?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff: don't use pathname-based diff drivers for symlinks
[git/dkf.git] / diff.c
blob95835d875ee392fdb5d0edbb52c37bbae06a8006
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 200;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34
35 static char diff_colors[][COLOR_MAXLEN] = {
36 GIT_COLOR_RESET,
37 GIT_COLOR_NORMAL, /* PLAIN */
38 GIT_COLOR_BOLD, /* METAINFO */
39 GIT_COLOR_CYAN, /* FRAGINFO */
40 GIT_COLOR_RED, /* OLD */
41 GIT_COLOR_GREEN, /* NEW */
42 GIT_COLOR_YELLOW, /* COMMIT */
43 GIT_COLOR_BG_RED, /* WHITESPACE */
44 GIT_COLOR_NORMAL, /* FUNCINFO */
45 };
46
47 static int parse_diff_color_slot(const char *var, int ofs)
48 {
49 if (!strcasecmp(var+ofs, "plain"))
50 return DIFF_PLAIN;
51 if (!strcasecmp(var+ofs, "meta"))
52 return DIFF_METAINFO;
53 if (!strcasecmp(var+ofs, "frag"))
54 return DIFF_FRAGINFO;
55 if (!strcasecmp(var+ofs, "old"))
56 return DIFF_FILE_OLD;
57 if (!strcasecmp(var+ofs, "new"))
58 return DIFF_FILE_NEW;
59 if (!strcasecmp(var+ofs, "commit"))
60 return DIFF_COMMIT;
61 if (!strcasecmp(var+ofs, "whitespace"))
62 return DIFF_WHITESPACE;
63 if (!strcasecmp(var+ofs, "func"))
64 return DIFF_FUNCINFO;
65 return -1;
66 }
67
68 static int git_config_rename(const char *var, const char *value)
69 {
70 if (!value)
71 return DIFF_DETECT_RENAME;
72 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
73 return DIFF_DETECT_COPY;
74 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
75 }
76
77 /*
78 * These are to give UI layer defaults.
79 * The core-level commands such as git-diff-files should
80 * never be affected by the setting of diff.renames
81 * the user happens to have in the configuration file.
82 */
83 int git_diff_ui_config(const char *var, const char *value, void *cb)
84 {
85 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
86 diff_use_color_default = git_config_colorbool(var, value, -1);
87 return 0;
88 }
89 if (!strcmp(var, "diff.renames")) {
90 diff_detect_rename_default = git_config_rename(var, value);
91 return 0;
92 }
93 if (!strcmp(var, "diff.autorefreshindex")) {
94 diff_auto_refresh_index = git_config_bool(var, value);
95 return 0;
96 }
97 if (!strcmp(var, "diff.mnemonicprefix")) {
98 diff_mnemonic_prefix = git_config_bool(var, value);
99 return 0;
100 }
101 if (!strcmp(var, "diff.noprefix")) {
102 diff_no_prefix = git_config_bool(var, value);
103 return 0;
104 }
105 if (!strcmp(var, "diff.external"))
106 return git_config_string(&external_diff_cmd_cfg, var, value);
107 if (!strcmp(var, "diff.wordregex"))
108 return git_config_string(&diff_word_regex_cfg, var, value);
109
110 return git_diff_basic_config(var, value, cb);
111 }
112
113 int git_diff_basic_config(const char *var, const char *value, void *cb)
114 {
115 if (!strcmp(var, "diff.renamelimit")) {
116 diff_rename_limit_default = git_config_int(var, value);
117 return 0;
118 }
119
120 switch (userdiff_config(var, value)) {
121 case 0: break;
122 case -1: return -1;
123 default: return 0;
124 }
125
126 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
127 int slot = parse_diff_color_slot(var, 11);
128 if (slot < 0)
129 return 0;
130 if (!value)
131 return config_error_nonbool(var);
132 color_parse(value, var, diff_colors[slot]);
133 return 0;
134 }
135
136 /* like GNU diff's --suppress-blank-empty option */
137 if (!strcmp(var, "diff.suppressblankempty") ||
138 /* for backwards compatibility */
139 !strcmp(var, "diff.suppress-blank-empty")) {
140 diff_suppress_blank_empty = git_config_bool(var, value);
141 return 0;
142 }
143
144 return git_color_default_config(var, value, cb);
145 }
146
147 static char *quote_two(const char *one, const char *two)
148 {
149 int need_one = quote_c_style(one, NULL, NULL, 1);
150 int need_two = quote_c_style(two, NULL, NULL, 1);
151 struct strbuf res = STRBUF_INIT;
152
153 if (need_one + need_two) {
154 strbuf_addch(&res, '"');
155 quote_c_style(one, &res, NULL, 1);
156 quote_c_style(two, &res, NULL, 1);
157 strbuf_addch(&res, '"');
158 } else {
159 strbuf_addstr(&res, one);
160 strbuf_addstr(&res, two);
161 }
162 return strbuf_detach(&res, NULL);
163 }
164
165 static const char *external_diff(void)
166 {
167 static const char *external_diff_cmd = NULL;
168 static int done_preparing = 0;
169
170 if (done_preparing)
171 return external_diff_cmd;
172 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
173 if (!external_diff_cmd)
174 external_diff_cmd = external_diff_cmd_cfg;
175 done_preparing = 1;
176 return external_diff_cmd;
177 }
178
179 static struct diff_tempfile {
180 const char *name; /* filename external diff should read from */
181 char hex[41];
182 char mode[10];
183 char tmp_path[PATH_MAX];
184 } diff_temp[2];
185
186 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
187
188 struct emit_callback {
189 int color_diff;
190 unsigned ws_rule;
191 int blank_at_eof_in_preimage;
192 int blank_at_eof_in_postimage;
193 int lno_in_preimage;
194 int lno_in_postimage;
195 sane_truncate_fn truncate;
196 const char **label_path;
197 struct diff_words_data *diff_words;
198 struct diff_options *opt;
199 int *found_changesp;
200 struct strbuf *header;
201 };
202
203 static int count_lines(const char *data, int size)
204 {
205 int count, ch, completely_empty = 1, nl_just_seen = 0;
206 count = 0;
207 while (0 < size--) {
208 ch = *data++;
209 if (ch == '\n') {
210 count++;
211 nl_just_seen = 1;
212 completely_empty = 0;
213 }
214 else {
215 nl_just_seen = 0;
216 completely_empty = 0;
217 }
218 }
219 if (completely_empty)
220 return 0;
221 if (!nl_just_seen)
222 count++; /* no trailing newline */
223 return count;
224 }
225
226 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
227 {
228 if (!DIFF_FILE_VALID(one)) {
229 mf->ptr = (char *)""; /* does not matter */
230 mf->size = 0;
231 return 0;
232 }
233 else if (diff_populate_filespec(one, 0))
234 return -1;
235
236 mf->ptr = one->data;
237 mf->size = one->size;
238 return 0;
239 }
240
241 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
242 {
243 char *ptr = mf->ptr;
244 long size = mf->size;
245 int cnt = 0;
246
247 if (!size)
248 return cnt;
249 ptr += size - 1; /* pointing at the very end */
250 if (*ptr != '\n')
251 ; /* incomplete line */
252 else
253 ptr--; /* skip the last LF */
254 while (mf->ptr < ptr) {
255 char *prev_eol;
256 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
257 if (*prev_eol == '\n')
258 break;
259 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
260 break;
261 cnt++;
262 ptr = prev_eol - 1;
263 }
264 return cnt;
265 }
266
267 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
268 struct emit_callback *ecbdata)
269 {
270 int l1, l2, at;
271 unsigned ws_rule = ecbdata->ws_rule;
272 l1 = count_trailing_blank(mf1, ws_rule);
273 l2 = count_trailing_blank(mf2, ws_rule);
274 if (l2 <= l1) {
275 ecbdata->blank_at_eof_in_preimage = 0;
276 ecbdata->blank_at_eof_in_postimage = 0;
277 return;
278 }
279 at = count_lines(mf1->ptr, mf1->size);
280 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
281
282 at = count_lines(mf2->ptr, mf2->size);
283 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
284 }
285
286 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
287 int first, const char *line, int len)
288 {
289 int has_trailing_newline, has_trailing_carriage_return;
290 int nofirst;
291 FILE *file = o->file;
292
293 if (o->output_prefix) {
294 struct strbuf *msg = NULL;
295 msg = o->output_prefix(o, o->output_prefix_data);
296 assert(msg);
297 fwrite(msg->buf, msg->len, 1, file);
298 }
299
300 if (len == 0) {
301 has_trailing_newline = (first == '\n');
302 has_trailing_carriage_return = (!has_trailing_newline &&
303 (first == '\r'));
304 nofirst = has_trailing_newline || has_trailing_carriage_return;
305 } else {
306 has_trailing_newline = (len > 0 && line[len-1] == '\n');
307 if (has_trailing_newline)
308 len--;
309 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
310 if (has_trailing_carriage_return)
311 len--;
312 nofirst = 0;
313 }
314
315 if (len || !nofirst) {
316 fputs(set, file);
317 if (!nofirst)
318 fputc(first, file);
319 fwrite(line, len, 1, file);
320 fputs(reset, file);
321 }
322 if (has_trailing_carriage_return)
323 fputc('\r', file);
324 if (has_trailing_newline)
325 fputc('\n', file);
326 }
327
328 static void emit_line(struct diff_options *o, const char *set, const char *reset,
329 const char *line, int len)
330 {
331 emit_line_0(o, set, reset, line[0], line+1, len-1);
332 }
333
334 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
335 {
336 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
337 ecbdata->blank_at_eof_in_preimage &&
338 ecbdata->blank_at_eof_in_postimage &&
339 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
340 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
341 return 0;
342 return ws_blank_line(line, len, ecbdata->ws_rule);
343 }
344
345 static void emit_add_line(const char *reset,
346 struct emit_callback *ecbdata,
347 const char *line, int len)
348 {
349 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
350 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
351
352 if (!*ws)
353 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
354 else if (new_blank_line_at_eof(ecbdata, line, len))
355 /* Blank line at EOF - paint '+' as well */
356 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
357 else {
358 /* Emit just the prefix, then the rest. */
359 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
360 ws_check_emit(line, len, ecbdata->ws_rule,
361 ecbdata->opt->file, set, reset, ws);
362 }
363 }
364
365 static void emit_hunk_header(struct emit_callback *ecbdata,
366 const char *line, int len)
367 {
368 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
369 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
370 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
371 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
372 static const char atat[2] = { '@', '@' };
373 const char *cp, *ep;
374 struct strbuf msgbuf = STRBUF_INIT;
375 int org_len = len;
376 int i = 1;
377
378 /*
379 * As a hunk header must begin with "@@ -<old>, +<new> @@",
380 * it always is at least 10 bytes long.
381 */
382 if (len < 10 ||
383 memcmp(line, atat, 2) ||
384 !(ep = memmem(line + 2, len - 2, atat, 2))) {
385 emit_line(ecbdata->opt, plain, reset, line, len);
386 return;
387 }
388 ep += 2; /* skip over @@ */
389
390 /* The hunk header in fraginfo color */
391 strbuf_add(&msgbuf, frag, strlen(frag));
392 strbuf_add(&msgbuf, line, ep - line);
393 strbuf_add(&msgbuf, reset, strlen(reset));
394
395 /*
396 * trailing "\r\n"
397 */
398 for ( ; i < 3; i++)
399 if (line[len - i] == '\r' || line[len - i] == '\n')
400 len--;
401
402 /* blank before the func header */
403 for (cp = ep; ep - line < len; ep++)
404 if (*ep != ' ' && *ep != '\t')
405 break;
406 if (ep != cp) {
407 strbuf_add(&msgbuf, plain, strlen(plain));
408 strbuf_add(&msgbuf, cp, ep - cp);
409 strbuf_add(&msgbuf, reset, strlen(reset));
410 }
411
412 if (ep < line + len) {
413 strbuf_add(&msgbuf, func, strlen(func));
414 strbuf_add(&msgbuf, ep, line + len - ep);
415 strbuf_add(&msgbuf, reset, strlen(reset));
416 }
417
418 strbuf_add(&msgbuf, line + len, org_len - len);
419 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
420 strbuf_release(&msgbuf);
421 }
422
423 static struct diff_tempfile *claim_diff_tempfile(void) {
424 int i;
425 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
426 if (!diff_temp[i].name)
427 return diff_temp + i;
428 die("BUG: diff is failing to clean up its tempfiles");
429 }
430
431 static int remove_tempfile_installed;
432
433 static void remove_tempfile(void)
434 {
435 int i;
436 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
437 if (diff_temp[i].name == diff_temp[i].tmp_path)
438 unlink_or_warn(diff_temp[i].name);
439 diff_temp[i].name = NULL;
440 }
441 }
442
443 static void remove_tempfile_on_signal(int signo)
444 {
445 remove_tempfile();
446 sigchain_pop(signo);
447 raise(signo);
448 }
449
450 static void print_line_count(FILE *file, int count)
451 {
452 switch (count) {
453 case 0:
454 fprintf(file, "0,0");
455 break;
456 case 1:
457 fprintf(file, "1");
458 break;
459 default:
460 fprintf(file, "1,%d", count);
461 break;
462 }
463 }
464
465 static void emit_rewrite_lines(struct emit_callback *ecb,
466 int prefix, const char *data, int size)
467 {
468 const char *endp = NULL;
469 static const char *nneof = " No newline at end of file\n";
470 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
471 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
472
473 while (0 < size) {
474 int len;
475
476 endp = memchr(data, '\n', size);
477 len = endp ? (endp - data + 1) : size;
478 if (prefix != '+') {
479 ecb->lno_in_preimage++;
480 emit_line_0(ecb->opt, old, reset, '-',
481 data, len);
482 } else {
483 ecb->lno_in_postimage++;
484 emit_add_line(reset, ecb, data, len);
485 }
486 size -= len;
487 data += len;
488 }
489 if (!endp) {
490 const char *plain = diff_get_color(ecb->color_diff,
491 DIFF_PLAIN);
492 emit_line_0(ecb->opt, plain, reset, '\\',
493 nneof, strlen(nneof));
494 }
495 }
496
497 static void emit_rewrite_diff(const char *name_a,
498 const char *name_b,
499 struct diff_filespec *one,
500 struct diff_filespec *two,
501 struct userdiff_driver *textconv_one,
502 struct userdiff_driver *textconv_two,
503 struct diff_options *o)
504 {
505 int lc_a, lc_b;
506 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
507 const char *name_a_tab, *name_b_tab;
508 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
509 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
510 const char *reset = diff_get_color(color_diff, DIFF_RESET);
511 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
512 const char *a_prefix, *b_prefix;
513 char *data_one, *data_two;
514 size_t size_one, size_two;
515 struct emit_callback ecbdata;
516 char *line_prefix = "";
517 struct strbuf *msgbuf;
518
519 if (o && o->output_prefix) {
520 msgbuf = o->output_prefix(o, o->output_prefix_data);
521 line_prefix = msgbuf->buf;
522 }
523
524 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
525 a_prefix = o->b_prefix;
526 b_prefix = o->a_prefix;
527 } else {
528 a_prefix = o->a_prefix;
529 b_prefix = o->b_prefix;
530 }
531
532 name_a += (*name_a == '/');
533 name_b += (*name_b == '/');
534 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
535 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
536
537 strbuf_reset(&a_name);
538 strbuf_reset(&b_name);
539 quote_two_c_style(&a_name, a_prefix, name_a, 0);
540 quote_two_c_style(&b_name, b_prefix, name_b, 0);
541
542 size_one = fill_textconv(textconv_one, one, &data_one);
543 size_two = fill_textconv(textconv_two, two, &data_two);
544
545 memset(&ecbdata, 0, sizeof(ecbdata));
546 ecbdata.color_diff = color_diff;
547 ecbdata.found_changesp = &o->found_changes;
548 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
549 ecbdata.opt = o;
550 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
551 mmfile_t mf1, mf2;
552 mf1.ptr = (char *)data_one;
553 mf2.ptr = (char *)data_two;
554 mf1.size = size_one;
555 mf2.size = size_two;
556 check_blank_at_eof(&mf1, &mf2, &ecbdata);
557 }
558 ecbdata.lno_in_preimage = 1;
559 ecbdata.lno_in_postimage = 1;
560
561 lc_a = count_lines(data_one, size_one);
562 lc_b = count_lines(data_two, size_two);
563 fprintf(o->file,
564 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
565 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
566 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
567 line_prefix, fraginfo);
568 print_line_count(o->file, lc_a);
569 fprintf(o->file, " +");
570 print_line_count(o->file, lc_b);
571 fprintf(o->file, " @@%s\n", reset);
572 if (lc_a)
573 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
574 if (lc_b)
575 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
576 if (textconv_one)
577 free((char *)data_one);
578 if (textconv_two)
579 free((char *)data_two);
580 }
581
582 struct diff_words_buffer {
583 mmfile_t text;
584 long alloc;
585 struct diff_words_orig {
586 const char *begin, *end;
587 } *orig;
588 int orig_nr, orig_alloc;
589 };
590
591 static void diff_words_append(char *line, unsigned long len,
592 struct diff_words_buffer *buffer)
593 {
594 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
595 line++;
596 len--;
597 memcpy(buffer->text.ptr + buffer->text.size, line, len);
598 buffer->text.size += len;
599 buffer->text.ptr[buffer->text.size] = '\0';
600 }
601
602 struct diff_words_style_elem
603 {
604 const char *prefix;
605 const char *suffix;
606 const char *color; /* NULL; filled in by the setup code if
607 * color is enabled */
608 };
609
610 struct diff_words_style
611 {
612 enum diff_words_type type;
613 struct diff_words_style_elem new, old, ctx;
614 const char *newline;
615 };
616
617 struct diff_words_style diff_words_styles[] = {
618 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
619 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
620 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
621 };
622
623 struct diff_words_data {
624 struct diff_words_buffer minus, plus;
625 const char *current_plus;
626 int last_minus;
627 struct diff_options *opt;
628 regex_t *word_regex;
629 enum diff_words_type type;
630 struct diff_words_style *style;
631 };
632
633 static int fn_out_diff_words_write_helper(FILE *fp,
634 struct diff_words_style_elem *st_el,
635 const char *newline,
636 size_t count, const char *buf,
637 const char *line_prefix)
638 {
639 int print = 0;
640
641 while (count) {
642 char *p = memchr(buf, '\n', count);
643 if (print)
644 fputs(line_prefix, fp);
645 if (p != buf) {
646 if (st_el->color && fputs(st_el->color, fp) < 0)
647 return -1;
648 if (fputs(st_el->prefix, fp) < 0 ||
649 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
650 fputs(st_el->suffix, fp) < 0)
651 return -1;
652 if (st_el->color && *st_el->color
653 && fputs(GIT_COLOR_RESET, fp) < 0)
654 return -1;
655 }
656 if (!p)
657 return 0;
658 if (fputs(newline, fp) < 0)
659 return -1;
660 count -= p + 1 - buf;
661 buf = p + 1;
662 print = 1;
663 }
664 return 0;
665 }
666
667 /*
668 * '--color-words' algorithm can be described as:
669 *
670 * 1. collect a the minus/plus lines of a diff hunk, divided into
671 * minus-lines and plus-lines;
672 *
673 * 2. break both minus-lines and plus-lines into words and
674 * place them into two mmfile_t with one word for each line;
675 *
676 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
677 *
678 * And for the common parts of the both file, we output the plus side text.
679 * diff_words->current_plus is used to trace the current position of the plus file
680 * which printed. diff_words->last_minus is used to trace the last minus word
681 * printed.
682 *
683 * For '--graph' to work with '--color-words', we need to output the graph prefix
684 * on each line of color words output. Generally, there are two conditions on
685 * which we should output the prefix.
686 *
687 * 1. diff_words->last_minus == 0 &&
688 * diff_words->current_plus == diff_words->plus.text.ptr
689 *
690 * that is: the plus text must start as a new line, and if there is no minus
691 * word printed, a graph prefix must be printed.
692 *
693 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
694 * *(diff_words->current_plus - 1) == '\n'
695 *
696 * that is: a graph prefix must be printed following a '\n'
697 */
698 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
699 {
700 if ((diff_words->last_minus == 0 &&
701 diff_words->current_plus == diff_words->plus.text.ptr) ||
702 (diff_words->current_plus > diff_words->plus.text.ptr &&
703 *(diff_words->current_plus - 1) == '\n')) {
704 return 1;
705 } else {
706 return 0;
707 }
708 }
709
710 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
711 {
712 struct diff_words_data *diff_words = priv;
713 struct diff_words_style *style = diff_words->style;
714 int minus_first, minus_len, plus_first, plus_len;
715 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
716 struct diff_options *opt = diff_words->opt;
717 struct strbuf *msgbuf;
718 char *line_prefix = "";
719
720 if (line[0] != '@' || parse_hunk_header(line, len,
721 &minus_first, &minus_len, &plus_first, &plus_len))
722 return;
723
724 assert(opt);
725 if (opt->output_prefix) {
726 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
727 line_prefix = msgbuf->buf;
728 }
729
730 /* POSIX requires that first be decremented by one if len == 0... */
731 if (minus_len) {
732 minus_begin = diff_words->minus.orig[minus_first].begin;
733 minus_end =
734 diff_words->minus.orig[minus_first + minus_len - 1].end;
735 } else
736 minus_begin = minus_end =
737 diff_words->minus.orig[minus_first].end;
738
739 if (plus_len) {
740 plus_begin = diff_words->plus.orig[plus_first].begin;
741 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
742 } else
743 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
744
745 if (color_words_output_graph_prefix(diff_words)) {
746 fputs(line_prefix, diff_words->opt->file);
747 }
748 if (diff_words->current_plus != plus_begin) {
749 fn_out_diff_words_write_helper(diff_words->opt->file,
750 &style->ctx, style->newline,
751 plus_begin - diff_words->current_plus,
752 diff_words->current_plus, line_prefix);
753 if (*(plus_begin - 1) == '\n')
754 fputs(line_prefix, diff_words->opt->file);
755 }
756 if (minus_begin != minus_end) {
757 fn_out_diff_words_write_helper(diff_words->opt->file,
758 &style->old, style->newline,
759 minus_end - minus_begin, minus_begin,
760 line_prefix);
761 }
762 if (plus_begin != plus_end) {
763 fn_out_diff_words_write_helper(diff_words->opt->file,
764 &style->new, style->newline,
765 plus_end - plus_begin, plus_begin,
766 line_prefix);
767 }
768
769 diff_words->current_plus = plus_end;
770 diff_words->last_minus = minus_first;
771 }
772
773 /* This function starts looking at *begin, and returns 0 iff a word was found. */
774 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
775 int *begin, int *end)
776 {
777 if (word_regex && *begin < buffer->size) {
778 regmatch_t match[1];
779 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
780 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
781 '\n', match[0].rm_eo - match[0].rm_so);
782 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
783 *begin += match[0].rm_so;
784 return *begin >= *end;
785 }
786 return -1;
787 }
788
789 /* find the next word */
790 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
791 (*begin)++;
792 if (*begin >= buffer->size)
793 return -1;
794
795 /* find the end of the word */
796 *end = *begin + 1;
797 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
798 (*end)++;
799
800 return 0;
801 }
802
803 /*
804 * This function splits the words in buffer->text, stores the list with
805 * newline separator into out, and saves the offsets of the original words
806 * in buffer->orig.
807 */
808 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
809 regex_t *word_regex)
810 {
811 int i, j;
812 long alloc = 0;
813
814 out->size = 0;
815 out->ptr = NULL;
816
817 /* fake an empty "0th" word */
818 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
819 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
820 buffer->orig_nr = 1;
821
822 for (i = 0; i < buffer->text.size; i++) {
823 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
824 return;
825
826 /* store original boundaries */
827 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
828 buffer->orig_alloc);
829 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
830 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
831 buffer->orig_nr++;
832
833 /* store one word */
834 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
835 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
836 out->ptr[out->size + j - i] = '\n';
837 out->size += j - i + 1;
838
839 i = j - 1;
840 }
841 }
842
843 /* this executes the word diff on the accumulated buffers */
844 static void diff_words_show(struct diff_words_data *diff_words)
845 {
846 xpparam_t xpp;
847 xdemitconf_t xecfg;
848 mmfile_t minus, plus;
849 struct diff_words_style *style = diff_words->style;
850
851 struct diff_options *opt = diff_words->opt;
852 struct strbuf *msgbuf;
853 char *line_prefix = "";
854
855 assert(opt);
856 if (opt->output_prefix) {
857 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
858 line_prefix = msgbuf->buf;
859 }
860
861 /* special case: only removal */
862 if (!diff_words->plus.text.size) {
863 fputs(line_prefix, diff_words->opt->file);
864 fn_out_diff_words_write_helper(diff_words->opt->file,
865 &style->old, style->newline,
866 diff_words->minus.text.size,
867 diff_words->minus.text.ptr, line_prefix);
868 diff_words->minus.text.size = 0;
869 return;
870 }
871
872 diff_words->current_plus = diff_words->plus.text.ptr;
873 diff_words->last_minus = 0;
874
875 memset(&xpp, 0, sizeof(xpp));
876 memset(&xecfg, 0, sizeof(xecfg));
877 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
878 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
879 xpp.flags = 0;
880 /* as only the hunk header will be parsed, we need a 0-context */
881 xecfg.ctxlen = 0;
882 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
883 &xpp, &xecfg);
884 free(minus.ptr);
885 free(plus.ptr);
886 if (diff_words->current_plus != diff_words->plus.text.ptr +
887 diff_words->plus.text.size) {
888 if (color_words_output_graph_prefix(diff_words))
889 fputs(line_prefix, diff_words->opt->file);
890 fn_out_diff_words_write_helper(diff_words->opt->file,
891 &style->ctx, style->newline,
892 diff_words->plus.text.ptr + diff_words->plus.text.size
893 - diff_words->current_plus, diff_words->current_plus,
894 line_prefix);
895 }
896 diff_words->minus.text.size = diff_words->plus.text.size = 0;
897 }
898
899 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
900 static void diff_words_flush(struct emit_callback *ecbdata)
901 {
902 if (ecbdata->diff_words->minus.text.size ||
903 ecbdata->diff_words->plus.text.size)
904 diff_words_show(ecbdata->diff_words);
905 }
906
907 static void free_diff_words_data(struct emit_callback *ecbdata)
908 {
909 if (ecbdata->diff_words) {
910 diff_words_flush(ecbdata);
911 free (ecbdata->diff_words->minus.text.ptr);
912 free (ecbdata->diff_words->minus.orig);
913 free (ecbdata->diff_words->plus.text.ptr);
914 free (ecbdata->diff_words->plus.orig);
915 if (ecbdata->diff_words->word_regex) {
916 regfree(ecbdata->diff_words->word_regex);
917 free(ecbdata->diff_words->word_regex);
918 }
919 free(ecbdata->diff_words);
920 ecbdata->diff_words = NULL;
921 }
922 }
923
924 const char *diff_get_color(int diff_use_color, enum color_diff ix)
925 {
926 if (diff_use_color)
927 return diff_colors[ix];
928 return "";
929 }
930
931 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
932 {
933 const char *cp;
934 unsigned long allot;
935 size_t l = len;
936
937 if (ecb->truncate)
938 return ecb->truncate(line, len);
939 cp = line;
940 allot = l;
941 while (0 < l) {
942 (void) utf8_width(&cp, &l);
943 if (!cp)
944 break; /* truncated in the middle? */
945 }
946 return allot - l;
947 }
948
949 static void find_lno(const char *line, struct emit_callback *ecbdata)
950 {
951 const char *p;
952 ecbdata->lno_in_preimage = 0;
953 ecbdata->lno_in_postimage = 0;
954 p = strchr(line, '-');
955 if (!p)
956 return; /* cannot happen */
957 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
958 p = strchr(p, '+');
959 if (!p)
960 return; /* cannot happen */
961 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
962 }
963
964 static void fn_out_consume(void *priv, char *line, unsigned long len)
965 {
966 struct emit_callback *ecbdata = priv;
967 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
968 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
969 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
970 struct diff_options *o = ecbdata->opt;
971 char *line_prefix = "";
972 struct strbuf *msgbuf;
973
974 if (o && o->output_prefix) {
975 msgbuf = o->output_prefix(o, o->output_prefix_data);
976 line_prefix = msgbuf->buf;
977 }
978
979 if (ecbdata->header) {
980 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
981 strbuf_reset(ecbdata->header);
982 ecbdata->header = NULL;
983 }
984 *(ecbdata->found_changesp) = 1;
985
986 if (ecbdata->label_path[0]) {
987 const char *name_a_tab, *name_b_tab;
988
989 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
990 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
991
992 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
993 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
994 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
995 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
996 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
997 }
998
999 if (diff_suppress_blank_empty
1000 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1001 line[0] = '\n';
1002 len = 1;
1003 }
1004
1005 if (line[0] == '@') {
1006 if (ecbdata->diff_words)
1007 diff_words_flush(ecbdata);
1008 len = sane_truncate_line(ecbdata, line, len);
1009 find_lno(line, ecbdata);
1010 emit_hunk_header(ecbdata, line, len);
1011 if (line[len-1] != '\n')
1012 putc('\n', ecbdata->opt->file);
1013 return;
1014 }
1015
1016 if (len < 1) {
1017 emit_line(ecbdata->opt, reset, reset, line, len);
1018 if (ecbdata->diff_words
1019 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1020 fputs("~\n", ecbdata->opt->file);
1021 return;
1022 }
1023
1024 if (ecbdata->diff_words) {
1025 if (line[0] == '-') {
1026 diff_words_append(line, len,
1027 &ecbdata->diff_words->minus);
1028 return;
1029 } else if (line[0] == '+') {
1030 diff_words_append(line, len,
1031 &ecbdata->diff_words->plus);
1032 return;
1033 }
1034 diff_words_flush(ecbdata);
1035 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1036 emit_line(ecbdata->opt, plain, reset, line, len);
1037 fputs("~\n", ecbdata->opt->file);
1038 } else {
1039 /* don't print the prefix character */
1040 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1041 }
1042 return;
1043 }
1044
1045 if (line[0] != '+') {
1046 const char *color =
1047 diff_get_color(ecbdata->color_diff,
1048 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1049 ecbdata->lno_in_preimage++;
1050 if (line[0] == ' ')
1051 ecbdata->lno_in_postimage++;
1052 emit_line(ecbdata->opt, color, reset, line, len);
1053 } else {
1054 ecbdata->lno_in_postimage++;
1055 emit_add_line(reset, ecbdata, line + 1, len - 1);
1056 }
1057 }
1058
1059 static char *pprint_rename(const char *a, const char *b)
1060 {
1061 const char *old = a;
1062 const char *new = b;
1063 struct strbuf name = STRBUF_INIT;
1064 int pfx_length, sfx_length;
1065 int len_a = strlen(a);
1066 int len_b = strlen(b);
1067 int a_midlen, b_midlen;
1068 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1069 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1070
1071 if (qlen_a || qlen_b) {
1072 quote_c_style(a, &name, NULL, 0);
1073 strbuf_addstr(&name, " => ");
1074 quote_c_style(b, &name, NULL, 0);
1075 return strbuf_detach(&name, NULL);
1076 }
1077
1078 /* Find common prefix */
1079 pfx_length = 0;
1080 while (*old && *new && *old == *new) {
1081 if (*old == '/')
1082 pfx_length = old - a + 1;
1083 old++;
1084 new++;
1085 }
1086
1087 /* Find common suffix */
1088 old = a + len_a;
1089 new = b + len_b;
1090 sfx_length = 0;
1091 while (a <= old && b <= new && *old == *new) {
1092 if (*old == '/')
1093 sfx_length = len_a - (old - a);
1094 old--;
1095 new--;
1096 }
1097
1098 /*
1099 * pfx{mid-a => mid-b}sfx
1100 * {pfx-a => pfx-b}sfx
1101 * pfx{sfx-a => sfx-b}
1102 * name-a => name-b
1103 */
1104 a_midlen = len_a - pfx_length - sfx_length;
1105 b_midlen = len_b - pfx_length - sfx_length;
1106 if (a_midlen < 0)
1107 a_midlen = 0;
1108 if (b_midlen < 0)
1109 b_midlen = 0;
1110
1111 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1112 if (pfx_length + sfx_length) {
1113 strbuf_add(&name, a, pfx_length);
1114 strbuf_addch(&name, '{');
1115 }
1116 strbuf_add(&name, a + pfx_length, a_midlen);
1117 strbuf_addstr(&name, " => ");
1118 strbuf_add(&name, b + pfx_length, b_midlen);
1119 if (pfx_length + sfx_length) {
1120 strbuf_addch(&name, '}');
1121 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1122 }
1123 return strbuf_detach(&name, NULL);
1124 }
1125
1126 struct diffstat_t {
1127 int nr;
1128 int alloc;
1129 struct diffstat_file {
1130 char *from_name;
1131 char *name;
1132 char *print_name;
1133 unsigned is_unmerged:1;
1134 unsigned is_binary:1;
1135 unsigned is_renamed:1;
1136 uintmax_t added, deleted;
1137 } **files;
1138 };
1139
1140 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1141 const char *name_a,
1142 const char *name_b)
1143 {
1144 struct diffstat_file *x;
1145 x = xcalloc(sizeof (*x), 1);
1146 if (diffstat->nr == diffstat->alloc) {
1147 diffstat->alloc = alloc_nr(diffstat->alloc);
1148 diffstat->files = xrealloc(diffstat->files,
1149 diffstat->alloc * sizeof(x));
1150 }
1151 diffstat->files[diffstat->nr++] = x;
1152 if (name_b) {
1153 x->from_name = xstrdup(name_a);
1154 x->name = xstrdup(name_b);
1155 x->is_renamed = 1;
1156 }
1157 else {
1158 x->from_name = NULL;
1159 x->name = xstrdup(name_a);
1160 }
1161 return x;
1162 }
1163
1164 static void diffstat_consume(void *priv, char *line, unsigned long len)
1165 {
1166 struct diffstat_t *diffstat = priv;
1167 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1168
1169 if (line[0] == '+')
1170 x->added++;
1171 else if (line[0] == '-')
1172 x->deleted++;
1173 }
1174
1175 const char mime_boundary_leader[] = "------------";
1176
1177 static int scale_linear(int it, int width, int max_change)
1178 {
1179 /*
1180 * make sure that at least one '-' is printed if there were deletions,
1181 * and likewise for '+'.
1182 */
1183 if (max_change < 2)
1184 return it;
1185 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1186 }
1187
1188 static void show_name(FILE *file,
1189 const char *prefix, const char *name, int len)
1190 {
1191 fprintf(file, " %s%-*s |", prefix, len, name);
1192 }
1193
1194 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1195 {
1196 if (cnt <= 0)
1197 return;
1198 fprintf(file, "%s", set);
1199 while (cnt--)
1200 putc(ch, file);
1201 fprintf(file, "%s", reset);
1202 }
1203
1204 static void fill_print_name(struct diffstat_file *file)
1205 {
1206 char *pname;
1207
1208 if (file->print_name)
1209 return;
1210
1211 if (!file->is_renamed) {
1212 struct strbuf buf = STRBUF_INIT;
1213 if (quote_c_style(file->name, &buf, NULL, 0)) {
1214 pname = strbuf_detach(&buf, NULL);
1215 } else {
1216 pname = file->name;
1217 strbuf_release(&buf);
1218 }
1219 } else {
1220 pname = pprint_rename(file->from_name, file->name);
1221 }
1222 file->print_name = pname;
1223 }
1224
1225 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1226 {
1227 int i, len, add, del, adds = 0, dels = 0;
1228 uintmax_t max_change = 0, max_len = 0;
1229 int total_files = data->nr;
1230 int width, name_width;
1231 const char *reset, *set, *add_c, *del_c;
1232 const char *line_prefix = "";
1233 struct strbuf *msg = NULL;
1234
1235 if (data->nr == 0)
1236 return;
1237
1238 if (options->output_prefix) {
1239 msg = options->output_prefix(options, options->output_prefix_data);
1240 line_prefix = msg->buf;
1241 }
1242
1243 width = options->stat_width ? options->stat_width : 80;
1244 name_width = options->stat_name_width ? options->stat_name_width : 50;
1245
1246 /* Sanity: give at least 5 columns to the graph,
1247 * but leave at least 10 columns for the name.
1248 */
1249 if (width < 25)
1250 width = 25;
1251 if (name_width < 10)
1252 name_width = 10;
1253 else if (width < name_width + 15)
1254 name_width = width - 15;
1255
1256 /* Find the longest filename and max number of changes */
1257 reset = diff_get_color_opt(options, DIFF_RESET);
1258 set = diff_get_color_opt(options, DIFF_PLAIN);
1259 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1260 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1261
1262 for (i = 0; i < data->nr; i++) {
1263 struct diffstat_file *file = data->files[i];
1264 uintmax_t change = file->added + file->deleted;
1265 fill_print_name(file);
1266 len = strlen(file->print_name);
1267 if (max_len < len)
1268 max_len = len;
1269
1270 if (file->is_binary || file->is_unmerged)
1271 continue;
1272 if (max_change < change)
1273 max_change = change;
1274 }
1275
1276 /* Compute the width of the graph part;
1277 * 10 is for one blank at the beginning of the line plus
1278 * " | count " between the name and the graph.
1279 *
1280 * From here on, name_width is the width of the name area,
1281 * and width is the width of the graph area.
1282 */
1283 name_width = (name_width < max_len) ? name_width : max_len;
1284 if (width < (name_width + 10) + max_change)
1285 width = width - (name_width + 10);
1286 else
1287 width = max_change;
1288
1289 for (i = 0; i < data->nr; i++) {
1290 const char *prefix = "";
1291 char *name = data->files[i]->print_name;
1292 uintmax_t added = data->files[i]->added;
1293 uintmax_t deleted = data->files[i]->deleted;
1294 int name_len;
1295
1296 /*
1297 * "scale" the filename
1298 */
1299 len = name_width;
1300 name_len = strlen(name);
1301 if (name_width < name_len) {
1302 char *slash;
1303 prefix = "...";
1304 len -= 3;
1305 name += name_len - len;
1306 slash = strchr(name, '/');
1307 if (slash)
1308 name = slash;
1309 }
1310
1311 if (data->files[i]->is_binary) {
1312 fprintf(options->file, "%s", line_prefix);
1313 show_name(options->file, prefix, name, len);
1314 fprintf(options->file, " Bin ");
1315 fprintf(options->file, "%s%"PRIuMAX"%s",
1316 del_c, deleted, reset);
1317 fprintf(options->file, " -> ");
1318 fprintf(options->file, "%s%"PRIuMAX"%s",
1319 add_c, added, reset);
1320 fprintf(options->file, " bytes");
1321 fprintf(options->file, "\n");
1322 continue;
1323 }
1324 else if (data->files[i]->is_unmerged) {
1325 fprintf(options->file, "%s", line_prefix);
1326 show_name(options->file, prefix, name, len);
1327 fprintf(options->file, " Unmerged\n");
1328 continue;
1329 }
1330 else if (!data->files[i]->is_renamed &&
1331 (added + deleted == 0)) {
1332 total_files--;
1333 continue;
1334 }
1335
1336 /*
1337 * scale the add/delete
1338 */
1339 add = added;
1340 del = deleted;
1341 adds += add;
1342 dels += del;
1343
1344 if (width <= max_change) {
1345 add = scale_linear(add, width, max_change);
1346 del = scale_linear(del, width, max_change);
1347 }
1348 fprintf(options->file, "%s", line_prefix);
1349 show_name(options->file, prefix, name, len);
1350 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1351 added + deleted ? " " : "");
1352 show_graph(options->file, '+', add, add_c, reset);
1353 show_graph(options->file, '-', del, del_c, reset);
1354 fprintf(options->file, "\n");
1355 }
1356 fprintf(options->file, "%s", line_prefix);
1357 fprintf(options->file,
1358 " %d files changed, %d insertions(+), %d deletions(-)\n",
1359 total_files, adds, dels);
1360 }
1361
1362 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1363 {
1364 int i, adds = 0, dels = 0, total_files = data->nr;
1365
1366 if (data->nr == 0)
1367 return;
1368
1369 for (i = 0; i < data->nr; i++) {
1370 if (!data->files[i]->is_binary &&
1371 !data->files[i]->is_unmerged) {
1372 int added = data->files[i]->added;
1373 int deleted= data->files[i]->deleted;
1374 if (!data->files[i]->is_renamed &&
1375 (added + deleted == 0)) {
1376 total_files--;
1377 } else {
1378 adds += added;
1379 dels += deleted;
1380 }
1381 }
1382 }
1383 if (options->output_prefix) {
1384 struct strbuf *msg = NULL;
1385 msg = options->output_prefix(options,
1386 options->output_prefix_data);
1387 fprintf(options->file, "%s", msg->buf);
1388 }
1389 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1390 total_files, adds, dels);
1391 }
1392
1393 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1394 {
1395 int i;
1396
1397 if (data->nr == 0)
1398 return;
1399
1400 for (i = 0; i < data->nr; i++) {
1401 struct diffstat_file *file = data->files[i];
1402
1403 if (options->output_prefix) {
1404 struct strbuf *msg = NULL;
1405 msg = options->output_prefix(options,
1406 options->output_prefix_data);
1407 fprintf(options->file, "%s", msg->buf);
1408 }
1409
1410 if (file->is_binary)
1411 fprintf(options->file, "-\t-\t");
1412 else
1413 fprintf(options->file,
1414 "%"PRIuMAX"\t%"PRIuMAX"\t",
1415 file->added, file->deleted);
1416 if (options->line_termination) {
1417 fill_print_name(file);
1418 if (!file->is_renamed)
1419 write_name_quoted(file->name, options->file,
1420 options->line_termination);
1421 else {
1422 fputs(file->print_name, options->file);
1423 putc(options->line_termination, options->file);
1424 }
1425 } else {
1426 if (file->is_renamed) {
1427 putc('\0', options->file);
1428 write_name_quoted(file->from_name, options->file, '\0');
1429 }
1430 write_name_quoted(file->name, options->file, '\0');
1431 }
1432 }
1433 }
1434
1435 struct dirstat_file {
1436 const char *name;
1437 unsigned long changed;
1438 };
1439
1440 struct dirstat_dir {
1441 struct dirstat_file *files;
1442 int alloc, nr, percent, cumulative;
1443 };
1444
1445 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1446 unsigned long changed, const char *base, int baselen)
1447 {
1448 unsigned long this_dir = 0;
1449 unsigned int sources = 0;
1450 const char *line_prefix = "";
1451 struct strbuf *msg = NULL;
1452
1453 if (opt->output_prefix) {
1454 msg = opt->output_prefix(opt, opt->output_prefix_data);
1455 line_prefix = msg->buf;
1456 }
1457
1458 while (dir->nr) {
1459 struct dirstat_file *f = dir->files;
1460 int namelen = strlen(f->name);
1461 unsigned long this;
1462 char *slash;
1463
1464 if (namelen < baselen)
1465 break;
1466 if (memcmp(f->name, base, baselen))
1467 break;
1468 slash = strchr(f->name + baselen, '/');
1469 if (slash) {
1470 int newbaselen = slash + 1 - f->name;
1471 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1472 sources++;
1473 } else {
1474 this = f->changed;
1475 dir->files++;
1476 dir->nr--;
1477 sources += 2;
1478 }
1479 this_dir += this;
1480 }
1481
1482 /*
1483 * We don't report dirstat's for
1484 * - the top level
1485 * - or cases where everything came from a single directory
1486 * under this directory (sources == 1).
1487 */
1488 if (baselen && sources != 1) {
1489 int permille = this_dir * 1000 / changed;
1490 if (permille) {
1491 int percent = permille / 10;
1492 if (percent >= dir->percent) {
1493 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1494 percent, permille % 10, baselen, base);
1495 if (!dir->cumulative)
1496 return 0;
1497 }
1498 }
1499 }
1500 return this_dir;
1501 }
1502
1503 static int dirstat_compare(const void *_a, const void *_b)
1504 {
1505 const struct dirstat_file *a = _a;
1506 const struct dirstat_file *b = _b;
1507 return strcmp(a->name, b->name);
1508 }
1509
1510 static void show_dirstat(struct diff_options *options)
1511 {
1512 int i;
1513 unsigned long changed;
1514 struct dirstat_dir dir;
1515 struct diff_queue_struct *q = &diff_queued_diff;
1516
1517 dir.files = NULL;
1518 dir.alloc = 0;
1519 dir.nr = 0;
1520 dir.percent = options->dirstat_percent;
1521 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1522
1523 changed = 0;
1524 for (i = 0; i < q->nr; i++) {
1525 struct diff_filepair *p = q->queue[i];
1526 const char *name;
1527 unsigned long copied, added, damage;
1528
1529 name = p->one->path ? p->one->path : p->two->path;
1530
1531 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1532 diff_populate_filespec(p->one, 0);
1533 diff_populate_filespec(p->two, 0);
1534 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1535 &copied, &added);
1536 diff_free_filespec_data(p->one);
1537 diff_free_filespec_data(p->two);
1538 } else if (DIFF_FILE_VALID(p->one)) {
1539 diff_populate_filespec(p->one, 1);
1540 copied = added = 0;
1541 diff_free_filespec_data(p->one);
1542 } else if (DIFF_FILE_VALID(p->two)) {
1543 diff_populate_filespec(p->two, 1);
1544 copied = 0;
1545 added = p->two->size;
1546 diff_free_filespec_data(p->two);
1547 } else
1548 continue;
1549
1550 /*
1551 * Original minus copied is the removed material,
1552 * added is the new material. They are both damages
1553 * made to the preimage. In --dirstat-by-file mode, count
1554 * damaged files, not damaged lines. This is done by
1555 * counting only a single damaged line per file.
1556 */
1557 damage = (p->one->size - copied) + added;
1558 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE) && damage > 0)
1559 damage = 1;
1560
1561 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1562 dir.files[dir.nr].name = name;
1563 dir.files[dir.nr].changed = damage;
1564 changed += damage;
1565 dir.nr++;
1566 }
1567
1568 /* This can happen even with many files, if everything was renames */
1569 if (!changed)
1570 return;
1571
1572 /* Show all directories with more than x% of the changes */
1573 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1574 gather_dirstat(options, &dir, changed, "", 0);
1575 }
1576
1577 static void free_diffstat_info(struct diffstat_t *diffstat)
1578 {
1579 int i;
1580 for (i = 0; i < diffstat->nr; i++) {
1581 struct diffstat_file *f = diffstat->files[i];
1582 if (f->name != f->print_name)
1583 free(f->print_name);
1584 free(f->name);
1585 free(f->from_name);
1586 free(f);
1587 }
1588 free(diffstat->files);
1589 }
1590
1591 struct checkdiff_t {
1592 const char *filename;
1593 int lineno;
1594 int conflict_marker_size;
1595 struct diff_options *o;
1596 unsigned ws_rule;
1597 unsigned status;
1598 };
1599
1600 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1601 {
1602 char firstchar;
1603 int cnt;
1604
1605 if (len < marker_size + 1)
1606 return 0;
1607 firstchar = line[0];
1608 switch (firstchar) {
1609 case '=': case '>': case '<': case '|':
1610 break;
1611 default:
1612 return 0;
1613 }
1614 for (cnt = 1; cnt < marker_size; cnt++)
1615 if (line[cnt] != firstchar)
1616 return 0;
1617 /* line[1] thru line[marker_size-1] are same as firstchar */
1618 if (len < marker_size + 1 || !isspace(line[marker_size]))
1619 return 0;
1620 return 1;
1621 }
1622
1623 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1624 {
1625 struct checkdiff_t *data = priv;
1626 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1627 int marker_size = data->conflict_marker_size;
1628 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1629 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1630 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1631 char *err;
1632 char *line_prefix = "";
1633 struct strbuf *msgbuf;
1634
1635 assert(data->o);
1636 if (data->o->output_prefix) {
1637 msgbuf = data->o->output_prefix(data->o,
1638 data->o->output_prefix_data);
1639 line_prefix = msgbuf->buf;
1640 }
1641
1642 if (line[0] == '+') {
1643 unsigned bad;
1644 data->lineno++;
1645 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1646 data->status |= 1;
1647 fprintf(data->o->file,
1648 "%s%s:%d: leftover conflict marker\n",
1649 line_prefix, data->filename, data->lineno);
1650 }
1651 bad = ws_check(line + 1, len - 1, data->ws_rule);
1652 if (!bad)
1653 return;
1654 data->status |= bad;
1655 err = whitespace_error_string(bad);
1656 fprintf(data->o->file, "%s%s:%d: %s.\n",
1657 line_prefix, data->filename, data->lineno, err);
1658 free(err);
1659 emit_line(data->o, set, reset, line, 1);
1660 ws_check_emit(line + 1, len - 1, data->ws_rule,
1661 data->o->file, set, reset, ws);
1662 } else if (line[0] == ' ') {
1663 data->lineno++;
1664 } else if (line[0] == '@') {
1665 char *plus = strchr(line, '+');
1666 if (plus)
1667 data->lineno = strtol(plus, NULL, 10) - 1;
1668 else
1669 die("invalid diff");
1670 }
1671 }
1672
1673 static unsigned char *deflate_it(char *data,
1674 unsigned long size,
1675 unsigned long *result_size)
1676 {
1677 int bound;
1678 unsigned char *deflated;
1679 z_stream stream;
1680
1681 memset(&stream, 0, sizeof(stream));
1682 deflateInit(&stream, zlib_compression_level);
1683 bound = deflateBound(&stream, size);
1684 deflated = xmalloc(bound);
1685 stream.next_out = deflated;
1686 stream.avail_out = bound;
1687
1688 stream.next_in = (unsigned char *)data;
1689 stream.avail_in = size;
1690 while (deflate(&stream, Z_FINISH) == Z_OK)
1691 ; /* nothing */
1692 deflateEnd(&stream);
1693 *result_size = stream.total_out;
1694 return deflated;
1695 }
1696
1697 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1698 {
1699 void *cp;
1700 void *delta;
1701 void *deflated;
1702 void *data;
1703 unsigned long orig_size;
1704 unsigned long delta_size;
1705 unsigned long deflate_size;
1706 unsigned long data_size;
1707
1708 /* We could do deflated delta, or we could do just deflated two,
1709 * whichever is smaller.
1710 */
1711 delta = NULL;
1712 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1713 if (one->size && two->size) {
1714 delta = diff_delta(one->ptr, one->size,
1715 two->ptr, two->size,
1716 &delta_size, deflate_size);
1717 if (delta) {
1718 void *to_free = delta;
1719 orig_size = delta_size;
1720 delta = deflate_it(delta, delta_size, &delta_size);
1721 free(to_free);
1722 }
1723 }
1724
1725 if (delta && delta_size < deflate_size) {
1726 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1727 free(deflated);
1728 data = delta;
1729 data_size = delta_size;
1730 }
1731 else {
1732 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1733 free(delta);
1734 data = deflated;
1735 data_size = deflate_size;
1736 }
1737
1738 /* emit data encoded in base85 */
1739 cp = data;
1740 while (data_size) {
1741 int bytes = (52 < data_size) ? 52 : data_size;
1742 char line[70];
1743 data_size -= bytes;
1744 if (bytes <= 26)
1745 line[0] = bytes + 'A' - 1;
1746 else
1747 line[0] = bytes - 26 + 'a' - 1;
1748 encode_85(line + 1, cp, bytes);
1749 cp = (char *) cp + bytes;
1750 fprintf(file, "%s", prefix);
1751 fputs(line, file);
1752 fputc('\n', file);
1753 }
1754 fprintf(file, "%s\n", prefix);
1755 free(data);
1756 }
1757
1758 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1759 {
1760 fprintf(file, "%sGIT binary patch\n", prefix);
1761 emit_binary_diff_body(file, one, two, prefix);
1762 emit_binary_diff_body(file, two, one, prefix);
1763 }
1764
1765 static void diff_filespec_load_driver(struct diff_filespec *one)
1766 {
1767 /* Use already-loaded driver */
1768 if (one->driver)
1769 return;
1770
1771 if (S_ISREG(one->mode))
1772 one->driver = userdiff_find_by_path(one->path);
1773
1774 /* Fallback to default settings */
1775 if (!one->driver)
1776 one->driver = userdiff_find_by_name("default");
1777 }
1778
1779 int diff_filespec_is_binary(struct diff_filespec *one)
1780 {
1781 if (one->is_binary == -1) {
1782 diff_filespec_load_driver(one);
1783 if (one->driver->binary != -1)
1784 one->is_binary = one->driver->binary;
1785 else {
1786 if (!one->data && DIFF_FILE_VALID(one))
1787 diff_populate_filespec(one, 0);
1788 if (one->data)
1789 one->is_binary = buffer_is_binary(one->data,
1790 one->size);
1791 if (one->is_binary == -1)
1792 one->is_binary = 0;
1793 }
1794 }
1795 return one->is_binary;
1796 }
1797
1798 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1799 {
1800 diff_filespec_load_driver(one);
1801 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1802 }
1803
1804 static const char *userdiff_word_regex(struct diff_filespec *one)
1805 {
1806 diff_filespec_load_driver(one);
1807 return one->driver->word_regex;
1808 }
1809
1810 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1811 {
1812 if (!options->a_prefix)
1813 options->a_prefix = a;
1814 if (!options->b_prefix)
1815 options->b_prefix = b;
1816 }
1817
1818 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1819 {
1820 if (!DIFF_FILE_VALID(one))
1821 return NULL;
1822
1823 diff_filespec_load_driver(one);
1824 if (!one->driver->textconv)
1825 return NULL;
1826
1827 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1828 struct notes_cache *c = xmalloc(sizeof(*c));
1829 struct strbuf name = STRBUF_INIT;
1830
1831 strbuf_addf(&name, "textconv/%s", one->driver->name);
1832 notes_cache_init(c, name.buf, one->driver->textconv);
1833 one->driver->textconv_cache = c;
1834 }
1835
1836 return one->driver;
1837 }
1838
1839 static void builtin_diff(const char *name_a,
1840 const char *name_b,
1841 struct diff_filespec *one,
1842 struct diff_filespec *two,
1843 const char *xfrm_msg,
1844 int must_show_header,
1845 struct diff_options *o,
1846 int complete_rewrite)
1847 {
1848 mmfile_t mf1, mf2;
1849 const char *lbl[2];
1850 char *a_one, *b_two;
1851 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1852 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1853 const char *a_prefix, *b_prefix;
1854 struct userdiff_driver *textconv_one = NULL;
1855 struct userdiff_driver *textconv_two = NULL;
1856 struct strbuf header = STRBUF_INIT;
1857 struct strbuf *msgbuf;
1858 char *line_prefix = "";
1859
1860 if (o->output_prefix) {
1861 msgbuf = o->output_prefix(o, o->output_prefix_data);
1862 line_prefix = msgbuf->buf;
1863 }
1864
1865 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1866 (!one->mode || S_ISGITLINK(one->mode)) &&
1867 (!two->mode || S_ISGITLINK(two->mode))) {
1868 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1869 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1870 show_submodule_summary(o->file, one ? one->path : two->path,
1871 one->sha1, two->sha1, two->dirty_submodule,
1872 del, add, reset);
1873 return;
1874 }
1875
1876 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1877 textconv_one = get_textconv(one);
1878 textconv_two = get_textconv(two);
1879 }
1880
1881 diff_set_mnemonic_prefix(o, "a/", "b/");
1882 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1883 a_prefix = o->b_prefix;
1884 b_prefix = o->a_prefix;
1885 } else {
1886 a_prefix = o->a_prefix;
1887 b_prefix = o->b_prefix;
1888 }
1889
1890 /* Never use a non-valid filename anywhere if at all possible */
1891 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1892 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1893
1894 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1895 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1896 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1897 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1898 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1899 if (lbl[0][0] == '/') {
1900 /* /dev/null */
1901 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1902 if (xfrm_msg)
1903 strbuf_addstr(&header, xfrm_msg);
1904 must_show_header = 1;
1905 }
1906 else if (lbl[1][0] == '/') {
1907 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1908 if (xfrm_msg)
1909 strbuf_addstr(&header, xfrm_msg);
1910 must_show_header = 1;
1911 }
1912 else {
1913 if (one->mode != two->mode) {
1914 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1915 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1916 must_show_header = 1;
1917 }
1918 if (xfrm_msg)
1919 strbuf_addstr(&header, xfrm_msg);
1920
1921 /*
1922 * we do not run diff between different kind
1923 * of objects.
1924 */
1925 if ((one->mode ^ two->mode) & S_IFMT)
1926 goto free_ab_and_return;
1927 if (complete_rewrite &&
1928 (textconv_one || !diff_filespec_is_binary(one)) &&
1929 (textconv_two || !diff_filespec_is_binary(two))) {
1930 fprintf(o->file, "%s", header.buf);
1931 strbuf_reset(&header);
1932 emit_rewrite_diff(name_a, name_b, one, two,
1933 textconv_one, textconv_two, o);
1934 o->found_changes = 1;
1935 goto free_ab_and_return;
1936 }
1937 }
1938
1939 if (!DIFF_OPT_TST(o, TEXT) &&
1940 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1941 (!textconv_two && diff_filespec_is_binary(two)) )) {
1942 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1943 die("unable to read files to diff");
1944 /* Quite common confusing case */
1945 if (mf1.size == mf2.size &&
1946 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
1947 if (must_show_header)
1948 fprintf(o->file, "%s", header.buf);
1949 goto free_ab_and_return;
1950 }
1951 fprintf(o->file, "%s", header.buf);
1952 strbuf_reset(&header);
1953 if (DIFF_OPT_TST(o, BINARY))
1954 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
1955 else
1956 fprintf(o->file, "%sBinary files %s and %s differ\n",
1957 line_prefix, lbl[0], lbl[1]);
1958 o->found_changes = 1;
1959 }
1960 else {
1961 /* Crazy xdl interfaces.. */
1962 const char *diffopts = getenv("GIT_DIFF_OPTS");
1963 xpparam_t xpp;
1964 xdemitconf_t xecfg;
1965 struct emit_callback ecbdata;
1966 const struct userdiff_funcname *pe;
1967
1968 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
1969 fprintf(o->file, "%s", header.buf);
1970 strbuf_reset(&header);
1971 }
1972
1973 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
1974 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
1975
1976 pe = diff_funcname_pattern(one);
1977 if (!pe)
1978 pe = diff_funcname_pattern(two);
1979
1980 memset(&xpp, 0, sizeof(xpp));
1981 memset(&xecfg, 0, sizeof(xecfg));
1982 memset(&ecbdata, 0, sizeof(ecbdata));
1983 ecbdata.label_path = lbl;
1984 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
1985 ecbdata.found_changesp = &o->found_changes;
1986 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
1987 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
1988 check_blank_at_eof(&mf1, &mf2, &ecbdata);
1989 ecbdata.opt = o;
1990 ecbdata.header = header.len ? &header : NULL;
1991 xpp.flags = o->xdl_opts;
1992 xecfg.ctxlen = o->context;
1993 xecfg.interhunkctxlen = o->interhunkcontext;
1994 xecfg.flags = XDL_EMIT_FUNCNAMES;
1995 if (pe)
1996 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
1997 if (!diffopts)
1998 ;
1999 else if (!prefixcmp(diffopts, "--unified="))
2000 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2001 else if (!prefixcmp(diffopts, "-u"))
2002 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2003 if (o->word_diff) {
2004 int i;
2005
2006 ecbdata.diff_words =
2007 xcalloc(1, sizeof(struct diff_words_data));
2008 ecbdata.diff_words->type = o->word_diff;
2009 ecbdata.diff_words->opt = o;
2010 if (!o->word_regex)
2011 o->word_regex = userdiff_word_regex(one);
2012 if (!o->word_regex)
2013 o->word_regex = userdiff_word_regex(two);
2014 if (!o->word_regex)
2015 o->word_regex = diff_word_regex_cfg;
2016 if (o->word_regex) {
2017 ecbdata.diff_words->word_regex = (regex_t *)
2018 xmalloc(sizeof(regex_t));
2019 if (regcomp(ecbdata.diff_words->word_regex,
2020 o->word_regex,
2021 REG_EXTENDED | REG_NEWLINE))
2022 die ("Invalid regular expression: %s",
2023 o->word_regex);
2024 }
2025 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2026 if (o->word_diff == diff_words_styles[i].type) {
2027 ecbdata.diff_words->style =
2028 &diff_words_styles[i];
2029 break;
2030 }
2031 }
2032 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2033 struct diff_words_style *st = ecbdata.diff_words->style;
2034 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2035 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2036 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2037 }
2038 }
2039 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2040 &xpp, &xecfg);
2041 if (o->word_diff)
2042 free_diff_words_data(&ecbdata);
2043 if (textconv_one)
2044 free(mf1.ptr);
2045 if (textconv_two)
2046 free(mf2.ptr);
2047 xdiff_clear_find_func(&xecfg);
2048 }
2049
2050 free_ab_and_return:
2051 strbuf_release(&header);
2052 diff_free_filespec_data(one);
2053 diff_free_filespec_data(two);
2054 free(a_one);
2055 free(b_two);
2056 return;
2057 }
2058
2059 static void builtin_diffstat(const char *name_a, const char *name_b,
2060 struct diff_filespec *one,
2061 struct diff_filespec *two,
2062 struct diffstat_t *diffstat,
2063 struct diff_options *o,
2064 int complete_rewrite)
2065 {
2066 mmfile_t mf1, mf2;
2067 struct diffstat_file *data;
2068
2069 data = diffstat_add(diffstat, name_a, name_b);
2070
2071 if (!one || !two) {
2072 data->is_unmerged = 1;
2073 return;
2074 }
2075 if (complete_rewrite) {
2076 diff_populate_filespec(one, 0);
2077 diff_populate_filespec(two, 0);
2078 data->deleted = count_lines(one->data, one->size);
2079 data->added = count_lines(two->data, two->size);
2080 goto free_and_return;
2081 }
2082 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2083 die("unable to read files to diff");
2084
2085 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2086 data->is_binary = 1;
2087 data->added = mf2.size;
2088 data->deleted = mf1.size;
2089 } else {
2090 /* Crazy xdl interfaces.. */
2091 xpparam_t xpp;
2092 xdemitconf_t xecfg;
2093
2094 memset(&xpp, 0, sizeof(xpp));
2095 memset(&xecfg, 0, sizeof(xecfg));
2096 xpp.flags = o->xdl_opts;
2097 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2098 &xpp, &xecfg);
2099 }
2100
2101 free_and_return:
2102 diff_free_filespec_data(one);
2103 diff_free_filespec_data(two);
2104 }
2105
2106 static void builtin_checkdiff(const char *name_a, const char *name_b,
2107 const char *attr_path,
2108 struct diff_filespec *one,
2109 struct diff_filespec *two,
2110 struct diff_options *o)
2111 {
2112 mmfile_t mf1, mf2;
2113 struct checkdiff_t data;
2114
2115 if (!two)
2116 return;
2117
2118 memset(&data, 0, sizeof(data));
2119 data.filename = name_b ? name_b : name_a;
2120 data.lineno = 0;
2121 data.o = o;
2122 data.ws_rule = whitespace_rule(attr_path);
2123 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2124
2125 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2126 die("unable to read files to diff");
2127
2128 /*
2129 * All the other codepaths check both sides, but not checking
2130 * the "old" side here is deliberate. We are checking the newly
2131 * introduced changes, and as long as the "new" side is text, we
2132 * can and should check what it introduces.
2133 */
2134 if (diff_filespec_is_binary(two))
2135 goto free_and_return;
2136 else {
2137 /* Crazy xdl interfaces.. */
2138 xpparam_t xpp;
2139 xdemitconf_t xecfg;
2140
2141 memset(&xpp, 0, sizeof(xpp));
2142 memset(&xecfg, 0, sizeof(xecfg));
2143 xecfg.ctxlen = 1; /* at least one context line */
2144 xpp.flags = 0;
2145 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2146 &xpp, &xecfg);
2147
2148 if (data.ws_rule & WS_BLANK_AT_EOF) {
2149 struct emit_callback ecbdata;
2150 int blank_at_eof;
2151
2152 ecbdata.ws_rule = data.ws_rule;
2153 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2154 blank_at_eof = ecbdata.blank_at_eof_in_preimage;
2155
2156 if (blank_at_eof) {
2157 static char *err;
2158 if (!err)
2159 err = whitespace_error_string(WS_BLANK_AT_EOF);
2160 fprintf(o->file, "%s:%d: %s.\n",
2161 data.filename, blank_at_eof, err);
2162 data.status = 1; /* report errors */
2163 }
2164 }
2165 }
2166 free_and_return:
2167 diff_free_filespec_data(one);
2168 diff_free_filespec_data(two);
2169 if (data.status)
2170 DIFF_OPT_SET(o, CHECK_FAILED);
2171 }
2172
2173 struct diff_filespec *alloc_filespec(const char *path)
2174 {
2175 int namelen = strlen(path);
2176 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2177
2178 memset(spec, 0, sizeof(*spec));
2179 spec->path = (char *)(spec + 1);
2180 memcpy(spec->path, path, namelen+1);
2181 spec->count = 1;
2182 spec->is_binary = -1;
2183 return spec;
2184 }
2185
2186 void free_filespec(struct diff_filespec *spec)
2187 {
2188 if (!--spec->count) {
2189 diff_free_filespec_data(spec);
2190 free(spec);
2191 }
2192 }
2193
2194 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2195 unsigned short mode)
2196 {
2197 if (mode) {
2198 spec->mode = canon_mode(mode);
2199 hashcpy(spec->sha1, sha1);
2200 spec->sha1_valid = !is_null_sha1(sha1);
2201 }
2202 }
2203
2204 /*
2205 * Given a name and sha1 pair, if the index tells us the file in
2206 * the work tree has that object contents, return true, so that
2207 * prepare_temp_file() does not have to inflate and extract.
2208 */
2209 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2210 {
2211 struct cache_entry *ce;
2212 struct stat st;
2213 int pos, len;
2214
2215 /*
2216 * We do not read the cache ourselves here, because the
2217 * benchmark with my previous version that always reads cache
2218 * shows that it makes things worse for diff-tree comparing
2219 * two linux-2.6 kernel trees in an already checked out work
2220 * tree. This is because most diff-tree comparisons deal with
2221 * only a small number of files, while reading the cache is
2222 * expensive for a large project, and its cost outweighs the
2223 * savings we get by not inflating the object to a temporary
2224 * file. Practically, this code only helps when we are used
2225 * by diff-cache --cached, which does read the cache before
2226 * calling us.
2227 */
2228 if (!active_cache)
2229 return 0;
2230
2231 /* We want to avoid the working directory if our caller
2232 * doesn't need the data in a normal file, this system
2233 * is rather slow with its stat/open/mmap/close syscalls,
2234 * and the object is contained in a pack file. The pack
2235 * is probably already open and will be faster to obtain
2236 * the data through than the working directory. Loose
2237 * objects however would tend to be slower as they need
2238 * to be individually opened and inflated.
2239 */
2240 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2241 return 0;
2242
2243 len = strlen(name);
2244 pos = cache_name_pos(name, len);
2245 if (pos < 0)
2246 return 0;
2247 ce = active_cache[pos];
2248
2249 /*
2250 * This is not the sha1 we are looking for, or
2251 * unreusable because it is not a regular file.
2252 */
2253 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2254 return 0;
2255
2256 /*
2257 * If ce is marked as "assume unchanged", there is no
2258 * guarantee that work tree matches what we are looking for.
2259 */
2260 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2261 return 0;
2262
2263 /*
2264 * If ce matches the file in the work tree, we can reuse it.
2265 */
2266 if (ce_uptodate(ce) ||
2267 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2268 return 1;
2269
2270 return 0;
2271 }
2272
2273 static int populate_from_stdin(struct diff_filespec *s)
2274 {
2275 struct strbuf buf = STRBUF_INIT;
2276 size_t size = 0;
2277
2278 if (strbuf_read(&buf, 0, 0) < 0)
2279 return error("error while reading from stdin %s",
2280 strerror(errno));
2281
2282 s->should_munmap = 0;
2283 s->data = strbuf_detach(&buf, &size);
2284 s->size = size;
2285 s->should_free = 1;
2286 return 0;
2287 }
2288
2289 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2290 {
2291 int len;
2292 char *data = xmalloc(100), *dirty = "";
2293
2294 /* Are we looking at the work tree? */
2295 if (s->dirty_submodule)
2296 dirty = "-dirty";
2297
2298 len = snprintf(data, 100,
2299 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2300 s->data = data;
2301 s->size = len;
2302 s->should_free = 1;
2303 if (size_only) {
2304 s->data = NULL;
2305 free(data);
2306 }
2307 return 0;
2308 }
2309
2310 /*
2311 * While doing rename detection and pickaxe operation, we may need to
2312 * grab the data for the blob (or file) for our own in-core comparison.
2313 * diff_filespec has data and size fields for this purpose.
2314 */
2315 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2316 {
2317 int err = 0;
2318 if (!DIFF_FILE_VALID(s))
2319 die("internal error: asking to populate invalid file.");
2320 if (S_ISDIR(s->mode))
2321 return -1;
2322
2323 if (s->data)
2324 return 0;
2325
2326 if (size_only && 0 < s->size)
2327 return 0;
2328
2329 if (S_ISGITLINK(s->mode))
2330 return diff_populate_gitlink(s, size_only);
2331
2332 if (!s->sha1_valid ||
2333 reuse_worktree_file(s->path, s->sha1, 0)) {
2334 struct strbuf buf = STRBUF_INIT;
2335 struct stat st;
2336 int fd;
2337
2338 if (!strcmp(s->path, "-"))
2339 return populate_from_stdin(s);
2340
2341 if (lstat(s->path, &st) < 0) {
2342 if (errno == ENOENT) {
2343 err_empty:
2344 err = -1;
2345 empty:
2346 s->data = (char *)"";
2347 s->size = 0;
2348 return err;
2349 }
2350 }
2351 s->size = xsize_t(st.st_size);
2352 if (!s->size)
2353 goto empty;
2354 if (S_ISLNK(st.st_mode)) {
2355 struct strbuf sb = STRBUF_INIT;
2356
2357 if (strbuf_readlink(&sb, s->path, s->size))
2358 goto err_empty;
2359 s->size = sb.len;
2360 s->data = strbuf_detach(&sb, NULL);
2361 s->should_free = 1;
2362 return 0;
2363 }
2364 if (size_only)
2365 return 0;
2366 fd = open(s->path, O_RDONLY);
2367 if (fd < 0)
2368 goto err_empty;
2369 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2370 close(fd);
2371 s->should_munmap = 1;
2372
2373 /*
2374 * Convert from working tree format to canonical git format
2375 */
2376 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2377 size_t size = 0;
2378 munmap(s->data, s->size);
2379 s->should_munmap = 0;
2380 s->data = strbuf_detach(&buf, &size);
2381 s->size = size;
2382 s->should_free = 1;
2383 }
2384 }
2385 else {
2386 enum object_type type;
2387 if (size_only)
2388 type = sha1_object_info(s->sha1, &s->size);
2389 else {
2390 s->data = read_sha1_file(s->sha1, &type, &s->size);
2391 s->should_free = 1;
2392 }
2393 }
2394 return 0;
2395 }
2396
2397 void diff_free_filespec_blob(struct diff_filespec *s)
2398 {
2399 if (s->should_free)
2400 free(s->data);
2401 else if (s->should_munmap)
2402 munmap(s->data, s->size);
2403
2404 if (s->should_free || s->should_munmap) {
2405 s->should_free = s->should_munmap = 0;
2406 s->data = NULL;
2407 }
2408 }
2409
2410 void diff_free_filespec_data(struct diff_filespec *s)
2411 {
2412 diff_free_filespec_blob(s);
2413 free(s->cnt_data);
2414 s->cnt_data = NULL;
2415 }
2416
2417 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2418 void *blob,
2419 unsigned long size,
2420 const unsigned char *sha1,
2421 int mode)
2422 {
2423 int fd;
2424 struct strbuf buf = STRBUF_INIT;
2425 struct strbuf template = STRBUF_INIT;
2426 char *path_dup = xstrdup(path);
2427 const char *base = basename(path_dup);
2428
2429 /* Generate "XXXXXX_basename.ext" */
2430 strbuf_addstr(&template, "XXXXXX_");
2431 strbuf_addstr(&template, base);
2432
2433 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2434 strlen(base) + 1);
2435 if (fd < 0)
2436 die_errno("unable to create temp-file");
2437 if (convert_to_working_tree(path,
2438 (const char *)blob, (size_t)size, &buf)) {
2439 blob = buf.buf;
2440 size = buf.len;
2441 }
2442 if (write_in_full(fd, blob, size) != size)
2443 die_errno("unable to write temp-file");
2444 close(fd);
2445 temp->name = temp->tmp_path;
2446 strcpy(temp->hex, sha1_to_hex(sha1));
2447 temp->hex[40] = 0;
2448 sprintf(temp->mode, "%06o", mode);
2449 strbuf_release(&buf);
2450 strbuf_release(&template);
2451 free(path_dup);
2452 }
2453
2454 static struct diff_tempfile *prepare_temp_file(const char *name,
2455 struct diff_filespec *one)
2456 {
2457 struct diff_tempfile *temp = claim_diff_tempfile();
2458
2459 if (!DIFF_FILE_VALID(one)) {
2460 not_a_valid_file:
2461 /* A '-' entry produces this for file-2, and
2462 * a '+' entry produces this for file-1.
2463 */
2464 temp->name = "/dev/null";
2465 strcpy(temp->hex, ".");
2466 strcpy(temp->mode, ".");
2467 return temp;
2468 }
2469
2470 if (!remove_tempfile_installed) {
2471 atexit(remove_tempfile);
2472 sigchain_push_common(remove_tempfile_on_signal);
2473 remove_tempfile_installed = 1;
2474 }
2475
2476 if (!one->sha1_valid ||
2477 reuse_worktree_file(name, one->sha1, 1)) {
2478 struct stat st;
2479 if (lstat(name, &st) < 0) {
2480 if (errno == ENOENT)
2481 goto not_a_valid_file;
2482 die_errno("stat(%s)", name);
2483 }
2484 if (S_ISLNK(st.st_mode)) {
2485 struct strbuf sb = STRBUF_INIT;
2486 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2487 die_errno("readlink(%s)", name);
2488 prep_temp_blob(name, temp, sb.buf, sb.len,
2489 (one->sha1_valid ?
2490 one->sha1 : null_sha1),
2491 (one->sha1_valid ?
2492 one->mode : S_IFLNK));
2493 strbuf_release(&sb);
2494 }
2495 else {
2496 /* we can borrow from the file in the work tree */
2497 temp->name = name;
2498 if (!one->sha1_valid)
2499 strcpy(temp->hex, sha1_to_hex(null_sha1));
2500 else
2501 strcpy(temp->hex, sha1_to_hex(one->sha1));
2502 /* Even though we may sometimes borrow the
2503 * contents from the work tree, we always want
2504 * one->mode. mode is trustworthy even when
2505 * !(one->sha1_valid), as long as
2506 * DIFF_FILE_VALID(one).
2507 */
2508 sprintf(temp->mode, "%06o", one->mode);
2509 }
2510 return temp;
2511 }
2512 else {
2513 if (diff_populate_filespec(one, 0))
2514 die("cannot read data blob for %s", one->path);
2515 prep_temp_blob(name, temp, one->data, one->size,
2516 one->sha1, one->mode);
2517 }
2518 return temp;
2519 }
2520
2521 /* An external diff command takes:
2522 *
2523 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2524 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2525 *
2526 */
2527 static void run_external_diff(const char *pgm,
2528 const char *name,
2529 const char *other,
2530 struct diff_filespec *one,
2531 struct diff_filespec *two,
2532 const char *xfrm_msg,
2533 int complete_rewrite)
2534 {
2535 const char *spawn_arg[10];
2536 int retval;
2537 const char **arg = &spawn_arg[0];
2538
2539 if (one && two) {
2540 struct diff_tempfile *temp_one, *temp_two;
2541 const char *othername = (other ? other : name);
2542 temp_one = prepare_temp_file(name, one);
2543 temp_two = prepare_temp_file(othername, two);
2544 *arg++ = pgm;
2545 *arg++ = name;
2546 *arg++ = temp_one->name;
2547 *arg++ = temp_one->hex;
2548 *arg++ = temp_one->mode;
2549 *arg++ = temp_two->name;
2550 *arg++ = temp_two->hex;
2551 *arg++ = temp_two->mode;
2552 if (other) {
2553 *arg++ = other;
2554 *arg++ = xfrm_msg;
2555 }
2556 } else {
2557 *arg++ = pgm;
2558 *arg++ = name;
2559 }
2560 *arg = NULL;
2561 fflush(NULL);
2562 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2563 remove_tempfile();
2564 if (retval) {
2565 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2566 exit(1);
2567 }
2568 }
2569
2570 static int similarity_index(struct diff_filepair *p)
2571 {
2572 return p->score * 100 / MAX_SCORE;
2573 }
2574
2575 static void fill_metainfo(struct strbuf *msg,
2576 const char *name,
2577 const char *other,
2578 struct diff_filespec *one,
2579 struct diff_filespec *two,
2580 struct diff_options *o,
2581 struct diff_filepair *p,
2582 int *must_show_header,
2583 int use_color)
2584 {
2585 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2586 const char *reset = diff_get_color(use_color, DIFF_RESET);
2587 struct strbuf *msgbuf;
2588 char *line_prefix = "";
2589
2590 *must_show_header = 1;
2591 if (o->output_prefix) {
2592 msgbuf = o->output_prefix(o, o->output_prefix_data);
2593 line_prefix = msgbuf->buf;
2594 }
2595 strbuf_init(msg, PATH_MAX * 2 + 300);
2596 switch (p->status) {
2597 case DIFF_STATUS_COPIED:
2598 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2599 line_prefix, set, similarity_index(p));
2600 strbuf_addf(msg, "%s\n%s%scopy from ",
2601 reset, line_prefix, set);
2602 quote_c_style(name, msg, NULL, 0);
2603 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2604 quote_c_style(other, msg, NULL, 0);
2605 strbuf_addf(msg, "%s\n", reset);
2606 break;
2607 case DIFF_STATUS_RENAMED:
2608 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2609 line_prefix, set, similarity_index(p));
2610 strbuf_addf(msg, "%s\n%s%srename from ",
2611 reset, line_prefix, set);
2612 quote_c_style(name, msg, NULL, 0);
2613 strbuf_addf(msg, "%s\n%s%srename to ",
2614 reset, line_prefix, set);
2615 quote_c_style(other, msg, NULL, 0);
2616 strbuf_addf(msg, "%s\n", reset);
2617 break;
2618 case DIFF_STATUS_MODIFIED:
2619 if (p->score) {
2620 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2621 line_prefix,
2622 set, similarity_index(p), reset);
2623 break;
2624 }
2625 /* fallthru */
2626 default:
2627 *must_show_header = 0;
2628 }
2629 if (one && two && hashcmp(one->sha1, two->sha1)) {
2630 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2631
2632 if (DIFF_OPT_TST(o, BINARY)) {
2633 mmfile_t mf;
2634 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2635 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2636 abbrev = 40;
2637 }
2638 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2639 find_unique_abbrev(one->sha1, abbrev));
2640 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2641 if (one->mode == two->mode)
2642 strbuf_addf(msg, " %06o", one->mode);
2643 strbuf_addf(msg, "%s\n", reset);
2644 }
2645 }
2646
2647 static void run_diff_cmd(const char *pgm,
2648 const char *name,
2649 const char *other,
2650 const char *attr_path,
2651 struct diff_filespec *one,
2652 struct diff_filespec *two,
2653 struct strbuf *msg,
2654 struct diff_options *o,
2655 struct diff_filepair *p)
2656 {
2657 const char *xfrm_msg = NULL;
2658 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2659 int must_show_header = 0;
2660
2661 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2662 pgm = NULL;
2663 else {
2664 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2665 if (drv && drv->external)
2666 pgm = drv->external;
2667 }
2668
2669 if (msg) {
2670 /*
2671 * don't use colors when the header is intended for an
2672 * external diff driver
2673 */
2674 fill_metainfo(msg, name, other, one, two, o, p,
2675 &must_show_header,
2676 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2677 xfrm_msg = msg->len ? msg->buf : NULL;
2678 }
2679
2680 if (pgm) {
2681 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2682 complete_rewrite);
2683 return;
2684 }
2685 if (one && two)
2686 builtin_diff(name, other ? other : name,
2687 one, two, xfrm_msg, must_show_header,
2688 o, complete_rewrite);
2689 else
2690 fprintf(o->file, "* Unmerged path %s\n", name);
2691 }
2692
2693 static void diff_fill_sha1_info(struct diff_filespec *one)
2694 {
2695 if (DIFF_FILE_VALID(one)) {
2696 if (!one->sha1_valid) {
2697 struct stat st;
2698 if (!strcmp(one->path, "-")) {
2699 hashcpy(one->sha1, null_sha1);
2700 return;
2701 }
2702 if (lstat(one->path, &st) < 0)
2703 die_errno("stat '%s'", one->path);
2704 if (index_path(one->sha1, one->path, &st, 0))
2705 die("cannot hash %s", one->path);
2706 }
2707 }
2708 else
2709 hashclr(one->sha1);
2710 }
2711
2712 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2713 {
2714 /* Strip the prefix but do not molest /dev/null and absolute paths */
2715 if (*namep && **namep != '/') {
2716 *namep += prefix_length;
2717 if (**namep == '/')
2718 ++*namep;
2719 }
2720 if (*otherp && **otherp != '/') {
2721 *otherp += prefix_length;
2722 if (**otherp == '/')
2723 ++*otherp;
2724 }
2725 }
2726
2727 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2728 {
2729 const char *pgm = external_diff();
2730 struct strbuf msg;
2731 struct diff_filespec *one = p->one;
2732 struct diff_filespec *two = p->two;
2733 const char *name;
2734 const char *other;
2735 const char *attr_path;
2736
2737 name = p->one->path;
2738 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2739 attr_path = name;
2740 if (o->prefix_length)
2741 strip_prefix(o->prefix_length, &name, &other);
2742
2743 if (DIFF_PAIR_UNMERGED(p)) {
2744 run_diff_cmd(pgm, name, NULL, attr_path,
2745 NULL, NULL, NULL, o, p);
2746 return;
2747 }
2748
2749 diff_fill_sha1_info(one);
2750 diff_fill_sha1_info(two);
2751
2752 if (!pgm &&
2753 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2754 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2755 /*
2756 * a filepair that changes between file and symlink
2757 * needs to be split into deletion and creation.
2758 */
2759 struct diff_filespec *null = alloc_filespec(two->path);
2760 run_diff_cmd(NULL, name, other, attr_path,
2761 one, null, &msg, o, p);
2762 free(null);
2763 strbuf_release(&msg);
2764
2765 null = alloc_filespec(one->path);
2766 run_diff_cmd(NULL, name, other, attr_path,
2767 null, two, &msg, o, p);
2768 free(null);
2769 }
2770 else
2771 run_diff_cmd(pgm, name, other, attr_path,
2772 one, two, &msg, o, p);
2773
2774 strbuf_release(&msg);
2775 }
2776
2777 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2778 struct diffstat_t *diffstat)
2779 {
2780 const char *name;
2781 const char *other;
2782 int complete_rewrite = 0;
2783
2784 if (DIFF_PAIR_UNMERGED(p)) {
2785 /* unmerged */
2786 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2787 return;
2788 }
2789
2790 name = p->one->path;
2791 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2792
2793 if (o->prefix_length)
2794 strip_prefix(o->prefix_length, &name, &other);
2795
2796 diff_fill_sha1_info(p->one);
2797 diff_fill_sha1_info(p->two);
2798
2799 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2800 complete_rewrite = 1;
2801 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2802 }
2803
2804 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2805 {
2806 const char *name;
2807 const char *other;
2808 const char *attr_path;
2809
2810 if (DIFF_PAIR_UNMERGED(p)) {
2811 /* unmerged */
2812 return;
2813 }
2814
2815 name = p->one->path;
2816 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2817 attr_path = other ? other : name;
2818
2819 if (o->prefix_length)
2820 strip_prefix(o->prefix_length, &name, &other);
2821
2822 diff_fill_sha1_info(p->one);
2823 diff_fill_sha1_info(p->two);
2824
2825 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2826 }
2827
2828 void diff_setup(struct diff_options *options)
2829 {
2830 memset(options, 0, sizeof(*options));
2831
2832 options->file = stdout;
2833
2834 options->line_termination = '\n';
2835 options->break_opt = -1;
2836 options->rename_limit = -1;
2837 options->dirstat_percent = 3;
2838 options->context = 3;
2839
2840 options->change = diff_change;
2841 options->add_remove = diff_addremove;
2842 if (diff_use_color_default > 0)
2843 DIFF_OPT_SET(options, COLOR_DIFF);
2844 options->detect_rename = diff_detect_rename_default;
2845
2846 if (diff_no_prefix) {
2847 options->a_prefix = options->b_prefix = "";
2848 } else if (!diff_mnemonic_prefix) {
2849 options->a_prefix = "a/";
2850 options->b_prefix = "b/";
2851 }
2852 }
2853
2854 int diff_setup_done(struct diff_options *options)
2855 {
2856 int count = 0;
2857
2858 if (options->output_format & DIFF_FORMAT_NAME)
2859 count++;
2860 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2861 count++;
2862 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2863 count++;
2864 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2865 count++;
2866 if (count > 1)
2867 die("--name-only, --name-status, --check and -s are mutually exclusive");
2868
2869 /*
2870 * Most of the time we can say "there are changes"
2871 * only by checking if there are changed paths, but
2872 * --ignore-whitespace* options force us to look
2873 * inside contents.
2874 */
2875
2876 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2877 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2878 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2879 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2880 else
2881 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2882
2883 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2884 options->detect_rename = DIFF_DETECT_COPY;
2885
2886 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2887 options->prefix = NULL;
2888 if (options->prefix)
2889 options->prefix_length = strlen(options->prefix);
2890 else
2891 options->prefix_length = 0;
2892
2893 if (options->output_format & (DIFF_FORMAT_NAME |
2894 DIFF_FORMAT_NAME_STATUS |
2895 DIFF_FORMAT_CHECKDIFF |
2896 DIFF_FORMAT_NO_OUTPUT))
2897 options->output_format &= ~(DIFF_FORMAT_RAW |
2898 DIFF_FORMAT_NUMSTAT |
2899 DIFF_FORMAT_DIFFSTAT |
2900 DIFF_FORMAT_SHORTSTAT |
2901 DIFF_FORMAT_DIRSTAT |
2902 DIFF_FORMAT_SUMMARY |
2903 DIFF_FORMAT_PATCH);
2904
2905 /*
2906 * These cases always need recursive; we do not drop caller-supplied
2907 * recursive bits for other formats here.
2908 */
2909 if (options->output_format & (DIFF_FORMAT_PATCH |
2910 DIFF_FORMAT_NUMSTAT |
2911 DIFF_FORMAT_DIFFSTAT |
2912 DIFF_FORMAT_SHORTSTAT |
2913 DIFF_FORMAT_DIRSTAT |
2914 DIFF_FORMAT_SUMMARY |
2915 DIFF_FORMAT_CHECKDIFF))
2916 DIFF_OPT_SET(options, RECURSIVE);
2917 /*
2918 * Also pickaxe would not work very well if you do not say recursive
2919 */
2920 if (options->pickaxe)
2921 DIFF_OPT_SET(options, RECURSIVE);
2922 /*
2923 * When patches are generated, submodules diffed against the work tree
2924 * must be checked for dirtiness too so it can be shown in the output
2925 */
2926 if (options->output_format & DIFF_FORMAT_PATCH)
2927 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2928
2929 if (options->detect_rename && options->rename_limit < 0)
2930 options->rename_limit = diff_rename_limit_default;
2931 if (options->setup & DIFF_SETUP_USE_CACHE) {
2932 if (!active_cache)
2933 /* read-cache does not die even when it fails
2934 * so it is safe for us to do this here. Also
2935 * it does not smudge active_cache or active_nr
2936 * when it fails, so we do not have to worry about
2937 * cleaning it up ourselves either.
2938 */
2939 read_cache();
2940 }
2941 if (options->abbrev <= 0 || 40 < options->abbrev)
2942 options->abbrev = 40; /* full */
2943
2944 /*
2945 * It does not make sense to show the first hit we happened
2946 * to have found. It does not make sense not to return with
2947 * exit code in such a case either.
2948 */
2949 if (DIFF_OPT_TST(options, QUICK)) {
2950 options->output_format = DIFF_FORMAT_NO_OUTPUT;
2951 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
2952 }
2953
2954 return 0;
2955 }
2956
2957 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
2958 {
2959 char c, *eq;
2960 int len;
2961
2962 if (*arg != '-')
2963 return 0;
2964 c = *++arg;
2965 if (!c)
2966 return 0;
2967 if (c == arg_short) {
2968 c = *++arg;
2969 if (!c)
2970 return 1;
2971 if (val && isdigit(c)) {
2972 char *end;
2973 int n = strtoul(arg, &end, 10);
2974 if (*end)
2975 return 0;
2976 *val = n;
2977 return 1;
2978 }
2979 return 0;
2980 }
2981 if (c != '-')
2982 return 0;
2983 arg++;
2984 eq = strchr(arg, '=');
2985 if (eq)
2986 len = eq - arg;
2987 else
2988 len = strlen(arg);
2989 if (!len || strncmp(arg, arg_long, len))
2990 return 0;
2991 if (eq) {
2992 int n;
2993 char *end;
2994 if (!isdigit(*++eq))
2995 return 0;
2996 n = strtoul(eq, &end, 10);
2997 if (*end)
2998 return 0;
2999 *val = n;
3000 }
3001 return 1;
3002 }
3003
3004 static int diff_scoreopt_parse(const char *opt);
3005
3006 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3007 {
3008 const char *arg = av[0];
3009
3010 /* Output format options */
3011 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3012 options->output_format |= DIFF_FORMAT_PATCH;
3013 else if (opt_arg(arg, 'U', "unified", &options->context))
3014 options->output_format |= DIFF_FORMAT_PATCH;
3015 else if (!strcmp(arg, "--raw"))
3016 options->output_format |= DIFF_FORMAT_RAW;
3017 else if (!strcmp(arg, "--patch-with-raw"))
3018 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3019 else if (!strcmp(arg, "--numstat"))
3020 options->output_format |= DIFF_FORMAT_NUMSTAT;
3021 else if (!strcmp(arg, "--shortstat"))
3022 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3023 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3024 options->output_format |= DIFF_FORMAT_DIRSTAT;
3025 else if (!strcmp(arg, "--cumulative")) {
3026 options->output_format |= DIFF_FORMAT_DIRSTAT;
3027 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3028 } else if (opt_arg(arg, 0, "dirstat-by-file",
3029 &options->dirstat_percent)) {
3030 options->output_format |= DIFF_FORMAT_DIRSTAT;
3031 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3032 }
3033 else if (!strcmp(arg, "--check"))
3034 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3035 else if (!strcmp(arg, "--summary"))
3036 options->output_format |= DIFF_FORMAT_SUMMARY;
3037 else if (!strcmp(arg, "--patch-with-stat"))
3038 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3039 else if (!strcmp(arg, "--name-only"))
3040 options->output_format |= DIFF_FORMAT_NAME;
3041 else if (!strcmp(arg, "--name-status"))
3042 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3043 else if (!strcmp(arg, "-s"))
3044 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3045 else if (!prefixcmp(arg, "--stat")) {
3046 char *end;
3047 int width = options->stat_width;
3048 int name_width = options->stat_name_width;
3049 arg += 6;
3050 end = (char *)arg;
3051
3052 switch (*arg) {
3053 case '-':
3054 if (!prefixcmp(arg, "-width="))
3055 width = strtoul(arg + 7, &end, 10);
3056 else if (!prefixcmp(arg, "-name-width="))
3057 name_width = strtoul(arg + 12, &end, 10);
3058 break;
3059 case '=':
3060 width = strtoul(arg+1, &end, 10);
3061 if (*end == ',')
3062 name_width = strtoul(end+1, &end, 10);
3063 }
3064
3065 /* Important! This checks all the error cases! */
3066 if (*end)
3067 return 0;
3068 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3069 options->stat_name_width = name_width;
3070 options->stat_width = width;
3071 }
3072
3073 /* renames options */
3074 else if (!prefixcmp(arg, "-B")) {
3075 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3076 return -1;
3077 }
3078 else if (!prefixcmp(arg, "-M")) {
3079 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3080 return -1;
3081 options->detect_rename = DIFF_DETECT_RENAME;
3082 }
3083 else if (!prefixcmp(arg, "-C")) {
3084 if (options->detect_rename == DIFF_DETECT_COPY)
3085 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3086 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3087 return -1;
3088 options->detect_rename = DIFF_DETECT_COPY;
3089 }
3090 else if (!strcmp(arg, "--no-renames"))
3091 options->detect_rename = 0;
3092 else if (!strcmp(arg, "--relative"))
3093 DIFF_OPT_SET(options, RELATIVE_NAME);
3094 else if (!prefixcmp(arg, "--relative=")) {
3095 DIFF_OPT_SET(options, RELATIVE_NAME);
3096 options->prefix = arg + 11;
3097 }
3098
3099 /* xdiff options */
3100 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3101 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3102 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3103 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3104 else if (!strcmp(arg, "--ignore-space-at-eol"))
3105 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3106 else if (!strcmp(arg, "--patience"))
3107 DIFF_XDL_SET(options, PATIENCE_DIFF);
3108
3109 /* flags options */
3110 else if (!strcmp(arg, "--binary")) {
3111 options->output_format |= DIFF_FORMAT_PATCH;
3112 DIFF_OPT_SET(options, BINARY);
3113 }
3114 else if (!strcmp(arg, "--full-index"))
3115 DIFF_OPT_SET(options, FULL_INDEX);
3116 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3117 DIFF_OPT_SET(options, TEXT);
3118 else if (!strcmp(arg, "-R"))
3119 DIFF_OPT_SET(options, REVERSE_DIFF);
3120 else if (!strcmp(arg, "--find-copies-harder"))
3121 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3122 else if (!strcmp(arg, "--follow"))
3123 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3124 else if (!strcmp(arg, "--color"))
3125 DIFF_OPT_SET(options, COLOR_DIFF);
3126 else if (!prefixcmp(arg, "--color=")) {
3127 int value = git_config_colorbool(NULL, arg+8, -1);
3128 if (value == 0)
3129 DIFF_OPT_CLR(options, COLOR_DIFF);
3130 else if (value > 0)
3131 DIFF_OPT_SET(options, COLOR_DIFF);
3132 else
3133 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3134 }
3135 else if (!strcmp(arg, "--no-color"))
3136 DIFF_OPT_CLR(options, COLOR_DIFF);
3137 else if (!strcmp(arg, "--color-words")) {
3138 DIFF_OPT_SET(options, COLOR_DIFF);
3139 options->word_diff = DIFF_WORDS_COLOR;
3140 }
3141 else if (!prefixcmp(arg, "--color-words=")) {
3142 DIFF_OPT_SET(options, COLOR_DIFF);
3143 options->word_diff = DIFF_WORDS_COLOR;
3144 options->word_regex = arg + 14;
3145 }
3146 else if (!strcmp(arg, "--word-diff")) {
3147 if (options->word_diff == DIFF_WORDS_NONE)
3148 options->word_diff = DIFF_WORDS_PLAIN;
3149 }
3150 else if (!prefixcmp(arg, "--word-diff=")) {
3151 const char *type = arg + 12;
3152 if (!strcmp(type, "plain"))
3153 options->word_diff = DIFF_WORDS_PLAIN;
3154 else if (!strcmp(type, "color")) {
3155 DIFF_OPT_SET(options, COLOR_DIFF);
3156 options->word_diff = DIFF_WORDS_COLOR;
3157 }
3158 else if (!strcmp(type, "porcelain"))
3159 options->word_diff = DIFF_WORDS_PORCELAIN;
3160 else if (!strcmp(type, "none"))
3161 options->word_diff = DIFF_WORDS_NONE;
3162 else
3163 die("bad --word-diff argument: %s", type);
3164 }
3165 else if (!prefixcmp(arg, "--word-diff-regex=")) {
3166 if (options->word_diff == DIFF_WORDS_NONE)
3167 options->word_diff = DIFF_WORDS_PLAIN;
3168 options->word_regex = arg + 18;
3169 }
3170 else if (!strcmp(arg, "--exit-code"))
3171 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3172 else if (!strcmp(arg, "--quiet"))
3173 DIFF_OPT_SET(options, QUICK);
3174 else if (!strcmp(arg, "--ext-diff"))
3175 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3176 else if (!strcmp(arg, "--no-ext-diff"))
3177 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3178 else if (!strcmp(arg, "--textconv"))
3179 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3180 else if (!strcmp(arg, "--no-textconv"))
3181 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3182 else if (!strcmp(arg, "--ignore-submodules"))
3183 handle_ignore_submodules_arg(options, "all");
3184 else if (!prefixcmp(arg, "--ignore-submodules="))
3185 handle_ignore_submodules_arg(options, arg + 20);
3186 else if (!strcmp(arg, "--submodule"))
3187 DIFF_OPT_SET(options, SUBMODULE_LOG);
3188 else if (!prefixcmp(arg, "--submodule=")) {
3189 if (!strcmp(arg + 12, "log"))
3190 DIFF_OPT_SET(options, SUBMODULE_LOG);
3191 }
3192
3193 /* misc options */
3194 else if (!strcmp(arg, "-z"))
3195 options->line_termination = 0;
3196 else if (!prefixcmp(arg, "-l"))
3197 options->rename_limit = strtoul(arg+2, NULL, 10);
3198 else if (!prefixcmp(arg, "-S"))
3199 options->pickaxe = arg + 2;
3200 else if (!strcmp(arg, "--pickaxe-all"))
3201 options->pickaxe_opts = DIFF_PICKAXE_ALL;
3202 else if (!strcmp(arg, "--pickaxe-regex"))
3203 options->pickaxe_opts = DIFF_PICKAXE_REGEX;
3204 else if (!prefixcmp(arg, "-O"))
3205 options->orderfile = arg + 2;
3206 else if (!prefixcmp(arg, "--diff-filter="))
3207 options->filter = arg + 14;
3208 else if (!strcmp(arg, "--abbrev"))
3209 options->abbrev = DEFAULT_ABBREV;
3210 else if (!prefixcmp(arg, "--abbrev=")) {
3211 options->abbrev = strtoul(arg + 9, NULL, 10);
3212 if (options->abbrev < MINIMUM_ABBREV)
3213 options->abbrev = MINIMUM_ABBREV;
3214 else if (40 < options->abbrev)
3215 options->abbrev = 40;
3216 }
3217 else if (!prefixcmp(arg, "--src-prefix="))
3218 options->a_prefix = arg + 13;
3219 else if (!prefixcmp(arg, "--dst-prefix="))
3220 options->b_prefix = arg + 13;
3221 else if (!strcmp(arg, "--no-prefix"))
3222 options->a_prefix = options->b_prefix = "";
3223 else if (opt_arg(arg, '\0', "inter-hunk-context",
3224 &options->interhunkcontext))
3225 ;
3226 else if (!prefixcmp(arg, "--output=")) {
3227 options->file = fopen(arg + strlen("--output="), "w");
3228 if (!options->file)
3229 die_errno("Could not open '%s'", arg + strlen("--output="));
3230 options->close_file = 1;
3231 } else
3232 return 0;
3233 return 1;
3234 }
3235
3236 static int parse_num(const char **cp_p)
3237 {
3238 unsigned long num, scale;
3239 int ch, dot;
3240 const char *cp = *cp_p;
3241
3242 num = 0;
3243 scale = 1;
3244 dot = 0;
3245 for (;;) {
3246 ch = *cp;
3247 if ( !dot && ch == '.' ) {
3248 scale = 1;
3249 dot = 1;
3250 } else if ( ch == '%' ) {
3251 scale = dot ? scale*100 : 100;
3252 cp++; /* % is always at the end */
3253 break;
3254 } else if ( ch >= '0' && ch <= '9' ) {
3255 if ( scale < 100000 ) {
3256 scale *= 10;
3257 num = (num*10) + (ch-'0');
3258 }
3259 } else {
3260 break;
3261 }
3262 cp++;
3263 }
3264 *cp_p = cp;
3265
3266 /* user says num divided by scale and we say internally that
3267 * is MAX_SCORE * num / scale.
3268 */
3269 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3270 }
3271
3272 static int diff_scoreopt_parse(const char *opt)
3273 {
3274 int opt1, opt2, cmd;
3275
3276 if (*opt++ != '-')
3277 return -1;
3278 cmd = *opt++;
3279 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3280 return -1; /* that is not a -M, -C nor -B option */
3281
3282 opt1 = parse_num(&opt);
3283 if (cmd != 'B')
3284 opt2 = 0;
3285 else {
3286 if (*opt == 0)
3287 opt2 = 0;
3288 else if (*opt != '/')
3289 return -1; /* we expect -B80/99 or -B80 */
3290 else {
3291 opt++;
3292 opt2 = parse_num(&opt);
3293 }
3294 }
3295 if (*opt != 0)
3296 return -1;
3297 return opt1 | (opt2 << 16);
3298 }
3299
3300 struct diff_queue_struct diff_queued_diff;
3301
3302 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3303 {
3304 if (queue->alloc <= queue->nr) {
3305 queue->alloc = alloc_nr(queue->alloc);
3306 queue->queue = xrealloc(queue->queue,
3307 sizeof(dp) * queue->alloc);
3308 }
3309 queue->queue[queue->nr++] = dp;
3310 }
3311
3312 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3313 struct diff_filespec *one,
3314 struct diff_filespec *two)
3315 {
3316 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3317 dp->one = one;
3318 dp->two = two;
3319 if (queue)
3320 diff_q(queue, dp);
3321 return dp;
3322 }
3323
3324 void diff_free_filepair(struct diff_filepair *p)
3325 {
3326 free_filespec(p->one);
3327 free_filespec(p->two);
3328 free(p);
3329 }
3330
3331 /* This is different from find_unique_abbrev() in that
3332 * it stuffs the result with dots for alignment.
3333 */
3334 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3335 {
3336 int abblen;
3337 const char *abbrev;
3338 if (len == 40)
3339 return sha1_to_hex(sha1);
3340
3341 abbrev = find_unique_abbrev(sha1, len);
3342 abblen = strlen(abbrev);
3343 if (abblen < 37) {
3344 static char hex[41];
3345 if (len < abblen && abblen <= len + 2)
3346 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3347 else
3348 sprintf(hex, "%s...", abbrev);
3349 return hex;
3350 }
3351 return sha1_to_hex(sha1);
3352 }
3353
3354 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3355 {
3356 int line_termination = opt->line_termination;
3357 int inter_name_termination = line_termination ? '\t' : '\0';
3358 if (opt->output_prefix) {
3359 struct strbuf *msg = NULL;
3360 msg = opt->output_prefix(opt, opt->output_prefix_data);
3361 fprintf(opt->file, "%s", msg->buf);
3362 }
3363
3364 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3365 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3366 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3367 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3368 }
3369 if (p->score) {
3370 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3371 inter_name_termination);
3372 } else {
3373 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3374 }
3375
3376 if (p->status == DIFF_STATUS_COPIED ||
3377 p->status == DIFF_STATUS_RENAMED) {
3378 const char *name_a, *name_b;
3379 name_a = p->one->path;
3380 name_b = p->two->path;
3381 strip_prefix(opt->prefix_length, &name_a, &name_b);
3382 write_name_quoted(name_a, opt->file, inter_name_termination);
3383 write_name_quoted(name_b, opt->file, line_termination);
3384 } else {
3385 const char *name_a, *name_b;
3386 name_a = p->one->mode ? p->one->path : p->two->path;
3387 name_b = NULL;
3388 strip_prefix(opt->prefix_length, &name_a, &name_b);
3389 write_name_quoted(name_a, opt->file, line_termination);
3390 }
3391 }
3392
3393 int diff_unmodified_pair(struct diff_filepair *p)
3394 {
3395 /* This function is written stricter than necessary to support
3396 * the currently implemented transformers, but the idea is to
3397 * let transformers to produce diff_filepairs any way they want,
3398 * and filter and clean them up here before producing the output.
3399 */
3400 struct diff_filespec *one = p->one, *two = p->two;
3401
3402 if (DIFF_PAIR_UNMERGED(p))
3403 return 0; /* unmerged is interesting */
3404
3405 /* deletion, addition, mode or type change
3406 * and rename are all interesting.
3407 */
3408 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3409 DIFF_PAIR_MODE_CHANGED(p) ||
3410 strcmp(one->path, two->path))
3411 return 0;
3412
3413 /* both are valid and point at the same path. that is, we are
3414 * dealing with a change.
3415 */
3416 if (one->sha1_valid && two->sha1_valid &&
3417 !hashcmp(one->sha1, two->sha1) &&
3418 !one->dirty_submodule && !two->dirty_submodule)
3419 return 1; /* no change */
3420 if (!one->sha1_valid && !two->sha1_valid)
3421 return 1; /* both look at the same file on the filesystem. */
3422 return 0;
3423 }
3424
3425 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3426 {
3427 if (diff_unmodified_pair(p))
3428 return;
3429
3430 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3431 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3432 return; /* no tree diffs in patch format */
3433
3434 run_diff(p, o);
3435 }
3436
3437 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3438 struct diffstat_t *diffstat)
3439 {
3440 if (diff_unmodified_pair(p))
3441 return;
3442
3443 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3444 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3445 return; /* no tree diffs in patch format */
3446
3447 run_diffstat(p, o, diffstat);
3448 }
3449
3450 static void diff_flush_checkdiff(struct diff_filepair *p,
3451 struct diff_options *o)
3452 {
3453 if (diff_unmodified_pair(p))
3454 return;
3455
3456 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3457 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3458 return; /* no tree diffs in patch format */
3459
3460 run_checkdiff(p, o);
3461 }
3462
3463 int diff_queue_is_empty(void)
3464 {
3465 struct diff_queue_struct *q = &diff_queued_diff;
3466 int i;
3467 for (i = 0; i < q->nr; i++)
3468 if (!diff_unmodified_pair(q->queue[i]))
3469 return 0;
3470 return 1;
3471 }
3472
3473 #if DIFF_DEBUG
3474 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3475 {
3476 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3477 x, one ? one : "",
3478 s->path,
3479 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3480 s->mode,
3481 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3482 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3483 x, one ? one : "",
3484 s->size, s->xfrm_flags);
3485 }
3486
3487 void diff_debug_filepair(const struct diff_filepair *p, int i)
3488 {
3489 diff_debug_filespec(p->one, i, "one");
3490 diff_debug_filespec(p->two, i, "two");
3491 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3492 p->score, p->status ? p->status : '?',
3493 p->one->rename_used, p->broken_pair);
3494 }
3495
3496 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3497 {
3498 int i;
3499 if (msg)
3500 fprintf(stderr, "%s\n", msg);
3501 fprintf(stderr, "q->nr = %d\n", q->nr);
3502 for (i = 0; i < q->nr; i++) {
3503 struct diff_filepair *p = q->queue[i];
3504 diff_debug_filepair(p, i);
3505 }
3506 }
3507 #endif
3508
3509 static void diff_resolve_rename_copy(void)
3510 {
3511 int i;
3512 struct diff_filepair *p;
3513 struct diff_queue_struct *q = &diff_queued_diff;
3514
3515 diff_debug_queue("resolve-rename-copy", q);
3516
3517 for (i = 0; i < q->nr; i++) {
3518 p = q->queue[i];
3519 p->status = 0; /* undecided */
3520 if (DIFF_PAIR_UNMERGED(p))
3521 p->status = DIFF_STATUS_UNMERGED;
3522 else if (!DIFF_FILE_VALID(p->one))
3523 p->status = DIFF_STATUS_ADDED;
3524 else if (!DIFF_FILE_VALID(p->two))
3525 p->status = DIFF_STATUS_DELETED;
3526 else if (DIFF_PAIR_TYPE_CHANGED(p))
3527 p->status = DIFF_STATUS_TYPE_CHANGED;
3528
3529 /* from this point on, we are dealing with a pair
3530 * whose both sides are valid and of the same type, i.e.
3531 * either in-place edit or rename/copy edit.
3532 */
3533 else if (DIFF_PAIR_RENAME(p)) {
3534 /*
3535 * A rename might have re-connected a broken
3536 * pair up, causing the pathnames to be the
3537 * same again. If so, that's not a rename at
3538 * all, just a modification..
3539 *
3540 * Otherwise, see if this source was used for
3541 * multiple renames, in which case we decrement
3542 * the count, and call it a copy.
3543 */
3544 if (!strcmp(p->one->path, p->two->path))
3545 p->status = DIFF_STATUS_MODIFIED;
3546 else if (--p->one->rename_used > 0)
3547 p->status = DIFF_STATUS_COPIED;
3548 else
3549 p->status = DIFF_STATUS_RENAMED;
3550 }
3551 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3552 p->one->mode != p->two->mode ||
3553 p->one->dirty_submodule ||
3554 p->two->dirty_submodule ||
3555 is_null_sha1(p->one->sha1))
3556 p->status = DIFF_STATUS_MODIFIED;
3557 else {
3558 /* This is a "no-change" entry and should not
3559 * happen anymore, but prepare for broken callers.
3560 */
3561 error("feeding unmodified %s to diffcore",
3562 p->one->path);
3563 p->status = DIFF_STATUS_UNKNOWN;
3564 }
3565 }
3566 diff_debug_queue("resolve-rename-copy done", q);
3567 }
3568
3569 static int check_pair_status(struct diff_filepair *p)
3570 {
3571 switch (p->status) {
3572 case DIFF_STATUS_UNKNOWN:
3573 return 0;
3574 case 0:
3575 die("internal error in diff-resolve-rename-copy");
3576 default:
3577 return 1;
3578 }
3579 }
3580
3581 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3582 {
3583 int fmt = opt->output_format;
3584
3585 if (fmt & DIFF_FORMAT_CHECKDIFF)
3586 diff_flush_checkdiff(p, opt);
3587 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3588 diff_flush_raw(p, opt);
3589 else if (fmt & DIFF_FORMAT_NAME) {
3590 const char *name_a, *name_b;
3591 name_a = p->two->path;
3592 name_b = NULL;
3593 strip_prefix(opt->prefix_length, &name_a, &name_b);
3594 write_name_quoted(name_a, opt->file, opt->line_termination);
3595 }
3596 }
3597
3598 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3599 {
3600 if (fs->mode)
3601 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3602 else
3603 fprintf(file, " %s ", newdelete);
3604 write_name_quoted(fs->path, file, '\n');
3605 }
3606
3607
3608 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3609 const char *line_prefix)
3610 {
3611 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3612 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3613 p->two->mode, show_name ? ' ' : '\n');
3614 if (show_name) {
3615 write_name_quoted(p->two->path, file, '\n');
3616 }
3617 }
3618 }
3619
3620 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3621 const char *line_prefix)
3622 {
3623 char *names = pprint_rename(p->one->path, p->two->path);
3624
3625 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3626 free(names);
3627 show_mode_change(file, p, 0, line_prefix);
3628 }
3629
3630 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3631 {
3632 FILE *file = opt->file;
3633 char *line_prefix = "";
3634
3635 if (opt->output_prefix) {
3636 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3637 line_prefix = buf->buf;
3638 }
3639
3640 switch(p->status) {
3641 case DIFF_STATUS_DELETED:
3642 fputs(line_prefix, file);
3643 show_file_mode_name(file, "delete", p->one);
3644 break;
3645 case DIFF_STATUS_ADDED:
3646 fputs(line_prefix, file);
3647 show_file_mode_name(file, "create", p->two);
3648 break;
3649 case DIFF_STATUS_COPIED:
3650 fputs(line_prefix, file);
3651 show_rename_copy(file, "copy", p, line_prefix);
3652 break;
3653 case DIFF_STATUS_RENAMED:
3654 fputs(line_prefix, file);
3655 show_rename_copy(file, "rename", p, line_prefix);
3656 break;
3657 default:
3658 if (p->score) {
3659 fprintf(file, "%s rewrite ", line_prefix);
3660 write_name_quoted(p->two->path, file, ' ');
3661 fprintf(file, "(%d%%)\n", similarity_index(p));
3662 }
3663 show_mode_change(file, p, !p->score, line_prefix);
3664 break;
3665 }
3666 }
3667
3668 struct patch_id_t {
3669 git_SHA_CTX *ctx;
3670 int patchlen;
3671 };
3672
3673 static int remove_space(char *line, int len)
3674 {
3675 int i;
3676 char *dst = line;
3677 unsigned char c;
3678
3679 for (i = 0; i < len; i++)
3680 if (!isspace((c = line[i])))
3681 *dst++ = c;
3682
3683 return dst - line;
3684 }
3685
3686 static void patch_id_consume(void *priv, char *line, unsigned long len)
3687 {
3688 struct patch_id_t *data = priv;
3689 int new_len;
3690
3691 /* Ignore line numbers when computing the SHA1 of the patch */
3692 if (!prefixcmp(line, "@@ -"))
3693 return;
3694
3695 new_len = remove_space(line, len);
3696
3697 git_SHA1_Update(data->ctx, line, new_len);
3698 data->patchlen += new_len;
3699 }
3700
3701 /* returns 0 upon success, and writes result into sha1 */
3702 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3703 {
3704 struct diff_queue_struct *q = &diff_queued_diff;
3705 int i;
3706 git_SHA_CTX ctx;
3707 struct patch_id_t data;
3708 char buffer[PATH_MAX * 4 + 20];
3709
3710 git_SHA1_Init(&ctx);
3711 memset(&data, 0, sizeof(struct patch_id_t));
3712 data.ctx = &ctx;
3713
3714 for (i = 0; i < q->nr; i++) {
3715 xpparam_t xpp;
3716 xdemitconf_t xecfg;
3717 mmfile_t mf1, mf2;
3718 struct diff_filepair *p = q->queue[i];
3719 int len1, len2;
3720
3721 memset(&xpp, 0, sizeof(xpp));
3722 memset(&xecfg, 0, sizeof(xecfg));
3723 if (p->status == 0)
3724 return error("internal diff status error");
3725 if (p->status == DIFF_STATUS_UNKNOWN)
3726 continue;
3727 if (diff_unmodified_pair(p))
3728 continue;
3729 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3730 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3731 continue;
3732 if (DIFF_PAIR_UNMERGED(p))
3733 continue;
3734
3735 diff_fill_sha1_info(p->one);
3736 diff_fill_sha1_info(p->two);
3737 if (fill_mmfile(&mf1, p->one) < 0 ||
3738 fill_mmfile(&mf2, p->two) < 0)
3739 return error("unable to read files to diff");
3740
3741 len1 = remove_space(p->one->path, strlen(p->one->path));
3742 len2 = remove_space(p->two->path, strlen(p->two->path));
3743 if (p->one->mode == 0)
3744 len1 = snprintf(buffer, sizeof(buffer),
3745 "diff--gita/%.*sb/%.*s"
3746 "newfilemode%06o"
3747 "---/dev/null"
3748 "+++b/%.*s",
3749 len1, p->one->path,
3750 len2, p->two->path,
3751 p->two->mode,
3752 len2, p->two->path);
3753 else if (p->two->mode == 0)
3754 len1 = snprintf(buffer, sizeof(buffer),
3755 "diff--gita/%.*sb/%.*s"
3756 "deletedfilemode%06o"
3757 "---a/%.*s"
3758 "+++/dev/null",
3759 len1, p->one->path,
3760 len2, p->two->path,
3761 p->one->mode,
3762 len1, p->one->path);
3763 else
3764 len1 = snprintf(buffer, sizeof(buffer),
3765 "diff--gita/%.*sb/%.*s"
3766 "---a/%.*s"
3767 "+++b/%.*s",
3768 len1, p->one->path,
3769 len2, p->two->path,
3770 len1, p->one->path,
3771 len2, p->two->path);
3772 git_SHA1_Update(&ctx, buffer, len1);
3773
3774 xpp.flags = 0;
3775 xecfg.ctxlen = 3;
3776 xecfg.flags = XDL_EMIT_FUNCNAMES;
3777 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3778 &xpp, &xecfg);
3779 }
3780
3781 git_SHA1_Final(sha1, &ctx);
3782 return 0;
3783 }
3784
3785 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3786 {
3787 struct diff_queue_struct *q = &diff_queued_diff;
3788 int i;
3789 int result = diff_get_patch_id(options, sha1);
3790
3791 for (i = 0; i < q->nr; i++)
3792 diff_free_filepair(q->queue[i]);
3793
3794 free(q->queue);
3795 DIFF_QUEUE_CLEAR(q);
3796
3797 return result;
3798 }
3799
3800 static int is_summary_empty(const struct diff_queue_struct *q)
3801 {
3802 int i;
3803
3804 for (i = 0; i < q->nr; i++) {
3805 const struct diff_filepair *p = q->queue[i];
3806
3807 switch (p->status) {
3808 case DIFF_STATUS_DELETED:
3809 case DIFF_STATUS_ADDED:
3810 case DIFF_STATUS_COPIED:
3811 case DIFF_STATUS_RENAMED:
3812 return 0;
3813 default:
3814 if (p->score)
3815 return 0;
3816 if (p->one->mode && p->two->mode &&
3817 p->one->mode != p->two->mode)
3818 return 0;
3819 break;
3820 }
3821 }
3822 return 1;
3823 }
3824
3825 void diff_flush(struct diff_options *options)
3826 {
3827 struct diff_queue_struct *q = &diff_queued_diff;
3828 int i, output_format = options->output_format;
3829 int separator = 0;
3830
3831 /*
3832 * Order: raw, stat, summary, patch
3833 * or: name/name-status/checkdiff (other bits clear)
3834 */
3835 if (!q->nr)
3836 goto free_queue;
3837
3838 if (output_format & (DIFF_FORMAT_RAW |
3839 DIFF_FORMAT_NAME |
3840 DIFF_FORMAT_NAME_STATUS |
3841 DIFF_FORMAT_CHECKDIFF)) {
3842 for (i = 0; i < q->nr; i++) {
3843 struct diff_filepair *p = q->queue[i];
3844 if (check_pair_status(p))
3845 flush_one_pair(p, options);
3846 }
3847 separator++;
3848 }
3849
3850 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
3851 struct diffstat_t diffstat;
3852
3853 memset(&diffstat, 0, sizeof(struct diffstat_t));
3854 for (i = 0; i < q->nr; i++) {
3855 struct diff_filepair *p = q->queue[i];
3856 if (check_pair_status(p))
3857 diff_flush_stat(p, options, &diffstat);
3858 }
3859 if (output_format & DIFF_FORMAT_NUMSTAT)
3860 show_numstat(&diffstat, options);
3861 if (output_format & DIFF_FORMAT_DIFFSTAT)
3862 show_stats(&diffstat, options);
3863 if (output_format & DIFF_FORMAT_SHORTSTAT)
3864 show_shortstats(&diffstat, options);
3865 free_diffstat_info(&diffstat);
3866 separator++;
3867 }
3868 if (output_format & DIFF_FORMAT_DIRSTAT)
3869 show_dirstat(options);
3870
3871 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
3872 for (i = 0; i < q->nr; i++) {
3873 diff_summary(options, q->queue[i]);
3874 }
3875 separator++;
3876 }
3877
3878 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
3879 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
3880 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
3881 /*
3882 * run diff_flush_patch for the exit status. setting
3883 * options->file to /dev/null should be safe, becaue we
3884 * aren't supposed to produce any output anyway.
3885 */
3886 if (options->close_file)
3887 fclose(options->file);
3888 options->file = fopen("/dev/null", "w");
3889 if (!options->file)
3890 die_errno("Could not open /dev/null");
3891 options->close_file = 1;
3892 for (i = 0; i < q->nr; i++) {
3893 struct diff_filepair *p = q->queue[i];
3894 if (check_pair_status(p))
3895 diff_flush_patch(p, options);
3896 if (options->found_changes)
3897 break;
3898 }
3899 }
3900
3901 if (output_format & DIFF_FORMAT_PATCH) {
3902 if (separator) {
3903 putc(options->line_termination, options->file);
3904 if (options->stat_sep) {
3905 /* attach patch instead of inline */
3906 fputs(options->stat_sep, options->file);
3907 }
3908 }
3909
3910 for (i = 0; i < q->nr; i++) {
3911 struct diff_filepair *p = q->queue[i];
3912 if (check_pair_status(p))
3913 diff_flush_patch(p, options);
3914 }
3915 }
3916
3917 if (output_format & DIFF_FORMAT_CALLBACK)
3918 options->format_callback(q, options, options->format_callback_data);
3919
3920 for (i = 0; i < q->nr; i++)
3921 diff_free_filepair(q->queue[i]);
3922 free_queue:
3923 free(q->queue);
3924 DIFF_QUEUE_CLEAR(q);
3925 if (options->close_file)
3926 fclose(options->file);
3927
3928 /*
3929 * Report the content-level differences with HAS_CHANGES;
3930 * diff_addremove/diff_change does not set the bit when
3931 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
3932 */
3933 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
3934 if (options->found_changes)
3935 DIFF_OPT_SET(options, HAS_CHANGES);
3936 else
3937 DIFF_OPT_CLR(options, HAS_CHANGES);
3938 }
3939 }
3940
3941 static void diffcore_apply_filter(const char *filter)
3942 {
3943 int i;
3944 struct diff_queue_struct *q = &diff_queued_diff;
3945 struct diff_queue_struct outq;
3946 DIFF_QUEUE_CLEAR(&outq);
3947
3948 if (!filter)
3949 return;
3950
3951 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
3952 int found;
3953 for (i = found = 0; !found && i < q->nr; i++) {
3954 struct diff_filepair *p = q->queue[i];
3955 if (((p->status == DIFF_STATUS_MODIFIED) &&
3956 ((p->score &&
3957 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
3958 (!p->score &&
3959 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
3960 ((p->status != DIFF_STATUS_MODIFIED) &&
3961 strchr(filter, p->status)))
3962 found++;
3963 }
3964 if (found)
3965 return;
3966
3967 /* otherwise we will clear the whole queue
3968 * by copying the empty outq at the end of this
3969 * function, but first clear the current entries
3970 * in the queue.
3971 */
3972 for (i = 0; i < q->nr; i++)
3973 diff_free_filepair(q->queue[i]);
3974 }
3975 else {
3976 /* Only the matching ones */
3977 for (i = 0; i < q->nr; i++) {
3978 struct diff_filepair *p = q->queue[i];
3979
3980 if (((p->status == DIFF_STATUS_MODIFIED) &&
3981 ((p->score &&
3982 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
3983 (!p->score &&
3984 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
3985 ((p->status != DIFF_STATUS_MODIFIED) &&
3986 strchr(filter, p->status)))
3987 diff_q(&outq, p);
3988 else
3989 diff_free_filepair(p);
3990 }
3991 }
3992 free(q->queue);
3993 *q = outq;
3994 }
3995
3996 /* Check whether two filespecs with the same mode and size are identical */
3997 static int diff_filespec_is_identical(struct diff_filespec *one,
3998 struct diff_filespec *two)
3999 {
4000 if (S_ISGITLINK(one->mode))
4001 return 0;
4002 if (diff_populate_filespec(one, 0))
4003 return 0;
4004 if (diff_populate_filespec(two, 0))
4005 return 0;
4006 return !memcmp(one->data, two->data, one->size);
4007 }
4008
4009 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4010 {
4011 int i;
4012 struct diff_queue_struct *q = &diff_queued_diff;
4013 struct diff_queue_struct outq;
4014 DIFF_QUEUE_CLEAR(&outq);
4015
4016 for (i = 0; i < q->nr; i++) {
4017 struct diff_filepair *p = q->queue[i];
4018
4019 /*
4020 * 1. Entries that come from stat info dirtiness
4021 * always have both sides (iow, not create/delete),
4022 * one side of the object name is unknown, with
4023 * the same mode and size. Keep the ones that
4024 * do not match these criteria. They have real
4025 * differences.
4026 *
4027 * 2. At this point, the file is known to be modified,
4028 * with the same mode and size, and the object
4029 * name of one side is unknown. Need to inspect
4030 * the identical contents.
4031 */
4032 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4033 !DIFF_FILE_VALID(p->two) ||
4034 (p->one->sha1_valid && p->two->sha1_valid) ||
4035 (p->one->mode != p->two->mode) ||
4036 diff_populate_filespec(p->one, 1) ||
4037 diff_populate_filespec(p->two, 1) ||
4038 (p->one->size != p->two->size) ||
4039 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4040 diff_q(&outq, p);
4041 else {
4042 /*
4043 * The caller can subtract 1 from skip_stat_unmatch
4044 * to determine how many paths were dirty only
4045 * due to stat info mismatch.
4046 */
4047 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4048 diffopt->skip_stat_unmatch++;
4049 diff_free_filepair(p);
4050 }
4051 }
4052 free(q->queue);
4053 *q = outq;
4054 }
4055
4056 static int diffnamecmp(const void *a_, const void *b_)
4057 {
4058 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4059 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4060 const char *name_a, *name_b;
4061
4062 name_a = a->one ? a->one->path : a->two->path;
4063 name_b = b->one ? b->one->path : b->two->path;
4064 return strcmp(name_a, name_b);
4065 }
4066
4067 void diffcore_fix_diff_index(struct diff_options *options)
4068 {
4069 struct diff_queue_struct *q = &diff_queued_diff;
4070 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4071 }
4072
4073 void diffcore_std(struct diff_options *options)
4074 {
4075 if (options->skip_stat_unmatch)
4076 diffcore_skip_stat_unmatch(options);
4077 if (!options->found_follow) {
4078 /* See try_to_follow_renames() in tree-diff.c */
4079 if (options->break_opt != -1)
4080 diffcore_break(options->break_opt);
4081 if (options->detect_rename)
4082 diffcore_rename(options);
4083 if (options->break_opt != -1)
4084 diffcore_merge_broken();
4085 }
4086 if (options->pickaxe)
4087 diffcore_pickaxe(options->pickaxe, options->pickaxe_opts);
4088 if (options->orderfile)
4089 diffcore_order(options->orderfile);
4090 if (!options->found_follow)
4091 /* See try_to_follow_renames() in tree-diff.c */
4092 diff_resolve_rename_copy();
4093 diffcore_apply_filter(options->filter);
4094
4095 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4096 DIFF_OPT_SET(options, HAS_CHANGES);
4097 else
4098 DIFF_OPT_CLR(options, HAS_CHANGES);
4099
4100 options->found_follow = 0;
4101 }
4102
4103 int diff_result_code(struct diff_options *opt, int status)
4104 {
4105 int result = 0;
4106 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4107 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4108 return status;
4109 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4110 DIFF_OPT_TST(opt, HAS_CHANGES))
4111 result |= 01;
4112 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4113 DIFF_OPT_TST(opt, CHECK_FAILED))
4114 result |= 02;
4115 return result;
4116 }
4117
4118 void diff_addremove(struct diff_options *options,
4119 int addremove, unsigned mode,
4120 const unsigned char *sha1,
4121 const char *concatpath, unsigned dirty_submodule)
4122 {
4123 struct diff_filespec *one, *two;
4124
4125 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES) && S_ISGITLINK(mode))
4126 return;
4127
4128 /* This may look odd, but it is a preparation for
4129 * feeding "there are unchanged files which should
4130 * not produce diffs, but when you are doing copy
4131 * detection you would need them, so here they are"
4132 * entries to the diff-core. They will be prefixed
4133 * with something like '=' or '*' (I haven't decided
4134 * which but should not make any difference).
4135 * Feeding the same new and old to diff_change()
4136 * also has the same effect.
4137 * Before the final output happens, they are pruned after
4138 * merged into rename/copy pairs as appropriate.
4139 */
4140 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4141 addremove = (addremove == '+' ? '-' :
4142 addremove == '-' ? '+' : addremove);
4143
4144 if (options->prefix &&
4145 strncmp(concatpath, options->prefix, options->prefix_length))
4146 return;
4147
4148 one = alloc_filespec(concatpath);
4149 two = alloc_filespec(concatpath);
4150
4151 if (addremove != '+')
4152 fill_filespec(one, sha1, mode);
4153 if (addremove != '-') {
4154 fill_filespec(two, sha1, mode);
4155 two->dirty_submodule = dirty_submodule;
4156 }
4157
4158 diff_queue(&diff_queued_diff, one, two);
4159 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4160 DIFF_OPT_SET(options, HAS_CHANGES);
4161 }
4162
4163 void diff_change(struct diff_options *options,
4164 unsigned old_mode, unsigned new_mode,
4165 const unsigned char *old_sha1,
4166 const unsigned char *new_sha1,
4167 const char *concatpath,
4168 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4169 {
4170 struct diff_filespec *one, *two;
4171
4172 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES) && S_ISGITLINK(old_mode)
4173 && S_ISGITLINK(new_mode))
4174 return;
4175
4176 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4177 unsigned tmp;
4178 const unsigned char *tmp_c;
4179 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4180 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4181 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4182 new_dirty_submodule = tmp;
4183 }
4184
4185 if (options->prefix &&
4186 strncmp(concatpath, options->prefix, options->prefix_length))
4187 return;
4188
4189 one = alloc_filespec(concatpath);
4190 two = alloc_filespec(concatpath);
4191 fill_filespec(one, old_sha1, old_mode);
4192 fill_filespec(two, new_sha1, new_mode);
4193 one->dirty_submodule = old_dirty_submodule;
4194 two->dirty_submodule = new_dirty_submodule;
4195
4196 diff_queue(&diff_queued_diff, one, two);
4197 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4198 DIFF_OPT_SET(options, HAS_CHANGES);
4199 }
4200
4201 void diff_unmerge(struct diff_options *options,
4202 const char *path,
4203 unsigned mode, const unsigned char *sha1)
4204 {
4205 struct diff_filespec *one, *two;
4206
4207 if (options->prefix &&
4208 strncmp(path, options->prefix, options->prefix_length))
4209 return;
4210
4211 one = alloc_filespec(path);
4212 two = alloc_filespec(path);
4213 fill_filespec(one, sha1, mode);
4214 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4215 }
4216
4217 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4218 size_t *outsize)
4219 {
4220 struct diff_tempfile *temp;
4221 const char *argv[3];
4222 const char **arg = argv;
4223 struct child_process child;
4224 struct strbuf buf = STRBUF_INIT;
4225 int err = 0;
4226
4227 temp = prepare_temp_file(spec->path, spec);
4228 *arg++ = pgm;
4229 *arg++ = temp->name;
4230 *arg = NULL;
4231
4232 memset(&child, 0, sizeof(child));
4233 child.use_shell = 1;
4234 child.argv = argv;
4235 child.out = -1;
4236 if (start_command(&child)) {
4237 remove_tempfile();
4238 return NULL;
4239 }
4240
4241 if (strbuf_read(&buf, child.out, 0) < 0)
4242 err = error("error reading from textconv command '%s'", pgm);
4243 close(child.out);
4244
4245 if (finish_command(&child) || err) {
4246 strbuf_release(&buf);
4247 remove_tempfile();
4248 return NULL;
4249 }
4250 remove_tempfile();
4251
4252 return strbuf_detach(&buf, outsize);
4253 }
4254
4255 size_t fill_textconv(struct userdiff_driver *driver,
4256 struct diff_filespec *df,
4257 char **outbuf)
4258 {
4259 size_t size;
4260
4261 if (!driver || !driver->textconv) {
4262 if (!DIFF_FILE_VALID(df)) {
4263 *outbuf = "";
4264 return 0;
4265 }
4266 if (diff_populate_filespec(df, 0))
4267 die("unable to read files to diff");
4268 *outbuf = df->data;
4269 return df->size;
4270 }
4271
4272 if (driver->textconv_cache) {
4273 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4274 &size);
4275 if (*outbuf)
4276 return size;
4277 }
4278
4279 *outbuf = run_textconv(driver->textconv, df, &size);
4280 if (!*outbuf)
4281 die("unable to read files to diff");
4282
4283 if (driver->textconv_cache) {
4284 /* ignore errors, as we might be in a readonly repository */
4285 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4286 size);
4287 /*
4288 * we could save up changes and flush them all at the end,
4289 * but we would need an extra call after all diffing is done.
4290 * Since generating a cache entry is the slow path anyway,
4291 * this extra overhead probably isn't a big deal.
4292 */
4293 notes_cache_write(driver->textconv_cache);
4294 }
4295
4296 return size;
4297 }
knittl's private git fork