?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Refactor --dirstat parsing; deprecate --cumulative and --dirstat-by-file
[git/mingw.git] / diff.c
blob3b0cda43f16ccf844b711e10f487b1763f0f88ab
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 = 400;
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 static struct diff_options default_diff_options;
35
36 static char diff_colors[][COLOR_MAXLEN] = {
37 GIT_COLOR_RESET,
38 GIT_COLOR_NORMAL, /* PLAIN */
39 GIT_COLOR_BOLD, /* METAINFO */
40 GIT_COLOR_CYAN, /* FRAGINFO */
41 GIT_COLOR_RED, /* OLD */
42 GIT_COLOR_GREEN, /* NEW */
43 GIT_COLOR_YELLOW, /* COMMIT */
44 GIT_COLOR_BG_RED, /* WHITESPACE */
45 GIT_COLOR_NORMAL, /* FUNCINFO */
46 };
47
48 static int parse_diff_color_slot(const char *var, int ofs)
49 {
50 if (!strcasecmp(var+ofs, "plain"))
51 return DIFF_PLAIN;
52 if (!strcasecmp(var+ofs, "meta"))
53 return DIFF_METAINFO;
54 if (!strcasecmp(var+ofs, "frag"))
55 return DIFF_FRAGINFO;
56 if (!strcasecmp(var+ofs, "old"))
57 return DIFF_FILE_OLD;
58 if (!strcasecmp(var+ofs, "new"))
59 return DIFF_FILE_NEW;
60 if (!strcasecmp(var+ofs, "commit"))
61 return DIFF_COMMIT;
62 if (!strcasecmp(var+ofs, "whitespace"))
63 return DIFF_WHITESPACE;
64 if (!strcasecmp(var+ofs, "func"))
65 return DIFF_FUNCINFO;
66 return -1;
67 }
68
69 static int parse_dirstat_params(struct diff_options *options, const char *params)
70 {
71 const char *p = params;
72 while (*p) {
73 if (!prefixcmp(p, "changes")) {
74 p += 7;
75 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
76 } else if (!prefixcmp(p, "files")) {
77 p += 5;
78 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
79 } else if (!prefixcmp(p, "noncumulative")) {
80 p += 13;
81 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
82 } else if (!prefixcmp(p, "cumulative")) {
83 p += 10;
84 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
85 } else if (isdigit(*p)) {
86 char *end;
87 options->dirstat_percent = strtoul(p, &end, 10);
88 p = end;
89 } else
90 return error("Unknown --dirstat parameter '%s'", p);
91
92 if (*p) {
93 /* more parameters, swallow separator */
94 if (*p != ',')
95 return error("Missing comma separator at char "
96 "%"PRIuMAX" of '%s'",
97 (uintmax_t) (p - params), params);
98 p++;
99 }
100 }
101 return 0;
102 }
103
104 static int git_config_rename(const char *var, const char *value)
105 {
106 if (!value)
107 return DIFF_DETECT_RENAME;
108 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
109 return DIFF_DETECT_COPY;
110 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
111 }
112
113 /*
114 * These are to give UI layer defaults.
115 * The core-level commands such as git-diff-files should
116 * never be affected by the setting of diff.renames
117 * the user happens to have in the configuration file.
118 */
119 int git_diff_ui_config(const char *var, const char *value, void *cb)
120 {
121 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
122 diff_use_color_default = git_config_colorbool(var, value, -1);
123 return 0;
124 }
125 if (!strcmp(var, "diff.renames")) {
126 diff_detect_rename_default = git_config_rename(var, value);
127 return 0;
128 }
129 if (!strcmp(var, "diff.autorefreshindex")) {
130 diff_auto_refresh_index = git_config_bool(var, value);
131 return 0;
132 }
133 if (!strcmp(var, "diff.mnemonicprefix")) {
134 diff_mnemonic_prefix = git_config_bool(var, value);
135 return 0;
136 }
137 if (!strcmp(var, "diff.noprefix")) {
138 diff_no_prefix = git_config_bool(var, value);
139 return 0;
140 }
141 if (!strcmp(var, "diff.external"))
142 return git_config_string(&external_diff_cmd_cfg, var, value);
143 if (!strcmp(var, "diff.wordregex"))
144 return git_config_string(&diff_word_regex_cfg, var, value);
145
146 if (!strcmp(var, "diff.ignoresubmodules"))
147 handle_ignore_submodules_arg(&default_diff_options, value);
148
149 return git_diff_basic_config(var, value, cb);
150 }
151
152 int git_diff_basic_config(const char *var, const char *value, void *cb)
153 {
154 if (!strcmp(var, "diff.renamelimit")) {
155 diff_rename_limit_default = git_config_int(var, value);
156 return 0;
157 }
158
159 switch (userdiff_config(var, value)) {
160 case 0: break;
161 case -1: return -1;
162 default: return 0;
163 }
164
165 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
166 int slot = parse_diff_color_slot(var, 11);
167 if (slot < 0)
168 return 0;
169 if (!value)
170 return config_error_nonbool(var);
171 color_parse(value, var, diff_colors[slot]);
172 return 0;
173 }
174
175 /* like GNU diff's --suppress-blank-empty option */
176 if (!strcmp(var, "diff.suppressblankempty") ||
177 /* for backwards compatibility */
178 !strcmp(var, "diff.suppress-blank-empty")) {
179 diff_suppress_blank_empty = git_config_bool(var, value);
180 return 0;
181 }
182
183 if (!prefixcmp(var, "submodule."))
184 return parse_submodule_config_option(var, value);
185
186 return git_color_default_config(var, value, cb);
187 }
188
189 static char *quote_two(const char *one, const char *two)
190 {
191 int need_one = quote_c_style(one, NULL, NULL, 1);
192 int need_two = quote_c_style(two, NULL, NULL, 1);
193 struct strbuf res = STRBUF_INIT;
194
195 if (need_one + need_two) {
196 strbuf_addch(&res, '"');
197 quote_c_style(one, &res, NULL, 1);
198 quote_c_style(two, &res, NULL, 1);
199 strbuf_addch(&res, '"');
200 } else {
201 strbuf_addstr(&res, one);
202 strbuf_addstr(&res, two);
203 }
204 return strbuf_detach(&res, NULL);
205 }
206
207 static const char *external_diff(void)
208 {
209 static const char *external_diff_cmd = NULL;
210 static int done_preparing = 0;
211
212 if (done_preparing)
213 return external_diff_cmd;
214 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
215 if (!external_diff_cmd)
216 external_diff_cmd = external_diff_cmd_cfg;
217 done_preparing = 1;
218 return external_diff_cmd;
219 }
220
221 static struct diff_tempfile {
222 const char *name; /* filename external diff should read from */
223 char hex[41];
224 char mode[10];
225 char tmp_path[PATH_MAX];
226 } diff_temp[2];
227
228 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
229
230 struct emit_callback {
231 int color_diff;
232 unsigned ws_rule;
233 int blank_at_eof_in_preimage;
234 int blank_at_eof_in_postimage;
235 int lno_in_preimage;
236 int lno_in_postimage;
237 sane_truncate_fn truncate;
238 const char **label_path;
239 struct diff_words_data *diff_words;
240 struct diff_options *opt;
241 int *found_changesp;
242 struct strbuf *header;
243 };
244
245 static int count_lines(const char *data, int size)
246 {
247 int count, ch, completely_empty = 1, nl_just_seen = 0;
248 count = 0;
249 while (0 < size--) {
250 ch = *data++;
251 if (ch == '\n') {
252 count++;
253 nl_just_seen = 1;
254 completely_empty = 0;
255 }
256 else {
257 nl_just_seen = 0;
258 completely_empty = 0;
259 }
260 }
261 if (completely_empty)
262 return 0;
263 if (!nl_just_seen)
264 count++; /* no trailing newline */
265 return count;
266 }
267
268 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
269 {
270 if (!DIFF_FILE_VALID(one)) {
271 mf->ptr = (char *)""; /* does not matter */
272 mf->size = 0;
273 return 0;
274 }
275 else if (diff_populate_filespec(one, 0))
276 return -1;
277
278 mf->ptr = one->data;
279 mf->size = one->size;
280 return 0;
281 }
282
283 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
284 static unsigned long diff_filespec_size(struct diff_filespec *one)
285 {
286 if (!DIFF_FILE_VALID(one))
287 return 0;
288 diff_populate_filespec(one, 1);
289 return one->size;
290 }
291
292 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
293 {
294 char *ptr = mf->ptr;
295 long size = mf->size;
296 int cnt = 0;
297
298 if (!size)
299 return cnt;
300 ptr += size - 1; /* pointing at the very end */
301 if (*ptr != '\n')
302 ; /* incomplete line */
303 else
304 ptr--; /* skip the last LF */
305 while (mf->ptr < ptr) {
306 char *prev_eol;
307 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
308 if (*prev_eol == '\n')
309 break;
310 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
311 break;
312 cnt++;
313 ptr = prev_eol - 1;
314 }
315 return cnt;
316 }
317
318 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
319 struct emit_callback *ecbdata)
320 {
321 int l1, l2, at;
322 unsigned ws_rule = ecbdata->ws_rule;
323 l1 = count_trailing_blank(mf1, ws_rule);
324 l2 = count_trailing_blank(mf2, ws_rule);
325 if (l2 <= l1) {
326 ecbdata->blank_at_eof_in_preimage = 0;
327 ecbdata->blank_at_eof_in_postimage = 0;
328 return;
329 }
330 at = count_lines(mf1->ptr, mf1->size);
331 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
332
333 at = count_lines(mf2->ptr, mf2->size);
334 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
335 }
336
337 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
338 int first, const char *line, int len)
339 {
340 int has_trailing_newline, has_trailing_carriage_return;
341 int nofirst;
342 FILE *file = o->file;
343
344 if (o->output_prefix) {
345 struct strbuf *msg = NULL;
346 msg = o->output_prefix(o, o->output_prefix_data);
347 assert(msg);
348 fwrite(msg->buf, msg->len, 1, file);
349 }
350
351 if (len == 0) {
352 has_trailing_newline = (first == '\n');
353 has_trailing_carriage_return = (!has_trailing_newline &&
354 (first == '\r'));
355 nofirst = has_trailing_newline || has_trailing_carriage_return;
356 } else {
357 has_trailing_newline = (len > 0 && line[len-1] == '\n');
358 if (has_trailing_newline)
359 len--;
360 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
361 if (has_trailing_carriage_return)
362 len--;
363 nofirst = 0;
364 }
365
366 if (len || !nofirst) {
367 fputs(set, file);
368 if (!nofirst)
369 fputc(first, file);
370 fwrite(line, len, 1, file);
371 fputs(reset, file);
372 }
373 if (has_trailing_carriage_return)
374 fputc('\r', file);
375 if (has_trailing_newline)
376 fputc('\n', file);
377 }
378
379 static void emit_line(struct diff_options *o, const char *set, const char *reset,
380 const char *line, int len)
381 {
382 emit_line_0(o, set, reset, line[0], line+1, len-1);
383 }
384
385 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
386 {
387 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
388 ecbdata->blank_at_eof_in_preimage &&
389 ecbdata->blank_at_eof_in_postimage &&
390 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
391 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
392 return 0;
393 return ws_blank_line(line, len, ecbdata->ws_rule);
394 }
395
396 static void emit_add_line(const char *reset,
397 struct emit_callback *ecbdata,
398 const char *line, int len)
399 {
400 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
401 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
402
403 if (!*ws)
404 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
405 else if (new_blank_line_at_eof(ecbdata, line, len))
406 /* Blank line at EOF - paint '+' as well */
407 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
408 else {
409 /* Emit just the prefix, then the rest. */
410 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
411 ws_check_emit(line, len, ecbdata->ws_rule,
412 ecbdata->opt->file, set, reset, ws);
413 }
414 }
415
416 static void emit_hunk_header(struct emit_callback *ecbdata,
417 const char *line, int len)
418 {
419 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
420 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
421 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
422 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
423 static const char atat[2] = { '@', '@' };
424 const char *cp, *ep;
425 struct strbuf msgbuf = STRBUF_INIT;
426 int org_len = len;
427 int i = 1;
428
429 /*
430 * As a hunk header must begin with "@@ -<old>, +<new> @@",
431 * it always is at least 10 bytes long.
432 */
433 if (len < 10 ||
434 memcmp(line, atat, 2) ||
435 !(ep = memmem(line + 2, len - 2, atat, 2))) {
436 emit_line(ecbdata->opt, plain, reset, line, len);
437 return;
438 }
439 ep += 2; /* skip over @@ */
440
441 /* The hunk header in fraginfo color */
442 strbuf_add(&msgbuf, frag, strlen(frag));
443 strbuf_add(&msgbuf, line, ep - line);
444 strbuf_add(&msgbuf, reset, strlen(reset));
445
446 /*
447 * trailing "\r\n"
448 */
449 for ( ; i < 3; i++)
450 if (line[len - i] == '\r' || line[len - i] == '\n')
451 len--;
452
453 /* blank before the func header */
454 for (cp = ep; ep - line < len; ep++)
455 if (*ep != ' ' && *ep != '\t')
456 break;
457 if (ep != cp) {
458 strbuf_add(&msgbuf, plain, strlen(plain));
459 strbuf_add(&msgbuf, cp, ep - cp);
460 strbuf_add(&msgbuf, reset, strlen(reset));
461 }
462
463 if (ep < line + len) {
464 strbuf_add(&msgbuf, func, strlen(func));
465 strbuf_add(&msgbuf, ep, line + len - ep);
466 strbuf_add(&msgbuf, reset, strlen(reset));
467 }
468
469 strbuf_add(&msgbuf, line + len, org_len - len);
470 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
471 strbuf_release(&msgbuf);
472 }
473
474 static struct diff_tempfile *claim_diff_tempfile(void) {
475 int i;
476 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
477 if (!diff_temp[i].name)
478 return diff_temp + i;
479 die("BUG: diff is failing to clean up its tempfiles");
480 }
481
482 static int remove_tempfile_installed;
483
484 static void remove_tempfile(void)
485 {
486 int i;
487 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
488 if (diff_temp[i].name == diff_temp[i].tmp_path)
489 unlink_or_warn(diff_temp[i].name);
490 diff_temp[i].name = NULL;
491 }
492 }
493
494 static void remove_tempfile_on_signal(int signo)
495 {
496 remove_tempfile();
497 sigchain_pop(signo);
498 raise(signo);
499 }
500
501 static void print_line_count(FILE *file, int count)
502 {
503 switch (count) {
504 case 0:
505 fprintf(file, "0,0");
506 break;
507 case 1:
508 fprintf(file, "1");
509 break;
510 default:
511 fprintf(file, "1,%d", count);
512 break;
513 }
514 }
515
516 static void emit_rewrite_lines(struct emit_callback *ecb,
517 int prefix, const char *data, int size)
518 {
519 const char *endp = NULL;
520 static const char *nneof = " No newline at end of file\n";
521 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
522 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
523
524 while (0 < size) {
525 int len;
526
527 endp = memchr(data, '\n', size);
528 len = endp ? (endp - data + 1) : size;
529 if (prefix != '+') {
530 ecb->lno_in_preimage++;
531 emit_line_0(ecb->opt, old, reset, '-',
532 data, len);
533 } else {
534 ecb->lno_in_postimage++;
535 emit_add_line(reset, ecb, data, len);
536 }
537 size -= len;
538 data += len;
539 }
540 if (!endp) {
541 const char *plain = diff_get_color(ecb->color_diff,
542 DIFF_PLAIN);
543 emit_line_0(ecb->opt, plain, reset, '\\',
544 nneof, strlen(nneof));
545 }
546 }
547
548 static void emit_rewrite_diff(const char *name_a,
549 const char *name_b,
550 struct diff_filespec *one,
551 struct diff_filespec *two,
552 struct userdiff_driver *textconv_one,
553 struct userdiff_driver *textconv_two,
554 struct diff_options *o)
555 {
556 int lc_a, lc_b;
557 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
558 const char *name_a_tab, *name_b_tab;
559 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
560 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
561 const char *reset = diff_get_color(color_diff, DIFF_RESET);
562 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
563 const char *a_prefix, *b_prefix;
564 char *data_one, *data_two;
565 size_t size_one, size_two;
566 struct emit_callback ecbdata;
567 char *line_prefix = "";
568 struct strbuf *msgbuf;
569
570 if (o && o->output_prefix) {
571 msgbuf = o->output_prefix(o, o->output_prefix_data);
572 line_prefix = msgbuf->buf;
573 }
574
575 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
576 a_prefix = o->b_prefix;
577 b_prefix = o->a_prefix;
578 } else {
579 a_prefix = o->a_prefix;
580 b_prefix = o->b_prefix;
581 }
582
583 name_a += (*name_a == '/');
584 name_b += (*name_b == '/');
585 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
586 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
587
588 strbuf_reset(&a_name);
589 strbuf_reset(&b_name);
590 quote_two_c_style(&a_name, a_prefix, name_a, 0);
591 quote_two_c_style(&b_name, b_prefix, name_b, 0);
592
593 size_one = fill_textconv(textconv_one, one, &data_one);
594 size_two = fill_textconv(textconv_two, two, &data_two);
595
596 memset(&ecbdata, 0, sizeof(ecbdata));
597 ecbdata.color_diff = color_diff;
598 ecbdata.found_changesp = &o->found_changes;
599 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
600 ecbdata.opt = o;
601 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
602 mmfile_t mf1, mf2;
603 mf1.ptr = (char *)data_one;
604 mf2.ptr = (char *)data_two;
605 mf1.size = size_one;
606 mf2.size = size_two;
607 check_blank_at_eof(&mf1, &mf2, &ecbdata);
608 }
609 ecbdata.lno_in_preimage = 1;
610 ecbdata.lno_in_postimage = 1;
611
612 lc_a = count_lines(data_one, size_one);
613 lc_b = count_lines(data_two, size_two);
614 fprintf(o->file,
615 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
616 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
617 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
618 line_prefix, fraginfo);
619 if (!o->irreversible_delete)
620 print_line_count(o->file, lc_a);
621 else
622 fprintf(o->file, "?,?");
623 fprintf(o->file, " +");
624 print_line_count(o->file, lc_b);
625 fprintf(o->file, " @@%s\n", reset);
626 if (lc_a && !o->irreversible_delete)
627 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
628 if (lc_b)
629 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
630 if (textconv_one)
631 free((char *)data_one);
632 if (textconv_two)
633 free((char *)data_two);
634 }
635
636 struct diff_words_buffer {
637 mmfile_t text;
638 long alloc;
639 struct diff_words_orig {
640 const char *begin, *end;
641 } *orig;
642 int orig_nr, orig_alloc;
643 };
644
645 static void diff_words_append(char *line, unsigned long len,
646 struct diff_words_buffer *buffer)
647 {
648 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
649 line++;
650 len--;
651 memcpy(buffer->text.ptr + buffer->text.size, line, len);
652 buffer->text.size += len;
653 buffer->text.ptr[buffer->text.size] = '\0';
654 }
655
656 struct diff_words_style_elem {
657 const char *prefix;
658 const char *suffix;
659 const char *color; /* NULL; filled in by the setup code if
660 * color is enabled */
661 };
662
663 struct diff_words_style {
664 enum diff_words_type type;
665 struct diff_words_style_elem new, old, ctx;
666 const char *newline;
667 };
668
669 static struct diff_words_style diff_words_styles[] = {
670 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
671 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
672 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
673 };
674
675 struct diff_words_data {
676 struct diff_words_buffer minus, plus;
677 const char *current_plus;
678 int last_minus;
679 struct diff_options *opt;
680 regex_t *word_regex;
681 enum diff_words_type type;
682 struct diff_words_style *style;
683 };
684
685 static int fn_out_diff_words_write_helper(FILE *fp,
686 struct diff_words_style_elem *st_el,
687 const char *newline,
688 size_t count, const char *buf,
689 const char *line_prefix)
690 {
691 int print = 0;
692
693 while (count) {
694 char *p = memchr(buf, '\n', count);
695 if (print)
696 fputs(line_prefix, fp);
697 if (p != buf) {
698 if (st_el->color && fputs(st_el->color, fp) < 0)
699 return -1;
700 if (fputs(st_el->prefix, fp) < 0 ||
701 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
702 fputs(st_el->suffix, fp) < 0)
703 return -1;
704 if (st_el->color && *st_el->color
705 && fputs(GIT_COLOR_RESET, fp) < 0)
706 return -1;
707 }
708 if (!p)
709 return 0;
710 if (fputs(newline, fp) < 0)
711 return -1;
712 count -= p + 1 - buf;
713 buf = p + 1;
714 print = 1;
715 }
716 return 0;
717 }
718
719 /*
720 * '--color-words' algorithm can be described as:
721 *
722 * 1. collect a the minus/plus lines of a diff hunk, divided into
723 * minus-lines and plus-lines;
724 *
725 * 2. break both minus-lines and plus-lines into words and
726 * place them into two mmfile_t with one word for each line;
727 *
728 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
729 *
730 * And for the common parts of the both file, we output the plus side text.
731 * diff_words->current_plus is used to trace the current position of the plus file
732 * which printed. diff_words->last_minus is used to trace the last minus word
733 * printed.
734 *
735 * For '--graph' to work with '--color-words', we need to output the graph prefix
736 * on each line of color words output. Generally, there are two conditions on
737 * which we should output the prefix.
738 *
739 * 1. diff_words->last_minus == 0 &&
740 * diff_words->current_plus == diff_words->plus.text.ptr
741 *
742 * that is: the plus text must start as a new line, and if there is no minus
743 * word printed, a graph prefix must be printed.
744 *
745 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
746 * *(diff_words->current_plus - 1) == '\n'
747 *
748 * that is: a graph prefix must be printed following a '\n'
749 */
750 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
751 {
752 if ((diff_words->last_minus == 0 &&
753 diff_words->current_plus == diff_words->plus.text.ptr) ||
754 (diff_words->current_plus > diff_words->plus.text.ptr &&
755 *(diff_words->current_plus - 1) == '\n')) {
756 return 1;
757 } else {
758 return 0;
759 }
760 }
761
762 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
763 {
764 struct diff_words_data *diff_words = priv;
765 struct diff_words_style *style = diff_words->style;
766 int minus_first, minus_len, plus_first, plus_len;
767 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
768 struct diff_options *opt = diff_words->opt;
769 struct strbuf *msgbuf;
770 char *line_prefix = "";
771
772 if (line[0] != '@' || parse_hunk_header(line, len,
773 &minus_first, &minus_len, &plus_first, &plus_len))
774 return;
775
776 assert(opt);
777 if (opt->output_prefix) {
778 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
779 line_prefix = msgbuf->buf;
780 }
781
782 /* POSIX requires that first be decremented by one if len == 0... */
783 if (minus_len) {
784 minus_begin = diff_words->minus.orig[minus_first].begin;
785 minus_end =
786 diff_words->minus.orig[minus_first + minus_len - 1].end;
787 } else
788 minus_begin = minus_end =
789 diff_words->minus.orig[minus_first].end;
790
791 if (plus_len) {
792 plus_begin = diff_words->plus.orig[plus_first].begin;
793 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
794 } else
795 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
796
797 if (color_words_output_graph_prefix(diff_words)) {
798 fputs(line_prefix, diff_words->opt->file);
799 }
800 if (diff_words->current_plus != plus_begin) {
801 fn_out_diff_words_write_helper(diff_words->opt->file,
802 &style->ctx, style->newline,
803 plus_begin - diff_words->current_plus,
804 diff_words->current_plus, line_prefix);
805 if (*(plus_begin - 1) == '\n')
806 fputs(line_prefix, diff_words->opt->file);
807 }
808 if (minus_begin != minus_end) {
809 fn_out_diff_words_write_helper(diff_words->opt->file,
810 &style->old, style->newline,
811 minus_end - minus_begin, minus_begin,
812 line_prefix);
813 }
814 if (plus_begin != plus_end) {
815 fn_out_diff_words_write_helper(diff_words->opt->file,
816 &style->new, style->newline,
817 plus_end - plus_begin, plus_begin,
818 line_prefix);
819 }
820
821 diff_words->current_plus = plus_end;
822 diff_words->last_minus = minus_first;
823 }
824
825 /* This function starts looking at *begin, and returns 0 iff a word was found. */
826 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
827 int *begin, int *end)
828 {
829 if (word_regex && *begin < buffer->size) {
830 regmatch_t match[1];
831 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
832 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
833 '\n', match[0].rm_eo - match[0].rm_so);
834 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
835 *begin += match[0].rm_so;
836 return *begin >= *end;
837 }
838 return -1;
839 }
840
841 /* find the next word */
842 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
843 (*begin)++;
844 if (*begin >= buffer->size)
845 return -1;
846
847 /* find the end of the word */
848 *end = *begin + 1;
849 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
850 (*end)++;
851
852 return 0;
853 }
854
855 /*
856 * This function splits the words in buffer->text, stores the list with
857 * newline separator into out, and saves the offsets of the original words
858 * in buffer->orig.
859 */
860 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
861 regex_t *word_regex)
862 {
863 int i, j;
864 long alloc = 0;
865
866 out->size = 0;
867 out->ptr = NULL;
868
869 /* fake an empty "0th" word */
870 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
871 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
872 buffer->orig_nr = 1;
873
874 for (i = 0; i < buffer->text.size; i++) {
875 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
876 return;
877
878 /* store original boundaries */
879 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
880 buffer->orig_alloc);
881 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
882 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
883 buffer->orig_nr++;
884
885 /* store one word */
886 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
887 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
888 out->ptr[out->size + j - i] = '\n';
889 out->size += j - i + 1;
890
891 i = j - 1;
892 }
893 }
894
895 /* this executes the word diff on the accumulated buffers */
896 static void diff_words_show(struct diff_words_data *diff_words)
897 {
898 xpparam_t xpp;
899 xdemitconf_t xecfg;
900 mmfile_t minus, plus;
901 struct diff_words_style *style = diff_words->style;
902
903 struct diff_options *opt = diff_words->opt;
904 struct strbuf *msgbuf;
905 char *line_prefix = "";
906
907 assert(opt);
908 if (opt->output_prefix) {
909 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
910 line_prefix = msgbuf->buf;
911 }
912
913 /* special case: only removal */
914 if (!diff_words->plus.text.size) {
915 fputs(line_prefix, diff_words->opt->file);
916 fn_out_diff_words_write_helper(diff_words->opt->file,
917 &style->old, style->newline,
918 diff_words->minus.text.size,
919 diff_words->minus.text.ptr, line_prefix);
920 diff_words->minus.text.size = 0;
921 return;
922 }
923
924 diff_words->current_plus = diff_words->plus.text.ptr;
925 diff_words->last_minus = 0;
926
927 memset(&xpp, 0, sizeof(xpp));
928 memset(&xecfg, 0, sizeof(xecfg));
929 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
930 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
931 xpp.flags = 0;
932 /* as only the hunk header will be parsed, we need a 0-context */
933 xecfg.ctxlen = 0;
934 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
935 &xpp, &xecfg);
936 free(minus.ptr);
937 free(plus.ptr);
938 if (diff_words->current_plus != diff_words->plus.text.ptr +
939 diff_words->plus.text.size) {
940 if (color_words_output_graph_prefix(diff_words))
941 fputs(line_prefix, diff_words->opt->file);
942 fn_out_diff_words_write_helper(diff_words->opt->file,
943 &style->ctx, style->newline,
944 diff_words->plus.text.ptr + diff_words->plus.text.size
945 - diff_words->current_plus, diff_words->current_plus,
946 line_prefix);
947 }
948 diff_words->minus.text.size = diff_words->plus.text.size = 0;
949 }
950
951 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
952 static void diff_words_flush(struct emit_callback *ecbdata)
953 {
954 if (ecbdata->diff_words->minus.text.size ||
955 ecbdata->diff_words->plus.text.size)
956 diff_words_show(ecbdata->diff_words);
957 }
958
959 static void free_diff_words_data(struct emit_callback *ecbdata)
960 {
961 if (ecbdata->diff_words) {
962 diff_words_flush(ecbdata);
963 free (ecbdata->diff_words->minus.text.ptr);
964 free (ecbdata->diff_words->minus.orig);
965 free (ecbdata->diff_words->plus.text.ptr);
966 free (ecbdata->diff_words->plus.orig);
967 if (ecbdata->diff_words->word_regex) {
968 regfree(ecbdata->diff_words->word_regex);
969 free(ecbdata->diff_words->word_regex);
970 }
971 free(ecbdata->diff_words);
972 ecbdata->diff_words = NULL;
973 }
974 }
975
976 const char *diff_get_color(int diff_use_color, enum color_diff ix)
977 {
978 if (diff_use_color)
979 return diff_colors[ix];
980 return "";
981 }
982
983 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
984 {
985 const char *cp;
986 unsigned long allot;
987 size_t l = len;
988
989 if (ecb->truncate)
990 return ecb->truncate(line, len);
991 cp = line;
992 allot = l;
993 while (0 < l) {
994 (void) utf8_width(&cp, &l);
995 if (!cp)
996 break; /* truncated in the middle? */
997 }
998 return allot - l;
999 }
1000
1001 static void find_lno(const char *line, struct emit_callback *ecbdata)
1002 {
1003 const char *p;
1004 ecbdata->lno_in_preimage = 0;
1005 ecbdata->lno_in_postimage = 0;
1006 p = strchr(line, '-');
1007 if (!p)
1008 return; /* cannot happen */
1009 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1010 p = strchr(p, '+');
1011 if (!p)
1012 return; /* cannot happen */
1013 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1014 }
1015
1016 static void fn_out_consume(void *priv, char *line, unsigned long len)
1017 {
1018 struct emit_callback *ecbdata = priv;
1019 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1020 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1021 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1022 struct diff_options *o = ecbdata->opt;
1023 char *line_prefix = "";
1024 struct strbuf *msgbuf;
1025
1026 if (o && o->output_prefix) {
1027 msgbuf = o->output_prefix(o, o->output_prefix_data);
1028 line_prefix = msgbuf->buf;
1029 }
1030
1031 if (ecbdata->header) {
1032 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1033 strbuf_reset(ecbdata->header);
1034 ecbdata->header = NULL;
1035 }
1036 *(ecbdata->found_changesp) = 1;
1037
1038 if (ecbdata->label_path[0]) {
1039 const char *name_a_tab, *name_b_tab;
1040
1041 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1042 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1043
1044 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1045 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1046 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1047 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1048 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1049 }
1050
1051 if (diff_suppress_blank_empty
1052 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1053 line[0] = '\n';
1054 len = 1;
1055 }
1056
1057 if (line[0] == '@') {
1058 if (ecbdata->diff_words)
1059 diff_words_flush(ecbdata);
1060 len = sane_truncate_line(ecbdata, line, len);
1061 find_lno(line, ecbdata);
1062 emit_hunk_header(ecbdata, line, len);
1063 if (line[len-1] != '\n')
1064 putc('\n', ecbdata->opt->file);
1065 return;
1066 }
1067
1068 if (len < 1) {
1069 emit_line(ecbdata->opt, reset, reset, line, len);
1070 if (ecbdata->diff_words
1071 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1072 fputs("~\n", ecbdata->opt->file);
1073 return;
1074 }
1075
1076 if (ecbdata->diff_words) {
1077 if (line[0] == '-') {
1078 diff_words_append(line, len,
1079 &ecbdata->diff_words->minus);
1080 return;
1081 } else if (line[0] == '+') {
1082 diff_words_append(line, len,
1083 &ecbdata->diff_words->plus);
1084 return;
1085 }
1086 diff_words_flush(ecbdata);
1087 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1088 emit_line(ecbdata->opt, plain, reset, line, len);
1089 fputs("~\n", ecbdata->opt->file);
1090 } else {
1091 /* don't print the prefix character */
1092 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1093 }
1094 return;
1095 }
1096
1097 if (line[0] != '+') {
1098 const char *color =
1099 diff_get_color(ecbdata->color_diff,
1100 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1101 ecbdata->lno_in_preimage++;
1102 if (line[0] == ' ')
1103 ecbdata->lno_in_postimage++;
1104 emit_line(ecbdata->opt, color, reset, line, len);
1105 } else {
1106 ecbdata->lno_in_postimage++;
1107 emit_add_line(reset, ecbdata, line + 1, len - 1);
1108 }
1109 }
1110
1111 static char *pprint_rename(const char *a, const char *b)
1112 {
1113 const char *old = a;
1114 const char *new = b;
1115 struct strbuf name = STRBUF_INIT;
1116 int pfx_length, sfx_length;
1117 int len_a = strlen(a);
1118 int len_b = strlen(b);
1119 int a_midlen, b_midlen;
1120 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1121 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1122
1123 if (qlen_a || qlen_b) {
1124 quote_c_style(a, &name, NULL, 0);
1125 strbuf_addstr(&name, " => ");
1126 quote_c_style(b, &name, NULL, 0);
1127 return strbuf_detach(&name, NULL);
1128 }
1129
1130 /* Find common prefix */
1131 pfx_length = 0;
1132 while (*old && *new && *old == *new) {
1133 if (*old == '/')
1134 pfx_length = old - a + 1;
1135 old++;
1136 new++;
1137 }
1138
1139 /* Find common suffix */
1140 old = a + len_a;
1141 new = b + len_b;
1142 sfx_length = 0;
1143 while (a <= old && b <= new && *old == *new) {
1144 if (*old == '/')
1145 sfx_length = len_a - (old - a);
1146 old--;
1147 new--;
1148 }
1149
1150 /*
1151 * pfx{mid-a => mid-b}sfx
1152 * {pfx-a => pfx-b}sfx
1153 * pfx{sfx-a => sfx-b}
1154 * name-a => name-b
1155 */
1156 a_midlen = len_a - pfx_length - sfx_length;
1157 b_midlen = len_b - pfx_length - sfx_length;
1158 if (a_midlen < 0)
1159 a_midlen = 0;
1160 if (b_midlen < 0)
1161 b_midlen = 0;
1162
1163 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1164 if (pfx_length + sfx_length) {
1165 strbuf_add(&name, a, pfx_length);
1166 strbuf_addch(&name, '{');
1167 }
1168 strbuf_add(&name, a + pfx_length, a_midlen);
1169 strbuf_addstr(&name, " => ");
1170 strbuf_add(&name, b + pfx_length, b_midlen);
1171 if (pfx_length + sfx_length) {
1172 strbuf_addch(&name, '}');
1173 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1174 }
1175 return strbuf_detach(&name, NULL);
1176 }
1177
1178 struct diffstat_t {
1179 int nr;
1180 int alloc;
1181 struct diffstat_file {
1182 char *from_name;
1183 char *name;
1184 char *print_name;
1185 unsigned is_unmerged:1;
1186 unsigned is_binary:1;
1187 unsigned is_renamed:1;
1188 uintmax_t added, deleted;
1189 } **files;
1190 };
1191
1192 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1193 const char *name_a,
1194 const char *name_b)
1195 {
1196 struct diffstat_file *x;
1197 x = xcalloc(sizeof (*x), 1);
1198 if (diffstat->nr == diffstat->alloc) {
1199 diffstat->alloc = alloc_nr(diffstat->alloc);
1200 diffstat->files = xrealloc(diffstat->files,
1201 diffstat->alloc * sizeof(x));
1202 }
1203 diffstat->files[diffstat->nr++] = x;
1204 if (name_b) {
1205 x->from_name = xstrdup(name_a);
1206 x->name = xstrdup(name_b);
1207 x->is_renamed = 1;
1208 }
1209 else {
1210 x->from_name = NULL;
1211 x->name = xstrdup(name_a);
1212 }
1213 return x;
1214 }
1215
1216 static void diffstat_consume(void *priv, char *line, unsigned long len)
1217 {
1218 struct diffstat_t *diffstat = priv;
1219 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1220
1221 if (line[0] == '+')
1222 x->added++;
1223 else if (line[0] == '-')
1224 x->deleted++;
1225 }
1226
1227 const char mime_boundary_leader[] = "------------";
1228
1229 static int scale_linear(int it, int width, int max_change)
1230 {
1231 /*
1232 * make sure that at least one '-' is printed if there were deletions,
1233 * and likewise for '+'.
1234 */
1235 if (max_change < 2)
1236 return it;
1237 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1238 }
1239
1240 static void show_name(FILE *file,
1241 const char *prefix, const char *name, int len)
1242 {
1243 fprintf(file, " %s%-*s |", prefix, len, name);
1244 }
1245
1246 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1247 {
1248 if (cnt <= 0)
1249 return;
1250 fprintf(file, "%s", set);
1251 while (cnt--)
1252 putc(ch, file);
1253 fprintf(file, "%s", reset);
1254 }
1255
1256 static void fill_print_name(struct diffstat_file *file)
1257 {
1258 char *pname;
1259
1260 if (file->print_name)
1261 return;
1262
1263 if (!file->is_renamed) {
1264 struct strbuf buf = STRBUF_INIT;
1265 if (quote_c_style(file->name, &buf, NULL, 0)) {
1266 pname = strbuf_detach(&buf, NULL);
1267 } else {
1268 pname = file->name;
1269 strbuf_release(&buf);
1270 }
1271 } else {
1272 pname = pprint_rename(file->from_name, file->name);
1273 }
1274 file->print_name = pname;
1275 }
1276
1277 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1278 {
1279 int i, len, add, del, adds = 0, dels = 0;
1280 uintmax_t max_change = 0, max_len = 0;
1281 int total_files = data->nr;
1282 int width, name_width;
1283 const char *reset, *add_c, *del_c;
1284 const char *line_prefix = "";
1285 struct strbuf *msg = NULL;
1286
1287 if (data->nr == 0)
1288 return;
1289
1290 if (options->output_prefix) {
1291 msg = options->output_prefix(options, options->output_prefix_data);
1292 line_prefix = msg->buf;
1293 }
1294
1295 width = options->stat_width ? options->stat_width : 80;
1296 name_width = options->stat_name_width ? options->stat_name_width : 50;
1297
1298 /* Sanity: give at least 5 columns to the graph,
1299 * but leave at least 10 columns for the name.
1300 */
1301 if (width < 25)
1302 width = 25;
1303 if (name_width < 10)
1304 name_width = 10;
1305 else if (width < name_width + 15)
1306 name_width = width - 15;
1307
1308 /* Find the longest filename and max number of changes */
1309 reset = diff_get_color_opt(options, DIFF_RESET);
1310 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1311 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1312
1313 for (i = 0; i < data->nr; i++) {
1314 struct diffstat_file *file = data->files[i];
1315 uintmax_t change = file->added + file->deleted;
1316 fill_print_name(file);
1317 len = strlen(file->print_name);
1318 if (max_len < len)
1319 max_len = len;
1320
1321 if (file->is_binary || file->is_unmerged)
1322 continue;
1323 if (max_change < change)
1324 max_change = change;
1325 }
1326
1327 /* Compute the width of the graph part;
1328 * 10 is for one blank at the beginning of the line plus
1329 * " | count " between the name and the graph.
1330 *
1331 * From here on, name_width is the width of the name area,
1332 * and width is the width of the graph area.
1333 */
1334 name_width = (name_width < max_len) ? name_width : max_len;
1335 if (width < (name_width + 10) + max_change)
1336 width = width - (name_width + 10);
1337 else
1338 width = max_change;
1339
1340 for (i = 0; i < data->nr; i++) {
1341 const char *prefix = "";
1342 char *name = data->files[i]->print_name;
1343 uintmax_t added = data->files[i]->added;
1344 uintmax_t deleted = data->files[i]->deleted;
1345 int name_len;
1346
1347 /*
1348 * "scale" the filename
1349 */
1350 len = name_width;
1351 name_len = strlen(name);
1352 if (name_width < name_len) {
1353 char *slash;
1354 prefix = "...";
1355 len -= 3;
1356 name += name_len - len;
1357 slash = strchr(name, '/');
1358 if (slash)
1359 name = slash;
1360 }
1361
1362 if (data->files[i]->is_binary) {
1363 fprintf(options->file, "%s", line_prefix);
1364 show_name(options->file, prefix, name, len);
1365 fprintf(options->file, " Bin ");
1366 fprintf(options->file, "%s%"PRIuMAX"%s",
1367 del_c, deleted, reset);
1368 fprintf(options->file, " -> ");
1369 fprintf(options->file, "%s%"PRIuMAX"%s",
1370 add_c, added, reset);
1371 fprintf(options->file, " bytes");
1372 fprintf(options->file, "\n");
1373 continue;
1374 }
1375 else if (data->files[i]->is_unmerged) {
1376 fprintf(options->file, "%s", line_prefix);
1377 show_name(options->file, prefix, name, len);
1378 fprintf(options->file, " Unmerged\n");
1379 continue;
1380 }
1381 else if (!data->files[i]->is_renamed &&
1382 (added + deleted == 0)) {
1383 total_files--;
1384 continue;
1385 }
1386
1387 /*
1388 * scale the add/delete
1389 */
1390 add = added;
1391 del = deleted;
1392 adds += add;
1393 dels += del;
1394
1395 if (width <= max_change) {
1396 add = scale_linear(add, width, max_change);
1397 del = scale_linear(del, width, max_change);
1398 }
1399 fprintf(options->file, "%s", line_prefix);
1400 show_name(options->file, prefix, name, len);
1401 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1402 added + deleted ? " " : "");
1403 show_graph(options->file, '+', add, add_c, reset);
1404 show_graph(options->file, '-', del, del_c, reset);
1405 fprintf(options->file, "\n");
1406 }
1407 fprintf(options->file, "%s", line_prefix);
1408 fprintf(options->file,
1409 " %d files changed, %d insertions(+), %d deletions(-)\n",
1410 total_files, adds, dels);
1411 }
1412
1413 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1414 {
1415 int i, adds = 0, dels = 0, total_files = data->nr;
1416
1417 if (data->nr == 0)
1418 return;
1419
1420 for (i = 0; i < data->nr; i++) {
1421 if (!data->files[i]->is_binary &&
1422 !data->files[i]->is_unmerged) {
1423 int added = data->files[i]->added;
1424 int deleted= data->files[i]->deleted;
1425 if (!data->files[i]->is_renamed &&
1426 (added + deleted == 0)) {
1427 total_files--;
1428 } else {
1429 adds += added;
1430 dels += deleted;
1431 }
1432 }
1433 }
1434 if (options->output_prefix) {
1435 struct strbuf *msg = NULL;
1436 msg = options->output_prefix(options,
1437 options->output_prefix_data);
1438 fprintf(options->file, "%s", msg->buf);
1439 }
1440 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1441 total_files, adds, dels);
1442 }
1443
1444 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1445 {
1446 int i;
1447
1448 if (data->nr == 0)
1449 return;
1450
1451 for (i = 0; i < data->nr; i++) {
1452 struct diffstat_file *file = data->files[i];
1453
1454 if (options->output_prefix) {
1455 struct strbuf *msg = NULL;
1456 msg = options->output_prefix(options,
1457 options->output_prefix_data);
1458 fprintf(options->file, "%s", msg->buf);
1459 }
1460
1461 if (file->is_binary)
1462 fprintf(options->file, "-\t-\t");
1463 else
1464 fprintf(options->file,
1465 "%"PRIuMAX"\t%"PRIuMAX"\t",
1466 file->added, file->deleted);
1467 if (options->line_termination) {
1468 fill_print_name(file);
1469 if (!file->is_renamed)
1470 write_name_quoted(file->name, options->file,
1471 options->line_termination);
1472 else {
1473 fputs(file->print_name, options->file);
1474 putc(options->line_termination, options->file);
1475 }
1476 } else {
1477 if (file->is_renamed) {
1478 putc('\0', options->file);
1479 write_name_quoted(file->from_name, options->file, '\0');
1480 }
1481 write_name_quoted(file->name, options->file, '\0');
1482 }
1483 }
1484 }
1485
1486 struct dirstat_file {
1487 const char *name;
1488 unsigned long changed;
1489 };
1490
1491 struct dirstat_dir {
1492 struct dirstat_file *files;
1493 int alloc, nr, percent, cumulative;
1494 };
1495
1496 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1497 unsigned long changed, const char *base, int baselen)
1498 {
1499 unsigned long this_dir = 0;
1500 unsigned int sources = 0;
1501 const char *line_prefix = "";
1502 struct strbuf *msg = NULL;
1503
1504 if (opt->output_prefix) {
1505 msg = opt->output_prefix(opt, opt->output_prefix_data);
1506 line_prefix = msg->buf;
1507 }
1508
1509 while (dir->nr) {
1510 struct dirstat_file *f = dir->files;
1511 int namelen = strlen(f->name);
1512 unsigned long this;
1513 char *slash;
1514
1515 if (namelen < baselen)
1516 break;
1517 if (memcmp(f->name, base, baselen))
1518 break;
1519 slash = strchr(f->name + baselen, '/');
1520 if (slash) {
1521 int newbaselen = slash + 1 - f->name;
1522 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1523 sources++;
1524 } else {
1525 this = f->changed;
1526 dir->files++;
1527 dir->nr--;
1528 sources += 2;
1529 }
1530 this_dir += this;
1531 }
1532
1533 /*
1534 * We don't report dirstat's for
1535 * - the top level
1536 * - or cases where everything came from a single directory
1537 * under this directory (sources == 1).
1538 */
1539 if (baselen && sources != 1) {
1540 if (this_dir) {
1541 int permille = this_dir * 1000 / changed;
1542 int percent = permille / 10;
1543 if (percent >= dir->percent) {
1544 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1545 percent, permille % 10, baselen, base);
1546 if (!dir->cumulative)
1547 return 0;
1548 }
1549 }
1550 }
1551 return this_dir;
1552 }
1553
1554 static int dirstat_compare(const void *_a, const void *_b)
1555 {
1556 const struct dirstat_file *a = _a;
1557 const struct dirstat_file *b = _b;
1558 return strcmp(a->name, b->name);
1559 }
1560
1561 static void show_dirstat(struct diff_options *options)
1562 {
1563 int i;
1564 unsigned long changed;
1565 struct dirstat_dir dir;
1566 struct diff_queue_struct *q = &diff_queued_diff;
1567
1568 dir.files = NULL;
1569 dir.alloc = 0;
1570 dir.nr = 0;
1571 dir.percent = options->dirstat_percent;
1572 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1573
1574 changed = 0;
1575 for (i = 0; i < q->nr; i++) {
1576 struct diff_filepair *p = q->queue[i];
1577 const char *name;
1578 unsigned long copied, added, damage;
1579 int content_changed;
1580
1581 name = p->two->path ? p->two->path : p->one->path;
1582
1583 if (p->one->sha1_valid && p->two->sha1_valid)
1584 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1585 else
1586 content_changed = 1;
1587
1588 if (!content_changed) {
1589 /*
1590 * The SHA1 has not changed, so pre-/post-content is
1591 * identical. We can therefore skip looking at the
1592 * file contents altogether.
1593 */
1594 damage = 0;
1595 goto found_damage;
1596 }
1597
1598 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1599 /*
1600 * In --dirstat-by-file mode, we don't really need to
1601 * look at the actual file contents at all.
1602 * The fact that the SHA1 changed is enough for us to
1603 * add this file to the list of results
1604 * (with each file contributing equal damage).
1605 */
1606 damage = 1;
1607 goto found_damage;
1608 }
1609
1610 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1611 diff_populate_filespec(p->one, 0);
1612 diff_populate_filespec(p->two, 0);
1613 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1614 &copied, &added);
1615 diff_free_filespec_data(p->one);
1616 diff_free_filespec_data(p->two);
1617 } else if (DIFF_FILE_VALID(p->one)) {
1618 diff_populate_filespec(p->one, 1);
1619 copied = added = 0;
1620 diff_free_filespec_data(p->one);
1621 } else if (DIFF_FILE_VALID(p->two)) {
1622 diff_populate_filespec(p->two, 1);
1623 copied = 0;
1624 added = p->two->size;
1625 diff_free_filespec_data(p->two);
1626 } else
1627 continue;
1628
1629 /*
1630 * Original minus copied is the removed material,
1631 * added is the new material. They are both damages
1632 * made to the preimage.
1633 * If the resulting damage is zero, we know that
1634 * diffcore_count_changes() considers the two entries to
1635 * be identical, but since content_changed is true, we
1636 * know that there must have been _some_ kind of change,
1637 * so we force all entries to have damage > 0.
1638 */
1639 damage = (p->one->size - copied) + added;
1640 if (!damage)
1641 damage = 1;
1642
1643 found_damage:
1644 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1645 dir.files[dir.nr].name = name;
1646 dir.files[dir.nr].changed = damage;
1647 changed += damage;
1648 dir.nr++;
1649 }
1650
1651 /* This can happen even with many files, if everything was renames */
1652 if (!changed)
1653 return;
1654
1655 /* Show all directories with more than x% of the changes */
1656 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1657 gather_dirstat(options, &dir, changed, "", 0);
1658 }
1659
1660 static void free_diffstat_info(struct diffstat_t *diffstat)
1661 {
1662 int i;
1663 for (i = 0; i < diffstat->nr; i++) {
1664 struct diffstat_file *f = diffstat->files[i];
1665 if (f->name != f->print_name)
1666 free(f->print_name);
1667 free(f->name);
1668 free(f->from_name);
1669 free(f);
1670 }
1671 free(diffstat->files);
1672 }
1673
1674 struct checkdiff_t {
1675 const char *filename;
1676 int lineno;
1677 int conflict_marker_size;
1678 struct diff_options *o;
1679 unsigned ws_rule;
1680 unsigned status;
1681 };
1682
1683 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1684 {
1685 char firstchar;
1686 int cnt;
1687
1688 if (len < marker_size + 1)
1689 return 0;
1690 firstchar = line[0];
1691 switch (firstchar) {
1692 case '=': case '>': case '<': case '|':
1693 break;
1694 default:
1695 return 0;
1696 }
1697 for (cnt = 1; cnt < marker_size; cnt++)
1698 if (line[cnt] != firstchar)
1699 return 0;
1700 /* line[1] thru line[marker_size-1] are same as firstchar */
1701 if (len < marker_size + 1 || !isspace(line[marker_size]))
1702 return 0;
1703 return 1;
1704 }
1705
1706 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1707 {
1708 struct checkdiff_t *data = priv;
1709 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1710 int marker_size = data->conflict_marker_size;
1711 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1712 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1713 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1714 char *err;
1715 char *line_prefix = "";
1716 struct strbuf *msgbuf;
1717
1718 assert(data->o);
1719 if (data->o->output_prefix) {
1720 msgbuf = data->o->output_prefix(data->o,
1721 data->o->output_prefix_data);
1722 line_prefix = msgbuf->buf;
1723 }
1724
1725 if (line[0] == '+') {
1726 unsigned bad;
1727 data->lineno++;
1728 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1729 data->status |= 1;
1730 fprintf(data->o->file,
1731 "%s%s:%d: leftover conflict marker\n",
1732 line_prefix, data->filename, data->lineno);
1733 }
1734 bad = ws_check(line + 1, len - 1, data->ws_rule);
1735 if (!bad)
1736 return;
1737 data->status |= bad;
1738 err = whitespace_error_string(bad);
1739 fprintf(data->o->file, "%s%s:%d: %s.\n",
1740 line_prefix, data->filename, data->lineno, err);
1741 free(err);
1742 emit_line(data->o, set, reset, line, 1);
1743 ws_check_emit(line + 1, len - 1, data->ws_rule,
1744 data->o->file, set, reset, ws);
1745 } else if (line[0] == ' ') {
1746 data->lineno++;
1747 } else if (line[0] == '@') {
1748 char *plus = strchr(line, '+');
1749 if (plus)
1750 data->lineno = strtol(plus, NULL, 10) - 1;
1751 else
1752 die("invalid diff");
1753 }
1754 }
1755
1756 static unsigned char *deflate_it(char *data,
1757 unsigned long size,
1758 unsigned long *result_size)
1759 {
1760 int bound;
1761 unsigned char *deflated;
1762 z_stream stream;
1763
1764 memset(&stream, 0, sizeof(stream));
1765 deflateInit(&stream, zlib_compression_level);
1766 bound = deflateBound(&stream, size);
1767 deflated = xmalloc(bound);
1768 stream.next_out = deflated;
1769 stream.avail_out = bound;
1770
1771 stream.next_in = (unsigned char *)data;
1772 stream.avail_in = size;
1773 while (deflate(&stream, Z_FINISH) == Z_OK)
1774 ; /* nothing */
1775 deflateEnd(&stream);
1776 *result_size = stream.total_out;
1777 return deflated;
1778 }
1779
1780 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1781 {
1782 void *cp;
1783 void *delta;
1784 void *deflated;
1785 void *data;
1786 unsigned long orig_size;
1787 unsigned long delta_size;
1788 unsigned long deflate_size;
1789 unsigned long data_size;
1790
1791 /* We could do deflated delta, or we could do just deflated two,
1792 * whichever is smaller.
1793 */
1794 delta = NULL;
1795 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1796 if (one->size && two->size) {
1797 delta = diff_delta(one->ptr, one->size,
1798 two->ptr, two->size,
1799 &delta_size, deflate_size);
1800 if (delta) {
1801 void *to_free = delta;
1802 orig_size = delta_size;
1803 delta = deflate_it(delta, delta_size, &delta_size);
1804 free(to_free);
1805 }
1806 }
1807
1808 if (delta && delta_size < deflate_size) {
1809 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1810 free(deflated);
1811 data = delta;
1812 data_size = delta_size;
1813 }
1814 else {
1815 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1816 free(delta);
1817 data = deflated;
1818 data_size = deflate_size;
1819 }
1820
1821 /* emit data encoded in base85 */
1822 cp = data;
1823 while (data_size) {
1824 int bytes = (52 < data_size) ? 52 : data_size;
1825 char line[70];
1826 data_size -= bytes;
1827 if (bytes <= 26)
1828 line[0] = bytes + 'A' - 1;
1829 else
1830 line[0] = bytes - 26 + 'a' - 1;
1831 encode_85(line + 1, cp, bytes);
1832 cp = (char *) cp + bytes;
1833 fprintf(file, "%s", prefix);
1834 fputs(line, file);
1835 fputc('\n', file);
1836 }
1837 fprintf(file, "%s\n", prefix);
1838 free(data);
1839 }
1840
1841 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1842 {
1843 fprintf(file, "%sGIT binary patch\n", prefix);
1844 emit_binary_diff_body(file, one, two, prefix);
1845 emit_binary_diff_body(file, two, one, prefix);
1846 }
1847
1848 static void diff_filespec_load_driver(struct diff_filespec *one)
1849 {
1850 /* Use already-loaded driver */
1851 if (one->driver)
1852 return;
1853
1854 if (S_ISREG(one->mode))
1855 one->driver = userdiff_find_by_path(one->path);
1856
1857 /* Fallback to default settings */
1858 if (!one->driver)
1859 one->driver = userdiff_find_by_name("default");
1860 }
1861
1862 int diff_filespec_is_binary(struct diff_filespec *one)
1863 {
1864 if (one->is_binary == -1) {
1865 diff_filespec_load_driver(one);
1866 if (one->driver->binary != -1)
1867 one->is_binary = one->driver->binary;
1868 else {
1869 if (!one->data && DIFF_FILE_VALID(one))
1870 diff_populate_filespec(one, 0);
1871 if (one->data)
1872 one->is_binary = buffer_is_binary(one->data,
1873 one->size);
1874 if (one->is_binary == -1)
1875 one->is_binary = 0;
1876 }
1877 }
1878 return one->is_binary;
1879 }
1880
1881 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1882 {
1883 diff_filespec_load_driver(one);
1884 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1885 }
1886
1887 static const char *userdiff_word_regex(struct diff_filespec *one)
1888 {
1889 diff_filespec_load_driver(one);
1890 return one->driver->word_regex;
1891 }
1892
1893 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1894 {
1895 if (!options->a_prefix)
1896 options->a_prefix = a;
1897 if (!options->b_prefix)
1898 options->b_prefix = b;
1899 }
1900
1901 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1902 {
1903 if (!DIFF_FILE_VALID(one))
1904 return NULL;
1905
1906 diff_filespec_load_driver(one);
1907 if (!one->driver->textconv)
1908 return NULL;
1909
1910 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1911 struct notes_cache *c = xmalloc(sizeof(*c));
1912 struct strbuf name = STRBUF_INIT;
1913
1914 strbuf_addf(&name, "textconv/%s", one->driver->name);
1915 notes_cache_init(c, name.buf, one->driver->textconv);
1916 one->driver->textconv_cache = c;
1917 }
1918
1919 return one->driver;
1920 }
1921
1922 static void builtin_diff(const char *name_a,
1923 const char *name_b,
1924 struct diff_filespec *one,
1925 struct diff_filespec *two,
1926 const char *xfrm_msg,
1927 int must_show_header,
1928 struct diff_options *o,
1929 int complete_rewrite)
1930 {
1931 mmfile_t mf1, mf2;
1932 const char *lbl[2];
1933 char *a_one, *b_two;
1934 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1935 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1936 const char *a_prefix, *b_prefix;
1937 struct userdiff_driver *textconv_one = NULL;
1938 struct userdiff_driver *textconv_two = NULL;
1939 struct strbuf header = STRBUF_INIT;
1940 struct strbuf *msgbuf;
1941 char *line_prefix = "";
1942
1943 if (o->output_prefix) {
1944 msgbuf = o->output_prefix(o, o->output_prefix_data);
1945 line_prefix = msgbuf->buf;
1946 }
1947
1948 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1949 (!one->mode || S_ISGITLINK(one->mode)) &&
1950 (!two->mode || S_ISGITLINK(two->mode))) {
1951 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1952 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1953 show_submodule_summary(o->file, one ? one->path : two->path,
1954 one->sha1, two->sha1, two->dirty_submodule,
1955 del, add, reset);
1956 return;
1957 }
1958
1959 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1960 textconv_one = get_textconv(one);
1961 textconv_two = get_textconv(two);
1962 }
1963
1964 diff_set_mnemonic_prefix(o, "a/", "b/");
1965 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1966 a_prefix = o->b_prefix;
1967 b_prefix = o->a_prefix;
1968 } else {
1969 a_prefix = o->a_prefix;
1970 b_prefix = o->b_prefix;
1971 }
1972
1973 /* Never use a non-valid filename anywhere if at all possible */
1974 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1975 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1976
1977 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1978 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1979 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1980 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1981 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1982 if (lbl[0][0] == '/') {
1983 /* /dev/null */
1984 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1985 if (xfrm_msg)
1986 strbuf_addstr(&header, xfrm_msg);
1987 must_show_header = 1;
1988 }
1989 else if (lbl[1][0] == '/') {
1990 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1991 if (xfrm_msg)
1992 strbuf_addstr(&header, xfrm_msg);
1993 must_show_header = 1;
1994 }
1995 else {
1996 if (one->mode != two->mode) {
1997 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1998 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1999 must_show_header = 1;
2000 }
2001 if (xfrm_msg)
2002 strbuf_addstr(&header, xfrm_msg);
2003
2004 /*
2005 * we do not run diff between different kind
2006 * of objects.
2007 */
2008 if ((one->mode ^ two->mode) & S_IFMT)
2009 goto free_ab_and_return;
2010 if (complete_rewrite &&
2011 (textconv_one || !diff_filespec_is_binary(one)) &&
2012 (textconv_two || !diff_filespec_is_binary(two))) {
2013 fprintf(o->file, "%s", header.buf);
2014 strbuf_reset(&header);
2015 emit_rewrite_diff(name_a, name_b, one, two,
2016 textconv_one, textconv_two, o);
2017 o->found_changes = 1;
2018 goto free_ab_and_return;
2019 }
2020 }
2021
2022 if (o->irreversible_delete && lbl[1][0] == '/') {
2023 fprintf(o->file, "%s", header.buf);
2024 strbuf_reset(&header);
2025 goto free_ab_and_return;
2026 } else if (!DIFF_OPT_TST(o, TEXT) &&
2027 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2028 (!textconv_two && diff_filespec_is_binary(two)) )) {
2029 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2030 die("unable to read files to diff");
2031 /* Quite common confusing case */
2032 if (mf1.size == mf2.size &&
2033 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2034 if (must_show_header)
2035 fprintf(o->file, "%s", header.buf);
2036 goto free_ab_and_return;
2037 }
2038 fprintf(o->file, "%s", header.buf);
2039 strbuf_reset(&header);
2040 if (DIFF_OPT_TST(o, BINARY))
2041 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2042 else
2043 fprintf(o->file, "%sBinary files %s and %s differ\n",
2044 line_prefix, lbl[0], lbl[1]);
2045 o->found_changes = 1;
2046 } else {
2047 /* Crazy xdl interfaces.. */
2048 const char *diffopts = getenv("GIT_DIFF_OPTS");
2049 xpparam_t xpp;
2050 xdemitconf_t xecfg;
2051 struct emit_callback ecbdata;
2052 const struct userdiff_funcname *pe;
2053
2054 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2055 fprintf(o->file, "%s", header.buf);
2056 strbuf_reset(&header);
2057 }
2058
2059 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2060 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2061
2062 pe = diff_funcname_pattern(one);
2063 if (!pe)
2064 pe = diff_funcname_pattern(two);
2065
2066 memset(&xpp, 0, sizeof(xpp));
2067 memset(&xecfg, 0, sizeof(xecfg));
2068 memset(&ecbdata, 0, sizeof(ecbdata));
2069 ecbdata.label_path = lbl;
2070 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2071 ecbdata.found_changesp = &o->found_changes;
2072 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2073 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2074 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2075 ecbdata.opt = o;
2076 ecbdata.header = header.len ? &header : NULL;
2077 xpp.flags = o->xdl_opts;
2078 xecfg.ctxlen = o->context;
2079 xecfg.interhunkctxlen = o->interhunkcontext;
2080 xecfg.flags = XDL_EMIT_FUNCNAMES;
2081 if (pe)
2082 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2083 if (!diffopts)
2084 ;
2085 else if (!prefixcmp(diffopts, "--unified="))
2086 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2087 else if (!prefixcmp(diffopts, "-u"))
2088 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2089 if (o->word_diff) {
2090 int i;
2091
2092 ecbdata.diff_words =
2093 xcalloc(1, sizeof(struct diff_words_data));
2094 ecbdata.diff_words->type = o->word_diff;
2095 ecbdata.diff_words->opt = o;
2096 if (!o->word_regex)
2097 o->word_regex = userdiff_word_regex(one);
2098 if (!o->word_regex)
2099 o->word_regex = userdiff_word_regex(two);
2100 if (!o->word_regex)
2101 o->word_regex = diff_word_regex_cfg;
2102 if (o->word_regex) {
2103 ecbdata.diff_words->word_regex = (regex_t *)
2104 xmalloc(sizeof(regex_t));
2105 if (regcomp(ecbdata.diff_words->word_regex,
2106 o->word_regex,
2107 REG_EXTENDED | REG_NEWLINE))
2108 die ("Invalid regular expression: %s",
2109 o->word_regex);
2110 }
2111 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2112 if (o->word_diff == diff_words_styles[i].type) {
2113 ecbdata.diff_words->style =
2114 &diff_words_styles[i];
2115 break;
2116 }
2117 }
2118 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2119 struct diff_words_style *st = ecbdata.diff_words->style;
2120 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2121 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2122 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2123 }
2124 }
2125 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2126 &xpp, &xecfg);
2127 if (o->word_diff)
2128 free_diff_words_data(&ecbdata);
2129 if (textconv_one)
2130 free(mf1.ptr);
2131 if (textconv_two)
2132 free(mf2.ptr);
2133 xdiff_clear_find_func(&xecfg);
2134 }
2135
2136 free_ab_and_return:
2137 strbuf_release(&header);
2138 diff_free_filespec_data(one);
2139 diff_free_filespec_data(two);
2140 free(a_one);
2141 free(b_two);
2142 return;
2143 }
2144
2145 static void builtin_diffstat(const char *name_a, const char *name_b,
2146 struct diff_filespec *one,
2147 struct diff_filespec *two,
2148 struct diffstat_t *diffstat,
2149 struct diff_options *o,
2150 int complete_rewrite)
2151 {
2152 mmfile_t mf1, mf2;
2153 struct diffstat_file *data;
2154
2155 data = diffstat_add(diffstat, name_a, name_b);
2156
2157 if (!one || !two) {
2158 data->is_unmerged = 1;
2159 return;
2160 }
2161
2162 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2163 data->is_binary = 1;
2164 data->added = diff_filespec_size(two);
2165 data->deleted = diff_filespec_size(one);
2166 }
2167
2168 else if (complete_rewrite) {
2169 diff_populate_filespec(one, 0);
2170 diff_populate_filespec(two, 0);
2171 data->deleted = count_lines(one->data, one->size);
2172 data->added = count_lines(two->data, two->size);
2173 }
2174
2175 else {
2176 /* Crazy xdl interfaces.. */
2177 xpparam_t xpp;
2178 xdemitconf_t xecfg;
2179
2180 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2181 die("unable to read files to diff");
2182
2183 memset(&xpp, 0, sizeof(xpp));
2184 memset(&xecfg, 0, sizeof(xecfg));
2185 xpp.flags = o->xdl_opts;
2186 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2187 &xpp, &xecfg);
2188 }
2189
2190 diff_free_filespec_data(one);
2191 diff_free_filespec_data(two);
2192 }
2193
2194 static void builtin_checkdiff(const char *name_a, const char *name_b,
2195 const char *attr_path,
2196 struct diff_filespec *one,
2197 struct diff_filespec *two,
2198 struct diff_options *o)
2199 {
2200 mmfile_t mf1, mf2;
2201 struct checkdiff_t data;
2202
2203 if (!two)
2204 return;
2205
2206 memset(&data, 0, sizeof(data));
2207 data.filename = name_b ? name_b : name_a;
2208 data.lineno = 0;
2209 data.o = o;
2210 data.ws_rule = whitespace_rule(attr_path);
2211 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2212
2213 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2214 die("unable to read files to diff");
2215
2216 /*
2217 * All the other codepaths check both sides, but not checking
2218 * the "old" side here is deliberate. We are checking the newly
2219 * introduced changes, and as long as the "new" side is text, we
2220 * can and should check what it introduces.
2221 */
2222 if (diff_filespec_is_binary(two))
2223 goto free_and_return;
2224 else {
2225 /* Crazy xdl interfaces.. */
2226 xpparam_t xpp;
2227 xdemitconf_t xecfg;
2228
2229 memset(&xpp, 0, sizeof(xpp));
2230 memset(&xecfg, 0, sizeof(xecfg));
2231 xecfg.ctxlen = 1; /* at least one context line */
2232 xpp.flags = 0;
2233 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2234 &xpp, &xecfg);
2235
2236 if (data.ws_rule & WS_BLANK_AT_EOF) {
2237 struct emit_callback ecbdata;
2238 int blank_at_eof;
2239
2240 ecbdata.ws_rule = data.ws_rule;
2241 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2242 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2243
2244 if (blank_at_eof) {
2245 static char *err;
2246 if (!err)
2247 err = whitespace_error_string(WS_BLANK_AT_EOF);
2248 fprintf(o->file, "%s:%d: %s.\n",
2249 data.filename, blank_at_eof, err);
2250 data.status = 1; /* report errors */
2251 }
2252 }
2253 }
2254 free_and_return:
2255 diff_free_filespec_data(one);
2256 diff_free_filespec_data(two);
2257 if (data.status)
2258 DIFF_OPT_SET(o, CHECK_FAILED);
2259 }
2260
2261 struct diff_filespec *alloc_filespec(const char *path)
2262 {
2263 int namelen = strlen(path);
2264 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2265
2266 memset(spec, 0, sizeof(*spec));
2267 spec->path = (char *)(spec + 1);
2268 memcpy(spec->path, path, namelen+1);
2269 spec->count = 1;
2270 spec->is_binary = -1;
2271 return spec;
2272 }
2273
2274 void free_filespec(struct diff_filespec *spec)
2275 {
2276 if (!--spec->count) {
2277 diff_free_filespec_data(spec);
2278 free(spec);
2279 }
2280 }
2281
2282 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2283 unsigned short mode)
2284 {
2285 if (mode) {
2286 spec->mode = canon_mode(mode);
2287 hashcpy(spec->sha1, sha1);
2288 spec->sha1_valid = !is_null_sha1(sha1);
2289 }
2290 }
2291
2292 /*
2293 * Given a name and sha1 pair, if the index tells us the file in
2294 * the work tree has that object contents, return true, so that
2295 * prepare_temp_file() does not have to inflate and extract.
2296 */
2297 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2298 {
2299 struct cache_entry *ce;
2300 struct stat st;
2301 int pos, len;
2302
2303 /*
2304 * We do not read the cache ourselves here, because the
2305 * benchmark with my previous version that always reads cache
2306 * shows that it makes things worse for diff-tree comparing
2307 * two linux-2.6 kernel trees in an already checked out work
2308 * tree. This is because most diff-tree comparisons deal with
2309 * only a small number of files, while reading the cache is
2310 * expensive for a large project, and its cost outweighs the
2311 * savings we get by not inflating the object to a temporary
2312 * file. Practically, this code only helps when we are used
2313 * by diff-cache --cached, which does read the cache before
2314 * calling us.
2315 */
2316 if (!active_cache)
2317 return 0;
2318
2319 /* We want to avoid the working directory if our caller
2320 * doesn't need the data in a normal file, this system
2321 * is rather slow with its stat/open/mmap/close syscalls,
2322 * and the object is contained in a pack file. The pack
2323 * is probably already open and will be faster to obtain
2324 * the data through than the working directory. Loose
2325 * objects however would tend to be slower as they need
2326 * to be individually opened and inflated.
2327 */
2328 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2329 return 0;
2330
2331 len = strlen(name);
2332 pos = cache_name_pos(name, len);
2333 if (pos < 0)
2334 return 0;
2335 ce = active_cache[pos];
2336
2337 /*
2338 * This is not the sha1 we are looking for, or
2339 * unreusable because it is not a regular file.
2340 */
2341 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2342 return 0;
2343
2344 /*
2345 * If ce is marked as "assume unchanged", there is no
2346 * guarantee that work tree matches what we are looking for.
2347 */
2348 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2349 return 0;
2350
2351 /*
2352 * If ce matches the file in the work tree, we can reuse it.
2353 */
2354 if (ce_uptodate(ce) ||
2355 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2356 return 1;
2357
2358 return 0;
2359 }
2360
2361 static int populate_from_stdin(struct diff_filespec *s)
2362 {
2363 struct strbuf buf = STRBUF_INIT;
2364 size_t size = 0;
2365
2366 if (strbuf_read(&buf, 0, 0) < 0)
2367 return error("error while reading from stdin %s",
2368 strerror(errno));
2369
2370 s->should_munmap = 0;
2371 s->data = strbuf_detach(&buf, &size);
2372 s->size = size;
2373 s->should_free = 1;
2374 return 0;
2375 }
2376
2377 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2378 {
2379 int len;
2380 char *data = xmalloc(100), *dirty = "";
2381
2382 /* Are we looking at the work tree? */
2383 if (s->dirty_submodule)
2384 dirty = "-dirty";
2385
2386 len = snprintf(data, 100,
2387 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2388 s->data = data;
2389 s->size = len;
2390 s->should_free = 1;
2391 if (size_only) {
2392 s->data = NULL;
2393 free(data);
2394 }
2395 return 0;
2396 }
2397
2398 /*
2399 * While doing rename detection and pickaxe operation, we may need to
2400 * grab the data for the blob (or file) for our own in-core comparison.
2401 * diff_filespec has data and size fields for this purpose.
2402 */
2403 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2404 {
2405 int err = 0;
2406 if (!DIFF_FILE_VALID(s))
2407 die("internal error: asking to populate invalid file.");
2408 if (S_ISDIR(s->mode))
2409 return -1;
2410
2411 if (s->data)
2412 return 0;
2413
2414 if (size_only && 0 < s->size)
2415 return 0;
2416
2417 if (S_ISGITLINK(s->mode))
2418 return diff_populate_gitlink(s, size_only);
2419
2420 if (!s->sha1_valid ||
2421 reuse_worktree_file(s->path, s->sha1, 0)) {
2422 struct strbuf buf = STRBUF_INIT;
2423 struct stat st;
2424 int fd;
2425
2426 if (!strcmp(s->path, "-"))
2427 return populate_from_stdin(s);
2428
2429 if (lstat(s->path, &st) < 0) {
2430 if (errno == ENOENT) {
2431 err_empty:
2432 err = -1;
2433 empty:
2434 s->data = (char *)"";
2435 s->size = 0;
2436 return err;
2437 }
2438 }
2439 s->size = xsize_t(st.st_size);
2440 if (!s->size)
2441 goto empty;
2442 if (S_ISLNK(st.st_mode)) {
2443 struct strbuf sb = STRBUF_INIT;
2444
2445 if (strbuf_readlink(&sb, s->path, s->size))
2446 goto err_empty;
2447 s->size = sb.len;
2448 s->data = strbuf_detach(&sb, NULL);
2449 s->should_free = 1;
2450 return 0;
2451 }
2452 if (size_only)
2453 return 0;
2454 fd = open(s->path, O_RDONLY);
2455 if (fd < 0)
2456 goto err_empty;
2457 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2458 close(fd);
2459 s->should_munmap = 1;
2460
2461 /*
2462 * Convert from working tree format to canonical git format
2463 */
2464 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2465 size_t size = 0;
2466 munmap(s->data, s->size);
2467 s->should_munmap = 0;
2468 s->data = strbuf_detach(&buf, &size);
2469 s->size = size;
2470 s->should_free = 1;
2471 }
2472 }
2473 else {
2474 enum object_type type;
2475 if (size_only) {
2476 type = sha1_object_info(s->sha1, &s->size);
2477 if (type < 0)
2478 die("unable to read %s", sha1_to_hex(s->sha1));
2479 } else {
2480 s->data = read_sha1_file(s->sha1, &type, &s->size);
2481 if (!s->data)
2482 die("unable to read %s", sha1_to_hex(s->sha1));
2483 s->should_free = 1;
2484 }
2485 }
2486 return 0;
2487 }
2488
2489 void diff_free_filespec_blob(struct diff_filespec *s)
2490 {
2491 if (s->should_free)
2492 free(s->data);
2493 else if (s->should_munmap)
2494 munmap(s->data, s->size);
2495
2496 if (s->should_free || s->should_munmap) {
2497 s->should_free = s->should_munmap = 0;
2498 s->data = NULL;
2499 }
2500 }
2501
2502 void diff_free_filespec_data(struct diff_filespec *s)
2503 {
2504 diff_free_filespec_blob(s);
2505 free(s->cnt_data);
2506 s->cnt_data = NULL;
2507 }
2508
2509 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2510 void *blob,
2511 unsigned long size,
2512 const unsigned char *sha1,
2513 int mode)
2514 {
2515 int fd;
2516 struct strbuf buf = STRBUF_INIT;
2517 struct strbuf template = STRBUF_INIT;
2518 char *path_dup = xstrdup(path);
2519 const char *base = basename(path_dup);
2520
2521 /* Generate "XXXXXX_basename.ext" */
2522 strbuf_addstr(&template, "XXXXXX_");
2523 strbuf_addstr(&template, base);
2524
2525 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2526 strlen(base) + 1);
2527 if (fd < 0)
2528 die_errno("unable to create temp-file");
2529 if (convert_to_working_tree(path,
2530 (const char *)blob, (size_t)size, &buf)) {
2531 blob = buf.buf;
2532 size = buf.len;
2533 }
2534 if (write_in_full(fd, blob, size) != size)
2535 die_errno("unable to write temp-file");
2536 close(fd);
2537 temp->name = temp->tmp_path;
2538 strcpy(temp->hex, sha1_to_hex(sha1));
2539 temp->hex[40] = 0;
2540 sprintf(temp->mode, "%06o", mode);
2541 strbuf_release(&buf);
2542 strbuf_release(&template);
2543 free(path_dup);
2544 }
2545
2546 static struct diff_tempfile *prepare_temp_file(const char *name,
2547 struct diff_filespec *one)
2548 {
2549 struct diff_tempfile *temp = claim_diff_tempfile();
2550
2551 if (!DIFF_FILE_VALID(one)) {
2552 not_a_valid_file:
2553 /* A '-' entry produces this for file-2, and
2554 * a '+' entry produces this for file-1.
2555 */
2556 temp->name = "/dev/null";
2557 strcpy(temp->hex, ".");
2558 strcpy(temp->mode, ".");
2559 return temp;
2560 }
2561
2562 if (!remove_tempfile_installed) {
2563 atexit(remove_tempfile);
2564 sigchain_push_common(remove_tempfile_on_signal);
2565 remove_tempfile_installed = 1;
2566 }
2567
2568 if (!one->sha1_valid ||
2569 reuse_worktree_file(name, one->sha1, 1)) {
2570 struct stat st;
2571 if (lstat(name, &st) < 0) {
2572 if (errno == ENOENT)
2573 goto not_a_valid_file;
2574 die_errno("stat(%s)", name);
2575 }
2576 if (S_ISLNK(st.st_mode)) {
2577 struct strbuf sb = STRBUF_INIT;
2578 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2579 die_errno("readlink(%s)", name);
2580 prep_temp_blob(name, temp, sb.buf, sb.len,
2581 (one->sha1_valid ?
2582 one->sha1 : null_sha1),
2583 (one->sha1_valid ?
2584 one->mode : S_IFLNK));
2585 strbuf_release(&sb);
2586 }
2587 else {
2588 /* we can borrow from the file in the work tree */
2589 temp->name = name;
2590 if (!one->sha1_valid)
2591 strcpy(temp->hex, sha1_to_hex(null_sha1));
2592 else
2593 strcpy(temp->hex, sha1_to_hex(one->sha1));
2594 /* Even though we may sometimes borrow the
2595 * contents from the work tree, we always want
2596 * one->mode. mode is trustworthy even when
2597 * !(one->sha1_valid), as long as
2598 * DIFF_FILE_VALID(one).
2599 */
2600 sprintf(temp->mode, "%06o", one->mode);
2601 }
2602 return temp;
2603 }
2604 else {
2605 if (diff_populate_filespec(one, 0))
2606 die("cannot read data blob for %s", one->path);
2607 prep_temp_blob(name, temp, one->data, one->size,
2608 one->sha1, one->mode);
2609 }
2610 return temp;
2611 }
2612
2613 /* An external diff command takes:
2614 *
2615 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2616 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2617 *
2618 */
2619 static void run_external_diff(const char *pgm,
2620 const char *name,
2621 const char *other,
2622 struct diff_filespec *one,
2623 struct diff_filespec *two,
2624 const char *xfrm_msg,
2625 int complete_rewrite)
2626 {
2627 const char *spawn_arg[10];
2628 int retval;
2629 const char **arg = &spawn_arg[0];
2630
2631 if (one && two) {
2632 struct diff_tempfile *temp_one, *temp_two;
2633 const char *othername = (other ? other : name);
2634 temp_one = prepare_temp_file(name, one);
2635 temp_two = prepare_temp_file(othername, two);
2636 *arg++ = pgm;
2637 *arg++ = name;
2638 *arg++ = temp_one->name;
2639 *arg++ = temp_one->hex;
2640 *arg++ = temp_one->mode;
2641 *arg++ = temp_two->name;
2642 *arg++ = temp_two->hex;
2643 *arg++ = temp_two->mode;
2644 if (other) {
2645 *arg++ = other;
2646 *arg++ = xfrm_msg;
2647 }
2648 } else {
2649 *arg++ = pgm;
2650 *arg++ = name;
2651 }
2652 *arg = NULL;
2653 fflush(NULL);
2654 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2655 remove_tempfile();
2656 if (retval) {
2657 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2658 exit(1);
2659 }
2660 }
2661
2662 static int similarity_index(struct diff_filepair *p)
2663 {
2664 return p->score * 100 / MAX_SCORE;
2665 }
2666
2667 static void fill_metainfo(struct strbuf *msg,
2668 const char *name,
2669 const char *other,
2670 struct diff_filespec *one,
2671 struct diff_filespec *two,
2672 struct diff_options *o,
2673 struct diff_filepair *p,
2674 int *must_show_header,
2675 int use_color)
2676 {
2677 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2678 const char *reset = diff_get_color(use_color, DIFF_RESET);
2679 struct strbuf *msgbuf;
2680 char *line_prefix = "";
2681
2682 *must_show_header = 1;
2683 if (o->output_prefix) {
2684 msgbuf = o->output_prefix(o, o->output_prefix_data);
2685 line_prefix = msgbuf->buf;
2686 }
2687 strbuf_init(msg, PATH_MAX * 2 + 300);
2688 switch (p->status) {
2689 case DIFF_STATUS_COPIED:
2690 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2691 line_prefix, set, similarity_index(p));
2692 strbuf_addf(msg, "%s\n%s%scopy from ",
2693 reset, line_prefix, set);
2694 quote_c_style(name, msg, NULL, 0);
2695 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2696 quote_c_style(other, msg, NULL, 0);
2697 strbuf_addf(msg, "%s\n", reset);
2698 break;
2699 case DIFF_STATUS_RENAMED:
2700 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2701 line_prefix, set, similarity_index(p));
2702 strbuf_addf(msg, "%s\n%s%srename from ",
2703 reset, line_prefix, set);
2704 quote_c_style(name, msg, NULL, 0);
2705 strbuf_addf(msg, "%s\n%s%srename to ",
2706 reset, line_prefix, set);
2707 quote_c_style(other, msg, NULL, 0);
2708 strbuf_addf(msg, "%s\n", reset);
2709 break;
2710 case DIFF_STATUS_MODIFIED:
2711 if (p->score) {
2712 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2713 line_prefix,
2714 set, similarity_index(p), reset);
2715 break;
2716 }
2717 /* fallthru */
2718 default:
2719 *must_show_header = 0;
2720 }
2721 if (one && two && hashcmp(one->sha1, two->sha1)) {
2722 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2723
2724 if (DIFF_OPT_TST(o, BINARY)) {
2725 mmfile_t mf;
2726 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2727 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2728 abbrev = 40;
2729 }
2730 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2731 find_unique_abbrev(one->sha1, abbrev));
2732 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2733 if (one->mode == two->mode)
2734 strbuf_addf(msg, " %06o", one->mode);
2735 strbuf_addf(msg, "%s\n", reset);
2736 }
2737 }
2738
2739 static void run_diff_cmd(const char *pgm,
2740 const char *name,
2741 const char *other,
2742 const char *attr_path,
2743 struct diff_filespec *one,
2744 struct diff_filespec *two,
2745 struct strbuf *msg,
2746 struct diff_options *o,
2747 struct diff_filepair *p)
2748 {
2749 const char *xfrm_msg = NULL;
2750 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2751 int must_show_header = 0;
2752
2753 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2754 pgm = NULL;
2755 else {
2756 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2757 if (drv && drv->external)
2758 pgm = drv->external;
2759 }
2760
2761 if (msg) {
2762 /*
2763 * don't use colors when the header is intended for an
2764 * external diff driver
2765 */
2766 fill_metainfo(msg, name, other, one, two, o, p,
2767 &must_show_header,
2768 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2769 xfrm_msg = msg->len ? msg->buf : NULL;
2770 }
2771
2772 if (pgm) {
2773 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2774 complete_rewrite);
2775 return;
2776 }
2777 if (one && two)
2778 builtin_diff(name, other ? other : name,
2779 one, two, xfrm_msg, must_show_header,
2780 o, complete_rewrite);
2781 else
2782 fprintf(o->file, "* Unmerged path %s\n", name);
2783 }
2784
2785 static void diff_fill_sha1_info(struct diff_filespec *one)
2786 {
2787 if (DIFF_FILE_VALID(one)) {
2788 if (!one->sha1_valid) {
2789 struct stat st;
2790 if (!strcmp(one->path, "-")) {
2791 hashcpy(one->sha1, null_sha1);
2792 return;
2793 }
2794 if (lstat(one->path, &st) < 0)
2795 die_errno("stat '%s'", one->path);
2796 if (index_path(one->sha1, one->path, &st, 0))
2797 die("cannot hash %s", one->path);
2798 }
2799 }
2800 else
2801 hashclr(one->sha1);
2802 }
2803
2804 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2805 {
2806 /* Strip the prefix but do not molest /dev/null and absolute paths */
2807 if (*namep && **namep != '/') {
2808 *namep += prefix_length;
2809 if (**namep == '/')
2810 ++*namep;
2811 }
2812 if (*otherp && **otherp != '/') {
2813 *otherp += prefix_length;
2814 if (**otherp == '/')
2815 ++*otherp;
2816 }
2817 }
2818
2819 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2820 {
2821 const char *pgm = external_diff();
2822 struct strbuf msg;
2823 struct diff_filespec *one = p->one;
2824 struct diff_filespec *two = p->two;
2825 const char *name;
2826 const char *other;
2827 const char *attr_path;
2828
2829 name = p->one->path;
2830 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2831 attr_path = name;
2832 if (o->prefix_length)
2833 strip_prefix(o->prefix_length, &name, &other);
2834
2835 if (DIFF_PAIR_UNMERGED(p)) {
2836 run_diff_cmd(pgm, name, NULL, attr_path,
2837 NULL, NULL, NULL, o, p);
2838 return;
2839 }
2840
2841 diff_fill_sha1_info(one);
2842 diff_fill_sha1_info(two);
2843
2844 if (!pgm &&
2845 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2846 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2847 /*
2848 * a filepair that changes between file and symlink
2849 * needs to be split into deletion and creation.
2850 */
2851 struct diff_filespec *null = alloc_filespec(two->path);
2852 run_diff_cmd(NULL, name, other, attr_path,
2853 one, null, &msg, o, p);
2854 free(null);
2855 strbuf_release(&msg);
2856
2857 null = alloc_filespec(one->path);
2858 run_diff_cmd(NULL, name, other, attr_path,
2859 null, two, &msg, o, p);
2860 free(null);
2861 }
2862 else
2863 run_diff_cmd(pgm, name, other, attr_path,
2864 one, two, &msg, o, p);
2865
2866 strbuf_release(&msg);
2867 }
2868
2869 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2870 struct diffstat_t *diffstat)
2871 {
2872 const char *name;
2873 const char *other;
2874 int complete_rewrite = 0;
2875
2876 if (DIFF_PAIR_UNMERGED(p)) {
2877 /* unmerged */
2878 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2879 return;
2880 }
2881
2882 name = p->one->path;
2883 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2884
2885 if (o->prefix_length)
2886 strip_prefix(o->prefix_length, &name, &other);
2887
2888 diff_fill_sha1_info(p->one);
2889 diff_fill_sha1_info(p->two);
2890
2891 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2892 complete_rewrite = 1;
2893 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2894 }
2895
2896 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2897 {
2898 const char *name;
2899 const char *other;
2900 const char *attr_path;
2901
2902 if (DIFF_PAIR_UNMERGED(p)) {
2903 /* unmerged */
2904 return;
2905 }
2906
2907 name = p->one->path;
2908 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2909 attr_path = other ? other : name;
2910
2911 if (o->prefix_length)
2912 strip_prefix(o->prefix_length, &name, &other);
2913
2914 diff_fill_sha1_info(p->one);
2915 diff_fill_sha1_info(p->two);
2916
2917 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2918 }
2919
2920 void diff_setup(struct diff_options *options)
2921 {
2922 memcpy(options, &default_diff_options, sizeof(*options));
2923
2924 options->file = stdout;
2925
2926 options->line_termination = '\n';
2927 options->break_opt = -1;
2928 options->rename_limit = -1;
2929 options->dirstat_percent = 3;
2930 options->context = 3;
2931
2932 options->change = diff_change;
2933 options->add_remove = diff_addremove;
2934 if (diff_use_color_default > 0)
2935 DIFF_OPT_SET(options, COLOR_DIFF);
2936 options->detect_rename = diff_detect_rename_default;
2937
2938 if (diff_no_prefix) {
2939 options->a_prefix = options->b_prefix = "";
2940 } else if (!diff_mnemonic_prefix) {
2941 options->a_prefix = "a/";
2942 options->b_prefix = "b/";
2943 }
2944 }
2945
2946 int diff_setup_done(struct diff_options *options)
2947 {
2948 int count = 0;
2949
2950 if (options->output_format & DIFF_FORMAT_NAME)
2951 count++;
2952 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2953 count++;
2954 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2955 count++;
2956 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2957 count++;
2958 if (count > 1)
2959 die("--name-only, --name-status, --check and -s are mutually exclusive");
2960
2961 /*
2962 * Most of the time we can say "there are changes"
2963 * only by checking if there are changed paths, but
2964 * --ignore-whitespace* options force us to look
2965 * inside contents.
2966 */
2967
2968 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2969 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2970 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2971 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2972 else
2973 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2974
2975 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2976 options->detect_rename = DIFF_DETECT_COPY;
2977
2978 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2979 options->prefix = NULL;
2980 if (options->prefix)
2981 options->prefix_length = strlen(options->prefix);
2982 else
2983 options->prefix_length = 0;
2984
2985 if (options->output_format & (DIFF_FORMAT_NAME |
2986 DIFF_FORMAT_NAME_STATUS |
2987 DIFF_FORMAT_CHECKDIFF |
2988 DIFF_FORMAT_NO_OUTPUT))
2989 options->output_format &= ~(DIFF_FORMAT_RAW |
2990 DIFF_FORMAT_NUMSTAT |
2991 DIFF_FORMAT_DIFFSTAT |
2992 DIFF_FORMAT_SHORTSTAT |
2993 DIFF_FORMAT_DIRSTAT |
2994 DIFF_FORMAT_SUMMARY |
2995 DIFF_FORMAT_PATCH);
2996
2997 /*
2998 * These cases always need recursive; we do not drop caller-supplied
2999 * recursive bits for other formats here.
3000 */
3001 if (options->output_format & (DIFF_FORMAT_PATCH |
3002 DIFF_FORMAT_NUMSTAT |
3003 DIFF_FORMAT_DIFFSTAT |
3004 DIFF_FORMAT_SHORTSTAT |
3005 DIFF_FORMAT_DIRSTAT |
3006 DIFF_FORMAT_SUMMARY |
3007 DIFF_FORMAT_CHECKDIFF))
3008 DIFF_OPT_SET(options, RECURSIVE);
3009 /*
3010 * Also pickaxe would not work very well if you do not say recursive
3011 */
3012 if (options->pickaxe)
3013 DIFF_OPT_SET(options, RECURSIVE);
3014 /*
3015 * When patches are generated, submodules diffed against the work tree
3016 * must be checked for dirtiness too so it can be shown in the output
3017 */
3018 if (options->output_format & DIFF_FORMAT_PATCH)
3019 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3020
3021 if (options->detect_rename && options->rename_limit < 0)
3022 options->rename_limit = diff_rename_limit_default;
3023 if (options->setup & DIFF_SETUP_USE_CACHE) {
3024 if (!active_cache)
3025 /* read-cache does not die even when it fails
3026 * so it is safe for us to do this here. Also
3027 * it does not smudge active_cache or active_nr
3028 * when it fails, so we do not have to worry about
3029 * cleaning it up ourselves either.
3030 */
3031 read_cache();
3032 }
3033 if (options->abbrev <= 0 || 40 < options->abbrev)
3034 options->abbrev = 40; /* full */
3035
3036 /*
3037 * It does not make sense to show the first hit we happened
3038 * to have found. It does not make sense not to return with
3039 * exit code in such a case either.
3040 */
3041 if (DIFF_OPT_TST(options, QUICK)) {
3042 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3043 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3044 }
3045
3046 return 0;
3047 }
3048
3049 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3050 {
3051 char c, *eq;
3052 int len;
3053
3054 if (*arg != '-')
3055 return 0;
3056 c = *++arg;
3057 if (!c)
3058 return 0;
3059 if (c == arg_short) {
3060 c = *++arg;
3061 if (!c)
3062 return 1;
3063 if (val && isdigit(c)) {
3064 char *end;
3065 int n = strtoul(arg, &end, 10);
3066 if (*end)
3067 return 0;
3068 *val = n;
3069 return 1;
3070 }
3071 return 0;
3072 }
3073 if (c != '-')
3074 return 0;
3075 arg++;
3076 eq = strchr(arg, '=');
3077 if (eq)
3078 len = eq - arg;
3079 else
3080 len = strlen(arg);
3081 if (!len || strncmp(arg, arg_long, len))
3082 return 0;
3083 if (eq) {
3084 int n;
3085 char *end;
3086 if (!isdigit(*++eq))
3087 return 0;
3088 n = strtoul(eq, &end, 10);
3089 if (*end)
3090 return 0;
3091 *val = n;
3092 }
3093 return 1;
3094 }
3095
3096 static int diff_scoreopt_parse(const char *opt);
3097
3098 static inline int short_opt(char opt, const char **argv,
3099 const char **optarg)
3100 {
3101 const char *arg = argv[0];
3102 if (arg[0] != '-' || arg[1] != opt)
3103 return 0;
3104 if (arg[2] != '\0') {
3105 *optarg = arg + 2;
3106 return 1;
3107 }
3108 if (!argv[1])
3109 die("Option '%c' requires a value", opt);
3110 *optarg = argv[1];
3111 return 2;
3112 }
3113
3114 int parse_long_opt(const char *opt, const char **argv,
3115 const char **optarg)
3116 {
3117 const char *arg = argv[0];
3118 if (arg[0] != '-' || arg[1] != '-')
3119 return 0;
3120 arg += strlen("--");
3121 if (prefixcmp(arg, opt))
3122 return 0;
3123 arg += strlen(opt);
3124 if (*arg == '=') { /* sticked form: --option=value */
3125 *optarg = arg + 1;
3126 return 1;
3127 }
3128 if (*arg != '\0')
3129 return 0;
3130 /* separate form: --option value */
3131 if (!argv[1])
3132 die("Option '--%s' requires a value", opt);
3133 *optarg = argv[1];
3134 return 2;
3135 }
3136
3137 static int stat_opt(struct diff_options *options, const char **av)
3138 {
3139 const char *arg = av[0];
3140 char *end;
3141 int width = options->stat_width;
3142 int name_width = options->stat_name_width;
3143 int argcount = 1;
3144
3145 arg += strlen("--stat");
3146 end = (char *)arg;
3147
3148 switch (*arg) {
3149 case '-':
3150 if (!prefixcmp(arg, "-width")) {
3151 arg += strlen("-width");
3152 if (*arg == '=')
3153 width = strtoul(arg + 1, &end, 10);
3154 else if (!*arg && !av[1])
3155 die("Option '--stat-width' requires a value");
3156 else if (!*arg) {
3157 width = strtoul(av[1], &end, 10);
3158 argcount = 2;
3159 }
3160 } else if (!prefixcmp(arg, "-name-width")) {
3161 arg += strlen("-name-width");
3162 if (*arg == '=')
3163 name_width = strtoul(arg + 1, &end, 10);
3164 else if (!*arg && !av[1])
3165 die("Option '--stat-name-width' requires a value");
3166 else if (!*arg) {
3167 name_width = strtoul(av[1], &end, 10);
3168 argcount = 2;
3169 }
3170 }
3171 break;
3172 case '=':
3173 width = strtoul(arg+1, &end, 10);
3174 if (*end == ',')
3175 name_width = strtoul(end+1, &end, 10);
3176 }
3177
3178 /* Important! This checks all the error cases! */
3179 if (*end)
3180 return 0;
3181 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3182 options->stat_name_width = name_width;
3183 options->stat_width = width;
3184 return argcount;
3185 }
3186
3187 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3188 {
3189 if (parse_dirstat_params(options, params))
3190 die("Failed to parse --dirstat/-X option parameter");
3191 /*
3192 * The caller knows a dirstat-related option is given from the command
3193 * line; allow it to say "return this_function();"
3194 */
3195 options->output_format |= DIFF_FORMAT_DIRSTAT;
3196 return 1;
3197 }
3198
3199 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3200 {
3201 const char *arg = av[0];
3202 const char *optarg;
3203 int argcount;
3204
3205 /* Output format options */
3206 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3207 options->output_format |= DIFF_FORMAT_PATCH;
3208 else if (opt_arg(arg, 'U', "unified", &options->context))
3209 options->output_format |= DIFF_FORMAT_PATCH;
3210 else if (!strcmp(arg, "--raw"))
3211 options->output_format |= DIFF_FORMAT_RAW;
3212 else if (!strcmp(arg, "--patch-with-raw"))
3213 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3214 else if (!strcmp(arg, "--numstat"))
3215 options->output_format |= DIFF_FORMAT_NUMSTAT;
3216 else if (!strcmp(arg, "--shortstat"))
3217 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3218 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3219 return parse_dirstat_opt(options, "");
3220 else if (!prefixcmp(arg, "-X"))
3221 return parse_dirstat_opt(options, arg + 2);
3222 else if (!prefixcmp(arg, "--dirstat="))
3223 return parse_dirstat_opt(options, arg + 10);
3224 else if (!strcmp(arg, "--cumulative"))
3225 return parse_dirstat_opt(options, "cumulative");
3226 else if (!strcmp(arg, "--dirstat-by-file"))
3227 return parse_dirstat_opt(options, "files");
3228 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3229 parse_dirstat_opt(options, "files");
3230 return parse_dirstat_opt(options, arg + 18);
3231 }
3232 else if (!strcmp(arg, "--check"))
3233 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3234 else if (!strcmp(arg, "--summary"))
3235 options->output_format |= DIFF_FORMAT_SUMMARY;
3236 else if (!strcmp(arg, "--patch-with-stat"))
3237 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3238 else if (!strcmp(arg, "--name-only"))
3239 options->output_format |= DIFF_FORMAT_NAME;
3240 else if (!strcmp(arg, "--name-status"))
3241 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3242 else if (!strcmp(arg, "-s"))
3243 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3244 else if (!prefixcmp(arg, "--stat"))
3245 /* --stat, --stat-width, or --stat-name-width */
3246 return stat_opt(options, av);
3247
3248 /* renames options */
3249 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3250 !strcmp(arg, "--break-rewrites")) {
3251 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3252 return error("invalid argument to -B: %s", arg+2);
3253 }
3254 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3255 !strcmp(arg, "--find-renames")) {
3256 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3257 return error("invalid argument to -M: %s", arg+2);
3258 options->detect_rename = DIFF_DETECT_RENAME;
3259 }
3260 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3261 options->irreversible_delete = 1;
3262 }
3263 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3264 !strcmp(arg, "--find-copies")) {
3265 if (options->detect_rename == DIFF_DETECT_COPY)
3266 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3267 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3268 return error("invalid argument to -C: %s", arg+2);
3269 options->detect_rename = DIFF_DETECT_COPY;
3270 }
3271 else if (!strcmp(arg, "--no-renames"))
3272 options->detect_rename = 0;
3273 else if (!strcmp(arg, "--relative"))
3274 DIFF_OPT_SET(options, RELATIVE_NAME);
3275 else if (!prefixcmp(arg, "--relative=")) {
3276 DIFF_OPT_SET(options, RELATIVE_NAME);
3277 options->prefix = arg + 11;
3278 }
3279
3280 /* xdiff options */
3281 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3282 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3283 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3284 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3285 else if (!strcmp(arg, "--ignore-space-at-eol"))
3286 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3287 else if (!strcmp(arg, "--patience"))
3288 DIFF_XDL_SET(options, PATIENCE_DIFF);
3289
3290 /* flags options */
3291 else if (!strcmp(arg, "--binary")) {
3292 options->output_format |= DIFF_FORMAT_PATCH;
3293 DIFF_OPT_SET(options, BINARY);
3294 }
3295 else if (!strcmp(arg, "--full-index"))
3296 DIFF_OPT_SET(options, FULL_INDEX);
3297 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3298 DIFF_OPT_SET(options, TEXT);
3299 else if (!strcmp(arg, "-R"))
3300 DIFF_OPT_SET(options, REVERSE_DIFF);
3301 else if (!strcmp(arg, "--find-copies-harder"))
3302 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3303 else if (!strcmp(arg, "--follow"))
3304 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3305 else if (!strcmp(arg, "--color"))
3306 DIFF_OPT_SET(options, COLOR_DIFF);
3307 else if (!prefixcmp(arg, "--color=")) {
3308 int value = git_config_colorbool(NULL, arg+8, -1);
3309 if (value == 0)
3310 DIFF_OPT_CLR(options, COLOR_DIFF);
3311 else if (value > 0)
3312 DIFF_OPT_SET(options, COLOR_DIFF);
3313 else
3314 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3315 }
3316 else if (!strcmp(arg, "--no-color"))
3317 DIFF_OPT_CLR(options, COLOR_DIFF);
3318 else if (!strcmp(arg, "--color-words")) {
3319 DIFF_OPT_SET(options, COLOR_DIFF);
3320 options->word_diff = DIFF_WORDS_COLOR;
3321 }
3322 else if (!prefixcmp(arg, "--color-words=")) {
3323 DIFF_OPT_SET(options, COLOR_DIFF);
3324 options->word_diff = DIFF_WORDS_COLOR;
3325 options->word_regex = arg + 14;
3326 }
3327 else if (!strcmp(arg, "--word-diff")) {
3328 if (options->word_diff == DIFF_WORDS_NONE)
3329 options->word_diff = DIFF_WORDS_PLAIN;
3330 }
3331 else if (!prefixcmp(arg, "--word-diff=")) {
3332 const char *type = arg + 12;
3333 if (!strcmp(type, "plain"))
3334 options->word_diff = DIFF_WORDS_PLAIN;
3335 else if (!strcmp(type, "color")) {
3336 DIFF_OPT_SET(options, COLOR_DIFF);
3337 options->word_diff = DIFF_WORDS_COLOR;
3338 }
3339 else if (!strcmp(type, "porcelain"))
3340 options->word_diff = DIFF_WORDS_PORCELAIN;
3341 else if (!strcmp(type, "none"))
3342 options->word_diff = DIFF_WORDS_NONE;
3343 else
3344 die("bad --word-diff argument: %s", type);
3345 }
3346 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3347 if (options->word_diff == DIFF_WORDS_NONE)
3348 options->word_diff = DIFF_WORDS_PLAIN;
3349 options->word_regex = optarg;
3350 return argcount;
3351 }
3352 else if (!strcmp(arg, "--exit-code"))
3353 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3354 else if (!strcmp(arg, "--quiet"))
3355 DIFF_OPT_SET(options, QUICK);
3356 else if (!strcmp(arg, "--ext-diff"))
3357 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3358 else if (!strcmp(arg, "--no-ext-diff"))
3359 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3360 else if (!strcmp(arg, "--textconv"))
3361 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3362 else if (!strcmp(arg, "--no-textconv"))
3363 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3364 else if (!strcmp(arg, "--ignore-submodules")) {
3365 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3366 handle_ignore_submodules_arg(options, "all");
3367 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3368 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3369 handle_ignore_submodules_arg(options, arg + 20);
3370 } else if (!strcmp(arg, "--submodule"))
3371 DIFF_OPT_SET(options, SUBMODULE_LOG);
3372 else if (!prefixcmp(arg, "--submodule=")) {
3373 if (!strcmp(arg + 12, "log"))
3374 DIFF_OPT_SET(options, SUBMODULE_LOG);
3375 }
3376
3377 /* misc options */
3378 else if (!strcmp(arg, "-z"))
3379 options->line_termination = 0;
3380 else if ((argcount = short_opt('l', av, &optarg))) {
3381 options->rename_limit = strtoul(optarg, NULL, 10);
3382 return argcount;
3383 }
3384 else if ((argcount = short_opt('S', av, &optarg))) {
3385 options->pickaxe = optarg;
3386 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3387 return argcount;
3388 } else if ((argcount = short_opt('G', av, &optarg))) {
3389 options->pickaxe = optarg;
3390 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3391 return argcount;
3392 }
3393 else if (!strcmp(arg, "--pickaxe-all"))
3394 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3395 else if (!strcmp(arg, "--pickaxe-regex"))
3396 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3397 else if ((argcount = short_opt('O', av, &optarg))) {
3398 options->orderfile = optarg;
3399 return argcount;
3400 }
3401 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3402 options->filter = optarg;
3403 return argcount;
3404 }
3405 else if (!strcmp(arg, "--abbrev"))
3406 options->abbrev = DEFAULT_ABBREV;
3407 else if (!prefixcmp(arg, "--abbrev=")) {
3408 options->abbrev = strtoul(arg + 9, NULL, 10);
3409 if (options->abbrev < MINIMUM_ABBREV)
3410 options->abbrev = MINIMUM_ABBREV;
3411 else if (40 < options->abbrev)
3412 options->abbrev = 40;
3413 }
3414 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3415 options->a_prefix = optarg;
3416 return argcount;
3417 }
3418 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3419 options->b_prefix = optarg;
3420 return argcount;
3421 }
3422 else if (!strcmp(arg, "--no-prefix"))
3423 options->a_prefix = options->b_prefix = "";
3424 else if (opt_arg(arg, '\0', "inter-hunk-context",
3425 &options->interhunkcontext))
3426 ;
3427 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3428 options->file = fopen(optarg, "w");
3429 if (!options->file)
3430 die_errno("Could not open '%s'", optarg);
3431 options->close_file = 1;
3432 return argcount;
3433 } else
3434 return 0;
3435 return 1;
3436 }
3437
3438 int parse_rename_score(const char **cp_p)
3439 {
3440 unsigned long num, scale;
3441 int ch, dot;
3442 const char *cp = *cp_p;
3443
3444 num = 0;
3445 scale = 1;
3446 dot = 0;
3447 for (;;) {
3448 ch = *cp;
3449 if ( !dot && ch == '.' ) {
3450 scale = 1;
3451 dot = 1;
3452 } else if ( ch == '%' ) {
3453 scale = dot ? scale*100 : 100;
3454 cp++; /* % is always at the end */
3455 break;
3456 } else if ( ch >= '0' && ch <= '9' ) {
3457 if ( scale < 100000 ) {
3458 scale *= 10;
3459 num = (num*10) + (ch-'0');
3460 }
3461 } else {
3462 break;
3463 }
3464 cp++;
3465 }
3466 *cp_p = cp;
3467
3468 /* user says num divided by scale and we say internally that
3469 * is MAX_SCORE * num / scale.
3470 */
3471 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3472 }
3473
3474 static int diff_scoreopt_parse(const char *opt)
3475 {
3476 int opt1, opt2, cmd;
3477
3478 if (*opt++ != '-')
3479 return -1;
3480 cmd = *opt++;
3481 if (cmd == '-') {
3482 /* convert the long-form arguments into short-form versions */
3483 if (!prefixcmp(opt, "break-rewrites")) {
3484 opt += strlen("break-rewrites");
3485 if (*opt == 0 || *opt++ == '=')
3486 cmd = 'B';
3487 } else if (!prefixcmp(opt, "find-copies")) {
3488 opt += strlen("find-copies");
3489 if (*opt == 0 || *opt++ == '=')
3490 cmd = 'C';
3491 } else if (!prefixcmp(opt, "find-renames")) {
3492 opt += strlen("find-renames");
3493 if (*opt == 0 || *opt++ == '=')
3494 cmd = 'M';
3495 }
3496 }
3497 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3498 return -1; /* that is not a -M, -C nor -B option */
3499
3500 opt1 = parse_rename_score(&opt);
3501 if (cmd != 'B')
3502 opt2 = 0;
3503 else {
3504 if (*opt == 0)
3505 opt2 = 0;
3506 else if (*opt != '/')
3507 return -1; /* we expect -B80/99 or -B80 */
3508 else {
3509 opt++;
3510 opt2 = parse_rename_score(&opt);
3511 }
3512 }
3513 if (*opt != 0)
3514 return -1;
3515 return opt1 | (opt2 << 16);
3516 }
3517
3518 struct diff_queue_struct diff_queued_diff;
3519
3520 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3521 {
3522 if (queue->alloc <= queue->nr) {
3523 queue->alloc = alloc_nr(queue->alloc);
3524 queue->queue = xrealloc(queue->queue,
3525 sizeof(dp) * queue->alloc);
3526 }
3527 queue->queue[queue->nr++] = dp;
3528 }
3529
3530 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3531 struct diff_filespec *one,
3532 struct diff_filespec *two)
3533 {
3534 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3535 dp->one = one;
3536 dp->two = two;
3537 if (queue)
3538 diff_q(queue, dp);
3539 return dp;
3540 }
3541
3542 void diff_free_filepair(struct diff_filepair *p)
3543 {
3544 free_filespec(p->one);
3545 free_filespec(p->two);
3546 free(p);
3547 }
3548
3549 /* This is different from find_unique_abbrev() in that
3550 * it stuffs the result with dots for alignment.
3551 */
3552 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3553 {
3554 int abblen;
3555 const char *abbrev;
3556 if (len == 40)
3557 return sha1_to_hex(sha1);
3558
3559 abbrev = find_unique_abbrev(sha1, len);
3560 abblen = strlen(abbrev);
3561 if (abblen < 37) {
3562 static char hex[41];
3563 if (len < abblen && abblen <= len + 2)
3564 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3565 else
3566 sprintf(hex, "%s...", abbrev);
3567 return hex;
3568 }
3569 return sha1_to_hex(sha1);
3570 }
3571
3572 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3573 {
3574 int line_termination = opt->line_termination;
3575 int inter_name_termination = line_termination ? '\t' : '\0';
3576 if (opt->output_prefix) {
3577 struct strbuf *msg = NULL;
3578 msg = opt->output_prefix(opt, opt->output_prefix_data);
3579 fprintf(opt->file, "%s", msg->buf);
3580 }
3581
3582 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3583 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3584 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3585 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3586 }
3587 if (p->score) {
3588 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3589 inter_name_termination);
3590 } else {
3591 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3592 }
3593
3594 if (p->status == DIFF_STATUS_COPIED ||
3595 p->status == DIFF_STATUS_RENAMED) {
3596 const char *name_a, *name_b;
3597 name_a = p->one->path;
3598 name_b = p->two->path;
3599 strip_prefix(opt->prefix_length, &name_a, &name_b);
3600 write_name_quoted(name_a, opt->file, inter_name_termination);
3601 write_name_quoted(name_b, opt->file, line_termination);
3602 } else {
3603 const char *name_a, *name_b;
3604 name_a = p->one->mode ? p->one->path : p->two->path;
3605 name_b = NULL;
3606 strip_prefix(opt->prefix_length, &name_a, &name_b);
3607 write_name_quoted(name_a, opt->file, line_termination);
3608 }
3609 }
3610
3611 int diff_unmodified_pair(struct diff_filepair *p)
3612 {
3613 /* This function is written stricter than necessary to support
3614 * the currently implemented transformers, but the idea is to
3615 * let transformers to produce diff_filepairs any way they want,
3616 * and filter and clean them up here before producing the output.
3617 */
3618 struct diff_filespec *one = p->one, *two = p->two;
3619
3620 if (DIFF_PAIR_UNMERGED(p))
3621 return 0; /* unmerged is interesting */
3622
3623 /* deletion, addition, mode or type change
3624 * and rename are all interesting.
3625 */
3626 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3627 DIFF_PAIR_MODE_CHANGED(p) ||
3628 strcmp(one->path, two->path))
3629 return 0;
3630
3631 /* both are valid and point at the same path. that is, we are
3632 * dealing with a change.
3633 */
3634 if (one->sha1_valid && two->sha1_valid &&
3635 !hashcmp(one->sha1, two->sha1) &&
3636 !one->dirty_submodule && !two->dirty_submodule)
3637 return 1; /* no change */
3638 if (!one->sha1_valid && !two->sha1_valid)
3639 return 1; /* both look at the same file on the filesystem. */
3640 return 0;
3641 }
3642
3643 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3644 {
3645 if (diff_unmodified_pair(p))
3646 return;
3647
3648 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3649 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3650 return; /* no tree diffs in patch format */
3651
3652 run_diff(p, o);
3653 }
3654
3655 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3656 struct diffstat_t *diffstat)
3657 {
3658 if (diff_unmodified_pair(p))
3659 return;
3660
3661 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3662 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3663 return; /* no useful stat for tree diffs */
3664
3665 run_diffstat(p, o, diffstat);
3666 }
3667
3668 static void diff_flush_checkdiff(struct diff_filepair *p,
3669 struct diff_options *o)
3670 {
3671 if (diff_unmodified_pair(p))
3672 return;
3673
3674 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3675 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3676 return; /* nothing to check in tree diffs */
3677
3678 run_checkdiff(p, o);
3679 }
3680
3681 int diff_queue_is_empty(void)
3682 {
3683 struct diff_queue_struct *q = &diff_queued_diff;
3684 int i;
3685 for (i = 0; i < q->nr; i++)
3686 if (!diff_unmodified_pair(q->queue[i]))
3687 return 0;
3688 return 1;
3689 }
3690
3691 #if DIFF_DEBUG
3692 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3693 {
3694 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3695 x, one ? one : "",
3696 s->path,
3697 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3698 s->mode,
3699 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3700 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3701 x, one ? one : "",
3702 s->size, s->xfrm_flags);
3703 }
3704
3705 void diff_debug_filepair(const struct diff_filepair *p, int i)
3706 {
3707 diff_debug_filespec(p->one, i, "one");
3708 diff_debug_filespec(p->two, i, "two");
3709 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3710 p->score, p->status ? p->status : '?',
3711 p->one->rename_used, p->broken_pair);
3712 }
3713
3714 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3715 {
3716 int i;
3717 if (msg)
3718 fprintf(stderr, "%s\n", msg);
3719 fprintf(stderr, "q->nr = %d\n", q->nr);
3720 for (i = 0; i < q->nr; i++) {
3721 struct diff_filepair *p = q->queue[i];
3722 diff_debug_filepair(p, i);
3723 }
3724 }
3725 #endif
3726
3727 static void diff_resolve_rename_copy(void)
3728 {
3729 int i;
3730 struct diff_filepair *p;
3731 struct diff_queue_struct *q = &diff_queued_diff;
3732
3733 diff_debug_queue("resolve-rename-copy", q);
3734
3735 for (i = 0; i < q->nr; i++) {
3736 p = q->queue[i];
3737 p->status = 0; /* undecided */
3738 if (DIFF_PAIR_UNMERGED(p))
3739 p->status = DIFF_STATUS_UNMERGED;
3740 else if (!DIFF_FILE_VALID(p->one))
3741 p->status = DIFF_STATUS_ADDED;
3742 else if (!DIFF_FILE_VALID(p->two))
3743 p->status = DIFF_STATUS_DELETED;
3744 else if (DIFF_PAIR_TYPE_CHANGED(p))
3745 p->status = DIFF_STATUS_TYPE_CHANGED;
3746
3747 /* from this point on, we are dealing with a pair
3748 * whose both sides are valid and of the same type, i.e.
3749 * either in-place edit or rename/copy edit.
3750 */
3751 else if (DIFF_PAIR_RENAME(p)) {
3752 /*
3753 * A rename might have re-connected a broken
3754 * pair up, causing the pathnames to be the
3755 * same again. If so, that's not a rename at
3756 * all, just a modification..
3757 *
3758 * Otherwise, see if this source was used for
3759 * multiple renames, in which case we decrement
3760 * the count, and call it a copy.
3761 */
3762 if (!strcmp(p->one->path, p->two->path))
3763 p->status = DIFF_STATUS_MODIFIED;
3764 else if (--p->one->rename_used > 0)
3765 p->status = DIFF_STATUS_COPIED;
3766 else
3767 p->status = DIFF_STATUS_RENAMED;
3768 }
3769 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3770 p->one->mode != p->two->mode ||
3771 p->one->dirty_submodule ||
3772 p->two->dirty_submodule ||
3773 is_null_sha1(p->one->sha1))
3774 p->status = DIFF_STATUS_MODIFIED;
3775 else {
3776 /* This is a "no-change" entry and should not
3777 * happen anymore, but prepare for broken callers.
3778 */
3779 error("feeding unmodified %s to diffcore",
3780 p->one->path);
3781 p->status = DIFF_STATUS_UNKNOWN;
3782 }
3783 }
3784 diff_debug_queue("resolve-rename-copy done", q);
3785 }
3786
3787 static int check_pair_status(struct diff_filepair *p)
3788 {
3789 switch (p->status) {
3790 case DIFF_STATUS_UNKNOWN:
3791 return 0;
3792 case 0:
3793 die("internal error in diff-resolve-rename-copy");
3794 default:
3795 return 1;
3796 }
3797 }
3798
3799 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3800 {
3801 int fmt = opt->output_format;
3802
3803 if (fmt & DIFF_FORMAT_CHECKDIFF)
3804 diff_flush_checkdiff(p, opt);
3805 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3806 diff_flush_raw(p, opt);
3807 else if (fmt & DIFF_FORMAT_NAME) {
3808 const char *name_a, *name_b;
3809 name_a = p->two->path;
3810 name_b = NULL;
3811 strip_prefix(opt->prefix_length, &name_a, &name_b);
3812 write_name_quoted(name_a, opt->file, opt->line_termination);
3813 }
3814 }
3815
3816 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3817 {
3818 if (fs->mode)
3819 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3820 else
3821 fprintf(file, " %s ", newdelete);
3822 write_name_quoted(fs->path, file, '\n');
3823 }
3824
3825
3826 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3827 const char *line_prefix)
3828 {
3829 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3830 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3831 p->two->mode, show_name ? ' ' : '\n');
3832 if (show_name) {
3833 write_name_quoted(p->two->path, file, '\n');
3834 }
3835 }
3836 }
3837
3838 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3839 const char *line_prefix)
3840 {
3841 char *names = pprint_rename(p->one->path, p->two->path);
3842
3843 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3844 free(names);
3845 show_mode_change(file, p, 0, line_prefix);
3846 }
3847
3848 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3849 {
3850 FILE *file = opt->file;
3851 char *line_prefix = "";
3852
3853 if (opt->output_prefix) {
3854 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3855 line_prefix = buf->buf;
3856 }
3857
3858 switch(p->status) {
3859 case DIFF_STATUS_DELETED:
3860 fputs(line_prefix, file);
3861 show_file_mode_name(file, "delete", p->one);
3862 break;
3863 case DIFF_STATUS_ADDED:
3864 fputs(line_prefix, file);
3865 show_file_mode_name(file, "create", p->two);
3866 break;
3867 case DIFF_STATUS_COPIED:
3868 fputs(line_prefix, file);
3869 show_rename_copy(file, "copy", p, line_prefix);
3870 break;
3871 case DIFF_STATUS_RENAMED:
3872 fputs(line_prefix, file);
3873 show_rename_copy(file, "rename", p, line_prefix);
3874 break;
3875 default:
3876 if (p->score) {
3877 fprintf(file, "%s rewrite ", line_prefix);
3878 write_name_quoted(p->two->path, file, ' ');
3879 fprintf(file, "(%d%%)\n", similarity_index(p));
3880 }
3881 show_mode_change(file, p, !p->score, line_prefix);
3882 break;
3883 }
3884 }
3885
3886 struct patch_id_t {
3887 git_SHA_CTX *ctx;
3888 int patchlen;
3889 };
3890
3891 static int remove_space(char *line, int len)
3892 {
3893 int i;
3894 char *dst = line;
3895 unsigned char c;
3896
3897 for (i = 0; i < len; i++)
3898 if (!isspace((c = line[i])))
3899 *dst++ = c;
3900
3901 return dst - line;
3902 }
3903
3904 static void patch_id_consume(void *priv, char *line, unsigned long len)
3905 {
3906 struct patch_id_t *data = priv;
3907 int new_len;
3908
3909 /* Ignore line numbers when computing the SHA1 of the patch */
3910 if (!prefixcmp(line, "@@ -"))
3911 return;
3912
3913 new_len = remove_space(line, len);
3914
3915 git_SHA1_Update(data->ctx, line, new_len);
3916 data->patchlen += new_len;
3917 }
3918
3919 /* returns 0 upon success, and writes result into sha1 */
3920 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3921 {
3922 struct diff_queue_struct *q = &diff_queued_diff;
3923 int i;
3924 git_SHA_CTX ctx;
3925 struct patch_id_t data;
3926 char buffer[PATH_MAX * 4 + 20];
3927
3928 git_SHA1_Init(&ctx);
3929 memset(&data, 0, sizeof(struct patch_id_t));
3930 data.ctx = &ctx;
3931
3932 for (i = 0; i < q->nr; i++) {
3933 xpparam_t xpp;
3934 xdemitconf_t xecfg;
3935 mmfile_t mf1, mf2;
3936 struct diff_filepair *p = q->queue[i];
3937 int len1, len2;
3938
3939 memset(&xpp, 0, sizeof(xpp));
3940 memset(&xecfg, 0, sizeof(xecfg));
3941 if (p->status == 0)
3942 return error("internal diff status error");
3943 if (p->status == DIFF_STATUS_UNKNOWN)
3944 continue;
3945 if (diff_unmodified_pair(p))
3946 continue;
3947 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3948 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3949 continue;
3950 if (DIFF_PAIR_UNMERGED(p))
3951 continue;
3952
3953 diff_fill_sha1_info(p->one);
3954 diff_fill_sha1_info(p->two);
3955 if (fill_mmfile(&mf1, p->one) < 0 ||
3956 fill_mmfile(&mf2, p->two) < 0)
3957 return error("unable to read files to diff");
3958
3959 len1 = remove_space(p->one->path, strlen(p->one->path));
3960 len2 = remove_space(p->two->path, strlen(p->two->path));
3961 if (p->one->mode == 0)
3962 len1 = snprintf(buffer, sizeof(buffer),
3963 "diff--gita/%.*sb/%.*s"
3964 "newfilemode%06o"
3965 "---/dev/null"
3966 "+++b/%.*s",
3967 len1, p->one->path,
3968 len2, p->two->path,
3969 p->two->mode,
3970 len2, p->two->path);
3971 else if (p->two->mode == 0)
3972 len1 = snprintf(buffer, sizeof(buffer),
3973 "diff--gita/%.*sb/%.*s"
3974 "deletedfilemode%06o"
3975 "---a/%.*s"
3976 "+++/dev/null",
3977 len1, p->one->path,
3978 len2, p->two->path,
3979 p->one->mode,
3980 len1, p->one->path);
3981 else
3982 len1 = snprintf(buffer, sizeof(buffer),
3983 "diff--gita/%.*sb/%.*s"
3984 "---a/%.*s"
3985 "+++b/%.*s",
3986 len1, p->one->path,
3987 len2, p->two->path,
3988 len1, p->one->path,
3989 len2, p->two->path);
3990 git_SHA1_Update(&ctx, buffer, len1);
3991
3992 if (diff_filespec_is_binary(p->one) ||
3993 diff_filespec_is_binary(p->two)) {
3994 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3995 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3996 continue;
3997 }
3998
3999 xpp.flags = 0;
4000 xecfg.ctxlen = 3;
4001 xecfg.flags = 0;
4002 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4003 &xpp, &xecfg);
4004 }
4005
4006 git_SHA1_Final(sha1, &ctx);
4007 return 0;
4008 }
4009
4010 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4011 {
4012 struct diff_queue_struct *q = &diff_queued_diff;
4013 int i;
4014 int result = diff_get_patch_id(options, sha1);
4015
4016 for (i = 0; i < q->nr; i++)
4017 diff_free_filepair(q->queue[i]);
4018
4019 free(q->queue);
4020 DIFF_QUEUE_CLEAR(q);
4021
4022 return result;
4023 }
4024
4025 static int is_summary_empty(const struct diff_queue_struct *q)
4026 {
4027 int i;
4028
4029 for (i = 0; i < q->nr; i++) {
4030 const struct diff_filepair *p = q->queue[i];
4031
4032 switch (p->status) {
4033 case DIFF_STATUS_DELETED:
4034 case DIFF_STATUS_ADDED:
4035 case DIFF_STATUS_COPIED:
4036 case DIFF_STATUS_RENAMED:
4037 return 0;
4038 default:
4039 if (p->score)
4040 return 0;
4041 if (p->one->mode && p->two->mode &&
4042 p->one->mode != p->two->mode)
4043 return 0;
4044 break;
4045 }
4046 }
4047 return 1;
4048 }
4049
4050 static const char rename_limit_warning[] =
4051 "inexact rename detection was skipped due to too many files.";
4052
4053 static const char degrade_cc_to_c_warning[] =
4054 "only found copies from modified paths due to too many files.";
4055
4056 static const char rename_limit_advice[] =
4057 "you may want to set your %s variable to at least "
4058 "%d and retry the command.";
4059
4060 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4061 {
4062 if (degraded_cc)
4063 warning(degrade_cc_to_c_warning);
4064 else if (needed)
4065 warning(rename_limit_warning);
4066 else
4067 return;
4068 if (0 < needed && needed < 32767)
4069 warning(rename_limit_advice, varname, needed);
4070 }
4071
4072 void diff_flush(struct diff_options *options)
4073 {
4074 struct diff_queue_struct *q = &diff_queued_diff;
4075 int i, output_format = options->output_format;
4076 int separator = 0;
4077
4078 /*
4079 * Order: raw, stat, summary, patch
4080 * or: name/name-status/checkdiff (other bits clear)
4081 */
4082 if (!q->nr)
4083 goto free_queue;
4084
4085 if (output_format & (DIFF_FORMAT_RAW |
4086 DIFF_FORMAT_NAME |
4087 DIFF_FORMAT_NAME_STATUS |
4088 DIFF_FORMAT_CHECKDIFF)) {
4089 for (i = 0; i < q->nr; i++) {
4090 struct diff_filepair *p = q->queue[i];
4091 if (check_pair_status(p))
4092 flush_one_pair(p, options);
4093 }
4094 separator++;
4095 }
4096
4097 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4098 struct diffstat_t diffstat;
4099
4100 memset(&diffstat, 0, sizeof(struct diffstat_t));
4101 for (i = 0; i < q->nr; i++) {
4102 struct diff_filepair *p = q->queue[i];
4103 if (check_pair_status(p))
4104 diff_flush_stat(p, options, &diffstat);
4105 }
4106 if (output_format & DIFF_FORMAT_NUMSTAT)
4107 show_numstat(&diffstat, options);
4108 if (output_format & DIFF_FORMAT_DIFFSTAT)
4109 show_stats(&diffstat, options);
4110 if (output_format & DIFF_FORMAT_SHORTSTAT)
4111 show_shortstats(&diffstat, options);
4112 free_diffstat_info(&diffstat);
4113 separator++;
4114 }
4115 if (output_format & DIFF_FORMAT_DIRSTAT)
4116 show_dirstat(options);
4117
4118 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4119 for (i = 0; i < q->nr; i++) {
4120 diff_summary(options, q->queue[i]);
4121 }
4122 separator++;
4123 }
4124
4125 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4126 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4127 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4128 /*
4129 * run diff_flush_patch for the exit status. setting
4130 * options->file to /dev/null should be safe, becaue we
4131 * aren't supposed to produce any output anyway.
4132 */
4133 if (options->close_file)
4134 fclose(options->file);
4135 options->file = fopen("/dev/null", "w");
4136 if (!options->file)
4137 die_errno("Could not open /dev/null");
4138 options->close_file = 1;
4139 for (i = 0; i < q->nr; i++) {
4140 struct diff_filepair *p = q->queue[i];
4141 if (check_pair_status(p))
4142 diff_flush_patch(p, options);
4143 if (options->found_changes)
4144 break;
4145 }
4146 }
4147
4148 if (output_format & DIFF_FORMAT_PATCH) {
4149 if (separator) {
4150 putc(options->line_termination, options->file);
4151 if (options->stat_sep) {
4152 /* attach patch instead of inline */
4153 fputs(options->stat_sep, options->file);
4154 }
4155 }
4156
4157 for (i = 0; i < q->nr; i++) {
4158 struct diff_filepair *p = q->queue[i];
4159 if (check_pair_status(p))
4160 diff_flush_patch(p, options);
4161 }
4162 }
4163
4164 if (output_format & DIFF_FORMAT_CALLBACK)
4165 options->format_callback(q, options, options->format_callback_data);
4166
4167 for (i = 0; i < q->nr; i++)
4168 diff_free_filepair(q->queue[i]);
4169 free_queue:
4170 free(q->queue);
4171 DIFF_QUEUE_CLEAR(q);
4172 if (options->close_file)
4173 fclose(options->file);
4174
4175 /*
4176 * Report the content-level differences with HAS_CHANGES;
4177 * diff_addremove/diff_change does not set the bit when
4178 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4179 */
4180 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4181 if (options->found_changes)
4182 DIFF_OPT_SET(options, HAS_CHANGES);
4183 else
4184 DIFF_OPT_CLR(options, HAS_CHANGES);
4185 }
4186 }
4187
4188 static void diffcore_apply_filter(const char *filter)
4189 {
4190 int i;
4191 struct diff_queue_struct *q = &diff_queued_diff;
4192 struct diff_queue_struct outq;
4193 DIFF_QUEUE_CLEAR(&outq);
4194
4195 if (!filter)
4196 return;
4197
4198 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4199 int found;
4200 for (i = found = 0; !found && i < q->nr; i++) {
4201 struct diff_filepair *p = q->queue[i];
4202 if (((p->status == DIFF_STATUS_MODIFIED) &&
4203 ((p->score &&
4204 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4205 (!p->score &&
4206 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4207 ((p->status != DIFF_STATUS_MODIFIED) &&
4208 strchr(filter, p->status)))
4209 found++;
4210 }
4211 if (found)
4212 return;
4213
4214 /* otherwise we will clear the whole queue
4215 * by copying the empty outq at the end of this
4216 * function, but first clear the current entries
4217 * in the queue.
4218 */
4219 for (i = 0; i < q->nr; i++)
4220 diff_free_filepair(q->queue[i]);
4221 }
4222 else {
4223 /* Only the matching ones */
4224 for (i = 0; i < q->nr; i++) {
4225 struct diff_filepair *p = q->queue[i];
4226
4227 if (((p->status == DIFF_STATUS_MODIFIED) &&
4228 ((p->score &&
4229 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4230 (!p->score &&
4231 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4232 ((p->status != DIFF_STATUS_MODIFIED) &&
4233 strchr(filter, p->status)))
4234 diff_q(&outq, p);
4235 else
4236 diff_free_filepair(p);
4237 }
4238 }
4239 free(q->queue);
4240 *q = outq;
4241 }
4242
4243 /* Check whether two filespecs with the same mode and size are identical */
4244 static int diff_filespec_is_identical(struct diff_filespec *one,
4245 struct diff_filespec *two)
4246 {
4247 if (S_ISGITLINK(one->mode))
4248 return 0;
4249 if (diff_populate_filespec(one, 0))
4250 return 0;
4251 if (diff_populate_filespec(two, 0))
4252 return 0;
4253 return !memcmp(one->data, two->data, one->size);
4254 }
4255
4256 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4257 {
4258 int i;
4259 struct diff_queue_struct *q = &diff_queued_diff;
4260 struct diff_queue_struct outq;
4261 DIFF_QUEUE_CLEAR(&outq);
4262
4263 for (i = 0; i < q->nr; i++) {
4264 struct diff_filepair *p = q->queue[i];
4265
4266 /*
4267 * 1. Entries that come from stat info dirtiness
4268 * always have both sides (iow, not create/delete),
4269 * one side of the object name is unknown, with
4270 * the same mode and size. Keep the ones that
4271 * do not match these criteria. They have real
4272 * differences.
4273 *
4274 * 2. At this point, the file is known to be modified,
4275 * with the same mode and size, and the object
4276 * name of one side is unknown. Need to inspect
4277 * the identical contents.
4278 */
4279 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4280 !DIFF_FILE_VALID(p->two) ||
4281 (p->one->sha1_valid && p->two->sha1_valid) ||
4282 (p->one->mode != p->two->mode) ||
4283 diff_populate_filespec(p->one, 1) ||
4284 diff_populate_filespec(p->two, 1) ||
4285 (p->one->size != p->two->size) ||
4286 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4287 diff_q(&outq, p);
4288 else {
4289 /*
4290 * The caller can subtract 1 from skip_stat_unmatch
4291 * to determine how many paths were dirty only
4292 * due to stat info mismatch.
4293 */
4294 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4295 diffopt->skip_stat_unmatch++;
4296 diff_free_filepair(p);
4297 }
4298 }
4299 free(q->queue);
4300 *q = outq;
4301 }
4302
4303 static int diffnamecmp(const void *a_, const void *b_)
4304 {
4305 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4306 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4307 const char *name_a, *name_b;
4308
4309 name_a = a->one ? a->one->path : a->two->path;
4310 name_b = b->one ? b->one->path : b->two->path;
4311 return strcmp(name_a, name_b);
4312 }
4313
4314 void diffcore_fix_diff_index(struct diff_options *options)
4315 {
4316 struct diff_queue_struct *q = &diff_queued_diff;
4317 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4318 }
4319
4320 void diffcore_std(struct diff_options *options)
4321 {
4322 if (options->skip_stat_unmatch)
4323 diffcore_skip_stat_unmatch(options);
4324 if (!options->found_follow) {
4325 /* See try_to_follow_renames() in tree-diff.c */
4326 if (options->break_opt != -1)
4327 diffcore_break(options->break_opt);
4328 if (options->detect_rename)
4329 diffcore_rename(options);
4330 if (options->break_opt != -1)
4331 diffcore_merge_broken();
4332 }
4333 if (options->pickaxe)
4334 diffcore_pickaxe(options);
4335 if (options->orderfile)
4336 diffcore_order(options->orderfile);
4337 if (!options->found_follow)
4338 /* See try_to_follow_renames() in tree-diff.c */
4339 diff_resolve_rename_copy();
4340 diffcore_apply_filter(options->filter);
4341
4342 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4343 DIFF_OPT_SET(options, HAS_CHANGES);
4344 else
4345 DIFF_OPT_CLR(options, HAS_CHANGES);
4346
4347 options->found_follow = 0;
4348 }
4349
4350 int diff_result_code(struct diff_options *opt, int status)
4351 {
4352 int result = 0;
4353
4354 diff_warn_rename_limit("diff.renamelimit",
4355 opt->needed_rename_limit,
4356 opt->degraded_cc_to_c);
4357 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4358 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4359 return status;
4360 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4361 DIFF_OPT_TST(opt, HAS_CHANGES))
4362 result |= 01;
4363 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4364 DIFF_OPT_TST(opt, CHECK_FAILED))
4365 result |= 02;
4366 return result;
4367 }
4368
4369 /*
4370 * Shall changes to this submodule be ignored?
4371 *
4372 * Submodule changes can be configured to be ignored separately for each path,
4373 * but that configuration can be overridden from the command line.
4374 */
4375 static int is_submodule_ignored(const char *path, struct diff_options *options)
4376 {
4377 int ignored = 0;
4378 unsigned orig_flags = options->flags;
4379 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4380 set_diffopt_flags_from_submodule_config(options, path);
4381 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4382 ignored = 1;
4383 options->flags = orig_flags;
4384 return ignored;
4385 }
4386
4387 void diff_addremove(struct diff_options *options,
4388 int addremove, unsigned mode,
4389 const unsigned char *sha1,
4390 const char *concatpath, unsigned dirty_submodule)
4391 {
4392 struct diff_filespec *one, *two;
4393
4394 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4395 return;
4396
4397 /* This may look odd, but it is a preparation for
4398 * feeding "there are unchanged files which should
4399 * not produce diffs, but when you are doing copy
4400 * detection you would need them, so here they are"
4401 * entries to the diff-core. They will be prefixed
4402 * with something like '=' or '*' (I haven't decided
4403 * which but should not make any difference).
4404 * Feeding the same new and old to diff_change()
4405 * also has the same effect.
4406 * Before the final output happens, they are pruned after
4407 * merged into rename/copy pairs as appropriate.
4408 */
4409 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4410 addremove = (addremove == '+' ? '-' :
4411 addremove == '-' ? '+' : addremove);
4412
4413 if (options->prefix &&
4414 strncmp(concatpath, options->prefix, options->prefix_length))
4415 return;
4416
4417 one = alloc_filespec(concatpath);
4418 two = alloc_filespec(concatpath);
4419
4420 if (addremove != '+')
4421 fill_filespec(one, sha1, mode);
4422 if (addremove != '-') {
4423 fill_filespec(two, sha1, mode);
4424 two->dirty_submodule = dirty_submodule;
4425 }
4426
4427 diff_queue(&diff_queued_diff, one, two);
4428 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4429 DIFF_OPT_SET(options, HAS_CHANGES);
4430 }
4431
4432 void diff_change(struct diff_options *options,
4433 unsigned old_mode, unsigned new_mode,
4434 const unsigned char *old_sha1,
4435 const unsigned char *new_sha1,
4436 const char *concatpath,
4437 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4438 {
4439 struct diff_filespec *one, *two;
4440
4441 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4442 is_submodule_ignored(concatpath, options))
4443 return;
4444
4445 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4446 unsigned tmp;
4447 const unsigned char *tmp_c;
4448 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4449 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4450 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4451 new_dirty_submodule = tmp;
4452 }
4453
4454 if (options->prefix &&
4455 strncmp(concatpath, options->prefix, options->prefix_length))
4456 return;
4457
4458 one = alloc_filespec(concatpath);
4459 two = alloc_filespec(concatpath);
4460 fill_filespec(one, old_sha1, old_mode);
4461 fill_filespec(two, new_sha1, new_mode);
4462 one->dirty_submodule = old_dirty_submodule;
4463 two->dirty_submodule = new_dirty_submodule;
4464
4465 diff_queue(&diff_queued_diff, one, two);
4466 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4467 DIFF_OPT_SET(options, HAS_CHANGES);
4468 }
4469
4470 void diff_unmerge(struct diff_options *options,
4471 const char *path,
4472 unsigned mode, const unsigned char *sha1)
4473 {
4474 struct diff_filespec *one, *two;
4475
4476 if (options->prefix &&
4477 strncmp(path, options->prefix, options->prefix_length))
4478 return;
4479
4480 one = alloc_filespec(path);
4481 two = alloc_filespec(path);
4482 fill_filespec(one, sha1, mode);
4483 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4484 }
4485
4486 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4487 size_t *outsize)
4488 {
4489 struct diff_tempfile *temp;
4490 const char *argv[3];
4491 const char **arg = argv;
4492 struct child_process child;
4493 struct strbuf buf = STRBUF_INIT;
4494 int err = 0;
4495
4496 temp = prepare_temp_file(spec->path, spec);
4497 *arg++ = pgm;
4498 *arg++ = temp->name;
4499 *arg = NULL;
4500
4501 memset(&child, 0, sizeof(child));
4502 child.use_shell = 1;
4503 child.argv = argv;
4504 child.out = -1;
4505 if (start_command(&child)) {
4506 remove_tempfile();
4507 return NULL;
4508 }
4509
4510 if (strbuf_read(&buf, child.out, 0) < 0)
4511 err = error("error reading from textconv command '%s'", pgm);
4512 close(child.out);
4513
4514 if (finish_command(&child) || err) {
4515 strbuf_release(&buf);
4516 remove_tempfile();
4517 return NULL;
4518 }
4519 remove_tempfile();
4520
4521 return strbuf_detach(&buf, outsize);
4522 }
4523
4524 size_t fill_textconv(struct userdiff_driver *driver,
4525 struct diff_filespec *df,
4526 char **outbuf)
4527 {
4528 size_t size;
4529
4530 if (!driver || !driver->textconv) {
4531 if (!DIFF_FILE_VALID(df)) {
4532 *outbuf = "";
4533 return 0;
4534 }
4535 if (diff_populate_filespec(df, 0))
4536 die("unable to read files to diff");
4537 *outbuf = df->data;
4538 return df->size;
4539 }
4540
4541 if (driver->textconv_cache && df->sha1_valid) {
4542 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4543 &size);
4544 if (*outbuf)
4545 return size;
4546 }
4547
4548 *outbuf = run_textconv(driver->textconv, df, &size);
4549 if (!*outbuf)
4550 die("unable to read files to diff");
4551
4552 if (driver->textconv_cache && df->sha1_valid) {
4553 /* ignore errors, as we might be in a readonly repository */
4554 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4555 size);
4556 /*
4557 * we could save up changes and flush them all at the end,
4558 * but we would need an extra call after all diffing is done.
4559 * Since generating a cache entry is the slow path anyway,
4560 * this extra overhead probably isn't a big deal.
4561 */
4562 notes_cache_write(driver->textconv_cache);
4563 }
4564
4565 return size;
4566 }
MinGW port of GIT