1/*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4#include "cache.h"
5#include "config.h"
6#include "tempfile.h"
7#include "quote.h"
8#include "diff.h"
9#include "diffcore.h"
10#include "delta.h"
11#include "xdiff-interface.h"
12#include "color.h"
13#include "attr.h"
14#include "run-command.h"
15#include "utf8.h"
16#include "userdiff.h"
17#include "submodule-config.h"
18#include "submodule.h"
19#include "hashmap.h"
20#include "ll-merge.h"
21#include "string-list.h"
22#include "argv-array.h"
23#include "graph.h"
24#include "packfile.h"
25#ifdef NO_FAST_WORKING_DIRECTORY
27#define FAST_WORKING_DIRECTORY 0
28#else
29#define FAST_WORKING_DIRECTORY 1
30#endif
31static int diff_detect_rename_default;
33static int diff_indent_heuristic = 1;
34static int diff_rename_limit_default = 400;
35static int diff_suppress_blank_empty;
36static int diff_use_color_default = -1;
37static int diff_color_moved_default;
38static int diff_context_default = 3;
39static int diff_interhunk_context_default;
40static const char *diff_word_regex_cfg;
41static const char *external_diff_cmd_cfg;
42static const char *diff_order_file_cfg;
43int diff_auto_refresh_index = 1;
44static int diff_mnemonic_prefix;
45static int diff_no_prefix;
46static int diff_stat_graph_width;
47static int diff_dirstat_permille_default = 30;
48static struct diff_options default_diff_options;
49static long diff_algorithm;
50static unsigned ws_error_highlight_default = WSEH_NEW;
51static char diff_colors[][COLOR_MAXLEN] = {
53 GIT_COLOR_RESET,
54 GIT_COLOR_NORMAL, /* CONTEXT */
55 GIT_COLOR_BOLD, /* METAINFO */
56 GIT_COLOR_CYAN, /* FRAGINFO */
57 GIT_COLOR_RED, /* OLD */
58 GIT_COLOR_GREEN, /* NEW */
59 GIT_COLOR_YELLOW, /* COMMIT */
60 GIT_COLOR_BG_RED, /* WHITESPACE */
61 GIT_COLOR_NORMAL, /* FUNCINFO */
62 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
63 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
64 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
65 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
66 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
67 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
68 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
69 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
70};
71static NORETURN void die_want_option(const char *option_name)
73{
74 die(_("option '%s' requires a value"), option_name);
75}
76static int parse_diff_color_slot(const char *var)
78{
79 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
80 return DIFF_CONTEXT;
81 if (!strcasecmp(var, "meta"))
82 return DIFF_METAINFO;
83 if (!strcasecmp(var, "frag"))
84 return DIFF_FRAGINFO;
85 if (!strcasecmp(var, "old"))
86 return DIFF_FILE_OLD;
87 if (!strcasecmp(var, "new"))
88 return DIFF_FILE_NEW;
89 if (!strcasecmp(var, "commit"))
90 return DIFF_COMMIT;
91 if (!strcasecmp(var, "whitespace"))
92 return DIFF_WHITESPACE;
93 if (!strcasecmp(var, "func"))
94 return DIFF_FUNCINFO;
95 if (!strcasecmp(var, "oldmoved"))
96 return DIFF_FILE_OLD_MOVED;
97 if (!strcasecmp(var, "oldmovedalternative"))
98 return DIFF_FILE_OLD_MOVED_ALT;
99 if (!strcasecmp(var, "oldmoveddimmed"))
100 return DIFF_FILE_OLD_MOVED_DIM;
101 if (!strcasecmp(var, "oldmovedalternativedimmed"))
102 return DIFF_FILE_OLD_MOVED_ALT_DIM;
103 if (!strcasecmp(var, "newmoved"))
104 return DIFF_FILE_NEW_MOVED;
105 if (!strcasecmp(var, "newmovedalternative"))
106 return DIFF_FILE_NEW_MOVED_ALT;
107 if (!strcasecmp(var, "newmoveddimmed"))
108 return DIFF_FILE_NEW_MOVED_DIM;
109 if (!strcasecmp(var, "newmovedalternativedimmed"))
110 return DIFF_FILE_NEW_MOVED_ALT_DIM;
111 return -1;
112}
113static int parse_dirstat_params(struct diff_options *options, const char *params_string,
115 struct strbuf *errmsg)
116{
117 char *params_copy = xstrdup(params_string);
118 struct string_list params = STRING_LIST_INIT_NODUP;
119 int ret = 0;
120 int i;
121 if (*params_copy)
123 string_list_split_in_place(¶ms, params_copy, ',', -1);
124 for (i = 0; i < params.nr; i++) {
125 const char *p = params.items[i].string;
126 if (!strcmp(p, "changes")) {
127 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
128 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
129 } else if (!strcmp(p, "lines")) {
130 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
131 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
132 } else if (!strcmp(p, "files")) {
133 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
134 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
135 } else if (!strcmp(p, "noncumulative")) {
136 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
137 } else if (!strcmp(p, "cumulative")) {
138 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
139 } else if (isdigit(*p)) {
140 char *end;
141 int permille = strtoul(p, &end, 10) * 10;
142 if (*end == '.' && isdigit(*++end)) {
143 /* only use first digit */
144 permille += *end - '0';
145 /* .. and ignore any further digits */
146 while (isdigit(*++end))
147 ; /* nothing */
148 }
149 if (!*end)
150 options->dirstat_permille = permille;
151 else {
152 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
153 p);
154 ret++;
155 }
156 } else {
157 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
158 ret++;
159 }
160 }
162 string_list_clear(¶ms, 0);
163 free(params_copy);
164 return ret;
165}
166static int parse_submodule_params(struct diff_options *options, const char *value)
168{
169 if (!strcmp(value, "log"))
170 options->submodule_format = DIFF_SUBMODULE_LOG;
171 else if (!strcmp(value, "short"))
172 options->submodule_format = DIFF_SUBMODULE_SHORT;
173 else if (!strcmp(value, "diff"))
174 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
175 else
176 return -1;
177 return 0;
178}
179static int git_config_rename(const char *var, const char *value)
181{
182 if (!value)
183 return DIFF_DETECT_RENAME;
184 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
185 return DIFF_DETECT_COPY;
186 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
187}
188long parse_algorithm_value(const char *value)
190{
191 if (!value)
192 return -1;
193 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
194 return 0;
195 else if (!strcasecmp(value, "minimal"))
196 return XDF_NEED_MINIMAL;
197 else if (!strcasecmp(value, "patience"))
198 return XDF_PATIENCE_DIFF;
199 else if (!strcasecmp(value, "histogram"))
200 return XDF_HISTOGRAM_DIFF;
201 return -1;
202}
203static int parse_one_token(const char **arg, const char *token)
205{
206 const char *rest;
207 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
208 *arg = rest;
209 return 1;
210 }
211 return 0;
212}
213static int parse_ws_error_highlight(const char *arg)
215{
216 const char *orig_arg = arg;
217 unsigned val = 0;
218 while (*arg) {
220 if (parse_one_token(&arg, "none"))
221 val = 0;
222 else if (parse_one_token(&arg, "default"))
223 val = WSEH_NEW;
224 else if (parse_one_token(&arg, "all"))
225 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
226 else if (parse_one_token(&arg, "new"))
227 val |= WSEH_NEW;
228 else if (parse_one_token(&arg, "old"))
229 val |= WSEH_OLD;
230 else if (parse_one_token(&arg, "context"))
231 val |= WSEH_CONTEXT;
232 else {
233 return -1 - (int)(arg - orig_arg);
234 }
235 if (*arg)
236 arg++;
237 }
238 return val;
239}
240/*
242 * These are to give UI layer defaults.
243 * The core-level commands such as git-diff-files should
244 * never be affected by the setting of diff.renames
245 * the user happens to have in the configuration file.
246 */
247void init_diff_ui_defaults(void)
248{
249 diff_detect_rename_default = 1;
250}
251int git_diff_heuristic_config(const char *var, const char *value, void *cb)
253{
254 if (!strcmp(var, "diff.indentheuristic"))
255 diff_indent_heuristic = git_config_bool(var, value);
256 return 0;
257}
258static int parse_color_moved(const char *arg)
260{
261 switch (git_parse_maybe_bool(arg)) {
262 case 0:
263 return COLOR_MOVED_NO;
264 case 1:
265 return COLOR_MOVED_DEFAULT;
266 default:
267 break;
268 }
269 if (!strcmp(arg, "no"))
271 return COLOR_MOVED_NO;
272 else if (!strcmp(arg, "plain"))
273 return COLOR_MOVED_PLAIN;
274 else if (!strcmp(arg, "zebra"))
275 return COLOR_MOVED_ZEBRA;
276 else if (!strcmp(arg, "default"))
277 return COLOR_MOVED_DEFAULT;
278 else if (!strcmp(arg, "dimmed_zebra"))
279 return COLOR_MOVED_ZEBRA_DIM;
280 else
281 return error(_("color moved setting must be one of 'no', 'default', 'zebra', 'dimmed_zebra', 'plain'"));
282}
283int git_diff_ui_config(const char *var, const char *value, void *cb)
285{
286 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
287 diff_use_color_default = git_config_colorbool(var, value);
288 return 0;
289 }
290 if (!strcmp(var, "diff.colormoved")) {
291 int cm = parse_color_moved(value);
292 if (cm < 0)
293 return -1;
294 diff_color_moved_default = cm;
295 return 0;
296 }
297 if (!strcmp(var, "diff.context")) {
298 diff_context_default = git_config_int(var, value);
299 if (diff_context_default < 0)
300 return -1;
301 return 0;
302 }
303 if (!strcmp(var, "diff.interhunkcontext")) {
304 diff_interhunk_context_default = git_config_int(var, value);
305 if (diff_interhunk_context_default < 0)
306 return -1;
307 return 0;
308 }
309 if (!strcmp(var, "diff.renames")) {
310 diff_detect_rename_default = git_config_rename(var, value);
311 return 0;
312 }
313 if (!strcmp(var, "diff.autorefreshindex")) {
314 diff_auto_refresh_index = git_config_bool(var, value);
315 return 0;
316 }
317 if (!strcmp(var, "diff.mnemonicprefix")) {
318 diff_mnemonic_prefix = git_config_bool(var, value);
319 return 0;
320 }
321 if (!strcmp(var, "diff.noprefix")) {
322 diff_no_prefix = git_config_bool(var, value);
323 return 0;
324 }
325 if (!strcmp(var, "diff.statgraphwidth")) {
326 diff_stat_graph_width = git_config_int(var, value);
327 return 0;
328 }
329 if (!strcmp(var, "diff.external"))
330 return git_config_string(&external_diff_cmd_cfg, var, value);
331 if (!strcmp(var, "diff.wordregex"))
332 return git_config_string(&diff_word_regex_cfg, var, value);
333 if (!strcmp(var, "diff.orderfile"))
334 return git_config_pathname(&diff_order_file_cfg, var, value);
335 if (!strcmp(var, "diff.ignoresubmodules"))
337 handle_ignore_submodules_arg(&default_diff_options, value);
338 if (!strcmp(var, "diff.submodule")) {
340 if (parse_submodule_params(&default_diff_options, value))
341 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
342 value);
343 return 0;
344 }
345 if (!strcmp(var, "diff.algorithm")) {
347 diff_algorithm = parse_algorithm_value(value);
348 if (diff_algorithm < 0)
349 return -1;
350 return 0;
351 }
352 if (!strcmp(var, "diff.wserrorhighlight")) {
354 int val = parse_ws_error_highlight(value);
355 if (val < 0)
356 return -1;
357 ws_error_highlight_default = val;
358 return 0;
359 }
360 return git_diff_basic_config(var, value, cb);
362}
363int git_diff_basic_config(const char *var, const char *value, void *cb)
365{
366 const char *name;
367 if (!strcmp(var, "diff.renamelimit")) {
369 diff_rename_limit_default = git_config_int(var, value);
370 return 0;
371 }
372 if (userdiff_config(var, value) < 0)
374 return -1;
375 if (skip_prefix(var, "diff.color.", &name) ||
377 skip_prefix(var, "color.diff.", &name)) {
378 int slot = parse_diff_color_slot(name);
379 if (slot < 0)
380 return 0;
381 if (!value)
382 return config_error_nonbool(var);
383 return color_parse(value, diff_colors[slot]);
384 }
385 /* like GNU diff's --suppress-blank-empty option */
387 if (!strcmp(var, "diff.suppressblankempty") ||
388 /* for backwards compatibility */
389 !strcmp(var, "diff.suppress-blank-empty")) {
390 diff_suppress_blank_empty = git_config_bool(var, value);
391 return 0;
392 }
393 if (!strcmp(var, "diff.dirstat")) {
395 struct strbuf errmsg = STRBUF_INIT;
396 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
397 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
398 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
399 errmsg.buf);
400 strbuf_release(&errmsg);
401 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
402 return 0;
403 }
404 if (starts_with(var, "submodule."))
406 return parse_submodule_config_option(var, value);
407 if (git_diff_heuristic_config(var, value, cb) < 0)
409 return -1;
410 return git_default_config(var, value, cb);
412}
413static char *quote_two(const char *one, const char *two)
415{
416 int need_one = quote_c_style(one, NULL, NULL, 1);
417 int need_two = quote_c_style(two, NULL, NULL, 1);
418 struct strbuf res = STRBUF_INIT;
419 if (need_one + need_two) {
421 strbuf_addch(&res, '"');
422 quote_c_style(one, &res, NULL, 1);
423 quote_c_style(two, &res, NULL, 1);
424 strbuf_addch(&res, '"');
425 } else {
426 strbuf_addstr(&res, one);
427 strbuf_addstr(&res, two);
428 }
429 return strbuf_detach(&res, NULL);
430}
431static const char *external_diff(void)
433{
434 static const char *external_diff_cmd = NULL;
435 static int done_preparing = 0;
436 if (done_preparing)
438 return external_diff_cmd;
439 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
440 if (!external_diff_cmd)
441 external_diff_cmd = external_diff_cmd_cfg;
442 done_preparing = 1;
443 return external_diff_cmd;
444}
445/*
447 * Keep track of files used for diffing. Sometimes such an entry
448 * refers to a temporary file, sometimes to an existing file, and
449 * sometimes to "/dev/null".
450 */
451static struct diff_tempfile {
452 /*
453 * filename external diff should read from, or NULL if this
454 * entry is currently not in use:
455 */
456 const char *name;
457 char hex[GIT_MAX_HEXSZ + 1];
459 char mode[10];
460 /*
462 * If this diff_tempfile instance refers to a temporary file,
463 * this tempfile object is used to manage its lifetime.
464 */
465 struct tempfile tempfile;
466} diff_temp[2];
467struct emit_callback {
469 int color_diff;
470 unsigned ws_rule;
471 int blank_at_eof_in_preimage;
472 int blank_at_eof_in_postimage;
473 int lno_in_preimage;
474 int lno_in_postimage;
475 const char **label_path;
476 struct diff_words_data *diff_words;
477 struct diff_options *opt;
478 struct strbuf *header;
479};
480static int count_lines(const char *data, int size)
482{
483 int count, ch, completely_empty = 1, nl_just_seen = 0;
484 count = 0;
485 while (0 < size--) {
486 ch = *data++;
487 if (ch == '\n') {
488 count++;
489 nl_just_seen = 1;
490 completely_empty = 0;
491 }
492 else {
493 nl_just_seen = 0;
494 completely_empty = 0;
495 }
496 }
497 if (completely_empty)
498 return 0;
499 if (!nl_just_seen)
500 count++; /* no trailing newline */
501 return count;
502}
503static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
505{
506 if (!DIFF_FILE_VALID(one)) {
507 mf->ptr = (char *)""; /* does not matter */
508 mf->size = 0;
509 return 0;
510 }
511 else if (diff_populate_filespec(one, 0))
512 return -1;
513 mf->ptr = one->data;
515 mf->size = one->size;
516 return 0;
517}
518/* like fill_mmfile, but only for size, so we can avoid retrieving blob */
520static unsigned long diff_filespec_size(struct diff_filespec *one)
521{
522 if (!DIFF_FILE_VALID(one))
523 return 0;
524 diff_populate_filespec(one, CHECK_SIZE_ONLY);
525 return one->size;
526}
527static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
529{
530 char *ptr = mf->ptr;
531 long size = mf->size;
532 int cnt = 0;
533 if (!size)
535 return cnt;
536 ptr += size - 1; /* pointing at the very end */
537 if (*ptr != '\n')
538 ; /* incomplete line */
539 else
540 ptr--; /* skip the last LF */
541 while (mf->ptr < ptr) {
542 char *prev_eol;
543 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
544 if (*prev_eol == '\n')
545 break;
546 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
547 break;
548 cnt++;
549 ptr = prev_eol - 1;
550 }
551 return cnt;
552}
553static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
555 struct emit_callback *ecbdata)
556{
557 int l1, l2, at;
558 unsigned ws_rule = ecbdata->ws_rule;
559 l1 = count_trailing_blank(mf1, ws_rule);
560 l2 = count_trailing_blank(mf2, ws_rule);
561 if (l2 <= l1) {
562 ecbdata->blank_at_eof_in_preimage = 0;
563 ecbdata->blank_at_eof_in_postimage = 0;
564 return;
565 }
566 at = count_lines(mf1->ptr, mf1->size);
567 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
568 at = count_lines(mf2->ptr, mf2->size);
570 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
571}
572static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
574 int first, const char *line, int len)
575{
576 int has_trailing_newline, has_trailing_carriage_return;
577 int nofirst;
578 FILE *file = o->file;
579 fputs(diff_line_prefix(o), file);
581 if (len == 0) {
583 has_trailing_newline = (first == '\n');
584 has_trailing_carriage_return = (!has_trailing_newline &&
585 (first == '\r'));
586 nofirst = has_trailing_newline || has_trailing_carriage_return;
587 } else {
588 has_trailing_newline = (len > 0 && line[len-1] == '\n');
589 if (has_trailing_newline)
590 len--;
591 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
592 if (has_trailing_carriage_return)
593 len--;
594 nofirst = 0;
595 }
596 if (len || !nofirst) {
598 fputs(set, file);
599 if (!nofirst)
600 fputc(first, file);
601 fwrite(line, len, 1, file);
602 fputs(reset, file);
603 }
604 if (has_trailing_carriage_return)
605 fputc('\r', file);
606 if (has_trailing_newline)
607 fputc('\n', file);
608}
609static void emit_line(struct diff_options *o, const char *set, const char *reset,
611 const char *line, int len)
612{
613 emit_line_0(o, set, reset, line[0], line+1, len-1);
614}
615enum diff_symbol {
617 DIFF_SYMBOL_BINARY_DIFF_HEADER,
618 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
619 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
620 DIFF_SYMBOL_BINARY_DIFF_BODY,
621 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
622 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
623 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
624 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
625 DIFF_SYMBOL_STATS_LINE,
626 DIFF_SYMBOL_WORD_DIFF,
627 DIFF_SYMBOL_STAT_SEP,
628 DIFF_SYMBOL_SUMMARY,
629 DIFF_SYMBOL_SUBMODULE_ADD,
630 DIFF_SYMBOL_SUBMODULE_DEL,
631 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
632 DIFF_SYMBOL_SUBMODULE_MODIFIED,
633 DIFF_SYMBOL_SUBMODULE_HEADER,
634 DIFF_SYMBOL_SUBMODULE_ERROR,
635 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
636 DIFF_SYMBOL_REWRITE_DIFF,
637 DIFF_SYMBOL_BINARY_FILES,
638 DIFF_SYMBOL_HEADER,
639 DIFF_SYMBOL_FILEPAIR_PLUS,
640 DIFF_SYMBOL_FILEPAIR_MINUS,
641 DIFF_SYMBOL_WORDS_PORCELAIN,
642 DIFF_SYMBOL_WORDS,
643 DIFF_SYMBOL_CONTEXT,
644 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
645 DIFF_SYMBOL_PLUS,
646 DIFF_SYMBOL_MINUS,
647 DIFF_SYMBOL_NO_LF_EOF,
648 DIFF_SYMBOL_CONTEXT_FRAGINFO,
649 DIFF_SYMBOL_CONTEXT_MARKER,
650 DIFF_SYMBOL_SEPARATOR
651};
652/*
653 * Flags for content lines:
654 * 0..12 are whitespace rules
655 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
656 * 16 is marking if the line is blank at EOF
657 */
658#define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
659#define DIFF_SYMBOL_MOVED_LINE (1<<17)
660#define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
661#define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
662#define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
663/*
665 * This struct is used when we need to buffer the output of the diff output.
666 *
667 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
668 * into the pre/post image file. This pointer could be a union with the
669 * line pointer. By storing an offset into the file instead of the literal line,
670 * we can decrease the memory footprint for the buffered output. At first we
671 * may want to only have indirection for the content lines, but we could also
672 * enhance the state for emitting prefabricated lines, e.g. the similarity
673 * score line or hunk/file headers would only need to store a number or path
674 * and then the output can be constructed later on depending on state.
675 */
676struct emitted_diff_symbol {
677 const char *line;
678 int len;
679 int flags;
680 enum diff_symbol s;
681};
682#define EMITTED_DIFF_SYMBOL_INIT {NULL}
683struct emitted_diff_symbols {
685 struct emitted_diff_symbol *buf;
686 int nr, alloc;
687};
688#define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
689static void append_emitted_diff_symbol(struct diff_options *o,
691 struct emitted_diff_symbol *e)
692{
693 struct emitted_diff_symbol *f;
694 ALLOC_GROW(o->emitted_symbols->buf,
696 o->emitted_symbols->nr + 1,
697 o->emitted_symbols->alloc);
698 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
699 memcpy(f, e, sizeof(struct emitted_diff_symbol));
701 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
702}
703struct moved_entry {
705 struct hashmap_entry ent;
706 const struct emitted_diff_symbol *es;
707 struct moved_entry *next_line;
708};
709static int next_byte(const char **cp, const char **endp,
711 const struct diff_options *diffopt)
712{
713 int retval;
714 if (*cp > *endp)
716 return -1;
717 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_CHANGE)) {
719 while (*cp < *endp && isspace(**cp))
720 (*cp)++;
721 /*
722 * After skipping a couple of whitespaces, we still have to
723 * account for one space.
724 */
725 return (int)' ';
726 }
727 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE)) {
729 while (*cp < *endp && isspace(**cp))
730 (*cp)++;
731 /* return the first non-ws character via the usual below */
732 }
733 retval = (unsigned char)(**cp);
735 (*cp)++;
736 return retval;
737}
738static int moved_entry_cmp(const struct diff_options *diffopt,
740 const struct moved_entry *a,
741 const struct moved_entry *b,
742 const void *keydata)
743{
744 const char *ap = a->es->line, *ae = a->es->line + a->es->len;
745 const char *bp = b->es->line, *be = b->es->line + b->es->len;
746 if (!(diffopt->xdl_opts & XDF_WHITESPACE_FLAGS))
748 return a->es->len != b->es->len || memcmp(ap, bp, a->es->len);
749 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_AT_EOL)) {
751 while (ae > ap && isspace(*ae))
752 ae--;
753 while (be > bp && isspace(*be))
754 be--;
755 }
756 while (1) {
758 int ca, cb;
759 ca = next_byte(&ap, &ae, diffopt);
760 cb = next_byte(&bp, &be, diffopt);
761 if (ca != cb)
762 return 1;
763 if (ca < 0)
764 return 0;
765 }
766}
767static unsigned get_string_hash(struct emitted_diff_symbol *es, struct diff_options *o)
769{
770 if (o->xdl_opts & XDF_WHITESPACE_FLAGS) {
771 static struct strbuf sb = STRBUF_INIT;
772 const char *ap = es->line, *ae = es->line + es->len;
773 int c;
774 strbuf_reset(&sb);
776 while (ae > ap && isspace(*ae))
777 ae--;
778 while ((c = next_byte(&ap, &ae, o)) > 0)
779 strbuf_addch(&sb, c);
780 return memhash(sb.buf, sb.len);
782 } else {
783 return memhash(es->line, es->len);
784 }
785}
786static struct moved_entry *prepare_entry(struct diff_options *o,
788 int line_no)
789{
790 struct moved_entry *ret = xmalloc(sizeof(*ret));
791 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
792 ret->ent.hash = get_string_hash(l, o);
794 ret->es = l;
795 ret->next_line = NULL;
796 return ret;
798}
799static void add_lines_to_move_detection(struct diff_options *o,
801 struct hashmap *add_lines,
802 struct hashmap *del_lines)
803{
804 struct moved_entry *prev_line = NULL;
805 int n;
807 for (n = 0; n < o->emitted_symbols->nr; n++) {
808 struct hashmap *hm;
809 struct moved_entry *key;
810 switch (o->emitted_symbols->buf[n].s) {
812 case DIFF_SYMBOL_PLUS:
813 hm = add_lines;
814 break;
815 case DIFF_SYMBOL_MINUS:
816 hm = del_lines;
817 break;
818 default:
819 prev_line = NULL;
820 continue;
821 }
822 key = prepare_entry(o, n);
824 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
825 prev_line->next_line = key;
826 hashmap_add(hm, key);
828 prev_line = key;
829 }
830}
831static int shrink_potential_moved_blocks(struct moved_entry **pmb,
833 int pmb_nr)
834{
835 int lp, rp;
836 /* Shrink the set of potential block to the remaining running */
838 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
839 while (lp < pmb_nr && pmb[lp])
840 lp++;
841 /* lp points at the first NULL now */
842 while (rp > -1 && !pmb[rp])
844 rp--;
845 /* rp points at the last non-NULL */
846 if (lp < pmb_nr && rp > -1 && lp < rp) {
848 pmb[lp] = pmb[rp];
849 pmb[rp] = NULL;
850 rp--;
851 lp++;
852 }
853 }
854 /* Remember the number of running sets */
856 return rp + 1;
857}
858/*
860 * If o->color_moved is COLOR_MOVED_PLAIN, this function does nothing.
861 *
862 * Otherwise, if the last block has fewer alphanumeric characters than
863 * COLOR_MOVED_MIN_ALNUM_COUNT, unset DIFF_SYMBOL_MOVED_LINE on all lines in
864 * that block.
865 *
866 * The last block consists of the (n - block_length)'th line up to but not
867 * including the nth line.
868 *
869 * NEEDSWORK: This uses the same heuristic as blame_entry_score() in blame.c.
870 * Think of a way to unify them.
871 */
872static void adjust_last_block(struct diff_options *o, int n, int block_length)
873{
874 int i, alnum_count = 0;
875 if (o->color_moved == COLOR_MOVED_PLAIN)
876 return;
877 for (i = 1; i < block_length + 1; i++) {
878 const char *c = o->emitted_symbols->buf[n - i].line;
879 for (; *c; c++) {
880 if (!isalnum(*c))
881 continue;
882 alnum_count++;
883 if (alnum_count >= COLOR_MOVED_MIN_ALNUM_COUNT)
884 return;
885 }
886 }
887 for (i = 1; i < block_length + 1; i++)
888 o->emitted_symbols->buf[n - i].flags &= ~DIFF_SYMBOL_MOVED_LINE;
889}
890/* Find blocks of moved code, delegate actual coloring decision to helper */
892static void mark_color_as_moved(struct diff_options *o,
893 struct hashmap *add_lines,
894 struct hashmap *del_lines)
895{
896 struct moved_entry **pmb = NULL; /* potentially moved blocks */
897 int pmb_nr = 0, pmb_alloc = 0;
898 int n, flipped_block = 1, block_length = 0;
899 for (n = 0; n < o->emitted_symbols->nr; n++) {
902 struct hashmap *hm = NULL;
903 struct moved_entry *key;
904 struct moved_entry *match = NULL;
905 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
906 int i;
907 switch (l->s) {
909 case DIFF_SYMBOL_PLUS:
910 hm = del_lines;
911 key = prepare_entry(o, n);
912 match = hashmap_get(hm, key, o);
913 free(key);
914 break;
915 case DIFF_SYMBOL_MINUS:
916 hm = add_lines;
917 key = prepare_entry(o, n);
918 match = hashmap_get(hm, key, o);
919 free(key);
920 break;
921 default:
922 flipped_block = 1;
923 }
924 if (!match) {
926 adjust_last_block(o, n, block_length);
927 pmb_nr = 0;
928 block_length = 0;
929 continue;
930 }
931 l->flags |= DIFF_SYMBOL_MOVED_LINE;
933 if (o->color_moved == COLOR_MOVED_PLAIN)
935 continue;
936 /* Check any potential block runs, advance each or nullify */
938 for (i = 0; i < pmb_nr; i++) {
939 struct moved_entry *p = pmb[i];
940 struct moved_entry *pnext = (p && p->next_line) ?
941 p->next_line : NULL;
942 if (pnext && !hm->cmpfn(o, pnext, match, NULL)) {
943 pmb[i] = p->next_line;
944 } else {
945 pmb[i] = NULL;
946 }
947 }
948 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
950 if (pmb_nr == 0) {
952 /*
953 * The current line is the start of a new block.
954 * Setup the set of potential blocks.
955 */
956 for (; match; match = hashmap_get_next(hm, match)) {
957 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
958 pmb[pmb_nr++] = match;
959 }
960 flipped_block = (flipped_block + 1) % 2;
962 adjust_last_block(o, n, block_length);
964 block_length = 0;
965 }
966 block_length++;
968 if (flipped_block)
970 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
971 }
972 adjust_last_block(o, n, block_length);
973 free(pmb);
975}
976#define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
978 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
979static void dim_moved_lines(struct diff_options *o)
980{
981 int n;
982 for (n = 0; n < o->emitted_symbols->nr; n++) {
983 struct emitted_diff_symbol *prev = (n != 0) ?
984 &o->emitted_symbols->buf[n - 1] : NULL;
985 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
986 struct emitted_diff_symbol *next =
987 (n < o->emitted_symbols->nr - 1) ?
988 &o->emitted_symbols->buf[n + 1] : NULL;
989 /* Not a plus or minus line? */
991 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
992 continue;
993 /* Not a moved line? */
995 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
996 continue;
997 /*
999 * If prev or next are not a plus or minus line,
1000 * pretend they don't exist
1001 */
1002 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
1003 prev->s != DIFF_SYMBOL_MINUS)
1004 prev = NULL;
1005 if (next && next->s != DIFF_SYMBOL_PLUS &&
1006 next->s != DIFF_SYMBOL_MINUS)
1007 next = NULL;
1008 /* Inside a block? */
1010 if ((prev &&
1011 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1012 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
1013 (next &&
1014 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
1015 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
1016 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1017 continue;
1018 }
1019 /* Check if we are at an interesting bound: */
1021 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
1022 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1023 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1024 continue;
1025 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
1026 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
1027 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
1028 continue;
1029 /*
1031 * The boundary to prev and next are not interesting,
1032 * so this line is not interesting as a whole
1033 */
1034 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1035 }
1036}
1037static void emit_line_ws_markup(struct diff_options *o,
1039 const char *set, const char *reset,
1040 const char *line, int len, char sign,
1041 unsigned ws_rule, int blank_at_eof)
1042{
1043 const char *ws = NULL;
1044 if (o->ws_error_highlight & ws_rule) {
1046 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1047 if (!*ws)
1048 ws = NULL;
1049 }
1050 if (!ws)
1052 emit_line_0(o, set, reset, sign, line, len);
1053 else if (blank_at_eof)
1054 /* Blank line at EOF - paint '+' as well */
1055 emit_line_0(o, ws, reset, sign, line, len);
1056 else {
1057 /* Emit just the prefix, then the rest. */
1058 emit_line_0(o, set, reset, sign, "", 0);
1059 ws_check_emit(line, len, ws_rule,
1060 o->file, set, reset, ws);
1061 }
1062}
1063static void emit_diff_symbol_from_struct(struct diff_options *o,
1065 struct emitted_diff_symbol *eds)
1066{
1067 static const char *nneof = " No newline at end of file\n";
1068 const char *context, *reset, *set, *meta, *fraginfo;
1069 struct strbuf sb = STRBUF_INIT;
1070 enum diff_symbol s = eds->s;
1072 const char *line = eds->line;
1073 int len = eds->len;
1074 unsigned flags = eds->flags;
1075 switch (s) {
1077 case DIFF_SYMBOL_NO_LF_EOF:
1078 context = diff_get_color_opt(o, DIFF_CONTEXT);
1079 reset = diff_get_color_opt(o, DIFF_RESET);
1080 putc('\n', o->file);
1081 emit_line_0(o, context, reset, '\\',
1082 nneof, strlen(nneof));
1083 break;
1084 case DIFF_SYMBOL_SUBMODULE_HEADER:
1085 case DIFF_SYMBOL_SUBMODULE_ERROR:
1086 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1087 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1088 case DIFF_SYMBOL_SUMMARY:
1089 case DIFF_SYMBOL_STATS_LINE:
1090 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1091 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1092 emit_line(o, "", "", line, len);
1093 break;
1094 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1095 case DIFF_SYMBOL_CONTEXT_MARKER:
1096 context = diff_get_color_opt(o, DIFF_CONTEXT);
1097 reset = diff_get_color_opt(o, DIFF_RESET);
1098 emit_line(o, context, reset, line, len);
1099 break;
1100 case DIFF_SYMBOL_SEPARATOR:
1101 fprintf(o->file, "%s%c",
1102 diff_line_prefix(o),
1103 o->line_termination);
1104 break;
1105 case DIFF_SYMBOL_CONTEXT:
1106 set = diff_get_color_opt(o, DIFF_CONTEXT);
1107 reset = diff_get_color_opt(o, DIFF_RESET);
1108 emit_line_ws_markup(o, set, reset, line, len, ' ',
1109 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1110 break;
1111 case DIFF_SYMBOL_PLUS:
1112 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1113 DIFF_SYMBOL_MOVED_LINE_ALT |
1114 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1115 case DIFF_SYMBOL_MOVED_LINE |
1116 DIFF_SYMBOL_MOVED_LINE_ALT |
1117 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1118 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1119 break;
1120 case DIFF_SYMBOL_MOVED_LINE |
1121 DIFF_SYMBOL_MOVED_LINE_ALT:
1122 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1123 break;
1124 case DIFF_SYMBOL_MOVED_LINE |
1125 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1126 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1127 break;
1128 case DIFF_SYMBOL_MOVED_LINE:
1129 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1130 break;
1131 default:
1132 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1133 }
1134 reset = diff_get_color_opt(o, DIFF_RESET);
1135 emit_line_ws_markup(o, set, reset, line, len, '+',
1136 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1137 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1138 break;
1139 case DIFF_SYMBOL_MINUS:
1140 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1141 DIFF_SYMBOL_MOVED_LINE_ALT |
1142 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1143 case DIFF_SYMBOL_MOVED_LINE |
1144 DIFF_SYMBOL_MOVED_LINE_ALT |
1145 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1146 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1147 break;
1148 case DIFF_SYMBOL_MOVED_LINE |
1149 DIFF_SYMBOL_MOVED_LINE_ALT:
1150 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1151 break;
1152 case DIFF_SYMBOL_MOVED_LINE |
1153 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1154 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1155 break;
1156 case DIFF_SYMBOL_MOVED_LINE:
1157 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1158 break;
1159 default:
1160 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1161 }
1162 reset = diff_get_color_opt(o, DIFF_RESET);
1163 emit_line_ws_markup(o, set, reset, line, len, '-',
1164 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1165 break;
1166 case DIFF_SYMBOL_WORDS_PORCELAIN:
1167 context = diff_get_color_opt(o, DIFF_CONTEXT);
1168 reset = diff_get_color_opt(o, DIFF_RESET);
1169 emit_line(o, context, reset, line, len);
1170 fputs("~\n", o->file);
1171 break;
1172 case DIFF_SYMBOL_WORDS:
1173 context = diff_get_color_opt(o, DIFF_CONTEXT);
1174 reset = diff_get_color_opt(o, DIFF_RESET);
1175 /*
1176 * Skip the prefix character, if any. With
1177 * diff_suppress_blank_empty, there may be
1178 * none.
1179 */
1180 if (line[0] != '\n') {
1181 line++;
1182 len--;
1183 }
1184 emit_line(o, context, reset, line, len);
1185 break;
1186 case DIFF_SYMBOL_FILEPAIR_PLUS:
1187 meta = diff_get_color_opt(o, DIFF_METAINFO);
1188 reset = diff_get_color_opt(o, DIFF_RESET);
1189 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1190 line, reset,
1191 strchr(line, ' ') ? "\t" : "");
1192 break;
1193 case DIFF_SYMBOL_FILEPAIR_MINUS:
1194 meta = diff_get_color_opt(o, DIFF_METAINFO);
1195 reset = diff_get_color_opt(o, DIFF_RESET);
1196 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1197 line, reset,
1198 strchr(line, ' ') ? "\t" : "");
1199 break;
1200 case DIFF_SYMBOL_BINARY_FILES:
1201 case DIFF_SYMBOL_HEADER:
1202 fprintf(o->file, "%s", line);
1203 break;
1204 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1205 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1206 break;
1207 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1208 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1209 break;
1210 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1211 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1212 break;
1213 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1214 fputs(diff_line_prefix(o), o->file);
1215 fputc('\n', o->file);
1216 break;
1217 case DIFF_SYMBOL_REWRITE_DIFF:
1218 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1219 reset = diff_get_color_opt(o, DIFF_RESET);
1220 emit_line(o, fraginfo, reset, line, len);
1221 break;
1222 case DIFF_SYMBOL_SUBMODULE_ADD:
1223 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1224 reset = diff_get_color_opt(o, DIFF_RESET);
1225 emit_line(o, set, reset, line, len);
1226 break;
1227 case DIFF_SYMBOL_SUBMODULE_DEL:
1228 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1229 reset = diff_get_color_opt(o, DIFF_RESET);
1230 emit_line(o, set, reset, line, len);
1231 break;
1232 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1233 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1234 diff_line_prefix(o), line);
1235 break;
1236 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1237 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1238 diff_line_prefix(o), line);
1239 break;
1240 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1241 emit_line(o, "", "", " 0 files changed\n",
1242 strlen(" 0 files changed\n"));
1243 break;
1244 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1245 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1246 break;
1247 case DIFF_SYMBOL_WORD_DIFF:
1248 fprintf(o->file, "%.*s", len, line);
1249 break;
1250 case DIFF_SYMBOL_STAT_SEP:
1251 fputs(o->stat_sep, o->file);
1252 break;
1253 default:
1254 die("BUG: unknown diff symbol");
1255 }
1256 strbuf_release(&sb);
1257}
1258static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1260 const char *line, int len, unsigned flags)
1261{
1262 struct emitted_diff_symbol e = {line, len, flags, s};
1263 if (o->emitted_symbols)
1265 append_emitted_diff_symbol(o, &e);
1266 else
1267 emit_diff_symbol_from_struct(o, &e);
1268}
1269void diff_emit_submodule_del(struct diff_options *o, const char *line)
1271{
1272 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1273}
1274void diff_emit_submodule_add(struct diff_options *o, const char *line)
1276{
1277 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1278}
1279void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1281{
1282 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1283 path, strlen(path), 0);
1284}
1285void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1287{
1288 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1289 path, strlen(path), 0);
1290}
1291void diff_emit_submodule_header(struct diff_options *o, const char *header)
1293{
1294 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1295 header, strlen(header), 0);
1296}
1297void diff_emit_submodule_error(struct diff_options *o, const char *err)
1299{
1300 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1301}
1302void diff_emit_submodule_pipethrough(struct diff_options *o,
1304 const char *line, int len)
1305{
1306 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1307}
1308static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1310{
1311 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1312 ecbdata->blank_at_eof_in_preimage &&
1313 ecbdata->blank_at_eof_in_postimage &&
1314 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1315 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1316 return 0;
1317 return ws_blank_line(line, len, ecbdata->ws_rule);
1318}
1319static void emit_add_line(const char *reset,
1321 struct emit_callback *ecbdata,
1322 const char *line, int len)
1323{
1324 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1325 if (new_blank_line_at_eof(ecbdata, line, len))
1326 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1327 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1329}
1330static void emit_del_line(const char *reset,
1332 struct emit_callback *ecbdata,
1333 const char *line, int len)
1334{
1335 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1336 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1337}
1338static void emit_context_line(const char *reset,
1340 struct emit_callback *ecbdata,
1341 const char *line, int len)
1342{
1343 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1344 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1345}
1346static void emit_hunk_header(struct emit_callback *ecbdata,
1348 const char *line, int len)
1349{
1350 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1351 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1352 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1353 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1354 static const char atat[2] = { '@', '@' };
1355 const char *cp, *ep;
1356 struct strbuf msgbuf = STRBUF_INIT;
1357 int org_len = len;
1358 int i = 1;
1359 /*
1361 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1362 * it always is at least 10 bytes long.
1363 */
1364 if (len < 10 ||
1365 memcmp(line, atat, 2) ||
1366 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1367 emit_diff_symbol(ecbdata->opt,
1368 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1369 return;
1370 }
1371 ep += 2; /* skip over @@ */
1372 /* The hunk header in fraginfo color */
1374 strbuf_addstr(&msgbuf, frag);
1375 strbuf_add(&msgbuf, line, ep - line);
1376 strbuf_addstr(&msgbuf, reset);
1377 /*
1379 * trailing "\r\n"
1380 */
1381 for ( ; i < 3; i++)
1382 if (line[len - i] == '\r' || line[len - i] == '\n')
1383 len--;
1384 /* blank before the func header */
1386 for (cp = ep; ep - line < len; ep++)
1387 if (*ep != ' ' && *ep != '\t')
1388 break;
1389 if (ep != cp) {
1390 strbuf_addstr(&msgbuf, context);
1391 strbuf_add(&msgbuf, cp, ep - cp);
1392 strbuf_addstr(&msgbuf, reset);
1393 }
1394 if (ep < line + len) {
1396 strbuf_addstr(&msgbuf, func);
1397 strbuf_add(&msgbuf, ep, line + len - ep);
1398 strbuf_addstr(&msgbuf, reset);
1399 }
1400 strbuf_add(&msgbuf, line + len, org_len - len);
1402 strbuf_complete_line(&msgbuf);
1403 emit_diff_symbol(ecbdata->opt,
1404 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1405 strbuf_release(&msgbuf);
1406}
1407static struct diff_tempfile *claim_diff_tempfile(void) {
1409 int i;
1410 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1411 if (!diff_temp[i].name)
1412 return diff_temp + i;
1413 die("BUG: diff is failing to clean up its tempfiles");
1414}
1415static void remove_tempfile(void)
1417{
1418 int i;
1419 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1420 if (is_tempfile_active(&diff_temp[i].tempfile))
1421 delete_tempfile(&diff_temp[i].tempfile);
1422 diff_temp[i].name = NULL;
1423 }
1424}
1425static void add_line_count(struct strbuf *out, int count)
1427{
1428 switch (count) {
1429 case 0:
1430 strbuf_addstr(out, "0,0");
1431 break;
1432 case 1:
1433 strbuf_addstr(out, "1");
1434 break;
1435 default:
1436 strbuf_addf(out, "1,%d", count);
1437 break;
1438 }
1439}
1440static void emit_rewrite_lines(struct emit_callback *ecb,
1442 int prefix, const char *data, int size)
1443{
1444 const char *endp = NULL;
1445 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1446 while (0 < size) {
1448 int len;
1449 endp = memchr(data, '\n', size);
1451 len = endp ? (endp - data + 1) : size;
1452 if (prefix != '+') {
1453 ecb->lno_in_preimage++;
1454 emit_del_line(reset, ecb, data, len);
1455 } else {
1456 ecb->lno_in_postimage++;
1457 emit_add_line(reset, ecb, data, len);
1458 }
1459 size -= len;
1460 data += len;
1461 }
1462 if (!endp)
1463 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1464}
1465static void emit_rewrite_diff(const char *name_a,
1467 const char *name_b,
1468 struct diff_filespec *one,
1469 struct diff_filespec *two,
1470 struct userdiff_driver *textconv_one,
1471 struct userdiff_driver *textconv_two,
1472 struct diff_options *o)
1473{
1474 int lc_a, lc_b;
1475 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1476 const char *a_prefix, *b_prefix;
1477 char *data_one, *data_two;
1478 size_t size_one, size_two;
1479 struct emit_callback ecbdata;
1480 struct strbuf out = STRBUF_INIT;
1481 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
1483 a_prefix = o->b_prefix;
1484 b_prefix = o->a_prefix;
1485 } else {
1486 a_prefix = o->a_prefix;
1487 b_prefix = o->b_prefix;
1488 }
1489 name_a += (*name_a == '/');
1491 name_b += (*name_b == '/');
1492 strbuf_reset(&a_name);
1494 strbuf_reset(&b_name);
1495 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1496 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1497 size_one = fill_textconv(textconv_one, one, &data_one);
1499 size_two = fill_textconv(textconv_two, two, &data_two);
1500 memset(&ecbdata, 0, sizeof(ecbdata));
1502 ecbdata.color_diff = want_color(o->use_color);
1503 ecbdata.ws_rule = whitespace_rule(name_b);
1504 ecbdata.opt = o;
1505 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1506 mmfile_t mf1, mf2;
1507 mf1.ptr = (char *)data_one;
1508 mf2.ptr = (char *)data_two;
1509 mf1.size = size_one;
1510 mf2.size = size_two;
1511 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1512 }
1513 ecbdata.lno_in_preimage = 1;
1514 ecbdata.lno_in_postimage = 1;
1515 lc_a = count_lines(data_one, size_one);
1517 lc_b = count_lines(data_two, size_two);
1518 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1520 a_name.buf, a_name.len, 0);
1521 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1522 b_name.buf, b_name.len, 0);
1523 strbuf_addstr(&out, "@@ -");
1525 if (!o->irreversible_delete)
1526 add_line_count(&out, lc_a);
1527 else
1528 strbuf_addstr(&out, "?,?");
1529 strbuf_addstr(&out, " +");
1530 add_line_count(&out, lc_b);
1531 strbuf_addstr(&out, " @@\n");
1532 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1533 strbuf_release(&out);
1534 if (lc_a && !o->irreversible_delete)
1536 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1537 if (lc_b)
1538 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1539 if (textconv_one)
1540 free((char *)data_one);
1541 if (textconv_two)
1542 free((char *)data_two);
1543}
1544struct diff_words_buffer {
1546 mmfile_t text;
1547 long alloc;
1548 struct diff_words_orig {
1549 const char *begin, *end;
1550 } *orig;
1551 int orig_nr, orig_alloc;
1552};
1553static void diff_words_append(char *line, unsigned long len,
1555 struct diff_words_buffer *buffer)
1556{
1557 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1558 line++;
1559 len--;
1560 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1561 buffer->text.size += len;
1562 buffer->text.ptr[buffer->text.size] = '\0';
1563}
1564struct diff_words_style_elem {
1566 const char *prefix;
1567 const char *suffix;
1568 const char *color; /* NULL; filled in by the setup code if
1569 * color is enabled */
1570};
1571struct diff_words_style {
1573 enum diff_words_type type;
1574 struct diff_words_style_elem new, old, ctx;
1575 const char *newline;
1576};
1577static struct diff_words_style diff_words_styles[] = {
1579 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1580 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1581 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1582};
1583struct diff_words_data {
1585 struct diff_words_buffer minus, plus;
1586 const char *current_plus;
1587 int last_minus;
1588 struct diff_options *opt;
1589 regex_t *word_regex;
1590 enum diff_words_type type;
1591 struct diff_words_style *style;
1592};
1593static int fn_out_diff_words_write_helper(struct diff_options *o,
1595 struct diff_words_style_elem *st_el,
1596 const char *newline,
1597 size_t count, const char *buf)
1598{
1599 int print = 0;
1600 struct strbuf sb = STRBUF_INIT;
1601 while (count) {
1603 char *p = memchr(buf, '\n', count);
1604 if (print)
1605 strbuf_addstr(&sb, diff_line_prefix(o));
1606 if (p != buf) {
1608 const char *reset = st_el->color && *st_el->color ?
1609 GIT_COLOR_RESET : NULL;
1610 if (st_el->color && *st_el->color)
1611 strbuf_addstr(&sb, st_el->color);
1612 strbuf_addstr(&sb, st_el->prefix);
1613 strbuf_add(&sb, buf, p ? p - buf : count);
1614 strbuf_addstr(&sb, st_el->suffix);
1615 if (reset)
1616 strbuf_addstr(&sb, reset);
1617 }
1618 if (!p)
1619 goto out;
1620 strbuf_addstr(&sb, newline);
1622 count -= p + 1 - buf;
1623 buf = p + 1;
1624 print = 1;
1625 if (count) {
1626 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1627 sb.buf, sb.len, 0);
1628 strbuf_reset(&sb);
1629 }
1630 }
1631out:
1633 if (sb.len)
1634 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1635 sb.buf, sb.len, 0);
1636 strbuf_release(&sb);
1637 return 0;
1638}
1639/*
1641 * '--color-words' algorithm can be described as:
1642 *
1643 * 1. collect the minus/plus lines of a diff hunk, divided into
1644 * minus-lines and plus-lines;
1645 *
1646 * 2. break both minus-lines and plus-lines into words and
1647 * place them into two mmfile_t with one word for each line;
1648 *
1649 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1650 *
1651 * And for the common parts of the both file, we output the plus side text.
1652 * diff_words->current_plus is used to trace the current position of the plus file
1653 * which printed. diff_words->last_minus is used to trace the last minus word
1654 * printed.
1655 *
1656 * For '--graph' to work with '--color-words', we need to output the graph prefix
1657 * on each line of color words output. Generally, there are two conditions on
1658 * which we should output the prefix.
1659 *
1660 * 1. diff_words->last_minus == 0 &&
1661 * diff_words->current_plus == diff_words->plus.text.ptr
1662 *
1663 * that is: the plus text must start as a new line, and if there is no minus
1664 * word printed, a graph prefix must be printed.
1665 *
1666 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1667 * *(diff_words->current_plus - 1) == '\n'
1668 *
1669 * that is: a graph prefix must be printed following a '\n'
1670 */
1671static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1672{
1673 if ((diff_words->last_minus == 0 &&
1674 diff_words->current_plus == diff_words->plus.text.ptr) ||
1675 (diff_words->current_plus > diff_words->plus.text.ptr &&
1676 *(diff_words->current_plus - 1) == '\n')) {
1677 return 1;
1678 } else {
1679 return 0;
1680 }
1681}
1682static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1684{
1685 struct diff_words_data *diff_words = priv;
1686 struct diff_words_style *style = diff_words->style;
1687 int minus_first, minus_len, plus_first, plus_len;
1688 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1689 struct diff_options *opt = diff_words->opt;
1690 const char *line_prefix;
1691 if (line[0] != '@' || parse_hunk_header(line, len,
1693 &minus_first, &minus_len, &plus_first, &plus_len))
1694 return;
1695 assert(opt);
1697 line_prefix = diff_line_prefix(opt);
1698 /* POSIX requires that first be decremented by one if len == 0... */
1700 if (minus_len) {
1701 minus_begin = diff_words->minus.orig[minus_first].begin;
1702 minus_end =
1703 diff_words->minus.orig[minus_first + minus_len - 1].end;
1704 } else
1705 minus_begin = minus_end =
1706 diff_words->minus.orig[minus_first].end;
1707 if (plus_len) {
1709 plus_begin = diff_words->plus.orig[plus_first].begin;
1710 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1711 } else
1712 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1713 if (color_words_output_graph_prefix(diff_words)) {
1715 fputs(line_prefix, diff_words->opt->file);
1716 }
1717 if (diff_words->current_plus != plus_begin) {
1718 fn_out_diff_words_write_helper(diff_words->opt,
1719 &style->ctx, style->newline,
1720 plus_begin - diff_words->current_plus,
1721 diff_words->current_plus);
1722 }
1723 if (minus_begin != minus_end) {
1724 fn_out_diff_words_write_helper(diff_words->opt,
1725 &style->old, style->newline,
1726 minus_end - minus_begin, minus_begin);
1727 }
1728 if (plus_begin != plus_end) {
1729 fn_out_diff_words_write_helper(diff_words->opt,
1730 &style->new, style->newline,
1731 plus_end - plus_begin, plus_begin);
1732 }
1733 diff_words->current_plus = plus_end;
1735 diff_words->last_minus = minus_first;
1736}
1737/* This function starts looking at *begin, and returns 0 iff a word was found. */
1739static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1740 int *begin, int *end)
1741{
1742 if (word_regex && *begin < buffer->size) {
1743 regmatch_t match[1];
1744 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1745 buffer->size - *begin, 1, match, 0)) {
1746 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1747 '\n', match[0].rm_eo - match[0].rm_so);
1748 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1749 *begin += match[0].rm_so;
1750 return *begin >= *end;
1751 }
1752 return -1;
1753 }
1754 /* find the next word */
1756 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1757 (*begin)++;
1758 if (*begin >= buffer->size)
1759 return -1;
1760 /* find the end of the word */
1762 *end = *begin + 1;
1763 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1764 (*end)++;
1765 return 0;
1767}
1768/*
1770 * This function splits the words in buffer->text, stores the list with
1771 * newline separator into out, and saves the offsets of the original words
1772 * in buffer->orig.
1773 */
1774static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1775 regex_t *word_regex)
1776{
1777 int i, j;
1778 long alloc = 0;
1779 out->size = 0;
1781 out->ptr = NULL;
1782 /* fake an empty "0th" word */
1784 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1785 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1786 buffer->orig_nr = 1;
1787 for (i = 0; i < buffer->text.size; i++) {
1789 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1790 return;
1791 /* store original boundaries */
1793 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1794 buffer->orig_alloc);
1795 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1796 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1797 buffer->orig_nr++;
1798 /* store one word */
1800 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1801 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1802 out->ptr[out->size + j - i] = '\n';
1803 out->size += j - i + 1;
1804 i = j - 1;
1806 }
1807}
1808/* this executes the word diff on the accumulated buffers */
1810static void diff_words_show(struct diff_words_data *diff_words)
1811{
1812 xpparam_t xpp;
1813 xdemitconf_t xecfg;
1814 mmfile_t minus, plus;
1815 struct diff_words_style *style = diff_words->style;
1816 struct diff_options *opt = diff_words->opt;
1818 const char *line_prefix;
1819 assert(opt);
1821 line_prefix = diff_line_prefix(opt);
1822 /* special case: only removal */
1824 if (!diff_words->plus.text.size) {
1825 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1826 line_prefix, strlen(line_prefix), 0);
1827 fn_out_diff_words_write_helper(diff_words->opt,
1828 &style->old, style->newline,
1829 diff_words->minus.text.size,
1830 diff_words->minus.text.ptr);
1831 diff_words->minus.text.size = 0;
1832 return;
1833 }
1834 diff_words->current_plus = diff_words->plus.text.ptr;
1836 diff_words->last_minus = 0;
1837 memset(&xpp, 0, sizeof(xpp));
1839 memset(&xecfg, 0, sizeof(xecfg));
1840 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1841 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1842 xpp.flags = 0;
1843 /* as only the hunk header will be parsed, we need a 0-context */
1844 xecfg.ctxlen = 0;
1845 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1846 &xpp, &xecfg))
1847 die("unable to generate word diff");
1848 free(minus.ptr);
1849 free(plus.ptr);
1850 if (diff_words->current_plus != diff_words->plus.text.ptr +
1851 diff_words->plus.text.size) {
1852 if (color_words_output_graph_prefix(diff_words))
1853 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1854 line_prefix, strlen(line_prefix), 0);
1855 fn_out_diff_words_write_helper(diff_words->opt,
1856 &style->ctx, style->newline,
1857 diff_words->plus.text.ptr + diff_words->plus.text.size
1858 - diff_words->current_plus, diff_words->current_plus);
1859 }
1860 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1861}
1862/* In "color-words" mode, show word-diff of words accumulated in the buffer */
1864static void diff_words_flush(struct emit_callback *ecbdata)
1865{
1866 struct diff_options *wo = ecbdata->diff_words->opt;
1867 if (ecbdata->diff_words->minus.text.size ||
1869 ecbdata->diff_words->plus.text.size)
1870 diff_words_show(ecbdata->diff_words);
1871 if (wo->emitted_symbols) {
1873 struct diff_options *o = ecbdata->opt;
1874 struct emitted_diff_symbols *wol = wo->emitted_symbols;
1875 int i;
1876 /*
1878 * NEEDSWORK:
1879 * Instead of appending each, concat all words to a line?
1880 */
1881 for (i = 0; i < wol->nr; i++)
1882 append_emitted_diff_symbol(o, &wol->buf[i]);
1883 for (i = 0; i < wol->nr; i++)
1885 free((void *)wol->buf[i].line);
1886 wol->nr = 0;
1888 }
1889}
1890static void diff_filespec_load_driver(struct diff_filespec *one)
1892{
1893 /* Use already-loaded driver */
1894 if (one->driver)
1895 return;
1896 if (S_ISREG(one->mode))
1898 one->driver = userdiff_find_by_path(one->path);
1899 /* Fallback to default settings */
1901 if (!one->driver)
1902 one->driver = userdiff_find_by_name("default");
1903}
1904static const char *userdiff_word_regex(struct diff_filespec *one)
1906{
1907 diff_filespec_load_driver(one);
1908 return one->driver->word_regex;
1909}
1910static void init_diff_words_data(struct emit_callback *ecbdata,
1912 struct diff_options *orig_opts,
1913 struct diff_filespec *one,
1914 struct diff_filespec *two)
1915{
1916 int i;
1917 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1918 memcpy(o, orig_opts, sizeof(struct diff_options));
1919 ecbdata->diff_words =
1921 xcalloc(1, sizeof(struct diff_words_data));
1922 ecbdata->diff_words->type = o->word_diff;
1923 ecbdata->diff_words->opt = o;
1924 if (orig_opts->emitted_symbols)
1926 o->emitted_symbols =
1927 xcalloc(1, sizeof(struct emitted_diff_symbols));
1928 if (!o->word_regex)
1930 o->word_regex = userdiff_word_regex(one);
1931 if (!o->word_regex)
1932 o->word_regex = userdiff_word_regex(two);
1933 if (!o->word_regex)
1934 o->word_regex = diff_word_regex_cfg;
1935 if (o->word_regex) {
1936 ecbdata->diff_words->word_regex = (regex_t *)
1937 xmalloc(sizeof(regex_t));
1938 if (regcomp(ecbdata->diff_words->word_regex,
1939 o->word_regex,
1940 REG_EXTENDED | REG_NEWLINE))
1941 die ("Invalid regular expression: %s",
1942 o->word_regex);
1943 }
1944 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1945 if (o->word_diff == diff_words_styles[i].type) {
1946 ecbdata->diff_words->style =
1947 &diff_words_styles[i];
1948 break;
1949 }
1950 }
1951 if (want_color(o->use_color)) {
1952 struct diff_words_style *st = ecbdata->diff_words->style;
1953 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1954 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1955 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1956 }
1957}
1958static void free_diff_words_data(struct emit_callback *ecbdata)
1960{
1961 if (ecbdata->diff_words) {
1962 diff_words_flush(ecbdata);
1963 free (ecbdata->diff_words->opt->emitted_symbols);
1964 free (ecbdata->diff_words->opt);
1965 free (ecbdata->diff_words->minus.text.ptr);
1966 free (ecbdata->diff_words->minus.orig);
1967 free (ecbdata->diff_words->plus.text.ptr);
1968 free (ecbdata->diff_words->plus.orig);
1969 if (ecbdata->diff_words->word_regex) {
1970 regfree(ecbdata->diff_words->word_regex);
1971 free(ecbdata->diff_words->word_regex);
1972 }
1973 FREE_AND_NULL(ecbdata->diff_words);
1974 }
1975}
1976const char *diff_get_color(int diff_use_color, enum color_diff ix)
1978{
1979 if (want_color(diff_use_color))
1980 return diff_colors[ix];
1981 return "";
1982}
1983const char *diff_line_prefix(struct diff_options *opt)
1985{
1986 struct strbuf *msgbuf;
1987 if (!opt->output_prefix)
1988 return "";
1989 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1991 return msgbuf->buf;
1992}
1993static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1995{
1996 const char *cp;
1997 unsigned long allot;
1998 size_t l = len;
1999 cp = line;
2001 allot = l;
2002 while (0 < l) {
2003 (void) utf8_width(&cp, &l);
2004 if (!cp)
2005 break; /* truncated in the middle? */
2006 }
2007 return allot - l;
2008}
2009static void find_lno(const char *line, struct emit_callback *ecbdata)
2011{
2012 const char *p;
2013 ecbdata->lno_in_preimage = 0;
2014 ecbdata->lno_in_postimage = 0;
2015 p = strchr(line, '-');
2016 if (!p)
2017 return; /* cannot happen */
2018 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
2019 p = strchr(p, '+');
2020 if (!p)
2021 return; /* cannot happen */
2022 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
2023}
2024static void fn_out_consume(void *priv, char *line, unsigned long len)
2026{
2027 struct emit_callback *ecbdata = priv;
2028 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2029 struct diff_options *o = ecbdata->opt;
2030 o->found_changes = 1;
2032 if (ecbdata->header) {
2034 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2035 ecbdata->header->buf, ecbdata->header->len, 0);
2036 strbuf_reset(ecbdata->header);
2037 ecbdata->header = NULL;
2038 }
2039 if (ecbdata->label_path[0]) {
2041 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2042 ecbdata->label_path[0],
2043 strlen(ecbdata->label_path[0]), 0);
2044 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2045 ecbdata->label_path[1],
2046 strlen(ecbdata->label_path[1]), 0);
2047 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2048 }
2049 if (diff_suppress_blank_empty
2051 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2052 line[0] = '\n';
2053 len = 1;
2054 }
2055 if (line[0] == '@') {
2057 if (ecbdata->diff_words)
2058 diff_words_flush(ecbdata);
2059 len = sane_truncate_line(ecbdata, line, len);
2060 find_lno(line, ecbdata);
2061 emit_hunk_header(ecbdata, line, len);
2062 return;
2063 }
2064 if (ecbdata->diff_words) {
2066 enum diff_symbol s =
2067 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2068 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2069 if (line[0] == '-') {
2070 diff_words_append(line, len,
2071 &ecbdata->diff_words->minus);
2072 return;
2073 } else if (line[0] == '+') {
2074 diff_words_append(line, len,
2075 &ecbdata->diff_words->plus);
2076 return;
2077 } else if (starts_with(line, "\\ ")) {
2078 /*
2079 * Eat the "no newline at eof" marker as if we
2080 * saw a "+" or "-" line with nothing on it,
2081 * and return without diff_words_flush() to
2082 * defer processing. If this is the end of
2083 * preimage, more "+" lines may come after it.
2084 */
2085 return;
2086 }
2087 diff_words_flush(ecbdata);
2088 emit_diff_symbol(o, s, line, len, 0);
2089 return;
2090 }
2091 switch (line[0]) {
2093 case '+':
2094 ecbdata->lno_in_postimage++;
2095 emit_add_line(reset, ecbdata, line + 1, len - 1);
2096 break;
2097 case '-':
2098 ecbdata->lno_in_preimage++;
2099 emit_del_line(reset, ecbdata, line + 1, len - 1);
2100 break;
2101 case ' ':
2102 ecbdata->lno_in_postimage++;
2103 ecbdata->lno_in_preimage++;
2104 emit_context_line(reset, ecbdata, line + 1, len - 1);
2105 break;
2106 default:
2107 /* incomplete line at the end */
2108 ecbdata->lno_in_preimage++;
2109 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2110 line, len, 0);
2111 break;
2112 }
2113}
2114static char *pprint_rename(const char *a, const char *b)
2116{
2117 const char *old = a;
2118 const char *new = b;
2119 struct strbuf name = STRBUF_INIT;
2120 int pfx_length, sfx_length;
2121 int pfx_adjust_for_slash;
2122 int len_a = strlen(a);
2123 int len_b = strlen(b);
2124 int a_midlen, b_midlen;
2125 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2126 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2127 if (qlen_a || qlen_b) {
2129 quote_c_style(a, &name, NULL, 0);
2130 strbuf_addstr(&name, " => ");
2131 quote_c_style(b, &name, NULL, 0);
2132 return strbuf_detach(&name, NULL);
2133 }
2134 /* Find common prefix */
2136 pfx_length = 0;
2137 while (*old && *new && *old == *new) {
2138 if (*old == '/')
2139 pfx_length = old - a + 1;
2140 old++;
2141 new++;
2142 }
2143 /* Find common suffix */
2145 old = a + len_a;
2146 new = b + len_b;
2147 sfx_length = 0;
2148 /*
2149 * If there is a common prefix, it must end in a slash. In
2150 * that case we let this loop run 1 into the prefix to see the
2151 * same slash.
2152 *
2153 * If there is no common prefix, we cannot do this as it would
2154 * underrun the input strings.
2155 */
2156 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2157 while (a + pfx_length - pfx_adjust_for_slash <= old &&
2158 b + pfx_length - pfx_adjust_for_slash <= new &&
2159 *old == *new) {
2160 if (*old == '/')
2161 sfx_length = len_a - (old - a);
2162 old--;
2163 new--;
2164 }
2165 /*
2167 * pfx{mid-a => mid-b}sfx
2168 * {pfx-a => pfx-b}sfx
2169 * pfx{sfx-a => sfx-b}
2170 * name-a => name-b
2171 */
2172 a_midlen = len_a - pfx_length - sfx_length;
2173 b_midlen = len_b - pfx_length - sfx_length;
2174 if (a_midlen < 0)
2175 a_midlen = 0;
2176 if (b_midlen < 0)
2177 b_midlen = 0;
2178 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2180 if (pfx_length + sfx_length) {
2181 strbuf_add(&name, a, pfx_length);
2182 strbuf_addch(&name, '{');
2183 }
2184 strbuf_add(&name, a + pfx_length, a_midlen);
2185 strbuf_addstr(&name, " => ");
2186 strbuf_add(&name, b + pfx_length, b_midlen);
2187 if (pfx_length + sfx_length) {
2188 strbuf_addch(&name, '}');
2189 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
2190 }
2191 return strbuf_detach(&name, NULL);
2192}
2193struct diffstat_t {
2195 int nr;
2196 int alloc;
2197 struct diffstat_file {
2198 char *from_name;
2199 char *name;
2200 char *print_name;
2201 unsigned is_unmerged:1;
2202 unsigned is_binary:1;
2203 unsigned is_renamed:1;
2204 unsigned is_interesting:1;
2205 uintmax_t added, deleted;
2206 } **files;
2207};
2208static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2210 const char *name_a,
2211 const char *name_b)
2212{
2213 struct diffstat_file *x;
2214 x = xcalloc(1, sizeof(*x));
2215 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2216 diffstat->files[diffstat->nr++] = x;
2217 if (name_b) {
2218 x->from_name = xstrdup(name_a);
2219 x->name = xstrdup(name_b);
2220 x->is_renamed = 1;
2221 }
2222 else {
2223 x->from_name = NULL;
2224 x->name = xstrdup(name_a);
2225 }
2226 return x;
2227}
2228static void diffstat_consume(void *priv, char *line, unsigned long len)
2230{
2231 struct diffstat_t *diffstat = priv;
2232 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2233 if (line[0] == '+')
2235 x->added++;
2236 else if (line[0] == '-')
2237 x->deleted++;
2238}
2239const char mime_boundary_leader[] = "------------";
2241static int scale_linear(int it, int width, int max_change)
2243{
2244 if (!it)
2245 return 0;
2246 /*
2247 * make sure that at least one '-' or '+' is printed if
2248 * there is any change to this path. The easiest way is to
2249 * scale linearly as if the alloted width is one column shorter
2250 * than it is, and then add 1 to the result.
2251 */
2252 return 1 + (it * (width - 1) / max_change);
2253}
2254static void show_graph(struct strbuf *out, char ch, int cnt,
2256 const char *set, const char *reset)
2257{
2258 if (cnt <= 0)
2259 return;
2260 strbuf_addstr(out, set);
2261 strbuf_addchars(out, ch, cnt);
2262 strbuf_addstr(out, reset);
2263}
2264static void fill_print_name(struct diffstat_file *file)
2266{
2267 char *pname;
2268 if (file->print_name)
2270 return;
2271 if (!file->is_renamed) {
2273 struct strbuf buf = STRBUF_INIT;
2274 if (quote_c_style(file->name, &buf, NULL, 0)) {
2275 pname = strbuf_detach(&buf, NULL);
2276 } else {
2277 pname = file->name;
2278 strbuf_release(&buf);
2279 }
2280 } else {
2281 pname = pprint_rename(file->from_name, file->name);
2282 }
2283 file->print_name = pname;
2284}
2285static void print_stat_summary_inserts_deletes(struct diff_options *options,
2287 int files, int insertions, int deletions)
2288{
2289 struct strbuf sb = STRBUF_INIT;
2290 if (!files) {
2292 assert(insertions == 0 && deletions == 0);
2293 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2294 NULL, 0, 0);
2295 return;
2296 }
2297 strbuf_addf(&sb,
2299 (files == 1) ? " %d file changed" : " %d files changed",
2300 files);
2301 /*
2303 * For binary diff, the caller may want to print "x files
2304 * changed" with insertions == 0 && deletions == 0.
2305 *
2306 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2307 * is probably less confusing (i.e skip over "2 files changed
2308 * but nothing about added/removed lines? Is this a bug in Git?").
2309 */
2310 if (insertions || deletions == 0) {
2311 strbuf_addf(&sb,
2312 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2313 insertions);
2314 }
2315 if (deletions || insertions == 0) {
2317 strbuf_addf(&sb,
2318 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2319 deletions);
2320 }
2321 strbuf_addch(&sb, '\n');
2322 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2323 sb.buf, sb.len, 0);
2324 strbuf_release(&sb);
2325}
2326void print_stat_summary(FILE *fp, int files,
2328 int insertions, int deletions)
2329{
2330 struct diff_options o;
2331 memset(&o, 0, sizeof(o));
2332 o.file = fp;
2333 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2335}
2336static void show_stats(struct diffstat_t *data, struct diff_options *options)
2338{
2339 int i, len, add, del, adds = 0, dels = 0;
2340 uintmax_t max_change = 0, max_len = 0;
2341 int total_files = data->nr, count;
2342 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2343 const char *reset, *add_c, *del_c;
2344 int extra_shown = 0;
2345 const char *line_prefix = diff_line_prefix(options);
2346 struct strbuf out = STRBUF_INIT;
2347 if (data->nr == 0)
2349 return;
2350 count = options->stat_count ? options->stat_count : data->nr;
2352 reset = diff_get_color_opt(options, DIFF_RESET);
2354 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2355 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2356 /*
2358 * Find the longest filename and max number of changes
2359 */
2360 for (i = 0; (i < count) && (i < data->nr); i++) {
2361 struct diffstat_file *file = data->files[i];
2362 uintmax_t change = file->added + file->deleted;
2363 if (!file->is_interesting && (change == 0)) {
2365 count++; /* not shown == room for one more */
2366 continue;
2367 }
2368 fill_print_name(file);
2369 len = strlen(file->print_name);
2370 if (max_len < len)
2371 max_len = len;
2372 if (file->is_unmerged) {
2374 /* "Unmerged" is 8 characters */
2375 bin_width = bin_width < 8 ? 8 : bin_width;
2376 continue;
2377 }
2378 if (file->is_binary) {
2379 /* "Bin XXX -> YYY bytes" */
2380 int w = 14 + decimal_width(file->added)
2381 + decimal_width(file->deleted);
2382 bin_width = bin_width < w ? w : bin_width;
2383 /* Display change counts aligned with "Bin" */
2384 number_width = 3;
2385 continue;
2386 }
2387 if (max_change < change)
2389 max_change = change;
2390 }
2391 count = i; /* where we can stop scanning in data->files[] */
2392 /*
2394 * We have width = stat_width or term_columns() columns total.
2395 * We want a maximum of min(max_len, stat_name_width) for the name part.
2396 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2397 * We also need 1 for " " and 4 + decimal_width(max_change)
2398 * for " | NNNN " and one the empty column at the end, altogether
2399 * 6 + decimal_width(max_change).
2400 *
2401 * If there's not enough space, we will use the smaller of
2402 * stat_name_width (if set) and 5/8*width for the filename,
2403 * and the rest for constant elements + graph part, but no more
2404 * than stat_graph_width for the graph part.
2405 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2406 * for the standard terminal size).
2407 *
2408 * In other words: stat_width limits the maximum width, and
2409 * stat_name_width fixes the maximum width of the filename,
2410 * and is also used to divide available columns if there
2411 * aren't enough.
2412 *
2413 * Binary files are displayed with "Bin XXX -> YYY bytes"
2414 * instead of the change count and graph. This part is treated
2415 * similarly to the graph part, except that it is not
2416 * "scaled". If total width is too small to accommodate the
2417 * guaranteed minimum width of the filename part and the
2418 * separators and this message, this message will "overflow"
2419 * making the line longer than the maximum width.
2420 */
2421 if (options->stat_width == -1)
2423 width = term_columns() - strlen(line_prefix);
2424 else
2425 width = options->stat_width ? options->stat_width : 80;
2426 number_width = decimal_width(max_change) > number_width ?
2427 decimal_width(max_change) : number_width;
2428 if (options->stat_graph_width == -1)
2430 options->stat_graph_width = diff_stat_graph_width;
2431 /*
2433 * Guarantee 3/8*16==6 for the graph part
2434 * and 5/8*16==10 for the filename part
2435 */
2436 if (width < 16 + 6 + number_width)
2437 width = 16 + 6 + number_width;
2438 /*
2440 * First assign sizes that are wanted, ignoring available width.
2441 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2442 * starting from "XXX" should fit in graph_width.
2443 */
2444 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2445 if (options->stat_graph_width &&
2446 options->stat_graph_width < graph_width)
2447 graph_width = options->stat_graph_width;
2448 name_width = (options->stat_name_width > 0 &&
2450 options->stat_name_width < max_len) ?
2451 options->stat_name_width : max_len;
2452 /*
2454 * Adjust adjustable widths not to exceed maximum width
2455 */
2456 if (name_width + number_width + 6 + graph_width > width) {
2457 if (graph_width > width * 3/8 - number_width - 6) {
2458 graph_width = width * 3/8 - number_width - 6;
2459 if (graph_width < 6)
2460 graph_width = 6;
2461 }
2462 if (options->stat_graph_width &&
2464 graph_width > options->stat_graph_width)
2465 graph_width = options->stat_graph_width;
2466 if (name_width > width - number_width - 6 - graph_width)
2467 name_width = width - number_width - 6 - graph_width;
2468 else
2469 graph_width = width - number_width - 6 - name_width;
2470 }
2471 /*
2473 * From here name_width is the width of the name area,
2474 * and graph_width is the width of the graph area.
2475 * max_change is used to scale graph properly.
2476 */
2477 for (i = 0; i < count; i++) {
2478 const char *prefix = "";
2479 struct diffstat_file *file = data->files[i];
2480 char *name = file->print_name;
2481 uintmax_t added = file->added;
2482 uintmax_t deleted = file->deleted;
2483 int name_len;
2484 if (!file->is_interesting && (added + deleted == 0))
2486 continue;
2487 /*
2489 * "scale" the filename
2490 */
2491 len = name_width;
2492 name_len = strlen(name);
2493 if (name_width < name_len) {
2494 char *slash;
2495 prefix = "...";
2496 len -= 3;
2497 name += name_len - len;
2498 slash = strchr(name, '/');
2499 if (slash)
2500 name = slash;
2501 }
2502 if (file->is_binary) {
2504 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2505 strbuf_addf(&out, " %*s", number_width, "Bin");
2506 if (!added && !deleted) {
2507 strbuf_addch(&out, '\n');
2508 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2509 out.buf, out.len, 0);
2510 strbuf_reset(&out);
2511 continue;
2512 }
2513 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2514 del_c, deleted, reset);
2515 strbuf_addstr(&out, " -> ");
2516 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2517 add_c, added, reset);
2518 strbuf_addstr(&out, " bytes\n");
2519 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2520 out.buf, out.len, 0);
2521 strbuf_reset(&out);
2522 continue;
2523 }
2524 else if (file->is_unmerged) {
2525 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2526 strbuf_addstr(&out, " Unmerged\n");
2527 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2528 out.buf, out.len, 0);
2529 strbuf_reset(&out);
2530 continue;
2531 }
2532 /*
2534 * scale the add/delete
2535 */
2536 add = added;
2537 del = deleted;
2538 if (graph_width <= max_change) {
2540 int total = scale_linear(add + del, graph_width, max_change);
2541 if (total < 2 && add && del)
2542 /* width >= 2 due to the sanity check */
2543 total = 2;
2544 if (add < del) {
2545 add = scale_linear(add, graph_width, max_change);
2546 del = total - add;
2547 } else {
2548 del = scale_linear(del, graph_width, max_change);
2549 add = total - del;
2550 }
2551 }
2552 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2553 strbuf_addf(&out, " %*"PRIuMAX"%s",
2554 number_width, added + deleted,
2555 added + deleted ? " " : "");
2556 show_graph(&out, '+', add, add_c, reset);
2557 show_graph(&out, '-', del, del_c, reset);
2558 strbuf_addch(&out, '\n');
2559 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2560 out.buf, out.len, 0);
2561 strbuf_reset(&out);
2562 }
2563 for (i = 0; i < data->nr; i++) {
2565 struct diffstat_file *file = data->files[i];
2566 uintmax_t added = file->added;
2567 uintmax_t deleted = file->deleted;
2568 if (file->is_unmerged ||
2570 (!file->is_interesting && (added + deleted == 0))) {
2571 total_files--;
2572 continue;
2573 }
2574 if (!file->is_binary) {
2576 adds += added;
2577 dels += deleted;
2578 }
2579 if (i < count)
2580 continue;
2581 if (!extra_shown)
2582 emit_diff_symbol(options,
2583 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2584 NULL, 0, 0);
2585 extra_shown = 1;
2586 }
2587 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2589}
2590static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2592{
2593 int i, adds = 0, dels = 0, total_files = data->nr;
2594 if (data->nr == 0)
2596 return;
2597 for (i = 0; i < data->nr; i++) {
2599 int added = data->files[i]->added;
2600 int deleted = data->files[i]->deleted;
2601 if (data->files[i]->is_unmerged ||
2603 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2604 total_files--;
2605 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2606 adds += added;
2607 dels += deleted;
2608 }
2609 }
2610 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2611}
2612static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2614{
2615 int i;
2616 if (data->nr == 0)
2618 return;
2619 for (i = 0; i < data->nr; i++) {
2621 struct diffstat_file *file = data->files[i];
2622 fprintf(options->file, "%s", diff_line_prefix(options));
2624 if (file->is_binary)
2626 fprintf(options->file, "-\t-\t");
2627 else
2628 fprintf(options->file,
2629 "%"PRIuMAX"\t%"PRIuMAX"\t",
2630 file->added, file->deleted);
2631 if (options->line_termination) {
2632 fill_print_name(file);
2633 if (!file->is_renamed)
2634 write_name_quoted(file->name, options->file,
2635 options->line_termination);
2636 else {
2637 fputs(file->print_name, options->file);
2638 putc(options->line_termination, options->file);
2639 }
2640 } else {
2641 if (file->is_renamed) {
2642 putc('\0', options->file);
2643 write_name_quoted(file->from_name, options->file, '\0');
2644 }
2645 write_name_quoted(file->name, options->file, '\0');
2646 }
2647 }
2648}
2649struct dirstat_file {
2651 const char *name;
2652 unsigned long changed;
2653};
2654struct dirstat_dir {
2656 struct dirstat_file *files;
2657 int alloc, nr, permille, cumulative;
2658};
2659static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2661 unsigned long changed, const char *base, int baselen)
2662{
2663 unsigned long this_dir = 0;
2664 unsigned int sources = 0;
2665 const char *line_prefix = diff_line_prefix(opt);
2666 while (dir->nr) {
2668 struct dirstat_file *f = dir->files;
2669 int namelen = strlen(f->name);
2670 unsigned long this;
2671 char *slash;
2672 if (namelen < baselen)
2674 break;
2675 if (memcmp(f->name, base, baselen))
2676 break;
2677 slash = strchr(f->name + baselen, '/');
2678 if (slash) {
2679 int newbaselen = slash + 1 - f->name;
2680 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2681 sources++;
2682 } else {
2683 this = f->changed;
2684 dir->files++;
2685 dir->nr--;
2686 sources += 2;
2687 }
2688 this_dir += this;
2689 }
2690 /*
2692 * We don't report dirstat's for
2693 * - the top level
2694 * - or cases where everything came from a single directory
2695 * under this directory (sources == 1).
2696 */
2697 if (baselen && sources != 1) {
2698 if (this_dir) {
2699 int permille = this_dir * 1000 / changed;
2700 if (permille >= dir->permille) {
2701 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2702 permille / 10, permille % 10, baselen, base);
2703 if (!dir->cumulative)
2704 return 0;
2705 }
2706 }
2707 }
2708 return this_dir;
2709}
2710static int dirstat_compare(const void *_a, const void *_b)
2712{
2713 const struct dirstat_file *a = _a;
2714 const struct dirstat_file *b = _b;
2715 return strcmp(a->name, b->name);
2716}
2717static void show_dirstat(struct diff_options *options)
2719{
2720 int i;
2721 unsigned long changed;
2722 struct dirstat_dir dir;
2723 struct diff_queue_struct *q = &diff_queued_diff;
2724 dir.files = NULL;
2726 dir.alloc = 0;
2727 dir.nr = 0;
2728 dir.permille = options->dirstat_permille;
2729 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2730 changed = 0;
2732 for (i = 0; i < q->nr; i++) {
2733 struct diff_filepair *p = q->queue[i];
2734 const char *name;
2735 unsigned long copied, added, damage;
2736 int content_changed;
2737 name = p->two->path ? p->two->path : p->one->path;
2739 if (p->one->oid_valid && p->two->oid_valid)
2741 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2742 else
2743 content_changed = 1;
2744 if (!content_changed) {
2746 /*
2747 * The SHA1 has not changed, so pre-/post-content is
2748 * identical. We can therefore skip looking at the
2749 * file contents altogether.
2750 */
2751 damage = 0;
2752 goto found_damage;
2753 }
2754 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2756 /*
2757 * In --dirstat-by-file mode, we don't really need to
2758 * look at the actual file contents at all.
2759 * The fact that the SHA1 changed is enough for us to
2760 * add this file to the list of results
2761 * (with each file contributing equal damage).
2762 */
2763 damage = 1;
2764 goto found_damage;
2765 }
2766 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2768 diff_populate_filespec(p->one, 0);
2769 diff_populate_filespec(p->two, 0);
2770 diffcore_count_changes(p->one, p->two, NULL, NULL,
2771 &copied, &added);
2772 diff_free_filespec_data(p->one);
2773 diff_free_filespec_data(p->two);
2774 } else if (DIFF_FILE_VALID(p->one)) {
2775 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2776 copied = added = 0;
2777 diff_free_filespec_data(p->one);
2778 } else if (DIFF_FILE_VALID(p->two)) {
2779 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2780 copied = 0;
2781 added = p->two->size;
2782 diff_free_filespec_data(p->two);
2783 } else
2784 continue;
2785 /*
2787 * Original minus copied is the removed material,
2788 * added is the new material. They are both damages
2789 * made to the preimage.
2790 * If the resulting damage is zero, we know that
2791 * diffcore_count_changes() considers the two entries to
2792 * be identical, but since content_changed is true, we
2793 * know that there must have been _some_ kind of change,
2794 * so we force all entries to have damage > 0.
2795 */
2796 damage = (p->one->size - copied) + added;
2797 if (!damage)
2798 damage = 1;
2799found_damage:
2801 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2802 dir.files[dir.nr].name = name;
2803 dir.files[dir.nr].changed = damage;
2804 changed += damage;
2805 dir.nr++;
2806 }
2807 /* This can happen even with many files, if everything was renames */
2809 if (!changed)
2810 return;
2811 /* Show all directories with more than x% of the changes */
2813 QSORT(dir.files, dir.nr, dirstat_compare);
2814 gather_dirstat(options, &dir, changed, "", 0);
2815}
2816static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2818{
2819 int i;
2820 unsigned long changed;
2821 struct dirstat_dir dir;
2822 if (data->nr == 0)
2824 return;
2825 dir.files = NULL;
2827 dir.alloc = 0;
2828 dir.nr = 0;
2829 dir.permille = options->dirstat_permille;
2830 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2831 changed = 0;
2833 for (i = 0; i < data->nr; i++) {
2834 struct diffstat_file *file = data->files[i];
2835 unsigned long damage = file->added + file->deleted;
2836 if (file->is_binary)
2837 /*
2838 * binary files counts bytes, not lines. Must find some
2839 * way to normalize binary bytes vs. textual lines.
2840 * The following heuristic assumes that there are 64
2841 * bytes per "line".
2842 * This is stupid and ugly, but very cheap...
2843 */
2844 damage = DIV_ROUND_UP(damage, 64);
2845 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2846 dir.files[dir.nr].name = file->name;
2847 dir.files[dir.nr].changed = damage;
2848 changed += damage;
2849 dir.nr++;
2850 }
2851 /* This can happen even with many files, if everything was renames */
2853 if (!changed)
2854 return;
2855 /* Show all directories with more than x% of the changes */
2857 QSORT(dir.files, dir.nr, dirstat_compare);
2858 gather_dirstat(options, &dir, changed, "", 0);
2859}
2860static void free_diffstat_info(struct diffstat_t *diffstat)
2862{
2863 int i;
2864 for (i = 0; i < diffstat->nr; i++) {
2865 struct diffstat_file *f = diffstat->files[i];
2866 if (f->name != f->print_name)
2867 free(f->print_name);
2868 free(f->name);
2869 free(f->from_name);
2870 free(f);
2871 }
2872 free(diffstat->files);
2873}
2874struct checkdiff_t {
2876 const char *filename;
2877 int lineno;
2878 int conflict_marker_size;
2879 struct diff_options *o;
2880 unsigned ws_rule;
2881 unsigned status;
2882};
2883static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2885{
2886 char firstchar;
2887 int cnt;
2888 if (len < marker_size + 1)
2890 return 0;
2891 firstchar = line[0];
2892 switch (firstchar) {
2893 case '=': case '>': case '<': case '|':
2894 break;
2895 default:
2896 return 0;
2897 }
2898 for (cnt = 1; cnt < marker_size; cnt++)
2899 if (line[cnt] != firstchar)
2900 return 0;
2901 /* line[1] thru line[marker_size-1] are same as firstchar */
2902 if (len < marker_size + 1 || !isspace(line[marker_size]))
2903 return 0;
2904 return 1;
2905}
2906static void checkdiff_consume(void *priv, char *line, unsigned long len)
2908{
2909 struct checkdiff_t *data = priv;
2910 int marker_size = data->conflict_marker_size;
2911 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2912 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2913 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2914 char *err;
2915 const char *line_prefix;
2916 assert(data->o);
2918 line_prefix = diff_line_prefix(data->o);
2919 if (line[0] == '+') {
2921 unsigned bad;
2922 data->lineno++;
2923 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2924 data->status |= 1;
2925 fprintf(data->o->file,
2926 "%s%s:%d: leftover conflict marker\n",
2927 line_prefix, data->filename, data->lineno);
2928 }
2929 bad = ws_check(line + 1, len - 1, data->ws_rule);
2930 if (!bad)
2931 return;
2932 data->status |= bad;
2933 err = whitespace_error_string(bad);
2934 fprintf(data->o->file, "%s%s:%d: %s.\n",
2935 line_prefix, data->filename, data->lineno, err);
2936 free(err);
2937 emit_line(data->o, set, reset, line, 1);
2938 ws_check_emit(line + 1, len - 1, data->ws_rule,
2939 data->o->file, set, reset, ws);
2940 } else if (line[0] == ' ') {
2941 data->lineno++;
2942 } else if (line[0] == '@') {
2943 char *plus = strchr(line, '+');
2944 if (plus)
2945 data->lineno = strtol(plus, NULL, 10) - 1;
2946 else
2947 die("invalid diff");
2948 }
2949}
2950static unsigned char *deflate_it(char *data,
2952 unsigned long size,
2953 unsigned long *result_size)
2954{
2955 int bound;
2956 unsigned char *deflated;
2957 git_zstream stream;
2958 git_deflate_init(&stream, zlib_compression_level);
2960 bound = git_deflate_bound(&stream, size);
2961 deflated = xmalloc(bound);
2962 stream.next_out = deflated;
2963 stream.avail_out = bound;
2964 stream.next_in = (unsigned char *)data;
2966 stream.avail_in = size;
2967 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2968 ; /* nothing */
2969 git_deflate_end(&stream);
2970 *result_size = stream.total_out;
2971 return deflated;
2972}
2973static void emit_binary_diff_body(struct diff_options *o,
2975 mmfile_t *one, mmfile_t *two)
2976{
2977 void *cp;
2978 void *delta;
2979 void *deflated;
2980 void *data;
2981 unsigned long orig_size;
2982 unsigned long delta_size;
2983 unsigned long deflate_size;
2984 unsigned long data_size;
2985 /* We could do deflated delta, or we could do just deflated two,
2987 * whichever is smaller.
2988 */
2989 delta = NULL;
2990 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2991 if (one->size && two->size) {
2992 delta = diff_delta(one->ptr, one->size,
2993 two->ptr, two->size,
2994 &delta_size, deflate_size);
2995 if (delta) {
2996 void *to_free = delta;
2997 orig_size = delta_size;
2998 delta = deflate_it(delta, delta_size, &delta_size);
2999 free(to_free);
3000 }
3001 }
3002 if (delta && delta_size < deflate_size) {
3004 char *s = xstrfmt("%lu", orig_size);
3005 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
3006 s, strlen(s), 0);
3007 free(s);
3008 free(deflated);
3009 data = delta;
3010 data_size = delta_size;
3011 } else {
3012 char *s = xstrfmt("%lu", two->size);
3013 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
3014 s, strlen(s), 0);
3015 free(s);
3016 free(delta);
3017 data = deflated;
3018 data_size = deflate_size;
3019 }
3020 /* emit data encoded in base85 */
3022 cp = data;
3023 while (data_size) {
3024 int len;
3025 int bytes = (52 < data_size) ? 52 : data_size;
3026 char line[71];
3027 data_size -= bytes;
3028 if (bytes <= 26)
3029 line[0] = bytes + 'A' - 1;
3030 else
3031 line[0] = bytes - 26 + 'a' - 1;
3032 encode_85(line + 1, cp, bytes);
3033 cp = (char *) cp + bytes;
3034 len = strlen(line);
3036 line[len++] = '\n';
3037 line[len] = '\0';
3038 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3040 line, len, 0);
3041 }
3042 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3043 free(data);
3044}
3045static void emit_binary_diff(struct diff_options *o,
3047 mmfile_t *one, mmfile_t *two)
3048{
3049 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3050 emit_binary_diff_body(o, one, two);
3051 emit_binary_diff_body(o, two, one);
3052}
3053int diff_filespec_is_binary(struct diff_filespec *one)
3055{
3056 if (one->is_binary == -1) {
3057 diff_filespec_load_driver(one);
3058 if (one->driver->binary != -1)
3059 one->is_binary = one->driver->binary;
3060 else {
3061 if (!one->data && DIFF_FILE_VALID(one))
3062 diff_populate_filespec(one, CHECK_BINARY);
3063 if (one->is_binary == -1 && one->data)
3064 one->is_binary = buffer_is_binary(one->data,
3065 one->size);
3066 if (one->is_binary == -1)
3067 one->is_binary = 0;
3068 }
3069 }
3070 return one->is_binary;
3071}
3072static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3074{
3075 diff_filespec_load_driver(one);
3076 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3077}
3078void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3080{
3081 if (!options->a_prefix)
3082 options->a_prefix = a;
3083 if (!options->b_prefix)
3084 options->b_prefix = b;
3085}
3086struct userdiff_driver *get_textconv(struct diff_filespec *one)
3088{
3089 if (!DIFF_FILE_VALID(one))
3090 return NULL;
3091 diff_filespec_load_driver(one);
3093 return userdiff_get_textconv(one->driver);
3094}
3095static void builtin_diff(const char *name_a,
3097 const char *name_b,
3098 struct diff_filespec *one,
3099 struct diff_filespec *two,
3100 const char *xfrm_msg,
3101 int must_show_header,
3102 struct diff_options *o,
3103 int complete_rewrite)
3104{
3105 mmfile_t mf1, mf2;
3106 const char *lbl[2];
3107 char *a_one, *b_two;
3108 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3109 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3110 const char *a_prefix, *b_prefix;
3111 struct userdiff_driver *textconv_one = NULL;
3112 struct userdiff_driver *textconv_two = NULL;
3113 struct strbuf header = STRBUF_INIT;
3114 const char *line_prefix = diff_line_prefix(o);
3115 diff_set_mnemonic_prefix(o, "a/", "b/");
3117 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
3118 a_prefix = o->b_prefix;
3119 b_prefix = o->a_prefix;
3120 } else {
3121 a_prefix = o->a_prefix;
3122 b_prefix = o->b_prefix;
3123 }
3124 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3126 (!one->mode || S_ISGITLINK(one->mode)) &&
3127 (!two->mode || S_ISGITLINK(two->mode))) {
3128 show_submodule_summary(o, one->path ? one->path : two->path,
3129 &one->oid, &two->oid,
3130 two->dirty_submodule);
3131 return;
3132 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3133 (!one->mode || S_ISGITLINK(one->mode)) &&
3134 (!two->mode || S_ISGITLINK(two->mode))) {
3135 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3136 &one->oid, &two->oid,
3137 two->dirty_submodule);
3138 return;
3139 }
3140 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
3142 textconv_one = get_textconv(one);
3143 textconv_two = get_textconv(two);
3144 }
3145 /* Never use a non-valid filename anywhere if at all possible */
3147 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3148 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3149 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3151 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3152 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3153 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3154 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3155 if (lbl[0][0] == '/') {
3156 /* /dev/null */
3157 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3158 if (xfrm_msg)
3159 strbuf_addstr(&header, xfrm_msg);
3160 must_show_header = 1;
3161 }
3162 else if (lbl[1][0] == '/') {
3163 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3164 if (xfrm_msg)
3165 strbuf_addstr(&header, xfrm_msg);
3166 must_show_header = 1;
3167 }
3168 else {
3169 if (one->mode != two->mode) {
3170 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3171 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3172 must_show_header = 1;
3173 }
3174 if (xfrm_msg)
3175 strbuf_addstr(&header, xfrm_msg);
3176 /*
3178 * we do not run diff between different kind
3179 * of objects.
3180 */
3181 if ((one->mode ^ two->mode) & S_IFMT)
3182 goto free_ab_and_return;
3183 if (complete_rewrite &&
3184 (textconv_one || !diff_filespec_is_binary(one)) &&
3185 (textconv_two || !diff_filespec_is_binary(two))) {
3186 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3187 header.buf, header.len, 0);
3188 strbuf_reset(&header);
3189 emit_rewrite_diff(name_a, name_b, one, two,
3190 textconv_one, textconv_two, o);
3191 o->found_changes = 1;
3192 goto free_ab_and_return;
3193 }
3194 }
3195 if (o->irreversible_delete && lbl[1][0] == '/') {
3197 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3198 header.len, 0);
3199 strbuf_reset(&header);
3200 goto free_ab_and_return;
3201 } else if (!DIFF_OPT_TST(o, TEXT) &&
3202 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3203 (!textconv_two && diff_filespec_is_binary(two)) )) {
3204 struct strbuf sb = STRBUF_INIT;
3205 if (!one->data && !two->data &&
3206 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3207 !DIFF_OPT_TST(o, BINARY)) {
3208 if (!oidcmp(&one->oid, &two->oid)) {
3209 if (must_show_header)
3210 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3211 header.buf, header.len,
3212 0);
3213 goto free_ab_and_return;
3214 }
3215 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3216 header.buf, header.len, 0);
3217 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3218 diff_line_prefix(o), lbl[0], lbl[1]);
3219 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3220 sb.buf, sb.len, 0);
3221 strbuf_release(&sb);
3222 goto free_ab_and_return;
3223 }
3224 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3225 die("unable to read files to diff");
3226 /* Quite common confusing case */
3227 if (mf1.size == mf2.size &&
3228 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3229 if (must_show_header)
3230 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3231 header.buf, header.len, 0);
3232 goto free_ab_and_return;
3233 }
3234 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3235 strbuf_reset(&header);
3236 if (DIFF_OPT_TST(o, BINARY))
3237 emit_binary_diff(o, &mf1, &mf2);
3238 else {
3239 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3240 diff_line_prefix(o), lbl[0], lbl[1]);
3241 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3242 sb.buf, sb.len, 0);
3243 strbuf_release(&sb);
3244 }
3245 o->found_changes = 1;
3246 } else {
3247 /* Crazy xdl interfaces.. */
3248 const char *diffopts = getenv("GIT_DIFF_OPTS");
3249 const char *v;
3250 xpparam_t xpp;
3251 xdemitconf_t xecfg;
3252 struct emit_callback ecbdata;
3253 const struct userdiff_funcname *pe;
3254 if (must_show_header) {
3256 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3257 header.buf, header.len, 0);
3258 strbuf_reset(&header);
3259 }
3260 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3262 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3263 pe = diff_funcname_pattern(one);
3265 if (!pe)
3266 pe = diff_funcname_pattern(two);
3267 memset(&xpp, 0, sizeof(xpp));
3269 memset(&xecfg, 0, sizeof(xecfg));
3270 memset(&ecbdata, 0, sizeof(ecbdata));
3271 ecbdata.label_path = lbl;
3272 ecbdata.color_diff = want_color(o->use_color);
3273 ecbdata.ws_rule = whitespace_rule(name_b);
3274 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3275 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3276 ecbdata.opt = o;
3277 ecbdata.header = header.len ? &header : NULL;
3278 xpp.flags = o->xdl_opts;
3279 xecfg.ctxlen = o->context;
3280 xecfg.interhunkctxlen = o->interhunkcontext;
3281 xecfg.flags = XDL_EMIT_FUNCNAMES;
3282 if (DIFF_OPT_TST(o, FUNCCONTEXT))
3283 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3284 if (pe)
3285 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3286 if (!diffopts)
3287 ;
3288 else if (skip_prefix(diffopts, "--unified=", &v))
3289 xecfg.ctxlen = strtoul(v, NULL, 10);
3290 else if (skip_prefix(diffopts, "-u", &v))
3291 xecfg.ctxlen = strtoul(v, NULL, 10);
3292 if (o->word_diff)
3293 init_diff_words_data(&ecbdata, o, one, two);
3294 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
3295 &xpp, &xecfg))
3296 die("unable to generate diff for %s", one->path);
3297 if (o->word_diff)
3298 free_diff_words_data(&ecbdata);
3299 if (textconv_one)
3300 free(mf1.ptr);
3301 if (textconv_two)
3302 free(mf2.ptr);
3303 xdiff_clear_find_func(&xecfg);
3304 }
3305 free_ab_and_return:
3307 strbuf_release(&header);
3308 diff_free_filespec_data(one);
3309 diff_free_filespec_data(two);
3310 free(a_one);
3311 free(b_two);
3312 return;
3313}
3314static void builtin_diffstat(const char *name_a, const char *name_b,
3316 struct diff_filespec *one,
3317 struct diff_filespec *two,
3318 struct diffstat_t *diffstat,
3319 struct diff_options *o,
3320 struct diff_filepair *p)
3321{
3322 mmfile_t mf1, mf2;
3323 struct diffstat_file *data;
3324 int same_contents;
3325 int complete_rewrite = 0;
3326 if (!DIFF_PAIR_UNMERGED(p)) {
3328 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3329 complete_rewrite = 1;
3330 }
3331 data = diffstat_add(diffstat, name_a, name_b);
3333 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3334 if (!one || !two) {
3336 data->is_unmerged = 1;
3337 return;
3338 }
3339 same_contents = !oidcmp(&one->oid, &two->oid);
3341 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3343 data->is_binary = 1;
3344 if (same_contents) {
3345 data->added = 0;
3346 data->deleted = 0;
3347 } else {
3348 data->added = diff_filespec_size(two);
3349 data->deleted = diff_filespec_size(one);
3350 }
3351 }
3352 else if (complete_rewrite) {
3354 diff_populate_filespec(one, 0);
3355 diff_populate_filespec(two, 0);
3356 data->deleted = count_lines(one->data, one->size);
3357 data->added = count_lines(two->data, two->size);
3358 }
3359 else if (!same_contents) {
3361 /* Crazy xdl interfaces.. */
3362 xpparam_t xpp;
3363 xdemitconf_t xecfg;
3364 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3366 die("unable to read files to diff");
3367 memset(&xpp, 0, sizeof(xpp));
3369 memset(&xecfg, 0, sizeof(xecfg));
3370 xpp.flags = o->xdl_opts;
3371 xecfg.ctxlen = o->context;
3372 xecfg.interhunkctxlen = o->interhunkcontext;
3373 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
3374 &xpp, &xecfg))
3375 die("unable to generate diffstat for %s", one->path);
3376 }
3377 diff_free_filespec_data(one);
3379 diff_free_filespec_data(two);
3380}
3381static void builtin_checkdiff(const char *name_a, const char *name_b,
3383 const char *attr_path,
3384 struct diff_filespec *one,
3385 struct diff_filespec *two,
3386 struct diff_options *o)
3387{
3388 mmfile_t mf1, mf2;
3389 struct checkdiff_t data;
3390 if (!two)
3392 return;
3393 memset(&data, 0, sizeof(data));
3395 data.filename = name_b ? name_b : name_a;
3396 data.lineno = 0;
3397 data.o = o;
3398 data.ws_rule = whitespace_rule(attr_path);
3399 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3400 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3402 die("unable to read files to diff");
3403 /*
3405 * All the other codepaths check both sides, but not checking
3406 * the "old" side here is deliberate. We are checking the newly
3407 * introduced changes, and as long as the "new" side is text, we
3408 * can and should check what it introduces.
3409 */
3410 if (diff_filespec_is_binary(two))
3411 goto free_and_return;
3412 else {
3413 /* Crazy xdl interfaces.. */
3414 xpparam_t xpp;
3415 xdemitconf_t xecfg;
3416 memset(&xpp, 0, sizeof(xpp));
3418 memset(&xecfg, 0, sizeof(xecfg));
3419 xecfg.ctxlen = 1; /* at least one context line */
3420 xpp.flags = 0;
3421 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
3422 &xpp, &xecfg))
3423 die("unable to generate checkdiff for %s", one->path);
3424 if (data.ws_rule & WS_BLANK_AT_EOF) {
3426 struct emit_callback ecbdata;
3427 int blank_at_eof;
3428 ecbdata.ws_rule = data.ws_rule;
3430 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3431 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3432 if (blank_at_eof) {
3434 static char *err;
3435 if (!err)
3436 err = whitespace_error_string(WS_BLANK_AT_EOF);
3437 fprintf(o->file, "%s:%d: %s.\n",
3438 data.filename, blank_at_eof, err);
3439 data.status = 1; /* report errors */
3440 }
3441 }
3442 }
3443 free_and_return:
3444 diff_free_filespec_data(one);
3445 diff_free_filespec_data(two);
3446 if (data.status)
3447 DIFF_OPT_SET(o, CHECK_FAILED);
3448}
3449struct diff_filespec *alloc_filespec(const char *path)
3451{
3452 struct diff_filespec *spec;
3453 FLEXPTR_ALLOC_STR(spec, path, path);
3455 spec->count = 1;
3456 spec->is_binary = -1;
3457 return spec;
3458}
3459void free_filespec(struct diff_filespec *spec)
3461{
3462 if (!--spec->count) {
3463 diff_free_filespec_data(spec);
3464 free(spec);
3465 }
3466}
3467void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3469 int oid_valid, unsigned short mode)
3470{
3471 if (mode) {
3472 spec->mode = canon_mode(mode);
3473 oidcpy(&spec->oid, oid);
3474 spec->oid_valid = oid_valid;
3475 }
3476}
3477/*
3479 * Given a name and sha1 pair, if the index tells us the file in
3480 * the work tree has that object contents, return true, so that
3481 * prepare_temp_file() does not have to inflate and extract.
3482 */
3483static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3484{
3485 const struct cache_entry *ce;
3486 struct stat st;
3487 int pos, len;
3488 /*
3490 * We do not read the cache ourselves here, because the
3491 * benchmark with my previous version that always reads cache
3492 * shows that it makes things worse for diff-tree comparing
3493 * two linux-2.6 kernel trees in an already checked out work
3494 * tree. This is because most diff-tree comparisons deal with
3495 * only a small number of files, while reading the cache is
3496 * expensive for a large project, and its cost outweighs the
3497 * savings we get by not inflating the object to a temporary
3498 * file. Practically, this code only helps when we are used
3499 * by diff-cache --cached, which does read the cache before
3500 * calling us.
3501 */
3502 if (!active_cache)
3503 return 0;
3504 /* We want to avoid the working directory if our caller
3506 * doesn't need the data in a normal file, this system
3507 * is rather slow with its stat/open/mmap/close syscalls,
3508 * and the object is contained in a pack file. The pack
3509 * is probably already open and will be faster to obtain
3510 * the data through than the working directory. Loose
3511 * objects however would tend to be slower as they need
3512 * to be individually opened and inflated.
3513 */
3514 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(oid->hash))
3515 return 0;
3516 /*
3518 * Similarly, if we'd have to convert the file contents anyway, that
3519 * makes the optimization not worthwhile.
3520 */
3521 if (!want_file && would_convert_to_git(&the_index, name))
3522 return 0;
3523 len = strlen(name);
3525 pos = cache_name_pos(name, len);
3526 if (pos < 0)
3527 return 0;
3528 ce = active_cache[pos];
3529 /*
3531 * This is not the sha1 we are looking for, or
3532 * unreusable because it is not a regular file.
3533 */
3534 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3535 return 0;
3536 /*
3538 * If ce is marked as "assume unchanged", there is no
3539 * guarantee that work tree matches what we are looking for.
3540 */
3541 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3542 return 0;
3543 /*
3545 * If ce matches the file in the work tree, we can reuse it.
3546 */
3547 if (ce_uptodate(ce) ||
3548 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3549 return 1;
3550 return 0;
3552}
3553static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3555{
3556 struct strbuf buf = STRBUF_INIT;
3557 char *dirty = "";
3558 /* Are we looking at the work tree? */
3560 if (s->dirty_submodule)
3561 dirty = "-dirty";
3562 strbuf_addf(&buf, "Subproject commit %s%s\n",
3564 oid_to_hex(&s->oid), dirty);
3565 s->size = buf.len;
3566 if (size_only) {
3567 s->data = NULL;
3568 strbuf_release(&buf);
3569 } else {
3570 s->data = strbuf_detach(&buf, NULL);
3571 s->should_free = 1;
3572 }
3573 return 0;
3574}
3575/*
3577 * While doing rename detection and pickaxe operation, we may need to
3578 * grab the data for the blob (or file) for our own in-core comparison.
3579 * diff_filespec has data and size fields for this purpose.
3580 */
3581int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3582{
3583 int size_only = flags & CHECK_SIZE_ONLY;
3584 int err = 0;
3585 /*
3586 * demote FAIL to WARN to allow inspecting the situation
3587 * instead of refusing.
3588 */
3589 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
3590 ? SAFE_CRLF_WARN
3591 : safe_crlf);
3592 if (!DIFF_FILE_VALID(s))
3594 die("internal error: asking to populate invalid file.");
3595 if (S_ISDIR(s->mode))
3596 return -1;
3597 if (s->data)
3599 return 0;
3600 if (size_only && 0 < s->size)
3602 return 0;
3603 if (S_ISGITLINK(s->mode))
3605 return diff_populate_gitlink(s, size_only);
3606 if (!s->oid_valid ||
3608 reuse_worktree_file(s->path, &s->oid, 0)) {
3609 struct strbuf buf = STRBUF_INIT;
3610 struct stat st;
3611 int fd;
3612 if (lstat(s->path, &st) < 0) {
3614 if (errno == ENOENT) {
3615 err_empty:
3616 err = -1;
3617 empty:
3618 s->data = (char *)"";
3619 s->size = 0;
3620 return err;
3621 }
3622 }
3623 s->size = xsize_t(st.st_size);
3624 if (!s->size)
3625 goto empty;
3626 if (S_ISLNK(st.st_mode)) {
3627 struct strbuf sb = STRBUF_INIT;
3628 if (strbuf_readlink(&sb, s->path, s->size))
3630 goto err_empty;
3631 s->size = sb.len;
3632 s->data = strbuf_detach(&sb, NULL);
3633 s->should_free = 1;
3634 return 0;
3635 }
3636 /*
3638 * Even if the caller would be happy with getting
3639 * only the size, we cannot return early at this
3640 * point if the path requires us to run the content
3641 * conversion.
3642 */
3643 if (size_only && !would_convert_to_git(&the_index, s->path))
3644 return 0;
3645 /*
3647 * Note: this check uses xsize_t(st.st_size) that may
3648 * not be the true size of the blob after it goes
3649 * through convert_to_git(). This may not strictly be
3650 * correct, but the whole point of big_file_threshold
3651 * and is_binary check being that we want to avoid
3652 * opening the file and inspecting the contents, this
3653 * is probably fine.
3654 */
3655 if ((flags & CHECK_BINARY) &&
3656 s->size > big_file_threshold && s->is_binary == -1) {
3657 s->is_binary = 1;
3658 return 0;
3659 }
3660 fd = open(s->path, O_RDONLY);
3661 if (fd < 0)
3662 goto err_empty;
3663 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3664 close(fd);
3665 s->should_munmap = 1;
3666 /*
3668 * Convert from working tree format to canonical git format
3669 */
3670 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, crlf_warn)) {
3671 size_t size = 0;
3672 munmap(s->data, s->size);
3673 s->should_munmap = 0;
3674 s->data = strbuf_detach(&buf, &size);
3675 s->size = size;
3676 s->should_free = 1;
3677 }
3678 }
3679 else {
3680 enum object_type type;
3681 if (size_only || (flags & CHECK_BINARY)) {
3682 type = sha1_object_info(s->oid.hash, &s->size);
3683 if (type < 0)
3684 die("unable to read %s",
3685 oid_to_hex(&s->oid));
3686 if (size_only)
3687 return 0;
3688 if (s->size > big_file_threshold && s->is_binary == -1) {
3689 s->is_binary = 1;
3690 return 0;
3691 }
3692 }
3693 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
3694 if (!s->data)
3695 die("unable to read %s", oid_to_hex(&s->oid));
3696 s->should_free = 1;
3697 }
3698 return 0;
3699}
3700void diff_free_filespec_blob(struct diff_filespec *s)
3702{
3703 if (s->should_free)
3704 free(s->data);
3705 else if (s->should_munmap)
3706 munmap(s->data, s->size);
3707 if (s->should_free || s->should_munmap) {
3709 s->should_free = s->should_munmap = 0;
3710 s->data = NULL;
3711 }
3712}
3713void diff_free_filespec_data(struct diff_filespec *s)
3715{
3716 diff_free_filespec_blob(s);
3717 FREE_AND_NULL(s->cnt_data);
3718}
3719static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3721 void *blob,
3722 unsigned long size,
3723 const struct object_id *oid,
3724 int mode)
3725{
3726 int fd;
3727 struct strbuf buf = STRBUF_INIT;
3728 struct strbuf template = STRBUF_INIT;
3729 char *path_dup = xstrdup(path);
3730 const char *base = basename(path_dup);
3731 /* Generate "XXXXXX_basename.ext" */
3733 strbuf_addstr(&template, "XXXXXX_");
3734 strbuf_addstr(&template, base);
3735 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
3737 if (fd < 0)
3738 die_errno("unable to create temp-file");
3739 if (convert_to_working_tree(path,
3740 (const char *)blob, (size_t)size, &buf)) {
3741 blob = buf.buf;
3742 size = buf.len;
3743 }
3744 if (write_in_full(fd, blob, size) != size)
3745 die_errno("unable to write temp-file");
3746 close_tempfile(&temp->tempfile);
3747 temp->name = get_tempfile_path(&temp->tempfile);
3748 oid_to_hex_r(temp->hex, oid);
3749 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3750 strbuf_release(&buf);
3751 strbuf_release(&template);
3752 free(path_dup);
3753}
3754static struct diff_tempfile *prepare_temp_file(const char *name,
3756 struct diff_filespec *one)
3757{
3758 struct diff_tempfile *temp = claim_diff_tempfile();
3759 if (!DIFF_FILE_VALID(one)) {
3761 not_a_valid_file:
3762 /* A '-' entry produces this for file-2, and
3763 * a '+' entry produces this for file-1.
3764 */
3765 temp->name = "/dev/null";
3766 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3767 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3768 return temp;
3769 }
3770 if (!S_ISGITLINK(one->mode) &&
3772 (!one->oid_valid ||
3773 reuse_worktree_file(name, &one->oid, 1))) {
3774 struct stat st;
3775 if (lstat(name, &st) < 0) {
3776 if (errno == ENOENT)
3777 goto not_a_valid_file;
3778 die_errno("stat(%s)", name);
3779 }
3780 if (S_ISLNK(st.st_mode)) {
3781 struct strbuf sb = STRBUF_INIT;
3782 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3783 die_errno("readlink(%s)", name);
3784 prep_temp_blob(name, temp, sb.buf, sb.len,
3785 (one->oid_valid ?
3786 &one->oid : &null_oid),
3787 (one->oid_valid ?
3788 one->mode : S_IFLNK));
3789 strbuf_release(&sb);
3790 }
3791 else {
3792 /* we can borrow from the file in the work tree */
3793 temp->name = name;
3794 if (!one->oid_valid)
3795 oid_to_hex_r(temp->hex, &null_oid);
3796 else
3797 oid_to_hex_r(temp->hex, &one->oid);
3798 /* Even though we may sometimes borrow the
3799 * contents from the work tree, we always want
3800 * one->mode. mode is trustworthy even when
3801 * !(one->oid_valid), as long as
3802 * DIFF_FILE_VALID(one).
3803 */
3804 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3805 }
3806 return temp;
3807 }
3808 else {
3809 if (diff_populate_filespec(one, 0))
3810 die("cannot read data blob for %s", one->path);
3811 prep_temp_blob(name, temp, one->data, one->size,
3812 &one->oid, one->mode);
3813 }
3814 return temp;
3815}
3816static void add_external_diff_name(struct argv_array *argv,
3818 const char *name,
3819 struct diff_filespec *df)
3820{
3821 struct diff_tempfile *temp = prepare_temp_file(name, df);
3822 argv_array_push(argv, temp->name);
3823 argv_array_push(argv, temp->hex);
3824 argv_array_push(argv, temp->mode);
3825}
3826/* An external diff command takes:
3828 *
3829 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3830 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3831 *
3832 */
3833static void run_external_diff(const char *pgm,
3834 const char *name,
3835 const char *other,
3836 struct diff_filespec *one,
3837 struct diff_filespec *two,
3838 const char *xfrm_msg,
3839 int complete_rewrite,
3840 struct diff_options *o)
3841{
3842 struct argv_array argv = ARGV_ARRAY_INIT;
3843 struct argv_array env = ARGV_ARRAY_INIT;
3844 struct diff_queue_struct *q = &diff_queued_diff;
3845 argv_array_push(&argv, pgm);
3847 argv_array_push(&argv, name);
3848 if (one && two) {
3850 add_external_diff_name(&argv, name, one);
3851 if (!other)
3852 add_external_diff_name(&argv, name, two);
3853 else {
3854 add_external_diff_name(&argv, other, two);
3855 argv_array_push(&argv, other);
3856 argv_array_push(&argv, xfrm_msg);
3857 }
3858 }
3859 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3861 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3862 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3864 die(_("external diff died, stopping at %s"), name);
3865 remove_tempfile();
3867 argv_array_clear(&argv);
3868 argv_array_clear(&env);
3869}
3870static int similarity_index(struct diff_filepair *p)
3872{
3873 return p->score * 100 / MAX_SCORE;
3874}
3875static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3877{
3878 if (startup_info->have_repository)
3879 return find_unique_abbrev(oid->hash, abbrev);
3880 else {
3881 char *hex = oid_to_hex(oid);
3882 if (abbrev < 0)
3883 abbrev = FALLBACK_DEFAULT_ABBREV;
3884 if (abbrev > GIT_SHA1_HEXSZ)
3885 die("BUG: oid abbreviation out of range: %d", abbrev);
3886 if (abbrev)
3887 hex[abbrev] = '\0';
3888 return hex;
3889 }
3890}
3891static void fill_metainfo(struct strbuf *msg,
3893 const char *name,
3894 const char *other,
3895 struct diff_filespec *one,
3896 struct diff_filespec *two,
3897 struct diff_options *o,
3898 struct diff_filepair *p,
3899 int *must_show_header,
3900 int use_color)
3901{
3902 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3903 const char *reset = diff_get_color(use_color, DIFF_RESET);
3904 const char *line_prefix = diff_line_prefix(o);
3905 *must_show_header = 1;
3907 strbuf_init(msg, PATH_MAX * 2 + 300);
3908 switch (p->status) {
3909 case DIFF_STATUS_COPIED:
3910 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3911 line_prefix, set, similarity_index(p));
3912 strbuf_addf(msg, "%s\n%s%scopy from ",
3913 reset, line_prefix, set);
3914 quote_c_style(name, msg, NULL, 0);
3915 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3916 quote_c_style(other, msg, NULL, 0);
3917 strbuf_addf(msg, "%s\n", reset);
3918 break;
3919 case DIFF_STATUS_RENAMED:
3920 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3921 line_prefix, set, similarity_index(p));
3922 strbuf_addf(msg, "%s\n%s%srename from ",
3923 reset, line_prefix, set);
3924 quote_c_style(name, msg, NULL, 0);
3925 strbuf_addf(msg, "%s\n%s%srename to ",
3926 reset, line_prefix, set);
3927 quote_c_style(other, msg, NULL, 0);
3928 strbuf_addf(msg, "%s\n", reset);
3929 break;
3930 case DIFF_STATUS_MODIFIED:
3931 if (p->score) {
3932 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3933 line_prefix,
3934 set, similarity_index(p), reset);
3935 break;
3936 }
3937 /* fallthru */
3938 default:
3939 *must_show_header = 0;
3940 }
3941 if (one && two && oidcmp(&one->oid, &two->oid)) {
3942 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3943 if (DIFF_OPT_TST(o, BINARY)) {
3945 mmfile_t mf;
3946 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3947 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3948 abbrev = 40;
3949 }
3950 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3951 diff_abbrev_oid(&one->oid, abbrev),
3952 diff_abbrev_oid(&two->oid, abbrev));
3953 if (one->mode == two->mode)
3954 strbuf_addf(msg, " %06o", one->mode);
3955 strbuf_addf(msg, "%s\n", reset);
3956 }
3957}
3958static void run_diff_cmd(const char *pgm,
3960 const char *name,
3961 const char *other,
3962 const char *attr_path,
3963 struct diff_filespec *one,
3964 struct diff_filespec *two,
3965 struct strbuf *msg,
3966 struct diff_options *o,
3967 struct diff_filepair *p)
3968{
3969 const char *xfrm_msg = NULL;
3970 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3971 int must_show_header = 0;
3972 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3975 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3976 if (drv && drv->external)
3977 pgm = drv->external;
3978 }
3979 if (msg) {
3981 /*
3982 * don't use colors when the header is intended for an
3983 * external diff driver
3984 */
3985 fill_metainfo(msg, name, other, one, two, o, p,
3986 &must_show_header,
3987 want_color(o->use_color) && !pgm);
3988 xfrm_msg = msg->len ? msg->buf : NULL;
3989 }
3990 if (pgm) {
3992 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3993 complete_rewrite, o);
3994 return;
3995 }
3996 if (one && two)
3997 builtin_diff(name, other ? other : name,
3998 one, two, xfrm_msg, must_show_header,
3999 o, complete_rewrite);
4000 else
4001 fprintf(o->file, "* Unmerged path %s\n", name);
4002}
4003static void diff_fill_oid_info(struct diff_filespec *one)
4005{
4006 if (DIFF_FILE_VALID(one)) {
4007 if (!one->oid_valid) {
4008 struct stat st;
4009 if (one->is_stdin) {
4010 oidclr(&one->oid);
4011 return;
4012 }
4013 if (lstat(one->path, &st) < 0)
4014 die_errno("stat '%s'", one->path);
4015 if (index_path(&one->oid, one->path, &st, 0))
4016 die("cannot hash %s", one->path);
4017 }
4018 }
4019 else
4020 oidclr(&one->oid);
4021}
4022static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
4024{
4025 /* Strip the prefix but do not molest /dev/null and absolute paths */
4026 if (*namep && **namep != '/') {
4027 *namep += prefix_length;
4028 if (**namep == '/')
4029 ++*namep;
4030 }
4031 if (*otherp && **otherp != '/') {
4032 *otherp += prefix_length;
4033 if (**otherp == '/')
4034 ++*otherp;
4035 }
4036}
4037static void run_diff(struct diff_filepair *p, struct diff_options *o)
4039{
4040 const char *pgm = external_diff();
4041 struct strbuf msg;
4042 struct diff_filespec *one = p->one;
4043 struct diff_filespec *two = p->two;
4044 const char *name;
4045 const char *other;
4046 const char *attr_path;
4047 name = one->path;
4049 other = (strcmp(name, two->path) ? two->path : NULL);
4050 attr_path = name;
4051 if (o->prefix_length)
4052 strip_prefix(o->prefix_length, &name, &other);
4053 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
4055 pgm = NULL;
4056 if (DIFF_PAIR_UNMERGED(p)) {
4058 run_diff_cmd(pgm, name, NULL, attr_path,
4059 NULL, NULL, NULL, o, p);
4060 return;
4061 }
4062 diff_fill_oid_info(one);
4064 diff_fill_oid_info(two);
4065 if (!pgm &&
4067 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4068 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4069 /*
4070 * a filepair that changes between file and symlink
4071 * needs to be split into deletion and creation.
4072 */
4073 struct diff_filespec *null = alloc_filespec(two->path);
4074 run_diff_cmd(NULL, name, other, attr_path,
4075 one, null, &msg, o, p);
4076 free(null);
4077 strbuf_release(&msg);
4078 null = alloc_filespec(one->path);
4080 run_diff_cmd(NULL, name, other, attr_path,
4081 null, two, &msg, o, p);
4082 free(null);
4083 }
4084 else
4085 run_diff_cmd(pgm, name, other, attr_path,
4086 one, two, &msg, o, p);
4087 strbuf_release(&msg);
4089}
4090static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4092 struct diffstat_t *diffstat)
4093{
4094 const char *name;
4095 const char *other;
4096 if (DIFF_PAIR_UNMERGED(p)) {
4098 /* unmerged */
4099 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4100 return;
4101 }
4102 name = p->one->path;
4104 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4105 if (o->prefix_length)
4107 strip_prefix(o->prefix_length, &name, &other);
4108 diff_fill_oid_info(p->one);
4110 diff_fill_oid_info(p->two);
4111 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4113}
4114static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4116{
4117 const char *name;
4118 const char *other;
4119 const char *attr_path;
4120 if (DIFF_PAIR_UNMERGED(p)) {
4122 /* unmerged */
4123 return;
4124 }
4125 name = p->one->path;
4127 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4128 attr_path = other ? other : name;
4129 if (o->prefix_length)
4131 strip_prefix(o->prefix_length, &name, &other);
4132 diff_fill_oid_info(p->one);
4134 diff_fill_oid_info(p->two);
4135 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4137}
4138void diff_setup(struct diff_options *options)
4140{
4141 memcpy(options, &default_diff_options, sizeof(*options));
4142 options->file = stdout;
4144 options->abbrev = DEFAULT_ABBREV;
4146 options->line_termination = '\n';
4147 options->break_opt = -1;
4148 options->rename_limit = -1;
4149 options->dirstat_permille = diff_dirstat_permille_default;
4150 options->context = diff_context_default;
4151 options->interhunkcontext = diff_interhunk_context_default;
4152 options->ws_error_highlight = ws_error_highlight_default;
4153 DIFF_OPT_SET(options, RENAME_EMPTY);
4154 /* pathchange left =NULL by default */
4156 options->change = diff_change;
4157 options->add_remove = diff_addremove;
4158 options->use_color = diff_use_color_default;
4159 options->detect_rename = diff_detect_rename_default;
4160 options->xdl_opts |= diff_algorithm;
4161 if (diff_indent_heuristic)
4162 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4163 options->orderfile = diff_order_file_cfg;
4165 if (diff_no_prefix) {
4167 options->a_prefix = options->b_prefix = "";
4168 } else if (!diff_mnemonic_prefix) {
4169 options->a_prefix = "a/";
4170 options->b_prefix = "b/";
4171 }
4172 options->color_moved = diff_color_moved_default;
4174}
4175void diff_setup_done(struct diff_options *options)
4177{
4178 int count = 0;
4179 if (options->set_default)
4181 options->set_default(options);
4182 if (options->output_format & DIFF_FORMAT_NAME)
4184 count++;
4185 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
4186 count++;
4187 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
4188 count++;
4189 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
4190 count++;
4191 if (count > 1)
4192 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4193 /*
4195 * Most of the time we can say "there are changes"
4196 * only by checking if there are changed paths, but
4197 * --ignore-whitespace* options force us to look
4198 * inside contents.
4199 */
4200 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
4202 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
4203 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
4204 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
4205 else
4206 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
4207 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
4209 options->detect_rename = DIFF_DETECT_COPY;
4210 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
4212 options->prefix = NULL;
4213 if (options->prefix)
4214 options->prefix_length = strlen(options->prefix);
4215 else
4216 options->prefix_length = 0;
4217 if (options->output_format & (DIFF_FORMAT_NAME |
4219 DIFF_FORMAT_NAME_STATUS |
4220 DIFF_FORMAT_CHECKDIFF |
4221 DIFF_FORMAT_NO_OUTPUT))
4222 options->output_format &= ~(DIFF_FORMAT_RAW |
4223 DIFF_FORMAT_NUMSTAT |
4224 DIFF_FORMAT_DIFFSTAT |
4225 DIFF_FORMAT_SHORTSTAT |
4226 DIFF_FORMAT_DIRSTAT |
4227 DIFF_FORMAT_SUMMARY |
4228 DIFF_FORMAT_PATCH);
4229 /*
4231 * These cases always need recursive; we do not drop caller-supplied
4232 * recursive bits for other formats here.
4233 */
4234 if (options->output_format & (DIFF_FORMAT_PATCH |
4235 DIFF_FORMAT_NUMSTAT |
4236 DIFF_FORMAT_DIFFSTAT |
4237 DIFF_FORMAT_SHORTSTAT |
4238 DIFF_FORMAT_DIRSTAT |
4239 DIFF_FORMAT_SUMMARY |
4240 DIFF_FORMAT_CHECKDIFF))
4241 DIFF_OPT_SET(options, RECURSIVE);
4242 /*
4243 * Also pickaxe would not work very well if you do not say recursive
4244 */
4245 if (options->pickaxe)
4246 DIFF_OPT_SET(options, RECURSIVE);
4247 /*
4248 * When patches are generated, submodules diffed against the work tree
4249 * must be checked for dirtiness too so it can be shown in the output
4250 */
4251 if (options->output_format & DIFF_FORMAT_PATCH)
4252 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
4253 if (options->detect_rename && options->rename_limit < 0)
4255 options->rename_limit = diff_rename_limit_default;
4256 if (options->setup & DIFF_SETUP_USE_CACHE) {
4257 if (!active_cache)
4258 /* read-cache does not die even when it fails
4259 * so it is safe for us to do this here. Also
4260 * it does not smudge active_cache or active_nr
4261 * when it fails, so we do not have to worry about
4262 * cleaning it up ourselves either.
4263 */
4264 read_cache();
4265 }
4266 if (40 < options->abbrev)
4267 options->abbrev = 40; /* full */
4268 /*
4270 * It does not make sense to show the first hit we happened
4271 * to have found. It does not make sense not to return with
4272 * exit code in such a case either.
4273 */
4274 if (DIFF_OPT_TST(options, QUICK)) {
4275 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4276 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4277 }
4278 options->diff_path_counter = 0;
4280 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
4282 die(_("--follow requires exactly one pathspec"));
4283 if (!options->use_color || external_diff())
4285 options->color_moved = 0;
4286}
4287static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4289{
4290 char c, *eq;
4291 int len;
4292 if (*arg != '-')
4294 return 0;
4295 c = *++arg;
4296 if (!c)
4297 return 0;
4298 if (c == arg_short) {
4299 c = *++arg;
4300 if (!c)
4301 return 1;
4302 if (val && isdigit(c)) {
4303 char *end;
4304 int n = strtoul(arg, &end, 10);
4305 if (*end)
4306 return 0;
4307 *val = n;
4308 return 1;
4309 }
4310 return 0;
4311 }
4312 if (c != '-')
4313 return 0;
4314 arg++;
4315 eq = strchrnul(arg, '=');
4316 len = eq - arg;
4317 if (!len || strncmp(arg, arg_long, len))
4318 return 0;
4319 if (*eq) {
4320 int n;
4321 char *end;
4322 if (!isdigit(*++eq))
4323 return 0;
4324 n = strtoul(eq, &end, 10);
4325 if (*end)
4326 return 0;
4327 *val = n;
4328 }
4329 return 1;
4330}
4331static int diff_scoreopt_parse(const char *opt);
4333static inline int short_opt(char opt, const char **argv,
4335 const char **optarg)
4336{
4337 const char *arg = argv[0];
4338 if (arg[0] != '-' || arg[1] != opt)
4339 return 0;
4340 if (arg[2] != '\0') {
4341 *optarg = arg + 2;
4342 return 1;
4343 }
4344 if (!argv[1])
4345 die("Option '%c' requires a value", opt);
4346 *optarg = argv[1];
4347 return 2;
4348}
4349int parse_long_opt(const char *opt, const char **argv,
4351 const char **optarg)
4352{
4353 const char *arg = argv[0];
4354 if (!skip_prefix(arg, "--", &arg))
4355 return 0;
4356 if (!skip_prefix(arg, opt, &arg))
4357 return 0;
4358 if (*arg == '=') { /* stuck form: --option=value */
4359 *optarg = arg + 1;
4360 return 1;
4361 }
4362 if (*arg != '\0')
4363 return 0;
4364 /* separate form: --option value */
4365 if (!argv[1])
4366 die("Option '--%s' requires a value", opt);
4367 *optarg = argv[1];
4368 return 2;
4369}
4370static int stat_opt(struct diff_options *options, const char **av)
4372{
4373 const char *arg = av[0];
4374 char *end;
4375 int width = options->stat_width;
4376 int name_width = options->stat_name_width;
4377 int graph_width = options->stat_graph_width;
4378 int count = options->stat_count;
4379 int argcount = 1;
4380 if (!skip_prefix(arg, "--stat", &arg))
4382 die("BUG: stat option does not begin with --stat: %s", arg);
4383 end = (char *)arg;
4384 switch (*arg) {
4386 case '-':
4387 if (skip_prefix(arg, "-width", &arg)) {
4388 if (*arg == '=')
4389 width = strtoul(arg + 1, &end, 10);
4390 else if (!*arg && !av[1])
4391 die_want_option("--stat-width");
4392 else if (!*arg) {
4393 width = strtoul(av[1], &end, 10);
4394 argcount = 2;
4395 }
4396 } else if (skip_prefix(arg, "-name-width", &arg)) {
4397 if (*arg == '=')
4398 name_width = strtoul(arg + 1, &end, 10);
4399 else if (!*arg && !av[1])
4400 die_want_option("--stat-name-width");
4401 else if (!*arg) {
4402 name_width = strtoul(av[1], &end, 10);
4403 argcount = 2;
4404 }
4405 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4406 if (*arg == '=')
4407 graph_width = strtoul(arg + 1, &end, 10);
4408 else if (!*arg && !av[1])
4409 die_want_option("--stat-graph-width");
4410 else if (!*arg) {
4411 graph_width = strtoul(av[1], &end, 10);
4412 argcount = 2;
4413 }
4414 } else if (skip_prefix(arg, "-count", &arg)) {
4415 if (*arg == '=')
4416 count = strtoul(arg + 1, &end, 10);
4417 else if (!*arg && !av[1])
4418 die_want_option("--stat-count");
4419 else if (!*arg) {
4420 count = strtoul(av[1], &end, 10);
4421 argcount = 2;
4422 }
4423 }
4424 break;
4425 case '=':
4426 width = strtoul(arg+1, &end, 10);
4427 if (*end == ',')
4428 name_width = strtoul(end+1, &end, 10);
4429 if (*end == ',')
4430 count = strtoul(end+1, &end, 10);
4431 }
4432 /* Important! This checks all the error cases! */
4434 if (*end)
4435 return 0;
4436 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4437 options->stat_name_width = name_width;
4438 options->stat_graph_width = graph_width;
4439 options->stat_width = width;
4440 options->stat_count = count;
4441 return argcount;
4442}
4443static int parse_dirstat_opt(struct diff_options *options, const char *params)
4445{
4446 struct strbuf errmsg = STRBUF_INIT;
4447 if (parse_dirstat_params(options, params, &errmsg))
4448 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4449 errmsg.buf);
4450 strbuf_release(&errmsg);
4451 /*
4452 * The caller knows a dirstat-related option is given from the command
4453 * line; allow it to say "return this_function();"
4454 */
4455 options->output_format |= DIFF_FORMAT_DIRSTAT;
4456 return 1;
4457}
4458static int parse_submodule_opt(struct diff_options *options, const char *value)
4460{
4461 if (parse_submodule_params(options, value))
4462 die(_("Failed to parse --submodule option parameter: '%s'"),
4463 value);
4464 return 1;
4465}
4466static const char diff_status_letters[] = {
4468 DIFF_STATUS_ADDED,
4469 DIFF_STATUS_COPIED,
4470 DIFF_STATUS_DELETED,
4471 DIFF_STATUS_MODIFIED,
4472 DIFF_STATUS_RENAMED,
4473 DIFF_STATUS_TYPE_CHANGED,
4474 DIFF_STATUS_UNKNOWN,
4475 DIFF_STATUS_UNMERGED,
4476 DIFF_STATUS_FILTER_AON,
4477 DIFF_STATUS_FILTER_BROKEN,
4478 '\0',
4479};
4480static unsigned int filter_bit['Z' + 1];
4482static void prepare_filter_bits(void)
4484{
4485 int i;
4486 if (!filter_bit[DIFF_STATUS_ADDED]) {
4488 for (i = 0; diff_status_letters[i]; i++)
4489 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4490 }
4491}
4492static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4494{
4495 return opt->filter & filter_bit[(int) status];
4496}
4497static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4499{
4500 int i, optch;
4501 prepare_filter_bits();
4503 /*
4505 * If there is a negation e.g. 'd' in the input, and we haven't
4506 * initialized the filter field with another --diff-filter, start
4507 * from full set of bits, except for AON.
4508 */
4509 if (!opt->filter) {
4510 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4511 if (optch < 'a' || 'z' < optch)
4512 continue;
4513 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4514 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4515 break;
4516 }
4517 }
4518 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4520 unsigned int bit;
4521 int negate;
4522 if ('a' <= optch && optch <= 'z') {
4524 negate = 1;
4525 optch = toupper(optch);
4526 } else {
4527 negate = 0;
4528 }
4529 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4531 if (!bit)
4532 return optarg[i];
4533 if (negate)
4534 opt->filter &= ~bit;
4535 else
4536 opt->filter |= bit;
4537 }
4538 return 0;
4539}
4540static void enable_patch_output(int *fmt) {
4542 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4543 *fmt |= DIFF_FORMAT_PATCH;
4544}
4545static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4547{
4548 int val = parse_ws_error_highlight(arg);
4549 if (val < 0) {
4551 error("unknown value after ws-error-highlight=%.*s",
4552 -1 - val, arg);
4553 return 0;
4554 }
4555 opt->ws_error_highlight = val;
4556 return 1;
4557}
4558int diff_opt_parse(struct diff_options *options,
4560 const char **av, int ac, const char *prefix)
4561{
4562 const char *arg = av[0];
4563 const char *optarg;
4564 int argcount;
4565 if (!prefix)
4567 prefix = "";
4568 /* Output format options */
4570 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4571 || opt_arg(arg, 'U', "unified", &options->context))
4572 enable_patch_output(&options->output_format);
4573 else if (!strcmp(arg, "--raw"))
4574 options->output_format |= DIFF_FORMAT_RAW;
4575 else if (!strcmp(arg, "--patch-with-raw")) {
4576 enable_patch_output(&options->output_format);
4577 options->output_format |= DIFF_FORMAT_RAW;
4578 } else if (!strcmp(arg, "--numstat"))
4579 options->output_format |= DIFF_FORMAT_NUMSTAT;
4580 else if (!strcmp(arg, "--shortstat"))
4581 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4582 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
4583 return parse_dirstat_opt(options, "");
4584 else if (skip_prefix(arg, "-X", &arg))
4585 return parse_dirstat_opt(options, arg);
4586 else if (skip_prefix(arg, "--dirstat=", &arg))
4587 return parse_dirstat_opt(options, arg);
4588 else if (!strcmp(arg, "--cumulative"))
4589 return parse_dirstat_opt(options, "cumulative");
4590 else if (!strcmp(arg, "--dirstat-by-file"))
4591 return parse_dirstat_opt(options, "files");
4592 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
4593 parse_dirstat_opt(options, "files");
4594 return parse_dirstat_opt(options, arg);
4595 }
4596 else if (!strcmp(arg, "--check"))
4597 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4598 else if (!strcmp(arg, "--summary"))
4599 options->output_format |= DIFF_FORMAT_SUMMARY;
4600 else if (!strcmp(arg, "--patch-with-stat")) {
4601 enable_patch_output(&options->output_format);
4602 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4603 } else if (!strcmp(arg, "--name-only"))
4604 options->output_format |= DIFF_FORMAT_NAME;
4605 else if (!strcmp(arg, "--name-status"))
4606 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4607 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4608 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4609 else if (starts_with(arg, "--stat"))
4610 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4611 return stat_opt(options, av);
4612 /* renames options */
4614 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
4615 !strcmp(arg, "--break-rewrites")) {
4616 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4617 return error("invalid argument to -B: %s", arg+2);
4618 }
4619 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
4620 !strcmp(arg, "--find-renames")) {
4621 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4622 return error("invalid argument to -M: %s", arg+2);
4623 options->detect_rename = DIFF_DETECT_RENAME;
4624 }
4625 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4626 options->irreversible_delete = 1;
4627 }
4628 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
4629 !strcmp(arg, "--find-copies")) {
4630 if (options->detect_rename == DIFF_DETECT_COPY)
4631 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4632 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4633 return error("invalid argument to -C: %s", arg+2);
4634 options->detect_rename = DIFF_DETECT_COPY;
4635 }
4636 else if (!strcmp(arg, "--no-renames"))
4637 options->detect_rename = 0;
4638 else if (!strcmp(arg, "--rename-empty"))
4639 DIFF_OPT_SET(options, RENAME_EMPTY);
4640 else if (!strcmp(arg, "--no-rename-empty"))
4641 DIFF_OPT_CLR(options, RENAME_EMPTY);
4642 else if (!strcmp(arg, "--relative"))
4643 DIFF_OPT_SET(options, RELATIVE_NAME);
4644 else if (skip_prefix(arg, "--relative=", &arg)) {
4645 DIFF_OPT_SET(options, RELATIVE_NAME);
4646 options->prefix = arg;
4647 }
4648 /* xdiff options */
4650 else if (!strcmp(arg, "--minimal"))
4651 DIFF_XDL_SET(options, NEED_MINIMAL);
4652 else if (!strcmp(arg, "--no-minimal"))
4653 DIFF_XDL_CLR(options, NEED_MINIMAL);
4654 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4655 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4656 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4657 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4658 else if (!strcmp(arg, "--ignore-space-at-eol"))
4659 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4660 else if (!strcmp(arg, "--ignore-blank-lines"))
4661 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4662 else if (!strcmp(arg, "--indent-heuristic"))
4663 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4664 else if (!strcmp(arg, "--no-indent-heuristic"))
4665 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4666 else if (!strcmp(arg, "--patience"))
4667 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4668 else if (!strcmp(arg, "--histogram"))
4669 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4670 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4671 long value = parse_algorithm_value(optarg);
4672 if (value < 0)
4673 return error("option diff-algorithm accepts \"myers\", "
4674 "\"minimal\", \"patience\" and \"histogram\"");
4675 /* clear out previous settings */
4676 DIFF_XDL_CLR(options, NEED_MINIMAL);
4677 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4678 options->xdl_opts |= value;
4679 return argcount;
4680 }
4681 /* flags options */
4683 else if (!strcmp(arg, "--binary")) {
4684 enable_patch_output(&options->output_format);
4685 DIFF_OPT_SET(options, BINARY);
4686 }
4687 else if (!strcmp(arg, "--full-index"))
4688 DIFF_OPT_SET(options, FULL_INDEX);
4689 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4690 DIFF_OPT_SET(options, TEXT);
4691 else if (!strcmp(arg, "-R"))
4692 DIFF_OPT_SET(options, REVERSE_DIFF);
4693 else if (!strcmp(arg, "--find-copies-harder"))
4694 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4695 else if (!strcmp(arg, "--follow"))
4696 DIFF_OPT_SET(options, FOLLOW_RENAMES);
4697 else if (!strcmp(arg, "--no-follow")) {
4698 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
4699 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
4700 } else if (!strcmp(arg, "--color"))
4701 options->use_color = 1;
4702 else if (skip_prefix(arg, "--color=", &arg)) {
4703 int value = git_config_colorbool(NULL, arg);
4704 if (value < 0)
4705 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4706 options->use_color = value;
4707 }
4708 else if (!strcmp(arg, "--no-color"))
4709 options->use_color = 0;
4710 else if (!strcmp(arg, "--color-moved")) {
4711 if (diff_color_moved_default)
4712 options->color_moved = diff_color_moved_default;
4713 if (options->color_moved == COLOR_MOVED_NO)
4714 options->color_moved = COLOR_MOVED_DEFAULT;
4715 } else if (!strcmp(arg, "--no-color-moved"))
4716 options->color_moved = COLOR_MOVED_NO;
4717 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4718 int cm = parse_color_moved(arg);
4719 if (cm < 0)
4720 die("bad --color-moved argument: %s", arg);
4721 options->color_moved = cm;
4722 } else if (!strcmp(arg, "--color-words")) {
4723 options->use_color = 1;
4724 options->word_diff = DIFF_WORDS_COLOR;
4725 }
4726 else if (skip_prefix(arg, "--color-words=", &arg)) {
4727 options->use_color = 1;
4728 options->word_diff = DIFF_WORDS_COLOR;
4729 options->word_regex = arg;
4730 }
4731 else if (!strcmp(arg, "--word-diff")) {
4732 if (options->word_diff == DIFF_WORDS_NONE)
4733 options->word_diff = DIFF_WORDS_PLAIN;
4734 }
4735 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4736 if (!strcmp(arg, "plain"))
4737 options->word_diff = DIFF_WORDS_PLAIN;
4738 else if (!strcmp(arg, "color")) {
4739 options->use_color = 1;
4740 options->word_diff = DIFF_WORDS_COLOR;
4741 }
4742 else if (!strcmp(arg, "porcelain"))
4743 options->word_diff = DIFF_WORDS_PORCELAIN;
4744 else if (!strcmp(arg, "none"))
4745 options->word_diff = DIFF_WORDS_NONE;
4746 else
4747 die("bad --word-diff argument: %s", arg);
4748 }
4749 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
4750 if (options->word_diff == DIFF_WORDS_NONE)
4751 options->word_diff = DIFF_WORDS_PLAIN;
4752 options->word_regex = optarg;
4753 return argcount;
4754 }
4755 else if (!strcmp(arg, "--exit-code"))
4756 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4757 else if (!strcmp(arg, "--quiet"))
4758 DIFF_OPT_SET(options, QUICK);
4759 else if (!strcmp(arg, "--ext-diff"))
4760 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
4761 else if (!strcmp(arg, "--no-ext-diff"))
4762 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
4763 else if (!strcmp(arg, "--textconv"))
4764 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
4765 else if (!strcmp(arg, "--no-textconv"))
4766 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
4767 else if (!strcmp(arg, "--ignore-submodules")) {
4768 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4769 handle_ignore_submodules_arg(options, "all");
4770 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
4771 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4772 handle_ignore_submodules_arg(options, arg);
4773 } else if (!strcmp(arg, "--submodule"))
4774 options->submodule_format = DIFF_SUBMODULE_LOG;
4775 else if (skip_prefix(arg, "--submodule=", &arg))
4776 return parse_submodule_opt(options, arg);
4777 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
4778 return parse_ws_error_highlight_opt(options, arg);
4779 else if (!strcmp(arg, "--ita-invisible-in-index"))
4780 options->ita_invisible_in_index = 1;
4781 else if (!strcmp(arg, "--ita-visible-in-index"))
4782 options->ita_invisible_in_index = 0;
4783 /* misc options */
4785 else if (!strcmp(arg, "-z"))
4786 options->line_termination = 0;
4787 else if ((argcount = short_opt('l', av, &optarg))) {
4788 options->rename_limit = strtoul(optarg, NULL, 10);
4789 return argcount;
4790 }
4791 else if ((argcount = short_opt('S', av, &optarg))) {
4792 options->pickaxe = optarg;
4793 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4794 return argcount;
4795 } else if ((argcount = short_opt('G', av, &optarg))) {
4796 options->pickaxe = optarg;
4797 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4798 return argcount;
4799 }
4800 else if (!strcmp(arg, "--pickaxe-all"))
4801 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4802 else if (!strcmp(arg, "--pickaxe-regex"))
4803 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4804 else if ((argcount = short_opt('O', av, &optarg))) {
4805 options->orderfile = prefix_filename(prefix, optarg);
4806 return argcount;
4807 }
4808 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4809 int offending = parse_diff_filter_opt(optarg, options);
4810 if (offending)
4811 die("unknown change class '%c' in --diff-filter=%s",
4812 offending, optarg);
4813 return argcount;
4814 }
4815 else if (!strcmp(arg, "--no-abbrev"))
4816 options->abbrev = 0;
4817 else if (!strcmp(arg, "--abbrev"))
4818 options->abbrev = DEFAULT_ABBREV;
4819 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4820 options->abbrev = strtoul(arg, NULL, 10);
4821 if (options->abbrev < MINIMUM_ABBREV)
4822 options->abbrev = MINIMUM_ABBREV;
4823 else if (40 < options->abbrev)
4824 options->abbrev = 40;
4825 }
4826 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4827 options->a_prefix = optarg;
4828 return argcount;
4829 }
4830 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4831 options->line_prefix = optarg;
4832 options->line_prefix_length = strlen(options->line_prefix);
4833 graph_setup_line_prefix(options);
4834 return argcount;
4835 }
4836 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4837 options->b_prefix = optarg;
4838 return argcount;
4839 }
4840 else if (!strcmp(arg, "--no-prefix"))
4841 options->a_prefix = options->b_prefix = "";
4842 else if (opt_arg(arg, '\0', "inter-hunk-context",
4843 &options->interhunkcontext))
4844 ;
4845 else if (!strcmp(arg, "-W"))
4846 DIFF_OPT_SET(options, FUNCCONTEXT);
4847 else if (!strcmp(arg, "--function-context"))
4848 DIFF_OPT_SET(options, FUNCCONTEXT);
4849 else if (!strcmp(arg, "--no-function-context"))
4850 DIFF_OPT_CLR(options, FUNCCONTEXT);
4851 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4852 char *path = prefix_filename(prefix, optarg);
4853 options->file = xfopen(path, "w");
4854 options->close_file = 1;
4855 if (options->use_color != GIT_COLOR_ALWAYS)
4856 options->use_color = GIT_COLOR_NEVER;
4857 free(path);
4858 return argcount;
4859 } else
4860 return 0;
4861 return 1;
4862}
4863int parse_rename_score(const char **cp_p)
4865{
4866 unsigned long num, scale;
4867 int ch, dot;
4868 const char *cp = *cp_p;
4869 num = 0;
4871 scale = 1;
4872 dot = 0;
4873 for (;;) {
4874 ch = *cp;
4875 if ( !dot && ch == '.' ) {
4876 scale = 1;
4877 dot = 1;
4878 } else if ( ch == '%' ) {
4879 scale = dot ? scale*100 : 100;
4880 cp++; /* % is always at the end */
4881 break;
4882 } else if ( ch >= '0' && ch <= '9' ) {
4883 if ( scale < 100000 ) {
4884 scale *= 10;
4885 num = (num*10) + (ch-'0');
4886 }
4887 } else {
4888 break;
4889 }
4890 cp++;
4891 }
4892 *cp_p = cp;
4893 /* user says num divided by scale and we say internally that
4895 * is MAX_SCORE * num / scale.
4896 */
4897 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4898}
4899static int diff_scoreopt_parse(const char *opt)
4901{
4902 int opt1, opt2, cmd;
4903 if (*opt++ != '-')
4905 return -1;
4906 cmd = *opt++;
4907 if (cmd == '-') {
4908 /* convert the long-form arguments into short-form versions */
4909 if (skip_prefix(opt, "break-rewrites", &opt)) {
4910 if (*opt == 0 || *opt++ == '=')
4911 cmd = 'B';
4912 } else if (skip_prefix(opt, "find-copies", &opt)) {
4913 if (*opt == 0 || *opt++ == '=')
4914 cmd = 'C';
4915 } else if (skip_prefix(opt, "find-renames", &opt)) {
4916 if (*opt == 0 || *opt++ == '=')
4917 cmd = 'M';
4918 }
4919 }
4920 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4921 return -1; /* that is not a -M, -C, or -B option */
4922 opt1 = parse_rename_score(&opt);
4924 if (cmd != 'B')
4925 opt2 = 0;
4926 else {
4927 if (*opt == 0)
4928 opt2 = 0;
4929 else if (*opt != '/')
4930 return -1; /* we expect -B80/99 or -B80 */
4931 else {
4932 opt++;
4933 opt2 = parse_rename_score(&opt);
4934 }
4935 }
4936 if (*opt != 0)
4937 return -1;
4938 return opt1 | (opt2 << 16);
4939}
4940struct diff_queue_struct diff_queued_diff;
4942void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4944{
4945 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4946 queue->queue[queue->nr++] = dp;
4947}
4948struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4950 struct diff_filespec *one,
4951 struct diff_filespec *two)
4952{
4953 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4954 dp->one = one;
4955 dp->two = two;
4956 if (queue)
4957 diff_q(queue, dp);
4958 return dp;
4959}
4960void diff_free_filepair(struct diff_filepair *p)
4962{
4963 free_filespec(p->one);
4964 free_filespec(p->two);
4965 free(p);
4966}
4967const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4969{
4970 int abblen;
4971 const char *abbrev;
4972 if (len == GIT_SHA1_HEXSZ)
4974 return oid_to_hex(oid);
4975 abbrev = diff_abbrev_oid(oid, len);
4977 abblen = strlen(abbrev);
4978 /*
4980 * In well-behaved cases, where the abbbreviated result is the
4981 * same as the requested length, append three dots after the
4982 * abbreviation (hence the whole logic is limited to the case
4983 * where abblen < 37); when the actual abbreviated result is a
4984 * bit longer than the requested length, we reduce the number
4985 * of dots so that they match the well-behaved ones. However,
4986 * if the actual abbreviation is longer than the requested
4987 * length by more than three, we give up on aligning, and add
4988 * three dots anyway, to indicate that the output is not the
4989 * full object name. Yes, this may be suboptimal, but this
4990 * appears only in "diff --raw --abbrev" output and it is not
4991 * worth the effort to change it now. Note that this would
4992 * likely to work fine when the automatic sizing of default
4993 * abbreviation length is used--we would be fed -1 in "len" in
4994 * that case, and will end up always appending three-dots, but
4995 * the automatic sizing is supposed to give abblen that ensures
4996 * uniqueness across all objects (statistically speaking).
4997 */
4998 if (abblen < GIT_SHA1_HEXSZ - 3) {
4999 static char hex[GIT_MAX_HEXSZ + 1];
5000 if (len < abblen && abblen <= len + 2)
5001 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
5002 else
5003 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
5004 return hex;
5005 }
5006 return oid_to_hex(oid);
5008}
5009static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
5011{
5012 int line_termination = opt->line_termination;
5013 int inter_name_termination = line_termination ? '\t' : '\0';
5014 fprintf(opt->file, "%s", diff_line_prefix(opt));
5016 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
5017 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
5018 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
5019 fprintf(opt->file, "%s ",
5020 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
5021 }
5022 if (p->score) {
5023 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
5024 inter_name_termination);
5025 } else {
5026 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5027 }
5028 if (p->status == DIFF_STATUS_COPIED ||
5030 p->status == DIFF_STATUS_RENAMED) {
5031 const char *name_a, *name_b;
5032 name_a = p->one->path;
5033 name_b = p->two->path;
5034 strip_prefix(opt->prefix_length, &name_a, &name_b);
5035 write_name_quoted(name_a, opt->file, inter_name_termination);
5036 write_name_quoted(name_b, opt->file, line_termination);
5037 } else {
5038 const char *name_a, *name_b;
5039 name_a = p->one->mode ? p->one->path : p->two->path;
5040 name_b = NULL;
5041 strip_prefix(opt->prefix_length, &name_a, &name_b);
5042 write_name_quoted(name_a, opt->file, line_termination);
5043 }
5044}
5045int diff_unmodified_pair(struct diff_filepair *p)
5047{
5048 /* This function is written stricter than necessary to support
5049 * the currently implemented transformers, but the idea is to
5050 * let transformers to produce diff_filepairs any way they want,
5051 * and filter and clean them up here before producing the output.
5052 */
5053 struct diff_filespec *one = p->one, *two = p->two;
5054 if (DIFF_PAIR_UNMERGED(p))
5056 return 0; /* unmerged is interesting */
5057 /* deletion, addition, mode or type change
5059 * and rename are all interesting.
5060 */
5061 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5062 DIFF_PAIR_MODE_CHANGED(p) ||
5063 strcmp(one->path, two->path))
5064 return 0;
5065 /* both are valid and point at the same path. that is, we are
5067 * dealing with a change.
5068 */
5069 if (one->oid_valid && two->oid_valid &&
5070 !oidcmp(&one->oid, &two->oid) &&
5071 !one->dirty_submodule && !two->dirty_submodule)
5072 return 1; /* no change */
5073 if (!one->oid_valid && !two->oid_valid)
5074 return 1; /* both look at the same file on the filesystem. */
5075 return 0;
5076}
5077static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5079{
5080 if (diff_unmodified_pair(p))
5081 return;
5082 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5084 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5085 return; /* no tree diffs in patch format */
5086 run_diff(p, o);
5088}
5089static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5091 struct diffstat_t *diffstat)
5092{
5093 if (diff_unmodified_pair(p))
5094 return;
5095 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5097 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5098 return; /* no useful stat for tree diffs */
5099 run_diffstat(p, o, diffstat);
5101}
5102static void diff_flush_checkdiff(struct diff_filepair *p,
5104 struct diff_options *o)
5105{
5106 if (diff_unmodified_pair(p))
5107 return;
5108 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5110 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5111 return; /* nothing to check in tree diffs */
5112 run_checkdiff(p, o);
5114}
5115int diff_queue_is_empty(void)
5117{
5118 struct diff_queue_struct *q = &diff_queued_diff;
5119 int i;
5120 for (i = 0; i < q->nr; i++)
5121 if (!diff_unmodified_pair(q->queue[i]))
5122 return 0;
5123 return 1;
5124}
5125#if DIFF_DEBUG
5127void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5128{
5129 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5130 x, one ? one : "",
5131 s->path,
5132 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5133 s->mode,
5134 s->oid_valid ? oid_to_hex(&s->oid) : "");
5135 fprintf(stderr, "queue[%d] %s size %lu\n",
5136 x, one ? one : "",
5137 s->size);
5138}
5139void diff_debug_filepair(const struct diff_filepair *p, int i)
5141{
5142 diff_debug_filespec(p->one, i, "one");
5143 diff_debug_filespec(p->two, i, "two");
5144 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5145 p->score, p->status ? p->status : '?',
5146 p->one->rename_used, p->broken_pair);
5147}
5148void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5150{
5151 int i;
5152 if (msg)
5153 fprintf(stderr, "%s\n", msg);
5154 fprintf(stderr, "q->nr = %d\n", q->nr);
5155 for (i = 0; i < q->nr; i++) {
5156 struct diff_filepair *p = q->queue[i];
5157 diff_debug_filepair(p, i);
5158 }
5159}
5160#endif
5161static void diff_resolve_rename_copy(void)
5163{
5164 int i;
5165 struct diff_filepair *p;
5166 struct diff_queue_struct *q = &diff_queued_diff;
5167 diff_debug_queue("resolve-rename-copy", q);
5169 for (i = 0; i < q->nr; i++) {
5171 p = q->queue[i];
5172 p->status = 0; /* undecided */
5173 if (DIFF_PAIR_UNMERGED(p))
5174 p->status = DIFF_STATUS_UNMERGED;
5175 else if (!DIFF_FILE_VALID(p->one))
5176 p->status = DIFF_STATUS_ADDED;
5177 else if (!DIFF_FILE_VALID(p->two))
5178 p->status = DIFF_STATUS_DELETED;
5179 else if (DIFF_PAIR_TYPE_CHANGED(p))
5180 p->status = DIFF_STATUS_TYPE_CHANGED;
5181 /* from this point on, we are dealing with a pair
5183 * whose both sides are valid and of the same type, i.e.
5184 * either in-place edit or rename/copy edit.
5185 */
5186 else if (DIFF_PAIR_RENAME(p)) {
5187 /*
5188 * A rename might have re-connected a broken
5189 * pair up, causing the pathnames to be the
5190 * same again. If so, that's not a rename at
5191 * all, just a modification..
5192 *
5193 * Otherwise, see if this source was used for
5194 * multiple renames, in which case we decrement
5195 * the count, and call it a copy.
5196 */
5197 if (!strcmp(p->one->path, p->two->path))
5198 p->status = DIFF_STATUS_MODIFIED;
5199 else if (--p->one->rename_used > 0)
5200 p->status = DIFF_STATUS_COPIED;
5201 else
5202 p->status = DIFF_STATUS_RENAMED;
5203 }
5204 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5205 p->one->mode != p->two->mode ||
5206 p->one->dirty_submodule ||
5207 p->two->dirty_submodule ||
5208 is_null_oid(&p->one->oid))
5209 p->status = DIFF_STATUS_MODIFIED;
5210 else {
5211 /* This is a "no-change" entry and should not
5212 * happen anymore, but prepare for broken callers.
5213 */
5214 error("feeding unmodified %s to diffcore",
5215 p->one->path);
5216 p->status = DIFF_STATUS_UNKNOWN;
5217 }
5218 }
5219 diff_debug_queue("resolve-rename-copy done", q);
5220}
5221static int check_pair_status(struct diff_filepair *p)
5223{
5224 switch (p->status) {
5225 case DIFF_STATUS_UNKNOWN:
5226 return 0;
5227 case 0:
5228 die("internal error in diff-resolve-rename-copy");
5229 default:
5230 return 1;
5231 }
5232}
5233static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5235{
5236 int fmt = opt->output_format;
5237 if (fmt & DIFF_FORMAT_CHECKDIFF)
5239 diff_flush_checkdiff(p, opt);
5240 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5241 diff_flush_raw(p, opt);
5242 else if (fmt & DIFF_FORMAT_NAME) {
5243 const char *name_a, *name_b;
5244 name_a = p->two->path;
5245 name_b = NULL;
5246 strip_prefix(opt->prefix_length, &name_a, &name_b);
5247 fprintf(opt->file, "%s", diff_line_prefix(opt));
5248 write_name_quoted(name_a, opt->file, opt->line_termination);
5249 }
5250}
5251static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5253{
5254 struct strbuf sb = STRBUF_INIT;
5255 if (fs->mode)
5256 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5257 else
5258 strbuf_addf(&sb, " %s ", newdelete);
5259 quote_c_style(fs->path, &sb, NULL, 0);
5261 strbuf_addch(&sb, '\n');
5262 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5263 sb.buf, sb.len, 0);
5264 strbuf_release(&sb);
5265}
5266static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5268 int show_name)
5269{
5270 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5271 struct strbuf sb = STRBUF_INIT;
5272 strbuf_addf(&sb, " mode change %06o => %06o",
5273 p->one->mode, p->two->mode);
5274 if (show_name) {
5275 strbuf_addch(&sb, ' ');
5276 quote_c_style(p->two->path, &sb, NULL, 0);
5277 }
5278 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5279 sb.buf, sb.len, 0);
5280 strbuf_release(&sb);
5281 }
5282}
5283static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5285 struct diff_filepair *p)
5286{
5287 struct strbuf sb = STRBUF_INIT;
5288 char *names = pprint_rename(p->one->path, p->two->path);
5289 strbuf_addf(&sb, " %s %s (%d%%)\n",
5290 renamecopy, names, similarity_index(p));
5291 free(names);
5292 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5293 sb.buf, sb.len, 0);
5294 show_mode_change(opt, p, 0);
5295}
5296static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5298{
5299 switch(p->status) {
5300 case DIFF_STATUS_DELETED:
5301 show_file_mode_name(opt, "delete", p->one);
5302 break;
5303 case DIFF_STATUS_ADDED:
5304 show_file_mode_name(opt, "create", p->two);
5305 break;
5306 case DIFF_STATUS_COPIED:
5307 show_rename_copy(opt, "copy", p);
5308 break;
5309 case DIFF_STATUS_RENAMED:
5310 show_rename_copy(opt, "rename", p);
5311 break;
5312 default:
5313 if (p->score) {
5314 struct strbuf sb = STRBUF_INIT;
5315 strbuf_addstr(&sb, " rewrite ");
5316 quote_c_style(p->two->path, &sb, NULL, 0);
5317 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5318 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5319 sb.buf, sb.len, 0);
5320 }
5321 show_mode_change(opt, p, !p->score);
5322 break;
5323 }
5324}
5325struct patch_id_t {
5327 git_SHA_CTX *ctx;
5328 int patchlen;
5329};
5330static int remove_space(char *line, int len)
5332{
5333 int i;
5334 char *dst = line;
5335 unsigned char c;
5336 for (i = 0; i < len; i++)
5338 if (!isspace((c = line[i])))
5339 *dst++ = c;
5340 return dst - line;
5342}
5343static void patch_id_consume(void *priv, char *line, unsigned long len)
5345{
5346 struct patch_id_t *data = priv;
5347 int new_len;
5348 /* Ignore line numbers when computing the SHA1 of the patch */
5350 if (starts_with(line, "@@ -"))
5351 return;
5352 new_len = remove_space(line, len);
5354 git_SHA1_Update(data->ctx, line, new_len);
5356 data->patchlen += new_len;
5357}
5358static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5360{
5361 git_SHA1_Update(ctx, str, strlen(str));
5362}
5363static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5365{
5366 /* large enough for 2^32 in octal */
5367 char buf[12];
5368 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5369 git_SHA1_Update(ctx, buf, len);
5370}
5371/* returns 0 upon success, and writes result into sha1 */
5373static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5374{
5375 struct diff_queue_struct *q = &diff_queued_diff;
5376 int i;
5377 git_SHA_CTX ctx;
5378 struct patch_id_t data;
5379 git_SHA1_Init(&ctx);
5381 memset(&data, 0, sizeof(struct patch_id_t));
5382 data.ctx = &ctx;
5383 for (i = 0; i < q->nr; i++) {
5385 xpparam_t xpp;
5386 xdemitconf_t xecfg;
5387 mmfile_t mf1, mf2;
5388 struct diff_filepair *p = q->queue[i];
5389 int len1, len2;
5390 memset(&xpp, 0, sizeof(xpp));
5392 memset(&xecfg, 0, sizeof(xecfg));
5393 if (p->status == 0)
5394 return error("internal diff status error");
5395 if (p->status == DIFF_STATUS_UNKNOWN)
5396 continue;
5397 if (diff_unmodified_pair(p))
5398 continue;
5399 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5400 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5401 continue;
5402 if (DIFF_PAIR_UNMERGED(p))
5403 continue;
5404 diff_fill_oid_info(p->one);
5406 diff_fill_oid_info(p->two);
5407 len1 = remove_space(p->one->path, strlen(p->one->path));
5409 len2 = remove_space(p->two->path, strlen(p->two->path));
5410 patch_id_add_string(&ctx, "diff--git");
5411 patch_id_add_string(&ctx, "a/");
5412 git_SHA1_Update(&ctx, p->one->path, len1);
5413 patch_id_add_string(&ctx, "b/");
5414 git_SHA1_Update(&ctx, p->two->path, len2);
5415 if (p->one->mode == 0) {
5417 patch_id_add_string(&ctx, "newfilemode");
5418 patch_id_add_mode(&ctx, p->two->mode);
5419 patch_id_add_string(&ctx, "---/dev/null");
5420 patch_id_add_string(&ctx, "+++b/");
5421 git_SHA1_Update(&ctx, p->two->path, len2);
5422 } else if (p->two->mode == 0) {
5423 patch_id_add_string(&ctx, "deletedfilemode");
5424 patch_id_add_mode(&ctx, p->one->mode);
5425 patch_id_add_string(&ctx, "---a/");
5426 git_SHA1_Update(&ctx, p->one->path, len1);
5427 patch_id_add_string(&ctx, "+++/dev/null");
5428 } else {
5429 patch_id_add_string(&ctx, "---a/");
5430 git_SHA1_Update(&ctx, p->one->path, len1);
5431 patch_id_add_string(&ctx, "+++b/");
5432 git_SHA1_Update(&ctx, p->two->path, len2);
5433 }
5434 if (diff_header_only)
5436 continue;
5437 if (fill_mmfile(&mf1, p->one) < 0 ||
5439 fill_mmfile(&mf2, p->two) < 0)
5440 return error("unable to read files to diff");
5441 if (diff_filespec_is_binary(p->one) ||
5443 diff_filespec_is_binary(p->two)) {
5444 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5445 GIT_SHA1_HEXSZ);
5446 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5447 GIT_SHA1_HEXSZ);
5448 continue;
5449 }
5450 xpp.flags = 0;
5452 xecfg.ctxlen = 3;
5453 xecfg.flags = 0;
5454 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
5455 &xpp, &xecfg))
5456 return error("unable to generate patch-id diff for %s",
5457 p->one->path);
5458 }
5459 git_SHA1_Final(oid->hash, &ctx);
5461 return 0;
5462}
5463int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5465{
5466 struct diff_queue_struct *q = &diff_queued_diff;
5467 int i;
5468 int result = diff_get_patch_id(options, oid, diff_header_only);
5469 for (i = 0; i < q->nr; i++)
5471 diff_free_filepair(q->queue[i]);
5472 free(q->queue);
5474 DIFF_QUEUE_CLEAR(q);
5475 return result;
5477}
5478static int is_summary_empty(const struct diff_queue_struct *q)
5480{
5481 int i;
5482 for (i = 0; i < q->nr; i++) {
5484 const struct diff_filepair *p = q->queue[i];
5485 switch (p->status) {
5487 case DIFF_STATUS_DELETED:
5488 case DIFF_STATUS_ADDED:
5489 case DIFF_STATUS_COPIED:
5490 case DIFF_STATUS_RENAMED:
5491 return 0;
5492 default:
5493 if (p->score)
5494 return 0;
5495 if (p->one->mode && p->two->mode &&
5496 p->one->mode != p->two->mode)
5497 return 0;
5498 break;
5499 }
5500 }
5501 return 1;
5502}
5503static const char rename_limit_warning[] =
5505N_("inexact rename detection was skipped due to too many files.");
5506static const char degrade_cc_to_c_warning[] =
5508N_("only found copies from modified paths due to too many files.");
5509static const char rename_limit_advice[] =
5511N_("you may want to set your %s variable to at least "
5512 "%d and retry the command.");
5513void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5515{
5516 if (degraded_cc)
5517 warning(_(degrade_cc_to_c_warning));
5518 else if (needed)
5519 warning(_(rename_limit_warning));
5520 else
5521 return;
5522 if (0 < needed && needed < 32767)
5523 warning(_(rename_limit_advice), varname, needed);
5524}
5525static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5527{
5528 int i;
5529 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5530 struct diff_queue_struct *q = &diff_queued_diff;
5531 if (WSEH_NEW & WS_RULE_MASK)
5533 die("BUG: WS rules bit mask overlaps with diff symbol flags");
5534 if (o->color_moved)
5536 o->emitted_symbols = &esm;
5537 for (i = 0; i < q->nr; i++) {
5539 struct diff_filepair *p = q->queue[i];
5540 if (check_pair_status(p))
5541 diff_flush_patch(p, o);
5542 }
5543 if (o->emitted_symbols) {
5545 if (o->color_moved) {
5546 struct hashmap add_lines, del_lines;
5547 hashmap_init(&del_lines,
5549 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5550 hashmap_init(&add_lines,
5551 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5552 add_lines_to_move_detection(o, &add_lines, &del_lines);
5554 mark_color_as_moved(o, &add_lines, &del_lines);
5555 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5556 dim_moved_lines(o);
5557 hashmap_free(&add_lines, 0);
5559 hashmap_free(&del_lines, 0);
5560 }
5561 for (i = 0; i < esm.nr; i++)
5563 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5564 for (i = 0; i < esm.nr; i++)
5566 free((void *)esm.buf[i].line);
5567 }
5568 esm.nr = 0;
5569}
5570void diff_flush(struct diff_options *options)
5572{
5573 struct diff_queue_struct *q = &diff_queued_diff;
5574 int i, output_format = options->output_format;
5575 int separator = 0;
5576 int dirstat_by_line = 0;
5577 /*
5579 * Order: raw, stat, summary, patch
5580 * or: name/name-status/checkdiff (other bits clear)
5581 */
5582 if (!q->nr)
5583 goto free_queue;
5584 if (output_format & (DIFF_FORMAT_RAW |
5586 DIFF_FORMAT_NAME |
5587 DIFF_FORMAT_NAME_STATUS |
5588 DIFF_FORMAT_CHECKDIFF)) {
5589 for (i = 0; i < q->nr; i++) {
5590 struct diff_filepair *p = q->queue[i];
5591 if (check_pair_status(p))
5592 flush_one_pair(p, options);
5593 }
5594 separator++;
5595 }
5596 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
5598 dirstat_by_line = 1;
5599 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5601 dirstat_by_line) {
5602 struct diffstat_t diffstat;
5603 memset(&diffstat, 0, sizeof(struct diffstat_t));
5605 for (i = 0; i < q->nr; i++) {
5606 struct diff_filepair *p = q->queue[i];
5607 if (check_pair_status(p))
5608 diff_flush_stat(p, options, &diffstat);
5609 }
5610 if (output_format & DIFF_FORMAT_NUMSTAT)
5611 show_numstat(&diffstat, options);
5612 if (output_format & DIFF_FORMAT_DIFFSTAT)
5613 show_stats(&diffstat, options);
5614 if (output_format & DIFF_FORMAT_SHORTSTAT)
5615 show_shortstats(&diffstat, options);
5616 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5617 show_dirstat_by_line(&diffstat, options);
5618 free_diffstat_info(&diffstat);
5619 separator++;
5620 }
5621 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5622 show_dirstat(options);
5623 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5625 for (i = 0; i < q->nr; i++) {
5626 diff_summary(options, q->queue[i]);
5627 }
5628 separator++;
5629 }
5630 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5632 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
5633 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5634 /*
5635 * run diff_flush_patch for the exit status. setting
5636 * options->file to /dev/null should be safe, because we
5637 * aren't supposed to produce any output anyway.
5638 */
5639 if (options->close_file)
5640 fclose(options->file);
5641 options->file = xfopen("/dev/null", "w");
5642 options->close_file = 1;
5643 options->color_moved = 0;
5644 for (i = 0; i < q->nr; i++) {
5645 struct diff_filepair *p = q->queue[i];
5646 if (check_pair_status(p))
5647 diff_flush_patch(p, options);
5648 if (options->found_changes)
5649 break;
5650 }
5651 }
5652 if (output_format & DIFF_FORMAT_PATCH) {
5654 if (separator) {
5655 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5656 if (options->stat_sep)
5657 /* attach patch instead of inline */
5658 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5659 NULL, 0, 0);
5660 }
5661 diff_flush_patch_all_file_pairs(options);
5663 }
5664 if (output_format & DIFF_FORMAT_CALLBACK)
5666 options->format_callback(q, options, options->format_callback_data);
5667 for (i = 0; i < q->nr; i++)
5669 diff_free_filepair(q->queue[i]);
5670free_queue:
5671 free(q->queue);
5672 DIFF_QUEUE_CLEAR(q);
5673 if (options->close_file)
5674 fclose(options->file);
5675 /*
5677 * Report the content-level differences with HAS_CHANGES;
5678 * diff_addremove/diff_change does not set the bit when
5679 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5680 */
5681 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5682 if (options->found_changes)
5683 DIFF_OPT_SET(options, HAS_CHANGES);
5684 else
5685 DIFF_OPT_CLR(options, HAS_CHANGES);
5686 }
5687}
5688static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5690{
5691 return (((p->status == DIFF_STATUS_MODIFIED) &&
5692 ((p->score &&
5693 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5694 (!p->score &&
5695 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5696 ((p->status != DIFF_STATUS_MODIFIED) &&
5697 filter_bit_tst(p->status, options)));
5698}
5699static void diffcore_apply_filter(struct diff_options *options)
5701{
5702 int i;
5703 struct diff_queue_struct *q = &diff_queued_diff;
5704 struct diff_queue_struct outq;
5705 DIFF_QUEUE_CLEAR(&outq);
5707 if (!options->filter)
5709 return;
5710 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5712 int found;
5713 for (i = found = 0; !found && i < q->nr; i++) {
5714 if (match_filter(options, q->queue[i]))
5715 found++;
5716 }
5717 if (found)
5718 return;
5719 /* otherwise we will clear the whole queue
5721 * by copying the empty outq at the end of this
5722 * function, but first clear the current entries
5723 * in the queue.
5724 */
5725 for (i = 0; i < q->nr; i++)
5726 diff_free_filepair(q->queue[i]);
5727 }
5728 else {
5729 /* Only the matching ones */
5730 for (i = 0; i < q->nr; i++) {
5731 struct diff_filepair *p = q->queue[i];
5732 if (match_filter(options, p))
5733 diff_q(&outq, p);
5734 else
5735 diff_free_filepair(p);
5736 }
5737 }
5738 free(q->queue);
5739 *q = outq;
5740}
5741/* Check whether two filespecs with the same mode and size are identical */
5743static int diff_filespec_is_identical(struct diff_filespec *one,
5744 struct diff_filespec *two)
5745{
5746 if (S_ISGITLINK(one->mode))
5747 return 0;
5748 if (diff_populate_filespec(one, 0))
5749 return 0;
5750 if (diff_populate_filespec(two, 0))
5751 return 0;
5752 return !memcmp(one->data, two->data, one->size);
5753}
5754static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
5756{
5757 if (p->done_skip_stat_unmatch)
5758 return p->skip_stat_unmatch_result;
5759 p->done_skip_stat_unmatch = 1;
5761 p->skip_stat_unmatch_result = 0;
5762 /*
5763 * 1. Entries that come from stat info dirtiness
5764 * always have both sides (iow, not create/delete),
5765 * one side of the object name is unknown, with
5766 * the same mode and size. Keep the ones that
5767 * do not match these criteria. They have real
5768 * differences.
5769 *
5770 * 2. At this point, the file is known to be modified,
5771 * with the same mode and size, and the object
5772 * name of one side is unknown. Need to inspect
5773 * the identical contents.
5774 */
5775 if (!DIFF_FILE_VALID(p->one) || /* (1) */
5776 !DIFF_FILE_VALID(p->two) ||
5777 (p->one->oid_valid && p->two->oid_valid) ||
5778 (p->one->mode != p->two->mode) ||
5779 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
5780 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
5781 (p->one->size != p->two->size) ||
5782 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
5783 p->skip_stat_unmatch_result = 1;
5784 return p->skip_stat_unmatch_result;
5785}
5786static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
5788{
5789 int i;
5790 struct diff_queue_struct *q = &diff_queued_diff;
5791 struct diff_queue_struct outq;
5792 DIFF_QUEUE_CLEAR(&outq);
5793 for (i = 0; i < q->nr; i++) {
5795 struct diff_filepair *p = q->queue[i];
5796 if (diff_filespec_check_stat_unmatch(p))
5798 diff_q(&outq, p);
5799 else {
5800 /*
5801 * The caller can subtract 1 from skip_stat_unmatch
5802 * to determine how many paths were dirty only
5803 * due to stat info mismatch.
5804 */
5805 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
5806 diffopt->skip_stat_unmatch++;
5807 diff_free_filepair(p);
5808 }
5809 }
5810 free(q->queue);
5811 *q = outq;
5812}
5813static int diffnamecmp(const void *a_, const void *b_)
5815{
5816 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
5817 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
5818 const char *name_a, *name_b;
5819 name_a = a->one ? a->one->path : a->two->path;
5821 name_b = b->one ? b->one->path : b->two->path;
5822 return strcmp(name_a, name_b);
5823}
5824void diffcore_fix_diff_index(struct diff_options *options)
5826{
5827 struct diff_queue_struct *q = &diff_queued_diff;
5828 QSORT(q->queue, q->nr, diffnamecmp);
5829}
5830void diffcore_std(struct diff_options *options)
5832{
5833 /* NOTE please keep the following in sync with diff_tree_combined() */
5834 if (options->skip_stat_unmatch)
5835 diffcore_skip_stat_unmatch(options);
5836 if (!options->found_follow) {
5837 /* See try_to_follow_renames() in tree-diff.c */
5838 if (options->break_opt != -1)
5839 diffcore_break(options->break_opt);
5840 if (options->detect_rename)
5841 diffcore_rename(options);
5842 if (options->break_opt != -1)
5843 diffcore_merge_broken();
5844 }
5845 if (options->pickaxe)
5846 diffcore_pickaxe(options);
5847 if (options->orderfile)
5848 diffcore_order(options->orderfile);
5849 if (!options->found_follow)
5850 /* See try_to_follow_renames() in tree-diff.c */
5851 diff_resolve_rename_copy();
5852 diffcore_apply_filter(options);
5853 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5855 DIFF_OPT_SET(options, HAS_CHANGES);
5856 else
5857 DIFF_OPT_CLR(options, HAS_CHANGES);
5858 options->found_follow = 0;
5860}
5861int diff_result_code(struct diff_options *opt, int status)
5863{
5864 int result = 0;
5865 diff_warn_rename_limit("diff.renameLimit",
5867 opt->needed_rename_limit,
5868 opt->degraded_cc_to_c);
5869 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5870 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5871 return status;
5872 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5873 DIFF_OPT_TST(opt, HAS_CHANGES))
5874 result |= 01;
5875 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5876 DIFF_OPT_TST(opt, CHECK_FAILED))
5877 result |= 02;
5878 return result;
5879}
5880int diff_can_quit_early(struct diff_options *opt)
5882{
5883 return (DIFF_OPT_TST(opt, QUICK) &&
5884 !opt->filter &&
5885 DIFF_OPT_TST(opt, HAS_CHANGES));
5886}
5887/*
5889 * Shall changes to this submodule be ignored?
5890 *
5891 * Submodule changes can be configured to be ignored separately for each path,
5892 * but that configuration can be overridden from the command line.
5893 */
5894static int is_submodule_ignored(const char *path, struct diff_options *options)
5895{
5896 int ignored = 0;
5897 unsigned orig_flags = options->flags;
5898 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5899 set_diffopt_flags_from_submodule_config(options, path);
5900 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5901 ignored = 1;
5902 options->flags = orig_flags;
5903 return ignored;
5904}
5905void diff_addremove(struct diff_options *options,
5907 int addremove, unsigned mode,
5908 const struct object_id *oid,
5909 int oid_valid,
5910 const char *concatpath, unsigned dirty_submodule)
5911{
5912 struct diff_filespec *one, *two;
5913 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5915 return;
5916 /* This may look odd, but it is a preparation for
5918 * feeding "there are unchanged files which should
5919 * not produce diffs, but when you are doing copy
5920 * detection you would need them, so here they are"
5921 * entries to the diff-core. They will be prefixed
5922 * with something like '=' or '*' (I haven't decided
5923 * which but should not make any difference).
5924 * Feeding the same new and old to diff_change()
5925 * also has the same effect.
5926 * Before the final output happens, they are pruned after
5927 * merged into rename/copy pairs as appropriate.
5928 */
5929 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5930 addremove = (addremove == '+' ? '-' :
5931 addremove == '-' ? '+' : addremove);
5932 if (options->prefix &&
5934 strncmp(concatpath, options->prefix, options->prefix_length))
5935 return;
5936 one = alloc_filespec(concatpath);
5938 two = alloc_filespec(concatpath);
5939 if (addremove != '+')
5941 fill_filespec(one, oid, oid_valid, mode);
5942 if (addremove != '-') {
5943 fill_filespec(two, oid, oid_valid, mode);
5944 two->dirty_submodule = dirty_submodule;
5945 }
5946 diff_queue(&diff_queued_diff, one, two);
5948 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5949 DIFF_OPT_SET(options, HAS_CHANGES);
5950}
5951void diff_change(struct diff_options *options,
5953 unsigned old_mode, unsigned new_mode,
5954 const struct object_id *old_oid,
5955 const struct object_id *new_oid,
5956 int old_oid_valid, int new_oid_valid,
5957 const char *concatpath,
5958 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5959{
5960 struct diff_filespec *one, *two;
5961 struct diff_filepair *p;
5962 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5964 is_submodule_ignored(concatpath, options))
5965 return;
5966 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5968 SWAP(old_mode, new_mode);
5969 SWAP(old_oid, new_oid);
5970 SWAP(old_oid_valid, new_oid_valid);
5971 SWAP(old_dirty_submodule, new_dirty_submodule);
5972 }
5973 if (options->prefix &&
5975 strncmp(concatpath, options->prefix, options->prefix_length))
5976 return;
5977 one = alloc_filespec(concatpath);
5979 two = alloc_filespec(concatpath);
5980 fill_filespec(one, old_oid, old_oid_valid, old_mode);
5981 fill_filespec(two, new_oid, new_oid_valid, new_mode);
5982 one->dirty_submodule = old_dirty_submodule;
5983 two->dirty_submodule = new_dirty_submodule;
5984 p = diff_queue(&diff_queued_diff, one, two);
5985 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5987 return;
5988 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5990 !diff_filespec_check_stat_unmatch(p))
5991 return;
5992 DIFF_OPT_SET(options, HAS_CHANGES);
5994}
5995struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5997{
5998 struct diff_filepair *pair;
5999 struct diff_filespec *one, *two;
6000 if (options->prefix &&
6002 strncmp(path, options->prefix, options->prefix_length))
6003 return NULL;
6004 one = alloc_filespec(path);
6006 two = alloc_filespec(path);
6007 pair = diff_queue(&diff_queued_diff, one, two);
6008 pair->is_unmerged = 1;
6009 return pair;
6010}
6011static char *run_textconv(const char *pgm, struct diff_filespec *spec,
6013 size_t *outsize)
6014{
6015 struct diff_tempfile *temp;
6016 const char *argv[3];
6017 const char **arg = argv;
6018 struct child_process child = CHILD_PROCESS_INIT;
6019 struct strbuf buf = STRBUF_INIT;
6020 int err = 0;
6021 temp = prepare_temp_file(spec->path, spec);
6023 *arg++ = pgm;
6024 *arg++ = temp->name;
6025 *arg = NULL;
6026 child.use_shell = 1;
6028 child.argv = argv;
6029 child.out = -1;
6030 if (start_command(&child)) {
6031 remove_tempfile();
6032 return NULL;
6033 }
6034 if (strbuf_read(&buf, child.out, 0) < 0)
6036 err = error("error reading from textconv command '%s'", pgm);
6037 close(child.out);
6038 if (finish_command(&child) || err) {
6040 strbuf_release(&buf);
6041 remove_tempfile();
6042 return NULL;
6043 }
6044 remove_tempfile();
6045 return strbuf_detach(&buf, outsize);
6047}
6048size_t fill_textconv(struct userdiff_driver *driver,
6050 struct diff_filespec *df,
6051 char **outbuf)
6052{
6053 size_t size;
6054 if (!driver) {
6056 if (!DIFF_FILE_VALID(df)) {
6057 *outbuf = "";
6058 return 0;
6059 }
6060 if (diff_populate_filespec(df, 0))
6061 die("unable to read files to diff");
6062 *outbuf = df->data;
6063 return df->size;
6064 }
6065 if (!driver->textconv)
6067 die("BUG: fill_textconv called with non-textconv driver");
6068 if (driver->textconv_cache && df->oid_valid) {
6070 *outbuf = notes_cache_get(driver->textconv_cache,
6071 &df->oid,
6072 &size);
6073 if (*outbuf)
6074 return size;
6075 }
6076 *outbuf = run_textconv(driver->textconv, df, &size);
6078 if (!*outbuf)
6079 die("unable to read files to diff");
6080 if (driver->textconv_cache && df->oid_valid) {
6082 /* ignore errors, as we might be in a readonly repository */
6083 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6084 size);
6085 /*
6086 * we could save up changes and flush them all at the end,
6087 * but we would need an extra call after all diffing is done.
6088 * Since generating a cache entry is the slow path anyway,
6089 * this extra overhead probably isn't a big deal.
6090 */
6091 notes_cache_write(driver->textconv_cache);
6092 }
6093 return size;
6095}
6096int textconv_object(const char *path,
6098 unsigned mode,
6099 const struct object_id *oid,
6100 int oid_valid,
6101 char **buf,
6102 unsigned long *buf_size)
6103{
6104 struct diff_filespec *df;
6105 struct userdiff_driver *textconv;
6106 df = alloc_filespec(path);
6108 fill_filespec(df, oid, oid_valid, mode);
6109 textconv = get_textconv(df);
6110 if (!textconv) {
6111 free_filespec(df);
6112 return 0;
6113 }
6114 *buf_size = fill_textconv(textconv, df, buf);
6116 free_filespec(df);
6117 return 1;
6118}
6119void setup_diff_pager(struct diff_options *opt)
6121{
6122 /*
6123 * If the user asked for our exit code, then either they want --quiet
6124 * or --exit-code. We should definitely not bother with a pager in the
6125 * former case, as we will generate no output. Since we still properly
6126 * report our exit code even when a pager is run, we _could_ run a
6127 * pager with --exit-code. But since we have not done so historically,
6128 * and because it is easy to find people oneline advising "git diff
6129 * --exit-code" in hooks and other scripts, we do not do so.
6130 */
6131 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
6132 check_pager_config("diff") != 0)
6133 setup_pager();
6134}