two, oid, oid_valid, mode);
6202 two->dirty_submodule = dirty_submodule;
6203 }
6204 diff_queue(&diff_queued_diff, one, two);
6206 if (!options->flags.diff_from_contents)
6207 options->flags.has_changes = 1;
6208}
6209void diff_change(struct diff_options *options,
6211 unsigned old_mode, unsigned new_mode,
6212 const struct object_id *old_oid,
6213 const struct object_id *new_oid,
6214 int old_oid_valid, int new_oid_valid,
6215 const char *concatpath,
6216 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
6217{
6218 struct diff_filespec *one, *two;
6219 struct diff_filepair *p;
6220 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
6222 is_submodule_ignored(concatpath, options))
6223 return;
6224 if (options->flags.reverse_diff) {
6226 SWAP(old_mode, new_mode);
6227 SWAP(old_oid, new_oid);
6228 SWAP(old_oid_valid, new_oid_valid);
6229 SWAP(old_dirty_submodule, new_dirty_submodule);
6230 }
6231 if (options->prefix &&
6233 strncmp(concatpath, options->prefix, options->prefix_length))
6234 return;
6235 one = alloc_filespec(concatpath);
6237 two = alloc_filespec(concatpath);
6238 fill_filespec(one, old_oid, old_oid_valid, old_mode);
6239 fill_filespec(two, new_oid, new_oid_valid, new_mode);
6240 one->dirty_submodule = old_dirty_submodule;
6241 two->dirty_submodule = new_dirty_submodule;
6242 p = diff_queue(&diff_queued_diff, one, two);
6243 if (options->flags.diff_from_contents)
6245 return;
6246 if (options->flags.quick && options->skip_stat_unmatch &&
6248 !diff_filespec_check_stat_unmatch(p))
6249 return;
6250 options->flags.has_changes = 1;
6252}
6253struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
6255{
6256 struct diff_filepair *pair;
6257 struct diff_filespec *one, *two;
6258 if (options->prefix &&
6260 strncmp(path, options->prefix, options->prefix_length))
6261 return NULL;
6262 one = alloc_filespec(path);
6264 two = alloc_filespec(path);
6265 pair = diff_queue(&diff_queued_diff, one, two);
6266 pair->is_unmerged = 1;
6267 return pair;
6268}
6269static char *run_textconv(const char *pgm, struct diff_filespec *spec,
6271 size_t *outsize)
6272{
6273 struct diff_tempfile *temp;
6274 const char *argv[3];
6275 const char **arg = argv;
6276 struct child_process child = CHILD_PROCESS_INIT;
6277 struct strbuf buf = STRBUF_INIT;
6278 int err = 0;
6279 temp = prepare_temp_file(spec->path, spec);
6281 *arg++ = pgm;
6282 *arg++ = temp->name;
6283 *arg = NULL;
6284 child.use_shell = 1;
6286 child.argv = argv;
6287 child.out = -1;
6288 if (start_command(&child)) {
6289 remove_tempfile();
6290 return NULL;
6291 }
6292 if (strbuf_read(&buf, child.out, 0) < 0)
6294 err = error("error reading from textconv command '%s'", pgm);
6295 close(child.out);
6296 if (finish_command(&child) || err) {
6298 strbuf_release(&buf);
6299 remove_tempfile();
6300 return NULL;
6301 }
6302 remove_tempfile();
6303 return strbuf_detach(&buf, outsize);
6305}
6306size_t fill_textconv(struct userdiff_driver *driver,
6308 struct diff_filespec *df,
6309 char **outbuf)
6310{
6311 size_t size;
6312 if (!driver) {
6314 if (!DIFF_FILE_VALID(df)) {
6315 *outbuf = "";
6316 return 0;
6317 }
6318 if (diff_populate_filespec(df, 0))
6319 die("unable to read files to diff");
6320 *outbuf = df->data;
6321 return df->size;
6322 }
6323 if (!driver->textconv)
6325 BUG("fill_textconv called with non-textconv driver");
6326 if (driver->textconv_cache && df->oid_valid) {
6328 *outbuf = notes_cache_get(driver->textconv_cache,
6329 &df->oid,
6330 &size);
6331 if (*outbuf)
6332 return size;
6333 }
6334 *outbuf = run_textconv(driver->textconv, df, &size);
6336 if (!*outbuf)
6337 die("unable to read files to diff");
6338 if (driver->textconv_cache && df->oid_valid) {
6340 /* ignore errors, as we might be in a readonly repository */
6341 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6342 size);
6343 /*
6344 * we could save up changes and flush them all at the end,
6345 * but we would need an extra call after all diffing is done.
6346 * Since generating a cache entry is the slow path anyway,
6347 * this extra overhead probably isn't a big deal.
6348 */
6349 notes_cache_write(driver->textconv_cache);
6350 }
6351 return size;
6353}
6354int textconv_object(const char *path,
6356 unsigned mode,
6357 const struct object_id *oid,
6358 int oid_valid,
6359 char **buf,
6360 unsigned long *buf_size)
6361{
6362 struct diff_filespec *df;
6363 struct userdiff_driver *textconv;
6364 df = alloc_filespec(path);
6366 fill_filespec(df, oid, oid_valid, mode);
6367 textconv = get_textconv(df);
6368 if (!textconv) {
6369 free_filespec(df);
6370 return 0;
6371 }
6372 *buf_size = fill_textconv(textconv, df, buf);
6374 free_filespec(df);
6375 return 1;
6376}
6377void setup_diff_pager(struct diff_options *opt)
6379{
6380 /*
6381 * If the user asked for our exit code, then either they want --quiet
6382 * or --exit-code. We should definitely not bother with a pager in the
6383 * former case, as we will generate no output. Since we still properly
6384 * report our exit code even when a pager is run, we _could_ run a
6385 * pager with --exit-code. But since we have not done so historically,
6386 * and because it is easy to find people oneline advising "git diff
6387 * --exit-code" in hooks and other scripts, we do not do so.
6388 */
6389 if (!opt->flags.exit_with_status &&
6390 check_pager_config("diff") != 0)
6391 setup_pager();
6392}