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#ifdef NO_FAST_WORKING_DIRECTORY
26#define FAST_WORKING_DIRECTORY 0
27#else
28#define FAST_WORKING_DIRECTORY 1
29#endif
30static int diff_detect_rename_default;
32static int diff_indent_heuristic = 1;
33static int diff_rename_limit_default = 400;
34static int diff_suppress_blank_empty;
35static int diff_use_color_default = -1;
36static int diff_color_moved_default;
37static int diff_context_default = 3;
38static int diff_interhunk_context_default;
39static const char *diff_word_regex_cfg;
40static const char *external_diff_cmd_cfg;
41static const char *diff_order_file_cfg;
42int diff_auto_refresh_index = 1;
43static int diff_mnemonic_prefix;
44static int diff_no_prefix;
45static int diff_stat_graph_width;
46static int diff_dirstat_permille_default = 30;
47static struct diff_options default_diff_options;
48static long diff_algorithm;
49static unsigned ws_error_highlight_default = WSEH_NEW;
50static char diff_colors[][COLOR_MAXLEN] = {
52 GIT_COLOR_RESET,
53 GIT_COLOR_NORMAL, /* CONTEXT */
54 GIT_COLOR_BOLD, /* METAINFO */
55 GIT_COLOR_CYAN, /* FRAGINFO */
56 GIT_COLOR_RED, /* OLD */
57 GIT_COLOR_GREEN, /* NEW */
58 GIT_COLOR_YELLOW, /* COMMIT */
59 GIT_COLOR_BG_RED, /* WHITESPACE */
60 GIT_COLOR_NORMAL, /* FUNCINFO */
61 GIT_COLOR_BOLD_MAGENTA, /* OLD_MOVED */
62 GIT_COLOR_BOLD_BLUE, /* OLD_MOVED ALTERNATIVE */
63 GIT_COLOR_FAINT, /* OLD_MOVED_DIM */
64 GIT_COLOR_FAINT_ITALIC, /* OLD_MOVED_ALTERNATIVE_DIM */
65 GIT_COLOR_BOLD_CYAN, /* NEW_MOVED */
66 GIT_COLOR_BOLD_YELLOW, /* NEW_MOVED ALTERNATIVE */
67 GIT_COLOR_FAINT, /* NEW_MOVED_DIM */
68 GIT_COLOR_FAINT_ITALIC, /* NEW_MOVED_ALTERNATIVE_DIM */
69};
70static NORETURN void die_want_option(const char *option_name)
72{
73 die(_("option '%s' requires a value"), option_name);
74}
75static int parse_diff_color_slot(const char *var)
77{
78 if (!strcasecmp(var, "context") || !strcasecmp(var, "plain"))
79 return DIFF_CONTEXT;
80 if (!strcasecmp(var, "meta"))
81 return DIFF_METAINFO;
82 if (!strcasecmp(var, "frag"))
83 return DIFF_FRAGINFO;
84 if (!strcasecmp(var, "old"))
85 return DIFF_FILE_OLD;
86 if (!strcasecmp(var, "new"))
87 return DIFF_FILE_NEW;
88 if (!strcasecmp(var, "commit"))
89 return DIFF_COMMIT;
90 if (!strcasecmp(var, "whitespace"))
91 return DIFF_WHITESPACE;
92 if (!strcasecmp(var, "func"))
93 return DIFF_FUNCINFO;
94 if (!strcasecmp(var, "oldmoved"))
95 return DIFF_FILE_OLD_MOVED;
96 if (!strcasecmp(var, "oldmovedalternative"))
97 return DIFF_FILE_OLD_MOVED_ALT;
98 if (!strcasecmp(var, "oldmoveddimmed"))
99 return DIFF_FILE_OLD_MOVED_DIM;
100 if (!strcasecmp(var, "oldmovedalternativedimmed"))
101 return DIFF_FILE_OLD_MOVED_ALT_DIM;
102 if (!strcasecmp(var, "newmoved"))
103 return DIFF_FILE_NEW_MOVED;
104 if (!strcasecmp(var, "newmovedalternative"))
105 return DIFF_FILE_NEW_MOVED_ALT;
106 if (!strcasecmp(var, "newmoveddimmed"))
107 return DIFF_FILE_NEW_MOVED_DIM;
108 if (!strcasecmp(var, "newmovedalternativedimmed"))
109 return DIFF_FILE_NEW_MOVED_ALT_DIM;
110 return -1;
111}
112static int parse_dirstat_params(struct diff_options *options, const char *params_string,
114 struct strbuf *errmsg)
115{
116 char *params_copy = xstrdup(params_string);
117 struct string_list params = STRING_LIST_INIT_NODUP;
118 int ret = 0;
119 int i;
120 if (*params_copy)
122 string_list_split_in_place(¶ms, params_copy, ',', -1);
123 for (i = 0; i < params.nr; i++) {
124 const char *p = params.items[i].string;
125 if (!strcmp(p, "changes")) {
126 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
127 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
128 } else if (!strcmp(p, "lines")) {
129 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
130 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
131 } else if (!strcmp(p, "files")) {
132 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
133 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
134 } else if (!strcmp(p, "noncumulative")) {
135 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
136 } else if (!strcmp(p, "cumulative")) {
137 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
138 } else if (isdigit(*p)) {
139 char *end;
140 int permille = strtoul(p, &end, 10) * 10;
141 if (*end == '.' && isdigit(*++end)) {
142 /* only use first digit */
143 permille += *end - '0';
144 /* .. and ignore any further digits */
145 while (isdigit(*++end))
146 ; /* nothing */
147 }
148 if (!*end)
149 options->dirstat_permille = permille;
150 else {
151 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%s'\n"),
152 p);
153 ret++;
154 }
155 } else {
156 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%s'\n"), p);
157 ret++;
158 }
159 }
161 string_list_clear(¶ms, 0);
162 free(params_copy);
163 return ret;
164}
165static int parse_submodule_params(struct diff_options *options, const char *value)
167{
168 if (!strcmp(value, "log"))
169 options->submodule_format = DIFF_SUBMODULE_LOG;
170 else if (!strcmp(value, "short"))
171 options->submodule_format = DIFF_SUBMODULE_SHORT;
172 else if (!strcmp(value, "diff"))
173 options->submodule_format = DIFF_SUBMODULE_INLINE_DIFF;
174 else
175 return -1;
176 return 0;
177}
178static int git_config_rename(const char *var, const char *value)
180{
181 if (!value)
182 return DIFF_DETECT_RENAME;
183 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
184 return DIFF_DETECT_COPY;
185 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
186}
187long parse_algorithm_value(const char *value)
189{
190 if (!value)
191 return -1;
192 else if (!strcasecmp(value, "myers") || !strcasecmp(value, "default"))
193 return 0;
194 else if (!strcasecmp(value, "minimal"))
195 return XDF_NEED_MINIMAL;
196 else if (!strcasecmp(value, "patience"))
197 return XDF_PATIENCE_DIFF;
198 else if (!strcasecmp(value, "histogram"))
199 return XDF_HISTOGRAM_DIFF;
200 return -1;
201}
202static int parse_one_token(const char **arg, const char *token)
204{
205 const char *rest;
206 if (skip_prefix(*arg, token, &rest) && (!*rest || *rest == ',')) {
207 *arg = rest;
208 return 1;
209 }
210 return 0;
211}
212static int parse_ws_error_highlight(const char *arg)
214{
215 const char *orig_arg = arg;
216 unsigned val = 0;
217 while (*arg) {
219 if (parse_one_token(&arg, "none"))
220 val = 0;
221 else if (parse_one_token(&arg, "default"))
222 val = WSEH_NEW;
223 else if (parse_one_token(&arg, "all"))
224 val = WSEH_NEW | WSEH_OLD | WSEH_CONTEXT;
225 else if (parse_one_token(&arg, "new"))
226 val |= WSEH_NEW;
227 else if (parse_one_token(&arg, "old"))
228 val |= WSEH_OLD;
229 else if (parse_one_token(&arg, "context"))
230 val |= WSEH_CONTEXT;
231 else {
232 return -1 - (int)(arg - orig_arg);
233 }
234 if (*arg)
235 arg++;
236 }
237 return val;
238}
239/*
241 * These are to give UI layer defaults.
242 * The core-level commands such as git-diff-files should
243 * never be affected by the setting of diff.renames
244 * the user happens to have in the configuration file.
245 */
246void init_diff_ui_defaults(void)
247{
248 diff_detect_rename_default = 1;
249}
250int git_diff_heuristic_config(const char *var, const char *value, void *cb)
252{
253 if (!strcmp(var, "diff.indentheuristic"))
254 diff_indent_heuristic = git_config_bool(var, value);
255 return 0;
256}
257static int parse_color_moved(const char *arg)
259{
260 switch (git_parse_maybe_bool(arg)) {
261 case 0:
262 return COLOR_MOVED_NO;
263 case 1:
264 return COLOR_MOVED_DEFAULT;
265 default:
266 break;
267 }
268 if (!strcmp(arg, "no"))
270 return COLOR_MOVED_NO;
271 else if (!strcmp(arg, "plain"))
272 return COLOR_MOVED_PLAIN;
273 else if (!strcmp(arg, "zebra"))
274 return COLOR_MOVED_ZEBRA;
275 else if (!strcmp(arg, "default"))
276 return COLOR_MOVED_DEFAULT;
277 else if (!strcmp(arg, "dimmed_zebra"))
278 return COLOR_MOVED_ZEBRA_DIM;
279 else
280 return error(_("color moved setting must be one of 'no', 'default', 'zebra', 'dimmed_zebra', 'plain'"));
281}
282int git_diff_ui_config(const char *var, const char *value, void *cb)
284{
285 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
286 diff_use_color_default = git_config_colorbool(var, value);
287 return 0;
288 }
289 if (!strcmp(var, "diff.colormoved")) {
290 int cm = parse_color_moved(value);
291 if (cm < 0)
292 return -1;
293 diff_color_moved_default = cm;
294 return 0;
295 }
296 if (!strcmp(var, "diff.context")) {
297 diff_context_default = git_config_int(var, value);
298 if (diff_context_default < 0)
299 return -1;
300 return 0;
301 }
302 if (!strcmp(var, "diff.interhunkcontext")) {
303 diff_interhunk_context_default = git_config_int(var, value);
304 if (diff_interhunk_context_default < 0)
305 return -1;
306 return 0;
307 }
308 if (!strcmp(var, "diff.renames")) {
309 diff_detect_rename_default = git_config_rename(var, value);
310 return 0;
311 }
312 if (!strcmp(var, "diff.autorefreshindex")) {
313 diff_auto_refresh_index = git_config_bool(var, value);
314 return 0;
315 }
316 if (!strcmp(var, "diff.mnemonicprefix")) {
317 diff_mnemonic_prefix = git_config_bool(var, value);
318 return 0;
319 }
320 if (!strcmp(var, "diff.noprefix")) {
321 diff_no_prefix = git_config_bool(var, value);
322 return 0;
323 }
324 if (!strcmp(var, "diff.statgraphwidth")) {
325 diff_stat_graph_width = git_config_int(var, value);
326 return 0;
327 }
328 if (!strcmp(var, "diff.external"))
329 return git_config_string(&external_diff_cmd_cfg, var, value);
330 if (!strcmp(var, "diff.wordregex"))
331 return git_config_string(&diff_word_regex_cfg, var, value);
332 if (!strcmp(var, "diff.orderfile"))
333 return git_config_pathname(&diff_order_file_cfg, var, value);
334 if (!strcmp(var, "diff.ignoresubmodules"))
336 handle_ignore_submodules_arg(&default_diff_options, value);
337 if (!strcmp(var, "diff.submodule")) {
339 if (parse_submodule_params(&default_diff_options, value))
340 warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
341 value);
342 return 0;
343 }
344 if (!strcmp(var, "diff.algorithm")) {
346 diff_algorithm = parse_algorithm_value(value);
347 if (diff_algorithm < 0)
348 return -1;
349 return 0;
350 }
351 if (!strcmp(var, "diff.wserrorhighlight")) {
353 int val = parse_ws_error_highlight(value);
354 if (val < 0)
355 return -1;
356 ws_error_highlight_default = val;
357 return 0;
358 }
359 return git_diff_basic_config(var, value, cb);
361}
362int git_diff_basic_config(const char *var, const char *value, void *cb)
364{
365 const char *name;
366 if (!strcmp(var, "diff.renamelimit")) {
368 diff_rename_limit_default = git_config_int(var, value);
369 return 0;
370 }
371 if (userdiff_config(var, value) < 0)
373 return -1;
374 if (skip_prefix(var, "diff.color.", &name) ||
376 skip_prefix(var, "color.diff.", &name)) {
377 int slot = parse_diff_color_slot(name);
378 if (slot < 0)
379 return 0;
380 if (!value)
381 return config_error_nonbool(var);
382 return color_parse(value, diff_colors[slot]);
383 }
384 /* like GNU diff's --suppress-blank-empty option */
386 if (!strcmp(var, "diff.suppressblankempty") ||
387 /* for backwards compatibility */
388 !strcmp(var, "diff.suppress-blank-empty")) {
389 diff_suppress_blank_empty = git_config_bool(var, value);
390 return 0;
391 }
392 if (!strcmp(var, "diff.dirstat")) {
394 struct strbuf errmsg = STRBUF_INIT;
395 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
396 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
397 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
398 errmsg.buf);
399 strbuf_release(&errmsg);
400 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
401 return 0;
402 }
403 if (starts_with(var, "submodule."))
405 return parse_submodule_config_option(var, value);
406 if (git_diff_heuristic_config(var, value, cb) < 0)
408 return -1;
409 return git_default_config(var, value, cb);
411}
412static char *quote_two(const char *one, const char *two)
414{
415 int need_one = quote_c_style(one, NULL, NULL, 1);
416 int need_two = quote_c_style(two, NULL, NULL, 1);
417 struct strbuf res = STRBUF_INIT;
418 if (need_one + need_two) {
420 strbuf_addch(&res, '"');
421 quote_c_style(one, &res, NULL, 1);
422 quote_c_style(two, &res, NULL, 1);
423 strbuf_addch(&res, '"');
424 } else {
425 strbuf_addstr(&res, one);
426 strbuf_addstr(&res, two);
427 }
428 return strbuf_detach(&res, NULL);
429}
430static const char *external_diff(void)
432{
433 static const char *external_diff_cmd = NULL;
434 static int done_preparing = 0;
435 if (done_preparing)
437 return external_diff_cmd;
438 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
439 if (!external_diff_cmd)
440 external_diff_cmd = external_diff_cmd_cfg;
441 done_preparing = 1;
442 return external_diff_cmd;
443}
444/*
446 * Keep track of files used for diffing. Sometimes such an entry
447 * refers to a temporary file, sometimes to an existing file, and
448 * sometimes to "/dev/null".
449 */
450static struct diff_tempfile {
451 /*
452 * filename external diff should read from, or NULL if this
453 * entry is currently not in use:
454 */
455 const char *name;
456 char hex[GIT_MAX_HEXSZ + 1];
458 char mode[10];
459 /*
461 * If this diff_tempfile instance refers to a temporary file,
462 * this tempfile object is used to manage its lifetime.
463 */
464 struct tempfile tempfile;
465} diff_temp[2];
466typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
468struct emit_callback {
470 int color_diff;
471 unsigned ws_rule;
472 int blank_at_eof_in_preimage;
473 int blank_at_eof_in_postimage;
474 int lno_in_preimage;
475 int lno_in_postimage;
476 sane_truncate_fn truncate;
477 const char **label_path;
478 struct diff_words_data *diff_words;
479 struct diff_options *opt;
480 struct strbuf *header;
481};
482static int count_lines(const char *data, int size)
484{
485 int count, ch, completely_empty = 1, nl_just_seen = 0;
486 count = 0;
487 while (0 < size--) {
488 ch = *data++;
489 if (ch == '\n') {
490 count++;
491 nl_just_seen = 1;
492 completely_empty = 0;
493 }
494 else {
495 nl_just_seen = 0;
496 completely_empty = 0;
497 }
498 }
499 if (completely_empty)
500 return 0;
501 if (!nl_just_seen)
502 count++; /* no trailing newline */
503 return count;
504}
505static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
507{
508 if (!DIFF_FILE_VALID(one)) {
509 mf->ptr = (char *)""; /* does not matter */
510 mf->size = 0;
511 return 0;
512 }
513 else if (diff_populate_filespec(one, 0))
514 return -1;
515 mf->ptr = one->data;
517 mf->size = one->size;
518 return 0;
519}
520/* like fill_mmfile, but only for size, so we can avoid retrieving blob */
522static unsigned long diff_filespec_size(struct diff_filespec *one)
523{
524 if (!DIFF_FILE_VALID(one))
525 return 0;
526 diff_populate_filespec(one, CHECK_SIZE_ONLY);
527 return one->size;
528}
529static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
531{
532 char *ptr = mf->ptr;
533 long size = mf->size;
534 int cnt = 0;
535 if (!size)
537 return cnt;
538 ptr += size - 1; /* pointing at the very end */
539 if (*ptr != '\n')
540 ; /* incomplete line */
541 else
542 ptr--; /* skip the last LF */
543 while (mf->ptr < ptr) {
544 char *prev_eol;
545 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
546 if (*prev_eol == '\n')
547 break;
548 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
549 break;
550 cnt++;
551 ptr = prev_eol - 1;
552 }
553 return cnt;
554}
555static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
557 struct emit_callback *ecbdata)
558{
559 int l1, l2, at;
560 unsigned ws_rule = ecbdata->ws_rule;
561 l1 = count_trailing_blank(mf1, ws_rule);
562 l2 = count_trailing_blank(mf2, ws_rule);
563 if (l2 <= l1) {
564 ecbdata->blank_at_eof_in_preimage = 0;
565 ecbdata->blank_at_eof_in_postimage = 0;
566 return;
567 }
568 at = count_lines(mf1->ptr, mf1->size);
569 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
570 at = count_lines(mf2->ptr, mf2->size);
572 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
573}
574static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
576 int first, const char *line, int len)
577{
578 int has_trailing_newline, has_trailing_carriage_return;
579 int nofirst;
580 FILE *file = o->file;
581 fputs(diff_line_prefix(o), file);
583 if (len == 0) {
585 has_trailing_newline = (first == '\n');
586 has_trailing_carriage_return = (!has_trailing_newline &&
587 (first == '\r'));
588 nofirst = has_trailing_newline || has_trailing_carriage_return;
589 } else {
590 has_trailing_newline = (len > 0 && line[len-1] == '\n');
591 if (has_trailing_newline)
592 len--;
593 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
594 if (has_trailing_carriage_return)
595 len--;
596 nofirst = 0;
597 }
598 if (len || !nofirst) {
600 fputs(set, file);
601 if (!nofirst)
602 fputc(first, file);
603 fwrite(line, len, 1, file);
604 fputs(reset, file);
605 }
606 if (has_trailing_carriage_return)
607 fputc('\r', file);
608 if (has_trailing_newline)
609 fputc('\n', file);
610}
611static void emit_line(struct diff_options *o, const char *set, const char *reset,
613 const char *line, int len)
614{
615 emit_line_0(o, set, reset, line[0], line+1, len-1);
616}
617enum diff_symbol {
619 DIFF_SYMBOL_BINARY_DIFF_HEADER,
620 DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
621 DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
622 DIFF_SYMBOL_BINARY_DIFF_BODY,
623 DIFF_SYMBOL_BINARY_DIFF_FOOTER,
624 DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
625 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
626 DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
627 DIFF_SYMBOL_STATS_LINE,
628 DIFF_SYMBOL_WORD_DIFF,
629 DIFF_SYMBOL_STAT_SEP,
630 DIFF_SYMBOL_SUMMARY,
631 DIFF_SYMBOL_SUBMODULE_ADD,
632 DIFF_SYMBOL_SUBMODULE_DEL,
633 DIFF_SYMBOL_SUBMODULE_UNTRACKED,
634 DIFF_SYMBOL_SUBMODULE_MODIFIED,
635 DIFF_SYMBOL_SUBMODULE_HEADER,
636 DIFF_SYMBOL_SUBMODULE_ERROR,
637 DIFF_SYMBOL_SUBMODULE_PIPETHROUGH,
638 DIFF_SYMBOL_REWRITE_DIFF,
639 DIFF_SYMBOL_BINARY_FILES,
640 DIFF_SYMBOL_HEADER,
641 DIFF_SYMBOL_FILEPAIR_PLUS,
642 DIFF_SYMBOL_FILEPAIR_MINUS,
643 DIFF_SYMBOL_WORDS_PORCELAIN,
644 DIFF_SYMBOL_WORDS,
645 DIFF_SYMBOL_CONTEXT,
646 DIFF_SYMBOL_CONTEXT_INCOMPLETE,
647 DIFF_SYMBOL_PLUS,
648 DIFF_SYMBOL_MINUS,
649 DIFF_SYMBOL_NO_LF_EOF,
650 DIFF_SYMBOL_CONTEXT_FRAGINFO,
651 DIFF_SYMBOL_CONTEXT_MARKER,
652 DIFF_SYMBOL_SEPARATOR
653};
654/*
655 * Flags for content lines:
656 * 0..12 are whitespace rules
657 * 13-15 are WSEH_NEW | WSEH_OLD | WSEH_CONTEXT
658 * 16 is marking if the line is blank at EOF
659 */
660#define DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF (1<<16)
661#define DIFF_SYMBOL_MOVED_LINE (1<<17)
662#define DIFF_SYMBOL_MOVED_LINE_ALT (1<<18)
663#define DIFF_SYMBOL_MOVED_LINE_UNINTERESTING (1<<19)
664#define DIFF_SYMBOL_CONTENT_WS_MASK (WSEH_NEW | WSEH_OLD | WSEH_CONTEXT | WS_RULE_MASK)
665/*
667 * This struct is used when we need to buffer the output of the diff output.
668 *
669 * NEEDSWORK: Instead of storing a copy of the line, add an offset pointer
670 * into the pre/post image file. This pointer could be a union with the
671 * line pointer. By storing an offset into the file instead of the literal line,
672 * we can decrease the memory footprint for the buffered output. At first we
673 * may want to only have indirection for the content lines, but we could also
674 * enhance the state for emitting prefabricated lines, e.g. the similarity
675 * score line or hunk/file headers would only need to store a number or path
676 * and then the output can be constructed later on depending on state.
677 */
678struct emitted_diff_symbol {
679 const char *line;
680 int len;
681 int flags;
682 enum diff_symbol s;
683};
684#define EMITTED_DIFF_SYMBOL_INIT {NULL}
685struct emitted_diff_symbols {
687 struct emitted_diff_symbol *buf;
688 int nr, alloc;
689};
690#define EMITTED_DIFF_SYMBOLS_INIT {NULL, 0, 0}
691static void append_emitted_diff_symbol(struct diff_options *o,
693 struct emitted_diff_symbol *e)
694{
695 struct emitted_diff_symbol *f;
696 ALLOC_GROW(o->emitted_symbols->buf,
698 o->emitted_symbols->nr + 1,
699 o->emitted_symbols->alloc);
700 f = &o->emitted_symbols->buf[o->emitted_symbols->nr++];
701 memcpy(f, e, sizeof(struct emitted_diff_symbol));
703 f->line = e->line ? xmemdupz(e->line, e->len) : NULL;
704}
705struct moved_entry {
707 struct hashmap_entry ent;
708 const struct emitted_diff_symbol *es;
709 struct moved_entry *next_line;
710};
711static int next_byte(const char **cp, const char **endp,
713 const struct diff_options *diffopt)
714{
715 int retval;
716 if (*cp > *endp)
718 return -1;
719 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_CHANGE)) {
721 while (*cp < *endp && isspace(**cp))
722 (*cp)++;
723 /*
724 * After skipping a couple of whitespaces, we still have to
725 * account for one space.
726 */
727 return (int)' ';
728 }
729 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE)) {
731 while (*cp < *endp && isspace(**cp))
732 (*cp)++;
733 /* return the first non-ws character via the usual below */
734 }
735 retval = (unsigned char)(**cp);
737 (*cp)++;
738 return retval;
739}
740static int moved_entry_cmp(const struct diff_options *diffopt,
742 const struct moved_entry *a,
743 const struct moved_entry *b,
744 const void *keydata)
745{
746 const char *ap = a->es->line, *ae = a->es->line + a->es->len;
747 const char *bp = b->es->line, *be = b->es->line + b->es->len;
748 if (!(diffopt->xdl_opts & XDF_WHITESPACE_FLAGS))
750 return a->es->len != b->es->len || memcmp(ap, bp, a->es->len);
751 if (DIFF_XDL_TST(diffopt, IGNORE_WHITESPACE_AT_EOL)) {
753 while (ae > ap && isspace(*ae))
754 ae--;
755 while (be > bp && isspace(*be))
756 be--;
757 }
758 while (1) {
760 int ca, cb;
761 ca = next_byte(&ap, &ae, diffopt);
762 cb = next_byte(&bp, &be, diffopt);
763 if (ca != cb)
764 return 1;
765 if (ca < 0)
766 return 0;
767 }
768}
769static unsigned get_string_hash(struct emitted_diff_symbol *es, struct diff_options *o)
771{
772 if (o->xdl_opts & XDF_WHITESPACE_FLAGS) {
773 static struct strbuf sb = STRBUF_INIT;
774 const char *ap = es->line, *ae = es->line + es->len;
775 int c;
776 strbuf_reset(&sb);
778 while (ae > ap && isspace(*ae))
779 ae--;
780 while ((c = next_byte(&ap, &ae, o)) > 0)
781 strbuf_addch(&sb, c);
782 return memhash(sb.buf, sb.len);
784 } else {
785 return memhash(es->line, es->len);
786 }
787}
788static struct moved_entry *prepare_entry(struct diff_options *o,
790 int line_no)
791{
792 struct moved_entry *ret = xmalloc(sizeof(*ret));
793 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[line_no];
794 ret->ent.hash = get_string_hash(l, o);
796 ret->es = l;
797 ret->next_line = NULL;
798 return ret;
800}
801static void add_lines_to_move_detection(struct diff_options *o,
803 struct hashmap *add_lines,
804 struct hashmap *del_lines)
805{
806 struct moved_entry *prev_line = NULL;
807 int n;
809 for (n = 0; n < o->emitted_symbols->nr; n++) {
810 struct hashmap *hm;
811 struct moved_entry *key;
812 switch (o->emitted_symbols->buf[n].s) {
814 case DIFF_SYMBOL_PLUS:
815 hm = add_lines;
816 break;
817 case DIFF_SYMBOL_MINUS:
818 hm = del_lines;
819 break;
820 default:
821 prev_line = NULL;
822 continue;
823 }
824 key = prepare_entry(o, n);
826 if (prev_line && prev_line->es->s == o->emitted_symbols->buf[n].s)
827 prev_line->next_line = key;
828 hashmap_add(hm, key);
830 prev_line = key;
831 }
832}
833static int shrink_potential_moved_blocks(struct moved_entry **pmb,
835 int pmb_nr)
836{
837 int lp, rp;
838 /* Shrink the set of potential block to the remaining running */
840 for (lp = 0, rp = pmb_nr - 1; lp <= rp;) {
841 while (lp < pmb_nr && pmb[lp])
842 lp++;
843 /* lp points at the first NULL now */
844 while (rp > -1 && !pmb[rp])
846 rp--;
847 /* rp points at the last non-NULL */
848 if (lp < pmb_nr && rp > -1 && lp < rp) {
850 pmb[lp] = pmb[rp];
851 pmb[rp] = NULL;
852 rp--;
853 lp++;
854 }
855 }
856 /* Remember the number of running sets */
858 return rp + 1;
859}
860/* Find blocks of moved code, delegate actual coloring decision to helper */
862static void mark_color_as_moved(struct diff_options *o,
863 struct hashmap *add_lines,
864 struct hashmap *del_lines)
865{
866 struct moved_entry **pmb = NULL; /* potentially moved blocks */
867 int pmb_nr = 0, pmb_alloc = 0;
868 int n, flipped_block = 1, block_length = 0;
869 for (n = 0; n < o->emitted_symbols->nr; n++) {
872 struct hashmap *hm = NULL;
873 struct moved_entry *key;
874 struct moved_entry *match = NULL;
875 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
876 int i;
877 switch (l->s) {
879 case DIFF_SYMBOL_PLUS:
880 hm = del_lines;
881 key = prepare_entry(o, n);
882 match = hashmap_get(hm, key, o);
883 free(key);
884 break;
885 case DIFF_SYMBOL_MINUS:
886 hm = add_lines;
887 key = prepare_entry(o, n);
888 match = hashmap_get(hm, key, o);
889 free(key);
890 break;
891 default:
892 flipped_block = 1;
893 }
894 if (!match) {
896 if (block_length < COLOR_MOVED_MIN_BLOCK_LENGTH &&
897 o->color_moved != COLOR_MOVED_PLAIN) {
898 for (i = 0; i < block_length + 1; i++) {
899 l = &o->emitted_symbols->buf[n - i];
900 l->flags &= ~DIFF_SYMBOL_MOVED_LINE;
901 }
902 }
903 pmb_nr = 0;
904 block_length = 0;
905 continue;
906 }
907 l->flags |= DIFF_SYMBOL_MOVED_LINE;
909 block_length++;
910 if (o->color_moved == COLOR_MOVED_PLAIN)
912 continue;
913 /* Check any potential block runs, advance each or nullify */
915 for (i = 0; i < pmb_nr; i++) {
916 struct moved_entry *p = pmb[i];
917 struct moved_entry *pnext = (p && p->next_line) ?
918 p->next_line : NULL;
919 if (pnext && !hm->cmpfn(o, pnext, match, NULL)) {
920 pmb[i] = p->next_line;
921 } else {
922 pmb[i] = NULL;
923 }
924 }
925 pmb_nr = shrink_potential_moved_blocks(pmb, pmb_nr);
927 if (pmb_nr == 0) {
929 /*
930 * The current line is the start of a new block.
931 * Setup the set of potential blocks.
932 */
933 for (; match; match = hashmap_get_next(hm, match)) {
934 ALLOC_GROW(pmb, pmb_nr + 1, pmb_alloc);
935 pmb[pmb_nr++] = match;
936 }
937 flipped_block = (flipped_block + 1) % 2;
939 }
940 if (flipped_block)
942 l->flags |= DIFF_SYMBOL_MOVED_LINE_ALT;
943 }
944 free(pmb);
946}
947#define DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK \
949 (DIFF_SYMBOL_MOVED_LINE | DIFF_SYMBOL_MOVED_LINE_ALT)
950static void dim_moved_lines(struct diff_options *o)
951{
952 int n;
953 for (n = 0; n < o->emitted_symbols->nr; n++) {
954 struct emitted_diff_symbol *prev = (n != 0) ?
955 &o->emitted_symbols->buf[n - 1] : NULL;
956 struct emitted_diff_symbol *l = &o->emitted_symbols->buf[n];
957 struct emitted_diff_symbol *next =
958 (n < o->emitted_symbols->nr - 1) ?
959 &o->emitted_symbols->buf[n + 1] : NULL;
960 /* Not a plus or minus line? */
962 if (l->s != DIFF_SYMBOL_PLUS && l->s != DIFF_SYMBOL_MINUS)
963 continue;
964 /* Not a moved line? */
966 if (!(l->flags & DIFF_SYMBOL_MOVED_LINE))
967 continue;
968 /*
970 * If prev or next are not a plus or minus line,
971 * pretend they don't exist
972 */
973 if (prev && prev->s != DIFF_SYMBOL_PLUS &&
974 prev->s != DIFF_SYMBOL_MINUS)
975 prev = NULL;
976 if (next && next->s != DIFF_SYMBOL_PLUS &&
977 next->s != DIFF_SYMBOL_MINUS)
978 next = NULL;
979 /* Inside a block? */
981 if ((prev &&
982 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
983 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK)) &&
984 (next &&
985 (next->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK) ==
986 (l->flags & DIFF_SYMBOL_MOVED_LINE_ZEBRA_MASK))) {
987 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
988 continue;
989 }
990 /* Check if we are at an interesting bound: */
992 if (prev && (prev->flags & DIFF_SYMBOL_MOVED_LINE) &&
993 (prev->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
994 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
995 continue;
996 if (next && (next->flags & DIFF_SYMBOL_MOVED_LINE) &&
997 (next->flags & DIFF_SYMBOL_MOVED_LINE_ALT) !=
998 (l->flags & DIFF_SYMBOL_MOVED_LINE_ALT))
999 continue;
1000 /*
1002 * The boundary to prev and next are not interesting,
1003 * so this line is not interesting as a whole
1004 */
1005 l->flags |= DIFF_SYMBOL_MOVED_LINE_UNINTERESTING;
1006 }
1007}
1008static void emit_line_ws_markup(struct diff_options *o,
1010 const char *set, const char *reset,
1011 const char *line, int len, char sign,
1012 unsigned ws_rule, int blank_at_eof)
1013{
1014 const char *ws = NULL;
1015 if (o->ws_error_highlight & ws_rule) {
1017 ws = diff_get_color_opt(o, DIFF_WHITESPACE);
1018 if (!*ws)
1019 ws = NULL;
1020 }
1021 if (!ws)
1023 emit_line_0(o, set, reset, sign, line, len);
1024 else if (blank_at_eof)
1025 /* Blank line at EOF - paint '+' as well */
1026 emit_line_0(o, ws, reset, sign, line, len);
1027 else {
1028 /* Emit just the prefix, then the rest. */
1029 emit_line_0(o, set, reset, sign, "", 0);
1030 ws_check_emit(line, len, ws_rule,
1031 o->file, set, reset, ws);
1032 }
1033}
1034static void emit_diff_symbol_from_struct(struct diff_options *o,
1036 struct emitted_diff_symbol *eds)
1037{
1038 static const char *nneof = " No newline at end of file\n";
1039 const char *context, *reset, *set, *meta, *fraginfo;
1040 struct strbuf sb = STRBUF_INIT;
1041 enum diff_symbol s = eds->s;
1043 const char *line = eds->line;
1044 int len = eds->len;
1045 unsigned flags = eds->flags;
1046 switch (s) {
1048 case DIFF_SYMBOL_NO_LF_EOF:
1049 context = diff_get_color_opt(o, DIFF_CONTEXT);
1050 reset = diff_get_color_opt(o, DIFF_RESET);
1051 putc('\n', o->file);
1052 emit_line_0(o, context, reset, '\\',
1053 nneof, strlen(nneof));
1054 break;
1055 case DIFF_SYMBOL_SUBMODULE_HEADER:
1056 case DIFF_SYMBOL_SUBMODULE_ERROR:
1057 case DIFF_SYMBOL_SUBMODULE_PIPETHROUGH:
1058 case DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES:
1059 case DIFF_SYMBOL_SUMMARY:
1060 case DIFF_SYMBOL_STATS_LINE:
1061 case DIFF_SYMBOL_BINARY_DIFF_BODY:
1062 case DIFF_SYMBOL_CONTEXT_FRAGINFO:
1063 emit_line(o, "", "", line, len);
1064 break;
1065 case DIFF_SYMBOL_CONTEXT_INCOMPLETE:
1066 case DIFF_SYMBOL_CONTEXT_MARKER:
1067 context = diff_get_color_opt(o, DIFF_CONTEXT);
1068 reset = diff_get_color_opt(o, DIFF_RESET);
1069 emit_line(o, context, reset, line, len);
1070 break;
1071 case DIFF_SYMBOL_SEPARATOR:
1072 fprintf(o->file, "%s%c",
1073 diff_line_prefix(o),
1074 o->line_termination);
1075 break;
1076 case DIFF_SYMBOL_CONTEXT:
1077 set = diff_get_color_opt(o, DIFF_CONTEXT);
1078 reset = diff_get_color_opt(o, DIFF_RESET);
1079 emit_line_ws_markup(o, set, reset, line, len, ' ',
1080 flags & (DIFF_SYMBOL_CONTENT_WS_MASK), 0);
1081 break;
1082 case DIFF_SYMBOL_PLUS:
1083 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1084 DIFF_SYMBOL_MOVED_LINE_ALT |
1085 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1086 case DIFF_SYMBOL_MOVED_LINE |
1087 DIFF_SYMBOL_MOVED_LINE_ALT |
1088 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1089 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT_DIM);
1090 break;
1091 case DIFF_SYMBOL_MOVED_LINE |
1092 DIFF_SYMBOL_MOVED_LINE_ALT:
1093 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_ALT);
1094 break;
1095 case DIFF_SYMBOL_MOVED_LINE |
1096 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1097 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED_DIM);
1098 break;
1099 case DIFF_SYMBOL_MOVED_LINE:
1100 set = diff_get_color_opt(o, DIFF_FILE_NEW_MOVED);
1101 break;
1102 default:
1103 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1104 }
1105 reset = diff_get_color_opt(o, DIFF_RESET);
1106 emit_line_ws_markup(o, set, reset, line, len, '+',
1107 flags & DIFF_SYMBOL_CONTENT_WS_MASK,
1108 flags & DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF);
1109 break;
1110 case DIFF_SYMBOL_MINUS:
1111 switch (flags & (DIFF_SYMBOL_MOVED_LINE |
1112 DIFF_SYMBOL_MOVED_LINE_ALT |
1113 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING)) {
1114 case DIFF_SYMBOL_MOVED_LINE |
1115 DIFF_SYMBOL_MOVED_LINE_ALT |
1116 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1117 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT_DIM);
1118 break;
1119 case DIFF_SYMBOL_MOVED_LINE |
1120 DIFF_SYMBOL_MOVED_LINE_ALT:
1121 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_ALT);
1122 break;
1123 case DIFF_SYMBOL_MOVED_LINE |
1124 DIFF_SYMBOL_MOVED_LINE_UNINTERESTING:
1125 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED_DIM);
1126 break;
1127 case DIFF_SYMBOL_MOVED_LINE:
1128 set = diff_get_color_opt(o, DIFF_FILE_OLD_MOVED);
1129 break;
1130 default:
1131 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1132 }
1133 reset = diff_get_color_opt(o, DIFF_RESET);
1134 emit_line_ws_markup(o, set, reset, line, len, '-',
1135 flags & DIFF_SYMBOL_CONTENT_WS_MASK, 0);
1136 break;
1137 case DIFF_SYMBOL_WORDS_PORCELAIN:
1138 context = diff_get_color_opt(o, DIFF_CONTEXT);
1139 reset = diff_get_color_opt(o, DIFF_RESET);
1140 emit_line(o, context, reset, line, len);
1141 fputs("~\n", o->file);
1142 break;
1143 case DIFF_SYMBOL_WORDS:
1144 context = diff_get_color_opt(o, DIFF_CONTEXT);
1145 reset = diff_get_color_opt(o, DIFF_RESET);
1146 /*
1147 * Skip the prefix character, if any. With
1148 * diff_suppress_blank_empty, there may be
1149 * none.
1150 */
1151 if (line[0] != '\n') {
1152 line++;
1153 len--;
1154 }
1155 emit_line(o, context, reset, line, len);
1156 break;
1157 case DIFF_SYMBOL_FILEPAIR_PLUS:
1158 meta = diff_get_color_opt(o, DIFF_METAINFO);
1159 reset = diff_get_color_opt(o, DIFF_RESET);
1160 fprintf(o->file, "%s%s+++ %s%s%s\n", diff_line_prefix(o), meta,
1161 line, reset,
1162 strchr(line, ' ') ? "\t" : "");
1163 break;
1164 case DIFF_SYMBOL_FILEPAIR_MINUS:
1165 meta = diff_get_color_opt(o, DIFF_METAINFO);
1166 reset = diff_get_color_opt(o, DIFF_RESET);
1167 fprintf(o->file, "%s%s--- %s%s%s\n", diff_line_prefix(o), meta,
1168 line, reset,
1169 strchr(line, ' ') ? "\t" : "");
1170 break;
1171 case DIFF_SYMBOL_BINARY_FILES:
1172 case DIFF_SYMBOL_HEADER:
1173 fprintf(o->file, "%s", line);
1174 break;
1175 case DIFF_SYMBOL_BINARY_DIFF_HEADER:
1176 fprintf(o->file, "%sGIT binary patch\n", diff_line_prefix(o));
1177 break;
1178 case DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA:
1179 fprintf(o->file, "%sdelta %s\n", diff_line_prefix(o), line);
1180 break;
1181 case DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL:
1182 fprintf(o->file, "%sliteral %s\n", diff_line_prefix(o), line);
1183 break;
1184 case DIFF_SYMBOL_BINARY_DIFF_FOOTER:
1185 fputs(diff_line_prefix(o), o->file);
1186 fputc('\n', o->file);
1187 break;
1188 case DIFF_SYMBOL_REWRITE_DIFF:
1189 fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
1190 reset = diff_get_color_opt(o, DIFF_RESET);
1191 emit_line(o, fraginfo, reset, line, len);
1192 break;
1193 case DIFF_SYMBOL_SUBMODULE_ADD:
1194 set = diff_get_color_opt(o, DIFF_FILE_NEW);
1195 reset = diff_get_color_opt(o, DIFF_RESET);
1196 emit_line(o, set, reset, line, len);
1197 break;
1198 case DIFF_SYMBOL_SUBMODULE_DEL:
1199 set = diff_get_color_opt(o, DIFF_FILE_OLD);
1200 reset = diff_get_color_opt(o, DIFF_RESET);
1201 emit_line(o, set, reset, line, len);
1202 break;
1203 case DIFF_SYMBOL_SUBMODULE_UNTRACKED:
1204 fprintf(o->file, "%sSubmodule %s contains untracked content\n",
1205 diff_line_prefix(o), line);
1206 break;
1207 case DIFF_SYMBOL_SUBMODULE_MODIFIED:
1208 fprintf(o->file, "%sSubmodule %s contains modified content\n",
1209 diff_line_prefix(o), line);
1210 break;
1211 case DIFF_SYMBOL_STATS_SUMMARY_NO_FILES:
1212 emit_line(o, "", "", " 0 files changed\n",
1213 strlen(" 0 files changed\n"));
1214 break;
1215 case DIFF_SYMBOL_STATS_SUMMARY_ABBREV:
1216 emit_line(o, "", "", " ...\n", strlen(" ...\n"));
1217 break;
1218 case DIFF_SYMBOL_WORD_DIFF:
1219 fprintf(o->file, "%.*s", len, line);
1220 break;
1221 case DIFF_SYMBOL_STAT_SEP:
1222 fputs(o->stat_sep, o->file);
1223 break;
1224 default:
1225 die("BUG: unknown diff symbol");
1226 }
1227 strbuf_release(&sb);
1228}
1229static void emit_diff_symbol(struct diff_options *o, enum diff_symbol s,
1231 const char *line, int len, unsigned flags)
1232{
1233 struct emitted_diff_symbol e = {line, len, flags, s};
1234 if (o->emitted_symbols)
1236 append_emitted_diff_symbol(o, &e);
1237 else
1238 emit_diff_symbol_from_struct(o, &e);
1239}
1240void diff_emit_submodule_del(struct diff_options *o, const char *line)
1242{
1243 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_DEL, line, strlen(line), 0);
1244}
1245void diff_emit_submodule_add(struct diff_options *o, const char *line)
1247{
1248 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ADD, line, strlen(line), 0);
1249}
1250void diff_emit_submodule_untracked(struct diff_options *o, const char *path)
1252{
1253 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_UNTRACKED,
1254 path, strlen(path), 0);
1255}
1256void diff_emit_submodule_modified(struct diff_options *o, const char *path)
1258{
1259 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_MODIFIED,
1260 path, strlen(path), 0);
1261}
1262void diff_emit_submodule_header(struct diff_options *o, const char *header)
1264{
1265 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_HEADER,
1266 header, strlen(header), 0);
1267}
1268void diff_emit_submodule_error(struct diff_options *o, const char *err)
1270{
1271 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_ERROR, err, strlen(err), 0);
1272}
1273void diff_emit_submodule_pipethrough(struct diff_options *o,
1275 const char *line, int len)
1276{
1277 emit_diff_symbol(o, DIFF_SYMBOL_SUBMODULE_PIPETHROUGH, line, len, 0);
1278}
1279static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
1281{
1282 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
1283 ecbdata->blank_at_eof_in_preimage &&
1284 ecbdata->blank_at_eof_in_postimage &&
1285 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
1286 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
1287 return 0;
1288 return ws_blank_line(line, len, ecbdata->ws_rule);
1289}
1290static void emit_add_line(const char *reset,
1292 struct emit_callback *ecbdata,
1293 const char *line, int len)
1294{
1295 unsigned flags = WSEH_NEW | ecbdata->ws_rule;
1296 if (new_blank_line_at_eof(ecbdata, line, len))
1297 flags |= DIFF_SYMBOL_CONTENT_BLANK_LINE_EOF;
1298 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_PLUS, line, len, flags);
1300}
1301static void emit_del_line(const char *reset,
1303 struct emit_callback *ecbdata,
1304 const char *line, int len)
1305{
1306 unsigned flags = WSEH_OLD | ecbdata->ws_rule;
1307 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_MINUS, line, len, flags);
1308}
1309static void emit_context_line(const char *reset,
1311 struct emit_callback *ecbdata,
1312 const char *line, int len)
1313{
1314 unsigned flags = WSEH_CONTEXT | ecbdata->ws_rule;
1315 emit_diff_symbol(ecbdata->opt, DIFF_SYMBOL_CONTEXT, line, len, flags);
1316}
1317static void emit_hunk_header(struct emit_callback *ecbdata,
1319 const char *line, int len)
1320{
1321 const char *context = diff_get_color(ecbdata->color_diff, DIFF_CONTEXT);
1322 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
1323 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
1324 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1325 static const char atat[2] = { '@', '@' };
1326 const char *cp, *ep;
1327 struct strbuf msgbuf = STRBUF_INIT;
1328 int org_len = len;
1329 int i = 1;
1330 /*
1332 * As a hunk header must begin with "@@ -<old>, +<new> @@",
1333 * it always is at least 10 bytes long.
1334 */
1335 if (len < 10 ||
1336 memcmp(line, atat, 2) ||
1337 !(ep = memmem(line + 2, len - 2, atat, 2))) {
1338 emit_diff_symbol(ecbdata->opt,
1339 DIFF_SYMBOL_CONTEXT_MARKER, line, len, 0);
1340 return;
1341 }
1342 ep += 2; /* skip over @@ */
1343 /* The hunk header in fraginfo color */
1345 strbuf_addstr(&msgbuf, frag);
1346 strbuf_add(&msgbuf, line, ep - line);
1347 strbuf_addstr(&msgbuf, reset);
1348 /*
1350 * trailing "\r\n"
1351 */
1352 for ( ; i < 3; i++)
1353 if (line[len - i] == '\r' || line[len - i] == '\n')
1354 len--;
1355 /* blank before the func header */
1357 for (cp = ep; ep - line < len; ep++)
1358 if (*ep != ' ' && *ep != '\t')
1359 break;
1360 if (ep != cp) {
1361 strbuf_addstr(&msgbuf, context);
1362 strbuf_add(&msgbuf, cp, ep - cp);
1363 strbuf_addstr(&msgbuf, reset);
1364 }
1365 if (ep < line + len) {
1367 strbuf_addstr(&msgbuf, func);
1368 strbuf_add(&msgbuf, ep, line + len - ep);
1369 strbuf_addstr(&msgbuf, reset);
1370 }
1371 strbuf_add(&msgbuf, line + len, org_len - len);
1373 strbuf_complete_line(&msgbuf);
1374 emit_diff_symbol(ecbdata->opt,
1375 DIFF_SYMBOL_CONTEXT_FRAGINFO, msgbuf.buf, msgbuf.len, 0);
1376 strbuf_release(&msgbuf);
1377}
1378static struct diff_tempfile *claim_diff_tempfile(void) {
1380 int i;
1381 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
1382 if (!diff_temp[i].name)
1383 return diff_temp + i;
1384 die("BUG: diff is failing to clean up its tempfiles");
1385}
1386static void remove_tempfile(void)
1388{
1389 int i;
1390 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
1391 if (is_tempfile_active(&diff_temp[i].tempfile))
1392 delete_tempfile(&diff_temp[i].tempfile);
1393 diff_temp[i].name = NULL;
1394 }
1395}
1396static void add_line_count(struct strbuf *out, int count)
1398{
1399 switch (count) {
1400 case 0:
1401 strbuf_addstr(out, "0,0");
1402 break;
1403 case 1:
1404 strbuf_addstr(out, "1");
1405 break;
1406 default:
1407 strbuf_addf(out, "1,%d", count);
1408 break;
1409 }
1410}
1411static void emit_rewrite_lines(struct emit_callback *ecb,
1413 int prefix, const char *data, int size)
1414{
1415 const char *endp = NULL;
1416 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
1417 while (0 < size) {
1419 int len;
1420 endp = memchr(data, '\n', size);
1422 len = endp ? (endp - data + 1) : size;
1423 if (prefix != '+') {
1424 ecb->lno_in_preimage++;
1425 emit_del_line(reset, ecb, data, len);
1426 } else {
1427 ecb->lno_in_postimage++;
1428 emit_add_line(reset, ecb, data, len);
1429 }
1430 size -= len;
1431 data += len;
1432 }
1433 if (!endp)
1434 emit_diff_symbol(ecb->opt, DIFF_SYMBOL_NO_LF_EOF, NULL, 0, 0);
1435}
1436static void emit_rewrite_diff(const char *name_a,
1438 const char *name_b,
1439 struct diff_filespec *one,
1440 struct diff_filespec *two,
1441 struct userdiff_driver *textconv_one,
1442 struct userdiff_driver *textconv_two,
1443 struct diff_options *o)
1444{
1445 int lc_a, lc_b;
1446 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
1447 const char *a_prefix, *b_prefix;
1448 char *data_one, *data_two;
1449 size_t size_one, size_two;
1450 struct emit_callback ecbdata;
1451 struct strbuf out = STRBUF_INIT;
1452 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
1454 a_prefix = o->b_prefix;
1455 b_prefix = o->a_prefix;
1456 } else {
1457 a_prefix = o->a_prefix;
1458 b_prefix = o->b_prefix;
1459 }
1460 name_a += (*name_a == '/');
1462 name_b += (*name_b == '/');
1463 strbuf_reset(&a_name);
1465 strbuf_reset(&b_name);
1466 quote_two_c_style(&a_name, a_prefix, name_a, 0);
1467 quote_two_c_style(&b_name, b_prefix, name_b, 0);
1468 size_one = fill_textconv(textconv_one, one, &data_one);
1470 size_two = fill_textconv(textconv_two, two, &data_two);
1471 memset(&ecbdata, 0, sizeof(ecbdata));
1473 ecbdata.color_diff = want_color(o->use_color);
1474 ecbdata.ws_rule = whitespace_rule(name_b);
1475 ecbdata.opt = o;
1476 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
1477 mmfile_t mf1, mf2;
1478 mf1.ptr = (char *)data_one;
1479 mf2.ptr = (char *)data_two;
1480 mf1.size = size_one;
1481 mf2.size = size_two;
1482 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1483 }
1484 ecbdata.lno_in_preimage = 1;
1485 ecbdata.lno_in_postimage = 1;
1486 lc_a = count_lines(data_one, size_one);
1488 lc_b = count_lines(data_two, size_two);
1489 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
1491 a_name.buf, a_name.len, 0);
1492 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
1493 b_name.buf, b_name.len, 0);
1494 strbuf_addstr(&out, "@@ -");
1496 if (!o->irreversible_delete)
1497 add_line_count(&out, lc_a);
1498 else
1499 strbuf_addstr(&out, "?,?");
1500 strbuf_addstr(&out, " +");
1501 add_line_count(&out, lc_b);
1502 strbuf_addstr(&out, " @@\n");
1503 emit_diff_symbol(o, DIFF_SYMBOL_REWRITE_DIFF, out.buf, out.len, 0);
1504 strbuf_release(&out);
1505 if (lc_a && !o->irreversible_delete)
1507 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
1508 if (lc_b)
1509 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
1510 if (textconv_one)
1511 free((char *)data_one);
1512 if (textconv_two)
1513 free((char *)data_two);
1514}
1515struct diff_words_buffer {
1517 mmfile_t text;
1518 long alloc;
1519 struct diff_words_orig {
1520 const char *begin, *end;
1521 } *orig;
1522 int orig_nr, orig_alloc;
1523};
1524static void diff_words_append(char *line, unsigned long len,
1526 struct diff_words_buffer *buffer)
1527{
1528 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
1529 line++;
1530 len--;
1531 memcpy(buffer->text.ptr + buffer->text.size, line, len);
1532 buffer->text.size += len;
1533 buffer->text.ptr[buffer->text.size] = '\0';
1534}
1535struct diff_words_style_elem {
1537 const char *prefix;
1538 const char *suffix;
1539 const char *color; /* NULL; filled in by the setup code if
1540 * color is enabled */
1541};
1542struct diff_words_style {
1544 enum diff_words_type type;
1545 struct diff_words_style_elem new, old, ctx;
1546 const char *newline;
1547};
1548static struct diff_words_style diff_words_styles[] = {
1550 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
1551 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
1552 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
1553};
1554struct diff_words_data {
1556 struct diff_words_buffer minus, plus;
1557 const char *current_plus;
1558 int last_minus;
1559 struct diff_options *opt;
1560 regex_t *word_regex;
1561 enum diff_words_type type;
1562 struct diff_words_style *style;
1563};
1564static int fn_out_diff_words_write_helper(struct diff_options *o,
1566 struct diff_words_style_elem *st_el,
1567 const char *newline,
1568 size_t count, const char *buf)
1569{
1570 int print = 0;
1571 struct strbuf sb = STRBUF_INIT;
1572 while (count) {
1574 char *p = memchr(buf, '\n', count);
1575 if (print)
1576 strbuf_addstr(&sb, diff_line_prefix(o));
1577 if (p != buf) {
1579 const char *reset = st_el->color && *st_el->color ?
1580 GIT_COLOR_RESET : NULL;
1581 if (st_el->color && *st_el->color)
1582 strbuf_addstr(&sb, st_el->color);
1583 strbuf_addstr(&sb, st_el->prefix);
1584 strbuf_add(&sb, buf, p ? p - buf : count);
1585 strbuf_addstr(&sb, st_el->suffix);
1586 if (reset)
1587 strbuf_addstr(&sb, reset);
1588 }
1589 if (!p)
1590 goto out;
1591 strbuf_addstr(&sb, newline);
1593 count -= p + 1 - buf;
1594 buf = p + 1;
1595 print = 1;
1596 if (count) {
1597 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1598 sb.buf, sb.len, 0);
1599 strbuf_reset(&sb);
1600 }
1601 }
1602out:
1604 if (sb.len)
1605 emit_diff_symbol(o, DIFF_SYMBOL_WORD_DIFF,
1606 sb.buf, sb.len, 0);
1607 strbuf_release(&sb);
1608 return 0;
1609}
1610/*
1612 * '--color-words' algorithm can be described as:
1613 *
1614 * 1. collect the minus/plus lines of a diff hunk, divided into
1615 * minus-lines and plus-lines;
1616 *
1617 * 2. break both minus-lines and plus-lines into words and
1618 * place them into two mmfile_t with one word for each line;
1619 *
1620 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
1621 *
1622 * And for the common parts of the both file, we output the plus side text.
1623 * diff_words->current_plus is used to trace the current position of the plus file
1624 * which printed. diff_words->last_minus is used to trace the last minus word
1625 * printed.
1626 *
1627 * For '--graph' to work with '--color-words', we need to output the graph prefix
1628 * on each line of color words output. Generally, there are two conditions on
1629 * which we should output the prefix.
1630 *
1631 * 1. diff_words->last_minus == 0 &&
1632 * diff_words->current_plus == diff_words->plus.text.ptr
1633 *
1634 * that is: the plus text must start as a new line, and if there is no minus
1635 * word printed, a graph prefix must be printed.
1636 *
1637 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
1638 * *(diff_words->current_plus - 1) == '\n'
1639 *
1640 * that is: a graph prefix must be printed following a '\n'
1641 */
1642static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
1643{
1644 if ((diff_words->last_minus == 0 &&
1645 diff_words->current_plus == diff_words->plus.text.ptr) ||
1646 (diff_words->current_plus > diff_words->plus.text.ptr &&
1647 *(diff_words->current_plus - 1) == '\n')) {
1648 return 1;
1649 } else {
1650 return 0;
1651 }
1652}
1653static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
1655{
1656 struct diff_words_data *diff_words = priv;
1657 struct diff_words_style *style = diff_words->style;
1658 int minus_first, minus_len, plus_first, plus_len;
1659 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
1660 struct diff_options *opt = diff_words->opt;
1661 const char *line_prefix;
1662 if (line[0] != '@' || parse_hunk_header(line, len,
1664 &minus_first, &minus_len, &plus_first, &plus_len))
1665 return;
1666 assert(opt);
1668 line_prefix = diff_line_prefix(opt);
1669 /* POSIX requires that first be decremented by one if len == 0... */
1671 if (minus_len) {
1672 minus_begin = diff_words->minus.orig[minus_first].begin;
1673 minus_end =
1674 diff_words->minus.orig[minus_first + minus_len - 1].end;
1675 } else
1676 minus_begin = minus_end =
1677 diff_words->minus.orig[minus_first].end;
1678 if (plus_len) {
1680 plus_begin = diff_words->plus.orig[plus_first].begin;
1681 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
1682 } else
1683 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
1684 if (color_words_output_graph_prefix(diff_words)) {
1686 fputs(line_prefix, diff_words->opt->file);
1687 }
1688 if (diff_words->current_plus != plus_begin) {
1689 fn_out_diff_words_write_helper(diff_words->opt,
1690 &style->ctx, style->newline,
1691 plus_begin - diff_words->current_plus,
1692 diff_words->current_plus);
1693 }
1694 if (minus_begin != minus_end) {
1695 fn_out_diff_words_write_helper(diff_words->opt,
1696 &style->old, style->newline,
1697 minus_end - minus_begin, minus_begin);
1698 }
1699 if (plus_begin != plus_end) {
1700 fn_out_diff_words_write_helper(diff_words->opt,
1701 &style->new, style->newline,
1702 plus_end - plus_begin, plus_begin);
1703 }
1704 diff_words->current_plus = plus_end;
1706 diff_words->last_minus = minus_first;
1707}
1708/* This function starts looking at *begin, and returns 0 iff a word was found. */
1710static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
1711 int *begin, int *end)
1712{
1713 if (word_regex && *begin < buffer->size) {
1714 regmatch_t match[1];
1715 if (!regexec_buf(word_regex, buffer->ptr + *begin,
1716 buffer->size - *begin, 1, match, 0)) {
1717 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
1718 '\n', match[0].rm_eo - match[0].rm_so);
1719 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
1720 *begin += match[0].rm_so;
1721 return *begin >= *end;
1722 }
1723 return -1;
1724 }
1725 /* find the next word */
1727 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
1728 (*begin)++;
1729 if (*begin >= buffer->size)
1730 return -1;
1731 /* find the end of the word */
1733 *end = *begin + 1;
1734 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
1735 (*end)++;
1736 return 0;
1738}
1739/*
1741 * This function splits the words in buffer->text, stores the list with
1742 * newline separator into out, and saves the offsets of the original words
1743 * in buffer->orig.
1744 */
1745static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
1746 regex_t *word_regex)
1747{
1748 int i, j;
1749 long alloc = 0;
1750 out->size = 0;
1752 out->ptr = NULL;
1753 /* fake an empty "0th" word */
1755 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
1756 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
1757 buffer->orig_nr = 1;
1758 for (i = 0; i < buffer->text.size; i++) {
1760 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
1761 return;
1762 /* store original boundaries */
1764 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
1765 buffer->orig_alloc);
1766 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
1767 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
1768 buffer->orig_nr++;
1769 /* store one word */
1771 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
1772 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
1773 out->ptr[out->size + j - i] = '\n';
1774 out->size += j - i + 1;
1775 i = j - 1;
1777 }
1778}
1779/* this executes the word diff on the accumulated buffers */
1781static void diff_words_show(struct diff_words_data *diff_words)
1782{
1783 xpparam_t xpp;
1784 xdemitconf_t xecfg;
1785 mmfile_t minus, plus;
1786 struct diff_words_style *style = diff_words->style;
1787 struct diff_options *opt = diff_words->opt;
1789 const char *line_prefix;
1790 assert(opt);
1792 line_prefix = diff_line_prefix(opt);
1793 /* special case: only removal */
1795 if (!diff_words->plus.text.size) {
1796 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1797 line_prefix, strlen(line_prefix), 0);
1798 fn_out_diff_words_write_helper(diff_words->opt,
1799 &style->old, style->newline,
1800 diff_words->minus.text.size,
1801 diff_words->minus.text.ptr);
1802 diff_words->minus.text.size = 0;
1803 return;
1804 }
1805 diff_words->current_plus = diff_words->plus.text.ptr;
1807 diff_words->last_minus = 0;
1808 memset(&xpp, 0, sizeof(xpp));
1810 memset(&xecfg, 0, sizeof(xecfg));
1811 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
1812 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
1813 xpp.flags = 0;
1814 /* as only the hunk header will be parsed, we need a 0-context */
1815 xecfg.ctxlen = 0;
1816 if (xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
1817 &xpp, &xecfg))
1818 die("unable to generate word diff");
1819 free(minus.ptr);
1820 free(plus.ptr);
1821 if (diff_words->current_plus != diff_words->plus.text.ptr +
1822 diff_words->plus.text.size) {
1823 if (color_words_output_graph_prefix(diff_words))
1824 emit_diff_symbol(diff_words->opt, DIFF_SYMBOL_WORD_DIFF,
1825 line_prefix, strlen(line_prefix), 0);
1826 fn_out_diff_words_write_helper(diff_words->opt,
1827 &style->ctx, style->newline,
1828 diff_words->plus.text.ptr + diff_words->plus.text.size
1829 - diff_words->current_plus, diff_words->current_plus);
1830 }
1831 diff_words->minus.text.size = diff_words->plus.text.size = 0;
1832}
1833/* In "color-words" mode, show word-diff of words accumulated in the buffer */
1835static void diff_words_flush(struct emit_callback *ecbdata)
1836{
1837 struct diff_options *wo = ecbdata->diff_words->opt;
1838 if (ecbdata->diff_words->minus.text.size ||
1840 ecbdata->diff_words->plus.text.size)
1841 diff_words_show(ecbdata->diff_words);
1842 if (wo->emitted_symbols) {
1844 struct diff_options *o = ecbdata->opt;
1845 struct emitted_diff_symbols *wol = wo->emitted_symbols;
1846 int i;
1847 /*
1849 * NEEDSWORK:
1850 * Instead of appending each, concat all words to a line?
1851 */
1852 for (i = 0; i < wol->nr; i++)
1853 append_emitted_diff_symbol(o, &wol->buf[i]);
1854 for (i = 0; i < wol->nr; i++)
1856 free((void *)wol->buf[i].line);
1857 wol->nr = 0;
1859 }
1860}
1861static void diff_filespec_load_driver(struct diff_filespec *one)
1863{
1864 /* Use already-loaded driver */
1865 if (one->driver)
1866 return;
1867 if (S_ISREG(one->mode))
1869 one->driver = userdiff_find_by_path(one->path);
1870 /* Fallback to default settings */
1872 if (!one->driver)
1873 one->driver = userdiff_find_by_name("default");
1874}
1875static const char *userdiff_word_regex(struct diff_filespec *one)
1877{
1878 diff_filespec_load_driver(one);
1879 return one->driver->word_regex;
1880}
1881static void init_diff_words_data(struct emit_callback *ecbdata,
1883 struct diff_options *orig_opts,
1884 struct diff_filespec *one,
1885 struct diff_filespec *two)
1886{
1887 int i;
1888 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1889 memcpy(o, orig_opts, sizeof(struct diff_options));
1890 ecbdata->diff_words =
1892 xcalloc(1, sizeof(struct diff_words_data));
1893 ecbdata->diff_words->type = o->word_diff;
1894 ecbdata->diff_words->opt = o;
1895 if (orig_opts->emitted_symbols)
1897 o->emitted_symbols =
1898 xcalloc(1, sizeof(struct emitted_diff_symbols));
1899 if (!o->word_regex)
1901 o->word_regex = userdiff_word_regex(one);
1902 if (!o->word_regex)
1903 o->word_regex = userdiff_word_regex(two);
1904 if (!o->word_regex)
1905 o->word_regex = diff_word_regex_cfg;
1906 if (o->word_regex) {
1907 ecbdata->diff_words->word_regex = (regex_t *)
1908 xmalloc(sizeof(regex_t));
1909 if (regcomp(ecbdata->diff_words->word_regex,
1910 o->word_regex,
1911 REG_EXTENDED | REG_NEWLINE))
1912 die ("Invalid regular expression: %s",
1913 o->word_regex);
1914 }
1915 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1916 if (o->word_diff == diff_words_styles[i].type) {
1917 ecbdata->diff_words->style =
1918 &diff_words_styles[i];
1919 break;
1920 }
1921 }
1922 if (want_color(o->use_color)) {
1923 struct diff_words_style *st = ecbdata->diff_words->style;
1924 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1925 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1926 st->ctx.color = diff_get_color_opt(o, DIFF_CONTEXT);
1927 }
1928}
1929static void free_diff_words_data(struct emit_callback *ecbdata)
1931{
1932 if (ecbdata->diff_words) {
1933 diff_words_flush(ecbdata);
1934 free (ecbdata->diff_words->opt->emitted_symbols);
1935 free (ecbdata->diff_words->opt);
1936 free (ecbdata->diff_words->minus.text.ptr);
1937 free (ecbdata->diff_words->minus.orig);
1938 free (ecbdata->diff_words->plus.text.ptr);
1939 free (ecbdata->diff_words->plus.orig);
1940 if (ecbdata->diff_words->word_regex) {
1941 regfree(ecbdata->diff_words->word_regex);
1942 free(ecbdata->diff_words->word_regex);
1943 }
1944 FREE_AND_NULL(ecbdata->diff_words);
1945 }
1946}
1947const char *diff_get_color(int diff_use_color, enum color_diff ix)
1949{
1950 if (want_color(diff_use_color))
1951 return diff_colors[ix];
1952 return "";
1953}
1954const char *diff_line_prefix(struct diff_options *opt)
1956{
1957 struct strbuf *msgbuf;
1958 if (!opt->output_prefix)
1959 return "";
1960 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
1962 return msgbuf->buf;
1963}
1964static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1966{
1967 const char *cp;
1968 unsigned long allot;
1969 size_t l = len;
1970 if (ecb->truncate)
1972 return ecb->truncate(line, len);
1973 cp = line;
1974 allot = l;
1975 while (0 < l) {
1976 (void) utf8_width(&cp, &l);
1977 if (!cp)
1978 break; /* truncated in the middle? */
1979 }
1980 return allot - l;
1981}
1982static void find_lno(const char *line, struct emit_callback *ecbdata)
1984{
1985 const char *p;
1986 ecbdata->lno_in_preimage = 0;
1987 ecbdata->lno_in_postimage = 0;
1988 p = strchr(line, '-');
1989 if (!p)
1990 return; /* cannot happen */
1991 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1992 p = strchr(p, '+');
1993 if (!p)
1994 return; /* cannot happen */
1995 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1996}
1997static void fn_out_consume(void *priv, char *line, unsigned long len)
1999{
2000 struct emit_callback *ecbdata = priv;
2001 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
2002 struct diff_options *o = ecbdata->opt;
2003 o->found_changes = 1;
2005 if (ecbdata->header) {
2007 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
2008 ecbdata->header->buf, ecbdata->header->len, 0);
2009 strbuf_reset(ecbdata->header);
2010 ecbdata->header = NULL;
2011 }
2012 if (ecbdata->label_path[0]) {
2014 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_MINUS,
2015 ecbdata->label_path[0],
2016 strlen(ecbdata->label_path[0]), 0);
2017 emit_diff_symbol(o, DIFF_SYMBOL_FILEPAIR_PLUS,
2018 ecbdata->label_path[1],
2019 strlen(ecbdata->label_path[1]), 0);
2020 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
2021 }
2022 if (diff_suppress_blank_empty
2024 && len == 2 && line[0] == ' ' && line[1] == '\n') {
2025 line[0] = '\n';
2026 len = 1;
2027 }
2028 if (line[0] == '@') {
2030 if (ecbdata->diff_words)
2031 diff_words_flush(ecbdata);
2032 len = sane_truncate_line(ecbdata, line, len);
2033 find_lno(line, ecbdata);
2034 emit_hunk_header(ecbdata, line, len);
2035 return;
2036 }
2037 if (ecbdata->diff_words) {
2039 enum diff_symbol s =
2040 ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN ?
2041 DIFF_SYMBOL_WORDS_PORCELAIN : DIFF_SYMBOL_WORDS;
2042 if (line[0] == '-') {
2043 diff_words_append(line, len,
2044 &ecbdata->diff_words->minus);
2045 return;
2046 } else if (line[0] == '+') {
2047 diff_words_append(line, len,
2048 &ecbdata->diff_words->plus);
2049 return;
2050 } else if (starts_with(line, "\\ ")) {
2051 /*
2052 * Eat the "no newline at eof" marker as if we
2053 * saw a "+" or "-" line with nothing on it,
2054 * and return without diff_words_flush() to
2055 * defer processing. If this is the end of
2056 * preimage, more "+" lines may come after it.
2057 */
2058 return;
2059 }
2060 diff_words_flush(ecbdata);
2061 emit_diff_symbol(o, s, line, len, 0);
2062 return;
2063 }
2064 switch (line[0]) {
2066 case '+':
2067 ecbdata->lno_in_postimage++;
2068 emit_add_line(reset, ecbdata, line + 1, len - 1);
2069 break;
2070 case '-':
2071 ecbdata->lno_in_preimage++;
2072 emit_del_line(reset, ecbdata, line + 1, len - 1);
2073 break;
2074 case ' ':
2075 ecbdata->lno_in_postimage++;
2076 ecbdata->lno_in_preimage++;
2077 emit_context_line(reset, ecbdata, line + 1, len - 1);
2078 break;
2079 default:
2080 /* incomplete line at the end */
2081 ecbdata->lno_in_preimage++;
2082 emit_diff_symbol(o, DIFF_SYMBOL_CONTEXT_INCOMPLETE,
2083 line, len, 0);
2084 break;
2085 }
2086}
2087static char *pprint_rename(const char *a, const char *b)
2089{
2090 const char *old = a;
2091 const char *new = b;
2092 struct strbuf name = STRBUF_INIT;
2093 int pfx_length, sfx_length;
2094 int pfx_adjust_for_slash;
2095 int len_a = strlen(a);
2096 int len_b = strlen(b);
2097 int a_midlen, b_midlen;
2098 int qlen_a = quote_c_style(a, NULL, NULL, 0);
2099 int qlen_b = quote_c_style(b, NULL, NULL, 0);
2100 if (qlen_a || qlen_b) {
2102 quote_c_style(a, &name, NULL, 0);
2103 strbuf_addstr(&name, " => ");
2104 quote_c_style(b, &name, NULL, 0);
2105 return strbuf_detach(&name, NULL);
2106 }
2107 /* Find common prefix */
2109 pfx_length = 0;
2110 while (*old && *new && *old == *new) {
2111 if (*old == '/')
2112 pfx_length = old - a + 1;
2113 old++;
2114 new++;
2115 }
2116 /* Find common suffix */
2118 old = a + len_a;
2119 new = b + len_b;
2120 sfx_length = 0;
2121 /*
2122 * If there is a common prefix, it must end in a slash. In
2123 * that case we let this loop run 1 into the prefix to see the
2124 * same slash.
2125 *
2126 * If there is no common prefix, we cannot do this as it would
2127 * underrun the input strings.
2128 */
2129 pfx_adjust_for_slash = (pfx_length ? 1 : 0);
2130 while (a + pfx_length - pfx_adjust_for_slash <= old &&
2131 b + pfx_length - pfx_adjust_for_slash <= new &&
2132 *old == *new) {
2133 if (*old == '/')
2134 sfx_length = len_a - (old - a);
2135 old--;
2136 new--;
2137 }
2138 /*
2140 * pfx{mid-a => mid-b}sfx
2141 * {pfx-a => pfx-b}sfx
2142 * pfx{sfx-a => sfx-b}
2143 * name-a => name-b
2144 */
2145 a_midlen = len_a - pfx_length - sfx_length;
2146 b_midlen = len_b - pfx_length - sfx_length;
2147 if (a_midlen < 0)
2148 a_midlen = 0;
2149 if (b_midlen < 0)
2150 b_midlen = 0;
2151 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
2153 if (pfx_length + sfx_length) {
2154 strbuf_add(&name, a, pfx_length);
2155 strbuf_addch(&name, '{');
2156 }
2157 strbuf_add(&name, a + pfx_length, a_midlen);
2158 strbuf_addstr(&name, " => ");
2159 strbuf_add(&name, b + pfx_length, b_midlen);
2160 if (pfx_length + sfx_length) {
2161 strbuf_addch(&name, '}');
2162 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
2163 }
2164 return strbuf_detach(&name, NULL);
2165}
2166struct diffstat_t {
2168 int nr;
2169 int alloc;
2170 struct diffstat_file {
2171 char *from_name;
2172 char *name;
2173 char *print_name;
2174 unsigned is_unmerged:1;
2175 unsigned is_binary:1;
2176 unsigned is_renamed:1;
2177 unsigned is_interesting:1;
2178 uintmax_t added, deleted;
2179 } **files;
2180};
2181static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
2183 const char *name_a,
2184 const char *name_b)
2185{
2186 struct diffstat_file *x;
2187 x = xcalloc(1, sizeof(*x));
2188 ALLOC_GROW(diffstat->files, diffstat->nr + 1, diffstat->alloc);
2189 diffstat->files[diffstat->nr++] = x;
2190 if (name_b) {
2191 x->from_name = xstrdup(name_a);
2192 x->name = xstrdup(name_b);
2193 x->is_renamed = 1;
2194 }
2195 else {
2196 x->from_name = NULL;
2197 x->name = xstrdup(name_a);
2198 }
2199 return x;
2200}
2201static void diffstat_consume(void *priv, char *line, unsigned long len)
2203{
2204 struct diffstat_t *diffstat = priv;
2205 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
2206 if (line[0] == '+')
2208 x->added++;
2209 else if (line[0] == '-')
2210 x->deleted++;
2211}
2212const char mime_boundary_leader[] = "------------";
2214static int scale_linear(int it, int width, int max_change)
2216{
2217 if (!it)
2218 return 0;
2219 /*
2220 * make sure that at least one '-' or '+' is printed if
2221 * there is any change to this path. The easiest way is to
2222 * scale linearly as if the alloted width is one column shorter
2223 * than it is, and then add 1 to the result.
2224 */
2225 return 1 + (it * (width - 1) / max_change);
2226}
2227static void show_graph(struct strbuf *out, char ch, int cnt,
2229 const char *set, const char *reset)
2230{
2231 if (cnt <= 0)
2232 return;
2233 strbuf_addstr(out, set);
2234 strbuf_addchars(out, ch, cnt);
2235 strbuf_addstr(out, reset);
2236}
2237static void fill_print_name(struct diffstat_file *file)
2239{
2240 char *pname;
2241 if (file->print_name)
2243 return;
2244 if (!file->is_renamed) {
2246 struct strbuf buf = STRBUF_INIT;
2247 if (quote_c_style(file->name, &buf, NULL, 0)) {
2248 pname = strbuf_detach(&buf, NULL);
2249 } else {
2250 pname = file->name;
2251 strbuf_release(&buf);
2252 }
2253 } else {
2254 pname = pprint_rename(file->from_name, file->name);
2255 }
2256 file->print_name = pname;
2257}
2258static void print_stat_summary_inserts_deletes(struct diff_options *options,
2260 int files, int insertions, int deletions)
2261{
2262 struct strbuf sb = STRBUF_INIT;
2263 if (!files) {
2265 assert(insertions == 0 && deletions == 0);
2266 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_NO_FILES,
2267 NULL, 0, 0);
2268 return;
2269 }
2270 strbuf_addf(&sb,
2272 (files == 1) ? " %d file changed" : " %d files changed",
2273 files);
2274 /*
2276 * For binary diff, the caller may want to print "x files
2277 * changed" with insertions == 0 && deletions == 0.
2278 *
2279 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
2280 * is probably less confusing (i.e skip over "2 files changed
2281 * but nothing about added/removed lines? Is this a bug in Git?").
2282 */
2283 if (insertions || deletions == 0) {
2284 strbuf_addf(&sb,
2285 (insertions == 1) ? ", %d insertion(+)" : ", %d insertions(+)",
2286 insertions);
2287 }
2288 if (deletions || insertions == 0) {
2290 strbuf_addf(&sb,
2291 (deletions == 1) ? ", %d deletion(-)" : ", %d deletions(-)",
2292 deletions);
2293 }
2294 strbuf_addch(&sb, '\n');
2295 emit_diff_symbol(options, DIFF_SYMBOL_STATS_SUMMARY_INSERTS_DELETES,
2296 sb.buf, sb.len, 0);
2297 strbuf_release(&sb);
2298}
2299void print_stat_summary(FILE *fp, int files,
2301 int insertions, int deletions)
2302{
2303 struct diff_options o;
2304 memset(&o, 0, sizeof(o));
2305 o.file = fp;
2306 print_stat_summary_inserts_deletes(&o, files, insertions, deletions);
2308}
2309static void show_stats(struct diffstat_t *data, struct diff_options *options)
2311{
2312 int i, len, add, del, adds = 0, dels = 0;
2313 uintmax_t max_change = 0, max_len = 0;
2314 int total_files = data->nr, count;
2315 int width, name_width, graph_width, number_width = 0, bin_width = 0;
2316 const char *reset, *add_c, *del_c;
2317 int extra_shown = 0;
2318 const char *line_prefix = diff_line_prefix(options);
2319 struct strbuf out = STRBUF_INIT;
2320 if (data->nr == 0)
2322 return;
2323 count = options->stat_count ? options->stat_count : data->nr;
2325 reset = diff_get_color_opt(options, DIFF_RESET);
2327 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
2328 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
2329 /*
2331 * Find the longest filename and max number of changes
2332 */
2333 for (i = 0; (i < count) && (i < data->nr); i++) {
2334 struct diffstat_file *file = data->files[i];
2335 uintmax_t change = file->added + file->deleted;
2336 if (!file->is_interesting && (change == 0)) {
2338 count++; /* not shown == room for one more */
2339 continue;
2340 }
2341 fill_print_name(file);
2342 len = strlen(file->print_name);
2343 if (max_len < len)
2344 max_len = len;
2345 if (file->is_unmerged) {
2347 /* "Unmerged" is 8 characters */
2348 bin_width = bin_width < 8 ? 8 : bin_width;
2349 continue;
2350 }
2351 if (file->is_binary) {
2352 /* "Bin XXX -> YYY bytes" */
2353 int w = 14 + decimal_width(file->added)
2354 + decimal_width(file->deleted);
2355 bin_width = bin_width < w ? w : bin_width;
2356 /* Display change counts aligned with "Bin" */
2357 number_width = 3;
2358 continue;
2359 }
2360 if (max_change < change)
2362 max_change = change;
2363 }
2364 count = i; /* where we can stop scanning in data->files[] */
2365 /*
2367 * We have width = stat_width or term_columns() columns total.
2368 * We want a maximum of min(max_len, stat_name_width) for the name part.
2369 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
2370 * We also need 1 for " " and 4 + decimal_width(max_change)
2371 * for " | NNNN " and one the empty column at the end, altogether
2372 * 6 + decimal_width(max_change).
2373 *
2374 * If there's not enough space, we will use the smaller of
2375 * stat_name_width (if set) and 5/8*width for the filename,
2376 * and the rest for constant elements + graph part, but no more
2377 * than stat_graph_width for the graph part.
2378 * (5/8 gives 50 for filename and 30 for the constant parts + graph
2379 * for the standard terminal size).
2380 *
2381 * In other words: stat_width limits the maximum width, and
2382 * stat_name_width fixes the maximum width of the filename,
2383 * and is also used to divide available columns if there
2384 * aren't enough.
2385 *
2386 * Binary files are displayed with "Bin XXX -> YYY bytes"
2387 * instead of the change count and graph. This part is treated
2388 * similarly to the graph part, except that it is not
2389 * "scaled". If total width is too small to accommodate the
2390 * guaranteed minimum width of the filename part and the
2391 * separators and this message, this message will "overflow"
2392 * making the line longer than the maximum width.
2393 */
2394 if (options->stat_width == -1)
2396 width = term_columns() - strlen(line_prefix);
2397 else
2398 width = options->stat_width ? options->stat_width : 80;
2399 number_width = decimal_width(max_change) > number_width ?
2400 decimal_width(max_change) : number_width;
2401 if (options->stat_graph_width == -1)
2403 options->stat_graph_width = diff_stat_graph_width;
2404 /*
2406 * Guarantee 3/8*16==6 for the graph part
2407 * and 5/8*16==10 for the filename part
2408 */
2409 if (width < 16 + 6 + number_width)
2410 width = 16 + 6 + number_width;
2411 /*
2413 * First assign sizes that are wanted, ignoring available width.
2414 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
2415 * starting from "XXX" should fit in graph_width.
2416 */
2417 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
2418 if (options->stat_graph_width &&
2419 options->stat_graph_width < graph_width)
2420 graph_width = options->stat_graph_width;
2421 name_width = (options->stat_name_width > 0 &&
2423 options->stat_name_width < max_len) ?
2424 options->stat_name_width : max_len;
2425 /*
2427 * Adjust adjustable widths not to exceed maximum width
2428 */
2429 if (name_width + number_width + 6 + graph_width > width) {
2430 if (graph_width > width * 3/8 - number_width - 6) {
2431 graph_width = width * 3/8 - number_width - 6;
2432 if (graph_width < 6)
2433 graph_width = 6;
2434 }
2435 if (options->stat_graph_width &&
2437 graph_width > options->stat_graph_width)
2438 graph_width = options->stat_graph_width;
2439 if (name_width > width - number_width - 6 - graph_width)
2440 name_width = width - number_width - 6 - graph_width;
2441 else
2442 graph_width = width - number_width - 6 - name_width;
2443 }
2444 /*
2446 * From here name_width is the width of the name area,
2447 * and graph_width is the width of the graph area.
2448 * max_change is used to scale graph properly.
2449 */
2450 for (i = 0; i < count; i++) {
2451 const char *prefix = "";
2452 struct diffstat_file *file = data->files[i];
2453 char *name = file->print_name;
2454 uintmax_t added = file->added;
2455 uintmax_t deleted = file->deleted;
2456 int name_len;
2457 if (!file->is_interesting && (added + deleted == 0))
2459 continue;
2460 /*
2462 * "scale" the filename
2463 */
2464 len = name_width;
2465 name_len = strlen(name);
2466 if (name_width < name_len) {
2467 char *slash;
2468 prefix = "...";
2469 len -= 3;
2470 name += name_len - len;
2471 slash = strchr(name, '/');
2472 if (slash)
2473 name = slash;
2474 }
2475 if (file->is_binary) {
2477 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2478 strbuf_addf(&out, " %*s", number_width, "Bin");
2479 if (!added && !deleted) {
2480 strbuf_addch(&out, '\n');
2481 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2482 out.buf, out.len, 0);
2483 strbuf_reset(&out);
2484 continue;
2485 }
2486 strbuf_addf(&out, " %s%"PRIuMAX"%s",
2487 del_c, deleted, reset);
2488 strbuf_addstr(&out, " -> ");
2489 strbuf_addf(&out, "%s%"PRIuMAX"%s",
2490 add_c, added, reset);
2491 strbuf_addstr(&out, " bytes\n");
2492 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2493 out.buf, out.len, 0);
2494 strbuf_reset(&out);
2495 continue;
2496 }
2497 else if (file->is_unmerged) {
2498 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2499 strbuf_addstr(&out, " Unmerged\n");
2500 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2501 out.buf, out.len, 0);
2502 strbuf_reset(&out);
2503 continue;
2504 }
2505 /*
2507 * scale the add/delete
2508 */
2509 add = added;
2510 del = deleted;
2511 if (graph_width <= max_change) {
2513 int total = scale_linear(add + del, graph_width, max_change);
2514 if (total < 2 && add && del)
2515 /* width >= 2 due to the sanity check */
2516 total = 2;
2517 if (add < del) {
2518 add = scale_linear(add, graph_width, max_change);
2519 del = total - add;
2520 } else {
2521 del = scale_linear(del, graph_width, max_change);
2522 add = total - del;
2523 }
2524 }
2525 strbuf_addf(&out, " %s%-*s |", prefix, len, name);
2526 strbuf_addf(&out, " %*"PRIuMAX"%s",
2527 number_width, added + deleted,
2528 added + deleted ? " " : "");
2529 show_graph(&out, '+', add, add_c, reset);
2530 show_graph(&out, '-', del, del_c, reset);
2531 strbuf_addch(&out, '\n');
2532 emit_diff_symbol(options, DIFF_SYMBOL_STATS_LINE,
2533 out.buf, out.len, 0);
2534 strbuf_reset(&out);
2535 }
2536 for (i = 0; i < data->nr; i++) {
2538 struct diffstat_file *file = data->files[i];
2539 uintmax_t added = file->added;
2540 uintmax_t deleted = file->deleted;
2541 if (file->is_unmerged ||
2543 (!file->is_interesting && (added + deleted == 0))) {
2544 total_files--;
2545 continue;
2546 }
2547 if (!file->is_binary) {
2549 adds += added;
2550 dels += deleted;
2551 }
2552 if (i < count)
2553 continue;
2554 if (!extra_shown)
2555 emit_diff_symbol(options,
2556 DIFF_SYMBOL_STATS_SUMMARY_ABBREV,
2557 NULL, 0, 0);
2558 extra_shown = 1;
2559 }
2560 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2562}
2563static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
2565{
2566 int i, adds = 0, dels = 0, total_files = data->nr;
2567 if (data->nr == 0)
2569 return;
2570 for (i = 0; i < data->nr; i++) {
2572 int added = data->files[i]->added;
2573 int deleted = data->files[i]->deleted;
2574 if (data->files[i]->is_unmerged ||
2576 (!data->files[i]->is_interesting && (added + deleted == 0))) {
2577 total_files--;
2578 } else if (!data->files[i]->is_binary) { /* don't count bytes */
2579 adds += added;
2580 dels += deleted;
2581 }
2582 }
2583 print_stat_summary_inserts_deletes(options, total_files, adds, dels);
2584}
2585static void show_numstat(struct diffstat_t *data, struct diff_options *options)
2587{
2588 int i;
2589 if (data->nr == 0)
2591 return;
2592 for (i = 0; i < data->nr; i++) {
2594 struct diffstat_file *file = data->files[i];
2595 fprintf(options->file, "%s", diff_line_prefix(options));
2597 if (file->is_binary)
2599 fprintf(options->file, "-\t-\t");
2600 else
2601 fprintf(options->file,
2602 "%"PRIuMAX"\t%"PRIuMAX"\t",
2603 file->added, file->deleted);
2604 if (options->line_termination) {
2605 fill_print_name(file);
2606 if (!file->is_renamed)
2607 write_name_quoted(file->name, options->file,
2608 options->line_termination);
2609 else {
2610 fputs(file->print_name, options->file);
2611 putc(options->line_termination, options->file);
2612 }
2613 } else {
2614 if (file->is_renamed) {
2615 putc('\0', options->file);
2616 write_name_quoted(file->from_name, options->file, '\0');
2617 }
2618 write_name_quoted(file->name, options->file, '\0');
2619 }
2620 }
2621}
2622struct dirstat_file {
2624 const char *name;
2625 unsigned long changed;
2626};
2627struct dirstat_dir {
2629 struct dirstat_file *files;
2630 int alloc, nr, permille, cumulative;
2631};
2632static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
2634 unsigned long changed, const char *base, int baselen)
2635{
2636 unsigned long this_dir = 0;
2637 unsigned int sources = 0;
2638 const char *line_prefix = diff_line_prefix(opt);
2639 while (dir->nr) {
2641 struct dirstat_file *f = dir->files;
2642 int namelen = strlen(f->name);
2643 unsigned long this;
2644 char *slash;
2645 if (namelen < baselen)
2647 break;
2648 if (memcmp(f->name, base, baselen))
2649 break;
2650 slash = strchr(f->name + baselen, '/');
2651 if (slash) {
2652 int newbaselen = slash + 1 - f->name;
2653 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
2654 sources++;
2655 } else {
2656 this = f->changed;
2657 dir->files++;
2658 dir->nr--;
2659 sources += 2;
2660 }
2661 this_dir += this;
2662 }
2663 /*
2665 * We don't report dirstat's for
2666 * - the top level
2667 * - or cases where everything came from a single directory
2668 * under this directory (sources == 1).
2669 */
2670 if (baselen && sources != 1) {
2671 if (this_dir) {
2672 int permille = this_dir * 1000 / changed;
2673 if (permille >= dir->permille) {
2674 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
2675 permille / 10, permille % 10, baselen, base);
2676 if (!dir->cumulative)
2677 return 0;
2678 }
2679 }
2680 }
2681 return this_dir;
2682}
2683static int dirstat_compare(const void *_a, const void *_b)
2685{
2686 const struct dirstat_file *a = _a;
2687 const struct dirstat_file *b = _b;
2688 return strcmp(a->name, b->name);
2689}
2690static void show_dirstat(struct diff_options *options)
2692{
2693 int i;
2694 unsigned long changed;
2695 struct dirstat_dir dir;
2696 struct diff_queue_struct *q = &diff_queued_diff;
2697 dir.files = NULL;
2699 dir.alloc = 0;
2700 dir.nr = 0;
2701 dir.permille = options->dirstat_permille;
2702 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2703 changed = 0;
2705 for (i = 0; i < q->nr; i++) {
2706 struct diff_filepair *p = q->queue[i];
2707 const char *name;
2708 unsigned long copied, added, damage;
2709 int content_changed;
2710 name = p->two->path ? p->two->path : p->one->path;
2712 if (p->one->oid_valid && p->two->oid_valid)
2714 content_changed = oidcmp(&p->one->oid, &p->two->oid);
2715 else
2716 content_changed = 1;
2717 if (!content_changed) {
2719 /*
2720 * The SHA1 has not changed, so pre-/post-content is
2721 * identical. We can therefore skip looking at the
2722 * file contents altogether.
2723 */
2724 damage = 0;
2725 goto found_damage;
2726 }
2727 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
2729 /*
2730 * In --dirstat-by-file mode, we don't really need to
2731 * look at the actual file contents at all.
2732 * The fact that the SHA1 changed is enough for us to
2733 * add this file to the list of results
2734 * (with each file contributing equal damage).
2735 */
2736 damage = 1;
2737 goto found_damage;
2738 }
2739 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
2741 diff_populate_filespec(p->one, 0);
2742 diff_populate_filespec(p->two, 0);
2743 diffcore_count_changes(p->one, p->two, NULL, NULL,
2744 &copied, &added);
2745 diff_free_filespec_data(p->one);
2746 diff_free_filespec_data(p->two);
2747 } else if (DIFF_FILE_VALID(p->one)) {
2748 diff_populate_filespec(p->one, CHECK_SIZE_ONLY);
2749 copied = added = 0;
2750 diff_free_filespec_data(p->one);
2751 } else if (DIFF_FILE_VALID(p->two)) {
2752 diff_populate_filespec(p->two, CHECK_SIZE_ONLY);
2753 copied = 0;
2754 added = p->two->size;
2755 diff_free_filespec_data(p->two);
2756 } else
2757 continue;
2758 /*
2760 * Original minus copied is the removed material,
2761 * added is the new material. They are both damages
2762 * made to the preimage.
2763 * If the resulting damage is zero, we know that
2764 * diffcore_count_changes() considers the two entries to
2765 * be identical, but since content_changed is true, we
2766 * know that there must have been _some_ kind of change,
2767 * so we force all entries to have damage > 0.
2768 */
2769 damage = (p->one->size - copied) + added;
2770 if (!damage)
2771 damage = 1;
2772found_damage:
2774 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2775 dir.files[dir.nr].name = name;
2776 dir.files[dir.nr].changed = damage;
2777 changed += damage;
2778 dir.nr++;
2779 }
2780 /* This can happen even with many files, if everything was renames */
2782 if (!changed)
2783 return;
2784 /* Show all directories with more than x% of the changes */
2786 QSORT(dir.files, dir.nr, dirstat_compare);
2787 gather_dirstat(options, &dir, changed, "", 0);
2788}
2789static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
2791{
2792 int i;
2793 unsigned long changed;
2794 struct dirstat_dir dir;
2795 if (data->nr == 0)
2797 return;
2798 dir.files = NULL;
2800 dir.alloc = 0;
2801 dir.nr = 0;
2802 dir.permille = options->dirstat_permille;
2803 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
2804 changed = 0;
2806 for (i = 0; i < data->nr; i++) {
2807 struct diffstat_file *file = data->files[i];
2808 unsigned long damage = file->added + file->deleted;
2809 if (file->is_binary)
2810 /*
2811 * binary files counts bytes, not lines. Must find some
2812 * way to normalize binary bytes vs. textual lines.
2813 * The following heuristic assumes that there are 64
2814 * bytes per "line".
2815 * This is stupid and ugly, but very cheap...
2816 */
2817 damage = DIV_ROUND_UP(damage, 64);
2818 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
2819 dir.files[dir.nr].name = file->name;
2820 dir.files[dir.nr].changed = damage;
2821 changed += damage;
2822 dir.nr++;
2823 }
2824 /* This can happen even with many files, if everything was renames */
2826 if (!changed)
2827 return;
2828 /* Show all directories with more than x% of the changes */
2830 QSORT(dir.files, dir.nr, dirstat_compare);
2831 gather_dirstat(options, &dir, changed, "", 0);
2832}
2833static void free_diffstat_info(struct diffstat_t *diffstat)
2835{
2836 int i;
2837 for (i = 0; i < diffstat->nr; i++) {
2838 struct diffstat_file *f = diffstat->files[i];
2839 if (f->name != f->print_name)
2840 free(f->print_name);
2841 free(f->name);
2842 free(f->from_name);
2843 free(f);
2844 }
2845 free(diffstat->files);
2846}
2847struct checkdiff_t {
2849 const char *filename;
2850 int lineno;
2851 int conflict_marker_size;
2852 struct diff_options *o;
2853 unsigned ws_rule;
2854 unsigned status;
2855};
2856static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
2858{
2859 char firstchar;
2860 int cnt;
2861 if (len < marker_size + 1)
2863 return 0;
2864 firstchar = line[0];
2865 switch (firstchar) {
2866 case '=': case '>': case '<': case '|':
2867 break;
2868 default:
2869 return 0;
2870 }
2871 for (cnt = 1; cnt < marker_size; cnt++)
2872 if (line[cnt] != firstchar)
2873 return 0;
2874 /* line[1] thru line[marker_size-1] are same as firstchar */
2875 if (len < marker_size + 1 || !isspace(line[marker_size]))
2876 return 0;
2877 return 1;
2878}
2879static void checkdiff_consume(void *priv, char *line, unsigned long len)
2881{
2882 struct checkdiff_t *data = priv;
2883 int marker_size = data->conflict_marker_size;
2884 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2885 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2886 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2887 char *err;
2888 const char *line_prefix;
2889 assert(data->o);
2891 line_prefix = diff_line_prefix(data->o);
2892 if (line[0] == '+') {
2894 unsigned bad;
2895 data->lineno++;
2896 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2897 data->status |= 1;
2898 fprintf(data->o->file,
2899 "%s%s:%d: leftover conflict marker\n",
2900 line_prefix, data->filename, data->lineno);
2901 }
2902 bad = ws_check(line + 1, len - 1, data->ws_rule);
2903 if (!bad)
2904 return;
2905 data->status |= bad;
2906 err = whitespace_error_string(bad);
2907 fprintf(data->o->file, "%s%s:%d: %s.\n",
2908 line_prefix, data->filename, data->lineno, err);
2909 free(err);
2910 emit_line(data->o, set, reset, line, 1);
2911 ws_check_emit(line + 1, len - 1, data->ws_rule,
2912 data->o->file, set, reset, ws);
2913 } else if (line[0] == ' ') {
2914 data->lineno++;
2915 } else if (line[0] == '@') {
2916 char *plus = strchr(line, '+');
2917 if (plus)
2918 data->lineno = strtol(plus, NULL, 10) - 1;
2919 else
2920 die("invalid diff");
2921 }
2922}
2923static unsigned char *deflate_it(char *data,
2925 unsigned long size,
2926 unsigned long *result_size)
2927{
2928 int bound;
2929 unsigned char *deflated;
2930 git_zstream stream;
2931 git_deflate_init(&stream, zlib_compression_level);
2933 bound = git_deflate_bound(&stream, size);
2934 deflated = xmalloc(bound);
2935 stream.next_out = deflated;
2936 stream.avail_out = bound;
2937 stream.next_in = (unsigned char *)data;
2939 stream.avail_in = size;
2940 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2941 ; /* nothing */
2942 git_deflate_end(&stream);
2943 *result_size = stream.total_out;
2944 return deflated;
2945}
2946static void emit_binary_diff_body(struct diff_options *o,
2948 mmfile_t *one, mmfile_t *two)
2949{
2950 void *cp;
2951 void *delta;
2952 void *deflated;
2953 void *data;
2954 unsigned long orig_size;
2955 unsigned long delta_size;
2956 unsigned long deflate_size;
2957 unsigned long data_size;
2958 /* We could do deflated delta, or we could do just deflated two,
2960 * whichever is smaller.
2961 */
2962 delta = NULL;
2963 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2964 if (one->size && two->size) {
2965 delta = diff_delta(one->ptr, one->size,
2966 two->ptr, two->size,
2967 &delta_size, deflate_size);
2968 if (delta) {
2969 void *to_free = delta;
2970 orig_size = delta_size;
2971 delta = deflate_it(delta, delta_size, &delta_size);
2972 free(to_free);
2973 }
2974 }
2975 if (delta && delta_size < deflate_size) {
2977 char *s = xstrfmt("%lu", orig_size);
2978 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_DELTA,
2979 s, strlen(s), 0);
2980 free(s);
2981 free(deflated);
2982 data = delta;
2983 data_size = delta_size;
2984 } else {
2985 char *s = xstrfmt("%lu", two->size);
2986 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER_LITERAL,
2987 s, strlen(s), 0);
2988 free(s);
2989 free(delta);
2990 data = deflated;
2991 data_size = deflate_size;
2992 }
2993 /* emit data encoded in base85 */
2995 cp = data;
2996 while (data_size) {
2997 int len;
2998 int bytes = (52 < data_size) ? 52 : data_size;
2999 char line[71];
3000 data_size -= bytes;
3001 if (bytes <= 26)
3002 line[0] = bytes + 'A' - 1;
3003 else
3004 line[0] = bytes - 26 + 'a' - 1;
3005 encode_85(line + 1, cp, bytes);
3006 cp = (char *) cp + bytes;
3007 len = strlen(line);
3009 line[len++] = '\n';
3010 line[len] = '\0';
3011 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_BODY,
3013 line, len, 0);
3014 }
3015 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_FOOTER, NULL, 0, 0);
3016 free(data);
3017}
3018static void emit_binary_diff(struct diff_options *o,
3020 mmfile_t *one, mmfile_t *two)
3021{
3022 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_DIFF_HEADER, NULL, 0, 0);
3023 emit_binary_diff_body(o, one, two);
3024 emit_binary_diff_body(o, two, one);
3025}
3026int diff_filespec_is_binary(struct diff_filespec *one)
3028{
3029 if (one->is_binary == -1) {
3030 diff_filespec_load_driver(one);
3031 if (one->driver->binary != -1)
3032 one->is_binary = one->driver->binary;
3033 else {
3034 if (!one->data && DIFF_FILE_VALID(one))
3035 diff_populate_filespec(one, CHECK_BINARY);
3036 if (one->is_binary == -1 && one->data)
3037 one->is_binary = buffer_is_binary(one->data,
3038 one->size);
3039 if (one->is_binary == -1)
3040 one->is_binary = 0;
3041 }
3042 }
3043 return one->is_binary;
3044}
3045static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
3047{
3048 diff_filespec_load_driver(one);
3049 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
3050}
3051void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
3053{
3054 if (!options->a_prefix)
3055 options->a_prefix = a;
3056 if (!options->b_prefix)
3057 options->b_prefix = b;
3058}
3059struct userdiff_driver *get_textconv(struct diff_filespec *one)
3061{
3062 if (!DIFF_FILE_VALID(one))
3063 return NULL;
3064 diff_filespec_load_driver(one);
3066 return userdiff_get_textconv(one->driver);
3067}
3068static void builtin_diff(const char *name_a,
3070 const char *name_b,
3071 struct diff_filespec *one,
3072 struct diff_filespec *two,
3073 const char *xfrm_msg,
3074 int must_show_header,
3075 struct diff_options *o,
3076 int complete_rewrite)
3077{
3078 mmfile_t mf1, mf2;
3079 const char *lbl[2];
3080 char *a_one, *b_two;
3081 const char *meta = diff_get_color_opt(o, DIFF_METAINFO);
3082 const char *reset = diff_get_color_opt(o, DIFF_RESET);
3083 const char *a_prefix, *b_prefix;
3084 struct userdiff_driver *textconv_one = NULL;
3085 struct userdiff_driver *textconv_two = NULL;
3086 struct strbuf header = STRBUF_INIT;
3087 const char *line_prefix = diff_line_prefix(o);
3088 diff_set_mnemonic_prefix(o, "a/", "b/");
3090 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
3091 a_prefix = o->b_prefix;
3092 b_prefix = o->a_prefix;
3093 } else {
3094 a_prefix = o->a_prefix;
3095 b_prefix = o->b_prefix;
3096 }
3097 if (o->submodule_format == DIFF_SUBMODULE_LOG &&
3099 (!one->mode || S_ISGITLINK(one->mode)) &&
3100 (!two->mode || S_ISGITLINK(two->mode))) {
3101 show_submodule_summary(o, one->path ? one->path : two->path,
3102 &one->oid, &two->oid,
3103 two->dirty_submodule);
3104 return;
3105 } else if (o->submodule_format == DIFF_SUBMODULE_INLINE_DIFF &&
3106 (!one->mode || S_ISGITLINK(one->mode)) &&
3107 (!two->mode || S_ISGITLINK(two->mode))) {
3108 show_submodule_inline_diff(o, one->path ? one->path : two->path,
3109 &one->oid, &two->oid,
3110 two->dirty_submodule);
3111 return;
3112 }
3113 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
3115 textconv_one = get_textconv(one);
3116 textconv_two = get_textconv(two);
3117 }
3118 /* Never use a non-valid filename anywhere if at all possible */
3120 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
3121 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
3122 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
3124 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
3125 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
3126 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
3127 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, meta, a_one, b_two, reset);
3128 if (lbl[0][0] == '/') {
3129 /* /dev/null */
3130 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, meta, two->mode, reset);
3131 if (xfrm_msg)
3132 strbuf_addstr(&header, xfrm_msg);
3133 must_show_header = 1;
3134 }
3135 else if (lbl[1][0] == '/') {
3136 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, meta, one->mode, reset);
3137 if (xfrm_msg)
3138 strbuf_addstr(&header, xfrm_msg);
3139 must_show_header = 1;
3140 }
3141 else {
3142 if (one->mode != two->mode) {
3143 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, meta, one->mode, reset);
3144 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, meta, two->mode, reset);
3145 must_show_header = 1;
3146 }
3147 if (xfrm_msg)
3148 strbuf_addstr(&header, xfrm_msg);
3149 /*
3151 * we do not run diff between different kind
3152 * of objects.
3153 */
3154 if ((one->mode ^ two->mode) & S_IFMT)
3155 goto free_ab_and_return;
3156 if (complete_rewrite &&
3157 (textconv_one || !diff_filespec_is_binary(one)) &&
3158 (textconv_two || !diff_filespec_is_binary(two))) {
3159 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3160 header.buf, header.len, 0);
3161 strbuf_reset(&header);
3162 emit_rewrite_diff(name_a, name_b, one, two,
3163 textconv_one, textconv_two, o);
3164 o->found_changes = 1;
3165 goto free_ab_and_return;
3166 }
3167 }
3168 if (o->irreversible_delete && lbl[1][0] == '/') {
3170 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf,
3171 header.len, 0);
3172 strbuf_reset(&header);
3173 goto free_ab_and_return;
3174 } else if (!DIFF_OPT_TST(o, TEXT) &&
3175 ( (!textconv_one && diff_filespec_is_binary(one)) ||
3176 (!textconv_two && diff_filespec_is_binary(two)) )) {
3177 struct strbuf sb = STRBUF_INIT;
3178 if (!one->data && !two->data &&
3179 S_ISREG(one->mode) && S_ISREG(two->mode) &&
3180 !DIFF_OPT_TST(o, BINARY)) {
3181 if (!oidcmp(&one->oid, &two->oid)) {
3182 if (must_show_header)
3183 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3184 header.buf, header.len,
3185 0);
3186 goto free_ab_and_return;
3187 }
3188 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3189 header.buf, header.len, 0);
3190 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3191 diff_line_prefix(o), lbl[0], lbl[1]);
3192 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3193 sb.buf, sb.len, 0);
3194 strbuf_release(&sb);
3195 goto free_ab_and_return;
3196 }
3197 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3198 die("unable to read files to diff");
3199 /* Quite common confusing case */
3200 if (mf1.size == mf2.size &&
3201 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
3202 if (must_show_header)
3203 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3204 header.buf, header.len, 0);
3205 goto free_ab_and_return;
3206 }
3207 emit_diff_symbol(o, DIFF_SYMBOL_HEADER, header.buf, header.len, 0);
3208 strbuf_reset(&header);
3209 if (DIFF_OPT_TST(o, BINARY))
3210 emit_binary_diff(o, &mf1, &mf2);
3211 else {
3212 strbuf_addf(&sb, "%sBinary files %s and %s differ\n",
3213 diff_line_prefix(o), lbl[0], lbl[1]);
3214 emit_diff_symbol(o, DIFF_SYMBOL_BINARY_FILES,
3215 sb.buf, sb.len, 0);
3216 strbuf_release(&sb);
3217 }
3218 o->found_changes = 1;
3219 } else {
3220 /* Crazy xdl interfaces.. */
3221 const char *diffopts = getenv("GIT_DIFF_OPTS");
3222 const char *v;
3223 xpparam_t xpp;
3224 xdemitconf_t xecfg;
3225 struct emit_callback ecbdata;
3226 const struct userdiff_funcname *pe;
3227 if (must_show_header) {
3229 emit_diff_symbol(o, DIFF_SYMBOL_HEADER,
3230 header.buf, header.len, 0);
3231 strbuf_reset(&header);
3232 }
3233 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
3235 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
3236 pe = diff_funcname_pattern(one);
3238 if (!pe)
3239 pe = diff_funcname_pattern(two);
3240 memset(&xpp, 0, sizeof(xpp));
3242 memset(&xecfg, 0, sizeof(xecfg));
3243 memset(&ecbdata, 0, sizeof(ecbdata));
3244 ecbdata.label_path = lbl;
3245 ecbdata.color_diff = want_color(o->use_color);
3246 ecbdata.ws_rule = whitespace_rule(name_b);
3247 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
3248 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3249 ecbdata.opt = o;
3250 ecbdata.header = header.len ? &header : NULL;
3251 xpp.flags = o->xdl_opts;
3252 xecfg.ctxlen = o->context;
3253 xecfg.interhunkctxlen = o->interhunkcontext;
3254 xecfg.flags = XDL_EMIT_FUNCNAMES;
3255 if (DIFF_OPT_TST(o, FUNCCONTEXT))
3256 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
3257 if (pe)
3258 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
3259 if (!diffopts)
3260 ;
3261 else if (skip_prefix(diffopts, "--unified=", &v))
3262 xecfg.ctxlen = strtoul(v, NULL, 10);
3263 else if (skip_prefix(diffopts, "-u", &v))
3264 xecfg.ctxlen = strtoul(v, NULL, 10);
3265 if (o->word_diff)
3266 init_diff_words_data(&ecbdata, o, one, two);
3267 if (xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
3268 &xpp, &xecfg))
3269 die("unable to generate diff for %s", one->path);
3270 if (o->word_diff)
3271 free_diff_words_data(&ecbdata);
3272 if (textconv_one)
3273 free(mf1.ptr);
3274 if (textconv_two)
3275 free(mf2.ptr);
3276 xdiff_clear_find_func(&xecfg);
3277 }
3278 free_ab_and_return:
3280 strbuf_release(&header);
3281 diff_free_filespec_data(one);
3282 diff_free_filespec_data(two);
3283 free(a_one);
3284 free(b_two);
3285 return;
3286}
3287static void builtin_diffstat(const char *name_a, const char *name_b,
3289 struct diff_filespec *one,
3290 struct diff_filespec *two,
3291 struct diffstat_t *diffstat,
3292 struct diff_options *o,
3293 struct diff_filepair *p)
3294{
3295 mmfile_t mf1, mf2;
3296 struct diffstat_file *data;
3297 int same_contents;
3298 int complete_rewrite = 0;
3299 if (!DIFF_PAIR_UNMERGED(p)) {
3301 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3302 complete_rewrite = 1;
3303 }
3304 data = diffstat_add(diffstat, name_a, name_b);
3306 data->is_interesting = p->status != DIFF_STATUS_UNKNOWN;
3307 if (!one || !two) {
3309 data->is_unmerged = 1;
3310 return;
3311 }
3312 same_contents = !oidcmp(&one->oid, &two->oid);
3314 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
3316 data->is_binary = 1;
3317 if (same_contents) {
3318 data->added = 0;
3319 data->deleted = 0;
3320 } else {
3321 data->added = diff_filespec_size(two);
3322 data->deleted = diff_filespec_size(one);
3323 }
3324 }
3325 else if (complete_rewrite) {
3327 diff_populate_filespec(one, 0);
3328 diff_populate_filespec(two, 0);
3329 data->deleted = count_lines(one->data, one->size);
3330 data->added = count_lines(two->data, two->size);
3331 }
3332 else if (!same_contents) {
3334 /* Crazy xdl interfaces.. */
3335 xpparam_t xpp;
3336 xdemitconf_t xecfg;
3337 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3339 die("unable to read files to diff");
3340 memset(&xpp, 0, sizeof(xpp));
3342 memset(&xecfg, 0, sizeof(xecfg));
3343 xpp.flags = o->xdl_opts;
3344 xecfg.ctxlen = o->context;
3345 xecfg.interhunkctxlen = o->interhunkcontext;
3346 if (xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
3347 &xpp, &xecfg))
3348 die("unable to generate diffstat for %s", one->path);
3349 }
3350 diff_free_filespec_data(one);
3352 diff_free_filespec_data(two);
3353}
3354static void builtin_checkdiff(const char *name_a, const char *name_b,
3356 const char *attr_path,
3357 struct diff_filespec *one,
3358 struct diff_filespec *two,
3359 struct diff_options *o)
3360{
3361 mmfile_t mf1, mf2;
3362 struct checkdiff_t data;
3363 if (!two)
3365 return;
3366 memset(&data, 0, sizeof(data));
3368 data.filename = name_b ? name_b : name_a;
3369 data.lineno = 0;
3370 data.o = o;
3371 data.ws_rule = whitespace_rule(attr_path);
3372 data.conflict_marker_size = ll_merge_marker_size(attr_path);
3373 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
3375 die("unable to read files to diff");
3376 /*
3378 * All the other codepaths check both sides, but not checking
3379 * the "old" side here is deliberate. We are checking the newly
3380 * introduced changes, and as long as the "new" side is text, we
3381 * can and should check what it introduces.
3382 */
3383 if (diff_filespec_is_binary(two))
3384 goto free_and_return;
3385 else {
3386 /* Crazy xdl interfaces.. */
3387 xpparam_t xpp;
3388 xdemitconf_t xecfg;
3389 memset(&xpp, 0, sizeof(xpp));
3391 memset(&xecfg, 0, sizeof(xecfg));
3392 xecfg.ctxlen = 1; /* at least one context line */
3393 xpp.flags = 0;
3394 if (xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
3395 &xpp, &xecfg))
3396 die("unable to generate checkdiff for %s", one->path);
3397 if (data.ws_rule & WS_BLANK_AT_EOF) {
3399 struct emit_callback ecbdata;
3400 int blank_at_eof;
3401 ecbdata.ws_rule = data.ws_rule;
3403 check_blank_at_eof(&mf1, &mf2, &ecbdata);
3404 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
3405 if (blank_at_eof) {
3407 static char *err;
3408 if (!err)
3409 err = whitespace_error_string(WS_BLANK_AT_EOF);
3410 fprintf(o->file, "%s:%d: %s.\n",
3411 data.filename, blank_at_eof, err);
3412 data.status = 1; /* report errors */
3413 }
3414 }
3415 }
3416 free_and_return:
3417 diff_free_filespec_data(one);
3418 diff_free_filespec_data(two);
3419 if (data.status)
3420 DIFF_OPT_SET(o, CHECK_FAILED);
3421}
3422struct diff_filespec *alloc_filespec(const char *path)
3424{
3425 struct diff_filespec *spec;
3426 FLEXPTR_ALLOC_STR(spec, path, path);
3428 spec->count = 1;
3429 spec->is_binary = -1;
3430 return spec;
3431}
3432void free_filespec(struct diff_filespec *spec)
3434{
3435 if (!--spec->count) {
3436 diff_free_filespec_data(spec);
3437 free(spec);
3438 }
3439}
3440void fill_filespec(struct diff_filespec *spec, const struct object_id *oid,
3442 int oid_valid, unsigned short mode)
3443{
3444 if (mode) {
3445 spec->mode = canon_mode(mode);
3446 oidcpy(&spec->oid, oid);
3447 spec->oid_valid = oid_valid;
3448 }
3449}
3450/*
3452 * Given a name and sha1 pair, if the index tells us the file in
3453 * the work tree has that object contents, return true, so that
3454 * prepare_temp_file() does not have to inflate and extract.
3455 */
3456static int reuse_worktree_file(const char *name, const struct object_id *oid, int want_file)
3457{
3458 const struct cache_entry *ce;
3459 struct stat st;
3460 int pos, len;
3461 /*
3463 * We do not read the cache ourselves here, because the
3464 * benchmark with my previous version that always reads cache
3465 * shows that it makes things worse for diff-tree comparing
3466 * two linux-2.6 kernel trees in an already checked out work
3467 * tree. This is because most diff-tree comparisons deal with
3468 * only a small number of files, while reading the cache is
3469 * expensive for a large project, and its cost outweighs the
3470 * savings we get by not inflating the object to a temporary
3471 * file. Practically, this code only helps when we are used
3472 * by diff-cache --cached, which does read the cache before
3473 * calling us.
3474 */
3475 if (!active_cache)
3476 return 0;
3477 /* We want to avoid the working directory if our caller
3479 * doesn't need the data in a normal file, this system
3480 * is rather slow with its stat/open/mmap/close syscalls,
3481 * and the object is contained in a pack file. The pack
3482 * is probably already open and will be faster to obtain
3483 * the data through than the working directory. Loose
3484 * objects however would tend to be slower as they need
3485 * to be individually opened and inflated.
3486 */
3487 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(oid->hash))
3488 return 0;
3489 /*
3491 * Similarly, if we'd have to convert the file contents anyway, that
3492 * makes the optimization not worthwhile.
3493 */
3494 if (!want_file && would_convert_to_git(&the_index, name))
3495 return 0;
3496 len = strlen(name);
3498 pos = cache_name_pos(name, len);
3499 if (pos < 0)
3500 return 0;
3501 ce = active_cache[pos];
3502 /*
3504 * This is not the sha1 we are looking for, or
3505 * unreusable because it is not a regular file.
3506 */
3507 if (oidcmp(oid, &ce->oid) || !S_ISREG(ce->ce_mode))
3508 return 0;
3509 /*
3511 * If ce is marked as "assume unchanged", there is no
3512 * guarantee that work tree matches what we are looking for.
3513 */
3514 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
3515 return 0;
3516 /*
3518 * If ce matches the file in the work tree, we can reuse it.
3519 */
3520 if (ce_uptodate(ce) ||
3521 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
3522 return 1;
3523 return 0;
3525}
3526static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
3528{
3529 struct strbuf buf = STRBUF_INIT;
3530 char *dirty = "";
3531 /* Are we looking at the work tree? */
3533 if (s->dirty_submodule)
3534 dirty = "-dirty";
3535 strbuf_addf(&buf, "Subproject commit %s%s\n",
3537 oid_to_hex(&s->oid), dirty);
3538 s->size = buf.len;
3539 if (size_only) {
3540 s->data = NULL;
3541 strbuf_release(&buf);
3542 } else {
3543 s->data = strbuf_detach(&buf, NULL);
3544 s->should_free = 1;
3545 }
3546 return 0;
3547}
3548/*
3550 * While doing rename detection and pickaxe operation, we may need to
3551 * grab the data for the blob (or file) for our own in-core comparison.
3552 * diff_filespec has data and size fields for this purpose.
3553 */
3554int diff_populate_filespec(struct diff_filespec *s, unsigned int flags)
3555{
3556 int size_only = flags & CHECK_SIZE_ONLY;
3557 int err = 0;
3558 /*
3559 * demote FAIL to WARN to allow inspecting the situation
3560 * instead of refusing.
3561 */
3562 enum safe_crlf crlf_warn = (safe_crlf == SAFE_CRLF_FAIL
3563 ? SAFE_CRLF_WARN
3564 : safe_crlf);
3565 if (!DIFF_FILE_VALID(s))
3567 die("internal error: asking to populate invalid file.");
3568 if (S_ISDIR(s->mode))
3569 return -1;
3570 if (s->data)
3572 return 0;
3573 if (size_only && 0 < s->size)
3575 return 0;
3576 if (S_ISGITLINK(s->mode))
3578 return diff_populate_gitlink(s, size_only);
3579 if (!s->oid_valid ||
3581 reuse_worktree_file(s->path, &s->oid, 0)) {
3582 struct strbuf buf = STRBUF_INIT;
3583 struct stat st;
3584 int fd;
3585 if (lstat(s->path, &st) < 0) {
3587 if (errno == ENOENT) {
3588 err_empty:
3589 err = -1;
3590 empty:
3591 s->data = (char *)"";
3592 s->size = 0;
3593 return err;
3594 }
3595 }
3596 s->size = xsize_t(st.st_size);
3597 if (!s->size)
3598 goto empty;
3599 if (S_ISLNK(st.st_mode)) {
3600 struct strbuf sb = STRBUF_INIT;
3601 if (strbuf_readlink(&sb, s->path, s->size))
3603 goto err_empty;
3604 s->size = sb.len;
3605 s->data = strbuf_detach(&sb, NULL);
3606 s->should_free = 1;
3607 return 0;
3608 }
3609 /*
3611 * Even if the caller would be happy with getting
3612 * only the size, we cannot return early at this
3613 * point if the path requires us to run the content
3614 * conversion.
3615 */
3616 if (size_only && !would_convert_to_git(&the_index, s->path))
3617 return 0;
3618 /*
3620 * Note: this check uses xsize_t(st.st_size) that may
3621 * not be the true size of the blob after it goes
3622 * through convert_to_git(). This may not strictly be
3623 * correct, but the whole point of big_file_threshold
3624 * and is_binary check being that we want to avoid
3625 * opening the file and inspecting the contents, this
3626 * is probably fine.
3627 */
3628 if ((flags & CHECK_BINARY) &&
3629 s->size > big_file_threshold && s->is_binary == -1) {
3630 s->is_binary = 1;
3631 return 0;
3632 }
3633 fd = open(s->path, O_RDONLY);
3634 if (fd < 0)
3635 goto err_empty;
3636 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
3637 close(fd);
3638 s->should_munmap = 1;
3639 /*
3641 * Convert from working tree format to canonical git format
3642 */
3643 if (convert_to_git(&the_index, s->path, s->data, s->size, &buf, crlf_warn)) {
3644 size_t size = 0;
3645 munmap(s->data, s->size);
3646 s->should_munmap = 0;
3647 s->data = strbuf_detach(&buf, &size);
3648 s->size = size;
3649 s->should_free = 1;
3650 }
3651 }
3652 else {
3653 enum object_type type;
3654 if (size_only || (flags & CHECK_BINARY)) {
3655 type = sha1_object_info(s->oid.hash, &s->size);
3656 if (type < 0)
3657 die("unable to read %s",
3658 oid_to_hex(&s->oid));
3659 if (size_only)
3660 return 0;
3661 if (s->size > big_file_threshold && s->is_binary == -1) {
3662 s->is_binary = 1;
3663 return 0;
3664 }
3665 }
3666 s->data = read_sha1_file(s->oid.hash, &type, &s->size);
3667 if (!s->data)
3668 die("unable to read %s", oid_to_hex(&s->oid));
3669 s->should_free = 1;
3670 }
3671 return 0;
3672}
3673void diff_free_filespec_blob(struct diff_filespec *s)
3675{
3676 if (s->should_free)
3677 free(s->data);
3678 else if (s->should_munmap)
3679 munmap(s->data, s->size);
3680 if (s->should_free || s->should_munmap) {
3682 s->should_free = s->should_munmap = 0;
3683 s->data = NULL;
3684 }
3685}
3686void diff_free_filespec_data(struct diff_filespec *s)
3688{
3689 diff_free_filespec_blob(s);
3690 FREE_AND_NULL(s->cnt_data);
3691}
3692static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
3694 void *blob,
3695 unsigned long size,
3696 const struct object_id *oid,
3697 int mode)
3698{
3699 int fd;
3700 struct strbuf buf = STRBUF_INIT;
3701 struct strbuf template = STRBUF_INIT;
3702 char *path_dup = xstrdup(path);
3703 const char *base = basename(path_dup);
3704 /* Generate "XXXXXX_basename.ext" */
3706 strbuf_addstr(&template, "XXXXXX_");
3707 strbuf_addstr(&template, base);
3708 fd = mks_tempfile_ts(&temp->tempfile, template.buf, strlen(base) + 1);
3710 if (fd < 0)
3711 die_errno("unable to create temp-file");
3712 if (convert_to_working_tree(path,
3713 (const char *)blob, (size_t)size, &buf)) {
3714 blob = buf.buf;
3715 size = buf.len;
3716 }
3717 if (write_in_full(fd, blob, size) != size)
3718 die_errno("unable to write temp-file");
3719 close_tempfile(&temp->tempfile);
3720 temp->name = get_tempfile_path(&temp->tempfile);
3721 oid_to_hex_r(temp->hex, oid);
3722 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", mode);
3723 strbuf_release(&buf);
3724 strbuf_release(&template);
3725 free(path_dup);
3726}
3727static struct diff_tempfile *prepare_temp_file(const char *name,
3729 struct diff_filespec *one)
3730{
3731 struct diff_tempfile *temp = claim_diff_tempfile();
3732 if (!DIFF_FILE_VALID(one)) {
3734 not_a_valid_file:
3735 /* A '-' entry produces this for file-2, and
3736 * a '+' entry produces this for file-1.
3737 */
3738 temp->name = "/dev/null";
3739 xsnprintf(temp->hex, sizeof(temp->hex), ".");
3740 xsnprintf(temp->mode, sizeof(temp->mode), ".");
3741 return temp;
3742 }
3743 if (!S_ISGITLINK(one->mode) &&
3745 (!one->oid_valid ||
3746 reuse_worktree_file(name, &one->oid, 1))) {
3747 struct stat st;
3748 if (lstat(name, &st) < 0) {
3749 if (errno == ENOENT)
3750 goto not_a_valid_file;
3751 die_errno("stat(%s)", name);
3752 }
3753 if (S_ISLNK(st.st_mode)) {
3754 struct strbuf sb = STRBUF_INIT;
3755 if (strbuf_readlink(&sb, name, st.st_size) < 0)
3756 die_errno("readlink(%s)", name);
3757 prep_temp_blob(name, temp, sb.buf, sb.len,
3758 (one->oid_valid ?
3759 &one->oid : &null_oid),
3760 (one->oid_valid ?
3761 one->mode : S_IFLNK));
3762 strbuf_release(&sb);
3763 }
3764 else {
3765 /* we can borrow from the file in the work tree */
3766 temp->name = name;
3767 if (!one->oid_valid)
3768 oid_to_hex_r(temp->hex, &null_oid);
3769 else
3770 oid_to_hex_r(temp->hex, &one->oid);
3771 /* Even though we may sometimes borrow the
3772 * contents from the work tree, we always want
3773 * one->mode. mode is trustworthy even when
3774 * !(one->oid_valid), as long as
3775 * DIFF_FILE_VALID(one).
3776 */
3777 xsnprintf(temp->mode, sizeof(temp->mode), "%06o", one->mode);
3778 }
3779 return temp;
3780 }
3781 else {
3782 if (diff_populate_filespec(one, 0))
3783 die("cannot read data blob for %s", one->path);
3784 prep_temp_blob(name, temp, one->data, one->size,
3785 &one->oid, one->mode);
3786 }
3787 return temp;
3788}
3789static void add_external_diff_name(struct argv_array *argv,
3791 const char *name,
3792 struct diff_filespec *df)
3793{
3794 struct diff_tempfile *temp = prepare_temp_file(name, df);
3795 argv_array_push(argv, temp->name);
3796 argv_array_push(argv, temp->hex);
3797 argv_array_push(argv, temp->mode);
3798}
3799/* An external diff command takes:
3801 *
3802 * diff-cmd name infile1 infile1-sha1 infile1-mode \
3803 * infile2 infile2-sha1 infile2-mode [ rename-to ]
3804 *
3805 */
3806static void run_external_diff(const char *pgm,
3807 const char *name,
3808 const char *other,
3809 struct diff_filespec *one,
3810 struct diff_filespec *two,
3811 const char *xfrm_msg,
3812 int complete_rewrite,
3813 struct diff_options *o)
3814{
3815 struct argv_array argv = ARGV_ARRAY_INIT;
3816 struct argv_array env = ARGV_ARRAY_INIT;
3817 struct diff_queue_struct *q = &diff_queued_diff;
3818 argv_array_push(&argv, pgm);
3820 argv_array_push(&argv, name);
3821 if (one && two) {
3823 add_external_diff_name(&argv, name, one);
3824 if (!other)
3825 add_external_diff_name(&argv, name, two);
3826 else {
3827 add_external_diff_name(&argv, other, two);
3828 argv_array_push(&argv, other);
3829 argv_array_push(&argv, xfrm_msg);
3830 }
3831 }
3832 argv_array_pushf(&env, "GIT_DIFF_PATH_COUNTER=%d", ++o->diff_path_counter);
3834 argv_array_pushf(&env, "GIT_DIFF_PATH_TOTAL=%d", q->nr);
3835 if (run_command_v_opt_cd_env(argv.argv, RUN_USING_SHELL, NULL, env.argv))
3837 die(_("external diff died, stopping at %s"), name);
3838 remove_tempfile();
3840 argv_array_clear(&argv);
3841 argv_array_clear(&env);
3842}
3843static int similarity_index(struct diff_filepair *p)
3845{
3846 return p->score * 100 / MAX_SCORE;
3847}
3848static const char *diff_abbrev_oid(const struct object_id *oid, int abbrev)
3850{
3851 if (startup_info->have_repository)
3852 return find_unique_abbrev(oid->hash, abbrev);
3853 else {
3854 char *hex = oid_to_hex(oid);
3855 if (abbrev < 0)
3856 abbrev = FALLBACK_DEFAULT_ABBREV;
3857 if (abbrev > GIT_SHA1_HEXSZ)
3858 die("BUG: oid abbreviation out of range: %d", abbrev);
3859 if (abbrev)
3860 hex[abbrev] = '\0';
3861 return hex;
3862 }
3863}
3864static void fill_metainfo(struct strbuf *msg,
3866 const char *name,
3867 const char *other,
3868 struct diff_filespec *one,
3869 struct diff_filespec *two,
3870 struct diff_options *o,
3871 struct diff_filepair *p,
3872 int *must_show_header,
3873 int use_color)
3874{
3875 const char *set = diff_get_color(use_color, DIFF_METAINFO);
3876 const char *reset = diff_get_color(use_color, DIFF_RESET);
3877 const char *line_prefix = diff_line_prefix(o);
3878 *must_show_header = 1;
3880 strbuf_init(msg, PATH_MAX * 2 + 300);
3881 switch (p->status) {
3882 case DIFF_STATUS_COPIED:
3883 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3884 line_prefix, set, similarity_index(p));
3885 strbuf_addf(msg, "%s\n%s%scopy from ",
3886 reset, line_prefix, set);
3887 quote_c_style(name, msg, NULL, 0);
3888 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
3889 quote_c_style(other, msg, NULL, 0);
3890 strbuf_addf(msg, "%s\n", reset);
3891 break;
3892 case DIFF_STATUS_RENAMED:
3893 strbuf_addf(msg, "%s%ssimilarity index %d%%",
3894 line_prefix, set, similarity_index(p));
3895 strbuf_addf(msg, "%s\n%s%srename from ",
3896 reset, line_prefix, set);
3897 quote_c_style(name, msg, NULL, 0);
3898 strbuf_addf(msg, "%s\n%s%srename to ",
3899 reset, line_prefix, set);
3900 quote_c_style(other, msg, NULL, 0);
3901 strbuf_addf(msg, "%s\n", reset);
3902 break;
3903 case DIFF_STATUS_MODIFIED:
3904 if (p->score) {
3905 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
3906 line_prefix,
3907 set, similarity_index(p), reset);
3908 break;
3909 }
3910 /* fallthru */
3911 default:
3912 *must_show_header = 0;
3913 }
3914 if (one && two && oidcmp(&one->oid, &two->oid)) {
3915 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
3916 if (DIFF_OPT_TST(o, BINARY)) {
3918 mmfile_t mf;
3919 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
3920 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
3921 abbrev = 40;
3922 }
3923 strbuf_addf(msg, "%s%sindex %s..%s", line_prefix, set,
3924 diff_abbrev_oid(&one->oid, abbrev),
3925 diff_abbrev_oid(&two->oid, abbrev));
3926 if (one->mode == two->mode)
3927 strbuf_addf(msg, " %06o", one->mode);
3928 strbuf_addf(msg, "%s\n", reset);
3929 }
3930}
3931static void run_diff_cmd(const char *pgm,
3933 const char *name,
3934 const char *other,
3935 const char *attr_path,
3936 struct diff_filespec *one,
3937 struct diff_filespec *two,
3938 struct strbuf *msg,
3939 struct diff_options *o,
3940 struct diff_filepair *p)
3941{
3942 const char *xfrm_msg = NULL;
3943 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
3944 int must_show_header = 0;
3945 if (DIFF_OPT_TST(o, ALLOW_EXTERNAL)) {
3948 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3949 if (drv && drv->external)
3950 pgm = drv->external;
3951 }
3952 if (msg) {
3954 /*
3955 * don't use colors when the header is intended for an
3956 * external diff driver
3957 */
3958 fill_metainfo(msg, name, other, one, two, o, p,
3959 &must_show_header,
3960 want_color(o->use_color) && !pgm);
3961 xfrm_msg = msg->len ? msg->buf : NULL;
3962 }
3963 if (pgm) {
3965 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3966 complete_rewrite, o);
3967 return;
3968 }
3969 if (one && two)
3970 builtin_diff(name, other ? other : name,
3971 one, two, xfrm_msg, must_show_header,
3972 o, complete_rewrite);
3973 else
3974 fprintf(o->file, "* Unmerged path %s\n", name);
3975}
3976static void diff_fill_oid_info(struct diff_filespec *one)
3978{
3979 if (DIFF_FILE_VALID(one)) {
3980 if (!one->oid_valid) {
3981 struct stat st;
3982 if (one->is_stdin) {
3983 oidclr(&one->oid);
3984 return;
3985 }
3986 if (lstat(one->path, &st) < 0)
3987 die_errno("stat '%s'", one->path);
3988 if (index_path(one->oid.hash, one->path, &st, 0))
3989 die("cannot hash %s", one->path);
3990 }
3991 }
3992 else
3993 oidclr(&one->oid);
3994}
3995static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3997{
3998 /* Strip the prefix but do not molest /dev/null and absolute paths */
3999 if (*namep && **namep != '/') {
4000 *namep += prefix_length;
4001 if (**namep == '/')
4002 ++*namep;
4003 }
4004 if (*otherp && **otherp != '/') {
4005 *otherp += prefix_length;
4006 if (**otherp == '/')
4007 ++*otherp;
4008 }
4009}
4010static void run_diff(struct diff_filepair *p, struct diff_options *o)
4012{
4013 const char *pgm = external_diff();
4014 struct strbuf msg;
4015 struct diff_filespec *one = p->one;
4016 struct diff_filespec *two = p->two;
4017 const char *name;
4018 const char *other;
4019 const char *attr_path;
4020 name = one->path;
4022 other = (strcmp(name, two->path) ? two->path : NULL);
4023 attr_path = name;
4024 if (o->prefix_length)
4025 strip_prefix(o->prefix_length, &name, &other);
4026 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
4028 pgm = NULL;
4029 if (DIFF_PAIR_UNMERGED(p)) {
4031 run_diff_cmd(pgm, name, NULL, attr_path,
4032 NULL, NULL, NULL, o, p);
4033 return;
4034 }
4035 diff_fill_oid_info(one);
4037 diff_fill_oid_info(two);
4038 if (!pgm &&
4040 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
4041 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
4042 /*
4043 * a filepair that changes between file and symlink
4044 * needs to be split into deletion and creation.
4045 */
4046 struct diff_filespec *null = alloc_filespec(two->path);
4047 run_diff_cmd(NULL, name, other, attr_path,
4048 one, null, &msg, o, p);
4049 free(null);
4050 strbuf_release(&msg);
4051 null = alloc_filespec(one->path);
4053 run_diff_cmd(NULL, name, other, attr_path,
4054 null, two, &msg, o, p);
4055 free(null);
4056 }
4057 else
4058 run_diff_cmd(pgm, name, other, attr_path,
4059 one, two, &msg, o, p);
4060 strbuf_release(&msg);
4062}
4063static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
4065 struct diffstat_t *diffstat)
4066{
4067 const char *name;
4068 const char *other;
4069 if (DIFF_PAIR_UNMERGED(p)) {
4071 /* unmerged */
4072 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, p);
4073 return;
4074 }
4075 name = p->one->path;
4077 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4078 if (o->prefix_length)
4080 strip_prefix(o->prefix_length, &name, &other);
4081 diff_fill_oid_info(p->one);
4083 diff_fill_oid_info(p->two);
4084 builtin_diffstat(name, other, p->one, p->two, diffstat, o, p);
4086}
4087static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
4089{
4090 const char *name;
4091 const char *other;
4092 const char *attr_path;
4093 if (DIFF_PAIR_UNMERGED(p)) {
4095 /* unmerged */
4096 return;
4097 }
4098 name = p->one->path;
4100 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
4101 attr_path = other ? other : name;
4102 if (o->prefix_length)
4104 strip_prefix(o->prefix_length, &name, &other);
4105 diff_fill_oid_info(p->one);
4107 diff_fill_oid_info(p->two);
4108 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
4110}
4111void diff_setup(struct diff_options *options)
4113{
4114 memcpy(options, &default_diff_options, sizeof(*options));
4115 options->file = stdout;
4117 options->abbrev = DEFAULT_ABBREV;
4119 options->line_termination = '\n';
4120 options->break_opt = -1;
4121 options->rename_limit = -1;
4122 options->dirstat_permille = diff_dirstat_permille_default;
4123 options->context = diff_context_default;
4124 options->interhunkcontext = diff_interhunk_context_default;
4125 options->ws_error_highlight = ws_error_highlight_default;
4126 DIFF_OPT_SET(options, RENAME_EMPTY);
4127 /* pathchange left =NULL by default */
4129 options->change = diff_change;
4130 options->add_remove = diff_addremove;
4131 options->use_color = diff_use_color_default;
4132 options->detect_rename = diff_detect_rename_default;
4133 options->xdl_opts |= diff_algorithm;
4134 if (diff_indent_heuristic)
4135 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4136 options->orderfile = diff_order_file_cfg;
4138 if (diff_no_prefix) {
4140 options->a_prefix = options->b_prefix = "";
4141 } else if (!diff_mnemonic_prefix) {
4142 options->a_prefix = "a/";
4143 options->b_prefix = "b/";
4144 }
4145 options->color_moved = diff_color_moved_default;
4147}
4148void diff_setup_done(struct diff_options *options)
4150{
4151 int count = 0;
4152 if (options->set_default)
4154 options->set_default(options);
4155 if (options->output_format & DIFF_FORMAT_NAME)
4157 count++;
4158 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
4159 count++;
4160 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
4161 count++;
4162 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
4163 count++;
4164 if (count > 1)
4165 die(_("--name-only, --name-status, --check and -s are mutually exclusive"));
4166 /*
4168 * Most of the time we can say "there are changes"
4169 * only by checking if there are changed paths, but
4170 * --ignore-whitespace* options force us to look
4171 * inside contents.
4172 */
4173 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
4175 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
4176 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
4177 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
4178 else
4179 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
4180 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
4182 options->detect_rename = DIFF_DETECT_COPY;
4183 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
4185 options->prefix = NULL;
4186 if (options->prefix)
4187 options->prefix_length = strlen(options->prefix);
4188 else
4189 options->prefix_length = 0;
4190 if (options->output_format & (DIFF_FORMAT_NAME |
4192 DIFF_FORMAT_NAME_STATUS |
4193 DIFF_FORMAT_CHECKDIFF |
4194 DIFF_FORMAT_NO_OUTPUT))
4195 options->output_format &= ~(DIFF_FORMAT_RAW |
4196 DIFF_FORMAT_NUMSTAT |
4197 DIFF_FORMAT_DIFFSTAT |
4198 DIFF_FORMAT_SHORTSTAT |
4199 DIFF_FORMAT_DIRSTAT |
4200 DIFF_FORMAT_SUMMARY |
4201 DIFF_FORMAT_PATCH);
4202 /*
4204 * These cases always need recursive; we do not drop caller-supplied
4205 * recursive bits for other formats here.
4206 */
4207 if (options->output_format & (DIFF_FORMAT_PATCH |
4208 DIFF_FORMAT_NUMSTAT |
4209 DIFF_FORMAT_DIFFSTAT |
4210 DIFF_FORMAT_SHORTSTAT |
4211 DIFF_FORMAT_DIRSTAT |
4212 DIFF_FORMAT_SUMMARY |
4213 DIFF_FORMAT_CHECKDIFF))
4214 DIFF_OPT_SET(options, RECURSIVE);
4215 /*
4216 * Also pickaxe would not work very well if you do not say recursive
4217 */
4218 if (options->pickaxe)
4219 DIFF_OPT_SET(options, RECURSIVE);
4220 /*
4221 * When patches are generated, submodules diffed against the work tree
4222 * must be checked for dirtiness too so it can be shown in the output
4223 */
4224 if (options->output_format & DIFF_FORMAT_PATCH)
4225 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
4226 if (options->detect_rename && options->rename_limit < 0)
4228 options->rename_limit = diff_rename_limit_default;
4229 if (options->setup & DIFF_SETUP_USE_CACHE) {
4230 if (!active_cache)
4231 /* read-cache does not die even when it fails
4232 * so it is safe for us to do this here. Also
4233 * it does not smudge active_cache or active_nr
4234 * when it fails, so we do not have to worry about
4235 * cleaning it up ourselves either.
4236 */
4237 read_cache();
4238 }
4239 if (40 < options->abbrev)
4240 options->abbrev = 40; /* full */
4241 /*
4243 * It does not make sense to show the first hit we happened
4244 * to have found. It does not make sense not to return with
4245 * exit code in such a case either.
4246 */
4247 if (DIFF_OPT_TST(options, QUICK)) {
4248 options->output_format = DIFF_FORMAT_NO_OUTPUT;
4249 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4250 }
4251 options->diff_path_counter = 0;
4253 if (DIFF_OPT_TST(options, FOLLOW_RENAMES) && options->pathspec.nr != 1)
4255 die(_("--follow requires exactly one pathspec"));
4256 if (!options->use_color || external_diff())
4258 options->color_moved = 0;
4259}
4260static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
4262{
4263 char c, *eq;
4264 int len;
4265 if (*arg != '-')
4267 return 0;
4268 c = *++arg;
4269 if (!c)
4270 return 0;
4271 if (c == arg_short) {
4272 c = *++arg;
4273 if (!c)
4274 return 1;
4275 if (val && isdigit(c)) {
4276 char *end;
4277 int n = strtoul(arg, &end, 10);
4278 if (*end)
4279 return 0;
4280 *val = n;
4281 return 1;
4282 }
4283 return 0;
4284 }
4285 if (c != '-')
4286 return 0;
4287 arg++;
4288 eq = strchrnul(arg, '=');
4289 len = eq - arg;
4290 if (!len || strncmp(arg, arg_long, len))
4291 return 0;
4292 if (*eq) {
4293 int n;
4294 char *end;
4295 if (!isdigit(*++eq))
4296 return 0;
4297 n = strtoul(eq, &end, 10);
4298 if (*end)
4299 return 0;
4300 *val = n;
4301 }
4302 return 1;
4303}
4304static int diff_scoreopt_parse(const char *opt);
4306static inline int short_opt(char opt, const char **argv,
4308 const char **optarg)
4309{
4310 const char *arg = argv[0];
4311 if (arg[0] != '-' || arg[1] != opt)
4312 return 0;
4313 if (arg[2] != '\0') {
4314 *optarg = arg + 2;
4315 return 1;
4316 }
4317 if (!argv[1])
4318 die("Option '%c' requires a value", opt);
4319 *optarg = argv[1];
4320 return 2;
4321}
4322int parse_long_opt(const char *opt, const char **argv,
4324 const char **optarg)
4325{
4326 const char *arg = argv[0];
4327 if (!skip_prefix(arg, "--", &arg))
4328 return 0;
4329 if (!skip_prefix(arg, opt, &arg))
4330 return 0;
4331 if (*arg == '=') { /* stuck form: --option=value */
4332 *optarg = arg + 1;
4333 return 1;
4334 }
4335 if (*arg != '\0')
4336 return 0;
4337 /* separate form: --option value */
4338 if (!argv[1])
4339 die("Option '--%s' requires a value", opt);
4340 *optarg = argv[1];
4341 return 2;
4342}
4343static int stat_opt(struct diff_options *options, const char **av)
4345{
4346 const char *arg = av[0];
4347 char *end;
4348 int width = options->stat_width;
4349 int name_width = options->stat_name_width;
4350 int graph_width = options->stat_graph_width;
4351 int count = options->stat_count;
4352 int argcount = 1;
4353 if (!skip_prefix(arg, "--stat", &arg))
4355 die("BUG: stat option does not begin with --stat: %s", arg);
4356 end = (char *)arg;
4357 switch (*arg) {
4359 case '-':
4360 if (skip_prefix(arg, "-width", &arg)) {
4361 if (*arg == '=')
4362 width = strtoul(arg + 1, &end, 10);
4363 else if (!*arg && !av[1])
4364 die_want_option("--stat-width");
4365 else if (!*arg) {
4366 width = strtoul(av[1], &end, 10);
4367 argcount = 2;
4368 }
4369 } else if (skip_prefix(arg, "-name-width", &arg)) {
4370 if (*arg == '=')
4371 name_width = strtoul(arg + 1, &end, 10);
4372 else if (!*arg && !av[1])
4373 die_want_option("--stat-name-width");
4374 else if (!*arg) {
4375 name_width = strtoul(av[1], &end, 10);
4376 argcount = 2;
4377 }
4378 } else if (skip_prefix(arg, "-graph-width", &arg)) {
4379 if (*arg == '=')
4380 graph_width = strtoul(arg + 1, &end, 10);
4381 else if (!*arg && !av[1])
4382 die_want_option("--stat-graph-width");
4383 else if (!*arg) {
4384 graph_width = strtoul(av[1], &end, 10);
4385 argcount = 2;
4386 }
4387 } else if (skip_prefix(arg, "-count", &arg)) {
4388 if (*arg == '=')
4389 count = strtoul(arg + 1, &end, 10);
4390 else if (!*arg && !av[1])
4391 die_want_option("--stat-count");
4392 else if (!*arg) {
4393 count = strtoul(av[1], &end, 10);
4394 argcount = 2;
4395 }
4396 }
4397 break;
4398 case '=':
4399 width = strtoul(arg+1, &end, 10);
4400 if (*end == ',')
4401 name_width = strtoul(end+1, &end, 10);
4402 if (*end == ',')
4403 count = strtoul(end+1, &end, 10);
4404 }
4405 /* Important! This checks all the error cases! */
4407 if (*end)
4408 return 0;
4409 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4410 options->stat_name_width = name_width;
4411 options->stat_graph_width = graph_width;
4412 options->stat_width = width;
4413 options->stat_count = count;
4414 return argcount;
4415}
4416static int parse_dirstat_opt(struct diff_options *options, const char *params)
4418{
4419 struct strbuf errmsg = STRBUF_INIT;
4420 if (parse_dirstat_params(options, params, &errmsg))
4421 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
4422 errmsg.buf);
4423 strbuf_release(&errmsg);
4424 /*
4425 * The caller knows a dirstat-related option is given from the command
4426 * line; allow it to say "return this_function();"
4427 */
4428 options->output_format |= DIFF_FORMAT_DIRSTAT;
4429 return 1;
4430}
4431static int parse_submodule_opt(struct diff_options *options, const char *value)
4433{
4434 if (parse_submodule_params(options, value))
4435 die(_("Failed to parse --submodule option parameter: '%s'"),
4436 value);
4437 return 1;
4438}
4439static const char diff_status_letters[] = {
4441 DIFF_STATUS_ADDED,
4442 DIFF_STATUS_COPIED,
4443 DIFF_STATUS_DELETED,
4444 DIFF_STATUS_MODIFIED,
4445 DIFF_STATUS_RENAMED,
4446 DIFF_STATUS_TYPE_CHANGED,
4447 DIFF_STATUS_UNKNOWN,
4448 DIFF_STATUS_UNMERGED,
4449 DIFF_STATUS_FILTER_AON,
4450 DIFF_STATUS_FILTER_BROKEN,
4451 '\0',
4452};
4453static unsigned int filter_bit['Z' + 1];
4455static void prepare_filter_bits(void)
4457{
4458 int i;
4459 if (!filter_bit[DIFF_STATUS_ADDED]) {
4461 for (i = 0; diff_status_letters[i]; i++)
4462 filter_bit[(int) diff_status_letters[i]] = (1 << i);
4463 }
4464}
4465static unsigned filter_bit_tst(char status, const struct diff_options *opt)
4467{
4468 return opt->filter & filter_bit[(int) status];
4469}
4470static int parse_diff_filter_opt(const char *optarg, struct diff_options *opt)
4472{
4473 int i, optch;
4474 prepare_filter_bits();
4476 /*
4478 * If there is a negation e.g. 'd' in the input, and we haven't
4479 * initialized the filter field with another --diff-filter, start
4480 * from full set of bits, except for AON.
4481 */
4482 if (!opt->filter) {
4483 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4484 if (optch < 'a' || 'z' < optch)
4485 continue;
4486 opt->filter = (1 << (ARRAY_SIZE(diff_status_letters) - 1)) - 1;
4487 opt->filter &= ~filter_bit[DIFF_STATUS_FILTER_AON];
4488 break;
4489 }
4490 }
4491 for (i = 0; (optch = optarg[i]) != '\0'; i++) {
4493 unsigned int bit;
4494 int negate;
4495 if ('a' <= optch && optch <= 'z') {
4497 negate = 1;
4498 optch = toupper(optch);
4499 } else {
4500 negate = 0;
4501 }
4502 bit = (0 <= optch && optch <= 'Z') ? filter_bit[optch] : 0;
4504 if (!bit)
4505 return optarg[i];
4506 if (negate)
4507 opt->filter &= ~bit;
4508 else
4509 opt->filter |= bit;
4510 }
4511 return 0;
4512}
4513static void enable_patch_output(int *fmt) {
4515 *fmt &= ~DIFF_FORMAT_NO_OUTPUT;
4516 *fmt |= DIFF_FORMAT_PATCH;
4517}
4518static int parse_ws_error_highlight_opt(struct diff_options *opt, const char *arg)
4520{
4521 int val = parse_ws_error_highlight(arg);
4522 if (val < 0) {
4524 error("unknown value after ws-error-highlight=%.*s",
4525 -1 - val, arg);
4526 return 0;
4527 }
4528 opt->ws_error_highlight = val;
4529 return 1;
4530}
4531int diff_opt_parse(struct diff_options *options,
4533 const char **av, int ac, const char *prefix)
4534{
4535 const char *arg = av[0];
4536 const char *optarg;
4537 int argcount;
4538 if (!prefix)
4540 prefix = "";
4541 /* Output format options */
4543 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch")
4544 || opt_arg(arg, 'U', "unified", &options->context))
4545 enable_patch_output(&options->output_format);
4546 else if (!strcmp(arg, "--raw"))
4547 options->output_format |= DIFF_FORMAT_RAW;
4548 else if (!strcmp(arg, "--patch-with-raw")) {
4549 enable_patch_output(&options->output_format);
4550 options->output_format |= DIFF_FORMAT_RAW;
4551 } else if (!strcmp(arg, "--numstat"))
4552 options->output_format |= DIFF_FORMAT_NUMSTAT;
4553 else if (!strcmp(arg, "--shortstat"))
4554 options->output_format |= DIFF_FORMAT_SHORTSTAT;
4555 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
4556 return parse_dirstat_opt(options, "");
4557 else if (skip_prefix(arg, "-X", &arg))
4558 return parse_dirstat_opt(options, arg);
4559 else if (skip_prefix(arg, "--dirstat=", &arg))
4560 return parse_dirstat_opt(options, arg);
4561 else if (!strcmp(arg, "--cumulative"))
4562 return parse_dirstat_opt(options, "cumulative");
4563 else if (!strcmp(arg, "--dirstat-by-file"))
4564 return parse_dirstat_opt(options, "files");
4565 else if (skip_prefix(arg, "--dirstat-by-file=", &arg)) {
4566 parse_dirstat_opt(options, "files");
4567 return parse_dirstat_opt(options, arg);
4568 }
4569 else if (!strcmp(arg, "--check"))
4570 options->output_format |= DIFF_FORMAT_CHECKDIFF;
4571 else if (!strcmp(arg, "--summary"))
4572 options->output_format |= DIFF_FORMAT_SUMMARY;
4573 else if (!strcmp(arg, "--patch-with-stat")) {
4574 enable_patch_output(&options->output_format);
4575 options->output_format |= DIFF_FORMAT_DIFFSTAT;
4576 } else if (!strcmp(arg, "--name-only"))
4577 options->output_format |= DIFF_FORMAT_NAME;
4578 else if (!strcmp(arg, "--name-status"))
4579 options->output_format |= DIFF_FORMAT_NAME_STATUS;
4580 else if (!strcmp(arg, "-s") || !strcmp(arg, "--no-patch"))
4581 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
4582 else if (starts_with(arg, "--stat"))
4583 /* --stat, --stat-width, --stat-name-width, or --stat-count */
4584 return stat_opt(options, av);
4585 /* renames options */
4587 else if (starts_with(arg, "-B") || starts_with(arg, "--break-rewrites=") ||
4588 !strcmp(arg, "--break-rewrites")) {
4589 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
4590 return error("invalid argument to -B: %s", arg+2);
4591 }
4592 else if (starts_with(arg, "-M") || starts_with(arg, "--find-renames=") ||
4593 !strcmp(arg, "--find-renames")) {
4594 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4595 return error("invalid argument to -M: %s", arg+2);
4596 options->detect_rename = DIFF_DETECT_RENAME;
4597 }
4598 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
4599 options->irreversible_delete = 1;
4600 }
4601 else if (starts_with(arg, "-C") || starts_with(arg, "--find-copies=") ||
4602 !strcmp(arg, "--find-copies")) {
4603 if (options->detect_rename == DIFF_DETECT_COPY)
4604 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4605 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
4606 return error("invalid argument to -C: %s", arg+2);
4607 options->detect_rename = DIFF_DETECT_COPY;
4608 }
4609 else if (!strcmp(arg, "--no-renames"))
4610 options->detect_rename = 0;
4611 else if (!strcmp(arg, "--rename-empty"))
4612 DIFF_OPT_SET(options, RENAME_EMPTY);
4613 else if (!strcmp(arg, "--no-rename-empty"))
4614 DIFF_OPT_CLR(options, RENAME_EMPTY);
4615 else if (!strcmp(arg, "--relative"))
4616 DIFF_OPT_SET(options, RELATIVE_NAME);
4617 else if (skip_prefix(arg, "--relative=", &arg)) {
4618 DIFF_OPT_SET(options, RELATIVE_NAME);
4619 options->prefix = arg;
4620 }
4621 /* xdiff options */
4623 else if (!strcmp(arg, "--minimal"))
4624 DIFF_XDL_SET(options, NEED_MINIMAL);
4625 else if (!strcmp(arg, "--no-minimal"))
4626 DIFF_XDL_CLR(options, NEED_MINIMAL);
4627 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
4628 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
4629 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
4630 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
4631 else if (!strcmp(arg, "--ignore-space-at-eol"))
4632 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
4633 else if (!strcmp(arg, "--ignore-blank-lines"))
4634 DIFF_XDL_SET(options, IGNORE_BLANK_LINES);
4635 else if (!strcmp(arg, "--indent-heuristic"))
4636 DIFF_XDL_SET(options, INDENT_HEURISTIC);
4637 else if (!strcmp(arg, "--no-indent-heuristic"))
4638 DIFF_XDL_CLR(options, INDENT_HEURISTIC);
4639 else if (!strcmp(arg, "--patience"))
4640 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
4641 else if (!strcmp(arg, "--histogram"))
4642 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
4643 else if ((argcount = parse_long_opt("diff-algorithm", av, &optarg))) {
4644 long value = parse_algorithm_value(optarg);
4645 if (value < 0)
4646 return error("option diff-algorithm accepts \"myers\", "
4647 "\"minimal\", \"patience\" and \"histogram\"");
4648 /* clear out previous settings */
4649 DIFF_XDL_CLR(options, NEED_MINIMAL);
4650 options->xdl_opts &= ~XDF_DIFF_ALGORITHM_MASK;
4651 options->xdl_opts |= value;
4652 return argcount;
4653 }
4654 /* flags options */
4656 else if (!strcmp(arg, "--binary")) {
4657 enable_patch_output(&options->output_format);
4658 DIFF_OPT_SET(options, BINARY);
4659 }
4660 else if (!strcmp(arg, "--full-index"))
4661 DIFF_OPT_SET(options, FULL_INDEX);
4662 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
4663 DIFF_OPT_SET(options, TEXT);
4664 else if (!strcmp(arg, "-R"))
4665 DIFF_OPT_SET(options, REVERSE_DIFF);
4666 else if (!strcmp(arg, "--find-copies-harder"))
4667 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
4668 else if (!strcmp(arg, "--follow"))
4669 DIFF_OPT_SET(options, FOLLOW_RENAMES);
4670 else if (!strcmp(arg, "--no-follow")) {
4671 DIFF_OPT_CLR(options, FOLLOW_RENAMES);
4672 DIFF_OPT_CLR(options, DEFAULT_FOLLOW_RENAMES);
4673 } else if (!strcmp(arg, "--color"))
4674 options->use_color = 1;
4675 else if (skip_prefix(arg, "--color=", &arg)) {
4676 int value = git_config_colorbool(NULL, arg);
4677 if (value < 0)
4678 return error("option `color' expects \"always\", \"auto\", or \"never\"");
4679 options->use_color = value;
4680 }
4681 else if (!strcmp(arg, "--no-color"))
4682 options->use_color = 0;
4683 else if (!strcmp(arg, "--color-moved")) {
4684 if (diff_color_moved_default)
4685 options->color_moved = diff_color_moved_default;
4686 if (options->color_moved == COLOR_MOVED_NO)
4687 options->color_moved = COLOR_MOVED_DEFAULT;
4688 } else if (!strcmp(arg, "--no-color-moved"))
4689 options->color_moved = COLOR_MOVED_NO;
4690 else if (skip_prefix(arg, "--color-moved=", &arg)) {
4691 int cm = parse_color_moved(arg);
4692 if (cm < 0)
4693 die("bad --color-moved argument: %s", arg);
4694 options->color_moved = cm;
4695 } else if (!strcmp(arg, "--color-words")) {
4696 options->use_color = 1;
4697 options->word_diff = DIFF_WORDS_COLOR;
4698 }
4699 else if (skip_prefix(arg, "--color-words=", &arg)) {
4700 options->use_color = 1;
4701 options->word_diff = DIFF_WORDS_COLOR;
4702 options->word_regex = arg;
4703 }
4704 else if (!strcmp(arg, "--word-diff")) {
4705 if (options->word_diff == DIFF_WORDS_NONE)
4706 options->word_diff = DIFF_WORDS_PLAIN;
4707 }
4708 else if (skip_prefix(arg, "--word-diff=", &arg)) {
4709 if (!strcmp(arg, "plain"))
4710 options->word_diff = DIFF_WORDS_PLAIN;
4711 else if (!strcmp(arg, "color")) {
4712 options->use_color = 1;
4713 options->word_diff = DIFF_WORDS_COLOR;
4714 }
4715 else if (!strcmp(arg, "porcelain"))
4716 options->word_diff = DIFF_WORDS_PORCELAIN;
4717 else if (!strcmp(arg, "none"))
4718 options->word_diff = DIFF_WORDS_NONE;
4719 else
4720 die("bad --word-diff argument: %s", arg);
4721 }
4722 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
4723 if (options->word_diff == DIFF_WORDS_NONE)
4724 options->word_diff = DIFF_WORDS_PLAIN;
4725 options->word_regex = optarg;
4726 return argcount;
4727 }
4728 else if (!strcmp(arg, "--exit-code"))
4729 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
4730 else if (!strcmp(arg, "--quiet"))
4731 DIFF_OPT_SET(options, QUICK);
4732 else if (!strcmp(arg, "--ext-diff"))
4733 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
4734 else if (!strcmp(arg, "--no-ext-diff"))
4735 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
4736 else if (!strcmp(arg, "--textconv"))
4737 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
4738 else if (!strcmp(arg, "--no-textconv"))
4739 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
4740 else if (!strcmp(arg, "--ignore-submodules")) {
4741 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4742 handle_ignore_submodules_arg(options, "all");
4743 } else if (skip_prefix(arg, "--ignore-submodules=", &arg)) {
4744 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
4745 handle_ignore_submodules_arg(options, arg);
4746 } else if (!strcmp(arg, "--submodule"))
4747 options->submodule_format = DIFF_SUBMODULE_LOG;
4748 else if (skip_prefix(arg, "--submodule=", &arg))
4749 return parse_submodule_opt(options, arg);
4750 else if (skip_prefix(arg, "--ws-error-highlight=", &arg))
4751 return parse_ws_error_highlight_opt(options, arg);
4752 else if (!strcmp(arg, "--ita-invisible-in-index"))
4753 options->ita_invisible_in_index = 1;
4754 else if (!strcmp(arg, "--ita-visible-in-index"))
4755 options->ita_invisible_in_index = 0;
4756 /* misc options */
4758 else if (!strcmp(arg, "-z"))
4759 options->line_termination = 0;
4760 else if ((argcount = short_opt('l', av, &optarg))) {
4761 options->rename_limit = strtoul(optarg, NULL, 10);
4762 return argcount;
4763 }
4764 else if ((argcount = short_opt('S', av, &optarg))) {
4765 options->pickaxe = optarg;
4766 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
4767 return argcount;
4768 } else if ((argcount = short_opt('G', av, &optarg))) {
4769 options->pickaxe = optarg;
4770 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
4771 return argcount;
4772 }
4773 else if (!strcmp(arg, "--pickaxe-all"))
4774 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
4775 else if (!strcmp(arg, "--pickaxe-regex"))
4776 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
4777 else if ((argcount = short_opt('O', av, &optarg))) {
4778 options->orderfile = prefix_filename(prefix, optarg);
4779 return argcount;
4780 }
4781 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
4782 int offending = parse_diff_filter_opt(optarg, options);
4783 if (offending)
4784 die("unknown change class '%c' in --diff-filter=%s",
4785 offending, optarg);
4786 return argcount;
4787 }
4788 else if (!strcmp(arg, "--no-abbrev"))
4789 options->abbrev = 0;
4790 else if (!strcmp(arg, "--abbrev"))
4791 options->abbrev = DEFAULT_ABBREV;
4792 else if (skip_prefix(arg, "--abbrev=", &arg)) {
4793 options->abbrev = strtoul(arg, NULL, 10);
4794 if (options->abbrev < MINIMUM_ABBREV)
4795 options->abbrev = MINIMUM_ABBREV;
4796 else if (40 < options->abbrev)
4797 options->abbrev = 40;
4798 }
4799 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
4800 options->a_prefix = optarg;
4801 return argcount;
4802 }
4803 else if ((argcount = parse_long_opt("line-prefix", av, &optarg))) {
4804 options->line_prefix = optarg;
4805 options->line_prefix_length = strlen(options->line_prefix);
4806 graph_setup_line_prefix(options);
4807 return argcount;
4808 }
4809 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
4810 options->b_prefix = optarg;
4811 return argcount;
4812 }
4813 else if (!strcmp(arg, "--no-prefix"))
4814 options->a_prefix = options->b_prefix = "";
4815 else if (opt_arg(arg, '\0', "inter-hunk-context",
4816 &options->interhunkcontext))
4817 ;
4818 else if (!strcmp(arg, "-W"))
4819 DIFF_OPT_SET(options, FUNCCONTEXT);
4820 else if (!strcmp(arg, "--function-context"))
4821 DIFF_OPT_SET(options, FUNCCONTEXT);
4822 else if (!strcmp(arg, "--no-function-context"))
4823 DIFF_OPT_CLR(options, FUNCCONTEXT);
4824 else if ((argcount = parse_long_opt("output", av, &optarg))) {
4825 char *path = prefix_filename(prefix, optarg);
4826 options->file = xfopen(path, "w");
4827 options->close_file = 1;
4828 if (options->use_color != GIT_COLOR_ALWAYS)
4829 options->use_color = GIT_COLOR_NEVER;
4830 free(path);
4831 return argcount;
4832 } else
4833 return 0;
4834 return 1;
4835}
4836int parse_rename_score(const char **cp_p)
4838{
4839 unsigned long num, scale;
4840 int ch, dot;
4841 const char *cp = *cp_p;
4842 num = 0;
4844 scale = 1;
4845 dot = 0;
4846 for (;;) {
4847 ch = *cp;
4848 if ( !dot && ch == '.' ) {
4849 scale = 1;
4850 dot = 1;
4851 } else if ( ch == '%' ) {
4852 scale = dot ? scale*100 : 100;
4853 cp++; /* % is always at the end */
4854 break;
4855 } else if ( ch >= '0' && ch <= '9' ) {
4856 if ( scale < 100000 ) {
4857 scale *= 10;
4858 num = (num*10) + (ch-'0');
4859 }
4860 } else {
4861 break;
4862 }
4863 cp++;
4864 }
4865 *cp_p = cp;
4866 /* user says num divided by scale and we say internally that
4868 * is MAX_SCORE * num / scale.
4869 */
4870 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
4871}
4872static int diff_scoreopt_parse(const char *opt)
4874{
4875 int opt1, opt2, cmd;
4876 if (*opt++ != '-')
4878 return -1;
4879 cmd = *opt++;
4880 if (cmd == '-') {
4881 /* convert the long-form arguments into short-form versions */
4882 if (skip_prefix(opt, "break-rewrites", &opt)) {
4883 if (*opt == 0 || *opt++ == '=')
4884 cmd = 'B';
4885 } else if (skip_prefix(opt, "find-copies", &opt)) {
4886 if (*opt == 0 || *opt++ == '=')
4887 cmd = 'C';
4888 } else if (skip_prefix(opt, "find-renames", &opt)) {
4889 if (*opt == 0 || *opt++ == '=')
4890 cmd = 'M';
4891 }
4892 }
4893 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
4894 return -1; /* that is not a -M, -C, or -B option */
4895 opt1 = parse_rename_score(&opt);
4897 if (cmd != 'B')
4898 opt2 = 0;
4899 else {
4900 if (*opt == 0)
4901 opt2 = 0;
4902 else if (*opt != '/')
4903 return -1; /* we expect -B80/99 or -B80 */
4904 else {
4905 opt++;
4906 opt2 = parse_rename_score(&opt);
4907 }
4908 }
4909 if (*opt != 0)
4910 return -1;
4911 return opt1 | (opt2 << 16);
4912}
4913struct diff_queue_struct diff_queued_diff;
4915void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
4917{
4918 ALLOC_GROW(queue->queue, queue->nr + 1, queue->alloc);
4919 queue->queue[queue->nr++] = dp;
4920}
4921struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
4923 struct diff_filespec *one,
4924 struct diff_filespec *two)
4925{
4926 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
4927 dp->one = one;
4928 dp->two = two;
4929 if (queue)
4930 diff_q(queue, dp);
4931 return dp;
4932}
4933void diff_free_filepair(struct diff_filepair *p)
4935{
4936 free_filespec(p->one);
4937 free_filespec(p->two);
4938 free(p);
4939}
4940const char *diff_aligned_abbrev(const struct object_id *oid, int len)
4942{
4943 int abblen;
4944 const char *abbrev;
4945 if (len == GIT_SHA1_HEXSZ)
4947 return oid_to_hex(oid);
4948 abbrev = diff_abbrev_oid(oid, len);
4950 abblen = strlen(abbrev);
4951 /*
4953 * In well-behaved cases, where the abbbreviated result is the
4954 * same as the requested length, append three dots after the
4955 * abbreviation (hence the whole logic is limited to the case
4956 * where abblen < 37); when the actual abbreviated result is a
4957 * bit longer than the requested length, we reduce the number
4958 * of dots so that they match the well-behaved ones. However,
4959 * if the actual abbreviation is longer than the requested
4960 * length by more than three, we give up on aligning, and add
4961 * three dots anyway, to indicate that the output is not the
4962 * full object name. Yes, this may be suboptimal, but this
4963 * appears only in "diff --raw --abbrev" output and it is not
4964 * worth the effort to change it now. Note that this would
4965 * likely to work fine when the automatic sizing of default
4966 * abbreviation length is used--we would be fed -1 in "len" in
4967 * that case, and will end up always appending three-dots, but
4968 * the automatic sizing is supposed to give abblen that ensures
4969 * uniqueness across all objects (statistically speaking).
4970 */
4971 if (abblen < GIT_SHA1_HEXSZ - 3) {
4972 static char hex[GIT_MAX_HEXSZ + 1];
4973 if (len < abblen && abblen <= len + 2)
4974 xsnprintf(hex, sizeof(hex), "%s%.*s", abbrev, len+3-abblen, "..");
4975 else
4976 xsnprintf(hex, sizeof(hex), "%s...", abbrev);
4977 return hex;
4978 }
4979 return oid_to_hex(oid);
4981}
4982static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
4984{
4985 int line_termination = opt->line_termination;
4986 int inter_name_termination = line_termination ? '\t' : '\0';
4987 fprintf(opt->file, "%s", diff_line_prefix(opt));
4989 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
4990 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
4991 diff_aligned_abbrev(&p->one->oid, opt->abbrev));
4992 fprintf(opt->file, "%s ",
4993 diff_aligned_abbrev(&p->two->oid, opt->abbrev));
4994 }
4995 if (p->score) {
4996 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
4997 inter_name_termination);
4998 } else {
4999 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
5000 }
5001 if (p->status == DIFF_STATUS_COPIED ||
5003 p->status == DIFF_STATUS_RENAMED) {
5004 const char *name_a, *name_b;
5005 name_a = p->one->path;
5006 name_b = p->two->path;
5007 strip_prefix(opt->prefix_length, &name_a, &name_b);
5008 write_name_quoted(name_a, opt->file, inter_name_termination);
5009 write_name_quoted(name_b, opt->file, line_termination);
5010 } else {
5011 const char *name_a, *name_b;
5012 name_a = p->one->mode ? p->one->path : p->two->path;
5013 name_b = NULL;
5014 strip_prefix(opt->prefix_length, &name_a, &name_b);
5015 write_name_quoted(name_a, opt->file, line_termination);
5016 }
5017}
5018int diff_unmodified_pair(struct diff_filepair *p)
5020{
5021 /* This function is written stricter than necessary to support
5022 * the currently implemented transformers, but the idea is to
5023 * let transformers to produce diff_filepairs any way they want,
5024 * and filter and clean them up here before producing the output.
5025 */
5026 struct diff_filespec *one = p->one, *two = p->two;
5027 if (DIFF_PAIR_UNMERGED(p))
5029 return 0; /* unmerged is interesting */
5030 /* deletion, addition, mode or type change
5032 * and rename are all interesting.
5033 */
5034 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
5035 DIFF_PAIR_MODE_CHANGED(p) ||
5036 strcmp(one->path, two->path))
5037 return 0;
5038 /* both are valid and point at the same path. that is, we are
5040 * dealing with a change.
5041 */
5042 if (one->oid_valid && two->oid_valid &&
5043 !oidcmp(&one->oid, &two->oid) &&
5044 !one->dirty_submodule && !two->dirty_submodule)
5045 return 1; /* no change */
5046 if (!one->oid_valid && !two->oid_valid)
5047 return 1; /* both look at the same file on the filesystem. */
5048 return 0;
5049}
5050static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
5052{
5053 if (diff_unmodified_pair(p))
5054 return;
5055 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5057 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5058 return; /* no tree diffs in patch format */
5059 run_diff(p, o);
5061}
5062static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
5064 struct diffstat_t *diffstat)
5065{
5066 if (diff_unmodified_pair(p))
5067 return;
5068 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5070 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5071 return; /* no useful stat for tree diffs */
5072 run_diffstat(p, o, diffstat);
5074}
5075static void diff_flush_checkdiff(struct diff_filepair *p,
5077 struct diff_options *o)
5078{
5079 if (diff_unmodified_pair(p))
5080 return;
5081 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5083 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5084 return; /* nothing to check in tree diffs */
5085 run_checkdiff(p, o);
5087}
5088int diff_queue_is_empty(void)
5090{
5091 struct diff_queue_struct *q = &diff_queued_diff;
5092 int i;
5093 for (i = 0; i < q->nr; i++)
5094 if (!diff_unmodified_pair(q->queue[i]))
5095 return 0;
5096 return 1;
5097}
5098#if DIFF_DEBUG
5100void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
5101{
5102 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
5103 x, one ? one : "",
5104 s->path,
5105 DIFF_FILE_VALID(s) ? "valid" : "invalid",
5106 s->mode,
5107 s->oid_valid ? oid_to_hex(&s->oid) : "");
5108 fprintf(stderr, "queue[%d] %s size %lu\n",
5109 x, one ? one : "",
5110 s->size);
5111}
5112void diff_debug_filepair(const struct diff_filepair *p, int i)
5114{
5115 diff_debug_filespec(p->one, i, "one");
5116 diff_debug_filespec(p->two, i, "two");
5117 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
5118 p->score, p->status ? p->status : '?',
5119 p->one->rename_used, p->broken_pair);
5120}
5121void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
5123{
5124 int i;
5125 if (msg)
5126 fprintf(stderr, "%s\n", msg);
5127 fprintf(stderr, "q->nr = %d\n", q->nr);
5128 for (i = 0; i < q->nr; i++) {
5129 struct diff_filepair *p = q->queue[i];
5130 diff_debug_filepair(p, i);
5131 }
5132}
5133#endif
5134static void diff_resolve_rename_copy(void)
5136{
5137 int i;
5138 struct diff_filepair *p;
5139 struct diff_queue_struct *q = &diff_queued_diff;
5140 diff_debug_queue("resolve-rename-copy", q);
5142 for (i = 0; i < q->nr; i++) {
5144 p = q->queue[i];
5145 p->status = 0; /* undecided */
5146 if (DIFF_PAIR_UNMERGED(p))
5147 p->status = DIFF_STATUS_UNMERGED;
5148 else if (!DIFF_FILE_VALID(p->one))
5149 p->status = DIFF_STATUS_ADDED;
5150 else if (!DIFF_FILE_VALID(p->two))
5151 p->status = DIFF_STATUS_DELETED;
5152 else if (DIFF_PAIR_TYPE_CHANGED(p))
5153 p->status = DIFF_STATUS_TYPE_CHANGED;
5154 /* from this point on, we are dealing with a pair
5156 * whose both sides are valid and of the same type, i.e.
5157 * either in-place edit or rename/copy edit.
5158 */
5159 else if (DIFF_PAIR_RENAME(p)) {
5160 /*
5161 * A rename might have re-connected a broken
5162 * pair up, causing the pathnames to be the
5163 * same again. If so, that's not a rename at
5164 * all, just a modification..
5165 *
5166 * Otherwise, see if this source was used for
5167 * multiple renames, in which case we decrement
5168 * the count, and call it a copy.
5169 */
5170 if (!strcmp(p->one->path, p->two->path))
5171 p->status = DIFF_STATUS_MODIFIED;
5172 else if (--p->one->rename_used > 0)
5173 p->status = DIFF_STATUS_COPIED;
5174 else
5175 p->status = DIFF_STATUS_RENAMED;
5176 }
5177 else if (oidcmp(&p->one->oid, &p->two->oid) ||
5178 p->one->mode != p->two->mode ||
5179 p->one->dirty_submodule ||
5180 p->two->dirty_submodule ||
5181 is_null_oid(&p->one->oid))
5182 p->status = DIFF_STATUS_MODIFIED;
5183 else {
5184 /* This is a "no-change" entry and should not
5185 * happen anymore, but prepare for broken callers.
5186 */
5187 error("feeding unmodified %s to diffcore",
5188 p->one->path);
5189 p->status = DIFF_STATUS_UNKNOWN;
5190 }
5191 }
5192 diff_debug_queue("resolve-rename-copy done", q);
5193}
5194static int check_pair_status(struct diff_filepair *p)
5196{
5197 switch (p->status) {
5198 case DIFF_STATUS_UNKNOWN:
5199 return 0;
5200 case 0:
5201 die("internal error in diff-resolve-rename-copy");
5202 default:
5203 return 1;
5204 }
5205}
5206static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
5208{
5209 int fmt = opt->output_format;
5210 if (fmt & DIFF_FORMAT_CHECKDIFF)
5212 diff_flush_checkdiff(p, opt);
5213 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
5214 diff_flush_raw(p, opt);
5215 else if (fmt & DIFF_FORMAT_NAME) {
5216 const char *name_a, *name_b;
5217 name_a = p->two->path;
5218 name_b = NULL;
5219 strip_prefix(opt->prefix_length, &name_a, &name_b);
5220 fprintf(opt->file, "%s", diff_line_prefix(opt));
5221 write_name_quoted(name_a, opt->file, opt->line_termination);
5222 }
5223}
5224static void show_file_mode_name(struct diff_options *opt, const char *newdelete, struct diff_filespec *fs)
5226{
5227 struct strbuf sb = STRBUF_INIT;
5228 if (fs->mode)
5229 strbuf_addf(&sb, " %s mode %06o ", newdelete, fs->mode);
5230 else
5231 strbuf_addf(&sb, " %s ", newdelete);
5232 quote_c_style(fs->path, &sb, NULL, 0);
5234 strbuf_addch(&sb, '\n');
5235 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5236 sb.buf, sb.len, 0);
5237 strbuf_release(&sb);
5238}
5239static void show_mode_change(struct diff_options *opt, struct diff_filepair *p,
5241 int show_name)
5242{
5243 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
5244 struct strbuf sb = STRBUF_INIT;
5245 strbuf_addf(&sb, " mode change %06o => %06o",
5246 p->one->mode, p->two->mode);
5247 if (show_name) {
5248 strbuf_addch(&sb, ' ');
5249 quote_c_style(p->two->path, &sb, NULL, 0);
5250 }
5251 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5252 sb.buf, sb.len, 0);
5253 strbuf_release(&sb);
5254 }
5255}
5256static void show_rename_copy(struct diff_options *opt, const char *renamecopy,
5258 struct diff_filepair *p)
5259{
5260 struct strbuf sb = STRBUF_INIT;
5261 char *names = pprint_rename(p->one->path, p->two->path);
5262 strbuf_addf(&sb, " %s %s (%d%%)\n",
5263 renamecopy, names, similarity_index(p));
5264 free(names);
5265 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5266 sb.buf, sb.len, 0);
5267 show_mode_change(opt, p, 0);
5268}
5269static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
5271{
5272 switch(p->status) {
5273 case DIFF_STATUS_DELETED:
5274 show_file_mode_name(opt, "delete", p->one);
5275 break;
5276 case DIFF_STATUS_ADDED:
5277 show_file_mode_name(opt, "create", p->two);
5278 break;
5279 case DIFF_STATUS_COPIED:
5280 show_rename_copy(opt, "copy", p);
5281 break;
5282 case DIFF_STATUS_RENAMED:
5283 show_rename_copy(opt, "rename", p);
5284 break;
5285 default:
5286 if (p->score) {
5287 struct strbuf sb = STRBUF_INIT;
5288 strbuf_addstr(&sb, " rewrite ");
5289 quote_c_style(p->two->path, &sb, NULL, 0);
5290 strbuf_addf(&sb, " (%d%%)\n", similarity_index(p));
5291 emit_diff_symbol(opt, DIFF_SYMBOL_SUMMARY,
5292 sb.buf, sb.len, 0);
5293 }
5294 show_mode_change(opt, p, !p->score);
5295 break;
5296 }
5297}
5298struct patch_id_t {
5300 git_SHA_CTX *ctx;
5301 int patchlen;
5302};
5303static int remove_space(char *line, int len)
5305{
5306 int i;
5307 char *dst = line;
5308 unsigned char c;
5309 for (i = 0; i < len; i++)
5311 if (!isspace((c = line[i])))
5312 *dst++ = c;
5313 return dst - line;
5315}
5316static void patch_id_consume(void *priv, char *line, unsigned long len)
5318{
5319 struct patch_id_t *data = priv;
5320 int new_len;
5321 /* Ignore line numbers when computing the SHA1 of the patch */
5323 if (starts_with(line, "@@ -"))
5324 return;
5325 new_len = remove_space(line, len);
5327 git_SHA1_Update(data->ctx, line, new_len);
5329 data->patchlen += new_len;
5330}
5331static void patch_id_add_string(git_SHA_CTX *ctx, const char *str)
5333{
5334 git_SHA1_Update(ctx, str, strlen(str));
5335}
5336static void patch_id_add_mode(git_SHA_CTX *ctx, unsigned mode)
5338{
5339 /* large enough for 2^32 in octal */
5340 char buf[12];
5341 int len = xsnprintf(buf, sizeof(buf), "%06o", mode);
5342 git_SHA1_Update(ctx, buf, len);
5343}
5344/* returns 0 upon success, and writes result into sha1 */
5346static int diff_get_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5347{
5348 struct diff_queue_struct *q = &diff_queued_diff;
5349 int i;
5350 git_SHA_CTX ctx;
5351 struct patch_id_t data;
5352 git_SHA1_Init(&ctx);
5354 memset(&data, 0, sizeof(struct patch_id_t));
5355 data.ctx = &ctx;
5356 for (i = 0; i < q->nr; i++) {
5358 xpparam_t xpp;
5359 xdemitconf_t xecfg;
5360 mmfile_t mf1, mf2;
5361 struct diff_filepair *p = q->queue[i];
5362 int len1, len2;
5363 memset(&xpp, 0, sizeof(xpp));
5365 memset(&xecfg, 0, sizeof(xecfg));
5366 if (p->status == 0)
5367 return error("internal diff status error");
5368 if (p->status == DIFF_STATUS_UNKNOWN)
5369 continue;
5370 if (diff_unmodified_pair(p))
5371 continue;
5372 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
5373 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
5374 continue;
5375 if (DIFF_PAIR_UNMERGED(p))
5376 continue;
5377 diff_fill_oid_info(p->one);
5379 diff_fill_oid_info(p->two);
5380 len1 = remove_space(p->one->path, strlen(p->one->path));
5382 len2 = remove_space(p->two->path, strlen(p->two->path));
5383 patch_id_add_string(&ctx, "diff--git");
5384 patch_id_add_string(&ctx, "a/");
5385 git_SHA1_Update(&ctx, p->one->path, len1);
5386 patch_id_add_string(&ctx, "b/");
5387 git_SHA1_Update(&ctx, p->two->path, len2);
5388 if (p->one->mode == 0) {
5390 patch_id_add_string(&ctx, "newfilemode");
5391 patch_id_add_mode(&ctx, p->two->mode);
5392 patch_id_add_string(&ctx, "---/dev/null");
5393 patch_id_add_string(&ctx, "+++b/");
5394 git_SHA1_Update(&ctx, p->two->path, len2);
5395 } else if (p->two->mode == 0) {
5396 patch_id_add_string(&ctx, "deletedfilemode");
5397 patch_id_add_mode(&ctx, p->one->mode);
5398 patch_id_add_string(&ctx, "---a/");
5399 git_SHA1_Update(&ctx, p->one->path, len1);
5400 patch_id_add_string(&ctx, "+++/dev/null");
5401 } else {
5402 patch_id_add_string(&ctx, "---a/");
5403 git_SHA1_Update(&ctx, p->one->path, len1);
5404 patch_id_add_string(&ctx, "+++b/");
5405 git_SHA1_Update(&ctx, p->two->path, len2);
5406 }
5407 if (diff_header_only)
5409 continue;
5410 if (fill_mmfile(&mf1, p->one) < 0 ||
5412 fill_mmfile(&mf2, p->two) < 0)
5413 return error("unable to read files to diff");
5414 if (diff_filespec_is_binary(p->one) ||
5416 diff_filespec_is_binary(p->two)) {
5417 git_SHA1_Update(&ctx, oid_to_hex(&p->one->oid),
5418 GIT_SHA1_HEXSZ);
5419 git_SHA1_Update(&ctx, oid_to_hex(&p->two->oid),
5420 GIT_SHA1_HEXSZ);
5421 continue;
5422 }
5423 xpp.flags = 0;
5425 xecfg.ctxlen = 3;
5426 xecfg.flags = 0;
5427 if (xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
5428 &xpp, &xecfg))
5429 return error("unable to generate patch-id diff for %s",
5430 p->one->path);
5431 }
5432 git_SHA1_Final(oid->hash, &ctx);
5434 return 0;
5435}
5436int diff_flush_patch_id(struct diff_options *options, struct object_id *oid, int diff_header_only)
5438{
5439 struct diff_queue_struct *q = &diff_queued_diff;
5440 int i;
5441 int result = diff_get_patch_id(options, oid, diff_header_only);
5442 for (i = 0; i < q->nr; i++)
5444 diff_free_filepair(q->queue[i]);
5445 free(q->queue);
5447 DIFF_QUEUE_CLEAR(q);
5448 return result;
5450}
5451static int is_summary_empty(const struct diff_queue_struct *q)
5453{
5454 int i;
5455 for (i = 0; i < q->nr; i++) {
5457 const struct diff_filepair *p = q->queue[i];
5458 switch (p->status) {
5460 case DIFF_STATUS_DELETED:
5461 case DIFF_STATUS_ADDED:
5462 case DIFF_STATUS_COPIED:
5463 case DIFF_STATUS_RENAMED:
5464 return 0;
5465 default:
5466 if (p->score)
5467 return 0;
5468 if (p->one->mode && p->two->mode &&
5469 p->one->mode != p->two->mode)
5470 return 0;
5471 break;
5472 }
5473 }
5474 return 1;
5475}
5476static const char rename_limit_warning[] =
5478N_("inexact rename detection was skipped due to too many files.");
5479static const char degrade_cc_to_c_warning[] =
5481N_("only found copies from modified paths due to too many files.");
5482static const char rename_limit_advice[] =
5484N_("you may want to set your %s variable to at least "
5485 "%d and retry the command.");
5486void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
5488{
5489 if (degraded_cc)
5490 warning(_(degrade_cc_to_c_warning));
5491 else if (needed)
5492 warning(_(rename_limit_warning));
5493 else
5494 return;
5495 if (0 < needed && needed < 32767)
5496 warning(_(rename_limit_advice), varname, needed);
5497}
5498static void diff_flush_patch_all_file_pairs(struct diff_options *o)
5500{
5501 int i;
5502 static struct emitted_diff_symbols esm = EMITTED_DIFF_SYMBOLS_INIT;
5503 struct diff_queue_struct *q = &diff_queued_diff;
5504 if (WSEH_NEW & WS_RULE_MASK)
5506 die("BUG: WS rules bit mask overlaps with diff symbol flags");
5507 if (o->color_moved)
5509 o->emitted_symbols = &esm;
5510 for (i = 0; i < q->nr; i++) {
5512 struct diff_filepair *p = q->queue[i];
5513 if (check_pair_status(p))
5514 diff_flush_patch(p, o);
5515 }
5516 if (o->emitted_symbols) {
5518 if (o->color_moved) {
5519 struct hashmap add_lines, del_lines;
5520 hashmap_init(&del_lines,
5522 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5523 hashmap_init(&add_lines,
5524 (hashmap_cmp_fn)moved_entry_cmp, o, 0);
5525 add_lines_to_move_detection(o, &add_lines, &del_lines);
5527 mark_color_as_moved(o, &add_lines, &del_lines);
5528 if (o->color_moved == COLOR_MOVED_ZEBRA_DIM)
5529 dim_moved_lines(o);
5530 hashmap_free(&add_lines, 0);
5532 hashmap_free(&del_lines, 0);
5533 }
5534 for (i = 0; i < esm.nr; i++)
5536 emit_diff_symbol_from_struct(o, &esm.buf[i]);
5537 for (i = 0; i < esm.nr; i++)
5539 free((void *)esm.buf[i].line);
5540 }
5541 esm.nr = 0;
5542}
5543void diff_flush(struct diff_options *options)
5545{
5546 struct diff_queue_struct *q = &diff_queued_diff;
5547 int i, output_format = options->output_format;
5548 int separator = 0;
5549 int dirstat_by_line = 0;
5550 /*
5552 * Order: raw, stat, summary, patch
5553 * or: name/name-status/checkdiff (other bits clear)
5554 */
5555 if (!q->nr)
5556 goto free_queue;
5557 if (output_format & (DIFF_FORMAT_RAW |
5559 DIFF_FORMAT_NAME |
5560 DIFF_FORMAT_NAME_STATUS |
5561 DIFF_FORMAT_CHECKDIFF)) {
5562 for (i = 0; i < q->nr; i++) {
5563 struct diff_filepair *p = q->queue[i];
5564 if (check_pair_status(p))
5565 flush_one_pair(p, options);
5566 }
5567 separator++;
5568 }
5569 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
5571 dirstat_by_line = 1;
5572 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
5574 dirstat_by_line) {
5575 struct diffstat_t diffstat;
5576 memset(&diffstat, 0, sizeof(struct diffstat_t));
5578 for (i = 0; i < q->nr; i++) {
5579 struct diff_filepair *p = q->queue[i];
5580 if (check_pair_status(p))
5581 diff_flush_stat(p, options, &diffstat);
5582 }
5583 if (output_format & DIFF_FORMAT_NUMSTAT)
5584 show_numstat(&diffstat, options);
5585 if (output_format & DIFF_FORMAT_DIFFSTAT)
5586 show_stats(&diffstat, options);
5587 if (output_format & DIFF_FORMAT_SHORTSTAT)
5588 show_shortstats(&diffstat, options);
5589 if (output_format & DIFF_FORMAT_DIRSTAT && dirstat_by_line)
5590 show_dirstat_by_line(&diffstat, options);
5591 free_diffstat_info(&diffstat);
5592 separator++;
5593 }
5594 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
5595 show_dirstat(options);
5596 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
5598 for (i = 0; i < q->nr; i++) {
5599 diff_summary(options, q->queue[i]);
5600 }
5601 separator++;
5602 }
5603 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
5605 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
5606 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5607 /*
5608 * run diff_flush_patch for the exit status. setting
5609 * options->file to /dev/null should be safe, because we
5610 * aren't supposed to produce any output anyway.
5611 */
5612 if (options->close_file)
5613 fclose(options->file);
5614 options->file = xfopen("/dev/null", "w");
5615 options->close_file = 1;
5616 options->color_moved = 0;
5617 for (i = 0; i < q->nr; i++) {
5618 struct diff_filepair *p = q->queue[i];
5619 if (check_pair_status(p))
5620 diff_flush_patch(p, options);
5621 if (options->found_changes)
5622 break;
5623 }
5624 }
5625 if (output_format & DIFF_FORMAT_PATCH) {
5627 if (separator) {
5628 emit_diff_symbol(options, DIFF_SYMBOL_SEPARATOR, NULL, 0, 0);
5629 if (options->stat_sep)
5630 /* attach patch instead of inline */
5631 emit_diff_symbol(options, DIFF_SYMBOL_STAT_SEP,
5632 NULL, 0, 0);
5633 }
5634 diff_flush_patch_all_file_pairs(options);
5636 }
5637 if (output_format & DIFF_FORMAT_CALLBACK)
5639 options->format_callback(q, options, options->format_callback_data);
5640 for (i = 0; i < q->nr; i++)
5642 diff_free_filepair(q->queue[i]);
5643free_queue:
5644 free(q->queue);
5645 DIFF_QUEUE_CLEAR(q);
5646 if (options->close_file)
5647 fclose(options->file);
5648 /*
5650 * Report the content-level differences with HAS_CHANGES;
5651 * diff_addremove/diff_change does not set the bit when
5652 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
5653 */
5654 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
5655 if (options->found_changes)
5656 DIFF_OPT_SET(options, HAS_CHANGES);
5657 else
5658 DIFF_OPT_CLR(options, HAS_CHANGES);
5659 }
5660}
5661static int match_filter(const struct diff_options *options, const struct diff_filepair *p)
5663{
5664 return (((p->status == DIFF_STATUS_MODIFIED) &&
5665 ((p->score &&
5666 filter_bit_tst(DIFF_STATUS_FILTER_BROKEN, options)) ||
5667 (!p->score &&
5668 filter_bit_tst(DIFF_STATUS_MODIFIED, options)))) ||
5669 ((p->status != DIFF_STATUS_MODIFIED) &&
5670 filter_bit_tst(p->status, options)));
5671}
5672static void diffcore_apply_filter(struct diff_options *options)
5674{
5675 int i;
5676 struct diff_queue_struct *q = &diff_queued_diff;
5677 struct diff_queue_struct outq;
5678 DIFF_QUEUE_CLEAR(&outq);
5680 if (!options->filter)
5682 return;
5683 if (filter_bit_tst(DIFF_STATUS_FILTER_AON, options)) {
5685 int found;
5686 for (i = found = 0; !found && i < q->nr; i++) {
5687 if (match_filter(options, q->queue[i]))
5688 found++;
5689 }
5690 if (found)
5691 return;
5692 /* otherwise we will clear the whole queue
5694 * by copying the empty outq at the end of this
5695 * function, but first clear the current entries
5696 * in the queue.
5697 */
5698 for (i = 0; i < q->nr; i++)
5699 diff_free_filepair(q->queue[i]);
5700 }
5701 else {
5702 /* Only the matching ones */
5703 for (i = 0; i < q->nr; i++) {
5704 struct diff_filepair *p = q->queue[i];
5705 if (match_filter(options, p))
5706 diff_q(&outq, p);
5707 else
5708 diff_free_filepair(p);
5709 }
5710 }
5711 free(q->queue);
5712 *q = outq;
5713}
5714/* Check whether two filespecs with the same mode and size are identical */
5716static int diff_filespec_is_identical(struct diff_filespec *one,
5717 struct diff_filespec *two)
5718{
5719 if (S_ISGITLINK(one->mode))
5720 return 0;
5721 if (diff_populate_filespec(one, 0))
5722 return 0;
5723 if (diff_populate_filespec(two, 0))
5724 return 0;
5725 return !memcmp(one->data, two->data, one->size);
5726}
5727static int diff_filespec_check_stat_unmatch(struct diff_filepair *p)
5729{
5730 if (p->done_skip_stat_unmatch)
5731 return p->skip_stat_unmatch_result;
5732 p->done_skip_stat_unmatch = 1;
5734 p->skip_stat_unmatch_result = 0;
5735 /*
5736 * 1. Entries that come from stat info dirtiness
5737 * always have both sides (iow, not create/delete),
5738 * one side of the object name is unknown, with
5739 * the same mode and size. Keep the ones that
5740 * do not match these criteria. They have real
5741 * differences.
5742 *
5743 * 2. At this point, the file is known to be modified,
5744 * with the same mode and size, and the object
5745 * name of one side is unknown. Need to inspect
5746 * the identical contents.
5747 */
5748 if (!DIFF_FILE_VALID(p->one) || /* (1) */
5749 !DIFF_FILE_VALID(p->two) ||
5750 (p->one->oid_valid && p->two->oid_valid) ||
5751 (p->one->mode != p->two->mode) ||
5752 diff_populate_filespec(p->one, CHECK_SIZE_ONLY) ||
5753 diff_populate_filespec(p->two, CHECK_SIZE_ONLY) ||
5754 (p->one->size != p->two->size) ||
5755 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
5756 p->skip_stat_unmatch_result = 1;
5757 return p->skip_stat_unmatch_result;
5758}
5759static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
5761{
5762 int i;
5763 struct diff_queue_struct *q = &diff_queued_diff;
5764 struct diff_queue_struct outq;
5765 DIFF_QUEUE_CLEAR(&outq);
5766 for (i = 0; i < q->nr; i++) {
5768 struct diff_filepair *p = q->queue[i];
5769 if (diff_filespec_check_stat_unmatch(p))
5771 diff_q(&outq, p);
5772 else {
5773 /*
5774 * The caller can subtract 1 from skip_stat_unmatch
5775 * to determine how many paths were dirty only
5776 * due to stat info mismatch.
5777 */
5778 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
5779 diffopt->skip_stat_unmatch++;
5780 diff_free_filepair(p);
5781 }
5782 }
5783 free(q->queue);
5784 *q = outq;
5785}
5786static int diffnamecmp(const void *a_, const void *b_)
5788{
5789 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
5790 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
5791 const char *name_a, *name_b;
5792 name_a = a->one ? a->one->path : a->two->path;
5794 name_b = b->one ? b->one->path : b->two->path;
5795 return strcmp(name_a, name_b);
5796}
5797void diffcore_fix_diff_index(struct diff_options *options)
5799{
5800 struct diff_queue_struct *q = &diff_queued_diff;
5801 QSORT(q->queue, q->nr, diffnamecmp);
5802}
5803void diffcore_std(struct diff_options *options)
5805{
5806 /* NOTE please keep the following in sync with diff_tree_combined() */
5807 if (options->skip_stat_unmatch)
5808 diffcore_skip_stat_unmatch(options);
5809 if (!options->found_follow) {
5810 /* See try_to_follow_renames() in tree-diff.c */
5811 if (options->break_opt != -1)
5812 diffcore_break(options->break_opt);
5813 if (options->detect_rename)
5814 diffcore_rename(options);
5815 if (options->break_opt != -1)
5816 diffcore_merge_broken();
5817 }
5818 if (options->pickaxe)
5819 diffcore_pickaxe(options);
5820 if (options->orderfile)
5821 diffcore_order(options->orderfile);
5822 if (!options->found_follow)
5823 /* See try_to_follow_renames() in tree-diff.c */
5824 diff_resolve_rename_copy();
5825 diffcore_apply_filter(options);
5826 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5828 DIFF_OPT_SET(options, HAS_CHANGES);
5829 else
5830 DIFF_OPT_CLR(options, HAS_CHANGES);
5831 options->found_follow = 0;
5833}
5834int diff_result_code(struct diff_options *opt, int status)
5836{
5837 int result = 0;
5838 diff_warn_rename_limit("diff.renameLimit",
5840 opt->needed_rename_limit,
5841 opt->degraded_cc_to_c);
5842 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5843 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
5844 return status;
5845 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
5846 DIFF_OPT_TST(opt, HAS_CHANGES))
5847 result |= 01;
5848 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
5849 DIFF_OPT_TST(opt, CHECK_FAILED))
5850 result |= 02;
5851 return result;
5852}
5853int diff_can_quit_early(struct diff_options *opt)
5855{
5856 return (DIFF_OPT_TST(opt, QUICK) &&
5857 !opt->filter &&
5858 DIFF_OPT_TST(opt, HAS_CHANGES));
5859}
5860/*
5862 * Shall changes to this submodule be ignored?
5863 *
5864 * Submodule changes can be configured to be ignored separately for each path,
5865 * but that configuration can be overridden from the command line.
5866 */
5867static int is_submodule_ignored(const char *path, struct diff_options *options)
5868{
5869 int ignored = 0;
5870 unsigned orig_flags = options->flags;
5871 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
5872 set_diffopt_flags_from_submodule_config(options, path);
5873 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
5874 ignored = 1;
5875 options->flags = orig_flags;
5876 return ignored;
5877}
5878void diff_addremove(struct diff_options *options,
5880 int addremove, unsigned mode,
5881 const struct object_id *oid,
5882 int oid_valid,
5883 const char *concatpath, unsigned dirty_submodule)
5884{
5885 struct diff_filespec *one, *two;
5886 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
5888 return;
5889 /* This may look odd, but it is a preparation for
5891 * feeding "there are unchanged files which should
5892 * not produce diffs, but when you are doing copy
5893 * detection you would need them, so here they are"
5894 * entries to the diff-core. They will be prefixed
5895 * with something like '=' or '*' (I haven't decided
5896 * which but should not make any difference).
5897 * Feeding the same new and old to diff_change()
5898 * also has the same effect.
5899 * Before the final output happens, they are pruned after
5900 * merged into rename/copy pairs as appropriate.
5901 */
5902 if (DIFF_OPT_TST(options, REVERSE_DIFF))
5903 addremove = (addremove == '+' ? '-' :
5904 addremove == '-' ? '+' : addremove);
5905 if (options->prefix &&
5907 strncmp(concatpath, options->prefix, options->prefix_length))
5908 return;
5909 one = alloc_filespec(concatpath);
5911 two = alloc_filespec(concatpath);
5912 if (addremove != '+')
5914 fill_filespec(one, oid, oid_valid, mode);
5915 if (addremove != '-') {
5916 fill_filespec(two, oid, oid_valid, mode);
5917 two->dirty_submodule = dirty_submodule;
5918 }
5919 diff_queue(&diff_queued_diff, one, two);
5921 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5922 DIFF_OPT_SET(options, HAS_CHANGES);
5923}
5924void diff_change(struct diff_options *options,
5926 unsigned old_mode, unsigned new_mode,
5927 const struct object_id *old_oid,
5928 const struct object_id *new_oid,
5929 int old_oid_valid, int new_oid_valid,
5930 const char *concatpath,
5931 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
5932{
5933 struct diff_filespec *one, *two;
5934 struct diff_filepair *p;
5935 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
5937 is_submodule_ignored(concatpath, options))
5938 return;
5939 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
5941 SWAP(old_mode, new_mode);
5942 SWAP(old_oid, new_oid);
5943 SWAP(old_oid_valid, new_oid_valid);
5944 SWAP(old_dirty_submodule, new_dirty_submodule);
5945 }
5946 if (options->prefix &&
5948 strncmp(concatpath, options->prefix, options->prefix_length))
5949 return;
5950 one = alloc_filespec(concatpath);
5952 two = alloc_filespec(concatpath);
5953 fill_filespec(one, old_oid, old_oid_valid, old_mode);
5954 fill_filespec(two, new_oid, new_oid_valid, new_mode);
5955 one->dirty_submodule = old_dirty_submodule;
5956 two->dirty_submodule = new_dirty_submodule;
5957 p = diff_queue(&diff_queued_diff, one, two);
5958 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
5960 return;
5961 if (DIFF_OPT_TST(options, QUICK) && options->skip_stat_unmatch &&
5963 !diff_filespec_check_stat_unmatch(p))
5964 return;
5965 DIFF_OPT_SET(options, HAS_CHANGES);
5967}
5968struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
5970{
5971 struct diff_filepair *pair;
5972 struct diff_filespec *one, *two;
5973 if (options->prefix &&
5975 strncmp(path, options->prefix, options->prefix_length))
5976 return NULL;
5977 one = alloc_filespec(path);
5979 two = alloc_filespec(path);
5980 pair = diff_queue(&diff_queued_diff, one, two);
5981 pair->is_unmerged = 1;
5982 return pair;
5983}
5984static char *run_textconv(const char *pgm, struct diff_filespec *spec,
5986 size_t *outsize)
5987{
5988 struct diff_tempfile *temp;
5989 const char *argv[3];
5990 const char **arg = argv;
5991 struct child_process child = CHILD_PROCESS_INIT;
5992 struct strbuf buf = STRBUF_INIT;
5993 int err = 0;
5994 temp = prepare_temp_file(spec->path, spec);
5996 *arg++ = pgm;
5997 *arg++ = temp->name;
5998 *arg = NULL;
5999 child.use_shell = 1;
6001 child.argv = argv;
6002 child.out = -1;
6003 if (start_command(&child)) {
6004 remove_tempfile();
6005 return NULL;
6006 }
6007 if (strbuf_read(&buf, child.out, 0) < 0)
6009 err = error("error reading from textconv command '%s'", pgm);
6010 close(child.out);
6011 if (finish_command(&child) || err) {
6013 strbuf_release(&buf);
6014 remove_tempfile();
6015 return NULL;
6016 }
6017 remove_tempfile();
6018 return strbuf_detach(&buf, outsize);
6020}
6021size_t fill_textconv(struct userdiff_driver *driver,
6023 struct diff_filespec *df,
6024 char **outbuf)
6025{
6026 size_t size;
6027 if (!driver) {
6029 if (!DIFF_FILE_VALID(df)) {
6030 *outbuf = "";
6031 return 0;
6032 }
6033 if (diff_populate_filespec(df, 0))
6034 die("unable to read files to diff");
6035 *outbuf = df->data;
6036 return df->size;
6037 }
6038 if (!driver->textconv)
6040 die("BUG: fill_textconv called with non-textconv driver");
6041 if (driver->textconv_cache && df->oid_valid) {
6043 *outbuf = notes_cache_get(driver->textconv_cache,
6044 &df->oid,
6045 &size);
6046 if (*outbuf)
6047 return size;
6048 }
6049 *outbuf = run_textconv(driver->textconv, df, &size);
6051 if (!*outbuf)
6052 die("unable to read files to diff");
6053 if (driver->textconv_cache && df->oid_valid) {
6055 /* ignore errors, as we might be in a readonly repository */
6056 notes_cache_put(driver->textconv_cache, &df->oid, *outbuf,
6057 size);
6058 /*
6059 * we could save up changes and flush them all at the end,
6060 * but we would need an extra call after all diffing is done.
6061 * Since generating a cache entry is the slow path anyway,
6062 * this extra overhead probably isn't a big deal.
6063 */
6064 notes_cache_write(driver->textconv_cache);
6065 }
6066 return size;
6068}
6069int textconv_object(const char *path,
6071 unsigned mode,
6072 const struct object_id *oid,
6073 int oid_valid,
6074 char **buf,
6075 unsigned long *buf_size)
6076{
6077 struct diff_filespec *df;
6078 struct userdiff_driver *textconv;
6079 df = alloc_filespec(path);
6081 fill_filespec(df, oid, oid_valid, mode);
6082 textconv = get_textconv(df);
6083 if (!textconv) {
6084 free_filespec(df);
6085 return 0;
6086 }
6087 *buf_size = fill_textconv(textconv, df, buf);
6089 free_filespec(df);
6090 return 1;
6091}
6092void setup_diff_pager(struct diff_options *opt)
6094{
6095 /*
6096 * If the user asked for our exit code, then either they want --quiet
6097 * or --exit-code. We should definitely not bother with a pager in the
6098 * former case, as we will generate no output. Since we still properly
6099 * report our exit code even when a pager is run, we _could_ run a
6100 * pager with --exit-code. But since we have not done so historically,
6101 * and because it is easy to find people oneline advising "git diff
6102 * --exit-code" in hooks and other scripts, we do not do so.
6103 */
6104 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
6105 check_pager_config("diff") != 0)
6106 setup_pager();
6107}